Split (1)

train · 100k rows

id stringlengths 3 9	source stringclasses 1 value	version stringclasses 1 value	text stringlengths 1.54k 298k	added stringdate 1993-11-25 05:05:38 2024-09-20 15:30:25	created stringdate 1-01-01 00:00:00 2024-07-31 00:00:00	metadata dict
49386266	pes2o/s2orc	v3-fos-license	Medical expenditure clustering and determinants of the annual medical expenditures of residents: a population-based retrospective study from rural China Objective To identify the characteristics of high-cost (HC) patients and the determinants of the annual medical expenditures of Chinese rural residents. Methods Medical expenditure clustering was performed by Lorentz curve and Gini index. T and X2 tests were performed to identify the characteristics of the respondents, and a multilevel regression model examined the determinants of their annual medical expenditures. Design A cluster sampling study was performed to identify those residents who availed healthcare services and to assign them to HC (top 5%), moderate-cost (top 30%) and low-cost (others) groups based on their annual medical expenditures. Setting The annual healthcare utilisation was calculated by using data from the population-based database of the 2014 New Rural Cooperative Medical System. Participants A total of 478 051 residents who availed healthcare services were recruited for the retrospective study in 2014. The annual medical expenditures of these residents were used as the research object. Results The total medical expenditures of Macheng city residents for the year 2014 have a Gini index of 0.81 and around 68.01% of these expenditures can be attributed to HC patients. Female residents (51.5%) and persons aged over 60 years (34.48%) who are suffering from diseases that are difficult to diagnose have a high tendency to accumulate high medical costs. The annual medical expenditures of people living in the same village or town tend to be approximated. Age, disease category, inpatient status, healthcare utilisation and utilisation level are identified as the determinants of annual medical expenditures. Conclusions The medical expenditures of rural residents are clustered at a remarkably high level, and HC patients are suffering from high economic burden. Therefore, policy-makers must guide these patients in seeking appropriate healthcare services and improve their management of healthcare quality to reduce the unnecessary healthcare utilisation of these patients. Trial registration number ChiCTR-OOR-14005563. Objective To identify the characteristics of high-cost (HC) patients and the determinants of the annual medical expenditures of Chinese rural residents. Methods Medical expenditure clustering was performed by Lorentz curve and Gini index. T and X 2 tests were performed to identify the characteristics of the respondents, and a multilevel regression model examined the determinants of their annual medical expenditures. Design A cluster sampling study was performed to identify those residents who availed healthcare services and to assign them to HC (top 5%), moderate-cost (top 30%) and low-cost (others) groups based on their annual medical expenditures. setting The annual healthcare utilisation was calculated by using data from the population-based database of the 2014 New Rural Cooperative Medical System. Participants A total of 478 051 residents who availed healthcare services were recruited for the retrospective study in 2014. The annual medical expenditures of these residents were used as the research object. results The total medical expenditures of Macheng city residents for the year 2014 have a Gini index of 0.81 and around 68.01% of these expenditures can be attributed to HC patients. Female residents (51.5%) and persons aged over 60 years (34.48%) who are suffering from diseases that are difficult to diagnose have a high tendency to accumulate high medical costs. The annual medical expenditures of people living in the same village or town tend to be approximated. Age, disease category, inpatient status, healthcare utilisation and utilisation level are identified as the determinants of annual medical expenditures. Conclusions The medical expenditures of rural residents are clustered at a remarkably high level, and HC patients are suffering from high economic burden. Therefore, policy-makers must guide these patients in seeking appropriate healthcare services and improve their management of healthcare quality to reduce the unnecessary healthcare utilisation of these patients. trial registration number ChiCTR-OOR-14005563. bACkgrOunD The rapid increase in health expenditures greatly impedes the development of the New Rural Cooperative Medical System (NRCMS), the largest basic social health insurance system in rural China that covers 603.46 million rural residents. Specifically, the health expenditures per capita in China have increased from ¥513.8 (US$83.6) in 2012 to ¥1279. 2 (208.2) in 2017 with an annual growth rate of 25.6%, which is much higher than the annual growth in fundraising per capita (16.02%). 1 Medical expenditure clustering is considered an important factor that motivates such rapid increase in health expenditures. 2 Studies on the distribution of diseases within a population have defined clustering as the uneven distribution of disease morbidity in time or space. 3 4 In line with strengths and limitations of this study ► This study is the first to introduce the medical expenditure clustering technique, the findings of which can supplement the results of previous research on high-cost patients. ► The annual medical expenditures of residents are seldom reported at the population level. This study was conducted in Macheng, a city in Hubei province with 889 160 residents according to the New Rural Cooperative Medical System database, which also stores the inpatient and outpatient records of these residents. ► The Lorentz curve and Gini index were used to cluster the annual medical expenditure data, and a three-level linear regression model was used to aggregate these data at the residential and town levels. ► The age, gender, hospitalisation, geographical and disease data of the sample were included in the regression model. However, some individual factors were not included in the model. Open access this definition, medical expenditure clustering indicates the uneven distribution of the medical expenditures of a given population. In recent years, researchers have shown great interest in medical expenditure clustering and have specifically focused on high-cost (HC) patients, who are known for accumulating high annual medical expenditures 5 and comprise the top 5% biggest spenders in healthcare. 2 Previous studies have revealed that the medical expenditures of HC patients exceed those of the entire population by 50%. 6 For instance, in 2014, HC patients account for 52.3% of the total medical expenditures in the USA. 5 These patients and their healthcare quality management have attracted much research attention because of their high healthcare utilisation rate and inappropriate utilisation of healthcare services. 7 Improving the healthcare quality of these patients can discourage such inappropriate utilisation, reduce their health expenditures, conserve social health insurance funds and promote horizontal equity. Medical expenditure clustering has also become a major concern in low-income and middle-income countries. In rural China, the rapid development of NRCMS significantly promoted the healthcare utilisation rates of residents. For instance, the annual hospitalisation rate in rural China increased from 8.7% in 2008 to 14.9% in 2017. 1 An empirical analysis of seven counties in China revealed that 78.6% of inpatient services in 2015 were distributed among one-third of all inpatients in the area. In addition, one patient in Qianjiang District used inpatient services 27 times in 2014. Another study involving 12 600 families in Jiangsu revealed that HC patients accounted for 44.9% of the total medical expenditures of the entire population. 8 Moreover, the medical expenditures clustered at the patient level are much higher than those clustered at the family level. We also suspect that China has a very high degree of medical expenditure clustering and that HC patients incur unnecessary medical expenditures because of the fragmented healthcare delivery system in rural China. Residents in rural China also seek healthcare services from a three-tier (village-town-county) healthcare delivery system where the higher tiers provide better services and charge higher medical costs. 9 Given that patients neither follow a specified order nor have limitations when seeking healthcare services and that general practitioners or consultants are unavailable in most parts of rural China, the residents of rural areas tend to make uninformed decisions when choosing among hospitals and various types of healthcare services, thereby leading to their inappropriate utilisation of healthcare services. For instance, some of these residents, especially HC patients, may be given inpatient services when they actually require outpatient services and may be unnecessarily admitted to higher-level hospitals, thereby incurring higher medical costs. 10 Yingchun revealed that the inappropriate hospital admission rate in five counties reached as high as 27.6% in 2014. 11 To reduce the economic burden of rural residents in China, several policies and strategies, such as the Tiered Healthcare System and Serious Illness Medical Insurance (2016), have been proposed. However, the effectivity of these programmes has received mixed feedback and the focus of these policies remains unsubstantiated, which may be ascribed to the difficulty in identifying high-demand and HC patients. 12 Waxmonsky et al argued that those HC patients with multiple or complex conditions, behavioural disorders or socioeconomic problems are particularly difficult to monitor. 13 Moreover, only few studies have investigated the characteristics and determinants of HC patients by using population-level data because of the lack of necessary data for exploring their annual medical expenditures. Therefore, the critical values, average expenditures and inpatient service utilisation of HC patients in rural China remain unclear. Previous studies reveal that HC patients often maintain a high level of medical expenditure for the following year. For instance, Robst 15 Therefore, while HC patients are assumed to possess certain characteristics, the distribution of their medical expenditures remains unclear. The growth of health expenditures must be controlled and the efficiency of insurance funds must be enhanced to improve the health insurance system of a specific area. Medical expenditure clustering can be used to monitor the cost efficiency of healthcare services. Based on the above findings, identifying HC patients is a necessary procedure and clustering the medical expenditures of a given population must be taken as the first step towards achieving such goal. This research focuses on medical expenditure clustering in particular as well as on the distribution and characteristics of HC patients and the determinants of the annual medical expenditures of residents in general. This study also aims to guide policy-makers and health planners in predicting and planning the future needs of these patients. MethODs study setting We calculated the annual medical expenditures of residents based on the outpatient and inpatient services that they have availed within a calendar year. A population-based retrospective study was performed in Macheng, a typical rural area in Hubei and county-level city in central China ( figure 1). Macheng has a total population of 889 160 and a GDP per capita of ¥22 758 (US$3704.83). Macheng has 2 county hospitals, 22 township hospitals and 207 village clinics. The residents of this city are enrolled in NRCMS, which reimburses the medical expenditures of these residents for any type of healthcare service (eg, outpatient and inpatient services in various medical institutions, including tertiary hospitals in urban areas). The healthcare Open access utilisation information of these residents is recorded in the NRCMS database whenever they request for a reimbursement. Therefore, this study uses the 2014 Macheng city data from the NRCMS database. Data processing Those residents of Macheng who availed healthcare services in the past were identified through a retrospective study. After screening the data, 478 051 of the 889 160 resident records stored in the NRCMS database were included in the sample and processed by using MS Excel 2010. First, the outpatient and inpatient cases were input into a single Excel sheet. Second, those cases under the same patient identifier (ID number) were sorted chronologically. Third, the annual medical (inpatient and outpatient) expenditures of each patient, the number of outpatient and inpatient cases and other information were recorded. Fourth, the annual healthcare utilisation cases were input into a new database, where each case represents the annual healthcare utilisation of a resident. Finally, the residents were sorted in a descending order according to the annual medical expenditures stated in their records. Those residents who occupied the top 5% and 6%-30% of the sample were included in the HC and moderate-cost (MC) groups, respectively, while the other patients were included in the low-cost (LC) group. These three groups represent the various degrees of medical expenditure clustering. The main programming techniques employed in this paper include Excel functions (eg, COUNTIF, SUMPRODUCT, LOOKUP and IF) and case processing technologies (eg, split columns and removal of duplicates). The outstanding diseases of each resident were marked and the original International Classification of Diseases (ICD)-10 disease codes were adjusted to broader ones (eg, the disease code for chronic obstructive pulmonary disease (COPD) was adjusted from J44.900 to J44). Township hospitals were divided into four levels according to their scale and service capacity. The distance from and arrival time to county hospitals were individually captured by using Google Maps. The exchange rate of the US$ against the RMB in 2014 was 6.1428. The land form of towns, the healthcare capacity of township hospitals and the sociological characteristics (eg, gender and age), arrival time to county hospitals, disease categories, healthcare utilisation (eg, annual length of stay (LOS)) and annual medical expenditures of the residents were collected to build the final database. 16 statistical analysis First, the medical expenditures of the residents were clustered by using the Gini coefficient and Lorentz curve. The Gini coefficient is a digitised representation of medical expenditure clustering. A larger Gini coefficient corresponds to a higher degree of medical expenditure clustering. Then, the characteristics of the residents in the HC, MC and LC groups were compared by conducting t-test and χ 2 test in the IBM SPSS Statistics V.22.0 software. At last, the determinants of annual medical expenditure were then examined by conducting a linear logistic regression analysis. Two key observations were obtained at this stage. First, the obtained data showed a hierarchical structure. Therefore, the determinants of annual medical expenditure were examined by conducting a multilevel linear logistic regression analysis using MLwiN V.2.30. 17 Second, the annual medical expenditure showed a skewed distribution as expected. Therefore, this variable was transformed to follow a normal distribution. The patient, village and town were assigned to levels 1, 2 and 3, respectively. After accepting the Log10 (x) translation, the following regression model was obtained: where β 0jk denotes the fixed-effects parameter while u oj and w ojk denote the random effects at the village and town levels, respectively. Patient and public involvement No patients or members of the public were involved in this research. The patient information was anonymised and deidentified before the analysis. results Medical expenditure clustering in the sample area Table 1 presents the clustering results for the medical expenditures of Macheng residents in 2014. Among these residents, the top 5% and 20% accounted for 68.01% and 90% of the total medical expenditures of the city, respectively. Figure 2 shows the Lorentz curve of the clustering results, which have a Gini coefficient of 0.814. Table 2 shows the medical expenditure distribution of the HC, MC and LC groups. The HC group has an average annual expenditure per capita of over ¥15 000 (US$2441.6) and a minimum expenditure of ¥4985.80 (US$811.61) while the LC group has a maximum expenditure of ¥347.29 (US$56.54). Figure 3 presents the expenditure composition of these groups. Table 3 shows the demographic characteristics of the patients. The three groups showed significant differences (p<0.001) for nine demographic items. Specifically, females accounted for 51.5% and 47.42% of the HC and LC groups, respectively. Most of the residents aged over 60 years were assigned to the HC group. The residents Open access in the HC and LC groups generally had small (4.01) and large (4.23) family sizes, respectively. Family size, distance from county hospitals and arrival time to county hospitals all showed the same change trends. Most of the residents living near a high-capacity township hospital were assigned to the HC group (15.19%). By contrast, 30.76% of the members in the LC group were living near a low-capacity township hospital. These three groups also showed similar distributions across varying geographical and traffic conditions. Specifically, those members of the LC group who were living in poor areas (mountains and county roads) showed high accident rates (12.54% and 32.36%, respectively). In addition, the members of the HC group were highly likely to develop cancer, circulatory, digestive and urinary diseases as well as haematological disorders while the members of the LC group often developed respiratory diseases. Table 4 shows the distribution of the healthcare utilisation of all Macheng residents in 2014. The residents in all three groups showed a decreasing trend in their average number of inpatient cases (2.36, 0.45 and 0.02 for the HC, MC and LC groups, respectively), average number of inpatient cases and annual LOS (25.69, 3.28 and 0.1). However, the opposite trend was observed in the annual LOS (0.07, 62.94 and 99.95) and average number of outpatient cases (8.28, 13.57 and 5.05) of those residents without prior hospitalisation experience in the HC, MC and LC groups. The healthcare utilisation of these residents also showed an inverted V-shaped distribution at the town and clinic levels. Specifically, the residents in the MC group had a high average number of outpatient cases at the town (3.69) and clinic (8.61) levels, but the residents in all three groups showed a decreasing trend in their average number of outpatient cases (1.76, 1.27 and 0.21 for the HC, MC and LC groups, respectively). Determinants of the annual medical expenditures of residents A three-level linear regression was performed where the patient, village and town were assigned to levels 1, 2 and 3, respectively. Table 5 displays the results for the explanatory variables that are used to fit the three-variance component models. Age, disease category and healthcare utilisation were identified as the major determinants of the annual medical expenditures of residents. No significant relationship was observed among gender, family size, distance to county hospitals, arrival time to county The three groups showed obvious differences in their expenditure structures. Hospitalisation expenditures accounted for over 95% of the medical expenditures of the HC group, thereby supporting the findings of Driessen et al. 18 Outpatient expenditures accounted for most of the medical expenditures of the LC group while outpatient and hospitalisation expenditures equally accounted for the medical expenditures of the MC group. The medical expenditures of the HC group showed a dispersed distribution and an SD that was much higher than the mean (¥22 817.95 vs ¥16 618.78, US$3714.58 vs US$2705.41). By contrast, the medical expenditures of the MC and LC groups showed a relatively concentrated distribution. Open access The HC group faces a very high economic burden. In fact, those residents with medical expenditures of over ¥5000 (US$813.96) were included in this group. This group had an annual medical expenditure per capita of over ¥16 000 (US$2604.68), of which only Open access around 50% were reimbursed by NRCMS. Therefore, HC patients had an average out-of-pocket expenditures over ¥8000 (US$1302.34), which was nearly 80% of the total consumer spending per capita of rural residents (¥10 129.8, US$1649.05). 1 This ratio can easily lead to a catastrophic health spending. Aggregating annual medical expenditures at the village and town levels The multilevel linear regression revealed a hierarchical structure (town-village-residents) in the collected data. The annual medical expenditures of residents were clustered at the town and village levels while the annual medical expenditures of residents living in the same village and town tend to be approximated. The distribution of annual medical expenditures as shown in table 3 significantly differs from that shown in table 5 in terms of geography, traffic condition and capacity of township hospitals. Such differences were mostly observed at the town level because the residents were living in towns with the same geography, traffic conditions and capacity of township hospitals. Meanwhile, 30 towns in the sample Open access showed differences in their social customs, geographical locations and capacity of township hospitals. These findings altogether show that those residents living in areas with favourable geographical conditions (eg, plains and national standard roads) and high-capacity hospitals tend to accumulate high medical expenditures. Similarly, the effects of distance from and arrival time to county hospitals were mostly observed at the village level. Given their convenient locations, those residents who are living near county hospitals have a higher probability to avail healthcare services in these institutions and accumulate higher medical expenditures compared with those who are living far from these hospitals. Determinants of annual medical expenditures Annual medical expenditures are directly determined by healthcare utilisation. Table 5 shows that a higher utilisation of healthcare services corresponds to higher medical expenditures. The village, town and county levels had regression coefficients of 0.031, 0.049 and 0.134 for outpatient cases, which were lower than the corresponding coefficients for inpatient cases. Similar to hospital utilisation, an increase in LOS corresponds to an increase in medical expenditures. In addition, a higher age corresponds to a higher expenditure. This case is particularly true for those residents aged above 60 years, who account for 34.48% of HC group, in other words, 8.5% of the elderly population was defined as HC, while only 5% of total population was defined as HC. Such high expenditures can be attributed to their poor physical condition and high tendency to develop one or multiple diseases. A WHO report revealed that an ageing population would increase the healthcare expenditures of a country, but the extent of such increase is less than expected. Although health-related needs often increase along with age, the relationship between healthcare utilisation and health expenditure varies as a person grows older. 19 In fact, the medical expenditures of people from high-income countries gradually decline after they reach the age of 75. Other studies even show that those people aged 80 years and above have a much lower share in the consumption of medical resources compared with the total population. In terms of gender, the average medical expenditures of female residents do not differ from those of male residents. The relationship between gender and medical expenditures also warrants further study. 20 Those patients with respiratory, urinary, endocrinology, skeletal and muscular diseases have the same degree of medical expenditure clustering. By contrast, those patients with cancer, circulatory, digestive, obstetrical and gynaecological diseases or haematological disorders have significantly high annual medical expenditures, with patients with cancer having the highest regression coefficient of 0.759. Developing diseases that are fatal and difficult to assess can easily lead to high medical expenditures. 21 These findings can be attributed to the limited ability of healthcare professionals at townships, most of whom lack specialisation in treating urinary and cardiovascular diseases or haematological disorders; therefore, those residents suffering from such diseases tend to be admitted to county hospitals or hospitals outside their respective counties. 22 redesigning the healthcare delivery system for the hC group The intense concentration of medical expenditures reveals a great imbalance in the healthcare utilisation of Macheng city residents. Therefore, the rural healthcare delivery system as well as the current efforts in reducing the costs and promoting the cost efficiency of healthcare services should focus on the HC patients. First, a monitoring mechanism for HC patients must be established and the NRCMS database can be used as a source of information that can facilitate the monitoring of these patients. Although HC patients are identified based on their medical expenditure or health claims, a patient can be predicted as HC high probability in advance based on several risk factors. Robst and Wodchis et al found that a patient identified as HC in a year is more than 40% likely to be identified as an HC patient in the following year given that these patients often maintain a high level of medical expenditure for the following year. 14 15 Therefore, those residents with a remarkably high healthcare utilisation are exposed to many risk factors and have been identified as HC patients in the previous year warrant special attention. Moreover, HC patients need guidance when choosing healthcare services and must be given priority access to primary healthcare. Second, enhancing the healthcare quality management of patients plays a vital role in discouraging their unnecessary utilisation of healthcare services, such as inappropriate admission to hospitals and excessive utilisation of inpatient services. 23 Third, exploring new mechanisms, such as the comprehensive management of patients with chronic diseases, setting a global budget for multilevel institutions, integrated management programmes and integrated delivery of medicine and nursing services to aged residents, can motivate doctors to deliver continued care to HC patients. In doing so, healthcare professionals can avoid unnecessary service duplications, improve their healthcare efficacy and help their patients take advantage of primary healthcare services. 24 COnClusiOn From the city-level population perspective, the medical expenditures of rural residents in China have an intense clustering level. Apart from demographic characteristics (eg, age and disease), healthcare utilisation was identified as a primary determinant of medical expenditure clustering. Therefore, policy-makers must guide HC patients in choosing healthcare services and improve their healthcare quality management to discourage their unnecessary utilisation of healthcare services. Doctors must also Open access be motivated to deliver continued care to this group of patients. liMitAtiOn This study has two limitations. First, hospitalisation information, geographical factors, referral status and diseases were included in the regression model while other individual factors, such as economic status, education and preference, were ignored. Second, several studies show that HC residents with a high tendency to develop multiple chronic diseases tend to use multidisciplinary and multi-institutional services. 25 However, this work only considered the main diseases of these patients as captured in the NRCMS database. These limitations may affect the stability of the findings and should be examined in further studies. Contributors YZ and SL participated in the conception, design, analyses and writing of the manuscript. NY participated in the data collection and statistical analysis. LZ helped draft, review and revise the manuscript. All authors gave their approval to publish this version of the manuscript. Funding This research is supported by the National Youth Natural Science Foundation of China (Grant No: 71603088).	2018-07-03T23:06:14.836Z	2018-06-01T00:00:00.000	{ "year": 2018, "sha1": "60a84cc04bb58585c76e41d3430d4cd85ce62254", "oa_license": "CCBYNC", "oa_url": "https://bmjopen.bmj.com/content/bmjopen/8/6/e022721.full.pdf", "oa_status": "GOLD", "pdf_src": "Highwire", "pdf_hash": "60a84cc04bb58585c76e41d3430d4cd85ce62254", "s2fieldsofstudy": [ "Medicine", "Political Science" ], "extfieldsofstudy": [ "Medicine" ] }
3061221	pes2o/s2orc	v3-fos-license	A Generic Model of Contracts for Embedded Systems We present the mathematical foundations of the contract-based model developed in the framework of the SPEEDS project. SPEEDS aims at developing methods and tools to support"speculative design", a design methodology in which distributed designers develop different aspects of the overall system, in a concurrent but controlled way. Our generic mathematical model of contract supports this style of development. This is achieved by focusing on behaviors, by supporting the notion of"rich component"where diverse (functional and non-functional) aspects of the system can be considered and combined, by representing rich components via their set of associated contracts, and by formalizing the whole process of component composition. Introduction Several industrial sectors involving complex embedded systems design have recently experienced drastic moves in their organization-aerospace and automotive being typical examples. Initially organized around large, vertically integrated companies supporting most of the design in house, these sectors were restructured in the 80's due to the emergence of sizeable competitive suppliers. OEMs performed system design and integration by importing entire subsystems from suppliers. This, however, shifted a significant portion of the value to the suppliers, and eventually contributed to late errors that caused delays and excessive additional cost during the system integration phase. In the last decade, these industrial sectors went through a profound reorganization in an attempt by OEMs to recover value from the supply chain, by focusing on those parts of the design at the core of their competitive advantage. The rest of the system was instead centered around standard platforms that could be developed and shared by otherwise competitors. Examples of this trend are AUTOSAR in the automotive industry [1], and Integrated Modular Avionics (IMA) in aerospace [2]. This new organization requires extensive virtual prototyping and design space exploration, where component or subsystem specification and integration occur at different phases of the design, including at the early ones [3]. Component based development has emerged as the technology of choice to address the challenges that result from this paradigm shift. In the particular context of (safety critical) embedded systems with complex OEM/supplier chains, the following distinguishing features must be addressed. First, the need for high quality, zero defect, software systems calls for techniques in which component specification and integration is supported by clean mathematics that encompasse both static and dynamic semantics-this means that the behavior of components and their composition, and not just their port and type interface, must be mathematically defined. Second, system design includes various aspects-functional, timeliness, safety and fault tolerance, etc.-involving different teams with different skills using heterogeneous techniques and tools. Third, since the structure of the supply chain is highly distributed, a precise separation of responsibilities between its different actors must be ensured. This is addressed by relying on contracts. Following [4] a contract is a component model that sets forth the assumptions under which the component may be used by its environment, and the corresponding promises that are guaranteed under such correct use. The semantic foundations that we present in this paper are designed to support this methodology by addressing the above three issues. At its basis, the model is a language-based abstraction where composition is by intersection. This basic model can then be instantiated to cover functional, timeliness, safety, and dependability requirements performed across all system design levels. No particular model of computation and communication is enforced, and continuous time dynamics such as those needed in physical system modeling is supported as well. On top of the basic model, we build the notion of a contract, which is central to our methodology, by distinguishing between assumptions and promises. This paper focuses on developing a generic compositional theory of contracts, providing relations of contract satisfaction and refinement called dominance, and the derivation of operators for the correct construction of complete systems. In addition to traditional parallel composition, and to enable formal multi-viewpoint analysis, our model includes boolean meet and join operators that compute conjunction and disjunction of contracts. We also introduce a new operator, called fusion, that combines composition and conjunction to compute the least specific contract that satisfies a set of specifications, while at the same time taking their interaction into account. The paper is organized as follows. The principles of our approach are presented in Section 2. Contracts and implementations are introduced in Section 3 and corresponding operations are studied in Section 4. The concept of rich component is formalized in Section 5, by introducing the contracts attached to it. In Section 6 we formalize the concept of designer responsibilities through the notion of controlled/uncontrolled port and we refine our theory of contracts accordingly. How we encompass the different viewpoints is sketched in Section 7 and related work is discussed in Section 8. Principles of Assume/Guarantee Reasoning The main element of our semantic model is a Heterogeneous Rich Component, or simply a component. A component consists of an interface, its expected behavior, and, optionally, one or more implementations. The interface is a set of ports and flows, used by the component to communicate with the rest of the system and with the environment. The expected behavior is described by one or several assumption/promise pairs, called contracts. Contracts can be combined together using three composition operators: greatest lower bound, parallel composition and fusion. The greatest lower bound is used to compose contracts referring to the same component and which use only variables and flows visible from the environment. Parallel composition is used to compute the contract resulting from the composition of several components. Fusion generalizes these two operators, and is capable of handling all cases. In particular, it is used to compose contracts whenever the greatest lower bound and parallel composition operators are inappropriate, for instance when contracts share local variables or flows. Thus, fusion is the implicit composition of contracts, whenever more than one contract is attached to a component. Implementations may be attached to a component, and are usually expressed as extended state machines, or as host tool models. We define several relations between components, contracts and implementations. The compatibility relation relates sets of components. A set of components are incompatible whenever for all environments, at least one of the assumption of at least one component is violated. Contract dominance relates assumptions and promises of two contracts. A contract dominates another when it has weaker assumptions and stronger promises. Satisfaction relates implementations to contracts. An implementation satisfies a contract whenever its behavior, modulo the assumptions, are consistent with the promises. Refinement relates implementations. An implementation refines another whenever it has fewer behaviors. INRIA Throughout this paper we shall need an abstract notion of "assertion". The only facts we need to know about assertions can be summarized as follows: Each assertion E possesses a set of ports and a set of variables that are the vehicle for interaction. An assertion is identified with the set of runs it accepts. A run assigns a history to each variable and port of the assertion. We assume that a proper notion of "complement" for an assertion E is available, denoted by ¬E. When seen as sets of runs, assertions compose by intersection-note that such an operation is monotonic w.r.t. inclusion of sets of runs. When performing this composition, we assume that the appropriate inverse projections have been performed to equalize the sets of ports and variables. Products are equivalently denoted by E 1 × E 2 or E 1 ∩ E 2 . Rich Components, Contracts, Implementations Definition 1 (Implementation). An implementation is simply an assertion, that is, a set of runs. We denote implementations by the symbol M (for "machine"). Implementations are ordered according to the runs they contain. Products preserve implementation refinement. A contract says that under certain assumptions, behaviors are guaranteed to be confined within a certain set. Definition 2 (Contract). A contract C is a pair (A, G), where A, the assumption, and G, the promise, are assertions over the same alphabet. Whenever convenient, we shall denote the assumption and promise of contract C by A C and G C . The interpretation of a contract is made precise by the following definition. Satisfaction can be checked using the following equivalent formulas, where ¬A denotes the set of all runs that are not runs of A: There exists a unique maximal implementation satisfying a contract C, namely: Definition 4 (Rich Component). A rich component is a tuple In ( Thus, every contract has an equivalent contract in canonical form, which can be obtained by replacing G with M C . Hence, working with contracts in canonical form does not limit expressiveness. The operation of computing the canonical form is well defined, since the maximal implementation is unique, and it is idempotent. In the following, we assume that all contracts are in canonical form. This assumption serves two purposes: (i) To have a unique representation of contracts, considered up to equivalence. (ii) To simplify the definition of contract composition operators. Example 1 (Running example: control/monitoring unit). Throughout this paper, we develop the system of Figure 1 to illustrate our notion of contract and its use. It consists of a control unit interacting with a monitoring unit. The system is subject to two independent faults, f 1 for the control unit, and f 2 for the monitoring unit. The nominal behavior of the system (when it should deliver y = a ∧ b at its output. When safe (f 2 = F), the monitoring unit ensures that, if the control unit gets faulty (f 1 = T), the overall system is shut down (y = F) unless a = T holds. Thus the overall system requirement is to maintain the Top Level Exception TLE = ¬a ∧ y false. This TLE may, however, get violated if the monitoring unit gets faulty too (f 2 = T). These requirements are summarized by the two contracts C, for the nominal mode, and C ′ for the exception mode: This separation of concerns into nominal and exception mode is similar to the separation of viewpoints (functional, timed, safety, etc) when handling components via their contracts. ⋄ Boolean algebra As usual, it will be extremely useful to have an algebra of contracts, providing ways of expressing more complex contracts from simpler ones. The following relation of dominance formalizes substituability for contracts and induces a boolean algebra of contracts, which provides such a logic of contracts. Dominance amounts to relaxing assumptions and reinforcing promises. Note that C C ′ and C ′ C together imply C = C ′ . Hence, dominance is a partial order relation. Furthermore, but the converse is not true. Property (5) implies the following result: The following theorem defines the boolean algebra over contracts, implied by . Its proof is straightforward and left to the reader. Theorem 7 (Boolean algebra of contracts). Let C 1 = (A 1 , G 1 ) and C 2 = (A 2 , G 2 ) be contracts. Then, the greatest lower bound of C 1 and C 2 , written Similarly, the least upper bound of C 1 and C 2 , written C = C 1 ⊔ C 2 , is given by The minimal and maximal contracts are ⊥ = (R, ∅) and ⊤ = (∅, R), respectively, where R denotes the set of all runs. Finally, the complement of C is the contract ¬C such that ¬C = (¬A, ¬G); it satisfies ¬C ⊓ C = ⊥ and ¬C ⊔ C = ⊤. Example 2 (Running example: greatest lower bound). The two contracts of (4) represent two viewpoints attached to a same component, corresponding to the nominal and exception modes, respectively. These two contracts involve the same set of ports. Combining them is by computing their greatest lower bound. Putting these two contracts in canonical form and then taking their greatest lower bound yields: This contract assumes that no double failure occurs. Its promise is the conjunction of the promises of C and C ′ . Expanding the promise of this global contract leads to a cumbersome formula, hardly understandable to the user, so we discard it. ⋄ Parallel composition Contract composition formalizes how contracts attached to different rich components should be combined to represent a single, compound, rich component. Let C 1 = (A 1 , G 1 ) and C 2 = (A 2 , G 2 ) be contracts. First, composing these two contracts amounts to composing their promises. Regarding assumptions, however, the situation is more subtle. Suppose first that the two contracts possess disjoint sets of ports and variables. Intuitively, the assumptions of the composite should be simply the conjunction of the assumptions of the rich components, since the environment should satisfy all the assumptions. In general, however, part of the assumptions A 1 will be already satisfied by composing C 1 with C 2 , acting as a partial environment for C 1 . Therefore, G 2 can contribute to relaxing the assumptions A 1 . And vice-versa. Whence the following definition: Note that the so defined contract is in canonical form. The following result expresses the compositionality of the implementation relation: Proof: The assumption of the lemma means that M i ⊆ M Ci , for i = 1, 2. Since the two contracts are in canonical form, we have M Ci = G i and the result follows directly from Definition 8. The following lemma relates greatest lower bound and parallel composition, it relies on the fact that we work with contracts in canonical form: Proof: Both sides of this relation possess identical promises. Thus the only thing to prove relates to the assumptions thereof. From Definition 8 and Theorem 7, the assumption of This proves the lemma. Example 3 (Running example: compositional reasoning). In Example 2, we have shown how to combine the two nominal and exception viewpoints, for the overall system of Figure 1. The system further decomposes into a control and monitoring unit. We would like to associate contracts to each of these components, for each of their viewpoint. Composing these contracts, we should recover the system's overall contract. Since the system is the parallel composition of control and monitoring units, we may reasonably expect that the parallel composition of contracts, for each of these components, should be used. However, we are also combining viewpoints INRIA for these two components and this sould be performed by the greatest lower bound. So, which is the correct answer? The new notion of contract fusion we shall introduce in the following section will provided the adequate answer. Prior to introducing this notion, we need to investigate what it means to eliminate ports in contracts. ⋄ Eliminating ports in contracts Elimination in contracts requires handling assumptions and promises differently. Definition 11 (Elimination). Let C = (A, G) be a contract and let p be any port. Define the elimination of p in C by: where A and G are seen as predicates. Note that we do not require that p be a port of C. Definition 11 is motivated by the following lemma: The following lemma relates elimination and greatest lower bounds: Lemma 13. For any two contracts C 1 and C 2 and any port p, we have: Proof: We have ∀p (A 1 ∪A 2 ) ⊇ (∀p A 1 ∪∀p A 2 ) and ∃p (G 1 ∩G 2 ) ⊆ (∃p G 1 ∩ ∃p G 2 ), which proves (6) as well as the promise part of (7). Regarding the assumption part of (7), we need to prove where . Which proves (8) and the lemma. Elimination trivially extends to finite sets of ports, we denote it by [C] P , where P is the considered set of ports. Set of contracts associated to a rich component We are now ready to address the case of synchronizing viewpoints when local ports are shared between viewpoints. More precisely, we shall formally define what it means to consider a set of contracts associated to a same rich component. Definition 14 (Fusion). Let (C i ) i∈I be a finite set of contracts and Q a finite set of ports. We define the fusion of (C i ) i∈I with respect to Q by where J ranges over the set of all subsets of I. The following particular cases of Definition 14 are of interest: Lemma 15. 1. When Q = ∅, the fusion reduces to the greatest lower bound: In 2. Assume that, for i = 1, 2: holds. Then: 3. Assume that, for i = 1, 2: holds, where C 1 C 2 = (A, G). Then: Condition (11) expresses that each rich component is a valid environment for the other rich components; in other words, the two contracts are attached to two rich components that together constitute a valid closed system. Condition (13) expresses that the restriction to Q of each component is a valid environment for the restriction to Q of the other component. This situation corresponds to two rich components interacting through ports belonging to Q, which are subsequently hidden from outside. Proof: We successively prove the three statements. Statement 1 results immediately from Lemma 10. To prove (12) in Statement 2, note that the two expressions only differ by their assumptions, since the promises of greatest lower bound and parallel composition are identical. For the assumptions, let C 1 C 2 = (A, G) and Regarding Statement 3, (13) implies ∀Q ((A 1 ∩ A 2 ) ∪ ¬G) ⊇ ∀Q (A 1 ∪ A 2 ). Whence (14) follows. The lesson is that fusion boils down to parallel composition for contracts attached to two different sub-components of a same compound component, whereas contracts attached to a same component and involving the same set of ports fuse via the operation of greatest lower bound. The general case lies in between and is given by formula (9). Finally, the various relations that we have established between greatest lower bound, parallel composition, and elimination, allows us to simplify the actual evaluation of the fusion in general. Corresponding heuristics to guide this remain to be developed. Definition 4 for rich components can now be completed. Example 4 (Running example: fusion of contracts). We shall perform a composability study for the two contracts C and C ′ , and then for their fusion [[C, Study of C Consider the following two contracts, for the control and monitoring unit, respectively C 1 = ( ¬f 1 , [x = a ∧ b] ) and C 2 = ( ¬ϕ , y = x ), where ϕ = ¬a ∧ x. Contract C 1 states that, if not faulty, the control unit guarantees that ¬ϕ holds, i.e., invariant a ∨ ¬x holds. Contract C 2 states that the monitoring unit guarantees that, if not faulty, y = x holds unless ϕ does not hold. Putting these two contracts in canonical form and then computing their fusion yields where P is some predicate (which we don't care about), from which we obtain, provided that C is put in canonical form, Study of C ′ Now, let us focus on the other contract C ′ , by proposing the following two local contracts, for the control and monitoring unit, respectively C ′ 1 = ( F , T ), and C ′ 2 = ( ¬f 2 , [y = x ∧ a] ). The first contract is trivial, and the second one states the invariant promised if the monitoring unit is not faulty. We first have Finally, (15)- (16) The remarkable point is that composability works both across components, and viewpoints/modes, i.e., we have [[C, The asymmetric role of ports So far we have ignored the role of ports and the corresponding splitting of responsibilities between the implementation and its environment, see the discussion in the introduction. Such a splitting of responsibilities avoids the competition between environment and implementation in setting the value of ports and variables. Intuitively, an implementation can only provide promises on the value of the ports it controls. On ports controlled by the environment, instead, it may only declare assumptions. Therefore, we will distinguish between two kinds of ports for implementations and contracts: those that are controlled and those that are uncontrolled. The latter property is formalized via the following notion of receptiveness: Definition 17 (Receptiveness). For E an assertion, and P ′ ⊆ P a subset of its ports, E is said to be P ′ -receptive if and only if for all runs σ ′ restricted to ports belonging to P ′ , there exists a run in σ of E such that σ ′ and σ coincide over P ′ . In words, E accepts any history offered to the subset P ′ of its ports. Note that: Lemma 18. If E is P ′ -receptive, then so is E ∪ E ′ for any E ′ having no extra ports or variables than those of E. In some cases, different viewpoints associated with a same rich component need to interact through some common ports. This motivates providing a scope for ports, by partitioning them into ports that are visible (outside the underlying component) and ports that are local (to the underlying component). Definition 19 (Profile). A profile is a 4-tuple π = (vis, loc, u, c), partitioning P as We are now ready to refine our theory of contracts by taking the asymmetric role of ports into account. INRIA Definition 20 (Implementation). An implementation is a pair M = (π, E), where π = (vis, loc, u, c) is a profile over a set P of ports, and E is a u-receptive assertion over P . The last requirement formalizes the fact that an implementation has no control over the values of ports set by the environment. Implementations refine as follows: Definition 21 (Implementation Refinement). For M and M ′ two implementations, say that M refines M ′ , written M M ′ , if and only if π = π ′ and E ⊆ E ′ . In defining parallel composition for implementations, we need to take into account controlled ports. Each implementation is responsible for its set of controlled ports, and, in our theory, such responsibility should not be shared. This motivates the following definition for our parallel composition of implementations associated with different rich components (whence the requirement loc 1 ∩ loc 2 = ∅ in this definition): Definition 22 (Parallel composition of implementations). Let M 1 and M 2 be two implementations such that loc 1 ∩ loc 2 = ∅. Then, M = M 1 \|\| M 2 is defined if and only if c 1 ∩ c 2 = ∅. In that case, E = E 1 × E 2 , and: Theorem 23. Implementation composition is monotonic relative to implementation refinement. Proof: Since profiles refine via identity, this results boils down to the well known monotonicity w.r.t. sets of runs. Definition 24 (Contract). A contract is a triple C = (π, A, G), where π = (vis, loc, u, c) is a profile over a set P of ports, and A and G are two assertions over P , respectively called the assumptions and promises of C. C is called consistent if G is u-receptive, and compatible if A if c-receptive. As pointed out in (3), contracts are in canonical form, meaning that G ⊇ ¬A. If this is not the case, we simply replace G by its most permissive version G∪¬A, which cannot per se break consistency, thanks to Lemma 18. The sets A and G are not required to be receptive. However, if G is not u-receptive, then the promises constrain the uncontrolled ports of the contract. This is against our policy of separation of responsibilities, since we stated that uncontrolled ports should remain entirely under the responsibility of the environment. Corresponding contracts are therefore called inconsistent. Definition 25 (Satisfaction). An implementation M satisfies contract C, written M \|= C, iff π M = π C and E M ⊆ G C . By Lemma 18, if contract C is consistent, then M C = G C ∪ ¬A C is still the maximal implementation satisfying C. We now turn to the relation of dominance and its consequences. Proof: This is a direct consequence of Theorem 7 Finally, it remains to define the parallel composition of contracts. Having done this we can directly borrow the definition 14 of fusion, for contracts enhanced with profiles. Definition 28 (Parallel composition of contracts). Let C 1 = (π 1 , A 1 , G 1 ) and C 2 = (π 2 , A 2 , G 2 ) be contracts. The parallel composition, or product, of C 1 and C 2 , written C = C 1 \|\| C 2 , is defined if and only if c 1 ∩ c 2 = ∅, and in that case is the contract C = (π, A, G) defined by: Unlike Definition 22, we do not require here that loc 1 ∩loc 2 = ∅. The reason is that we wish to encompass the composition of different viewpoints attached to a same rich component. (For contracts attached to different rich components, however, we do have loc 1 ∩ loc 2 = ∅.) With parallel composition, we can formalize the notion of contract compatibility. Recall that a contract is compatible whenever A is c-receptive. If not, then there exists a sequence of values on the controlled ports that are refused by all acceptable environments. However, by our definition of satisfaction, implementations are allowed to output such sequence. Unreceptiveness, in this case, implies that a hypothetical environment that wished to prevent a violation of the assumptions should actually prevent the behavior altogether, something it cannot do since the port is controlled by the contract. Therefore, unreceptive assumptions denote the existence of an incompatibility internal to the contract, that cannot be avoided by any environment. This justifies the following definition. Definition 29 (Compatibility). Two contracts C 1 and C 2 are compatible if and only the assumption of their parallel composition is receptive with resepct to the controlled ports. Assumptions may become unreceptive as a result of a parallel composition even if they are not so individually. This is because the set of controlled ports after a composition is strictly larger than before the composition. In particular, ports that were uncontrolled may become controlled, because they are controlled by the other contract. INRIA Note that consistency and compatibility may not be preserved by Boolean operations. For example, one obtains an inconsistent contract when taking the greatest lower bound of two contracts, one of which promises that certain behaviors will never occur in response to a certain input, while the other promises that the remaining behaviors will not occur in response to the same input. Both contracts have legal responses to the input, but their intersection is empty, thus making the combination unreceptive. In this case, inconsistency is due to two contracts making inconsistent promises. Addressing Multiple Viewpoints An important question is: can our abstract notion of "assertion" encompass the different functional and non-functional viewpoints of system design? Since assertions are just sets of runs, we can, in particular, accomodate hybrid automata following [5]. So seemingly, we can in particular support functional, timeliness, safety, as these can be modeled by specific subclasses of hybrid automata. A closer investigation reveals that we need to deal with classes of models that are stable under parallel composition (defined by intersection), union, and complement. Taking complements is a delicate issue: hybrid automata are not closed under complementation; in fact, no model class is closed under complementation beyond deterministic automata. To account for this fact, various countermeasures can be considered. First, the designer has the choice to specify either E or its complement ¬E (e.g., by considering observers). However, the parallel composition of contracts requires manipulating both E and its complement ¬E, which is the embarrasing case. To get compact formulas, our theory was developed using canonical forms for contracts, systematically. Not enforcing canonical forms provides room for flexibility in the representation of contracts, which can be used to avoid manipulating both E and ¬E at the same time. A second idea is to redefine an assertion as a pair (E,Ē), whereĒ is an approximate complement of E, e.g., involving some abstraction. In doing so, one of the two characteristic properties of complements, namely E ∩Ē = ∅ or E ∪Ē = ⊤, do not hold. However, either necessary of sufficient conditions for contract dominance can be given. The above techniques are the subject of ongoing work and will be reported elsewhere. Related Work The notion of contract derives from the theory of abstract data types, first suggested by Meyer in the context of the programming language Eiffel [6]. In his work, Meyer introduces preconditions and postconditions as assertions for the methods of a class, and invariants for the class itself. Preconditions correspond to the assumptions under which the method operates, while postconditions express the promises at method termination, provided that the assumptions are satisfied. Invariants must be true at all states of the class regardless of any assumption. To guarantee safe substitutability, a subclass is only allowed to weaken the assumptions and to strengthen the promises. Similar ideas were in fact, already present in earlier work by Dill, although phrased in less explicit terms [7]. Dill proposes an asynchronous model based on sets of sequences and parallel composition (trace structures). Behaviors (traces) can be either accepted as successes, or rejected as failures. The failures, which are still possible behaviors of the system, correspond to unacceptable inputs from the environment, and are therefore the complement of the preconditions. Safe substitutability is expressed as trace containment between the successes and failures of the specification and the implementation. Wolf later extended the same technique to a discrete synchronous model [8]. More recently, De Alfaro and Henzinger have proposed Interface Automata which are similar to synchronous trace structures, where failures are implicitly all the traces that are not accepted by an automaton representing the component [9]. Composition is defined on automata, rather than on traces, and requires a procedure to restrict the state space that is equivalent to the process called autofailure manifestation of Dill and Wolf. A more general approach along the lines proposed by Dill and Wolf is the work by Negulescu with Process Spaces [10], and by Passerone with Agent Algebra [11], both of which extend the algebraic approach to generic behaviors introduced by Burch [12]. Our notion of contract supports speculative design in which distributed teams develop partial designs concurrently and synchronize by relying on the notions of rich component [4] and associated contracts. We define assumptions and promises in terms of behaviors, and use parallel composition as the main operator for decomposing a design. This choice is justified by the reactive nature of embedded software, and by the increasing use of component models that support structured concurrency. We developed our theory on the basis of assertions, i.e., languages of generic "runs". To achieve the generality of a (mathematical) metamodel we have complemented this by developing a concrete model for such assertions, that encompasses the different viewpoints of the design [13]. In our approach, behaviors are decomponsed into assumptions and promises, as in Process Spaces, a representation that is more intuitive than, albeit equivalent to, the one based on the successes and failures of asynchronous trace structures. Unlike Process Spaces, however, we explicitly consider inputs and outputs, which we generalize to the concept of controlled and uncontrolled signals. This distinction is essential in our framework to determine the exact role and responsibilities of users and suppliers of components. This is concretized in our framework by a notion of compatibility which depends critically on the particular partition of the signals into inputs and outputs. We also extend the use of receptiveness of asynchronous trace structures, which is absent in Process Spaces, to define formally the condition of compatibility of components for open systems. Our refinement relation between contracts, which we call dominance to distinguish it from refinement between implementations of the contracts, follows the usual scheme of weakening the assumption and strengthening the guarantees. The order induces boolean operators of conjunction and disjunction, which resembles those of asynchronous trace structures and Process Spaces. In addition, we also define a new fusion operator that combines the operation of composition and conjunction for a set of contracts. This operator is introduced to make it easier for the user to express the interaction between contracts related to different viewpoints of a component. INRIA The model that we present in this paper is based on execution traces, and is therefore inherently limited to representing linear time properties. The branching structure of a process whose semantics is expressed in our model is thus abstracted, and the exact state in which non-deterministic choices are taken is lost. Despite this, the equivalence relation that is induced by our notion of dominance between contracts is more distinguishing than the traditional trace containment used when executions are not represented as pairs (assumptions, promises). This was already observed by Dill, with the classic example of the vending machine [7], see also Brookes et al. on refusal sets [14]. There, every accepted sequence of actions is complemented by the set of possible refusals, i.e., by the set of actions that may not be accepted after executing that particular sequence. Equivalence is then defined as equality of sequences with their refusal sets. Under these definitions, it is shown that the resulting equivalence is stronger than trace equivalence (equality of trace sets), but weaker than observation equivalence [15,16]. A precise characterization of the relationships with our model, in particular with regard to the notion of composition, is deferred to future work. Conclusion We have presented mathematical foundations for the contract-based model developed in the framework of the SPEEDS project. Our generic mathematical model of contract supports "speculative design". This is achieved by focusing on behaviors, by supporting the notion of rich component where diverse (functional and non-functional) aspects of the system can be considered and combined, by representing rich components via their set of associated contracts, and by formalizing the whole process of component composition through the general mechanism of contract fusion. These foundations support the Heterogeneous Rich Component (HRC) metamodel under development in SPEEDS [13]. Future work includes the development of effective algorithms to handle contracts, coping with the problems raised by complementation.	2007-06-13T02:16:32.000Z	2007-06-11T00:00:00.000	{ "year": 2007, "sha1": "180d18f43064ecd3c5b768f87e308110d2e15095", "oa_license": null, "oa_url": null, "oa_status": null, "pdf_src": "Arxiv", "pdf_hash": "b909fa8af0cc36aa5d2f233770d3608015c7ea38", "s2fieldsofstudy": [ "Computer Science" ], "extfieldsofstudy": [ "Computer Science" ] }
266166202	pes2o/s2orc	v3-fos-license	Modeling hydro, nuclear, and renewable electricity generation in India: An atom search optimization-based EEMD-DBSCAN framework and explainable AI Background and objective Tracking clean electricity generation in developing economies is highly challenging owing to the influence of turbulent external factors. Clean electricity is a significant enabler of striving toward environmental sustainability. In this research, we aim to model hydro, nuclear, and renewable electricity generation in India through applied predictive modeling. We also strive to uncover the influence of the critical determinants responsible for clean electricity growth. Methodology We propose a granular predictive framework comprising ensemble empirical mode decomposition, clustering applications in spatial data based on density, including noise, and atom search optimization-based novel optimization methodology to predict absolute figures of clean energy generation. The framework uses a series of socio-economic factors reflecting household demand and industrial growth in India as explanatory variables. Results The rigorous scrutiny of the predictive framework specifies hydro electricity generation is relatively more predictable during the time horizon influenced by the COVID-19 pandemic. The deployment of dedicated explainable artificial intelligence (AI) tools suggests an increased adoption of clean electricity in selected industrial sectors in India, which broadly governs the evolutionary pattern. Conclusion The underlying research is the first of its kind to fathom the daily temporal dynamics of clean electricity generation in the Indian context. Consideration of three distinct clean electricity sources during highly volatile time regimes underscores the contribution of the work. The predictive framework survives a stringent performance check, which justifies the robustness of the same. Demand in different industrial sectors in India profoundly influences the growth toward clean electricity. Background and objective: Tracking clean electricity generation in developing economies is highly challenging owing to the influence of turbulent external factors.Clean electricity is a significant enabler of striving toward environmental sustainability.In this research, we aim to model hydro, nuclear, and renewable electricity generation in India through applied predictive modeling.We also strive to uncover the influence of the critical determinants responsible for clean electricity growth.Methodology: We propose a granular predictive framework comprising ensemble empirical mode decomposition, clustering applications in spatial data based on density, including noise, and atom search optimization-based novel optimization methodology to predict absolute figures of clean energy generation.The framework uses a series of socio-economic factors reflecting household demand and industrial growth in India as explanatory variables.Results: The rigorous scrutiny of the predictive framework specifies hydro electricity generation is relatively more predictable during the time horizon influenced by the COVID-19 pandemic.The deployment of dedicated explainable artificial intelligence (AI) tools suggests an increased adoption of clean electricity in selected industrial sectors in India, which broadly governs the evolutionary pattern. Conclusion: The underlying research is the first of its kind to fathom the daily temporal dynamics of clean electricity generation in the Indian context.Consideration of three distinct clean electricity sources during highly volatile time regimes underscores the contribution of the work.The predictive framework survives a stringent performance check, which justifies the robustness of the same.Demand in different industrial sectors in India profoundly influences the growth toward clean electricity. Introduction Renewable energy sources are vital for a sustainable low-carbon society [1][2][3].The power in the form of electricity manifests the status of household and industrial activities of a nation [4].The nexus of electricity supply and demand is largely governed by setting reasonable electricity prices.The predictability of electricity prices for boosting the market has been introspected thoroughly in the literature [5,6].Nevertheless, it is equally essential to delve into the dynamics of the electricity generation process to anticipate the futuristic movements of energy reserve, excess power, growth of the business, etc.A precise prediction of electricity generation from renewable sources explicitly can be translated into practical implications for tracking the overall economic affairs [7][8][9].An accurate forecast of electricity generation can assist policymakers in managing the energy mix to ensure grid stability, security of supply, etc., by adequately integrating renewable energy sources. Mining the inherent pattern generation and energy consumption has been chiefly confined at the micro level, covering buildings, commercial processes, vehicles, etc.The said research category needs to be revised to measure the appetite for clean energy intake in nations in the context of de-carbonization.On the other hand, as stated, electricity price forecasting has seen considerable traction in literature [10][11][12][13], predominantly owing to thwarting the supply-demand vagary in extreme weather conditions, geopolitical conflicts, etc.The influence of renewables on electricity price prediction has also been explored [14,15].Renewable electricity generation is linked to societal implications, too [16].Crozier and Baker [17] exemplified the utility of renewable electricity in cross-border interconnection in terms of power trading.Renewable resources have been marked to be necessary for electricity mix transformation in Germany [18].Kan et al. [19] elicited the utility of hydro power for low-cost renewable power generation.Khosravi et al. [20] highlighted how investment in nuclear power could help stabilize the future electricity prices in Finland, enabling the transition away from coal by 2029.The effectiveness of hydro and nuclear assets in generating sustainable power is widely acknowledged in the literature [21][22][23][24].The paucity of research to model clean electricity generation at the macro scale is amply apparent.Considering the disruptive impact of the COVID-19 pandemic on production and business, and most recently, the unprecedented military conflict between Russia-Ukraine, it is arduous and practically necessary to closely introspect the pattern and dependence structure of clean energy generation from different renewable sources.The lack of substantial research in developing a predictive analytics framework is a significant roadblock to comprehending the interplay of clean energy generation with socio-economic factors. The current work endeavors to fill the research void by creating a novel modeling framework to predict and decode clean energy generation patterns at the country level.The contribution of this research can be divided into three main categories.Firstly, we strive to predict India's daily hydro, nuclear, and renewable electricity generation, explicitly covering the COVID-19 pandemic timeline.Predicting the structure of electrical power generation of selected resources is the first of its kind compared to the existing literature.Secondly, the current work profoundly delves into the dependence of the chosen electricity generation components on a series of socioeconomic factors in the Indian context.Activities of commoners captured through Google search volume index data are used as determinants of electricity generation, which act as representations of demand.Additionally, the financial outlook of several sectoral indices is used as explanatory variables to incorporate the demand for electricity for different industrial purposes.The aforesaid combination of explanatory variables effectively accounts for the influence of both household and industrial demand on electricity generation.Thirdly, we also contribute to the methodology front by designing a hybrid granular predictive structure and vis-a-vis deploying state-of-the-art explainable artificial intelligence tools are employed to interpret how selected features influence predictions, facilitating the derivation of significant implications. The methodological framework propounds a granular predictive structure driven by decomposition to achieve the goals.The daily time observations of raw electricity generation in India from hydro, nuclear, and renewable resources are initially disentangled into granular sub-series using the ensemble empirical mode decomposition (EEMD) procedure.Afterward, the sub-series are clustered into high and low-frequency counterparts.The Hurst exponent and fuzzy entropy estimates of the respective sub-series are considered for clustering.To perform the time series clustering, the density-based spatial clustering of applications with noise (DBSCAN) has been adopted.Subsequently, predictive exercises are carried out on clustered high and low-frequency sub-series separately to yield component-wise predictions, aggregating to determine the final predictions.We design an ensemble of ensembles (EoE) type predictive structures infusing random forest (RF), bagging, gradient boosting (GB), and atom search optimization (ASO) in an optimization setup.Breaking down complex time series data into high and low-frequency segments has proven beneficial for forecasting volatile data.However, our novel predictive framework stands out due to its seamless integration of diverse tools, ensuring exceptionally accurate predictions.Rigorous numerical and statistical tests validate its superiority over benchmark models.Additionally, we derive crucial insights into how specific socio-economic factors impact electricity generation.We utilize explainable AI methods to uncover both global and local dependency structures.Using permutation feature evaluation, we pinpoint key contributors to clean energy generation in India.Then, accumulation local effect plots reveal the influence patterns of individual features, aiding strategic decision-making for managing hydro, renewable, and nuclear energy generation.Finally, the local interpretable model-agnostic explanations (LIME) framework offers insights into the prediction process at a local level.The deployment of XAI methodologies on top of the proposed granular predictive structure underscores the underlying research's effort to offer practical insights apart from the theoretical contributions. The subsequent sectioning of this article unfolds as follows: Section 2 delineates related literature, aiming to identify research gaps and contextualize this work.Section 3 meticulously details the research methodology, describing individual tools and the procedural flow.Section 4 focuses on discussing data sources and the statistical properties of variables.Detailed results and comprehensive analyses are presented in Section 5. Section 6 critically examines the implications of the findings and proposes avenues for future research.Finally, Section 7 encapsulates the manuscript's conclusion. I. Ghosh et al. Literature review Tracking and predicting the electricity generation process has predominantly been carried out by analyzing the impact of different process parameters, climate factors, and physical microstructural properties [25].Alternatively, using past historical records either as raw explanatory features or technical indicators for predicting electricity and energy generation patterns has been documented, too [26].Implications manifested in the form of empirical research on the properties of renewable electricity have been highlighted in the literature.Sravan et al. [27] pointed out that Indian electricity consumption did not influence renewable electricity production.It was found that natural resource rents and greenhouse gas emissions affected renewable electricity generation positively.We summarize the research trend of clean electricity consumption and its impact on holistic development in developing economies in Table 1 to properly position the underlying work. The aforesaid literature clearly points out the close interlinkage among clean electricity generation, economic development, and societal reforms.Hence, the selection of the pertinent explanatory features in the Indian context will be critical to predict hydro, nuclear, and renewable electricity generation precisely.Also, the literature above primarily considers one specific source of clean electricity.We aim to extend the same by selecting three different sources of clean electricity generation for comprehensive modeling.We now enunciate the critical findings of the cognate research on the prediction of power and electricity generation. Esfetang and Kazemzadeh [36] developed a hybrid predictive framework combining wavelet transformation, neural network, and weight-improved particle swarm optimization (WIPSO) for precisely modeling the generation of electrical power in wind farms.Meteorological characteristics played a critical role in the entire exercise.Guo et al. [37] explored the correlation between climatic, hydrological, and socio-economic factors to forecast monthly hydroelectric generation, electricity demand, and greenhouse gas emissions.The methodological framework comprised of artificial neural network (ANN) fine-tuned through an enhanced electromagnetic field optimization algorithm (IEFO)which outclassed several benchmark tools in terms of accuracy.S ¸ahin et al. [38] relied upon seasonal grey and machine learning-based models for estimating monthly electricity production in France, Germany, Spain, Turkey, and the UK.The methodological frameworks transpired to draw highly accurate forecasts and provided a share of renewables in total electricity generation.Research by Jiang et al. [39] successfully decoded the predictability of energy generation in electric buses using a Markov-based Gaussian Process Regression model (M-GPR), where the deployment of process-specific external variables appeared to be highly effective.Jin et al. [40] adopted deep reinforcement learning and the Markov Decision Process (MDP), closely tracking power fluctuations and predicting the overall energy consumption in buildings with high precision.The framework demonstrated statistical superiority over various competing models.Li et al. [41] designed a hybrid back propagation neural network (BPNN) optimized by the improved particle swarm (IPSO) algorithm for electricity consumption prediction during the COVID-19 pandemic regime effectively.The predictive structure utilized medical information, public opinion, policy data, and historical records of electricity consumption during the pandemic for drawing forecasts.Lu et al. [42] successfully developed a hybrid predictive framework incorporating an improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise and Support Vector Machine (SVM) for predicting daily electricity demand in the US during the COVID-19 pandemic.It was revealed that the daily infection rate largely influenced the prediction accuracy.Shi et al. [43] propounded a sliding window and dual-channel convolutional neural network for capturing temporal characteristics focused on accurately estimating both coal and electricity consumption within a specific 5-min interval during the cement calcination process.The proposed methodology outperformed a series of machine and deep learning algorithms. The careful scrutiny of the past literature clearly elicits the scarcity of research to model clean energy generation patterns for specific resources.The past research is mostly restricted to pattern discovery of conventional coal-based energy consumption and generation pattern.However, owing to the growing need to focus on clean electricity generation, it becomes imperative to analyze the predictability of the same in practical terms, it's essential to assess the dependence structure between the energy generation and factors representing household and industrial demand to draw actionable insights.The prevailing predictive modeling literature of Table 1 Pertinent literature. Authors Endeavor and Methodology Location Tiwari et al. [28] Confirmed the correlation between electricity consumption and economic growth using the Granger causality test within a vector autoregression framework India Liu et al. [29] Propounded a quantum genetic algorithm-inspired fractional grey polynomial model to accurately estimate the electricity consumption India and China Rajkumari [30] Explored the interplay between electricity consumption and economic reforms through Granger causality assessment and forecasted electricity demand using the Holt-Winters smoothening method India Wassie and Ahlgreen [31] Delved into the influence of household size and configuration, income distribution, load pattern, etc., on solar electricity consumption using statistical modelling. East Africa Xu et al. [32] Expounded on the strong direct and indirect nexus between clean electricity growth, ecological awareness and economic welfare using structural equal modelling China Abbas et al. [33] Investigated how green finance, environmental tax policies, and geopolitical risk factors contribute to shaping investments within the renewable electricity ecosystem China Casati et al. [34] Created a social clean energy access index aimed at assessing the societal effects of clean electricity and determining optimal countries for investing in renewable energy infrastructure Sub-Saharan Africa Espoir et al. [35] Studied the influence of renewable electricity consumption on broader economic and financial growth employing both linear regression and kernel-based regularized least square models Africa I. Ghosh et al. consumption and generation patterns sheds little light on the same.Hence, the current research's strategic alignment to address these potential gaps is well-justified.Considering the success of machine and deep learning models in forecasting future trends, the methodology outlined in this paper efficiently harnesses these models to construct a detailed predictive model.Subsequently, the underlying dependency patterns are revealed through eXplainable AI (XAI) methodologies. Research methodology Here, we elucidate the methodological framework designed to perform predictive modeling and subsequently explain the impact of considered explainable variables on chosen clean electricity generation series.As mentioned, the granular framework initially decomposes the underlying series into granular subseries using the EEMD method, which is subjected to the DBSCAN clustering algorithm to form the high and low-frequency series.Clustering is performed based on fuzzy entropy (FENT) and Hurst exponent (HEXP) values of the decomposed series.We explain the components used for the said restructuring at first. Ensemble empirical mode decomposition (EEMD) EEMD [44] is a method that modifies the traditional empirical mode decomposition (EMD) approach.It is applied to separate signals that are nonlinear and nonstationary into a component called intrinsic mode function (IMFs) and another component that is the residual.We briefly outline IMFs and EEMD steps. The steps of generating IMFs: Step 1. Find all the local minimums and maximums of the time series X(t) and use cubic spline to interpolate them to form lower envelop L(t) and upper envelop U(t). Step 2. Compute the mean of lower and upper envelopes as equation ( 1): Obtain a local detail as eq.( 2): Step 3. Perform Steps 1 and 2 until conditions in steps 3.1 and 3.2 are satisfied: Step 3.1.The value of M(t) limits to 0. Step 3.2.The difference between local extrema and zero crossings is at most 1. The first IMF, F 1 (t), is given by Z(t) provided it satisfies Steps 3.1 and 3.2.Alternatively, the maximum number of iterations can be defined instead of performing Steps 3.1and 3.2 for extracting F 1 (t).The original series is reduced by the extracted IMF to create the residual series (eq.( 3)): Step 4. Repeat Steps 1 to 3 until the desired number of IMFs is extracted, and the error component has at most two local extrema or termination criteria reached.Finally, eq. ( 4) expresses the original time series X(t) after extracting n number of IMFs as: The EEMD removes the mode mixing problem of the EMD, resulting in IMF, containing signals spanning a wide band of frequency.The steps of generating EEMD: Step 1. Perturb the time series X(t) by adding noise components to generate the series (eq.( 5)): Where ε(t) represents independent Gaussian white noise, and I denotes the number of trials. Step 2. Use EMD on individual transformed series to get the respective IMFs and residuals (eq.( 6)). Step 3. Compute the average outcome of each trial to cancel out the effect of uncorrelated white noise while preserving the meaningful information and getting back the original series (eq.( 7)). Where I. Ghosh et al. Density-based spatial clustering of applications with noise (DBSCAN) It is a density-based clustering algorithm capable of automatically estimating an optimal number of clusters of any shape in a dataset [45].It is reliant upon two important parameters, namely, epsilon, the distance of the vecinity around a data point, and minPts, referring to the number of points within the radius of epsilon to construct clusters based on the density.The parameter epsilon is used to classify the data points into core and noise points.The core points must exceed the minPts.For a data point p if another point q lies within the epsilon neighborhood, then p is treated as a core point, and the connection between the points is referred to as directly density reachable.The point p is categorized as density reachable from q if a succession of points (p 1 , …, p n ; p 1 = p; p n = q) exists such that p i+1 is directly density reachable from p i . The algorithmic procedures of DBSCAN are enunciated below: Step 1. Traverse the data points to discover the core points with respect to the figures of epsilon and minPts. Step 2. Starting from any core point, mark the directly density-reachable and the density-reachable points to form a cluster. The leftover points not assigned to any clusters are treated as noise. Hurst exponent (HEXP) This work utilizes the rescaled range (R/S) analysis-based [46] Hurst exponent [47] figure to identify the high-frequency and low-frequency counterparts of the decomposed series.The procedural steps to estimate the same are jotted down as follows. Step 1: At first, the underlying time series (R N ) of length N is segmented into d groups of continuous subseries of length n. Step 3: The cumulative deviation from the mean of the subseries is, thereafter, calculated as eq.( 8): Step 4: The range (R d ) is determined as eq.( 9): Step 5: The standard deviation (S d ) of the respective subseries is calculated in eq (10). Step 6: The rescaled range mean figure for the underlying sub-series is determined as eq (11): The R/S statistic and Hurst coefficient (H) are asymptotically related in eq. 12 Wherein C is a constant Finally, the Hurst exponent (H) value is estimated by applying a standard ordinary least squares regression (OLS) on eq. ( 13). An H value of 0.5 implies the underlying time series perfectly follows an independent and identically distributed (iid) Gaussian Random Walk model.On the flip side, if its value is greater than 0.5, a persistent presence trend characterized by long-memory dependence is concluded.Time series characterized by long-memory dependence ideally suggests the dominance of trend component, inferring low-frequency traits.The presence of high-frequency components is inferred if the estimated figure of H is less than 0.5, which suggests strong dominance of an anti-persistent pattern.We determine the values of H of respective decomposed series of selected energy generation indicators. Fuzzy entropy Entropy is an effective reflector of the extent of disorder in thermodynamic systems.It has been successfully used to gauge the degree of complexity and volatility of time series data [48].In this work, we rely upon the fuzzy entropy measure used by Ref. [49] for forecasting Carbon futures for clustering decomposed components of respective energy generation series into sub-series.The procedural steps are outlined below. Step 1: For a given time series, x(t) with embedding dimension m, the m dimensional vector representation is expressed as eq.( 14): Where 1 ≤ i ≤ T − m + 1. Step 2: Eq. ( 15) shows the distance between two vectors, X(i), X(j): Step 3: The similarity (D m ij ) between X m i and X m j is determined by estimating the fuzzy membership function as eq.( 16): where n is a parameter, and r represents a tolerance parameter. Step 4: The following functions are calculated on top of the fuzzy similarity score (eqs ( 17) and ( 18)): Step 5: Finally, Eq. ( 19) computes the fuzzy entropy of the time series, x(t): The value of r, denoting the fuzzy function extent of the border, is set as 0.25σ SD , where σ SD denotes the time series standard deviation, the embedding dimension, m is set to be 3, and the value of n assigned to 2 in this work.The figures of fuzzy entropy for respective decomposed subseries alongside the trend component are estimated.Higher entropy values indicate high variability and thereby show high-frequency components.Fig. 1 depicts the integrated process to generate the aggregate series from hydro/nuclear/renewable electricity generation series. The ASO-based predictive structure is applied to the reconstructed aggregate series of electricity generation from respective clean resources.We run the predictive framework on both aggregate high and low component series separately for individual series.The estimated forecasts on both counterparts are added to obtain the final predictions.The ASO algorithm is used in an optimization setup to combine predictions obtained from the individual ensemble models, RF, Bagging, and GB.Thus, the approach can alternatively be classified as EoE framework.The descriptions of the constituent components are elaborated below. Atom search optimization (ASO) Propounded by Zhao et al. [50], ASO is a population-based metaheuristic search algorithm mimicking molecular dynamics.It is inspired by the natural atomic motion, subjected to interaction forces and geometric constraints.Each atom represents a potential solution in search space and traverses toward the target via acceleration.Atoms interact with each other in the feasible domain to obtain the best solution.The motion of an atom is described using equations ( 20) and (21). I. Ghosh et al. Where i denotes an atom in the population, d= 1, …, D; D accounts for the number of decision variables, v d i , and x d i are the speed and location of the ith atom, t stands for the iteration, rand is random number in the range [0, 1], and a d i is the acceleration component.Considering the interaction force and geometric constraints, the acceleration of atoms is estimated using the principles of Newtonian mechanics as expressed in equation (22). The interaction fore and the geometric constraint between the atoms are represented by F d i and G d i , α represents the depth weight, β stands for multiplier weight, h ij is an adaptive ratio of Euclidean distance between atom i and j, to the distance at which the interparticle potential is 0, T is the maximum number of repetitions.The interaction force, geometric constant, and mass of the respective particles are determined using equations ( 23)- (26). Here, F d ij (t) is the interaction force between atoms i and j, and K best denotes the collection of best K atoms as per the fitness.Thus, individual atoms are engaged in transfusing information with K neighbors.The best and the worst fitness values at iteration t are denoted by Fit best (t) and Fit worst (t).The neighborhood size is controlled over the iterations as follows (eq (27)): N denotes the population size.The methodological framework to perform predictive modeling attempts to utilize the capacity of the ASO algorithm, which has been highly regarded for resolving complex NP-Complete optimization problems [51,52], to intelligently traverse the search space for combing three ensemble learning algorithms.The rationale for the selection of the ASO algorithm also covers the endeavor to explicate its theoretical potential in building a sophisticated granular time series forecasting methodology.The faster convergence and ability to penetrate the farthest solution space by thwarting local optimal make it ideal for complex optimization problems.The ASO-based EoE structure alternatively can be regarded to be an ensemble of ensembles owing to the flow of operations.The novelty of the predictive frame lies in adopting ASO in building the EoE structure for predicting a pattern of paramount significance in the context of sustainability.Here, we enunciate a brief overview of the three ensemble models. Random forest (RF) Breiman [53] developed RF, a typical ensemble machine learning algorithm predominantly used for predictive modeling.It uses a set of base learning algorithms in parallel for making the initial predictions.Generally, the orthodox decision tree is used as base learners.In this work, the classical regression tree (CART) has been used for that role.Each base learner is constructed based on a bootstrapped data segment of the entire training data samples.At each node of the selected tree, the best feature for the split is determined based on a randomly chosen subset of features.The final prediction is obtained by averaging the individual base learners estimates.These algorithms have been successfully applied in the predictive modeling of financial time series [54,55]. Bagging Originated by Breiman [56], bagging is another ensemble machine learning algorithm wherein a series of CART is deployed as the base learners in parallel.Similar to RF, the base learners are built on bootstrapped samples.However, it does not select features in underlying base learners on a random subset of features.Unlike RF, all features are evaluated to identify the most suitable one, thereby striving to reduce the variance of unstable learning in CART.The final predicted value is the average of all the individual learners.Bagging has been successfully used for complex pattern mining [57]. Gradient boosting (GB) Propounded by Schapire and Singer [58], Boosting is also an ensemble machine-learning technique that applies a series of base learning algorithms forward-stage-wise to produce the final predictions.The GB is a variant of the traditional boosting algorithm that uses gradient-driven error rate for the identification of training samples.As individual learners, classical regression trees are trained in each stage sequentially in a forward direction.The sequential ensemble approach assists in reducing the inaccuracy of the prediction.Akin to RF and bagging, GB has been found to be highly effective in mining complex patterns [59,60]. The entire simulation has been performed using Python programming language, wherein the 'GridSearchCV' utility of the 'sklearn' library is used for parameter tuning of respective methods.The 'mealpy' library is used for implementing the ASO algorithm. Optimization framework for drawing final predictions As discussed, the ASO framework is utilized to combine the energy consumption forecasts produced by RF, Bagging, and GB to fetch the final prediction.The ASO-based structure computes the weighted average of individual predictions for drawing the final outcome.The mathematical formulation is narrated in equation (28).29)). For finding the optimal weights, we propose the following optimization problem (eq.( 30)): The objective function mentioned in equation ( 6) can be represented by following an equivalent optimization framework (eq.( 31)-( 38)). subject to The ASO algorithm is deployed to iteratively fetch the near-optimal values of the respective weights to augment the accuracy of the final predictions.To assess the predictive performance, the present research utilizes four metrics as defined next. Performance indicators The following criteria are used to measure how well the proposed XAI framework performs.Let (Y t ) denotes the observed series and ( Ŷt ) denote the estimated series.Then the four measures, Nash-Sutcliffe Efficiency (NSE), Index of Agreement (IA), Theil Index (TI), and Directional Predictive Accuracy (DA) are defined as follows (eq.( 39)-( 42)): A predictive modeling approach will be highly efficient if the NSE, IA, and DA values approach to one and the TI value approach to zero.The Diebold-Mariano (DM) test for equal predictability analysis is used in this work to statistically ascertain the competing models' relative efficiency.The DM test can evaluate the accuracy differences between multiple forecasting models using meansquared residual.We compare the models using mean square prediction error (MSPE) as the loss function. The integrated approach is sufficient to gauge the extent of predictability of the underlying, but it reduces model interpretability.To overcome this limitation, the current work uses dedicated XAI tools, explained below. Explainable artificial intelligence (XAI) To understand how the selected explanatory features affect the respective electricity generation series, the current work invokes several XAI models outlined below to accomplish the task. Permutation feature importance Breiman [53] originally developed orthodox permutation feature importance to understand the effect of the explanatory features in a random forest model.Fisher et al. [61] later extended it to use it as a model agnostic tool.In the updated scheme, the importance of any feature is calculated by randomly changing its original values and measuring the impact on the overall predictive accuracy of the model.A higher error means a higher importance of the feature. Accumulation local effect plot The accumulation local effect (ALE) plot was initially introduced by Apley and Zhu [62], which measures the average impact of chosen independent variables on the predictions of developed machine learning models by exposing the black box operations.ALE plots are quicker to build and give correct interpretation in the case of correlated features, unlike the PDP plots. The local-dependence (LD) profile for a model f() and predictor X j was later developed by Apley and Zhu [63] as (eq.( 43)): Basically, it is the expected value of the model predictions over the conditional distribution of X − j given X j = z. The LD profile is calculated as follows (eq.( 44)): Where N j represents the set of observations utilized for determining the conditional distribution of X − j ⃒ ⃒ X j = z, with the values of X j close to z. A smooth estimator for the LD profile can be written as (eq.( 45)): The weight components w i (z) indicates the distance between z and x j i .The Accumulated-local (AL) profile for the model f() and predictor X j is estimated as (eq.( 46)): I. Ghosh et al. Where , z 0 shows a value near the lower bound of the distribution X j , and c is the constant. The local variation of the model due to X j is measured by q j ( x j\|=v ) .The amount of changes is averaged to calculate the accumulation of local effects.The mentioned formulation is very effective in exploring contributions for correlated feature. Local interpretable model-agnostic explanations Local Interpretable Model-Agnostic Explanations (LIME) is an explainable AI model to understand machine learning models locally, developed by Ribiero et al. [64].LIME generates a novel dataset based on a learned model by modifying the input variables' figures and getting the target variable's predictions.It then attempts to map the relationship between the target and input variables on the new dataset using more interpretable machine learning models, viz., decision trees, LASSO, etc.It is mostly used to assess the local influence structure of the explanatory feature set. Data description & statistical properties The CEIC global database [65] is used for collating data of variables for conducting the experiments.The samples of daily hydro, nuclear, and renewable energy generation in a million units (MU) of India from July 15, 2019, to June 30, 2022, are compiled to test the predictability.Renewable electricity is extracted from sources comprising solar, biomass, etc.The sample duration of the study duly covers the COVID-19 pandemic time horizon, which underscores the contribution of the study.The following exhibits, Figs.2-4, provide the visual depiction of the original series and respective histograms. The evolutionary temporal pattern appears to exhibit cyclic and steep nonlinear movements over time.Table 2 shows the main statistical properties of the underlying energy generation series. The measure of dispersion, Std.Dev.Indicates a relatively high degree of variation for hydro and renewable electricity generation than nuclear-based one implying considerably higher demand fluctuation.All underlying time series emerge to be nonparametric, as manifested by the SW and AD test statistics.On the other hand, insignificant ADF test statistic figures suggest nonstationary behavior.Finally, as apparent from the visual inspection, the presence of nonlinearity conforms with the outcome of Terasvirta's NN Test. The daily Google search trends in India on six topics, Government Subsidy (Subsidy), Udemy, Zoom, Amazon, Coursera, and Unemployment, are hosted in the CEIC database, which is collated to reflect the sentiment of the household and gauges the engagement with different activities striving on electricity consumption.The search topics are selected considering the study's timeline and the Indian context.We have carefully attempted to incorporate the degree of engagement in online learning and work from home, buying behavior, effects of apprehension, and reliance on government support of common people to represent the state of household in influencing electricity intake through the chosen search indicators.The utilization of the six indicators is meaningful in the COVID-19 pandemic scenario.To incorporate the demand side effects of industrial production, daily closing prices of sectoral indices, namely, Power, Energy, Automobile (Auto), Fast Moving Consumer Goods (FMCG), Capital Goods (Cap_Goods), Telecommunication (Telecom), Healthcare, and Consumer Durable (Con_Durable) have been chosen as explanatory features.The data on sectoral market prices are collected from the official portal of Investing.com[66].Deploying sectoral indices account for the overall industrial growth and appetite for energy intake in India. Results & analyses In this section, we elucidate the detailed outcomes of predictions and interpretation. Predictive modeling As discussed, the first stage of the predictive modeling commences with the decomposing of the original daily electricity generation series for identifying the high and low-frequency counterparts subsequently through the clustering framework.Figs.5-7 represent the outcome of the decomposition process and the resultant IMFs of respective series through the EEMD methodology. After completion of the decomposition of the respective series, we estimate the values of HEXP and FENT of the underlying subseries.The estimated figures are treated as features for DBSCAN clustering.Table 3 reports the results of the granular time series After the successful segregation of decomposed components into high and low-frequency buckets, we aggregate to form composite high and low-frequency categories of select clean electricity generation series as visually shown in Figs.1-3.The ASO-based predictive setup is applied for drawing predictions.The whole samples of the respective series have been split into two different data partitions, 70%-30 %, and 85%-15 %, for training and testing the model, respectively.The partition has been forward-looking, which has been reported as an ideal setup for analyzing the predictability of the financial time series [55].Consequently, the test phase considerably covers the COVID-19 horizon, enabling the volatile phase to evaluate the predictability.Tables 4 and 5 report the outcome of predictive exercises in terms of performance metrics.The figures of the NSE and IA for individual electricity generation series have emerged to be on the higher side, greater than 0.97 on training and 0.96 on the test segment.In addition, the TI values have turned out to be reasonably low in both segments.Therefore, inference can be drawn that the proposed ASO-based granular predictive framework has successfully decoded the inherent pattern of hydro, nuclear, and renewable electricity generation and estimated accurate predictions.The computed DA is close to 1 in both training and test segments, which exemplifies the predictive model's capacity to forecast the trend movements precisely.The effectiveness in estimating accurate directional changes is of paramount practical relevance as quick and precise anticipation of futuristic trends is useful for drawing roadmaps to control clean electricity generation.Among the three electricity generation series, hydro electricity generation is more predictable, followed by nuclear and renewable.We next assess the predictability in an 85%-15 % setup. Similar results to that of 70%-30 % data split have prevailed for this setup as well.High NSE, IA, and DA values can be observed in both training and test segments, while considerably low TI values are linked with train and test sections.Marginal improvement in overall accuracy is apparent, too, as manifested by the respective performance metrics.An increase in more training samples largely accounts for the phenomenon.The propounded ASO-driven granular forecasting methodology can precisely predict both absolute figures and trends.Hydro electricity generation in the said setup has appeared to be relatively more predictable, too, followed by nuclear and renewable counterparts.The quality of prediction on both configurations truly rationalizes the utility and capability of the predictive structure in modeling clean electricity generation in India.As the samples of the study suitably cover the COVID-19 pandemic regime and the ongoing Russia-Ukraine military conflict, the ability to derive predictions of supreme accuracy in turbulent times is truly established.Although the result of the predictive exercises confirms the utility of the methodological framework in precisely estimating the clean electricity trend, it is essential to conduct a comparative performance evaluation against several benchmark models to explain the usage holistically. Comparative performance evaluation We consider the individual ensemble machine learning models, RF, Bagging, and GB, as competing models to gauge the advantage of the ASO in combining the forecasts from individual models.The respective electricity generation series are decomposed using the EEMD procedure to facilitate the forecasting process of the competing models.The RF, Bagging, and GB are applied to draw forecasts on the decomposed series, which are aggregated to produce the final output.The DBSCAN-based clustering framework is not used to identify the high and low-frequency counterparts.The said design effectively captures the contribution of the clustering component on the overall performance of the granular predictive methodology.Lastly, support vector regression (SVR) has been used as a standalone machine learning-based competing model.As stated earlier, the DM statistical test is invoked to compare the efficacy of the proposed predictive structure over the competing models.The test involves paired comparisons, so the order of the pair members is important for the final interpretation.The competing models are numbered in parentheses to show the order.If the test statistic is positive and significant, the model with number 2 in parentheses is considered to have statistically better forecasts than the model with number 1. Conversely, if the test statistic is negative and significant, the opposite is true, i.e., the model with number 1 in parentheses is statistically better than the model with number 2. Finally, if the test statistic is not significant, it is assumed that there is no significant difference between the models in prediction accuracy.Tables 6 and 7 outline the outcome. The outcome of the pairwise DM tests provides an outright indication of the superiority of the proposed predictive model over all competing models for hydro, nuclear, and renewable electricity generation prediction processes.Therefore, the utility of the ASOdriven optimization approach on top of clustered granular series for forecasting future figures can be concluded to augment the quality of predictions significantly.It should be noted that merely decomposing the series by EEMD procedure to facilitate the prediction task may be unable to explain abrupt random variations.However, the three competing models, EEMD-RF, EEMD-Bagging, and EEMD-GB, have emerged to be superior to the standalone SVR model, which clearly emphasizes the advantage of EEMD methodology for modeling complex time series at a granular level. A similar outcome is apparent in the 85%-15 % setup, exemplifying the statistical superiority of the proposed ASO-driven EoE framework over the competing ones.Incorporating EEMD significantly improved the accuracy, as manifested by the relatively superior form of EEMD-RF, EEMD-Bagging, and EEMD-GB over standalone SVR. Thus, the overall comparative evaluation by DM test across both setups espouses the efficacy of the proposed granular predictive architecture.The deployment of a clustering-based decomposition process and subsequent systematic combination of the individual ensemble learning models through metaheuristic search algorithms enable the methodology to outshine the competing frameworks.The robustness of the developed forecasting architecture to predict clean electricity generation in India during the turmoil regime is proven.We next proceed to modeling by XAI methodologies for interpreting the role of the explanatory features. Model interpretation through XAI We aim to derive feature interpretation globally by permutation feature evaluation, ALE plots, and locally by LIME plots.The findings are discussed sequentially. Outcome of permutation feature evaluation Figs. 8-10 display global feature ranking for explaining the variability of the electricity generation from three different sources through the lens of permutation feature evaluation. It can be noticed that both Google search trends and sectoral market indicators have emerged to feature in the top four important features to track hydro electricity generation.Thus, dependence on household and industrial demands emerges to be equal.Predictive prowess of Google search trends on Subsidy, Unemployment transpires to be comparatively lower.Interestingly, the impact of search trends linked to strict academic programs in the form of online education manifested by Coursera and Udemy appears to be reciprocal in nature.Zoom, on the other hand, can be used in academic and official engagements simultaneously. The dominance of sectoral outlook over Google search trends in exerting predictive influence on nuclear electricity generation is apparent.The four most significant characteristics are closing prices of the Power, Telecom, and Energy sectors.Likewise, the subdued impact of Google search trends, Udemy, Unemployment, and Subsidy, can be observed in the previous scenario.Hence nuclear Table 7 Result of DM test for 70%-30 % setup.electricity is primarily consumed in industrial production mainly. Clear supremacy of the influence of industrial demand on renewable electricity over household demand is evident as the former category of explanatory features occupies the top four important feature list.Similar to nuclear electricity demand, the reliance of the telecom and power sector on renewable counterparts is evident as well.The impact of Google search trends on Udemy intensifies in explaining the variation of the underlying series, while Subsidy remains relatively less important.Overall, the permutation feature evaluation reveals critical findings in identifying the sectors that significantly absorb clean electricity and gauging the impact of Outcome of ALE-based feature evaluation We now estimate the ALEs of individual features to explain their contribution to the respective electricity generation series at a more profound scale.The ALE exhibits of determinants of nuclear electricity generation unveil interesting insights too.The magnitude of influence of the Google search indicators can be seen to be relatively lower than that of the sectoral ones.Thus, the domination of the industrial demand over the household over nuclear electricity generation, as observed in permutation feature evaluation, is validated.The impact of several features remains stagnant over different intervals.On the other hand, decreasing predictive power of Telecom, Subsidy, and FMCG with an increase of their respective values can be noticed. The ALE plots of renewable electricity generation suggest a strong positive influence on Power and Udemy as their values cross a threshold.The influence of Unemployment diminishes as its search pattern intensifies.A spike in the contribution of Coursera bounded to a specific interval is apparent.No substantial difference in the predictive contribution of the remaining search indicators could be found, as ALE plots do not document sharp increases or decreases.As deemed critical in the permutation feature evaluation, Power, Cap_Goods, and Telecom are linked to increasing positive influence, whereas Con_Durable exerts negative predictive power. In this nutshell, the utilization of the permutation feature evaluation and ALE plots simultaneously caters to the proper interpretation of functionalities of the underlying explanatory variables globally.The key contributory features have been discovered in conjunction with explaining the dependence structure. Outcome of LIME-based feature evaluation We, now, proceed to delve into the prediction process at the local level applying the LIME plots on four randomly chosen data samples for select clean electricity generation series.Figs.20-22 depict the outcome.The vertical axis annotates the explanatory features, while the horizontal axis represents their respective contributions. It is amply apparent that local level feature ranking for predicting hydro electricity generation is not uniform across the samples The outcome of the local feature evaluation for renewable electricity generation has demonstrated similarity to its counterparts as a significant difference in the contribution pattern to that of the global scale is imminent.Interestingly, Udemy has featured in the top four feature list in two samples, wherein no search indicators were deemed to be highly important in explaining the overall variability of the renewable electricity generation pattern.The predictive prowess of the Con_Durable has experienced a drastic reduction for the selected data instances. Therefore, the overall assessment through the ALE plots underscores the utility of all underlying features in precisely tracking the clean electricity generation pattern.It is equally important to emphasize explaining the short-run fluctuation owing to practical implications.As the current work is the first of its kind, the inclusion of Google search trend indicators for making is unique.Although on a global scale, domination of the sectoral indices reflecting industrial growth and demand has relatively outperformed the search indices in terms of predictive prowess, the utility of the latter is proven in the local-level prediction processes.Thus, the findings of XAI methodologies also rationalize the selection of explanatory features. Discussion The findings of the present research echo that the chosen daily electricity generation from the chosen renewable sources is highly predictable, implying steady growth and reliance on the same for household and industrial activities in the Indian context.Initially, the daily electricity generation series from the chosen sources has been found to follow a long-memory dependence structure.Hence, the demand for the same can be inferred to exhibit high volatile phases followed by high volatile regimes and low volatile phases followed by low volatile regimes.Thus, the eventual consumption patterns are unlikely to experience sporadic movements, which can be used for regulatory frameworks.The hydro electricity generation has been observed to be relatively more predictable, which is followed by nuclear and renewable electricity generation.The strong predictive influence of selected socio-economic factors manifested by the financial outlook of Indian sectoral indices and Google search trend indicators suggest tracking daily clean electricity demand can be facilitated by gauging industrial production and user engagement in various activities.The prosperity of Edtech companies, viz.Coursera, Udemy, etc., have garnered high traffic, which indirectly catalyzed the transition toward clean electricity owing to increased reliance on computational power.On the other hand, different industrial sectors do not spur daily clean electricity generation The other aspect of the contribution, i.e., the propounded forecasting structure, is of immense practical implications as the same survives a series of numerical and statistical tests.The robustness of the framework is apparent as the dynamics can be explained precisely during the COVID-19 pandemic and Russia-Ukraine conflict regimes.The efficiency of the framework in minutely estimating the directional changes as manifested by high DA figures on both training and test segments has been established.The said characteristics of the framework could be suitably used to anticipate peak or fall in demand in short-run horizons, enabling effective resource planning at power plants.Accurate estimation of absolute figures and directional changes is of paramount relevance for risk management in uncertain periods.The success of the ASO metaheuristic algorithm in augmenting the predictive performance of the proposed granular forecasting structure is evident, which significantly contributes to the superiority of the proposed model over benchmark competing methodologies.The property handling nonlinear and nonstationary time series augments the potential of the ASO-based EEMD-DBSCAN granular approach.Overall, the effectiveness of the methodological framework can easily be extended to carry out predictive modeling of financial markets, wherein the stakes are even higher. The scope of underlying research is restricted to three clean electricity resources in the Indian context with chosen macro indicators as explanatory variables.As stated, the previous research has primarily been carried out in micro setups to gauge power generation or consumption patterns.Hence, the research findings are useful to mitigate the demand and supply gap at the country level and strategize increased adoption of clean electricity.Nevertheless, the spectrum of the research can easily be extended into cross-country comparisons of clean electricity production processes to gauge whether the degree of predictability differs across developed and developing economies.It is also possible to introspect the behavioral pattern in different regions of a country in a state-wise manner for deeper inspection.A few micro-process-specific variables can be incorporated into the research framework to explain the leftover variability of chosen time series.Comparative modeling of daily nonrenewable and clean electricity generation in developed and developing economies can be carried out to deeply comprehend the consumption dynamics and influence of cognate macroeconomic and other factors.The capacity of the predictive framework can be utilized for analyzing complex financial time series during Black Swan events.On the methodology front, a comparison of the capability of state-of-the-art deep learning algorithms with ensemble machine learning techniques in deriving predictions through the ASO-based granular forecasting framework can also be examined in the future. Conclusions The underlying research endeavors to critically analyze the pattern of clean electricity generation in the Indian context through the applied predictive analysis lens.We advance a novel EEMD-DBSCAN-based decomposition methodology and integrate the same with the ASO-based EoE forecasting structure to delve into the predictability of hydro, nuclear, and clean electricity generation patterns for accomplishing the objectives.Additionally, the current work strives to uncover the deeper insights pertinent to the predictive influence of the industrial and household demand governing clean electricity growth. The overall findings of the present research suggest that the daily clean electricity generation in India from hydro, nuclear, and renewable sources can indeed be predicted with a high degree of accuracy.The utility of utilizing socio-economic factors by incorporating Google search indicators and sectoral outlook as explanatory variables for tracking clean electricity generation is established.The study is highly relevant for maintaining sustainability and reducing carbon emissions.The values of the DA indicator for the prediction of the three energy generation series have emerged to be higher than 0.9, which signifies the effectiveness of the research framework in accurately determining the immediate directional changes in electricity generation.The quality of forecasts on both setups, 70%-30 % and 85%-15 % covering chaotic and volatile regimes testify to the validity of the model.The designed granular structure has transpired to yield statistically superior forecasts over 4 competing models, which rationalizes its relative efficiency.We critically delve into the predictive interplay of hydro, nuclear, and renewable electricity generation with important socio-economic factors by incorporating Google search indicators and sectoral outlook.The global and local level feature interpretation reflects the nature and importance of the explanatory variables.Among the sectors, Cap_Goods, Con_Durable, Power, and Energy have emerged to drive clean electricity growth in India prudently.It, nonetheless, is necessary to monitor the growth and prosperity of other select sectors as well as to explain the abrupt variation of clean electricity demand, which can streamline the generation process in the long run.The relatively better predictability of hydro electricity generation over the other two counterparts implies a stable intake of the same in running a business and day-to-day activities.The accuracy of the predictions, specifically during the COVID-19 regime's emphasis on clean electricity generation in India, remained less perturbed despite heavy disruptions and lockdowns.From a policymaking perspective, providing incentives or rolling out appropriate schemes to migrate to hydro, nuclear, and renewable power for sectors that are linked to comparatively low dependence on the same can be chalked out.In the Indian context, fear of unemployment and the tendency to avail subsidy offerings have emerged not to be critical in shaping the clean electricity generation process, indicating a presence of reasonably effective governance.Industrial production predominantly consumes nuclear and renewable electricity.Nevertheless, the short-run electricity appetite of household operations at the individual level is apparent too. The present work significantly contributes to the methodological front for conducting predictive analysis of clean electricity generation.Our work propounds a robust granular predictive framework that not only produces highly accurate forecasts but scrupulously survives validation and comparative performance assessments.Utilizing the DBSCAN-based clustering framework for disentangling the original series into high and low-frequency subseries immensely facilitates the training process.Subsequently, the seamless integration of three ensemble machine learning models applying the ASO-based optimization setup significantly improves the predictive accuracy.The framework is, therefore, classified as a significant addition to the granular methodological spectrum.The profound dependence of clean electricity generation on both sets of factors has been established, which explains the apparently I. Ghosh et al. nonlinear and nonstationary pattern.The practical implications of the predictive framework in facilitating energy trading further underscore the contribution. ith sample by RF, Bagging, and GB methods; Y Fin i represents the final predicted figure by the ASOE model; and w 1 , w 2 , w 3 are respective weighted contributions of the respective ensemble models.The weighted average framework follows the following constraint (eq.( I .Ghosh et al. I .Ghosh et al. evaluation.The impact of Zoom and Coursera increases as a steep jump in contributions can be observed after they cross threshold figures.It basically indicates longer engagement in online education sharpens the electricity appetite of the household.Amongst the industrial sector, a bullish phase was observed in Energy, FMCG, and Healthcare linked to higher hydro electricity generation.Telecom and Con_Durable, nonetheless, exhibit opposite behavior. I .Ghosh et al. Table 2 Key statistical characteristics. I.Ghosh et al. Table 3 Outcome of granular time series clustering. I.Ghosh et al.	2023-12-12T16:02:33.859Z	2023-12-01T00:00:00.000	{ "year": 2023, "sha1": "159206f8cba79df91c9cf64e925e26ada5711b60", "oa_license": "CCBY", "oa_url": "https://doi.org/10.1016/j.heliyon.2023.e23434", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "6a2f896e12028889dc1e04b68b3f49a4f2000a88", "s2fieldsofstudy": [ "Engineering", "Environmental Science", "Computer Science" ], "extfieldsofstudy": [] }
79164647	pes2o/s2orc	v3-fos-license	OZONE NUCLEOLYSIS IN LUMBAR INTERVERTEBRAL DISC HERNIATION: NON-RANDOMIZED PROSPECTIVE ANALYSIS STUDY DESIGN: Non-randomized, prospective analysis of 68 patients of lumbar disc herniation treated with ozone nucleolysis. OBJECTIVE: To assess the patients with lumbar disc herniation treated with intradiscal ozone, pre and post ozone nucleolysis, for pain using Visual Analog Scale (VAS) functional & disability score using Japanese Orthopedic Association (JOA) Clinical Symptom Score. SUMMARY OF BACKGROUND DATA: Ozone therapy for disc herniation is becoming popular because of its minimal invasive, lesser recurrences and remarkably fewer side effects. Successful outcomes of ozone therapy have been reported from various European & Indian centers. METHODS: A series of 68 patients were treated with ozone therapy for lumbar disc herniation from January 2009 to January 2012. The procedure is done under C-arm guidance under local anesthesia by “Single sitting double injection technique”. All patients were assessed using VAS for radiation pain & back pain, Clinical Symptom Score of the Japanese Orthopaedic Association (JOA) for a Patient with Lumbar Disc Herniation, pre op and post op, on day one, after a week, two weeks, first month, third months, sixth month one year second year. Were classified them as Good, Moderate & Poor outcome. RESULTS: Out of 68 patients 89.7% (61/68) patients had good outcome, 7.35% (5/68) patients had moderate outcome, 2.95% (2/68) had poor outcome. Intra-op in 1 patient where ozone spread in Para spinal muscles but had no postoperative problem.4 patients had mild nausea, 2 had mild headache & No infection. CONCLUSIONS: Ozone nucleolysis is a new, minimally invasive procedure done under local anesthesia & has shown effective results in the treatment of contained intervertebral disc herniation with no side effects. Newer trend of minimal invasive technique, which is cost effective with minimal hospital stay and least complications, for discogenic low back pain has been in research for some time now, giving rise to modern techniques came into being that are minimally invasive and share the principle of acting directly on disc content without accessing spinal channel. Image guided percutanous intradiscal injection techniques are chemodiscolysis by chimopapain or by Ozone nucleolysis. Chemodiscolysis by chimopapain is based on an enzymatic dissolution of the nucleus. Although purified prepared enzymes are used, anaphylactic shock can occur. 4 Ozone is a tri-atomic Oxygen molecule, O3, with a different molecular structure from Oxygen. Its name is derived from the Greek word "ozein" meaning "to smell". Ozone, an emerging medical drug, is used in disc hernaition because of its property to break down the proteo-glycan bridges in the nucleus pulposus of disc along with an anti-inflammatory action. 5 In this study we aim to assess the effectiveness of ozone nucleolysis in lumbar intervertebral disc herniation. This study is done over a period of three years, wherein 68 individuals with complaints of low backache and radiation pain due to disc herniation were taken as subjects and were treated with ozone injection in the herniated disc followed up for 2 years. Ozone & How Ozone Acts: The action of ozone is due to the active oxygen atom or the free radical liberated from breaking down of ozone molecule. As ozone is injected into the disc the active oxygen atom attaches with the proteo-glycan bridges in the jelly-like material of nucleus pulposus. They are broken down and they no longer capable of holding water. As a result disc shrinks and mummified and there is decompression of nerve roots. It is almost equivalent to surgical discectomy and so the procedure is called ozone discectomy or ozonucleolysis. It has an anti-inflammatory action due to inhibition of formation of inflammation producing substances, tissue oxygenation is increased due to increased 2, 3 diphosphoglycerate level in the red blood cells. All these leads to decompression of nerve roots, decreased inflammation of nerve roots, and increased oxygenation to the diseased tissue for repair work. Ozone has a dose-related biological action. At high concentrations (30-70µg/ml 02), it may cause alterations of tissue structure; at medium concentrations (20-30µg/ml 02) it seems to affect the regulation of the immune system and at low concentrations (<20µg/ml 02) it improves the microcirculation. 5, 6,7 The dose of ozone is crucial and must not exceed the capacity of antioxidant enzyme and glutathione to prevent accumulation of the superoxide anion and hydrogen peroxide, which can cause cell membrane degradation. 8,9 The specific feature of oxygen ozone therapy noted in disc specimens was dehydration of the fibrillary matrix of the nucleus pulposus, revealing collagen fibers and signs of regression (Vacuole formation and fragmentation)-a sort of disk "mummification." 10 MATERIAL & METHODS: A series of 68 patients were treated with ozone therapy for lumbar disc herniation from January 2009 to January 2012. All patients age group 18 yrs to 60 yrs who presented with clinical signs of disc compressing the nerve root & whose MRI showed evidence of contained disc herniation [ Figure 1], with duration of symptoms more than six weeks and had failed to respond conservative treatment with Non-steroidal anti-inflammatory medical therapy and physical therapy were included. Exclusion Criteria: Previous lumbar spine surgery, possible pregnancy. Current fracture, infection, and/or deformity (greater than 15 degrees of lumbar scoliosis, using Cobb's measure technique) of the spine. Calcified or migrated herniated disc, signifiant glucose-6-phosphate-dehydrogenase deficit, active hyperthyroidism. cauda equina syndrome or progressive neurological deficit (Usually requiring urgent surgery), loss of control of urination & defecation. All patients were orally informed about the procedure and potential risks of treatment. Written informed consent was obtained from all patients. The procedure followed was approved by the committee on Human Experimentation of our institution. The procedure is done under C-arm guidance under local anesthesia. All patients were assessed using VAS for radiation pain, Clinical Symptom Score of the Japanese Orthopaedic Association (JOA) for a Patient with Lumbar Disc Herniation 11 , pre op and post op, day one, after a week, two weeks, first month, third months, sixth month one year &second year. Were classified as good, moderate & poor outcome according to [ Table 1] which included VAS for radiation pain & JOA score. PROCEDURE: The patient is taken to the operation theatre lying on oblique position with a pillow under lower abdomen. The side of pain turned upwards. The area is prepared and draped in sterile manner. It is done usually under local anaesthesia Intravenous antibiotic ceftriaxone 1G should be given 30minutes before the procedure. An intravenous line & Pulse oxymeter are secured before the procedure. The procedure is done under C-arm guidance. C-arm first should be focused to a pure anterior-posterior view to view the diseased disc. Then C-arm is cranially tilted to abolish any double end-plates and thereby getting widest possible view of disc space. Then C-arm is focused in a away such that facet joint come at the center of the end plates. On the side of maximum pain, the needle entry point is marked just lateral to the superior pars/articular pillar exactly at the center of the disc [ Figure 2]. Local 2% xylocaine is infiltrated, then 22 G Chiba, 178mm long spinal needle is introduced into the diseased disc under fluoroscopic guidance [ Figure 3]. The position of needle tip is confirmed in AP & lateral view. The ozone oxygen mixture gas is collected from the ozone generator at 6 Lt's of oxygen, concentration of 40µgm ozone in a 20 cc syringe. Initial 10 to 15 cc is injected with needle placed just ahead of midline. As Ozone gas is injected in the disc, the disc appears radiolucent and illuminated [ Figure 4 a, b].Usually after injecting 10 to 15 cc there is resistance the syringe is held in place under pressure for 5 min. Then the trochar is placed back and waited for 10 minutes. Again fresh ozone is collected from ozone generator and 5 to 10 cc of ozone is injected in the disc with needle 2-3mm posterior of the midline. So this act as 2 injection in one sitting, as the half-life of ozone is around 20 minutes" single sitting double injection". Then needle is removed and aseptic spray is applied. Patient is advised bed rest for 2 hrs. later patient can mobilize according to his comfort. RESULTS: Of 68 patients 41 were male & 27 female. Mean age was 34 years.L4-5 disc protrusion in 38 pts, L5-S1disc protrusion in 23 Pts &doth L4-L5and L5-S1dics protrusion in 7 patients. Data analysed by using SPSS (Statistical Package for social sciences) version 17.0 We have used paired ttest to find out the significance between the pre-operative and post-operative pain score and JOA score at 5% level of significance (95% C.I.) if the p-value < 0.05 then there is significant difference between pre & post-operative score. Patients with VAS score for radiation pain [ DISCUSSION: Disc Herniation is defined as a localized displacement of disc material beyond the limits of the intervertebral disc space. Partial tear of annular ligament with herniation of the nucleolus pulposus is termed as contained disc herniation. Diagnosing contained disc herniation on MRI becomes crucial as ozonucleolysis as its treatment has better prognosis as compared to disc protrusion. Clinical signs and symptoms of disc herniation are not only caused by mechanical compression but also by biochemical factors that play an important role in inflammatory sensitization and immune response in the epidural environment of the nerve root. 12,13,14 Pain and inflammation developed from the pressure of the herniated material on the posterior longitudinal ligament and the dura mater, may ultimately affect the nerve roots. The patients who fail to respond to medical treatment including analgesics and physiotherapy are one who require decompression of nerve roots either by surgery or some percutaneous intradiscal or foraminal procedure. Use of medical ozone for treatment of low back pain was advocated by orthopaedic surgeon Verga in the 1980s, treated about 8000 of disc herniation patients over 15 years, in whom relapse of pain had occurred in less than 2% of cases. Muto suggested intradiscal injection of ozone for disc hernia in 1998 under CT guidance. 15 Leonardi popularized fluoroscopy guided ozone injection into the intervertebral disc. Nowadays, ozone therapy is understood to be a genuine treatment method in complementary medicine, encouraging the scientific dialog between traditional medicine and complementary methods. Critical discussions have activated basic research in the field of ozone therapy including the revision of highly complicated treatment methods. It has been published in Anesthesia and pain journals that up to 85% of disc surgeries can be avoided with these non-surgical interventions. Success rate is about 88% which is comparable to surgical discectomy (50% to 90%). Complications are remarkably low and much less than surgery. In Our Study: We had good result in 87.9% of the patients selected for the study, who had complaints of low back ache lasting longer than 6 weeks, refractory to all conservative treatment with NSAIDs and physiotherapy. VAS for radiation pain & JOA pre operatively as well as on various specified intervals post operatively had shown significant difference. Ozonucleolysis caused a more significant improvement of radiation pain in those patients who had complaints of low backache along with radicular pain. Patients with moderate (7.35%) outcome used occasional analgesic and physical therapy. Two patients (2.95%) had poor outcome who underwent surgical discectomy. Bonetti et Al also reported excellent results in 74.4% patients after six months. 16 A success rate of 70.3% was noted by Cosma F. Andreula et Al. 10 Successful outcomes of ozone therapy have been reported from various European centers. Other studies as Treatment of Intraforaminal and intradiscal injections of a steroid, an anesthetic, and O2-O3 are more effective at 6 months than injections of only a steroid and an anesthetic in the same sites. 17 Technique: 'Single sitting double injection technique' in which 15 -25 cc of freshly prepared oxygenozone mixture was given twice intra discally with a time gap of 10 minutes. This technique was used to increase the tissue destruction property of ozone and considering its half-life that is about 20 minutes. We used a 178 mm long 22 guage Cheba needle. The advantage of using a thin bore needle is that it reduces tissue injury and patient discomfort. All patients were discharged & made to walk the same day or the next day as per patients comfort. Most of the studies used radio opaque dyes for confirmation of needle in the disc & its containment. The use of intradiscal contrast reduces the discal absorption of ozone and the space available for the action of ozone. Hence, no intradiscal dye was used in our study. Under C-arm guidance. At the time of injection of ozone, the patient complaints of increase in pain symptom initially and reduction eventually. This is because the intradiscal pressure rises with the injection of ozone and reduces with tissue destruction that it causes. We used ozone in concentration of 40 microgram per millilitre of oxygen to increase its tissue destruction property. Paoloni et Al in their study used a concentration of 20 mc/ml 18 .Bonetti et Al used a concentration of 25mc/ml. [12] Viebahn reported that the nontoxic concentration of ozone varies from one to 40 microgram per millilitre of oxygen and concentration should not exceed 40 mc/. 19 Complications of ozone therapy are very rare. They include post-procedural transient muscle spasm & burning pain. No patients had infection or discitis (Very rare due to the bactericidal effect of ozone). Intra-op, in one of our patients where ozone had spread in Para spinal muscles. Patient had mild burning pain for 7to 8 minutes but had no postoperative problem. Review with literature showed intramuscular ozone therapy in treatment of acute back pain has shown to reduce disability & the intake of analgesic drugs in a study with no complications. 19 A case being reported of the complication: Acute fatal septicemia should be considered among the major complications of the oxygen-ozone therapy in the treatment of a herniated lumbar disc. The patient developed a fulminant septicemia. An abdominal-pelvic and chest computed tomography and blood culture led to the diagnosis of pyogenic lumbar muscle involvement, accompanied with septic pulmonary embolism secondary to Escherichia coli infection. 20 In our experience the failure has been mostly related to patients with some amount of spinal canal stenosis, recurrent herniated disc and small descending herniated disk of the lateral recess with significant compression of the nerve sheath. On the basis of our results and the assessment of our failures, we recommend careful selection of patients to avoid broadening the indications for treatment, thereby ensuring a high success rate. Ozone nucleolysis being a minimally invasive technique allows for early mobilization of the patient as well as early recovery, reduced hospital stay and working days lost of the population. Ozone nucleolysis has a success rate of about 87.9% and remarkably few side effects. CONCLUSION: Ozone nucleolysis is a new, minimally invasive procedure done under local anesthesia & has shown effective results in the treatment of contained lumbar intervertebral disc herniation with remarkably few side effects. Ozone therapy is a good option to treat lumbar disc herniation that has failed to respond to conservative management, before recourse to medications or surgery. Post-operative at 6 months Good Moderate Poor Vas	2019-03-16T13:13:52.733Z	2015-05-06T00:00:00.000	{ "year": 2015, "sha1": "dd1a387018398e2a4eccb6ff8853d4597ed6cfcd", "oa_license": null, "oa_url": "https://doi.org/10.14260/jemds/2015/938", "oa_status": "GOLD", "pdf_src": "MergedPDFExtraction", "pdf_hash": "d62d35ed936472a15eecc4f9dfc75f5f82c1ed41", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
13931879	pes2o/s2orc	v3-fos-license	Prevalence and molecular characterization of Clostridium difficile isolated from European Barn Swallows (Hirundo rustica) during migration Background Clostridium difficile is an important bacterial pathogen of humans and a variety of animal species. Birds, especially migratory passerine species, can play a role in the spread of many pathogens, including Clostridium difficile. Barn Swallows (Hirundo rustica) nest in close proximity to human habitats and their biology is closely associated with cattle farming. Therefore, we hypothesized that Barn Swallows can be the reservoir of Clostridium difficile. Results Barn Swallows (n = 175) were captured on their autumn migration across Europe to sub-Saharan Africa. Droppings were collected from juvenile (n = 152) and adult birds (n = 23). Overall prevalence of Clostridium difficile was 4% (7/175); 4.6% (7/152) in juvenile birds and 0/23 in adults. Clostridium difficile ribotypes 078, 002 and 014 were identified, which are commonly found in farm animals and humans. Three new Clostridium difficile ribotypes were also identified: SB3, SB159 and SB166, one of which was toxigenic, harbouring genes for toxins A and B. Conclusions Results of this study indicate that Barn Swallows might play a role in national and international dissemination of Clostridium difficile and could serve as a source for human and animal infection. Clostridium difficile ribotype 078 was identified, which has been reported as an emerging cause of community-associated Clostridium difficile infection in humans. Based on this and other studies, however, it is more likely that Barn Swallows have a more indicative than perpetuating role in Clostridium difficile epidemiology. Background Clostridium difficile (CD) is present in almost all human environments and is a potential zoonotic pathogen that can be isolated from a variety of animal species [1][2][3][4][5]. It is the most commonly diagnosed cause of antimicrobialand hospital-associated diarrhoea [6] and is an emerging cause of community-associated disease [7,8]. The presence of CD in animals, including those that come in close contact with humans as pets or through the food chain [1,9], and the significant overlap between the types of CD isolated from humans and animals [10][11][12], has lead to suggestion that CD might be a zoonotic and foodborne pathogen [9,12]. Recognition of different potential sources of CD transmission remains a pressing clinical and investigative quest. Migratory birds travel between continents twice each year and are responsible for the transmission of several pathogens [13] and/or disease vectors [14]. The world's largest bird migration system is the Palaearctic-African flyway, which involves an estimated 2 billion passerines and nearpasserines migrating between the European continent and sub-Saharan Africa [15]. An epidemiological study on mostly migrating passerine birds in Europe suggested that they are unlikely to serve as a carrier or reservoir of CD [16]. However, the study [16] was conducted in the region where Barn Swallows (Hirundo rustica) were not congregating during their migration south, and included only garden birds unlikely to dwell in habitats intensively cultivated by humans. On the other hand, Barn Swallows, come in close contact with human habitats [17,18], as do House Sparrows (Passer domesticus), which are a non-migrating passerine bird, that have been associated with CD when sampled on pig farms [19]. Barn Swallows are among most common and widespread migratory birds in the Palaearctic. In Europe they most commonly nest in cattle farms inside the barns or stables, which had a significant evolutionary effect on Barn Swallow biology with regards to their nesting habits. Any changes in their environment significantly affect their population [17,18]. Its habit of nesting in buildings associated with human habitation has made this species one of the most familiar bird species in the world [20] as well as a potential source for international and local CD perpetuation. The aims of the present study were to determine the prevalence of CD in Barn Swallows, and to characterize CD isolates. Methods Sampling of Barn Swallows was conducted once in September 2011 in the central part of Slovenia (46°0′ 2.12″ N, 14°22′ 58.27″ E) in an area identified as a Barn Swallow congregation point during their autumn migration across Europe (average week temperature 21.8°C, average week humidity 69.5%). One hundred seventy-five (n = 175) Barn Swallows were captured with mist nets and placed individually in clean custom made hot pressed plastic containers with air holes. Bird faeces, if present, were collected from containers using sterile gloves (Ansell Ltd, UK) after the birds were removed for standard general inspection, ringing and release to continue with their migration. Faeces were transferred into 2 mL sterile tubes (Eppendorf Tubes, Germany) and stored at -20°C until analysed for the presence of CD spores. The study was carried out with animal ethics approval by the Ministry of the Environment and Spatial Planning (document No.: 35601-10/2010-6). Clostridium difficile culture Entire samples were inoculated into 9 mL of CD moxalactam norfloxacin (CDMN) enrichment broth (Oxoid Ltd; Nepean, ON Canada) containing 0.1% sodium taurocholate. Anaerobic incubation at 37°C lasted for 7 days. An aliquot of broth was transferred to a new vial and an equal amount of anhydrous alcohol was added to each sample. This was followed by a 60 min incubation at room temperature, centrifugation (3800 g for 10 min) and inoculation of the pellets onto CDMN (Oxoid Ltd; Nepean, ON Canada) agar. Further incubation in an anaerobic chamber for 2 days at 37°C was followed by another 3 days if necessary. Isolation and identification of CD was based on the characteristic morphology and odour of the colonies, Gram stain and the presence of the L-proline aminopeptidase activity (Remel Inc, Lenexa, KS, USA). One single colony for each isolate was subcultured and stored at -80°C and re-cultured prior to molecular analysis. Molecular analysis Extraction of DNA was done on pure cultures after they were obtained from blood agar with a Chelex resin-based DNA extraction commercial kit (InstaGene Matrix, Bio-Rad Laboratories, USA), following the manufacturer's instructions. Extracted DNA was used as template for further molecular analysis. Ribotyping was performed as previously described by [21]. Ribotpyes were assigned visually based on comparison with an internal library of ribotypes as well as reference strains from the Cardiff ECDC reference library. Testing for genes encoding toxins A (tcdA) and B (tcdB) was performed by PCR as previously described by [22]. The presence of the gene encoding the binding component of the binary toxin (cdtA) was detected as described in [23]. Results A total of 175 samples were taken from 152 juvenile and 23 adult Barn Swallows. Seven (7/175; 4%) samples were positive for CD; 7/152 (4.61%) juveniles and 0/23 adults. Five (5/7; 71%) isolates were toxigenic. All five toxigenic isolates possessed tcdA and tcdB while two also possessed cdtA. Clostridium difficile isolates that were toxin A, B and CDT (Clostridium difficile toxin/binary toxin) positive, were identified as ribotype 078. Single isolates of ribotypes 002 and 014 were also identified. The remaining three ribotypes, one of which was toxigenic, had not been previously identified in this laboratory or documented in the available Cardiff ECDC collection (Table 1). Discussion This is the first study investigating Barn Swallow as a possible source of CD for farm animals and humans. In the past, there has been increasing concern about animals as potential CD reservoirs and sources of human exposure [12,24]. Of particular note is the finding of ribotype 078. This strain is common in food animals [4,25,26] and has been reported as an emerging and increasing cause of community-associated Clostridium difficile infection (CDI) in humans [25,27,28]. Two of the three other toxigenic ribotypes identified in this study (002, 014) can be found both in humans and animals [26,29]. Three new CD ribotypes were also identified (SB3, SB159, SB166), one possessing A + B + CDT-(SB166). It was interesting that CD was only found in juvenile birds. The sample population in this study was predominantly juvenile birds, which is in concordance with the expected ratio of juvenile birds on migration (>80%) [16,30]. In most studied animal species, CD tends to predominate in younger animals [31,32]. There is no evidence that CD causes disease in Barn Swallows. Currently, the epidemiology of community-associated CDI is poorly understood. Food animals and food have been suggested as sources of human exposure [9,33]; however, other potential forms of exposure in the community must be considered. While the high prevalence of CD in some farm animal groups clearly indicates that they could be reservoirs of CD, the potential for other animals to act as a source of CD from farms to the broader human or animal population is intriguing. The biology of Barn Swallows potentially makes them a very efficient vector for CD dissemination. They are the most common and widespread migratory bird in Europe and cohabit with farm animals and humans [17,18,20]. Passerine birds were previously not determined to be the source of CD; however, only passerines unlikely to come in close contact with humans have been investigated [16]. In contrary, House Sparrows, a non-migrating passerine, had previously been associated with CD on pig farms in the Netherlands [19]. European Barn Swallows, the species studied here, preferably nest within the cattle barns during warm months of the year, and spend the winter in Sub-Saharan Africa. Some can migrate to Arabia and to the Indian sub-continent [17,18]. In our study, a total of 4% of all captured and sampled Barn Swallows were positive for CD, which reflects the prevalence of adult cattle in some reports. Given the presence of CD in the bovine population, with reported prevalence of 2.4-6.3% in adult cattle and 7.6-51% in calves [3,5,34] and the nesting locations of Barn Swallows, it is certainly plausible that these birds could acquire CD on farms. Interestingly, to date, including this study, CD was isolated only from wild birds that are associated with intensively farmed habitats [16,19]. Therefore, it is more likely that Barn Swallows have a more indicative than perpetuating role in CD epidemiology. Conclusions Results of this study showed that Barn Swallows could potentially be a source of CD in humans and animals. During their congregation at migration destinations interindividual and interspecies horizontal transmission can occur [13]. However, based on this and previous studies on the prevalence of CD in wild passerine birds [16,19], and studies associated with farm animals and humans [10,27,31,32,35], Barn Swallows may more realistically be an indicator and not the source of the contamination of CD in the environment. It seems that intensive farming and hospital environment are the sources of CD, and only humans and/or animals associated with such environment perpetuate CD. Competing interests Authors have no competing interests to declare. Authors' contributions All of the authors contributed to the conception design, interpretation of data, drafting the article and critical revision of the manuscript. PB, TT and MV collected samples from Barn Swallows. JSW and JR performed bacterial culture and molecular analysis. All authors approved the final version of this manuscript.	2017-06-26T23:18:16.603Z	2014-02-08T00:00:00.000	{ "year": 2014, "sha1": "23b8e6c8d7c92320b5e495608a3484de09d91af8", "oa_license": "CCBY", "oa_url": "https://bmcvetres.biomedcentral.com/track/pdf/10.1186/1746-6148-10-40", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "4095c519c71e1e5763be083c77aceb27323a1713", "s2fieldsofstudy": [ "Biology", "Medicine" ], "extfieldsofstudy": [ "Biology", "Medicine" ] }
220251318	pes2o/s2orc	v3-fos-license	Down syndrome: A curative prospect? Experimental work regarding corrective actions on chromosomes and genes, and control of gene products is yielding promising results. It opens the way to advances in dealing with the etiological aspects of Down syndrome and may lead to important changes in the life of individuals affected with this condition. A small number of molecules are being investigated in pharmacological research that may have positive effects on intellectual functioning. Studies of the pathological consequences of the amyloid cascade and the TAU pathology in the etiology of Alzheimer disease (AD), which is more frequent and occuring earlier in life in persons with Down syndrome (DS), are presented. The search for biological markers of AD and ways for constrasting its early manifestations are also discussed. Introduction The prevalence rate of intellectual disability in the general population is estimated to be between 1 and 3%. Chromosome abnormalities (numerical as well as structural) are responsible for up to 28% of intellectual disabilities [1], among which aneuploidies are well represented. DS is the most frequent autosomal trisomy (chromosome 21-C21) occurring naturally in approximately 1 in 700 live births. C21 actually is the smallest of human chromosomes. It should bear number 22 as chromosomes are numbered in size order and there are 22 pairs of autosomes plus one pair of sex chromosomes, XX for females and XY for males. The misplacement of C21 in chromosome order of magnitude goes back to an early time in molecular genetics where microscopes did not have the discriminative power that they acquired later, and the numerical mistake has not been corrected. Human C21, technically called Hsa21, harbors around 250 protein-coding genes (the precise number varying depending on genome annotation) and between 165 and 404 non-coding RNA (ribonucleic acid) genes regulating gene expression. Overexpression of proteins linked to gene triplication determines a constellation of abnormalities involving heart, nervous system and gastro-intestinal tract. Impaired brain development causing structural anomalies and decreased volumes of frontal and temporal cortices, hippocampus, cerebellum, and brain stem, is typically observed. Anomalies of neural connectivity are the rule. However, almost every aspect of the DS phenotype is subject to an important degree of interindividual variability due to its polygenic nature and interactions with environmental factors [2]. DS appears in several forms. The most common one (95% of cases) is standard trisomy 21-TR21-(karyotype 47 + 21). It is characterized by a triplication of Hsa21 in every cell of the body. In mosaic TR21 (1 to 2% of the cases), only a portion of the cells carries one extra Hsa21. The proportion depends on the moment when the triplication of Hsa21 occurs (first, second, or third cell division). In Robertsonian (centric fusion; nonreciprocal) translocations, the participating chromosomes (pairs 13, 14, 15, 21, or 22) break at their centromeres and the long arms (q segments) fuse to form a single large chromosome with a single centromere. The short segments (p) are usually lost. DS maps to a region on the long arm of Hsa21 covering an area of 37 to 44 megabytes corresponding to band 21q22 and containing around 225 genes [3]. Ait Yahya-Graison et al. [4] supply a figure of 200-250 genes between 21q21 and 21q22.3. A smaller region in Hsa21, labelled DSCR (Down Syndrome Critical Region), involving bands 21q21.1 and 21q22.2, which includes about 50 protein-coding genes and a larger number of non-coding elements, may arguably harbor most of the critical determinants of the phenotype of the condition [5]. Other analyses suggest that there may be several critical regions for different phenotypes and not just one region for the whole phenotype. For example, heart defects in DS map to a particular area of 5.2 megabytes within Hsa21 [6,2]. Other specific sub-regions of Hsa21 are associated with hypotonia [3] and with one form of acute leukemia (megakaryoblastic) accounting for about 20% of the cases of leukemia in children with DS [7]. It is also possible that the overexpression of a number of genes on Hsa21 in TR21 leads to a genetic imbalance that deregulates the expression of other genes in other regions of Hsa21 or even in the entire genome [8]. The expression levels of genes located outside of Hsa21 may also be altered as some of the Hsa21 genes have a regulatory role extending beyond this chromosome [9]. Pelleri et al. [10] and Pelleri et al. [11] published the results of a reanalysis of 132 cases of partial segmental TR21. The main result is that there is one highly restricted region in DSCR, which they dubbed HR (highly restricted)-DSCR, of only 34 kilobases. It is located in the distal part of the 21q22.13 sub-band. Duplication of this region is shared by all subjects with DS but is absent in people without DS. A caveat is that this region contains no known gene. As to the identification of the genetic determinants located in HR-DSCR, the authors speculate that unknown microRNAs (miRNAs) in this region could be involved in DS pathology with the capability to regulate a large number of protein-coding genes. As a result, the HR-DSCR could carry longer-range interactions with other chromosomes. The cause of standard TR21 is chromosomal nondisjunction mostly during meiosis I in the maternal egg. Paternal nondisjunction occurs during meiosis II in spermatogenesis. Translocations involving Hsa21 may occur de novo during syngamy or be inherited from parental genotypes (in about one quarter of the cases). Although these parents as heterozygous carriers are phenotypically normal (they have no genetic material in excess or deficit and their translocation is equilibrated), they have a 10% risk of having a child with DS if the mother carries the translocation and 2.5% risk if the father is the carrier [12]. This paper analyses ongoing work in genetics, epigenetics, and cognitive pharmacology relevant to the neurobiology of DS. In spite of a large number of conceptual and methodological limitations of these studies, the preliminary character of most of the findings reported, and the biologcal gaps between animal (murine) and human cognition, it would seem that this trend of research holds a potential for improving the neurobiology and possibly behavioral aspects of people affected with DS. Neurocognitive development and functioning in persons with DS have been the object of much research worldwide over the last 75 years. There is no need to represent all or parts of this huge literature in the present context nor to propose a detailed analysis of what is known of the cognitive processes and limitations of people with DS. (See, for example, [13]; for analyses of the language difficulties of persons with DS, one may consult [14,15]). Genetic and epigenetic approaches Chromosome and gene correction and reduction of excess protein production contributed by the triplication of a number of genes along the long arm of Hsa21 are the objects of ambitious experimental attempts. Chromosome correction At least three different techniques were reported to have met success in removing one extra Hsa21 in trisomic cells. Takahashi et al. [16] demonstrated that induced pluripotent stem cells (iPSCs) can be generated from adult human dermal fibroblasts through genetic engineering using transcription factors. Pluripotent stem cells are capable of self-renewing and differentiating into a limited set of specialized cells in the body such as blood, liver, heart, or brain cells, but not all types of cells as is the case for embryonic or so-called omnipotent or multipotent stem cells originating from the inner mass of the embryonic blastocyst. On this basis, Li et al. [17] generated iPSCs from fibroblasts obtained from adults with DS. They introduced a TKNEO fusion transgene carried by a modified adenovirus at the locus 21q21.3 of the gene APP (amyloid-bêta precursor protein) into one copy of Hsa21. This gene was chosen due to its location on Hsa21 and high expression in iPSCs.The operation resulted in spontaneous loss of an entire copy of Hsa21 in a large majority of clones while point mutations, epigenetic silencing, and TKNEO deletions occurred at lower frequencies in the five experiments undertaken. No damage to other chromosomes was observed. Disomic cells proliferated faster in a co-culture than their trisomic counterparts doubling their population on average in about 37 ± 0.7 hrs against 45 ± 0.09 hrs for trisomic counterparts. The authors suggest that iPSCs offer a promising way to study human trisomy because they can be derived from the somatic cells of individuals with trisomy. Their approach could also be used to eliminate unwanted trisomies arising in stem cell cultures. Jiang et al. [18] took advantage of a natural phenomenon to correct TR21. Nature has evolved a mechanism to compensate for the difference in dosage of X-linked gene copies between mammalian females and males. In humans, the formulae for the sex chromosomes are XY for males and XX for females. However, the Y chromosome is much smaller than its X counterpart. It contains only a few dozen genes compared to about 3000 for the X. Natural X dosage reduction in females is driven by a particular large non-coding RNA, named XIST (for X-inactive specific transcript) produced exclusively from the inactive X chromosome. This RNA inactivates the DNA (deoxyribonucleic acid) of this chromosome through methylation and chromatin modification turning it into a Barr body. Jiang et al. [18] reprogrammed fibroblasts obtained from males with DS into iPSCs. They inserted a transgene XIST at locus 21q22 of the gene DYRK1A in one of the three Hsas21. This silenced this chromosome in 85% of the clones treated. Silencing of a dozen genes on the inactivated Hsa21 was confirmed. No alterations of the other chromosomes were observed. A few sub-clones remained in the 245 colonies of cells treated showing either one Hsa21 fused with the XIST RNA, two Hsas21 in the same state, or the three Hsas21 fused with the XIST RNA. As in the experiment of Li et al. [17], disomic cells exhibited a capacity for in vitro proliferation above trisomic counterparts. Another RNA in mammalian females, antagonist of XIST, named TSIX (anagram for XIST), has been identified. XIST and TSIX neutralize each other on the X chromosome that remains active, whereas the expression of TSIX is stopped on the inactivated X chromosome. In the experiment of Jiang et al. [18], the fibroblasts were obtained from males with DS. It is not known what would happen in the same situation using fibroblasts from females with DS. Applying the above XIST dosage compensation technique, one could end up with some cells in the colonies treated exhibiting one Hsa21 and one X chromosome inactivated, and others where one Hsa21 and the two X chromosomes would be inactivated. Natural silencing of one chromosome X in females is never complete and the choice of the genes remaining active is random. However, Jiang et al. [18] reported that the global expressivity of the two active Hsa21 was reduced by 20, 15, and 19%, respectively in the three clones tested, which is close to the 22% usually observed in disomic iPSCs that lack the third Hsa21 altogether. This suggests that the XIST RNA inserted in the extra Hsa21 covers key regions of this chromosome actually preventing transcription factors from reading the sequence of nucleic acids. XIST appears to induce a robust dosage compensation of most Hsa genes overexpressed in TR21. Amano et al. [19] normalized the karyotypes in a culture of mouse embryonic stem cells engineered to become aneuploid or polyploid, using a biologic made of a mammalian-specific gene, ZSCAN4 (zinc finger and scan domain) containing 4 transcription factors regularly expressed in preimplantation embryos and occasionally in stem cells, which were encoded for delivery in a synthetic messenger RNA (mRNA) and Sendai virus vector. They tested this biologic on iPSCs generated from fibroblasts obtained from individuals with DS and carrying standard TR21. Within a few weeks, chromosome examination showed the emergence of up to 24% and then 40% of cells with a normal karyotype. These findings were confirmed by whole-exome sequencing. Similar results were obtained for cells with TR18, known as Edwards syndrome. The authors suggest that introduction of human ZSCAN4-mRNAs into cells may have the ability to remove chromosome extracopies without affecting the remainder of the genome. They speculate that a ZSCAN4-mediated mechanism detects unpaired chromosomes during cell division (meiosis or mitosis) and detaches them from the rest of the replication apparatus. An unexpected result was reported by Inoue et al. [20]. They established independent iPSC lines derived from amniotic fluid obtained from a female fetus with DS associated with polyhydroaminios at 29 weeks of gestation for the purpose of reducing amniotic fluid. Karyotypic analyses confirmed that all iPSCs contained Hsa21 trisomy in all the cell lines. They then continuously cultivated the iPSCs lines for 70 weeks. At that time, normal Hsa21 diploids were observed in 20% of the cells. The experiment was repeated several times to make sure that trisomy rescue was not due to Hsa21mosaicism. Expression counts based on gene chip analyses performed on the diploid and the TR21 iPSCs indicated that the expression levels of the genes for DYRK1A, SOD1, ETS2, APP, and DSCR1 decreased to two-thirds in diploid iPSCs Hsa21 compared to TR21 iPSCs. This implies that the revertant cells had regained normal genic expression. A sample of the two types of iPSCs cells was differentiated into neural stem cells (NSCs) as to morphology and neural marker expression. It was observed that the expression levels of genes APP and DSCR1 in TR21 NSCs were higher than in diploid NSCs. The authors attribute the observed spontaneous reversion from trisomic cells to disomy without genetic manipulation, chemical treatment, or exposure to radiation, to mitotic chromosome nondisjunction during iPSCs long-time cultivation. Trisomic rescue may be tissue-dependent [21]. This could mean that the tissue environment of aneuploid cells is a key variable in chromosome regulation. Although the way towards clinical applications of the above technologies may still be far off, these results obtained suggest the feasability of normalizing operations on trisomic cells at least in vitro. Acting on genes The genetic "scissor" CRISPR-Cas9 can perform precise cutting on the DNA and RNA ribbons. It is possible to remove a specific portion of a chromosome and see what happens when particular genes are removed. Enzymes that have the ability to catalyze larger DNA or RNA molecules are employed. One may envisage removing or inactivating triplicated genes located on Hsa21. Partial corrections may already have a phenotypic effect. This is suggested by the milder phenotype characteristic of those cells of TR21 mosaicism that have a normal karyotype. Efforts are being made to identify the particular set of genes whose overexpression determines the brain alterations observed in persons with DS. Several genes may interact in impairing neurological development. Some genes on Hsa21 may also have an impact on other genes in the genome. A number of genes on Hsa21 show dosage effect in TR21. This means that their expression is increased (more or less 50% in general) in the cells or tissues of persons with DS. This increase can be measured by biological activity (enzymatic, for example), quantity of proteins produced, or mi-RNAs. Ait Yahia-Graison et al. [4] counted 120 genes expressed in lymphoblastoid cells derived from 10 persons with DS (3 women and 7 men) and 11 controlled individuals (4 women and 7 men). Twenty-two percent of these genes were overexpressed in DS cells in correspondence with the genedosage effect and 7% were amplified beyond this level. Fifteen percent were highly variable among individuals with DS, which may account at least partially for the phenotypic variability observed in the syndrome. Fifty-seven genes were found to be compensated by decreased or increased transcription, which the authors consider an indication of genetic robustness. However, there exists important variations in gene expression in the DS condition. As already suggested [22], it is unlikely that all genes on Hsa21 would produce equally marked biological harm when triplicated because this would have lethal consequences for the affected persons. As suggested in DS, DS-derived iPSCs, and DS mouse models, overexpression of genes DYRK1A, APP, EURL involved in various cell functions and structural aspects of neurogenesis, of OLIG1/2 responsible for myelinating cells and oligodendrocyte differentiation, and of ERG, and RCAN1 affecting the central nervous system, is probably among the most noxious mechanisms in T21 brain etiopathology [23]. For example, DYRK1A transgenic mice exhibit neurogenesis alterations and brain and behavioral abnormalities comparable to those of human beings with T21 [24]. Thomazeau et al. [25] found that overexpression of DYRK1A increases the number of spines on oblique dendrites of pyramidal neurons in the pre-frontal brain of adult mice trangenic for gene DYRK1A. Li et al. [26] observed that perturbation of EURL mRNA levels in mice C57BL/6 impairs progenitor proliferation and neuronal differentiation, and reduces the dendritic spine densities of cortical neurons. These authors objectified similar features in tissue samples from human fetuses between 16-19 weeks of gestation. Chakrabarti et al. [27] established that overexpression of OLIG1 and OLIG2 in the forebrain of mice Ts65Dn leads to defective neurogenesis. Manley and Anderson [28] observed that OLIG2 gene dosage alters cerebral cortical interneuron development and contributes to cognitive disability in mice. Ishihara et al. [29] observed that ERG gene triplication contributes to dysregulation of the homeostatic proportion of the populations of immune cells in the embryonic brain and decreases prenatal cortical neurogenesis in a mouse model. There are a large number of RNAs, some protein-coding and others non-coding. Hsa21 harbors four major types of micro RNAs (miRNA-99a, miRNA-125b, miRNA-155, and miRNA-802). These are short non-coding RNAs mediating post-transcriptional gene silencing, whose overproduction is involved in the etiopathology of T21. Selective inactivation of all or at least some of these RNAs could be an efficient strategy for improving the DS phenotype [30]. Other RNAs can be used to silence other genes (see [31], for an example). Wang et al. [32] documented an association between T21 and abnormally low levels of proteins SNX27 in cells. This protein protects neurons from excess quantities of neurotransmitter glutamate. Increased production of miRNA-155, linked to the triplication of several genes on Hsa21, is correlated with a reduced amount of SNX27. Wang et al. [32] used mice genetically modified to produce fewer proteins SNX27 (mice named SNX27+/−). These mice showed a learning and memory handicap associated with reduced synaptic recycling of glutamate in spite of a globally normal neuroanatomy. Synapses operate in association with specific neurotransmitters. They are produced presynaptically and recuperated post-synaptically. Wang et al. [32] engineered a therapy at the level of the hippocampus for restoring normal levels of protein SNX27 and post-synaptic glutamate recycling. This resulted in better learning and memory in the treated mice. Murine models are useful tools in genetic research and engineering. However, TR21 in humans is orders-of-magnitude more complex than the corresponding mouse models, especially related to cognitive functions. Mice Ts65Dn and Ts1Cje, for example, are genetically modified to correspond to the triplication of Hsa21 in humans (see [30] for a differential analysis of various mouse models). Ts65Dn partially mimics the DS human condition including the developmental delay and memory deficit. Ts1Cje mice exhibit drastic limitations in spatial learning and present craniofacial alterations. Genes in mice orthologous to Hsa21 are distributed on Mmu chromosomes 10 (39 genes), 16 (112 genes), and 17 (19 genes). In these regions, the orthologous genes are syntenically conserved. Ts65Dn mice have genes corresponding approximately to 60% of the genes harbored by Hsa21 [33]. A recent study by Aziz et al. [34] suggest that there are important phenotypic differences between cytogenetically distinct mouse models of DS. Using the same longitudinal experimental design, they compared Ts65Dn, Ts1Cje, and a third variety of mice, Dp(16)1/Yey. This last variety, engeneered more recently, duplicates a 23.3 megabytes segment of Hsa21(119 genes). This means that, although these mouse models are extremely valuable for the study of genotypic/phenotypic features in DS, they are incomplete. Mice with full T21, i.e., with all their genes ortholog to those on Hsa21 triplicated, can be created by complex crossing of genetic lines (for example, [35]). But they are difficult to produce, expansive, and short living. Transgenic mice are mice whose genomes have been modified by the transfer of a gene or a chromosome from another species. Fillat et al. [31] employed an adenovirus-associated AAV2/1-shDYRK1A as vector to carry a miRNA into the hippocampus of experimental mice Ts65Dn for reducing the expressivity of gene DYRK1A and in the control group a virus with a sequence that does not interfere with the gene DYRK1A (AAV2/1-scDyrk1A(SC). Separately, a lentivirus LY-anti-miR155-802 was used to increase the expression of another gene, MECP2, located on the X chromosome at the locus Xq28. In both cases, a statistically significant improvement (p < 0.01) was observed in the learning ability of the treated mice on the Morris aquatic labyrinth against the euploid mice. This research suggests that a modification of genes located on other chromosomes than the 21st, for example a downregulation of MECP2, may also have a role in the DS phenotype. It also opens a biological way to address the reduced or suppressed function of the gene MECP2 on the active X chromosome that is causally involved in the etiology of Rett syndrome; a degenerative condition in female children and adults characterized by serious difficulties and limitations in language, motor behavior, and intellectual capability. Mouse chimeras are used in experimental neurology to investigate cell development. They are obtained by introduction of targeted embryonic stem cells into blastocysts (early embryonic stage) resulting in animals that have two (or more) populations of genetically distinct cells coming from different zygotes. Xu et al. [36] observed that iPSCs derived from persons with DS overproduce OLIG2 ventral forebrain neural progenitors, which favors excess production of sub-classes of GABAergic interneurons (neurons organizing circuits between efferent and afferent neuronal bodies). Transferred in neural chimeric mice, this causes impaired memory recognition. Short hairpin RNAs (shRNAs) were used to reverse abnormal OLIG2 expression. This reduced interneuron production in DS iPSCs and chimeric mouse brains, which in turn reduced behavioral deficits in the latter. The implication drawn by the authors is that altered OLIG2 expression may underlie neurodevelopmental abnormalities and cognitive defects in persons with DS. Chakrabarti et al. [27] removed one allele of each triplicated gene within the genome of Ts65Dn mice by breeding Ts65Dn females with OLIG1/2 double heterozygous males to normalise the dosage of genes OLIG1 and OLIG2 in the forebrain of the resulting pups. Returning Ts65Dn animals to disomy for OLIG1 and OLIG2 genes had the effect of normalising neurogenesis. This restituted a normal balance between excitatory and inhibitory neurons in the central nervous system alleviating a defect in synaptic plasticity due to the overinhibition phenotype suspected to be one of the underlying causes of cognitive deficit in Ts65Dn mice. Ishihara et al. [29] restored the perturbed proportions of immune cells in Ts1Cje mice embryo brains through a genetic manipulation rendering these mice disomic for the ERG gene (but otherwise trisomic on a Ts1Cje background). The neurogenesis defects observed in Ts1Cje were reduced in the ERG modified mice embryos. This finding suggests that ERG gene triplication may contribute to a dysregulation of the proportions of immune cells in the DS embryo, which perturbates pre-natal cortical neurogenesis. Acting on genes may not be without risk, however. The immunological systems have evolved to contrast viral aggression. It is imperative to make sure that reprogrammed cells are not confounded with infected ones and end up as targets for destruction. Gene correction may become technically possible in DS in the middle-term, including in prenatal stages. It should be envisaged with extreme caution as major risks exist for mothers and fetuses given the toxicity of the viruses, possible adverse immunological reactions, and the risks of tumor generation [37]. Acting on gene products Besides modifying genes, it is possible to operate on the proteins and enzymes encoded by these genes. Nakano-Kobayashi et al. [38] reported that the oral administration of a growth inducer identified in their screen of neural stem cells (NSCs), named Algernon (for "altered generation of neurons"), which they claim has an inhibitory activity against the expression of gene DYRK1A, rescued NSC proliferation and increased the number of newborn neurons in Ts65Dn mice derived neurospheres and in human NSCs derived from human fibroblasts with DS. When administered to pregnant Ts65Dn dams, the medication induced improved cortical formation and prevented the development of abnormal behavior in the Ts65Dn mouse offsprings. The authors suggest that ALGERNON pre-natal therapy may have the capacity of preventing structural and developmental aspects of the DS neurogenic phenotype. However, no precise information is given on the nature and chemical composition of the algernon molecular compound. A catechin molecule named epigallocathechin 3-gallate (EGCG), a polyphenol of green tea with antioxidant properties, has generated much interest in recent years. It is a natural inhibitor of the enzyme encoded by gene DYRK1A. Experiments with murine models support the efficiency of EGCG for rescuing various aspects of neurogenesis. For example, in the course of the study mentioned before, Thomazeau et al. [25] administered drinking water containing 25% green tea decaffeinated extract (0.08 mg/ml) and 25% glucose to male mice Tg189N3 (mBACtgDyrk1a) aged 4-6 months for 4 to 6 weeks. These mice with a third copy of DYRK1A expressed more DYRK1A brain proteins compared to wild-type littermates. Green tea extracts contained 45% EGCG. Daily doses ranged between 120 and 200 mg per kilo of weight. Thomazeau et al. observed a normalization of spine density in deep layer pyramidal cells of the prefrontal cortex associated with a rescue of long-term potentiation of synapses in memory structures. This study suggests that the origin of the morphological and functional DYRK1A-related deficits in the pre-frontal cortex is not merely developmental. Continued overexpression of DYRK1A in adult times may also be detrimental to brain pre-frontal functioning. Contrasting DYRK1A excessive activity also in adolescents and adults years could be justified in DS. Hibaoui et al. [39] induced iPSCs from primary fetal skin fibroblasts obtained post mortem from discordant monozygotic twins, one with T21 (twin-DS) and the other normal (twin-N) in an attempt to control for the genomic background. Both the iPSCs and the NPSCs (i.e., iPSCs derived into neural cells) exhibited defects associated with changes in the architecture and density of neurons, glial cells and with the expression of genes involved in neurogenesis. In particular, a two-fold increase in DYRK1A enzymes was found in NPSCs. Hibaoui et al. [39] reported rescuing neurogenesis impairment in these cells through DYRK1A inhibition using EGCG or a short-hairpin RNA silencing (shRNA); the latter one in order to exclude a simple anti-oxidant effect of EGCG. EGCG improved the number of NPSCs derived from twin-DS-iPSCs by promoting cell proliferation and preventing apoptosis. When these NPSCs were further induced into mature neurons, DYRK1A inhibition through EGCC or shRNA treatment improved the expression of neuronal markers B3-TUBULIN and MAP2, an indication of improved neurogenesis. A reduced expression of several genes, up to 30%, in NPSCs derived from the twin-DS-IPSCs, was promoted by restoring DYRK1A expression to near normal levels by shRNA, confirming that the DYRK1A gene when overexpressed is a major contributor to impaired neurogenesis in DS. Moreover the genetic profiling of the twin-DS-iPSCs compared to the twin-N-iPSCs revealed that TR21 not only affects the expression of trisomic but also of disomic genes. As many as 96 genes related to brain functions were found to be downregulated. This finding suggests that TR21 actually determines an alteration of the whole set of RNAs resulting from gene expression, as suggested by Lyle et al. [3]. Corresponding data have been published by Guedj et al. [40]. They submitted transgenic Ts65Dn and Ts1Cje mice carrying an additional copy of gene DYRK1A to a diet rich in green tea polyphenols from gestation to adult age. The study included four groups of mice: wild-type and trangenic, fed either with water or green tea. Green tea infusions were given daily. The chronic polyphenol-based diet rescued major features of the transgenic phenotype as demonstrated by significant statistical differences between the groups in brain weight, brain magnetic resonance imagery, and volume of the thalamus-hypothalamus region (p < 0.05). The above studies and others (see Stagni et al. [41], for a summary of additional works) suggest the interest of an EGCG strategy for improving neurogenesis in mice genetically modified or transgenic for the gene DYRK1A. It must be noted, however, that these studies differ from each other in several respects that may interact with the effect of EGCG: for example, length of treatment, age of the animals, dosage, administration of other green tea extracts in ill-defined combinations with EGCG (see Stagni et al. [41], for an overall discussion). This prevents solid conclusions to be drawn. Very few studies have been conducted for testing the effect of EGCG at the human level. In a first experiment, De la Torre et al. [42] witnessed positive effects on visual recognition in a group of young adults with DS following three months of daily treatment with EGCG (green tea extracts, low in caffeine, 9 mg per kilo of body weight, administered orally). No adverse event linked to the medication was observed. However, three months after the end of treatment, the participants' performance returned to pre-intervention level. In a second study, De la Torre et al. [43] tested the effect of a daily EGCG treatment (with a dosage identical to the preceding study), lasting for one year, coupled with a behavioral program of cognitive training. The sample included 84 adults with DS (about as many women as men) aged 16 to 34 years. They were divided into two groups: one treated with EGCG and undergoing cognitive training; the other receiving a placebo and the same cognitive training as the first group. Long-term administration of the medication in the experimental group appeared to be well tolerated A battery of neuropsychological tests was administered at the end of the study. Participants treated with EGCG and exposed to cognitive training showed a statistically significant superiority (p < 0.05) in two cognitive tasks (memory and visual recognition) and in several adaptive tasks (daily routines abilities). A retest 16 months later showed partial persistence of the effects measured at the end of the intervention. The above data suggest that EGCG may improve brain defects in TR21 mice, genetically modified or transgenic. It is questionable, however, whether the product actually eliminates these defects. At the human level, 12 months of treatment with green tea extracts containing EGCG plus cognitive training seems to have a positive effect on cognitive performance, albeit of a modest magnitude, and this effect is partially retained after discontinuation of the treatment. Xicota et al. [44] tested the effect of EGCG on the lipid profile of individuals with DS. It is known that many of these persons tend to have higher rates of obesity. A double-blind clinical trial compared the effect of EGCG administered during 12 months to a placebo on the body lipidic composition of 77 young adults with DS. Individuals receiving a placebo showed the expected increase over time in body weight and body mass index (ratio of body weight to squared body height). A similar increase was not observed in participants receiving an EGCG treatment. However, the group difference was statistically significant (p < 0.05) only for the male subjects. An anonymous survey has been conducted by Long et al. [45] on parents' attitudes regarding the administration of green tea extracts containing EGCG. Parents who give green tea extracts to their children with DS are mostly younger, highly educated and they tend to consult scientific sources. The individuals with DS who receive green tea extracts are characterized as less severely disabled to begin with. Most caregivers who do not give green teas extracts report to be concerned about potential negative side effects and they have doubts regarding treatment effectiveness. Early diagnosis The DS brain starts at a disadvantage and the disadvantage is cumulative. It is essential that the attempts at improving neurogenesis and neural connectivity be conducted as early as possible in life. Major phenotypic features of DS can already be traced back to the fetal period. Early intervention implies early diagnosis of the condition. Prenatal screening for DS is possible from the eleventh week of pregnancy through the analysis of fetal DNA fragments or mRNAs in maternal blood samples. Ninety-nine percent reliability can be reached in combining ultrasound, blood analysis, cardiac rhythm, and nuchal translucency [46][47][48]. A less expensive technique has been experimented by Shan et al. [49]. They analyzed the peptidome of urine samples of pregnant women carrying fetuses with DS and with normal karyotypes. A classification model was constructed based on candidate peptides that could differentiate fetuses with DS from controls reaching a sensitivity of 95.7% and a specificity of 70%. This suggests that a maternal urinary peptidome could offer a prospect for a noninvasive biomarker screening of fetal DS. Amniocentesis and chorionic villus sampling remain the only fully reliable diagnosis techniques, but they are intrusive and there is a risk of a miscarriage between 0.5 and 1% [50]. Progress is being made in the development of noninvasive prenatal diagnostic methods. Asim et al. [2] have published an analysis of the advantages and disadvantages of a series of molecular methods for prenatal diagnosis of DS (e.g., cytogenetics analysis, fluorescence in situ hybridization). A noninvasive diagnostic technique has been experimented by Zbucka-Kretowska et al. [51]. They studied the expression level of miRNAs in the plasma of 198 pregnant women with fetal DS at 15-18 weeks of gestation and 12 women with uncomplicated pregnancies who delivered healthy newborns at term. Out of 800 miRNAs analyzed by Zbucka-Kretowska et al. [51], six were upregulated and seven downregulated in plasma samples of women with fetal DS. The genes regulated by these miRNAs are involved in central nervous system development, congenital abnormalities, and heart defects. This study opens the way towards designing a panel of miRNAs for a nonintrusive technique of prenatal DS diagnosis. Genetic corrections in the early stages of prenatal development could become possible in the future. The ultimate target is embryonic gene correction inducing normalized genetic expression in most body cells. The technology exists pending further refinements and security checks. It will leads to the development of complexes of mosaic cells, part normal, part aneuploidic, as there is no way to modify the entire embryo's line cell in the present state of gene editing technology. It is known that neurogenesis continues in the ventricular and subventricular zones of cerebral cortex during the third trimester of pregnancy [52]. Later corrective intervention in fetuses with DS therefore could still have a therapeutic interest. Molecular pharmacology The neurobiological consequences of DS result in reduction of synaptic density and plasticity. Much attention has been devoted recently to the neurotransmitters. Gotti et al. [53] have published a review of studies on the alterations of brain circuits that can be identified in murine models of DS. It shows that different neurotransmission systems are downgraded or upgraded in several cerebral regions including the hippocampus, the locus coeruleus, and the frontal cortex. Major neurotransmitter systems involved in brain functions are the cholinergic system (i.e., an excitatory neurotransmitter), the noradrenergic system (with the mostly excitatory neurotransmitter noradrenaline), the glutamate system (i.e., brain's major excitatory neurotransmitter), the GABAergic system (i.e., brain's major inhibitory neurotransmitter), and the serotoninergic system (also inhibitory). Drugs are being developed for reducing the neurotransmission deficits caused by TR21. Interesting results have been obtained with the Ts65Dn murine model of TR21. Several substances have been shown to rescue at least partially learning and memory deficits. In some cases, attempts have been made to extent the intervention to humans with DS. Cholinergic system Inhibitory agents of acetylcholinesterase have been tested which catalyzes acetylcholine postsynaptically reducing it to its basic constituents and allowing neurons to return to a state of rest following activation. Administration of physostigmine has been demonstrated to inhibit acetylcholinesterase and increase concentration of acetylcholine in the synaptic area in Ts65Dn mice at 4 months of age (but no longer at 10 and 16 months) and to improve learning and memory [54]. Maternal choline acetyltransferase supplementation from conception to weaning has been showed to improve spatial learning in mice derived from Ts65Dn dams and tested on a water maze at 14-18 months. Hippocampal tissues were examined for intensity of choline acetyltransferase immunoreactivity. The increased innervation produced by choline supplementation appears to improve hippocampal function [55]. At the human level, Heller et al. [56] have reported an open-label case series of six nondemented adults subjects with DS (aged 20-41 years) receiving 5 then 10 mg donepezil over 24 weeks. Despite transient side-effects expectable from cholinergic overstimulation, all subjects tolerated well the 10 mg dosage. A small but significant improvement (p < 0.05) in expressive language was observed at 24 weeks in four sub-tests of the Clinical Evaluation of Language Function-Revised (CELF-R). However, the small sample of subjects and the lack of a control group render the outcome of this experiment difficult to interpret. Heller et al. [57] reported results from a 16-week pilot clinical trial on the effects of donepezil on the language of 5 children with DS aged 8 to 13 years. The drug was dosed orally at 2.5 mg once daily for 8 weeks and at 5 mg for the remaining 8 weeks. Two language measures were used: the Test of Problem Solving (TOPS) and the Clinical Evaluation of Language Fundamentals (CELF-3). Medication effects were measured by changes from the baseline to performance at weeks 8 and 16. No subject experienced serious adverse effects from the administration of the medication. T-tests yielded no significant indication of change in language performance between base line and TOPS scores at 8 and 16 weeks. However, a significant improvement (p < 0.003) was registered in the mean CELF-3 performance from baseline to week 16. Results from this study are also difficult to interpret owing to the lack of a control group and the small sample of subjects. Kishnani et al. [58] tested the efficiency of donepezil in a sample of 123 young adults with DS (aged 18-35 years), who had no evidence of AD, in a 12-week randomized, double-blind, and placebo-controlled study. Experimental subjects were treated with doses of 5 mg per kilo of weight for 6 weeks and then 10 mg/kilo in the following 6 weeks. Cognitive measures included the Severe Impairment Battery (SIB), the Rivermead Behavioral Memory Test for Children, and the Clinical Evaluation of Language Fundamentals (CELF-3). Additionally the Vineland Adaptive Behavior Scales were administered. Donepezil appeared safe. However, two subjects had to be withdrawn from the double-blind phase for hypertension rated as possibly problematic by the investigators. Outcomes suggested efficacy of the drug in some but not all subjects, which the authors estimate consistent with the phenotypical variability in DS. However, improvements were also observed in the placebo group, particularly on the SIB, during the double-blind phase of the study, preventing clear-cut conclusions regarding the cognitive specificity of this molecule. Kishnani et al. [59] assessed the efficiency and safety of donepezil with 129 children and adolescents with DS (aged 10-17 years) in a 10-week, randomized, double-blind, placebo-controlled, and multicenter study. Participants received a dose of 2.5 mg/kilo donepezil in the first part of the experiment, increased every 14 days until reaching 10 mg/kilo. Measures included the Vineland-II Adaptive Behavior Scales Parent/Caregiver Rating Form (VABS II/PCRF). The medication appeared to be well tolerated. During the double-blind phase, the VABS II/PCRF scores improved significantly (p < 0.001) in both the treated and control groups but with no significant difference between groups. Rivastigmine was tested by Heller and associates [60,61] in two open-label studies with 10 children and adolescents with DS (8 boys and 2 girls) aged 10 to 17 years. Doses at the beginning of the intervention were 1.5 mg/kilo increased to 3 mg/kilo and 4.5 mg/kilo in the following weeks. Five subjects reported no adverse events with the medication, but five others signalled transient vomiting, diarrhea, fatigue, or insomnia related to cholinergic enhancement. After 16 weeks, there were statistically significant gains in expressive language on the Test of Verbal Expression and Reasoning (TOVER; p < 0.02), as well as for language measures (narrative memory and immediate memory for names on the Developmental Neuropsychological Assessment Test (NEPSY; p < 0.02). There were also significant improvements in attention on the Leiter-R Attention Sustained Tests A and B (p < 0.01 and p < 0.02, respectively). Five participants (4 boys and 1 girl) continued the treatment for another 38 months. Longer-term use of rivastigmine appeared to have no adverse effect on overall health. A comparison of median performance change between those who continued the treatment for 38 months versus those who did not yielded no statistically significant difference. However, two subjects demonstrated important improvements in adaptive function (measured on the VABS) over the longer-term period with continued rivastigmine administration. The results of the above studies suggest that the effects of donepezil and rivastigmine administration are not consistent across individuals with DS. Many participants show little to no gains but a small subset of individuals seem to respond positively to the cholinergic intervention. Noradrenergic system A reduction of the concentration of thyroid hormones is observed in the locus coeruleus of Ts65Dn mice from 6 weeks of age in comparison with normal mice. This can be hypothesized to reduce the supply of noradrenaline in the hippocampus. Ts65Dn mice treated with formoterol that stimulates the adrenergic receptors appear to have a more favorable neurogenesis [62]. Neurotransmitter glutamate Molecules have been tested that target the receptors NMDA (N-methyl-D-aspartate) of the neurotransmitter glutamate in order to reduce an excess of activation due to this neurotransmitter. Costa et al. [63] observed that acute injections of the NMDA receptor antagonist memantine improves learning in Ts65Dn mice. Administered to young human adults with DS, in preliminary studies, memantine has shown encouraging clinical signs [64]. GABA system Also in the pipeline of cognitive pharmacotherapy, is the GABA neurotransmitter system (gamma aminobutyric acid). GABA receptor antagonists picrotoxin and pentylenetetrazole have been tested by Fernandez et al. [65]. Two weeks of daily injection for each drug appears to rescue maze learning and object recognition in Ts65Dn mice. Pentylenetetrazole is controversial for possible clinical application in humans because it is known to be convulsant. A longitudinal validation study of several cognitive scales (for example, the Leiter Performance Scale-Revised and the Clinical Evaluation of Language Fundamentals-Preschool-2) currently used to assess memory, executive function, and language in individuals with DS [66] served for optimizing trial design and endpoint selection in a clinical trial testing basmisanil, a negative allosteric modulator of GABA receptor alpha5-subtype. The trial revealed no improvement of cognitive abilities in individual with DS. Serotoninergic system A series of studies with Ts65Dn mice shows that chronic treatment with fluoxetine, an antidepressant antagonist of serotonin synaptic recapture, from postnatal days, rescues abnormalities in neurogenesis and stimulates the production of neurons and their incorporation into functional networks. Drug concentration must be carefully controlled for higher levels of fluoxetine in daily consumption may provoke seizures [62]. Guidi et al. [67] administered fluoxetine to pregnant mice Ts65Dn. After delivery, the mice whose mothers had been given fluoxetine exhibited normal brain neuronal proliferation and dendritic growth at 45 days. Fluoxetine is known as a molecule interacting with the class of enzymes histone deacetylase that catalyzes a reduction of the acetyl group thereby reducing genic expression. Another series of biochemical agents have nootropic (nonspecific brain boosting) and antioxidant effects. They contrast continued production in the cells of biological residus of oxygen reduction (more or less 2% of the regular oxygen intake) that are harmful to the biochemistry of the body and participate to cellular aging in the long term. Overexpression of gene SOD-1 in persons with DS leads to aggregation of hydrogen peroxide within cells and oxidative damage. Vitamin E has been tested as antioxidant therapy with positive results on memory and basal forebrain pathology in mouse models of DS [68] but no positive indication in humans with DS [69]. Piracetam increases brain oxygenation and improves GABA neurotransmission. Lobaugh et al. [70] assessed the cognitive and adaptive effects of piracetam in a double-blind study with 18 children with DS (aged 7-13 years). The experimental group received 80-100 mg/kg piracetam per day for 15 weeks. No statistically significant benefit was observed in the treated sujects in comparison with the control group in a series of tests measuring learning, attention, and memory, as well as on the adaptive scales. Treatment induced side effects of irritability and poor sleep in 7 subjects. An antibiotic commonly used to treat acne, minocycline, seems to have neuroprotective effects and inhibit neuron apoptosis. Three months of treatment in 10-month old Ts65Dn mice significantly improved performance in cognitive tasks (p < 0.05) [71]. Chen et al. [72] observed that glial cells (astrocytes) supplying nutrients to the neurons and detoxifying the extracellular milieu in neutralizing excess glutamate, induced in vitro from stem cells derived from fibroblasts of persons with DS, do not favor in vivo neurogenesis when transplanted into iPSCs mice brains. Minocycline corrects this deficiency by modulating the expression of gene S100B, located on chromosome 21 at locus 21q22.3. This gene regulates various cellular processes including the glial function. Broze et al. [73] tested the ability of hydroxyurea, a derivative of urea, to activate neural pathways in Ts65Dn mice. Treatment was initiated when the mice were three-month old and lasted three months. A significant improvement of memory retention of spatial information was measured in the treated animals. On the whole, experiments with Ts65Dn mice suggest that different drugs can rescue learning and memory in cognitive tasks and in some cases improve neurogenesis. However, there is a need for standardization of the experimental protocols. In some studies, acute single doses of drugs are administered. In other studies chronic administration extends from a few weeks to several months. The ages of the treated mice vary from early pre-natal to adulthood. Sex variation is rarely carefully controlled. A conundrum with present-day murine models concerns the non-Hsa21 ortholog genes in genetically modified mice. It is estimated that there are 50 of these genes in Ts65Dn mice. When overexpressed they should contribute to the phenotype of these animals. Also the molecular basis of drug responses must be better defined before envisaging pre-clinical and clinical trials [62]. The neurobiological gap between mice and human beings cannot be underestimated. Any generalization must be considered with maximum caution and the conclusions envisaged only hypothetically. Cognitive pharmacotherapy in DS is still in infancy and the results obtained so far with some of the drugs tested are preliminary. Drug safety and the occurrence of adverse events need to be better assessed and controlled. Alzheimer disease Alzheimer disease (AD) occurs more frequently and at an earlier age in persons with DS than in the general population. It affects approximately 20% of persons with DS beyond 40 years of age, 40% beyond 50 years, 80% and even more beyond 60 years. In some cases, the degenerative process is relatively slow and may occur over several years. In other cases, the pathological involution is relatively rapid. In the general population, cases of AD are found most often in persons beyond 70 years of age, although familial (hereditary) cases may have an earlier onset. The evolution of the AD pathology is generally the same in persons with DS as in the general population except for the temporal aspects. However, there are a few differences that deserve additional investigations. For example, symptoms of depression are more pronounced and epileptic seizures more frequent in persons with DS, particularly in the early phases of the disease [74]. Doran et al. [75] submitted one man with DS when he was between 66 and 72 years to neuropsychological testing, neurological examination, amyloid PET imaging, and a series of related measurements. This person had partial T21 lacking triplication of the APP gene. The clinical phenotype was typical for DS. No dementia was ascertained on neurological examination. Post mortem neuropathological findings showed only a single neuritic plaque and neurofibrillary degeneration consistent with normal aging but not with AD. The authors suggest that APP could have an obligatory role in the clinical and neuropathological findings of AD in DS. However, the role of the APP gene is not clear in the evolution of normal persons developing AD, given that less than 10% of the mutations affecting this gene are responsible for early-onset AD. It could be that the APP gene takes part in, but is not needed for AD pathology to occur [76]. Important quantities of nontoxic amyloid-alpha proteins are produced naturally in the brain. In DS, a phenomenon not yet identified switches this production to toxic peptides amyloid-bêta 40 and 42 that cannot be dissolved in brain fluids. These peptides initiate the formation of amyloid plaques. They are tiny structures of 40 microns diameter composed of degenerated axons and dendrites, glial cells, and astrocytes centered around amyloid deposits. The plaques aggregate around the junctions between neurons disturbing transmission. They then circumvent neuronal bodies. This seems to accelerate a tauopathy, which may start independently of the amyloid pathophysiology. It is characterized by the aggregation within the brain neurons of a protein called TAU made of helicoidal filaments of 10 nm diameter, responsible for the neurofibrillation of the cells'cytoplasm. TAU is a natural protein involved in the construction of the cell microtubules. In the AD pathology, TAU proteins become hyperphosphorylated and that destroys the neuronal tissue. The amyloid plaques gradually invade all cortical layers, mid-brain structures, brain trunk, and cerebellum. Neurofibrillation follows a route going from entorhinal and hippocampal cortices in the medial temporal lobe to adjacent temporal cortex, frontal cortex, and then the entire brain. Alternativeley, it is also suggested that there is an early involvement of the frontal cortex and the locus coeruleus in both amyloid and tau pathology [77,78]. The primum movens of the amyloid cascade remains unknown. A massive destruction of the Meynert nucleus, located in brain basis and supplying the departure point of the cholinergic neurons, could be at its origin [79,80]. Prionic mechanisms are also involved in the etiology of AD. Prion proteins fold abnormally, spread through tissues and determine other proteins to unfold abnormally. This leads to infection and cerebral lesions. At least two forms of the bêta-amyloid proteins central to the pathology of AD act as prions [81]. It is suspected that TAU proteins may also display prion-like autoreplication properties [82]. Not all persons with DS seemingly develop AD. Compensatory mechanisms may operate. It is known that other genes on chromosomes other than the 21 st may influence the determinism and the course of AD in aging persons without DS. A gene encoding apolipoprotein E (associated with cholesterol metabolism and intervening in a number of neurophysiological processes), located on C19, exists in three allelic variants. APOE3 is the most frequent one in the population, followed by APOE4 and APOE2. Higher dosages of APOE3 and APOE4 are correlated with increased and earlier risk for AD. They are associated with higher concentrations of protein amyloid-bêta, amyloidbêta peptides, and phosphorilated TAU in the brain [83,84]. This may be true for persons with DS as well. Raha-Chowdhury et al. [85] reported that in a cohort of persons with DS diagnosed with early dementia, the APOE4 genotype appears to be prevalent. In people without DS, variant APOE2 seems to have a protective action [86]. A variant form of the APOE3 gene, labelled APOE3ch (for Christchurch in New Zealand, where it was first identified) may also have a protective role (overall against the TAU pathology) particularly when present in two copies in the genome [87]. Allelic variants of the gene SORL1 (sortilin receptor, located on C1) have been identified as additional risk factors for later AD onset [88]. When SORL1 is overexpressed, there is an increase in brain amyloid deposits. Alterations of genes PSEN1 (pre-senilin) on chromosome 14 and PSEN2 on chromosome 1 also appear to be involved at various moments in the amyloid cascade; most often in the early stages in the hereditary forms of the PSEN1 mutation, called PSEN E280A [89]. Vascular cholesterol may also be involved in interaction with genotype APOE [90]. The first molecule proposed (and approved by the US. Federal Drug Administration, FDA) for trying to control the early manifestations of AD was tacrine (tetrahydroaminoacridine), an inhibitor of acetylcholine-esterase. This increases intracerebral concentration of the neurotransmitter acetylcholine by reducing its post-synaptic degradation. However, the heavy hepatic toxicity of tacrine led the pharmaceutical firms to retire the product from the market. Other inhibitors of acetylcholine-esterase have been developed with no particular hepatic toxicity. They include donepezil, rivastigmine, and galantamine [91] which have been approved by the FDA. As indicated, a moderate cholinergic deficit exists in DS. It is increased in AD. In Ts65Dn mice, overexpression of the amyloid-bêta molecule has been shown to be associated with degeneration of cholinergic and noradrenergic neurons particularly in the brain hippocampal region [92]. On this basis, it make sense to try reducing the natural elimination of neurotransmitter acetylcholine with donepezil, rivastigmine, or galantamine, in chronic administration. Acetylcholine-esterase inhibitors are generally well tolerated by the patients. When side effects occur, they include nausea, vomiting, loss of appetite, and diarrhea, but are mostly transitory. So far, however, attempts to rescue cognitive functioning in persons with DS affected with AD in using acetylcholine-esterase have met only with limited success and primarily in patients at the early stages of the pathology [64]. Memantine (also approved by the FDA) is a molecule that regulates the activity of neurotransmitter glutamate. It acts on the NMDA receptors for regulating the quantity of calcium ions entering neurons in the propagation of the neural influx. Excess glutamate increases the activity of NMDA receptors causing an excess of calcium penetrating the cells and altering neural transmission. Similar side effects to those observed with cholinergic medication have been observed with chronic administration of memantine. Clinical trials with memantine have met with limited success in improving cognitive functioning in persons with DS and AD [93]. Immunotherapy addressing the first leg of the amyloid cascade has been tried with transgenic mice [94]. The strategy was to inoculate the APP protein in the vascular system, which stimulated the production of antibodies acting against the injected antigene but also against the amyloid plaques in the brain of the treated mice. A drastic reduction of the amyloid plaques was observed to be associated with cognitive improvements. Clinical trials with human subjects at the early stages of AD have been undertook. They determined a reduction of the amyloid charge in the brain of the persons treated but no improvement of the clinical symptoms. Lithium has been tested by Matsunaga et al. [95] in an attempt to limit cognitive deterioration in persons with DS at the first stages of AD. Positive results have been reported in comparison with a control group. However, AD incipiens in DS is characterized by episodes of depression. Lithium is known to be a mood stabilizer. It is not clear whether this molecule acts mostly as an antidepressant or whether it may constitute a genuine antagonist to early cognitive deterioration in AD. Antioxidant molecules, including nicotinamide, levocarnitine, and lipoid acid, have been tried for reducing oxidative stress, which is thought to play a role in the etiology of AD, but seemingly without clear results [96]. A long series of other molecules (e.g., phenserine, xaliproden, bapineuzumab, huperzine, intravenous hemoglobin, methylthioninium chloride, raloxifene) are in various phases of clinical trials [97]. They may lead to new pharmacological options for the treatment of AD. Among the primary suspects in the etiopathology of AD, one finds the gene DYRK1A. Overexpression of this gene is thought to be one of the major culprits of the hyperphosphorylation of the TAU protein leading to the neurofibrillation of neuronal bodies. Normalizing DYRK1A dosage by breeding Ts65Dn mice with a triplication of this gene with mice trisomic for the same DNA segment but without the gene DYRK1A, yielded Ts65Dn mice with normal dosage of this gene and a normal concentration of protein APP in cerebral cortex, hippocampus, and cerebellum [98]. Kawakubo et al. [99] treated fibroblasts obtained from persons with DS and AD pathology with harmine, an inhibitor of protein DYRK1A. Results indicated an important increase in the concentration of the enzyme neprilysin and correlatively a decrease in the concentration of DYRK1A proteins in the fibroblasts. Accumulation of amyloid plaques in people's brains intensifies some 5 to 20 years before a significant cognitive decline is observed. Global cortical atrophy, increased concentration of TAU proteins, and neurofibrillation of the cells' cytoplasm can be objectified 1 to 5 years before a diagnosis of AD is warranted. Current work has focused on validating a list of biological markers of early AD. It includes: (1) Concentration of proteins APP and TAU in blood plasma [100][101][102]; (2) Blood expression of MTRNR2L12, a gene located on C3 (at 3q11.2), almost identical to the mitochondrial gene MT-RNR2, which encodes the micropeptide HUMANIN considered to be a protective factor in familial AD [103]; (3) Blood plasma neurofilament light chains [104,105]; (4) Levels of amyloid-bêta peptides and phosphorilated TAU in neuron-derived exosomes (small extracellular vesicles secreted by the cells) [106]. Blood tests could predict onset of AD in normally aging and people with DS up to 10 years in advance in combining measures of amyloid-bêta proteins, proteins IRS-1 (involved in insulin signaling in the brain and commonly defective in people with AD), the presence of genetic variants APOE3 and APOE4 in blood plasma, and differences in miRNAs levels. The question is still pending, however, because the proportion of amyloid-bêta proteins in blood although correlating positively with the presence of amyloid plaques in the brain, does not ineluctabily means beginning AD. A proportion of aging people have them without developing the pathology. Neurological examination by transcranial magnetic stimulation offers a promise for revealing synaptic dysfunctions linked to AD and predicting cognitive decline in the early phases of the disease [107]. Assuming the validity of the amyloid cascade hypothesis, a genuine curative strategy for AD is to prevent accumulation of the amyloid plaques, proliferation of TAU proteins, and neurofibrillation of neurons'cytoplasm. A number of molecules and drugs are being tested that may prove efficient, such as inhibitors of the APP protein, vaccines and antibodies, inhibitors of the TAU protein, as well as a number of other biochemical agents thought to be able to boost brain defenses against APP and TAU toxicity. Nawa et al. [108] derived dermal fibroblasts from patients with DS. The cells showed a severe limitation of proliferation and signs of premature senescence accompanied by perturbation of homeostasis leading to accumulation of protein aggregates. They treated these cells with sodium 4phenylbutyrate (4-PBA), a drug used to regulate urea cycle disorders, and observed a decrease in the protein aggregates of the fibroblasts. Studies involving protein VPS35 (vacuolar sorting 35) that has the capability of altering the amyloid preprotein/amyloid-bêta metabolism, are also relevant. Li et al. [109] observed that triple transgenic mice (3xTg) overexpressing VPS35 exhibit better spatial learning and short-term memory compared to control animals. This improvement is associated with a significant reduction of amyloid-bêta levels and TAU phosphorylation. In vitro studies revealed reduced synaptic pathology and neuroinflammation. Vagnozzi et al. [110] reported that an enzyme labelled TPT-172 can induce higher levels of protein VPS35. In vitro studies show that overexpression of VPS35 leads to a reduction of critical TAU levels in neurons. In contrast, silencing gene VPS35 is associated with an accumulation of TAU proteins. In vivo experiments with a transgenic mouse model of TAU pathology show that downregulation of VPS35 leads to an increase in the density of TAU proteins and a reduction of synaptic integrity. It appears that active cathepsin D encoded by gene CTSD on chromosome 4 is the agent mediating the VPS35 effect on TAU accumulation due to its role in degrading toxic proteins in the brain. These latter studies open new perspectives for biochemically contrasting the second leg of the amyloid-cascade, i.e., the one concerned with TAU toxicity and neurofibrillation. Conclusion Important developments are taking place in the field of DS with a perspective of improvement in the life and cognitive functioning of persons affected with the condition. Genetic therapy is still largely experimental and mostly restricted to in vitro manipulations. It will require several additional experimental studies before being in capacity to be assessed pre-clinically and tried clinically. Cognitive pharmacotherapy is a very active field not the least regarding intellectual disabilities and DS. Conclusive outcomes have been relatively rare so far. However, a number of molecules are in the research pipeline. Some of them could prove effective as adjunct treatments for boosting brain development and neurotransmission in persons with DS in particular. However, short-and longerterm negative secondary effects need to be controlled and the drugs possible toxicity further assessed. Biological research on Alzheimer disease has identified a significant part of the causal chain leading to brain degeneration. However, pharmacological treatments have not demonstrated particular effectiveness except in a limited way at the early stages of the illness. Further research on AD in aging persons without DS will benefit to persons with DS and vice versa. AD represents an important area of biological and pharmacological research and for good reasons. In the United States alone, at present time, there are 4.7 million persons diagnosed with AD. It is estimated that 20% of the persons aged beyond 80 years are affected with AD and 40% beyond 90 years. An increase is expected in coming decades associated with possible further gains in average life expectancy. In this respect, it is relevant to mention that recent statistics in the United Kingdom suggest a higher prevalence of AD in women (65% of the cases), possibly linked to longer life expectancy; presently 72 years 8 months in women versuss 68 years 4 months for men (according to the United Nations World Population Project, revision 2015; The Guardian, July 17, 2019). Conflict of interest The author declares no conflict of interest.	2020-06-25T09:05:45.979Z	2020-05-29T00:00:00.000	{ "year": 2020, "sha1": "ee6b4294c6d21b6301cb217837d8c8faaac7e6d3", "oa_license": "CCBY", "oa_url": "https://doi.org/10.3934/neuroscience.2020012", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "60f3ac393381d809aa15ea6b36273de4eb7d493b", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
55430371	pes2o/s2orc	v3-fos-license	Comparative study of morbidities in sawmills workers from central India: a cross sectional study Wood has been harvested and processed for centuries all over the world for various purposes. 1 In India it is used mainly as a fuel and construction material but before it can be used as a construction material it is subjected to sawing. Sawmill is a facility where logs are cut into boards. 2 Large population is employed in sawmill industry in India. The main problem encountered in the sawmill environment is the respirable dust (<10 μm). 2 Around 2 million people are exposed to the wood dust every year. INTRODUCTION Wood has been harvested and processed for centuries all over the world for various purposes. 1 In India it is used mainly as a fuel and construction material but before it can be used as a construction material it is subjected to sawing. Sawmill is a facility where logs are cut into boards. 2 constituents. Rastogi et al have reported restrictive type of respiratory impairment in workers exposed to sawdust of mango and seesam. Occupational exposure to wood dust may be a cause of respiratory diseases such as hypersensitivity pneumonitis (allergic alveolitis), rhinitis, asthma and chronic obstructive lung disease. 3,4 Further the sawmill workers are also exposed to other types of health hazards such as injuries from handling wood logs, saw blade and unguarded machine parts, high noise exposure etc. There is paucity of studies on sawmill workers and their health problems in India. This study therefore aimed at studying various morbidities in sawmills workers by comparing them in control group. METHODS A cross sectional study was conducted among Sawmill workers in Nagpur city in central India. Study duration was from September 2013 to December 2015. Inclusion criterion was saw mill workers exposed to saw dust as study group. Log transporters in sawmills whose exposure to wood dust is minimal were excluded. Approval from the Institutional Ethics Committee was sought before the start of study. List of registered sawmills was obtained from the department of forest. A pilot study was carried out on 50 sawmill workers to assess the feasibility; test the proforma and get an idea about the prevalence of common morbidities in sawmill workers. Two sawmills were selected randomly. All the workers in selected sawmills except log transporters were enrolled in the study i.e. 180. Large homogenous group of local government workers were purposively selected by matching them with the selected sawmill workers in terms of socio-demographic characteristics such as age, sex (frequency matching). So 180 sawmill workers and 180 comparison group (SRPF) workers were selected. The necessary permission for carrying out the study was obtained from the owner of the sawmill and appropriate authority in case of local government workers after apprising them about the nature and purpose of the study. After obtaining the necessary consent and explaining the nature and purpose of the study, data was collected as per the guidelines of international labour organization, at suitable timings convenient to both workers as well as the appropriate authority in order to seek their maximum cooperation. There were various departments with different job descriptions in sawmills. When the tree logs transported from the forest, the log loaders transport the logs to the milling machine (log transporters); after the logs are cut into wide planks by the machine operators (cutters), stacked directly (stackers). There were laborers who assist in various operations (helpers). There was also an administrative staff that serves as the supervisor /manager /clerk / technician of the sawmill industries. Information regarding socio-demographic characteristics collected using predesigned proforma by interview technique. Thorough general, systemic examination and anthropometry was done to assess health status. Height of the subjects was measured using stadiometer (Seca) to the nearest 0.1 cm. 5 Body weight was measured in light clothing, without shoes to the nearest 0.1 kg using the electronic weighing machine. 6 Hearing function assessment done by Rinne's test and Absolute bone conduction test using tuning fork. 7 Operational definition of some important terms Physical deformity: Acquired loss or malpositioning of finger, toes, limbs or other body parts. Musculoskeletal disorders: Person with symptoms as myalgia, pain in the limbs/back/joints is labelled as having musculoskeletal disorders. Bronchial asthma: Recurrent episodes of wheezing, breathlessness, chest tightness with or without cough particularly at night or in the early morning with or without decrease in FEV 1 or known cases those who are taking asthma medication. Chronic bronchitis: This is a productive cough that lasts for three months or more per year for at least two years. Respiratory morbidity: If any one of asthma, chronic bronchitis, T.B., rhinitis and acute respiratory infection is present among study subjects then respiratory morbidity is said to be present. Cardiovascular morbidity: Any one of hypertension, ischemic heart disease, piles and varicose veins is present among study subjects then cardiovascular morbidity is said to be present. Statistical analysis The data was analyzed using Epi info 7.0.1. The quantitative variables were expressed in terms of mean and standard deviation and the qualitative variables were expressed in percentages. Chi square test was used to find the difference between two proportions. Fischer exact test and chi square test for trend was used wherever applicable. Unadjusted odd's ratio was calculated for various morbidities. RESULTS Sawmills were located in the outskirts of Nagpur city. Main wood species used for sawing was Tectona grandis (Teak). In total, 360 workers were employed in the study, 180 of them sawmill workers (study group) and 180 local government workers as a control group. Socio-demographic characteristics of study subjects mentioned in the Table 1. It shown that majority of the study subjects were males (87%) and belonged to younger age group (28-37 years). But the two groups differ in socioeconomic status and education. Most of the sawmill workers were from lower socioeconomic class and had low level of education as compared to control group. Mean duration of work (years)= 7.59±5.89. Most of the sawmill workers working in cutters (31.66%) and helpers (57.22%) section while least (around 5%) were employed in the administrative section ( Table 2). Study of distribution of sawmill workers according to duration of work (Table 3) shown that mean duration of work (years) was 7.59 with standard deviation of 5.89. Majority of the workers (47.78%) from our study had work exposure less than 5 years. 8.89% of the workers had work exposure of more than 16 years. Most commonly sawmill workers were suffered from musculoskeletal disorders (Table 4) and it was significantly more than comparison group (p<0.05). Some of the other significantly common morbidities among sawmill workers (p<0.05) were conjunctivitis, rhinitis, injuries, hearing impairment, ARI's, bronchial asthma, chronic bronchitis, contact dermatitis, abdominal hernia, varicose veins and pulmonary tuberculosis. When we studied the distribution of sawmill workers according to morbidities and duration of work ( Table 5) we found that cardiovascular morbidity, hernia and hearing impairment had shown significantly (p<0.05) increasing trend with duration of work. Study of morbidities according to category of work ( Table 6), conclude that cutters were significantly (p<0.05) more prone to injuries than the workers in other section, while the other morbidities did not show any significant difference between various categories of work. DISCUSSION In our study workers employed in sawmills were exposed to high concentration of sawdust besides the other risk factors. In the present study majority of the sawmill workers were young productive males in the age group of 28-37 years with mean age 33. Ismaila et al, Bello et al found that majority of the participants were belong to the age group of 25-44 years. [8][9][10][11][12] It indicates that this occupation demands young productive workforce. Most of the participants from the study group had low level of education. Similar findings were reported by Yusuff, et al and Wani, et al. 13,14 Majority of the sawmill workers had low socioeconomic profile and were employed in the wood cutters or helper category so they were more consistently exposed to the sawdust than the workers in administrative category like technician, clerk etc. In our study most of the workers had sawdust exposure of <5 years with a mean duration of exposure (years)= 7.59±5.89 years. These findings are consistent with the findings of Oluwatosin et al, Ugheoke et al. 15 This could be due to chronic dust exposure which impairs the phagocytic activity of alveolar macrophages and also affects the mucociliary performance. When the dust particles are inhaled, scavenger cells like macrophages dissolve the dust by surrounding it, but if there is dust overload, the macrophages fail to completely clear the dust; consequently the dust particles lodge in and irritate the respiratory mucosa. [18][19][20] Bello et al found high prevalence of injury i.e. 83%. 12 This may be due to sawmill in our study were large and most of tasks which prone to injury were mechanized so we may have found less prevalence of injury (40%) in our study. 21 Alamgir et al found significant negative trend for injuries i.e. injuries decreased as the duration of work increased. 22 Serious injuries decreases as the working duration increases as the skill increases but in our study most of the injuries were minor and occurred due to wood handling, that does not depend on skill, further skilful tasks were mechanized so we could not find any trend for injury. Mandryk et al found significant dose-response relationships for respiratory morbidity but in our study we did not find significant trend for respiratory morbidities. 23 This may be due to that respiratory disorders discussed in our study take years to manifest and most of the subjects in our study had <5 years of work exposure. CONCLUSION Musculoskeletal disorders were the commonest morbidity among sawmill workers. Respiratory morbidities and other allergic effects of wood dust were other common morbidities among sawmill workers. Cardiovascular morbidities, hearing impairment and hernia were chronic morbidities and they have shown increasing trend with the duration of work. Cutters were more prone to injuries than the sawmill workers working in the other section. Controlling the dust in occupational environment, safety education, periodic medical examination and occupational training could solve majority of these problems.	2019-03-18T13:58:51.836Z	2018-06-22T00:00:00.000	{ "year": 2018, "sha1": "b7151a81725d93adda65f444e691c5505767e155", "oa_license": null, "oa_url": "https://www.ijcmph.com/index.php/ijcmph/article/download/2679/2175", "oa_status": "GOLD", "pdf_src": "MergedPDFExtraction", "pdf_hash": "846e68b8430697587cba6cf69529d60346463177", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
11598558	pes2o/s2orc	v3-fos-license	Neutral-Lipid Analysis Reveals Elevation of Acylglycerols and Lack of Cholesterol Esters in Plasmodium falciparum-Infected Erythrocytes ABSTRACT Here we show that blood-stage Plasmodium falciparum organisms accumulate a high mass of triacylglycerol and diacylglycerol. However, we failed to detect cholesterol esters, a second neutral lipid species reported to be important for a related apicomplexan, Toxoplasma gondii. Evidence for P. falciparum and T. gondii homologues of acyl coenzyme A:diacylglycerol acyltransferase suggests that acylglycerols may be the conserved neutral lipids in apicomplexans. Plasmodium falciparum is a protozoan parasite belonging to the phylum Apicomplexa. During the blood stages of infection, its merozoite stage enters the red cell to become an intracellular ring-stage parasite. After 24 h, the ring develops through the trophozoite and multinucleate schizont stages. At the end of the intraerythrocytic cycle, as many as 16 daughter merozoites emerge to reinfect red cells and thus maintain the asexual cycle. Parasite-induced synthesis of both sphingolipids and phospholipids as well as utilization of host cholesterol is critical for intraerythrocytic proliferation of P. falciparum (6,(8)(9)(10)15). Fatty acid synthesis and host-derived cholesterol are known to be important for intracellular replication of a related apicomplexan, Toxoplasma gondii (3,4). Studies with T. gondii have also shown the presence of neutral lipid droplets and the production of cholesterol esters (CE) that can be blocked by inhibitors of mammalian acyl coenzyme A cholesterol acyltransferase (ACAT); this inhibition is detrimental to parasite growth (16). However, the presence of a second major neutral lipid class, acylglycerols (triacylglycerol [TAG] and diacylglycerol [DAG]), was not examined in T. gondii. Further, CE, acylglycerols, their stage-dependent accumulation, and the genes underlying their biosynthesis remain poorly defined for P. falciparum. Uninfected erythrocytes, as well as P. falciparum-infected erythrocytes at the ring, trophozoite, and schizont stages of growth, isolated from in vitro culture were extracted, and neutral lipids were analyzed by thin-layer chromatography (TLC) (see materials and methods information available online at http://bigbird.pathology.northwestern.edu/Nawabi/supplemen-tal_Methods). Uninfected erythrocytes contained no detectable levels of TAG. However, accumulation of high levels of TAG and DAG was detected in infected erythrocytes, with maximum levels seen at the terminal schizont stage (Fig. 1A). The identities of TAG and DAG were confirmed by twodimensional TLC analysis (data not shown). Interestingly, there was no difference between levels of CE detected in infected erythrocytes and those in their uninfected counterparts (Fig. 1A). Since uninfected erythrocyte membranes do not contain CE, these data suggest that infected cells also fail to accumulate esters. To further investigate this, we used the more sensitive Amplex red cholesterol assay (see materials and methods information online). This kit converts CE to cholesterol by the action of an esterase activity. Thus, measurement of cholesterol in samples before and after esterase treatment indicates levels of esters present in them. Using this assay, we found no change in the levels of cholesterol, in either infected erythrocytes (6.89 Ϯ 0.1 g/5 ϫ 10 6 cells) or uninfected erythrocytes (6.83 Ϯ 0.1 g/5 ϫ 10 6 cells) before and after esterase treatment. Together, these data suggest that CE are absent and acylglycerols are the major neutral lipids found in P. falciparum-infected erythrocytes. Biosynthetic studies using [ 3 H]oleic acid coupled to bovine serum albumin indicated active incorporation of radiolabel into TAG and DAG but not CE (Fig. 1B). This incorporation of radiolabel is not seen in uninfected erythrocytes, suggesting that acylglycerols are actively synthesized by P. falciparum. This may underlie the observed TAG and DAG accumulation in infected erythrocytes. In contrast, no new synthesis of CE is seen, suggesting that they are not produced upon infection. Staining infected erythrocytes with the dye Nile red (which binds lipid droplets) indicates that acylglycerols are present in large droplets within the parasite (see the figure available online at http://bigbird.pathology .northwestern.edu/Nawabi/supplemental_Figure1). Since TAG was the major acylglycerol detected, we searched the P. falciparum genome for genes that may be involved in its formation. The best-characterized pathway for TAG biosynthesis is the Kennedy pathway, in which glycerol-3-phosphate is acylated three times, the final acyl transfer being mediated by the enzyme acyl coenzyme A:DAG acyltransferase (DGAT) (7,11). We were able to identify a gene (GenBank accession no. CAB39042) with over 50% (the highest homology was within the conserved C terminus) homology to DGAT1 family members from other eukaryotes. Two very conserved regions with the consensus sequences FGDRXFY(R/K)DWWN and IERXLKL, unique to DGAT1 family members, are both present in CAB39042 (1). This consensus is not present in a second family of DGAT proteins, designated DGAT2. Moreover, the DGAT1 family is not closely related to the DGAT2 family. The DGAT1 family shows significant similarities with the ACAT1 family, but a phylogenetic tree ( Fig. 2A) constructed by the method of parsimony (see materials and methods information online) reveals that despite the high homology between DGAT1 and ACAT families, CAB39042 (PfDGAT) is more closely related to the DGAT1 family of acyltransferases. In our search we were also able to identify a putative plasmodial lecithin-cholesterol acyltransferase (LCAT) (GenBank accession no. NP_703950). LCAT is the enzyme that transfers an acyl chain from phosphatidylcholine (lecithin) to the 3-hydroxyl group of cholesterol, producing lysophosphatidylcholine and CE. The conserved C-terminal domain of NP_703950 is 90% homologous (ϳ25% identity) to the consensus domain of LCAT. Phospholipid:DAG acyltransferase (PDAT), an enzyme homologous to LCAT, is also able to synthesize TAG in plants and yeasts (5,13). Thus, it was important to establish whether NP_703950 is more homologous to PDAT or LCAT. Our percent similarity analyses revealed that NP_703950 is more similar to LCAT (17 and 16.4% similar to human and mouse enzymes) than to PDAT (11.4 and 13.2% similar to Arabidopsis and yeast enzymes). Further, phylogenetic analysis revealed closest homology within the LCAT family ( Fig. 2A). Thus, it is highly unlikely that the TAG found in P. falciparuminfected erythrocytes is synthesized by this putative LCAT. Our phylogenetic analysis makes PfDGAT an attractive candidate for the enzyme responsible for TAG biosynthesis. Thus, Northern blot analysis was undertaken (see materials and methods information online) to determine whether PfDGAT is expressed during blood-stage infection. Shown in Fig. 2B is a blot obtained by using a 900-bp fragment of PfDGAT as a probe. Low levels of transcripts were seen in rings with a marked elevation in trophozoite stages. This expression closely parallels the stage-specific accumulation of TAG seen in Fig. 1. These data strongly suggest that the plasmodial genome encodes a DGAT that is expressed and may be responsible for the high levels of TAG present in infected erythrocytes. Previous studies with T. gondii have shown that an inhibitor of mammalian ACAT, Sandoz 58-035, blocked T. gondii growth and the production of CE (16). Although we failed to detect cholesterol esterification activity, PfDGAT is highly homologous to the ACAT and DGAT families (2,12), and therefore, we treated P. falciparum-infected red cells with Sandoz 58-035 for 2 h at a concentration (10 g/ml) known to inhibit 99% of ACAT activity in mammalian cells (14,17). However, we failed to induce any change in the incorporation of [ 3 H]oleic acid into either TAG or its intermediates or show any deleterious effects on plasmodial growth (data not shown). This is consistent with the lack of CE production in P. falciparum. The absence of CE in P. falciparum-infected erythrocytes was surprising. It suggests that the putative plasmodial LCAT is not functional in blood stages of the parasite. Despite the Our studies show that TAG is a major lipid species stored in lipid droplets in the late trophozoite and schizont stages of P. falciparum. It may be utilized to store acyl groups for phospholipid synthesis, glycosyl phosphatidyl inositol synthesis, and possibly for beta-oxidation (although evidence for the complete pathway in P. falciparum is still not available). Sonda et al. suggested that two expressed sequence tags may reflect the presence of ACAT homologues in T. gondii (16). However, with a higher level of completion of the T. gondii genome, we can now ascribe both to a single open reading frame (unique identifier, TGG_3996; toxodb.org). A BLAST search using human ACAT1 or DGAT1 as the query against the T. gondii database revealed that TGG_3996 is more homologous to human DGAT1 (3.6e Ϫ15 ) than to human ACAT1 (9.9e Ϫ6 ). Further, using this open reading frame as a query in BLAST resulted in hits for DGAT1 homologues but not ACAT homologues. This suggests that TAG may also be present in T. gondii. A separate T. gondii ACAT homologue underlying cholesterol esterification has yet to be identified. Given that TAG is produced at such high levels in P. falciparum, an organism that causes human malaria, it is tempting to speculate that blocking its synthesis may be detrimental to infection. We thank Yvonne Lange (Rush Medical Center, Chicago, Ill.) for the generous gift of the mammalian ACAT inhibitor Sandoz 58-035 and for her critical reading of the manuscript. This work was supported by American Heart Association predoctoral fellowship 0215249Z (to P.N.) and NIH grant AI39075 (to K.H.)	2018-04-03T00:22:54.812Z	2003-10-01T00:00:00.000	{ "year": 2003, "sha1": "157dcae79b7610a9e5736633b60d6f5b94b5d1be", "oa_license": null, "oa_url": "https://ec.asm.org/content/2/5/1128.full.pdf", "oa_status": "GOLD", "pdf_src": "Highwire", "pdf_hash": "8f4d021d8158fbfe7d5f0a23f7920ce41bb17a51", "s2fieldsofstudy": [ "Biology", "Chemistry" ], "extfieldsofstudy": [ "Biology", "Medicine" ] }
207462740	pes2o/s2orc	v3-fos-license	Thyrocervical artery - jugular fistula following internal jugular venous catheterization Arteriovenous fistula (AVF) is an anomalous communication between an artery and a vein, caused by an iatrogenic or traumatic etiology. Surgically created upper limb AVF remains the preferred vascular access for patients on maintenance hemodialysis. Nonetheless central vein cannulation for hemodialysis is a common procedure done in patients who need hemodialysis. We incidentally detected a thyrocervical artery - jugular fistula in a patient on maintenance hemodialysis. He underwent a successful intra arterial coil embolization of the feeding vessel. Review of literature has shown that, a thyrocervical artery - internal jugular vein arteriovenous fistula following a central venous catheterization has not been reported so far. We report, a case of an iatrogenic fistula between the thyrocervical branch of subclavian artery and the right internal jugular vein following a failed attempt at internal jugular venous catheterization for HD, which was incidentally detected after a period of 1 year.To the best of our knowledge a thyrocervical artery; internal jugular vein arteriovenous fistula following central venous catheterization has not been described in the literature so far. Case Report A 32-year-old male was diagnosed to have immunoglobulin a (IgA) nephropathy and systemic hypertension in 2010.He progressed to end-stage renal disease within 2 years and was initiated on HD at another center through a non-tunneled uncuffed left internal jugular catheter.He gave history of an attempted right internal jugular catheterization for HD 1 year ago at another center, but the procedure was unsuccessful.He had been undergoing intermittent HD through left radiocephalic fistula for the last 1 year.He was admitted in our center for pretransplant evaluation.Clinically he appeared poorly Introduction Placement of an uncuffed double lumen catheter for initiating hemodialysis (HD) is a common interventional procedure performed by a nephrologist.Right internal jugular vein is the preferred initial access site for catheter placement using the Seldinger technique.It is thought to be safer than the subclavian or the femoral vein sites.Although central venous catheterization is frequently performed, various complications of central venous catheter insertion have been reported.Arteriovenous fistula following central venous catheterization is a very rare complication, [1] with only isolated case reports of carotid jugular fistula.Ultrasound-guided central venous catheterization could reduce the complication rate and increase success rate.wave pattern.An angiogram was done which showed an abnormal connection between the thyrocervical branch of right subclavian artery and right internal jugular vein, resulting in dilatation of the internal jugular vein [Figure 1].Feeder artery was embolized using multiple 035 and 018 coils [Figure 2].Check angiogram after the procedure showed collapse of the fistulous track [Figure 3] and follow-up angiogram after 48 h shows that the arteriovenous fistula has been completely closed. Discussion Although arteriovenous fistula is the vascular access of choice for HD, central venous catheters are commonly used in patients with renal failure requiring dialysis.The preferred choice of access for HD would be right internal jugular vein, femoral veins, left internal jugular vein, and finally subclavian vein in that order. Central venous catheterization has a complication rate ranging from 5 to 19%. [2]Complications of central vein catheterization can be classified into mechanical, infectious, and thrombotic complications.Mechanical complications include arterial puncture, hematoma, pneumothorax, hemothorax, cardiac arrhythmia, and malpositioned catheter.The most common predictor of complication is an unsuccessful insertion attempt, with complications rates as high as 28%. [3]Vascular injuries during central vein catheter insertion include a wide spectrum of complications, with arterial puncture being the most common.Even though arterial puncture is generally self-limiting, it can lead to substantial morbidity or death.Forty percent of carotid punctures are associated with a hematoma. [4]aditionally, the site of the needle insertion for central venous catheter insertion is determined by the landmark method; by palpating the structures in relationship to the concerned vein.The use of ultrasound for central venous catheterization increases success rate, while simultaneously decreasing procedural time and complication rate.Central vein catheter placement under ultrasound guidance reduces the risk of carotid puncture (7-10%) to virtually zero. [5]Therefore, it should be employed especially in patients who are obese with short neck, those with coagulopathy or on anticoagulant therapy, severe hypovolemia, respiratory distress, and with altered regional anatomy.An iatrogenic carotid jugular fistula is an extremely rare complication.In the literature, the occurrence of carotid jugular arteriovenous fistula has been reported as a rare complication following cannulation of the internal jugular vein. [6]Long standing arteriovenous fistula causes venous engorgement of the collateral veins producing a continuous machine like murmur, diminished distal arterial pulse, and ipsilateral limb edema of the extremity; and if left untreated, the fistula may lead to high output heart failure, arrhythmias, or thromboembolic episodes. Our patient developed a fistula between the thyrocervical branch of subclavian artery and the right internal jugular vein following the failed internal jugular venous catheterization done 1 year ago and clinically he had features of high output cardiac failure.Although color Doppler, computed tomography (CT), or magnetic resonance imaging (MRI) can identify the arteriovenous fistula, in most cases conventional angiography is still required for accurate localization of lesion and tailoring of the surgical or endovascular treatment. [7]Surgical options include partial resection, ligation, and primary repair.Surgical corrections are more common and less expensive in the treatment of arteriovenous fistulas, but carry high perisurgical mortality and morbidity rates in patients on HD or with multisystem problems.Therefore, endovascular procedures are becoming widely popular in the recent years due to the development of new endovascular techniques. Stent placement [8] or transcutaneous intravascular embolization [9] with metallic coils, covered stent implants, detachable balloons, or (N-butyl-2-cyanoacrylate) have been utilized frequently for endovascular treatment of arteriovenous fistulas.Endovascular repair of such injuries is a safe and reasonable treatment option with lower morbidity and mortality when compared to open surgical approach.Our patient successfully underwent intra-arterial coil embolization of the feeding vessel. Conclusion Iatrogenic arteriovenous fistula is a rare complication of internal jugular vein catheterization.Considering the long-term harmful effects of an arteriovenous fistula, prompt measures are to be taken to close the fistula.Intra-arterial coil embolization is a simple and safe option for management, as compared to surgical methods.The use of Doppler-guided venous catheterization is likely to reduce the mechanical complications of central vein catheterization for HD. Figure 1 : Figure 1: (a) Graphical representation of the normal anatomy of the subcalvian artery and its branches.(b) Shows the patient's angiogram with the abnormal arteriovenous fistula (curved arrow) between the thyrocervical branch of subclavian artery and the internal jugular vein b a	2018-04-03T02:23:40.293Z	2014-05-01T00:00:00.000	{ "year": 2014, "sha1": "a5dea589f9934a5ae38429367d292504a7f6b5b1", "oa_license": "CCBYNCSA", "oa_url": "https://doi.org/10.4103/0971-4065.132019", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "3326c8283fbc4458c9305b14e7250832d694f2dd", "s2fieldsofstudy": [ "Medicine", "Biology" ], "extfieldsofstudy": [ "Medicine" ] }
228638451	pes2o/s2orc	v3-fos-license	Brazilian guidelines for the management of brain-dead potential organ donors. The task force of the AMIB, ABTO, BRICNet, and the General Coordination of the National Transplant System Objective To contribute to updating the recommendations for brain-dead potential organ donor management. Method A group of 27 experts, including intensivists, transplant coordinators, transplant surgeons, and epidemiologists, joined a task force formed by the General Coordination Office of the National Transplant System/Brazilian Ministry of Health (CGSNT-MS), the Brazilian Association of Intensive Care Medicine (AMIB), the Brazilian Association of Organ Transplantation (ABTO), and the Brazilian Research in Intensive Care Network (BRICNet). The questions were developed within the scope of the 2011 Brazilian Guidelines for Management of Adult Potential Multiple-Organ Deceased Donors. The topics were divided into mechanical ventilation, hemodynamic support, endocrine-metabolic management, infection, body temperature, blood transfusion, and use of checklists. The outcomes considered for decision-making were cardiac arrest, number of organs recovered or transplanted per donor, and graft function/survival. Rapid systematic reviews were conducted, and the quality of evidence of the recommendations was assessed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system. Two expert panels were held in November 2016 and February 2017 to classify the recommendations. A systematic review update was performed in June 2020, and the recommendations were reviewed through a Delphi process with the panelists between June and July 2020. Results A total of 19 recommendations were drawn from the expert panel. Of these, 7 were classified as strong (lung-protective ventilation strategy, vasopressors and combining arginine vasopressin to control blood pressure, antidiuretic hormones to control polyuria, serum potassium and magnesium control, and antibiotic use), 11 as weak (alveolar recruitment maneuvers, low-dose dopamine, low-dose corticosteroids, thyroid hormones, glycemic and serum sodium control, nutritional support, body temperature control or hypothermia, red blood cell transfusion, and goal-directed protocols), and 1 was considered a good clinical practice (volemic expansion). Conclusion Despite the agreement among panel members on most recommendations, the grade of recommendation was mostly weak. The observed lack of robust evidence on the topic highlights the importance of the present guideline to improve the management of brain-dead potential organ donors. Introduction Organ donation for transplantation is a complex process led by several health care professionals responsible for a sequence of actions and procedures that begin with identifying a potential organ donor and end with organ procurement surgery and distribution. The progress of this process is essential to increase the deceased-donor pool, and to decrease the growing disparity between the number of patients on transplant waiting lists and the availability of organs [1,2]. The organ donation process includes the identification of the potential donor, diagnosis of brain death, family support and interview, evaluation of donor eligibility criteria, clinical management of the potential organ donor, and organ procurement and distribution [2,3]. Given the marked clinical instability that occurs in patients who progress to brain death, the application of potential donor-management strategies aiming at hemodynamic stabilization is crucial to avoid loss of organs due to hypoperfusion or loss of donors due to cardiac arrest. Also, the control of ventilatory support, body temperature, and endocrine-metabolic functions contributes to improving the quality of organs and clinical outcomes in transplant recipients [1,2,4,5]. Despite the lack of evidence on some aspects of the clinical management of potential organ donors, the recommendations presented in this guideline intend to promote a general approach to mitigate the disparity between supply and demand of organs for transplantation. Objective To provide recommendations to guide the clinical management of brain-dead potential organ donors aiming to reduce the rate of cardiac arrest of the potential donor and to improve organ viability for transplantation. Method The present document provides a partial update on the 2011 Brazilian Guidelines for Management of Adult Potential Multiple-Organ Deceased Donors [6][7][8]. The working group consisted of physicians, nurses, pharmacists, physical therapists, epidemiologists, methodologists, and transplant system managers. The contributions of each participant are shown in Additional file 1, and the respective conflict-of-interest disclosures are shown in Additional file 2. The target audience of this guideline is health care professionals, especially physicians and nursing staff working in adult ICUs and emergency departments, who are involved in the care of adult individuals with known or suspected brain death. The clinical issues addressed by the guideline were defined by coordinators of the working group and the methodologists in a face-to-face meeting held in March 2016, after reviewing the recommendations of the 2011 Brazilian Guidelines for Management of Adult Potential Multiple-Organ Deceased Donors [6][7][8]. The issues were prioritized according to the perception of their impact on medical management and variability in clinical practice and divided into the following major topics: (1) ventilatory support; (2) hemodynamic support; (3) endocrine, metabolic and nutritional management; (4) specific aspects that include infection and sepsis, red blood cell transfusion, and body temperature control; and (5) goaldirected therapy. For each clinical issue, operational questions were developed and framed using the PICO (population-intervention-comparison-outcome) format. The population of interest consists of potential organ donors with known or suspected brain death [3], hereafter referred to as potential donors. The outcomes considered for decision-making were cardiac arrest, the number of organs recovered or transplanted per donor, and graft function or graft survival. For each clinical issue, rapid systematic reviews [9,10] were conducted using the following search strategy: (1) Review of the reference lists of Brazilian guidelines [6][7][8] and the Society of Critical Care Medicine (SCCM) [11] statement on the management of the potential organ donor; (2) Review of related topics in the DynaMed and UpToDate databases; and (3) PubMed search focusing on systematic reviews and clinical trials published until October 2016 and until January 2017. Quality of evidence was assessed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system [12]. The recommendations were prepared and submitted to two face-to-face expert panels held in November 2016, and February 2017. For each recommendation, the direction of the course of action was discussed (whether to perform or not to perform the proposed action), and the strength of the recommendation was classified as strong or weak according to the GRADE system [12]. After the last panel meeting, a new systematic search covering the period from October 2016 to May 2020 was carried out to identify new evidence that could potentially modify the recommendations. From June to July 2020, a Delphi process was performed with the panelists to present the results of the literature update and review the direction and strength of the recommendations. Results A total of 19 recommendations were drawn from the expert panel. Of these, 7 were classified as strong, 11 as weak, and 1 was considered as good clinical practice. Table 1 shows a summary of the recommendations. Ventilatory support 1. We recommend using a lung-protective ventilation strategy in all potential donors (low level of evidence, strong recommendation). Summary of evidence In potential donors, an initially normal or near-normal lung function (PaO 2 /FiO 2 ≥ 300) may deteriorate due to common complications in critical patients, such as pulmonary contusion, lung injury following blood transfusion, pneumonia, atelectasis, and mechanical ventilation-related iatrogenic injuries [13][14][15][16][17][18]. In addition, approximately 30-45% of potential donors develop acute respiratory distress syndrome (ARDS; PaO 2 /FiO 2 < 300), and only 15-20% of the lungs are suitable for transplantation at the end of the procurement process [13,15,17]. The lung-protective ventilation strategy in potential donors with normal lungs and the apnea testing performed with continuous positive airway pressure (CPAP) have been associated with an increase in eligibility for lung donation [18][19][20]. Remarks The protective ventilation strategy for healthy lungs consists of the combination of a tidal volume of 6-8 mL/kg and PEEP of 8-10-cm H 2 O. To promote adequate blood oxygenation, FiO 2 and PEEP must be adjusted to obtain a SaO 2 > 90%. To avoid atelectasis, the apnea test with 10 cm H 2 O CPAP can be performed using a closed-circuit system in potential donors with preserved lungs who are candidates for lung procurement, or even when hypoxemic respiratory failure is present. Also, the same procedure can be considered on those who have failed the test due to hypoxemia after disconnection. 2. We suggest not using alveolar recruitment maneuvers routinely in potential donors (very low level of evidence, weak recommendation). Summary of evidence Although alveolar recruitment maneuvers have been suggested for the ventilatory management of organ donors with lung injury (PaO 2 / FiO 2 < 300) [13-16, 18, 20], and these maneuvers could reduce hypoxemia after apnea testing, contributing to increasing the viability of pulmonary grafts [14][15][16][17][18]20], a randomized clinical trial showed unfavorable outcomes in critically ill patients [21]. Besides, no randomized studies have demonstrated their efficacy in the population of potential donors. Remarks Performing alveolar recruitment maneuvers in hemodynamically stable potential donors is probably feasible in units with experience in the management of ARDS. In cases of hypoxemia refractory to the lung-protective ventilation strategy, however, alveolar recruitment maneuvers should not be performed routinely. Their use is not indicated in hemodynamically unstable potential donors. Hemodynamic support Volemic expansion and vasopressors 3. We recommend performing initial volemic expansion in hemodynamically unstable potential donors with hypovolemia or responsive to fluids according to fluid responsiveness assessment (good clinical practice). 4. We recommend administering norepinephrine or dopamine to control blood pressure in potential donors who remain hypotensive after volemic expansion (very low level of evidence, strong recommendation). Summary of evidence Potential donor hypotension is associated with a higher incidence of postoperative liver graft dysfunction and longer hospital stay in liver transplant recipients [22,23]. Targeting a mean arterial pressure (MAP) ≥ 65 mm Hg has also been associated with reduced occurrence of cardiac arrest in potential donors [22,24]. Intravascular volume expansion guided by ventricular filling pressures or respiratory pulse pressure variation (PPV) in hemodynamically unstable potential donors is associated with faster recovery of renal graft function and reduced circulating levels of inflammatory cytokines [22,25]. A randomized trial detected no difference between usual fluid management or fluid management directed by a PPV and cardiac index. On the other hand, there was a trend toward an increase in the number of organs transplanted per donor among unstable potential donors responsive to fluids (p = 0.059) [26]. Conversely, avoiding fluid overload after the initial volume resuscitation to stabilize blood pressure seems to be beneficial. This approach is associated with a greater number of organs transplanted per donor and a greater number of lungs transplanted without reducing the number of other donated organs or impairing survival in the heart, liver, pancreas, or kidney transplant recipients [19,[27][28][29]. If hypotension persists after adequate volume resuscitation, adrenergic vasopressors should be used to achieve adequate blood-pressure levels [30]. There is no difference in clinical outcomes in studies comparing norepinephrine and dopamine [31][32][33]. Disruption of vagal activity secondary to brain death may result in atropinerefractory bradycardia. In these cases, adrenergic drugs as isoproterenol, epinephrine, and dopamine have been suggested as positive chronotropic agents to treat bradycardia in potential donors. Considering the predominance of noradrenaline action on alpha-1 receptors, its infusion usually occurs without significant increase in heart rate. Hence, dopamine or epinephrine may be more convenient for the treatment of hypotension due to a positive chronotropic effect [6,34,35]. Remarks Obtaining an MAP ≥ 65 mm Hg as a bloodpressure target contributes to the perfusion of organs that are intended to be preserved for transplantation [22][23][24]. Hypovolemia is very frequent in potential organ donors and should be considered when hypotension is present. The initial infusion of crystalloids (e.g., 30 mL/ kg) in potential donors who are hypovolemic or responsive to fluids (when any fluid responsiveness assessment parameter is already available) contributes to blood-pressure control by improving tissue perfusion [24][25][26]. If the blood-pressure target is not achieved with the initial volume expansion, norepinephrine or dopamine infusion should be started immediately. The use of dopamine is likely advantageous for cases of bradycardia with signs of low cardiac output [6,34,35], but the arrhythmogenic potential of dopamine should be considered [39]. 5. We suggest not using low-dose dopamine for renal protection in potential donors (very low level of evidence, weak recommendation). Summary of evidence A cohort study of 93 heart transplant recipients showed that pretreatment with low-dose dopamine (4 μg/kg/min) in heart donors was associated with higher graft survival 3 years after transplantation (87.0 vs. 67.8%, p < 0.03) [40]. A randomized-controlled trial of 264 organ donors reported that the administration of low-dose dopamine reduced the need for hemodialysis in recipients (OR 0.54; 95% CI 0.35-0.83), but with no benefits for kidney graft survival after 3 years [41]. In the 5-year follow-up analysis of 487 renal transplant recipients from the same trial, the researchers failed to show a significant advantage of dopamine administration in potential donors to long-term kidney graft survival, although time of dopamine infusion and graft failure were exposure-related (HR 0.96; 95% CI 0.92-1.00, per hour) [42]. The same group reported that low-dose dopamine did not negatively affect the short-or long-term outcomes after liver transplants [43]. Remarks Although the administration of low-dose dopamine in potential donors reduces the need for multiple dialysis sessions, the long-term benefits for heart and kidney graft survival are unclear. The panel considered the potential arrhythmogenic effect of dopamine, which may imply a greater risk of loss of potential donors due to cardiac arrest before organ procurement. Endocrine and electrolyte management Hormones 6. We recommend combining arginine vasopressin (AVP) in potential donors receiving norepinephrine or dopamine to control blood pressure (low level of evidence, strong recommendation). Summary of evidence The use of AVP in brain-dead potential donors contributes to reducing the need for adrenergic vasopressors and is associated with a lower incidence of cardiovascular deterioration and cardiac arrest [44][45][46][47][48], in addition to contributing to the control of plasma hyperosmolarity [46]. AVP infusion allows, in some cases, complete discontinuation of adrenergic vasopressors without causing adverse effects on the function of organs transplanted [48,49]. Finally, AVP infusion seems to be associated with a greater number of donated organs and a lower rate of graft refusal due to organ dysfunction [45]. Remarks The administration of an initial 1 IU AVP bolus followed by infusion of 0.5 IU/h to 2.4 UI/h helps to maintain blood pressure in potential donors requiring vasopressors, and contributes to the control of polyuria and normovolemia in the presence of diabetes insipidus [44-46, 48, 49]. AVP should be started at the same time of adrenergic vasopressor infusion. 7. We recommend administering AVP or 1-deamino-8-d-arginine vasopressin (DDAVP) to control polyuria in potential donors with diabetes insipidus (low level of evidence, strong recommendation). Summary of evidence The analysis of the database of a randomized clinical trial that evaluated 487 renal graft recipients showed better control of daily urine output (p < 0.001) and a lower need for fluids in the DDAVP group (p < 0.001). DDAVP was associated with improved renal graft survival (85.4% vs. 73.6%, p = 0.003) after 2 years, with no differences in acute rejections (OR 1.32; 95% CI 0.70-2.49) or delayed graft function (OR 0.97; 95% CI 0.57-1.65) [50]. Remarks DDAVP acts exclusively on V2 receptors and is indicated to control polyuria (urine output > 4 mL/ kg/h) in potential donors with diabetes insipidus who maintain adequate blood pressure without adrenergic vasopressors. AVP is preferred to control polyuria in potential donors with diabetes insipidus who need adrenergic vasopressors. The combination of AVP and DDAVP may be considered in refractory cases [51]. Although the intranasal route is feasible, DDAVP should preferably be administered intravenously, at a dose of 1-2 µg every 2-4 h [8,13,15], until a urine output < 4 mL/kg/h has been achieved [50][51][52][53]. 8. We suggest using low-dose corticosteroids in potential donors receiving norepinephrine or dopamine to control blood pressure (low level of evidence, weak recommendation). Summary of evidence A small retrospective study reported that administration of 15-mg/kg methylprednisolone was associated with higher PaO 2 /FiO 2 values (p = 0.01) and a greater number of lungs transplanted (p < 0.01) [54]. Conversely, a before-and-after study comparing 15-mg/kg methylprednisolone with 300-mg hydrocortisone found no difference in the oxygenation and hemodynamic stability of the potential donor or in the number of organs transplanted [55]. A recent small randomized-controlled trial showed that a single dose of 15 mg/kg/day of methylprednisolone administered to the potential organ donor may negatively affect the graft function by increasing the antigenicity of the kidneys before transplantation. This negative effect was not noticed among brain-dead donors who received 15 mg/ kg/day of methylprednisolone followed by 100 mg every 2 h until organ harvesting [56]. Eleven randomizedcontrolled trials analyzed in a systematic review did not support the use of high-dose corticosteroids in the management of potential donors [57]. On the other hand, a randomized multicenter cluster study including 259 individuals compared the administration of low-dose hydrocortisone (300 mg/day) with no corticosteroids. The doses (p = 0.03) and duration of infusion (p < 0.001) of vasopressors were lower in the intervention group, and the complete vasopressor withdrawal was 4.7 times more frequent in the corticosteroid group [58]. Remarks Despite conflicting evidence, the use of corticosteroids is of low cost and a low risk to potential donors and may have a positive effect on hemodynamic outcomes; therefore, their use is indicated in these patients. Current evidence does not suggest ventilatory or hemodynamic benefits associated with corticosteroid therapy at high doses compared with low doses (i.e., 100 mg every 8 h). Higher doses should be avoided. 9. We suggest not using thyroid hormones routinely in potential donors (very low level of evidence, weak recommendation). Summary of evidence Administration of thyroid hormones in potential donors did not add any benefit, such as a reduction in vasopressor use, an increase in cardiac index, or an increase in organ procurement for transplantation [59][60][61][62][63][64][65]. Observational studies had suggested an increase in heart procurement, which was not confirmed in randomized clinical trials [66,67], even in brain-dead organ donors with hemodynamic instability and/or impaired cardiac function [68,69]. Remarks Brain death is associated with a drop in circulating thyroid hormone levels, which could contribute to hemodynamic instability; however, there is no evidence to support the use of thyroid hormones in potential donors, given their costs and risks. 10. We suggest performing glycemic control in potential donors (very low level of evidence, weak recommendation). Summary of evidence Four observational studies evaluated the effect of potential donor hyperglycemia on post-transplant pancreatic function [70][71][72][73]. One study showed a correlation between donor blood glucose immediately before organ retrieval and HbA1C 1 year after transplantation [73], and another study found an association between hyperglycemia and graft loss (HR 1.4; p = 0.03) [74]. Two studies showed no association between potential donor blood glucose and post-transplant pancreatic graft function [70][71][72]. One observational study found an association between glycemic control and creatinine of the potential donor before organ retrieval [75]. Conversely, there is no evidence that hyperglycemia is associated with liver graft dysfunction [76]. A study of 1611 potential donors reported that a glucose level < 180 mg/dL was an independent predictor of four or more organs transplanted per donor (OR 1.35; 95% CI 1.01-1.82) [77]. A set of potential donor care measures, including glycemic control, was associated with achieving ≥ 3 organs transplanted per donor (OR 1.9; 95% CI 1.35-2.68), but it was not possible to assess the isolated effect of glycemic control [78]. Remarks Very-low-quality evidence suggests that a glucose level < 180 mg/dL is associated with a greater number of organs transplanted. Blood glucose should be monitored in all potential donors at least every 6 h, targeting levels of 140-180 mg/dL, and intravenous insulin infusion can be used to this end. Westphal et al. Ann. Intensive Care (2020) 10:169 Electrolytes 11. We suggest maintaining serum sodium levels below 155 mEq/dL in potential donors (very low level of evidence, weak recommendation). Summary of evidence. Five descriptive observational studies were identified (n = 5733), which evaluated only graft viability/function. In four of these studies (n = 5545), there was no negative effect of donor hypernatremia above 155 mEq/L on liver or heart graft function [79][80][81][82]. In only one study (n = 188), hypernatremia was associated with more cases of early graft loss [83]. Some authors have suggested that deceased-donor hypernatremia may be a factor for worse prognosis of graft function, but these findings have not been universally confirmed [79][80][81][82][83][84][85]. Changes in natremia may reflect inadequate volume management, especially in the presence of diabetes insipidus, one of the reasons for its correction [11]. Remarks Hypernatremia is often associated with hypovolemia, and should be controlled with volume expansion, replacement of hypotonic fluids, and control of polyuria with AVP or DDAVP. Serum sodium should be monitored, targeting levels < 155 mg/dL. 12. We recommend maintaining serum potassium levels between 3.5 and 5.5 mEq/L in potential donors (very low level of evidence, strong recommendation). Summary of evidence There are no studies that directly evaluate the effect of hyper-or hypokalemia in potential donors. A comparison of potassium levels in ICU patients showed that hyperkalemia was more common in patients who died (9.2% vs. 0.9%, p < 0.001) and that serum potassium concentration could be a predictor of death in critically ill patients [86]. Remarks Despite the absence of studies directly evaluating the effects of potential donor serum potassium levels, potassium is a determining factor in the resting potential of electrically sensitive cells. Changes in potassium levels are related to cardiac arrhythmias and may compromise the management of potential donors. Potassium levels should be monitored, and usual correction measures should be implemented, targeting serum levels between 3.5 and 5.5 mEq/L. 13. We recommend maintaining serum magnesium levels above 1.6 mEq/L in potential donors (very low level of evidence, strong recommendation). Summary of evidence Studies on the influence of serum magnesium levels were found in critically ill patients, but none in potential donors [87][88][89][90][91][92]. Two observational studies and one randomized study identified an association between hypomagnesemia and higher mortality in critically ill patients [87,88,91], in addition to a greater likelihood of QT interval prolongation (OR 42.8; 95% CI 14.5-126.2) [88]. This association of hypomagnesemia with mortality was reinforced in a systematic review [89]. In addition to being arrhythmogenic, hypomagnesemia appears to be associated with nonrecovery of renal function in patients with acute kidney injury (70% vs. 31%, p = 0.003) [92]. Remarks Hypomagnesemia is associated with cardiac arrhythmias and worse prognosis in critically ill patients, with no direct evidence in brain-dead potential donors. However, this is a low-cost procedure, and in the ICU setting, routine monitoring until normalization of magnesium levels is a common practice, which may be beneficial for potential donors. Magnesium levels should be monitored, and magnesium sulfate should be administered, as usual, targeting serum levels above 1.6 mEq/L. Other aspects of potential donor management Nutritional support 14. We suggest maintaining nutritional support in potential donors if well tolerated (very low level of evidence, weak recommendation). Summary of evidence Although there is no evidence on nutritional support, different guidelines recommend continuing nutritional support of the donor in the absence of contraindications [7,9,51]. Possible benefits include increased liver glycogen reserves, which could positively influence the liver graft [93,94], and maintenance of intestinal mucosal trophism, which could reduce the potential for bacterial translocation. Remarks For brain-dead individuals requiring ICU management for prolonged periods (e.g., brain-dead pregnant women; prolongation of the diagnostic process or the family decision for donation), it is reasonable that energy expenditure should be estimated or measured [95], considering that baseline energy expenditure is 15-30% lower in brain-dead individuals than in other critically ill patients [96]. Thus, in individuals already receiving full nutritional support, energy intake may be reduced once brain death is established. A minimum energy intake (e.g., 500 kcal) could be considered in potential donors who had not been on enteral feeding before brain death was diagnosed, taking into account its potential benefit in the maintenance of intestinal mucosal trophism. However, it does not seem appropriate to start enteral feeding when the organs are likely to be harvested within a short period or in the presence of any of the usual contraindications to initiate/maintain enteral feeding (e.g., gastrointestinal tract obstruction, ileus, vomiting/aspiration of gastric contents, severe hemodynamic instability, and high doses of vasopressors). 15. We recommend using antibiotics in potential donors with infection or sepsis (low level of evidence, strong recommendation). Infection and sepsis Summary of evidence Different observational studies evaluated the transmission of bacterial infection in organ donors with culture-proven infection. The most commonly observed microorganisms were Staphylococcus aureus, Streptococcus sp., Klebsiella sp., and Acinetobacter baumannii. Bacterial transmission is rarely observed [97][98][99][100][101][102][103][104], provided that donors with evidence of infection receive appropriate antibiotic therapy [97-102, 105, 106]. The duration of donor antibiotic therapy ranged from 24 to 96 h in different studies [97,99,102,105]. Also, different authors have reported maintaining the same antibiotics administered to the donors in the transplant recipients, for periods ranging from 7 to 14 days [98,100,105,106]. The presence of donor infection had no impact on the survival of grafts or transplant recipients [97-102, 105, 106]. Remarks The risk of transmission of bacterial infection from organ donors to recipients is low, and donor infection does not appear to negatively affect outcomes. The risks are lower with appropriate antibiotic therapy in the donor for at least 24 h, followed by maintenance of the antibiotic in the recipient for 7-14 days [97-102, 105, 106]. Some donors have subclinical bacteremia at the time of organ procurement; therefore, cultures should be collected from blood and different sites in all donors, and the recipient antibiotic therapy should be directed by the results of culture [99,[107][108][109][110]. Body temperature control 16. We suggest maintaining body temperature above 35 °C in hemodynamically unstable potential donors (very low level of evidence, weak recommendation). 17. We suggest inducing moderate hypothermia (34-35 °C) in potential donors without hemodynamic instability (low level of evidence, weak recommendation). Summary of evidence Delayed renal graft function was evaluated in a randomized-controlled trial that compared hypothermia (34-35 °C) versus usual management (36.5-37.5 °C) in 370 potential donors without hemodynamic instability. The main result was a reduction in delayed renal graft function among recipients (OR 0.62; 95% CI 0.43-0.92). There was no difference in the number of organs transplanted per donor, adverse events, or cardiac arrest [111]. Two retrospective cohort studies nested in the randomized dopamine trial demonstrated that spontaneous donor hypothermia was associated with lower creatinine levels before organ procurement without effect on kidney graft survival [112], and with an unfavorable clinical course after heart transplant [113]. In a clinical population of post-cardiac arrest patients, i.e., patients at increased risk of hemodynamic instability, a meta-analysis of five clinical trials found a higher risk of recurrent arrest in patients with induced hypothermia (< 35 °C) in prehospital management (RR 1.23; 95% CI 1.02-1.48) [114]. Remarks Hypothermia is a low-cost intervention [115] associated with better renal graft function, but it can increase the risk of cardiac arrest in the potential donor [111,114]. The risk appears to be low in hemodynamically stable potential donors, in whom the use of hypothermia can be justified by improved graft viability. In the presence of hemodynamic instability [111], normothermia (> 35 °C) should be maintained in potential donors to reduce the risk of cardiac arrest [114]. Induction of moderate hypothermia (34-35 °C) is considered a simple (application of ice packs) and inexpensive approach, but it is important to monitor core temperature, which is not available in all ICUs. 18. We suggest transfusing packed red blood cells in potential donors with hemoglobin levels < 7 g/dL (very low level of evidence, weak recommendation). Summary of evidence The systematic literature search identified 1 descriptive observational study that evaluated function in 1884 renal grafts from 1006 brain-dead donors. Among donors, 52% received blood transfusion. Renal grafts from transfused donors had a lower rate of delayed graft function than those from nontransfused donors (26% vs. 34%, p < 0.001). The criteria defining the need for blood transfusion were not identified [116]. Remarks Anemia can compromise oxygen transport and delivery to the organs that are intended to be preserved for transplantation. However, we are unaware of the hemoglobin concentration necessary to contribute to adequate oxygen transport and delivery in potential donors. Considering the high cost and frequent shortage of blood products for transfusion, the decision to transfuse should not differ from the usual practice in other critically ill patients. 19. We suggest using a goal-directed protocol during the management of potential donors (very low level of evidence, weak recommendation). Summary of evidence Although there is no consistent evidence about an individual treatment that will improve the number and quality of donated organs [117], observational studies have reported that combining different treatments through the application of a potential donormanagement protocol is associated with a higher organ yield for transplantation [24,78,[118][119][120][121][122][123][124], lower incidence of delayed renal graft function [111], greater eligibility for lung donation [19,28], and lower incidence of donor losses due to cardiac arrest [19,24,28,119,120]. In general, the outcomes are associated with the number of goals achieved during potential donor management, including ventilatory, hemodynamic, and endocrinemetabolic management goals [24,78,[121][122][123]. In seven studies, the use of a checklist helped implement the goaldirected protocols and may have positively influenced the results [19,28,78,121,[124][125][126]. Remarks The application of a potential donor-management protocol guided by a clinical goal-directed checklist may contribute to increasing the number of organs transplanted per donor, influence graft function, and reduce donor losses due to cardiac arrest. General considerations and future directions The present guideline aimed to provide parameters to optimize the clinical management of potential donors based on the available evidence, aiming to improve the quality of organs for transplantation and to reduce donor losses. However, it is well known that it may take years for a large-scale translation of the best scientific evidence into effective practice. Thus, establishing clinical protocols can help to reduce the time required to incorporate best practices. The use of a goal-directed checklist can play an important role in the adjustment of approaches and adherence to the best evidence in complex procedures [127][128][129][130]. This guideline evaluated a broad volume of treatments and we performed rigorous PICO-driven research to provide the recommendations based on standardized rapid review methods [9,10]. Potential limitations are the low or very low certainty in the evidence identified for many of the questions, and indirect evidence that did not change after the systematic review update. However, management recommendations are consistent with similar documents recently published [11,131,132]. Several challenges regarding ethical, infrastructure, and operational issues are faced while planning and conducting studies that involve potential organ donors, which results in few randomized clinical trials [133]. The scarcity of studies with such methodological strength implies uncertainties about some interventions such as low-dose dopamine and moderate hypothermia, which, despite appearing to be related to renal graft benefit, may result in cardiac arrhythmias and hemodynamic instability. In this context, developing clinical trials in this field of medical knowledge may be helpful to understand some important aspects in the management of the potential organ donor. Some observational studies have reported that the application of a checklist to guide the management of brain-dead potential donors may help to reduce the rate of cardiac arrest in potential donors and increase the number of organs recovered per donor [24,78,119,121,122,124,126,134,135]. In this context, we used the main recommendations of the present guideline to develop an evidence-based clinical goal-directed checklist (Additional file 3) with the purpose of providing transplant coordinators and ICU professionals with essential information to optimize the care of potential donors. However, because the available studies highlighting the role of potential donor-management checklists are observational, there is insufficient evidence to support the systematic use of checklists in the management of potential donors. Therefore, we proposed the Donation Network to Optimize Organ Recovery Study (DONORS; NCT03179020), which is a parallel cluster randomizedcontrolled multicenter trial that aims to test the effectiveness of the implementation of a checklist containing goals and recommendations of care in reducing organ donor losses due to cardiac arrest and increasing the number of organs recovered per donor [136]. The implementation of the checklist should be preceded by the appropriate training of intensive care teams and transplant coordinators. We suggest applying the checklist at the bedside immediately after the first clinical examination for the diagnosis of brain death, repeating the application, ideally, every 6 h until organ procurement for transplantation. We also suggest that a member of the transplant coordination office or a designated professional of the ICU or emergency department applies the checklist at the bedside. The same individual will also be responsible for personally prompting the physician in charge to modify the clinical management if any inappropriate aspect of care, according to the checklist, is noted.	2020-12-14T14:36:25.330Z	2020-12-01T00:00:00.000	{ "year": 2020, "sha1": "c02afb8bbd7cf691b256b16b3ff062a6d6fc57a3", "oa_license": "CCBY", "oa_url": "https://annalsofintensivecare.springeropen.com/track/pdf/10.1186/s13613-020-00787-0", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "c02afb8bbd7cf691b256b16b3ff062a6d6fc57a3", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
255600534	pes2o/s2orc	v3-fos-license	Exploring the prognostic value of HK3 and its association with immune infiltration in glioblastoma multiforme Background: Hexokinase 3 (HK3) is one of the key enzymes involved in glucose phosphorylation (the first step in most glucose metabolic pathways). Many studies have demonstrated the vital role of dysregulation of HK3 in several tumors. However, there is a need for in-depth characterization of the role of HK3 in glioblastoma multiforme (GBM). Methods: All data were sourced from The Cancer Genome Atlas (TCGA) and Chinese Glioma Genome Atlas (CGGA). Kaplan-Meier analysis and univariate regression were applied for survival analysis. Gene set enrichment analysis (GSEA) was used for enrichment analysis. Tumor Immune Single Cell Hub (TISCH) database was applied for single-cell analysis. Tumor Immune Dysfunction and Exclusion (TIDE) analysis was applied to evaluate the immune response. Results: HK3 expression was upregulated in GBM and correlated with poor prognosis. The high HK3 expression group was primarily enriched in adaptive immune response, chemokine signaling pathway, and cytokine-cytokine receptor interaction. The high HK3 expression group showed significantly greater enrichment of the majority of immune cells and immune-related pathways. HK3 showed significant correlation with most immune cells, especially macrophages (p < .001, R = .81). TISCH analysis showed that HK3 was predominantly expressed in macrophages in most cancers. HK3 showed significant correlation with most immune-related genes, such as PD-1 (p < .001, R = .41), PDL-1 (p < .001, R = .27), and CTLA-4 (p < .001, R = .29). TIDE analysis revealed that the low HK3 expression group has a lower TIDE score and may benefit from immunotherapy. Drug sensitivity analysis showed that patients with high HK3 expression frequently showed drug resistance. Conclusion: HK3 was associated with poor prognosis and may serve as a biomarker of macrophages in GBM. HK3 was also associated with immune response and drug resistance. Our findings may provide novel insights for GBM immunotherapy. Introduction Glioblastoma multiforme (GBM) is considered as the most frequent primary tumor of the nervous system (Ostrom et al., 2019). Patients with GBM have poor prognoses (median overall survival [OS]: 15 months; 5-year survival rate: <5%) (Tan et al., 2020). Several novel therapies, such as immune checkpoint inhibitors, anti-angiogenesis, tumor vaccines, and Tumor Treating Fields (TTFields), have been tried in GBM (Rodríguez-Camacho et al., 2022); however, most of these therapies are ineffective (Mun et al., 2018). Therefore, understanding the molecular mechanisms and exploring more effective targets for GBM are key research imperatives (Uddin et al., 2022). Metabolic reprogramming is a core feature of tumors characterized by upregulation of glycolysis (Biswas, 2015;Pavlova and Thompson, 2016). Metabolic reprogramming plays an essential role in tumors (Pavlova and Thompson, 2016). Acyl-CoA-binding protein was found to drive GBM tumorigenesis by sustaining fatty acid oxidation (Duman et al., 2019). Isocitrate dehydrogenases (IDH)1/2 was found to drive GBM progression by producing an oncometabolite (Chang et al., 2019). Metabolic programming was shown to help maintain the proliferation of stem cell-like tumor cells in GBM (Suvà et al., 2014). Metabolic reprogramming has also been shown to be involved in the immune activity (Xia et al., 2021). Tumor depletion of glucose limits the metabolism of T cells, resulting in their diminished mTOR activity, glycolytic capacity, and interferon (IFN)-γ production (Chang et al., 2015). Hexokinases are the key enzymes in metabolism (Cavalcante et al., 2016), including HK1, HK2, and HK3 (Wilson, 2003). High expression of HK1 has been shown to be associated with poor prognosis in the context of various tumors, such as colorectal and ovarian cancer (Graziano et al., 2017;Li T. et al., 2020;Jiang et al., 2021). A study demonstrated downregulation of HK1 in GBM (Wolf et al., 2011a). HK2 was shown to promote the motility and proliferation of human ovarian cancer cells by activating Akt1/ p-Akt1 (Tian et al., 2022). Moreover, HK2 was shown to promote GBM progression by glycolysis (Wolf et al., 2011b). HK3 was found to be up-regulated in GBM; however, its role has not been extensively investigated. In this study, we performed a comprehensive analysis of the role of HK3 in GBM using The Cancer Genome Atlas (TCGA), Chinese Glioma Genome Atlas (CGGA), Tumor Immune Single Cell Hub (TISCH), and Human Protein Atlas (HPA) databases. In particular, we performed HK3 gene expression analysis, survival analysis, immune infiltration analysis, single-cell RNA sequencing analysis, and functional enrichment analysis to investigate the prognostic and immunological significance of HK3 in GBM. Materials and methods Datasets collection and pre-processing RNA-sequencing data and associated clinical information of 33 types of cancers were sourced from the UCSC Xena (http:// xena.ucsc.edu/). The CGGA-325 dataset was sourced from the CGGA (http://www.cgga.org.cn/) (Zhao et al., 2021). Samples with incomplete clinical information were excluded. Genotype-tissue expression (GTEx) was available from the UCSC Xena. GTEx includes RNA-sequencing data of 207 cerebral cortex samples. After merging the TCGA-GBM and GTEx, the "normalizeBetweenArrays" function of the "limma" package was applied to eliminate the batch effect . All RNA-sequencing data were normalized (Fragments Per Kilobase Million, FPKM) and log2 (FPKM+1) transformed. Survival analysis Univariate cox regression analysis was used to determine the prognostic value of HK3. Kaplan-Meier survival curves were plotted using the "surv" and "survminer" R packages, and between-group differences in survival were assessed using the log-rank test. Frontiers in Genetics frontiersin.org Immune-related analysis The enrichment scores of 16 immune cells and 13 immunerelated pathways were computed using the "GSVA" and "GSEABase" R packages. The correlation between gene expression and immune cells was assessed using the TIMER 2.0 web application (http://timer.cistrome.org/) . The correlation between HK3 and immune-related genes was determined using Spearman's correlation analysis (Zhu et al., 2021). Tumor Immune Dysfunction and Exclusion (TIDE) analysis includes MSI Expr Sig, Dysfunction, Exclusion, and TIDE scores, which were computed by uploading gene expression data through the web application (http://tideere.dfci.harvard.edu). Patients with low TIDE scores may benefit from immunotherapy . HK3 expression in immune cells The "RNA immune cell" is a part of the Human Protein Atlas (HPA) which was applied to explore the gene expression in immune cells (https://www.proteinatlas.org/) (Karlsson et al., 2021). In this part, the Monaco dataset includes 29 immune cells, and the Schmiedel dataset includes 15 immune cells in peripheral blood (Schmiedel et al., 2018;Monaco et al., 2019). The TISCH is a convenient web application for single-cell analysis (http://tisch.comp-genomics.org/home/), which includes 190 single-cell datasets . Drug sensitivity analysis The "OncoPredict" R package based on Genomics of Drug Sensitivity in Cancer (www.cancerrxgene.org/) was used to calculate the half-maximal inhibitory concentration (IC50) of drugs (Maeser et al., 2021). Cell lines and culture The human glioblastoma cell lines (U87, A172, and U373) and normal human astrocytes (NHA) cell line SVGp12 employed in the study were purchased from American Type Culture Collection. Cells were cultured in Dulbecco's modified Eagle's medium (DMEM; HyClone, Logan, United States) with 10% fetal bovine serum (FBS; Gibco, NY, United States) and 1% penicillin-streptomycin (HyClone, Logan, United States) in 37°C incubators with 5% CO 2 . Statistical analysis All statistical analyses were performed using R software (version 4.2.1). Spearman's test was used for correlation analysis. p values <.05 were considered indicative of statistical significance. Expression of HK3 in cancers HK3 expression was explored at the pan-cancer level using TIMER2.0. The results showed dysregulation of HK3 expression in 14 types of human cancers ( Figure 1A). HK3 mRNA expression in GBM samples was significantly higher than that in normal samples ( Figures 1B, C). qRT-PCR showed that HK3 mRNA expression was significantly higher in A172 and U373 cells than in NHA cells, but not elevated in U87 cells ( Figure 1D). The protein encoded by HK3 was also upregulated in GBM ( Figure 1E). Prognostic value of HK3 in cancers The prognostic value of HK3 is tumor-specific. Univariate Cox regression analysis showed that the expression of HK3 was associated with poor OS in GBM, kidney renal clear cell carcinoma (KIRC), brain lower grade glioma (LGG), thymoma (THYM), uveal melanoma (UVM), kidney chromophobe (KICH), acute myeloid leukemia (LAML), testicular germ cell tumors (TGCT), and liver hepatocellular carcinoma (LIHC), but associated with favorable OS in skin cutaneous melanoma (SKCM) (Figure 2A). Kaplan-Meier analysis showed that high expression of HK3 was associated with poor OS in GBM, LGG, KIRC, LAML, and UVM, but associated with favorable OS in SKCM ( Figure 2B). The HK3 expression was also associated with poor OS of GBM in the CGGA-325 dataset (p < .05, Supplementary Figure S1 Frontiers in Genetics frontiersin.org 04 HK3 in different tumors, and high expression of HK3 was usually associated with poor survival. Function enrichment analysis in GBM To elucidate the biological function and pathways with HK3 involvement, we performed enrichment analysis in the TCGA-GBM. The top five gene ontology (GO) items enriched in the high HK3 expression group were activation of the immune response, acute inflammatory response, adaptive immune response, adaptive immune response based on somatic recombination of immune receptors built from immunoglobulin superfamily domains, and alpha-beta T cell activation ( Figure 4A). In comparison, the top five GO items enriched in the low expression HK3 group were chromosome segregation, nuclear chromosome segregation, condensed chromosome, chromosome centromeric region, and kinetochore ( Figure 4B). The top five KEGG items enriched in the high HK3 expression group were the chemokine signaling pathway, cytokine-cytokine receptor interaction, hematopoietic cell lineage, lysosome, and nod like receptor signaling pathway ( Figure 4C). The top five KEGG items enriched in the low HK3 expression group were cell cycle, ribosome terpenoid backbone, biosynthesis, notch signaling pathway, and spliceosome ( Figure 4D). Relationship between HK3 and immune characteristics in GBM We next performed GSEA analysis to investigate the immunological significance of HK3. In the TCGA and CGGA cohorts, the high HK3 expression group showed significantly greater enrichment of the majority of immune cells and immune-related pathways compared to the low HK3 expression group (Figures 5A, B). HK3 expression showed a significant correlation with immune cell infiltration in both cohorts ( Figures 5C, D) (Supplementary Table S1). Interestingly, among all genes of Table S2). Furthermore, increased macrophage infiltration was associated with poor OS in GBM in the TCGA and CGGA cohorts (Supplementary Figure S2). Analysis of TIMER 2.0 database showed that HK3 expression was associated with M2 macrophages, but not M1 (Supplementary Figure S3). HK3 showed a significant association with macrophage M2 polarization-related genes, such as CD163, VSIG4, MS4A4A, ITGAM, MRC1, and ITGAX (Supplementary Figure S3). Association of HK3 with macrophages Since the expression of HK3 showed a significant association with macrophages (p < .001, R = .81), we further explored whether HK3 expression in macrophages was specific. We investigated the HK3 expression in immune cells in plasma through the HPA. In the Monaco dataset, HK3 was predominantly expressed in basophils, neutrophils, and monocytes/macrophages ( Figure 6A). In the Schmiedel dataset, HK3 was predominantly expressed in monocytes/macrophages ( Figure 6B). We further explored the expression of HK3 at the single cell level through the TISCH. HK3 was found to be predominantly expressed in macrophages in GBM ( Figure 6C). HK3 was also predominantly expressed in macrophages in ovarian serous cystadenocarcinoma (OV), pancreatic adenocarcinoma (PAAD), SKCM, LIHC, non-small cell lung cancer (NSCLC), head and neck squamous cell carcinoma (HNSC), KIRC, carcinoma of colon and rectum (CRC), breast invasive carcinoma (BRCA), chronic lymphocytic leukemia (CLL), and cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC) (Supplementary Figures S4-S7). Co-expression of HK3 with immunerelated genes We next assessed the co-expression of HK3 with immune-related genes. The results showed a positive correlation of HK3 with most Frontiers in Genetics frontiersin.org 07 Association of HK3 with immune response and drug sensitivity We applied the TIDE analysis to explore the association between HK3 expression and the immune response. In the TCGA cohort, the Exclusion-score and the MSI Expr Sigscore in the low HK3 expression group were higher than that in high HK3 expression group, while Dysfunction-score and TIDE-score were lower in the low HK3 expression group (Figures 8A-D). HK3 showed a significant association with the TIDE score (p < .001, R = .48) ( Figure 8E), and its correlation value ranked in the top 5% of all genes in GBM (Supplementary Table S4). Drug sensitivity analysis showed that the high HK3 expression group was frequently associated with resistance to drugs, such as axitinib, carmustine, cyclophosphamide, dactinomycin, and nelarabine ( Figure 9). Frontiers in Genetics frontiersin.org Discussion GBM is a highly aggressive malignancy and the current treatments do not lead to satisfactory outcomes. Therefore, identification of new therapeutic targets in the context of GBM is a key imperative (Koh et al., 2022). Metabolic reprogramming is crucial for cancer and has emerged as a promising therapeutic target (Martínez-Reyes and Chandel, 2021). HK3 is one of the key enzymes involved in metabolism, but few studies have focused on its role in GBM. In the present study, we performed a comprehensive analysis of the prognostic and immunological significance of HK3 in GBM. Our findings may provide a basis for future studies. We first explored HK3 expression and its prognostic value at the pan-cancer level using the TCGA database. Our results demonstrated dysregulation of HK3 in several tumors and its frequent correlation with poor prognosis. In previous studies, HK3 was shown to promote colorectal cell proliferation through epithelial mesenchymal transition . HK3 was also shown to promote metastasis of colorectal cancer via the nuclear factor κB/Snail/Hexokinase-3 signaling axis . Moreover, HK3 was found to prevent apoptosis in colorectal cancer and melanoma cells (Kudryavtseva et al., 2016). Downregulation of HK3 expression impaired neutrophil differentiation and increased sensitivity to anthracyclines in acute promyelocytic leukemia (Federzoni et al., 2012). High expression of HK3 was associated with poor OS in kidney cancer . We applied enrichment analysis to investigate the potential involvement of HK3 in the biological functions and pathways in GBM. Intriguingly, the top five GO and KEGG items enriched in the high HK3 expression group were all related to immune activity, such as immune response, adaptive immune response, chemokine signaling pathway (Ortiz Zacarías et al., 2021), and cytokine-cytokine receptor interaction (Zheng et al., 2022). Thus, we explored the relationship between HK3 and immune activity. We observed consistently greater enrichment of most immune cells and immune-related pathways in the high HK3 expression group. These results suggest the potential involvement of HK3 in shaping the GBM tumor microenvironment. Tumor-associated macrophages (TAMs) account for half of all non-tumor cells in GBM (Charles et al., 2012). TAMs are heterogeneous cellular populations composed of microglia, blood-derived infiltrating macrophages, and monocytes that cross the compromised blood-brain barrier (Zhang H. et al., 2021). Infiltration of TAMs has been shown to be associated with poor OS in the context of most tumors, such as PDAC , GBM (Huang et al., 2020), and BLCA (Wu et al., 2020). Studies have demonstrated a crucial role of TAMs in GBM. SLIT2/ROBO signaling in TAM drives GBM Frontiers in Genetics frontiersin.org immunosuppression and vascular dysmorphia (Geraldo et al., 2021). Pleiotrophin secreted by TAMs was shown to promote PTPRZ1 signaling in GBM stem cells leading to tumor growth (Shi et al., 2017). Colony-stimulating factor 1 receptor (CSF-1R) was found to inhibit macrophage polarization and block glioma progression (Pyonteck et al., 2013). Macrophage-associated pgk1 phosphorylation promoted GBM aerobic glycolysis and tumorigenesis (Zhang et al., 2018). In the present study, we observed a significant correlation of HK3 with macrophages in the GBM (p < .001, R > .8); in addition, HK3 was predominantly expressed in macrophages in most types of cancers at the singlecell level. Thus, HK3 may serve as a biomarker for macrophages. In previous studies, HK3 expression was found to be 100-fold higher in macrophages and granulocytes compared to other immune cells (Seiler et al., 2022). HK3 was also identified as a biomarker of macrophages in clear cell renal cell carcinoma . Moreover, HK3 showed an association with M2 macrophages and T cell dysfunction in GBM (Ji et al., 2022). Immune checkpoint inhibitors represent a great breakthrough in cancer treatment, but these drugs are not effective against GBM (Chokshi et al., 2021). Although anti-PD-1 treatment was shown to increase T cell activity in GBM, TAMs account for half of immune cells, and their Frontiers in Genetics frontiersin.org immunosuppressive effects resulted in treatment failure (de Groot et al., 2020;Lee et al., 2021). Several preclinical and clinical studies have identified TAM as a promising target for cancer immunotherapy (Zhang S. Y. et al., 2020). Differentiation of CD34 progenitor cells into macrophages using macrophage colony stimulating factor (M-CSF) was shown to induce a 6-fold increase in HK3 expression (Seiler et al., 2022). The generation of central nervous system macrophages relies on the transcription factor PU.1 (Goldmann et al., 2016), and HK3 was considered as transcriptional target of PU.1 (Federzoni et al., 2012). Thus, HK3 may be involved in macrophage differentiation and could be a potential target for anti-TAMs; however, further studies are required to verify this hypothesis. In our study, patients with low HK3 expression had lower macrophage infiltration and TIDE scores, indicating that these patients may potentially benefit from immunotherapies. Our study demonstrated increased expression of HK3 expression in GBM tissues and high expression of HK3 was associated with poor prognosis. HK3 showed a significant association with macrophage infiltration and may serve as a biomarker of macrophages. HK3 was also associated with immune response and drug-resistance. In conclusion, this study provides a comprehensive understanding of the role of HK3 in GBM, which may help provide novel insights for developing GBM immunotherapy. Our study has some limitations. The underlying mechanism of the relationship of HK3 with macrophage differentiation and immunotherapy is unclear. Further studies are required to explore this aspect. Conclusion HK3 was associated with poor prognosis of GBM and may serve as a biomarker of macrophages in GBM. HK3 was also associated with immune response and drug resistance. Our findings provide novel insights for development of GBM immunotherapy. Data availability statement The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.	2023-01-12T14:28:25.288Z	2023-01-12T00:00:00.000	{ "year": 2022, "sha1": "42bfa0dff23d1525cc0199f6ccd4328c68c8ac31", "oa_license": null, "oa_url": null, "oa_status": null, "pdf_src": "Frontier", "pdf_hash": "42bfa0dff23d1525cc0199f6ccd4328c68c8ac31", "s2fieldsofstudy": [ "Biology", "Medicine" ], "extfieldsofstudy": [] }
203511839	pes2o/s2orc	v3-fos-license	THE IMPACT OF DEMOCRACY, CORRUPTION AND MILITARY EXPENDITURE ON ENVIRONMENTAL DEGRADATION: EVIDENCE FROM TOP SIX ASEAN COUNTRIES Purpose of the study: The current study aims to examine the relationship between corruption, democracy, military expenditure and environmental degradation in a panel of six ASEAN countries including Malaysia, Indonesia, Philippines, Thailand, Singapore and Vietnam using a panel data from 1995 to 2017. Methodology: In addition, the current study is unique in applying the sophisticated methods of panel Fully Modified Ordinary Least Square (FMOLS) and Dynamic Ordinary Least Square (DOLS) that have been adopted in several earlier quality research. Main Findings: The results of panel estimations conclude that corruption, military expenditure, and democracy have a noteworthy and significant impact on carbon dioxide emission in ASEAN countries. The results of FMOLS and DOLS confirm that there is a positive and significant impact of military expenditure and corruption on carbon dioxide emission. However, we found a negative and significant impact of democracy on carbon dioxide emission in all selected ASEAN countries. Implications: In general, the consequences of both statistical estimations affirm that corruption, democracy, and military expenditure are the critical and noteworthy determinants of carbon dioxide emission in ASEAN nations. INTRODUCTION The growing environmental changes around the globe are persistently changing the living dynamics and potential concerns of the people all around the World (Okon, 2016;Diekmann and Franzen, 2019). The extensive carbon emanations from several domestic and corporate processes are playing their part in disrupting ecological conditions. Keeping in mind the inevitable role of power utilization, the economies are striving to explore the eco-friendly ways of generating power through renewable sources. However, apart from economic indicators, there are certain socio-political measures that exert influence on environmental quality ( The contribution of political stability is fundamental for bringing confidence and reliability in the country's policies. The notion of good governance is highly dependent on fair and unbiased political systems that can work in favor of generating productive opportunities for the economies and environment. In this regard, the role of corruption is greatly important in disrupting environment, both directly and indirectly. As a direct impact, corruption can influence environment in disrupting ecological policies and regulations especially in cases where it leads to disregard ecological laws and enhancements in the culture of subornation (Winbourne, 2002;Sundström, 2013;Alkali and Imam, 2016;Gibson, 2016). In terms of indirect impact, the corruption can affect environmental degradation through augmenting the levels of income inequality, poverty and permitting diminution of natural resources (Frhd et Likewise, the stability of the political regime is considered crucial for ensuring environmental quality. In similar context, democracy or political freedom is vital in influencing environment through sound and stable eco-policies and effective legislation. In addition, corporate sector is considered more influential in the presence of democratic political regime that helps economic goals of future sustainability (Nekooei et al., 2015;Khan and Ali, 2017). Similarly, it is also viewed that political privileges and choice of information encourage the objectives of sustainable development and thus bring positive impact on environmental quality (Rehman et al., 2012). In addition, the freedom of speech in democratic regimes tends to supplement the motives of ecological interest groups that ultimately enhance societal awareness and motivate ecological legislation (Parks and Roberts, 2010;Cossiga, 2018). The growing technological advancements and militarization are also referred as the critical inducer of environmental degradation. Many studies argued that greater levels of greenhouse gases emanations are the outcome of augmented militarization (Clark et al., 2010;Jorgenson et al., 2012;Bildirici, 2018). At present, the increasing race for militarization is causing abruption in the atmosphere by enhancing the levels of pollutions by emitting toxic emissions and higher energy consumption, therefore crucial to be investigated (Bildirici, 2017a; 2017b; 2017c). Therefore, acknowledging the vitality of socio-political indicators in influencing environmental quality (Bae et al., 2017), the current study aims to examine the relationship between corruption, democracy, military spending and environmental degradation in a panel of six ASEAN countries including Malaysia, Indonesia, Philippines, Thailand, Singapore, and Vietnam. The importance of present study lies in exploring the potential impact of military expenditures, level of democracy and corruption in growing trends of carbon emission in the sampled nations. In addition, against the orthodox methods that can raise the queries on the authentication and reliability of the derived findings, the current study is unique in applying the sophisticated methods of panel Fully Modified Ordinary Least Square (FMOLS) and Dynamic Ordinary Least Square (DOLS) that have been adopted in several earlier quality research. Thus, the results obtained from rigorous adopted methodology in exploring the critical dynamic of social-political and environmental association would be helpful to establish the trustworthy role of democracy, military spending, and corruption in changing the levels of environmental quality in ASEAN region. The remaining of the investigation is outlays as below. Section-2 portrays the relevant studies on the association of democracy, military expenditure and corruption with environmental degradation across the globe. Section-3 defines the adopted methods of the current investigation. Section-4 displays the derived results and their interpretations. Lastly, Section-5 concludes the research findings and offers future recommendations. LITERATURE REVIEW Many studies explored the direct and indirect link between several socio-political measures with environmental. In recent years, the shift of interest in identifying the potential link between socio-political indicators such as corruption, democracy, military spending, etc. have been recognized in several studies. Among them, Duong (2016) inspected the association between corruption and carbon emanation in a panel investigation. Utilizing the sample of forty-two economies, the study applied the methods of three-stage least squares (3SLS) to identify the long-run connection among the variables. The outcomes of the study reported that there existed a positive significant relationship between corruption and carbon emanation in the sampled economies. Similarly, Habib et al. (2018) also examined the association between corruption and carbon emanation in a panel of African countries. Utilizing the sample of eighteen economies between the period of 1992 to 2013, the study applied the methods of Generalized Method of Moments (GMM) to identify the long-run connection among the variables. The outcomes of the study reported that there existed a positive significant relationship between corruption and carbon emanation in the African region. For the countries of Asia-Pacific Economic Cooperation (APEC), Zhang et al. (2016) examined the association between corruption and carbon emanation. Utilizing the sample of nineteen economies between the period of 1992 to 2012, the study applied the methods of Panel quantile regression to identify the long-run connection among the variables. The outcomes of the study reported mixed findings. In particular, the outcome confirmed that there existed a negative association between corruption and carbon emanation in low carbon emitting economies. However, the findings failed to find any significant relationship between carbon emission and corruption in high carbon emitting nations. METHODOLOGY In the current examination, we utilize yearly information on corruption, democracy, military expenditure, and carbon dioxide emission. Corruption perception index is used as a proxy of corruption which is represented by (CPI). Moreover, democracy index is used as a proxy of democracy which is explained by (DEMO). Also, military expenditure is denoted by (MILEX) and finally carbon dioxide emission is used as a proxy of environmental degradation which is used as a (CO2). The data of corruption perception index is collected from the website of Transparency International. However, the information of democracy index is collected from the website of Polity IV project. Finally, the data on military expenditure and carbon dioxide emission is collected from the website of World Bank (World Development Indicator). The current investigation is analyzing the effect of corruption, military expenditure, and democracy on environmental degradation in ASEAN nations including Thailand, Singapore, Indonesia, Malaysia, Vietnam, and the Philippines. In the current investigation, we utilize corruption, military expenditure, and democracy as a determinant of carbon dioxide emission in different ASEAN nations. Additionally, the current investigation covers the time range from 1995 to 2017 for selected ASEAN nations. In the current investigation, the carbon dioxide emission is exhibited according to the below equation: Where, CO2 it is the carbon dioxide emission and measure in (kilo tons of oil equivalent), CPI it is corruption perception index and measured from the range of 0 (highly corrupt) to 100 (very clean). MILEX it is a military expenditure which is measured in (US Dollars) and DEMO it is the democracy index which is measured from the range of 0 (no democracy) to 10 (full democracy). At long last, I speak about the numbers of nations use in the current study and t explains the timeframe of the current investigation. In a recent study, we explore the long-term connection among the factors by consuming a panel long-run relationship. Correspondingly, current study explains the long-run impact of corruption, military expenditure, and democracy on carbon dioxide emission by utilizing FMOLS and DOLS approaches. At last, we utilize a new method of heterogeneous panel causality analysis to deal with inspect the possible causal connection among corruption, military expenditure, democracy and carbon dioxide emission in top six ASEAN countries. DATA ANALYSIS AND DISCUSSION Descriptive Statistics Primarily, we investigate the essential statistics of the selected number of ASEAN nations. The outcomes associated with the estimations are displayed in This methodology is utilized to research the pattern of integration of the components. For example, if every single chosen factor is not stationary of the level, for instance I(0), by then this implies the majority of the variables have a unit root issue at level and are stationary at first differential arrangement. Consequently, it might be observed that all the chose variable in the present examination might have a connection in the longer-term period. for all variables, implying that each of the variables is non-stationary at level and become stationary at first differential series. Therefore, the results of unit root confirm that each of the selected variables is showing up non-stationary property at the level series and exhibiting stationary properties at the first difference stage. In general, all components are integrated at I(1). Thusly, there must be a sign of long-run association between the components in long run. Panel Cointegration tests We additionally apply Kao (2003) and Pedroni (2004) panel long-run relationship procedures to examine the long-run connection between corruption, democracy, military expenditure, and carbon dioxide emission in the ASEAN countries. So as to run this approach, every factor ought to be stationary at first differential, for example, I(1). In light of the existence of long-run relationship, the long-run estimation remains calculated. In a cross-sectional examination, the error fluctuation changes over the groups which affect the consistency of the parameters. So as to adapt up this issue the generalized least squares strategy (GLS) could be used. In any case, the difference consistency still happens, for example, the relationship of the squared residuals with the regressor in each group. As such, to deal with the issue giving the issue of heteroskedasticity, we finally apply fully modified ordinary least square (FMOLS) and dynamic ordinary least square (DOLS) strategy. The cointegration among all factors confirms by using the Kao test. As saw from the outcomes presented in table-4, the null hypothesis is rejected and the alternative hypothesis is accepted, i.e., long-run relationship exists between corruption, military expenditure, democracy and carbon dioxide emission in the top six ASEAN nations. Long Run Estimations Present investigation utilizes pooled conventional least squares (OLS) to look at the effect of corruption, military expenditure, and democracy on carbon dioxide emission in top six ASEAN nations. Moreover, Pedroni (2001a;2001b) opposed that as a result of regression result, inconsistent controls could affect the presence of sequential correlation and endogeneity issue among the regressor. Likewise, to deal with these issues, the present examination uses the FMOLS procedure. This system focuses on the non-parametric strategy so as to choose the issue of endogeneity and sequential correlation (Ametorwo, 2016;Sharif et al., 2019). In like way, we use FMOLS and DOLS methodologies to examine the long haul association among corruption, military expenditure, democracy, and carbon dioxide emission in top six ASEAN countries. The long-run association between the variable is examined by utilizing the FMOLS and DOLS approaches. These approaches were displayed by Phillips and Hansen (1990) and sometime later adjusted by the Pedroni (2001). We select these methods since they talk about to endogeneity and autocorrelation problems and provide healthy outcomes. Likewise, we investigate the long-run checks by taking the FMOLS and DOLS coefficients. The consequences of FMOLS and DOLS have been shown to in table-5. The long-run coefficient evaluated utilizing two novel strategies which are particularly proportional and basic at the 10% essentialness level. The results of FMOLS and DOLS attest that every one of the determinants of carbon dioxide emission, corruption, military expenditure and democracy in the best six ASEAN countries. The consequences of the long-run coefficient further recommend that every one of the determinants considered in this investigation significantly affect carbon dioxide emission in ASEAN nations. Table-5 explains that the results of panel estimations conclude that the long-run impact on carbon dioxide emission by corruption perception index to 0.226; a per unit variation in the corruption will impact on carbon dioxide emission by 0.226 unit. The outcomes further suggested that military expenditure have also a significant and positive impact on carbon dioxide emission. The outcomes confirm that a per unit increase in military expenditure causes 0.332 unit change in carbon dioxide emission. Finally, results also confirmed that democracy has also negative and significant impact on carbon dioxide emission. Results suggested that a per unit increase in democracy causes 0.393 unit decrease in carbon dioxide emission in ASEAN countries. In general, the consequences of FMOLS and DOLS affirm that corruption, democracy, and military expenditure are the critical and noteworthy determinants of carbon dioxide emission in ASEAN nations. The outcomes recommend that ASEAN countries need an environmental efficiency and good governance system which is free of corruption also when they are cooperating with one another. The outcomes additionally affirmed that the more the corruption and military consumption the more will be the carbon dioxide emission in these ASEAN countries. CONCLUSION The growing environmental changes around the globe are persistently changing the living dynamics and potential concerns of the people all around the World. The extensive carbon emanations from several domestic and corporate processes are playing their part in disrupting ecological conditions. Keeping in mind the inevitable role of power utilization, the economies are striving to explore the eco-friendly ways of generating power through renewable sources. However, apart from economic indicators, there are certain socio-political measures that exert influence on environmental quality. Therefore, acknowledging the influencing nature of environmental indicators, the current study aims to examine the relationship between corruption, democracy, military expenditure and environmental degradation in a panel of six ASEAN countries including Malaysia, Indonesia, Philippines, Thailand, Singapore, and Vietnam. The importance of present study lies in exploring the potential impact of military expenditures, level of democracy and corruption in growing trends of carbon emission in the sampled nations. In addition, against the orthodox methods that can raise the queries on the authentication and reliability of the derived findings, the current study is unique in applying the sophisticated methods of panel Fully Modified Ordinary Least Square (FMOLS) and Dynamic Ordinary Least Square (DOLS) that have been adopted in several earlier quality research. Thus, the results obtained from rigorous adopted methodology in exploring the critical dynamic of social-political and environmental association would be helpful to establish the trustworthy role of democracy, military spending, and corruption in changing the levels of environmental quality in ASEAN region. The results of panel estimations conclude that corruption, military expenditure, and democracy have a noteworthy and significant impact on carbon dioxide emission in ASEAN countries. The results of FMOLS and DOLS confirm that there is a positive and significant impact of military expenditure and corruption on carbon dioxide emission, however, we found a negative and significant impact of democracy on carbon dioxide emission in all selected ASEAN countries. In general, the consequences of both statistical estimations affirm that corruption, democracy, and military expenditure are the critical and noteworthy determinants of carbon dioxide emission in ASEAN nations. The outcomes recommend that ASEAN countries need an environmental efficiency and good governance system which is free of corruption also when they are cooperating with one another. The outcomes additionally affirmed that the more the corruption and military consumption the more will be the carbon dioxide emission in these ASEAN countries.	2019-09-19T09:13:34.699Z	2019-09-11T00:00:00.000	{ "year": 2019, "sha1": "17c8799055011813eea1831b47016bcfa07a30cd", "oa_license": null, "oa_url": "https://doi.org/10.18510/hssr.2019.7443", "oa_status": "GOLD", "pdf_src": "MergedPDFExtraction", "pdf_hash": "664354c7333e87a186a6d4e771e38b8e5b82a214", "s2fieldsofstudy": [ "Business" ], "extfieldsofstudy": [ "Economics" ] }
221724103	pes2o/s2orc	v3-fos-license	Treatment of Mandible Fractures Using a Miniplate System: A Retrospective Analysis Three-dimensional (3D) mini plate systems are used in the treatment of mandibular fractures. The system is advantageous in comparison to conventional plates due to the stabilization of tension and compression areas, improved initial stability, and biomechanical behavior. The aim of this retrospective study was to evaluate the use of a 3D miniplate system for the treatment of patients with mandibular fractures. Patients with mandibular fractures treated with a 3D plate system at the Department of Oral and Maxillofacial Surgery, University Hospital Münster, during a period of 5 years, were included in this study. Mandibular fracture conditions and minor and major post-operative complications were reported. In total, 336 patients and 391 mandibular fractures were assessed. The most common fracture site was anterior mandible, and 155 cases involved a tooth-bearing area. Minor complications were seen in 8.03% of cases, whereas only 1.49% of patients suffered from major complications. The treatment of mandible fractures using 3D miniplates resulted in fracture reduction with a low complication rate. Introduction In the management of mandible fractures, it is essential to restore occlusion, temporomandibular function, and facial esthetics [1,2]. Treatments range from non-surgical management to surgical procedures with osteosynthesis plates, according to the fracture complexity [3]. In cases where surgical procedure is required, internal rigid fixation techniques promote better stabilization and consolidation of fractured bone. Nevertheless, a successful treatment is influenced by various factors, such as the location and complexity of the fracture, and the plate system and its geometry [4]. Infection, paresthesia, and malocclusions are reported as possible post-operative complications after osteosynthesis of mandible fractures [3][4][5][6]. To ensure successful osteosynthesis with minimal invasive approach, three-dimensional (3D) miniplate systems can be used [7][8][9][10][11]. The system comprises two miniplates joined by interconnecting crossbars and fixed to the bone with monocortical screws [9,10]. When placed in the fracture line, the 3D miniplate system ensures fracture stabilization, regardless of plate thickness and geometry [10]. The placement of screws in a square pattern on either side of the fracture creates a solid platform that increases resistance to twisting and rotation of the longitudinal axis of the plate [10]. Simultaneous stabilization of tension and compression areas, improved initial stability, and biomechanical behavior are the main advantages of the 3D miniplate system in comparison to conventional plates [9,11]. In addition, this system can be easily operated by the patient due to its shape and low profile, which enables reduced palpability and contouring. Compared to the conventional system, previous studies using 3D miniplates showed a reduction in 58% of post-operative complications [12,13]. However, few studies reported the clinical efficacy of 3D miniplate systems for treating mandibular fractures. The aim of this retrospective study was to evaluate the use of the 3D miniplate system for the treatment of patients with mandibular fractures. Experimental Section Patients who had attended the Department of Oral and Maxillofacial Surgery at the University Hospital Münster over a period of five years were included in this study. Inclusion criteria were the presence of mandibular fractures treated with a 3D miniplate system (Trauma 2.0, Medartis, Basel, Switzerland). The system is adapted for craniofacial fractures and ensures the semi-rigid fixation of fractured bone. Exclusion criteria consisted of the use of alternative techniques, condylar fracture, or minimal fractures treated with conservative approaches. Pre-and Intra-Operative Data Assessment Patient data (gender, age, weight, medical history), surgical protocol, and radiographic exams were recorded. Fracture assessment was performed with panoramic radiograph, occlusal, and Clementschitsch view projections. Based on pre-operative panoramic radiographs, the localization of jaw fractures was classified as follows: (1) right and left mandibular angles, (2) right and left mandible body, (3) anterior mandible. The degree of displacement was classified as either major or minor. The presence of teeth at the fracture line, dental trauma, and additional soft-tissue injuries, as well as cranial, thoracic, and abdominal traumas were also assessed. In addition, the operation time was recorded. Surgical Procedure Surgical protocol was followed as briefly described: In order to access bone, an intra-oral vestibular incision was made and the mucoperiosteal flap was raised. After reducing bone fragments into their anatomical position, intermediate intermaxillary fixation (IMF) was performed using wire ligature or IMF screws to ensure adequate occlusion. Fracture osteosynthesis of the fracture was performed using the 3D locking miniplate system (Trauma 2.0, Medartis, Basel, Switzerland). Plates were adapted to the bone anatomy and fixed to the cortical bone using monocortical 2.0 mm screws ( Figure 1). After IMF removal and proof of occlusion, incision closure was performed with 3.0 Vycril sutures. When necessary, a postoperative IMF was applied to stabilize the occlusion using wire ligature, IMF screws, or head-chin bandages. Patients were hospitalized for 7 to 14 days. During this period, patients received a liquid diet or enteral nutrition, and a soft diet was recommended for the first 30 days after the procedure. Analgesic was prescribed in case of pain, and antibiotic therapy (Amoxiclav 875/125 mg; in case of penicillin allergy: Clindamicin 600 mg) was prescribed for a period of 5-7 days to prevent bacterial infection. sutures. When necessary, a postoperative IMF was applied to stabilize the occlusion using wire ligature, IMF screws, or head-chin bandages. Patients were hospitalized for 7 to 14 days. During this period, patients received a liquid diet or enteral nutrition, and a soft diet was recommended for the first 30 days after the procedure. Analgesic was prescribed in case of pain, and antibiotic therapy (Amoxiclav 875/125 mg; in case of penicillin allergy: Clindamicin 600 mg) was prescribed for a period of 5-7 days to prevent bacterial infection. Post-Operative Data Assessment Treatment protocol, presence of multiple fractures, and the need for additional plate systems were assessed. Additionally, gaps between fractured segments and the presence of minor and major post-operative complications were recorded. Post-operative pain was assessed using the Visual Analogue Scale (VAS). Categories of pain were defined as follows: 1-3 = mild pain, 4-6 = moderate pain, and 7-10 = severe pain. Pre-and Intra-Operative Data Assessment Over five years, 582 cases of mandible fractures were surgically treated. Of these, 336 patients received a 3D miniplate (Trauma 2.0, Medartis, Basel, Switzerland). and were included in this study. Fifty-five patients showed a second fracture, totalizing 391 treated mandibular fractures. Participants consisted of male (73.5%) and female (26.5%) patients. At the time of the surgical osteosynthesis, most patients (46.72%) were between 21-40 years old. Among 70 patients that presented a comorbidity (e.g., hypertension), 31.9% presented cardiovascular disease and a total of 22 patients required medication (Table 1). Fracture Assessment The main fracture characteristics are reported in Table 2. Brutality acts (38.7%) and accidents (35.41%) were the main causes of fractures. The most common fracture sites were the anterior mandible, mandibular angle, and mandible body. In addition to these conditions, 290 patients presented a second mandible fracture. Additionally, 37 patients presented with other facial fractures, of which 11 (11%) had zygomatic fractures. Most fractures (86.4%) involved the dental area, and in 155 cases, fractures were located in the tooth-bearing area. Furthermore, 36 patients reported a recent third molar extraction on the fracture site. Dental trauma occurred in 93 cases, as follows: 33 enamel-dentin fracture, 11 avulsion teeth, 7 subluxation teeth and 5 horizontal fractures. In addition, 107 patients received intermaxillary fixation to treat condylar fractures, and seven patients were treated with a head-chin bandage. Postoperative antibiotic therapy was prescribed for 320 patients (95.2%). Figure 2 shows the post-operative panoramic radiograph of a double mandibular fracture treated with 3D miniplates. In 314 patients, the plate was removed after four months, whereas for six patients, the second surgical procedure occurred in two months. The mean operation time (standard deviation) was 80 (60.72) min. In addition, 107 patients received intermaxillary fixation to treat condylar fractures, and seven patients were treated with a head-chin bandage. Postoperative antibiotic therapy was prescribed for 320 patients (95.2%). Figure 2 shows the post-operative panoramic radiograph of a double mandibular fracture treated with 3D miniplates. In 314 patients, the plate was removed after four months, whereas for six patients, the second surgical procedure occurred in two months. The mean operation time (standard deviation) was 80 (60.72) min. Post-Operative Data Assessment In 87.79% of cases, there was a minor displacement of bone fragments, and the gap between fractured sites was less than 1 mm. Furthermore, in 89.3% of cases, no displacement was seen in the panoramic radiography. Post-operative minor complications were observed in 8.03% of the patients, which were treated with antibiotic therapy and local procedures. Only 1.48% of the patients presented major complications (Table 3). No plate fracture was reported, and no plate was lost due to complications, although screw loosening occurred in 1.78% of cases. Due to infectious or technical complications, the plate had to be replaced in five patients (1.5%). Post-Operative Data Assessment In 87.79% of cases, there was a minor displacement of bone fragments, and the gap between fractured sites was less than 1 mm. Furthermore, in 89.3% of cases, no displacement was seen in the panoramic radiography. Post-operative minor complications were observed in 8.03% of the patients, which were treated with antibiotic therapy and local procedures. Only 1.48% of the patients presented major complications (Table 3). No plate fracture was reported, and no plate was lost due to complications, although screw loosening occurred in 1.78% of cases. Due to infectious or technical complications, the plate had to be replaced in five patients (1.5%). Discussion The present study shows a retrospective assessment of mandibular fractures treated with 3D miniplates. According to our results, post-operative complications occurred in 8.03% of cases, most of which were treated by antibiotic therapy and local management. This finding is consistent with previous studies evaluating 3D miniplates, in which the percentage of complications reached 6%. Sensibility disturbance was reported by a high number of patients. This may be related to the nerve injury during bone manipulation and plating, since only few of them reported sensibility disturbance after plate removal. Moreover, wound dehiscence, which tends to occur when the plate is placed close to the incision [10,14], was the most prevalent post-operative complication among patients observed in this study. Out of the 336 patients assessed, only five of them presented major complications and underwent a second surgery to replace the titanium plates. The low rate of major complications is consistent with the findings of Singh et al. (2016) [10], who reported no complications in 20 patients treated with either locking or non-locking 3D plates. Currently, the treatment of mandibular fractures is founded on two principles. The first one claims that the plate must provide sufficient rigidity to avoid displacement between fragments during functional movements. This is achieved by a rigid fixation using solid plates [15,16]. Conversely, the Champy technique suggests the use of semi-rigid fixation, on which tensional forces are neutralized when miniplates are placed according to the ideal line of osteosynthesis [17,18]. Three-dimensional miniplates were introduced as a modification of Champy´s method, on which two plates are connected by interconnecting crossbars, and separated by a distance of 4-5 mm from each other. Due to their malleable format, they can be anatomically adapted without distortion. Thus, tensile and compressive forces are simultaneously neutralized, providing bone alignment and undisturbed healing. In addition, due to their shape and low profile, these plates allow reduced palpability and contouring in the fractured jaw area, decreasing the risk of paresthesia [19]. A recent meta-analysis showed that 3D and conventional miniplates present a similar complication rate [18]. In regard to post-operative infection, wound dehiscence and paresthesia, both methods provided similar results [18]. Conversely, post-operative complications, such as mal-occlusion and trismus, appear to occur more frequently when fractures are fixed using conventional miniplates [9,20]. These findings are in agreement with Mohd et al. (2019) [21], who showed a lower mobility and improved occlusion restoration for parasymphisis fractures treated with 3D miniplates. Reduced rates of occlusion-related complications may be associated with improved plate stability. However, as a comparison among different plate systems was beyond the scope of this study, no conclusion regarding the best fixation method can be drawn from these results. Another advantage of miniplates is the easy handling, which contributes to the shortening of intra-operative time. The present study shows an acceptable mean operation time, which ranges from 12 to 630 min, according to the clinical condition and complexity of the case. Moreover, the treatment of anterior mandibular fracture using 3D miniplates reduced operative time [22]. This might be related to the capability of achieving a stable condition using 3D miniplates, as well as the simultaneous stabilization of tension and compression zones provided by the plate. Nonetheless, a limitation of this system is the need to remove the entire plate when a failure occurs. In this case, two separate plates allow an easier management of plate failure. In this study, most of the fractures treated with a 3D miniplate system showed a gap of less than 1 mm and no displacement between bone fragments. These factors play an important role in the final outcome, since one of the main purposes of the mandibular fracture treatment is to provide fixation and stabilization of the fractured bone [16]. The optimal results are possibly related to the fact that only few cases showed unfavorable displacement of bone segments. A previous systematic review showed a success rate of 88.9% for treatment of mandibular fractures with 3D miniplates. This result relies on the area of the fracture zone and fractured bone displacement. Regarding mandibular angle fractures, lower success rates have been reported [23]. In this study, a large number of fractures were located at the mandibular angle, which might be considered unfavorable due to the strong displacement caused by muscular movements [23,24]. This confirms previous findings in the literature, which classify mandibular symphysis and mandibular angle as the most fractured sites [3]. In regard to the patients' data, our study provides further evidence for the higher prevalence of male patients as a result of brutality acts or accidents. In 155 cases, the fracture occurred in the tooth-bearing area, although only five required tooth removal. Further extractions involved dental trauma indications. According to the literature, indications for tooth extraction are related to the teeth's condition, and the impossibility of reducing and fixating the fracture. Whereas fractured teeth, periodontal diseases, or caries are conditions that may require tooth extraction, the complication rate does not increase if a healthy tooth is maintained [25]. Several treatment-and patient-related aspects can lead to complications, such as patient medical conditions, surgical managements, and patient compliance. These factors should be considered individually to provide a true relation between post-operative complications and the plate system. A limitation of the study is the bias related to the study design, since the choice to evaluate a 3D miniplate system for surgical treatment of mandibular fractures introduces selection bias. Furthermore, measurement of displacement is restricted to bidimensional images. For further studies, the authors recommend a 3D assessment of plating to obtain accurate results. Randomized trials comparing different plate systems and taking into consideration the case complexity should be performed to indicate the appropriate plate system according to individual situations. Conclusions Within the limitations of this study, the treatment of mandible fractures using 3D miniplates resulted in fracture reduction with a low complication rate and an adequate surgical procedure time.	2020-09-16T13:06:19.337Z	2020-09-01T00:00:00.000	{ "year": 2020, "sha1": "e7b05b38425ab1e4e6a5c16d1d3f344a57955459", "oa_license": "CCBY", "oa_url": "https://www.mdpi.com/2077-0383/9/9/2922/pdf", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "26191c8d7493f0add7545885bd095c931c344f66", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
248451759	pes2o/s2orc	v3-fos-license	Sleep quality in persons with mental disorders: Changes during inpatient treatment across 10 diagnostic groups Sleep disturbances have been documented across a range of mental disorders, particularly depression. However, studies that have examined sleep quality in large samples of different diagnostic groups and that report how sleep quality changes during inpatient treatment have been scarce. This retrospective, observational study examined changes in sleep quality during inpatient treatment at a psychosomatic hospital in Germany from admission to discharge as a function of 10 diagnostic groups. Data of 11,226 inpatients were analysed who completed the Pittsburgh Sleep Quality Index as part of the routine diagnostic assessment at admission and discharge. All diagnostic groups showed impaired sleep quality (Pittsburgh Sleep Quality Index score > 5). Patients with trauma‐related disorders had the lowest sleep quality and patients with obsessive–compulsive disorder had the highest sleep quality. While sleep quality significantly improved in each diagnostic group, changes differed in size, with patients with trauma‐related disorders showing the smallest improvement and patients with eating disorders showing the largest improvement. The current study documents impaired sleep quality in inpatients with mental disorders and shows that sleep problems are a transdiagnostic feature in this population. Results also resonate with earlier suggestions that sleep disturbances represent a key feature of trauma‐related disorders in particular and the need for trauma‐specific sleep interventions. Although sleep quality significantly improved during disorder‐specific inpatient treatment in all diagnostic groups, average scores were still clinically elevated at discharge. Thus, a future avenue would be to examine whether adding sleep‐specific treatment elements fosters both short‐ and long‐term success in the treatment of mental disorders. Summary Sleep disturbances have been documented across a range of mental disorders, particularly depression. However, studies that have examined sleep quality in large samples of different diagnostic groups and that report how sleep quality changes during inpatient treatment have been scarce. This retrospective, observational study examined changes in sleep quality during inpatient treatment at a psychosomatic hospital in Germany from admission to discharge as a function of 10 diagnostic groups. Data of 11,226 inpatients were analysed who completed the Pittsburgh Sleep Quality Index as part of the routine diagnostic assessment at admission and discharge. All diagnostic groups showed impaired sleep quality (Pittsburgh Sleep Quality Index score > 5). Patients with trauma-related disorders had the lowest sleep quality and patients with obsessive-compulsive disorder had the highest sleep quality. While sleep quality significantly improved in each diagnostic group, changes differed in size, with patients with trauma-related disorders showing the smallest improvement and patients with eating disorders showing the largest improvement. The current study documents impaired sleep quality in inpatients with mental disorders and shows that sleep problems are a transdiagnostic feature in this population. Results also resonate with earlier suggestions that sleep disturbances represent a key feature of trauma-related disorders in particular and the need for trauma-specific sleep interventions. Although sleep quality significantly improved during disorder-specific inpatient treatment in all diagnostic groups, average scores were still clinically elevated at discharge. Thus, a future avenue would be to examine whether adding sleep-specific treatment elements fosters both short-and long-term success in the treatment of mental disorders. (Baglioni et al., 2016;Lijun et al., 2012). However, studies that have examined sleep quality in large samples of different diagnostic groups and that report how sleep quality changes during psychotherapeutic treatment have been scarce (Schennach, Feige, Riemann, Heuser, & Voderholzer, 2019). Therefore, this report examines = 8.6 (SD changes in sleep quality during inpatient treatment from admission to discharge as a function of 10 diagnostic groups in more than 11,000 patients with mental disorders. which are converted to seven component scores, internal reliability of which was ω = 0.758 at admission and ω = 0.783 at discharge. Higher total scores represent lower sleep quality. Changes in PSQI total scores from admission to discharge as a function of diagnostic groups (based on ICD-10 categories; Table 1) were examined with analyses of variance and paired t-tests. The data that support the findings of this study are openly available at https://osf.io/82wmp. F I G U R E 1 Mean sum scores of the Pittsburgh Sleep Quality Index at admission and discharge as a function of diagnostic groups. Error bars indicate the standard error of the mean. Note that higher scores represent lower sleep quality In line with previous findings, the current study documents impaired sleep quality in inpatients with mental disorders (Baglioni et al., 2016;Lijun et al., 2012;Schennach et al., 2019). Specifically, average PSQI scores were above the cut-off score of 5-indicating poor sleep quality (Buysse et al., 1989)-in all diagnostic groups, suggesting that sleep problems are a transdiagnostic feature in this population. Of note, however, is that results are limited to inpatients that are typically treated in psychosomatic hospitals in Germany and, thus, are not representative for the entire population of inpatients with mental disorders (including, e.g., psychiatric inpatients with substance use or psychotic disorders). In contrast to previous reports, which found the lowest selfreported sleep quality and the most severe alterations in polysomnographic variables in patients with depression (Baglioni et al., 2016;Lijun et al., 2012), patients with trauma-related disorders showed the lowest sleep quality in our sample. However, this result is in line with previous findings from our hospital (Schennach et al., 2019), and with earlier suggestions that sleep disturbances represent a key feature of post-traumatic stress disorder (Germain, 2013). Furthermore, patients with trauma-related disorders showed the smallest improvements in sleep quality from admission to discharge. Thus, the current findings resonate with recent suggestions that highlight the need for traumaspecific sleep interventions in the treatment of trauma-related disorders (Miller, Brownlow, & Gehrman, 2020). Finally, although disorder-specific inpatient treatment significantly improved sleep quality in all diagnostic groups in the current study, average scores were still clinically elevated at discharge. Thus, a future avenue would be to examine whether low sleep quality at discharge is predictive of poorer long-term treatment outcome, and whether adding sleepspecific treatment elements fosters both short-and long-term success in the treatment of mental disorders. Such transdiagnostic, sleep-specific treatment programs for inpatients with mental disorders are currently under development (Sheaves et al., 2018), and preliminary feasibility trials suggest that they have the potential to improve sleep quality and possibly other health outcomes as well (Schneider et al., 2020).	2022-05-01T06:23:17.446Z	2022-04-29T00:00:00.000	{ "year": 2022, "sha1": "2cabfad94da86c77b9f6fc01e39bfd102298101e", "oa_license": "CCBY", "oa_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/jsr.13624", "oa_status": "HYBRID", "pdf_src": "Wiley", "pdf_hash": "a8814cfe731e538cca78ca87026540dc8a812618", "s2fieldsofstudy": [ "Psychology", "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
30996969	pes2o/s2orc	v3-fos-license	In vivo antitussive activity of Coccin-ia grandis against irritant aerosol and sulfur dioxide-induced cough model in rodents In vivo anti-tussive activity of Coccinia grandis against irritant aerosol and sulfur dioxide-induced cough model in rodents Introduction Coccinia grandis Voigt.(Cucurbitaceae), commonly known as 'Little gourd' and as 'Kovai' (Hindi), is a climbing perennial herb with a tuberous rootstock producing annual stems up to several meters long, hispid. The plant has been extensively used in Ayurvedic and Unani practice in the Indian subcontinent (Chopra et al., 1956;Varier and Sala, 1994).The entire plant is reported to use for the treatment of syphilis, sores and bacterial infections (De Boer et al., 2005).The plant also possesses potent antidiabetic (Kumar et al., 1993;Singh et al., 1985;Mukerjee et al., 1972) and anti-dyslipidemic activity (Singh et al., 2007).Indigenous people use various parts of the plant to get relief from asthma and cough (Varier and Sala., 1994).To substantiate its activity against cough and cold attack, the present study was designed to investigate the effect of C. grandis fruit extract on irritant aerosol-induced coughing in male guinea pigs (Forsberg et al., 1988) and sulfur dioxide-induced cough reflex in mice (Miyagoshi et al., 1986). Plant material and extraction The fruits of C. grandis were collected from Birla Institute of Technology, Mesra campus, Ranchi during September, 2006.The plant materials were identified by the Department of Natural Products, Regional Research Laboratory, Bhubaneswar, India.A voucher specimen [No.CNH/1-1 (44)] had been kept in our laboratory for future reference.The fruits were shade-dried, powdered, passed through a 40-mesh sieve and finally subjected to extraction with methanol in a soxhlet apparatus.The solvent was removed under vacuum and a solid mass (15.3% w/w with respect to dry starting material) so obtained was stored in a desiccator and used for further experimental studies.The methanol extract of the fruit was subjected for preliminary phytochemical screening to show the presence of steroid, alkaloid, glycoside, tannin, triterpenoid, carbohydrates and reducing sugar. Experimental animals The experiments were carried out in male guinea pigs (400-450 g) and Swiss albino mice (30-40 g) of either sex.Animals were kept in the animal house at 26 ± 2°C at relative humidity 44-55% and light-dark cycles of 10 and 14 hours, respectively.Animals were provided with rodent diet and water ad libitum.The animal experiment was performed according to the institute's ethical committee approval and guidelines Registration No. 621/02/ac/CPCSEA of Birla Institute of Technology, Mesra, India. Protocol for irritant aerosol-induced antitussive evaluation Guinea pigs, five in each group, were used in the study by the method described elsewhere (Forsberg et al., 1988).Unanesthetized and unrestrained animals were placed individually in a transparent perspex chamber, dimensions 30 × 20 × 20 cm and exposed to a nebulized aqueous solution of i.e. 0.1 g/mL of citric acid for 7 min.The aerosol was produced by airflow of 8 L/min through a Wright nebulizer.The aerosol particles had a mass median aerodynamic diameter of 0.9 µm.The output of nebulizer was 0.7 ± 0.04 mL solution per min and continued for 7 min.The same nebulizer was used throughout the experiment.During the last 5 min of the exposure, the animals were watched continuously by a trained observer, and the numbers of coughs were determined.Coughs could easily be distinguished from sneeze since there was a clear difference in sound as well as in behavior of the animals (Forsberg et al., 1988). The above protocol was performed 10 min after exposing animals to aerosols of the following solutions for a period of 7 min (n=5): a) normal saline (baseline measurement); b) codeine solution (0.03 g/mL, positive control); c) 2.5% methanol extract; d) 5% methanol extract. All the experiments were performed randomly with 2 hours resting period between each two experiments. Sulfur dioxide-induce antitussive evaluation Antitussive effect against sulfur dioxide-induced cough was evaluated by the method as described by Miyagoshi et al., 1986.The experimental model is shown in Figure 1, where, V1 is 500 mL three-necked flask containing aqueous saturated solution of sodium hydrogen sulfite. By opening the stopcock of a burette V2, the concentrated sulfuric acid was introduced to generate sulfur dioxide gas.The chemical reaction that occurred in the flask A is: Previously, SO2 gas was filled in V1 and V3 gas reservoirs, and then by opening the cocks 3 and 2, pressure in the gas reservoir V3 was elevated which was recorded by the water manometer V4.Then the stopcock 2 was closed and stopcock 4 was opened slightly till the pressure in V4 (11 mm i.d.) reached 75 mm water, when the stopcock was closed.The procedure was operated in a draught. The mice were divided into five groups, each containing 10 mice.One group served as a control group receiving only 2% v/v aqueous Tween 80 solution (10 mL/kg, per orally).Three groups were used for methanol extract of fruits (100, 200 and 400 mg/kg per orally) and the remaining group was used for standard drug codeine phosphate (10, 20 and 40 mg/kg per orally).Both the extract and codeine phosphate were suspended separately in 2% v/v aqueous Tween 80 solution.Initially, the cough responses of all groups of animals were observed (0 min) by placing the animals individually in a desiccator V5.The cocks 3, 6 and 5 were opened in order and when the pressure in V4 became 0 mm of water, all the cocks were closed immediately.A certain amount of SO2 gas (5 mL which was kept constant throughout the experiment) was introduced in the desiccator in this way.After 1 min of introduction of the gas, the mice were taken out of the desiccator and the frequency of cough was observed for 5 min in an open-ended filter funnel with a stethoscope at the tip in which the mice were confined.In this way the frequency of cough was observed for all animal groups at 0 min (before the drug administration) and at 30, 60, 90, 120 min interval (after the drug administration). Statistical analysis The frequency of cough produced by irritant-aerosol in guinea pigs were analyzed by one-way ANOVA followed by Bonferroni's multiple comparison tests to compare all pairs of columns.The results are expressed as means ± standard error of mean (S.E.M.).The data obtained from sulfur dioxide-induced experiment were Aerosol-induced antitussive evaluation In the present study the antitussive effects of extracts from C. grandis were evaluated using a standard method used previously by several investigators (Forsberg et al., 1988;Karlsson et al., 1990).The results are depicted in Figure 2.Both concentrations of methanol extract (2.5% and 5%) and the prototype drug, codeine phosphate (0.03 g/mL) caused significant reduction in cough number compared to base line value (p<0.001, Figure 2).The antitussive effects 5% methanol extract also significantly greater than that of codeine (p<0.01, Figure 2), but 2.5% methanol extract was not significant.In addition, the antitussive effects of higher concentrations of methanol extract of C. grandis were significantly greater than those of lower concentrations (p<0.01). Sulfur dioxide-induced antitussive evaluation The effect of C. grandis extract on sulfur dioxideinduced cough in experimental animals is shown in Table I.It was observed that on exposure of the experimental animals (mice) to sulfur dioxide gas, the frequency of cough of control group remains more or less constant, i.e. it varies between 57.6 ± 0.2 and 62.9 ± 0.3 (mean ± SEM).But both in case of codeine phosphate and extract of C. grandis on oral administration, frequency of cough decreased in a dose-related manner.It was found that both the standard drug and extract (at different doses) produce maximum inhibition of cough reflex at 90 min after drug administration.A dose-dependent inhibition of cough was observed with methanol extract and results were also comparable with the effect produced by codeine phosphate, a prototype antitussive agent.The results obtained with 100, 200, and 400 mg/kg doses of extract were statistically significant (p<0.01)throughout the time span of the experiment.The highest inhibition of cough (56.7%) was produced by the extract of the 400 mg/kg dose level at 90 min of the experiment, whereas codeine phosphate (40 mg/kg) showed maximum 73.0% inhibition.The percentage inhibition of cough in codeine phosphate was 68.5-70.5% between 30 min to Discussion Coughing is a normal physiological response to an irritation of the laryngo-tracheo-bronchial system caused by mechanical or chemical stimulation.It may be painful and fatiguing and require suppression by antitussive drugs.In animals, coughing has been elicited by mechanical (Tedeschi et al., 1959) or chemical irritation (Turner, 1968) and by electrical stimulation (Cavanagh et al., 1976) of tracheal mucosa or by nerve stimulation (Pickering and James, 1979).Of all these methods, chemical or mechanical stimulation are more similar to the physiological event and also the experimental models generally used in man. In the present study, the antitussive activity of methanol extract has been compared with that of codeine against coughing-induced in two different animal species by chemicals stimulation (irritant citric acid aerosol and sulphur dioxide gas) stimulation.The extracts showed marked antitussive effect.The extract showed significant inhibition of cough like the standard drug (codeine phosphate) in dose-dependent manner; thus the extract might be acting via the central nervous system, but the exact mechanism of action can not be withdrawn from the preliminary study. Conclusion The extract of C. grandis produced a significant antitussive effect and thus established the claim of using the plant as an anti-cough agent in ancient folklore medicine. Financial Support Self-funded Figure 2 : Figure 2: Cough number obtained in the presence of low (2.5%, w/v) and high (5%, w/v) concentrations of methanol extract of Coccinia grandis fruits compared to those obtained in the presence of saline (baseline) and codeine phosphate (0.03 g/mL).Statistical differences between plant extracts, standard (codeine) with baseline values; a p<0.001.Statistical differences between cough number obtained in presence of plant extracts with that of codeine; ns: non significant difference, b p<0.001.Statistical difference between both the extract concentrations; c p<0.01	2017-09-08T16:00:37.337Z	2009-04-18T00:00:00.000	{ "year": 2009, "sha1": "551a06e5ba86265bf7e47942183950ba93eba421", "oa_license": "CCBY", "oa_url": "https://www.banglajol.info/index.php/BJP/article/download/1537/2048", "oa_status": "GOLD", "pdf_src": "ScienceParsePlus", "pdf_hash": "85155d74a191cb37aa45ca20cce931bcc4ba16fe", "s2fieldsofstudy": [ "Environmental Science", "Medicine" ], "extfieldsofstudy": [ "Chemistry" ] }
16906204	pes2o/s2orc	v3-fos-license	Evidence of Extensive Homologous Recombination in the Core Genome of Rickettsia The important role of homologous recombination has been extensively demonstrated to be fundamental for genetic variation in bacterial genomes. In contrast to extracellular or facultative intracellular bacteria, obligate intracellular bacteria are considered to be less prone to recombination, especially for their core genomes. In Rickettsia, only antigen-related genes were identified to have experienced homologous recombination. In this study, we employed evolutionary genomic approaches to investigate the impact of recombination on the core genome of Rickettsia. Phylogenetic network and phylogenetic compatibility matrix analyses are clearly consistent with the hypothesis that recombination has occurred frequently during Rickettsia evolution. 28% of Rickettsia core genes (194 out of 690) are found to present the evidence of recombination under four independent statistical methods. Further functional classification shows that these recombination events occur across all functional categories, with a significant overrepresentation in the cell wall/membrane/envelope biogenesis, which may provide a molecular basis for the parasite adaptation to host immunity. This evolutionary genomic analysis provides insight into the substantial role of recombination in the evolution of the intracellular pathogenic bacteria Rickettsia. Introduction Rickettsia spp. are obligate intracellular bacteria that are classified as gram-negative bacteria and belong to alpha subgroup of proteobacteria. They often live in arthropods, such as ticks, mites, louse, and fleas and can stably coexist within the host population. Some Rickettsia species can infect humans through bites or feces of the vectors, and then cause mild to serious diseases. It has been reported that various kinds of diseases caused by Rickettsia can be found worldwide. Based on phenotypic and phylogenetic evidence [1], Rickettsia have been classified into three different groups, including the spotted fever group (e.g., R. africae, R. conorii, and R. massiliae), the typhus group (e.g., R. prowazekii and R. typhi), and the unclassified group (e.g., R. bellii). Characterizing the evolutionary mechanism of Rickettsia is a key step towards understanding its adaptation to different hosts and the design of more effective drugs [2]. Homologous recombination, involved in transferring of a specific DNA fragment from one strain into the homologous region of another strain, is considered to be fundamental for generating genetic variation and to be one of the most important evolutionary mechanisms in bacterial evolution [3]. Meanwhile, it was believed that recombination is largely restricted to extracellular or facultative intracellular bacteria, such as Escherichia coli [4], Mycobacterium [5], Helicobacter pylori [6], and Streptococcus [7]. By contrast, obligate intracellular bacteria are often considered to be less prone to recombination [8,9], and recombination cases are rarely reported. Especially, their core genomes are thought to be free of recombination. Recently, however, a completely different scenario was addressed in Wolbachia genome [10]. Based on four housekeeping genes (gltA, dnaA, ftsZ, and groEL) as well as other previously reported cases of Wolbachia recombination, Baldo et al. suggested that Wolbachia could be subject to widespread recombination. More recently, by investigating of multiple loci dispersed throughout the Chlamydia trachomatis chromosome, Gomes et al. [11] showed that recombination is widely spread in Chlamydia trachomatis. However, Rickettsia, as a group of important obligate intracellular bacteria, are often considered to have evolved with little impact of recombination [12]. Till now, only the antigens have been reported to present the evidence of recombination [13]. In this study, a number of evolutionary genomic methods have been employed to reveal the evidence of recombination in the evolution of Rickettsia. As a result, significant evidence for recombination is found among Rickettsia core genomes through phylogenetic network reconstruction and phylogenetic compatibility matrix analyses. We also found that 28% of the core genes may have been under recombination, which is unexpected for intracellular bacteria. Functional classification showed that recombination events occurred across all functional categories but with a significant bias in the genes involved in cell wall/membrane/envelope biogenesis. This evolutionary genomic analysis provided insight into the important role of recombination in Rickettsia. [14], which has been proven to be a very powerful tool in identifying orthologous families from multiple genomes, and it has been widely adopted in recent studies [15,16]. Among the identified families, only the family with oneto-one orthologous relationship was considered for further analyses. The functional category of each orthologous family was obtained by BLASTing to the COG database (http://www.ncbi.nih.gov/COG/) with an E-value of 10 −5 . Materials and Methods Multiple sequence alignment was performed at the amino acid level using the ClustalW program with the default settings [17]. The corresponding nucleotide sequence alignment was generated according to the protein alignment using the tranalign program implemented in the EMBOSS package [18]. To construct a phylogenetic compatibility matrix, all the orthologous gene sets were firstly concatenated into a single alignment. Then, the position of sequences in trees generated for each region was compared with that for all the other regions to produce a phylogenetic compatibility matrix using the TreeOrderScan program [19]. The phylogenetic network of the concatenated DNA sequences was constructed using the Neighbor-net method [20] implemented in the SplitsTree version 4.8 [21]. Identification of potential recombination evidence from each aligned orthologous group (at a nucleotide level) was performed using the RDP [22], MaxChi [23], Chimaera [24], and GENECONV [25] methods implemented in the RDP3 program [26]. These four programs have been indicated to be powerful in detecting recombination event and have been widely used by many studies [10,27]. Results and Discussion A core genome is a set of orthologous genes, which are shared by all members of a group of bacteria. Using the Markov Clustering algorithm [14], a total of 690 orthologous sets are found in the 12 Rickettsia genomes (See supplementary Table S1 in supplementary material available online at doi: 10.1155/2009/510270). This core set represents a large proportion (690/835) of the total number of ORFs in the smallest genome (R. prowazekii Madrid E) and nearly a half (690/1512) in the largest genome (R. felis URRWXCal2). In spite of the differences on the use of orthologous identification method and the number of genomes, the core set size of Rickettsia estimated here is quite similar to previous report [12]. Functional classification of these core sets of genes based on the COG database shows that they are present in all primary functions (Figure 1); the majority of which encode components of information-processing systems involved in translation, ribosomal structure, and biogenesis. However, there are also ∼100 genes being classified into the class of "Functionally unknown" or "No hit." A strategy for recombination detection is to construct a phylogenetic network for the target genes, which can generalize the phylogenetic tree through allowing the representation of conflicting signals or alternative phylogenetic histories [28]. To observe the phylogenetic incongruence among different gene fragments, all 690 core-set genes were concatenated into a single alignment. Using the Neighbornet method [20], we found that multiple recombination events have occurred in the Rickettsia lineage ( Figure 2). Meanwhile, based on the Phi test [29], it also strongly evidenced the presence of recombination in the Rickettsia's core genes (P .00 001). Another useful strategy for detection of recombination is to generate phylogenetic trees from different parts of a gene [19]. If adjacent regions yield significantly different tree topologies, then recombination may have occurred. To visualize phylogenetic incongruence of different parts of a gene, a phylogenetic compatibility matrix was generated. As shown in Figure 3, the matrix shows a considerable number of phylogeny violations among Rickettsia, indicating the wide spread of recombination. Among them, about 3.4% of the phylogeny violation frequency is above 0.6 and 11.7% from 0.4 to 0.6. To detect individual recombination event within the orthologous gene set, four recombination detection methods were used, including RDP [22], MaxChi [23], Chimaera [24], and GENECONV [25]. To ensure the reliability of our estimated results, only the recombination events that are supported by at least two independent tests were taken into account. As a result, a total of 194 core genes were found to be prone to recombination (supplementary Table S2), of which 60 gene sets were supported by four approaches and 45 gene sets could be identified by three methods. Hence, it is suggested that recombination should be prevalent than ever expected in Rickettsia's core genes. Functional classification of these recombined genes clearly shows that none of the functional categories can be free from recombination (Figure 1), including the categories of translation ribosomal structure and biogenesis category. Interestingly, recombination preferentially occurs in genes involved in cell wall/membrane/envelope biogenesis. Such an elevated rate of recombination may be related with the coevolutionary arms race between Rickettsia and their host cells and may confer a selective advantage to survive in different niches. Taken together, our results strongly suggest that recombination may have played an important role in Rickettsia evolution. Since this study is only designed to detect the evidence of recombination events during the evolution of Rickettsia, the recombination breakpoints, parental and daughter sequences, as well as the possible reasons of widespread recombination in the core genome of Rickettsia are not considered. Combined with previous reports on Wolbachia [10] and Chlamydia trachomatis [11], we may have to update the evolutionary scenario of obligate intracellular bacteria, of which they can also be subject to extensive recombination events during evolution. The identification of the genes involved in recombination may prompt further investigation into their biological functions.	2018-04-03T02:19:24.684Z	2009-05-25T00:00:00.000	{ "year": 2009, "sha1": "60694ff7fedaba83a7b6c2e440c8fb57dcb9dfe1", "oa_license": "CCBY", "oa_url": "http://downloads.hindawi.com/journals/ijg/2009/510270.pdf", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "e841ccf6d30800e67f982aef6dbf7e1332c4438d", "s2fieldsofstudy": [ "Biology", "Environmental Science" ], "extfieldsofstudy": [ "Biology", "Medicine" ] }
254625775	pes2o/s2orc	v3-fos-license	Annulation of Perimidines with 5-Alkynylpyrimidines en Route to 7-Formyl-1,3-Diazopyrenes Unusual rearrangements were shown to accompany Brønsted acid-assisted peri-annulations of 1H-perimidines with 5-alkynylpyrimidines. These transformations take different routes depending on the nature of acetylene precursor, and lead to the formation of 7-formyl-1,3-diazopyrenes. Introduction Due to their unique physicochemical properties, pyrene derivatives have emerged as some of the most privileged structures in the design of organic fluorescent materials. Pyrene motifs attract attention of many research groups studying photochemistry and molecular electronics. Examples demonstrating utilization of pyrenes in the manufacturing of organic light-emitting diodes (OLED) [1-3], organic filed-effect transistors (OFET) [2,[4][5][6], organic photo-voltaic devises (OPVs) [2,7,8], hole-conductive materials for solar cells [9], and other photoconductive covalent organic building blocks are omnipresent in literature [10]. Major advances were made in the development of pyrene-based fluorescent probes [11] for the analytical detection of copper [12,13] and other heavy metals [13], as well as picric acid [14]. These versatile synthons also possess a great intercalating ability to selectively bind to DNA in cellular nuclei [15,16]. The undisputed advantages of pyrene derivatives are outweighed by one significant drawback-their low solubility in common organic solvents-which complicates synthesis of advanced synthetic precursors for the manufacturing of photosensors and electronic devices. Another problem is associated with high carcinogenicity of these compounds and their slow metabolism, which severely limits their application in medicinal and pharmaceutical chemistry. Both issues could be addressed by incorporation of nitrogen atoms in the pyrene structure, simultaneously providing a powerful tool for the fine-tuning of photochemical and electrochemical properties of the resulting products. Our research group has a pioneering expertise in the development of synthetic methods for peri-annulation of carbo-and azacyclic compounds [17][18][19][20]. 1H-Perimidines 1 are typically employed as model substrates in these investigations, since they are characterized by increased electron density in the peri-position, making them excellent nucleophilic synthons. Reactions of 1H-perimidines with chalcones 2 [21] and pyrimidines 4 [19], which proceed in acidic media and afford derivatives of 1,3-diazapyrenes 3, deserve a special note as an expeditious one-step route to 1,3-diazopyrenes (Scheme 1). Herein, we disclose an alternative approach to 1,3-diazopyrenes 6 or 7 via annulation of 1H-perimidines 1 with 5-alkynylpyrimidines 5 (Scheme 1). This reaction allows for selective installation of the formyl group at C-7 amenable for further synthetic modifications. allows for selective installation of the formyl group at C-7 amenable for further synthetic modifications. To evaluate this idea, we carried out a reaction between 2-phenyl-1H-perimidine (1a) with 5-(hept-1-yn-1-yl)pyrimidine (5a) in methanesulfonic acid at room temperature. Contrary to our expectations, the reaction did not afford product 15. Instead, 1,3-diazapyrene 16a possessing a n-hexyl substituent at C-6 and an aldehyde moiety at C-7 was obtained as a sole isolable product in modest yield (Scheme 4). The reaction proceeded to completion consuming both starting materials 1a and 5a, but a significant amount of product decomposed, as indicated by the formation of notable amounts of polymeric tars. The same outcome was observed in the reaction of alkyne 5a with other perimidines (1b-g) affording a series of 7-formyl-1,3-diazapyrenes in low to moderate yield (Scheme 4). To evaluate this idea, we carried out a reaction between 2-phenyl-1H-perimidine (1a) with 5-(hept-1-yn-1-yl)pyrimidine (5a) in methanesulfonic acid at room temperature. Contrary to our expectations, the reaction did not afford product 15. Instead, 1,3-diazapyrene 16a possessing a n-hexyl substituent at C-6 and an aldehyde moiety at C-7 was obtained as a sole isolable product in modest yield (Scheme 4). The reaction proceeded to completion consuming both starting materials 1a and 5a, but a significant amount of product decomposed, as indicated by the formation of notable amounts of polymeric tars. The same outcome was observed in the reaction of alkyne 5a with other perimidines (1b-g) affording a series of 7-formyl-1,3-diazapyrenes in low to moderate yield (Scheme 4). Next, we explored the possibility to perform this reaction with perimidine 1a using 5-(phenylethynyl)pyrimidine (5b) as the electrophilic component. Interestingly, the reaction took a different route leading to the formation of 1,3-diazopyrene 17a bearing a benzylamine moiety at C-6 (Scheme 4). Similarly to the example above, this reaction was rather general with respect to a variety of perimidines 1a-f,g affording the corresponding 1,3-diazopyrenes 17a-f,g as sole isolable products in moderate yield. The material balance in both of these reactions was far from perfect due to significant polymerization of the products. However, in all cases, the polymers were easily separated via a simple filtration through a short path silica gel column. It was rationalized that formation of 7-formyl-1,3-diazapyrenes 16 and 17 may occur via two related cascade transformations depicted in Scheme 5. The initially produced dihydroquinazolino [6,7,8-gh]perimidines 15 (Scheme 3) are unstable under strongly acidic conditions, but their further reactivity is strongly dependent on the nature of the substituent at C-10a. n-Hexyl-substituted derivative 15a undergoes electrocyclic cleavage of the dehydropyrimidine ring to establish aromaticity of 1,3-diazapyrene core (Scheme 5). The resulting intermediate 18 bearing a masked aldehyde functionality in a form of acyclic formamidine moiety is highly susceptible to acidic hydrolysis. The removal of this protecting group should afford 7-formyl-1,3-diazapyrene 16 (Scheme 5). Benzyl-substituted analog 15b reacts via an alternative mechanistic pathway due to a much greater migratory aptitude of the benzyl group. Since 15b has four nitrogen atoms with nearly identical basicity, it can form several protonated species coexisting in a dynamic equilibrium in the strongly acidic reaction medium. One of such forms (19) is an intermediate in tautomerization of 10,10a-dihydro-(15) into 1,10a-dihydroquinazolino [6,7,8-gh]perimidine 20 (Scheme 5). Protonation of the latter triggers 1,2-migration of the benzyl moiety to N-10a to furnish 21. A significant energy release accompanying aromatization of 1,3-diazopyrene serves as a strong driving force for this transformation. Furthermore, formation of a more basic sp 3 -hybridized nitrogen atom should be favored in an acidic medium. The non-aromatic heterocyclic ring in 21 is essentially a (methyleneamino)methanamine, which collapses under acidic hydrolysis conditions to furnish the benzylamine substituent at C-6 and an aldehyde group at C-7 in product 17. Next, we explored the possibility to perform this reaction with perimidine 1a using 5-(phenylethynyl)pyrimidine (5b) as the electrophilic component. Interestingly, the reaction took a different route leading to the formation of 1,3-diazopyrene 17a bearing a benzylamine moiety at C-6 (Scheme 4). Similarly to the example above, this reaction was rather general with respect to a variety of perimidines 1a-f,g affording the corresponding 1,3-diazopyrenes 17a-f,g as sole isolable products in moderate yield. The material balance in both of these reactions was far from perfect due to significant polymerization of the products. However, in all cases, the polymers were easily separated via a simple filtration through a short path silica gel column. General The NMR spectra, 1 H, and 13 C were measured in solutions of CDCl3 or DMSO-d6 on a Bruker AVANCE-III HD instrument (at 400.40 or 100.61 MHz, respectively). The residual solvent signals were used as internal standards in DMSO-d6 (2.50 ppm for 1 H, and 40.45 ppm for 13 C nuclei) or in CDCl3 (7.26 ppm for 1 H, and 77.16 ppm for 13 C nuclei). The high-resolution mass spectra were registered with a Bruker Maxis spectrometer (electrospray ionization, in MeCN solution, using HCO2Na-HCO2H for calibration). See Supplementary Materials for NMR (Figures S1-S32) and HRMS ( Figures S33-S48) spectral charts. The IR spectra were measured on FT-IR spectrometer Shimadzu IR Affinity-1S equipped with an ATR sampling module. The melting points were measured with a Stuart SMP30 apparatus. The reaction progress and purity of isolated compounds were controlled by TLC on ALUGRAM Xtra SIL G UV 254 plates. The column chromatography was performed with Macherey Nagel Silica gel 60 (particle size: 0.063-0.2 mm). The pyrimidines were synthesized by published methods [22,23] and the synthesis of 5- General The NMR spectra, 1 H, and 13 C were measured in solutions of CDCl 3 or DMSO-d 6 on a Bruker AVANCE-III HD instrument (at 400.40 or 100.61 MHz, respectively). The residual solvent signals were used as internal standards in DMSO-d 6 (2.50 ppm for 1 H, and 40.45 ppm for 13 C nuclei) or in CDCl 3 (7.26 ppm for 1 H, and 77.16 ppm for 13 C nuclei). The high-resolution mass spectra were registered with a Bruker Maxis spectrometer (electrospray ionization, in MeCN solution, using HCO 2 Na-HCO 2 H for calibration). See Supplementary Materials for NMR (Figures S1-S32) and HRMS (Figures S33-S48) spectral charts. The IR spectra were measured on FT-IR spectrometer Shimadzu IR Affinity-1S equipped with an ATR sampling module. The melting points were measured with a Stuart SMP30 apparatus. The reaction progress and purity of isolated compounds were controlled by TLC on ALUGRAM Xtra SIL G UV 254 plates. The column chromatography was performed with Macherey Nagel Silica gel 60 (particle size: 0.063-0.2 mm). The pyrimidines were synthesized by published methods [22,23] and the synthesis of 5-ethynylpyrimidines is described in our recent report [24]. All other reagents and solvents were purchased from commercial vendors and used as received.	2022-12-14T16:17:11.213Z	2022-12-01T00:00:00.000	{ "year": 2022, "sha1": "c26a46acd224e45102ab7feb95239f9a03af1124", "oa_license": "CCBY", "oa_url": "https://www.mdpi.com/1422-0067/23/24/15657/pdf?version=1670822561", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "8296d6cb7baef73934232aef12250893b37e3508", "s2fieldsofstudy": [ "Chemistry" ], "extfieldsofstudy": [ "Medicine" ] }
124143249	pes2o/s2orc	v3-fos-license	A 2D inverse problem of predicting boiling heat transfer in a long fin A method for the determination of local values of the heat transfer coefficient on non-isothermal surfaces was analyzed on the example of a long smooth-surfaced fin made of aluminium. On the basis of the experimental data, two cases were taken into consideration: one-dimensional model for Bi < 0.1 and two-dimensional model for thicker elements. In the case when the drop in temperature over the thickness could be omitted, the rejected local values of heat fluxes were calculated from the integral of the equation describing temperature distribution on the fin. The corresponding boiling curve was plotted on the basis of temperature gradient distribution as a function of superheat. For thicker specimens, where Bi > 0.1, the problem was modelled using a 2-D heat conduction equation, for which the boundary conditions were posed on the surface observed with a thermovision camera. The ill-conditioned inverse problem was solved using a method of heat polynomials, which required validation. Introduction Heat transfer in the nucleate boiling regime, especially the heat flux density, has been a subject of interest since the 1950s. Although much research has been devoted to the problem, no satisfactory explanation of the physical phenomena accompanying the process has been found. That is mainly due to the fact that the physics of mass, momentum and energy transfer between the surface of the solid body and the wetting fluid is difficult to explain and describe. Abstract A method for the determination of local values of the heat transfer coefficient on non-isothermal surfaces was analyzed on the example of a long smooth-surfaced fin made of aluminium. On the basis of the experimental data, two cases were taken into consideration: one-dimensional model for Bi < 0.1 and two-dimensional model for thicker elements. In the case when the drop in temperature over the thickness could be omitted, the rejected local values of heat fluxes were calculated from the integral of the equation describing temperature distribution on the fin. The corresponding boiling curve was plotted on the basis of temperature gradient distribution as a function of superheat. For thicker specimens, where Bi > 0.1, the problem was modelled using a 2-D heat conduction equation, for which the boundary conditions were posed on the surface observed with a thermovision camera. The ill-conditioned inverse problem was solved using a method of heat polynomials, which required validation. List of symbols A, B Matrices a i Coefficients Bi Biot Cross-sectional area Another reason is the nucleation, growth and detachment of vapour bubbles, which, at appropriately high wall superheat, lead to a crisis called the first critical point. A wide review of physical models describing the boiling process and a means of predicting the operation of smooth surface heat exchangers were presented by Dhir [1]. In the last 80 years, a number of empirical and mechanism based correlations have been proposed for nucleate pool boiling. Empirical correlations differ from each other substantially with respect to the functional form of the heat flux density depending on fluid and surface properties. Additionally, all of them require knowledge of the density of active sites, bubble diameter, its frequency of departure, etc. However, because of the complex nature of the coexisting processes, it has not been possible to develop comprehensive formulas for these parameters. The experimental data concerning boiling heat transfer described in the literature focus on isothermal surfaces. This is primarily due to the measuring technique, which involves mounting a specimen on the heating element. The power is determined directly by measuring the electricity supplied, as done by Ahn et al. [2,3] and Naphon et al. [4], or indirectly by applying a gradient method including measurement of temperature at several points along the specimen axis, as done by Ahn et al. [5]. The final result of the experiment is a boiling curve, which shows the relationship between the heat flux density and the heat transfer coefficient as a function of wall superheat. The generalized results are illustrated in the form of correlations for the Nusselt number. In the case of nucleate boiling heat transfer on smooth surfaces, the Rohsenow equation is used, and values of the required constants are selected for various surface-fluid combinations, Pioro [6] and Pioro et al. [7]. The factors involved in boiling heat transfer include the type, state, shape and orientation of the surface, the type of the fluid, possibly the amount of foreign particles, etc. However, the correlation relationships determined by different researchers even under the same or similar geometric and material conditions will differ. For instance, Kolev [8] analyzes a number of different correlation equations describing heat transfer for water boiling at ambient pressure on smooth surfaces. The values of the heat flux density calculated by means of these relationships differ, which can be explained by the fact that the measurements were conducted under different conditions and the surfaces differed in state. Sixteen different correlations to predict the boiling heat transfer coefficient are analyzed and compared to evaluate their use in evaporator design by Adib and Vasseur [9]. From this review, it appears that it is difficult to predict the proper value of the heat flux on the basis of the existing formulas found in the literature. A large number of correlations available in the literature are seldom satisfactory. Using the available literature and experimental test results, some researchers try to find a universal correlation for the boiling heat transfer coefficients, as Shah [10]. Although they make progress in prediction accuracy, so far such attempts are still not satisfactory for general purposes. Physical phenomena accompanying nucleate boiling are also considered theoretically. A 2D numerical investigation of boiling with a energy balance simulation is applied by Orzechowski et al. [11]. Measured and numerically calculated values of heat transfer coefficient demonstrate good congruence, which indicates a correct selection of measurement methods and a cogent analysis of the experimental results. The application of the heat transfer theory to pool boiling is discussed by Roh [12]. The study provides an explicit illustration of the kinetics of phase transition mechanisms at the interface of two different phase materials. Considerable divergence of results discussed in the literature indicates that the designed heat transfer elements should be measured separately. This is particularly essential in the case of non-isothermal surfaces, where the boiling heat transfer processes differ significantly from those observed on surfaces with uniformly distributed temperature, as shown by Orzechowski [13]. A typical example, discussed by Krikkis [14], Liaw et al. [15] and Gonzalez-Fernandez et al. [16], is a fin, on the surface of which different boiling regimes can be distinguished. Similar conclusions are sought analytically by Turkyilmazoglu [17] and Kundu et al. [18]. They determine the performance of different fin geometries by analyzing the temperature-dependent thermal conductivity of the fin material, the variable heat transfer coefficient under different boiling regimes or surface conditions. A original technique for determining the transient heat transfer coefficients from measured temperatures inside solids is presented in [19]. Numerical examples are given as verification of the method. Maintaining the heat transfer rate and the desired operating temperature is an important task when a reliable operation of machinery components is to be provided. The proper design of the heat exchangers can produce an efficient solution to the problem. Chen et al. [20] applied an inverse method together with FLUENT software and the experimental temperature distribution to investigate the non-uniform heat transfer coefficient on the fin in the heat sink for various fin spacings. Torabi and Aziz [21] investigated the thermal performance of T-shaped fins. They accounted for temperature-dependent thermal conductivity, surface emissivity, convection heat transfer coefficient, and non-zero convection and radiation sink temperatures. The problem was solved numerically. Konda Reddy et al. [22] presented a methodology for the estimation of temperaturedependent heat transfer coefficient for a vertical rectangular fin operating in a steady state under natural convection regime. Heat transfer coefficient was considered as a power 1 3 law function of temperature excess. Thermal performance of variable diameter micro-pin fins was calculated by Dίez et al. [23]. On the basis of an exact solution of smooth pin fins for three selected profiles (hyperbolic, trapezoidal and concave parabolic ones), the approximate series solution was given. The effect of surface roughness on efficiency increase was evaluated. Three different types of fins made of aluminium and copper, at a fixed fin array volume were discussed by Huang et al. [24] while looking for the most efficient solutions to heat sinks. Results obtained by using the Levenberge-Marquardt method to solve 3-D inverse problem were based on the numerical experiments. Estimation of unknown parameters in a rectangular fin satisfying a predefined temperature are discussed by Das et al. [25]. The study is useful for selecting various unknowns in a conductive-convective and radiative fin for required temperature and efficiency. Grysa et al. [26] presented an approximate method of solving nonlinear inverse stationary heat conduction identification of boundary condition. The obtained results were compared with the direct solution of the problem. A numerical study of the inverse problem of determining time-variable surface heat flux in a plane wall, with constant or temperature dependent thermal properties, is given by Zueco et al. [27]. Direct and inverse heat conduction problems are solved with the network simulation method. Both the accuracy and effectiveness of the method has been discussed. Considerations such as those mentioned above, aim to determine the performance and optimum design of fin assemblies. The aim of this study was to derive an efficient algorithm for the determination of local values of the boiling heat transfer coefficient using a specimen made of aluminium. Experimental procedure and measurements Investigations were conducted at the test facility, specially arranged for the task. The diagram of the facility is presented in Fig. 1. The experiment involved a long fin-shaped specimen with a width of 12 mm, an active length of 90 mm and a different thickness, which was placed between two textolite plates. One of the specimen surfaces was in contact with the boiling fluid, and the other was surrounded by quiescent atmospheric air. The test facility was equipped with two independent heating systems. The main heating element, whose power was controlled with an autotransformer, was located at the fin base. The auxiliary heating system, made of spirally coiled resistance wire, was positioned on the bottom of the vessel. A constant supply of the electric power to the auxiliary system was maintained throughout all measurement series. The task of the auxiliary system was to sustain the boiling process in the whole volume of the liquid. Supplying power to the auxiliary system was preceded by a measurement intended to examine the effect produced by the power supply on the results of investigations. Within the whole range of the power supply, no such effect was observed. The test facility was equipped with a cooling and condensate recovery system. Vapours of the boiling liquid condensed in the system of two tap water-cooled condensers, and then flowed, by gravity, into the vessel. The outer surface of the examined element was observed with a thermal imaging camera, as shown in Fig. 2. Since high emissivity of the observed surface is required to reduce the uncertainty of the IR temperature measurement, the outer surface was coated in black paint. The optical homogeneity of the observed surface and the emission coefficient in the camera spectral range were determined during the initial calibration process. The experiments were performed using a VIGO System V50 camera, operating in the longwave infrared range (8-11 μm), equipped with a 384 × 288 microbolometer FPA (Focal Plane Array), whose sensitivity at a temperature of 30 °C is <0.08 K. The main heater was mounted at the base of the specimen, where the predetermined time-invariant electric power was supplied. The heater outer surface was shielded so that the emitted radiation would not affect the temperature measurement. The tests were conducted for refrigerant FC 72 at ambient pressure, which after evaporation was directed onto the cooling system in order to recover the condensate. The specimen had to be adequately fixed so that the temperature field observed with the thermal imaging camera could be treated as a one-dimensional quantity. The result of the measurement was the distribution of temperature along the length for the central line of the fin. The test was carried out during refrigerant FC 72 boiling at ambient pressure on a nonisothermal smooth surface of an aluminium fin. Exemplary measurement results for a smooth surface specimen 3 mm in thickness are shown in Fig. 2. The results were obtained for three different heating powers supplied to the fin base. Measurements of the electrical power applied to the main heater do not reflect the heat dissipated by the fin. It is thus difficult to estimate the amount of heat lost to the surroundings. Moreover, the changes are uneven as they are dependent on the changes in load. It can be assumed that for sufficiently thin specimens the Biot number is small, i.e. Bi < 0.1, since there is no change in temperature with thickness, it can be omitted and the total heat conducted at the base of the fin Q b can be calculated using Fourier's law. The necessary temperature gradient can be determined from the measurement data shown in Fig. 2. The same experimental set-up was used for a 12 mm thick smooth specimen made of aluminium. It had identical chemical composition and surface morphology as the 3 mm specimen. The electrical power supplied to the main heater was adjusted to keep the temperature of the outer surface at the fin base as close as possible to that reported for the thinner sample. Figure 3 shows the axial temperature distributions in the two fins. Direct observation of the boiling process at the fin surface shows that for the two samples heat transfer at the base occurs in the fully developed nucleate boiling regime in the vicinity of the critical point. The heat transfer coefficient for nucleate boiling is frequently described by the Rohsenow correlation, which indicates its exponential dependence on the wall superheat [6]. In this case, the total amount of heat dissipated by the fin with a constant cross-section is proportional to the square root of the surface at the base, i.e. [28]). The heat transfer surfaces of both fins are of equal dimensions, have the same spatial orientation and morphology, and fins differ only in thickness. With the assumption of constant temperature in the fin crosssection, the amount of heat dissipated by the 12 mm thick fin should be twice as high as the amount of heat dissipated by the 3 mm thick fin. Thus, according to Fourier's law, the derivative of the temperature at the base should be twice as small as that for the thicker fin. However, the values measured were different; they were −2412.8 and −1519.1 K/m for the thinner and thicker fins, respectively, while their ratio was 0.63. This large discrepancy indicates that the 1-D analysis cannot be used for fins with a relatively large thickness. Thus, the problem must be solved numerically. Problem formulation and the method of solution The distribution of temperature on the outer surface measured with the thermovision camera is the evidence of the heat transfer taking place on the surface invisible to the camera. As shown in Fig. 3 the surface temperature distributions along the centrelines of fins of 3 and 12 mm, however, are not the same, which results only from different thicknesses of the samples. The different slopes of the temperature distribution for both specimens indicate that a thicker specimen does not satisfy the Biot number condition and 1-D approximation cannot be applied automatically. One-dimensional inverse problem The method for determining a local value of the heat transfer coefficient over a non-isothermal fin surface can be used only with one-dimensional systems, i.e. ones where the drop in temperature over the thickness can be omitted. This assumption is true for numbers Bi < 0.1. At nucleate boiling regime, for extended surfaces, the heat transfer coefficients reach values even higher than 0.1 MW/m 2 K. Thus, using a one-dimensional model is possible only for elements with a small thickness that are made of high thermal conductivity materials. For such large values of heat transfer coefficients, a copper fin <0.4 mm in thickness is required to strictly satisfy the Biot number criterion. Boiling heat transfer of FC72 on the smooth surface of the aluminium sample requires that the sample thickness should be a little below 4 mm. The performance of a straight fin with a constant crosssection is analysed on the basis of one-dimensional expression for the temperature distribution: where m 1 2 = P/λF, with P denoting the perimeter of the fin cross-section and F the cross-sectional area, λ thermal conductivity, q α local value of the heat flux rejected by the fin lateral side, α heat transfer coefficient. In inverse problems, the temperature distribution is also used to determine the heat transfer coefficient and the heat flux density. The right-hand side of Eq. 1 is dependent only on the wall superheat. Thus, the integration gives where C is an integration constant calculated, if necessary, from the boundary condition. Assuming that the result of integration of the right-hand side of Eq. (2) is described by the holomorphic function, its value can be represented by a power series as shown in (2). Differentiation of both sides of Eq. (2) is replaced by the following expression for the heat flux density: where m 2 1 is defined in Eq. (1) while coefficients b i result from polynomial fit (see Fig. 5). Formulation and solution of a 2-D ill-posed problem In the case of thicker fins, where the differences in temperature between the outer and inner surfaces are large, the assumption of a one-dimensional model leads to considerable errors. A good solution is to apply a method of noninvasive measurement to determine the temperature field on the accessible surface and calculate the heat transfer surface dθ dx area. Then, the local values of the heat transfer coefficient and heat flux are established. The conditions of the measurement described above require that the heat transfer coefficients should be determined on the basis of at least the two-dimensional Laplace equation for OABC area shown in Fig. 4. Two of the boundary conditions on the side observed by the thermovision camera, i.e. the temperature distribution and the dissipated heat flux, are generally easy to estimate. The other boundary conditions need to be determined individually. Thus, that constitutes not only an inverse but also an ill-posed problem. The problem described in this study was solved using heat harmonic polynomials (sometimes called heat polynomials). This method, proposed by Ciałkowski et al. [29], makes it necessary to find an exact solution in the form of an approximation satisfying the two-dimensional heat conduction equation and approximately satisfying the boundary conditions. For a fin operating in the system presented in Fig. 4, the modelling was based on the stationary two-dimensional heat conduction equation, θ the wall superheat. It is expected that the solution to θ(x,y) should be looked for in the OABC area. Heat is removed from the side surface, BC, at pool boiling of the surrounding medium. The aim of the calculations is to determine some local values of the heat flux density of this energy. The four boundary conditions required to solve the problem are: • for the thermovision camera OA side (see Fig. 4): (4) where x m is the coordinate of the experimental points. Condition (5) represents the distribution of discrete surface temperature measured along the central line of the fin, OA, at M points with the x m coordinates. Additionally, it is assumed that the heat flux rejected to the surroundings along OA is negligible in comparison with the flux generated from the BC surface. The phenomenon of boiling heat transfer is observed along the BC side of the fin, while the OA edge is in contact with quiescent atmospheric air. The heat transfer coefficient is several orders of magnitude lower for OA in comparison with BC. The estimations show that for refrigerant boiling at atmospheric pressure the heat flux dissipated by OA is <1 % of the heat flux dissipated by BC, which confirms the correctness of assumption (6). Moreover, it is a condition of the fin central symmetry. • The temperature distribution in the fin cross-section at the base, OC, is assumed to be constant: • For the fin top (the AB side in Fig. 4), it is assumed that: where θ b is the wall superheat at the fin base. Satisfying condition (8) requires using a fin with a sufficient length, L. It should be noted that the condition on the BC side is not known, whereas for y = 0, two conditions are given. The problem analyzed here is not only inverse but also ill-posed. To solve it, a method of resolving functions (in this case, some kind of polynomials for the local temperature) needs to be employed. For a given differential equation, it is always necessary to determine a generating function dependent on the parameter p. In accordance with [29], the following function was used: because it satisfies the differential Eq. (4). The coefficients of the Taylor series expansion of the above function relative to the parameter p are the desired polynomials solving the differential equation. The series expansion of the exponential function (9) is given by the following formula: This gives two sequences of harmonic polynomials, the socalled heat polynomials, one being a real part and the other an imaginary part of the complex function, (x + iy) j : The same order polynomials u and w are used interchangeably to obtain a sequence of harmonic polynomials V j , each of which satisfies the Laplace equation and can be treated as a particular solution of Eq. (4): Its general solution is approximated by means of a linear combination of the heat polynomials, which constitute the approximation Θ(x,y) of the exact solution having the form of: c j coefficients, N degree of the polynomial approximating the exact solution. Expression (13) satisfies the Laplace Eq. (4); in a general case, however, it does not exactly satisfy the boundary conditions. One can only demand that the acceptable error be as negligible as possible. The condition assumed in this way allows us to select unknown coefficients c j . It is thus essential to express the error functional as: where M is the total number of measured points.The following terms of the above error functional refer to the differences in boundary conditions (5)-(8) and value of function (13) which represents the approximation of the exact solution of Laplace Eq. (4).All components required in the above expression should be determined from the assumed form of the approximation (13). Then, by differentiating functional (14) in relation to all coefficients c j , the following system of linear equations is obtained: Thus, the problem is solved by applying the N + where The elements of matrices A and B are defined by the following formulae: As a result, the desired coefficients of the linear combination of functions (12) are obtained, which, introduced in Eq. (13), form an approximation of the exact solution. By using the obtained relationships, the thermal field in the analyzed area can be determined. Method validation The temperature measured on the outer surface OA (see Fig. 4) provides information on the processes occurring on the inner surface of the fin, not accessible to the thermovision camera. As a result, the temperature measured is different from the actual one occurring during intensive heat transfer. The thermal field on the fin surface is measured at many points with a thermovision camera. That makes it possible to apply typical procedures of smoothing and then numerical differentiation available in a number of calculation packages. For the 3 mm thick aluminium fin, the results of the analysis of the measurement data from Fig. 2 are shown in Fig. 5. Using the polynomial approximation to the data given in Fig. 5, all coefficients of Eq. (2) could be determined and then the local heat fluxes q α and heat transfer coefficients could also be established. The corresponding boiling curve for the investigated surface is drawn in Fig. 7. Here and hereafter, the boiling curve is defined locally, i.e. for the local values of heat flux and heat transfer coefficients from superheat. Detailed analysis of the method and its uncertainties are discussed by Orzechowski [28]. The major factor affecting the proper determination of the heat flux density is the limited accuracy of the measuring device. The smaller are the differences between the temperature values, the greater effect of the measurement accuracy is noted. In this study, a regime around the critical heat flux and free convection boiling was analyzed. The determination of the heat flux density or heat transfer coefficient from error-prone measurement data is very difficult and requires validation of the method, especially if the problems are ill-posed. The correctness of the developed algorithm is generally verified by a numerical experiment, in which a similar problem is solved analytically. A more reliable method should use, if possible, the experimental data. For that reason, an additional independent test was conducted at the same test facility using the same devices and under the same conditions. The experiments involved measuring the distribution of temperature on the aluminium specimen with a thickness of 3 mm. For aluminium, whose conductivity is very high, and, for the small thickness of the specimen, the condition Bi < 0.1 was well satisfied nearly over the entire wall superheat range, and the heat transfer could be treated as a one-dimensional problem. The boiling curve was plotted using the method described in Sect. 3, and the results were compared with analogous data using the proposed method of heat polynomials. Figure 6 shows the temperature measured on the outer surface of the aluminium fin and the temperature calculated for the inverse ill-posed problem using the method of heat polynomials with the boundary conditions (5)-(8). In Fig. 6b, a difference in the measured and calculated temperature is shown. Its mean value is close to zero, and nearly all points are located in the range ±0.2 K. Only a few points are positioned outside that range. The probable reason is the local change in the emissivity of the paint, but even with this, the standard deviation for this set is low and equals 0.102 K. The proposed numerical method used to solve the inverse problem allows us, for instance, to determine the surface temperature and the local heat flux as seen in Fig. 7 as the function of wall superheat in FC-72 boiling on the aluminium fin at ambient pressure. For the sake of comparison, a similar relationship devised using a 1-D model is shown. It is also worth noting that in this particular range, the maximum Bi ≈ 0.08 is found at the wall superheat ~20 K. As can be seen from the two figures above, the good convergence of the calculation results obtained with two different methods using the same experimental data confirms that the method proposed for solving the ill-conditioned inverse problem is correct. Similarly, the relationship between the local heat flux removed from the surface and the wall superheat is determined from the distribution of temperature measured at the outer surface of the 12 mm fin, according to the 1-D procedure. This leads to an exponential dependence of q on θ, as shown by the straight line in Fig. 7. The constant exponent can then be determined from its slope. The exponent value 1.52 indicates the presence of only nucleate boiling on the whole fin surface. Moreover, this exponent is twice as small as that given in the Rohsenow correlation [6]. The boiling curves plotted for the two specimens in accordance with the 1-D procedure differ in shape and value. This falsely suggests that the physical phenomena present during the phase change heat transfer differ for the two specimens made of the same material and having the same surface morphology. Since this is not true, proper determination of the boiling heat transfer processes in thick fins requires using a numerical method, which, even for illposed problems, gives correct results. Discussion The local values of heat flux density derived in Sect. 4 can be used to calculate the heat transfer coefficient for each wall superheat. As its maximum value is about 4800 W/ m 2 K, the Biot number for the 3 mm thick fin is smaller than 0.1. For a fin with the same physical parameters but Fig. 6 Results obtained for a 3 mm thick aluminium fin: a measured and calculated wall superheats at the outer OA and inner CB lines (see Fig. 4), θ(x,0) and θ(x,h) respectively, b temperature differences between the measured and calculated points at the surface viewed through a thermovision camera (y = 0) Fig. 7 Measured and calculated heat flux for a thin aluminium fin a greater thickness, e.g. 12 mm, the temperature changes through the thickness cannot be neglected while determining the heat transfer conditions. The Biot number for the 12 mm thick fin is about 0.32, so the problem must be at least two-dimensional. Systems with an easy-to-observe surface can have the thermal field measured; the boiling heat transfer coefficient can be calculated using the algorithm described in Sect. 3.2. For the distribution of temperature measured at the fin surface (Fig. 3), it was essential to assume the boundary conditions (5)- (8), develop an approximation of Θ(x,y), and calculate the coefficients c j from the best fit condition. The function determined in this way describes the thermal field in the OABC area (see Fig. 4). Figure 8 presents the measured and calculated values of the wall superheat for two surfaces of the 12 mm thick fin, i.e. y = 0 and y = h (see Fig. 4). By comparison, during boiling heat transfer, e.g. by natural convection or by forced convection with no phase change, large amounts of heat flux are removed. The largest are in the nucleate boiling regime. As shown in Fig. 7, for the smooth aluminium sample with a thickness of 12 mm, the process starts at a wall superheat of about 8 K at a distance of about 3 cm from the base (see Fig. 8). With a rise in the wall superheat in the nucleate boiling regime, there is an increase in the heat transfer coefficient. Heat flux removal increases locally, resulting in an increase in the difference in temperature between the outer surface (y = 0) and the inner surface (y = h). The value exceeds 5 K at points located near the base of the 12 mm thick fin. This fact cannot be ignored in the analysis of heat transfer in fins with such a geometry. Figure 9 shows a gradient of the wall superheat over the outer surfaces (y = 0 and y = h) along the fin length. The calculated values result from the assumed boundary conditions. It should be noted that a good convergence of the measurement results with the assumed values is found. However, some disagreement with Eq. 7 is observed for the temperature at the base of the fin. As can be seen from Fig. 10, the distribution of the wall superheat for x = 0 changes with the height. The variability of this quantity is a result of a very high heat transfer coefficient, and it seems to correct the assumed boundary condition. Figure 10 shows the numerically calculated distribution of temperature in the 12 mm thick fin in selected cross-sections. There is a significant change in temperature in the fin cross section to the height of about 30 mm from the base. Near the surface, where nucleate boiling exists, the temperature drops considerably. Beyond this area, the decrease in temperature is insignificant. In the numerical calculations, the approximation of the exact solution (13) is employed to determine the thermal field, which can be used to calculate the temperature gradients. In accordance with the Fourier law, this gradient, multiplied by the coefficient of heat conductivity, determines the heat flux density. On the other hand, the one calculated in the perpendicular direction on the inner edge (y = h) is equal to the boiling heat transfer on the fin: (19) q α = − ∂Θ ∂y y=h and α = q α Θ The calculation result in the perpendicular direction on the inner edge (y = h) of the fin is presented in the form of a boiling curve in Fig. 11. The boiling curve is defined locally, i.e. for the local values of heat flux and heat transfer coefficients from wall superheat. The temperatures are taken at the inner surfaces of the fins. For the 3 mm thick fin, the temperature values at the inner and outer surfaces are nearly the same. For the thicker fin, the difference is significant, as shown in Figs. 8 and 10. The total heat dissipation for the two fins is given in Figs. 2 and 3, respectively. The analysis of heat transfer for refrigerant FC 72 boiling at ambient pressure over an aluminium fin shows that two steady boiling regimes on the surface: nucleate boiling and free convection are found. The boiling curve can be divided into three characteristic regimes. In the low superheat regime (up to about 6.5 K), free convection at a nearly constant value of the heat transfer coefficient is observed. In the range of the wall superheat higher than 7 K, two subregimes can be distinguished. When the wall superheat is in the range from about 7 K to ~8.5 K, single bubbles lift off the surface and cause more intensive mixing of the fluid close to the fin surface. When the wall superheat grows above 9 K, more nucleation sites become active and the nucleation boiling regime occurs. As a result, a rapid rise in the heat flux density and the heat transfer coefficient is observed. Due to the very high coefficient of heat transfer in the nucleate boiling regime, in order to ensure a good representation of one-dimensional heat conduction, it is necessary to prepare the fin of a material with high thermal conductivity and low thinness. Both of those affect the value of the Biot number, which should be as small as possible. Therefore, studies were performed for aluminium fins with a thickness of 3 mm, for which at the maximum heat transfer coefficient, Bi ≈ 0.08 and gradually decreases with drop of the local heat transfer coefficient. Figure 6a shows the wall superheat along the height of the fin. The figure presents measurement result and two lines for y = 0 and y = h, which were determined numerically by the given algorithm. From the comparison of the temperature difference from the side of thermovision camera shown in Fig. 6b, the errors of the method could be drawn out. From the set under consideration, it can be seen that almost all values are within the range ±0.2 K and the standard deviation equals 0.102 K. For the fin made of aluminium with a thickness of 12 mm, temperature distribution in its cross OABC was calculated (see Fig. 4), and the results are shown in Figs. 8 and 10. Temperature difference between inner and outer surfaces at the base of the fin is considerable and reaches slightly above 5.5 K. The difference of these values below 0.5 K occurs only at a distance greater than 35 mm from the base, where free convection boiling is observed. Figure 9 shows the temperature gradients. Its value in the direction of the y-axis is almost zero, which indicates the correct adoption of the boundary condition (6). The other two curves in this figure are the result of good agreement of measured and calculated temperatures at the surface, as discussed above. For the commercial surface with lots of microcavities, the nucleate boiling regime starts at a wall superheat of about 9 K and, as more nucleation sites become active, and grows rapidly with the superheat increase, as shown in the boiling curve in Fig. 11. For superheat above 8 K, the heat transfer coefficient gradually increases and its maximum is about ~6.2 kW/m 2 K at the wall superheat of 20 K. The increased generation of vapor causes bubble interference and coalescence, and the vapour escapes in forms of jets or columns, which subsequently merge into slugs. This is accompanied by a much slower increase in the heat transfer coefficient and at the wall superheat above ~23.5 K, a consequent decrease in the heat flux density is observed. Conclusions The aim of the experimental research was to measure the local values of the heat flux and the boiling heat transfer coefficient on a non-isothermal fin surface. The thermal conductivity of aluminium under nucleate boiling conditions is relatively high, but 12 mm thickness of the fin makes the Biot number much higher than 0.1. The study involved comparing two fins with the same surface morphology but different thicknesses. For the thinner fin, the Biot number was smaller than 0.1, while for the thicker fin the number was much greater. The boiling curves plotted for the two specimens in accordance with the 1-D procedure described in Sect. 2.1 are clearly different from each other in shape and value (see Fig. 7). The Fig. 11 Boiling curve for FC 72 at ambient pressure on an aluminium fin maximum difference is more than 31 % at a wall superheat of 18 K. Since it is impossible to use a 1-D model of heat transfer for such a system, the calculations were conducted using a two-dimensional heat conduction equation, which is an ill-conditioned problem due to the measurement method applied. To validate the method of calculation, a series of measurements was performed, whose purpose was to estimate the local heat transfer coefficient under conditions as close as possible to one-dimensional temperature distribution and to compare those results with the analogous ones obtained using the method in the inverse problem. The performance of 1-D fin with a constant cross-section is analysed on the basis of a one-dimensional expression for the temperature distribution. With an assumption that the outcome of integration is an holomorphic function, its value is represented by a power series. By differentiating this expression, functional dependences for the local heat fluxes and heat transfer coefficients are obtained. In this way the corresponding boiling curve for the investigated surface is given and thoroughly discussed in the paper. The results of numerical calculations are the temperature values inside the area and on the invisible side (see Fig. 10). They are described by the formula (13). On this basis, the local heat flux density and corresponding heat transfer coefficient are evaluated, as shown in Figs. 7 and 11. There, a correlation between the values calculated assuming a one-dimensional model of heat transport phenomena in the fin is also presented. Visible reduction in heat transfer coefficient in the high heat fluxes range is the result of polynomial approximation used in the model 1-D and cannot be here the basis for inference about the location of the first critical point of boiling. To sum up, the analysis performed shows that the proposed method provides accurate results coincident with 1D ones. The obtained results of this study show that the method proposed is accurate and efficient for solving the engineering heat transfer problems where the thermal field could be measured and corresponding heat fluxes are sought. 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.	2019-04-21T13:06:31.611Z	2016-10-01T00:00:00.000	{ "year": 2016, "sha1": "b88c6a06a58992333cdf56d26e7e6fe7748b3e13", "oa_license": "CCBY", "oa_url": "https://link.springer.com/content/pdf/10.1007/s00231-015-1733-x.pdf", "oa_status": "HYBRID", "pdf_src": "Adhoc", "pdf_hash": "037c1eeb51037b0d3576ac7e47ba7611f2d64c41", "s2fieldsofstudy": [ "Engineering", "Physics" ], "extfieldsofstudy": [ "Materials Science" ] }
7876451	pes2o/s2orc	v3-fos-license	Fourth-order Contour Mode Zno-on-soi Disk Resonators for Mass Sensing Applications In this work, we have investigated the design, fabrication and testing of ZnO-on-SOI fourth-order contour mode disk resonators for mass sensing applications. This study aims to unveil the possibility for real-time practical mass sensing applications by using high-Q ZnO-on-SOI contour-mode resonators while taking into account their unique modal characteristics. Through focused ion beam (FIB) direct-write metal deposition techniques, the effects of localized mass loading on the surface of three extensional mode devices have been investigated. Ten microfabricated 40 m-radius disk resonators, which all have a 20 m-thick silicon device layer and 1 m-thick ZnO transducer layer but varied anchor widths and numbers, have exhibited resonant frequencies ranging from 84.9 MHz to 86.7 MHz with Q factors exceeding 6000 (in air) and 10,000 (in vacuum), respectively. It has been found that the added mass at the nodal locations leads to noticeable Q-factor degradation along with lower induced frequency drift, thereby resulting in reduced mass sensitivity. All three measured devices have shown a mass sensitivity of ~1.17 Hz· fg −1 at the maximum displacement points with less than 33.3 ppm of deviation in term of fractional frequency change. This mass sensitivity is significantly higher than 0.334 Hz· fg −1 at the nodal points. Moreover, the limit of detection (LOD) for this resonant mass sensor was determined to be 367 ag and 1290 ag (1 ag = 10 −18 g) for loaded mass at the maximum and minimum displacement points, accordingly. Introduction Miniaturized ultrasensitive devices for medical point-of-care and industrial portable systems are in huge demand. Devices such as MEMS/NEMS mass sensors have rapidly evolved to a point that has enabled us to weigh single atoms with unparalleled precision [1]. In order to widely deploy these devices for real-world consumer applications, comprehensive studies need to be performed to understand the technological limits of every design. The MEMS/NEMS devices have great potential for a broad spectrum of ultrasensitive mass sensing applications, such as label-free monitoring of biological interactions [2], detection of gas molecules and volatile organic compounds [3], thickness measurement of the deposited thin films, and so on. The working principle of this type of MEMS/NEMS resonant sensor is based upon measureable infinitesimal mechanical perturbation converted into the electrical domain via a variety of transduction mechanisms such as electrostatic, piezoelectric, and optical detection [4]. Some of the most widely studied resonant mass sensors to date are quartz crystal microbalance (QCM), surface acoustic wave (SAW) resonators, and capacitively transduced NEMS/MEMS resonators [5]. However, these technologies still face a number of challenges that limit their ultimate performance in mass sensing applications and seamless integration with CMOS IC technology including relatively low quality factors, low resonance frequencies, lack of scalability, need of vacuum environment, high motional resistance, and use of complex fabrication techniques [6][7][8]. Lateral extensional mode ZnO-on-SOI MEMS resonators are an alternative and emerging technology that has promised a well-balanced solution to most of the aforementioned problems. These devices were initially developed for wireless transceivers because they have a great potential to satisfy the very stringent specifications of wireless communications standards such as low insertion loss, high frequency selectivity and low phase noise [9]. In the past few years, a great deal of effort has been devoted to study the performance characteristics of all the aforementioned resonant sensing technologies by using direct write techniques like focus ion beam (FIB) deposition, as well as other physical deposition techniques such as e-beam evaporation and sputtering. With these techniques, several research groups have reported the most important sensing parameters such as the mass sensitivity and resolution as summarized in Table 1. In this work, we present the sensitivity analysis of mass loading at both maximum and minimum displacement points of the fourth-order lateral-extensional mode of a ZnO-on-SOI disk resonator. The devices have been fabricated on a silicon-on-insulator (SOI) wafer by using a recently developed 5-photomask CMOS compatible batch microfabrication process [14]. In particular, ten piezoelectrically-transduced 40 m-radius ZnO-on-SOI disk resonators operating in the fourth-order contour mode have exhibited resonance frequencies of roughly 86 MHz with measured Q factors exceeding 6000 (in air) and 10,000 (in vacuum). Three of the devices have been tested as mass sensors which demonstrated a sensitivity of 1.17 Hz· fg −1 and 0.334 Hz· fg −1 (1 fg = 10 −15 g) induced by the loaded mass at the maximum and minimum displacement points accordingly. Aside from the reduced sensitivity, it is observed that the loaded mass on the nodal points also leads to significant Q-factor degradation, which greatly affects the overall performance of the device in terms of increased motional impedance and lowered resolution. On the contrary, the loaded mass on the maximum displacement locations has led to much more repeatable and linear mass sensitivity and with less than 33.3 ppm of deviation in term of mass-loading induced fractional frequency change. The systematic investigation of the localized mass loading effect by this work is expected to expand our knowledge base and greatly improve our understanding of the behavior of the resonant mass sensor while providing crucial design criteria for the implementation of binding sites for optimum sensor performance. Design and Operation The ZnO-on-SOI lateral-extensional mode resonator was equipped with two split top electrodes in order to match the strain field patterns of the fundamental and the fourth-order extensional contour modes while operating in a 2-port configuration as seen in Figure 1. When the added infinitesimal mass is much smaller than the device's equivalent dynamic mass at the corresponding location, the mass sensitivity of the device can be related to the induced frequency shift, which is proportion to the ratio between the mode frequency and the equivalent mass given by: where ∆ is the measurable resonance frequency shift, ∆ is the change due to the loaded mass, 4 is the resonance frequency of the fourth-order contour mode and is the equivalent dynamic mass of the resonator at the attachment location of the loaded mass. It is worthwhile mentioning that the fourth-order contour mode of the device was employed because it offers the highest frequency-Q product for this electrode and anchor configuration for disk resonators with radius smaller than 100 µm . A shown in Figure 2a,b, only two resonant modes can be efficiently actuated based on the electrode and disk resonator mode shape design, which are the first-order and the fourth-order contour modes. As shown in Figure 1, the correlations between the electrode shape, size of the resonant body and the position of the anchors versus the modal strain fields strongly influence the electromechanical coupling and the anchor-related energy dissipation, the fourth-order contour resonant mode is anticipated to provide the highest frequency-Q product. As shown in Figure 2b, the first four contour modes are identified for a 88 µm -radius disk resonator from its measured wideband frequency response with frequency range spanning from 30 kHz to 50 MHz. The resonant characteristics of these contour modes were explored by COMSOL® modal simulation using an equivalent acoustic velocity = 7.153 m/s found from Equation (3) for the ZnO-on-SOI disk resonator. As seen in Table 2, the simulated resonant frequencies for the first four contour modes of a 88 m-radius disk resonator match closely with the measured values, which verifies their coexistence and superior frequency-Q product of the fourth contour mode studied by this work. For the resonant mass sensor application, which monitor mass-loading induced frequency shift, both high frequency and high Q can be leveraged to achieve high sensitivity and high resolution simultaneously. The required radius for the disk resonator for the desired resonance frequency at this corresponding 4th order contour mode can be obtained by [15]: where 4 is the frequency scaling factor related to the fourth-order resonant mode ( 4 = 0.493), is the radius of the resonator disk and is the equivalent acoustic velocity. For the ZnO-on-SOI resonator made of stacked structural layers, the equivalent acoustic velocity can be expressed as: where , , , are the equivalent Young's modulus, density, Poisson's ratio and the thickness of the corresponding layer of the stacked structural materials, including top/bottom metal electrode layer, ZnO piezoelectric transducer layer and silicon device layer. It is important to note that the equivalent dynamic mass of the resonator body tends to be much smaller than its physical mass, as different infinitesimal elements contribute based on the level of their kinetic energies. In another word, the regions adjacent to the nodal locations make significantly smaller contributions as compared to those of maximum displacement areas. Hence, the equivalent dynamic mass of the disk-shaped resonator body can be approximated by dividing its total kinetic energy by one-half of the square velocity at a specific circumferential location given by [16]: where 1 is the Bessel function of the first kind, Rdisk is the radius of the resonator disk, 4 is the forth-order angular resonant frequency. The equivalent stiffness of the resonator can be determined by: As the device goes into resonance in a 2-port configuration, the electromechanical coupling efficiency between the electrical and acoustic energies at the terminals of the two split top electrodes can be calculated by the total charge induced through the piezoelectric effects underneath the electrodes at the maximum displacement region. The electromechanical coupling coefficient can be expressed as: where 31 is the transverse piezoelectric strain coefficient, which equals to −4.7 pC/N for sputtered ZnO thin films [17]. In order to model the behavior of a two-port piezoelectrically-transduced MEMS disk resonator, a lumped element circuit model as shown in Figure 3 is employed with all the crucial parameters extracted from the measured characteristics. The two transformers represent electromechanical signal conversion by the piezoelectric transducers at the input and output ports, while the capacitors Cod and Cos represent the static stray capacitances of the driving and sensing electrodes. The series LCR electrical equivalent circuit can be derived from the corresponding mechanical model elements (i.e., stiffness, mass, damping) using Equations (7)- (9) to depict the frequency-domain mechanical resonance behavior of the resonating body. The substrate feedthrough capacitance Cf accounts for the RF signal leakage between the input and output electrodes through the piezoelectric transducer layer and the resonator body. The element values of the lumped-element LCR equivalent circuit can be expressed as follow: where mre, Kre represent the equivalent mass, equivalent stiffness, and unloaded quality factor at the fourth-order resonant mode, respectively. The electromechanical coupling coefficient, ηre, represents the energy conversion efficiency between the electrical and mechanical domains at both input and output terminals of the two-port resonator device. Device Microfabrication Processes The sensors of this work were microfabricated through a five UV photo-mask process, starting with a SOI wafer with a highly doped 20 μm-thick silicon device layer, by using the cleanroom facility of the Nanotechnology Research and Education Center (NREC) at the University of South Florida. The process begins with a pre-releasing step by using a 49% HF solution to undercut the device regions through strategically placed release holes, which were defined by dry etching through the silicon device layer. Then, the 200 nm-thick platinum bottom electrodes are deposited and patterned by lift-off as shown in Figure 4a, followed by the reactive sputtering of a 500 nm-thick c-axis aligned ZnO film using a AJA Orion 5 magnetron RF sputtering system (Figure 4b). The sputtering process parameters of ZnO thin films have been optimized to obtain a strong preferential c-axis orientation along the film normal direction (002) along with increased deposition rate of ~250 nm/h. In situ post-process annealing at 400 °C has been performed for 1 h with Ar:O2 (1:1) forming gases at 30 mTorr of pressure under a vacuum environment. The via-hole contacts to the bottom electrodes are subsequently opened by a wet chemical etching of ZnO in a mixture of 1:100 HCl:H2O, followed by sputter deposition and lift-off patterning of the 200 nm-thick platinum top electrodes (Figure 4c). Thereafter, the body of the resonator is then defined by a sequential reactive ion etching (RIE) and deep reactive ion etching (DRIE) of stacked ZnO and silicon device layers (Figure 4d). The devices are then protected with a layer of photoresist and diced into individual chips. Finally, the batch-fabricated devices are released by a sequential acetone and methanol solvent rinsing followed by critical CO2 drying. Table 3. Measured frequency characteristics of ten disk resonator devices operating at the fourth-order contour mode, which all have a 40 μm radius, 20 μm-thick silicon device layer and 5 μm-long supporting anchors but different anchor numbers and widths. The measured results of ten 40 m-radius disk resonator devices operating at the fourth-order contour mode under ambient conditions are shown in Table 3. We can observe that there is a small variation of the loaded Q factors and the resonant frequencies partially due to discrepancies of thin-film depositions and UV contact photolithography. Specifically, physical vapor deposition systems such as the e-beam evaporator and sputtering systems used in this work tend to have limited thickness uniformity throughout a 10 cm diameter substrate. Similarly, the lamp of UV contact lithographic exposure systems lamps often produce a light intensity with limited uniformity thus causing geometry variation across different areas of the wafer. Despite the limited fabrication tolerance of a university cleanroom facility, the yield of these devices is greater than 90% and the variations of resonant frequency and loaded Q factor among ten measured devices are less than 1.9 MHz and 1000, respectively. It is worth mentioning that these devices have different numbers of supporting anchors strategically positioned at nodal points together with their anchor width ranging from 3 m to 6 m. Even though the radius of the resonator body plays a paramount role in setting the resonance frequency, a fraction of the observed variations of frequencies and Q factors amongst these devices can be ascribed to the differences in anchor sizes and numbers. Results and Discussion After the initial on-wafer probing test, the diced chip was then mounted onto a printed circuit board (PCB) chip carrier while the device under test was connected to the external ports using a ball wire bonder. The device was tested and the electrical parameters were obtained as shown in Figure 5. As shown, the anti-resonance peak observed by the two port ground-signal-ground (GSG) on-wafer probing measurement is mitigated by the extra parasitics introduced by the PCB chip carrier and wire bonds. Using the lumped-element equivalent circuit model as shown in Figure 3, the element values of the LCR electrical characteristics based on the on-chip probed and wire-bonded measurements were obtained as seen in the indented table in Figure 5. The device was then tested inside a vacuum chamber to provide a controlled environment (Figure 6a). The frequency responses of the devices were captured in the form of S-parameters using a vector network analyzer. As shown in Figure 6b, the device has exhibited a resonant frequency of 85.4 MHz and a loaded Q factor of 6967 and 10,638 in air and in vacuum, respectively. The measured sensitivity of this device is 1.17 Hz· fg −1 in response to the loaded mass at its maximum displacement locations, which is on par with the theoretically extrapolated value of 1.379 Hz· fg −1 for a model-predicted equivalent dynamic mass of 31.1 ng based on Equation (3). The same device has exhibited more than 3X lower mass-loading sensitivity of 0.334 Hz· fg −1 at the minimum displacement locations adjacent to the nodal points. The sensitivity and effects of mass loading versus modal location on the surface of the contour mode resonator were thoroughly explored through a series of depositions that lead to well-controlled infinitesimal increment of the loaded mass at the maximum and minimum displacement locations. Figure 5. Frequency responses of a fourth-order contour-mode disk resonator measured using an on-chip ground-signal-ground (GSG) probing and a PCB chip carrier with bond wires. In this work, depositions of micro-pellets composed of platinum-gallium alloy composite were performed on predetermined areas (i.e., maximum or minimum displacement locations) using a Quanta 200 3D dual beam focused ion beam (FIB) equipped with a Gas Injection System (GIS). The FIB platinum-gallium deposition process was calibrated until repeatable micro-pellets were obtained by using a low beam current of 10 pA to mitigate the gallium contamination as shown in Figure 7b. It is important to note that even after a thorough calibration, the sizes, shapes and compositions of micro-pellets formed by FIB tend to vary slightly due to the change in FIB chamber conditions, which is a source of uncertainty for the mass sensitivity characterization in this work. By using C5H5Pt(CH3)3 as an organometallic precursor, the typical FIB deposition leads to a micro-pellet that consists of C (45%-55%), O (5%), Pt (40%-50%), and Ga (5%-7%) [18]. As shown in Figure 7a, the FIB deposited micro-pellets were characterized using a high-resolution atomic force microscope (AFM), which revealed a nominal volume of 0.665 μm 3 corresponding to 7 pg of effective mass based on the PT/Ga alloy density of 10.2 g· cm −3 . Figure 8a, four consecutive micro-pellet depositions were performed to produce a discrete increment of 7 pg per micro-pellet as the extra loaded mass at the maximum displacement points near the disk circumference, which yield an average mass sensitivity of 1.17 Hz· fg −1 . To minimize the disturbance of the lateral external mode by the mass-loading effect, the micro-pellets were added one by one on the opposite sides of the disk resonator to retain the balance. As shown in Figure 8c,d, the extra loaded masses introduced by the micro-pellets in this fashion made negligible impacts on the Q factor and the effective motional impedance while resulting in consistent resonant frequency drop for each micro-pellet deposited. On the opposite side, it can be observed that when the loaded mass by the micro-pellets were added at locations adjacent to the nodal points of the disk resonator in its fourth contour mode as shown in Figure 9a, the mass loading significantly affects the energy dissipation and therefore the Q factor and motional impedance the device while producing much less severe resonance frequency shift and poor mass sensitivity of 0.334 Hz· fg −1 , thus negatively impacting the achievable resolution of the contour-mode resonator based mass sensor. Aside from the sensitivity, another important performance metric for a resonant mass sensor is the linearity of its output signal in response to the infinitesimal amount of loaded mass. By introducing FIB deposited identically-sized micro-pellets at strategically-chosen maximum displacement locations, highly repeatable frequency shifts linearly proportional to the loaded mass have been achieved while retaining the other key performance parameters, such as Q factor and motional impedance. Meanwhile, the binding site at which the external analyte is attached to the resonator body is proven to play a crucial role in affecting the effective sensitivity, linearity and repeatability in the resonant response of the lateral extensional mode sensor. For instance, when the micro-pellets were deposited adjacent to central and circumferential nodal points as seen in Figure 9a, we observed a much more nonlinear mass-loading induced frequency shift along with noticeable Q-factor degradation and motional impedance variation. We attribute this to the increased energy dissipation due to the higher energy losses for an unbalanced and distorted resonating microstructure. To the best of our knowledge, this is the first time such an effect has been observed because previous works have only focused on exploring a single region of the maximum displacement area [11,12,19]. As seen in Figure 9c,d, the total loaded mass of 35 pg from 5 FIB micro-pellets near the central nodal points (i.e., location 5-9) has essentially reduced the loaded Q factor from 10,000 to merely 6000 along with an abrupt increase of the motional impedance from 3.5 kΩ to 4.5 kΩ. As shown in The subsequent FIB depositions of two more micro-pellets at location 10 and 11 near the circumferential nodal points as seen in Figure 9a have led to slightly higher frequency shifts but also inconsistent behavior. We observed that the micro-pellet deposited at location 10 disrupts the resonant mode to cause damping, Q-factor degradation and an increased motional impedance of 5.1 kΩ. We also observed that the second micro-pellet at location 11 recovered the balance of the modal vibration thus bringing the Q-factor and motional impedance back to their nominal values of 6000 and 4.5 kΩ prior to these FIB micro-pellet depositions. In some of the prior works, the entire surface of the resonator body was employed as the binding site in an indiscriminate fashion for weighing target analytes, which could lead to severe uncertainty and degraded sensitivity as evidenced by the measurement results shown in Figures 8 and 9 [20,21]. Based on our new findings, it is clear that the strategic design of the binding location of target analytes should be deemed a crucial step for adapting contour mode MEMS resonators for practical mass sensing applications. To evaluate the repeatability of these mass sensor devices, two more devices were tested as presented in Figure 10, where four micro-pellets were deposited at the maximum displacement areas of the fourth-order contour mode. Figure 10 presents the fractional frequency changes for all three sensors as a function of the loaded mass from FIB deposited micro-pellets at their corresponding maximum displacement points. As seen in Figure 10a, the sensors demonstrated consistent and close-to-linear mass-loading induced frequency shifts with very small variations based upon the counter balance effect due to the inaccurate placements of the micro-pellet deposition. As shown in Figure 10b, a device-to-device discrepancy less than 33.3 ppm was observed, which can be largely attributed to the fabrication tolerance (device-to-device geometry variation) and the placement precision of each added micro-pellets with respect to modal locations of the desired resonance mode. In addition, the composition of the micro-pellets could change slightly as a consequence of variation of the FIB chamber conditions. The mass resolution of the resonant sensor can be obtained by determining the short-term resolution through the slope of the phase at the resonance given by [19]: where ∂φ ∂ \| 0 is the slope of the phase at the resonance frequency and ∆φ is the zero span phase noise. The measured zero span phase noise at the nominal resonance was 0.042° and the slope at resonance was calculated to be 0.09717°· Hz −1 as shown in Figure 11. Therefore, the short term frequency noise for the sensor was estimated to be 0.43 Hz. Based on this value, the point mass sensitivity of this device can be estimated to be 367 ag and 1290 ag at the maximum and minimum displacement areas, respectively. It is important to note here that the Q factor plays a crucial role on the stability of the sensor signal. As the resonator Q degrades as a result mass loading on the nodal points, so will the slope of the phase at resonance, which results in reduced mass resolution. Conclusion We have investigated the design, fabrication and testing of ZnO-on-SOI fourth-order contour mode disk resonators for mass sensing applications. Using direct-write metal deposition techniques, the effects of localized mass loading on the surface of the device have been explored. The ten microfabricated 40 m-radius disk resonators have exhibited resonant frequencies between 84.9 MHz and 86.7 MHz with Q factors exceeding 6000 (in air) and 10,000 (in vacuum), respectively. It was observed that the location of the FIB-deposited micro-pellets with respect to the mode shape of the resonator body strongly affects the sensor output characteristics. In addition, the repeatability of the devices was demonstrated where the frequency shift due to added mass was close-to-linear and almost identical across the three sensors. The three measured devices have shown a consistent mass sensitivity of 1.17 Hz· fg −1 at the maximum displacement points with deviation of less than 33.3 ppm in term of fractional frequency change. On the other hand, much lower mass sensitivity of 0.334 Hz· fg −1 has been observed at the nodal points. Also, the limit of detection (LOD) for this resonant mass sensor was determined to be 367 ag and 1290 ag (1 ag = 10 −18 g) for loaded mass at the maximum and minimum displacement points, respectively. This indicates that resonant sensors developed from this technology have sensitivities and mass resolutions largely on par with those of the best reported mass sensors, as shown in Table 1. The loaded mass by the analytes adjacent to nodal locations severely impacted the basic sensor performance metrics, resulting in significant Q-factor degradation and an increase of motional impedance. On the contrary, the mass loading at maximum displacement locations mainly induced a resonance frequency shift rather than influencing other output characteristics such as Q-factor, insertion loss and motional impedance. It is therefore important to carry out resonant sensor design and implementation while mapping the analyte attachment location against the actual resonator mode shape. In essence, this study has verified that microfabricated high-Q ZnO-on-SOI contour-mode resonators are a viable candidate for implementing ultrasensitive sensors for a range of real-time practical mass sensing applications.	2015-09-18T23:22:04.000Z	2016-09-23T00:00:00.000	{ "year": 2016, "sha1": "4069e639037fdc3afaee01229b693ff5427b66a8", "oa_license": "CCBY", "oa_url": "https://www.mdpi.com/2076-0825/5/4/24/pdf", "oa_status": "GOLD", "pdf_src": "Crawler", "pdf_hash": "01a0423000007bc691d679d2ddbcc28b0d433e60", "s2fieldsofstudy": [ "Engineering", "Materials Science" ], "extfieldsofstudy": [ "Computer Science" ] }
104941797	pes2o/s2orc	v3-fos-license	Phytochemical Screening , LC-MS Studies and Antidiabetic Potential of Methanol Extracts of Seed Shells of Archidendron bubalinum ( Jack ) Background: Some Malaysia and Indonesia people believed that root and seed shell of Archidendron bubalinum can treat diabetes. However, seed shell of Archidendron bubalinum has not yet to be scientifically proven and confirmed their ability to treat diabetes. The study of the potential of this seed shell was also scarcely available. Objective: The recent work was aimed to investigate the phytochemical screening of methanol extracts of seed shells of Archidendron bubalinum and to evaluate their chemical compositions and antidiabetic activities. Material and Methods: The methods of phytochemical screening were including alkaloids, flavonoids, tannins, polyphenols, saponins, and terpenoids. Their chemical compositions were determined by Liquid Chromatography-Mass Spectrometry (LC-MS) and antidiabetic activities were performed by α-glucosidase inhibitory method. Results: The phytochemical screening showed that methanol extracts of seed shells of Archidendron bubalinum contain flavonoids, tannins, polyphenols, and terpenoids. This extracts exhibited antidiabetic activity with IC50 7.77 μg/mL. This result was supported by LC-MS analysis which showed the presence of phlorizin and astilbin, in which these compounds had high inhibitory activity against α-glucosidase or diabetes. Conclusion: LC-MS analysis revealed the presence of polyphenol compounds namely phlorizin and astilbin in which had high α-glucosidase inhibitory activity, might largely contribute in the antidiabetic activity. middle income countries, including China, India, Indonesia, Pakistan, Egypt and Mexico. 3Diabetes is a well-known metabolic disorder, which is characterized by an abnormal postprandial increase of blood glucose level.The control of postprandial hyperglycemia is believed to be important in the treatment of diabetes.α-glucosidase secreted from intestinal chorionic epithelium is responsible for the degradation of carbohydrates.α-glucosidase inhibitors slow down the process of digestion and absorption of carbohydrates by competitively blocking the activity of α-glucosidase.Consequently, the peak concentration of postprandial blood glucose is reduced and the blood sugar level comes under control. 4Some of α-glucosidase inhibitors, such as acarbose that obtained from natural sources, can effectively control blood glucose levels after food intake and have been used clinically in the treatment of diabetes. 5Only a few α-glucosidase inhibitors are commercially available.All of them contain sugar moieties and their synthesis involves tedious multistep procedures.Moreover, clinically they have been associated with serious gastrointestinal side effects.Therefore, it is necessary to search for alternatives that can display α-glucosidase inhibitory activity but without side reactions. 6-glucosidase inhibitors can be used as a new class of antidiabetic drug.411 compounds exhibiting α-glucosidase inhibitors activity were summarized and isolated them from medical plants.The compound classes isolated include: 61 terpenes, 37 alkaloids, 49 quinines, 103 flavonoids, 37 phenols, 73 phenylpropanoids, 8 steroids, and 43 other types of compounds.] The species Archidendron bubalinum belongs to the family Fabaceae or Leguminosae in the major groups of Angiosperms.It contains about 126 species and is a large and economically important family of flowering plants.Fabaceae is the most common family found in tropical rainforests and in dry forests. 9The species is indigenous to Sumatra in Indonesia, Peninsular Malaysia and Thailand.The vernacular name is kabau or julang jaling (Indonesia), kerdas (Malaysia) and nieng-nok (Thailand).The strong pungent smelling seeds have an odor like jering (Archidendron jiringa) and petai (Parkia speciosa).The bark of Archidendron bubalinum is used as febrifuge.Young seeds can be eaten in raw as traditional vegetable salad and can be used traditionally as a diuretic but excessive consumption can lead to poisonous and kidney problem. 10 Archidendron bubalinum seeds were found to have high moisture content, low fat content, high protein content but being limited in valine, methionine and tyrosine. 11Archidendron bubalinum contained of anti-nutrient: tannins, trypsin inhibitors and hemaglutinin.Studies in Thailand have been reported that crude extracts of Archidendron bubalinum seeds had superoxide scavenging activities. 124] The study of the potential of this seed is scarcely available and little attention is given due to limited resources.Archidendron bubalinum has not yet to be scientifically proven and confirmed their ability and chemical constituents to treat diabetes. 14he objective of this study is to identify the types of compounds found in the methanol extracts from seed shells of Archidendron bubalinum from Lampung Indonesia and to investigate its potential antidiabetic activity. MATERIALS AND METHODS The seeds of Archidendron bubalinum were collected from uncultivated farm lands located at Lampung, Sumatera Island, one of province in Indonesia.The seeds were peeled to separate the seed shells and seed (Figure 1).The seed shells were collected, thoroughly washed and then sun dried for 4 h.The dried seed shells were soaked in methanol with ratio 1:10 (w/v) and extracted by maceration methods for 24 h.Then, the supernatant was filtered by filter paper.Soaking process was repeated once again in the same sample and the supernatant was filtered.All the supernatant were collected together and concentrated using rotary evaporator at 50 o C. Alkaloids determination About 100 mg of methanol extracts was mixed in 5 ml of 10% hydrochloric acid then pH was adjusted to 8 with dilute ammonia solution.The mixture was added with 20 ml chloroform and then dried.The extracts were tested with Dragendorff reagent and Mayer's reagent.The formation of precipitate was indicating the presence of alkaloids. Flavonoids determination About 100 mg of methanol extracts was boiled in 100 ml of distilled water for 5 min and filtered.5 ml of the filtrate was mixed with 0.1 mg magnesium powder, 1 ml concentrated hydrochloric acid and 1 ml amyl alcohol then shaked vigorously.The colour changed to yellow or red or orange, indicating the presence of flavonoid. Terpenoids and steroids determination About 100 mg of methanol extracts were mixed with 25 ml diethyl ether and shaked vigorously.The diethyl ether layer was separated then added 2-3 drops of Lieberman-Burchard reagents.A bluish colour of the interface was formed, indicating positive results for the presence of terpenoids while greenish colour, indicating positive results for the presence of steroids. Saponins determination About 100 mg of methanol extracts were added with 10 ml of boiled distilled water, cooled and shaked vigorously for 10 min.The foam formation was observed and will be stable if added with a few drops of hydrochloric acid, indicating the presence of saponins. Tannins and polyphenols determination About 100 mg of methanol extracts were extracted with 1 ml ethanol and 1 ml distilled water.The filtrate was mixed with a few drops of 1% FeCl 3 and into another filtrate added with gelatin salt.Then observe the colour changing, if a green or blue or purple colour was appeared, indicating the presence of tannins. Liquid Chromatography-Mass Spectrometry (LC-MS) studies The chemical constituents of the methanol extracts were determined using LC-MS.LC-MS analysis was performed using Mariner Bio spectrometry equipped with a binary pump.The HPLC was interfaced with a Q-TOF mass spectrometer fitted with an ESI source.Full-scan mode from m/z 100 to 1200 was performed with a source temperature of 140°C.HPLC column Phenomenex 5μ C8, (150 × 2 mm i.d.) was used for the analysis.Solvent was methanol with 0.3% formic acid.Solvents were delivered at a total flow rate of 0.1 mL/min.The solvent was run by isocratic elution.The MS spectra were acquired in the positive ion mode.The temperature of the drying gas (N 2 ) was 350 o C, at a gas flow rate of 6 mL/min, and a nebulizing pressure (N 2 ) of 25 psi.About 0.5 g of sample extracts was diluted with methanol and filtered with 0.22 µm nylon filter S79 prior to analysis.A 5 µl volume of the extracts were injected onto the analytical column for analysis.The mass fragmentations were identified by using spectrum database for organic compounds in SDBS application. α-Glucosidase inhibition test 16,17 A 250 μL solution of p-nitrophenyl-α-D-glucopyranoside 5 mM and 495 μL phosphate buffer 0.1M pH 7 was added to the reaction tube containing 5 μL of the sample solution in DMSO with a concentrations variation of 100, 50, 25, and 10 μg/mL.After homogeneous, the solution was pre incubated for 5 min at 37°C, the reaction was initiated by the addition of 250 μL α-glucosidase solution (0.062 units), incubation was continued for 15 min.The reaction was stopped by the addition of 1 ml of Na 2 CO 3 0.2 M. The activity of the enzyme was measured, based on the reading of p-nitrophenol absorbance at λ 400 nm.Quersetin was used as a reference standard with concentration of 10, 7.5, 5, 2.5, and 1 µg/mL.% inhibition was measured by using the equation: Phytochemical screening The phytochemical screening of methanol extracts of Archidendron bubalinum seed shells revealed the presence of flavonoids, tannins, total phenols, and terpenoids, while alkaloids and were not detected (Table 1).These results are the same as in the husks sample analysis.However, this extract provides a different bioactivity, in which this extract gives positive results against antidiabetes (Table 3).This is interesting so it proceeds with the analysis of phytoconstituents using LC-MS Liquid Chromatography-Mass Spectrometry (LC-MS) studies LC-MS analysis of methanol extracts of Archidendron bubalinum seed shells had detected five peaks with the retention time 0.41, 1.28, 1.90, 2,63, and 2.80 minutes (Figure 2).Then each peak was fragmented, resulting 5 fragmentation spectrum with candidates mass (m/z) 365, 436, 436, 247 and 450.(Figure 3 and Table 2) The LC-MS spectrums interpretation was performed using a spectrum database for organic compounds in SDBS application.The results of spectrum interpretation on methanol extracts of Archidendron bubalinum seed shells, indicating that there are substance of phlorizin and astilbin with the retention time of 1.90 and 2.80, respectively.These results were confirmed by each of the fragmentation pattern as figure-4 and figure-5 Based on the results of antidiabetic test in Table 3, Figure 6, it can be seen that the methanol extracts of Archidendron bubalinum seed shells had high antidiabetic activity with IC 50 of 7.77+0.11µg/mL.This value was almost same as IC 50 of quercetin that was 6.04+0.14µg/mL.] Phlorizin is a natural product and dietary constituent found in a number of fruit trees.It has been used as a pharmaceutical and tool for physiology research for over 150 years. 21Both the phlorizin which was alkaloid compound, as well as astilbin have many phenolic groups in their molecules.Polyphenols were often present in polar glycosides and were easily soluble in polar solvents.The antioxidant and antidiabetic properties of these compounds were related to the presence of phenolic groups that can donate hydrogen atoms to a free radical so that the radicals become less reactive.The antidiabetic mechanism of methanol extracts of Archidendron bubalinum seed shells depended on the content of the compounds therein.Methanol extracts of Archidendron bubalinum seed shells was useful as an antidiabetic through several mechanisms: 1. Methanol extracts of Archidendron bubalinum seed shells were rich in flavonoids and phenolic compounds such as phlorozin and astilbin.Phenolic compounds can donate their hydrogen atoms and function as free radical inhibitors, where phenolic compounds will protect organs such as the pancreas from free radicals attack.(Figure 7) 2. Another mechanism was by competitively inhibiting α-glycosidase activity, these inhibitors help to prevent the fast breakdown of sugars and thereby control the blood sugar level. CONCLUSION Phytochemical screening showed that methanol extracts from Archidendron bubalinum seed shells containing many compounds either flavonoids or phenolic compounds.Based on the LC-MS analysis which was interpreted by SDBS application, it can be identified that some of the compounds in the methanol extracts were phlorizin and astilbin.From the results of antidiabetic test, it was proved that methanol extracts of Archidendron bubalinum seed shells had high antidiabetic activity with IC 50 of 7.77+0.11µg/mL which was almost same with IC 50 of quercetin standard, 6.04+0.11µg/mL.High antidiabetic activity of methanol extracts of Archidendron bubalinum seed shells were coming from the compounds of phlorizin and astilbin.From these results, it can be concluded that methanol extracts of Archidendron bubalinum seed shells had great potential as biomedicine for diabetes diseases. mentioned. 18 The spectrum of fragmentation mass of phlorizin was written by m/z 436 of molecular ion of C 21 H 24 O 10 resulted from phlorizin substance when it was exposed with 70 eV of energy.Molecular ion of C 21 H 24 O 10 was experiencing fragmentation and McLafferty reorganization by releasing C 6 H 10 O 5 and producing fragments by m/z 274 originating from C 15 H 14 O 5 .(Figure 4) The mass spectrum of astilbin was written by m/z 450 originating from the molecular ion of C 21 H 22 O 11 which was formed when it was exposed by 70 eV of electron.Molecular ion of C 21 H 22 O 11 was experiencing fragmentation and McLafferty reorganization by releasing C 6 H 10 O 5 and producing fragments by m/z 288 originating from C 15 H 12 O 6 .The pattern of fragmentation of astilbin can be seen in Figure 5. Figure 6 : Figure 6: Graph of the relationship between concentration and % inhibition for IC 50 determination.A) Quercetin B) Methanol extracts. Figure 7 : Figure 7: Reaction of phenolic compounds with free radicals.	2019-04-10T13:12:35.486Z	2018-11-12T00:00:00.000	{ "year": 2018, "sha1": "e81c923d956ac1e61fd55004d9f0624e0f768700", "oa_license": "CCBY", "oa_url": "https://www.phcogj.com/sites/default/files/PharmacognJ-10-6s-77.pdf", "oa_status": "HYBRID", "pdf_src": "Anansi", "pdf_hash": "e81c923d956ac1e61fd55004d9f0624e0f768700", "s2fieldsofstudy": [ "Medicine", "Environmental Science" ], "extfieldsofstudy": [ "Materials Science" ] }
119226480	pes2o/s2orc	v3-fos-license	The magnetic moments of the proton and the antiproton Recent exciting progress in the preparation and manipulation of the motional quantum states of a single trapped proton enabled the first direct detection of the particle's spin state. Based on this success the proton magnetic moment $\mu_p$ was measured with ppm precision in a Penning trap with a superimposed magnetic field inhomogeneity. An improvement by an additional factor of 1000 in precision is possible by application of the so-called double Penning trap technique. In a recent paper we reported the first demonstration of this method with a single trapped proton, which is a major step towards the first direct high-precision measurement of $\mu_p$. The techniques required for the proton can be directly applied to measure the antiproton magnetic moment $\mu_{\bar{p}}$. An improvement in precision of $\mu_{\bar{p}}$ by more than three orders of magnitude becomes possible, which will provide one of the most sensitive tests of CPT invariance. To achieve this research goal we are currently setting up the Baryon Antibaryon Symmetry Experiment (BASE) at the antiproton decelerator (AD) of CERN. The local relativistic quantum field theories involved in the Standard Model of particle physics are CPT invariant [1]. As a consequence, in collisions particles and their antimatter mirror images annihilate completely. This is in strong contradiction to the striking imbalance of matter and antimatter observed in our Universe, which inspires comparisons of the fundamental properties of particles and their antiparticles with highest precision [2]. Any measured difference would shed light on this contradiction [3]. In BASE [4,5] we focus on the measurements of the magnetic moments of the proton µ p and the antiproton µp with ppb precision or better. We store a single (anti)proton in a Penning trap [6] and measure the ratio of its cyclotron frequency ν c = 1/(2π) · (e/m p ) · B and its spin precession (Larmor) frequency ν L = g/2 · ν c . B is the magnetic field of the Penning trap, e/m the particle's charge-to-mass ratio and g the so-called g-factor. The measured frequency ratio gives the dimensionless magnetic moment µ p,p in units of the nuclear magneton µ N = (eh)/(2m). The cyclotron frequency ν c is measured via image current detection [7], while ν L is determined by using the continuous Stern-Gerlach effect [8]. In this detection scheme the spin magnetic moment is coupled to the particle's axial oscillation frequency ν z , which is achieved by superimposing a so-called magnetic bottle B z = B 2 z 2 − ρ 2 /2 to the homogeneous magnetic field B of the trap, where B 2 characterizes the strength of the magnetic inhomogeneity and z, ρ are cylindrical coordinates. A spin quantum jump shifts ν z by Therefore, from a sequence of alternating axial frequency measurements and resonantly induced spin quantum transitions, the spin flip probability as a function of the spin-flip drive frequency ν rf is obtained, and thus the Larmor frequency ν L [9]. This principle has been applied with great success in the widely recognized measurements of the electron and the positron magnetic moments [10] providing one of the most stringent tests of CPT symmetry. However, the (anti)proton magnetic moment is 658 times smaller than that of the electron, which constitutes a significant challenge. Typical axial frequencies in Penning traps are in the order of 1 MHz and the axial frequency shift induced by a spin flip is ∆ν z,SF ≈ 0.4 µHz·B 2 . Thus, to resolve single proton spin flips we developed a Penning trap with B 2 ≈ 300 000 T/m 2 . With this apparatus we reported the first observation of proton spin-transitions [11] and measured the particle's magnetic moment with a relative precision of 8.9 ppm [12]. By applying the same method, an independent research group reported a precision of 2.5 ppm for the proton [13] and 4.4 ppm for the antiproton [14]. All these experiments were carried out in a Penning trap with the superimposed magnetic bottle. However, this produces an unavoidable line broadening of the spin resonance, ultimately limiting experimental precision to the ppm level. To overcome this problem, a group at Mainz developed the double Penning trap technique [15]. In this elegant scheme the precision measurement of ν c and ν L , and the detection of the spin state are separated to two traps, a precision trap (PT) with a homogeneous magnetic field, and an analysis trap (AT) with the strong magnetic bottle. In the BASE experiment the magnetic field in the precision trap is about 75 000 times more homogeneous than that in the analysis trap which reduces the width of the spin-line significantly, thus enabling high-precision measurements. By applying this technique in measurements of the g-factor of the electron bound to highly charged ions experimental precisions below the ppb level were achieved [16,17,18,19]. Our double Penning trap is shown in Fig. 1. The cylindrical electrodes are made out of oxygen-free copper. They are gold-plated to prevent oxidation. The precision trap has an inner diameter of 7 mm while that of the analysis trap is 3.6 mm. This small diameter is required to produce the strong magnetic bottle by a central ring electrode which is made of ferromagnetic Co/Fe material. The two traps are connected by transport electrodes. Adequate voltage ramps applied to these electrodes are used to shuttle the (anti)proton adiabatically between the traps. Radio frequency (rf) drives applied to small coils mounted close to each trap produce oscillating magnetic fields which are used to induce spin quantum transitions [20]. The whole setup is placed in a hermetically sealed vacuum chamber with a volume of about 1 l, which is cooled to 4 K. When cooled to cryogenic temperatures in this volume typically background pressures in the order of 10 −16 mbar are achieved. This prevents particle loss due to charge exchange or annihilation with the background gas. To measure the particle's eigenfrequencies highly sensitive detection systems are used [21]. They consist of coils with inductance L and high quality factors Q, coupled to ultra lownoise amplifiers. Together with the trap capacitance C p they form parallel tuned circuits with resonance frequencies ν res . Once tuned to resonance with the particle's respective oscillation frequency ν k , image currents induced in the trap electrodes by the particle's oscillation lead to a voltage drop across the effective detection resistor R p = 2πν k LQ. The signals are amplified by cryogenic ultra low-noise amplifiers and from a fast Fourier transform (FFT) the particle's resonance frequencies are obtained. By interaction with R p the particle is cooled resistively. The cooling time constant is where D is a trap specific length, 7.6 mm in the PT and about 14 mm in the AT. Once cooled to thermal equilibrium, the particle shorts the thermal Johnson noise of the detector [22], resulting in a dip in the FFT-spectrum. In this case the particle's eigenfrequency is obtained from a best fit to the data. We obtain an axial frequency resolution of 60 mHz in a measurement time of about 60 s. For the application of the double-trap technique we prepare a single proton in the AT and determine its spin state (see Fig. 2). Next, we transport the single proton to the precision trap, and while a spin-flip drive is applied, the cyclotron frequency is measured. In a last step the particle is transported back to the AT and the spin state is analyzed. It is obvious that for the application of this technique "single spin flip resolution" is required [23]. This means, that the spin state of the proton has to be identified unambiguously. However, this is difficult. The strong magnetic bottle B 2 couples in addition to the spin magnetic moment as well the magnetic moment of the orbital angular magnetic momentum of the radial modes to the axial oscillation frequency. Thus, small fluctuations of the magnetron and the cyclotron energy cause axial frequency fluctuations. A cyclotron quantum transition ∆n + = ±1 causes an axial frequency shift of ∆ν z = ±63 mHz and a transition of the magnetron quantum number ∆n − = ±1 leads to ∆ν z = ±54 µHz. Thus, to clearly observe spin transitions it is crucial to keep axial frequency fluctuations induced by radial quantum transitions small in comparison to the frequency shift ∆ν z,SF = 171 mHz. As an example we consider in the further discussion cyclotron quantum jumps. They are due to electric dipole transitions driven by spurious noise on the trap electrodes. The transition rate δn + /dt is proportional to the quantum number n + , as described in detail in [23]. We measured the axial frequency fluctuation as a function of the proton's average cyclotron quantum number. The details of the calibration and preparation procedure of n + are described in [24]. From this measurement we obtained the data shown in Fig. 3. The heating-rate increases as a function of the average cyclotron quantum number, which confirms our semi-quantitative understanding of the axial frequency stability. To apply the double Penning trap technique we prepare the particle with an average cyclotron quantum number below n + ≈ 1500. There, the cyclotron transition rate is below 0.03/s. This is achieved by resistive cooling of the modified cyclotron mode in the precision trap which takes about 2 hours. Once the cyclotron quantum state is prepared, we achieve a spin flip fidelity of 75 %. This means, that three out of four detected spin transitions are correctly identified [23]. For a demonstration of the double-trap technique in the precision trap resonant spin flips are driven with a drive amplitude strong enough to saturate the spin transition. For a background measurement we applied the whole double-trap scheme, however no spin flip drive was applied in the PT. Results of these measurements are shown in Fig. 4. For the resonantly driven spin flips we obtain a spin flip probability of 50 %, while in the background measurement only 25 % are obtained. This is due to the 75 % spin flip fidelity in the AT. Under these conditions, when measuring a full Larmor resonance curve in the PT, the maximum to baseline ratio of the resulting spin line would be only about 25 %. This reduced "contrast" affects the achievable experimental precision in the g-factor resonance. A possible way to improve the contrast of the resonance is the application of a phase sensitive axial frequency measurement technique [25]. By using this method we measured the axial frequency in the precision trap within a few seconds, achieving spin flip resolution [26]. Once implemented in the analysis trap we expect to obtain spin flip fidelities of 100 %. The filled squares are measured data, the solid line is a fit. Once the proton's cyclotron quantum state is prepared below n + ≈ 1500, we achieve single spin flip resolution with a fidelity of about 75 %. To apply our methods to the antiproton we are currently setting up the BASE experiment [4,5] . Spin flip probability as a function of measuring cycles. The filled circles represent data were spin flips were driven in the PT. The filled squares represent results of a background measurement where the double-trap sequence has been applied, however no spin flips were induced in the PT. When the spin flip drive was turned on a probability of 50 % is measured, corresponding to saturation. When no spin flip drive is applied a probability of 25 % is measured, which is due to the finite fidelity. at the antiproton decelerator of CERN [27,28]. Essentially, the antiproton magnetic moment can be measured by applying exactly the same techniques as the ones described above. However loading of antiprotons into our Penning trap apparatus requires several modifications of the trap and the implementation of the experiment into the high energy physics infrastructure of CERN. A new transfer beamline for the 5.3 MeV antiprotons provided by the AD has been designed and verified by the AD group. The shielded experimental zone where the apparatus will be installed is currently being constructed. In addition to the PT and the AT which are required for the gfactor measurement a so-called catching trap as well as a degrader structure will be included into our apparatus, which is shown in Fig. 5. The degrader structure is a sequence of thin stainless steel (SUS), Be and Al foils of variable thickness. Once the foil sequence is carefully chosen and the thickness of each foil is adjusted carefully, about 1 per-mille of the incident antiprotons are decelerated to energies below 15 keV. Downstream to the degrader a specifically designed catching trap (CT) will be placed. 15 kV catching pulses can be applied to the trap electrodes to catch the degraded antiprotons. These are standard methods which were developed and optimize in the late 1980ies [29]. Once the particles are trapped, electron cooling [30] and resistive cooling will be used to prepare the 100 µeV particle energies, which are required for our high precision measurement. To become independent from accelerator shutdown and power cuts, our CT will be operated on battery based voltage sources. We plan to load several AD antiproton shots into this trap and techniques will be developed to prepare single antiprotons into the AT/PT cycle to perform the precision measurement of the antiproton magnetic moment µp. Together with the planned measurement of the ground state hyperfine splitting of antihydrogen [31], constraints on the antiproton sub-structure will be obtained. Besides the precision measurement of the antiproton magnetic moment we plan as well to improve the value of the proton to antiproton charge-to-mass ratio. This number has been measured in the late 1990ies and is known with a relative precision of about 90 ppt [32]. However, in recent years advanced mass spectrometry techniques have been developed and mass measurements at the level of 5 ppt were reported [33]. By applying these techniques the values of the proton to antiproton charge-to-mass ratio might be improved by another order of magnitude. In this paper the status of the BASE experiment was presented. By using a Penning trap with a strong superimposed magnetic inhomogeneity we detected spin flips with a single trapped proton and measured the particle's magnetic moment with a relative precision of 8.9 ppm. With an improved apparatus we achieved single spin flip resolution. Based on this success we reported on the first demonstration of the double-trap method with a single trapped proton, which is a major step towards a direct high precision measurement of µ p on the ppb-level. Currently we are setting up the BASE experiment at CERN to apply our techniques to a single trapped antiproton.	2018-12-02T00:17:30.410Z	2014-02-08T00:00:00.000	{ "year": 2014, "sha1": "4a9048eb4749a8f505348c5e07970f481c2d6525", "oa_license": "CCBY", "oa_url": "http://iopscience.iop.org/article/10.1088/1742-6596/488/1/012033/pdf", "oa_status": "GOLD", "pdf_src": "Arxiv", "pdf_hash": "4a9048eb4749a8f505348c5e07970f481c2d6525", "s2fieldsofstudy": [ "Physics" ], "extfieldsofstudy": [ "Physics" ] }
254179191	pes2o/s2orc	v3-fos-license	Markers of Renal Complications in Beta Thalassemia Patients with Iron Overload Receiving Chelation Agent Therapy: A Systematic Review Objective The emerging renal complications in beta-thalassemia patients have raised the global exchange of views. Despite better survival due to blood transfusion and iron chelation therapy, the previously unrecognized renal complication remain a burden of disease affecting this population —the primary concern on how iron overload and chelation therapy correlated with renal impairment is still controversial. Early detection and diagnosis is crucial in preventing further kidney damage. Therefore, a systematic review was performed to identify markers of kidney complications in beta thalassemia patients with iron overload receiving chelation therapy. Methods Searches of PubMed, Scopus, Science Direct, and Web of Science were conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) to identify studies of literature reporting renal outcome in β-TM patients with iron overload and receiving chelation therapy. The eligible 17 studies were obtained. Results uNGAL/NGAL, uNAG/NAG, uKIM-1 are markers that can be used as predictor of renal tubular damage in early renal complications, while Cystatin C and uβ2MG showed further damage at the glomerular level. Discussion and Conclusion The renal complication in beta-thalassemia patients with iron overload receiving chelating agent therapy may progress to kidney disease. Early detection using accurate biological markers is a substantial issue that deserves further evaluation to determine prevention and management. Introduction Beta thalassemia is one of the most frequent haemoglobin disorders inherited in an autosomal recessive manner bringing about worldwide health problems, especially in the tropical belt. Southeast Asia accounts for about 50% of thalassemia carriers in the world. Beta-thalassemia results from reduced or lost-globin chain synthesis due to mutations in the betaglobin gene. 1,2 Indonesia, with more than 200 million people, has a beta thalassemia carrier frequency of 6-10%, becoming an issue to anticipate by the government. The high demand for blood transfusions and chelating therapy marks their linear effect on the improved prognosis of beta-thalassemia patients. 2 Despite a better survival rate due to supportive therapy, the beta-thalassemia population still have an increased risk for various complications and thus remains a challenge in affecting their quality of life. The four most common complications are cardiac, endocrine, hepatic, and renal. [3][4][5][6][7] The emerging renal complications reported in betathalassemia patients have been correlated with the nature of the disease (eg, chronic anemia, hypoxia), iron overload due to regular blood transfusions, and chelating therapy. Altered vascular resistance and increased renal plasma flow, hyperfiltration, and renal tubular dysfunction were some common mechanisms in diminished renal function of betathalassemia patients. Iron overload beta-thalassemia patients were found to have distinct markers of kidney injury such as serum beta2-Μ, urinary calcium/creatinine, urine 2-M/creatinine, urinary NAG, urinary NAGL, urinary a1microglobulin, and urinary RBP. [8][9][10][11] Moreover, renal complications also occurred significantly in those who received chelating agent therapy. 3,10,[12][13][14] This study aims to identify markers of renal complications in beta-thalassemia patients with iron overload and chelation agent therapy. Methods This study was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). Literature Search A comprehensive search of the electronic literature was done using the online databases PubMed, Scopus, Science Direct, and Web of Science. We used the following search term: ([MeSH]beta-Thalassemia] OR (thalassemia) OR (thalassemia beta major) AND (renal outcome) OR (renal complication) OR (renal function) OR (renal damage) OR (kidney injury) OR (kidney function) from the Cochrane Collaboration's search strategy for randomized control trials. Our search approach does not use a filter and has no year restrictions. The first search on each database was done in March, and the final search was done on May 29, 2022. Figure 1 contains the PRISMA flow for literature search. Eligibility and Study Selection All randomized control trials, cohort studies, case-control studies, and cross-sectional studies reporting renal outcome in patients with TM connected to chelation therapy were considered. We only list publications written in English. Unrecoverable full text studies and duplicate research were eliminated. Authors EP and EDA carried out the search. The full text of any articles that satisfied the inclusion criteria was retrieved by three writers (PZR, SDS, and AEP) after they separately screened the title and abstracts. The eligibility of full text articles was examined by both writers. The Mendeley program, a free online tool for managing references, was used during the selection process to eliminate duplicate studies and analyze the abstract and full text. Data Extraction Data extraction was carried out by two reviewers (BAM, CW), and any differences were settled by team consensus. Using an excel predesign table, the studies that satisfied the relevance and eligibility of our aforementioned criteria were extracted. All of collected data were: 1) A summary of the studies that were included, including methodological information about the site, sample, interventions, outcomes, and results. 2) Baseline features of the studies. 3) Renal outcome or function related to the previously described use of chelation therapy. a significant difference in uβ2MG 20,23,25 in beta thalassemia patients with iron overload and receiving chelating agents. Increase in uNGAL/NGAL; uNAG/NAG; uKIM-1; Cystatin C; This uβ2MG/β2MG is associated with early tubular and glomerular dysfunction. Three studies suggest that deferasirox is a chelating agent that causes an increase in uKIM-1, uNGAL, and u uβ2MG. 3,10,23 uNGAL/NGAL The proximal tubule, distal tubule, and loop of Henle segments all contain epithelial cells that express the 25-kDa lipocalin iron-carrying protein known as neutrophil gelatinase-associated lipocalin (NGAL), which is released by active neutrophils. Administration of chelating agents in beta thalassemia patients actuates a damage to the proximal tubular kidney. 2 Smolkin 2008 In beta thalassemia patients, UNAG and UNAG/creatinine ratio markers can be used as markers to determine the frequency of transfusion and the effect of chelating agent therapy on ferritin levels. 3 Economou 2010 Glomerular and tubular dysfunction in iron overload beta thalassemia patients receiving chelating agents was observed through the elevation of cystatin C, proteinuria, hypercalciuria and beta2-microglobulin. 4 Hamed 2010 Chelating agents in beta thalassemia patients have been shown stimulate glomerular and tubular dysfunction. Smolkin 2008 In beta thalassemia patients, UNAG and UNAG/creatinine ratio markers can be used as markers to determine the frequency of transfusion and the effect of chelating agent therapy on ferritin levels. 6 Mansi 2013 Kidney damage occurs in beta thalassemia patients receiving chelating agents characterized by increased urea, creatinine, uric acid, urinary sodium and potassium. 7 Ali 2014 Elevated cystatin C, serum creatinine, and serum ferritin and high uACR indicate kidney damage in beta thalassemia 8 Sen 2015 NAG and NGAL can be used as markers of renal injury in beta thalassemia patients experiencing iron overload and receive therapeutic chelating agents. 9 Behairy 2017 Cystatin-C and beta-2 microglobulin are specific and sensitive early biomarkers for monitoring glomerular and tubular dysfunction in children with beta-TM. 10 Bekhit 2017 NAG is excreted following renal damage, hence used as marker of the toxicity index in beta thalassemia patients receiving chelating agents. 11 Annayev 2018 Kidney damage was concluded after an increase in Cystatin C and beta microglobulin in beta thalassemia patients receiving iron chelating agents. 12 ElAlfy 2018 Iron overload in beta thalassemia patients with long-term use of chelating agents revealed differences in serum cystatin C, uACR, and urinary B2 microglobulin, and B2 microglobulin/albumin ratio values. 13 Badeli 2019 Administration of deferasirox led to kidney damage, characterized by an increase in uNGAL, inflammatory factors, sort of IL18, in beta thalassemia patients. 14 Fouad 2019 uNGAL levels, uNGAL/creatinine ratio, eGFR and urinary albumin creatinine ratio were distinctive between group with iron overload and added chelating agents. 15 Nafea 2019 UKIM-1 is a biomarker used to detect worsening eGFR in beta-thalassemia patients with iron overload and routine deferasirox. 16 Bilir 2020 Iron overload marked by high serum ferritin, is associated with proteinuria, increased cystatin C, and urinary protein/ Cr in beta thalassemia patients routinely receiving chelating agents. 17 Capolongo 2020 The administration of chelating agents in patients with beta thalassemia iron overload induced renal tubular damage and required early detection to prevent further kidney damage. 18 Mahmoud 2021 Serum cystatin-C is a good predictive marker in the evaluation of glomerular dysfunction. Urine concentrations of NAG and KIM-1 represent sensitive, specific, and highly predictive early biomarkers for acute renal injury in patients with beta TM when subclinical kidney injury or dysfunction is expected before serum creatinine increases. Tanous 2021 The use of deferasirox for approximately 10 years induced the decline of eGFR and was negatively correlated with uNAG in beta thalassemia patients with iron overload. Detection of NGAL can be done through serum or urine withdrawal. However, in renal tubular damage, NGAL is often upregulated. 26 NGAL is a low-concentration protein that binds iron-siderophore complexes and is found in a variety of cell types. NGAL is discharged in urine when proximal tubular damage inhibits the reabsorption or increase the synthesis. An abrupt increase in NGAL is brought on by acute kidney injury (AKI). High NGAL levels are also present in patients with other abnormalities, such as lupus nephritis, immunoglobulin A nephropathy, and urinary tract infections. 27,28 Significant relationships between urinary NGAL (uNGAL) and proteinuria were documented in chronic kidney disease. 29 Recent case-control studies revealed that β -TM patients receiving deferasirox and regular blood transfusions had considerably greater levels of uNGAL and NGAL. 9 It is interesting to note that in β-TM patients, combined values of albumin/creatinin ratio and the uNGAL/creatinin ratio may be a more accurate predictor of kidney impairment and taken into account as potential indicators of renal failure. Additionally, as previously demonstrated, uNGAL somehow can predict renal function since it can estimate eGFR, evaluate the course of CKD, and serve as a surrogate measure for baseline eGFR. 30 NGAL is highly associated with early renal complication regardless of the creatinine level, both in major and intermediate thalassemia. 31 In both young and adult beta-thalassemia patients, NGAL level elevates due to iron overload and prolonged use of chelating agents. Since renal injury might elevate NGAL to a significant level, it is wisely advised to routinely monitor the NGAL level. Another highlighted feature of NGAL is that it immediately elevates right after kidney injury, even before serum creatinine, urinary N-acetyl glucosaminidase, and 2-microglobulin levels are detected. uNAG/NAG Tubular damage is indicated by an increase in the level of N-acetyl-beta-d-glucosamine (NAG), a well-known biomarker for proximal tubular injury. 32 According to Aldudak et al, people with beta thalassemia had higher levels of the tubular damage indicators, such as NAG, malondialdehyde, and b2-microglobulin excreted in their urine. 17 In addition, Jalali et al showed that b-TM patients had significantly higher urine NAG levels than controls, with high NAG levels being the norm in most cases. 33 Besides, participants in Sen et al's study who had renal proximal tubular damage also showed significantly elevated levels of uNAG/Cr and uNGAL/Cr. 11 The proximal tubular dysfunction that results from thalassemia itself may be caused by chronic hypoxia brought on by persistent anemia, by iron deposition, or by iron chelation. 4,19,23,34 According to some studies, the absence of a link between urine indicators and hemoglobin, haematocrit and ferritin levels may be due to very early tubular dysfunction, as seen by the rise of only NAG and NGAL, but not KIM-1, L-FABP, or urinary electrolytes. 11,19,34,35 uKIM-1 A transmembrane protein called urinary human kidney injury molecule-1 (uKIM-1) is found in the proximal renal tubules within 24 hours of acute tubular necrosis after renal ischemia. Even while serum creatinine concentrations are unaffected by exposure to certain nephrotoxic drugs, urinary KIM-1 may still be detected. 36 According to the findings of the Nafea et al study, young thalassemia patients receiving deferasirox therapy had evident subclinical nephrotoxicity when compared to patients receiving other chelation therapies. This was demonstrated by the statistically significantly higher levels of serum creatinine, eGFR, and UKIM-1/Cr. The serum levels of phosphorus, magnesium, creatinine, and blood sugar all showed a significant positive association with UKIM-1/Cr, however the blood hemoglobin level showed a significant negative correlation. 3 According to Badeli et al's study results, the deferasirox group had significantly greater levels of uIL-18, uNGAL, uNGAL/CREA, uKIM-1/CREA, and BUN than the control group. 10 Deferasirox treatment led to partial necrosis in the renal tubules and increased urinary NGAL, Cystatin C, KIM-1, protein, and glucose production, as demonstrated by Sánchez González et al in an animal study. 37 Cystatin C and uβ2MG Cystatin C, not excreted by the renal tubules or reabsorbed into the serum, is a sensitive biomarker for glomerular filtration rate (GFR). All cells in the body continuously produce cystatin C and its production is unaffected by changes in age, sex, gender, or muscle mass. Cystatin C is a reliable and early marker of glomerular dysfunction in the pediatric population. 22,36,38 A low-molecular-weight protein called urinary β2 microglobulin (u β2MG) is freely filtered by glomeruli, reabsorbed by renal tubules, and then eliminated. Due to its continual production, both are thought to be a more reliable endogenous measure of early glomerular filtration rate (GFR) affection than creatinine. 39 For monitoring glomerular and tubular dysfunction in kids with β -TM, β2MG is a sensitive early biomarker. 23,24 β2MG levels are very low in healthy people; it rises in inflammatory, immunologic, and cancerous conditions. 36 Other than that, Cystatin C and β2MG have strong correlation with age and creatinine clearance. 40 Early detection of glomerular disease will decrease the rate of renal failure and mortality. Referring to a study by ElAllfy et al, thalassemia patients had significantly greater serum levels of cystatin C than healthy controls. 20 Positive correlations were found between serum cystatin C and indirect bilirubin, LDH and serum ferritin. Additionally, there was no connection between any of these indicators and the kidney function tests (serum creatinine, urea, and uACR) or between serum cystatin C and u2MG. 20 Elbedewy et al 41 Behairy et al discovered that serum cystatin C and uβ2MG were negatively correlated with creatinine clearance, hemoglobin, and estimated GFR in children with β -TM while both markers were positively correlated with urea, creatinine, serum ferritin, UACR, duration of chelation therapy, and frequency of blood transfusion per year. As indicators for glomerular and tubular dysfunction, cystatin C and β2MG have good sensitivity and good specificity. 24 Cystatin C and β2M, as well as serum ferritin and liver iron deposition were found to be significantly positively correlated, according to Kacar et al. 43 Serum ferritin levels were discovered to be associated with cystatin C and β2M levels by Uzun et al 44 found that serum ferritin is correlated with level of cystatin C and β2 M. The risk of glomerular and tubular dysfunction may increase with iron buildup in the body. Conclusion The renal complication in beta-thalassemia patients with iron overload receiving chelating agent therapy may progress to kidney disease. Early detection using accurate biological markers is a substantial issue that deserves further evaluation to determine prevention and management. Ethics This study was approved by Airlangga Hospital's ethical board with certificate number 024/KEP/2022. All analyses for the present study were based on previous published research, thus no patient consent was required. Disclosure The authors report no financial or other conflicts of interest in this review article.	2022-12-04T05:11:16.849Z	2022-11-01T00:00:00.000	{ "year": 2022, "sha1": "e36f078165016b915bf4c6647e0a52304ca75546", "oa_license": "CCBYNC", "oa_url": "https://www.dovepress.com/getfile.php?fileID=85734", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "e36f078165016b915bf4c6647e0a52304ca75546", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
267564823	pes2o/s2orc	v3-fos-license	Constructing [2.2]Paracyclophane-Based Ultrasensitive Optical Fluorescent-Phosphorescent Thermometer with Cucurbit[8]uril Supramolecular Assembly The pursuit of developing novel approaches to fully organic and eﬃcient phosphorescent materials is in high demand. The optical activity of such functional, organic phosphorescent/ﬂuorescent materials may exhibit great temperature dependence, allowing their application as advanced, highly sensitive molecular thermometers. In this study, a rational strategy involving host–guest complexation and polymerization of [2.2]paracyclophane (PCP) based molecules with cucurbit[8]uril (CB8) to suppress the molecular motion and promote temperature-dependent phosphorescence is presented. The rigid cavity of CB8 provides an ideal microenvironment to host PCP molecules 1 and 2, signiﬁcantly enhancing the photophysical performance after complexation. Co-polymerizing phosphors 1 and 2 with acrylamide is an eﬃcient method for improving phosphorescence. Incorporating CB8 into the resulting P-1 and P-2 polymers enhances phosphorescence performance. Importantly, the obtained materials exhibit a big structure-dependent spectral shift and change of phosphorescence lifetimes with temperature, allowing novel, phosphorescence-based, and purely organic optical thermometers to be developed. The practical applications of PCP-based luminescent materials in temperature sensing via a multi-parameter approach are showcased, i.e., using ﬂuorescence spectral shift and changes in bandwidth, as well as phosphorescence lifetimes, exhibiting thermal sensitivity of ≈ 17.7 cm − 1 ° C − 1 , 47.8 cm − 1 ° C, and 5.2% ° C − 1 , respectively. Introduction Temperature detection is a basic and crucial technique in scientific and industrial fields and our daily lives.With the fast development of optical (luminescence) thermometry, many promising results were reported, benefiting the development of aerospace technology, biomedicine, superconductors, polar explorations, etc. [1][2][3][4][5][6] Luminescent temperature sensors mainly encompass inorganic materials and nanoparticles, such as Ln 3+/2+ -doped structures, materials based on d-block metal ions, and quantum dots.[8] In general, the benefits of optical organic thermometers are their great susceptibility to temperature changes, high luminescence quantum yield, and brightness, as well as the possibility of advanced and specific modification of their structure according to the requirements of given applications.[11] Organic molecular thermometers, which are based on temperature-responsive fluorescent materials, enabling fast photophysical response, outstanding spatial and temporal resolution, good biocompatibility, and high sensitivity, are regarded as one of the most valuable thermometric tools in the cryogenic range.However, designing and developing novel optical thermometers with high sensitivity is still challenging.An efficient strategy to significantly improve the thermal sensitivity of optical sensors may be to exploit the thermal response of the lifetime of phosphorescence in organic phosphor compounds, which may be extremely long for phosphorescence (up to hundreds of milliseconds or even seconds) and highly sensitive to temperature changes.Note that fluorescence phenomena, on the other hand, are very fast, lasting typically in the range of a few nanoseconds.This is due to fluorescence typically originating from allowed singlet-singlet (S 1 -S 0 ) transitions, [12][13][14] whereas phosphorescence is related to the forbidden triplet-singlet (T 1 -S 0 ) transitions. [15,16]Another promising strategy is to use the spectral shift of the emission bands as a thermometric parameter, which is typically much more susceptible to temperature changes in the case of organic compounds compared with inorganic materials. Advances in highly efficient and stable room-temperature phosphorescence materials have attracted considerable attention in various fields, such as display technology, [17][18][19] optoelectronics, [19] optical sensors technology, [20][21][22] etc. [23,24] This is due to the advantages of organic phosphorescent materials, such as facile modification, biocompatibility, very long emission lifetime, large Stokes shifts, rich triplet states, as well as high quantum yield and brightness, which are not fully covered in fluorescent materials. [25]To improve phosphorescence by enhancing the intersystem crossing (ISC) rate between the excited singlet and triplet states, two main strategies have been proposed: I) introducing heavy atoms to promote spin-orbit coupling and reducing the splitting energy of the S 1 and T 1 states (ΔE ST ); II) suppression of the undesired nonradiative transitions from the triplet state by restraining the molecular motions, e.g. by crystallization, polymerization, H-aggregation (molecular selfassembly), and host-guest interaction. [24]2.2]Paracyclophane (PCP) is a compact molecule with two slightly curved benzene decks constrained by ethylene bridges in a coplanar conformation.With the short deck distance of 3.09 Å, smaller than the van der Waals distance between the layers of graphite (3.35 Å), strong transannular electronic communication between the benzene decks can occur.PCP-based molecules have been successfully exploited in material science to develop organic light-emitting diodes (OLEDs) and nonlinear optical materials. [26,27]However, PCP-based phosphorescent materials were rarely reported in the literature. [28]o construct a PCP-based phosphorescent thermometer, the supramolecular host-guest assembly provides an alternative approach, wherein phosphor motion can be restrained, and phosphorescence quenching by oxygen can be prevented, thus promoting the phosphorescence phenomenon even at relatively high temperatures. [29]In this study, we utilized cucurbit [8]uril (CB8), a rigid macrocyclic host synthesized by polymerization of glycoluril and formaldehyde as such a host.Additionally, by copolymerizing PCP phosphors with acrylamide, phosphorescence properties were also improved.Furthermore, the polymerized phosphor molecules can be again encapsulated within the CB8 cavities, further enhancing the phosphorescence effect.Finally, we demonstrated the potential application of the encapsulated PCP-based organic fluorophores for optical temperature detection, which may culminate in a new approach for producing novel phosphorescent-fluorescent thermometers.The developed molecular optical thermometers allow multiple-parameter temperature sensing, i.e., detection based on several parameters, including emission spectral position, FWHM, and lifetime, with a relatively high sensitivity of ≈17.7 cm −1 °C−1 , 47.8 cm −1 °C−1 , and 5.2% °C−1 , respectively. Design and Strategy We have previously reported that CB8 is an ideal host to accommodate methyl-pyridinium-paracyclophane iodide (MPCP-I) in aqueous media complementarily. [30]However, iodide counter ion usually weakens the photophysical performance due to the charge transfer in pyridinium iodides. [31]Therefore, methylpyridinium-paracyclophane chloride (sample 1) was synthesized to access a better photophysical performance (Figures S1-S3, Supporting Information).To investigate the heavy atom effect, which can also influence the photophysical properties, methylbromo-pyridinium-paracyclophane chloride (sample 2) was designed, synthesized, and fully characterized (Figures S6-S8, Supporting Information).On the one hand, encapsulation of the PCP chromophores in the CB8 cavity restrains the molecular motions, promoting the ISC from excited singlet (S1) to triplet (T1) state and reducing the nonradiative decay channels.On the other hand, the rigid shell of the CB8 protects it from the diffusion of molecular oxygen and other possible triplet-state quenchers.Moreover, polymerizing PCP chromophores into a more rigid structure may promote overall phosphorescence performance.In the same fashion, introducing the CB8 host to the PCP polymer may bring a further positive effect on the phosphorescence.Figure 1 depicts the synthesis, encapsulation, polymerization scheme, and expected phosphorescence enhancement effects.Please note that the detailed mechanism in the materials requires a systematic, complicated study.Thus, here we present the simplified, general energy level diagram for the organic fluorophores studied (Figure 1), highlighting the ground and excited singlet and triplet states, photo-excitation from the ground S 0 to the excited S 1 state, fluorescence, energy transfer from the excited singlet (S 1 ) to triplet (T 1 ) state, and finally phosphorescence from the excited T 1 to the ground S 0 state. Host-Guest Interactions of PCP Chromophores with CB8 The binding behavior of the synthesized PCP derivatives (samples 1 and 2) with CB8 was initially investigated by nuclear magnetic resonance (NMR) spectroscopy.As shown in Figure 2a, when an equal molar equivalent of the CB8 was introduced into the sample 1 dissolved in D 2 O, the protons from PCP aromatic region (c-i), pyridine (b), and PCP bridge (Figure S11, Supporting Information) were observed significantly upfield shifted, suggesting the deep encapsulation of sample 1 within CB8 cavity.The same phenomenon was also observed for sample 2, as shown in Figure 2b, where protons b-h and from the PCP bridge (Figure S13, Supporting Information) were upfield shifted.The formation of the 1:1 host-guest complexes was further proved by electrospray ionization (ESI) mass spectrometry in Figures S12 and S14 (Supporting Information), where identical peaks with massto-charge ratio (m/z) of 1626.10 for [1+CB8] + and 1704.75 for [2+CB8] + were observed. To gain direct evidence of the complexation of the PCP molecules within the CB8 cavity, a single crystal was obtained by slow evaporation of sample 1 mixed together with CB8 (dissolved in Milli-Q water) and further analyzed by X-ray crystallography (Figure S15 and Table S1, Supporting Information).Due to the disorder of many water molecules non-covalently interacting with the CB8 moiety and the Cl counter ions, their clear identification was hampered.However, the general crystal structure of sample 1 encapsulated in CB8 can be determined, as shown in Figure 2c,d.Indeed, the PCP moiety from molecule 1 is completely stabilized within the CB8 cavity, while the pyridinium part is exposed to the carbonyl portal of the CB8 structure, exhibiting various weak noncovalent bindings (Figure 2c,d), which is consistent with the NMR analysis. The binding constant for MPCP-I has been reported in our previous study as 3.89 (± 0.99) × 10 12 m −1 . [30]To investigate the binding constant of molecule 2 with CB8, a competitive binding assay was performed utilizing memantine hydrochloride (Mem) as a competitive guest, whose binding constant with CB8 was reported [30] as 8.28 (± 0.38) × 10 12 m −1 .As shown in Figure S16 (Supporting Information), when over one molar equivalent of Mem was introduced to the mixture of compounds 2 and CB8, the proton signals of CB8 from 2•CB8 and Mem•CB8 were distinctively separated at 5.46 ppm for 2•CB8 and 5.54 ppm for Mem•CB8.Based on the binding equations, the binding constant of molecule 2 with CB8 was calculated as 1.13 (± 0.16) × 10 12 m −1 , which ensures good stability of the formed complex structure. Photophysical Properties of PCP and CB8 Assembly at Ambient Conditions Further, we studied the photophysical properties of the obtained liquid samples, i.e., 1, 2, 1•CB8, and 2•CB8 dissolved in aqueous media, as shown in Figures S17-S22 and Table S2 (Supporting Information).Upon the addition of CB8 to sample 1, the bathochromic-shifted absorption spectra suggest the formation of 1•CB8 assembly, while the observed photoluminescence (PL) intensity was significantly enhanced (Figure S17, Supporting Information).The luminescence quantum yield (Ф) and fluorescence lifetimes were increased accordingly (Figures S19 and S20, and Table S2, Supporting Information), which was caused by the complexation of molecule 1 with CB8 moiety, promoting the radiative decay (emission) and limiting the nonradiative decay process (see Table S2 was also found in the case of the complexation of sample 2 with CB8 (Figures S18,S21,S22, and Table S2, Supporting Information). It has been reported that polymerization could improve the photophysical properties of organic fluorophores by reducing the molecular motions and preventing aggregation-caused quenching processes. [32]Therefore, the polymeric materials P-1 and P-2 were designed and synthesized, where the phosphors 1 and 2 were copolymerized with acrylamide (details in Supporting Information).Furthermore, the CB8 compound has been introduced into the P-1 and P-2 in aqueous solutions, and the formation of the P-1•CB8 and P-2•CB8 complex structures was proved by the NMR studies (Figures S23 and S24, Supporting Information) where the aromatic peaks from the PCP are upfield shifted when the CB8 is introduced to polymers P-1 and P-2.Their photophysical properties were studied and listed in Figures S25-S30 and Table S2 (Supporting Information).Compared to the monochromophores (monomers), polymerization of molecules 1 and 2 greatly enhanced their photophysical properties.The effect of the complexation of chromophores with CB8 is also scaled up in the complex, encapsulated polymer structures P-1•CB8 and P-2•CB8. Afterward, we investigated in detail the PL performance of the synthesized materials in the solid state, where the effect of water on the chromophores can be minimized.The excitation and emission spectra, luminescence quantum yield, and phosphorescence decay curves were measured to compare the PL properties.As shown in Figure 3a, the PL excitation band is located in the range of ≈250-500 nm and dominates in the excitation spectra of all synthesized materials, while the center of the band varies.As shown in Figure 3b, it can be seen that all the emission spectra of the samples are dominated by a broad band located in the range of 370-800 nm, centered ≈500 nm, which is consistent with the emission spectra recorded in the aqueous media.Please note that these observed blue emission bands are associated with the fluorescence of the samples, whereas the phosphorescence bands are bathochromic-shifted to the yellow-orange region, as will be discussed later.the encapsulation of the CB8 cavity and polymerization of the chromophores can largely improve the QY values of the samples.Using sample 1 as an example (QY = 3.87%), sample 1•CB8 (QY of 27.7%) is ≈7.2 times higher than sample 1, indicating that the encapsulation of the CB8 cavity significantly elevated QY.Second, P-1 (QY = 77.3%)shows ≈20 times higher QY than sample 1, i.e., polymerization of the chromophores also provides the additional possibility to increase the QY value and improve the PL performance.Noteworthy, P-1•CB8 presents a very high QY of 86.3%, which greatly improved the QY (≈22 times) compared to sample 1.On the other hand, in the series of sample 2, the same QY enhancement effect can also be observed.Namely, in the case of the P-2 (QY = 22.9%), the polymerization of the chromophores results in ≈13.5 times higher QY than sample 2 (QY = 1.79%).Whereas the QY of sample 2•CB8 (QY of 16.0%) is ≈8.9 times higher than sample 2, the complex, encapsulated polymer structure P-2•CB8 exhibits QY of 28.8%, which is ≈16 times higher than sample 2. Such results indicate that encapsulating PCP organic chromophores in the CB8 cavity and their copolymerization with acrylamide can tremendously boost the QY values. To explore the influence of the encapsulation in the CB8 cavity and polymerization of the PCP chromophores, the measured phosphorescence decay curves and the determined phosphorescence lifetimes are illustrated in Figure 3d and the corresponding inset.As shown, the determined phosphorescence lifetimes at room temperature for the samples 1, 2, P-1, P-2, 1•CB8, 2•CB8, P-1•CB8, and P-2•CB8 are 0.13, 0.08, 6.17, 3.61, 2.46, 1.98, 8.86, and 4.83 ms, respectively, so in extreme situations they differ by 2-orders of magnitude.The influence of the encapsulation in the CB8 cavity and polymerization of the PCP chromophores on the PL QY is similar to that of the PL lifetime.In other words, the encapsulation and polymerization can largely prolong the PL lifetime, e.g., in sample 2 with 2•CB8, the emission lifetime was prolonged ≈25 times; for sample P-2 (polymer), the PL lifetime was prolonged ≈45 times, compared to sample 2 (monomer); from sample 2 to sample P-2•CB8, the lifetime was prolonged 60 times (Table S3, Supporting Information).The increase in the lifetime is thus correlated with the increase in QY (see above).The results suggest that the encapsulation in the CB8 cavity and polymerization of the organic chromophores can suppress nonradiative electronic relaxation channels and thus enhance the QY and lifetime.This is probably due to the spatial confinement/restriction of the guest molecules (1 and 2) by the CB8 cavity, which largely suppresses molecular motions (vibrations, rotations, and inter-collisions). Red-Shift-Based thermometry To investigate the temperature detection properties of the obtained PCP materials, we first studied the thermostability by thermogravimetric analysis (TGA).The results (Figure S32 stable below 200 °C.After this, the PL properties of all the synthesized samples, from cryogenic temperature to high temperature, were measured.Please note, for optimizing the thermometric performance, we selected the operating temperature range, where the centroid or FWHM presents only the monotonic change, which is crucial in optical sensing applications.To determine the temperature-dependent spectroscopic parameters and the related thermometric performances, the corresponding spectra were converted to the energy scale using the Jacobian transformation, [33] minimizing the quantitative error of the emission spectra analysis.As disclosed in Figure 4a,b, the intensity of sample 1 increases with a temperature below −70 °C; after that, it shows a declining trend with further elevating the temperature.For sample 1•CB8, it exhibits a similar tendency, and the emission spectra' highest intensity appeared at 10 °C.In the selected temperature range, which can be accessed for optical thermometry, samples 1 and 1•CB8 emission bands show significant spectral redshift (toward lower energy) with temperature rising.As shown in Figure 4c-f, to study the optical thermometric properties, the determined band centroid ( centroid ) and bandwidth (fullwidth half maximal; FWHM) of sample 1 and 1•CB8 can be fitted to third-order polynomial functions in a relatively broad temperature range, i.e. where A 0 , A 1 , A 2 , and A 3 are constants and T is temperature.Please note that the fitting function for other samples is also third or fourth polynomial function to analyze the thermometric properties.Herein, these fitting constants are shown in Table S4 (Sup-porting Information).In the case of sample 1 (see Figure 4c,e), the centroid increases from 520.6 nm at −90 °C to 543.0 nm at 130 °C with a monotonic shift of 22.4 nm (≈−708.5 cm −1 ).FWHM changed from 114.9 nm at −80 °C to 129.0 nm at 70 °C, with a band-broadening of 14.1 nm (239.7 cm −1 ).On the other hand, for sample 1•CB8, the centroid increases from 522.9 nm at −70 °C to 530.0 nm at 140 °C, leading to a spectral shift of 7.1 nm (≈−283.4cm −1 ).The FWHM decreases from 115.8 nm at −150 °C to 111.7 nm at 20 °C, leading to a total decrease in FWHM of 4.1 nm, i.e., ≈−296.3cm −1 . Thermal sensitivity is the rate of change of the thermometric parameter in response to the temperature variation.To quantitatively investigate the performances of the developed PCP materials for temperature sensing, the absolute sensitivity, i.e., S a, is estimated using the following functions: The S a curves based on centroid and FWHM versus temperature of samples 1 and 1•CB8 are shown in Figure 5a-d, respectively.The S a MAX based on centroid and FWHM of samples 1 are 4.37, 6.51 cm −1 °C−1 , i.e., 0.14 and 0.24 nm °C−1 , respectively.Those of sample 1•CB8 are 1.71 cm −1 °C−1 (0.05 nm °C−1 ) and 3.02 cm −1 °C−1 (0.07 nm °C−1 ), respectively.It can also be seen that the temperature-dependent color of both samples is slightly changed from cold cyan to warm cyan (see Figure 5c,f).Importantly, in the selected temperature ranges, the emission bands of 1 and 1•CB8 demonstrate a significant redshift, indicating their potential in optical thermometry. Blue-Shift-Based Thermometry To investigate the temperature-sensing properties of other synthesized PCP samples, the emission spectra of samples 2 and 2•CB8 at various temperatures are recorded, and corresponding temperature-sensing performances are illustrated in Figures 6 and Figure 7.The emission bands of samples 2 and 2•CB8 show significant spectral blue-shift (toward higher energy) with rising temperatures, opposite to samples 1 and 1•CB8.As shown in Figure 6a,b, the intensity of the emission band of sample 2 shows a monotonically decline tendency with increasing temperature.Meanwhile, sample 2•CB8 exhibits a decreasing tendency but a much lower rate, which is beneficial for optical sensing.The thermometric properties based on centroid and FWHM of samples 2 and 2•CB8 were studied in detail, shown in Figure 4c-f.In the selected temperature range, the centroid of sample 2 blue-shifts from 499.0 nm at −130 °C to 480.1 nm at −50 °C (total shift of −18.9 nm, i.e., 1046.1 cm −1 ), while FWHM increases from 118.8 nm at −190 °C to 156.9 nm at −90 °C (total FWHM change of 38.1 nm, i.e., 1950.6 cm −1 ). On the other hand, the centroid of sample 2•CB8 exhibits a blue shift from 546.2 nm at −190 °C to 519.9 nm at 10 °C, i.e., a total shift of −26.3 nm (744.1 cm −1 ).The FWHM change from 143.0 nm at −120 °C to 100.9 nm at 10 °C, that is, total FWHM change of 42.1 nm, i.e., −1301.8cm −1 .The temperaturedependent centroid or FWHM of samples 2 and 2•CB8 can be well fitted to polynomial function (third or fourth order) in the selected temperature range to study the thermometric properties.As mentioned before, the restriction of the guest molecules by CB8 largely suppresses the molecular motions, and the CB8 cage can shield the triplet states and the guest molecules from quenchers, thus weakening the thermal response of the optical properties of developed samples.It is clear that the luminescence properties of the "bare" molecule, i.e., sample 2, are more susceptible to temperature changes; however, the optical thermometer having the molecular "cage," i.e., sample 2•CB8 shows monotonic changes in broader T-range, it has better QY and high intensity, still having a satisfactory thermal response. The temperature sensing performances as a function of the temperature of samples 2 and 2•CB8 are illustrated in Figure 7a-f.In the case of sample 2, the S a value based on centroid initially shows an increasing tendency, reaching the optimal S a (S a MAX ) of 0.33 nm °C−1 (≈17.7 cm −1 °C−1 ) at −90 °C, and then a decreasing trend is observed with further temperature elevation (see Figure 7a).The S a value based on FWHM demonstrates a monotonically increasing tendency, achieving the maximal value of 0.61 nm °C−1 (≈47.8 cm −1 °C−1 ) at −90 °C (see Figure 7c).Please note that such a large shift rate is one of the highest compared to the previously described organic or inorganic optical thermometer based on spectral band shift, which will be discussed in the following paragraph.Interestingly, the emission color of the UV-irradiated sample demonstrates a significant visual color-tuning with increasing temperature, changing from yellow to green, then to cyan, and finally to deep blue, as shown in Figure 7e.For the sample 2•CB8 (see Figure 7b,d,f), it is clear that the S a based on centroid shows a continuously increasing trend with temperature rising, reaching 0.26 nm °C−1 (6.4 cm −1 °C−1 ) of the S a MAX , while the S a based on FWHM shows a maximum of 0.43 nm °C−1 (13.8 cm −1 °C−1 ) at −45 °C.The emission color of the sample 2•CB8 is tuned from yellowish green to cyan, then 7f).Importantly, the measurable temperature ranges of samples 2 and 2•CB8 are located in the cryogenic range (down to −190 °C), indicating that these two samples can be applied as highly sensitive temperature sensing materials in the cryogenic range. A detailed explanation of a mechanism governing the observed structure-dependent red-and blue-shifts of the emission bands is beyond this work's scope and requires further investigations, including complex theoretical calculations.However, here we note that the observed redshift (for the compounds without heavy atoms) is caused by the decreasing energy difference between the ground and excited (emitting) states, whereas the blue-shift (for the compounds with heavy atoms) implies an increase of that energy gap with temperature.Please note that the temperatureinduced spectral shifts in materials are often associated with their unit cell expansion with temperature or, more rarely, contraction in the case of structures with a negative thermal expansion coefficient (static contributions).On the other hand, the temperature elevation enhances electron-phonon interactions (vibrational contribution), as well as intra-and intermolecular interaction, which may contribute to the observed spectral shifts and changes in the width of the emission bands with temperature, making the whole mechanism even more complex for the organic compounds studied. Moreover, in our case, the presence of heavy atoms, which in general should promote the spin-orbit coupling of the phosphorescent molecules, resulted in a sight deterioration of their phosphorescence performance, including emission lifetimes and quantum yields.It may be associated with competing charge transfer processes in pyridinium bromides, similarly as in the case of iodide counter ions, which typically weakens the phosphorescence performance of various organic materials.For comparison, the temperature sensing performances of all obtained samples are summarized in Table 1.Please note that the temperature sensing performances of the samples P-1, P-1•CB8, P-2, and P-2•CB8 are presented in Figure S33-S36 (Supporting Information).It is shown that the S a MAX values obtained for samples 1 and 2 are 6.51 and 47.8 cm −1 °C−1 , which are much higher than the ones for the samples P-1 and P-2, i.e., 4.82 and 2.59 cm −1 °C−1 , respectively, indicating that the polymerization of the chromophores leads to smaller susceptibility to temperature changes.In general, the bare molecules, i.e., samples 1 and 2, present a higher thermal sensitivity than the 1•CB8 and 2•CB8 ones, respectively, confirming that the encapsulation in the CB8 cavity restricts the molecular motions of the organic molecules studied, and the CB8 can shield the triplet states of the guest chromophores from exposing directly to the stimulus of the external environment, thus slightly weaken their thermal response.However, as was already mentioned, the luminescent thermometers with molecular "cage" exhibit monotonic changes in the broader T-range; they have significantly better QY and much higher signal intensity.Hence, these benefits compensate very well for a bit smaller sensitivity and indicate that materials with molecular cages are much better and reasonable choices for their implementation as optical thermometers.Such multi-parameter thermometers can avoid potential biases in temperature detection in case of possible spectral interferences, which is common in real temperature sensing applications. Lifetime-Based Optical Thermometry To explore the temperature sensing properties based on the emission lifetime, the temperature-dependent phosphorescence decay curves of the selected samples with superior QY, i.e., 2•CB8, P-1•CB8 and P-2•CB8, were measured in the broad temperature range (from −190 to 140 °C) and presented in Figure 8a-c.Phosphorescence was chosen over fluorescence in this case due to the long-lasting nature of phosphorescence facilitating its applications for optical temperature sensing.The emission decay curves of the samples at low temperatures (see Figure 8d) are very long.That is why, for the P-1•CB8 sample at the lowest temperature values (where the phosphorescence lasted extremely long), due to technical limitations, the decay curves seem to be cut at ≈1 s (Figure 8a).Nonetheless, we have used the following expression to give an estimation of the average lifetimes: where I(t) is the acquired intensity as a function of the time.As shown in Figure 8a-c, the phosphorescence decay curves of all samples show strong temperature dependence, and the corresponding lifetimes monotonically shorten with increasing temperature.This shortening is due to the well-established thermal quenching processing, including the enhanced nonradiative relaxation probability with temperature, increased fluorescence probability concerning the competing phosphorescence phenomenon, back-energy transfer, improved electron-phonon coupling, etc.Initially, the determined lifetimes shorten slowly in the temperature range from −190 to −50 °C, but afterward, till ≈140 °C, their values decrease at much higher rates (see Figure 8d-f).The determined phosphorescence lifetimes as a function of temperature, selected in the range of the biggest Note that the phosphorescence decay seems to be faster in Figure 8h, in contrast to Figure 8g, simply due to a much better time resolution of the measurements (recorded every 10 ms), which comes at the cost of scarifying exposure time, compared to the digital photographs. In difference to thermometry based on band-shift and FWHM, the most reliable figure of merit for sensing based on the emission lifetimes is relative sensitivity (S r ), which is usually ex-pressed in % and shows how the measured parameter changes per 1 K of absolute temperature. Please note that combined with the thermometry mentioned above based on centroid and FWHM, the developed materials can be used as multi-parameter, remote thermometers, which can realize good thermal sensitivity in the whole measured temperature range, including the cryogenic and high-temperature regions, which overcomes the drawbacks of single-parameter thermometers.Thermometric performances of the developed molecular temperature sensors, including their sensitivity (S a for centroid and FWHM, S r for a lifetime) and operating temperature range, are compared with other luminescent thermometers in Table 2. Compared with other thermometers, the developed organic sensors show 3 approaches for temperature determination, and in each category, they are one of the thermometers with the highest sensitivity.In particular, in the category of FWHM, sample 2 shows a sensitivity of one order of magnitude higher than other reported thermometers.These results indicate the great potential of the developed molecular sensors as superior and supersensitive, noninvasive optical thermometers. Conclusion A novel rational strategy related to host-guest complexation, encapsulation, and polymerization of PCP-based molecules with CB8 organic cage to suppress the molecular motions and promote efficient phosphorescence was studied.The rigid and hydrophobic CB8 cavity provides an ideal microenvironment to host PCP molecules 1 and 2, thus largely improving the photophysical performance after complexation.Moreover, copolymerizing phosphors 1 and 2 with acrylamide is an efficient method to improve the phosphorescence phenomenon.Incorporating a CB8 molecular cage into the resulting P-1 and P-2 polymers leads to enhanced phosphorescence performance.The designed and synthesized organic phosphors demonstrate superior thermally induced spectral shift, FWHM change, and a huge variation of luminescence lifetimes, developing novel, fluorescence-and phosphorescence-based molecular optical thermometers with great sensitivity.Depending on the material structure and their encapsulation in organic cages, they exhibit either red-or blue-shifts with temperature and bright, yellow phosphorescence lasting several seconds in the cryogenic T-range.The practical applications of PCP-based phosphorescent materials in temperature sensing were demonstrated via a multi-parameter approach, i.e., spectral-shift, bandwidth, and phosphorescence lifetime, highlighting their high thermal sensitivity of ≈17.7 cm −1 °C−1 , ≈47.8 cm −1 °C−1 , and 5.2% °C−1 .Such a strategy allowed to cover a very broad temperature range, including cryogenic and high-temperature regions, i.e., from −190 to 140 °C, which is very beneficial in optical temperature sensing applications. Figure 1 . Figure 1.The scheme shows the synthesis strategies to boost the phosphorescence phenomena, utilizing encapsulation of the PCP-based materials in the CB8 and their polymerization.The inset shows the scheme of the simplified, general energy level diagram of the phosphorescence and fluorescence. As shown in the Commission Internationale de I'Eclairage (CIE) diagram (Figure S31, Supporting Information), different samples show different emission colors, changing from deep blue to green.Moreover, the luminescence quantum yield (QY) values for all synthesized materials are shown in Figure 3c.In comparison, Figure 3 . Figure 3. a) Excitation spectra of the materials 1, 2, 1•CB8, 2•CB8, P-1, P-2, P-1•CB8, and P-2•CB8 in solid state form; b) Emission spectra of the corresponding samples; c) Luminescence QY values of the synthesized compounds materials in solid state; d) Phosphorescence decay curves and the determined emission lifetimes (inset) for the developed materials in solid state. Figure 4 . Figure 4. a, b) The emission spectra of the selected samples, i.e., 1 and 1•CB8, as a function of temperature.c, d) The determined centroid of the emission band as a function of temperature for the samples 1 and 1•CB8.e, f) The FWHM of the emission band was determined as a function of temperature for samples 1 and 1•CB8. Figure 5 . Figure 5. a, b) The absolute temperature sensitivity (Sa) was determined based on the centroid of samples 1 and 1•CB8.c, d) The S a is based on the FWHM of samples 1 and 1•CB8.f) The digital photographs of samples 1 and 1•CB8 under UV excitation as a function of temperature. Figure 6 . Figure 6.a, b) The emission spectra of the selected samples, i.e., 2 and 2•CB8, as a function of temperature.c, d) The determined centroid of the emission band and the determined FWHM of the band as a function of temperature for samples 2 and 2•CB8, respectively. Figure 7 . Figure 7. a, b) The determined absolute temperature sensitivity (S a ) based on the centroid of the samples 2 and 2•CB8.c, d) The S a is based on the FWHM of samples 2 and 2•CB8.f) The digital photographs of samples 2 and 2•CB8 under UV light as a function of temperature. Figure 8 . Figure 8. a, c) The phosphorescence decay curves for the selected samples, i.e., 2•CB8, P-1•CB8 and P-2•CB8, recorded in temperature range from −190 to 140 °C.d, f) The determined phosphorescence lifetimes as a function of temperature for the corresponding samples.g) Digital photographs showing time-resolved phosphorescence (after UV exposure; laser off) of the selected samples as different temperature values.h) The representative time-resolved emission spectra of the sample P-2•CB8 measured at −190 °C.i) The determined relative sensitivity (S r ) values as a function of temperature. Table 1 . The temperature sensing performances, i.e., S a MAX , T, and T-range, based on centroid and FWHM of the different obtained samples. Table 2 . Temperature sensing performances, including maximal sensitivities for different organic and inorganic luminescent thermometers based on the band-shift, FWHM, LIR, and emission lifetimes.Sensitivities marked with asterisks (*) are estimated from figures in the references.	2024-02-09T16:27:26.530Z	2024-02-05T00:00:00.000	{ "year": 2024, "sha1": "0a998a85c68d829a5fffce472d91839c67888c3b", "oa_license": "CCBY", "oa_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/adfm.202313517", "oa_status": "HYBRID", "pdf_src": "ScienceParsePlus", "pdf_hash": "fc154edba34abc560b56909809ee910efc9f1995", "s2fieldsofstudy": [ "Materials Science", "Chemistry" ], "extfieldsofstudy": [] }
209718162	pes2o/s2orc	v3-fos-license	Research on Adsorption of Cr(VI) in Water by Potassium Chloride Modified Peanut Shell Carbon In order to remove Cr (VI) in water effectively, peanut shell carbon was modified by potassium chloride. The results showed that the removal rate of modified peanut shell carbon was 50.3%, increasing by 17.73% compared with peanut shell carbon. The removal and adsorption effects of modified shell carbon on Cr (VI) under different adsorption conditions were investigated and which including pH, adsorbent dosage, adsorption time and the ion concentration. The results showed that the best adsorption removal efficient of Cr (VI) in water can be obtained with pH of 4, adsorption dosage of 0.3g, the adsorption time of 150 min, and the Cr (VI) concentration of 8μg/ml. The removal rate of modified peanut shell carbon reached 99.86% under the condition. The adsorption process fitted the Freundlich adsorption isotherm and conformed to the quasi-second order kinetic equation. Introduction The accumulation of heavy metals in the environment is a serious problem through artificial pollution such as mining, metal smelting and processing, discharge of chemical wastewater, abuse of pesticides and fertilizers, and disposal of various kinds of garbage. Chromium pollution mainly comes from leather preparations, chromium plating parts of metal, industrial pigments, rubber and ceramic raw materials, etc. Chromium mainly exists in the form of Cr(III) and Cr (VI) in the environment. Among them, Cr(VI) has strong toxicity, carcinogenicity and mutagenicity，which can actively migrate and transform among water, soil, atmosphere and biology. It is one of the priority heavy metals monitored on the blacklist of environmental priority pollutants in China, and is also listed as one of the internationally recognized carcinogenic metals [1][2] . Heavy metal pollution has the characteristics of concealment, long-term and irreversibility, so the treatment of heavy metal in wastewater has been highly valued by scientific researchers. There are many methods to treat heavy metals in wastewater. Traditional methods mainly include chemical precipitation, ions exchange, electrochemical method, and membrane filtration method, etc. Adsorption method has attracted wide attention because of its simple operation and strong adsorption capacity. Biochar is a stable carbon rich product synthesized through pyrolysis of biomass. Due to its abundant pore structure and adsorption capacities, biochar has been widely applied to study heavy-metal wastewater treatment [3][4] . Peanut shells are by-products of peanut production. With the exception of a small amount of raw materials used as roughage and fiberboard, most of them are discarded, which not only causes great waste of resources, but also easily leads to environmental pollution. The conversion of peanut shell into activated carbon can improve its application value and solve environmental pollution [5] . There-fore, biochar produced by peanut shell and modified by potassium chloride were selected as adsorbent to study the adsorption properties of Cr (VI) in wastewater. Preparation of Peanut Shell carbon and modified by potassium chloride The peanut shells were carbonized in a continuous vertical biomass carbonization kiln, and the reaction temperature was adjusted to 110-120℃ for 2 hours and 350℃-450℃ for 2 hours. And then peanut shell carbon was prepared and removed from the kiln, and grinded, passed through 80 mesh screen. This prduct was called the original bichar. And then took the original bichar and modified by 0.3mol/L potassium chloride solution [6] . This product was called the modified bichar. The characteristics of organic carbon, specific surface area, cation exchange capacity, alkaline functional group and pH of the original bichar was analyzed [7] . Adsorption experiments When modified peanut shell carbon were added into Cr(VI) solution, the effects of change of the solution pH, absorption time, adsorbent dose and concentration of Cr(VI) on removal rate and adsorption capacity of Cr(VI) were studied. Use diphenylcarbazide (DPC) spectrometry to determine the content of Cr(VI) in the filtrate. The removal rateηand adsorption capacity q were calculated according to the formula as follows: In the formula: C0-the concentration of Cr(VI) in solution before adsorption(μg/ml); C-concentration of Cr(VI) in solution after adsorption (μg/ml) q(mg/g) = (2) In the formula: W-dosage of adsorbents, g; V-volume of solution, ml Analysis of basic Properties of Peanut Shell carbon Table1 Characteristic of peanut shell carbon According to the results in Table 1, peanut shell carbon contained high specific surface area, organic carbon and CEC, as well as rich functional groups. The pH of peanut shell carbon was more than 9, which was alkaline. These properties indicated that peanut shell carbon contained more adsorption sites and had ion exchange capacity. Comparison the effect of original and modified biochar on Cr(VI) removal rate The effect of original and modified biochar on the adsorption of Cr(VI) in water was studied when pH value was 4, absorbent time was 120 min, the concentration of Cr(VI) was 10μg/ml, and absorbent dosage was 0.1g. The removal rate of Cr(VI) in solution of original and modified biochar was 32.6% and 50.3% respectively. The modification of peanut shell carbon by potassium chloride had certain effect. The peanut carbon had negative surface charge which generated electrostatic repulsion to Cr (Ⅵ) in solution with CrO4 2mostly, Cr2O7 2and HCrO4 -. Peanut carbon modified by potassium chloride made its structure loose, which was advantageous to the Cr (Ⅵ) adsorption. Meanwhile, the adsorption of potassium changed the surface charge property of peanut shell carbon and increased the adsorption capacity of Cr (Ⅵ). Effect of pH on Cr(VI) removal rate The pH value of the solution was adjusted to be 2, 4, 5 and 7 respectively with 0.5g of modified peanut shell carbon added into Cr (VI) concentration of 10μg /ml of 25ml in conical bottles. And then the conical bottles were oscillated at 150r/min for 2 h at the temperature of 25℃, filtering, measuring Cr (VI) concentration of the filtrate, and calculating the removal rate of Cr (VI) . The results showed that in the process of increasing pH from 2 to 7, the adsorption removal rate of Cr (VI) in the solution decreased from 95.63% to 43.64%. Cr(VI) in water on biochar adsorption was mainly due to physical adsorption because biochar had a large specific surface area, and these disordered structures had a high surface energy, which can well absorb metal ions in wastewater [8] . Apart from the physical adsorption, there existed electrostatic adsorption [9] . When the pH value was in the range of 1~6, the Cr (VI) in the solution was mainly in the form of CrO4 2and HCrO4 -. A large amount of H + in solution combined with amino, hydroxyl and other functional groups on the surface of modified peanut shell carbon to form -NH4 + and -OH2 + , and adsorbed the CrO4 2and HCrO4in the solution. When pH > 6, the number of positive adsorption centers on bichar surface decreased, resulting in the decrease of electrostatic adsorption. Although the Cr (VI) removal rate was the best at pH 2, the following experiment of solution pH was controlled at 4 because of saving the reagents and the difficulty in treating acidic solution. Effect of adsorbent dosage on Cr(VI) removal rate The modified peanut shell carbon of 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 and 1.0g were added Figure 1. Effects of dosages on Cr(VI) removal rate and adsorption capacity into Cr (VI) solution with concentration of 10μg /ml and volume of 25ml, adjusting the pH value of the solution to 4, oscillating at 150r/min for 2 h at the temperature of 25℃, filtering, measuring Cr (VI) concentration of the filtrate, and calculating the removal rate and the adsorption capacity. The effect of adsorbent dosage was shown in figure 1. The fig.1 showed that when the dosage of adsorbents was the range of 0.1g~ 0.3g, the adsorption removal rate of Cr (VI) in the solution increased from 49.32% to 80.23%, and when it was more than 0.3g, the removal rate decreased from 80.23% to 40.76%. The adsorption capacity had been declining from 1.19 mg.g -1 to 0.09 mg.g -1 . As the amount of adsorbent increased, the total number of adsorption sites increased too, leading to increase in adsorption capacity. However, when the amount of adsorbent increased to a certain extent, the effective adsorption sites decreased due to contact and combination between the adsorbents. Therefore, modified peanut shell carbon of 0.3g was selected as the best dosage for adsorption of Cr (VI) in the following experiment. Effect of adsorption time on Cr(VI) removal rate The modified peanut shell carbon of 0.3g was added into Cr (VI) solution of 10 ug / mL of 25ml, adjusting the pH value of the solution to 4, oscillating at 150r/min for 30, 60, 90, 120, 150 and 180 min at the temperature of 25℃, filtering, measuring Cr (VI) concentration of the filtrate, and calculating the removal rate and the adsorption capacity. Figure 2. Effects of adsorption time on Cr(VI) removal rate and adsorption capacity Figure 2 showed that when the adsorption time was in the range of 0~90min, the removal rate and adsorption capacity increase rapidly. The growth rate became slowly in 90-150min period, reached the peak at 150min, and began to show a downward trend in 150min-180min period. The adsorption surface area changed with the adsorbent dosage. At beginning of adsorption process, the removal rate and adsorption capacity were on the rise because the adsorption kinetics was obvious. With the increase of adsorbent time, most of the adsorption sites on the surface of the adsorbents were occupied by Cr(VI), and it gradually entered the micropores. So the increase of removal rate and adsorption capacity became slowly. When the adsorption time was 150min, the removal rate reached 85.95%. Therefore, 150min was selected as the best absorbent time for adsorption of Cr (VI) by modified peanut shell carbon in the following experiment. Adsorption isotherm The adsorption isotherm is an important basis for studying the adsorption process. It is often used to describe the equilibrium relationship, affinity and adsorption capacity [10] . The adsorption isotherm experiment was carried out by fixing the modified peanut shell carbon of 0.3g and changing the initial content of Cr(VI) solution. The obtained data were fitted by Langmuir linear isotherm and Freundlich linear isotherm, respectively. The results were shown in table 2. Table 2 showed that the adsorption isotherm of Cr(VI) by the modified carbon was fitted by Langmuir model and Freundlich model. The adsorption of Cr (VI) by modified peanut shell carbon was more suitable to be described by Freundlich isotherm model because R 2 was closer to 1. As the Freundlich model is half empirical equation, which is based on the adsorption equilibrium model established on the polyphase surface. So it can be inferred that there was physical adsorption during the adsorption process of modified peanut shell carbon, and the physical adsorption process was multi-molecular layer [11] . Adsorption kinetics studies In order to better describe the kinetic characteristics of Cr(VI) adsorption by modified peanut carbon, quasi-level 1 kinetic equation and quasi-level 2 kinetic equation were used to fit the experimental data, and the kinetic parameters were shown in table 3. 3 showed that the adsorption of Cr (VI) by modified peanut shell carbon was more in accordance with the quasi-second-order kinetic equation which meant the adsorption of Cr(VI) had two process. At the first process, adsorption speed was fast with Cr(VI) diffusion to the surface of adsorbent and adsorption on the surface . At the second process, Cr(VI) entered into the mesopore and micropore of the adsorbent and the adsorption speed was slow [12] . Conclusions Biochar was prepared from peanut shells and modified by 0.3mol/L potassium chloride solution. The removal rate and adsorption capacity of Cr(VI) in water by modified peanut shell carbon were studied by controlling factor, such as pH, adsorbent dose, adsorption time and concentration of Cr(VI). Some conclusions can be drawn as follows: ⑴ The removal rate of Cr(VI) in the solution of peanut shell carbon and modified by potassium chloride was 32.6% and 50.3% respectively. The modification of potassium chloride had certain effects on remove of Cr(VI) in water. ⑵ The optimum conditions for adsorption of Cr(VI) in water by modified peanut shell carbon were as follows: The removal rate of Cr(VI) in water reached 99.86% when pH was 4, the dosage was 0.3 g, the adsorption time was 150 min, and the initial concentration of Cr (VI) ion was 8μg/ml. (3) Adsorption models showed that the removal behaviors of Cr(VI) by modified peanut shell carbon followed Freundlich and quasi-second-order kinetic equation, and the linear correlation coefficient was 0.96 and 0.99 respectively, which meant the adsorption was physical process and had multi-molecular layer.	2019-11-14T17:03:43.150Z	2019-11-08T00:00:00.000	{ "year": 2019, "sha1": "ff9047a7dce0aaba659d16cfed3fdb5e66039d8a", "oa_license": null, "oa_url": "https://doi.org/10.1088/1755-1315/330/4/042033", "oa_status": "GOLD", "pdf_src": "IOP", "pdf_hash": "12ebbf5464f0fd836aa5cc66cb676e96d33fd12b", "s2fieldsofstudy": [ "Agricultural And Food Sciences", "Engineering" ], "extfieldsofstudy": [ "Physics", "Chemistry" ] }
54079931	pes2o/s2orc	v3-fos-license	Mycelial compatibility groups and pathogenic diversity in Sclerotium rolfsii populations from sugar beet crops in Mediterranean‐type climate regions The population structure of Sclerotium rolfsii from autumn‐sown sugar beet crops in Mediterranean‐type climate regions of Chile, Italy, Portugal and Spain was determined by analyses of mycelial compatibility groups (MCGs) and pathogenicity to 11 economically important plant species. Twelve MCGs (i–xii) were identified among 459 S. rolfsii isolates. MCG iii was the most prevalent group in all countries except Italy. MCG i, the most abundant group (64·7% of isolates) was identified in Portugal and Spain. The remaining MCGs were restricted to various regions within one country (ii, vi, ix) or different countries (v), or to specific localities (iv, vii, viii, x, xi, xii). MCGs iv, vii and x each comprised one isolate. Fields extensively sampled in southern Spain were infected with one to three MCGs. Plant species differed in susceptibility to MCG tester isolates with a MCG by species interaction. Cluster analyses allowed selection into five MCG groupings and grouped plant species into species‐groups 1 (broccoli, chickpea, sunflower, tomato) and 2 (cotton, pepper, sugar beet, watermelon). MCG groupings 1 (i, ix), 2 (ii, iii, vi, viii) and 5 (x, xii) were moderately virulent to species‐group 1 and mildly virulent to species‐group 2. MCG groupings 3 (iv, v, xi) and 4 (vii) were mildly virulent to both species‐groups. Across MCG groups, species were rated highly susceptible (chickpea, sunflower), susceptible (cotton, pepper, tomato, watermelon), moderately resistant (broccoli, melon, sugar beet) and resistant (corn, wheat). Establishing the MCG population structure and virulence variability among S. rolfsii isolates should help in the management of sclerotium root rot diseases. Introduction Sclerotium rolfsii is the mycelial stage of the basidiomycete Athelia rolfsii, a worldwide soilborne plant pathogenic fungus that lacks a conidial stage. The host range of S. rolfsii comprises over 500 plant species, mostly dicotyledons, but also some monocotyledons (Harlton et al., 1995;Okabe et al., 1998;Cilliers et al., 2000). The main disease symptoms caused by this pathogen include crown and root rot, stem canker or damping-off. The resulting disease is referred to as southern blight, southern stem rot, or sclerotium root rot (Aycock, 1966;Punja, 1985). S. rolfsii produces abundant coarse, white mycelia on infected host tissues and forms sclerotia (Punja, 1985). The fungus overwinters in soil by means of sclerotia, as well as by mycelia in infected plants or infested plant debris (Punja, 1985). It can form a sexual stage, but this rarely occurs in nature and its role in the life cycle of the fungus is unknown (Nalim et al., 1995). Sclerotium rolfsii has been reported from nearly every country between northern and southern latitudes of 38°. However, disease epidemics are most severe in warm temperate and subtropical regions that favour sclerotial germination, mycelial growth, infection and subsequent disease development (Aycock, 1966;Punja, 1985). These climatic conditions occur in the Mediterranean basin, southern California and Chile, where outbreaks of S. rolfsii often occur on autumn-sown sugar beet (Beta vulgaris) crops. Disease surveys conducted in Andalusia, southern Spain, in 2004 showed that sclerotium root rot of sugar beet is widespread in autumn-sown crops in the southernmost part of the region, causing 5-80% yield loss (AIM-CRA, 2005). Similarly, the disease severely affects sugar beet crops in central and southern Portugal (M. Paim, Associaçã o para Desenvolvimento da Beterraba, Portugal, personal communication) and the central area of Chile (regions VI to VII) (R. Paillalef-Monnrad, IANSAGRO, Chile, personal communication). As a cosmopolitan soilborne plant pathogen, knowledge of its genetic diversity and distribution in infested areas would be useful for the management of sclerotium root rot diseases. Characterization of fungal populations into mycelial compatibility groups (MCGs) has been widely used for indirectly assessing the genetic variability among isolates of fungal plant pathogens (Leslie, 1993). MCGs play an important role both in defining field populations of fungi and facilitating genetic exchange in fungal species. This is particularly relevant for fungi for which sexual reproduction has a minimal impact on the disease cycle (Kohn et al., 1991;Leslie, 1993) because MCGs help to preserve the identity of genetically dissimilar mycelia and restrict the exchange of cytoplasm, genetic material and extrachromosomal elements (Caten, 1972). Mycelial incompatibility results in separate and distinct gene pools, which may differ in ecological, physiological and pathological traits. Mycelial compatibility studies have been used to estimate genetic variability or relatedness within many fungal species of the Basidiomycota such as Sclerotium cepivorum (Earnshaw et al., 2000) and Rhizoctonia solani (Carling et al., 2002;Ciampi et al., 2008), as well as ascomycetes such as Fusarium spp. (Puhalla, 1985;Elias et al., 1993;Chulze et al., 2000;Bayraktar et al., 2010), Cryphonectria spp. (van Heerden & Wingfield, 2001), Verticillium dahliae (Korolev et al., 2000;Collado-Romero et al., 2006) and Sclerotinia sclerotiorum (Kohn et al., 1991). Mycelial compatibility was demonstrated to occur between field isolates of S. rolfsii, and has been used to designate MCGs (Punja & Grogan, 1983) as well as to assess the genetic variability among S. rolfsii populations (Punja & Grogan, 1983;Harlton et al., 1995;Nalim et al., 1995;Cilliers et al., 2000Cilliers et al., , 2002Okabe & Matsumoto, 2000;Almeida et al., 2001;Punja & Sun, 2001;Sarma et al., 2002;Adandonon et al., 2005). These studies suggested that MCGs in S. rolfsii can be associated with either the host source or geographical area of isolates, but also a single MCG can comprise isolates from diverse geographical origins and host sources. The variability of mycelial compatibility among S. rolfsii populations occurring in particular geographic regions or specific crops has also been demonstrated. Studies in Brazil (Almeida et al., 2001), India (Sarma et al., 2002) and South Africa (Cilliers et al., 2002) indicated that a considerable genetic variability existed among S. rolfsii isolates from various hosts and localities without clear correlation between the MCG to which they belonged and their host source or geographic origin (Almeida et al., 2001;Sarma et al., 2002), although a certain relationship between MCGs and host source of isolates was suggested for S. rolfsii populations in South Africa (Cilliers et al., 2002). On the other hand, in peanut, wide variation was reported among isolates within a field in Texas, USA (Nalim et al., 1995) and Japan (Okabe & Matsumoto, 2000), with a single peanut field harbouring up to three (Okabe & Matsumoto, 2000) or five (Nalim et al., 1995) MCGs. Control of sclerotium root rot diseases is difficult because of the extensive mycelial growth, persistence in soil, genetic variability of populations and wide host range of S. rolfsii. In sugar beet, no resistant cultivars are available and attempts to control the disease with fungicides or biological control agents have been unsuccessful (Lal et al., 1997). However, a degree of resistance to S. rolfsii was found in some cultivars of cowpea (Fery & Dukes, 2002), alfalfa (Pratt & Rowe, 2002), peanut (Branch & Brenneman, 1999), pepper, sweet potato (Dukes et al., 1983) and chickpea (Akram et al., 2008). Reports from California and India indicate that the pathogen inoculum density can be effectively reduced by rotations of sugar beet with crops that are slightly susceptible to S. rolfsii, such as alfalfa, asparagus, barley, corn and wheat (Schneider & Whitney, 1996). This same strategy was also effective with rotations of susceptible carrot with sweet potato or buckwheat (Jenkins & Averre, 1986) and peanut with bahiagrass, corn, cotton (Johnson et al., 1999) or wheat (Minton et al., 1991). Knowledge about the genetic and virulence diversity in local populations of S. rolfsii associated with different crops is a key component for the management of sclerotium root rot diseases, particularly through the use of host resistance and crop rotation in a given region. In addition, knowledge of the host range of S. rolfsii populations present in a given area is essential for recommending suitable crops as an alternative to sugar beet production. Unfortunately, information of that nature is lacking in areas of Mediterranean-type climate. Furthermore, the sugar beet production area in the Mediterranean Basin has steadily decreased in Italy, Portugal and southern Spain during the last few years, being replaced by vegetable crops that can be potentially affected by sclerotium root rot. In the present study, a large collection of S. rolfsii isolates from sclerotium root rot-affected sugar beet crops in four countries with a Mediterranean-type climate and eight intensively sampled fields in southern Spain were used to: (i) assess S. rolfsii MCG diversity and geographical distribution; (ii) determine within-field diversity and prevalence of S. rolfsii MCGs in sugar beet fields of southern Spain; (iii) determine any correlation that might exist between MCGs of S. rolfsii isolates from sugar beet and their pathogenicity and virulence to 11 agricultural crops; and (iv) determine the pathogenic variability within MCGs on three selected susceptible hosts. Materials and methods Sampling and isolation of the pathogen Two sampling strategies were used to account for variability of S. rolfsii on different spatial scales. Sugar beet field plots located in southern Spain were intensively sampled to determine variability at field level. For that purpose, an area of c. 0AE06 ha in each field plot was divided into a regular grid of 80 (4-· 2-m) quadrants, and one affected sugar beet root was sampled from each quad-rant. For the second spatial scale, whole fields in Chile, Italy and Portugal were arbitrarily selected to determine pathogen variability among fields with at least four arbitrarily chosen affected roots sampled from each field. In all cases, the sampled area was representative of the sugar beet fields affected by sclerotium root rot in each country. For isolation of S. rolfsii, root tissues with symptoms were washed under running tap water, surface-sterilized in 0AE5% NaOCl for 1 min and blotted dry between sterile filter papers. Pieces of surface-sterilized tissues (2-4 mm 2 ) were plated onto potato dextrose agar (PDA, Difco Laboratories) amended with 0AE25 mL 85% lactic acid and incubated at 25 ± 1°C in the dark for 2-5 days. Pure cultures forming sclerotia were obtained from each root sample. Single-sclerotia cultures obtained from pure cultures were incubated to form abundant sclerotia and were allowed to dry at room temperature. Sclerotia from dry cultures were collected, dried and stored in paper bags at room temperature until use. Determination of mycelial compatibility groups A single sclerotium from each of the S. rolfsii isolates was placed on a PDA plate and incubated in the dark at 25 ± 1°C for 5 days. Then, small plugs from the inner colony were transferred to modified Patterson's medium (MPM) in new plates (90-mm diameter) and incubated for 5 days in the same conditions. This MPM allows improved assessment of mycelial interactions among S. rolfsii isolates (E. Remesal & J. A. Navas-Cortés, unpublished results). Mycelial discs from the edges of actively growing MPM colonies of isolates were paired on MPM plates. Three isolates were paired on a 90-mmdiameter plate by placing discs 50 mm apart and incubating at 25 ± 1°C in the dark. There were four replicated plates per pairing. All isolates were paired with themselves to ensure self-compatibility. All pairings were repeated twice. Pairings were examined macroscopically after 5 and 10 days of incubation for the presence of an antagonism zone in the region of mycelial contact (Punja & Grogan, 1983) and the presence of a coloured red line in the reaction zone between the colonies ( Fig. 1a,b). Mycelial compatibility reactions showed no antagonism and were distinguished by merging colonies with no detectable coloured red line between the interacting colonies ( Fig. 1c). MCGs were identified based on data from compatible reactions among tested isolates. Because of the high number of S. rolfsii isolates and difficulty of pairing them in all possible combinations, MCGs were defined first at field plot level, pairing isolates from a plot in all possible combinations. Then, one tester isolate was arbitrarily selected from each of the identified MCGs for further analyses. Representative tester isolates from all the identified MCGs in each field plot were paired in all possible combinations (Nalim et al., 1995). The diversity of MCGs in each country and ⁄ or field plot in the study was estimated by the Shannon-Weaver diversity index: H¢ = )Rp i · lnp i , where p i is the frequency of the ith MCG (Shannon & Weaver, 1963). Experimental design Three experiments (I-III) were carried out in a growth chamber to determine the reaction of plant species to S. rolfsii isolates. Experiments I and II were first performed to determine reaction of the 11 plant species to inoculation with 12 tester isolates representative of each of the 12 identified MCGs. Because of space limitations in the growth chamber, experiments I and II consisted of six and five plant species, respectively. To test for reproducibility of reactions to the 12 MCG tester isolates, sunflower and tomato were common plant species in the two experiments. In experiment III, a total of 23 S. rolfsii isolates representative of five of the more abundant MCGs identified, which comprised 5% of the total number of isolates in the study, were used to assess within-MCG variation in pathogenicity and virulence. Isolates tested included 12 isolates for MCG i, four for MCG iii, three for MCG v and two for each of MCG vi and MCG xii. For this experi-ment, sunflower, tomato and watermelon were selected as susceptible hosts based on the consistency of their reaction to tester MCG isolates in experiments I and II. In all three experiments, there were eight replications (eight pots, one plant per pot) of each treatment combination of plant species and S. rolfsii isolate. Each experiment was repeated partially and preliminary analyses of common treatments indicated no significant differences (P > 0AE05) among experiments (see below); consequently, data were pooled for further analyses. For each plant species, seeds were surface-sterilized in 2% (v ⁄ v) NaOCl for 30 s, and germinated on autoclaved layers of filter paper in moist chambers at 25 ± 1°C in the dark for 2 days. Germinated seeds, selected for uniformity (length of radicle = 0AE5-1 cm), were sown in sterile plastic pots filled with an autoclaved (121°C, 1 h, twice, on two consecutive days) soil mixture (silt:peat, 2:1). Sclerotium rolfsii inoculum, plant inoculation and growth conditions Inoculum for experiments consisted of infested oat (Avena sativa) seeds. One hundred grams of seeds were moistened in 50 mL water and autoclaved in 1-L flasks for 60 min at 121°C. The sterilized seeds were then infested using 10 mycelial discs, 5 mm in diameter, from the growing edge of a S. rolfsii colony on PDA and incubated at 25 ± 1°C in the dark for 2 weeks. Plants were inoculated at the two-leaf stage. For inoculation, two selected oat seeds heavily and homogenously colonized with mycelia of an isolate were buried at a depth of 0AE5 cm and a distance of 0AE5-1 cm from the plant stem. Plants in pots with non-infested seeds served as controls. Plants were incubated in a walk-in growth chamber adjusted to 28 ± 1°C with a 14-h photoperiod of fluorescent light of 360 lE m )2 s )1 and 60-90% relative humidity for 18 days. Plants were watered as needed to maintain field capacity and fertilized weekly with 100 mL 0AE1% hydro-sol fertilizer solution (Haifa Chemicals, Ltd, 20-5-32 N-P-K + micronutrients). tion. S. rolfsii isolates were classified in terms of virulence based on the disease incidence (DI) that they induced: highly virulent (DI ‡ 66%), moderately virulent (66 < DI ‡ 33%), mildly virulent (33 < DI > 0%) and non-pathogenic (DI = 0%). Similarly, plant species were classified according to DI values as: highly susceptible (DI ‡ 66%), susceptible (66 < DI ‡ 33%), moderately resistant (33 < DI > 0%) or resistant (DI = 0%). The number of dead plants per treatment was used for statistical analyses. Data were analysed with the GENMOD procedure using the binomial distribution and the logit as link function in SAS (version 9AE2, SAS Institute Inc.) (Agresti, 2002). Least squares means were computed for all response variables to allow multiple comparisons between ⁄ among treatments by using the SAS macro MULT (https://www.uni-hohenheim.de/bioinformatik/beratung/ toolsmacros/sasmacros/mult.sas) (Piepho, 2004). Linear single-degree-of-freedom contrasts were computed to test the effect of selected experimental treatment combinations. The stability of disease reactions in the plantspecies by MCG combination was explored by cluster analyses of the number of dead plants to identify associated groupings of MCGs and plant species in two separate analyses. To find functional groupings of correlated MCGs or plant species, respectively, an agglomerative clustering based on the Spearman correlation matrix was performed among MCGs and among plant species using the Ward clustering method. For MCGs, the optimum number of clusters was estimated on the basis of the average silhouette width according to the Mantel statistic. Thus, the number of clusters in which the within-group mean intensity of the link of the objects (MCG) to their groups was highest (i.e. with the largest average silhouette width) indicated the optimum cluster number. A new dendrogram was then produced representing the identified groupings of MCGs (Borcard et al., 2011). All calculations for cluster analyses were made using R version 2AE13AE0 (R Foundation for Statistical Computing, http://www.R-project.org/) with the CLUSTER (Maechler et al., 2005) and VEGAN (Oksanen et al., 2011) packages. Disease assessment and data analyses The degree of virulence variation within each of five MCGs tested in experiment III was estimated by a separate likelihood analysis using the GENMOD procedure of SAS, as described above, followed by multiple comparison of mean number of dead plants induced by the S. rolfsii isolates tested. In addition, the disease reaction induced by the tester isolate of each of the MCGs was compared with that of each of the rest of isolates within the same MCG using linear single-degree-of-freedom contrast at P < 0AE05. Homogeneity of disease reactions across experiments The consistency of results between experiments I and II was tested by a preliminary analysis in which the number of dead plants for treatments common to both experiments (i.e. the plant species sunflower and tomato inoculated with each of 12 MCG-representative isolates) were analysed with the GENMOD procedure of SAS as described above. The likelihood ratio test indicated no significant effects of the two experimental runs (v 2 = 0AE35, P = 0AE5536), or their interaction with MCG isolate (v 2 = 9AE22, P = 0AE6017), plant species (v 2 = 2AE21, P = 0AE1367), or both factors (v 2 = 13AE14, P = 0AE2844). Similarly, homogeneity of disease reactions in experiments I, II and III was tested with the same methodology as indicated above using the number of dead plants of common treatments in all three experiments (i.e. the plant species sunflower and tomato inoculated with isolates representative of each of MCGs i, iii, v, vi and xii included in experiment III). The likelihood ratio test indicated no significant effects of the three experimental runs (v 2 = 0AE27, P = 0AE8740), or their interactions with MCG isolate (v 2 = 4AE25, P = 0AE8336), plant species (v 2 = 2AE05, P = 0AE3595), or both factors (v 2 = 11AE53, P = 0AE1732). Mycelial compatibility groups Twelve MCGs were identified among the 459 isolates of S. rolfsii tested in the study. MCGs were designated in roman numerals from i to xii. The MCGs of isolates and their geographical locations are shown in Table 1. Mycelia compatibility among isolates of a MCG was characterized by their mycelia intermingling at the zone of interaction without aversion (Fig. 1c). Conversely, isolates assigned to different MCGs showed aversion at the interaction zone, together with thinned mycelium and formation of a red reaction line (Fig. 1a,b). All 459 isolates were self-compatible, showing a reaction similar to that described for compatible isolates. Replicated experiments and replicated pairings within each experiment produced identical results. Distribution and frequency of the 12 identified MCGs differed among sampling regions and countries. MCGs i and iii were the most prevalent among sampled isolates, representing 64AE71 and 18AE95% of the isolates, respectively. MCG i was identified only in the Iberian Peninsula, being present in all eight locations sampled in southern Spain and one location sampled in central Portugal (Table 1; Fig. 2). MCG iii was less abundant than MCG i but was the most widely distributed MCG, being identified in 10 of the 18 sampled locations from all sampled countries [Chile (two), Portugal (five) and southern Spain (three)] except Italy (Table 1; Fig. 2). MCG v comprised 5AE66% of the isolates and was restricted to two distant field plots located in Vejer de la Frontera (Cá diz province) and Vila Franca de Xira (Lisbon province) in southern Spain and central Portugal, respectively. MCG ii comprised 3AE92% of the isolates from two locations in Có rdoba and Seville provinces in southern Spain. The remaining eight MCGs accounted for 6AE75% of the isolates and were locally distributed. MCGs ix, vi, xii, viii and xi comprised 10, nine, four, three and two isolates, respectively; among these five multimember MCGs, MCGs viii, ix and xii were identified only in three loca-tions in Chile, MCG vi in two locations in Portugal, and MCG xi in one location in Spain. MCGs iv, vii and x each consisted of a single isolate, identified in Bari (southern Italy), Colchagua (Chile) and Chillá n (Chile), respectively. Regarding sampled countries, six, five and four MCGs were identified in Chile (MCGs iii, vii, viii, ix, x and xii), Spain (MCGs i, ii, iii, v and xi) and Portugal (MCGs i, iii, v and vi), respectively. The isolate sampled in southern Italy was assigned to MCG iv and occurred in this location only. Accordingly, the highest Shannon-Weaver H¢ index of diversity, 1AE49, was estimated in Chile, decreasing to 0AE95 and 0AE74 in Portugal and Spain, respectively. In southern Spain, MCG i was present in all locations sampled, whereas MCGs iii and v were identified in Cá diz, in the southwest of the region, MCG xi was found only at Huelva in the northwest of the region, and MCGs ii and iii occurred at Có rdoba and Seville, in the centre of the region in the Guadalquivir valley (Fig. 2). At the field level, the eight intensively sampled fields in four provinces in southern Spain showed a high degree of MCG homogeneity (Table 1; Fig. 3). The number of MCGs per field ranged from one to three and MCG i was present in all fields. A single field located at Posadas (Có rdoba province) consisted of MCGs i, ii and iii, 12AE9, 54AE8 and 32AE3% of a total of 31 isolates, respectively, and yielding a Shannon-Weaver index of H¢ = 0AE96. Two MCGs were identified in five fields; they included MCG i together with either MCGs ii (1AE3% of isolates, H¢ = 0AE07), iii (two fields, 14AE0 and 89AE3% of isolates, H¢ = 0AE40 and H¢ = 0AE34, respectively), v (45AE7% of isolates, H¢ = 0AE69) or xi (3AE1% of isolates, H¢ = 0AE14) ( Table 1; Fig. 3). In those fields, S. rolfsii isolates of a MCG tended to form discrete clusters of different sizes distributed along the plot, but there was no discernible spatial pattern for individual MCGs (Fig. 3). Pathogenic diversity Susceptible reactions consisted of initial pale green and wilted leaves, followed by complete collapse and death of plants. Most plant mortality occurred within the first 8 days after inoculation. Symptoms developed neither in non-inoculated controls nor in resistant plants. Overall, MCGs iii and MCGs vi had the widest host range, with their isolates being pathogenic to all plant species tested, except for wheat and corn, which were resistant to all 12 MCGs. Isolates of MCGs i, ii, vii, ix and x were pathogenic to eight plant species, whereas isolates of MCGs viii, xi and xii were pathogenic to seven plant species, and those of MCGs iv and v were pathogenic to six (Fig. 4a). Three of the 12 tested plant species, chickpea, pepper and watermelon, were infected by all 12 MCGs, whereas cotton and sunflower were infected by 11 MCGs, tomato by 10, sugar beet by nine, broccoli by eight and melon by only six MCGs (Fig. 4b). Melon plants were not infected by six of the 12 MCGs tested. DI values were low for the remaining MCGs, reaching 12AE5% in inoculations with the representative isolate of each of MCGs iii to vi, and 25 and 37AE5% with those of MCG ii and MCG i, respectively, and were not included in the statistical analyses (data not shown). Global likelihood ratio analysis of the number of dead plants for the combinations of isolates representative of each of MCGs i to xii and the remaining eight plant species indicated a significant effect of MCG (v 2 = 33AE37, P = 0AE0005), plant species (v 2 = 68AE04, P < 0AE0001) and the MCG by plant-species interaction (v 2 = 104AE17, P < 0AE0213) (data not shown). To further analyse the MCG by plant-species interaction, functional groupings of MCGs were first identified using clustering silhouettes. In this analysis, four to seven cluster groupings could be appropriate according to the silhouettes' mean width. This allowed selection of five functional MCG groupings among the 12 MCGs of the study which included: MCG grouping 1 (MCGs i and ix), MCG grouping 2 (MCGs ii, iii, vi and viii), MCG grouping 3 (MCGs iv, v and xi), MCG grouping 4 (MCG vii) and MCG grouping 5 (MCGs x and xii) (Fig. 5a). Similarly, plant species could be grouped into two plant species (SP) groups, namely SP group 1 (chickpea, broccoli, sunflower and tomato) and SP group 2 (cotton, pepper, sugar beet and watermelon) according to the two-first-order cluster groups identified in the cluster analysis (Fig. 5b). Thereafter, a new analysis was performed using MCG groupings and SP groups as independent variables, which indicated significant effects of both factors (P £ 0AE0495), as well as of their interaction (P = 0AE0012) ( Table 2). This interaction was the result of disease incidence induced by isolates of MCG groupings 1, 2 and 5 being higher (P £ 0AE0319) on SP group 1 than on SP group 2, whereas isolates of MCG groupings 3 and 4 induced similar disease incidence (P ‡ 0AE0902) on SP groups 1 and 2 (Table 2). Additionally, a separate analysis was performed for each SP group to further assess the effects of MCG groupings on disease incidence induced by S. rolfsii isolates on the plant species included in each SP group. Results indicated no significant (P ‡ 0AE05) MCG-grouping by plant-species interaction for the two SP groups (Table 3). Overall, for SP group 1, disease reaction to inoculation was significantly influenced both by the MCG grouping and plant species (P < 0AE0001) but not by their interaction (P = 0AE2026) ( Table 3). DI was highest (P < 0AE05) for isolates of MCG groupings 1, 2 and 5, which showed moderate virulence, and decreased significantly (P < 0AE05) for MCG groupings 3 and 4, which were mildly virulent (Fig. 6a). Among plant species, DI was highest (P < 0AE05) on chickpea, which was categorized as highly susceptible, decreased significantly (P < 0AE05) on sunflower, which was categorized as susceptible, and was least (P < 0AE05) on tomato and broccoli, which were considered moderately resistant with no significant differences (P ‡ 0AE05) between these two species (Fig. 6a). No significant differences (P = 0AE0681) occurred among MCG groupings on SP group 2 (Table 3; Fig. 6b). Across plant species in SP group 2, disease incidence was highest (P < 0AE05) on cotton and pepper, which showed susceptible reactions, and decreased significantly (P < 0AE05) to a moderately susceptible reaction on sugar beet. DI on watermelon was intermediate between those occurring on the three species of SP group 2 and did not differ significantly (P ‡ 0AE05) from any of them (Fig. 6b). Finally, the disease reaction induced by isolates of MCGs included in each of the four multimember MCG groupings (i.e. MCG groupings 1, 2, 3 and 5) on SP groups 1 and 2 was fairly homogeneous (P ‡ 0AE0566). This supports the consistency of the five MCG groupings identified in the cluster analysis. Within-MCG pathogenic variability There was little variation in the disease reaction induced on sunflower, tomato and watermelon by S. rolfsii isolates belonging to the same MCG (Fig. 7). Thus, likelihood analyses showed no significant differences (P ‡ 0AE5350) among S. rolfsii isolates belonging to either MCG iii, v, vi or xii (Table 4). Moreover, disease reactions induced by isolates of each of these four MCGs were not significantly different (P ‡ 0AE05) from those in experiments I and II. An exception to this occurred for isolates of MCG i, which showed significant differences between each other (P < 0AE0001) (Table 4, Fig. 7). In any case, of the 12 MCG i isolates tested, only the disease reaction induced by isolate Sr033 differed significantly (P = 0AE0068) from that induced by the MCG i type isolate Sr078 (Fig. 7). Comparison of the disease reaction induced by this isolate with those induced by the rest of the MCG i isolates indicated that Sr033 was more virulent to all three plant species (Fig. 7). For all five MCGs tested, there were no significant (P ‡ 0AE4638) isolate by plant-species interactions (Table 4). terranean-type climate were characterized by means of MCG analysis and pathogenicity to 11 economically important plant species cultivated in those regions. It was demonstrated that differences in virulence occur among isolates from certain locations and that these differences can be associated with genetic diversity based on MCGs of isolates. In addition, variation at field level was assessed by intensive sampling of plots at several locations in southern Spain. A system of mycelial incompatibility was previously demonstrated to occur among field isolates of S. rolfsii and was used to designate MCGs (Punja & Grogan, 1983). This study was able to identify up to 12 MCGs among 459 S. rolfsii isolates originating from a wide geographic range but a single host crop. This is apparently novel work in S. rolfsii because previously published work has involved S. rolfsii isolates from: (i) widely distant geographic areas and diverse host sources (Punja & Grogan, 1983;Harlton et al., 1995;Punja & Sun, 2001); (ii) a restricted region and diverse hosts (Cilliers et al., 2000(Cilliers et al., , 2002Almeida et al., 2001;Sarma et al., 2002); and (iii) a single region and host plant (Nalim et al., 1995;Okabe & Matsumoto, 2000;Adandonon et al., 2005). Two abundant and prevalent MCGs (MCGs i and iii) were found. Some other MCGs were: (a) less prevalent but present in various regions of different countries (MCG v) or regions within a country (MCGs ii, vi and ix); (b) restricted to specific localities (MCGs viii, xi and xii); or (c) single-isolate member groups (MCGs iv, vii and x). The results agree with most previous reports of a single S. rolfsii MCG comprising isolates from widely distant geographic areas of Brazil (Almeida et al., 2001), India (Sarma et al., 2002), South Africa (Cilliers et al., 2000) and the USA (Harlton et al., 1995;Punja & Sun, 2001). In this same context, the existence of MCGs at high frequencies with wide geographic distribution was also reported for S. sclerotiorum infecting canola (Brassica napus) through Canada, where a single MCG representing 18% of the isolates was found in all three distant provinces sampled (Kohli et al., 1992). In southern Spain, S. rolfsii was reported infecting sugar beet crops in some localities in Có rdoba and Seville provinces in the centre of Andalusia in the early 1940s (Benlloch, 1943). By 1944, the pathogen had also been reported in Cá ceres province in the Extremadura region of southwest Spain, close to the Portuguese border (Domínguez García-Tejero, 1951). Interestingly, this coincided with when the crop sowing date was brought forward from spring to early winter (Morillo-Velarde et al., 2003), which appears more favourable for disease development (Aycock, 1966;Punja, 1985). In autumn sowings, several important changes were implemented in the husbandry of sugar beet, which included complete mechanization of harvesting using heavy equipment. Thus, movement of equipment across fields, together with transportation of harvested roots from fields to distant processing factories, may have facilitated dispersal of the pathogen in soil particles and could be responsible for the rapid spread of MCGs i and iii through Andalusia (Fig. 2). On the other hand, sugar beet has been extensively grown in Portugal since the late 1990s (FAOSTAT, 2010), and sclerotium root rot was of great concern in the major growing areas of central and southern Portugal in 2003, at which time a disease prevalence of 21% of fields Sclerotium rolfsii isolates representative of mycelial compatibility groups (MCGs) grouped into five MCG groupings (x-axis) and eight plant species (y-axis) grouped into species-group (SP) 1 (a) and SP group 2 (b). Plant species: Br, broccoli; Ck, chickpea; Ct, cotton; Pp, pepper; Sb, sugar beet; Sf, sunflower; To, tomato; and Wm, watermelon. For each SP group, MCG groupings or plant species followed by a different letter showed significant differences on the least squares means at P < 0AE05. MCG groupings with an asterisk indicate a significantly higher least squares mean for disease incidence induced on SP group 1, compared to that on SP group 2 at P < 0AE05. accounted for 16AE6% of the cultivated area being affected (M. Paim, Associaçã o para Desenvolvimento da Beterraba, Portugal, personal communication). In Chile, sugar beet has been grown since the early 1950s and S. rolfsii was first reported in the central regions affecting diverse crops in 1981 (Esterio & Auger, 1982) and on sugar beet in the Metropolitan region and V region in 1983 (Acuñ a, 1985). Thereafter, the pathogen spread to the southern regions reaching the VIII region in 1999 (R. Paillalef-Monnrad, IANSAGRO, Chile, personal communication). MCG iii was the most prevalent group in the current study, being present in two of four, three of eight and all five locations sampled in Chile, Spain and Portugal, respectively. This suggests that MCG iii may have been spread extensively within and among cultivated areas by cultural practices, vehicle movements, planting material, containers, etc. The existence of S. rolfsii isolates of the same MCG in different distant countries was also reported by Harlton et al. (1995), who identified MCG 8 in the USA (Alabama, California and Maryland) and Pakistan; and MCG 1 in the USA (California, Georgia and North Carolina) and Mexico. Similarly, Kull et al. (2004) reported MCG 8 of S. sclerotiorum in soyabean fields in several states in the USA, on soyabean in Switzerland and on canola in Canada. The present results are also consistent with those of other authors, suggesting a correlation between MCG and geographical source of isolates, as S. rolfsii isolates originating from the same locality tended to form a distinct MCG (Punja & Grogan, 1983;Harlton et al., 1995;Cilliers et al., 2000Cilliers et al., , 2002Punja & Sun, 2001;Sarma et al., 2002). In fact, three of the 12 MCGs identified (MCGs ii, vi and ix) were found only at two close localities and six were present at a single locality. This was the case of MCG iv, which was identified exclusively in Bari (Italy), MCGs vii, viii, x and xii, which were present in each of four different localities in Chile, and MCG xi, which was identified only in Huelva (Spain) (Fig. 2). In a comparable study, Nalim et al. (1995) found that a rather small number of MCGs existed within single peanut fields in Texas. Localization of unique S. sclerotiorum MCGs was also observed in vegetable-growing regions in New Zealand (Carpenter et al., 1999) and on winter canola in Ontario, Canada (Kohn et al., 1991). Explanations for unique low-frequency MCGs in a sampling area may include recent MCG introductions or random mutation events. The emergence of new genotypes localized in single fields may be an indication of MCGs or clones becoming adapted to specific field microclimates or hosts, and are less likely to be the result of genetic exchange and recombination (Ben-Yephet & Bitton, 1985;Hambleton Figure 7 Two-way plot showing results of likelihood analyses of disease reaction estimated by the number of dead plants of sunflower (Sf), tomato (To) and watermelon (Wm) (x-axis) to infection by 23 Sclerotium rolfsii isolates representative of five of the more prevalent mycelial compatibility groups (MCG) (y-axis) identified in this study. Isolates tested included 12 isolates for MCG i, four isolates for MCG iii, three for MCG v, and two for each of MCG vi and MCG xii. The type isolate for each MCG is underlined. For each MCG, isolates followed by a different letter showed significant differences in the least squares means at P < 0AE05. Isolates with an asterisk indicate significantly different least squares means from that of the type isolate at P < 0AE05. , 2002). The existence of single-member MCGs was also reported with variable frequency in different geographic areas and ⁄ or host sources (Punja & Grogan, 1983;Harlton et al., 1995;Nalim et al., 1995;Cilliers et al., 2000;Okabe & Matsumoto, 2000;Almeida et al., 2001;Punja & Sun, 2001;Sarma et al., 2002). The population of the pathogen in Chile was particularly diverse. The area sampled for the study in that country was similar in size to that in southern Spain. However, MCGs identified in Chile accounted for half of the 12 MCGs identified in the study, of which five were found only in this country. Moreover, a single location in Chile could contain up to three different MCGs, as occurred in Linares (MCGs viii, ix and xii) and Chillá n (MCGs iii, ix and x). The existence of considerable genetic variability among S. rolfsii isolates was also reported in Brazil, where 13 compatibility groups were identified using 30 isolates from 14 states and 11 hosts (Almeida et al., 2001). At field-plot level, a relatively low number of MCGs was found within each of the eight sugar beet fields sampled in the Guadalquivir Valley in southern Spain. In these fields, three, two and one MCGs were identified in one, five and two fields, respectively. MCG i was always present, being unique and predominant in two and four fields, respectively (Figs 2 and 3). Interestingly, the number of MCGs in a field was not related to the number of S. rolfsii isolates sampled in it. Thus, for a similar sampled area the number of isolates ranged from 28 to 80, the field harbouring three MCGs being represented by 31 isolates. These results do not fully agree with those reported by Nalim et al. (1995), who studied four peanut fields in Texas and found one, three and five MCGs in one, two and one field, respectively. However, none of the MCGs reported by Nalim et al. (1995) were prevalent across fields, as only one MCG was present in two fields, while the others were specific to each field. In contrast, the results of the present study are closer to those of Okabe & Matsumoto (2000) in peanut fields in Japan. These authors analysed seven peanut fields and identified four MCGs, with up to three MCGs occurring in a single field and one MCG being dominant in most of the fields sampled. In similar studies on S. sclerotiorum, Kohn et al. (1991) found much higher variability within two canola fields in Ontario, Canada, where they identified 26 and six MCGs from 30 and 33 isolates, respectively, with two MCGs being predominant in one field, and only one MCG being common to the two fields. Similar results were described for the MCG population structure of S. sclerotiorum on soyabean in Argentina and Illinois (USA) (Kull et al., 2004). The present observations and those of Nalim et al. (1995) and Okabe & Matsumoto (2000) are coherent with a 'founder effect' and may reflect the mode of reproduction of S. rolfsii, which does not produce conidia and spreads by mycelial growth. Whilst S. rolfsii does have a perfect stage, its role in nature is not fully understood. In contrast, sexual reproduction in S. sclerotiorum and a greater selection for specific MCGs in populations of this fungus com-pared with that in S. rolfsii may account for the smaller number of S. rolfsii MCGs found per field in this study. Thus, in the six sugar beet fields in southern Spain comprising more than one S. rolfsii MCG, the MCGs appeared scattered within distinct disease foci with no discernible spatial pattern. This agrees with reports on S. rolfsii in peanut fields in Texas (Nalim et al., 1995) or S. sclerotiorum on soyabean in Illinois (Kull et al., 2004). The results demonstrate that high pathogenic variability occurs among isolates representative of S. rolfsii MCGs as well as in the degree of susceptibility among the tested plant species. To our knowledge this is the first genetic and virulence characterization of S. rolfsii populations. Overall, MCGs vi and ii were highly virulent, MCGs i, iii, ix and x moderately virulent, and MCGs iv, v, vii, viii, xi and xii were mildly virulent. However, that range in virulence was not directly related to the pathogenic spectrum of a given MCG. The results also showed a low level of pathogenic variability among S. rolfsii isolates within a MCG. In fact, for five of the most abundant MCGs identified in this study, isolates within a MCG were similarly virulent to three plant species. Only one out of 12 MCG i isolates tested caused a disease reaction that differed significantly from that induced by the MCG i type isolate, Sr078, being more virulent to all three plant species, sunflower, tomato and watermelon (Fig. 7). These results suggest that a correlation may exist between MCG typing and virulence trait. Among plant species, chickpea and sunflower were highly susceptible; cotton, pepper, tomato and watermelon moderately susceptible; broccoli and sugar beet mildly susceptible; and corn and wheat were resistant. A degree of variability in host plant reaction to a S. rolfsii isolate from sugar beet in Pakistan was reported by Yaqub & Shahzad (2005): mungbean, sugar beet and sunflower were highly susceptible; cabbage, lentil, tomato and sweet pumpkin were mildly susceptible; and cauliflower was completely resistant. To some extent, these results agree with those of Flores- Moctezuma et al. (2006) who evaluated the reactions of 12 plant species to a set of 20 non-MCG-typed S. rolfsii isolates collected from different agricultural and ecological areas in Mexico. These authors found that gooseberry, lentil, pumpkin, radish and tomato were susceptible to all tested isolates, whereas the degree of susceptibility of the remaining seven species varied widely with the isolate. In the present study, sugar beet cv. Markus was susceptible, moderately susceptible and resistant to isolates from one, eight and three MCGs, respectively. Similarly, Sharma et al. (1991) found that only one of five S. rolfsii isolates from different sugar beet-growing areas in northwest India was highly virulent to sugar beet in artificial inoculations. The results of this study have important implications for the management of sclerotium root rot of sugar beet in Mediterranean-type climates, particularly in relation to crop rotation and choice of appropriate crops for use in infested soils previously cropped to sugar beet. Crop rotations are essential to maintain crop productivity and reduce the build-up of soilborne plant pathogens and diseases, which can devastate crops grown in multiple consecutive years (Cook, 2000). Generally, the degree of control is based on the level of susceptibility or resistance of crops involved and the sequences of cropping. In the present study, monocotyledonous species (wheat and corn) were resistant to all MCGs tested and thus are suitable for rotation with susceptible crops. These results agree with those of Boyle (1967), who indicated a reduction in the incidence of sclerotium root rot in peanut fields following a monocotyledonous crop, as well as with those of Minton et al. (1991), who reported that disease incidence in a wheat-peanut rotation was lower than that in a fallow-peanut rotation. Similarly, disease incidence was the lowest in peanut following 2 years of bahiagrass, intermediate following 2 years of corn or cotton, and highest in continuous peanut (Johnson et al., 1999). Nevertheless, results of this present study suggest that S. rolfsii isolates in any of the identified MCGs were able to cause different levels of plant mortality in at least six of the nine susceptible plant species tested. This, together with the occurrence of up to three different MCGs within a field plot or locality differing in pathogenic and virulence profiles, make it difficult to recommend crop rotation as a sole control measure. It would be of interest to evaluate the risk of occurrence of severe epidemics on a given crop based on the prevalence of a given MCG of the pathogen. This would be of importance for southern Spain and Portugal, where vegetable crops such as carrots, peppers, tomato, etc. are becoming alternative crops for sugar beet as a result of the sugar beet production area being drastically reduced or nearly eliminated, respectively. Consequently, sclerotium root rot diseases can become an important threat for vegetable crop production in these areas, as was recently reported for pepper fields in southern Spain (Remesal et al., 2010). Studies are in progress to assess the molecular diversity existing within S. rolfsii MCGs and to determine any genetic relationships that might exist among them.	2018-11-28T04:41:56.829Z	2011-10-27T00:00:00.000	{ "year": 2011, "sha1": "194ac20c66ee168d5c6f767de66c32679811dc5c", "oa_license": "implied-oa", "oa_url": "https://europepmc.org/articles/pmc7192328?pdf=render", "oa_status": "GREEN", "pdf_src": "PubMedCentral", "pdf_hash": "194ac20c66ee168d5c6f767de66c32679811dc5c", "s2fieldsofstudy": [ "Environmental Science", "Agricultural and Food Sciences" ], "extfieldsofstudy": [ "Biology", "Medicine" ] }
16478376	pes2o/s2orc	v3-fos-license	NEDD9 Is a Positive Regulator of Epithelial-Mesenchymal Transition and Promotes Invasion in Aggressive Breast Cancer Epithelial to mesenchymal transition (EMT) plays an important role in many biological processes. The latest studies revealed that aggressive breast cancer, especially the triple-negative breast cancer (TNBC) subtype was frequently associated with apparent EMT, but the mechanisms are still unclear. NEDD9/HEF1/Cas-L is a member of the Cas protein family and was identified as a metastasis marker in multiple cancer types. In this study, we wished to discern the role of NEDD9 in breast cancer progression and to investigate the molecular mechanism by which NEDD9 regulates EMT and promotes invasion in triple-negative breast cancer. We showed that expression of NEDD9 was frequently upregulated in TNBC cell lines, and in aggressive breast tumors, especially in TNBC subtype. Knockdown of endogenous NEDD9 reduced the migration, invasion and proliferation of TNBC cells. Moreover, ectopic overexpression of NEDD9 in mammary epithelial cells led to a string of events including the trigger of EMT, activation of ERK signaling, increase of several EMT-inducing transcription factors and promotion of their interactions with the E-cadherin promoter. Data presented in this report contribute to the understanding of the mechanisms by which NEDD9 promotes EMT, and provide useful clues to the evaluation of the potential of NEDD9 as a responsive molecular target for TNBC chemotherapy. Introduction Breast cancer is a heterogeneous disease, classifiable into five major biologically subtypes, i.e., luminal A, luminal B, basal, ERBB2-overexpressing and normal-like [1,2,3]. Importantly, this molecular taxonomy has significant clinical value because some of the molecular phenotypes (especially Her2 and basal-like) show aggressiveness and unfavorable prognosis [2,3]. The triplenegative breast cancer (TNBC), which accounts for 15%-20% of total breast cancer patients, shares many similarities with the basal subgroup [4,5]. It refers to any breast cancers that do not express the genes for estrogen receptor (ER), progesterone receptor (PR) and Her2/neu. The bulk of data indicate that this subgroup of patients may have a poorer prognosis than those who with hormone receptor-positive or Her2/neu-positive genotypes [6,7]. Therapy of TNBC has been a challenge to the physicians because it is resistant to many effective therapeutic approaches. So far, much of the research interest has been focused on identification of new biomarkers of TNBC, but the understanding of its molecular events is still limited. Results from only a few studies suggested that FOXC2, ID1 and LSD1 were involved in the metastasis of TNBC [8,9,10]. Weinberg and colleagues found that expression of FOXC2 was induced in cells undergoing epithelial-mesenchymal transition (EMT), and FOXC2 was correlated with the highly aggressive basal-like subtype of human breast cancers [10]. Moreover, LBX1, an EMT inductor, was shown to be upregulated in the triple-negative basal-like subtype [11]. These studies implicated that EMT played a critical role in the invasion and metastasis of TNBC. A number of studies suggest that carcinoma cells often activate a trans-differentiation program termed the epithelial-mesenchymal transition (EMT) to acquire the ability to execute the multiple steps of the invasion-metastasis cascade [12,13]. During an EMT, epithelial cells lose cell-cell contacts and cell polarity, express the mesenchymal markers, and undergo major changes in the cytoskeleton that enables cells to acquire a mesenchymal appearance with increased motility and invasiveness [14,15,16]. EMT process can be induced by several crucial signaling pathways including the TGF-b [17], Wnt [18] and Notch [19]. Certain developmental factors, such as Snail, Slug, ZEB1 and FOXC2, were also demonstrated to regulate EMT [20]. In recent years, NEDD9 has been confirmed to contribute to the development of several cancer types [21,22]. Recent studies showed that Nedd9-null genetic background significantly limited mammary tumor initiation in the MMTV-polyoma virus middle T genetic model, suggesting that NEDD9 expression played an important role in breast cancer [23]. Despite these diverse reports, the precise functions and the mechanistic action of NEDD9 have not been well defined. In this study, we demonstrate that NEDD9 is a potent activator of EMT. In mammary epithelial cells, NEDD9 can activate ERK signaling, increase the expression of EMTinducing transcription factors Snail/Slug and their interactions in vivo with the E-cadherin promoter. Ectopic overexpression of NEDD9 led to morphological transformation, induced characteristic molecular features of EMT and enhanced cellular migration, invasion and proliferation. Analysis of NEDD9 expression across different cancers revealed an apparent correlation of this gene with the aggressive human triple-negative breast cancer. Ethics statement Written informed consent was obtained from all participants involved. We obtained ethics approval from the ethics committees at The Tumor Hospital of Jilin Province and The Bethune Hospital of Jilin University. Tissue specimens and cell culture Breast carcinoma tissues were obtained from the Tumor Hospital of Jilin Province and the Bethune Hospital of Jilin University. Samples were frozen in liquid nitrogen immediately after surgical removal and maintained at 280uC until use. All human tissues were collected using the protocols approved by the Ethics Committee of the Jilin Tumor Hospital. The normal human breast epithelia cell lines and the human breast cancer cell lines were obtained from the Institute of Cell Biology, Shanghai, China. Plasmid constructs and transfection The E-cadherin promoter plasmid was a gift from Dr. Ji-Hshiung Chen (Graduate Institute of Molecular and Cell Biology, Tzu Chi University, Taiwan). The LZRS-Ires-Nedd9 plasmid was generously provided by Dr. Lynda Chin (Department of Dermatology, Harvard Medical School, Boston). Nedd9 cDNA was cloned using the following primers: 59-CCGCTCGA-GATGTGGACAAGGAATCTTATGGC-39 (sense) and 59-CC-GGAATTCAGAACGTTGCCATCTCCAGCAAAGA-39 (antisense). The resultant DNA fragment was inserted to pEGFP-N1 vector at XhoI and EcoRI sites. Short interfering RNA (siRNA) targeting the Nedd9 sequence (GAAGCTCTATCAAGTGCCA) was synthesized. Oligonucleotide that represents the siRNA was cloned into the pSuper-neo vector (Oligoengine) between EcoRI and HindIII sites following the manufacturer's instructions. NEDD9-GFP and pEGFP-N1 were transfected to the MDCK and MCF10A cells using FuGENE HD (Roche) following the manufacture's instructions. Cells were selected with G418 for more than two weeks to establish NEDD9-MDCK, EGFP-MDCK, NEDD9-MCF10A, EGFP-MCF10A. NEDD9 siRNA and control siRNA were transfected to the MDA-MB-231 cells using Amaxa nucleofector kits (Lonza). Cells were selected with G418 for more than two weeks to establish NEDD9 siRNA-MDA-MB-231, control siRNA-MDA-MB-231. Wound-healing assay Cells (1610 6 cells per well) were seeded on 6-well plates. 24 hr later, cell layers were wounded in serum-free medium with 1% bovine serum albumin using a sterile 200 ml pipette tip. After washing away the suspended cells, cells were subjected to serum starvation. The progress of migration was photographed in six regions, immediately and during 2 days after wounding (0/12/24/ 48 hr), under an inverted microscope. RNA extraction, reverse transcription and quantitative RT-PCR Total RNA was isolated using the Trizol reagent (Invitrogen) following manufacturer's instructions. One microgram RNA was used for cDNA synthesis using a reverse transcriptase reaction kit (Promega). Quantitative real-time RT-PCR was carried out on an ABI Prism 7000 Sequence Detection System (Applied Biosystems), and SYBR Green (TOYOBO) was used as a double-stranded DNA-specific fluorescent dye. The PCR primer sequences were mentioned in the Methods S1. Immunofluorescence Cells were grown on glass cover-slips in a six-well plate and washed three times with PBS then fixed in 4% formaldehyde solution and permeabilized with 0.1% Triton X-100 in PBS for 5 min. Cells were blocked with 2% BSA in PBS for 30 min at room temperature. Cover-slips were incubated with respective primary antibodies at 1:100 dilutions for 1 hr and then washed with PBS and incubated for 1 hr with TRITC-conjugated secondary antibodies at 1:50 dilutions (Zhongshan, China). Cells were further washed in PBS and mounted with Vectashield mounting medium containing 49, 6-diamidino-2-phenylindole (DAPI; Vector Laboratories) and were analyzed using fluorescence microscopy. Photographs were taken under a Nikon microscope with a fluorescein isothiocyanate filter. Cell migration and invasion assay In vitro cell migration assays were performed as described previously [25]. Images of three random 610 fields were captured from each membrane and the number of migratory cells was counted. The means of triplicate assays for each experimental condition were used. Similar inserts coated with Matrigel were used to determine invasive potential in the invasion assay. MTT assay Cell proliferation was assessed by using the MTT [3-(4, 5dimethylthiazol-2-yl)-2, 5-diphenyl tetrazolium bromide] assay. Cells were plated at 1610 5 cells/well on 96-well plates. At 24, 48 hr after transfection, 20 ml of MTT (5 mg/ml) was added to each well; the samples were incubated for 4 hr at 37uC and then sub-cultured to the medium with 100 ml dimethyl sulfoxide (DMSO). The absorbance of each well was determined at 492 nm. Survival percentage (%) was calculated relative to the control. Colony formation assay Cells were plated in 10-cm tissue culture plates 24 hr before transfection. pEGFP-N1 control vector or NEDD9-GFP expression vector was transfected. 24 hr later, the transfected cells were diluted, re-plated, and selected in 10-cm plates containing 1 mg/L G418 for 12 days. Colonies were stained with crystal violet (Sigma-Aldrich). Luciferase reporter assay Reporter gene assays were done as previously described [24]. Briefly, 5610 4 cells were seeded in 24-well tissue culture plates 24 hr before transfection. The E-cadherin promoter luciferase reporter was transfected at 100 ng/well and the Renilla luciferase control plasmid pREP7-RLuc was cotransfected at 50 ng/well as an internal control reporter. For reporter assays in HEK293T cells, b-catenin was used to activate the reporter gene. Increasing amounts of NEDD9-GFP expression vector were transfected into cells. Thirty hours post transfection, cells were washed and lysed in passive lysis buffer (Promega) and the transfection efficiency was normalized to the paired Renilla luciferase activity by using the Dual Luciferase Reporter Assay System (Promega) according to the manufacture's instructions. Chromatin immunoprecipitation assay The protocol for chromatin immunoprecipitation (ChIP) was described elsewhere [24]. Briefly, the chromatin solution was precleared with 50 ml of protein A-agarose beads (Upstate Biotechnology). The soluble fraction was collected and 5 mg of antibodies was added. The precipitated chromatins were analyzed by PCR. Primer 1 and 2 were used to amplify the E-cadherin promoter regions from 2600 to 2329 and 2359 to 263, respectively. A human negative control was designed. The primers Statistical analysis Student test was used to calculate the statistical significance of the experimental results. The significance level was set as P,0.05 and *P,0.01. Error bars denote the standard deviations (SDs). NEDD9 was overexpressed in human aggressive breast cancers NEDD9 was expressed in several types of tumors [21,22,26]. This observation prompted us to investigate whether NEDD9 was also overexpressed in breast cancers. We analyzed 20 breast tumor samples, together with their adjacent normal tissues, from breast cancer patients. The results revealed a statistically significant increase in NEDD9 mRNA expression in tumors, compared with their adjacent normal mammary tissues (Fig. 1A). We then employed immunohistochemistry (IHC) to assess the NEDD9 protein expression in the paraffin-embedded mammary tissue sections from 84 breast cancer patients in parallel with the surrounding normal breast epithelia. The results indicated that, while normal mammary epithelial cells displayed none or weak NEDD9 staining (Fig. 1B, a), breast carcinoma cells were positive for NEDD9 staining in cytoplasm and/or in nucleus (Fig. 1B, b and c). Further analysis of the data revealed that NEDD9 expression was associated with several adverse prognostic markers, including estrogen receptor (ER) negativity and high tumor grade (Table S1). Interestingly, high levels of NEDD9 expression were associated with aggressive breast cancers, including ER 2 /PR 2 /Her2 2 subtype of invasive ductal breast cancers and Her2/neu-positive breast cancers (Table 1 and Table S1). As shown in Table 1, 31.82% of the TNBC tumors and 24.00% of Her2 + subtype tumors exhibited high levels of NEDD9 expression, whereas only 11.62% of the common ER + subtype of tumors showed high expression of NEDD9 protein. To date, only a few distinct molecular markers have been identified that are uniquely associated with TNBC [10,27,28,29]. NEDD9 expression therefore may prove to be a useful diagnostic marker for this subtype. Moreover, western blotting analysis of immunoreactive NEDD9 in established mammary epithelial cell lines indicated that the levels of NEDD9 in aggressive breast cancer cell lines were considerably higher than those in MCF10A cells derived from normal mammary epithelial cells (Fig. 1C). Collectively, these data suggest that NEDD9 is dominantly overexpressed in human aggressive breast cancer. NEDD9 was a positive regulator of migration, invasion and proliferation in highly aggressive TNBC cells The above results indicated that the endogenous NEDD9 mRNA level was barely detectable in MCF10A cells, but was expressed in several invasive breast cancer cell lines. This suggests that NEDD9 may also play a role in breast cancer migration and invasion. To validate this, we examined the function of NEDD9 in breast cancer by repressing its expression in two highly aggressive TNBC cell lines, MDA-MB-231 [30] and HCC1937 [31]. To test whether constitutive NEDD9 expression in MDA-MB-231 cells contributes to their oncogenicity, we knocked down the endoge-nous NEDD9 by a specific siRNA. The efficiency of this gene silencing protocol was confirmed by western blotting (Fig. 2A). As shown in Fig. 2B and C, using scratch wound assay, we showed that the NEDD9 siRNA-MDA-MB-231 cells had only completed a half closure at 24 hr compared to the control siRNA-MDA-MB-231 cells. The trans-well migration assays also demonstrated that the control cells migrated approximately 2 times faster than NEDD9 siRNA-MDA-MB-231 cells (Fig. 2D). To further validate the roles of NEDD9 in regulating cell invasion, we performed trans-well Matrigel invasion assay to assess the ability of cells to invade through the Matrigel layer. As shown in Fig. 2E, NEDD9 siRNA-MDA-MB-231 cells invaded much slower than control cells. Moreover, results from MTT assays revealed that stable expression of NEDD9-siRNA inhibited the proliferation of MDA-MB-231 cells (Fig. 2F). Colony formation assays also confirmed that knockdown of NEDD9 expression markedly decreased the number of MDA-MB-231 cell colonies (Fig. 2G, H). In order to rule out the false positive results of migration and invasion caused by proliferation inhibition, we carried out representative assays in the presence of 12 mM mitomycin C (MMC), an alkylating agent which inhibits DNA synthesis. As a result, we found no significant difference between groups treated with and without MMC (Figure. S1). These findings suggested that inhibition of NEDD9 expression reduced the migration, invasion and proliferation of MDA-MB-231 cells. Similar results were obtained with HCC1937 human breast cancer cells when NEDD9 expression was repressed by siRNA ( Figure. S2). We next examined the function of NEDD9 in mammary epithelial cells. We showed that overexpression of NEDD9 in MCF10A cell increased cell migration (Fig. 3B-D) and invasion (Fig. 3E). NEDD9 promoted epithelial-mesenchymal transition (EMT) NEDD9 has been shown to be a target gene of TGF-b cell signaling [32,33], which is an important signaling in epithelialmesenchymal transition (EMT), and we found that NEDD9 overexpression in MCF10A mammary epithelial cells changed morphology of the cells (Fig. 4E). So we tested whether NEDD9, on its own, is sufficient to induce the EMT program in a Maden-Darby canine kidney epithelial cell line (MDCK) which has been a widely used model to study epithelial cell biology [34,35]. First, the NEDD9-MDCK cells were generated and confirmed by immunoblotting (Fig. 4A). As can be seen in Fig. 4B, after ectopic NEDD9 expression, MDCK cells displayed a spindle-like, fibroblastic morphology, one of the main characteristics of EMT. At the molecular level, expression of both epithelial and mesenchymal molecular markers was confirmed by western blotting and immunofluorescence (Fig. 4C). It can be seen that the epithelial markers E-cadherin, Occludin and b-catenin were significantly reduced in NEDD9-MDCK cells. Meanwhile, Ecadherin and Occludin were lost from the cell membranes, as revealed by immunofluorescence. In contrast, the mesenchymal (Fig. 4C). We then examined whether NEDD9 could induce EMT process in MCF10A human mammary epithelial cells. The NEDD9-MCF10A cells were generated and confirmed by immunoblotting (Fig. 4D). Similarly, we found that the morphology of the cells changed from epithelial to mesenchymal-like (Fig. 4E); and overexpression of NEDD9 in MCF10A cells caused the reduction of the epithelial markers and increase of the mesenchymal markers (Fig. 4F). Next, we tested whether suppression of endogenous NEDD9 expression is sufficient to reverse the EMT progression. We showed that after knockdown of NEDD9 in MDA-MB-231 cells, the mesenchymal markers N-cadherin, Vimentin and Fibronectin were reduced, whereas the epithelial markers E-cadherin, Occluding and b-catenin were increased, as revealed by western blotting (Fig. 4G). Similar results were obtained by using immunocytochemistry. Concurrently, the NEDD9 siRNA-MDA-MB-231 cells displayed an egg-shaped, epithelial-like morphology (Fig. 4H), consistent with the increase of epithelial markers and the decrease of mesenchymal molecular markers. These results suggest that suppression of NEDD9 in TNBC cells not only reduced migration and invasion but also partially reversed the EMT process. Finally, we tested whether NEDD9 contributes to EMT in vivo. We assessed the correlation between the level of NEDD9 and that of the mesenchymal markers, such as Vimentin and Fibronectin in 32 aggressive breast tumors. We found that positive expression of NEDD9 was significantly associated with the expression of Vimentin and Fibronectin (Fig. 5). Overall, these results demonstrate that NEDD9 is a regulator of EMT. NEDD9 promoted EMT through the ERK-Snail/Slug signaling To further investigate the molecular events involved in NEDD9-induced EMT, we first tested whether NEDD9 is capable of interfering with E-cadherin promoter activity by using a human E-cadherin proximal regulatory promoter luciferase reporter gene plasmid. Transient expression of NEDD9 in HEK293T cells resulted in strong downregulation of the activities of reporter gene (Fig. 6A). Similar results were obtained with MCF10A cell line (Fig. 6B). Moreover, mRNA (Fig. 6C) and protein (Fig. 4F) expression levels of E-cadherin were also reduced upon NEDD9 overexpression in MCF10A cells. These data indicate that NEDD9 is a potential factor that downregulates E-cadherin expression during EMT. Since NEDD9 is not a transcription factor, we wondered whether Snail, Slug, Twist or ZEB, the transcription factors known to repress E-cadherin in various cell systems [36,37,38,39], act in cooperation with NEDD9. RT-PCR analysis in NEDD9-MCF10A and control cells revealed similar expression levels of ZEB1, ZEB2, ID2 and Twist (Fig. 6D). On the other hand, evident increase in Snail and Slug expression was noted in NEDD9 overexpressing MCF10A cells in contrast to control cells, and western analysis confirmed the increased Snail and Slug protein levels (Fig. 6D). Moreover, ChIP assays revealed that overexpression of NEDD9 enhanced the interaction between the Snail family and E-cadherin promoter (Fig. 6E). It has been reported that activation of the extracellular-signal regulated kinase (ERK) can induce Snail and Slug expression [40,41]. We then tested if this signaling pathway functions in the cell model we used. We found that treatment of NEDD9-MCF10A cells with 20 mM the specific ERK kinase inhibitor U0126 indeed caused a recovery of E-cadherin expression (Fig. 6F). We further explored the possible involvement of ERK signaling cascade in the NEDD9-mediated regulation of Snail and Slug expression. Western analysis revealed a significant increase of activated ERK (p-ERK) levels in NEDD9-MCF10A cells compared with control cells (Fig. 6G), suggesting that ERK acted downstream of NEDD9. Together, these data suggest that NEDD9 is able to activate the ERK cascade and subsequently to induce Snail and Slug upregulation that bound to E-cadherin promoter to inhibit its expression. This finally contributes to the EMT and invasive phenotypes of human breast adenocarcinoma cells. Discussion A noticeable point arising from this study is that high levels of NEDD9 expression were associated with the aggressive breast cancers, including TNBC and ERBB2-overexpressing subtypes (Fig. 1, Table 1; Table S1). As shown in Table 1, 31.82% of the triple-negative tumors and 24.00% of Her2 + tumors exhibited high levels of NEDD9 expression, whereas only 11.62% of the luminal ER + subtype of tumors showed high expression of NEDD9 protein. High expression of NEDD9 may therefore be a potential diagnostic marker for these subtypes of breast cancer. Our data are compatible with the pro-oncogenic role identified for NEDD9 overexpression in glioblastoma, melanoma, and lung cancers [21,22,42]. These studies implicated that NEDD9 may have a promotive effect for the carcinogenic process. The process of EMT has been shown to provide carcinoma cells with many of the phenotypes required to execute multiple steps of the invasion-metastasis cascade. When examining the NEDD9 and mesenchymal markers, such as Vimentin and Fibronectin in 32 aggressive breast tumors, we found that positive expression of NEDD9 was associated significantly with the expression of Vimentin and Fibronectin (Fig. 5), implicating that NEDD9 may play a role in EMT in vivo in aggressive breast tumors. Moreover, we discovered that ectopic expression of NEDD9 promoted migration and invasion in MCF10A cells, and knockdown of NEDD9 in highly aggressive TNBC cells reduced their migration, invasion and proliferation (Fig. 2, 3). These data suggested that NEDD9 was a regulator of the migration, invasion and proliferation in breast cancer cells. In contrast to our findings, a previous study suggested that NEDD9 acted as an inhibitor of migration, as the authors reported that siRNAmediated NEDD9 depletion promoted cell migration in breast epithelial cells [43]. In line with our results, Fashena et al found that NEDD9 production induced crescent morphology and cell spreading in MCF7 cell lines [44], and Izumchenko et al reported that the Nedd9 2/2 mice significantly limited the mammary tumor initiation in the MMTV-polyoma virus middle T genetic model [23]. These data support our finding that the NEDD9 is positively correlated with breast cancer progression. However, the authors detected no significant differences when they examined a number of hallmarks of TGF-b-induced EMT in Nedd9 2/2 tumors and tumor-derived cell lines [23]. These discrepancies may probably be due to the fact that NEDD9 is required at early stages in the EMT process, but downregulated after the metastatic cancers undergo a reverse mesenchymalepithelial transition (MET) process. Further investigation to address this issue is required. Our Boyden-chamber trans-well assays determined that the ectopic expression of NEDD9 in MCF10A cells turned this epithelial cell lineage into a highly migratory (Fig. 3D) and invasive (Fig. 3E) phenotypes. Cancer cells degrade the nearby extracellular matrix by using secreted matrix metalloproteinases (MMPs) [45]. As gauged by gelatin zymogram assay and western blotting, we found that MMP9 ( Figure. S3), but not MMP2 (data not shown), was upregulated in response to the expression of NEDD9 in MCF10A cells. These observations reinforced the notion that NEDD9 can serve as an organizer of mesenchymal differentiation during an EMT. Significantly, our data suggest that NEDD9 expression activated the ERK cascade and subsequently induced Snail and Slug upregulation resulted in a potentiated binding of Snail and Slug to E-cadherin promoter (Fig. 6E). These events eventually contributed to the initiation of EMT and invasion in human breast adenocarcinoma cells. Consistent with our results, Storci et al reported that the tumor tissues expressing high levels of Slug mRNA displayed a basal-like breast carcinoma phenotype [46]. Shin et al described that ERK2 specifically regulated EMT in MCF10A cells [47]. ERKs are effectors of MAPK cascade activated by Ras/Raf. Meanwhile, Kim et al demonstrated that NEDD9-dependent tumor promotion was partly dependent on Ras/Raf pathway activation [22]. These studies suggested a close relationship between NEDD9 and Ras signaling in tumor growth. Moreover, our ChIP assays revealed that overexpression of NEDD9 enhanced the recruitment of histone deacetylase HDAC1/HDAC2 repressor complex to E-cadherin promoter ( Figure. S4), indicating that the regulation of E-cadherin repression by overexpression of NEDD9 may involve several epigenetic repressors, although this assumption requires further experiments to confirm. To summarize, we validated in this study that the expression of NEDD9 was frequently upregulated in highly aggressive TNBC breast cancer cell lines as well as in aggressive breast tumors, including ERBB2-positive and triple-negative subtypes. In vitro, knockdown of NEDD9 reduced the mesenchymal molecular markers, increased the epithelial markers and inhibited the invasion and migration of aggressive TNBC cells. Ectopic overexpression of NEDD9 in MCF10A mammary epithelial cells led to a morphological transformation towards the mesenchymal phenotype, together with the expression of mesenchymal markers, and consequently resulted in an enhanced cell migration, invasion and proliferation. Moreover, ectopic expression of NEDD9 activated ERK signaling, upregulated the expression of the EMT-inducing transcription factors Snail and Slug, and promoted their interactions in vivo with the E-cadherin promoter. Results from this study contribute to the understanding of the mechanisms by which NEDD9 promotes the epithelial-mesenchymal transition. Also, this study provides useful clues to the evaluation of the potentiality of NEDD9 as a responsive molecular target for TNBC therapeutics. the effect of NEDD9 overexpression on MMP-9. Cell lysates were prepared and subjected to western analysis using an anti-MMP-9 polyclonal antibody. b-actin was used as the loading reference. (TIF) Figure S4 ChIP assays at the E-cadherin promoter. Increased binding of HDAC1 and HDAC2 at the E-cadherin promoter in the presence of NEDD9. Primer 1 and 2 were used to amplify the E-cadherin promoter regions from 2600 to 2329 and 2359 to 263, respectively. (TIF)	2014-10-01T00:00:00.000Z	2011-07-28T00:00:00.000	{ "year": 2011, "sha1": "dec621b7a3969123fb644d8469a1137b8acfd63c", "oa_license": "CCBY", "oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0022666&type=printable", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "dec621b7a3969123fb644d8469a1137b8acfd63c", "s2fieldsofstudy": [ "Biology", "Medicine" ], "extfieldsofstudy": [ "Biology", "Medicine" ] }
139611989	pes2o/s2orc	v3-fos-license	Comparative Study of Radiation Shielding Parameters for Binary Oxide Glasses Melt and quench technique was used for the preparation of glassy samples of the composition 0.40 MO-0.60 B2O3 where MO contains oxides of lithium, sodium, potassium and bismuth. XCOM computer program is used for the evaluation of gamma-ray shielding parameters of the prepared glass samples. Further the values of mass attenuation coefficients and effective atomic number at various photon energies is calculated. Values of half value layers have been compared with concretes. Density and molar volume studies have been carried out to get the idea regarding rigidity of glass network. INTRODUCTION The study of interaction of nuclear radiations with matter is the important research area for the development of materials which can be used in high radiation environment.These radiation shielding materials have great importance for many scientific, engineering and medical applications.The data based on mass attenuation coefficient and half value layer is very useful for the purpose to identify the various radiation shielding materials.In the environment of high radiation exposure concrete is used as radiation shielding material because it is cheap and it can be molded easily into any desired design.In spite of all these advantages some limitation are also associated with the concretes like it is not transparent to visible light thus restricting one to see through it.Secondly when it is exposed to the radiations for a longer period of time its mechanical strength is reduced.So it is desired to have materials which are transparent to visible light and have better shielding properties in terms of lesser volume requirement.Now a days heavy metal glasses are proving to be the promising candidates as an alternate to the conventional shielding material like concrete [1][2][3][4][5] .These have considerable technological applications due to their density, high refractive index and low melting point [6][7] . Sample preparation, X-ray and density measurements Glass samples of the chemical compositions given in table1 have been prepared by using the melt quenching technique All the chemicals used in the sample preparation were of analytical reagent (AR) grade having percentage purity of 99.9%.Appropriate amounts of Li 2 CO 3 , Na 2 CO 3 , K 2 CO 3 , Bi 2 O 3 , and H 3 BO 3 were mixed thoroughly.Mixture was then taken in porcelain crucible and was melted and the melt was kept at the same temperature for 2 hours.Melt was quickly poured into preheated moulds in the temperature region of 200 to 250 o C followed by cooling down to room temperature. Densities of the samples at room temperature were measured by using Archmede's principle with Benzene as immersion liquid.Glass samples were weighed in air and when immersed in Benzene at 20 0 C. The density was calculated by using the formula. Where ρ is the density of glass sample, a is weight of sample in air, b is weight of glass sample in Benzene and 0.787 is the density of Benzene at 20 0 C. Chemical composition along with density and molar volume of the prepared glass samples is given in table 1. To confirm the amorphous nature glass samples XRD measurements were carried out on the prepared samples.Samples were crushed to fine powder form for X-ray diffraction studies.A Philips PW 1710 diffractrometer was used during the investigations.Radiation used was CuKa.The pattern was recorded at a scanning rate (2θ/s) of 0.030 with start angle (2θ) at 5.010 0 and end angle (2θ) at 60.0 0 .Absence of crystallization peak in XRD data shows that prepared samples are amorphous. Linear attenuation coefficients and effective atomic number The mass attenuation coefficient of prepared samples is evaluated by using WIN XCOM [3][4] developed by NIST.Several authors and researchers have experimentally measured mass attenuation coefficients of different materials including silicate and borate glasses.Excellent agreement has been observed between the experimental and theoretical data obtained by XCOM software.In light of this it is possible to estimate mass attenuation coefficient for our glass system with reasonable accuracy by XCOM software [8][9][10] .p= a/(a-b)x 0.787 (1) HVL have been determined by mass attenuation coefficient by using the relation. Effective atomic number can be calculated by using relation. where the average atomic cross-section, σ t,a is evaluated by using equation. σ t,m = (μ/p)M/N A ,Where M=n j A j is the mole mass,N A is Avogadro's constant and n j and A j are the number of formula units and atomic weight, respectively of constituent elements. Average electronic cross-section, σ t,el is given by where f i = n i / Σn j is fractional abundance of element i with respect to number of atoms and Z i is atomic number of i element. Density and molar volume Density and molar volume values of the prepared samples are shown in table1.The values of density increase from GS1 to GS4 showing the effect of changing a light metal atom with the heavier metal atom.The values of molar volume also show linear increase depicting that structure becomes more and more open as light metal atom is changed with heavier one. 11The value of molar volume is maximum for the composition corresponding to Bi.It can be due to the fact that Bi ions play dual role in the glass structure.At higher mole fraction the role of Bi ions in the glass structure is of glass former. Linear attenuation coefficients The values of mass attenuation coefficients for the various samples are shown in table 2 from energy range 1KeV to 100 GeV.It can be observed that the mass attenuation coefficient values are very high in low energy range for which photoelectric effect is dominant.The values of mass attenuation coefficients as evident from table 2 shows rapid decrease attaining minimum value in the intermediate energy range. Effective atomic numbers and half value layer The scattering and absorption of gamma radiations are related to the effective atomic number of the materials.These values are calculated using equation 3 and variation for the prepared samples is shown in Fig. 1.It can be seen from the figure that for all the samples the Z eff steadily increases up to 10 -2 MeV, then it steadily decreases up to 10 0 MeV after which it increases up to 10 2 MeV after this energy the values of Z eff almost remains constant.The variation of Z eff , value with energy for the samples materials may can be assigned due to the relative domination of different mechanisms mainly photoelectric effect, coherent scattering and incoherent scattering.At low energies, where photoelectric effect dominates, Z eff is more and at higher energies, where scattering dominates, Z eff is less 14 .Therefore, the Z eff for total gamma ray interaction varies from a higher value at lower energies to a lower value at higher energies depending on the relative domination of the partial gamma ray interaction processes.Effective atomic numbers increases linearly as we change the light metal atom with heavier metal atom.Value of effective number is maximum for the sample containing Bi atom.i.e. sample GS4 .Half value layer parameter is calculated from the linear attenuation coefficient using equation 2. Values of HVL are given in table 3. It can be seen that HVL decreases with replacement of light metal with the heavy metal [14][15][16] .HVL is not only composition dependent but it also dependent upon the density of material which is further related to the structural arrangements of constituents of composite material.This parameter is very important for the materials to be used as the radiation shield.For this purpose, HVL values of the prepared glasses have been compared with some standard radiation shielding concretes (Table 4).For this glass sample GS4 is taken for the comparison purpose since its HVL value is minimum among all the prepared glass samples.Fig. 2 shows comparative plot of the HVL value for various concretes and GS4 sample.It can be clearly seen that for the prepared sample (GS4), value of HVL at all the energies is lesser than the concretes. CONCLUSIONS Bi containing grasses are found to have lowest HVL values in all the prepared samples.It is concluded that the Bi containing glasses can be used as an alternate to the radiation shielding materials.Further these are transparent to visible light and require lesser volume as compared to concretes which are opaque and require much larger volume. Fig. 1 . Fig. 1.Plot of Effective atomic number for the prepared glass samples as a function of photon energy.	2018-03-28T05:38:04.130Z	2017-10-25T00:00:00.000	{ "year": 2017, "sha1": "892c9cc61cd9c3f8d5df95154f283c1ce6ab3b68", "oa_license": "CCBY", "oa_url": "http://www.orientjchem.org/pdf/vol33no5/OJC_Vol33_No5_p_%202324-2328.pdf", "oa_status": "HYBRID", "pdf_src": "Anansi", "pdf_hash": "892c9cc61cd9c3f8d5df95154f283c1ce6ab3b68", "s2fieldsofstudy": [ "Materials Science" ], "extfieldsofstudy": [ "Materials Science" ] }
203626820	pes2o/s2orc	v3-fos-license	Kinematics of Hot Subdwarfs from the Gaia DR2 Catalogue We have studied the kinematic properties of the candidates for hot subdwarfs (HSDs) selected by Geier et al. from the Gaia DR2 catalogue. We have used a total of 12 515 stars with relative trigonometric parallax errors less than 30\%. The HSDs are shown to have different kinematics, depending on their positions on the celestial sphere. For example, the sample of low-latitude $(\|b\|<20^\circ)$ HSDs rotates around the Galactic center with a linear velocity $V_0=221\pm5$ km s$^{-1}$. This suggests that they belong to the Galactic thin disk. At the same time, they lag behind the local standard of rest by $\Delta V_\odot\sim16$ km s$^{-1}$ due to the asymmetric drift. The high-latitude ($\|b\|\geq20^\circ$) HSDs rotate with a considerably lower velocity, $V_0=168\pm6$ km s$^{-1}$. Their lagging behind the local standard of rest is already $\Delta V_\odot\sim40$ km s$^{-1}$. Based on the entire sample of 12 515 HSDs, we have found a positive rotation around the $x$ axis significantly differing from zero with an angular velocity $\omega_1=1.36\pm0.24$ km s$^{-1}$ kpc$^{-1}$. We have studied the samples of HSDs that are complete within $r<1.5$ kpc. Based on them, we have traced the evolution of the parameters of the residual velocity ellipsoid as a function of both latitude $\|b\|$ and coordinate $\|z\|.$ The following vertical disk scale heights have been found: $h=180\pm6$ and $290\pm10$ pc from the low- and high-latitude HSDs, respectively. A new estimate of the local stellar density $\Sigma_{out}=53\pm4~M_{\odot}$/kpc$^2$ kpc$^{-2}$ has been obtained for $z_{out}=0.56$ kpc from the high-latitude HSDs. INTRODUCTION The hot subdwarfs (sdO, sdB, sdOB) are located between the main sequence and the white dwarf sequence on the Hertzsprung-Russell diagram. Their peculiarity is core helium burning. Such stars were first detected by Humason and Zwicky (1947) while studying superbright blue stars near the North Galactic Pole. Further spectroscopic studies revealed a hydrogen underabundance in many hot subdwarfs. Their temperature and surface gravity measurements (Greenstein and Sargent 1974) allowed the proper positions of such stars on the Hertzsprung-Russell diagram to be determined. More specifically, they occupy a compact region at the blue end of the horizontal branch, thereby being a late evolutionary stage of massive stars. The formation of hot subdwarfs (HSDs) is explained, for example, by the loss of hydrogen by a red giant from its outer layers before the reactions involving helium begin in the core. The cause of this loss is not completely clear, but it is hypothesized that this is the result of interaction between the stars in a binary system. It is assumed that single HSDs can result from a merger of two white dwarfs. At present, there is extensive literature on the investigation of single HSDs (Randall et al. 2016; Latour et al. 2018) and binary systems with HSDs ; Vos et al. 2018). Their study can help in understanding the evolutionary peculiarities of other galaxies, the presence of an ultraviolet excess, in particular (Han et al. 2007), galactic globular clusters (Lei et al. 2016), the stellar evolution (Fontaine et al. 2012), and can help in searching for the progenitors of type Ia supernovae (Geier et al. 2007). Only with the appearance of highly accurate astrometric data from the Gaia space experiment (Brown et al. 2016(Brown et al. , 2018Lindegren et al. 2018) has a largescale statistical analysis of various Galactic subsystems, in particular, a statistical and kinematic analysis of HSDs, become possible. The methods of searching for and identifying of such stars in largescale surveys are being developed (Bu et al. 2017); the first big catalogues of HSDs containing the necessary data on tens of thousands of stars have appeared (Geier et al. 2019). Samples of protoplanetary nebulae (PPNe) and planetary nebulae (PNe) were studied in Bobylev and Bajkova (2017a) and Bobylev and Bajkova (2017b), respectively. Based on three laws of the density distribution, we determined the vertical disk scale height and estimated the Galactic rotation parameters. Many HSDs are at a more advanced evolutionary stage compared to PPNe and PNe, which is important for studying the various stages of stellar evolution. The goal of this paper is to analyze the kinematics and spatial distribution of a large sample of HSD candidates. Such a sample of 39 800 stars has recently been produced by Geier et al. (2019) using data from the Gaia DR2 catalogue (Brown et al. 2018;Lindegren et al. 2018) and a number of photometric sky surveys. DATA In this paper we use the catalogue by Geier et al. (2019). It contains 39 800 HSD candidates selected from the Gaia DR2 catalogue in combination with several ground-based multiband photometric sky surveys. HSDs of spectral types O and B, late-B stars in the blue part of the horizontal branch, hot post-AGB stars, and PN central stars are expected to constitute the majority of candidates. The authors of the catalogue believe that the contamination by cooler stars is about 10%. The selected HSDs are distributed over the entire celestial sphere. According to the estimates by Geier et al. (2019), the catalogue is complete to a distance of 1.5 kpc, except for the narrow region near the Galactic plane and the zones with the Magellanic Clouds. For each star this catalogue gives the trigonometric parallax π and two proper motion components, µ α cos δ and µ δ . There are no radial velocities. Extensive photometric information is also available. Of course, the radial velocities have been measured for several hundred HSDs. Note, for example, the catalogue of radial velocities for such stars in close binary systems by or . In those cases where it is possible to construct the orbit of a binary or multiple system, the systemic velocity V γ is usually determined with an error of 1-5 km s −1 . However, for such a large number as 39 800 HSDs no highly accurate radial velocities have been measured so far. In this paper, to determine the kinematic parameters, we took stars with relative trigonometric parallax errors less than 30%. In addition, the zero-point correction ∆π was added to the Gaia DR2 parallaxes; thus, the new stellar parallax is π + 0.050 mas. The necessity of applying this systematic correction ∆π = −0.029 mas (here the minus means that the correction should be added to the Gaia DR2 stellar parallaxes to reduce them to the standard) was first pointed out by Lindegren et al. (2018) and confirmed by Arenou et al. (2018). Slightly later, having analyzed various highly accurate sources of distance scales, Stassun METHODS In this paper we consider stars only with the proper motions, because there is no information about the radial velocities in the catalogue by Geier et al. (2019). Thus, two tangential velocity components are known from observations, V l = 4.74rµ l cos b and V b = 4.74rµ b along the Galactic longitude l and latitude b, respectively, expressed in km s −1 . Here, the coefficient 4.74 is the ratio of the number of kilometers in an astronomical unit to the number of seconds in a tropical year, and r = 1/π is the stellar heliocentric distance in kpc that we calculate via the stellar parallax π. The proper motion components µ l cos b and µ b are expressed in mas yr −1 . To determine the parameters of the Galactic rotation curve, we use the equations derived from Bottlinger's formulas in which the angular velocity Ω is expanded into a series to terms of the second order of smallness in r/R 0 : where R is the distance from the star to the Galactic rotation axis (cylindrical radius vector): Ω 0 is the angular velocity of Galactic rotation at the solar distance R 0 , the parameters Ω ′ 0 and Ω ′′ 0 are the corresponding derivatives of the angular velocity, and V 0 = \|R 0 Ω 0 \|. Apart from the rotation around the Galactic z axis (described by the parameter Ω), in this paper we also consider the angular velocities of rotation around the x and y axes described by the parameters ω 1 and ω 2 , respectively. Note that in this paper the signs of the unknowns are taken in such a way that the positive rotations occur from the x axis to y, from y to z, and from z to x. Thus, the angular velocity of Galactic rotation will be negative, in contrast to many other papers (Rastorguev et The interest in ω 1 and ω 2 stems from the following factors. First, this can be a slight residual rotation of the Gaia DR2 reference frame relative to the inertial reference frame. Second, this can be a reflection of the Galactic warp or other processes causing the vertical oscillations of stars in the Galaxy. A uniform distribution of the stars being analyzed over the celestial sphere is required for a reliable determination of the parameters ω 1 and ω 2 . Dwarfs and subdwarfs are well suited for solving this problem. In the system of conditional equations (1)-(2) eight unknowns are to be determined: U ⊙ , V ⊙ , W ⊙ , Ω 0 , Ω ′ 0 , Ω ′′ 0 , ω 1 and ω 2 . We find their values by solving the conditional equations by the least-squares method (LSM). Weights of the form w l = S 0 / S 2 0 + σ 2 V l and w b = S 0 / S 2 0 + σ 2 V b , are used, where S 0 is the "cosmic" dispersion (for each sample we find its value in advance to be close to the error per unit weight σ 0 , obtained by presolving the equations), σ V l and σ V b are the errors in the corresponding observed velocities. S 0 is comparable to the root-mean-square residual σ 0 (the error per unit weight) calculated by solving the conditions equations (1)-(2). The solution is sought in several iterations by applying the 3σ criterion to exclude the stars with large residuals. Note several papers devoted to determining the mean Galactocentric distance of the Sun using its individual determinations made in the last decade by independent methods. For example, R 0 = 8.0 ± 0.2 kpc (Valleé 2017), R 0 = 8.40 ± 0.4 kpc (de Grijs and Bono 2017) or R 0 = 8.0 ± 0.15 kpc (Camarillo et al. 2018). Based on these reviews, in this paper we adopt R 0 = 8.0 ± 0.15 kpc. Residual Velocity Ellipsoid We estimated the dispersion of the stellar residual velocities using the following well-known method (Ogorodnikov 1965). Let U, V, W be the velocities along the coordinate axes x, y, and z. Let us consider the six second-order moments a, b, c, f, e, d : which are the coefficients of the equation for the surface and also the components of the symmetric tensor of moments of the residual velocities: To determine the values in this tensor in the absence of radial-velocity data, the following three equations are used: which are solved by the least-squares method for the six unknowns a, b, c, f, e, and d. The eigenvalues of the tensor (6) λ 1,2,3 are then found from the solution of the secular equation The eigenvalues of this equation are equal to the reciprocals of the squares of the semiaxes of the velocity moment ellipsoid and, at the same time, the squares of the semiaxes of the residual velocity ellipsoid: We found the directions of the main axes of the tensor (10), L 1,2,3 and B 1,2,3 , from the relations: Exponential Density Distribution In the case of an exponential distribution of the density ρ along the z coordinate axis, we have where ρ 0 is the normalization constant; z ⊙ is the mean value calculated from the z coordinates of the sample stars, which reflects the well-known fact of the Sun's elevation above the Galactic plane; and h is the vertical scale height. The observed histogram of the distribution of stars along the z axis is described by an analogous expression: where N 0 is the normalization coefficient. The surface density of gravitating matter Σ out (z out ) within a distance z out from the Galactic plane z = 0 can be found from the following relation: where the gravitational constant G is taken to be equal to one. Following the approach by Korchagin et al. (2003) based on the solution by Spitzer (1942) for a self-gravitating disk, we have 1 ρ Given z out , the mean vertical velocity v 2 z , the scale height h, and the Oort constants A and B, we can estimate the surface density from the relation RESULTS AND DISCUSSION Galactic Rotation Parameters We have run into the fact that the HSDs under consideration have different kinematic properties, depending on their positions in the Galaxy. Therefore, the entire sample was divided into two groups, depending on the absolute value of the Galactic latitude: low-latitude (\|b\| < 20 • ) and high-latitude (\|b\| ≥ 20 • ) stars. The kinematic parameters found for these stars are given in Table 1. Apart from the eight sought for kinematic parameters, it gives the following: the mean distance of the sample of stars r, the error per unit weight σ 0 that we estimate when seeking the LSM solution of the system of conditional equations (1)-(2), the Oort constants A and B that we calculate based on the following relations: and the linear rotation velocity of the Galaxy at the solar distance V 0 = \|R 0 Ω 0 \|. When calculating the kinematic parameters presented in Table 1, we took the stars from the interval 0.5-6 kpc and rejected the stars with large (more than 60 km s −1 ) random measurement errors of their proper motions. The separation in relative parallax error was made for the following reasons. On the one hand, using the stars with relative trigonometric parallax errors less than 30% allows us to use a large number of objects and to estimate the sought-for parameters with smaller errors. On the other hand, the sample of stars with parallax errors less than 15% allows more local parameters, in particular, more reliable estimates of the velocities (U, V, W ) ⊙ and Ω 0 to be obtained. First note the parameters that describe the kinematics of the most rapidly rotating Galactic subsystems. For example, based on 130 masers with measured VLBI trigonometric parallaxes, Rastorguev The linear velocity V 0 = 229 ± 6 km s −1 inferred from the low-latitude HSDs (the upper part of Table 1) suggests that this population of stars belongs to the thin disk. At the same time, the velocity V ⊙ ∼ 28 km s −1 derived from them shows that they lag behind the local standard of rest by ∼ 16 km s −1 due to the so-called asymmetric drift. One of the reliable present-day determinations of the parameters of the peculiar solar motion N ⋆ is the number of stars used, r is the mean distance of the sample of stars, and σ 0 is the error per unit weight. For all of the older Galactic objects their lagging behind the local standard of rest increases, the velocity V ⊙ rises, due to the asymmetric drift. As can be seen from the last column in the lower part of Table 1, the lagging of the high-latitude HSDs behind the local standard of rest is already more than 40 km s −1 . The first and second derivatives of the angular velocity Ω ′ 0 and Ω ′′ 0 are also determined satisfactorily (in agreement with our analysis of young thin-disk objects) from the lowlatitude HSDs. Nevertheless, the absolute value of the Oort constant A is less than that of the Oort constant B for these stars as well, suggesting that the linear rotation velocity in the neighborhood increases with R. For the youngest Galactic objects, for example, maser sources or OB stars, the reverse situation is a typical one, i.e., the absolute value of the Oort constant A is greater than that of the Oort constant B The second derivative Ω ′′ 0 is not determined from the high-latitude HSDs, while the first derivative Ω ′ 0 differs greatly from its value found from younger objects. Rotation around the x and y Axes From the data in Table 1 we can unambiguously conclude that there is no rotation around the y axis (ω 2 ) differing significantly from zero in our samples. In contrast, there is a rotation around the x axis (ω 1 ) differing significantly from zero in the samples of high-latitude HSDs and the "all stars" sample. The nature of this rotation has not yet been established. First, this can be interpreted as a residual rotation of the Gaia DR2 reference frame relative to the reference frame of extragalactic sources. In this case, ω 1 = 1.36 ± 0.24 km s −1 kpc −1 from the first column in the lower part of Table 1 found from all stars is of value. Taking into account the mean distance of this sample, we obtain ω 1 = 0.15 ± 0.03 mas yr −1 . Second, this can be a manifestation of some large-scale physical process characteristic of the high-latitude HSDs. Note also that the angular velocities ω 1 and ω 2 cannot be determined reliably from the low-latitude objects, because sin b in Eq. (1) is close to zero. Therefore, the parameters ω 1 and ω 2 are actually found only from one Eq. (2). Lindegren et al. (2018) concluded that the Gaia DR2 reference frame has no rotation relative to the reference frame of quasars within 0.15 mas yr −1 (along three axes), with the effect being most pronounced in the region of bright (G < 12 m ) stars. Having analyzed ∼6 million stars from the Gaia DR2 catalogue, Tsvetkov and Amosov (2019) found ω 1 ∼ 0.7 ± 0.11 km s −1 kpc −1 significantly differing from zero, which is stably preserved depending on the mean distance of the sample. When studying ∼900 open star clusters with data from the Gaia DR2 catalogue, Bobylev and Bajkova (2019) detected a rotation of this entire sample around the Galactic x axis with an angular velocity ω 1 = 0.48 ± 0.15 km s −1 kpc −1 . The idea that the kinematics of the observed stars is affected by the large-scale Galactic warp underlies the second approach. Based on the simplest solid-body rotation model, for example, Miyamoto and Zhu (1998) Parameters of the Residual Velocity Ellipsoid To form the moments (4) for each subsample, the corresponding velocities U ⊙ , V ⊙ , and W ⊙ were taken from Table 1. In addition, the stellar velocities were corrected for the Galactic rotation. For this purpose, we use the parameters of the rotation curve derived from the low-latitude HSDs (the upper part of Table 1). Based on our sample of 4181 low-latitude stars with relative trigonometric parallax errors less than 15%, we found the following residual velocity dispersions: and the following orientation parameters of this ellipsoid: Obviously, the orientation of this residual velocity ellipsoid coincides with the x, y, z coordinate axes with a high accuracy. From the low-latitude HSDs, but with relative trigonometric parallax errors less than 30%(7156 stars) we found the following residual velocity dispersions: while the orientation of this ellipsoid is The orientation of this ellipsoid is also close to the directions of the x, y and z coordinate axes; only L 1 is determined with a large error. Note the paper by Rowell and Hambly (2011), where the various properties of the thin and thick disks as well as the halo are discussed in sufficient detail. In particular, for the exponential law (15) they think the scale height h = 250 kpc to be typical for the thin disk. In their Table 4 2018) found the following dispersions: (σ U , σ V , σ W ) = (33, 28, 23) ± (4, 2, 2) km s −1 and (σ U , σ V , σ W ) = (57, 38, 37) ± (6, 5, 4) km s −1 for the thin and thick disks, respectively. They showed that for various thin-disk stellar groupings the vertex deviation (l uv ) in the UV plane changes in a very wide range, from −5 • to +40 • , while the tilt angle (l uw ) in the UW plane varies from −10 • to +15 • . We may conclude that the low-and high-latitude HSDs exhibit the thin-and thick-disk kinematics, respectively. Of course, the presence of halo stars in the samples considered must not be ruled out. However, their influence on the kinematics is statistically negligible. Parameters of the Density Distribution The values of z ⊙ and h for the low-and high-latitude HSDs used below to construct the exponential distribution (15) are given in the upper part of Table 2. The results of the analysis of various Galactic subsystems obtained by other authors are given in the middle part of this table. N ⋆ is the number of stars used, r is the mean distance of the sample of stars. Sample 1 in Bobylev and Bajkova (2017a) consists of relatively young protoplanetary nebulae with luminosities higher than 5000L ⊙ , while sample 2 includes older protoplanetary nebulae with luminosities of 4000L ⊙ or 3500L ⊙ . As can be seen from Table 2, the estimate of h = 190±4 pc obtained from the low-latitude HSDs is close to the value found from planetary nebulae (Bobylev and Bajkova 2017b). In contrast, the estimate of h = 700±8 obtained from the high-latitude HSDs exceeds the value found from sample 2 of protoplanetary nebulae (Bobylev and Bajkova 2017a). Estimating the Surface Density A complete stellar sample is required to estimate the surface density. According to the estimates by Geier et al. (2019), except for the narrow region near the Galactic plane, the catalogue of HSDs is complete to a distance of 1.5 kpc, which we take as the completeness boundary. Thus, in this section we consider the low-and high-latitude HSDs selected under the condition r < 1.5 kpc. The results of their analysis are presented in Tables 2 and 3 as well as in Figs. 3 and 4. As can be seen from Table 2, the scale height h remains almost constant for the lowlatitude HSDs under various sample production conditions. In contrast, h for the highlatitude HSDs h depends strongly on the constraints used. We can see that the moderate h = 290 ± 10 kpc is consistent with the results of the analysis of white dwarfs (Vennes et al. 2002) and planetary nebulae (Bobylev and Bajkova 2017b). As can be seen from Table 3, the third axis in all ellipsoids is close to the direction to the Galactic Pole. Therefore, in Eq. (17) we may set v 2 z = σ 2 3 . This is apparently the only way of estimating the vertical velocity, because in our case we have no information about the radial velocities and, therefore, we cannot directly calculate the space velocities U, V, N ⋆ is the number of stars used, r is the mean distance of the sample of stars. and W. Let us make an estimate for the mean \|z\| given in Table 3. We see that \|z\| = 0.17 kpc for the low-latitude HSDs, which is of no real interest for the estimate of Σ out due to the small z. From the high-latitude HSDs at σ π /π < 30% and z out = \|z\| = 0.56 kpc we find Σ out = 53 ± 4 M ⊙ kpc −2 . For comparison, we can give some values of the surface density obtained by other authors. Korchagin et al. (2003) found Σ out = 46 ± 2 M ⊙ kpc −2 for z out = 0.35 kpc based on a sample of ∼1500 red giants from the Hipparcos catalogue (1997). Based on giants from the Hipparcos catalogue (1997), Holmberg and Flynn (2004) estimated the local density of the visible matter, Σ out = 56 ± 6 M ⊙ kpc −2 , and the entire gravitating matter, including the dark matter disk and halo, Σ out = 74 ± 6 M ⊙ kpc −2 , for z out =1.1 kpc. Based on ∼16 000 G dwarfs from the SEGUE catalogue ( Dependence of Kinematic Parameters on \|z\| In this section we consider the dependence of some kinematic parameters on the coordinate \|z\| based on a complete sample of HSDs. For this purpose, we used a sample of HSDs with errors σ π /π¡30%, which we divided into four nonoverlapping parts, depending on \|z\|. The results are presented in Table 4. Equations (1)-(2) were solved with five unknowns: U ⊙ , V ⊙ , W ⊙ , Ω 0 , and Ω ′ 0 . Here we did not divide the stars into low-and high-latitude ones. We see good agreement between the parameters of the residual velocity ellipsoid inferred from both high-latitude HSDs (Table 3) and HSDs with large \|z\| (Table 4). In particular, CONCLUSIONS We studied the kinematics of hot dim stars from the catalogue by Geier et al. (2019). These are 39 800 HSD candidates selected by them from the Gaia DR2 catalogue using data from several multiband photometric sky surveys. In this paper we analyzed more than 12 500 proper motions of stars with relative trigonometric parallax errors less than 30%. Thus,we investigated a huge set of present-day highly accurate data that is of great value for a statistical analysis. The stars were divided into two parts, depending on the Galactic latitude, into lowlatitude (\|b\| < 20 • ) and high-latitude (\|b\| ≥ 20 • ) samples. Such a breakdown is needed to construct the histograms in a wide z range and to determine the parameters of the exponential density distribution from them. These two samples were shown to have completely different kinematics. We determined the Galactic rotation parameters from relatively more distant objects whose parallax errors do not exceed 30%. In that case, at least the first derivative of the angular velocity of Galactic rotation is determined quite reliably. In contrast, for the subsequent determination of the residual velocity ellipsoid parameters we already use more reliable data, objects with trigonometric parallax errors no greater than 15%. For example, the sample of low-latitude HSDs rotates around the Galactic center with a linear velocity V 0 = 221 ± 5 km s −1 . This suggests that they belong to the Galactic thin disk. They lag behind the local standard of rest by ∆V ⊙ ∼ 16 km s −1 due to the asymmetric drift. The high-latitude HSDs rotate with a considerably lower velocity, V 0 = 168 ± 6 km s −1 , which is more typical for the thick disk. Their lagging behind the local standard of rest is ∆V ⊙ ∼ 40 km s −1 . A joint analysis of the entire sample of 12 515 stars with parallax errors less than 30% revealed a positive rotation around the x axis differing significantly from zero with an angular velocity ω 1 = 1.36±0.24 km s −1 kpc −1 . Expressed in different units, given the mean distance of the sample stars, this angular velocity is ω 1 = 0.15 ± 0.033 mas yr −1 . The latter value is consistent with the estimates by Lindblad et al. (2018) for the residual rotation of the Gaia DR2 reference frame relative to the inertial reference frame. Based on a sample of 3584 high-latitude HSDs with parallax errors less than 15%, we found (σ 1 , σ 2 , σ 3 ) = (51.9, 46.6, 34.8) ± (1.1, 1.8, 0.8) km s −1 , with the third axis of this ellipsoid (24) being deflected from the direction to the Galactic Pole by 15 ± 4 • . Such a deflection is also confirmed by the direction of the second axis B 2 = −15 ± 2 • , with the error having been estimated here along each axis independently. We studied two samples of HSDs that are complete within r < 1.5 kpc. The following vertical disk scale heights were derived: h = 180 ± 6 pc and 290 ± 10 pc from the lowand high-latitude HSDs, respectively. Here, we obtained a new estimate of the local stellar density Σ out = 53 ± 4 M ⊙ kpc −2 for z out = 0.56 kpc from the high-latitude HSDs. Based on a sample that is complete with errors σ π /π¡30%, we traced the evolution of the kinematic parameters as a function of the coordinate \|z\|. We estimated the gradient of the circular rotation velocity V 0 along \|z\|, ∆V 0 ∆\|z\| = −64 ± 5 km s −1 kpc −1 . All of the above results lead us to conclude that the low-and high-latitude HSDs exhibit the kinematics of the thin and thick disks, respectively. This suggests that most of the lowlatitude HSDs "inherited" the kinematic properties of relatively young massive precursor stars. The high-latitude HSDs are less homogeneous kinematically; their properties depend strongly on the positions above the Galactic plane and the heliocentric distance. Kinematically more "heated" objects that have already receded quite far from the Galactic plane are apparently their precursors.	2019-10-01T19:39:13.000Z	2019-09-01T00:00:00.000	{ "year": 2019, "sha1": "427c4a49aff434280fcd206cf891e283624f1b97", "oa_license": null, "oa_url": "http://arxiv.org/pdf/1910.00632", "oa_status": "GREEN", "pdf_src": "Arxiv", "pdf_hash": "427c4a49aff434280fcd206cf891e283624f1b97", "s2fieldsofstudy": [ "Physics" ], "extfieldsofstudy": [ "Physics" ] }
243985749	pes2o/s2orc	v3-fos-license	Gradient-push algorithm for distributed optimization with event-triggered communications Decentralized optimization problems consist of multiple agents connected by a network. The agents have each local cost function, and the goal is to minimize the sum of the functions cooperatively. It requires the agents communicate with each other, and reducing the cost for communication is desired for a communication-limited environment. In this work, we propose a gradient-push algorithm involving event-triggered communication on directed network. Each agent sends its state information to its neighbors only when the difference between the latest sent state and the current state is larger than a threshold. The convergence of the algorithm is established under a decay and a summability condition on a stepsize and a triggering threshold. Numerical experiments are presented to support the effectiveness and the convergence results of the algorithm. I. INTRODUCTION In recent years, distributed optimization techniques over a multi-agent network have attracted considerable attention since they play an essential role in engineering problems in distributed control [1], [2], signal processing [3], [4] and the machine learning problems [5]- [7]. In distributed optimization, many agents have their own local cost and try to find a minimizer of the sum of those local cost functions in a collaborative way that each agent only uses the information from its neighboring agents where the neighborhood structure is depicted as a graph, often undirected or directed. There has been a significant interest in consensus-based distributed gradient method. One fundamental work is [9] which developed the distributed gradient descent on undirected graph. This algorithm consists of local gradient step and consensus step based on communication between neighboring agents. The convergence property of the algorithm has been studied in the works [8]- [11]. There are also various distributed algorithms containing the distributed dual averaging method [12], consensus-based dual decomposition [13], [14], and the alternating direction method of multipliers (ADMM) based algorithms [15], [16]. These algorithms work with doublystochastic matrix associated with the undirected graph. The gradient-push algorithm was introduced in [17] to solve the distributed optimization for directed graph which utilizes push-sum algorithms [18], [19]. The communication of this algorithm is represented by column stochastic matrix, which requires each agent to know its out-degree at each time, without having the information of the number of agents. This algorithm has influenced a significant impact of later works. The work [20] studied the algorithm with gradient having a noise. The time varying distributed optimization was also considered [21] using the gradient-push algorithm. Recently, stochastic gradient-push algorithm was designed for large scale deep learning problem [22]. This work was also extended in [23] further to quantized communication settings. We also refer to [24], [25] where the authors studied the asynchronous version of the gradient-push algorithm. Regardless of the types of graphs, these distributed algorithms require each agent to communicate with their neighbors at every iteration, which leads to a overhead in restricted environments. Power consumption by communication may become more significant than that by computation of control inputs or optimization algorithms [26]. Recently, the eventtriggering approach has appeared as a promising paradigm to reduce the communication load in distributed systems. In the distributed detection problem over sensor network [27], [28], each sensor censors its local data and send the updated data to the fusion center only when the data is informative. For distributed control problems, agents send their coordinate information only when a triggering condition is satisfied [29], [30]. For the distributed optimization problems, recent works [31]- [35] developed distributed optimization algorithms with event-triggered communication to overcome the communication overhead of distributed systems. Lu-Li [32] designed the distributed gradient descent with event-triggered communication for the distributed optimization on the whole space, and it was further studied in Li-Mu [36] to establish a convergence rate. For the distributed optimization on bounded domain, Kajiyama et al [31] designed the projected distributed gradient descent with event-triggered communication. Liu et al [37] extended the work to the case with constant stepsize. Cao-Basar [38] studied the online distributed problem using the distributed event-triggered gradient method. In these algorithms, each agent sends its state only when the difference between the current state and the latest sent state is larger than a threshold, therefore reducing possible unnecessary network utilization. The consensus-based distributed optimization algorithms with event-triggering communication mentioned above have been proposed for the undirected graph. In this work, we are interested in developing a distributed optimization on directed graph involving the even-triggered communication. Precisely we propose the gradient-push algorithm incorpo-rating the event-triggered communication. In the proposed algorithm, each agent only sends its state information when the differences between the latest sent states and the current states are larger than a triggering thresholds. We prove that the algorithm solves the distributed optimization under suitable decays and summability conditions on the stepsize and the triggering thresholds. The numerical experiments are given for the proposed algorithm, supporting the theoretical results. The rest of the paper is organized as follows. In Section 2, we state the problem and introduce the algorithm with its convergence results. Section 3 is devoted to provide a consensus estimate, which is essentially used in Section 4 to prove the convergence results. In section 5, we present numerical results of the proposed algorithm. Before ending this section, we state several notations used in this paper. For a matrix A ∈ R n×m , a ij or x T x denotes the standard Eculidean norm. In addition, for X ∈ R m×d given by A. Problem statement We consider the distributed optimization problem, which consists of m agents connected by a network that collaboratively minimize a global cost function given by the sum of local private cost functions. Formally, the problem is described by where f i : R d → R is a local convex cost function only known to agent i ∈ V = {1, 2, · · · m}. We let f * be the optimal value of problem (1) and denote by X * the set of optimal solutions, i.e., which is assumed to be nonempty. We make the following standard assumption on the local cost functions. We set D = max 1≤i≤m D i . The communication pattern among agents in (1) at each time t ∈ N ∪ {0} is characterized by a directed graph G(t) = (V, E(t)), where each node in V represents each agent, and each directed edge (i, j) ∈ E(t) means that i can send messages to j. In this work, we consider a sequence of graphs {G(t)} t≥0 satisfying the following assumption. Assumption II.2. The sequence of graph {G(t)} t≥0 is uniformly strongly connected, i.e., there exists a value B ∈ N such that the graph with edge set ∪ We define in-neighbors and out-neighbors of node i, respectively, as N in The matrix A(t) ∈ R m×m is column stochastic and we recall some useful properties of this matrix from [17,Corollary 2]. Corollary 2). Suppose that the graph sequence {G(t)} is uniformly strongly connected. Then, the following statements are valid. 1) For each integer s ≥ 0, there is a stochastic vector φ(s) such that for all i, j and t ≥ s for some values C 0 ≥ 1 and λ ∈ (0, 1) depending on the graph sequence. 2) The following inequality holds. Here we denote by A(t : s) the matrix given as In addition, we consider the following assumptions on the stepsize and the thresholds for the trigger conditions. Assumption II.6. The sequence of event-triggering thresholds {ζ(t)} t≥0 is monotonically non-increasing and satisfies Note that ∞ t=1 t 3/2 ζ(t) < ∞ implies that there exists a finite M such that ∞ t=0 ζ(t) = M . If we set a new sequence {ζ(t)} t∈N byζ(t) = ζ(t)/(M + 1), then it satisfy ∞ t=1ζ (t) < 1. Hence if we have a sequence {ζ(t)} t∈N satisfying ∞ t=1 t 3/2 ζ(t) < ∞, then we may divide the sequence by a positive constant to satisfy Assumption II.6. One example of the sequence that satisfies Assumption II.6 is ζ(t) = 1 3t 3 . For T ≥ 0 we set the following variables Algorithm 1 Distributed Event-Triggered gradient-push algorithm Require: Initialize x i (0) arbitraily, y i (0) = 1 for all i ∈ {1, · · · , m}. Setx i (0) = x i (0) 1: for t = 0, 1, · · · , do 2: Compute the new action as Do not send and setx Do not send and setŷ i (t + 1) =ŷ i (t) 12: end if 13: end for (10) These variables will appear in the statements and proofs of the convergence result for Algorithm 1. Before finishing this subsection, we define the following constant whose positivity is proved in Lemma III.2 under the Assumption II.6. B. Main results Our first result establishes the convergence ofẑ i (t) to the optimal solutions for an arbitrary stepsize α(t) satisfying Assumption II.4, and event-triggering thresholds τ (t) and ζ(t) satisfying Assumption II.5 and II.6. Theorem II.7. Suppose that Assumptions II.1 -II.6 hold. Then the sequence {ẑ i (t)} t∈N for 1 ≤ i ≤ n of the Algorithm 1 satisfies the following property: lim t→∞ẑ i (t) = x * for all i and for some x * ∈ X * . Next we consider the Algorithm 1 with specific stepsize α(t) = 1/ √ t. This stepsize does not satisfy Assumption II.4, but we may obtain an explicit convergence rate as in the following result. Before stating the result, we give some notations which are used throughout the paper. We define the constants m ζ and B ζ by and These constants are well-defined if {ζ(s)} s≥0 is summable since we have the inequality \|1 T m θ(s)\| ≤ mζ(s) from the triggering condition. Theorem II.8. Suppose that Assumption II.1, II.2, II.5 and II.6 hold. Let Then we have for each T ≥ 0 and i = 1, · · · , m, the following estimate and x * ∈ X * . In addition, the right hand side is bounded by O(log(T + 1)/ √ T + 1). Here K(t) = β(t)/m ζ denotes a bound for asymptotic behavior of y i (t) stated in Lemma III.1. III. PROPERTIES OF THE SEQUENCE {y i (t)} t∈N AND DISAGREEMENT IN AGENT ESTIMATES A. Properties of the sequence {y i (t)} t∈N A convergence property of {y i (t)} t∈N and their positive uniform lower bound are key points in proving our main results. Let us first look at the case without event-triggering (ζ(t) = 0), which means all in-neighbors of agent i share the y i (t) with this agent for every time step. In this case, since y(t) =ŷ(t) for all t ∈ N, it holds that by (6) in Algorithm 1. Hence we can directly show that y(t) converges to φ(t) and has a uniform lower bound Q using Lemma II.3. In the event-triggered case, y(t) can be written as where θ(s) =ŷ(s) − y(s). Therefore the convergence and uniform lower boundedness property may not hold due to the additional term The following lemmas shows that y(t) has a positive uniform lower bound δ and converges to m ζ φ(t) instead of φ(t) under Assumption II.6. Lemma III.1. Suppose that Assumption II.6 holds. Then we have and the following estimate holds: where In addition, we have Using this in (12), we get which proves (16). Next we prove (17). For each s ≥ 0, by definition we have where we have set θ(0) = 0. Using this iteratively gives the following formula Using the above inequality, we obtain Hence we have Now we estimate the second and third terms in the right hand side of (19). Using (3) we have and Putting the estimates (20) and (21) in (19), we get This proves the second assertion of the lemma. Now we shall show that lim t→∞ t 3/2 β(t) = 0. Since λ ∈ (0, 1), it suffices to show that This fact follows directly from the fact that ∞ t=0 t 3/2 ζ(t) < ∞ and the following inequality The proof is done. Lemma III.2. Suppose that Assumption II.6 holds. Then the value δ ∈ R defined in (11) is positive. Proof. Note that from (3) and (4), we have Using the above inequality and Lemma III.1, we deduce for each 1 ≤ i ≤ m the following estimate Since β(t) converges to zero as t goes to infinity and λ ∈ (0, 1), there exists a time T ∈ N and a constantδ > 0 such that for any t ≥ T , Note that by Assumption II.2, each matrix A(t) has no zero row. This fact, together with the definition ofŷ and (6), for any t ∈ N we have Therefore, combining (22) with (23), we conclude that δ defined in (11) satisfies The proof is done. B. Disagreement in Agent Estimates In this subsection, we derive a bound of the disagreement in agent estimates {ẑ i (t)} m i=1 that will be used in the proofs of the main theorems. In the case without event-triggering (τ (t) = ζ(t) = 0), the paper [17] proved that ẑ i (t+1)−x(t) converges to zero for the stepsize satisfying Assumption II.4 as t goes to infinity. For the event-triggered case, the following proposition shows that the values {ẑ i (t)} m i=1 approach B ζx (t) instead ofx(t) as t goes to infinity due to the effect of the threshold ζ(t) for the triggering condition of {y i (t)} m i=1 . Proposition III.3. For any t ≥ 1 we have where K(t) = β(t)/m ζ and for t = 0 we have where the constant δ > 0 satisfies y i (t) > δ for all t > 0. To prove Proposition III.3, we consider a variable w i (t + 1) ∈ R d which is a companion to the variableŵ i (t + 1) ∈ R n defined as and their difference e i (t + 1) =ŵ i (t + 1) − w i (t + 1). Then we may rewrite the gradient step (8) as Summing up (26) for 1 ≤ i ≤ m and using that A(t) is column-stochastic, we havē (27) Now we find a bound of e i (t + 1) which is the difference between w i (t + 1) andŵ i (t + 1) associated to the eventtriggering τ (t). Lemma III.4. The quantity e i (t + 1) defined in (25) satisfies where Proof. By using the triggering condition, we have which proves (28). Summing this over 1 ≤ i ≤ m and using that A(t) is column stochastic, we find The proof is finished. Now we are ready to prove Proposition III.3. Proof of Proposition III.3. We regard x k (t) as a row vector in R 1×d and define the variables x(t) ∈ R m×d , ∇f (ẑ(t)) ∈ R m×d , and e(t) ∈ R m×d as . . . Using Assumption II.1 and (29) we have the following bound Since A(t) is column stochastic we have 1 T m A(t) = 1 T m , and combine this with (31) to have Combining (31) and (33) yields where φ(t) is the stochastic vector satisfying (3). By Lemma III.1, for y(t) := (y 1 (t), · · · , y m (t)) T ∈ R m×1 we have where r(t) satisfies r(t) ∞ ≤ β(t). Combining this with (34), we obtain By applying (3) here, we deduce where K(t) = β(t)/m ζ . From (33) we find the following estimate Combining this with (35) and using (32), we obtain By applying Lemma III.2, (28) and the above inequality to the norm of (30), we obtain It remains to estimate the case t = 0. By the algorithm, we havê Using this we find The proof is finished. By utilizing Proposition III.3, we analyze the relation betweenẑ i (t) and B ζx (t) under the assumptions on {α(t)} t∈N , {τ (t)} t∈N and {ζ(t)} t∈N of the main theorems. To do this, we first recall a useful lemma from [8]. Proof. We recall from Proposition III.3 the following inequality We notice that lim t→∞ t 3/2 K(t) = 0 by Lemma III.1. From this and the boundedness of α(s) and τ (s), it easily follows that In addition, by Assumptions II.4, II.5 and II.6, we know that lim s→∞ α(s + 1) = 0, lim s→∞ τ (s) = 0 and lim s→∞ ζ(s) = 0. Using this fact with Lemma III.5 in the right hand side of (37), we deduce which completes the proof. Proof. By Proposition III.3, we have The terms involving K(t) are fit to the inequality of the lemma. Let us estimate each summation not involving K(t) in the right hand side. Using that α(t) ≤ 1, the first term is bounded with The fourth term is bounded using We estimate the second term using In order to estimate the third term, we estimate where [a] denotes the largest integer not larger than a ∈ R. Using this we derive Putting the above estimates (39)-(42) in (38), we obtain which finishes the proof. IV. CONVERGENCE ESTIMATES In this section we prove our main results, namely Theorems II.7 and II.8. In Section 3, we obtained the bound of the disagreement in agent estimates. Especially, Corollary III.6 and III.7 investigate the difference between the stateẑ i (t) in the Algorithm 1 and B ζx (t) in (27). Based upon these results, Theorem II.7 and II.8 can be proved by comparing the cost values computed at the points B ζx (t) and x * . Lemma IV.1. Suppose Assumption II.1 holds. Then for any t ≥ 0 and x ∈ R d we have Proof. By convexity, we have where (24) is used in the last equality. Summing up the above inequality from i = 1 to i = m, we find that Now we estimate each term in the right hand side. First using the equality a, b = 1 Using (27) along with (28) and (2), we estimate the right-most term as We apply (28) again to estimate and use (2) to deduce Combining the above estimates on I,II and III, we have Finally we observe that (2) gives us the estimate Summing up the above two inequalities, we obtain the desired estimate. A. Proof of Theorem II.7 We recall the following lemma for proving Theorem II.7. Lemma IV.2 ( [17] Lemma 7). Consider a minimization problem min x∈R d f (x), where f : R d → R is a continuous function. Assume that the solution X * of the problem is nonempty. Let {x(t)} be a sequence such that for all x ∈ X * and for all t ≥ 0, Then the sequence {x(t)} converges to some solution x * ∈ X * By manipulating the estimate in Lemma IV.1, we obtain the following estimate which is suitable for applying Lemma IV.2. Corollary IV.3. Under the same assumptions in Lemma IV.1 we have where and Proof. We use Young's inequality to find Applying this to (43), we get Dividing both sides by m 2α(t+1)B ζ , it follows that Rearranging this we obtain the desired estimate. Now we are ready to prove Theorem II.7. Proof of Theorem II.7. By Lemmas IV.2 and Corollary IV.3 it is enough to prove ∞ t=1 (c(t) + d(t)) < ∞, where and It follows that ∞ t=0 c(t) < ∞ by Assumptions II.4 and II.5. Next we will show that ∞ t=0 d(t) < ∞. By Proposition III.3, it suffices to show that and The latter one is proved in Lemma IV.4 below. We proceed to prove (46). Using the Cauchy-Schwarz inequality, we have By rearranging and using the decreasing property of α(t) in Assumption II.4, we find Similarly, due to Assumption II.5, the last term is bounded as Gathering the above estimates, we find that ∞ t=1 (c(t) + d(t)) < ∞. Hence by Lemma IV.2, the sequence {B ζx (t)} converges to some solution x * ∈ X * . Finally, we apply Corollary III.6 to conclude that each sequence {ẑ i (t)}, i = 1, · · · , n, converges to the same solution x * . The proof is done. Lemma IV.4. Suppose that Assumption II.6 and Assumption II.5 hold. Then for the stepsize {α(t + 1)} t≥0 satisfying Assumption II.4 or α(t) = 1/ √ t, we have Proof. From Lemma III.1, we know that lim t→∞ t 3/2 K(t) = 0. Using this fact the summability of τ (s), it easily follows that Next, for α(t) satisfying Assumption II.4, we observe that For α(t) = 1/ √ t, by using that The proof is done. where Using this inequality, we deduce the following estimate The proof is done. To prove Theorem II.8, we first modify Lemma IV.1 which states the boundedness of . Lemma IV.6. Suppose that all the conditions are same as in Theorem II.8. Then we have for any T ∈ N, where J 1 (T ), J 2 (T ), and J 3 (T ) are defined in Theorem II.8 Proof. We recall from (48) the following inequality Dividing both sides of (48) by mH(t) 2α(t+1) , we get Summing this from t = 0 to t = T we obtain This, together with the fact that H(−1) = 1 and H(t) ≥ 1, gives T t=0 2α(t + 1) We find and by definition (10) Finally we estimate the last term of the right hand side of (51) using Corollary III.7 as follows: Test 1. Here we fix α(t) = 1/t 0.52 which satisfies Assumption II.4 and ζ(t) = 1/(3t 3 ) which satisfies Assumption II.6, and consider various choices of τ (t). We measure the relative distance between the states z i (t) and the optimal point x * the value We set the termination time k f as the first time k ∈ N when R d (k) < 10 −2 . And we let N x and N y be the average of total number of triggers for all agents until the termination time associated with τ (t) and ζ(t), respectively. Table I indicates the average of those values depending on τ (t) and ζ(t) in 100 trials. We first look at the effect of ζ(t), the threshold for variables y i (t). Table I shows that an existence of the threshold (ζ = 0) does not bring a big difference in the termination time if we compare the cases ζ(t) = 0 and ζ(t) = 1/(3t 3 ) with same τ (t), but there is a big improvement in the number of triggers for y i (t). Next we discuss the values N x and k f of Table I in terms of τ (t), the threshold for variables x i (t). As in Table I, some cases give us similar or worse results compared to the cases τ (t) = 0. For τ (t) = 1/t 1.1 , the number of triggers is decreased by more than 70%, and the termination time increased by more than 530%. For τ (t) = 1/t 1.7 , both the termination time and the number of triggers increased by almost 36% when ζ(t) = 0 and remain similar when ζ(t) = 1/(3t 3 ) compared to the cases τ (t) = 0. For τ (t) = 1/t 1.5 , the termination time is almost same, the number of triggers decreased by more than 20%. These results show that the proposed gradient-push with event-triggered communication with proper τ (t) and ζ(t) can diminish the number of communications to achieve the convergence compared to the gradient-push algorithm without triggering. The threshold functions τ (t) and ζ(t) should be chosen carefully considering the characteristics of the given optimization problem. VI. CONCLUSION In this work, we considered the gradient-push algorithm with event-triggered communication for distributed optimization problem whose agents are connected by directed graphs. We showed that by the algorithm each agent's state converges to a common minimizer under a diminishing and summability condition on the stepsize and the triggering function. Numerical simulations have been conducted to support the convergence results. It would be interesting further about how to choose an optimal triggering function for reducing the communication burden, reflecting various elements such as the connectivity of graphs and the number of agents.	2021-11-12T02:15:42.430Z	2021-11-11T00:00:00.000	{ "year": 2021, "sha1": "600575ce66e9b72a40631770b7105e69157e4c28", "oa_license": null, "oa_url": null, "oa_status": null, "pdf_src": "Arxiv", "pdf_hash": "600575ce66e9b72a40631770b7105e69157e4c28", "s2fieldsofstudy": [ "Computer Science" ], "extfieldsofstudy": [ "Mathematics" ] }
18175465	pes2o/s2orc	v3-fos-license	Revision surgery after pregnancy in a patient with congenital kyphoscoliosis Abstract Rationale: Rod breakage during pregnancy and delivery has never been described in a patient who has undergone surgery for congenital scoliosis (CS). Here, we present an unusual but significant case of revision surgery. Patient concerns: A 29-year-old woman presented with low back pain during pregnancy after posterior osteotomy, correction and fusion at T9 to L5 for CS. Radiographs during follow-up, 4 months after the patient gave birth, demonstrated rod breakage. Diagnoses: Rod breakage after orthopaedic surgery of congenital kyphoscoliosis Interventions: The patient was taken into the operating room for replacement of the broken rods, recovery of sagittal balance, bone graft fusion, and improvement of stability by cross-connection. The patient recovered fully by the 3-month postoperative follow-up. Outcomes: In follow-up, the instruments were in good condition, the orthopedic effect was not lost, and low back pain relief was observed. Lessons: We opine that the rod breakage during pregnancy resulted from weight gain and a lack of an anterior approach to the supportive bone graft. Therefore, female patients with spinal surgery should visit the hospital for advice before pregnancy. Introduction The etiology of congenital scoliosis (CS) is unknown, although a variety of factors have been implicated in the development of vertebral abnormalities. [1] The mainstay of surgical treatment remains early diagnosis before severe curvature and deformity develop. Occasionally, patients present with large deformities that require more significant procedures. [2] Statistics indicate that the female:male ratio is 1.4:1, [3] meaning that more than half of patients are women. Therefore, many patients who are treated with surgery will face pregnancy and delivery. Previous studies have suggested that pregnancy does not significantly increase fused scoliosis curvatures or the remaining unfused curvatures. [4,5] However, there have been no reports concerning this situation in CS. Here, we present a case of revision surgery in a female patient with congenital kyphoscoliosis with an unusual presentation after pregnancy and delivery. Case report We present the case of a 29-year-old woman who was diagnosed with congenital kyphoscoliosis in childhood. She remained untreated until the deformity increased and her low back pain worsened sharply. At her first outpatient visit, plain radiographs of the spine showed the following in the coronal plane: lumbar scoliosis with a Cobb angle of 55°(L1-L3), deviation of the apical vertebra to the central sacral vertebral line (CSVL) of 46.61 mm, and location of C7 in the CSVL. In the sagittal plane, lumbar kyphosis with a Cobb angle of 93°(T11-L2), deviation of the most kyphotic vertebra to the C7 vertical line of 112.95 mm, and a distance between the C7 vertical line and the posterior superior margin of the sacrum of À15.53 mm were observed (Fig. 1, Table 1), suggesting a need for surgical correction. After hospitalization of the patient, further examination was performed. On clinical physical examination, the surface lacked abnormal hair and hyperpigmentation, and a pelvic tilt to the left was observed. Left lumbar and severe kyphoses were noted. Shoulder imbalance, stretching and bending of the back with mild activity, and mild tenderness of the lumbar muscle were also observed. Limb muscle tension and myodynamia and feeling in the limbs and saddle area were all normal. No reflection or pathological characteristics were abnormal either. In February 2015, due to increasingly severe back pain (beginning in the fourth month), the patient had to go to the hospital during pregnancy. To avoid radiation, the patient refused to undergo plain radiographs. We suggested that she should undergo reexamination after delivery. In April 2016 (4 months after delivery: eutocia, no anesthesia, male baby, 2.9, and 62.5 kg before delivery), plain radiographs of the spine showed rod breakage, increased kyphoscoliosis, and spinal instability at follow-up (Fig. 6). Her plain radiographs of the spine also showed the following in the coronal plane: lumbar scoliosis with a Cobb angle of 32°(L1-L3), deviation of the apical vertebra to the CSVL of 69.04 mm, and a distance between C7 and the CSVL of 24.91 mm. In the sagittal plane, lumbar kyphosis with a Cobb angle of 54°(T11-L2), deviation of the most kyphotic vertebra to the C7 vertical line of 87.1 mm, and a distance between the C7 vertical line and the posterior-superior margin of the sacrum of À3.46 mm were observed, suggesting a need for revision surgery. On April 25, 2016, a posterior revision was performed. First, the broken rods were removed after exposure. Second, suitable rods were reinstalled, and sagittal balance was recovered. Third, the reprocessed right partial ilium and artificial bone were mixed to perform bone graft fusion on the back side. Lastly, crossconnection was used to improve the stability between L1 and L2. Spinal cord monitoring was performed during surgery. The bleeding volume was approximately 300 mL. Plain radiographs of the spine showed the following in the coronal plane: lumbar scoliosis with a Cobb angle of 37°(L1-L3), deviation of the apical vertebra to the CSVL of 57.19 mm, and a distance between C7 and the CSVL of 22.26 mm. In the sagittal plane, lumbar kyphosis with a Cobb angle of 20°(T11-L2), deviation of the most kyphotic vertebra to the C7 vertical line of 62.62 mm, and a distance between the C7 vertical line and the posterior-superior margin of the sacrum of À17.38 mm were observed using the Expedium spinal system (Johnson & Johnson) (Fig. 7). At the 3-month postoperative follow-up, the instruments were in good condition, the orthopedic effect was not lost, and low back pain relief was observed. Relative parameters are listed in Table 1 ( Fig. 8). Discussion Vertebral anomalies causing CS are classified on the basis of failures of formation, segmentation, or both. The natural history depends on the type of anomaly and the location of the anomaly. [6] Congenital vertebral anomalies have the potential to progress, so careful assessment and monitoring are essential, and early intervention may be desirable. [7] The mainstay of treatment is either observation or, in the case of curve progression (>10°/y), surgery. [8] The hallmark of surgical treatment is early intervention before the development of large curvatures. The surgical options for congenital spine deformities are numerous and depend on the type of anomaly, the degree of deformity, and the age of the patient. Posterior hemivertebra resection, correction, and fusion are together the mainstream method of treatment for adult patients. [2,9] Given that congenital vertebral deformity has the characteristic of progressing quickly, the best treatment time is before puberty. Therefore, most female patients undergo orthopedic surgery before adulthood. These female patients may think about the following questions: Will orthopedic surgery affect pregnancy? Will pregnancy aggravate the curvature? Danielsson and Nachemson used a contrast method for 136 surgically treated women and 111 brace-treated women. The results showed that there was no correlation between progression of the major or lumbar curve and the number of pregnancies or between curve progression and the age at first pregnancy. [5] In addition, Betz et al reported that the age of the patient at the time of the first pregnancy did not influence the risk of progression and that the stability of the curve before pregnancy did not decrease the risk of its progression during pregnancy. In patients who had undergone spinal fusion, progression in the unfused portion of the spine was negligible. [10] Lebel et al [11] suggested that scoliosis is not a risk factor for adverse pregnancy outcomes, and specifically labor dystocia. However, the case described here showed an exception-rod breakage. The patient had symptoms of low back pain in the fourth month of pregnancy, 2 years postoperation. In particular, dual-side connecting rods broke at the level of the osteotomy (L1/ L2) after delivery. The Cobb angle of T11 to L2 in the sagittal plane increased to 54°(it was 15°after the first surgery), and the distance between the kyphosis and the C7 vertical line increased to 87.10 mm (it was 14.85 mm after the first surgery); these findings demonstrate serious loss of balance and serious vertebral instability in the sagittal plane. What was the cause of the rod breakage? An average of 30°to 40°correction can initially be achieved with 1-level PSO [12][13][14] ; as the current case involved sagittal lumbar kyphosis with a Cobb angle of 55°, the operative plan of L1/L2 use of PSO was appropriate. [2] On the other side, the kyphoscoliosis dramatically improved after operation (preoperation T11-L2 angle of 93°, postoperation T11-L2 angle of 15°). The correction was satisfactory and was not lost by the 3-month postoperative follow-up. Moreover, there was no back pain after the operation. Adding-on could be ruled out because the patients' bones were fully mature. [15] In our opinion, the weight gain of pregnancy was the most likely reason for the rod breakage. Generally, data show an average gestational weight gain of 19.7 ± 5.1 kg. [16] Here, the patient's weight was 46 kg before pregnancy, and her weight was 62.5 kg before delivery, indicating a 16.5-kg increase. Abdominal weight gain could have increased the stress concentration and overtaxed the instruments, resulting in kyphosis progression and, ultimately, rod breakage. At last, due to the different fixation strategies between idiopathic and CS, with resection of a deformed vertebra in CS, the osteotomy zone cannot be implanted with pedicle screws, and stress concentration will appear in this zone. Does pregnancy increase the incidence of adverse events postoperatively? There is a lack of literature on this subject. Can a patient avoid rod breakage if an anterior approach is added to support the bone graft in the first operation? We are confident that we can reduce the risk if we add this feature. This is regrettable, however. In recent years, with the improvement of medical technology, more and more female scoliosis patients have undergone treatment with surgery. The abovementioned idiopathic scoliosis was not a risk factor for delivery in a previous study, and the curve will not progress in pregnancy. However, the present study's subject was CS, and research on the relationship between CS and delivery is lacking. To a certain extent, this article reflects a preliminary study. Our research specifically suggests that for women with CS, connecting rods can easily break in the osteotomy zone during pregnancy and delivery if fusion is not sufficient in the osteotomy zone. For female patients with congenital spinal deformity who wish to get pregnant, we suggest attempting sufficient fusion in the osteotomy zone, such as via anterior fixation support and reinforcement with cross-connection. Spine X-rays should be taken before pregnancy and after delivery. Measurement of the sagittal and coronal parameters preoperation can help to evaluate balance and the fusion situation in the osteotomy zone. Postoperative images can specifically be compared with preoperative images. Whether controlling one's weight during pregnancy can reduce the occurrence of adverse events has not been studied, but we speculate that weight control is a method that can reduce the stress in the osteotomy area, thus reducing the occurrence of adverse events. This topic needs further research.	2018-04-03T02:06:51.796Z	2016-12-01T00:00:00.000	{ "year": 2016, "sha1": "744d0b6c53578e3b1afb1801765aea6548310f0b", "oa_license": "CCBYSA", "oa_url": "https://doi.org/10.1097/md.0000000000005624", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "744d0b6c53578e3b1afb1801765aea6548310f0b", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
220920001	pes2o/s2orc	v3-fos-license	On regularization methods based on R\'enyi's pseudodistances for sparse high-dimensional linear regression models Several regularization methods have been considered over the last decade for sparse high-dimensional linear regression models, but the most common ones use the least square (quadratic) or likelihood loss and hence are not robust against data contamination. Some authors have overcome the problem of non-robustness by considering suitable loss function based on divergence measures (e.g., density power divergence, gamma-divergence, etc.) instead of the quadratic loss. In this paper we shall consider a loss function based on the R\'enyi's pseudodistance jointly with non-concave penalties in order to simultaneously perform variable selection and get robust estimators of the parameters in a high-dimensional linear regression model of non-polynomial dimensionality. The desired oracle properties of our proposed method are derived theoretically and its usefulness is illustustrated numerically through simulations and real data examples. Introduction We consider the high-dimensional linear regression model (LRM) given by where X i = (X i1 , .., X ip ) T are the explanatory variables, β = (β 1 , .., β p ) T ∈ R p is the vector of unknown regression coefficients and U i s are random noise with U = (U 1 , ..., U n ) ∈ R n being normally distributed with null mean vector and variance covariance matrix σ 2 I n . Assume that the explanatory variables are stochastic in nature; in other words, X T i , Y i , i = 1, ..., n are independent and identically distributed. Without loss of generality, we may assume that the model does not have any intercept terms by mean-centering all the response and covariates. We denote by X the (n × p)-dimensional matrix X = (X 1 , .., X n ) T . Therefore, we can write (1) in a matricial form by being Y = (Y 1 , ..., Y n ) T . We shall assume, in the context of sparse high-dimensional LRM, that the number of explanatory variables, p, is greater than the number of observations. More concretely, in this paper, we consider nonpolynomial dimensionality, i.e., log p = O(n α ) for some α ∈ (0, 1); see Fan and Lv (2010). In many applications most explanatory variables do not provide relevant information to predict the response, i.e., most of the true regression coefficients are zero. In this situation we say that the regression parameter β is sparse, in the sense that many of its elements are zero and the corresponding LRM is called "sparse high-dimensional LRM". Regularization methods for sparse high-dimensional data analysis are characterized by loss functions measuring data fits and penalty terms constraining model parameters. In LRM, regularization estimates of the parameter vector (β, σ) ∈ R p+1 is obtained by minimizing a criterion function or objective function of the form which consists of a data fit functional L n (β, σ), called loss function, and a penalty function p j=1 p λn (\|β j \|) , assessing the physical plausibility of β. The loss function measures how well β fits the observed set of data; on the other hand, the penalty is used to control the complexity of the fitted model in order to avoid overfitting. A regularization parameter λ n (λ n ≥ 0) regulates the penalty. From a practical point of view, the regularization parameter is chosen using some information criterion, e.g., AIC or BIC, or sorts of cross-validation. The former emphasizes the model's fit to the data, while the latter is more focused on its predictive performance. If L n (β, σ) corresponds to the loss function associated to an M-estimator, the minimizers of an objective function like (3) are called "penalized regression M-estimators". Such an estimator verifies the oracle properties, see Fan and Li (2001), if it estimates zero components of the true parameter vector exactly as zero with probability approaching one as sample size increases. Let us consider the l q norm β q = p j=1 \|β j \| q 1/q . The most common data fit functional is the quadratic loss function If we consider jointly with (4) the penalty function p j=1 p λn (\|β j \|) = λ β q q for a given λ, where q > 0, its minimization leads us to Bridge estimators (Frank and Friedman, 1993). For q = 2, we get the Ridge estimator considered in , while for q = 1, we get the well-known LASSO estimator introduced by Tibshirani (1996). However, Zou (2006) provided some examples where the LASSO is inconsistent for variable selection. Estimators obtained using l 2 penalty function or smooth penalty functions, in general, are unable to detect the null regression coefficients, see Fan and Li (2001). On the other hand, l 1 penalty function produces sparse estimators for the regressions parameters. Knight and Fu (2000) showed that, the estimators corresponding to a penalty function with q < 1 have the oracle properties, but for q = 1, the asymptotic distribution of the LASSO estimator corresponding to zero coefficients of the true parameter vector can put positive probability at zero. More details about the previous regularization procedures can be seen in Bühlmann and van de Geer (2011) as well as in the reviews by Fan and Lv (2010) and Tibshirani (2011). To address the problem of high false positives in LASSO, there have been several generalizations of it yielding consistent estimator of the active set under much weaker conditions. Some of the most popular are: the adaptive LASSO (Zou, 2006); the relaxed LASSO (related to the adaptive LASSO discussed by Meinshausen, 2007); the group LASSO (Yuan and Lin, 2006); Multi-step adaptive LASSO, considered in Bühlmann and Meier (2008); Dantzig selector (Candes and Tao, 2007); Fused LASSO (Tibshirani et al., 2005); Graphical LASSO, studied in Yuan and Lin (2007) and Friedman et al. (2007); etc. A further limitation of the estimators based on minimizing the objective function, Q n,λ (β, σ) , with quadratic loss function is their lack of robustness with regard to outliers. Alfons et al. (2013) established that the breakpoint of the LASSO estimator is 1/n, i.e., only one single outlier can make the estimate completely unreliable. Subsequently different procedures are developed for obtaining sparse estimators that limit the impact of contamination in the data. In general, these procedures rely on the intuition that a loss function yielding robust estimators in simple (classical) statistical set-up (Hampel et al., 1986) should also define robust estimators when it is penalized by a deterministic function. More concretely, the idea is to replace the quadratic loss function The Rényi's pseudodistance (RP) was introduced for the first time in Jones et al. (2001) and later additional properties were studied in Broniatowski et al. (2012). In this paper, we shall consider a loss function based on RP, to which we call RP loss function, jointly with non-concave penalties in order to simultaneously perform variable selection and to obtain robust estimators of β and σ in high-dimensional LRM with nonpolynomial dimensionality. This RP loss function was earlier considered for a low-dimensional LRM with p < n in Castilla et al. (2020) establishing their nice robust properties. Here we present a nonconcave penalized version of the PR loss function for the LRM. This method achieves simultaneously robust parameter estimation and variable selection in an ultra-high dimensional setting. It is worthwhile to note that Kawashima and Fujisawa (2017) considered the γ−divergence loss function, which has the same expression as the RP loss function for the LRM, but they only considered the LASSO penalty function (with no theory). Considering nonconcave penalties is the most important (empirical) difference with respect to Kawashima and Fujisama 's work, where only LASSO penalty was contemplate. Additionally, we also develop detailed theory of the proposed estimators, proving their oracle model selection property as well as consistency and asymptotic normality of the non-zero estimates. Performances of the proposed estimators are illustrated and compared with the state-of-the-art procedures via extensive simulation studies and interesting real data examples. For brevity, all the proofs are presented in the Online Supplement along with additional numerical results. The R codes for the computation of the proposed estimator is also provided in the Online Supplement enabling any practitioner to apply this procedure in future researches. 2 The proposed RP based regularization method in sparse highdimensional LRM The RP loss function is the empirical distribution function corresponding to the random sample u 1 , ..., u n from U 1 , ..., U n ; here I(·) denotes the indicator function. The probability mass function associated to G β n (u) is given by p β On the other hand, U i is normally distributed with mean zero and variance σ 2 . Therefore, the density function for U i is given by If we denote by P β,σ the measure of probability associated to the density function f β,σ (u) and by P β n the measure of probability associated to the empirical distribution function G β n (u), the RP between P β,σ and P β n , in accordance with Formula (7) in Broniatowski et al. (2012), can be written by where α is a non-negative tuning parameter controlling the compromises between efficiency and robustness. Taking into account that for α > 0. For α = 0 it is given by the limit as We are going to simplify the expression of R α P β,σ , P β n . It is immediate to see that, Therefore we have, An estimator for β and σ can be defined by minimizing R α P β,σ , P β n with respect to β and σ, i.e. for α > 0, and for α = 0 we get the maximum likelihood estimator (MLE). Remark 1 If we consider the loss function given by the RP between the density function associated to our model, f β,σ (y\|x), and the true density function for the model g(y\|x), the loss function associated with the RP is . If we assume that the distribution function of the random variable X is given by G(x), under some regularity Ghosh and Basu (2013) proposed, on the basis of the density power divergence (DPD), to minimize the expectation of the DPD expression between g(y\|x) and f β,σ (y\|x). In our situation we can consider the same but using the RP instead of DPD, i.e, where g(y, x) denotes the joint density function. Now, expression (8) can be approximated by Based on (6) and Remark 1, we can consider the loss function for the LRM based on RP by Again, we can observe that for α = 0, L α n (β, σ) coincides with the negative loglikelihood function. Therefore, the MLE is a particular case of the minimum RP estimator. Based on (9) the estimating equations are given for α > 0 by It is clear that the estimating equations of the minimum RP estimator, for α > 0, can be written as where and Thus, the minimum RP estimator is an M-estimator and its asymptotic distribution can be obtained on the basis of the asymptotic distribution of an M-estimator (see Maronna, et al., 2006). More details about the asymptotic distribution can be found in Broniatosky et al. (2012). Non-concave penalty functions Several penalty functions have been considered in regularization methods for high-dimensional LRM. In addition to the l i -penalties (i = 1, 2) associated to LASSO and Ridge methods, respectively, we can define the l 0 -penalty as p λ (\|β j \|) = λ I(β j = 0), or consider the l q -penalty functions given by p λ (\|β j \|) = λ \|β j \| q , which have been examined for this purpose over the choices 0 < q < 2. Some combinations of such penalties are also used; for example, the combination of l 1 and l 2 penalties are referred to as the elastic net penalty. The l 1 penalty is increasing and therefore imposes larger penalty for larger \|β j \|; hence it induces biased estimator for β even when the true β is sufficiently large. To remedy this flaw, the nonconcave penalties, such as SCAD (smoothly clipped absolute deviation) considered by Fan (1997) and Fan and Li (2001) and MCP (minimax concave penalty) introduced by Zhang (2010), transmit from l 1 function to constant function as β increases, in the sense that p λ (\|β j \|) is an absolute linear function around the 0 and it becomes a constant when \|β j \| is larger than some threshold. Fan and Li (2001) advocated three characteristics properties of a "good" penalty function, namely Unbiasedness, Sparsity and Continuity. It has been verified that the l q -penalty with q > 1 does not satisfy the sparsity condition, whereas the l 1 -penalty does not satisfy the unbiasedness condition; also the concave l q -penalty having 0 ≤ q < 1 does not satisfy the continuity condition. In other words, none of the l q -penalties satisfy the three conditions simultaneously. The SCAD penalty verifies the three properties and the MCP penalty verifies the unbiasedness and sparsity but not continuity. In this paper we shall consider non-concave penalties p λ (.) that admits a decomposition of the form whereJ λ (\|s\|) is a differentiable concave function. It is immediate to see that the penalties SCAD and MCP verify the decomposition (13) with the functionJ λ (\|s\|) being given, respectively, bỹ The proposed estimation procedure The criterion function for the nonconcave penalized RP estimator has the form with L α n (β, σ) the loss function and p λ (.) any nonconcave penalty function. Using the expression of L α n (β, σ) in (9), Q α n,λ (β, σ) is given by In the following, the estimator obtained by minimizing the objective function (15) with respect to β and σ will be called Minimum Non-concave Penalized RP estimator (MNPRPE). We could also define, in the same way, the statistical functional corresponding to the MNPRPE. For this purpose, let G be the true distribution function of the random vector (Y, X) and g(y, x) the corresponding density function, which can be expressed by g(y, x) = g(y/x)g(x). Given a random sample (y 1 , x 1 ), ..., (y n , x n ) from (Y, X) we shall denote by G n (y, x) = n i=1 1 n I(y i ≤ y, x i ≤ x) its empirical distribution function. Here, the inequality x i ≤ x refers to the vector ordering in R p . We then define the MNPRPE functional, T α (G), at the true joint distribution function, G, as the minimizer of with andp λ (β) = (p λ (β 1 ), .., p λ (β p )) T the penalty function. We denote the resulting penalized M-estimator as T α (G) = (β * , σ * ) T , with β * ∈ R p and σ * ∈ R, and the MNPRPE will be T α (G n ) with Influence function of the MNPRPE We compute the IF of the MNPRPE, following the notation of Avella-Medina (2017), depending on whether the penalty function is twice differentiable or not. For example, the l 2 −penalty function is twice differentiable but l 1 , SCAD and MCP penalty functions are not twice differentiable. We pay special attention to these last two non-concave penalties. We follow the same steps as in Section 3 in Ghosh and Majunder (2020). Note that equality (16) is equivalent to equation (1) in Avella-Medina (2017) with L * α (β, σ) = L(Z, θ). Then, the IF of the functional T α (G), corresponding to the MNPRPE, is the Gateaux derivative given by (Hampel, 1974) where G ε = (1 − ε)G + ε∆ (yt,xt) being ε the contamination proportion and ∆ (yt,xt) the distribution that assigns mass 1 at point (y t , x t ) and 0 elsewhere. Clearly, the IF describes the effect of an infinitesimal contamination, at the point (y t , x t ), on the estimate, standardized by the mass of contamination. Twice differentiable functions In case we assume that the penalty functionp λ (β) = (p λ (β 1 ), .., p λ (β p )) T is twice differentiable, we shall use Lemma 1 in Avella Medina (2017) in order to get the IF of the MNPRPE functional. Non-concave penalty functions Then, i) Denote β * = T β α (G) and assume that it has no null components (p ≤ n). Then, the IF of the MNPRPE functional T α (G) is given by ii) If β * has s (s < n) non zero components, i.e., β * = (β * 1 ) T , 0 T p−s T ( where β * 1 contains all and only s-non-zero elements of β * ), the corresponding partition of the MNPRPE functional T α (G) by Then, whenever the associated quantities exists, the IF of T β 2,α (G) is identically zero and the IF of T β 1,α (G) T , T σ α (G) T is given by Note that the boundedness of the IF of the model parameters does not depend on the penalty function. We shall assume in accordance with Fan and Lv (2011) and Ghosh and Majunder (2020) that the penalty functionp λ (β) = p j=1 p λ (β j ) verify the following condition: (C1) p λ (s) is increasing, continuously differentiable and concave in s ∈ [0, ∞). Also p λ (s)/λ is an increasing function of λ with ρ(p λ ) := p λ (0+)/λ being positive and independent of λ. It is not difficult to see, Li and Fan (2009), that the penalties 1 , SCAD and MCP verify condition (C1). Following Lv and Fan (2009) and Zhang (2010) we define the local and maximum concavity of a penalty function: Definition 5 The local concavity of the penalty function p λ (·) at b = (b 1 , .., b p ) T ∈ R p is defined as and the maximum concavity is defined as ξ( . It is not difficult to establish, using Condition (C1), that ξ(p λ , b) ≥ 0. Additionally, ξ(p λ ) ≥ 0 and using the mean-value theorem and assuming that the second derivative of p λ (·) is continuous, Let θ = (β, σ) be the unknown parameters of the LRM and we denote, following Ghosh and Majunder (2020), r i (θ) = y i − x T i β /σ, i = 1, ..., n and r(θ) = (r 1 (θ), ..., r n (θ)) T . We shall establish necessary and sufficient conditions for the existence of a local minimizer of the objective function, Q α n,λ (θ), given in (14). Theorem 6 Assume that the penalty function verifies Condition C1. Then, θ T = β, σ , is a strict minimizer of the objective function, Q α n,λ (θ), given in (14), for a fixed α ≥ 0, if and only if, where β 1 is the subvector of β formed by all noncero components, is the corresponding partition of x i in such a way that the number of components of x 1i coincides with the components of β 1 , the matrices T and Λ min (A) denotes the minimum eigenvalue of the symmetric matrix A. Now we are going to give some conditions in order to establish the oracle properties of MNPRPE, θ. It is necessary to introduce some notation: Assume that the first s components of β 0 are non-zero and the vector β 0 can be written as β T 0 = (β S0 , 0 p−s ) with β S0 ∈ R s . In the following we denote β T = (β S , β N ) and where X S ∈ R n×s and X N ∈ R n×(p−s) and we define the following matrices: where 1 n = (1, ..., 1) T ∈ R n . Based on this notation, Equation (22) can be written as (A2) The design matrix X verifies: denote the second order derivative with respect to (δ, σ) and δ T * = (δ, 0 p−s , σ) , b s is a diverging sequence of positive numbers depending on s and hence depend on n, d n = min j∈S \|β 0,j \|/2 and A ∞ the maximum of 1 norm of each row of A. (A3) Assume that d n ≥ log n/n τ and b s = o min n 1/2−τ log n, n τ (s + 1) log n for τ ∈ (0, 0.5]. In addition, assume if s = O(n τ0 ) that the regularization parameter λ satisfy Based on the previous assumptions we are going to establish a weak oracle property of the MNPRPE. Note that these assumptions are in line with those used by Ghosh and Majumder (2020). We start with the following proposition. Proposition 7 For all a ∈ R n and 0 < ε < \|\|a\|\|2 \|\|a\|\|∞ , we have, In Ghosh and Majunder (2020), the result in Proposition 7 is considered as an assumption, namely (A4); however, in our case, it always holds as can be seen from the proof of Proposition 7. To establish the asymptotic normality, we need an additional assumptions related to the Liapunov condition. We define the following matrices We now need to assume the following additional assumption. Computational Algorithm In this section, we discuss algorithms for minimizing the penalized objetive function Q n (β, σ), given in (14), with nonconcave penalties like SCAD and MCP. Recall that, the efficient algorithms for the least squares regression and group LASSO penalties, usually use penalty, which iteratively bounds the loss, resulting in a weighted least squared regression. In this paper, we propose to combine MM-algorithm and coordinate descent minimization for the RP loss function and the non-concave penalties including SCAD and MCP. The proposed method is iterative, and it updates the estimates of the parameters β and σ separately at each step. Before describing our proposal, let us briefly mention the MM and the coordinate descent algorithm to understand the underlying reasonings. MM-algorithm The MM optimization algorithm iteratively updates a current solution by finding a surrogate function that majorizes the objective function. Optimizing the surrogate function will then drive the actual objective function downward until a local minimum is reached (Hunter and Lange, (2004)). Mathematically, let be h(ν) a real-valued objective function. A function h M M (ν\|ν (m) ) is said to majorize Then, in the MM-algorithm, the next updated solution is obtained as The process is repeated for m = 0, 1, 2, ... until convergence is reached. The notation h M M (ν (m) \|ν (m) ) emphasizes the dependence of the current solution ν (m) , a crucial requirement of the MM-algorithm. . Proof. The first equation on (34) ensures both functions match at ν (m) , while the second one guarantees the stricly downward. The objective function h(ν) monotonically decreases at each step and hence, the MM-algorithm converges to a local minimum of h(ν). The descent property (35) lends an MM-algorithm remarkable numerical stability. Note that, in view of (35), ν (m+1) is not necessary a minimizer of h M M (ν\|ν (m) ), but it will suffices if only Thus, any other simple iterative algorithms may be used to minimize the majorization function. Moreover, Hunter and Lange (2004) proved that MM-algorithms boast a linear rate of convergence Kawashima and Fujisawa (2017) constructed the majorization function for γ−divergence loss function, which coincides with our RP loss for the LRM, by Jensen's inequality where κ(ν) is a convex function and z, ν and ν (m) are postive vectors. In our case of the RP loss function Therefore, it is enough to minimize the majorization function h M M (β, σ\|β (m) , σ (m) ) to downward L α n (β, σ) at each step. Note that only the last term of h M M depends on β. The same is also true to their penalized versions as required to compute the MNPRPE. Coordinate descent algorithm In order to minimize the majorization function in each step during the computation of the MNPRPE, we use the popular coordinate descent algorithm which optimizes the objective function with respect to every single parameter at a time, iteratively cycling through all parameters until convergence. This algorithm is specially appropriate for very high-dimensional problems, as each pass over the parameters requires only O(np) operations, and computational burden increases only linearly with p. While working with penalized objective functions with SCAD or MCP penalties, showed that, for the squared error loss in a univariate penalized regression, the minimization problem has an explicit solution. Consider the soft-thresholding operator (Donoho and Johnstone (1994)) and the simple linear regression model Given a random sample ((y 1 , x 1 ), .., (y n , x n )) and assuming for simplicity that the explanatory variable x is centered, the objective function for penalized least squares regression is If the p λ is the MCP penalty, then the minimizer of the objective function in (38) has the explicit form and for the SCAD penalty, the corresponding minimizer of (38) has the form where z = 1 n x T y is the solution of unpenalized univariate least squares regression. Coordinate descent minimization considers, on each iteration, p simple linear regression problems, and optimizes with respect to each and every parameter separately employing the univariate solution. Introducing the notation (−j) to refer to the portion that remains after the j-th column or element is removed, the partial residuals of x j are r −j = y − X −j β −j , where β is the most recently updated value of β. Thus, for given xed value of parameters { β k : k = j}, at a current estimates β, we wish to partially minimize the objective function with respect to β j yielding to the simple linear regression problem Therefore, the Coordinate Descent Algorithm is constructed as follows: 1. Set m = 0. Fix initial value β 0 , tuning parameter λ and tolerance ε (for convergence). Breheny and Huang (2011) showed that coordinate descent algorithm for the penalized squared loss with SCAD or MCP (with parameter a > 2 or a > 1 respectively) downward the objetive function at each iteration, . Furthermore, the sequence is guaranteed to converge to a point that is both a local minimum and a global coordinate-wise minimum of U n,λ . Remark 12 If the simple regression problem has not an explicit solution, but the penalty admits a decomposition as in (13), then Concave-Convex Procedure (CCCP) (An and Tao, (1997); Yuille and Rangarajan, (2003) ) may be used to bound the penalty with a convex approximation at which univariate regression possess an explicit solution and the coordinate descent algorithm can be applied (Lee (2015) [40]). In this case, the convergence of the method is guaranteed by the convexity of the objective function. The proposed algorithm for computation of the MNPRPE We propose to combine both optimization algorithms in order to compute the proposed MNPRPE of the regression parameter β along with the error variance σ at each step. Let us consider the objective function Q n (β, σ) defined in (14), and denote by ( β , σ (m) ) the current estimates at step m, m = 1, 2, .. We first apply MM-algorithm to bound L n (β, σ) as in (37). Then, the function to minimize is a weighted version of mean squared loss, so iterative coordinate descent algorithm can be used to update the current solution of β as β (m+1) . The convergence of the method is guaranteed by the convergence of both algorithms, as both decrease its objective function in each iteration. Next to obtain σ (m+1) , the update for σ, we consider the following derivative Note that the equation does not have an explicit solution. So, we should approximate it defining and then σ (m+1) is obtained as The full algorithm is described on the following pseudocode. and σ (0) , tuning parameter λ and tolerances ε 1 ,ε 2 (for convergence). Calculate µ as follows. and update σ (m) using , σ (m) ) ≤ ε 2 : Stop Else : set m ← m + 1 and go to step 2. 2013)). We dene a robust version of the HBIC as: and select the optimal λ that minimizes the HBIC over a pre-determined set of values. .1 Experimental Set-up We now present an extensive simulation study so as to evaluate the robustness and efficiency of the proposal MNPRPE under the LRM. We also estimate the regression parameters (β, σ) using other exiting robust and non-robust methods of high-dimensional LRM to compare their performances with our proposed method. The data are generated from the LRM (1) Error (EE) of the estimate σ as follows. Finally, in order to examine the efficiency loss against non-robust methods in absence of any contamination, as well as compare the performance in the presence of contamination in the data, we consider different scenarios: • Absence of contamination (pure data) • Contaminated data -Y -outliers : We add 20 to the response variables of a random 10% of samples. -X-outliers : We add 20 to each of the elements in the rst 10 rows of X for a random 10% of samples. Competing methods In order to compare our results with existing competitors, we calculate the same performance for measures the For the proposed MNPRPE we use the two most common penalties: SCAD and MCP. The results are very similar and hence, for brevity, we only report the ndings for the SCAD penalty. RLARS is used to initialize the computation of the MNPRPE and HBIC criterion (40) is applied to choose the regularizer parameter λ. Results Tables 1-6 summarize the simulation results for p = 500 covariates; the results for p = 100 and p = 200 are presented in the Online Supplement for brevity. The results evidence that our MNPRPE selects the true model better than any other method, and it is also more accurate in the estimation of the vector β. However, its indisputable advantage is its accuracy on the estimation of σ. The estimation error on σ is lower than that of any other method for all values of α. On the other hand, the optimum value of α hover around α = 0.3, in keeping with the best values for the LDPD-lasso. Finally, from results it is apparent that the use of nonconcave penalization improves the global performance of the method. To examine the performance of the proposed method with increasing dimensions, Figure 2 shows the mean root square error (RMSE) in prediction against the number of covariates in absence of contamination and 10% of Y −outliers respectively. The RMSE is calculated as RMSE( β) = 1 n y test − X test β 2 2 In both cases low values of the tuning parameter α register lower error. Moreover, the behavior of the method for the different tuning parameters is similar for any number of covariates, suggesting that the election of α should only be based on the compromise between efficiency and robustness (as described previously) . Table 9 contains the error measures as employed before, but now for difference between the predictions obtained from the model fitted with the (full) contaminated and the clean data for each method; the lower the values of these error measures, greater the stability is for the corresponding method. The difference on estimation when deleting outlier observation is lower for the MNPRPE than for any other method, illustrating its robustness. Conclusions In this paper we have presented a robust estimating method for the LRM in ultra-high dimensional settings. In order to have the matrix J α (G; β, σ) it is necessary to get Therefore, Now we are going to get the expectation of the random vector. We shall use E Y,X = E X E Y \|X . First we calculate the conditional expectations, Therefore we have A.2.2 Proof Theorem 6 Necessary condition: The classical optimization theory establishes that if θ T = ( β, σ) is a local minimizer of the objective function Q α n (θ), then it verifies the Karush-Kuhn-Tucker (KKT) conditions, i.e., there exists some where defined in Equation (9) of the main paper. Therefore we have It is clear that Equations (18) and (20) of the statement are verified. On the other hand, and Equation (19) of the statement is also verified. Sufficient condition: We shall assume that conditions (18)-(21) of the main paper are verified. We first constrain Q α n (θ) on the subspace B ⊂R s × R + . Assumption (21) of the statement establishes that Q α n (θ) is strictly concave in a neighborhood N 0 ⊂ B centered at θ. This fact, jointly with (18) and (20) of the statement, establish that θ, as a critical point of Q α n (θ) in B, is the unique minimizer of Q α n (θ) in the ball N 0 . Now it is necessary to prove that θ T = ( β, σ) is indeed a strict local minimizer of Q α n (θ) on R p × R + . We consider a sufficiently small ball N 1 ⊂ R p × R + centered at θ such that B ∩ N 1 ⊂ N 0 . Let γ 2 be the projection of γ 1 onto B. Then γ 2 ∈ N 0 and Q α n ( θ) < Q α n (γ 2 ) if γ 2 = θ, since θ is the strict minimizer of Q α n (θ) in N 0 , and it will be enough to prove that Q α n (γ 2 ) < Q α n (γ 1 ) for any γ 1 ∈ N 1 \ N 0 . On the basis of the mean-value theorem, where γ 0 lies on the line segment jointly γ 2 and γ 1 . The components of the vector γ 1 − γ 2 coincide in B ∩ N 1 because γ 2 is the projection of γ 1 onto B, and γ 2j = 0 for s < j < p + 1 because it belongs to B. Moreover, sg(γ 0,j ) = sg(γ 1,j ) if s < j < p + 1. Therefore, we have where γ 12 are the non null components of (γ 1 − γ 2 ). By γ 1 ∈ N 1 − N 0 we have γ 12 = 0. From concavity of p λ (s) , applying Condition (C1) of the main paper, we have that p λ (s) is decreasing in s ∈ [0, ∞). Therefore by Assumption (19) of the statement of the Theorem and continuity of p λ (s) , there exist Reducing the ball if it is necessary, we assume that N 1 ⊂ B( θ, δ), and therefore \|γ 0,j \| < δ, s < j < p + 1. A.2.3 Proof Proposition 7 Let Z 1 , .., Z n be independent bounded random variables with Hoedings inequality establishes We define, where φ 1,α (u) = u exp −α 2 u 2 . It can be shown that the function φ 1,α (u) is bounded, and so are the variables Z i . On the other hand, Now, for any a = (a 1 , .., a n ) ∈ R n , (a 1 Z i , .., a n Z n ) are n independent bounded random variables. Applying Hoedings inequality, we have , or equivalently, using that a1 \|\|a\|\|2 Z i , .., an \|\|a\|\|2 Z n have the same bounds, . 2 . It is clear that A∩B ⊂ ζ n . Then, The last inequality follows by Markov inequality. Using triangular inequality, Assumptions (A2) * and (A3) * of the main paper as well as that the function p λ is an increasing function we have, In a similar way, it is possible to see that E \|d In the same way that in Theorem 8, Pr(ζ 2 ) ≤ 1 − 2(p − s) exp(−c 1 u 2 n ) ≤ 1 − 2(p − s) exp(−n 1−2τ * log n) that tends to 1 when n → ∞, because log p = O(n α ). A second order Taylor expansion like in (49), jointly with the assumptions corresponding the the design matrix given in (A2) * of the main paper and the bound given for the regularization parameter, λ, we have, which shows that inequality. For any unit vector a ∈ R q , we consider the asymptotic distribution of the linear combination a T u n = The random variables r i (θ 0 ) = yi−x T i β 0 σ0 are independent, therefore the random variables ξ i are independent too with mean 0 and By hypothesis n i=1 var(ξ i ) → n→∞ a T Ga. Finally, by Assumption (A5) of the main paper and using the Cauchy- . A.3 Additional Numerical Results Tables 1021 present the simulation results, under the set-up discussed in Section 6 of the main paper, for p = 100 and p = 200 covariates. For each number of covariates, we consider pure data, 10% Y outliers and X outliers respectively. A.4 Example R Code for computation of the MNPRPE The present R code is provided to help the reader to implement the MNPRPE. This code has been used to obtain the results in the simulation study in Section 6 and to fit the model in the numerical example of glioblastoma gene expression analysis studied in Section 8 of the main paper. The code is inspired from the Robust and Sparse Regression via Gamma-Divergence (gamreg) package, created by Takayuki Kawashima (2017), and it	2020-08-03T01:00:46.235Z	2020-07-31T00:00:00.000	{ "year": 2020, "sha1": "4230e9fd47e377cc6b29a741259891a531914d78", "oa_license": null, "oa_url": null, "oa_status": null, "pdf_src": "Arxiv", "pdf_hash": "4230e9fd47e377cc6b29a741259891a531914d78", "s2fieldsofstudy": [ "Computer Science" ], "extfieldsofstudy": [ "Mathematics" ] }
36351251	pes2o/s2orc	v3-fos-license	Laryngeal epithelial thickness: a comparison between optical coherence tomography and histology Objectives: Optical coherence tomography, an imaging modality using near‐infrared light, produces cross‐sectional tissue images with a lateral pixel resolution of 10 μm. However, normative data is first needed on epithelial thickness for lesion characterisation, and, to date, little exists. The purpose of our study is to measure normal laryngeal epithelial thickness by in vivo optical coherence tomography, and compare these values to those obtained from fixed ex‐vivo laryngectomy specimens. Objectives: Optical coherence tomography, an imaging modality using near-infrared light, produces crosssectional tissue images with a lateral pixel resolution of 10 lm. However, normative data is first needed on epithelial thickness for lesion characterisation, and, to date, little exists. The purpose of our study is to measure normal laryngeal epithelial thickness by in vivo optical coherence tomography, and compare these values to those obtained from fixed ex-vivo laryngectomy specimens. Design and Setting: Prospective at a single medical center in California, United States. Participants: A total of 116 patients undergoing operative endoscopy. Main outcome measures: Optical coherence tomography images of clinically normal laryngeal subsites were selected. Calibrated measurements of epithelial thickness at various laryngeal subsites were recorded. Measurements of epithelial thickness from corresponding areas were obtained using optical micrometry on histologically normal regions of 15 total laryngectomy specimens. Descriptive statistics were performed. Results: Mean epithelial optical coherence tomography thicknesses were: true vocal cords (81 lm), false vocal cords (78 lm), subglottis (61 lm), aryepiglottic folds (111 lm), laryngeal epiglottis (116 lm) and lingual epiglottis (170 lm). Epithelial thicknesses in fixed tissues were: true vocal cords (103 lm), false vocal cords (79 lm), aryepiglottic folds (205 lm) subglottis (61 lm), laryngeal epiglottis (38 lm) and lingual epiglottis (130 lm). Conclusions: Optical coherence tomography does not have the artifacts associated with conventional histologic techniques. The inevitable development of office-based optical coherence tomography devices will increase the precision of laryngeal measurements and contribute to the clinical application of this technology in diagnosing laryngeal disease. Optical coherence tomography (OCT) is a new imaging modality analogous to ultrasound, with the exception that light from low-coherence sources rather than sound is used to create cross-sectional images of tissue, 1 with an axial resolution approaching that of light microscopy (approximately 7 lm). This technology is already in clinical use by ophthalmologists for examination of both retina and cornea. 2 Applications in the areas of cardiology, dermatology, urology, gastroenterology, and dentistry are in development. [3][4][5][6][7] In otolaryngology, OCT has been used to image the mucosa of the upper aerodigestive tract, the middle ear, the cochlea, and the thyroid gland. [8][9][10][11] OCT imaging is becoming increasingly important technology for the upper airway, especially in the larynx, where perhaps is the most promising application of this technology within the upper aerodigestive tract, because OCT can identify key structural features such as the basement membrane. 8,[12][13][14] This is particularly appealing for the management of early laryngeal cancers, as distinguishing between benign lesions and early carcinoma can be difficult without biopsy. Laryngeal microsurgery carries numerous risks including scarring and permanent vocal quality changes, and these concerns combined with the difficulty of diagnosing cancer on clinical examination alone have made the decision to pursue surgery often a considerable undertaking. Unlike advanced solid tumors of the head and neck whose boundaries and extension to other structures can be determined or estimated using CT or MR imaging, there are no imaging techniques in clinical use that can provide surgeons with the resolution needed to identify the axial spread of early cancer through the basement membrane and into the lamina propria. At present this information on tumor behavior at the tissue structural level requires a biopsy. OCT has potential to provide a means to image with high resolution (7 lm) the vertical extension of early cancers and other neoplasms through the basement membrane of the delicate laryngeal mucosa. The potential clinical utility of OCT in the larynx has already been demonstrated in several publications. 8,12,14 However, while detailed OCT images of both benign and malignant laryngeal disease continue to be acquired and published, little is known about the normative microstructure of the laryngeal mucosa. This is important because the dimensions of the epithelium are generally thicker in early T1 cancers, dysplastic lesions, and benign disease, which may mimic cancer on clinical examination. 8 If OCT is to evolve into a viable in vivo imaging modality to provide valuable information on the cross-sectional anatomy of the larynx, then it is absolutely necessary to know the dimensions of the tissue layers in the normal larynx. Just as in previous decades when CT and MRI images were compared with gross anatomy dissections and large whole body cross-sectional slices, 15 with OCT the thickness of the different layers of the laryngeal mucosa measured in vivo using OCT must be compared with measurements made in ex-vivo histological specimens. The purpose of this study is to accomplish this basic task. It is important to emphasise that although the ideal study would be to correlate the in vivo OCT images with in vivo laryngeal histological specimens, but this is unethical, because it requires performing total laryngectomies on healthy patients after in vivo OCT images were taken. Ethical considerations This study was approved by the Institutional Review Board of the University of California, Irvine. Subjects Optical coherence tomography imaging was performed in 116 patients undergoing surgical endoscopy of the upper aerodigestive tract under general anesthesia at the University of California, Irvine (UCI) Medical Center. The larynxes of 83 patients were imaged with OCT. While many patients exhibited some degree of laryngeal disease, thus the clinical indication for endoscopy, not all subsites in every patient were abnormal. Only normal subsites were used for this study, and laryngeal subsites with macroscopically abnormal epithelium (e.g., clinical appearance of carcinoma, inflammation, or acute trauma produced by intubation or suspension) on endoscopic examination were excluded. Thus OCT images were analysed from only 64 patients for this study. Information obtained from OCT imaging did not alter clinical decision-making in any way during the course of the study. Six of these 64 patients later underwent total laryngectomy, which provided direct histologic comparison with corresponding OCT images. All specimens were removed from the individuals undergoing salvage surgery for squamous cell carcinoma of the larynx following radiation therapy failure. To increase the histologic measurements database, an additional nine total laryngectomy specimens were obtained from the Department of Pathology at UCI Medical Center. Histologic preparations from tumor-uninvolved regions of these specimens as well as the six referred to above, for a total of 15, were examined with light microscopy. Table 1 outlines the characteristics of the two patient populations: one which underwent OCT, and the other from which laryngectomy specimens were procured. OCT measurements The OCT system used and the optical principles governing its performance have been described previously, 16,17 and will only be briefly reviewed here. A low coherence light source with central wavelength of k = 1310 nm was used (AFC BT 1020, JDS Uniphase, San Jose, CA, USA). The axial resolution of this system in tissue is 7 lm and is determined by the coherence length of the optical source. The lateral resolution is diffraction-limited (10 lm). A handheld probe containing the OCT fiber was placed through the laryngoscope bore (with or without the use of laryngeal suspension depending on the clinical need) in near, or more often, light, contact with the area of interest. Raster scanned images were generated by controlled motion of the imaging fiber using a piezoelectric stage (Model 663.4pr, Physik Instrumente, Tustin, CA, USA). In constructing these images a refractive index of 1.4, chosen to approximate that of most soft tissues in the body, was assumed. 18,19 Signals were acquired up to a depth of 1.6 mm while the lateral extent of each image was determined by the length over which the fiber was translated by the stage, typically 6 mm. OCT images were acquired with the left and right side of each image representing proximal (superior) and distal (inferior) respectively. OCT imaging was performed simultaneously with endoscopic visualisation of the laryngeal subsite under investigation. Due to operating room time constraints, not all laryngeal subsites were imaged in all patients, however, for each subsite imaged, more than one image was often acquired (see Table 2). The OCT images were digitally captured and catalogued in a computer database into one of six laryngeal subsites: true vocal cords, false vocal cords, aryepiglottic folds, laryngeal epiglottis, lingual epiglottis and subglottis. Digital morphometry was performed using Adobe Photoshop (Adobe System, San Jose, CA, USA) to obtain measurements of epithelial thickness. For each image, five measurements were made at 1 mm intervals and mean epithelial thickness was calculated. Measurements were not performed in laryngeal subsites in which the demarcation between epithelium and underlying lamina propria could not be identified conclusively. Histological measurements Laryngectomy specimens were prepared for histological exam according to standard protocol of the Department of Pathology at our institution which consists of immersion in neutral buffered 10% formalin solution for 48 h, followed by embedding in paraffin, sectioning, and staining with hematoxylin and eosin. Each specimen was examined under light microscopy and epithelial thickness measured by dropping a line perpendicular to the lumen surface and measuring a distance to the basement membrane, including any cilia. This was repeated every 1 mm. Histologically abnormal areas were not measured. Areas in which the quality of the specimen was judged inadequate to accurately determine epithelial thickness (e.g., tissue layers had separated or relevant structures could not be discerned) were also excluded. In order to standardise the artifact from preparation of the specimens, only permanent sections were measured; frozen sections were excluded. Due to the policies of the Department of Pathology regarding the retention ⁄ archival of slides in the specimen storage facility, not all subsites were available for examination on every patient (the remaining slides were destroyed prior to the time of this study). However, of the slides available, more than one section per subsite per patient was often present (see Table 2). Optical coherence tomography morphometry and light microscopy morphometry, or measurement of distances on each image ⁄ slide, were performed by two separate researchers working independently of one another. Neither researcher was privy to the findings of the other until the time of statistical analysis. Epithelial thickness values were average by subsite for each patient in both OCT and histologic measurements. If several histology slides were examined of a single subsite for a particular patient, averages per slide were in turn averaged to produce one mean per subsite per patient. These values were then used to calculate an overall mean per subsite for all patients in each arm. Standard errors for each value were also derived. Figure 1 shows a representative histological coronal section of a true vocal cord in the axial plane as viewed by light microscopy. The bracket marks the extent of the epithelium, with an arrow demonstrating the locus of the basement membrane. Figure 2 shows an OCT image in the coronal plane of the true vocal cords of the same patient. Again, the epithelium and its basement membrane are plainly evident and are marked with a bracket and an arrow respectively. In this patient, average thickness of the true vocal cord epithelium is 153 lm by light microscopy and 161 lm by OCT. Table 3 gives the mean epithelial thickness values for each subsite by both modalities and the standard error of the mean associated with each value. Figure 3 graphically depicts the mean thickness at each subsite obtained by light microscopy (x-axis) versus the mean thickness obtained by OCT (y-axis) with their respective standard errors for each subsite shown as cross-bars. Synopsis key ⁄ new findings Previous OCT studies done in the larynx have largely focused on ex vivo qualitative identification of histological features. 12,13,20 Our study is the only to date to establish a quantitative comparison between epithelial thickness in vivo with OCT, and ex vivo with light microscopy in the larynx. Comparisons with other studies Optical coherence tomography accurately depicts glands, vasculature, tissue layers and other structures. 8,13,14,20 In contrast, we have compared quantitative measurements obtained by both OCT and light microscopy in the upper aerodigestive tract. Similar studies have been done elsewhere in the body, particularly the eye. Chauhann and Marshall found that the total thickness of both fixed and fresh bovine retinas as determined by OCT had a strong linear correlation to measurements obtained by light microscopy, although this work has been criticised as describing artifact. 21 Wirbelauer et al. found that the thickness of diseased human corneas measured by in vivo OCT was directly proportional to that measured by light microscopy of the same specimens when fixed ex vivo. 22 In the upper aerodigestive tract, Wong et al. measured epithelial thickness by OCT, but did not correlate these findings with histology. 8 Strengths of the study Histology is the gold standard for tissue diagnosis. However, the use of histological slides examined with light microscopy for measuring normal epithelial thickness (as well as that of neoplasms) is not without its disadvantages. For microscopic examination, tissue must first be removed and then subjected to fixation, sectioning and staining, which may result in artifacts. Johnson et al. found that canine labiobuccal mucosal margins on histological slides were reduced to half their in vivo size, and tongue mucosal margins on histological slides similarly shrunk by one-third. 23 Formalin fixation shrinks laryngeal tissues grossly by 9-42% (in the false vocal cord and true vocal cord respectively) relative to freshly excised normal canine larynxes. 24 Further manipulation for histologic preparation such as embedding and sectioning may result in up to 20% additional reduction in size. 24 Moreover, different laryngeal subsites vary in the degree to which they are affected, creating distortion. 24 In contrast, OCT images areas of interest in vivo and in real time, without these artifacts produced by histologic preparation. Given that histologic preparation generally results in tissue shrinkage, our findings are somewhat surprising in that the OCT measurements are similar or even smaller than light microscopy measurements for several subsites. Wirbelauer et al. found that thickness of corneas measured in vivo by OCT was, in fact, greater than the thickness of the same corneas measured ex vivo by light microscopy, by about 9%. 22 This is keeping with a setting of histologic shrinkage. However, Chauhann and Marshall, working only with ex vivo bovine retinas, found the opposite: that light microscopy measurements were often larger than OCT by a comparable factor. 21 Our study appears to more closely reflect the findings of Chauhann. Chief among the reasons for this is likely the specific technique of OCT image acquisition. The in vivo study was performed on awake patients via slit lamp, with no direct contact to the tissues measured, and thus no pressure exerted, whereas the ex vivo study did not have to avoid direct contact given that there would be no concern for irritation of the cornea and patient comfort. 21,22 Although our work was performed in vivo, it frequently involved direct contact of the probe with the tissue under investigation. Even slight pressures created by the probe may be sufficient to alter the measurements, and it is impossible to gauge the amplitude of pressures exerted by the hand-held probe. There are several other points to bear in mind when analyzing these results. First, the study populations in the two arms are different. The OCT group is a heterogeneous population that includes individuals both with and without a history of laryngeal cancer, smoking (the biggest risk factor for laryngeal cancer), and radiation therapy (which may cause permanent damage in the tissues irradiated). In contrast, every patient in the light microscopy arm of this study underwent total laryngectomy for carcinoma, which recurred or persisted following radiation treatment. It is possible that their disease and previous therapies produced an increase in epithelial thickness via edema, inflammation, or fibrosis. 25 In the subsites of the larynx not affected by cancer, these alterations may be difficult to appreciate clinically but nonetheless present; differences of 10-20 mm are imperceptible to the naked eye. As noted in the methods, areas with histologically obvious abnormalities such as dysplasia or inflammation on microscopy were excluded from measurements in an attempt to reduce the potential impact of these effects. Ideally, we could measure the epithelial thickness by light microscopy of larynxes taken from young healthy individuals, but, of course, this is unethical because it will require that healthy individuals without any laryngeal disease underwent total laryngectomy for further histological examination. So what it is feasible to measure are larynxes from patients that underwent total laryngectomy for malignant disease of the larynx, although these larynxes are usually distorted due to radiation or surgical artifact. Another source of possible error in the use of OCT concerns variability of the refractive index. In processing the backscattering optical signal, we assume that the refractive index of laryngeal soft tissue is uniformly 1.4. While practical use of the OCT device demands this assumption be made, it is, nonetheless, a broad assumption as tissue is heterogeneous and will have actual refractive index variations due to glands, ducts, dense connective tissue, and different degrees of hydration. For example, Knuttel and Boehlau-Godau found that the presence of an epidermal sweat gland containing a significant amount of water may lower the refractive index at that point to 1.37, relative to 1.43 in the surrounding tissues. 19 This is relevant to our study in a variety ways. Pressures exerted intraoperatively by instrumentation, such as the endotracheal tube or laryngoscope, may result in altered tissue perfusion. Post-radiation changes may vary from edema to fibrosis. 25 Either may, in turn, significantly dynamically affect the hydration level of the tissue, and thus the refractive index. For instance, significant fibrosis and decreased tissue perfusion would both result in an increased refractive index. Underestimating the refractive index would result in an artificially thinned epithelium as measured by OCT, and may, in part, explain our data. Likewise, trauma from intubation or instrumentation (such as the forces exerted on tissue during suspension) could also produce the opposite effect, swelling, and lead to an over-estimation of epithelial thickness. Such changes when obvious were excluded from our analysis, however subtle changes may have been difficult to detect clinically. Finally, OCT measurements are subject to a certain degree of bias as a consequence of imaging being performed in areas most accessible to the probe. This is due to clinical limitations on the amount of time available to image patients while under anesthesia. Hence, there is a preponderance of true and false vocal cord measurements in contrast to measurements of the lingual surface of the epiglottis which would generally require repositioning of the laryngoscope. In contrast, laryngectomy renders all regions of the organ readily accessible to the probe. It may be valuable in the future to image a larynx during surgery, and then immediately after each step in the histologic processing cascade to determine the effect perfusion, formalin fixation, ethanol dehydration and paraffin embedding on the epithelial thickness measured. Nevertheless, morphometry of histological specimens will always be subject to highly variable shrinkage artifacts, and these can vary considerably depending on the nature of the tissue (e.g., lamina propria versus epithelium) and the details of the chemical processing route. This is a contributing factor to the large histologic measurement error bars in Fig. 3. In contrast, in vivo OCT may more accurately represent the true morphology of the targeted region since no chemical process (fixation) or volume contraction (dehydration) is involved. Our findings suggest that the impact of the probe contacting the tissue surface still remains a potential source of artifact, but we are presently working on the development and evaluation of a clinical OCT device mounted on a surgical microscope which may eliminate this effect. Likewise, the placement of the laryngoscope or endotracheal tube may cause subtle swelling in tissue; we are therefore currently engaged in the design of an OCT device that works in tandem with a rigid videostroboscopy system, though this is a formidable optomechanical engineering task. Clinical applicability of the study Regardless, the clinical impact of OCT in laryngology will be its ability to interrogate and image the epithelium and its demarcation with the superficial lamina propria. Knowledge of the normal thickness of these layers is critically important as they thicken in pathologic conditions, in particular dysplasia and chronic inflammation. Previous decades saw the verification of CT and MR imaging as clinically useful tools through the rigorous comparison of these modalities to macroscopic anatomy; OCT images must be likewise legitimised by reference to histology. Conclusion This study reports the in vivo thickness of the laryngeal epithelium in six distinct subsites measured using OCT, a non-invasive imaging modality. OCT does not have the artifacts associated with conventional histologic techniques and can be performed in living tissues. Knowledge of the normal thickness of the epithelium may be the first sign of early malignancy. The inevitable development of office-based OCT devices will increase the demand for precise measurements of the laryngeal microstructure and its correlation with histopathology. This innovation will likely accelerate the application of the OCT technology in both diagnosing laryngeal disease and monitoring its progressions.	2017-06-17T07:50:27.117Z	2009-10-01T00:00:00.000	{ "year": 2009, "sha1": "28f15c8c3ef5ea437ccd2d67821dbd3f4635f486", "oa_license": "CCBY", "oa_url": "https://escholarship.org/content/qt8716n9x0/qt8716n9x0.pdf?t=phez6h", "oa_status": "GREEN", "pdf_src": "Wiley", "pdf_hash": "28f15c8c3ef5ea437ccd2d67821dbd3f4635f486", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
244897324	pes2o/s2orc	v3-fos-license	HSV-1 reactivation is associated with an increased risk of mortality and pneumonia in critically ill COVID-19 patients Background Data in the literature about HSV reactivation in COVID-19 patients are scarce, and the association between HSV-1 reactivation and mortality remains to be determined. Our objectives were to evaluate the impact of Herpes simplex virus (HSV) reactivation in patients with severe SARS-CoV-2 infections primarily on mortality, and secondarily on hospital-acquired pneumonia/ventilator-associated pneumonia (HAP/VAP) and intensive care unit-bloodstream infection (ICU-BSI). Methods We conducted an observational study using prospectively collected data and HSV-1 blood and respiratory samples from all critically ill COVID-19 patients in a large reference center who underwent HSV tests. Using multivariable Cox and cause-specific (cs) models, we investigated the association between HSV reactivation and mortality or healthcare-associated infections. Results Of the 153 COVID-19 patients admitted for ≥ 48 h from Feb-2020 to Feb-2021, 40/153 (26.1%) patients had confirmed HSV-1 reactivation (19/61 (31.1%) with HSV-positive respiratory samples, and 36/146 (24.7%) with HSV-positive blood samples. Day-60 mortality was higher in patients with HSV-1 reactivation (57.5%) versus without (33.6%, p = 0.001). After adjustment for mortality risk factors, HSV-1 reactivation was associated with an increased mortality risk (hazard risk [HR] 2.05; 95% CI 1.16–3.62; p = 0.01). HAP/VAP occurred in 67/153 (43.8%) and ICU-BSI in 42/153 (27.5%) patients. In patients with HSV-1 reactivation, multivariable cause-specific models showed an increased risk of HAP/VAP (csHR 2.38, 95% CI 1.06–5.39, p = 0.037), but not of ICU-BSI. Conclusions HSV-1 reactivation in critically ill COVID-19 patients was associated with an increased risk of day-60 mortality and HAP/VAP. Supplementary Information The online version contains supplementary material available at 10.1186/s13054-021-03843-8. Several authors described an impact of HSV tracheobronchitis or pneumonia on mortality [29][30][31]. Moreover, HSV detection in lower respiratory tract and true HSV bronchopneumonitis were associated with poorer outcomes [27,32]. Data on the association between HAP/VAP and HSV reactivation are scarce [27,29]. In March 2020, during the first COVID-19 wave [8], we observed blood and respiratory samples positive for HSV in several patients. Therefore, we decided to systematically monitor on a weekly basis HSV reactivation in blood of critically ill COVID-19 patients admitted to our ICU. Furthermore, each on-demand respiratory samples was also processed for HSV detection. Our primary objective was to describe patients with HSV reactivation and to evaluate its impact on mortality in critically ill COVID-19 patients. Our secondary objective was to investigate the impact of HSV reactivation on severe healthcare-associated ICU-infections (i.e., HAP/ VAP and ICU-acquired bloodstream infections [ICU-BSI]) in COVID-19 patients. Study design and population We conducted an observational study using prospectively collected data from patients hospitalized at the Bichat university hospital, a large reference center for COVID-19 patients in the Northern Paris Region. We included all COVID-19 patients who underwent HSV tests in blood and lower respiratory tract (i.e., in mechanically-ventilated patients) samples on a regular basis. Data collection The study used data from the French prospectively collected multicenter OutcomeRea ® database. In accordance with French law, the OutcomeRea ® database was approved by the French Advisory Committee for Data Processing in Health Research and the French National Commission for Data Protection and Liberties (registration number 8999262). The database protocol was approved by the Institutional Review Board of the Clermont-Ferrand University Hospital, France. As the research objective of this study was not detailed in the original database protocol, this particular data analysis was then approved by the ethical committee of the Society of Resuscitation of French Language (SRLF), referenced as CE SRLF 21-83. Data storage and data analysis followed MR003 or MR004 methodology as appropriate and had been approved by French legal authorities. Informed consent was not required since the study did not modify patient management, and data were anonymously processed. Oral information was given to the patients or closest family about the possible use of data entered for research purposes, especially about determinants of prognostic of ICU patients. Demographic, clinical and microbiological data on Polymerase Chain Reaction (PCR)-tested COVID-19 patients were prospectively collected at ICU admission and daily by clinicians. The following variables at ICU admission were included: demographics, chronic disease/comorbidities, Sequential Organ Failure Assessment (SOFA) scores, body temperature, blood levels of C-Reactive Protein (CRP), D-Dimers, Lactate Dehydrogenase (LDH), and ferritin, ventilation status, and use of immune modulatory agents. Clinical and biological parameters were recorded daily. We also recorded daily data on antiviral agents (including acyclovir, ganciclovir, valganciclovir and valaciclovir) and immunomodulator (corticosteroids, Interleukin-6 receptor antagonists [IL6-ra] and interleukin-1 receptor antagonists [IL1-ra]) medications. Patients were monitored daily during the ICU stay and followed after ICU discharge until day 60 after admission. Study procedures Blood HSV specimens were prospectively collected on ICU admission and weekly. All mechanically-ventilated COVID-19 patients underwent regularly lower respiratory tract samples for HSV test until removal of the endotracheal tube or death. Endotracheal aspirates, bronchoalveolar fluid, and samples through plugged telescoping catheter were all considered as lower respiratory tract samples. HSV-1 and HSV-2 were routinely screened. All specimens were sent to an accredited reference virology laboratory. Blood and lower respiratory HSV were detected by an automated real-time PCR technology "Simplexa ® direct assay HSV1/HSV2 Diasorin ® ". HSV-1 and HSV-2 DNA were detected by real-time PCR targeting the HSV polymerase genes. Results for HSV-1 and HSV-2 were provided as positive or negative with PCR cycle threshold (Ct) values. Definitions and outcomes HSV reactivation was defined as a HSV-positive PCR in blood or respiratory samples. The primary outcome was the 60-day mortality. Secondary outcomes were HAP/VAP and ICU-BSI. HAP was defined as a pneumonia occurring in patients hospitalized for at least 48 h. VAP was defined as a pneumonia occurring in ICU patients mechanically ventilated for at least 48 h [33]. The following three criteria had to be fulfilled: (i) new or progressive and persistent infiltrates or consolidation or cavitation; (ii) pathological body temperature (< 36 °C or > 38 °C) or white blood cell count < 4 or > 12 × 10 3 cells/mm 3 ; (iii) either the new onset/increase of purulent aspirates or worsening gas exchange [33][34][35]. Endotracheal aspirates, bronchoalveolar fluid, and plugged telescoping catheter samples with semi-quantitative cultures were assessed for microbiological diagnostic with positivity thresholds of 10 6 , 10 4 and 10 3 cfu/mL, respectively. ICU-BSI was defined as a bacteremic episode occurring > 48 h after ICU admission. Typical skin contaminants (i.e., coagulase-negative staphylococci [CoNS]) were included only if ≥ 2 blood cultures showed the same phenotype on separate samples drawn within a 48-h period, or ≥ 1 blood culture positive for clinical sepsis, without any other infectious process, and with an antibacterial agent just initiated by the attending physician [36]. Statistical analysis Characteristics of patients were described as median (interquartile range [IQR]) or count (percent) for qualitative and quantitative variables, respectively. HSV positive and negative patients were compared with Chi-square, Fisher and Wilcoxon tests, as appropriate. The association between HSV-1 reactivation and mortality was estimated using univariable and multivariable Cox models. The variable of interest (HSV-1 reactivation) was forced as a time-dependent variable in the multivariable model. The following covariates were included in the models: age, chronic disease, SOFA score, type of ventilation, use of corticosteroids (i.e., well-known mortality risk factors in critically ill COVID-19 patients). To avoid overfitting with mechanical ventilation variables, only extra-respiratory components of SOFA score were taken into account in the multivariate model. Then, a hazard ratio (HR) for HSV-1 reactivation was derived; a HR > 1 indicated an increased risk for HSV-1 reactivation. The follow-up lasted 60 days. We performed subgroup analyses for blood and respiratory HSV-1 reactivation, respectively. Moreover, we tested the impact of acyclovir therapy on day-60 mortality by sorting HSV-1 positive patients into two groups according to acyclovir treatment. To evaluate the effect of HSV-1 reactivation on healthcare-associated infections (i.e., HAP/VAP and ICU-BSI), we used cause-specific hazards models. Time zero (T0) was the ICU admission. To determine the risk of healthcare-associated infections, the basis assumption was that censoring was not associated with an altered chance of the event occurring at any given moment. In the current analysis, ICU-mortality was considered as competing event with HAP/VAP or ICU-BSI. The risk of HSV-1 reactivation (versus non-HSV-1) as a timedependent variable was then estimated using Cox causespecific hazards models. Risk of HAP/VAP or ICU-BSI was expressed in cause-specific hazard ratios (csHR). We performed subgroup analysis for blood and respiratory HSV-1 reactivations, respectively. p-values < 0.05 were considered to be significant. Statistical analyses were performed using SAS 9.4 (Cary, North Carolina, USA). Study population Of the 153 SARS-CoV2 patients admitted for more than 48 h from February 2020 to February 2021 and who underwent at least one HSV screening, respiratory and blood samples were tested from 61/153 (39.9%) and 146/153 (95.4%) patients, respectively (Fig. 1). HSV results were positive in respiratory samples for 19/61 (31.1%) patients and blood samples for 36/146 (24.7%) patients. Fifteen of 153 (9.8%) patients had a HSV-1 reactivation both in blood and respiratory tract: HSV-1 reactivation occurred first in blood for 7/15, in respiratory tract for 6/15 and simultaneously for 2/15 patients. The median time from ICU admission to the first HSVpositive sample was 14 days (IQR 9.5-18). Importantly, HSV-1 was recovered from all HSV-positive specimens, no reactivation with HSV-2 was observed. Overall, 40/153 (26.1%) patients had at least one positive sample with HSV-1. In total, 481 samples were collected, including 120 respiratory samples: 91 (81.3%) from bronchoalveolar lavage, the other ones from upper tracheal secretions. HSV-1 reactivation and mortality Primary and secondary outcomes are described in Table 2. The median length of ICU stay was 23 (IQR 17-40) days and 9 (IQR 6-15) days for patients with and without HSV-1 reactivation (p = 0.001), respectively. Eighty-nine patients (58.2%) required invasive mechanical ventilation or Extra-Corporeal Membrane Oxygenation (ECMO). After ICU discharge, only one clinical HSV-1 infection (skin infection) was detected and treated with acyclovir. Among respiratory samples, univariable and multivariable Cox models did not show any association between mortality and HSV-1 reactivation (Fig. 2 and Additional file 1: Table E2). Among patients with HSV-1 reactivation, acyclovir was administered to twenty-eight (70%) patients, 3 (IQR 2-5) days after positivity in median. After adjustment for mortality factors and using acyclovir as a time-dependent covariate, our Cox model showed an increased risk of mortality at day 60 for patients treated with acyclovir (HR 4.37, 95% CI 2.12-9.02, p < 0.001). Discussion Using high-quality prospectively collected data from all COVID-19 patients admitted to a large French COVID-19 reference center, we showed that mortality and HAP-VAP risks were increased in patients with HSV-1 reactivation. However, the risk for ICU-BSI was not increased in patients with HSV-1 reactivation. In our cohort none of the COVID-19 patients had HSV-2 reactivation. Data on HSV reactivation in COVID-19 patients are scarce. Only one small study described HSV reactivation in lower respiratory tract of COVID-19 patients [37] without specifically investigating its impact on mortality or ICU-acquired infections. Furthermore, to our knowledge, no data on HSV detection in blood samples of COVID-19 patients were available. Interestingly, we observed higher rates of HSV positive PCR in the blood compared to previous studies including non-COVID-19 patients [38,39]. This finding may be explained by immune alterations with profound leucopenia [2,40] and a large use of immunomodulators in COVID-19 patients [11,14]. We showed that HSV-1 reactivation in COVID-19 patients was associated with an increased mortality. Impact of HSV reactivation on prognostic among critically ill non-COVID-19 patients is well known. Several authors showed an increased morbidity and mortality [25,29,30,41]. However, data on HSV-1 reactivation among COVID-19 patients are scant [37]. Le Balch et al. showed that HSV-1 reactivation among severe COVID-19 patients was associated with a longer duration of invasive mechanical ventilation and a longer length of stay, but without any impact on mortality. However, due to the low number of included patients (n = 38), adjusting for confounding factors was impossible. To the best of our knowledge, we performed the first study that prospectively investigated HSV-1 reactivation in COVID-19 patients. The relevance of HSV-1 reactivation and its real impact on mortality remains controversial [42]. On the one hand, HSV could be a simple bystander and its presence not reflect a direct pathogenicity [43][44][45]. Due to the observational nature of our study, we could not determine causality between HSV and mortality since many confounders both related with HSV-1 reactivation and mortality might have been unmeasured. It may be especially the case for immune paralysis that might have induced both HSV-1 reactivation and increased the mortality risk [46][47][48]. In our study, most patients did not demonstrate reactivation until about two weeks. Patients who had HSV-1 reactivation had higher levels of inflammation at admission, suggesting a different immune response comparing to patients without HSV-1 reactivation. Recently, Seeßle et al. [49] showed an increase in activated CD8 cells followed by a drop in interferon stimulated gene expression occurring after HSV-1 reactivation, also occurring in a similar time frame to what was seen in our study. This suggests some form of immune failure leading to mortality with HSV-1 reactivation just being an epiphenomenon. However, since our analyses were adjusted for severity of critical illness, a direct pathogenic role of HSV may be hypothesized. A recent meta-analysis [50] showed a decreased mortality among HSV patients receiving acyclovir. In light of these considerations, it could be possible that HSV is a true pathogen inducing pulmonary parenchyma damages. However, the only randomized trial evaluating the use of acyclovir in HSV reactivation [28] showed a marginal and non-significant impact on mortality, thus weakening this hypothesis. Of note, our data showed that adequate early acyclovir therapy was not associated with a prognostic improvement of HSV-1 positive patients. Although our study was not designed to evaluate the impact of acyclovir and these data were issued from a complementary analysis, it suggested that HSV-1 reactivation was more a marker of poor immune status than a disease that would need specific therapy. Further studies are necessary to determine if early acyclovir therapy could improve outcomes. We observed that HSV-1 reactivation was associated with an increased risk of HAP/VAP. Interestingly, only a few studies investigated the association between HSV reactivation and VAP in critically ill patients. Daubin et al. [51] performed a prospective observational study and determined the incidence of VAP among intubated patients in ICU. HSV-1 was recovered in only 31% of VAP patients. To date, the association between HAP/VAP and HSV has not been studied in COVID-19 patients. We showed an association between HSV-1 reactivation and HAP/VAP which remained significant after adjustment on several factors. On the one hand, our findings suggested that HSV-1 reactivation could induce lung injuries which could promote bacterial superinfections. On the other hand, it is conceivable that lesions induced by superinfections may promote HSV reactivation in lung and/or blood. The causality relationship between HAP/VAP and HSV reactivation remains to be elucidated. Overall, COVID-19 exposes critically ill patients to an increased risk of superinfections and HSV reactivation by induced immunosuppression and severe lung damages. HSV-1 reactivation among COVID-19 patients could help clinicians to identify patients at risk of developing superinfections. Moreover, frequent microbiological samples may be collected in order to early identify HAP/VAP. Our study has several limitations. First, we performed an observational study and residual confounding cannot be excluded. Therefore, associations between HSV-1 and mortality or HAP/VAP should be interpreted with caution. However, our multivariable models allowed us to adjust for several factors. It may be debatable to do not use IL1-ra or IL6-ra exposition as an adjustment covariate for multivariable Cox Models. Actually, Tocilizumab seems to improve outcomes in recent studies [13,52]. Nevertheless, only 31 (20.3%) patients received Interleukin-receptor antagonists ([IL-ra] i.e., Anakinra or Tocilizumab) in the first two ICU days in our cohort. The association between IL-ra and mortality using univariables Cox models was not significant (HR 1.04, 95% CI 0.54-2.01, p = 0.90). Therefore, we did not include IL-ra as adjustment covariate. Second, HSV reactivation in respiratory tract was tested only in intubated patients, which may result in a selection of patients with poorer prognostic. Nevertheless, HSV reactivation in blood was also tested in non-intubated patients, thus mitigating this selection bias. Moreover, our subgroup analyses showed an increased risk of mortality and HAP/VAP for HSV-1 reactivation in blood, thus suggesting that blood HSV-1 reactivation played a predominant role (i.e., compared to respiratory HSV-1 reactivation). Third, in our main analysis, we pooled respiratory and blood samples, thus assuming that HSV positive blood PCR was mostly due to a pulmonary reactivation. Our analyses showed a predominant relevance of blood HSV reactivation. It may be debatable that a HSV reactivation in blood may be associated with a respiratory HSV reactivation. Fourth, median time to HSV-1 reactivation was shorter than median length of ICU stay for patients without HSV-1 reactivation (14 days versus 9 days, respectively). Some patients without HSV-1 reactivation leaved the ICU before HSV-1 reactivation could occur. HSV-1 was not routinely tested after ICU discharge so a selection bias may occur. Nevertheless, even if ICU HSV screening was not performed after ICU discharge, we reviewed all patient courses and we detected only one case with a HSV infection (HSV skin infection): we suppose that event did not substantially modify the results. Moreover, no clinical or biological signs of HSV-1 reactivation were described so we could reasonably assume that no patient had HSV-1 reactivation after ICU discharge. Fifth, our analysis suggested that blood sampling may be sufficient to monitor HSV-1 reactivation. Nevertheless, we have to nuance that notion. We had a potential selection bias caused by the lack of respiratory samples for non-intubated patients which may result in a selection of patients with poorer prognosis. Moreover, our subgroup analysis for HSV-1 reactivation in respiratory tract showed a non-significant trend for healthcare-associated ICU-infections with a HR > 1, particularly for bloodstream infection. Sixth, we performed a monocentric study, thus limiting the generalizability to other ICU settings. However, a large sample size with high quality data was collected. Finally, our findings should be limited to critically ill patients. The rate of respiratory HSV reactivation in non-critically ill patients remained unresolved and further studies are needed in this context. Conclusion Critically ill COVID-19 patients frequently reactivate HSV-1 but not HSV-2. HSV-1 reactivation in critically ill COVID-19 patients was associated with an increased risk of day-60 mortality and HAP/VAP occurrence. The	2021-12-06T14:35:30.337Z	2021-12-01T00:00:00.000	{ "year": 2021, "sha1": "f9d3d9d3b18e4ee4401986e5a63450db086365aa", "oa_license": "CCBY", "oa_url": "https://ccforum.biomedcentral.com/track/pdf/10.1186/s13054-021-03843-8", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "f9d3d9d3b18e4ee4401986e5a63450db086365aa", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
261397122	pes2o/s2orc	v3-fos-license	Rebound increase in microRNA levels at the end of 5-FU-based therapy in colorectal cancer patients Treatment with 5-fluorouracil (5-FU) based therapy is still used for colorectal cancer (CRC). Epigenetics has become a focus of study in cancer because of its reversibility besides its known regulatory functions. In this study, we will monitor the change in microRNAs (miRNAs) levels with 5-FU-based therapy at baseline and after 3 and 6 months of treatment to be correlated with their prognostic potential. The expression levels of 5 miRNAs, namely miRNA223-3p, miRNA20a-5p, miRNA17-5p, miRNA19a-3p, and miRNA7-5p, were measured in the peripheral blood of 77 CRC patients, along with the expression of 3 proteins PTEN, ERK, and EGFR. At baseline, CRC patients had significantly higher levels of circulating miRNAs than healthy controls. This level was reduced after 3 months of 5-FU-based therapy, then increased after 6 months significantly in responder patients compared to non-responders. MiRNA19a-3p showed that significant pattern of change in the subgroups of patients with high ERK, EGFR, and PTEN protein levels, and its 6 months level after 5-FU-based therapy showed significance for the hazard of increased risk of disease recurrence and progression. Patients and methods Patients.This is a prospective study with 77 pathologically proven CRC patients and 20 age and sex-matched healthy control individuals.All patients had normal organ functions, and their demographics were recorded.All patients were chemo naive and received 5-FU-based therapy in doses and duration described in the protocols of CRC treatment followed by the Medical Oncologists in the Egyptian National Cancer Institute, Cairo University.At baseline, 77 patients were approved to participate in the study.Peripheral blood samples were collected from all patients and control groups to measure peripheral miRNA and protein levels.From those 77 CRC patients, 60 and 41 patients were sampled again after 3 and 6 months of 5-FU-based therapy. In patients with stage II disease with high-risk features, adjuvant capecitabine was administered orally as a single agent in a 1,250 mg/m2 dose twice/day for two weeks, followed by a 1-week rest period, given as 3-week cycles for a total of eight cycles.While stage III and metastatic patients received XELOX in 3-week treatment cycles in which intravenous oxaliplatin 130 mg/m2 was given on day 1, followed by oral capecitabine 1,000 mg/ m2 twice daily for two weeks and a one-week rest. The study was approved by the Institutional Human Research Ethics Committee of the Egyptian National Cancer Institute, Cairo University, Number 00004025, with IRB review Number 2010014019.3.Written informed consent was obtained from all participants for using their blood samples in this study.Patients were followed for 3 years till the primary endpoint of event-free survival (EFS), calculated from the date of primary treatment till the date of relapse or progressive disease, and the secondary end-point of overall survival (OS) calculated from the date of diagnosis till the date of death.Living patients or patients lost to follow-up were censored on the last known alive date. Serum samples collection. A whole peripheral blood sample (5 ml) was collected in an anti-coagulantfree Vacuette® blood collection tube (Greiner Bio-One, Kremsmünster, Austria) to isolate serum before the start of treatment (baseline) and then after 3 and 6 months of 5-FU-based therapy.The blood serum was separated by centrifugation at 1400 rpm for 10 min, and the clear serum supernatant was stored in RNase-free Eppendorf tubes at -80 °C for further use. RNA extraction and qRT-PCR. Total RNA, including miRNAs, was extracted from 200µL of thawed serum using the miRNeasy Mini Kit (Qiagen, Valencia, CA, USA, Cat.No. 217004) according to the manufacturer's instructions.The RNA content (ng/µl) was measured using the NanoDropTM (Thermo Fisher Scientific, USA).According to the manufacturer's protocol, cDNA Synthesis was performed using miScript II RT kit (Qiagen, Valencia, CA, USA, Cat.No. 218161) on 100 ng of total RNA in a final volume of 20 µl.Quantitative real-time PCR (qRT-PCR) was performed using miScript SYBRÒ Green PCR kit (Qiagen, Valencia, CA, USA, Cat.No. 218075).Mature miRNAs expression levels were generated in 96-well arrays using the custom miScript miRNA RT-PCR for miRNA223-3p, miRNA20a-5p, miRNA17-5p, miRNA19a-3p, miRNA7-5p (Qiagen, Valencia, CA, USA, Cat.No. CM1HS0064C) according to the provided instructions as follows: 15 min at 95 °C for 1 cycle, 15 s at 94 °C, 30 s at 55 °C, and 34 s at 70 °C for 45 cycles using AB7500 Fast Real-Time PCR system.Threshold cycle data were analyzed using the RT2 Profiler Software (Version 3.4; SABiosciences).To obtain serum levels of studied miRNAs, relative gene expression levels were normalized to those obtained from the amplification of RNU6, and the relative quantity of miRNA was calculated according to 2 -ΔΔCT and log2 -ΔΔCT method. ERK, EGFR and PTEN protein levels..The ERK, EGFR, and PTEN protein concentration was determined using ELISA kits (Sunlong Biotech Co., LTD, Hangzhou, Zhejiang, China).Procedures were carried out following the manufacturer's instructions.The concentration of the markers in plasma samples was calculated by comparing the samples' optical density (OD) to the corresponding plotted standard curves. Statistical analysis. Data management and analysis were done using IBM SPSS Statistics for Windows version 24.0 SPSS Inc., Chicago, IL, USA.The Kolmogorov-Smirnov and Levene's tests assessed the normal distribution and variance homogeneity of data.Numerical data were summarized using median and interquartile range (IQR).Categorical data were summarized as counts and percentages.Comparison of miRNAs between patients and the control group, or between the subgroups of patients, were done using the Mann-Whitney test (when comparing two independent groups) and Kruskal-Wallis (when comparing more than two independent groups).The change in the miRNAs and proteins levels under the effect of 5-FU-based therapy over time was tested for significance with Friedman test when comparing more than two dependent groups, and Wilcoxon matched test when comparing two dependent groups only.Associations between categorical variables were performed using Pearson's chi-square.Spearman's test was used to detect the strength of the correlation between the tested markers.For diagnostic sensitivity and specificity evaluation, receiver operating characteristic (ROC) curves were constructed, and the areas under the curves (AUC) were estimated.The cumulative survival rate, and median levels of OS and EFS, were estimated with Kaplan-Meier method and log-rank test for subgroups` survival curves comparisons.The hazard of disease recurrence and progression was estimated using Cox Proportion Hazard Model for the levels of miRNAs at baseline and after 3 and 6 months of 5-FU-based therapy.P-values are two-sided and were considered significant at 0.05 levels in all analyses except in subgroup analysis; the significance level decreased to 0.01. Institutional review board statement. Informed consent statement. Informed consent was obtained from all subjects involved in the study.S1 shows the demographics of 77 CRC patients accepted to participate in this study.At baseline, the median age of this group of patients was 47 years.The male/female ratio was 1:0.93.Most patients had good performance status according to ECOG classification (83.1%), and with positive family history in 16.9%.Colon and rectal cancers were diagnosed in 45 and 32 patients, respectively.Adenocarcinoma was the dominant pathological type, with histological variants of mucinous, signet ring and neuroendocrine in 16, 11 and 1 patients respectively.Most patients had grade II tumors [59 (67%)].According to TNM classification, T3 was identified in 33 patients (43%), positive nodal disease was presented in 27 patients (35%), and 21 patients had metastatic disease.Surgical intervention was performed on 55 patients.All patients received chemotherapy, 5-FU-based, either in the adjuvant or metastatic settings, alone or in combination with other chemotherapeutic agents.At the end of therapy, 54.5% of patients were in complete remission.Baseline patients were monitored during their journey of treatment with 5-FU-based therapy.60 and 41 of them were sampled after 3 and 6 months, respectively, however the count and % of patients remained insignificant among all subgroups. Paradoxically increased levels of miRNAs at the 6 months of treatment with 5-FU-based therapy, after initial reduction at the 3 months.After 3 months of 5-FU-based therapy, all studied miRNAs were down-regulated in CRC patients (n = 60) compared to their expression levels before the start of therapy.A rebound increase in the expression level of all miRNAs was observed after 6 months of therapy (n = 41).This trend was significant with miRNA223-3p, miRNA19a-3p, and miRNA7-5p (P = 0.002, 0.001, and 0.048, respectively).While the fluctuating trend was insignificant with miRNA20a-5p, miRNA17-5p, and the 3 measured proteins ERK, EGFR, and PTEN (Fig. S1).Spearman correlation did not show significance between all circulating miRNAs and the 3 proteins, just a weak inverse relation was exhibited after 3 months of 5-FU therapy clearly observed in metastatic patients (Fig. S2). Responder patients showed significant trend of reduction then increase in their miRNA levels over the treatment period with 5-FU-based therapy, and that change in miRNA19a-3p level was significant in patients with high basal protein levels.Doing inter-treatment comparison by stratifying the patients according to the treatment regimen either single 5-FU versus combination of 5-FU with oxaliplatin, showed significance in the trend of initial reduction then increase with the levels of miRNA223-3p (P = 0.008), with preferential significance toward the oxaliplatin combination (Table S2).However, we noticed www.nature.com/scientificreports/ a drop in the levels of all microRNAs in the subgroup of patients receiving oxaliplatin either at baseline (while assigning the patients to the appropriate protocol) or during the treatment cycles. Responder patients tend to show a significant trend of change in miRNA levels than non-responders over the treatment period.The associated changes in miRNA levels with 5-FU-based-therapy in responder patients were significant by P-values of 0.011, 0.006, 0.0002, 0.016, and 0.017 compared to the P values documented with non-responders of 0.031, 0.543, 0.166, 0.579, and 0.2104 with miRNA223-3p, miRNA19a-3p, miRNA17-5p, miRNA7-5p, respectively, Fig. 2. Patients with basal levels of EGFR > 6.53, ERK > 0.92, and PTEN > 1407 showed a significant reduction of miRNA19a-3p at 3 months and then increase at the end of treatment compared to patients with basal protein levels equal or lower than the median values of these proteins (P = 0.008, 007 and 0.002 for patients with the high level of proteins compared to P = 0.046, 0.048 and 0.042 for patients with the low level of proteins, Fig. 3). Adding to the above, there were multiple significances in that trend of rebound change in miRNA levels over the treatment period associated certain subgroups of patients.Significant rebound elevation in the expression of miRNA223-3p at the end of 5-FU-based therapy was observed in male patients (P < 0.001), smokers (P = 0.004), with colonic site disease (P = 0.002), patients with T3 tumors (P < 0.001), and those with positive lymph nodes (P = 0.001), Table S3.The re-bound increase of miRNA19a-3p in Table S5 was significant in male patients and those with colonic tumor (P = 0.003 and 0.004, respectively), while miRNA7-5p showed rebound significance in young patients (age ≤ 47 years, P = 0.006, Table S7). Significant high miRNA levels in the non-metastatic subgroup of patients at the 3 months sampling time.Non-metastatic patients have significantly higher miRNA levels after 3 months of 5-FUbased therapy than metastatic patients.After 3 months of 5-FU therapy, all miRNAs, except miRNA7-5p, showed significantly higher levels of expression in non-metastatic patients than metastatic ones (P = 0.0036, 0.0045, www.nature.com/scientificreports/0.0019, 0.007, and 0.0.584 for miRNA223-3p, miRNA20a-5p, miRNA19a-3p, miRNA17-5p and miRNA7-5p, respectively).However, after 6 months of 5-FU-based therapy, the difference in the miRNA expression levels become insignificant between the non-metastatic and metastatic patients, Fig. 4.Separately, by measuring the levels of miRNAs at this time point of 3 months, we found significant association between high miRNA17-5p and miRNA7-5p with low ERK protein level (Table 1).Also, miRNA17-5p level was significantly high with low EGFR, but high PTEN protein levels (Table 1). The level of miRNA19a-3p after 6 months of 5-FU therapy was significant predictor for the hazard of disease recurrence and progression.Tables S8 and S9 show the change in miRNA levels, either by increase or decrease at the end of 5-FU-baed therapy relative to their baseline expression level.Of the 77 patients, 18 males experienced an increase in miRNA223-3p at the end of therapy (P = 0.009).A significant percentage of patients who got increased levels of miRNA20a-5p and miRNA7-5p were diagnosed with T3 tumors, P = 0.006 and 0.002, respectively. After 3 years of follow-up, patients` OS and EFS were estimated in the subgroups of CRC patients with low and high basal expression levels of miRNAs (Supplementary Figs.S3-S12).Generally, better survival rates were associated CRC patients with a high expression level of miRNAs before treatment, but the difference has not reached significance in all miRNAs. Estimating the hazard of disease recurrence and progression for the levels of five miRNAs at baseline and after 3 and 6 months of treatment with 5-FU-based therapy (Fig. 5).A trend of increasing the hazard ratio toward the positive side was observed with miRNA19a-3p and miRNA17-5p, in which LogHR of − 0.14 and 0.13 were estimated for miRNA19a-3p and miRNA17-5p at baseline increased to be 0.01 and 0.28 after 3 months of treatment, respectively.Then at the end of therapy, the logHR increased to 0.66 and 0.5 for same both miRNA19a-3p and miRNA17-5p, respectively.That level of miRNA19a-3p at the end of treatment (6 months) was significantly associated with an increased hazard of disease recurrence and progression (P = 0.031). Discussion CRC progress in a multistep pattern from early adenoma to late-stage adenocarcinoma through the accumulation of many genetic and epigenetic events initiated by APC mutation and chronic inflammation 21 .MiRNAs take place in this pathogenesis pathway either as promoters or inhibitors at all levels of CRC progression and at different expression levels 4,[22][23][24] .Our 77 CRC patients showed significant upregulation of their serum miRNAs compared to matched sex and age-healthy controls.This high baseline expression of miRNAs had diagnostic potential and was associated with better median survival values.Web literature databases are rich with articles identifying the diagnostic and prognostic potentials of miRNAs in cancer in general and in CRC in particular 7,8,14,25-28 .Epigenetic markers like miRNAs retain the advantage of being changeable with treatment, and their levels in the peripheral blood of CRC patients almost match the corresponding levels in the tumors 29 .So, monitoring the pattern of change in their level during the treatment period will produce an honest reflection of disease progression and patient response to therapy 30,31 . The use of 5-FU-based therapy is still the cornerstone in CRC treatment, despite the significant induction of its target, thymidylate synthase (TYMS), after 6 months of its usage, as we recorded before in a previous publication 32 .Digital karyotyping identified amplification of an approximately 100-kb region on chromosome 18p11.32that contains the coding of TYMS in metastatic CRC patients subjected to prior treatment with 5-FUbased therapy 33 .This TYMS amplification is suggested to be controlled with some non-coding RNAs.Recently a review of long non-coding RNA (MALAT-1) highlighted its cooperation with co-factor complex (YAP1/TCF4/ β-catenin) to control specific groups of miRNAs and TYMS expression 34 . Measuring the level of miRNAs with 5-FU-based therapy in this study exhibited a pattern of fluctuated expression from an initial reduction of their levels followed by an increase to the baseline or higher than the baseline level.It is known that 5-FU impacts RNA 35 .It causes splicing defects in intron-containing mRNA, rRNA, and tRNA leading to distorted transcription and post-transcriptional modification of uracil residues in RNA 35 .Tumor suppressor miRNAs, like miR375-3p, increased the chemosensitivity to 5-FU therapy through its direct targeting of TYMS 36 .Another suppressor miRNA, miRNA149, increased the chemosensitivity to 5-FU therapy by targeting Forkhead Box Transcription Factor (FOXM1) 37 .However, the oncogenic miRNA135b and miR-NA182 induced resistance to 5-FU by targeting ST6GALNAC2 via PI3K/AKT pathway 38 .So, the fluctuating changes of miRNAs with 5-FU therapy could be advantageous to monitor the patient's response to therapy and stratify them according to their treatment outcome. Our data showed higher expression of miRNA223-3p level in patients ≤ 47 years than older age patients, matching the general association of reduced miRNA levels with age in healthy individuals.A study on 5221 healthy adults found that most peripheral miRNA levels were down expressed in older individuals 39 .In a microarray analysis of the serum of non-small cell lung cancer patients, miRNA223 was found to be down-expressed compared to healthy individuals, and its high level suppresses the expression of EGFR protein 40 .Such described tumor suppressor effect of miRNA223-3p matches our observation for the 3 months level of miRNA223-3p in which metastatic patients were showed to have lower miRNA223-3p compared to non-metastatic patients. From the miRNAs which showed constant protective indication in its negative levels of logHR at baseline and after 5-FU-based therapy was miRNA20a-5p (logHR = − 0.23, − 0.11, − 1.64, baseline, 3 months, 6 months of 5-FU-based therapy respectively).Also, when its level reduced after 3 months of 5-FU-based therapy, that was associated with metastasis.In research conducted by Dalmasso et al. in 2014, the upregulation of miRNA20a-5p was related to gut microbiota-induced senescence to colonic cancer cells through targeting SUMO Specific Peptidase 1 (SENP1) and inducing the SUMOylation of P53 41 .The rest of the research suggests the pro-tumorigenic effects of miRNA20a leading to the induction of epithelial-mesenchymal transition (EMT) 42 and regulation of TRAIL-induced apoptosis 43 .The treatment with 5-FU caused the fluctuated level of miRNA20a-5p by initial reduction, then induction.In our study, we claim that the 3 months level of miRNAs could correspond to the patients` sensitive phase to 5-FU-based therapy.Then the induction of miRNA levels after 6 months of 5-FUbased therapy could correspond to the phase at which CRC patients showed resistance to 5-FU-based therapy.The knockdown of miRNA20a sensitized CRC cell lines to cisplatin therapy through activation of ROS/ASK1/ JNK pathway 44 , recommending its service as a follow-up response marker by Xiao et al. 45 . In contrast, miRNA19a-3p showed a change in the association of its level before and after 5-FU-based therapy with the hazard of disease recurrence and progression.At baseline, the recorded logHR was = − 0.14 (95% CI = − 0.10-0.28),then after 3 and 6 months of therapy logHR increased to 0.01 and 0.66, respectively.That 6 months level showed significance with the increase in the hazard of disease recurrence and progression, P = 0.031.Interestingly, the reduced level of miRNA19a-3p after three months of therapy, which resemble the responsive phase of treatment, was associated with metastatic patients and those diagnosed at stage IV of the diagnosis, suggesting the initial tumor suppressor effect of that miRNA.However, the pattern of change of miRNA19a-3p level with therapy; of 3 months` reduction then 6 months` induction, showed significance in male patients, those with colonic site disease, and those with high ERK, EGFR, and PTEN protein levels, suggesting the conversion of miRNA19a-3p's function to be oncogenic during the resistant phase of treatment. In a global miRNAs screen (miRNome) conducted by Uhlmann et al., they validate miRNA193a-3p along with miRNA124 and miRNA147 as tumor suppressors that co-target EGFR-driven breast cancer 46 .Similar co-operativity was seen between miRNA193a-3p and miRNA193a-5p to target EGFR in non-small cell lung cancer 47 , and the overexpression of miRNA19a exhibited anti-angiogenesis effects in inverse relation with KRAS expression in CRC cells 48 .The metastasis derived by miRNA19 in CRC is believed to happen because of the inhibition of transglutaminase-2 49 and T-cell intracellular antigen 1 50 , suggesting an oncomeres` functions related to this miRNA.Another mRNA interaction site, testis-specific protein Y-encoded-like 5 (TSPYL5), which is related to miRNA19a as a target of suppression, leading to an increase in the aggressiveness of CRC disease 51 .The suppression of either miRNA19a or TSPYL5 induced cellular apoptosis and accumulation of HT29 cells at the G0/G1 phase of the cell cycle 51 .Finally, miRNA19a was found to be part of the construct of circulating exosomes in CRC patients 52 .Its high expression was associated with disease recurrence and overall poorer prognosis of the disease 52 . Even when the hazard ratio was positive with the miRNA17a-5p level before and after therapy, its low level after 3 months of therapy was associated with metastasis like other miRNAs, which suggests the prevalence of the tumor suppressor function of miRNAs 53 .Most research measured miRNA17 suggesting its oncogenicity.Upregulation of miRNA17 was an indicator of poor prognosis in rectal cancer patients 54 , and in colon cancer cells through targeting Par4 55 , vimentin 56 , and CCNG2 in recurrent head and neck squamous cell carcinoma 57 . CRC patients with low basal expression of PTEN and EGFR showed significantly higher expression of miRNA7-5p, which exhibited a fluctuated hazard ratio with 5-FU-based therapy of positive, then negative, and then to positive again.In concordance, the tumor suppressor function of miRNA7-5p was predominant in the literature.It has a direct effect on EGFR to suppress its level and an indirect effect on phospho-Akt to decrease tumor growth in vitro and in vivo 58 .Also, it was found that XRCC2, a key component in the homologous recombination repair pathway, is a target of miRNA7-5p, assisting in the apoptosis and chemo-sensitization of CRC cells 59 .So, the use of mimetics of miRNA7 was recommended to reduce multidrug resistance in various tumors 60 . Patients with low ERK protein level (≤ 0.92), showed to have higher miRNA17a-5p and miRNA7-5p levels after 3 months of 5-FU-based therapy than the subgroup of patients with low baseline level of ERK.It is demonstrated that ERK has upstream effect on miRNAs through inhibiting pre-miRNA nuclear export 61 .Also, capbinding protein 4EHP is found to be a mutual control of ERK phosphorylation and miRNA145 expression 62 .MiRNAs inhibition to endocytosis repressed embryonic stem cell differentiation through inhibition of ERK signaling 63 . In general, 5-FU therapy-induced changes to the levels of circulating miRNAs associated with a certain subgroup of patients and with disease recurrence and progression.This pattern of change could be used as a marker of patients` response to therapy in studies recruiting larger number of patients.Also, further mechanistic studies should be conducted to explore this fluctuation of the tumor suppressor function of miRNAs over the treatment period.Is it a loss of function under the effect of 5-FU and/or oxaliplatin that could be reversed to sensitize CRC patients to treatment again or not? Conclusions Serum miRNAs are measured in elevated levels in CRC patients before treatment with 5-FU-based therapy. From them, miRNA223-3p showed the best sensitivity, specificity, and AUC diagnostic parameters.Like most global epigenetic events such as DNA methylation 64 , 5-FU-based therapy changes the level of miRNAs over time.A significant trend of initial reduction and then increase was observed in miRNA223-3p, miRNA19a-3p, and miRNA7-5p after the 3 and 6 months of treatment with 5-FU-based therapy, respectively.That trend was significant with miRNA20a-5p and miRNA19a-3p in responder patients, and the change in miRNA19a-3p level over therapy was significant in patients with high basal EGFR, ERK, and PTEN protein levels.MiRNA223a-3p level was the most affected by adding oxaliplatin to the 5-FU treatment regimen.At the 3 months sampling point, we found significant elevations in miRNA223-3p, miRNA20a-5p, miR-NA19a-3p, and miRNA17-5p in non-metastatic patients relative to those diagnosed with metastasis.At the same time of 3 months, we found significant associations between high miRNA17-5p and miRNA7-5p with low ERK protein, and high miRNA17-5p level with low EGFR but high PTEN levels.This time point could be defined as the responsive phase to 5-FU-based therapy before experiencing the acute induction of TYMS as we showed before, and also suggest for the dominant tumor suppressor nature of miRNAs in cancer. By moving to the 6 months treatment with 5-FU-based therapy, we noticed an increase in the hazard ratio of miRNA19a-3p, and its 6 months levels reached significance for predicting the hazard of disease recurrence and progression. https://doi.org/10.1038/s41598-023-41030-7 Figure 3 . Figure 3. Change in miRNA19a-3p level in response to three and six months of 5-FU therapy in patients with low and high basal levels of proteins.(A) EGFR, (B) ERK (C) PTEN.The levels of miRNAs are expressed as log2 −ΔΔCT , and the P-values are displayed in the figures. Figure 5 . Figure 5.The hazard of disease recurrence and progression after 3 years of follow-up for 77, 60 and 41 CRC patients at baseline, after 3 months of 5-FU therapy, after 6 months of 5-FU therapy, respectively.Data presented as median logHR (95% CI) of disease recurrence and progression after 3 years of follow-up.Abbreviations: HR: hazard ratio, CI: confidence interval. The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Human Research Ethics Committee of the Egyptian National Cancer Institute, Cairo, Egypt, Number 00004025, with IRB review Number 2010014019.3. Table 1 . Change in miRNA levels with low and high EGFR, ERK, and PTEN protein levels.Data presented as medians and IQR of miRNAs level.Significant P-values ≤ 0.01 are displayed in bold-italic font.IQR interquartile range.	2023-09-01T06:16:12.188Z	2023-08-30T00:00:00.000	{ "year": 2023, "sha1": "e1ed9e1d9dbf15403ea4adfb9b844cfb8e7523a3", "oa_license": "CCBY", "oa_url": "https://www.nature.com/articles/s41598-023-41030-7.pdf", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "64c36f708adc32d8f00414fa9ba1c5d2c0a8592c", "s2fieldsofstudy": [ "Biology", "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
229438244	pes2o/s2orc	v3-fos-license	Early Warning Signals for Bearing Failure Using Detrended Fluctuation Analysis : Prognostic techniques play a critical role in predicting upcoming faults and failures in machinery or a system by monitoring any deviation in the operation. This paper presents a novel method to analyze multidimensional sensory data and use its characteristics in bearing health prognostics. Firstly, detrended fluctuation analysis (DFA) is exploited to evaluate the long ‐ range correlations in ball bearing vibration data. The results reveal the existence of the crossover phenomenon in vibration data with two scaling exponents at the short ‐ range and long ‐ range scales. Among several data sets, applying the DFA method to vibration signals shows a consistent increase in the short ‐ range scaling exponent toward bearing failure. Finally, Kendall’s tau is used as a ranking coefficient to quantify the trend in the scaling exponent. It was found that the Kendall’s tau coefficient of the vibration scaling exponent could provide an early warning signal (EWS) for bearing failure. Introduction Rolling element bearings (REBs) are one of the essential elements in rotary machinery, such as pumps, compressors, gearboxes, and electric motors. Usually, REBs are susceptible to catastrophic failures when they are operating under abnormal conditions, including heavier loading, corrosion, overheating, etc. As a result of frequent wear and friction, they will experience a gradual degradation and functionality loss during their operation. Evidently, REBs go through different stages (mild, moderate, and serious) of functionality loss before their failure. Prognostics and diagnostics have been active research areas in recent years [1,2]. Diagnostic techniques evaluate the fault severity during operation to help in planning immediate reactive maintenance and avoiding any future failure. Usually, once a bearing fault has been detected, the time remaining before bearing failure is relatively very short. So, reactive maintenance often cannot be completed before the failure. On the other hand, prognostic techniques provide an early warning for bearing breakdown to plan the needed maintenance ahead of time. That permits a continuous and reliable operation of the mechanical system while enabling proactive maintenance. The prognostic techniques can be classified into three categories: physics-based models (PbMs), data-driven model (DDMs), and hybrid models. The PbM approach depends on the mechanical component models to determine its deterioration over time. Moreover, the hybrid model approach is a combination of the PbM and DDM approaches. The main challenge in these two approaches is the difficulty of deriving the mathematical models accurately. On the other hand, the DDM approach extracts the features in the data collected from the mechanical component to determine its proximity to a failure. It is considered a trusted approach in predicting future failures. In general, the DDM approach has two subcategories: the machine-learning approach [3][4][5] and the statistical-based approach. The statistical-based approach is based on several techniques in the time and frequency domains. In the time domain, the root mean square (RMS) is used to detect the starting point of bearing fault from its vibration data [6]. Furthermore, advanced statistical features, such as kurtosis [7], are applied to predict the remaining useful life (RUL) of bearings. On the other hand, novel frequency-domain statistical parameters, such as envelope spectrum and spectral kurtosis (SK) [8,9], are proposed to detect bearing fault features. Medjaher et al. have studied the correlation between nominal and degraded vibration signals to estimate the RUL of bearings [10]. Li et al. have used an improved exponential model and a particle filter (PF) to predict bearing failure [11]. The adoption of one statistical model does not explain the discrepancy between the different degradation rates in the vibration data. Wang et al. have developed a multi-mode PF to track the different stages of bearing performance degradation [12]. Ahmad et al. have proposed adaptive predictive models to understand the trend in the vibration data and predict bearing failure [13]. Singleton et al. have estimated the extended Kalman filter parameters to predict bearing failure under different operating conditions [14]. Soualhi et al. have used Hilbert-Huang transform to monitor ball bearings and track their degradation [15]. The statistical techniques have different approaches to predict bearing failure up to a certain level. However, the used statistical features in these approaches have no clear trend of degradation or monotonic change over time before bearing failure. Moreover, despite signal denoising, the statistical parameters keep fluctuating over time and sometimes go through a sudden spike when bearings near their end of life. From the prognostics point of view, these drawbacks limit the usage of these statistical methods in practical applications. In this paper, using detrended fluctuation analysis (DFA), we propose a new health indicator (scaling exponent) to anticipate bearing degradation. DFA has been used extensively in diagnostic techniques to identify and classify faults. As vibration signals are known to be non-stationary, DFA is one of the reliable methods to study the long-range correlations in these signals. Hu et al. have used DFA and variational model decomposition (VMD) to diagnose and classify the acquired vibration signals at different fault severities [16]. Wang et al. have applied DFA to gear vibrational signals for fault diagnostic purposes and concluded that DFA is capable of identifying gear faults to a reasonable extent [17]. Lin et al. have applied DFA to roller bearing fault data to diagnose the faults and showed that DFA delivers excellent extraction of features of non-stationary and nonlinear data [18]. The main contributions of this paper can be summarized in three main points. Firstly, the DFA method combined with Kendall's tau introduce a novel health indictor of bearing degradation. The proposed indicator is corroborated using vibration signals collected from 17 bearings under different operating conditions. Secondly, the novel combined approach shows evidence of critical slowing down in the bearing vibration signals. The existence of the critical slowing down could provide a universal EWS for the failure of bearings and other components in rotary machinery. Finally, most of the classical health indicators used in the prognostics have shown a fluctuating trend toward bearing failure; however, the proposed method reveals a persistent trend in the scaling exponent of the vibration signals at an early stage of bearing degradation. The rest of the paper is organized as follows. Section 2 provides a description of the testing platform and the vibration data sets. In Section 3, the DFA method and the Kendall's tau coefficient are explained. Section 4 summarizes the results and the proposed approach. Finally, Section 5 presents the conclusions of the paper. Experimental Configuration The experimental platform PRONOSTIA is a test bed used in the IEEE PHM 2012 prognostic challenge to collect real data from ball bearings under accelerated degradation tests [19]. These tests provide real experimental data sets that characterize the degradation of the complete operational life in only a few hours. Under different operating conditions, vibration data are collected from several ball bearings before their failure. In the experiment, the ball bearings (NSK 6804DD) are examined throughout the different tests. As shown in Figure 1, the experimental platform PRONOSTIA consists of three main parts: a rotating module, a loading module, and a measurement module. The three modules are summarized below:  The rotating module: it is composed of an asynchronous motor, a primary shaft, a speed reducer (gearbox), and a secondary shaft with a speed sensor. The asynchronous motor has a rated power of 250 W and an output speed of 2830 rpm. The gearbox consists of two pulleys connected by a timing belt. The main function of the gearbox is maintaining the speed of the secondary shaft at less than 2000 rpm while delivering a rated torque.  The loading module: this has a pneumatic jack that generates a force acting on a lever arm. Then, the lever arm amplifies this force and transmits it as a radial force to the ball bearing under test. Adjusting the lever arm can help create different loading conditions on the bearing.  The measurement module: different types of sensors are used to measure the radial force applied on the bearing, the rotation speed of the secondary shaft holding the bearing, and the torque delivered to the bearing. In addition, two high-frequency accelerometers (DYTRAN 3035B) are mounted perpendicularly to measure the horizontal and the vertical accelerations during the degradation test. In the measurement system, the temperature of the tested bearing is measured via one temperature sensor (PT100). The main components in the experimental platform are shown in Figure 2. In the experiment, the ball bearings (NSK 6804DD) are deep groove, radial, and double rubber sealed. As shown in Figure 2, the dimensions of the bearings are 20 32 7 mm. In addition, all the technical data of the ball bearing are shown in Table 1. Vibration Data Sets In this work, accelerated degradation tests were conducted on 17 ball bearings running in the experimental platform PRONOSTIA. There are several conditions that may cause a failure in a bearing, including high running speed and excessive dynamic loading. In these tests, several loads greater than or equal to the maximum dynamic load (4000 N) were applied to different bearings. All the operating conditions of the bearings are shown in Table 2. Since the experimental data are extracted from the laboratory tests instead of the real-life field, the need for different considerations (de-noising, time synchronous averaging, and order tracking) is minimal. All the bearings were running under a radial load greater than or equal to the maximum dynamic load (4000 N) to accelerate the degradation process. Even though several other factors (lubricant degradation, wear debris, high temperature, etc.) may play a role in bearing degradation, especially at the final stage before the failure, the dynamic overloading of the bearing is the dominant factor in the degradation. During these tests, the vibration data of the bearing were only monitored via the horizontal and vertical accelerometers. In the future, it would be interesting to monitor other factors, such as subsurface stresses and lubricant condition [20][21][22], and study their effect on the vibration data. The vibration data of a rolling bearing can be measured via displacement, velocity, or acceleration sensors. In this paper, the vibration data sets were measured using an accelerometer (DYTRAN 3035B). The sampling rate of the accelerometer was 25.6 kHz and the data were measured for 0.1 s duration every 10 s. The accelerometer is a device capable of measuring bearing vibration using acceleration. The main advantages of the accelerometer are its lightweight and compact design, and wide-range frequency response. However, it is known to have higher sensitivity at high frequencies compared to low frequencies. On the other hand, the anderometer measures bearing vibration using a velocity sensor [23]. It can overcome the accelerometer's disadvantages by working at low signal magnification and measuring at low frequencies up to 30 Hz [24]. DFA Method Long-range correlation has been proven to exist in various systems from different research fields, including physics [25,26], engineering [27], economy [28], medicine [29], etc. The long-range correlated data possess unique statistical characteristics compared to short-range correlated and uncorrelated data. One of the main features of these data is an autocorrelation function that decays hyperbolically with time. The slow decay of the autocorrelation function shows a significant correlation between the data samples as the lag increases. Several methods and techniques have been derived to quantify the level of long-range memory in stationary and non-stationary time series. One of the robust methods for non-stationary time series is the DFA method. In [30], the method was first introduced to study the existence of long-range correlations in DNA nucleotides. DFA evaluates the strength of long-range correlations by calculating the scaling exponent ( ). The steps to compute the scaling exponent ( ) of a time series, , with samples are summarized below: 1. Subtract the mean of the time series and compute the integrated time series, , as shown in Equation (1). ̅ (1) 2. Divide the integrated series into equally spaced boxes, . Inside each box, detrend the series segment by subtracting it from its best linear fit, . Repeat this step for different box sizes ( ). 3. Compute the fluctuation function ( ), as shown in Equation (2), by finding the RMS value of the detrended time series: 1 (2) 4. Plot the fluctuation function ( ) versus the box size (n) on log-log scale. The scaling exponent ( ) represents the slope of the best linear fit. All the steps of the DFA method are shown in Figure 3. The DFA is applied on a simulated series with a scaling exponent equal to 0.6, as shown in Figure 3a. The integrated and detrended time series at a box size of 100 samples are shown in Figure 3b,c. In Figure 3d, the estimated scaling exponent represents the slope of the best linear fit of the fluctuation function versus the box size on a log-log scale. The estimated scaling exponent ( 0.59) is within the accepted margin of error. The scaling exponent ( ) provides a quantitative measure of the degree of long-range correlation in the time series. The long-range correlated time series has a scaling exponent between 0.5 and 1. The increases in these series are likely to be followed by further increases and the decreases are likely to be followed by further decreases. For 0 0.5, the time series is called anticorrelated with a decrease expected to follow the increases, and vice versa. The uncorrelated time series has a scaling exponent of 0.5. Kendall's Tau Coefficient The Kendall's tau ( ) is a rank correlation coefficient that measures the statistical association between ranked variables. The ranked variables should have ordered and numbered data samples. It is a non-parametric test as it does not rely on any assumptions with respect to the distributions of variables. For ). Any pairs with ( ) or ( ) are not considered in calculating the coefficient. The Kendall's tau coefficient can be computed using Equation (3). The Kendall's tau values vary between 1 and 1. The positive value (0 1) of the coefficient indicates a simultaneous increase in the rank of both variables. The negatively signed coefficients refer to a rank increase in one variable and rank decrease in the other one. A Kendall's tau close to zero shows no association between the two variables. Three time series with their calculated Kendall's tau coefficients are shown in Figure 4. Results and Discussion In this paper, a novel approach is proposed to extract the hidden statistical characteristics of the vibration measurements of bearings. As part of the run-to-failure experiment on several bearings, several data sets of horizontal and vertical vibrations signals were collected. DFA can help in understanding the statistical and fractal characteristics of the vibration signals by calculating their scaling exponents ( ). It is argued that the dynamics of the scaling exponent could provide an early warning for bearing failure. Evidently, bearing vibrations are non-stationary under normal and faulty operating conditions [31]. The non-stationarity in the time series manifests itself in a time-varying mean and variance over different segments of the series. Since DFA is one of the robust methods to evaluate the long-range correlation in non-stationary data, it was applied to two segments of the vibration signals (horizontal and vertical), as shown in Figure 5. The fluctuation functions in Figure 5b,d have two linear regions with different slopes at the short-range (10-50 samples) and long-range (100-500 samples) scales. This behavior of the fluctuation function is known as the crossover phenomenon [32]. It may result from the sinusoidal trends in the vibrations signals. Moreover, these trends could be the culprit of the proven non-stationarity in the vibration measurements. In the horizontal vibration signal, the scaling exponents at the short range ( ) and long range ( ) were 0.72 and 0.42, respectively. The calculated scaling exponents show that the horizontal signal possessed long-range correlation at the short range, and it was anti-correlated at the long-range scale. On the other hand, the vertical vibration signal had scaling exponents of 0.41 and 0.27 at the shortrange and long-range scales, respectively. This signal had anti-correlated behavior on both scales. Even though all the scaling exponents were below 1, similar to the stationary time series ( 1), it is evident that the vibrations signals were non-stationary as a result of the existing trends and the clear crossover phenomenon. After examining the long-range correlations in the two segments (2560 samples each) of the horizontal and vertical vibration signals, it is important to study the long-range correlations in the complete run-to-failure vibration signals. It would be interesting to see how the evolution of different stages of bearing failure was reflected in scaling exponent dynamics. To that end, the scaling exponents were computed in the complete horizontal and vertical run-to-failure vibration signals from data set 1_1, as shown in Figures 6a and 7a. Each signal consisted of 2800 segments (2560 samples) and total duration of 7.77 h (28,000 s). It was easier to track the changes in the scaling exponent by dividing each vibration signal into 6 smaller time series consisting of 500 segments (5000 s) each, except the last one, which had 300 segments. Then, the scaling exponents ( and ) were estimated for each segment inside these series. The average fluctuation functions inside each of these 6 smaller time series are shown in Figures 6b and Figure 7b. In both figures, the results show a persistent increase in the average scaling exponents ( ) at the short range by moving from the first time series (pink) to the last one (green). Although the average scaling exponents ( ) show a decreasing trend, it will be shown later that the scaling exponent ( ) behaves differently in other data sets. For the horizontal and vertical vibration signals, the means and standard deviations of the scaling exponents ( and ) are shown in Table 3. In the horizontal vibration signal, the average scaling exponent ( ) changed from 0.43 (anticorrelated) at the beginning of the experiment to 0.98 (long-range correlated) before the bearing failure. The change in the scaling exponent ( ) at the short range could be related to a shift in the mechanical system dynamics as part of the run-to-failure experiment. Similarly, in the vertical vibration signal, the average scaling exponent ( ) had a gradual increase with time, starting from 0.18 in the first time series and increasing up to 0.47 in the last time series. In Figures 6c,d and 7c,d, the plots show all the scaling exponents ( and ) over time toward bearing failure for the horizontal and vertical vibration signals. The increase in the scaling exponent ( ) is clear in scaling exponent samples and its smoothed version (red-color line). Using the Kendall's tau ( ) coefficient, the trend in the scaling exponent ( ) can be quantified by analyzing the statistical association between the scaling exponent and time. The calculated Kendall's tau coefficients of the scaling exponent ( ) were 0.67 and 0.74 in the horizontal and vertical vibration signals, respectively. Based on the significant values of the Kendall's tau before bearing failure, it appears that this rise was related to accelerated failure conditions and the proximity to bearing failure. Tracking the Kendall's tau over time could be an efficient tool in the preventative maintenance field. Usually, complex and dynamic systems undergo hidden changes as they approach a major critical transition. These changes are known as the critical slowing down phenomenon where the system requires more time to recover from any small disturbance [33]. The slower recovery can be examined in the system state by calculating the lag-1 autocorrelation or the scaling exponent. As the system is approaching the critical transition, it shows a persistent and gradual increase in the autocorrelation and the scaling exponent [34,35]. The critical slowing down could be the culprit of the change in the scaling exponent ( ) before bearing failure. Therefore, the Kendall's tau of the scaling exponent at the short-range scale could provide a reliable measure of the proximity to bearing failure. To validate the proposed method, DFA was applied on several vibration signals from 17 data sets to calculate the scaling exponents over all the segments. Then, the trend in the scaling exponent for each vibration signal was examined by calculating the Kendall's tau coefficient. Table 4 illustrates the Kendall's tau of the scaling exponents ( and ) for the 17 data sets with a total of 34 horizontal and vertical vibration signals. In addition, the plots of the scaling exponent ( ) for the data sets are shown in Figures 8-10. The Practically, a large portion of machine breakdowns originates from bearing failures; therefore, it is of great importance to monitor bearing condition. The main goals of bearing monitoring are decreasing maintenance cost and maintaining a smooth and continuous running of the system. As part of predictive maintenance, the proposed method can be integrated as a new detection technique in bearing condition monitoring through its vibration measurements. Moreover, ball bearings are mostly used in low-and medium-speed applications, such as industry, automotive, agriculture, etc. So, the proposed method has the potential to be used in different rotating equipment in the industrial sector (electric motors, pumps, compressors, assembly lines, and elevators) and the automotive sector (engines, steering wheel, driveshaft, driveline, and gearboxes). Conclusions Using DFA, it was shown that long-range correlations exist in bearing vibration signals before their failure. DFA is a reliable method to quantify the long-range correlations in long time series with a linear fluctuation function at different time scales. In several data sets of bearing vibration signals, the crossover phenomenon emerged with two scaling exponents at the short and long ranges. The estimated short-range scaling exponents showed a persistent and gradual increase toward failure among the different data sets. Then, the Kendall's tau coefficient was used to measure the trend in the scaling exponents. The results call for adoption of the Kendall's tau of the scaling exponent as a novel health indicator of the bearings. Furthermore, the proposed method has the potential to be integrated as a new detection technique of bearing failure in condition monitoring systems. In the future, the effect of other abnormal conditions (lubricant degradation, wear debris, high temperature, etc.) on the vibration signal dynamics will be examined. In addition, it would be interesting to investigate the vibration signals of other elements in mechanical systems using the proposed approach.	2020-12-03T09:04:54.800Z	2020-11-27T00:00:00.000	{ "year": 2020, "sha1": "1c337fc666cbf2e4ce506d90fbd818bd97aa16bb", "oa_license": "CCBY", "oa_url": "https://www.mdpi.com/2076-3417/10/23/8489/pdf", "oa_status": "GOLD", "pdf_src": "ScienceParsePlus", "pdf_hash": "4edf55e578b85996cd7c6ee00606be0746fb11d3", "s2fieldsofstudy": [ "Engineering" ], "extfieldsofstudy": [ "Geology" ] }
255990613	pes2o/s2orc	v3-fos-license	Inhibition of DEC2 is necessary for exiting cell dormancy in salivary adenoid cystic carcinoma Patients were prone to have poor prognosis once dormant tumor cells being reactivated. However, the molecular mechanism of tumor cell dormancy remains poorly understood. This study aimed to investigate the function of DEC2 in the dormancy of salivary adenoid cystic carcinoma (SACC) in vitro and vivo. The function of DEC2 in tumor dormancy of SACC was investigated in nude mice by establishing primary and lung metastasis model. Meanwhile, the interaction between hypoxia and SACC dormancy and the role of DEC2 were demonstrated through CoCl2 induced hypoxia–mimicking microenvironments. Furthermore, the expression of DEC2 was detected by immunohistochemical staining in primary SACC samples with and without recurrence. In the primary SACC, DEC2 overexpression inhibited cell proliferation, increased cell population arrested in G0/G1 phase, and participated in dormancy regulation, which limited tumor growth. Intriguingly, in the model of lung metastasis, the level of DEC2 was reduced significantly and resulted in dormancy exit and growth resumption of SACC cells. Then, we found that DEC2 may associate with hypoxia in contributing to tumor dormancy, which might provide a possible cue to explain the different roles of DEC2 in primary and metastasis lesions. And overexpression of DEC2 induced dormancy and promoted migration and invasion through activating EMT program. Finally, DEC2 positive expression was shown to be significantly correlated with recurrence and dormancy of SACC patients. These findings provide a novel insight into the role of DEC2 gene in tumor dormancy and metastasis. Background Salivary adenoid cystic carcinoma (SACC) is one of the most common malignancies characterized with slow growth, but high incidence of local recurrence and metastasis [1,2]. Although the 5-year disease-free survival rate generally reaches 90%, it drops to 10% after 20 years owing to potential local recurrence and hematogenous distant metastases [3]. Therefore, it is necessary to elucidate the molecular mechanisms of recurrence and metastasis, which will provide molecular targets for the treatment of SACC patients. Tumor dormancy was a period in tumor progression in which residual disease existed but remained asymptomatic clinically for years or even decades [4]. It may appear during the formation of primary tumor, after dissemination of primary tumor cells or in the micrometastasis [5,6]. Tumor cell dormancy is defined at cellular level which is characterized with cells that are not divided and arrested in G0/G1 cell phase [7]. The dormant tumor cells could escape immune surveillance and chemo-radiotherapies, remaining undetectable for long periods [6]. Dormant tumor cells are emerging as a critical cause for recurrence and metastasis once they escape this state [8]. However, how tumor cells enter into dormancy and what processes govern their exit in human cancers including SACC are still absent and remain unclear. Differentiated embryonic chondrocyte gene 2 (DEC2, also known as BHLHE41/BHLHB3/Sharp1) is one of the basic helix-loop-helix (bHLH) transcription regulators, which has been demonstrated to play important roles in regulating circadian rhythms, cell proliferation, hypoxia reaction, immune responses as well as malignant tumor progression [9][10][11][12]. Recent evidence demonstrated that inhibition of DEC2 and nuclear receptor subfamily 2 group F member 1(NR2F1) resulted in increased growth of breast cancer cells [13] and downregulation of DEC2 interrupted tumor cell dormancy [14], indicating DEC2 might involve in tumor dormancy. And TGF-β2 activated a [ERK/p38] low signaling ratio, which resulted in induction of DEC2 and dormancy of malignant disseminated tumor cells (DTCs) in the bone marrow of head and neck squamous cell carcinoma (HNSCC). However, the role and its molecular mechanism of DEC2 in the dormancy and malignant progression of SACC remain unclear. Here, we found that DEC2 induced tumor dormancy of the primary SACC and in the model of lung metastasis, DEC2 positive tumor cells manifested enhanced migration and invasion and formed more metastases and the level of DEC2 was reduced significantly with the resumption of cell proliferation. Then, DEC2 may associate with HIF1α in contributing to tumor dormancy, which might provide a possible cue to explain the different roles of DEC2 in primary and metastasis lesions. Our findings demonstrated that overexpression of DEC2 contributed to the dormancy of tumor and low expression reawakened cell dormancy of SACC, which may provide important implications for the therapy of patients. Xenografts Lentivirus-transfected SACC-83 cells were subcutaneously injected s.c. (5 × 10 6 cells/200 μL PBS/mouse) into the flank of 6-week-old nude female mice (Laboratory Animal Center of Sichuan University, Chengdu, China) and examined every 3 to 5 days for tumor appearance. Tumor growth was then measured once a week until 35 days after inoculation by determining the tumor volumes using caliper measurements. Lentivirus-transfected SACC-83 cells (5 × 10 6 cells/200 μL PBS/mouse) were injected via tail vein of nude female mice to establish the model of lung metastasis. The nude mice were weighed weekly. Lung tissues were excised after 4 weeks for immumohistochemical staining. All animal experiments were approved by the Institutional Ethics Committee of the West China Medical Center, Sichuan University, China. Immunohistochemistry Paraffin-embedded sections were cut into 4um and deparaffinized in xylene and rehydrated, and endogenous peroxidase was blocked with 3%H 2 O 2 . Antigen retrieval was accomplished by 0.01 mol/L citrate buffer solution (pH 6.0) in a 700 W microwave oven for 15 min. After incubation with 5% normal goat serum for 20 min, the slides were exposed overnight at 4°C to the rabbit anti-DEC2 (1:150; Proteintech), rabbit anti-Ki-67(1:800; Proteintech), rabbit anti-NR2F1 (1:200; Proteintech). Sections were then incubated with biotinylated goat antirabbit IgG (Zhongshan Goldenbridge Biotechnology) for 1 h, and streptavidin-peroxidase for 30 min. The 0.02% diaminobenzidine tetrahydrochloride was used as a chromogen, and the slides were counterstained with hematoxylin. The percentage of positive cells was estimated using an image analysis system (Leica). Cell lines and cell culture Two SACC cell lines, SACC-83 and SACC-LM, were obtained from the State Key Laboratory of Oral Disease, Sichuan University. Cells were cultured in RPMI 1640 medium (Gibco) supplemented with 10% heatinactivated FCS (Hyclone), 2 mmol/L L-glutamine, 25 mmol/L HEPES, and 100 units/mL penicillin and streptomycin at 37°Cin 5% CO 2 . For hypoxic treatment, cells were exposed to 0.1% O 2 with 5% CO 2 at 37°C with hypoxia chamber. Cloning, Lentivirus preparation, and plasmids The targeted cDNA of DEC2 was cloned into the pEZ-Lv201 and constructed into EX-W0115-Lv201 plasmid and negative control plasmid EX-NEG-LV201. After sequencing verification, it was packaged into virus. Human DEC2, BC_025968, total 1449 bp. P/Puro plasmid vector and transfected into cells by Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. The stable transfected cells were selected with puromycin. Transient siRNA knockdowns SiRNAs targeting DEC2 and Slug and their control siR-NAs were purchased from Genechem. The target sequences were as following: DEC2 siRNA-1: CUCCCU AUAUCCCAAUGGATT, UCCAUUGGGAUAUAGG GAGTT; DEC2 siRNA-2: CGAGGAAGAACUAUGAAC ATT, UGUUCAUAGUUCUUCCUCGTT; Quantitative real-time RT-PCR Total RNA was extracted from cells using the trizol (Invitrogen, Carlsbad, CA) and were quantified with the NanoDrop ND-1000 Spectrophotometer (Thermo Scien-tificInc., Waltham, MA). PCR amplification of the cDNA template was done using Thunderbird SYBR qPCR mix (TOYOBO) on ABI PRISM 7300 sequence detection system (Applied Biosystems) according to the manufacturer's protocol. The resulting cDNA was diluted and used as a template for Quantitative real-time PCR using LightCycler (Roche Diagnostics GmbH, Mannheim, Germany). β-actin was used as the housekeeping gene to normalize the target gene expression. The sequences of PCR primers were showed in Supplementary Table S1. Immunofluorescence staining SACC cells were cultured in 12-well cell culture plates. Upon reaching 70% confluence, cells were washed in cold PBS and fixed in 4% paraformaldehyde for 30 min, permeabilized in 0.25% Triton X-100 in PBS for 15 min, and blocked with 1% bovine serum albumin prepared in PBS for 30 min. Lastly, cells were incubated overnight with mouse anti-Ki-67 (1:100 dilution), and then incubated with FITC or TRITC-conjugated goat anti-mouse IgG (1:500; Zhongshan Goldenbridge Biotechnology) at 37°C for 1 h. Cells were visualized using the Olympus Fluoview confocal microscope (Tokyo, Japan), and fluorescence images were taken. Wound healing assay SACC cells were seeded and cultured in 6-well plates. Upon reaching 80% confluence, cells were wounded by scratching with a pipette tip and incubated with medium containing no FBS for 24 h. Then, they were photographed under phase-contrast microscopy. In vitro cell invasion assay Invasion of cells was assessed using Matrigel-coated membrane (24-well insert, pore size, 8 μm; BD Biosciences). About 5 × 10 4 cells were plated in the top chamber in serum free medium, and medium with serum was used as a chemo-attractant in the lower chamber. After 48 h of incubation, cells remaining on the top chamber were removed using a cotton swab. Traversed cells on the lower surface of the membrane were fixed in 4% paraformaldehyde and stained with 1% Crystal Violet; five fields per filter were counted. Cell proliferation assays The cell proliferation was assessed by Cell Counting Kit-8 (CCK-8, Dojindo) assay. Cells were seeded in 96-well plates in triplicate and the proliferation assay was performed after 24 h incubation. 10ul of CCK-8 solution was added into per well and the absorbance reading was measured at 450 nm after 30 min of incubation at 37°C. The above experiments were repeated the next 4 days. Apoptosis detection by FCM Cell apoptosis was performed by combined application of Annexin V-FITC and propidium iodide (BD Biosciences Clontech, USA). Cells were washed by PBS and adjusted to1 × 10 6 cells/ml with 4°C PBS. One hundred microliter of suspensions was added to each labeled Falcon tube (12 mm × 75 mm, polystyrene round-bottom); 10 μl of Annexin V-FITC and 10 μl propidium iodide (20 μg/ml) were added into the above labeled tube, which was incubated for 30 min at room temperature in the dark environment; and then 400 μl PBS binding buffer was added to each tube which was analyzed using FCM analysis (BD Biosciences Clontech, USA). Glucose consumption test Glucose consumption was detected using a glucose assay kit (Nobio, China). About 1 × 10 5 cells/well was seeded in 6-well plates. The test was performed according to the manufacturer's protocol. The experiments were performed at least three times. Cell senescence detection Senescent cells were measured using a senescence βgalactosidase staining kit (Beyotime, China). Cells were seeded in 6-well plates (1 × 10 5 cells/well). The staining was performed according to the manufacturer's instructions. The cells were then observed under an Olympus BX51 microscope and were analyzed using ImageJ software. Clinical samples collection and study The cohort was assembled from 70 patients who were histologically diagnosed with SACC and underwent resection of their tumors at West China Hospital of Stomatology, Sichuan University, between 2005 and 2015. Exclusion criteria included preoperative chemotherapy, hormone therapy or radiotherapy. All samples of SACC were collected at the time of surgery. All the paraffinembedded sections were confirmed histologically with blind method by two pathologists. The protocol of the study was approved by the Institutional Ethics Committee of the West China Medical Center, Sichuan University, China. The pathologic characteristics of the tumors and clinical data of the patients were summarized in Table 1. TUNEL assay Terminal deoxynucleotidyl transferase-mediated dUTP nick and labeling (TUNEL) Kit (KeyGEN) was to test cell apoptosis. Negative was graded as 0 to 10% within 4-6 microscopic fields at × 400 magnification; and the positive was graded as more than 10% as well. Statistical analyses All statistical analyses were conducted using GraphPad Prism. Data was plotted with GraphPad Prism software. A value of P < 0.05 was considered statistically significant. All experiments were performed independently at least three times. High expression of DEC2 induced tumor dormancy of the primary SACC in nude mice Our previous findings have demonstrated that atRA treatment can be used to induce dormancy in SACC cells, and its effects on cancer dormancy resulted directly from modulation of NR2F1 [15,16]. To further examine the molecular mechanism of tumor dormancy of SACC, atRA treated SACC-83 and the control cells were injected subcutaneously into nude mice, respectively. As shown in Fig. 1a, tumor volume in atRA treated group was much smaller than the control group after subcutaneously injecting. Interestingly, immunohistochemistrical staining showed the positive expression of DEC2 in the xenograft of atRA treated group and the negative expression in the control group (Fig. 1b). The data showed that upregulation of DEC2 involved in atRA treatment contributing to tumor dormancy of the xenograft model. We subsequently investigated the expression of DEC2 in the dormant state induced by atRA in SACC cells by qRT-PCR and Western blot. The results demonstrated that DEC2 was also involved in the dormant state induced by atRA besides NR2F1 (Fig. 1c). 35 4 Immunohistochemistry results showed that the nuclear staining of DEC2 was detected in 17 of 70 and it was significantly associated with tumor site, histological subtype, recurrence and metastasis. Neither was related to age, gender or T stage To address the different roles of DEC2 in the primary lesion and metastasis location of SACC nude mice, we stably overexpressed DEC2 in SACC-83 cells by using lentivirus infection, as confirmed by real-time PCR, WB and immunofluorescence (Fig. 1d). DEC2 overexpression and vector SACC-83 cells were injected respectively into nude mice. As shown in Fig. 1e, DEC2 overexpressed groups exerted a slower tumor growth than the vector groups. And tumor volume in DEC2-overexpressed group of the xenograft nude mice model was smaller than the vector in 5 weeks after subcutaneously injecting SACC-83 cells. This indicated that DEC2 might induce dormancy and inhibit tumor growth of the primary SACC. Immunohistochemistrical staining showed the positive expression of DEC2 and weak positive expression of Ki-67 in the xenograft of DEC2-overexpressed group. And the vector group displayed high expression of Ki-67 (Fig. 1f). These results suggested that DEC2 overexpression contributed to tumor dormancy of the primary SACC. Low level of DEC2 exited dormancy to promote lung metastasis of SACC in nude mice We next established lung metastasis model in nude mice by injecting DEC2 overexpression and vector SACC-83 cells into tail vein. As shown in Fig. 2a, the weight of nude mice in DEC2-overexpressed group was significantly lower than the vector group. Only 20% (1/5) of nude mice implanted with vector cells produced lung metastasis, and 100% (5/5) of the mice injected by DEC2 overexpressed cells developed lung metastases (Fig. 2b), indicating that DEC2 expression promoted metastasis of SACC to the target organ of lung tissue. HE staining confirmed the tumor metastatic lumps in lung tissues of DEC2 overexpressed group (Fig. 2c). Then how did the dormant SACC cells with DEC2 overexpression form c Fold change in mRNA expression of NR2F1, P27, P53, DEC2, CDK4, P38 and ERK, and the protein expression of DEC2 after atRA treatment compared to control group. d DEC2 over expression was confirmed by QPCR, WB and immunofluorescence in SACC-83 cells. e and f SACC-83 cells with or without DEC2 overexpression were injected into nude mice subcutaneously to observe tumor development. Tumor volumes were determined (e) and the expression of DEC2, HIF1α and Ki-67 were detected by immunohistochemical staining (f). Error bars represent the mean ± SD of triplicate experiments. (P < 0 .05, P < 0.01, *P < 0 .001) more metastases in lung of nude mice? We hypothesized that these dormant tumor cells were reactivated into proliferative ones by the oxygen-rich environment in the lung tissue accompanied by downregulation of DEC2. So we further explored the expression of DEC2 by immunohistochemical staining and found the positive expression of Ki-67, while the negative expression of DEC2 in the metastases (Fig. 2d). This confirmed that DEC2 promoted lung metastasis of SACC in nude mice, and these dormant tumor cells resumed proliferation with downregulation of DEC2, leading to the occurrence of metastases. High expression of DEC2 induced dormancy and promoted migratory and invasive abilities of SACC cells We further examined the function of DEC2 in vitro and found that cell proliferation was inhibited after DEC2 overexpression according to immunofluorescence staining of Ki-67 and CCK-8 assays (Fig. 3a and b).This change in proliferation was confirmed by metabolic capacity, which showed that glucose consumption of DEC2 overexpressed SACC-83 cells was much lower than the control group (Fig. 3c). Moreover, we tested cell cycle by flow cytometry analysis and found that DEC2 overexpression increased cell population arrested in G0/G1 phase in SACC-83 cells (Fig. 3d). Meanwhile, no significant difference of cell apoptosis and senescence were observed between DEC2 overexpressed SACC-83 cells and the control group ( Fig. 3e and f). And we got similar results in SACC-LM cells (Additional file 1: Fig. S1A-C). We next silenced DEC2 using siRNA in SACC cells and the silence efficiency of DEC2 was verified by mRNA and protein expression patterns in Additional file 1: Fig. S1D. And it was demonstrated that knockdown of DEC2 promoted tumor cells proliferation and glucose consumption (Additional file 1: Fig. S1E-F). The above results indicated that overexpression of DEC2 could inhibit proliferation and induce dormancy of SACC cells. Then, we detected the effect of DEC2 on the migration and invasion of SACC cells by wound-healing and transwell invasion assays. As shown in Fig. 3g, DEC2 overexpressed SACC cells had higher migratory and invasive behaviors than SACC vector. And knockdown of DEC2 in SACC cells prominently impaired the migration and invasion ability of tumor cells (Fig. 3h). The results suggested that DEC2 overexpression dramatically increased the migratory and invasive behaviors of SACC cells. We next explored the expression of HIF1α and P27 in DEC2 overexpressed SACC-83 cells and found that the mRNA levels of HIF1α and P27 were all increased. In contrast, they were both decreased in DEC2 silenced SACC-83 cells (Fig. 3i). These results indicated that DEC2 may associate with HIF1α in contributing to tumor dormancy, which might provide a possible cue to explain the different roles of DEC2 in primary and metastasis lesions. High expression of DEC2 was necessary for CoCl 2induced dormancy Many studies have displayed the close relationship between hypoxia and tumor dormancy [17][18][19], and here we have demonstrated that DEC2 was positively correlated with HIF1α expression in SACC cells. Hence, we proposed that due to the abundance of oxygen in the lung tissue, DEC2-overexpressed dormant cells transferred here are reactivated to form metastases. Here, we first found the proliferation of SACC-83 cells was suppressed in a dose-dependent manner up to 500 μM of CoCl 2 (Fig. 4a). And the ratio of positive Ki-67 was markedly reduced (Fig. 4b) and cell population arrested in G0/G1 phase was increased in CoCl 2 treated SACC-83 cells (Fig. 4c), indicating that a population of these hypoxic cells may enter into dormant state. Then, cell growth analysis suggested that while SACC-83 cells proliferation were suppressed during 7-day CoCl 2 treatment, they recovered growth after removal of CoCl 2 (Fig. 4d), demonstrating that growth-inhibited SACC-83 cells under CoCl 2 treatment were dormant instead of , metabolism (c) and cell cycle (d) between DEC2 over-expressed and vector SACC-83 cells was detected by immunofluorescence, CCK-8 assay, glucose consumption and Flow cytometry, respectively. e and f Cell apoptosis test (e) and senescent analysis (f) in both DEC2 over-expressed and vector cells. g and h Migration and invasion abilities of DEC2 over-expressed/vector SACC-83 cells (g) and SACC-83/SACC-83 + siDEC2 cells (h) were detected by transwell and scratch assays, respectively. i The mRNA levels of HIF1α and P27 in DEC2 overexpression and silence SACC-83 cells were tested by QPCR. Error bars represent the mean ± SD of triplicate experiments. (P < 0.05, P < 0.01, P < 0 .001) senescent and apoptotic. Cell growth curves were similar between 0.1% hypoxia condition and CoCl 2 treatment, indicating that tumor dormancy induced by CoCl 2 is similar to true hypoxia. And cell proliferation was suppressed during 7-day 0.1% hypoxia condition, then resumed growth after changing to normoxia condition (Fig. 4e). Similar results were observed in SACC-LM cells (Additional file 1: Fig. S2A-F). This indicated that hypoxia under certain concentration may push cancer cells to enter into dormant state. Furthermore, we found that the expression of DEC2, NR2F1, P53 and P27, hypoxia gene HIF1α and EMT transcription factor slug were all upregulated in SACC-83 cells treated by CoCl 2 (Fig. 4f and g). Interestingly, DEC2 expression presented dose-dependent increase up to 500uM of CoCl 2 , which indicated that DEC2 contributed to hypoxia-induced dormant process (Fig. 4h). The same results were displayed in SACC-LM cells (Additional file 1: Fig. S2G and H). We further silenced DEC2 using siRNA in CoCl 2 treated SACC-83 cells and found that suppression of DEC2 resulted in reversible proliferation of dormant SACC-83 cells under CoCl 2 treatment (Fig. 4i). Consistent with the reversible growth, cell population arrested in G0/G1 phase was reduced after DEC2 silencing during CoCl 2 treatment in SACC cells (Fig. 4j). Furthermore, the expression of Ki-67 was also up-regulated (Additional file 1: Fig. S2I). Knockdown of DEC2 impeded cell invasiveness and d Cell growth analysis of SACC-83 cells treated with 500 μM CoCl 2 for 7 days (from day 4 to day 10) and then recovered in normal media. e Cell growth analysis of SACC-83 cells treated with 0.1% O 2 for 7 days (from day 4 to day 10) and then recovered into normoxia environment. f The expression of DEC2, HIF1α, P27, P53, NR2F1, CDK4, P38, ERK, Slug, Snail, Twist, Zeb, E-cadherin and N-cadherin were determined by QPCR in SACC-83/LM cells treated by CoCl 2 and the control group. g The protein levels of DEC2, HIF1α and Slug in CoCl 2 treatment and control SACC-83 cells. h DEC2 expression at different concentrations of CoCl 2 in SACC-83. i and j Proliferation capacity and cell cycle were determined by CCK-8 assay (i) and flow cytometry (j). k Migration and invasion abilities of CoCl 2 treatment and CoCl 2 + siDEC2 SACC-83 cells were detected by Transwell and scratch assays, respectively. l The mRNA levels of Slug and HIF-1α in CoCl 2 or CoCl 2 + siDEC2 treatment SACC-83 cells. Error bars represent the mean ± SD of triplicate experiments (P < 0 .05, P < 0.01, P < 0 .001) metastatic potential increased by CoCl 2 treatment (Fig. 4k), accompanying downregulation of Slug and HIF-1α (Fig. 4l). Thus, our loss-of-function study showed that the suppression of DEC2 could inhibit cell dormancy induced by CoCl 2 treatment, indicating that different hypoxia states in primary and metastasis lesion may regulate DEC2-induced dormancy to lead to keep or reawaken cell dormancy state. Slug-induced EMT was involved in DEC2-induced dormancy Recently, it was proposed that EMT positive tumor cells always characterized by low proliferation rate or quiescence and EMT program may have a potential role during tumor dormancy [20,21]. We next analyzed the expressions of dormant related genes including NR2F1, P27, P53, HIF1α, CDK4, P38, ERK, E-cadherin, Ncadherin and EMT transcription factors (Snail/Slug/ Twist/Zeb) in SACC-83 and SACC-LM cells with DEC2-overexpressed vector. The results showed that HIF1α, P27, P53, Snail, Slug and N-cadherin levels were significantly boosted after DEC2 overexpression, whereas CDK4 and E-cadherin were downregulated. But the protein expression of Snail did not change significantly. Moreover, the ERK/p38 signaling ratio was also decreased. NR2F1, Twist and Zeb were not significantly affected after DEC2 overexpression ( Fig. 5a and b). We also silenced DEC2 in SACC-83 and SACC-LM cells and the opposite trend was observed (Additional file 1: Fig. S3). This showed that EMT had involved in the dormancy of SACC cells. In order to further examine the role of EMT program during DEC2 induced tumor dormancy, we silenced Slug using siRNA in DEC2 overexpressed SACC-83 cells. The SACC-83 vector cells had an epithelial morphology, whereas DEC2-overexpression group exhibited a more spindle-cell morphology with tentacles. And suppression of Slug in DEC2 overexpressed cells reversed their c Cell morphologies were changed after Slug knockdown in DEC2 overexpressed SACC-83 cells. d and e Cell proliferation (d) and metabolism (e) were determined by CCK-8 assay and glucose consumption in DEC2 hi /DEC2 hi + siSlug/vector SACC-83 cells. f Transwell chamber and scratch assays for DEC2 hi or DEC2 hi + siSlug SACC-83 cells. Error bars represent the mean ± SD of triplicate experiments (P < 0 .05, P < 0.01, P < 0 .001) spindle changes (Fig. 5c), proliferation capacity (Fig. 5d) and glucose consumption (Fig. 5e), indicating their dormant state was partial reversed. Additionally, downregulated Slug expression suppressed their migration and invasion significantly (Fig. 5f). These results showed that DEC2 may promote SACC dormancy by regulating EMT directly through Slug and thus enhance their migration and invasion ability. High expression of DEC2 associated with tumor dormancy in the primary SACC patients To investigate the significance of DEC2 in human SACC cases, we used a cohort of 70 primary SACC samples obtained from clinical patients and 10 normal salivary gland samples. Immunohistochemistry results showed that the nuclear staining of DEC2 was detected in 17 of 70 (24%) in SACC and 7 out of 10 (70%) in normal salivary gland samples, respectively (Fig. 6a). Then, we focused on the cases with DEC2 positive expression, which occupied the small part of SACC patients. We found that 14 out of DEC2 positive specimens had low positive expression of Ki-67 (5-10%) (Fig. 6b, P < 0.01) and the negative of TUNEL (Fig. 6c, P < 0.05) and senescence tests (Fig. 6d, P > 0.05). In addition, SACC specimens were simultaneously immunostained for NR2F1, relevant to tumor dormancy. We found that DEC2 expression was significantly associated with NR2F1 (Fig. 6e, P < 0.001). This indicated that DEC2 positive tumor cells were neither proliferative nor apoptotic, which was consistent with the dormant phenotype of SACC patients. We next investigated the expression of DEC2 and clinicopathologic parameters of SACC patients. The results showed that DEC2 was significantly associated with tumor site (P = 0.0438), histological subtype (P = 0.009), recurrence (P = 0.005) and metastasis (P = 0.002, Table 1). But there was no significant association of DEC2 positive expression with age, gender and T stage. These suggested that positive expression of DEC2 in a few SACC cases was related to recurrence and metastasis of Fig. 6 High expression of DEC2 associated with tumor dormancy in the primary SACC patients. a and b Expression levels of DEC2 and Ki-67 were analyzed by Immunohistochemical staining. c and d TUNEL and senescence tests of SACC between recurrence and no recurrence were shown. e Expression of NR2F1 was detected by Immunohistochemical staining. Student's paired t test was used to analyze the differences between the metastasis and no metastasis specimens (P < 0 .05, P < 0.01, *P < 0 .001). s Schematic figure illustrated dormancy during tumor progression and the functions of DEC2 in that (Round cell: Proliferative tumor cell, main DEC2 low ; Fusiform cells: Dormant tumor cell, main DEC2 hi ) [22][23][24] SACC patients, although there was negative expression of DEC2 in most of SACC patients. Schematic illustration was shown in Fig. 6f. Discussion Tumor dormancy, mentioned in 1864 and described in 1954 by Hadfield as a temporary arrest in mitosis [25] has been defined as a clinical term [26]. It has been demonstrated that tumor dormancy was implicated in the invasion and metastasis of EMT program in many types of tumors. Here, we demonstrated that DEC2 participated in SACC dormancy under a high expression condition in xenograft of nude mice. And the formation of lung metastases was accompanied by low level of DEC2, which then reawakened dormancy and promoted proliferation of metastases. Then, we further addressed these different roles of DEC2 in primary and metastasis lesions and found that high expression of DEC2 involved in CoCl 2 -induced dormancy, indicating that different hypoxia states in primary and metastasis lesion may regulate DEC2-induced dormancy to keep or exit cell dormancy state. Further, overexpression of DEC2 induced the entry of SACC into dormancy, mediated by activating the expression of Slug, which then drived EMT program, contributing to growth arrest and dormancy, as well as enhanced migration and invasion capabilities. Our current finding of low level of DEC2 inducing SACC cells to exit dormancy in the second lesion reveals the role of DEC2 in regulating tumor cell dormancy has involved in hypoxia condition and EMT program. DEC2 is one of the basic helix-loop-helix-Orange transcription factors, featured with a basic DNA binding domain, a helix-loop-helix (HLH) dimerization domain, and Orange extended dimerization domain [27]. And it has been implicated in regulating many types of biochemical processes, including circadian rhythm [28], cell proliferation and differentiation [29], apoptosis, hypoxia response, and EMT of tumor cells [28,30]. It was demonstrated that DEC2 inhibited tumor cells proliferation in esophageal cancer and osteosarcoma [31,32]. In this study, we demonstrated that DEC2 induced tumor dormancy of SACC both in vivo and vitro. It was supported by previous publication, which showed that the expression of DEC2 drived tumor dormancy in HNSCC and breast cancer [13,33]. Our previous study has demonstrated that atRA treatment could also drive tumor dormancy of SACC by upregulating NR2F1 [15]. Sosa et al. proposed that NR2F1 was an important node in tumor dormancy induction and maintenance [34]. In the present study, we found that DEC2 was also participated in atRA induced dormancy. DEC2 was increased in atRA treated dormant cells and these cells were activated partially after DEC2 knockdown. But the expression changes of DEC2 did not affect NR2F1. The reason for this may be that this dormant process was mediated by complicated signal networks including DEC2 and NR2F1, but they did not regulate each other directly. Hypoxia was a poor-prognosis microenvironmental feature of solid tumor, and Fluegen et al. proposed that primary tumor hypoxic microenvironment promoted the production of dormant tumor cells and resulted in chemo-radiotherapy resistance [16,35]. It has also been demonstrated that hypoxic stress induced breast cancer dormancy, but the relationship and molecular mechanism within hypoxia and dormancy of SACC is still ambiguity [22]. Previous studies have shown that the expression of DEC2 and HIF1α were positively correlated during the progression of human osteosarcoma [36]. We found that high expression levels of HIF1α and P27 were all participated in the dormancy process regulated by DEC2 in SACC. And we also verified the validity of CoCl 2 -based model in vitro for researching the relationship between tumor progression and hypoxic stress. We next detected the relationship between hypoxia and dormancy of SACC and the function of DEC2 during this process. The results showed that CoCl 2 induced hypoxia-mimicking microenvironments can drive dormancy in SACC cells and high expression of DEC2 was necessary for the above dormant state. Furthermore, the dormant tumor cells could be reawakened when the microenvironment changed from hypoxia to normal oxygen. Consistent with the above results, we proposed that in mouse xenograft model, DEC2overexpresed dormant SACC cells transferred into lung tissue and reactivated colony growth because of the abundant oxygen microenvironment. Exiting from dormancy always accompanied by DEC2 downregulation and resulted in the formation of numerous lung metastases and poor prognosis. It has been proposed that EMT-positive cells may enter into dormant state [5,20,37], so we investigated the role of DEC2 in EMT inducing dormancy. In the present study, we initially found that DEC2 overexpressed dormant tumor cells displayed upregulation of Slug, one of the EMT transcription factors. Knockdown of Slug reversed their dormant state and suppressed migration and invasion. Therefore, we hypothesized that the clock gene DEC2 could drive tumor dormancy through inducing EMT program and thus promoting tumor cells ability of migration and invasion in SACC. Consistently, it was shown that tumor cells with the features of invasiveness which have undergone EMT always manifested characteristics of dormancy. And reexpression of E-cadherin always accompanied by proliferative activity which further indicated the critical roles of EMT during tumor dormancy [38]. Jiang et al. have also proposed that PRRX1 can drive the transition of EMT, and dormant state of cancer cells through miR-642b-3p in head and neck squamous cell carcinoma [39]. Therefore, cellular phenotype affects the process of tumor dormancy. Furthermore, it was reported that DEC2, one of the clock genes, promoted tumor metastasis and drived tumor dormancy [40]. To our knowledge, this is the first research suggesting that DEC2 induced EMT process to facilitate tumor dormancy through the control of Slug. Additionally, we also investigated the expression of DEC2 in the primary SACC patients and normal salivary gland. The results showed that high expression of DEC2 was related with tumor dormancy and tumor site, histological subtype, and recurrence. We proposed the primary reason of poor prognosis caused by tumor dormancy is that dormant tumor cells are unstable and can resume proliferation as the external microenvironment changes. And just because of that, the purpose of tumor therapy might attempt to maintain the dormant state of tumor cells and prevent dormancy exit and growth resumption. Neophytou et al. proposed that the ultimate goal is to prolong the dormant period of metastatic tumor cells in breast cancer [41]. And it was also demonstrated that during tumor treatment, the Hippo signaling pathway contributes the initiation and stabilization of tumor dormancy [42]. But there are still questions that need to be answered to detect how DEC2 affect the growth state of tumor cells in different microenvironment. Conclusion In this study, we showed two different dormancy fates of SACC cells by different expressions of DEC2 in the primary and metastasis of SACC. Several groups have reported that there are different oxygen concentrations in primary and metastasis tumors [43][44][45]. Our findings suggest the possibility that different hypoxia states in primary and metastasis lesion may regulate DEC2induced dormancy to keep or awaken cell dormancy state. And DEC2 promoted the dormancy, EMT, migration and invasion of SACC cells, in which transcription factor Slug played an important role. A deep understanding of DEC2 function in tumor dormancy may provide a novel insight for improved treatment of SACC.	2023-01-19T21:35:26.441Z	2021-05-14T00:00:00.000	{ "year": 2021, "sha1": "ce74e4aa996ccaa86cf1749831462260a08d1d4b", "oa_license": "CCBY", "oa_url": "https://jeccr.biomedcentral.com/counter/pdf/10.1186/s13046-021-01956-0", "oa_status": "GOLD", "pdf_src": "SpringerNature", "pdf_hash": "ce74e4aa996ccaa86cf1749831462260a08d1d4b", "s2fieldsofstudy": [ "Biology" ], "extfieldsofstudy": [] }
372208	pes2o/s2orc	v3-fos-license	SEQADAPT: an adaptable system for the tracking, storage and analysis of high throughput sequencing experiments Background High throughput sequencing has become an increasingly important tool for biological research. However, the existing software systems for managing and processing these data have not provided the flexible infrastructure that research requires. Results Existing software solutions provide static and well-established algorithms in a restrictive package. However as high throughput sequencing is a rapidly evolving field, such static approaches lack the ability to readily adopt the latest advances and techniques which are often required by researchers. We have used a loosely coupled, service-oriented infrastructure to develop SeqAdapt. This system streamlines data management and allows for rapid integration of novel algorithms. Our approach also allows computational biologists to focus on developing and applying new methods instead of writing boilerplate infrastructure code. Conclusion The system is based around the Addama service architecture and is available at our website as a demonstration web application, an installable single download and as a collection of individual customizable services. Background This paper introduces a flexible and loosely coupled data management system for high throughput sequencing experiments. The system is designed to face the challenges of research, and is required as the versatility and applicability of high throughput sequencing experiments is growing rapidly. The system can be overlaid on top of existing software, and can be used to integrate different specialized algorithms. There already exist a number of commercial solutions (Geospiza's GeneSifter [1], Genomatix Genome Analyzer [2,3]), and non-commercial solutions (Galaxy [4], Cis-Genome [5], ChIP-Seq Analysis Server [6]) for the management and analysis of high throughput sequencing information. The main drawback to these solutions is that they focus on providing static "one stop shop" solutions, which are designed to fit known markets, using well-established methods. While these static systems are useful for non-technical researchers in a production science environment, they lack flexibility for the research scientist who wishes to use cutting edge methods and tools. The existing systems tend to focus on well-established applications for high throughput sequencing: experiments where the technology is seen as a more accurate "digital" equivalent to microarrays (e.g. RNA-Seq), experiments to determine protein binding (e.g. ChIP-Seq), or large scale genome assembly projects. However, high throughput sequencing has the potential of becoming ubiquitous across many avenues of investigation. This potential is due to both an increase in our understanding of systems biology and the capabilities of the new generation of instruments. As the field is constantly evolving new discoveries are continually being made, including new medically related functionality of small RNAs [7], new families of RNA [8], and signaling through extra-cellular RNAs [9]. New techniques and instruments are also being developed that provide insight into these new facets, due to an increase in throughput (e.g. multiplexing [10,11] and long reads [12]) and sophistication (e.g. BS-Seq and targeted approaches). For these reasons, any sequencing software infrastructure used in the research environment must be easily adaptable. By this we mean it must have the ability to be readily changed for new usage. For example, we can expect each research area to require different mechanisms for normalization and replication strategies, sample and experiment vocabularies, and analysis algorithms. Generally within research each project requires a large amount of de novo analysis development and customization to support: new technology strategies such as allowing for multiplexing or integrating with new instrumentation; informatics strategies, to allow for data and system integration; and new computational strategies, to support analysis and datamining tasks. Additionally, each laboratory will have their own demands in terms of experiment QA, annotations and integration with processes (e.g. preferred desktop analysis tools) and integration with other data types. Therefore, it is important that the research community have access to a system that is: • Open. The system must be distributed as an open software project as many users will need to modify the system to meet their specific needs. • Standardized. The system should follow widely used standards for both software development and data exchange. This will ensure that the code base will be easier to maintain and have greater connectivity with external systems and tools. • Adaptable. The system must be easily adaptable without requiring a detailed understanding of the aspects of the internal software architecture. In this way, significant modifications can be implemented efficiently and quickly. • Deployable. The system must be easy to rapidly deploy and modify. A system that is cumbersome or overly complex wastes the end user's development time with unnecessary setup and technical details. SeqAdapt follows these principles, and provides a standardized and modular architecture which is easy to use, adapt and maintain. The underlying enterprise architecture, Addama [13] has been designed to provide the adaptability required to enable the rapid development needed within research driven science. Implementation To meet the demands of researchers we have developed SeqAdapt, a solution that is able to: scale to meet the requirements of the research environment, use best practices for mainstay applications (e.g. ChIP-Seq), and be readily adapted to new usage. The system is built using a general software infrastructure to support Adaptable Data Management (Addama). SeqAdapt integrates external sample tracking software (e.g. SLIMseq [14]), workflows for executing analyses (e.g. the MACS algorithm [15]) and robust data management (e.g. JCR) to provide a modular and adaptable system for high throughput sequencing experiments. Due to the data volumes involved with high throughput sequencing a software infrastructure is often required to facilitate storage, management and analysis. We have used the Addama system to provide the necessary support for the creation of a workflow encompassing the entire process (see Figure 1) that is complete, lightweight and easily adapted to changing requirements. This solution allows for changes in the underlying sequencing technology while still providing the ability to plug in new processing methods. A pluggable architecture is important as the technology, data formats, and processing methods are changing rapidly in the field of sequencing. Performance and the ability to scale up as datasets grow in size is also a requirement with the management of sequencing data. To allow for flexible deployment, we use a Service Oriented Architecture (SOA), where distributed services provide the required functions. As these services grow, they can be deployed on the appropriate computational infrastructure (e.g. compute farms, cloud computing environments or dedicated servers). Results The SeqAdapt default download consists of a preconfigured bundle of services. Each of these services can be interchanged with alternative implementations. Additionally the underlying system makes it possible to develop custom applications, and to rapidly integrate new analysis tools (see Integrating New Analyses section). The software infrastructure is comprised of three main services: a sample tracking system, a data management system and a process management system (see Figure 1). The sample tracking system's functionality includes sample submission, annotation with controlled vocabularies and file management. The data management system uses Addama to organize the data and to trigger new analyses. The process management system (the Addama Robot) is a lightweight and generic system for executing processing pipelines and persisting inputs and outputs. The Addama Robot allows for analyses to be run directly on a dedicated machine which has been configured for a specific analysis (e.g. has all the data files, dependencies and resources required for the specific analysis). Typically the analyses are run on either a server machine or, if the analysis is still under in-house development, on the specific developers computer. The robot is responsible for monitoring jobs and the transparent transportation of required data between the repository and the analysis environment. The data management system uses standard open technologies including: content repositories which are used to generically store all experimental information; information indexing services, which provide for search capabilities across all data and metadata stored; and service registries, which allow for run-time discovery of different content repositories and associated services. Addama also provides an abstraction so that a set of interlinked content repositories can be accessed through a single web application layer. This layer is exposed with a JSON based RESTful web service. The process management system coordinates jobs using a Java Message Service (JMS). With this configuration, the system can scale from a single computer to a distributed set of execution agents on multiple servers to listening to a JMS message queue. All components of the SeqAdapt system are loosely coupled to allow for easy replacement with alternative systems. These alternative systems could include different sample tracking systems (e.g. Sequencescape [16]) different persistence stores (e.g. RDBMS) and different analysis tools (e.g. RNA-Seq tools). This system allows for rapid integration of scientific algorithms using the standardized Addama framework. The integration system is designed to be flexible, and allows for any command line analysis tools to be "plugged in". The framework is suited for developing small-scale analyses as well as for large scale processing that requires scaled-up distribution. Further, all components needed for this system are provided in an easy-toinstall package. The default download for SeqAdapt has been set up for Chip-Seq analysis, and will process data from the Illumina Genome Analyzer II using the MACS algorithm (see the Availability section for download details). The default install allows the user to submit an analysis, monitor its progress, and then view the result files. Integrating New Analyses As discussed above, the enterprise system can be extended in a number of ways. As the system is based upon distributed loosely coupled services, these services can be replaced with ones offering the same interface. Each of the major components are built against technology standards (e.g. REST Web Services serve out JSON, the repository is standardized to the JCR specification). New analysis pipelines can be added by writing a mapping module and then registering the analysis with Addama. A benefit of the Addama systems is that a script can execute in a preexisting development environment, eliminating the time consuming task of replicating software installations on a processing server. When algorithms have reached a mature state and are used widely, the system scales up to have the execution agent Figure 1 Architecture of the default SeqAdapt system. The default SeqAdapt system consist of a number of services and analysis programs that can be replaced to allow for a high level of customization. Default sample tracking, data management, and analysis components are available. These components can be replaced with other systems to suit individual needs. All interfaces between the components are standardized using REST to enable interoperability. The default system uses SLIMseq for initially entering of sample/experiment annotation information, and publishes this information using JSON (over REST). An Addama service accesses the SLIMseq web service and stores the information in a JCR repository. When runs have been completed SLIMseq is updated and this information is pulled into Addama. To run an analysis the Addama Robot system is used, this system allows for any command line utility to be prepared, triggered and monitored. When an analysis is complete the results are automatically pulled into an Addama repository. Customized applications can also be written against the Addama web services. installed on many servers. A key benefit of this system is that it manages all of the non-scientific functionality needed in this type of processing. This freedom from writing boilerplate infrastructure code allows the computational biologist to focus on developing the needed scientific software. The Addama Robot (see Figure 2) allows for the rapid integration of these tools in a relatively short period of time. Integration of these tools requires that the developer has a rudimentary understanding of Addama, and also understands how the specific analysis tool works (in terms of data formats). By way of example we have integrated the ERANGE [17] RNA-Seq analysis tool. Once ERANGE was installed the integration work was completed in less than a day. Any analysis script that can be run from the command line can be integrated in the same manner. The steps involved in such integration are: 1. Define input location. This is done by providing a command-line executable wrapper script. This script will define all of the inputs to the ERANGE analysis and execute it. It will read the inputs from a local JSON file downloaded from the Addama service by the Robot. 2. Control outputs by configuring the wrapper script to write all results to an "/outputs" directory. This directly will be in the same location as the script, the creation of the directory will be handled by the Robot as well. Similarly any log information (e.g. errors, debugging messages) should be written to a "/log" directory. 3. Register the script by configuring the Addama Robot. The robot uses a properties file that defines the wrapper script that is to be executed, and a local path where each run will be output. Update these properties to reflect the locations of the ERANGE script and the directory where it write inputs, outputs and log messages. 4. Enable user submissions. To make submitting simple for the user, an optional web application may be developed. This application will take the expected inputs and send them to the Addama system via the REST interface. This same page can also be used to query Addama for the results of the Robot analysis and display those for the user as well. The robot automates the tasks that are required to integrate the analysis with the enterprise system. When the analysis is triggered the robot is responsible for the delivery of the inputs to the analysis, starting the analysis and monitoring the outputs. When completed the outputs, and any associated logs are loaded back, into Addama. Walkthrough A walk-through showing the default workflow for SeqAdapt is given in Figures 3,4,5 and 6. This walk-through shows how the system can be used to capture information about a Chip-Seq experiment, store the results and then analyze the reads using MACS. Conclusions The default system can be used "as-is" to support ChIP-Seq analysis, however it can also be rapidly customized Figure 2 Integration of analysis tools. An automated system is used to integrate new analysis tools with Addama. The robot system is responsible for delivering the correct inputs to an analysis script, monitoring the process while it runs, and then for publishing the results of the analysis back. Web applications can then be built on top of standard Addama APIs for providing customizable input information for specific scripts, and then for visualizing the results of the analyses. The loose coupling of each of the web applications from the underlying analysis script makes the system robust to change and (relatively) easy to maintain -especially when the scripts are under constant revision. Figure 3 Step 1: Sample Entry. Sample information is entered into the custom ordering system SLIMseq. Figure 4 Step 2: Browsing. A repository stores all the annotations and sample data files so they can be searched and browsed. Figure 5 Step 3: Analysis. Data stored within the repository can be analyzed with any integrated analysis. In this case MACS can be triggered and the results automatically stored back in Addama. to suit new usage. The requirement for such flexibility is due to the fact that it is rarely possible to foresee all the usages with which a new instrument technology, such as high throughput sequencing, will be applied. To be able to meet the requirements of a rapidly evolving research environment, as is found in most scientific institutions, an adaptable data management system is required. The modular framework described in this article is designed to provide such adaptability. The major advantage of the SeqAdapt/Addama system is that "ad hoc" tools can be rapidly integrated. In this instance "ad hoc" means tools that do not adhere to predefined standards and have been built without integration in mind (e.g. not provided as web services). The system is available from the Institute for Systems Biology's informatics web site, as: a demonstration system running as a web application, a preconfigured install for quick deployment and as a full download containing all the separate services (see Availability, below). If access control is a requirement then the full download should be used. Figure 6 Step 4: Browsing. The results of an analysis are stored in Addama and can be integrated with other data files and viewed in applications such as IGV. Availability and Requirements Project name: SeqAdapt. Licence: Apache 2.0. Any restrictions to use by non-academics: None. Full installation instructions and software downloads are available at the project web site. Abbreviations JSON (JavaScript Object Notation): A simple object serialization format originally designed to work with JavaScript. JSON is growing in popularity over other text based formats (e.g. XML) due to its lightweight and simplistic nature. JSON encoders and parsers are available in most languages; SOA (Service Oriented Architecture): A software architecture method where modules of software are separated into services that expose an interface of functionality. These modules can be combined in several ways to create larger distributed applications; Web Service: A software system that supports interaction over a network. Web services are typically exposed using HTTP; JMS (Java Message Service): A Java API standard that allows Java applications to exchange messages. It allows for distributed communication that is loosely coupled and asynchronous; REST (REpresentational State Transfer): An architecture for distributed content, such as the World Wide Web. A RESTful web service uses standard HTTP constructs (e.g. GET, POST, DELETE) to provide services, and communicates back to the client using HTTP response codes (e.g. 200 -OK, 404 -Not Found) along with content; JCR (Java Content Repository): A type of object store that is suited towards the storage, searching and retrieval of hierarchical data. A JCR can store both metadata and as well as files;	2017-06-20T04:34:33.289Z	2010-07-14T00:00:00.000	{ "year": 2010, "sha1": "5b1436a540f38641045e3cd7e54c32ba8501ef44", "oa_license": "CCBY", "oa_url": "https://bmcbioinformatics.biomedcentral.com/track/pdf/10.1186/1471-2105-11-377", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "5b1436a540f38641045e3cd7e54c32ba8501ef44", "s2fieldsofstudy": [ "Computer Science" ], "extfieldsofstudy": [ "Computer Science", "Medicine" ] }
10356691	pes2o/s2orc	v3-fos-license	Microscopic and macroscopic polarization within a combined quantum mechanics and molecular mechanics model A polarizable quantum mechanics and molecular mechanics model has been extended to account for the difference between the macroscopic electric field and the actual electric field felt by the solute molecule. This enables the calculation of effective microscopic properties which can be related to macroscopic susceptibilities directly comparable with experimental results. By seperating the discrete local field into two distinct contribution we define two different microscopic properties, the so-called solute and effective properties. The solute properties account for the pure solvent effects, i.e., effects even when the macroscopic electric field is zero, and the effective properties account for both the pure solvent effects and the effect from the induced dipoles in the solvent due to the macroscopic electric field. We present results for the linear and nonlinear polarizabilities of water and acetonitrile both in the gas phase and in the liquid phase. For all the properties we find that the pure solvent effect increases the properties whereas the induced electric field decreases the properties. Furthermore, we present results for the refractive index, third-harmonic generation ~THG!, and electric field induced second-harmonic generation ~EFISH! for liquid water and acetonitrile. We find in general good agreement between the calculated and experimental results for the refractive index and the THG susceptibility. For the EFISH susceptibility, however, the difference between experiment and theory is larger since the orientational effect arising from the static electric field is not accurately described. © 2005 American Institute of Physics. @DOI: 10.1063/1.1831271 # I. INTRODUCTION 4][5][6][7] The main reason for this is that TD-DFT provides a level of accuracy which in most cases is sufficient at a lower computational requirement than other methods.[10][11][12][13][14][15][16][17] In recent years applications of TD-DFT to calculate properties of molecules in solution has also been presented. 18 -27Among the methods for treating solvent effect on molecules are the combined quantum mechanical and molecular mechanics ͑QM/MM͒ models. 28 -39The QM/MM model has recently been extended to study molecular response properties in solution within TD-DFT. 18 -20,22-24An example of such a QM/MM method is the discrete solvent reaction field ͑DRF͒ model which we have recently developed. 18 -20The model combines TD-DFT ͑QM͒ description of the solute molecule with a classical ͑MM͒ description of the discrete solvent molecules.The latter are represented using distributed atomic charges and polarizabilities. An important feature of the model is the inclusion of polarizabilities in the MM part which allows for the solvent molecules to be polarized by the solute and by interactions with other solvent molecules.Van Duijnen et al. 50applied basically the same polarizable QM/MM formalism, implemented in a conventional wave function package, to calculate the static ͑hyper͒polarizability of acetone in ten different solvent.Also Kongsted et al. 41 stressed the importance of inclusion of polarizabilities in the calculations of response properties.Another important feature of the model is the inclusion of a short range damping of the interactions.This has been included in two ways.The first is the use of the modi-fied dipole interaction model of Thole 42 which avoids the''polarization catastrophe'' by introducing smeared out dipoles which mimic the overlapping of the charge distributions at short distances.The second way is that at short distances between the QM and the MM part the QM/MM interactions are damped to account for the short range repulsion in an approximate way.This is done by replacing the point charge by a Gaussian charge distribution with a unit width and point dipoles smeared out in a similar manner. 18,19,43n this work we will extend the QM/MM formalism to also include the so-called local field factors, i.e., the difference between the macroscopic electric field and the actual electric field felt by the solute.This will enable the calculation of effective microscopic properties which can be related to the macroscopic susceptibilities.The macroscopic susceptibilities can then be compared directly with experimental results.There exist in the literature some other approaches to calculate effective properties and relate these to the macroscopic properties, 40,44 -50 which differ from this work in the way the solvent was represented.First we will describe the theoretical framework and then we will present numerical results for liquid water and liquid acetonitrile.These liquids are chosen since there exist several theoretical and experimental studies of the microscopic and macroscopic properties. II. THEORY The basic concept of nonlinear optics is the expansion of the total macroscopic polarization in a material in powers of the macroscopic electric field where the expansion coefficients define the macroscopic ͑nonlinear͒ susceptibilities.Similarly, the total microscopic polarization ͑dipole moment͒ is expanded in terms of the total microscopic electric field with expansion coefficients defining the microscopic ͑nonlin-ear͒ polarizabilities.However, in the literature several different conventions exist for describing nonlinear optical properties 51 which differ in the numerical coefficients used.Therefore, in order to compare values obtained in different conventions it is important to correct for the differences in the numerical factors used.This has been clarified by Willetts et al. 51 but remains a problem, because often it is not stated explicitly which convention is used.In this work we will use a perturbation series expansion for the macroscopic polarization, which is often used for experimental properties, and a Taylor series expansion for the microscopic polarization, which frequently is used for theoretical properties. A. The macroscopic polarization The macroscopic polarization of a material in the presence of a macroscopic electric field F mac is expressed as a power series in the field strength as 52,53 where P 0 is the permanent polarization, (1) the linear optical susceptibility, (2) the second-order nonlinear optical susceptibility, and (3) the third-order nonlinear optical sus-ceptibility.The subscripts I,J,K,L, . . .denotes space-fixed axes and the Einstein summation convention is used for repeated subscripts.If we consider the macroscopic field to be a superposition of a static and an optical component, the macroscopic polarization can be expressed as 52,53 The Fourier amplitudes of the polarization are then given in terms of the frequency-dependent susceptibilities as 52,53 P I s ϭ␦ s ,0 ͑4͒ where the output frequency is given as the sum of input frequencies s ϭ ͚ a a .The numerical coefficients K (Ϫ s ; a ,...) arise from the Fourier expansion of the electric field and polarization and ensures that all susceptibilities of the same order have the same static limit.A tabulation of the coefficients can be found in Ref. 54 and 55.The frequency-dependent susceptibilities can then be found from Eq. ͑4͒ by differentiation which gives the linear susceptibility the second-order nonlinear susceptibility and the third-order nonlinear susceptibility . B. The microscopic polarization Similarly the microscopic polarization ͑dipole moment͒ can be expanded in terms oscillating at different frequencies as 44,51 The microscopic dipole moment is then usually given by a Taylor expansion as 44,51 ͑9͒ where ␣ 0 is the permanent electric dipole moment, ␣ ␣␤ (Ϫ s ; s ) is the polarizabillity, ␤ ␣␤␥ (Ϫ s ; a , b ) is the first hyperpolarizability, and, ␥ ␣␤␥␦ (Ϫ s ; a , b , c ) is the second hyperpolarizability.The numerical coefficients K(Ϫ s ; a ,...) are the same as for the macroscopic polarization and again this ensures that all ͑hyper͒polarizabilities of the same order have the same static limit.The subscripts ␣,␤,␥,... denote molecule-fixed axes and again the Einstein summation convention is used for repeated subscripts.The microscopic polarization is expanded in terms of the actual total electric field F b ,␥ tot felt by the molecule.In the condensed phase the actual electric field felt by the molecule is different from the macroscopic electric field.Therefore, in order to express the macroscopic properties in terms of the microscopic properties we need to relate the actual electric field at a molecule to the macroscopic electric field. C. The local electric field The concept of relating the actual electric field, often called the internal or local field, to the macroscopic field dates back to the work of Lorentz. 56,57Lorentz 56 derived a simple relation between the internal electric field, the macroscopic electric field, and the macroscopic polarization of the system, and due to its simplicity Lorentz local field theory is still used. 52,53,57,58The central idea is that only close to the molecule we need to consider explicitly the field from nearby molecules, so the total system is separated into a macroscopic region far from the molecule and a microscopic region close to the molecule.The molecules in the macroscopic region can then be described by the average macroscopic properties.Therefore, inside a macroscopically small, but microscopically large, virtual cavity V we subtract the contribution from the macroscopic electric field and replace it by the correct discrete local field, where F s ,␣ pol is the macroscopic electric field in the cavity V and F s ,␣ disc (⍀) is the discrete electric field in the cavity V which depends on the local configuration ⍀ of the molecules inside the cavity.Since we are not allowing the macroscopic region to adjust to the presence of the cavity the polarization remains homogeneous. 57This approach neglects that a static electric field tends to orient molecules with a permanent dipole 57,59 and therefore a correction due to Onsager 59 is often used for a static electric fields.Lorentz showed 56 that for a cubic arrangement of identical particles the discrete field was zero.This is also true on average for a completely random distribution where there is no correlation between the induced dipoles and the position of the molecules. 57For a spherical cavity the macroscopic field is simply given in terms of the macroscopic polarization 56,57 and the total electric field can be written as where we have split the discrete electric field F disc into two different contributions, F ind and F perm .The first term arises from the interactions of the macroscopic electric field with the other molecules in the cavity, i.e., accounts for the induced polarization of the surrounding molecules due to the electric field.The second term accounts for the interactions between the molecules when there is no electric field present, i.e., arises from the permanent charge distribution of the surrounding molecules.However, depending on the theoretical model used for describing the microscopic region, this splitting of the discrete electric field is not always possible nor necessary.The last two terms depend strongly on the local configuration of the molecules in the cavity and are inherently microscopic in nature and it is therefore better to treat these fields explicitly within the microscopic model used. D. The effective molecular properties Instead of expanding the induced dipole moment in terms of the total field, Eq. ͑9͒, we expand it in terms of an effective macroscopic electric field This expansion in terms of the effective field defines the so-called effective properties 44 as ͑13͒ These effective properties give an induced dipole moment due to the effective macroscopic electric field which is identical to the induced dipole moment in Eq. ͑9͒ and are the properties which we will relate to the experimental susceptibilities.This means that the microscopic contributions to the total field are incorporated into the effective properties.These effective properties could be compared with experi-mental results corrected for differences between the total field and the macroscopic electric field by the Lorentz/ Onsager local field method. 44 E. The solute molecular properties Since we have separated the discrete field into the two contributions mentioned above we can also choose to expand the induced dipole moment in terms of the field arising directly from the macroscopic electric field, where the field arising from the interactions between the molecules when there is no macroscopic field is incorporated into the properties.This gives an expansion which defines the so-called solute properties 44 as ͑15͒ These solute properties relate to the macroscopic properties corrected for the field from the dipoles of all other molecules induced by the macroscopic field in addition to the Lorentz local field.This corresponds to a thought experiment where the macroscopic field is allowed to propagate inside the cavity without being modified by interactions with the molecules. F. Relating the macroscopic and the microscopic polarization The macroscopic polarization is related to the average microscopic dipole moment per molecule by 52,58 ͑16͒ where N d is the number density and the brackets, ͗͘, denote orientational averaging, and relate the molecule-fixed axes to the space-fixed axes. 58,60Inserting the expansion of the dipole moment in terms of the effective macroscopic field, Eq. ͑13͒, we can express the macroscopic polarization in terms of the effective ͑hyper͒polarizabilities as, ͑17͒ We see that the averaging is done on the product of the ͑hyper͒polarizabilities and the effective fields.This is exactly the reason why the total electric field was split into an effec-tive macroscopic part and a microscopic part which was incorporated into the ͑hyper͒polarizabilities in Eq. ͑13͒ by expanding the dipole moment in terms of the effective field. Since the effective field is macroscopic we can take it outside the averaging and express the macroscopic polarization in terms of orientational averages of the effective ͑hyper͒polarizabilities as G. Local field factors Comparing Eq. ͑18͒ with Eq. ͑4͒ we see that we have to consider derivatives of the effective electric field with respect to the macroscopic electric field.This is usually done by introducing the so-called local field factors which relate the macroscopic field to the effective field. Using the definition of the effective electric field, Eq. ͑12͒, we obtain a local field factor where ⑀ (1) ( a ) is the optical dielectric constant at frequency a .This is the Lorentz form of the local field factors.However, as mentioned above this does not account for the fact that for a static macroscopic field the molecules tend to orient.Onsager 59 analyzed this problem and suggested the following form for the local field factor, to be used for static electric fields, where ⑀ (1) (0) is the dielectric constant and n (1) (0) is the refractive index at zero frequency.We see that the Onsager field factor reduces to the Lorentz factor for optical fields by using the relation n (1) ( a ) 2 ϭ⑀ (1) ( a ). H. Orientational averaging In order to relate the molecule-fixed axes to the spacefixed axes we need also to consider molecular rotations ͑or orientations͒.The orientational averaging and thermal averaging of the dipole moment will be done using classical theory and is given by 60 where ⌽ ␣I is the cosine of the angle between the molecular axis ␣ and the laboratory axis I.The angular dependent part of the energy in the presence of the electric field is given by We note that it is only the solute properties which are responsible for the change in the energy due to the electric field.If we expand the exponential and only terms of the order (kT) Ϫ1 are retained we get By combining the definitions of the susceptibilities in Eqs.͑5͒, ͑6͒, and ͑7͒ with the expression for the macroscopic polarization in terms of the effective ͑hyper͒polarizabilities we can obtain a link between the macroscopic and the microscopic properties. I. Refractive index The macroscopic quantity determining the refractive index is the linear optical polarization due to an optical electric field.By inserting the definition of the effective field, Eq. ͑12͒, into Eq.18 and using the definition of the linear susceptibilitiy, Eq. ͑5͒, we obtain ͑25͒ It is noted that there is no contribution from the rotation of the dipole moment since the optical field is considered to be oscillating faster than the permanent dipole moments can be oriented.The linear susceptibility can then be written in terms of the mean effective polarizability by rewriting Eq. ͑25͒ as which is the standard expression for the susceptibility corrected for the Lorentz local field, 52,61 although using the effective rather than the gas phase polarizability.The susceptibility is related to the refractive index or the optical dielectric constant of the system as which is the familiar Lorentz-Lorenz or Clausius-Mossotti equation, 56,57,61 again with the effective polarizability instead of the gas phase polarizability. J. Dielectric constant In a dielectric constant measurement the polarization due to a static electric field is measured and the corresponding susceptibility is given by The first term is the rotational contribution arising from the permanent dipole moment and is given by The second term is the isotropic orientation average of the polarizability, often referred to as the mean polarizability, given by 60 and denoted ␣ ¯.By combining these two terms the equation for the linear susceptibility can be rewritten as ͑31͒ The dielectric constant is then related to the linear susceptibility through the usual equation ⑀ (1) ͑ 0 ͒ϭ1ϩ4 ZZ (1) ͑ 0;0͒.͑32͒ It should be noted that the susceptibility in Eq. ͑31͒ depends on the dielectric constant through the Onsager local field factor, L 0 , and is therefore not a defining equation.However, this will allow us to make an estimate of the dielectric constant. K. Third-harmonics generation The first nonlinear susceptibility we will consider is the third-order nonlinear susceptibility which arises from three optical electric fields corresponding to the THG experiments.The THG susceptibility is then obtained by inserting Eq. ͑12͒ into Eq.͑18͒ and using the definition of the susceptibilitiy, Eq. ͑7͒.This gives the susceptibility as where we see that we have a contribution both from the linear susceptibility and from the third-order nonlinear susceptibility.The isotropic orientation average of the second hyperpolarizability, often referred to as the mean or parallel second hyperpolarizability, is given by 60 ͑34͒ We can then rewrite the THG susceptibility as where we have used the relation between the linear susceptibility and the dielectric constant in Eq. ͑27͒.Using Eq. ͑26͒ we can express the term with the effective polarizability in terms of the dielectric constant at 3. This allows us to express the third-order nonlinear susceptibility as which is the form for the nonlinear susceptibility well known from standard Lorentz local field theory with nϩ1 local field corrections, where n is the number of applied fields. L. Electric field induced second-harmonic generation The second nonlinear susceptibility we will consider is the third-order nonlinear susceptibility arising from two optical and one static electric field and corresponds to the electric field induced second-harmonic generation ͑EFISH͒ experiments.The EFISH susceptibility is then obtained by inserting Eq. ͑12͒ into Eq.͑18͒ and using the definition of the susceptibilitiy, Eq. ͑7͒.This gives the susceptibility as We see that the EFISH susceptibility consists of three terms and compared with the THG expression there is a term depending on the effective first hyperpolarizability.The second term is a rotational contribution analogous with the dielectric constant and is ͑38͒ where the mean hyperpolarizability ␤ ¯ʈ in the direction of the dipole moment, here the z axis, is introduced This allows us to rewrite the EFISH susceptibility as M. The discrete solvent reaction field model In the QM/MM method the solvent molecules ͑MM͒ are treated with a classical force field and the interactions between the solute and solvent are described with an effective operator.In the QM/MM method the total ͑effective͒ Hamiltonian for the system is written as 28 -39 H ˆϭH ˆQM ϩH ˆQM/MM ϩH ˆMM , ͑41͒ where H ˆQM is the quantum mechanical Hamiltonian for the solute, H ˆQM/MM describes the interactions between solute and solvent, and H ˆMM describes the solvent-solvent interactions. We have recently developed such a method for studying solvent effect on molecular properties which we denoted the DRF ͑Refs.18 -20͒ where the QM part is treated using DFT. Within the discrete solvent reaction field model the QM/MM operator at a point r i is given by where the first term el is the electrostatic operator and describes the Coulombic interaction between the QM system and the permanent charge distribution of the solvent molecules.The second term pol , is the polarization operator and describes the many-body polarization of the solvent molecules, i.e., the change in the charge distribution of the solvent molecules due to interaction with the QM part and other solvent molecules.The charge distribution of the solvent is represented by atomic point charges and the many-body polarization term is represented by induced atomic dipoles at the solvent molecules. For a collection of atomic polarizabilities in an electric field, assuming linear response, the induced atomic dipole at site s is given by where ␣ a,␣␤ is a component of the atomic polarizability tensor at site s, which for an isotropic atom gives ␣ s,␣␤ ϭ␦ ␣␤ ␣ s , and is the screened dipole interaction tensor. 42,62,18Here we neglect the frequency dependence of the classical part, i.e., the atomic polarizability is frequency independent, but the model can easily be extended to include also this effect. 43,63 s,␤ init () is the initial electric field at site s and is in this work extended to also include the macroscopic electric field.The initial field then consist of four terms where F t,␤ QM,el () is the field arising from the frequencydependent electronic charge distribution of the QM part, F t,␤ QM,nuc the field from the QM nuclei, F t,␤ M M,q the field from the point charges at the solvent molecules, and F t,␤ mac () the macroscopic electric field.The inclusion of the macroscopic electric field in Eq. ͑44͒ describes the induced dipole moments in the solvent due to macroscopic electric field.This was the reason for splitting the discrete electric field in Eq. ͑11͒ into a part induced by the macroscopic electric field and a part existing even without a macroscopic field.Therefore, if the macroscopic electric field is included in Eq. ͑44͒ we will be calculating the effective properties and if it is excluded we calculate the solute properties. We can now calculate and distinguish between different effects in going from microscopic properties to macroscopic properties.Since the permanent discrete electric field in Eq. ͑11͒ is always present we will associate this with a pure solvent effect, i.e., the solute properties include this solvent effect.The effective properties then includes the effects of a microscopic induced field in the solvent due to the macroscopic electric field.Finally, by combining the effective properties with the macroscopic local field factors ͑Lorentz/ Osager͒ we can obtain the macroscopic susceptibilities. III. COMPUTATION DETAILS 7][68] The details of the implementation are described in Refs.18 -20 and is in this work only extended to include the macroscopic electric field in Eq. ͑44͒. The basis set used for water consists of a large eventempered basis set of Slater-type orbitals with orbital exponent ϭ␣␤ i , iϭ1,...,n ͑details given in Ref. 19͒.For acetonitrile the standard TZP basis set was augmented with firstorder and second-order field induced polarization functions taken from Ref. 69.All the calculations were done using the BP-GRAC ͑gradient-regulated asymptotic connection BP͒ potentials. 13,14The BP-GRAC potential sets the highest ordered molecular orbital ͑HOMO͒ level at the first ionization potential ͑IP͒ and therefore requires the IP as input.For water IPϭ0.45 a.u. and for acetonitrile IPϭ0.46143 a.u. was used.The values for IP were obtained from calculations using the SAOP potential for which it has been shown that the HOMO level corresponds well with the experimental IP. 70he molecular dynamics ͑MD͒ simulations were performed with the discrete reaction field polarizable force field 35,62,[71][72][73][74] using the DRF90 program. 71For both water and acetonitrile a MD-simulation of 50 ps with a timestep of 1 fs was performed at 298K, using a Nose-Hoover thermostat 75 ͑with ϭ1 ps) to keep the temperature constant and a soft wall force potential 71 to keep the particles inside the simulation box.After every 0.5 ps, the configuration of solvent molecules was kept and the QM/MM calculations were performed.In the simulation of water, 256 molecules were placed in a spherical box of 23.12 bohrs; for acetonitrile, 128 molecules were placed in a spherical box of 26.15 bohrs.The sizes of the simulation boxes were chosen so that the simulated macroscopic densities correspond to the experimental values of 0.998 and 0.786 kg/l, respectively.In the simulations, the molecules were treated as rigid bodies using quaternions. 76Standard atomic polarizabilities 62 for all atoms were used in the simulations.For acetonitrile MDC-d charges were used 77 that were obtained from DFT calculations in a TZ2P basis set with the ADF program, while for water charges were fitted to reproduce the experimental dipole moment.For the van der Waals interaction a 6-12 Lennard-Jones potential were used for water and the standard DRF90 potential for acetonitrile.For water the Lennard-Jones parameters were taken from Ref. 78 and adjusted to match the point charges and atomic polarizabilities used in this work.The new parameters obtained are Rϭ1.7385Å and ⑀ϭ0.2900 kcal/mol located on the oxygen atom.By inspecting radial distribution functions, it was checked that the solvent shells around the central molecule were correctly represented. The atomic parameters, i.e., point charges and atomic polarizabilities, needed for the solvent molecules in the QM/MM calculations are: For water the point charges are q H ϭ0.3295 and q O ϭϪ0.6590 a.u. and the atomic polarizabilities are ␣ H ϭ0.0690 and ␣ O ϭ9.3005 a.u.For acetonitrile the point charges are q C1 ϭ0.288 340 a.u, q C2 ϭϪ0.009 643 a.u, q H ϭ0.017 028 and q N ϭϪ0.329781, where C2 is the carbon atom attached to the nitrogen.The atomic polarizabilities are ␣ C ϭ8.6959, ␣ H ϭ2.8382, and ␣ N ϭ3.5042 a.u. IV. RESULTS In the following we will present microscopic and macroscopic properties for the two liquids water and acetonitrile.The solute and the effective properties will be presented as averaged over the 101 different solvent configurations.The standard deviation will also be displayed to indicate the average fluctuation in the properties due to the different solvent configuration.All microscopic properties will be given in atomic units ͑a.u.͒ whereas the macroscopic susceptibilities will be presented in cgs units ͑esu͒. A. Gas phase results In Table I we present DFT results for , ␣ ¯(Ϫ;), ␤ ¯ʈ(Ϫ2;,), and ␥ ¯ʈ(Ϫ2;,,0) for water and acetonitrile in the gas phase.The results have been calculated at ϭ0.0428 a.u.(ϭ1064 nm) for water and ϭ0.0885 ( ϭ514.5 nm) for acetronitrile and are compared both with experimental and ab initio coupled cluster single doubles ͑CCSD͒ results.In general we find good agreement between the DFT results and the CCSD results for all properties.The largest difference of ϳ25% between the calculated values is in ␥ ¯ʈ and ␤ ¯for acetonitrile.The DFT results for ␤ ¯ʈ is lower than the CCSD results whereas for ␥ ¯ʈ the opposite is found.If we compare with results obtained from EFISH experiments we see that for water there is an excellent agreement between the calculated and the experimental results for all properties.In the case of acetonitrile we see that for ␤ ¯ʈ the DFT is in better agreement with the experiment whereas for ␥ ¯it is the CCSD results.However, as mentioned in the theory section the measured quantity in the EFISH experiment is ⌫ ¯ʈϭ z ␤ ¯ʈ/3k b T ϩ␥ ¯ʈ .Therefore, this value is also reported for the different methods in Table I.Again, we see that there is good agreement between theory and experiment for water.For acetonitrile the DFT and CCSD results are within 10% and 20%, respectively, of the experimental results. B. Microscopic response properties In Table II we present the dipole moment of acetonitrile and water in the gas phase and in the liquid phase.For both water and acetonitrile we see that there is a large enhancement of the dipole moment in going from the gas phase to the liquid phase.The enhancement for water is ϳ40% and for acetonitrile ϳ25%.From Table II we see that the dipole moment is completely determined by the z component both in the gas phase and in the liquid phase.Since the liquid phase dipole moment is obtained from an averaging over 101 configurations the fact that the other component of the dipole moment in the liquid phase is zero indicates that the averaging is close to isotropic.The standard deviations of the dipole moment is also presented in Table II and amounts to ϳ5% for both water and acetonitrile.In Fig. 1 we display the dipole moment of the individual configurations for water.As can be seen from the figure there is strong dependence on the configurations and the dipole moment oscillate between 2.2 and 2.8 debye.In this work there is no difference between the solute and the effective dipole moment because the MD simulations are done without the static electric field present.Therefore, the orientational effect due to the electric field is not accounted for in the MD simulations.In classical MD simulations the orientational effects on the dielectric constant can be obtained from the fluctuation in the dipole moments of the molecules. 79However, this approach will not work in the present QM/MM simulations since there is only one space-fixed molecule in the QM part. The frequency-dependent polarizability components, mean value and anisotropy of water and acetonitrile both in the gas and liquid phase are presented in Table III.For the liquid phase we present both the solute and the effective properties.We also present both the average of the anisotropy ͗⌬␣͘ and the anisotropy of the average polarizability ⌬͗␣͘.First we note that the solvent effects are not very large both for water and acetonitrile.In both cases the solute properties are larger than the gas phase values.The mean value of the effective properties are very close to the gas phase values.However, for water the components of the polarizability tensor is different for the effective and the gas phase properties.The fact that the effective mean polarizability is close to the gas phase values shows that the discrete electric field in Eq. ͑11͒ is, to first order, close to zero.If we compare ͗⌬␣͘ with ⌬͗␣͘ we see that for water they are very different.The reason for this is that for water the anistropy is very small and therefore the off-diagonal tensor components becomes important.The off-diagonal elements are on average equal to zero due to the isotropic sampling and, therefore, the anisotropy of the averaged polarizability is small and close to the gas phase value.We also note that the fluctuations are slightly larger for the effective properties than for the solute properties.In the Figs.2͑a͒ and 2͑b͒ we display, respectively, the solute and effective mean polarizability of water for the individual configurations.The solute polarizability oscillates between 10 and 10.6 a.u.whereas the effective polarizability oscillations between 9.6 and 10.4 a.u.In the case of the solute polarizability all results are larger than the gas phase value whereas for the effective polarizability some of the configurations give a polarizability smaller than in the gas phase. In Table IV we present the frequency-dependent ͑SHG͒ first hyperpolarizability at the frequency, ϭ0.0428 a.u., for water and acetonitrile in the gas and liquid phase.For the liquid phase we present both the solute and the effective properties.First we note that for the first hyperpolarizability the solvent effects are very large.For water this leads to a change in sign for the mean first hyperpolarizability.For acetonitrile the solute and effective mean hyperpolarizability are, respectively, a factor of 5 and 6 larger than the gas phase value.In both cases we see that the fluctuations due to the different solvent configurations are very large.The first hyperpolarizability is therefore extremely sensitive to the local structure of the solvent.In Figs.3͑a͒ and 3͑b͒ we display, respectively, the solute and effective mean first hyperpolarizability of water for the individual configurations.For both the solute and the effective mean first hyperpolarizability the fluctuations are large: they oscillates between Ϫ5 and 23 a.u.again illustrating the strong sensitivity to the solvent configurations. In Table V we present the frequency-dependent ͑EFISH͒ second hyperpolarizability for water and acetonitrile in the gas and liquid phase.For the liquid phase we present both the solute and the effective properties.For water we see that the solute second hyperpolarizability is slightly larger than the vacuum results, whereas the effective second hyperpolarizability is smaller.For acetonitrile both the solute and the effective second hyperpolarizability is larger than the vacuum results.For both water and acetonitrile the effective second hyperpolarizability is smaller than the solute in agreement with the trend found for the linear polarizability.In Figs.4͑a͒ and 4͑b͒ we display, respectively, the solute and effective mean second hyperpolarizability of water for the individual configurations.The solute mean second hyperpolarizability oscillates between 1800 and 2600 a.u.whereas the effective mean second hyperpolarizability oscillates between 1300 and 1900 a.u.The solute properties are in general above the gas phase value whereas the effective second hyperpolarizability is always lower than the gas phase value. For all the properties we find that the pure solvent effect, i.e., the difference between the gas phase and the solute properties, increases the properties.On the other hand the induced electric field, i.e., the difference between the solute and effective properties, decreases the properties.Therefore, the electric field induced in the solvent due to the macroscopic electric field produces a screening of the electric field whereas the field from the charge distribution of the solvent molecules produces an enhancement of the electric field.The fact that the first hyperpolarizability is nearly unaffected by the induced electric field is likely to arise from the sensitivity to the short range screening. 20 C. Macroscopic response properties 1. Refractive index We have calculated the refractive index of liquid water and acetonitrile using Eq.͑27͒ and the effective polarizability presented above.For water we used a number density of N d ϭ0.3338ϫ10 23 cm Ϫ1 and N d ϭ0.1153ϫ10 23 cm Ϫ1 was used for acetonitrile.The results for the refractive index of water are n()ϭ1.334and n(2)ϭ1.342,where ϭ0.0428 a.u., obtained from an effective polarizability of 9.95 and 10.17 a.u., respectively.The results are in good agreement with the experimentally determined refractive index at ϭ0.0428 of nϭ1.326 ͑Ref.80͒ and at ϭ0.077 of nϭ1.333. 81or acetonitrile the calculated refractive indices are n ϭ1.362, 1.377, and 1.434 at a frequency of ϭ0.000, 0.0885, and 0.1770 a.u., calculated from effective polariz-abilities of 31.02,32.16 and 36.4 a.u., respectively.The experimental results for the refractive index at ϭ0.0856 is n()ϭ1.347and n(2)ϭ1.384. 82Again, the calculated results are in agreement with the experiments although in this case the calculated results are somewhat larger than the experimental results. It is well known that the Lorentz-Lorenz equation often ͑but not always͒ gives a good relation between the gas phase polarizability and the refractive index.As described in the theory section it is not the gas phase polarizability but rather the effective polarizabilities that should be used in the Lorentz-Lorenz equations.We have shown that by including both the solvent effects and the effect of the local field induced in the solvent due to the electric field in the calculation of the liquid phase polarizability we obtain a value of the effective polarizability close to the gas phase value.It is therefore not surprising that the refractive index are in agreement with the experimental results. Local field factors In order to calculate the nonlinear susceptibilities we need to consider the local field factors described in the theory section.From the refractive index calculated above we can obtain the optical local field factors given by Eq. ͑20͒.For water at ϭ0.0428 we obtain a local field factors of L ϭ1.26 and L 2 ϭ1.27.For acetonitrile the local field factor at ϭ0.000 a.u. is L ϭ1.28, and at ϭ0.0856 the local field factors are L ϭ1.30 and L 2 ϭ1.35. Furthermore, to calculate the EFISH susceptibility in Eq. ͑40͒ we also need to consider the static local field factor given in Eq. ͑21͒.The static local field factor depends on the dielectric constant of the liquid.However, since we have not included the orientation effect due to the static electric field in the calculation it is not likely that we can calculate this quantity correctly.In fact the calculation of the dielectric constant, in particular for water is a highly complicated problem. 57Although the Onsager local field factor is introduced to account for some of the orientational effect it is not included in a consistent manner. We can, however, calculate the Onsager local field factors from the experimental data.For water using the experimental dielectric constant of ⑀ 1 (0)ϭ78 and the refractive index nϭ1.326for water we obtain a static local field factor of L 0 ϭ1.86.We can use this local field factor to estimate the dielectric constant using the susceptibility obtained from Eq. ͑31͒.This give a dielectric constant of ⑀ (1) (0)ϳ43.This is significantly lower than the experimental value.For acetonitrile the experimental dielectric constant is 37.5 and the refractive index is 1.339.Using this we obtain a local field factor of L 0 ϭ1.85.Again using this local field factor we estimate the dielectric constant of acetonitrile to be ⑀ (1) (0) ϳ57.For acetonitrile the estimate is significantly larger than the experimental value.Since the refractive index is predicted correctly and the largest contribution to the dielectric constant is from the dipole moment it is the orientational term which is not described accurately. THG susceptibility Since we calculate the effective second hyperpolarizability by finite differentiations of the frequency-dependent polarizability we cannot obtain the THG second hyperpolarizability directly.The THG second hyperpolarizability can however be obtained by calculating the EFISH second hyperpolarizability at different frequency and then use dispersion formulas 2 to estimate the THG second hyperpolarizability.Here we will estimate the THG susceptibility directly from the EFISH second hyperpolarizability as The estimated THG susceptibility for water is then (3) ϭ1.07ϫ10 Ϫ14 esu at ϭ0.0428 a.u.The experimental result for water relative to the reference value of fused silica is 0.90ϫ SiO 2 (3) measured at the same frequency. 80In the original experimental work 80 a reference value for fused silica of SiO 2 (3) ϭ3.11ϫ10 Ϫ14 esu ͑Ref.83͒ was used, however, recently a value of 1.43ϫ10 Ϫ14 esu ͑Ref.84 and 85͒ has been measured and is believed to be more accurate.Adopting the latter reference value the THG susceptibility for water is (3) ϭ1.29ϫ10 Ϫ14 esu.The estimated THG susceptibility is somewhat lower but in good agreement with the experimental result.Part of the difference can be attributed to the lower dispersion arising from the EFISH second hyperpolarizability compared with the THG second hyperpolarizability.However, it is likely that the effective second hyperpolarizability is too small.The THG susceptibility of liquid water has also been calculated using different continuum and discrete local field models giving results in the range (3) ϳ1 -2ϫ10 Ϫ14 esu depending on the local field model used and in good agreement with the results presented here. 45he experimental result for acetonitrile measured at ϭ0.0239 a.u. is (3) ϭ2.54ϫ10 Ϫ14 esu using the old reference value of SiO 2 (3) ϭ2.79ϫ10 Ϫ14 esu. 83The new reference value at ϭ0.0239 a.u. is SiO 2 (3) ϭ1.15ϫ10 Ϫ14 esu. 84,85Correcting for the differences between the two reference values gives a THG susceptibilty of (3) ϭ1.05ϫ10 Ϫ14 esu.Since we have not calculated the effective second hyperpolarizability for acetonitrile at this frequency we will use the static result to estimate the THG susceptibility instead.The static effective second hyperpolarizability for acetonitrile is ␥ ϭ4891.1 a.u.Using this value the static THG susceptibility is (3) ϭ1.27ϫ10 Ϫ14 esu.The result is larger than but in agreement with the experimental result. EFISH susceptibility Finally, we will use the effective mean first and second hyperpolarizability to calculate the EFISH susceptibility given in Eq. ͑40͒.This gives for water an EFISH susceptibility of (3) ϭ4.1ϫ10 Ϫ14 esu at ϭ0.0428 a.u.For liquid water there has been one EFISH experiment at ϭ0.0428 a.u. 86In that work a value of (3) ϭ17.6 ϫ10 Ϫ14 esu was reported.The measurement was done in the X convention, 51 i.e., no numerical factors in the expansion of the polarization, and relative to a quartz crystal reference where a value of d 11 ϭ0.8ϫ10Ϫ9 esuϭ0.335pm/V was adopted.However, the currently accepted value for quartz is d 11 ϭ0.30pm/V. 87,88Correcting for the difference in the reference value and conventions gives (3) ϭ10.5ϫ10Ϫ14 esu (17.6ϫ2/3ϫ0.3/0.335),i.e., significantly larger than the calculated value.Since the larger contribution comes from the term depending on the dipole moment and first hyperpolarizability it is likely that the-not well describedorientational contribution is responsible for the difference.The agreement between theory and experiment for the THG susceptibility also support this conclusion. For acetonitrile we calculate the EFISH susceptibility at ϭ0.0885 a.u. to be (3) ϭ35.7ϫ10Ϫ14 esu.The experimental result 89 at the same frequency is (3) ϭ13.6 ϫ10 Ϫ14 esu ͑corrected with 2/3 to convert it to the convention used here͒.The calculated value is much larger than the experimental results in agreement with the trend found for the dielectric constant.There has recently been a study of both the microscopic and macroscopic properties of acetonitrile using ab initio method combined with an Onsager model for the solvation where in general a good agreement was found between experiments and the calculated results. 46 V. CONCLUSIONS We have in this work presented an extension of the QM/MM formalism to include the so-called local field factors, i.e., the difference between the macroscopic electric field and the actual electric field felt by the solute molecule.This enables the calculation of effective microscopic properties which can be related to the macroscopic susceptibilities directly comparable with experimental results.By separating the discrete local field into two distinct contribution we can define two different microscopic properties, the so-called solute and effective properties.In the solute properties the pure solvent effect, i.e., effects even when the macroscopic electric field is zero, are accounted for and in the effective both the pure solvent effect and the effect from the induced dipoles in the solvent is accounted for.We have presented results for linear and nonlinear polarizabilities of water and acetonitrile both in the gas phase and in the liquid phase.For all the properties we see that the pure solvent effect, i.e., the difference between the gas phase and the solute properties, gives an increase in the properties.For the induced electric field, i.e., the difference between the solute and effective properties, a decrease in the properties was found.Therefore, the electric field induced in the solvent due to the macroscopic electric field produced a screening of the electric field whereas the field from the charge distribution of the solvent molecules produces an enhancement of the electric field.Furthermore, we have presented results for the refractive index, third-harmonic generation ͑THG͒, and electric field induced second harmonic generation ͑EFISH͒ for pure water and acetonitrile.We find in general good agreement between the calculated and experimental values for the refractive index and the THG susceptibility.For the EFISH susceptibility the differences between experiment and theory is larger due to neglect of the orientational effect arising from the static electric field. FIG. 1 . FIG. 1.The dipole moment of water in Debye.Top line is the averaged results and the bottom line is the gas phase value. a Results are taken from ͒ b Results are taken from ͒ c TABLE II.The dipole moments of water and acetonitrile in gas and liquid phase in atomic units. TABLE III . The frequency-dependent polarizability of water and acetonitrile in the gas and liquid phase in atomic units.For water the frequency is ϭ0.0428 and for acetonitrile ϭ0.0885 a.u. TABLE IV . The first hyperpolarizability ␤(Ϫ2;) for water and acetonitrile in the gas and liquid phase.For water the frequency is ϭ0.0428 and for acetonitrile ϭ0.0885 a.u.All results are in atomic units. TABLE V . The second hyperpolarizability, ␥(Ϫ2;,0), for water and acetonitrile in the gas and liquid phase.For water the frequency is ϭ0.0428 and for acetonitrile ϭ0.0885 a.u.All results are in atomic units.	2017-06-16T11:56:17.785Z	2005-01-15T00:00:00.000	{ "year": 2005, "sha1": "06ff72650e15ba5560b71f14dc81b20d25052a55", "oa_license": "CC0", "oa_url": "https://dugi-doc.udg.edu/bitstream/10256/8419/1/MicroscopicMacroscopic.pdf", "oa_status": "GREEN", "pdf_src": "Anansi", "pdf_hash": "be1d49b040b25031e5a4d9bdf9d98f6fd5bdaa8b", "s2fieldsofstudy": [ "Chemistry", "Physics" ], "extfieldsofstudy": [ "Chemistry", "Medicine" ] }
226521546	pes2o/s2orc	v3-fos-license	Decomposition of Certain Complete Graphs and Complete Multipartite Graphs into Almost-bipartite Graphs and Bipartite Graphs In his classical paper [14], Rosa introduced a hierarchical series of labelings called ρ, σ, β and α labeling as a tool to settle Ringel’s Conjecture [13] which states that if T is any tree with m edges then the complete graph K2m+1 can be decomposed into 2m + 1 copies of T . Inspired by the result of Rosa [14] many researchers significantly contributed to the theory of graph decompositions using graph labelings. In this direction, in 2004, Blinco et al. [6] introduced γ-labeling as a stronger version of ρ-labeling. A function g defined on the vertex set of a graph G with n edges is called a γ-labeling if (i) g is a ρ-labeling of G, (ii) G is a tripartite graph with vertex tripartition (A,B,C) with C = {c} and b̄ ∈ B such that {b̄, c} is the unique edge joining an element of B to c, (iii) g(a) < g(v) for every edge {a, v} ∈ E(G) where a ∈ A, (iv) g(c)− g(b̄) = n. Further, Blinco et al. [6] proved a significant result that the complete graph K2cn+1 can be cyclically decomposed into c(2cn + 1) copies of any γ-labeled graph with n edges, where c is any positive integer. Recently, in 2013, Anita Pasotti [4] introduced a generalisation of graceful labeling called d-divisible graceful labeling as a tool to obtain cyclic G-decompositions in complete multipartite graphs. Let G be a graph of size e = d . m. A d-divisible graceful labeling of the graph G is an injective function g : V (G) → {0, 1, 2, . . . , d(m + 1) − 1} such that {\|g(u) − g(v)\|/{u, v} ∈ E(G)} = {1, 2, . . . , d(m+ 1)− 1}\{m+ 1, 2(m+ 1), . . . , (d− 1)(m+ 1)}. A d-divisible graceful labeling of a bipartite graph G is called as a d-divisible α-labeling of G if the maximum value of one of the two bipartite sets is less than the minimum value of the other one. Further, Anita Pasotti [4] proved a significant result that the complete multipartite graph K( e d +1)×2dc can be cyclically decomposed into copies of d-divisible α-labeled graph G, where e is the size of the graph G and c is any positive integer (K( e d +1)×2dc contains e d + 1 parts each of size 2dc). Motivated by the results of Blinco et al. [6] and Anita Pasotti [4], in this paper we prove the following results. i) For t ≥ 2, disjoint union of t copies of the complete bipartite graph Km,n, where m ≥ 3, n ≥ 4 plus an edge admits γ-labeling. ii) For t ≥ 2, t-levels shadow graph of the path Pdn+1 admits d-divisible α-labeling for any admissible d and n ≥ 1. Further, we discuss related open problems. Abstract In his classical paper [14], Rosa introduced a hierarchical series of labelings called ρ, σ, β and α labeling as a tool to settle Ringel's Conjecture [13] which states that if T is any tree with m edges then the complete graph K 2m+1 can be decomposed into 2m + 1 copies of T . Inspired by the result of Rosa [14] many researchers significantly contributed to the theory of graph decompositions using graph labelings. In this direction, in 2004, Blinco et al. [6] introduced γ-labeling as a stronger version of ρ-labeling. A function g defined on the vertex set of a graph G with n edges is called a γ-labeling if (i) g is a ρ-labeling of G, Further, Blinco et al. [6] proved a significant result that the complete graph K 2cn+1 can be cyclically decomposed into c(2cn + 1) copies of any γ-labeled graph with n edges, where c is any positive integer. Recently, in 2013, Anita Pasotti [4] introduced a generalisation of graceful labeling called d-divisible graceful labeling as a tool to obtain cyclic G-decompositions in complete multipartite graphs. Let G be a graph of size e = d . Further, Anita Pasotti [4] proved a significant result that the complete multipartite graph K ( e d +1)×2dc can be cyclically decomposed into copies of d-divisible α-labeled graph G, where e is the size of the graph G and c is any positive integer (K ( e d +1)×2dc contains e d + 1 parts each of size 2dc). Motivated by the results of Blinco et al. [6] and Anita Pasotti [4], in this paper we prove the following results. i) For t ≥ 2, disjoint union of t copies of the complete bipartite graph K m,n , where m ≥ 3, n ≥ 4 plus an edge admits γ-labeling. ii) For t ≥ 2, t-levels shadow graph of the path P dn+1 admits d-divisible α-labeling for any admissible d and n ≥ 1. Further, we discuss related open problems. Introduction Terms which are not defined here can be found in [15]. In an attempt to settle the Ringel's conjecture [13] which states that if T is any tree with m edges then the complete graph K 2m+1 can be decomposed into 2m + 1 copies of T , in his classical paper [14], Rosa introduced a series of labelings called α, β, σ, ρ-labeling. Let G be a graph with n edges. A one-to-one function g from V (G) to {0, 1, 2, . . . , n} is called a β-labeling of G if {\|g(u) − g(v)\|/{u, v} ∈ E(G)} = {1, 2, . . . , n}. A β-labeling g of a graph G with n edges is called an α-labeling if there exists an integer k such that for every edge {u, v} ∈ E(G) either g(u) ≤ k < g(v) or g(v) ≤ k < g(u). Given two vertices u and v by uv we denote the edge {u, v}. It is clear that α-labeling is a stronger version of β-labeling. β-labeling was later called as graceful labeling by Golomb [12] and this term is most widely used now. ρ-labeling is weaker version of graceful labeling. The precise definition of ρ-labeling is given below. Let G be a graph with n edges. A one-to-one function g from V (G) to {0, 1, 2, . . . , 2n} is called a Further, Rosa [14] proved the following two significant theorems. Theorem 1.1. Let G be a graph with n edges. Then there exists a cyclic G-decomposition of the complete graph K 2n+1 if and only if G has a ρ-labeling. If G is a graph with n edges that has an α-labeling, then the complete graph K 2cn+1 can be cyclically decomposed into subgraphs isomorphic to G, where c is an arbitrary natural number. The interesting part of α-labeled graphs with n edges is that they not only decompose complete graphs K 2cn+1 but also decompose the complete bipartite graphs K an,bn . This interesting result proved by El-Zanati and Vanden Eynden [9] is precisely stated in the following theorem. Theorem 1.3. If a graph G with n edges has an α-labeling then there exists a cyclic decomposition of the complete bipartite graph K an,bn into subgraphs isomorphic to G, where a and b are arbitrary positive integers. These results attracted many researchers to significantly contribute in theory of graph decompositions using graph labelings. It is clear from the definition of α-labeling that if a graph G admits α-labeling then it must be necessarily bipartite. This restriction prompted Blinco et al. [6] to introduce γ-labeling in order to achieve cyclic G-decompositions in K 2cn+1 , where G is a non-bipartite graph, c is any positive integer and n is the number of edges of the graph G. A function g defined on the vertex set of a graph G with n edges is called a γ-labeling if (i) g is a ρ-labeling of G, (ii) G is a tripartite graph with vertex tripartition (A, B, C) with C = {c} andb ∈ B such that {b, c} is the unique edge joining an element of B to c, Motivated by the above result of Blinco et al. [6], the almost-bipartite graphs P n + e, n ≥ 4, K m,n +e, m ≥ 2, n > 2, C 2k+1 , k ≥ 2, C 2m +e, m > 2, C 3 ∪C 4m , m > 1, C 2k+1 ∪C 4n+2 , k ≥ 1, n ≥ 1 are found to have γ-labeling (refer [5], [6], [7], [8], [10]). (A graph is said to be almost-bipartite if the removal of a particular edge makes the graph bipartite). For survey on γ-labeling refer the survey on graph labelings by Gallian [11]. Motivated by the results of Blinco et al. [6], in this paper we prove that for t ≥ 2, disjoint union of t copies of the complete bipartite graph K m,n , where m ≥ 3, n ≥ 4 plus an edge admits γ-labeling. Recently, in 2013, Anita Pasotti [4] introduced a generalisation of graceful labeling called d-divisible graceful labeling as a tool to obtain cyclic G-decomposition in complete multipartite graphs. Let G be a graph of size e = d . m. An injective function (m + 1)} is called as a d-divisible graceful labeling of the graph G. A d-divisible graceful labeling of a graph G can exist only if d is a divisor of the size e of G, hence, for this reason, any divisor d of e is said to be admissible for the existence of a d-divisible graceful labeling of G. A d-divisible graceful labeling of a bipartite graph G is called as a d-divisible α-labeling of G if the maximum value of one of the two bipartite sets is less than the minimum value of the other one. can be cyclically decomposed into copies of the d-divisible graceful labeled graph G, where e is the size of the graph G. Theorem 1.6. (Anita Pasotti [4]) The complete multipartite graph K ( e d +1)×2dc can be cyclically decomposed into copies of the d-divisible α-labeled graph G, where e is the size of the graph G and c is any positive integer. In the literature survey [11], one can observe that a very few families of graphs are identified to have d-divisible α-labeling. Anita Pasotti [4] has proved that path and star admit d-divisible α-labeling for any admissible d. She [3] also proved that for any integer k ≥ 1 and m ≥ 2, C 4k × P m admits (2m − 1)-divisible α-labeling. In [1] and [2], Anna Benini and Anita Pasotti proved the following results. A hairy cycle of size e admits an e-divisible α-labeling if and only if it is bipartite. The hairy cycle H(2t, λ) admits d-divisible α-labeling for any admissible d. The ladder L 2k has 2-divisible α-labeling if and only if k is even. Inspired by the decomposition theorems proved by Anita Pasotti, in this paper we prove that for t ≥ 2, t-levels shadow graph of the path P dn+1 admits d-divisible α-labeling for any admissible d and n ≥ 1. t-levels shadow graph of a graph is defined as follows. t-levels shadow graph of a graph G, denoted S t (G) is obtained by taking t ≥ 2 copies G 1 , G 2 , . . . , G t of G and joining each vertex v ij in G i to the copies of its adjacent vertices in G i+1 , for 1 ≤ j ≤ n and 1 ≤ i ≤ t − 1, where n = \|V (G)\|. 2 γ-labeling of disjoint union of complete bipartite graphs plus an edge In this section we prove that disjoint union of t copies of the complete bipartite graph K m,n , where m ≥ 3 and n ≥ 4 plus an edge admits γ-labeling. Theorem 2.1. For t ≥ 2, disjoint union of t copies of a complete bipartite graph with one part containing at least three vertices and another part containing at least four vertices, plus an edge admits γ-labeling. Proof. Consider the complete bipartite graph K m,n , where m ≥ 3, n ≥ 4. . . , v in } be the two parts of the i-th copy K i m,n of the complete bipartite graph K m,n . Clearly, U and V are the two parts of the disjoint union of the t copies of K m,n , denoted by Join the vertices v 11 and v 12 by an edgeê. Denote the new graph thus obtained by ( First we define the labels of the vertices in the set U in the following way. For 1 ≤ j ≤ m, define g(u 1j ) = 2(j − 1) and g(u 2j ) = 2j + 1. Now we define the labels of the vertices in the set V in the following manner. We define the labels of the vertices v 2k , for k, 2 ≤ k ≤ n in two cases depending on n is even or odd. Case 1. n is even define the labels of the vertices v ik , for each k, 2 ≤ k ≤ n in the following way. We prove that the vertex labels of the graph ( t i=1 K i m,n ) +ê are distinct depending on n is even or odd. Hence the vertex labels of the graph ( t i=1 K i m,n ) +ê are distinct. Observation 2. Edge labels of ( The edge v 11 v 12 has the label N . We prove that the edge labels of t i=1 K i m,n are distinct in two cases depending on n is even or odd. Case i. n is even The labels of the edges in the first copy K 1 m,n can be arranged as a sequence, S 11 : ((N − 1, N − 2, N − 3, . . . , N + 2m + 1 − mn, N + 2m − mn), (2m, 2m − 1, . . . , 2, 1)). For each i, 2 ≤ i ≤ t, the labels of the edges in the i th copy K i m,n can be arranged as a sequence, The labels of the edges in the above sequences together with the label of the edge v 11 v 12 , \|g(v 11 ) − g(v 12 )\| = N can be rearranged as a monotonic decreasing sequence S : (N, N − 1, N − 2, . . . , 3, 2, 1). Thus the edge labels are distinct when n is even. Hence the edge labels of the graph ( Observation 3. g is a γ-labeling. In order to prove that g is a γ-labeling, we partition the vertex set V (( d-divisible α-labeling of t-levels shadow graph of path In this section we prove that for t ≥ 2, t-levels shadow graph of the path P dn+1 , S t (P dn+1 ) with d ≥ 1, n ≥ 1 admits d-divisible α-labeling for all d ≥ 1. Theorem 3.1. For t ≥ 2, the t-levels shadow graph of the path P dn+1 , S t (P dn+1 ) with d ≥ 1 and n ≥ 1 admits d-divisible α-labeling for all d ≥ 1. For the convenience, we let Then the t-levels shadow graph of the path P dn+1 , S t (P dn+1 ) has the vertex set Therefore, \|V (S t (P dn+1 ))\| = t\|V (P dn+1 )\| = t(dn + 1). By the definition of the t-levels shadow graph of the path P dn+1 , the graph S t (P dn+1 ) can be visualised as t copies of the path P dn+1 and a pair of t − 1 copies of P dn+1 which connect the vertices of the copies G i and G i+1 of the path P dn+1 , 1 ≤ i ≤ t − 1. Therefore, \|E(S t (P dn+1 ))\| = tdn + 2(t − 1)dn = (3t − 2)dn. For all the remaining vertices of S t (P dn+1 ) we define g depending on d = 1 and d > 1. When d = 1 define g as follows. where s = dn, if dn + 1 is even dn + 1, if dn + 1 is odd, then the above sequence forms a strictly increasing sequence. Hence the vertex labels of S t (P dn+1 ) are distinct. From the above arrangement of vertex labels observe that dn+1) ), when dn + 1 is even; while when dn + 1 is odd, We prove that the edge labels of S t (P dn+1 ) are distinct depending on d = 1 and d > 1. Case 1. d = 1 When n is even, the edges of the graph S t (P dn+1 ) can be arranged as the following sequence, When n is odd, the edges of S t (P dn+1 ) can be arranged as the following sequence, Then from the definition of g for both the cases we have the corresponding edge label sequence, (N, N − 1, N − 2, . . . , 3, 2, 1). Hence, it is clear that the edge labels are distinct. Therefore, when d = 1, g is a 1-divisible α-labeling of S t (P dn+1 ). That is, g is an α-labeling of the graph S t (P dn+1 ). Case 2. d > 1 In order to show that the edge labels of the edges of S t (P dn+1 ) are distinct, we partition the edge set of S t (P dn+1 ) into d subsets of the edge set of S t (P dn+1 ) and they are arranged as d sequences. Consequently, their corresponding edge labels are also arranged as d sequences. When n is even then we consider the first edge sequence to be the following sequence When n is odd then we consider the first edge sequence to be the following sequence Then from the definition of g for both the cases we have the corresponding edge label sequence, When n is even then we consider the second edge sequence to be the following sequence When n is odd then we consider the second edge sequence to be the following sequence Then from the definition of g for both the cases we have the corresponding edge label sequence, When n is even then we consider the third edge sequence to be the following sequence When n is odd then we consider the third edge sequence to be the following sequence Then from the definition of g for both the cases we have the corresponding edge label sequence, In general, we consider the j th edge sequence, for 4 ≤ j ≤ d − 2 depending on n and j. Case i. n is even or n is odd and j is even Then we consider the j th edge sequence to be the following sequence Case ii. n and j are odd Then we consider the j th edge sequence to be the following sequence Then from the definition of g for all the above cases we have the corresponding edge label sequence, S j : ((d−j)(3t−2)n+d−(j +1), (d−j)(3t−2)n+d−(j +2), (d−j)(3t−2)n+d−(j +3), . . . , (d−(j +1))(3t−2)n+d−(j +2), (d−(j +1))(3t−2)n+d−(j −1), (d−(j +1))(3t−2)n+d−j). Now we consider the (d − 1) th edge sequence depending on n is even or odd. S d ). Then we observe that S forms a monotonically decreasing sequence. Also observe that none of the terms (d − 1)((3t − 2)n + 1), (d − 2)((3t − 2)n + 1), . . . , 3((3t − 2)n + 1), 2((3t − 2)n + 1), (3t − 2)n + 1 appear in the combined sequence S. Thus, g is a d-divisible α-labeling of S t (P dn+1 ) for any admissible d > 1. Therefore the graph S t (P dn+1 ) admits d-divisible α-labeling for any admissible d. Discussion In this section we pose two open problems for further research. In Theorem 2.1 we have proved that for t ≥ 2, disjoint union of t copies of the complete bipartite graph K m,n plus an edge, ( t i=1 K i m,n ) +ê admits γ-labeling. In this direction investigating the following question will be useful for achieving a generalised result. Is it true that disjoint union of t copies of an α-labeled graph G plus an edge, t ≥ 2, admits γ-labeling? In Theorem 3.1 we have proved that for t ≥ 2, the t-levels shadow graph of the path P dn+1 with d ≥ 1, n ≥ 1 admits d divisible α-labeling for all d ≥ 1. It is evident that the path P dn+1 admits α-labeling for all d ≥ 1, n ≥ 1. This observation tempts us to ask the following question to understand d-divisible α-labeled graphs. What are the α-labeled graphs whose t-levels shadow graph admits d divisible α-labeling for all values of d?	2020-08-06T09:08:24.239Z	2020-01-01T00:00:00.000	{ "year": 2020, "sha1": "19dc2842dbeb3a529c1867dbb359f0c8aaa2e816", "oa_license": "CCBY", "oa_url": "https://digitalcommons.georgiasouthern.edu/cgi/viewcontent.cgi?article=1127&context=tag", "oa_status": "GOLD", "pdf_src": "Anansi", "pdf_hash": "833d7bdac51ed97dbf457cdabc74f525950d1153", "s2fieldsofstudy": [ "Mathematics" ], "extfieldsofstudy": [ "Mathematics" ] }
243657024	pes2o/s2orc	v3-fos-license	Objectively Measured Association Between Air Pollution and Physical Activity and Sedentary Behavior Among College Students in Beijing Background: Air pollution has become a major environmental health risk factor, notably in China. Air pollution potentially has the impact of human populations’ health behavior. Gaps in scientic literature remain regarding more accurately estimates the relationship between air pollution and sedentary behavior in China. The purpose of this study is to examine the association between hourly air pollution AP on hourly physical activity (PA) and sedentary behavior (SB) among college students in Beijing, China. The secondary aim was to examine such associations varied at specic times. Methods: A total of 340 participants were recruited from the Tsinghua University, in Beijing, China. Accelerometers provided PA measures, including moderate-to-vigorous physical activity (MVPA), walking steps, energy expenditure and sedentary time for 7 consecutive days. Corresponding AP data by the Beijing Municipal Ecological Environment Bureau in the closed site (Wan Liu site) at Tsinghua University were collected to include average hourly air quality index (AQI) and PM 2.5 (µg/m³). Associations were estimated using linear individual xed-effect regressions. Results: A one level increase in hourly air quality index (AQI) was associated with a reduction in one-hour MVPA by 0.083 minutes (95% CI =[-0.137, -0.029]), 8.8 walking steps (95% CI = [-15.0, -2.6]), and 0.65 kcals of energy expenditure (95% CI =-[1.03, -0.27]). A 10µg/m³ increase in AP concentration in hourly PM 2.5 was associated with a reduction in one-hour MVPA by 0.021 minutes (95% CI = [-0.033, -0.010]), 2.2 walking steps (95% CI = [-3.5, -0.9]), 0.170 kcals of energy expenditure (95% CI = [-0.250, -0.089]), and an increase in one-hour SB by 0.045 (95% CI = [0.005, 0.0845]). At a specic time, there are stronger negative associations of AQI and PM 2.5 to PA at 8 am, 4 pm, 5 pm and 7 pm. Similarly, stronger positive associations were found at one hour AQI and PM 2.5 with SB at 8 am, 9 am, 11 am, and 7 pm. Conclusions: AP may discourage PA and increases SB among freshman students living in Beijing, China. The impact of AP on PA and sedentary behavior at a specic time may be different. Introduction Air pollution (AP) has become a major environmental health risk factor for overall health worldwide. AP levels in China have increased rapidly due to the industrialization use of fossil fuel and population growth [1]. Previous evidences has showed that exposure to AP has been detrimental to various health outcomes e.g., cardiovascular disease, stroke, lung cancer, respiratory disease, all-cause mortality, sleep apnea, and depression [2][3][4][5][6][7][8][9][10]. Strong evidence indicates that engaging in regular physical activity (PA) has many health bene ts including increased effective weight management, reduced risk of all-cause mortality, and prevention and management of chronic diseases, e.g., cardiovascular diseases, diabetes, colon and breast cancer, hypertension, coronary heart disease, and osteoarthritis [11][12][13][14]. While there are several health bene ts to PA, performing PA under high levels of AP may accelerate the risks for adverse health effects such as asthma attacks and heart or lung pathologies [15][16][17]. Outdoor parks and playgrounds are common places to perform PA in China [18], and the presence of AP in China may further discourage young adults from engaging in regular PA and exercise [19]. Substantial research has measured the effects of AP on health outcomes, but few studies have determined its impact on health-related behaviors, speci cally PA and SB behavior. There are three major gaps in the scienti c literature that need for further investigation. First, existing research on the associations between AP, PA and SB are limited. Second, most previous studies have used a self-reported survey method [26], therefore, existing studies were subject to social desirability bias and limited by the frequency of PA performance and SB data (i.e., week by week, month by month, or year by year). This study is the rst to provide precise, objectively measured data through use of the digital accelerometers to measure PA and SB among university freshman students in China. A digital accelerometer omits the chance for social desirability bias while providing minute-by-minute measurements of PA and SB. Third, to date, all previous studies estimated the impact of AP on health-related behaviors week on a week, month by month, or year by year basis. There is currently no study which investigates the relationship between AP and PA and SB at a speci c time. This study reported the associations between AP, PA and SB among university freshman students living in Beijing, China. Objectively measured data were collected by using of the digital accelerometers during this study. The nal sample size was 340 participants. Hourly AP data and minute-by-minute PA and SB data were measured. We hypothesized that in response to high levels of AP, university freshmen reduced their PA behaviors and increased their SB. We also hypothesized that such associations varied at a speci c time. Participants and Sampling Procedure The study was conducted from the period of November 2017 to April 2018. All freshmen students are required to take the same physical education class in Tsinghua University. 344 freshmen students who were enrolled in the required physical education class were recruited as participants. Participants were . Upon acceptance, the subjects were asked to visit our lab to get a wGT3X-BT device tted. Height and weight were measured to the nearest 0.1 kg using a Seca beam scale (Tsinghua Tongfang, S5000, China). Each participant completed one paper-pencil based health survey on demographic status (age, gender, ethnicity, lifestyle smoking, drinking, and mental and physical health conditions). Of the 344 recruited, 340 subjects had completed the study (4 participants' data were excluded due to 2 devices errors, and 2 missing accelerometers that were not returned). All participants gave informed consent, and the study was approved by the Tsinghua University Institutional Review Board (IRB #2017DX02_11). PA and SB measurement The participants were instructed to wear a wGT3X-BT accelerometer over the right hip on a waist band for at least 10 hours a day (excluding sleep time) over 7 consecutive days. The participants removed the device it only during showers, bathing, swimming or other water activities. The recording epoch was set to record by one minute. Hourly PA and SB were calculated using the minutes' data. For all subjects, absolute time in moderate-to-vigorous physical activity (MVPA), walking steps, kcals in energy expenditure and absolute time spent in SB were estimated using the device. Nonwear hour periods were de ned as 60 consecutive minutes of zero activity intensity count at 1 METs and 0 kcals in one hour [27]. Environmental measures AP data and other environmental measures were selected including-air quality index (AQI), PM 2.5, and average daytime temperature (°C). Hourly AQI and PM 2.5 data were provided by the Beijing Municipal Ecological Environment Bureau. The data collection site is Wan Liu was the Hai Dian district site, which is approximately 5 km from Tsinghua University. The air quality measurement data we collected are concurrent with PA measurements. Daytime temperature was retrieved from the China Meteorological Administration. The AQI is an index for reporting daily air quality [28]. It indicates you how clean or polluted the surrounding air is, and which associated health effects might be of concern for you. The AQI focuses on health effects that may be experienced within a few hours or days after breathing polluted air. The China Meteorological Administration calculates the AQI using the following ve major air pollutants: groundlevel ozone, particle pollution (also known as particulate matter), carbon monoxide, sulfur dioxide, and nitrogen dioxide. Statistical Methods Descriptive statistics were computed including means, SD, and percentages. All descriptive statistics were compared for the characteristics of the overall sample. Chi-square tests were conducted to compare categorical variables. ANOVA tests and t-test were used to compare for continuous variables. One-way repeated measures ANOVA tests were conducted to compare the differences in hourly time between 7 am and 11 pm. Adjusted linear individual xed-effect regressions were performed to reveal the associations between the accelerometer data and AP data. The data of participant controlled for age, gender, BMI, selfrated physical health, self-rated mental health, smoking, drinking, temperature, and temporal order for participants. The key independent variables were AQI and PM 2.5 during this time. We performed six levels (1)(2)(3)(4)(5)(6) in AQI and used 10 µg/m³ in PM 2.5 in the model analysis. Individual-level time-variant covariates and environmental measures including average daytime temperature were controlled for the aforementioned. Each outcome variable was then analyzed using separate regression and were strati ed by speci c time (from 7 am to 11 pm). Compared to the conventional pooled cross-sectional regression, individual xed-effect regression was selected due to its used of within-individual variations in hourly PA and SB to identify the impacts of AP concentration, thus removing potential omitted variable bias due to differences in time-invariant individual characteristics such as habits, and personal preferences. Table 1 presents the characteristics of the participants. Among the 340 student participants, accounting for more than two-thirds (70.6%) were male. Participants complied to wearing a wGT3X accelerometer for 42291 hours over 7 continuous days. The mean age of the participants was 18.4 (SD = 1.0). The mean participants' BMI was 21.6 kg/m 2 (SD = 3.1) where male BMI was signi cantly higher than females' (p < 0.001). Only 0.88% of participants were reported smokers and 2.9% was reported drinkers. The mean selfrated physiological health score was 5.5 (SD = 1.8) and the mean self-rated mental health score was 6.5 (SD = 1.9). The PA and SB Variations. Table 3 presents the mean variations of participants' PA and SB in the study. As illustrated, there are large variations in PA and SB. For example, the mean minutes of the participant's one-hour MVPA was 3.5 (SD = 4.5). One-hour mean minutes of MVPA ranged largely from 1.6 (SD = 2.7) at 11 pm to 6.1 (SD = 8.3) at 5 pm in the participants (p < 0.001). The mean steps of the participant's one-hour walking was 394 (SD = 627). Table 1 shows that one-hour mean steps of walking ranged largely from 147 (SD = 303) at 11 pm to 710 (SD = 947) at 5 pm in the participants (p < 0.001). Similarly, the mean participant's energy expenditure in one-hour was 19.8 (SD = 39.1). One-hour mean energy expenditure in kcal ranged largely from 9.2 (SD = 19.5) at 11 pm to 38.0 (SD = 66.2) at 5 pm among the participants (p < 0.001). Whereas the mean minutes of participant's one-hour SB was 31.5 (SD = 19.7). One-hour mean minutes of SB ranged largely from 38.6 (SD = 18.2) at 11 pm to 25.4 (SD = 17.1) at 12 pm among the participants (p < 0.001). Impact of AQI on PA and SB. Table 4 shows the estimated effects of air quality index (AQI) on individual-level outcomes of hourly PA and SB using linear individual xed-effect regressions. AQI was found to be signi cant and negatively associated with one-hour PA among participants. A one level increase in AQI was linked to a signi cant reduction in minutes of one-hour MVPA, in steps of one-hour walking, and in kcals of one-hour energy Table 1 (i.e., age, gender, BMI, drinking status, drinking status, self-rated physical health, self-rated mental health, temporal order for participants) and environmental variables (average temperature). * P<0.05; P<0.01; * P<0.001 Descriptive Statistics The impact of AQI on individual-level one-hour PA at a speci c time was different. AQI was found to be more negatively associated with participants' one-hour PA at 8 am, 4 pm, 5 pm and 7 pm. Speci cally, a one level increase in AQI was linked with a signi cantly reduction in steps of one-hour walking at 8 am, , respectively. However, AQI was found to be positively associated with participants' one-hour PA at 10 am and 3 pm. There was no signi cant relationship between AQI and one-hour SB among participants. Impact of PM 2.5 on PA and SB. Table 1 (i.e., age, gender, BMI, drinking status, drinking status, self-rated physical health, self-rated mental health, temporal order for participants) and environmental variables (average temperature). * P<0.05; P<0.01; * P<0.001 The impact of PM 2.5 on individual-level one-hour PA at speci c time was also different. PM 2.5 was found to be more negatively associated with participants' one-hour PA at 8 am, 4 pm, 5 pm and 7 pm. The impact of PM 2.5 on individual-level one-hour SB at a speci c time was also different. PM 2.5 was found to be more positively associated with participants' one-hour SB at 9 am, 11 pm, 5 pm and 7 pm. Speci cally, a 10 µg/m³ increase in PM 2.5 was linked with a signi cantly increase in minutes of one-hour Discussion The purpose of this study was to examine the impact of AP level on PA and SB among university freshman students in Beijing, China from November 2017 to April 2018 using objectively-measured PA and SB. Our study found a signi cantly negative relationship between AP and PA and a positive relationship between AP and SB. With a one level increase in AQI and a 10 µg/m³ increase in PM 2.5 , hourly total minutes of MVPA, walking steps and kcals of energy expenditure were signi cantly reduced. A 10 µg/m³ increase in PM 2.5 was associated with a signi cantly increase in SB among participants. The impact of AP on individual-level one-hour PA and SB behavior at a speci c time was different. To our best knowledge, this is the rst study to use objective methods to determine the effect of hourly AP on PA and SB. In addition, this is the rst study to estimate the impact of AP on PA and SB at a speci c time. Our ndings on the negative relationship between AP and PA are consistent with existing literature [19,[30][31][32][33]. In our study, we found that a one-hour AQI increase one level was associated with a decrease by 9 walking steps in one hour. This study additionally found that a 10 µg/m³ increase in PM 2.5 was linked with reduction by 2 walking steps in one hour. Two previous U.S. studies linked one unit (< 10 µg/m³) monthly average PM 2.5 increase of AP to be associated with decreasing 0.46% leisure time PA using a cross sectional study from the US Behavioral Risk Factor Surveillance System (BRFSS) survey [31,34]. Evidence from our previous follow-up studies also found a one unit (44.72-56.6 µg/m³) increase in PM 2.5 to discourage outdoor PA 110.67 PASE scores among older adults in China [30] and to reduce 32.45 weekly MVPA among Chinese college students [19,35]. However, these previous studies were limited by the potential social bias of self-report measures of PA and therefore were have not been able to examine the hourly effects of AP on objective PA. Only two studies with objectively measured data have reported that the association between AP and PA. Consistent with this study, a study of 153 middle-age adult users of an exercise app reported that AQI increase was associated with participates' reduction in outdoor PA, such as running, biking, and walking [33]. With this current analysis, we can more precisely use accelerometers to estimate the impact of AQI and PM 2.5 on MVPA, energy expenditure, and steps rather than use an exercise app associated with PA. Inconsistent with this study, another study showed that PM 2.5 increase had no impact on PA in Beijing among 40 Han Chinese participants in the mean age of 31 years using GT3X accelerometers [36]. A possible explanation for this difference could be that the study had a relatively small sample size and could not account for differences among participants. Based on 340 participants' wGT3X accelerometers data, we can more con dently suggest that an increase in AQI and PM 2.5 increase was associated with a reduction of PA in MVPA, energy expenditure, and walking steps. This study con rmed ndings from previous studies regarding the positive correlation between AP and SB [37][38][39]. This nding suggests that a one-hour 10 µg/m³ PM 2.5 increase was associated with an increase in SB by 0.045 minutes in one hour. Consistent with our previous research, an increase in AP concentration in PM 2.5 by one unit (81.16 µg/m³) was associated with an increase in total weekly hours of SB by 6.24 hours among a large sample (12,174) of university freshmen in China based on a cohort study survey [37]. To our knowledge, this is one of the rst studies to investigate the impact of AP on SB by hourly use objectively measured GT3X accelerometers. Because of con icting results in this emerging area and somewhat preliminary of this nding, additional investigations are necessary to fully explore the effect of AP on SB among different groups. The impact of AP on individual-level PA and SB at a speci c time was different. This study is the rst to examine the association between AP on PA and SB at a speci c time. Stronger negative associations of AQI and PM 2.5 AP with MVPA, walking steps and energy expenditure in the morning before 8 am, at 4 pm, at 5 pm and at 7 pm were found. Similarly, stronger positive associations of one hour AQI and one hour PM 2.5 on SB at 8 am, 9 am, 11 am, and 7 pm were found. In this study, all participants were recruited from freshmen. Typically, a substantial proportion of freshmen do not have class before 8 am in the morning, and/or after 4 pm. Participants in this study could choose PA or SB behavior according to their own preferences of activities during leisure time. Similar time-speci c relationships between built environment and PA were found in the previous studies [40,41]. However, it is interesting that positive associations of AQI and PM 2.5 were found with MVPA, walking steps and energy expenditure in the morning at 10 am and 3 pm. Yet, negative associations of AQI and PM 2.5 were found with SB. This could be explained by the fact that there are classes, including physical education class, for a substantial proportion of a freshman students' schedule between 10 am in the morning and 3 pm in the afternoon. Freshmen often engage in PA when traveling to class (e.g. walking or bicycling to or from class) and may perform more exercise in physical education class, regardless of AP. Therefore, 10 am in the morning and 3 pm in the afternoon associations between AP and PA or SB observed in this study are logical. Thus, this study con rms that the impact of AP on PA and SB are different from patterns of in time-speci c associations. This nding is closer to the 'true' potential effects of AP on PA and SB. First, the strengths of this study reside in its objectively measured PA-related behavior and precise reporting of data. Most existing studies on the impact of AP on PA and SB have used subjective methods, allowing for uncontrolled confounding bias due to self-report and limited frequency of PA data. Second, this is one of the rst studies to measure the impact of AP on PA and SB by one hour using objective methods. Third, this is the rst study to examine time-speci c results on the relationship between AP and objectively measured PA and SB. However, a few limitations to this study should be noted. First, on the one hand, we didn't monitor indoor air pollution and may have bias in this study. A subject's physical activity participation may not vary by air pollution if an individual is active indoors. On the other hand, all freshmen (the subjects of this study) in Tsinghua University live in the same campus, live in the similar dormitories, take similar transportation (e.g., bicycle), and have similar classrooms (6th teaching building in Tsinghua). The subjects sharing similar classrooms, transportation means, and dormitories may partially offset the disruption the in uence of indoor air pollution levels. Second, we could not identify the speci c types of PA and SB through using accelerometers to assess PA and SB. Third, all participants were recruited by a convenience sampling. Freshman students from one university cannot represent all university students in Beijing, China or nationwide therefore limiting the generalizability of the study's ndings. Future studies are warranted to produce more generalized estimates. Conclusions In conclusion, the present study aimed to examine the relationship between AP and hourly behavioral modi cation to relate objectively-assessed PA and SB among university freshmen in Beijing, China. A negative association between hourly AQI and PM 2.5 and hourly PA in minutes of MVPA, walking steps and energy expenditure among study participants were found. A positive relationship between hourly AQI and Trend for walking steps in speci c time in one day among the freshman.	2020-06-18T09:06:18.563Z	2020-06-17T00:00:00.000	{ "year": 2020, "sha1": "4cbee252cfcd2ff6c6b664973e8419144eebb458", "oa_license": "CCBY", "oa_url": "https://www.researchsquare.com/article/rs-33667/v1.pdf?c=1631857961000", "oa_status": "GREEN", "pdf_src": "ScienceParsePlus", "pdf_hash": "0739065e7a48daf99a49647887a259fc68574c15", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [] }
14349388	pes2o/s2orc	v3-fos-license	A Molecular Evolution Approach to Study the Roles of Tropomyosin in Fission Yeast Tropomyosin, a coiled-coil protein that binds along the length of the actin filament, is a universal regulator of the actin cytoskeleton. We have taken a bioinformatics/proteomic approach to studying structure-function relationships in this protein. The presence of a single, essential tropomyosin gene, cdc8, in fission yeast, Schizosaccharomyces pombe, enables a systems-based approach to define the residues that are important for cellular functions. Using molecular evolution methodologies we identified the most conserved residues and related them to the coiled coil structure. Mutants in which one or more of 21 of the most conserved surface residues was mutated to Ala were tested for the ability to rescue growth of a temperature-sensitive cdc8 mutant when overexpressed at the restrictive temperature. Based on altered morphology of the septum and actin cytoskeleton, we selected three sets of mutations for construction of mutant cdc8 strains using marker reconstitution mutagenesis and analysis of recombinant protein in vitro: D16A.K30A, V114S.E117A.H118A and R121A.D131A.E138A. The mutations have sequence-specific effects on cellular morphology including cell length, organization of cytoskeletal structures (actin patches, actin cables and contractile rings), and in vitro actin affinity, lending credence to the proteomic approach introduced here. We propose that bioinformatics is a valid analysis tool for defining structure-function relationships in conserved proteins in this model organism. Introduction The actin cytoskeleton has ancient origins. Eukaryotic actin and prokaryotic homologues share an overall folding pattern, a bound nucleotide, and some common protofilament interactions despite poor sequence homology [1,2]. Within eukaryotes actin shows remarkable sequence and structural conservation; human b-actin and S. pombe actin (Act1p) are 90% identical. Just as the actin sequence is conserved, so are the functions and states of the actin cytoskeleton throughout eukaryotes [3]: actin monomers, filaments and bundles that function in cellular motility, intracellular transport, cell shape and formation of the contractile ring during cytokinesis. Accordingly, numerous actin binding proteins and their functions are universal to eukaryotes: myosins, Arp2/3 complex, ADF-cofilin, formins and crosslinking proteins such as aactinin and fimbrin, to name a few. One class of actin binding proteins, the tropomyosins (Tm), has been reported only in animals and fungi. Tropomyosin is a twochained a-helical coiled coil that binds end-to-end along the helical actin filament, stabilizing it and regulating its interactions with other proteins [4,5]. Tropomyosin regulates diverse actin filament functions including myosin-dependent contraction and motility in muscle, yeast and animal non-muscle cells. By binding and stabilizing actin filaments and protecting them from severing, branching by Arp2/3 complex and from certain crosslinkers, Tm controls actin filament dynamics. It can positively regulate formins [6,7,8,9] and make the interaction of myosin with actin cooperative and in some cases processive [10,11,12,13]. We postulate that residues required for universal Tm functions are conserved during evolution. In order to test this hypothesis we have taken a bioinformatics/proteomic approach to studying structure-function relationships in this protein using molecular evolution methodologies to construct phylogenetic trees. Based on the trees, the substitution rates at individual codons are calculated to identify the most conserved sites. In a previous study we constructed phylogenetic trees for coding regions of animal Tm genes and showed that the most conserved residues in striated muscle Tm have roles in actin binding and actomyosin regulation, two Tm functions that can be measured in vitro [14,15,16]. The results have been used to test and modify available thin filament models [14,16,17,18,19]. Other cellular Tm functions, such as assembly of higher order structures, regulation of cellular protrusions, vesicular transport, cellular polarity determination and regulation of cellular contraction and cytokinesis are not readily analyzed in vitro. The combination of bioinformatics and proteomics to analyze function in a model system offers such an opportunity. The complexity of the mammalian Tm gene family that results from four genes, alternate promoters and alternatively expressed exons encoding 40 or more isoforms [20] makes in vivo or even in situ studies impractical. For this reason, we turned to the fission yeast model organism, Schizosaccharomyces pombe that has a single, essential Tm gene, cdc8 [21,22]. Most classes of actin binding and regulatory proteins are represented in fission yeast, but the number of isoforms is small compared to vertebrates (1 actin, 5 myosins, and 2 formins, for example). Nevertheless, they account for the formation of the three actin structures in vegetative cells: cables, contractile rings and patches [23]. Fission yeast Tm (Cdc8p) is predominantly localized in actin cables and contractile rings with smaller amounts in actin patches [21,24,25]. Although there is only one Tm isoform, post-translational N-terminal acetylation affects actin affinity in vitro and localization within the cell [25,26], adding a mechanism for diversity beyond the genomic simplicity. Here we present the first stage in the implementation of a proteomic approach to investigate structure-function relationships in Tm regulation of the actin cytoskeleton in fission yeast, using the organism as a bioassay. We identified the most evolutionarily conserved residues and made mutations in the 21 most conserved surface residues that are most likely to be involved in binding other proteins. We screened for cellular function by assaying the ability of each mutant protein, when overexpressed, to rescue growth of a cdc8 ts mutant at the restrictive temperature. Although all rescued growth, certain mutants had morphological defects that infer altered assembly of cytoskeletal structures as well as determination of cellular polarity. Based on the initial screen, we selected three mutants for further study with the creation of mutant cdc8 strains using homologous gene replacement and in vitro analyses of recombinant protein. The mutations affect the cellular morphology in specific ways including cell length, organization and positioning of actin cytoskeletal structures and change the in vitro actin affinity. The results establish the validity of the proteomic approach for the study of structure-function relationships and set a framework to establish the molecular basis of actin regulation by tropomyosin in a cellular system. Results The first step in our bioinformatics approach to the study of structure-function relationships in fission yeast Tm, Cdc8p, was to identify the most conserved residues using molecular evolution methodology. We then tested the importance of conserved surface residues for cellular function using overexpression and gene replacement strategies, as well as in vitro functional assays. The report here establishes the validity of this proteomics approach as a way to screen for functionally important residues and paves the way to a detailed dissection and understanding of cytoskeletal mechanisms in fission yeast. Evolutionary Analysis In our evolutionary analysis of animal Tm genes we searched for homologous sequences in protein databases in an attempt to determine how deep in the phylogenetic tree we could identify Tm [15]. A BLAST search using the N-terminal 14 residues of rat striated muscle a-Tm, a highly conserved region in animal Tms, identified a sequence in Phaeosphaeria nodorum that encodes a 161 amino acid protein, the length of known fungal Tms, with a coiled coil propensity similar to confirmed Tms. Searching further with residues 1-24 of Phaeosphaeria nodorum revealed other known fungal Tm sequences. After aligning the sequences that were annotated as Tms, we selected conserved regions of 40 amino acids that we used to search the NCBI database using BLAST for additional fungal Tm sequences. Using this search method, we collected and evaluated 29 expressed sequences from 27 fungal species that had annotated genomic data available (as of the end of 2008, Table S1in File S1). Requirements for inclusion were conservative and included location of the known coding sequence at the 59 end of the open reading frame, a length consistent with known Tm sequences, and the absence of internal proline. We were unable to identify homologous sequences in protists or plants. The redundancy of the coiled coil sequence heptapeptide repeat limited the depth of the search. The protein sequences were aligned for phylogenetic tree construction ( Figure S1 in File S1). All species have a gene encoding a 161-residue protein, a length sufficient to span four subunits in the actin filament. Two species, N. glabrata and S. cerevisiae have two genes that encode proteins of 161 and 199 residues, respectively. The longer form reflects an insertion of 38 residues after codon 35 and corresponds to the addition of an actin binding period, enabling the longer form to span five actin subunits in the filament. The assembled dataset was used to construct trees using maximum-likelihood (GARLI 1.0) [27] and Bayesian (MrBayes v3.2) [28] approaches for phylogenetic analysis ( Figure 1, Bayesian tree). The consensus trees from the two approaches resulted in similar topologies with good branch support, .60% posterior probability for all major branches (except one 52% node) and bootstrap values .60% (except one 51% node). The tree corresponds well with known fungal phylogeny (http://tolweb.org/Fungi/). Based on the trees, the rates of evolution at individual codons were determined using PAML 4.1 [29] where v = dN/dS, dN and dS are the nonsynonymous and synonymous substitution rates for codons, respectively. v,1 represents residues under negative selection, v = 1, neutral selection, and v.1, positive selection. We plotted the v values against the 161 codons of cdc8 ( Figure 2A; Table S2 in File S1. Tropomyosin is highly conserved overall since v,0.2 for all residues. The codons with average v#0.02 (an arbitrary cutoff, ,10% of the maximal v) were selected as highly conserved sites and are highlighted in the S. pombe cdc8 protein sequence ( Figure 2B). The distribution of the highly conserved residues was similar in the fungal and animal bioinformatics analyses [15]. Of the sites, 31.2% (50 residues of 160, exclusive of the initial methionine) were highly conserved in fungi, 31.8% in animals. Considered in terms of the heptapeptide ''a,b,c,d,e,f,g'' repeat characteristic of coiled coils, the largest fraction was in e and g residues, often paired oppositely charged residues involved in inter-chain ion pairs in the coiled coil (15.6% in fungi, 12.4% in animals). The a and d residues that form the hydrophobic core between the two chains of the coiled coil are 3.8% evolutionarily-conserved in fungi and 8.5% in animals. The difference reflects the lower thermal stability of Cdc8p relative to mammalian striated muscle Tms (T M = 33.1uC for AS-Cdc8p vs 49.4uC for AS-rat skeletal a-Tm; [30], commensurate with a lower growing temperature. The core may be more critical for conserving structure and function in animal than in fungal Tms: the animal Tms have conserved interface Ala clusters and a conserved charged residue at a d position (D137) that is not present in fungal Tms. Fungal Tms have a four-residue interruption (a stammer) in the heptad repeat [31] that would alter interface packing and the supercoil [32]. On the other hand, the larger fraction of conserved e and g sites in fungi may partially compensate by contributing to the stability of the coiled coil. Both animal and fungal Tms have similar fractions of conserved b, c, and f surface sites (11.9% in fungi, 11.0% in animals), residues that are most available for interacting with other proteins and are therefore the focus of our proteomic screens [14,15,16], and the present work. Despite the above parallels with animal Tms, the fungal and animal Tms are too evolutionarily distant for meaningful alignment. Proteomic Screen: Rescue of a cdc8 ts Mutant at Restrictive Temperature Our initial screen for the functional importance of conserved cdc8 sites was to test the ability of mutant protein to rescue a cdc8 ts mutant at the restrictive temperature. Based on the evolutionary analysis (v# 0.02, or close to the value), evaluation of the amino acid sequences for conservation of amino acid class, and our previous results with vertebrate Tms [14,15,16], we selected 21 surface sites for mutagenesis, underlined in Figure 2B. All sites are b, c or f positions in the coiled-coil heptad repeat because they are the most likely to be involved in binding other proteins and the least likely to influence folding and stability. We introduced one or more Ala mutations (except Val114Ser) into cdc8 cloned in the pREP41x vector under control of the thiamine-repressible nmt promotor [33,34] for overexpression in S. pombe to generate a set of 16 mutants ( Figure 3A). In another study E6A and D16A rescued growth of cdc8-110, another temperature sensitive cdc8 mutant, at the restrictive temperature, but growth with D16A was somewhat slower than wildtype [30]. The transfection of wildtype and several Cdc8p mutants into wildtype S. pombe (SP6 strain) had no obvious effects on growth or morphology, either on agar or in suspension in the absence of thiamine. Despite the ability to rescue growth, overexpression of mutant protein in the ts strain (cdc8-27) resulted in a variety of phenotypes that include abnormal contractile ring and septum formation, actin cable organization and nuclear and septum localization, summarized in Table S3 in File S1. Figure 3C shows the results for the ts mutant (vector only) and for rescue by overexpression of wildtype and three Cdc8p mutants. In all, the percentage of cells with $2 nuclei increases during the first 3-4 hours after transfer to the restrictive temperature, the time required for the first cell division. After that, the percentage of cells with $2 nuclei continues to increase in cells with vector only because cdc8-27 cells become multinucleated being unable to undergo cytokinesis, while all others return to normal or near normal values ( Figure 3C; Table S3 in File S1). Cells expressing the following Cdc8p mutants have a higher percentage of cells in division: R86A.E93A.R103-A.E104A, R121A, and R121A.D131A.E138A, but most dividing cells have two nuclei after 17-24 hours. The percentage of cells with .2 nuclei was variable, complicated by the presence of some cells that we presume lost the vector and reverted to the multinucleated ts phenotype. To screen for correct formation and localization of the septum, we stained the cells with Calcofluor, together with DAPI to visualize the nuclei. Figure 3D shows the fraction of normally dividing, septated cells with normal septa at 25uC (t = 0), and at times after transfer to the restrictive temperature for overexpression of three of the mutants. At t = 0 nearly all septa were normal. After 4 hours, ,30% of the cells with 'vector only' had normal septa. In cells over-expressing wildtype or mutant Cdc8p, after 4 hours there were ,50% normal septa reflecting the residual ts phenotype (multiple, poorly organized septa and abnormal morphologies). After 17-24 hours, almost no cdc8-27 cells were normal; most had .2 nuclei, abnormal and multiple septa, and were branched. By 17 hours most cells expressing wildtype and E104A had normal septa while D16A and R121A.D131A.E138A had an elevated fraction of cells with an abnormal morphology and number of septa ( Figure 3D). Certain other mutants divided normally based on nuclear number, yet had a higher percentage of abnormalities including shape, position or number of septa (E6A, D16A, E107A.R110A, for example; Table S3 in File S1). Two mutants (E82A, V114S.E117A.H118A) were so variable from one transformation to the next that we did not report values in Table S3 in File S1. In general, however, overexpression of mutant Cdc8p had poor penetrance; most cells were morphologically similar to wildtype. Since Cdc8p is required for actin cable and contractile ring formation, we stained the cells with phalloidin to visualize the actin cytoskeleton after transfer to the restrictive temperature ( Figure 3B). Cells overexpressing wildtype Cdc8p showed long straight actin cables, polar distribution of actin patches and a discrete contractile ring typical of the S. pombe actin cytoskeleton. In cdc8-27 (with empty vector), after four hours at 35uC, when many cells still have one or two nuclei, there were no contractile rings or actin cables, and the polarization of actin patches was poor, as others have reported with this and other cdc8 ts mutants [36,37,38]. Most cells overexpressing Cdc8p mutants had normal actin cytoskeletons, but there was variability, including poorly organized actin cables, failure to assemble a normal contractile ring, abnormal positioning of nuclei and the contractile ring or weak polarization of actin patches, depending on the Cdc8p mutant. One mutant, R121A.D131A.E138A, had a consistent phenotype: poorly formed cables, failure to form a coherent contractile ring and irregular polarization of actin patches. Nevertheless, these cells could complete cell division because cells with .2 nuclei were rare after 17-24 hours at 35uC. While we were disappointed that overexpression of mutant Cdc8p did not result in more consistent and severe phenotypes, we did find morphologies associated with specific mutations, implying that the sites mutated are involved in specific Cdc8p functions. The penetrance of abnormal phenotypes resulting from overexpression of Cdc8p mutants was generally poor but the results aided in the selection of sites for further study using gene replacement and in vitro analysis of recombinant protein. As Tm is a twochained coiled coil, we expect the mild phenotypes result from heterodimer formation of molecules with one chain encoded by cdc8-27 and one by the pREP-cdc8 vector. Immunoblots showed that Cdc8p was present in cdc8-27 cells after 24 hours at the restrictive temperature and therefore available for heterodimer formation. The only mutant with a consistent phenotype (R121A.D131A.E138A) has sites close to the cdc8-27 mutation (E129K, confirmed in our laboratory) [22], which may explain the absence of a functional heterodimer. Reasonably functional Cdc8p could be formed with heterodimers of Cdc8-E129K with the other mutant proteins. We suggest that E16A is less effective than wildtype in rescuing growth of cdc8-110 at the restrictive temperature, as reported by [30], because cdc8-110 has mutations in the vicinity of Glu16 (A18T, E31K;Hitchcock-DeGregori, unpublished) that may impair formation of a functional heterodimer. Proteomic Screen: Gene Replacement to Test Phenotypes of Site-specific cdc8 Mutants Strategy. The construction of strains carrying mutations in the cdc8 gene allows us to test the hypothesis that mutation of conserved sites will result in specific phenotypes. Based on our results with rescue of the ts phenotype ( Figure 3, Table S3 in File S1), we selected three sets of mutations to introduce into the cdc8 gene: D16A.K30A, V114S.E117A.H118A, and R121A.D131A.E138A, hereafter referred to as 16.30, 114.117.118 and 121.131.138. Making mutations in the cdc8 gene overcomes the problem of heterodimer formation with an endogenous Tm as well as the generic caveats with overexpression. Furthermore creation of mutant cdc8 strains will enable future genetic experiments. The 16.30 strain isolated and characterized in the analyses described below was diploid. The FACS analysis of 16.30 in Figure 4 was of a colony from the original isolate, which was a spontaneous haploid that arose after completion of the rest of the study. Because cdc8 is an essential gene, we took advantage of the facility of homologous recombination in S. pombe and used marker reconstitution mutagenesis, a reverse genetic approach developed to create mutant strains [39]. Whereas the approach has been used to introduce conditionally-defective (ts) alleles into specific genes [39,40], we are using directed, specific mutagenesis and therefore have no selection method other than the nutritional marker (his5 + ). We first constructed a strain that integrated his5Dc:ura4 + next to the 39 end of the coding region of cdc8 in a his5-D1 ura4-D18 strain ( Figure S2 in File S1). Using homologous recombination, we then incorporated cdc8 (wildtype or with the desired mutations) and the C-terminal region of his5 (his5 c ) to place his5 + proximal to cdc8. Recombinants were selected for the ability to grow in the absence of histidine and uracil. Mutants were backcrossed twice to parental strains with cdc8 + or cdc8-27 to ensure tight and stable linkage to his5 + and cdc8. The presence of mutations was confirmed by genomic sequencing of the entire cdc8 gene. FACS analysis indicated that the114.117.118 and 121.131.138 strains are diploid and homozygous (Figure 4). We have been unable to isolate these two mutants as haploid strains. The 16.30 strain isolated and characterized in the analyses described below was diploid. The FACS analysis of 16.30 in Figure 4 was of a colony from the original isolate, which was a spontaneous haploid that arose after completion of the rest of the study. Cell morphology. Our screen of basic parameters of cell growth, size and organization of the actin cytoskeleton shows mutant-specific phenotypes and validates the approach introduced in this work. The wildtype and mutant cells in mid-log growth were analyzed for nuclear number, septum structure and location, overall morphology, and cell length ( Figure 5). Wildtype cells show the expected cell shape and size and fraction of binuclear cells ( Figure 5A, 5F, 5G) [41]. The septa are well formed and located at the midline. All three mutants exhibit altered shapes and septum morphologies that distinguish them from wildtype and from each other. Both 16.30 and 114.117.118 show a small increase in the fraction of binuclear cells ( Figure 5F, not statistically significant), supported by the presence of a minor second peak in the FACS analysis ( Figure 4B, 4C). Most septa are well formed and located at the midline, but the septa in both strains are often wavy ( Figure 5B, 5C). Together, the results indicate a delay in cell division relative to nuclear replication and separation. The 121.131.138 mutant has a significantly elevated fraction of binuclear cells and abnormal septa, consistent with delayed cell division ( Figure 5D, 5F). The presence of a large secondary peak in the FACS (about 50% of the primary peak) supports this conclusion ( Figure 4D). In this strain, many septa fail to form discrete structures or form double septa, leading to cells with blebs or branches and nuclearfree cellular fragments after cell division ( Figure 5D). Some septa are not perpendicular to the cellular axis. In addition, Calcofluorstainable material can be found between daughter cells that remain attached and at the new end at the completion of cell division. Similar morphologies were observed when 121.131.138 was overexpressed in the ts background ( Figure 3B). These features made scoring of the cells difficult (,13% of the cells could not be scored for nuclear number), but the cells can eventually divide. Growth rate. To learn if the differences in cell length and fraction of cells with two nuclei are related to growth we measured the growth rates of the three mutant strains compared to wildtype ( Figure 5E). The cell doubling times were calculated from the slope of the logarithmic phase of growth. Wildtype and 16.30 have doubling times of 2.5 hours. The doubling time for 114.117.118 is 2.9 hours and 121.131.138 is 3.3 hours. Assuming similar doubling times for haploid and diploid cells, the slower growth rate of 121.131.138 is consistent with impaired and delayed cell division as reflected by an increased fraction of binuclear cells. The abnormal position and appearance of septa ( Figure 5D, 5F) and the absence of well-formed contractile rings would both contribute to a slower rate of cell division ( Figure 6D). By comparison, the longer doubling time of 114.117.118 together with shorter length ( Figure 5G) suggests the mutations may affect both size control and timing. Actin cytoskeleton. The actin cytoskeleton is abnormal in all three mutants ( Figure 6). In wildtype cells F-actin, visualized using Alexa-phalloidin, is found in actin cables, actin patches and the contractile ring. The actin patches have the expected unipolar or bipolar distribution according to the stage of the cell cycle, discrete contractile rings form at the midline and there are long, straight actin cables, often running most of the cell length. The cdc8 mutations affect cable and contractile ring organization in site-specific ways. All three mutant strains show variable polar distribution of actin-containing patches, and the actin cables are wavy and reticular compared to those in wildtype cells ( Figure 6B-6D). The effects of the mutations on the actin cytoskeleton are least severe in 16.30 ( Figure 6C). Cables are variable in diameter and length, ranging from long and straight (similar to wildtype) to wavy and reticular. The cables in 114.117.118 are thinner and more reticular than in 16.30. In 121.131.138 the cables are not as prominent or consistent as in wildtype or the other two mutant strains. However, the presence of cables and at least partial polar distribution of actin patches indicates that Cdc8p does retain function in all three mutants, in contrast to the cdc8 ts mutants in which actin cables and contractile rings are absent at the restrictive temperature, and polarization of the actin patches is lost ( Figure 3B) [36,37,38]. The contractile rings are abnormal in all three mutants ( Figure 6). The least severe is 16.30, and the most severe is 121.131.138, inferring defects at different stages of contractile ring assembly and function. Nevertheless, the rings are able to constrict. In 121.131.138 the contractile rings predominate as assemblies of what appear to be nodes, similar to the very first stages when myosin assembles into the ring [42]. Often there are two contractile rings, followed by two septa that form on both sides of the midline. Even in cells in which the ring is more discrete, it appears to be an aggregation of nodes, with additional nodes remaining in the cortex near the midline that are not incorporated into the contractile ring. The poorly formed contractile rings in 16.30 and 114.117.118 ( Figure 6B, C) have the appearance of rings at early anaphase in wildtype cells when the filaments are not packed and fused, just as the nuclei are beginning to separate, early in spindle pole elongation [36]. However, in both strains, the rings appear to be in the primary stages of formation, even where the nuclei are well separated, as in late anaphase B (spindle pole elongation). Eventually, the rings become more compact without residual cortical nodes, and closure takes place, although in both strains they are less well formed than in wildtype cells. The defect in 114.117.118 is more severe than in 16.30 though double rings and septa, typical for 121.131.138, were never observed. Mutations of Conserved cdc8 Residues Reduce Actin Affinity To begin to understand the mechanism for the effect of cdc8 mutations in living cells, we expressed recombinant Cdc8p in E. coli and measured actin affinity and thermal stability. Cdc8 was modified to introduce AlaSer (AS) at the N terminus, a strategy to increase actin affinity of recombinant Tms that has proven to be effective for striated muscle as well as yeast Tms [43,44]. Without the N-terminal extension, unacetylated Cdc8p bound poorly to skeletal muscle Factin, about 500-fold weaker ( Figure 7A, inset), in agreement with results from the Maytum laboratory (personal communication) but in contrast to a previous report [25]. The effect of the N-terminal AlaSer on the affinity of Cdc8p for actin is similar to that reported for Nacetylation of skeletal muscle Tm [45]. The actin affinity of recombinant AS-wildtype Cdc8p and its variants was measured by cosedimentation with F-actin. AS-16.30 bound cooperatively to F-actin, with ,2.4 fold reduced affinity compared to AS-wildtype ( Figure 7A; K app of ASwildtype = 6.3610 6 and AS-121.131.138 was too weak to measure at the highest AS-Cdc8p concentrations our assay allows. Even though the intracellular actin concentration is about 106 higher than in our binding assays [46], and operationally even higher because of crowding [47], the reduced actin affinity of AS-114.117.118 and AS-121.131.138 is likely to be a major reason for their cellular dysfunction. The ability of the cdc8 strains with these mutations to form actin cables indicates residual function that may be supported by other actin binding proteins in the cell. On the Thermal Stability of Tropomyosins The thermal stability of wildtype and mutant Cdc8p was determined using circular dichroism to measure the ellipticity at 222 nm between 0-60uC ( Figure 7B). There was no major change in stability of the mutants; the small differences from wildtype did not correlate with the effects of the mutations on actin affinity. The main unfolding transition of AS-121.131.138 was slightly less cooperative than wildtype AS-Cdc8p. Discussion In this paper, we present and validate an evolution-based proteomic approach to study structure-function relationships in the conserved, essential cytoskeletal protein, tropomyosin (Cdc8p) in fission yeast. We have identified evolutionarily conserved residues and illustrated their specific importance both in living cells and in vitro. The approach will be extended to other conserved surface residues of Cdc8p. The amenability of fission yeast for systems-based and genetic approaches to elucidate function will allow us to combine cellular studies with in vitro assays to dissect the specific mechanism of dysfunction in different mutants. Tropomyosin is involved in numerous cellular processes in fission yeast [23] and more roles for this cytoskeletal regulator may be discovered. For example, Cdc8p is a positive regulator of fission yeast formin (cdc12) [6] and of mammalian formins [7]. As with animal myosins, fission yeast myosins can be positively or negatively regulated by Tm. Myo51p, Myo52p, Myo2p, Myp2p are positively regulated by Cdc8p, whereas Myo1p is negatively regulated [48,49]. Myo2p and Myp2p, together with Cdc8p are required for in vitro contraction of the contractile ring [50].The approach taken in the present work provides the framework to define the mechanisms by which Tm carries out these regulatory functions and to understand the consequences in living cells. The advantage of our directed approach based on evolutionary analysis, versus isolation of temperature-sensitive mutants that selects for proteins that are unstable or do not function at high growth temperatures, is that we will identify residues primarily involved in binding rather than folding. The strategy has been successful in identifying animal Tm residues required for binding actin and cooperative regulation of myosin [14,15,16], but cellular studies are difficult in a mammalian system. The three cdc8 mutants we report here have specific effects on the actin cytoskeleton with reference to actin patch localization, cable and contractile ring morphology, inferring the involvement of the mutated sites in particular cellular functions. All three mutants exhibit delays in contractile ring assembly relative to nuclear division that may relate to transport of materials to the midline andcontractile ring, and/or assembly of the ring. The most severe phenotype is seen in 121.131.138 in which the rings appear to be aggregates of nodes rather than filamentous structures. Cytokinesis nodes that are viewed as coalescing into contractile rings contain Mid1p that is important for establishing the midline of the cell in the first stage of contractile ring assembly [51,52]. The nodes contain numerous other proteins including Cdc12p (formin) and Myo2p and its light chains that are positively regulated by Cdc8p [23,52,53,54,55,56,57,58,59]. The delayed contractile ring assembly, patchy appearance, and the residual cortical nodes around the better-formed rings suggest that 121.131.138 is impaired in early stages of ring assembly, and may be a factor in the longer cell doubling time of this mutant. This finding resonates with the SCPR model of contractile ring assembly [54,60] in which Cdc8p has a role in stabilizing newly formed actin filaments. The 121.131.138 strain shows double contractile rings and septa, and oblique, wavy contractile rings resembling mid1D and other strains where genes that control positioning of the division plane at the cellular midline are disrupted [61]. The effects of the 16.30 and 114.117.118 mutations on contractile ring appearance are less severe than 121.131.138. The contractile rings are often filamentous in appearance, similar to the aster-like actin structures that Arai and Mabuchi observed at metaphase [36]. These structures may represent actin cables that are nucleated by formin-Cdc12p in other regions of the cell [62] rather than by a SCPR mechanism. Assembly of the nonmedial cables requires Cdc12p [6] and compaction into a ring involves Myo2p [56] and Myo51p [63], which are regulated by Cdc8p [6,12,48,58,59]. An alternate explanation for the contractile ring abnormalities is that actin-Cdc8p filaments containing mutant Cdc8p are more susceptible than wildtype to severing by cofilin-Adf1p [64], a process that is inhibited by Tm [65,66]. The cell length of the 114.117.118 mutant is shorter than that expected for diploid wildtype cells indicating that the mutation influences size control. Construction of cdc8 mutant strains that express fluorescentlylabeled myosins, formin, and other cytoskeletal and cell divisionrelated proteins will show whether localization and function of these proteins are affected in the strains expressing mutant Cdc8p. Such analyses, coupled with in vitro functional studies would help determine if the cdc8 mutations affect contractile ring assembly, constriction or both, and suggest residues on Cdc8p that may interact with or regulate the relevant cytoskeletal proteins and lead to the development of mechanistic models. Bioinformatics Phylogenetic trees were constructed from the Tm gene sequences using two different approaches for phylogenetic analyses: maximum-likelihood (GARLI) [27] and Bayesian (MrBayes) [28]. The evolutionary model was JTT+G+F, determined using Prottest [67]. GARLI 1.0 was used for the maximumlikelihood analysis [27]. Branch support was obtained with 1500 bootstrap replicates with 1 search replicate per bootstrap. The 50% majority consensus tree was obtained in PAUP* 4.0 [68]. The Bayesian analysis was done using MrBayes v3.2 [28] with two independent runs of 4 simultaneous chains for 500,000 generations and a sample frequency of 10. The average standard deviation of split frequencies was,0.01 at the end of the runs. The consensus tree was constructed after discarding 25% of the samples. Branch support was obtained as posterior probabilities. PAML version 4.1 [29] was used to calculate the substitution rates (v) at individual codon sites using variable site codon models M3 and M7 with 4 rate categories (K = 4) under CODEML with the Bayesian (MrBayes) tree. CODEML was first run under model M0 (no variation in v among sites) with codon substitution model F61. The branch lengths obtained under model M0 were used for the analysis with variable site models M3 and M7 in order to reduce computational times. The v values were obtained as the posterior mean, where v = v0PP0+v1PP1+v2*PP2 and so on for each v category, where PP0, PP1 etc. are the posterior probabilities of categories v0, v1 etc, respectively. The average v value for each site was calculated from the posterior means of the v values obtained using the two variable site codon models from the Bayesian tree (Table S2 in File S1). In all cases v,1, meaning there is no neutral or positive selection for any codon. Fission Yeast Strains, Plasmid and Genetic Methods Yeast growth and general methods. Strains were grown in EMM (Edinburgh minimal medium, Sunrise Sciences, San Diego, CA) with the appropriate selective supplements (0.0225% adenine, lysine, histidine, uracil) or YEA medium (0.5% yeast extract, 3% dextrose, 0.015% adenine) following standard growth, genetic and cell biology protocols in [69,70] and other online resources. The strains used and created in this study are listed in Table S5 in File S1. The standard growth temperature was 30uC, 35uC was the restrictive temperature for temperature-sensitive strains, and 25uC was used for permissive growth of temperature-sensitive strains and for other selected procedures. Cells were counted using a hemocytometer. Cellular transformation was carried out using the lithium acetate method [71] and purified plasmid or PCRgenerated fragments. Genomic DNA was purified according to [72] with modifications. The breaking buffer was 2% Triton X-100, 1% SDS, 100 mM NaCl, 10 mM TrisHCl, 8.0, 1 mM EDTA; we included an additional chloroform extraction, and RNase A digestion (10 mg/ ml, 30 min at 37uC) in the final step. The cdc8 sequence was verified in all strains and plasmids used in this study using appropriate primers by sequencing (Genescript, Piscataway, NJ or Genewiz, South Plainfield, NJ). Overexpression in S pombe. Cdc8p was overexpressed in wildtype (SP6) and temperature-sensitive cells (cdc8-27) [35] using pREP41X, under control of the nmt promoter (medium strength) [33,34]. Mutations were introduced into pREP41X-cdc8 (cloned between XhoI and BamH1 sites, gift of D. Kovar, University of Chicago), by Mutagenex (Hillsborough, NJ). Following transformation, cells were selected on EMM leu 2 medium, and expression was repressed by the addition of 10 mM thiamine. The protocol for inducing expression is described in the legend to Figure 3. Marker reconstitution mutagenesis. To make gene replacements of the cdc8 + with selected mutations, we used the method developed by Tang et al. [39] ( Figure S2 in File S1). First a fusion PCR fragment was amplified from genomic DNA prepared from a wildtype strain (SP6) using four primers (MRM-P1-MRM-P4, Table S4 in File S1), digested with SalI and BglII and cloned into the SalI and BglII sites of p208H5cdU4 + to create p208H5c-cdc8fusion. This plasmid was then digested with PvuII and transformed into YS007 (h-his5-D1 ura4-D18 ) [73] to place the N-terminal region of his5 + and ura4 + next to the 39 end of cdc8. Transformants were selected on minimal medium lacking uracil, shown to be 5-Fluorootic acid sensitive, and backcrossed twice to confirm the linkage of ura4 + and cdc8 + . Insertion of his5c next to the 39 UTR of cdc8 was confirmed by genomic sequencing. These strains are SH6, SH7 and SH12, SH13, the latter having been backcrossed to cdc8-27, a temperature-sensitive strain. FACS analysis. The ploidy of confirmed strains was analyzed using FACS (M. O'Connell, Mt. Sinai School of Medicine). Cells were grown to mid-log phase and fixed in 70% ethanol. The cells were rehydrated in 50 mM sodium citrate, 100 mg/ml RNAase A and incubated overnight at 36uC. Cells were sonicated for 30 seconds (Bioruptor, Diagenode, Inc., Denville, NJ). Propidium iodide was added to 10 mg/ml. The sample was analyzed on a Becton Dickinson FACScan with Cell Quest Pro software. Microscopy Staining. For microscopy cells were fixed at mid-log phase. For staining with DAPI and Calcofluor cells were fixed and stored in 70% ethanol [70,74]. To visualize the septum, ethanol-fixed cells were washed in PBS and resuspended in 5 ml 50-100 mg/ml Calcofluor (Sigma Life Science, St. Louis, MO) in 50 mM sodium citrate, 100 mM sodium phosphate, pH 6.0) and incubated at ambient temperature for 5 minutes in the dark. For counterstaining with DAPI to visualize the nuclei, the cells were washed and resuspended in 2-5 ml PBS with the addition of 0.5 ml 50 mg/ml DAPI (Sigma Life Science, St. Louis, MO). The samples were incubated for 5 min in the dark at ambient temperature. One ml of stained cells were mixed on a slide with 0.5 ml 1 mg/ml phenylenediamine in 50% glycerol, covered with a poly-l-lysine coated coverslip and sealed with clear nail polish. Filamentous actin was visualized using Alexa Fluor 488 phalloidin (Life Technologies, Grand Island, NY) [38]. Mid-log phase cells were fixed in 3.7% fresh paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA) for 5 min. at the growth temperature, washed three times in PBS, and stored in PBS with 0.01% NaN 3 at 4uC for 1 week or less. To stain with phalloidin, 2 ml of fixed cells were permeabilized by vortexing in 100 ml 1% Triton X-100 in PBS for 1 min, and washed 36 with PBS. Alexaphalloidin was added to the permeabilized cells (4 ml of 0.2 U/ml Alexa-phallodin) and incubated with gentle agitation for 50 minutes at room temperature. Counterstaining with DAPI and preparation of the slides was as described above. Fluorescence microscopy. Epifluorescence images were captured on a Nikon Optiphot 2 microscope fitted for epifluorescence using an Ushio USH-1020H mercury lamp with a Model HB-10101AF power supply and a DS epifluorescence illuminator with Chroma FITC and DAPI filters for imaging Alexa Fluor 488 phalloidin and Calcofluor/DAPI, respectively. We used a Nikon E-Plan 1006/1.25 Ph4DL oil immersion objective lens and a CoolSNAP fx camera (Photometrix, Tuscon, AZ). Exposure was controlled by a Uniblitz Model VMM-D1 shutter driver (Vincent Associates, Rochester, NY). IP Lab 4.0.8 (Scanalytics, Inc.) was used to control the microscope and its external devices. Depending on the stain intensity in the examined field, the exposure times were between 100 and 400 ms. Images of the same field were taken at different focal planes to visualize the full thickness of the cells. Binning was set to 161. The images were adjusted for brightness and contrast and merged using Image J 1.43 u [75]. Scale bars for all images were obtained by using an AO micrometer with 2 mm divisions subdivided into units of 10 m. Image analysis. We quantified the nuclear number and cell length in Calcofluor/DAPI stained micrographs using Image J 1.43 u [75]. The length of the cells was measured in pixels and converted to m using a conversion factor of 1 pixel = 0.0535 m obtained using the AO micrometer above. The distributions were plotted by rounding the length of cells to the nearest ten pixels and sorting them into groups. Expression, Purification and Analysis of Recombinant Cdc8p Wildtype Cdc8p was expressed in E. coli BL21(DE3) using pJC20-cdc8 (gift of D. Mulvihill, University of Kent) to make unacetylated Fission Yeast Tropomyosin	2017-04-20T13:43:20.714Z	2013-10-22T00:00:00.000	{ "year": 2013, "sha1": "d280491bd3d949786fa90929c71d43e7a5a7399f", "oa_license": "CCBY", "oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0076726&type=printable", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "d280491bd3d949786fa90929c71d43e7a5a7399f", "s2fieldsofstudy": [ "Biology" ], "extfieldsofstudy": [ "Biology", "Medicine" ] }
229931490	pes2o/s2orc	v3-fos-license	The importance of cell culture parameter standardization: an assessment of the robustness of the 2102Ep reference cell line ABSTRACT Work undertaken using the embryonic carcinoma 2102Ep line, highlighted the requirement for robust, well-characterized and standardized protocols. A systematic approach utilizing ‘quick hit’ experiments demonstrated variability introduced into culture systems resulting from slight changes to culture conditions (route A). This formed the basis for longitudinal experiments investigating long-term effects of culture parameters including seeding density and feeding regime (route B).Results demonstrated that specific growth rates (SGR) of passage 59 (P59) cells seeded at 20,000 cells/cm2 and subjected to medium exchange after 48h prior to reseeding at 72h (route B2) on average was marginally higher than, P55 cells cultured under equivalent conditions (route A1); whereby SGR values were (0.021±0.004) and (0.019±0.004). Viability was higher in route B2 over 10 passages with average viability reported as (86.3%±8.1) compared to route A1 (83.3±8.8). The metabolite data demonstrated both culture route B1 (P57 cells seeded at 66,667 cells/cm2) and B2 had consistent-specific metabolite rates (SMR) for glucose, but SMR values of route B1 was consistently lower than route B2 (0.00001 mmol, cell-1.d-1 and 0.000025).Results revealed interactions between phenotype, SMR and feeding regime that may not be accurately reflected by growth rate or observed morphology. This implies that current schemes of protocol control do not adequately account for variability, since key cell characteristics, including phenotype and SMR, change regardless of standardized seeding densities. This highlights the need to control culture parameters through defined protocols, for processes that involve culture for therapeutic use, biologics production, and reference lines. Introduction The progression of cell therapies from basic research to clinical products has prompted excitement in the regenerative medicine field; however, challenges still remain to be addressed prior to clinical realization [1][2][3]. These include the comprehensive characteriza tion of cell therapy products (CTPs), since CTPs are innately complex due to their biological nature. As a result, their characterization is equally complex; unli ke traditional pharmaceutical products which have sta ndards that can be easily be produced and used to compare product batches, CTPs currently do not [4][5][6]. Therefore, reference cell lines are utilized as the closest proxy in CTP manufacturing. However, these reference lines are equally complex and dynamic as the CTPs they are being used to assess. This is exac erbated by the lack of consensus protocol standardization used to culture reference lines, which inherently produces an additional level of uncontrolled variability to the product manufacturing quality control (QC) process [7,8]. Traditional pharmaceuticals assure high purity levels, however, purity presents a challenge for CTPs [5,9], due to the inherent nature of cells which interact with intrinsic and/or extrinsic factors. Unpredictable process variables pose problems where high purity yield is critical to functionality and efficacy. For instan ce, only ventral midbrain dopaminergic neurons (vm DA) would be required to ensure efficacy and innervat ion of the correct cell type into the appropriate regions of the dorsal striatum, for the treatment of Parkinson's disease [10]. Understandably regulators expect proof of high purity, to assure that only cells of interest are procured for patients [11,12]. Achieving high purity with such complex products remains challenging, i.e. ensuring all cells are differentiated or manipulated to the desired state and crucially, retain that state from bench to patient. This is important from a safety perspective and is scrutinized by regulators since undesired cells may potentially cause unwanted/unforeseen effects. Impurities present another bottleneck which can hinder clinical use and/or commercialization of CTPs. Impurities usually fall into two broad categories: product-related (e.g. cell fragments); and process-related (antibiotics, cell culture reagents etc.) [9,13]. Regulators tend to, understandably, have a strict view on all impurities and suggest that they should be addressed in the risk analysis [14]. In addition, impurities including metabolites should be tested and/or have their removal demonstrated through validation. For the most part, the issue of impurities can be addressed by using monoclonal antibodies to remove the cells and leave behind the unwanted fragments using methods such as fluorescenceactivated cell sorting (FACS). Furthermore, testing for the absence of bacteria, fungi and mycoplasma is essential [5,9]. In some cases, if accredited, inhouse tests can be used to screen the product and this can be integrated into the process quality management of the development and manufacturing process [15]. Importantly, tests and interventions to remove impurities must consider that cells cannot undergo rigorous sterilization and purification steps such as irradiation; that occurs during the manufacturing process of traditional pharmaceuticals. Thus, acceptable limits have been established for (most) CTP process-related impurities; however, the onus is on the manufacturer to set specifications regarding productrelated impurities based on factors such: as preclinical and clinical safety studies, the capability of the process to remove the impurities and also related products/historical data [14,16,17]. Knowledge of the identity of a product is required, it is crucial that the correct cells are identified and only those that are desired are used for the product. If undesired cells remain in the product and are administered to the patient, they may alter the product's mechanism of action or result in spontaneous and potentially tumorigenic differentiation, i.e. with remnant pluripotent cells. Cluster of differentiation (CD) markers are one of the most prominent methods of ide ntifying cells, via their cell surface molecules [14,18]. However, like many aspects of CTPs there are nuances that present some identification challenges. For instan ce, similar CD marker combinations can be found on very different cell types; and/or the marker expression profiles can be transient over time. The European Medicines Agency (EMA) says that 'identity of the cellular components should be based on phenotypic and/or genotypic markers [19]'. As a result, it means that the test methods used need to be specific for the cells of interest. Hence why, when addressing phenotype, relevant analyses should be used such as gene expression, antigen presentation and specific biochemical activity in an orthogonal manner [5,20]. For allogeneic CTPs, it is imperative that the identity profile should include histocompatibility markers, as the cells will be heavily scrutinized by immune system. Tumorigenesis presents trepidation with CTPs, particularly those using heavily manipulated or genetically edited cells, since they may undergo transformation, whereby chromosomal instability is possible [8,19]; thus, impeding authorization. When using unestablished cell lines, karyology tests should be considered to investigate the tumorigenic potential of the cells [19,21,22]. Novel approaches to ensuring CTP safety are being devised, for instance the introduction of a suicide gene can reduce the risk of tumorgenicity for products differentiated from pluripotent cells [23,24]. To this end, the present work highlights the require ment for standardization of culture protocols, especially when the cells are used for QC of cell-based products. This work was executed using the Embryo nic Carcinoma (EC) 2102Ep cell line derived from primary human testicular teratocarcinoma [25]. An initial procedure provided by the National Institute for Biological Standards and Control (NIBSC) was followed; however, cell growth inconsistencies including population doubling times and growth rates were recorded. Therefore, it is was hypothesized that a ma jor contributing factor to such inconsistency was due to the ambiguity of an undefined protocol that is susceptible to operator interpretation/intervention. Study-based examples whereby understanding and application of ambiguous protocols that may introduce cell growth inconsistencies into a culture system include operator-specific interpretation of unspecified time points for medium exchange, feeding volume and passage points, that are not based on the number of viable cells/cm [2] in the culture vessels. Common place use of split ratio can create inconsistencies rega rding when/if the operator intervenes which may ultimately lead to lower yields of the final products and/or process failure. This issue is relevant to many published protocols [7,[25][26][27][28][29][30]. For instance, a range of nonspecific passage timings, i.e. 3-4 days, are suggested for cells to reach 'adequate' confluency; obse rved confluency in itself being highly subjective [31,32]. Additionally, the use of defined seeding densities is absent in many protocols, instead split ratios are prescribed, typically with a range of ratios, i.e. 1:3 or 1:5 [33][34][35][36]. Conversely, well-defined feeding and culture protocols in general prevent operator interpre tation and therefore would facilitate optimal cell gro wth and a desired process outcome. The results of a study carried out by Senger et al. revealed that optimization of fed-batch parameters, harvest time and feeding strategies of Chinese hamster ovary (CHO) cell cultures via well-defined protocols significantly favored cell growth by controlling the metabolite rate in stirred tank reactors [37]. More recent studies have demonstrated that controlling the glycan moieties of antibody therapeutics and improving antibody productivity are highly desirable in maintaining batch-to-batch culture consistency [38]. Kotidis et al. employed several strategies to preserve cellular health and productivity while enhancing antibody quality. Their Quality by Design (QbD) approach consisted of a modeling platform and well-defined protocols that allowed for the quantification of the impact of gly cosylation precursor feeding on cellular growth, and metabolism in addition to antibody productivity. The platform was then used to optimize feeding strategies that enhanced the final concentration of glycosylated antibody by >90% while sustaining the integrity of viable cell density or final antibody titer [38]. The specific objectives and aim of this work were to identify attributes of a reference cell line that would inform its application including variability in important biological markers under normal laboratory processing variation and key influe ncing culture parameters. Materials and methods The laboratory setting where these experiments were undertaken run under an industry-style quality system that segregates different cells cultures in to defend, segregated areas and equipment with robust cleaning and maintenance protocols. The facility operates under a quality system based on ISO 9001:2015 quality system principles [39]. In Vitro cell culture All reagents and consumables were obtained from Fisher Scientific, UK unless otherwise stated. EC 2102 Ep (Passage 54, GlobalStem, USA) cells were thawed from cryopreservation and cultured in T25cm 2 tissueculture flasks to create a working cell bank. Each vial contained 6 × 10 6 cells in 1 ml of Cryostor® CS10 (cat# C2874-100ML, Sigma Aldrich, UK) which were stored in liquid nitrogen. The cells were cultured at 37°C, 5% CO2 using Gibco Dulbecco's Modified Eagle Medium (DMEM) high glucose with GlutaMax™ medium (cat# 10569010) and supplemented with fetal bovine serum (FBS, 10% v/v, cat# 10100139) herein referred to as growth medium. A systematic experimental approach was applied to two parameters: cell seeding density and feeding regime. Initial experiments over three passages were implemented to investigate the effect of these parameters on cell characteristics, with a focus on growth rate, cell viability, phenotype, and specific metabolic rates SMR. These experiments fed into a longer-term 10 passage (30 days) experimental protocols which were utilized to elucidate the long-term effects of the seeding density and feeding regime on the characteristics mentioned above, with the addition of gene expression. The 10-passage duration was implemented in order to compare the overall experimental outcomes to the work of Andrews (1982) that alluded to seeding density being the major driving force for variation in growth and unwanted cell differentiation in EC 2102Ep cells. Experiment A: four-way comparison of the effects of medium of exchange 2102Ep cells (P59) were expanded in a T75cm 2 tissueculture flask, the growth medium was exchanged following at 48 h. Cells were passaged following 72 h and were seeded into two T75cm 2 flasks for another expan sion passage (P60). The cells were then separated into four different routes of culture ( Figure 1): Route A1 and A3, cells were seeded in duplicate at 20,000 and 66,667 cells/cm 2 , respectively in 5 ml growth medium. Following 48 h the growth medium was exchanged. The cells were then passaged following at 72 h; cell counts were obtained in duplicate for each flask and cultured for a further two passages. Route A2 and A4, cells were seeded in duplicate at 20,000 and 66,667 cells/cm 2 , respectively, in 5 ml growth medium. No medium exchange took place prior to passage. The cells were passaged following 72 h culture, cell counts were obtained in duplicate for each flask and cultured for a further two passages. Experiment B: longitudinal comparison of two protocol culture conditions 2102Ep cells (P55) were cultured in T75cm 2 tissueculture flasks; the growth medium was exchanged at48h. The cells were harvested and passaged following 72 h into two T75cm 2 flasks for another expansion passage (P56). The cells were pooled then separated into two routes of culture in T25cm [2] tissue-culture flasks, herein referred to as B1 and B2 ( Figure 1): Route B1. Cells were seeded in triplicate at 66,667 cells/cm 2 in 5 ml of growth medium. Medium was exchanged at 48 h. The cells were harvested and passaged following 72 h (P57). Cell counts were performed in duplicate for each flask and cultured for a further nine passages. Route B2: Cells were seeded in triplicate at 20,000 cells/ cm 2 in 5 ml growth medium. No medium exchange took place prior to passage. The cells were then passaged following 72 h culture (P57). Cell counts were obtained in duplicate for each flask and cultured for a further nine passages. Four analytical time points were selected over the 10 passages, two at an early stage of the experiment and two at a later stage (passages 1, 3, 7 and 9). Cell counting Cell counts and viability (via acridine orange uptake and DAPI exclusion) were obtained using an automated mammalian cell counter (NucleoCounter NC-3000, Chemometec, Denmark). The results were used to obtain the specific growth rate (μ) and the Popula tion Doublings (P d ) using equations previously published by Heathman et al. [40]. Metabolite analysis Spent media samples, 500 μl, were collected then stored at −20°C prior to analysis. For Experiment A, media samples were collected at two time-points: following medium exchange at 48 h (A1 only) and prior to cell passage. In experiment B, spent media samples were collected on the medium exchange and passage time points for B1 and B3. Samples for route B2 and B4 were collected only on the passage time point. Spent media samples were analyzed for glucose using the Cedex Bio-HT (Roche, Germany). The results were used to obtain the SMR mmol.cell −1 .d −1 (SMR) using an equation previously published by Heathman et al. [40]. Flow cytometry All reagents were obtained from BD Biosciences (Oxford, UK) unless otherwise stated. Flow cytometry immunophenotyping samples were collected and fixed on the day of harvest. Experiment A samples we re collected at all three passages. Experiment B samp les were collected at passage 1, 3, 7 and 9. A minimum of 1 × 10 6 cells were collected from culture route and fixed for 20 min in the dark at room temperature (RT, BD Cytofix™). The cells were washed with phosphate buffered saline (PBS) and centrifuged at 500xG (Aacc uSpin Micro 17, Fisher Scientific, UK) for 5 min; this step was repeated twice. The cells were permeabilised for 10 min in the dark at RT (a BD Perm/Wash™). Cell staining was performed using pre-conjugated antibodies Oct3/4-PerCp-Cy5.5, SSEA-1-PE and SSEA-4-Alexa 647, for 30 min in the dark at RT, a sample with the respective isotype controls was used to account for nonspecific binding. Following 30 min two washes were performed using the permeabilising buffer. The cells were resuspended in cell stain buffer prior to analysis. In total, 250 μl of each sample was used for analysis using flow cytometry (BD FACSCanto™ II, BD Biosciences, USA). PCR PCR samples were collected upon cell harvesting and stored at −80° C until analysis. For experiment B sam ples were collected at passage 1, 3, 7 and 9. Gene expre ssion quantification was performed using Quantitative RT-PCR. Total RNA was isolated using the RNeasy Mini Kit (Qiagen, Cat #74104) following a defined manufacturer's protocol. RNA yield and purity were determined using the NanoDrop™ 2000 Spectrophoto meter (Thermo Scientific, UK). RNA integrity was assessed using an Agilent 2100 Bioanalyser (Agilent Technologies, Germany). 1 µl of total RNA was reverse transcribed using the QuantiTect Reverse Transcription Kit (Qiagen, Cat #205311) according to the defined manufacturer instructions. 1 µl of the cDNA synthesis reaction was used as template for each real-time PCR using VeriQuest Fast SYBR Green qPCR Master Mix. PCR was run in a StepOnePlus™ Real-Time PCR System (Applied Biosystems, USA) for 40 cycles at 95°C for 3 seconds to denature and 60°C for 30 seconds to anneal. The relative amounts of PCR pro duct were quantified using the relative threshold cycle (ΔΔCt) method corrected for efficiency for each amplification. The gene quantities for each sample were normalized against the geometric mean of expression of the housekeeping genes GAPDH and β-actin [25]. 8. Cell time Cell time can be applied to quantify the capability of a given volume of medium to sustain the growth of a given number of cells for specific period. Cell time (CT) permits the analysis of cell growth medium capacity/medium exhaustion. CT is expressed in cell hours/days. Where N O is the initial cell density, k is the specific growth rate and t is the time of culture (hours/ days). It is the area underneath the curve of a [cell] vs. time graph (Equation 1). Statistics Unless otherwise noted, statistical significance was determined using two-way analysis of variance using Graphpad Prism 7 Version 7.0d (CA, USA). Statistical significance was assigned as indicated in the figure legends. '' indicates p < 0.05, '' indicates p < 0.01, '' indicates p < 0.001, and '' indicates p < 0.0001. Tukey and Sidak's multiple comparisons tests were used to compare means between groups. Experiment A; n= 3 for each route, n= 2 for cell counts, n= 2 for metabolite analysis. Experiment B; n= 2 for each route, n= 2 for cell counts, n= 2 for metabolite analysis. Results The aim of this work was to apply parameter chan ges to an in-house protocol adapted from NIBSC by utilizing defined seeding densities and timedefined passage points. This was to investigate the impact of cell culture parameter changes and highlight the need for standardization; lack of standardization can result in process output variation, i.e. variability of growth rates and cell phenotype. It was hypothesized that a major contributing factor to such inconsistency was due to the ambiguity of an undefined protocol that is susceptible to operator interpretation/intervention. The use of defined seeding densities and timedefined passage points were applied to minimize human-based sources of variation including observed confluency and uncontrolled/undefined parameters including split ratios to have a more defined and standardized protocol. The in-house protocol was then applied to a series of 'quick hit' experiments performed over three passages to realize the impact of different protocol parameters on characteristic variation including cell metabolic rate, cell-specific growth rates and phenotype. A longitudinal experiment set over 10 passages employed a streamlined version of the in-house protocol, which removed the medium ex change step and used a defined seeding density of 20,000 cells/cm 2 , that differed from the original NIBSC protocol (66,667 cells/cm 2 ), to experimentally compare the effect of the changes in protocol. 1. Experiment A: four-way comparison of the effects of medium of exchange. SSEA-1 and SSEA-4 expression were both stable with minimal variation throughout the three passages, apa rt from route A2. The cells exhibited variation in Oct3/ 4 expression in all conditions as expression decreases, most notably at passage three, regardless of whether they were subjected to medium exchange at 48 h; the same trend was observed across all culture routes independent of the seeding density (Figure 2c). Cultu re route A2 exhibited a higher rate of metabolism ( Figure 2b); other significant differences between routes are noted in Table 1. SGR values were similar across all culture routes, independent of culture conditions and density, apart from route A4 at passage number three (Figure 2a). Experiment B: longitudinal comparison of two protocol culture conditions The growth rate data obtained demonstrated that over nine passages, culture route B2 on average had marginally higher SGR values (0.021 ± 0.004) compared to A1 (0.019 ± 0.004), the SGR values of both routes fluctuated throughout the 10 passages (Figure 3a). The fluctuation trend was similar for both routes with regards to SGR values, Pd values and cell viability. Ro ute B2 had a marginally higher average cell viability (86.3% ± 8.1) compared to route A1 (83.3 ± 8.8) over the 10 passages (Figure 3b). The metabolite data demonstrated that route B1 had a consistent SMR for glucose metabolism over the 10 passages, appearing to be independent of the SGR value (Figure 3a & C). Flow cytometry marker expression analysis showed no significant change in SSEA-4 expression over the 10 passages or between route B1 and B2 (p = 0.07 and p = 0.10, respectively) (figure 3f). However, for route B2 there was expression of SSEA-1 which is a negative marker, this was observed from passage cycle three onwards. Oct3/4 expression in route B2 decreased by 52% and ~30% in B1 at passage cycle seven, which increased back to >95% in route B1 and only to 79.6% by passage cycle nine in route B2 (Figure 3c). The PCR performed for experiment A demonstrated that for a selection of genes analyzed, DPPA4, POU5F1 and REX1, there is no significant difference between the two different routes throughout the 10 passages ( Figure 4). However, DNMT3B, NANOG and SOX2 exhibit significant differences between the two routes and between passage cycles as they have higher expression in route B2 (Figure 4a,c,f). TDGF is the only gene that had a significant difference at passage cycle three, being much lower than the route B1 condition ( p = 0.018). DNMT3B, NANOG and SOX2 showed differences at passage cycle seven and/or passage nine between the two routes, in all cases route B2 having a higher fold change. (b) Glucose SMR over three passages demonstrating that route conditions B2 and B4 that have no medium exchange following 48 h have higher SMR compared to route A1 and A3, significant differences between routes at each passage shown on the graph highlight that seeding density, feeding regime and a combination of both parameters result in a different SMRs (n = 2). (c) Oct3/4 marker expression over three passages, expression levels are similar between routes, significant difference observed between passages 2 and 3 for all routes (p < 0.05) except route A4. (d) SSEA-1 marker expression over three passages, inset shows that percentage expression levels are below 1.5%. (e) SSEA-4 marker expression over three passages, significant differences are observed only within route A2 at the different passages (p < 0.05). N.B. Oct 3/4 and SSEA-4 are positive markers for pluripotent human embryonic stems, SSEA-1 is a negative marker associated with pluripotency of human embryonic stem cells. Table 1. Summary of the significance values at each passage, comparing the SMRs of the different culture routes using Tukey's multiple comparisons test. Both seeding density and feeding regime are shown to result in significant difference between the four routes. The highest significant differences between routes, within passages, are observed when both parameters are changed, e.g. A2 vs A3 (n = 2 for each condition). Discussion In the context of the growing cell therapy area, it is important to develop robust protocols that are easily transferable and comparable, to ensure high-quality cell products CTPs that satisfy regulatory requirements without being burdensome regarding the man ufacturing process [7,8,17]. As such, before truly fully automated and closed systems can be employ ed, it is important to have standardization in manual and semi-automated protocols that are currently in use. This work set out to demonstrate that defined, specific protocols with specific operating ranges for the variables studied are that are not open to interpretation are necessary to maintain consistency and reduce variability, particularly in the context of reference standards and reference cell lines [27,28,41]. The simultaneous culture of the 2102Ep cells using various culture routes was utilized to explore the effect that changes in protocol parameters have had on the outcome of cell culture process character ization and outcomes. Effect of variable culture conditions on cell growth variability Previous work performed in-house has demonstra ted that despite using an embryonic carcinoma cell lines like such as 2102Ep, that are considered to be 'robust' [42], the growth is inconsistent; however, it is unclear what specifically causes variability in cell culture systems. Environmental and culture factors can have a major effect on growth and cell viability including (sub-optimal) temperature, level of dissolved oxygen, nutrient degradation/depletion, and waste product accumulation [43]. Other prevalent types of inconsistency arise during the measurement and characterization of cell growth. Contributing factors may include a lack of/inadequate calibration of cell counting instruments, poorly defined cell culture and feeding protocols, and a lack of a standa rdized cell counting instruments in different laboratories [42]. In addition, a lack of QC systems and inline sensing may result in fundamental and accumulative inconsistences [44] including but not limited to: poor quality of starting materials/stock culture, unsystematic freeze/thaw processing and inadequate mixing of cells, media, and/or other reagents that contribute to poor cell growth and/or cell count inconsistencies [45]. Other pivotal factors to consider are high-passaged cells, incorrect-formulated /poor quality media, buffer and/or serum and microbial contamination (particularly mycoplasma [45]). The present work hypothesized that the root cause of variability is due to the lack of standardized controlled culture protocols, which results in differences in culture growth and characterization outcomes. It is evident from previous experiments performed by the authors using the 2102Ep cells that, in terms of growth rate the cells are not sensitive to the wide range of densities investigated . Gene expression analysis of experiment B at three passage points over the 10 passages (normalized to the cells at passage 1), expression was measured by real-time quantitative reverse transcription-polymerase chain reaction (qRT-PCR). Fold change was calculated using the relative quantity of each gene was calculated by the ΔΔCt method, using a correction for the amplification efficiency of that gene, and normalized to the geometric mean of two housekeeping genes: glyceraldehyde-3-phosphate dehydrogenase and β-actin. Significant differences in fold change at different passage cycles were seen only for A-DNMT3B (**p = 0.0007), NANOG (p = 0.01) and SOX2 (p = 0.004). Significant differences between route A1 and A2 were seen only for A -DNMT3B (p = 0.0032) and F-SOX2 (p = 0.006). Differences between the two routes within a passage cycle were only observed for G-TDGF (passage cycle 3, p = 0.018), A-DNMT3B (passage cycle 7, p = 0.0221; passage cycle 9, , p = 0.0073), and F-SOX2 (passage cycle 9, p = 0.0191). Error bars indicate standard deviation (n= 3 for all data presented). N.B. Cells from passage 1 were used as the control sample for the qRT-PCR fold change analysis, therefore no data is presented for passage 1 in the graphs. (5,000 to 93,000 cells/cm 2 ) see Supplementary Data 1 and 2), illustrating that at least in the case of 2102Ep cells, density is not a limiting factor of cell growth within the relatively wide ranges tested. Instead, it is assumed that the differences observed are a result resultant of the cell system dynamics, as a consequence of protocol parameters such as undefined seeding densities due to split ratio passaging and inconsistencies in feeding regime. Effect of seeding density and feeding regime on phenotype and SMR Density has been reported to be a major parameter that can influence undesirable cell differentiation [26,46]. Andrews (1982) previously demonstrated that when 2102Ep cells are seeded at low density (1,300 cell/cm 2 ) phenotypical changes arise as evidenced by the expression of SSEA-1 which is a negati ve pluripotency marker [25,26]. The present work employed a systematic approach to ascertain if density was the predominant factor for differentiation observ ed by Andrews (1982). The initial experimental design explored feeding regimes as a potential protocol parameter that may attribute to variation, either in isolation or as a combined effect with seeding density. The longitudinal experimental approach investigated the long-term effects of changing both, seeding density and feeding regime parameters. Akin to the work by Andrews (1982), it was observed that changes in seeding density resulted in changes to phenotype and metabolism suggesting cell differentiation; however, it was also observed that the feeding regime employed also resulted in other cell characteristic differences. These differences were assessed using secondary meth ods of appraisal that are based on SGR and specific consumptions/production of measured metabolites. These analyses revealed that SMR is significantly affe cted by feeding regime regardless of seeding density, illustrating that controlling seeding in isolation is insufficient for ensuring consistent cell characteristics during cell culture. A limitation of the SMR calculation is that it assumes both a constant cell growth rate and constant consumption/production rate over the measurement period. Any deviation from this will introduce some inaccuracy but will still provide a be tter estimate than simpler approaches such as dividing change in concentration by average cell density which do not account for the exponential nature of cell growth; thus, it is routinely applied as a standard engineering approach. Although Andrews (1982) prescribes the route B1 density (66,667 cells/cm 2 ) as an optimal seeding density, it is evident that even at this density variation is observed through the passages and experiments carried out. It can be postulated that the variation is due to the innate biological of the cells and the complex cell system dynamics that are involved in culture [4,47,48]. This further illustrates the necessity of protocol standardization to avoid the compounding of different sources of variation. Stability of cell growth, phenotype, and gene expression in response to streamlined culture protocols Experiment B was a longitudinal study investigating the stability of the two routes used to assess the effect of feeding regime on cell growth performance. The in-house defined NIBSC protocol condition, route B1, was compared to a streamlined protocol using a lower starting seeding density. The route B2 condition was chosen as previous in-house experiments revealed that cells grown at 20,000 cells/cm 2 with 5 ml of growth medium performed well in terms of cell viability and SGR (see Supplementary Data 1 and 2). The cell time for this condition was then calculated, a value of 2.22 × 10 7 cell hours per 5 ml of growth medium over 72 h was obtained, which was a more than sufficient medium capacity to sustain the cells over a 72 h culture period without medium exchange. From a manufacturing perspective, this minimal inter vention protocol employed is ideal as it reduces operator manipulation and reagent use, which are simple means of lowering operating costs [4,9]. The objective of experiment B was to elucidate whe ther a cell time-defined protocol, route B2, would per mit cells to be cultured in a streamlined manner and retain the same characterization outcomes as the original protocol in terms of growth dynamics, phenotypic expression and metabolism. Josephson et al. (2007) previously cultured 2102Ep for several passages and reported no notable characteristic cell chan ges, which made them an ideal human embryonic ste m cell reference line candidate [25]. As such the exper iment cultured the cells for 10 passages whilst tracking the growth and metabolic rates along with gene expre ssion and phenotypic markers. Over the 10 passages of experiment B, a fluctuation trend in growth rate and viability was observed. Interestingly, this fluctuation was synched synchronized between the two different culture routes ( Figure 3a,b), suggesting that there is an artifact that synchronizes the two routes in terms of cell growth and viability that is independent of the culture route. It is unclear whether this is an innate feature of the cells (this synchronized behavior is conserved even over 10 passages of the same cell stocks), which alludes to the behavior being an innate cell feature. Alternatively, although not considered to be the sole cause, it could be deemed hypothesized that the synchronized beha vior is subject to the result of a measurement system error. This notion is supported by observation of the behavior across independent culture trains. However, different characterization systems (cell counting, imm unophenotyping, gene-and metabolic analysis) with distinct manipulation techniques in terms of sample handling were used. Therefore, it is unlikely that meas urement error would occur across all techniques, sug gesting the synchronized behavior observed is an effe ct attribute of the cells themselves; especially since the gene expression of DNMT3B and SOX2 also shows the synchronized supports the synched behavior that is seen in the SGR and cell viability (this was illustra ted by the no difference in fold change between the two culture routes, and yet there was a significant fold change difference between the passages in the case of DNTM3B, NANOG and SOX2 expression). While there is no difference in gene regulation between the two routes for DPPA4, POU5F1 and REX1 which all had consistently similar levels between routes and between passages (Figure 4b,d,e). The high levels of consistency with the routes throughout the passages showed that when a defined protocol is followed, less variation during culture is observed. SGR differences were observed for like-for-like conditions between the two experiments, i.e. the A3 condition had a higher SGR at passage three (0.018SGR/ h ± 0.004) than at passage 1 (0.015 SGR/h), whilst the opposite was true for the B1 condition. This difference can be attributed to the cells being seeded at different passages. However, this difference is not observed in the SMR, where there cycling trend is similar between the two experiments (Figure 2b,c), suggesting that the observed difference in SGR can be attributed to the innate cell biological variation. Further experimentation would be required to fully ascertain the cause of the observed SGR differences. The notably higher SMR for route B2 was unexpected as this disparity in SMR has not been previously seen in other published literature or previous work of the authors to the same magnitude. This beha vior was observed in experiments A and B, suggesting that changes in SMR are due to the cells being cultured following a streamlined protocol without medium exchange. This implies that SMRs and phenoty pic expression are affected by the culture conditions, i.e. nutrient availability based on feeding parameters. The change in expression of SSEA-1 and Oct3/4 furt her demonstrates that culture conditions have an influ ence on marker expression. This illustrates the impact of changing the culture environment by manipulating nutrient levels; regardless of previous culture conditions, the cells are the significantly impacted. The change in expression of SSEA-1 was immediately evident following the first passage, illustrating that feeding regime can impact the cell metabolic behavior and phenotypic expression (Figure 3c,d), but not necessarily the growth performance ( Figure 3a). This implies that the potential cause of SMR discrepancy is due to the lack of medium exchange following 48 h, which is substantiated by experiment A, as conditions not subjected to medium exchange had higher SMRs compared to the cells that were subjected to medium exchange. Evidenced through experiment A are the differences between the routes at the three passages, the most significant differences are indicated (Figure 2), other significant differences are reported in Table 1. This highlights that differences in culture protocol based on seeding density, feeding regime, or both, result in notable SMR variation with in the same stock of cells cultured simultaneously. The highest significant differences between routes, within passages, are observed when both parameters are changed, e.g. A2 vs. A3, where both the seeding density and feed regime are changed. Cell stability and differentiation in reference cell lines The EP2102 cells are putatively considered as a reference line, thus it would be assumed that there would be stability in the characterization of the cells. However, over the 10 passages some unexpected behaviors were observed; of particular interest is the presence of SSEA-1 which is a negative marker for pluripotent stem cells, and the decrease in expression of the positive pluripotency marker Oct3/4. The presence of SSEA-1 and the significant decrease in Oct-3/ 4 (up to 52% in route B2 passage cycle 7) implies that the cells are differentiating (Figure 3c) [25,26,46]. Interestingly, the significant differences in gene expre ssion are observed at passage seven and/or nine, which are the same passage cycles that demonstrate notable differences in phenotypic marker expression through flow cytometry, particularly Oct3/4. It is unclear what causes the change in behavior, in either cell SMR or marker expression [28,49,50]. As previously mentioned, the phenotypic change that has was observed in the experiments could be attributed to cell differentiation. This is a possible effe ct of the culture conditions, the literature regarding embryonic and pluripotent cells suggests that seeding at low densities causes differentiation, much like the conditions that the cells were cultured under [49][50][51]. The gene expression negative fold changes observed for DNTM3B and SOX2 suggested that the cells were differentiating as they are key pluripotency markers. Furthermore, the experimental data showed evidence of higher metabolite rate and phenotypic change, which is often concomitant with cell differentiation and/or the presence of a different cell population [52][53][54][55]. However, there is evidence within the experiments that is contrary to the cells differentiating, suggesting it is the protocol parameters, i.e. medium exchange that are causing variation as exemplified in experiment A (Figure 2b). For instance, as previously stated, no morphological change was observed under light microscopy throughout the 10 passages. Another proposed explanation of the observed phenotypic change is that the change is linked to cell death. This rational is reinforced by the fact that the phenotypic change is not cumulative over time, suggesting that the expression is presented just before the cells start to die resulting in non-cumulative expression of SSEA-1. However, cell viability and SGR data show evidence that is contrary to this, as cells in route B2 had on average, higher cell viability and growth rate which would not be expected with cell death (Figure 3a,b). Furthermore, at passage cycle seven there is a drastic decrease in Oct3/4, yet in the next passage there is no significant decrease in cell number or cell viability. Density had been reported to be a crucial factor in some cell culture protocols, in particular, those using embryonic cell lines, as it is mentioned to have an influence on metabolic behavior and both directed and spontaneous differentiation [51,56--56-60]. In experiment B the observed differences in behavior could be attributed to the seeding densities used, as the seeding density was one of the main parameters that was altered between the two routes. Nonetheless, density is unlikely to be the only major driving force as previous experiments have shown no drastic difference in SMR and growth based on density (see online resource 1A, 1 C and 2A). Furthermore, the results from experiment A demonstrated that growth medium exchange, the second parameter, has a significant effect on SMR regardless of density. This implies that cell metabolism based on feeding regime has a notable impact on SMR and marker expression I addition to seeding density. It is well understood and recognized that metabolic rates are a function of nutrient availability [61,62]; it is therefore unsu rprising that conditions with a higher cell load exhibit a lower metabolic rate measured by SMR as nutrient depletion will result in decreased nutrient availability overall. Thus, if these parameters are not adequately controlled the resultant cell out put quality would be highly variable and incomparable between different process runs and undo ubtedly between different sites, which would compromise the use of the cells be it as a reference standard, or therapeutic product. This highlights the necessity of standardized culture procedures to maintain consistency and reduce variability. The requirement for standardized culture protocols for maintenance of cell quality The above rationalizations highlight that even in a cell line that is considered as a stable reference point [25,26,30], differences can be observed due to changes in parameters and conditions, over time. These different parameters, such as when to perform medium exchanges and the use of seeding densities that vary, i.e. through the use of nonstandardized seeding densities or split ratios, are often left to the operator's discretion in many protocols. Evidently, this results in noteworthy effects on cell behavior and characterization outcomes. Split ratios are not best practice as innately cells will grow differently from passage to passage. For example, a split ratio of 1:3 can be drastically different from passage to passage, particularly if the cells grow at significantly different rates. In ad dition, split ratios that are based on observed confluency, are also likely to cause variation in cell gro wth dynamics. This is due to the subjective nature of observed confluency and its lack of accurate representation of cell number, therefore potentially resulting in non-ideal metabolite profiles that can influence cell growth behavior, producing further cell system inconsistency. Differing levels of variability and non-conformity are detrimental to the successful use of reference cell lines, especia lly for those intended to be a standardized QC reference. During the process development and training, operators should record possible areas for error, process deviation and associated risks using a centralized database, overseen by a qualified person (QP) [42]. All operators are assumed to have successfully completed standardized training and assessment of cell culture practice, liquid handling, aseptic technique and equipment operation. Also, they must be able to demonstrate competence in the assessment of any process against quality attribute (QA) criteria using appropriate assays and techniques [63]. Protocols should be written to minimize operator's discretion and limit the content to the technical ability to run the assays and equipment. Furthermore, detailed standard operating protocols (SOP) inclusive of risk mitigation strategies should be referred to for all procedures and equipment (including maintenance/calibration) [42]. This is a requirement of all laboratory users within not only our facility but all institutions working to Good Laboratory Practice (GLP) guidelines [27,64,65]. This is of relevant importance as the use of reference cell line has seen prominence due to the increase in chimeric antigen receptor T-cell (CAR-T) usage [66,67], which use them as QC reference standards for flow cytometry [4]. Interestingly, flow cytometry analysis revealed marker profile differences; even when the cells were at the prescribed density of 66,667 cells/cm [2], showing that despite a consistent culture process and protocol there is inherent variation in marker profiles through the 10 passages. This further emphasizes the need to standardize cell seeding within protocols to minimize and control variation as much as possible. This can be achieved by obtaining cell counts, ensuring that the input and output cells numbers are used to maintain optimal cul ture conditions for growth, without entering regions of metabolic instability. Consequently, this allows for greater control and consistency of the culture system as the seeding density and nutrient levels are predefined. The use of cell time, which is a concept that can be used to quantify the capability of a given volume of medium to sustain the growth of a given number of cells for a specific period, ensures that the culture system given the set density and nutrient availability will not enter a region of metabolic strain. This is important since imbalances in key nutrients including glucose levels have previously been shown to result in limitation of cell growth [7,60], and importantly for reference lines, impact the stability of marker expression. Metabolic strain is potentially problematic if not controlled, as reference lines are used for QC of CTPs need to demonstrate their stability over time, using gene and immunocytochemistry marker expression. It has been illustrated that cell system behavior is affected by protocol parameters; experiments A and B demonstrated that feeding regimes (medium exchange) had a significant effect on SMR and phenotypic marker expression when compared to seeding density. This highlights that differences in culture parameters do cause variation; however, it was evident that when parameters were controlled that less variation occurred. Prior experiments detailed in the supplementary information (Online resource 1) showed that there was no difference in SGR between different seeding densities when the feeding regime was maintained throughout the culture period. It has been demonstrated through a range of experiments that density was not the sole factor that can cause variation especially when the flasks were seeded at the same density. Instead it was clear that other protocol parameters themselves or when compounded with seeding density resulted in variation. Here it has been demonstrated that both density and the feeding regime had a major effect on cell culture variation. Lower densities have been shown to have distinctly different behaviors compared to higher-density cultures in terms of their SMR, a combinatory combined effect of density and feeding regime has been shown to cause variation in SGR, SMR, phenotype and expression of some genes. However, it is unclear under what mechanism medium exchange or lack of medium exchange influences cell characteristics such as SMR and phenotypic marker expression. This work provided a demonstration of biological variation arising when culturing a reference cell line with pluripotent attributes under normal laboratory operating controls. It also identified the extent to which key culture parameters contributed to this variation and therefore are critical to control for appropriate application of a reference cell line. Systematic culture trends and patterns observed in biological markers were identified that, if not accounted for, would lead to noise in ruler line applications. These attributes were identified for a range of key markers including cell number, metabolic behavior, surface markers and gene expression. Conclusion The present work illustrates how culture conditions impact cell characteristics, notably, SMR and phenotype. This demonstrates the complex interactions bet ween gene expression, phenotype and feeding regimes that cannot be accurately represented by growth rate and cell counts alone. The lack of robust, well-defined, and standardized protocols results in compounded variation in culture systems. This is due to differences in inter-lab /individual decision-making processes based upon protocol parameters, including observed confluency to gauge when to perform a cell passage. This highlights that non-standardized, ambiguous protocols can easily generate differences over time. At present, this work is yet to be repeated and the results confirmed in other laboratory environments. However, it will provide the basis for others to do so in the near future since standardized protocols are crucial not only from a uniformity perspective but also for ensuring process reproducibility and comparability; particularly in decentralized manufacturing of cell-based products. This is integral to the pragmatic and successful use of reference lines in cell therapy manufacturing and the manufacturing of CTPs in general. The present work demonstrates that culture conditions have a significant impact on cell characteristics, notably on specific metabolic rate SMR and phenotypic marker expression. Thus, demonstrating that there is a. This demonstrates the complex interactions between gene expression, cell phenotype and the feeding regimes that cannot be accurately represe nted by growth rate and cell counts alone. The lack of robust, well-defined and standardized cell-culture protocols result in compounded variation in the culture systems. This is due to differences in inter-lab and individual decision-making processes based upon protocol parameters, such as including observed cell confluency as a to gauge of when to perform a cell passage. This highlights that use of non-standardized, ambiguous protocols can easily produce differences over time, illustrating the need for standardization. standardized protocols. Such standardized Standard ized protocols are crucial not only from a uniformity point of view perspective but also for ensuring the reproducibility and comparability of a process particularly in decentralized manufacturing of cell-based products. The point of reproducibility is integral to the pragmatic and successful use of reference cell lines in cell therapy manufacturing and the manufacturing of CTPs in general. Research Highlights • Subtle, often unintentional variations in culture parameters have a significant impact on cultured cells. • The importance of balancing key nutrients, gene expression, phenotype and feeding regimes was shown. • Standardization is crucial to ensure manufacturing process reproducibility and comparability. • This is integral to the pragmatic and successful use of reference lines in cell therapy manufacturing. • Our experimental methodical approach demon strates the complex interactions and balance required between key nutrients, gene expression, cell phenotype and feeding regimes emplo yed that cannot be accurately represented by growth rate and cell counts alone. • Standardised protocols are crucial not only from a uniformity perspective, but also for to ensure process reproducibility and comparability, particularly in the decentralised manufacturing of cell-based products. • This is integral to the pragmatic and successful use of reference cell lines in cell therapy manufacturing.	2021-01-01T07:29:47.063Z	2020-12-30T00:00:00.000	{ "year": 2021, "sha1": "97231d9b684d1da72cc0e579320b252fc88ba820", "oa_license": "CCBY", "oa_url": "https://www.tandfonline.com/doi/pdf/10.1080/21655979.2020.1870074?needAccess=true", "oa_status": "GOLD", "pdf_src": "TaylorAndFrancis", "pdf_hash": "2faa1921a147205eb19d558f7cf4d9b9cbea5287", "s2fieldsofstudy": [ "Biology" ], "extfieldsofstudy": [ "Medicine" ] }
220327751	pes2o/s2orc	v3-fos-license	Deformity progression in congenital posteromedial bowing of the tibia: a report of 44 cases Background congenital posteromedial bowing of tibia (CPMBT) is a very rare birth defect, characterized by shortened bowed leg and ankle deformity. We described a single institution experience in the management of CPMBT. Methods we identified 44 CPMBT in 44 children. The age at presentation was 5.5 ± 5.6 years and the mean age at the final review was 10.1 ± 4.8 years. Radiographic evaluation included the antero-posterior and lateral inter-physeal angle (AP-IPA and L-IPA), the limb length discrepancy (LLD), the morphology of the distal tibia and the lateral distal tibial angle (LDTA). During the study period, 26 children underwent surgical treatment. Results the estimated curves showed a progressive spontaneous correction of both AP-IPA and L-IPA during growth, but a progressive increase of the LLD. The L-IPA showed a more predictable behaviour while the AP-IPA showed a scattered correction, with a wider variation of the estimated final angle. The final LDTA was 85.3° ± 4.2° and was correlated with the L-IPA (r = 0.5; p = 0.02). Among the 26 children who underwent surgical treatment, 23 cases had limb lengthening, 1 case had contralateral epiphysiodesis, 1 child underwent tibial osteotomy, 1 patient was treated by hemiepiphysiodesis of the distal tibia to correct ankle valgus deformity. Conclusions our study described the largest case series of CPMBT. A combination of surgical treatments, in a staged surgical process, should be tailored to the developmental characteristics of this abnormality. An experience-based algorithm of treatment is also proposed. Further studies are needed to understand which is the best strategy to correct this deformity during childhood. Level of evidence level IV prognostic study. Background Congenital posteromedial bowing of the tibia (CPMBT) is a very rare birth defect, firstly fully described in 1949 by Heyman and Herndon [1]. It has been generally considered a benign, self-solving condition, in contrast to the anterolateral bowing, associated with congenital pseudarthrosis of the tibia, and the anteromedial bowing, associated with fibular hemimelia [2]. CPMBT is obvious at birth, with a notable shortening and bowing of the leg and the foot resting dorsiflexed against the tibial shaft [2][3][4][5]. Although the cause of CPMBT remains unknown, a potential role of amniotic strains has been hypothesized [6]. This condition is generally unilateral and not associated with other abnormalities [1,7]. Several authors demonstrated spontaneous improvement of the tibial bowing within the first 3-4 years of life. Conversely, the limb length discrepancy (LLD) increases with age, until it reaches 4-7 cm at skeletal maturity [2,3]. Studies have shown that the amount of the tibial bowing at birth is positively correlated to the LLD at maturity [2,3,8]. Moreover, some authors noticed residual ankle valgus at maturity, but its incidence and relationship with the leg deformity have not been established [2]. To date, several therapeutic options have been proposed for the management of CPMBT, but the treatment of choice remains controversial. Therefore, we investigated a series of children presenting with CPMBT, treated at a single institution. We aimed to explain the behaviour of CPMBT during growth and the relationship between the tibial bowing, the leg shortening and the ankle deformity. These aspects could be useful in order to suggest a possible rationale of treatment. Case series The present study is a retrospective analysis of medical charts and radiographs of children affected by CPMBT, who were admitted at the Department of Pediatric Orthopedics and Traumatology from 1972 to 2016. Our institution is a tertiary referral center for pediatric orthopedics and traumatology, highly specialized in the treatment of complex deformities of the lower limb. All the charts and radiographs were analyzed by independent observers, who were not involved in the decision process about treatment and surgical management of the patients. During the study period, 44 CPMBT were identified in 44 children, 27 boys and 17 girls. All children had unilateral involvement and no cases were excluded from the present study. The right side was affected in 25 children whereas the left side was affected in 19. The age at presentation was 5.5 ± 5.6 years (range 0-15) and the mean age at the final review was 10.1 ± 4.8 (range 0-16). Twenty-six patients underwent surgery during the study period, while the follow-up is still ongoing and surgery not yet planned for the remaining eighteen patients. Overall, Twenty-five children reached skeletal maturity at the time of final review; of them, twenty-four received definitive surgical treatment and one boy is waiting for tibial lengthening. Radiographic evaluation The following variables were measured on serial sequential radiographs in order to assess the initial deformity and the spontaneous remodeling:1) the anteroposterior interphyseal angle (AP-IPA) and the lateral interphyseal angle (L-IPA), that are the angles measured between a line perpendicular to the proximal physis and a line perpendicular to the distal physis, on true anteroposterior and lateral views of the leg, respectively [2] Positive AP-IPA indicates medial bowing, while negative AP-IPA indicates lateral bowing. Positive L-IPA indicates posterior bowing, while negative L-IPA indicates anterior bowing (Fig. 1); 2) the leg length discrepancy (LLD) measured on long standing radiographs. The difference was expressed as crude length (LLDcm) and as percentage shortening as compared with the opposite side (LLD%); 3) the level of the distal fibular growth plate was graded according to the method of Malhotra et al. [9] ( Fig. 1S), while the extent of the wedging of the distal tibial epiphysis was graded according to the method of Shapiro et al. (Fig. 2S) [10]. 4) the lateral distal tibial angle (LDTA), medial proximal tibial angle (MPTA), anatomical lateral distal femoral angle (aLDFA) hip-knee-ankle angle (HKA) and mechanical axis deviation (MAD) measured on long standing radiographs in children approaching skeletal maturity [11,12]. Statistical analysis Data were entered in Excel, SPSS and nlme package in R. Continuous variables were expressed as mean ± standard deviation (SD), while dichotomous or ordinal variables were expressed as percentage and 95% confidence interval (CI). Exploratory univariable and multivariable analyses were performed to assess the relationships among the parameters of CPMBT. Normality was tested using the χ2 test for categorical variables and the Kolmogorov-Smirnov test for continuous variables. The Spearman's Rho correlation test was used to investigate the relationships among continuous variables. The differences between groups were determined using the Fisher exact test for categorical variables and the independent sample t-tests for continuous variables with normal distribution. Variables with skewed distributions (Kolmogorov-Smirnov test, p < 0.05) were tested with the Mann-Whitney U-test. In order to predict the spontaneous progression of both the deformity and the length discrepancy among patients over time, generalized linear mixed models, fitted by the maximum likelihood, were performed. A separate model was implemented for each parameter using the subject "patient" as a random factor. The best fitting model was chosen using the ANOVA R function, that compares the AIC, BIC and logLik values. A p-value of < 0.05 was considered statistically significant, and all reported p-values were 2-sided. Results The estimated curves showed a progressive spontaneous correction of both the AP-IPA and the L-IPA (Fig. 2a-b and Table 1), and a slight decrease in LLD% (Fig. 2c), but a progressive increase of the LLDcm (Fig. 2d), until a final estimated discrepancy of 4.3 cm (95% CI 3.7-5.0). Concerning the predictive model of spontaneous correction, we found that the best fitting curve was loglinear for the AP-IPA and LLDcm, exponential for the L-IPA and linear for the LLD%. Yet, the L-IPA showed a more predictable behaviour while the AP-IPA showed a scattered correction, with a wider variation of the estimated final angle (Table 1). The first case of CPMBT was diagnosed in our hospital in 1972 and was treated in 1978, at the age of 6 years. The patient sustained tibial lengthening by unilateral Wagner external fixator, then he underwent plate fixation after 2 months. The patient further developed delayed union treated with autologous bone graft and a subsequent fracture of the regenerated bone, treated by intra-medullary nailing. Two patients had intermediate surgical treatment during the study period. In one patient, a severe tibial deformity persisted at the age of 5.8 years, with an AP-IPA of 26°, a L-IPA of 6°a nd a LLD 5 cm (18%) (Fig. 4a-c). The patient was treated by corrective tibial osteotomy at the apex of the deformity, stabilized with two Kirschner wires (Fig. 4d). At the latest follow up, 3.5 years after surgery, the patient had a residual deformity with an AP-IPA of 9°a L-IPA of 5°a residual LLD of 5 cm (16%) and an ankle valgus deformity of 70° (Fig. 4e). To date, the patient is waiting for final correction and leg lengthening. One patient underwent temporary medial hemiepiphysiodesis of the distal tibia at the age of 10.4 years, for correcting ankle valgus (LDTA = 79°). At the latest follow-up visit, 2 years after the operation, the ankle axis was restored (LDTA = 90°) and the patient is currently waiting for final lengthening, to equalize a residual LLD of 4 cm (11%). One patient, who presented with residual LLD of 2.2 cm (7%) and no relevant deformities of the tibial and ankle axis, was treated at the age of 11.4 years with contralateral epiphysiodesis of the proximal tibia. The patient achieved limb length equalization 30 months after surgery, without further complications. Twenty-two patients were treated at skeletal maturity (13.7 ± 1.8 years). The pre-operative radiographic data are summarized in Table 2. All these patients received leg lengthening by circular external fixation. We never bridged the ankle, but preferred foot splints attached to the fixator by elastic bands, in order to promote immediate full weight bearing and active rehabilitation of the ankle. Bifocal lengthening was accomplished in one case (Fig. 5), in which the deformity was corrected through the distal osteotomy, while lengthening was carried out at the proximal osteotomy, in order to reduce the time in frame, due to the better healing potential of the metaphyseal osteotomy. In this group, the final limb equalization was achieved in all patients, but we experienced 11 complications in 8 patients. According to Lascombes et al. [13], there were 2 grade IIa complications (2 operations to change or modify the frame) and 9 grade IIIa complications (1 delayed bone healing, 3 fibular nonunion, 1 knee joint stiffness; 2 achilles tendon shortening; 2 malalignment with residual anterior bowing of the tibia). Moreover the final LDTA was 83.3°± 5.2°(range 72°-90°), with 9 patients out of 14 showing a residual LDTA < 85°. Discussion This study reports the largest series of children with CPMBT in the current literature. To date, about 200 cases of CPMBT have been reported in the available literature [1-5, 7, 8, 14-18]. For years CPMBT has been considered a benign self-solving condition, due to the virtually absent risk of fracture or pseudarthrosis and the natural tendency to the spontaneous resolution, with minimal residual deformity. Nonetheless, we noticed that the spontaneous correction of the bowing was sometimes incomplete, the LLD was frequently wide and a residual ankle valgus could persist at the end of growth. We found some interesting correlation between the AP-IPA and the L-IPA, and between the AP-IPA and the LLD%. In other words, the greater is the angular deformity at birth, the wider will be the LLD at skeletal maturity. This finding is confirmed by previous reports [2,3,8,18]. We found that the posterior bow corrected more efficiently than the medial bow, but the spontaneous correction of the posterior bow was accompanied by an increase of the ankle valgus. This phenomenon was also noticed by other authors [2,3] and could be explained by a possible pathogenetic hypothesis of CPMBT, namely the "amniotic band theory" [6]. According to this theory, CPMBT is caused by a partial rupture of the amniotic sac, during pregnancy. This rupture can lead to a focal growth arrest of the antero-lateral distal portion of the leg, causing the typical clinical aspect of CPMBT. Then, after birth, the diaphyseal deformity of the leg improves spontaneously, according to the Wolff's law, thus the leg corrects more efficiently in the sagittal plane (plane of walking) rather than in the coronal plane. It is plausible that part of the spontaneous diaphyseal improvement is determined by an asymmetric growth of the distal tibial physis, leading to consequent ankle valgus [2,9,10]. However, our study cannot definitely confirm this hypothesis, since we did not find significant correlations among Malothra, Shapiro scores and LDTA; further studies are needed to clarify this effect. The management of CPMBT is still debated since there is no complete evidence about the optimal treatment strategy. Given the tendency to the spontaneous correction of the deformity, some authors suggest conservative treatment consisting of manipulation, serial casting, orthoses and shoe lifts; then, a limb equalization is proposed during late childhood, if needed [4,7,14]. Currently, there is no evidence that the use of braces and orthoses may improve the angular correction in CPMBT, since it occurs spontaneously and independently by the use of orthoses. The main goal of braces and orthoses is to aid walking and balance, while the patient is too young for the surgical treatment. There are three reasons for surgical intervention in CPMBT: equalize the LLD, correct the ankle valgus and correct the residual bowing of the tibia [2]. Nonetheless, there are no clearly defined guidelines for the surgical treatment of CPMBT. Regarding the LLD, we found that the final discrepancy at skeletal maturity averaged 4.3 cm, corresponding to 13% of the length of the unaffected leg; this ratio, remains rather constant during growth, as confirmed by many previous reports [2][3][4][5]18]. In our series, all the children who reached skeletal maturity underwent limb lengthening by circular external fixator while only one child underwent contralateral epiphysiodesis. Albeit contralateral epiphysiodesis has been recommended in CPMBT, due to the lower risk of complications compared to limb lengthening [3], aesthetical issues can raise due to the loss of body height. Moreover, recent concern has mounted regarding the potential risk of compromising the morphology of the proximal tibia, when a large, congenital LLD must be addressed [19,20]. Therefore, we suggest to reserve this treatment only for children in which the LLD% at 10-12 years is less than 10%, corresponding about 2 to 3 cm. Regarding the tibial bowing, the majority of cases improved spontaneously by the end of growth. Our behaviour consisted in a "waiting strategy", using braces and orthoses until skeletal maturity, then, correcting in a single stage the length and the potential residual bowing. In our opinion, this strategy should reduce the risks for the patient and the costs for the health service. Nonetheless, a more pronounced reduction of the angular deformity was noticed during the first 6 years of life; thereafter, the rate of spontaneous correction decreased markedly. Many authors reported that, in CPMBT, the greatest rate of correction is observed during the first year of life, then, rapidly decreases until the age of four [2-4, 15, 18]. This aspect may have practical implications, because an extreme bowing of tibia in a school-age child might hamper even the possibility of using braces to aid walking. In this scenario, some authors suggested early corrective osteotomy at the apex of the deformity, by the age of 3 to 6 years [2-4, 15, 16]. It has been argued that the intense periosteal activity at this age allows for early bone healing of the diaphyseal osteotomy; furthermore, the overgrowth of the tibia due to the physeal stimulation and the tibial straightening could potentially contribute to the leg length equalization [15]. We treated only one case by early corrective osteotomy of the bowed tibia: although we achieved a rapid healing of the osteotomy and a perfect alignment of the tibia, we observed a progressive partial relapse of the bowing, an important ankle valgus, while the LLD remained unchanged. These Table 2 finding are consistent with those reported by Johari et al. [4], suggesting that the early tibial osteotomy should be proposed only in case of severe, disabling bowing, as an intermediate treatment, to avoid complex bracing and allow walking with simple foot orthosis or shoe lift. Based on our experience, we recommend early tibial osteotomy, if the AP-IPA does not decrease below 21°b y the age of 6 years (corresponding to more than two standard deviations of AP-IPA within our cohort). Nonetheless, further studies are needed to confirm this threshold. Another reason to recommend early tibial osteotomy, is related to the possibility to perform intramedullary lengthening by telescopic nails at skeletal maturity [21]. This technique has been reported as safe, effective and more tolerated by the patients, in comparison with external fixation. Nevertheless, intramedullary nailing is more simple, safer and more effective when applied to a straight tibia rather than to a bowed tibia. Yet, simultaneous correction of the bowing and lengthening by circular external fixation is not simple, requires high compliance by the child and the parents, high Although we have about 40-year experience with the Ilizarov technique [22], we experienced a relevant rate of moderate and severe complications, in line with other reports: Kaufman et al. [5] reported 17 mild to severe complications in 11 CPMBT treated by external fixation; Johari et al. [4] described complications in all the 6 cases treated by external fixation; Wright et al. [18] reported 16 complications in 17 children treated by external fixation. Probably, computer-assisted hexapods devices could be more effective in achieving lengthening and correction compared to the conventional Ilizarov method, reducing the lengthening index for faster correction, but potential advantages must be balanced by costs, since the Ilizarov device is much less expensive, compared to Hexapod [23]. Furthermore, we suggest to perform the lengthening procedures closer to the skeletal maturity, since we did not experience any recurrence of the limb length inequality, in contrast with other authors, who reported relapse of the limb length inequality, if the lengthening procedure was performed during growth [2][3][4][5]18]. Finally, in our series about one third of children with CPMBT presented a valgus ankle by the end of growth (LDTA < 85°). This issue has been reported previously [2,4]. Although the normal range of the LDTA has been established [11,12], an exact cut-off to define a pathologic deformity has not been clearly defined, with proposed values varying from 5°to 10°of valgus [20,21,24,25]. In our series, only one patient underwent distal tibial hemiepiphysiodesis to treat ankle valgus. Nonetheless, it is our opinion that a medial distal tibial hemiepiphysiodesis should be performed if the valgus inclination of the distal tibial articular surface persists by the age of 10 to 11 years, when sufficient growth potential is still present [18,19]. This simple procedure can effectively realign the ankle by minimally invasive surgery, reducing complications and avoiding more demanding surgeries at the end of growth [24][25][26][27][28][29][30]. The residual "S" shape within the long axis of the tibia could not require further correction, if the joint lines of knee and ankle are parallel to the floor. Based on our findings and on the available literature, we identified crucial steps and thresholds for an experience-based algorithm of surgical decision-making ( Fig. 6): 1) diaphyseal tibial osteotomy could be performed at 5-6 years of age, in case of severely disabling bowing deformity (generally AP-IPA > 21°); 2) distal tibial medial hemiepiphysiodesis could be indicated at 10-11 years, in case of important ankle valgus (LDTA < 85°); 3) minor LLD (discrepancy ≤10%, corresponding to about 3 cm) can be addressed by contralateral epiphysiodesis at the age of 11-12 years; 4) severe LLD (discrepancy > 10%) should be treated close to skeletal maturity (13-15 years) by leg lengthening, possibly combined with simultaneous correction of the residual bowing. Limitations Despite our study describes the largest series of CPMBT in the available literature, some limitations must be highlighted. We conducted a retrospective analysis of cases collected across more than 40 years, sometimes with incomplete information and missing data. Twelve cases out of forty-four had no sufficient radiographic follow-up available for assessing the progression of the deformity. Furthermore, the cases were not collected uniformly at birth and 16 children had the first radiographic evaluation when they were older than 5 years. The majority of long leg radiographs were taken without the patient balanced on blocks. Although this unbalanced position should not significantly influence the assessment of the leg length discrepancy, it could introduce a source of bias in assessing the mechanical axis of the lower limb. Finally, almost half of cases did not reach skeletal maturity, thus not undergoing the final correction. We aim to maintain a close follow-up of those patients who have not received the definitive treatment and update our results within five to 10 years. These limitations were encountered in all previous reports about CPMBT [2][3][4][5]18], emphasizing the difficulty in obtaining complete information about rare abnormalities, where less than one case is diagnosed per year, even in highly specialized institutions. In order to address this issues and minimize biases, we used mixed effect models, a complex statistical approach that allows to maximize the prediction power from small and heterogeneous groups of subjects with missing data; nonetheless, we are aware that our results could be weakened by poor statistical power and precision. Conclusions Despite its supposed benignity, CPMBT is a complex deformity, in which the spontaneous correction of the angular deformity is inconstant and incomplete, the LLD is generally wide, and a substantial ankle valgus can be observed by the end of growth. A combination of surgical treatments (osteotomies, epiphysiodesis or hemiepiphysiodesis, leg lengthening) in a staged multistep surgical process, should be tailored on the developmental characteristics of the deformity. This approach may accomplish the correction of the deformity, minimizing complications and failure, and helping surgeons and parents in the decision process. Further studies are needed to understand which is the best strategy to address this rare deformity during childhood.	2020-07-04T13:59:16.354Z	2020-03-20T00:00:00.000	{ "year": 2020, "sha1": "9f6ebbb2a59d24f2889967149c9b143301a956ce", "oa_license": "CCBY", "oa_url": "https://bmcmusculoskeletdisord.biomedcentral.com/track/pdf/10.1186/s12891-020-03408-w", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "9f6ebbb2a59d24f2889967149c9b143301a956ce", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
116883921	pes2o/s2orc	v3-fos-license	A Microfluidic System for Studying the Effects of Disturbed Flow on Endothelial Cells Arterial endothelium experience physical stress associated with blood flow and play a central role in maintaining vascular integrity and homeostasis in response to hemodynamic forces. Blood flow within vessels is generally laminar and streamlined. However, abrupt changes in the vessel geometry due to branching, sharp turns or stenosis can disturb the laminar blood flow, causing secondary flows in the form of vortices. Such disturbed flow patterns activate pro-inflammatory phenotypes in endothelial cells, damaging the endothelial layer and can lead to atherosclerosis and thrombosis. Here, we report a microfluidic system with integrated ridge-shaped obstacles for generating controllable disturbed flow patterns. This system is used to study the effect of disturbed flow on the cytoskeleton remodeling and nuclear shape and size of cultured human aortic endothelial cells. Our results demonstrate that the generated disturbed flow changes the orientation angle of actin stress fibers and reduces the nuclear size while increases the nuclear circularity. INTRODUCTION Endothelial cells, lining the inner surface of blood vessels, are in direct contact with the flowing blood, and their response to physiological as well as pathological flow dynamics affects vascular health (Chiu and Chien, 2011;Baratchi et al., 2017a). Endothelial cells experience various flow patterns and hemodynamic forces across the vascular system. Endothelial cells lining along the straight segments of the arterial trees experience a high-shear laminar flow essential for their physiological functions such as flow-mediated dilation (Chiu and Chien, 2011). In contrary, the endothelial cells lining along the arterial branches and curvatures experience a low-shear disturbed flow due to the presence of secondary flows in the form of vortices, which causes endothelial dysfunction and atherosclerosis (Caro et al., 1969(Caro et al., , 1971Bharadvaj et al., 1982;Suo et al., 2008). Flow-induced alignment of the endothelial cells has been reported previously (Alenghat and Ingber, 2002;Wang et al., 2013;Yoshino et al., 2017). Recent advances in endothelial mechanobiology have also demonstrated that different classes of mechanoreceptors control the physiological function of endothelial cells, including cytoskeleton remodeling, gene expression, cell viability index and calcium homeostasis that could consequently control myogenic tone (Ingber, 2006;Chatzizisis et al., 2007;Shemesh et al., 2015). Shear-induced activation of various proteins such as RAC1 (Tzima et al., 2002) and Syndecan 4 (Baeyens et al., 2014) as well as enzymes such as RhoGTPases (Kroon et al., 2017) have also been reported to control the alignment and cytoskeleton remodeling of endothelial cells. A variety of models have been developed to study the effect of disturbed flow on endothelial cells. These include in vivo models through surgical intervention (Katoh et al., 2007;Harding et al., 2018), ex vivo models using the endothelium at naturally occurring disturbed flow regions of the vessel (Katoh et al., 2007), and in vitro models using emerging microfluidic systems (Estrada et al., 2011). Among these models, microfluidic systems offer unprecedented advantages such as reduced cost and complexity of experiments, decreased volume of reagents (Katt et al., 2016;Ho et al., 2018;Yaman et al., 2018), and importantly provide predictable and controllable disturbed flow patterns based on the geometric specifications of the system (Rezvan et al., 2011;Balaguru et al., 2016). Despite these advantages, the majority of existing microfluidic systems generate localized disturbed flow patterns and thus are not suitable for studying the gene and protein expression of cultured cells under disturbed flow (Rezvan et al., 2011;Balaguru et al., 2016). Here, we developed a microfluidic system with an array of ridge-shaped obstacles patterned along its entire surface. This allows for the generation of vortices and, in turn, low-shear disturbed flow regions in the cavities located between the successive ridges. This feature is used to quantify the effect of disturbed flow on the actin cytoskeleton remodeling, nucleus shape and size of cultured human aortic endothelial cells. Fabrication of 2D Parallel-Plate Flow Chambers Ridged flow chambers were fabricated in two parts: (i) the main PDMS block consisting of a 500 µm tall channel and (ii) a PDMS film consisting of an array of 100 µm tall ridges. The main PDMS block was fabricated using a silicon wafer mold diced into a 5 mm × 50 mm × 500 µm section and adhered on to a 4-inch silicon wafer and enclosed with a 6 mm tall Teflon barrier (22 × 60 × 6 mm) (Figure 1a i ). PDMS (Sylgard R 184, 10:1 w/w base to curing agent) was poured into the Teflon barrier and cured at 80 • C for 30 min. The PDMS block was then peeled off the mold and inlet, and outlet holes were punched using a 4 mm biopsy punch (Harris Unicore) (Figure 1a ii ). The PDMS film with ridges was fabricated using a mold consisting of an array of rectangular grooves (100 µm × 100 µm × 5 mm) separated by 400 µm gaps. The mold was patterned in SU-8 3050 photoresist (MicroChem) using a high-resolution chrome mask (Figure 1b i ). PDMS was then poured onto the mold and spun at 100 rpm for 30 s before curing at 85 • C for 5 min. The resulting PDMS film was then peeled off the mold revealing the array of rectangular obstacles. The PDMS film was cut to 22 × 60 mm prior to the assembly of the two parts (Figure 1b ii ). The PDMS main block and PDMS film with ridges were manually aligned and clamped together using a PMMA clamp resulting in a ridged flow chamber with rectangular obstacles (Figure 1c). A third mold, similar to the PDMS film mold, but without ridges was used to fabricate flat PDMS films for use as the control. Computational Fluid Dynamics (CFD) CFD simulations were conducted to predict the formation of vortices along the ridged microfluidic channel. This involved solving the differential equations governing the balance of mass and momentum, also known as Navier-Stokes equations. Simulations were performed using ANSYS Fluent software (ANSYS Inc.). Simulations were conducted in 3D and under steady-state conditions, considering the cell culture medium as an incompressible and Newtonian liquid. Flow was considered laminar due to its low Reynolds number. Boundary conditions included desired flow rates at the inlet, ambient pressure at the outlet and no-slip at the walls. The density and dynamic viscosity of the cell culture medium at 37 • C were considered as 998 kg/m 3 and 7 × 10 −4 Pa.s, respectively. Cell Culture and in-vitro Generation of Laminar vs. Disturbed Flow HAECs (Lonza) at early passages of 2 to 5 were used in this study. HAECs were cultured inside the microfluidic channels at the density of 5 × 10 6 cell/mL to produce a confluent layer of endothelial cells within 24 h. The volume of liquid inside the channel was ∼100 µL. The channels were pre-coated with 10 µg/mL MaxGel human extracellular matrix extracts (Sigma) according to the manufacturer's specification. Flow experiments were carried out for 16 h at 37 • C in a humidified atmosphere with 5% CO 2 . The flow was applied through the channels using a peristaltic pump (OINA QP6 LAB High Accuracy) at flow rates of 20 mL/min and 6 mL/min. Image Processing and Analysis of Stress Fibers The orientation of the cells and stress fibers was determined using an automated image processing algorithm written in MATLAB, as detailed elsewhere (Karlon et al., 1999;Kaunas et al., 2005;Ranade et al., 2014). Briefly, the algorithm computed the intensity gradients of the image in the horizontal and vertical directions by convolving the image with a spatial gradient filter. By treating these gradients as components of a vector field, the algorithm computed pixel-by-pixel magnitude and direction information. Local dominant orientations were then determined for each pixel by constructing a histogram of orientations from the magnitude and direction information and choosing the orientation with the largest value. Each histogram was constructed using a small subregion of 20 × 20 pixels, centered around the pixel of interest, and evaluating their deviation from a set of angles ranging from −89 to 90 degrees relative to the horizontal. For more details, we refer the reader to Kroon et al. (2017). Statistical significance was assessed with a global Watson's U2 test, and statistics were computed using the circular statistic toolbox (Berens, 2009). Quantification of Nuclear Shape and Size Image processing and calculations of nuclear shape and size were performed using NIS element software (Nikon Instruments Inc.). For statistical analysis, one-way ANOVA was performed using Prism 7.02 (GraphPad software) and P < 0.05 was considered significant. Characterization of Disturbed Flow Microfluidic structures with the ridged or grooved patterns induce secondary flows in the form of vortices. This feature has been utilized for enhancing passive mixing of liquids (Stroock et al., 2002) as well as cell-based studies involving sorting (Yan et al., 2017), capturing (Manbachi et al., 2008;Stott et al., 2010;Hsiao et al., 2016;Wang et al., 2017) and forced tethering (Choi et al., 2014) of target cells. An extensive set of numerical simulations were performed to analyze the disturbed flow patterns inside our microfluidic channel with integrated ridge-shaped obstacles. Our simulations indicated the formation of two vortices along the two corners of the cavity located between the successive ridges, which are referred to as "leading" and "trailing" vortices (Figure 1d). Our experiments confirmed the generation of "leading" and "trailing" vortices at the two corners of the cavity (Supplementary Video 1). Experiments were performed at a flow rate of 5 mL/min using a suspension of 1 µm polystyrene particles to facilitate the visualization of vortices. Increasing the flow rate of the cell culture medium led to the expansion of the vortices. In this manner, the two vortices occupied almost the entire cavity region at flow rates higher than 15 mL/min. Further increase of the flow rate led to merging of "leading" and "trailing" vortices. The magnitude and distribution of the wall shear stress were governed by the configuration of the vortices (Figure 1e). Numerical simulations revealed the existence of three localized shear peaks at 5 mL/min, corresponding to the "leading" vortex, vortex-free zone and "trailing" vortex. A similar pattern was observed at 10 mL/min. Two localized shear peaks were obtained at 15 mL/min due to the diminishing of the vortex-free zone. In comparison, only one localized shear peak was obtained at 20 mL/min due to the merging of "leading" and "trailing" vortices. The shear stress induced by the "leading" vortex increased linearly with respect to the flow rate (Figure 1f). A similar trend was obtained for the control channel with no ridges (Figure 1f). Based on these results, the cellular experiments were performed at a flow rate of 20 mL/min to ensure (i) the endothelial cells cultured inside the "channel with ridges" experience a low shear stress under homogenous disturbed flow conditions while (ii) the endothelial cells cultured inside the "channel without ridges" experience physiological shear stress under laminar flow conditions (Chiu and Chien, 2011). Disturbed Flow Affects the Orientation of Actin Stress Fibers and Nuclear Shape Change Flow shear stress controls different endothelial phenotypic characteristics, including cell morphology, cytoskeleton remodeling and gene expression (Caro et al., 1969(Caro et al., , 1971Bharadvaj et al., 1982;Suo et al., 2008). Cytoskeleton proteins play important roles in maintaining the shape and integrity of the cells as well as transduction of shear stress from the luminal surface of endothelial cells to the cytosol (Loufrani and Henrion, 2008). Actin stress fibers consist of bundles of 10-30 actin filaments that are held together by actin crosslinking protein known as α-Actinin (Pellegrin and Mellor, 2007). Stress fibers are very important transducers of shear stress and transmit the stress to various intracellular locations (Franke et al., 1984;Wechezak et al., 1985). HAECs were cultured overnight inside the channels with ridge-shaped obstacles (disturbed flow), without ridges (laminar flow) as well as inside Petri dishes (static condition). The flow rate of the medium was set to 20 mL/min. Under this condition, the cells cultured under disturbed flow experienced low shear stress with a maximum of 3 dyne/cm 2 , whereas the cells cultured under laminar flow experienced physiological shear stress with a magnitude of ∼10 dyne/cm 2 , as detailed in the previous section. First, we studied the orientation of stress fibers developed under the laminar flow, disturbed flow, and static condition. Under the laminar flow, HAECs stress fibers were highly orientated to the direction of flow (R 2 = 0.93), with 71.59 ± 3.2% of stress fibers having an orientation angle of 0-30 • . In contrary, in the presence of the disturbed flow, the majority of stress fibers were oriented perpendicular to the direction of flow (R 2 = 0.92), with 59.74 ± 1.9% of stress fibers having an orientation angle of 60-90 • (Figure 2A). In comparison, under static condition, the orientation of stress fibers did not follow any specific trend (Figure 2A). This trend can be clearly seen in the contours of the stress fiber orientation angle, obtained by our automated image processing algorithm. The trend shows a transition from green-blue colors (corresponding to the orientation angles in the range of ±30 • ) toward red-purple colors (corresponding to orientation angles outside the range of ±70 • ) ( Figure 2B). This trend can also be seen in the histograms of the stress fiber orientation angle, which clearly shows the significant changes of stress fiber orientation under disturbed flow condition (Figure 2C). Consistent results were obtained in five independent experiments ( Figure 2D). Next, we conducted a control experiment, by applying a flow rate of 6 mL/min through a straight microfluidic channel without ridges, to induce a shear stress of 3 dyne/cm 2 . Our experiments indicated that stress fibers were aligned perpendicular to the direction of the flow, similar to what we observed under the disturbed flow conditions inside the channel with ridges. This observation suggests that the response of HAECs to the disturbed flow is dependent on the magnitude of shear stress, rather than the flow pattern of the small vortices formed between the ridges (Supplementary 1). Alignment of endothelial cells to the direction of flow has been observed previously (Alenghat and Ingber, 2002;Wang et al., 2013;Yoshino et al., 2017). It has been demonstrated that endothelial cells have the threshold of 10 dyne/cm 2 to align to the direction of flow which is similar to what we have also observed in our preliminary studies (Wang et al., 2013;Ostrowski et al., 2014). Further, porcine valvular endothelial cells have been demonstrated to orient perpendicular to the direction of flow in response to the shear stress of 20 dyne/cm 2 in comparison to the aortic endothelial cells that are aligned in parallel to the flow direction at the same shear stress level (Butcher et al., 2004;Baratchi et al., 2017b;Nguyen et al., 2018). Morphology of endothelial cells (shape and size) and actin cytoskeleton is reported to be different at the branch point of the aorta where the blood flow is disturbed compared to the regions of the aorta where blood flow is laminar (Katoh et al., 2007). This can be attributed to the different responses of endothelial cells to different shear levels, as reported in our work. The nucleus in eukaryotic cells is the site of transcriptional regulation and receives the mechanical stress that is transmitted via the cytoskeleton. Both intra and extracellular forces affect the nuclear shape and structure (Dahl et al., 2008). To evaluate this phenomenon, we compared the nuclear shape index and size of HAECs under the laminar flow, disturbed flow and static conditions. Disturbed flow significantly increased the nucleus circularity index of HAECs compared to the laminar flow (P < 0.001) and static condition (P < 0.001) (Figure 2E). In comparison, the nuclear size had a significantly smaller area under the disturbed flow compared to the laminar flow (P < 0.01) and static condition (P < 0.01) (Figure 2F). The observed nuclear shape change in the presence of different flow dynamics can be attributed to the change in the mechanically induced signaling or macromolecular conformational changes related to change in gene expression, in the presence of different flow dynamics (Davies, 2009). CONCLUSIONS Here, we reported the microfluidic-based in vitro model for generating the disturbed flow that mimics the pathological flow patterns of arterial branch point and curvatures. The observed flow disturbance is due to the formation of vortices along the ridges. We showed that the expansion of vortices and the magnitude of the wall shear stress can be tuned by varying the flow rate of the cell culture medium through the system. At the flow rate of 20 mL/min, the vortices filled the entire cavity region between the neighboring ridges, inducing a maximum wall shear stress of 3 dyne/cm 2 along the bottom surface of the channel where endothelial cells were cultured. To demonstrate the capability and efficacy of this model, we studied the effect of the disturbed flow on endothelial cytoskeleton remodeling and stress fiber formation as well as its nuclear shape and size. We used a MATLAB code to quantify the orientation of actin stress fibers under the laminar flow, disturbed flow and static condition. Our results indicated that the generated disturbed flow affects the morphology and cytoskeleton remodeling of HAECs. Under the laminar flow, endothelial cells were aligned to the direction of flow and formed actin stress fibers, whereas the low-shear disturbed flow caused endothelial cells to orient perpendicular to the direction of flow. Furthermore, the HACEs exhibited a significantly higher nucleus circularity index and smaller nuclear size in the presence of disturbed flow. These observations demonstrated the suitability of the presented microfluidic system for studying the effect of disturbed flow on the biology of endothelial cells, providing unique opportunities for evaluating the effect of the disturbed flow on the expression and function of different mechanoreceptors in endothelial cells. AUTHOR CONTRIBUTIONS FT-L fabricated the microfluidic device. PT conducted experiments and analyzed the results. CG wrote the MATLAB code for automated processing of images. NN conducted experiments. EP wrote the manuscript. KK performed numerical simulations and wrote the manuscript. SB designed the study, performed experiments, wrote the manuscript and supervised the work.	2019-04-17T15:33:47.523Z	2019-04-17T00:00:00.000	{ "year": 2019, "sha1": "b2cc8f8614e79c44d18111b374cad712e5be96b9", "oa_license": "CCBY", "oa_url": "https://www.frontiersin.org/articles/10.3389/fbioe.2019.00081/pdf", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "b2cc8f8614e79c44d18111b374cad712e5be96b9", "s2fieldsofstudy": [ "Engineering", "Medicine" ], "extfieldsofstudy": [ "Chemistry", "Medicine" ] }
219634235	pes2o/s2orc	v3-fos-license	Soil loss as a desertification risk indicator: mapping and simulation in the Salitre River Sub-Basin, Northeast Brazil Discussions on desertification frequently highlight soil erosion as a striking feature of this phenomenon. In particular, the high spatial density of gullies represents a strong indication of the formation of desertification hotspots. In this study, through field activities and Monte Carlo simulations, we estimated the volume of soil loss by linear erosion on the slopes of the middle course of the Salitre river in the North of Bahia State. This estimative contributes to the recognition of a desertification process in the studied local. The lengths of the gullies and rills, visible through high-spatial-resolution satellite images, were vectorized. The width and depth of the Linear Erosion Features (LEFs) were measured through field study and recorded via geoprocessing. Statistical treatment was applied to the data to indicate the probability of occurrence of the width and depth classes. Subsequently, the Monte Carlo simulation was used to indicate the volume of soil removed from the slopes by the linear erosion process. Several ramified systems of LEFs are identified and mapped. Monte Carlo simulation fits the measured data well. Estimates indicate that linear erosion event eroded approximately 450,000 m3 of soil in an area of 2,000 hectares, which indicates extreme land degradation. INTRODUCTION Despite being hugely controversial, discussions about the desertification problem focus on environmental degradation in areas that are vulnerable and pressed by anthropic action. The United Nations Convention to Combat Desertification (UNCCD) defines the desertification phenomenon as the degradation of the lands in arid, semiarid, and dry sub-humid areas, resulting from various factors, including natural causes and human activities, mainly associated with the inadequate use of soils, water, and vegetation (United Nations, 1994). Desertification is associated mainly with the degradation of soil, vegetation, and changes in climate and hydrological conditions (D'Odorico et al., 2013). These aspects yield significant changes in ecosystems, impacting in the biodiversity and socio-economic activities. Several studies performed in different parts around the world, like Italy (Salvati et al., 2016), Greece (Karamesouti et al., 2018), China (Huang and Siegert, 2006), Irã (Sarparast et al., 2018), Mexico (Becerril-Piña et al., 2015), Africa (Capozzi et al., 2018), and Brazilian Northeast (Vieira et al., 2015), show that the desertification process can be caused by anthropogenic exploitation on the natural resources. The desertification process produces effects on regional and global scales. For instance, in a regional scale, the soil degradation in a river basin will alter the essential components of the hydrological cycle, such as water infiltration and storage in the soil, runoff, and evapotranspiration. On the other hand, a non-local consequence of desertification is the production and transport of dust to other regions, inducing a large-scale effect (Johnson et al., 2011;Middleton, 2017Middleton, , 2018. Finally, the deteriorating livelihoods resulting from diminishing crop productivity, recurrent climatic extremes, and political instability may result in large-scale human migrations, with significant environmental, socio-economic, and political consequences (Myers, 1993;Bates, 2002;Lu et al., 2016). The discussions on desertification frequently highlight soil degradation as a marked characteristic of the process . Soil erosion and sedimentation have been proposed as valuable indicators of desertification (Wasson et al., 2002;Krause et al., 2003;Dawelbait and Morari, 2011). Soils support (directly or indirectly) most forms of life on Earth, and the loss of soil resources, which is accelerated by the development of agriculture and livestock grazing, has historically threatened food security and induced the collapse of some societies (Diamond, 2005). The generation of desertification indicators for a region demands the conduction of studies on soil erosion, especially on linear erosion in the form of rills and gullies. The existence of rills on a slope indicates a more advanced stage of erosion because the formation of channels implies that a large amount of soil has been removed (Bigarella, 2003;Descroix et al., 2008). Bigarella (2003) states that rills constitute ephemeral and discontinuous features. Nevertheless, the concentrated runoff normally deepens the rills, causing the collapse of the interior walls of the incision, and the headwaters advance upstream, removing many particles, and dragging them downstream. This fact means that the formation of rills is promptly followed by its evolution, and consequent increment in soil losses (Poesen et al., 2003;Chaplot et al., 2005). Gullies are relatively permanent features on the slopes (more extended and more profound than rills) and are associated with the processes of accelerated erosion, and consequently, with the instability of the landscape (Poesen et al., 2003). The gully formation is a consequence of both the natural landscape evolution and the impact of anthropic action. Indeed, gullies represent signs of instability on a slope, caused by alterations in the environment, and humans are just one of the agents of these changes. Nevertheless, the formation of gullies is accelerated when there are quick changes in the landscape caused by human action (Chaplot et al., 2005). Additionally, it is a broad and advanced stage of land degradation and slope dissection, represented by the downstream Rev Bras Cienc Solo 2020;44:e0190159 movement of sediments, which causes an imbalance in the quality of soils, local biota, and fluvial channels. On a river basin and/or sub-basin level, the description of the quantity, shape, and length of the existing rills and gullies require exhaustive measurements performed at the site. This is a complex task that involves uncertainties generated by the variability in the shape of the transverse sections of the rills and gullies (Casalí et al., 2015). The set of values of the cross-section shapes of the existing rills and gullies in an area will hardly follow a normal distribution, making the use of average and mean values inadequate to estimate the eroded soil volume. One alternative to attenuate these problems is the use of satellite images (Dawelbait and Morari, 2011) integrated with non-parametric statistical methods such as the Monte Carlo method (Gimenez et al., 2004;Sidorchuk, 2005;Huang and Siegert, 2006;Lal et al., 2015;Cuomo et al., 2016). Mathematical simulation, in combination with cartographic studies, is becoming one of the main approaches for the quantitative description of degradation processes in agro landscapes (Kulik et al., 2013). Based on the above, this work aimed to identify LEFs and estimating the volume of soil loss due to linear erosion occurring on the slopes near the middle course of the Salitre river. This estimative allowed us to recognize that there is a desertification process under development at this hydrographic sub-basin of the semiarid of Bahia, Brazil. Indeed, this work contributes to reinforcing the findings of Santos (2016), which has recognized this area under the action of an accelerated desertification process. Our results determine one of the most critical indicators of desertification, the linear soil erosion (Avni, 2005). Although several studies discuss the desertification associated with the erosive phenomenon (Avni, 2005;Salvati et al., 2015;Dai et al., 2017;Sarparast et al., 2018;Zweig et al., 2018), desertification processes occurring in the Brazilian Northeast are poorly studied. Therefore, the mapping of LEFs and the estimation of soil loss are crucial to be used as an indicator of desertification processes. In this context, from obtained data in field activities, we propose to use the Monte Carlo method to estimate the volume of soil loss from the formation of ravines and gullies. Monte Carlo calculations consist of an exciting tool for this issue because its randomness allows inserting uncertainty in the evaluated parameters. Therefore, the main novelty of the proposed methodology is that the estimation of soil loss in a vast region having soil with the same physical properties can be estimated from obtaining geometrical data of a fraction of the observed ravines and gullies. Studied area The experimental region is located in a rural area of Campo Formoso municipality, state of Bahia-Brazil, along the middle course of the Salitre River sub-basin (Figure 1), close to the bed of the main river (between the coordinates: 10° 10' 25" S, 40° 44' 55" W and 40° 45' 16" W, 10° 19' 28" S), corresponding to a narrow deep valley depression. The Salitre River emerges in the Northern portion of Chapada Diamantina and ends in the São Francisco River (Governo do Estado da Bahia, 2017). The central part of the sub-basin of the Salitre river is characterized by a planed surface, established on a carbonate plateau (Brito Neves et al., 2012), where the watercourses are well carved and with narrow floodplains (Silva, 2006). The geological materials are mainly carbonated rocks of the Caatinga Formation, with calcrete lithologies (Brito Neves et al., 2012;Borges et al., 2016) The studied area consists of a narrow valley bottom depression following the Salitre River ( Figure 1), from South to North, along the carbonate plateau. Within this depression, there are steeper slopes, above 10 %, and embedded drainage, with an altimetric variation of approximately 60 m, between the edge of the plateau and the Salitre river bed (USGS, 2014). The delimited space consists of a polygon, which was able to include the morphologically lowered area and with a significant presence of LEFs. Besides, according to Carlini (2013) and Santos (2016), this region presents a low density of vegetation cover and extensive patches with exposed soils. Additionally, the region has a semiarid climate with a mean annual temperature above 25 °C and rainfall ranging from 400 to 500 mm yr -1 , concentrated between November and March (Governo do Estado da Bahia, 2017). The potential evapotranspiration at Salitre Weather Station (9° 30' S, 40° 38' W) is around 1,900 mm yr -1 (Embrapa, 2019). The aridity indices (ratio between precipitation and potential evapotranspiration) of middle course locations of the Salitre are between 0.25 and 0.36 (Santos, 2016). The aridity index below 0.5 demonstrates reduction and high evapotranspiration, being widely used as one of the desertification risk indicators (Spinoni et al., 2015). Mapping of linear erosion features using high-resolution satellite images The area was visualized using the program Google Earth, which allowed obtaining previous plan-altimetric information, to determine the sites and collect topographic data at the field. Through a navigation GPS, a travel itinerary was developed such that we could transversely cross the Salitre River (East-West/West-East) by constructing transects in three different portions of the area (South, Central, and North), aiming at understanding the topographical context of the area ( Figure 2). The mapping of the LEFs and the delimitation of the degraded area was performed using high-resolution orthorectified images from the satellite WorldView-2. The images had a resolution of 0.5 cm and were dated May 15, 2015. We then proceeded with the digitization (linear vector) of the LEFs visualized in the image, using the program Quantum GIS 2.8. The adopted methodology is based on the visual interpretation of the primary elements (location, tone/color) and secondary elements (size, shape, and texture). Rills and gullies were identified from the continuous lines highlighted on the image, especially by the differences in coloration, tonality, texture, and shape. Small and discontinuous erosion channels along the slope were not considered due to the limit of the spatial resolution of the images. The data on location, quantity, and length of the erosion lines were directly extracted through the QGIS, and it served as the basis for the calculations soil losses caused by the linear erosion process on the slopes. The next step was the random selection of 11 LEFs (distributed in the area) for field measurement of width, depth, and shape variables. The selected LEFs are shown in figure 2, in which we also present the length and location of each LEF measured. Soil loss estimation Aiming to determine the eroded soil volume, the values of length of the rills and gullies identified in the area were obtained via satellite images. At the site, the following variables we have obtained some experimental data to characterize LEFs (based on cross-sections) for soil loss estimation: (i) width of extreme points of the LEFs (WE); (ii) width inside the LEFs (WI); (iii) depth of WE (DE); (iv) depth of WI (DI); and (v) shape ( Figure 3). The performed measurements were performed in intervals of 15 m along the lines of 11 LEFs, randomly sampled to ensure proper statistical distribution of the above-described quantities. Each site of measurement of the variables formed a sampling point, which was georeferenced. The width and depth were measured along a section transverse to the length of the incision in the soil. Visual interpretation and photographic records were used to determine whether the walls of the incisions were U-shaped (walls close to 90°) or V-shaped (inclined walls). Figure 3 illustrates the measurement of the variables in the sampling sites. The variables were measured from a total of 201 cross-sections (sampling). The obtained data were subjected to the Kolmogorov-Smirnov test to verify the hypothesis of normal distribution. After confirming the hypothesis of non-normality of the obtained values of WE, WI, DE, and DI, Monte Carlo simulations (Rubinstein and Kroese, 2007;Landau and Binder, 2014) were performed to simulate the occurrence of all the variables in the entire extension of the linear features existing in the area. From the 201 cross sections sampled, a total of 9,701 sections were simulated for the whole studied area. The Monte Carlo method was developed as follows: initially, the total length of the gullies/ rills and the distance between the cross-sections of data collection were inserted, and, after that, the averaged value of each frequency range for the variables WE, WI, DE, and DI were given as initial data. With this data, the program performs four random selections of a number between 0 and 1. The random number is divided into 13 ranges, following the percentages of occurrence obtained during the field activities. The first, second, third, and fourth selections determine the probability of the gully section presents WE, WI, DE, and DI respectively. A fifth number is randomly selected to determine whether the simulated cross-section corresponds to a U-or V-shaped rill/gully, using the frequency measured for these two shapes. To have more confidence in the simulated data, this procedure was repeated using 20 different seeds for generating the random numbers, until all cross-sections of the studied area have been simulated. After obtaining all the values of WE, WI, DE, and DI, the geometry of two consecutive cross-sections of the gullies were approximated using two overlapped trapezoidal solids (Figure 4). The volume of this geometrical structure is given by equation 1. Eq. 1 in which L is the distance between two measured cross-sections, A M and A m are, respectively, the areas of the trapezia at the extremities M and m of the trapezoidal solid. The sub-indices S and I indicate the upper and lower trapezia, respectively. The upper trapezium is considered as the one formed by a geometrical structure whose sides are given by WE, WI, and (DE-DI). The lower one will be the trapezium with sides WI, B, and DI, where B = min (WI-10 cm, 15 cm) depends on the shape of the erosion feature. In other words, if it is U-shaped, the side WI-10 cm is considered; if it is V-shaped, the shortest side of the trapezium will be equal to 15 cm. After calculating the geometrical parameters and volumes from the 20 different seeds, we have averaged the 20 obtained values aiming at having more confident values. The verification of Monte Carlo simulation was done using the standard Nash-Sutcliffe Efficiency (NSE) statistics (Nash and Sutcliffe, 1970), as recommended by Moriasi et al. (2007). Nash-Sutcliffe Efficiency indicates how well the plot of observed relative frequencies of "WE, WI, DE, and DI" versus simulated probability of occurrences of these variables fits the 1:1 line Eq. 2 in which Y i obs is the ith observed relative frequencies of "WE, WI, DE, and DI"; Y i sim is the ith probability of "WE, WI, DE, and DI" occurrences simulated by Monte Carlo; Y mean is the mean of observed relative frequencies, and n is the total ranges number used in the Monte Carlo simulations. The NSE can assume values between -∞ and 1, with NSE = 1 being the optimal value. Values between 0 and 1 are generally viewed as acceptable levels of simulation performance, whereas values <0 indicates that the mean observed value is a better predictor than the simulated value (Moriasi et al., 2007). RESULTS The most noticeable aspects of land degradation in the studied area are the marks left on the soil by the intense linear erosion, which moves large portions of sediments from the slopes to the fluvial channel, and damages the seasonal dynamics of the Salitre River. Figure 5 shows the low density of vegetable cover and the large patches of exposed soil in the analyzed area, which allow clear visualization of the linear erosion features (LEFs). The presence of various LEFs can be observed, and the depth of the channels is estimated. The map shown in figure 6 presents the result of the digitization of the LEFs. From a subtle analysis of the map, 734 LEFs in a complex system of ramifications were observed. The length of such linear erosion ranges from 14 m (for sub-lines on the high slopes) to 1,788 m (for main lines), resulting in a mean length of 199 m. It was also observed that only 3 % of the gullies/rills exceeded 800 m, 25 % were between 200 and 799 m, and 71 % were small, with the extension lower than 200 m. The sum of the LEFs in the mapped area totaled 145.53 km of visible gullies and rills, over an area of 20.78 km 2 (2,078 ha). Besides, 275 points of connection (confluence) between the LEFs and the fluvial environments in the area were accounted for. Thus, there are 275 systems of erosion channels (most of which are individual channels and do not have associated sub-channels) depositing sediments in the Salitre River and ephemeral tributaries. The vertical perspective (depth) and the width of the LEFs obtained from the satellite images encourage us to obtain numerical data on the width and depth of a percentage of the erosion features. Therefore, measurements were performed at the field to determine the width, shape, and depth for a total of 11 lines, accounting for a length of 4,117 m, which represented 1.50 % (amount of lines) and 2.83 % (length) of the total found in the interpretation of the satellite images. Although small in percentage, this sampling portion is a significant representation of the condition of the entire area when the pedological, climatic, geological, topographic, phytogeographic, and land use conditions are similar in the entire area. Thus, each of the WE (width of extreme points of the LEFs), WI (width inside the LEFs), DE (depth from WE), and DI (depth from WI) variables were grouped into 13 classes, according to the probabilities of occurrence. For all observed variables, the Monte Carlo simulation fit the measured data well, as judged visually in figure 7 and by the calculated NSE values, which are: NSE = 0.9 for WE; NSE = 0.7 for WI; NSE = 0.8 for DE, and NSE = 0.6 for DI. The predominance of low values for all variables yields an asymmetric distribution with non-normal behavior, as evidenced by the analysis of figure 7, which shows the data obtained through field study and simulated by Monte Carlo. Based on measured data the non-normality of the hypothesis is confirmed by the Kolmogorov-Smirnov test with 95 % confidence, and it is not possible to use the averaged values to estimate the volume of soil loss due to linear erosion. In figure 7a (width of extreme points of the LEFs -WE), the class with the highest observed relative frequencies and highest probability of occurrence simulated by Monte Carlo is the one with the mean value of 4.5 m. It was also noticed that in both -observed and simulated WE values -more than half of the values were within the range of 2.5 to 5.5 m. Therefore, these classes had the most influential WE range in the Monte Carlo simulations. In the case of WI, the class with the highest probability of occurrence is the one with an average value of 2.6 m. The figure 7b indicates that more than half of the analyzed LEFs have WI ranging from 0.4 to 2.6 m, and these were the most important values in the Monte Carlo simulation. The mean (m), standard deviation (± m), and coefficient of variation (%) of the measured and simulated WI data were respectively: 2.96 ± 1.92 m, 64.48 %; and 2.63 ± 1.33 m; 50.57 %. For the DE variable, the class with the highest probability of occurrence is the one with the mean value of 0.8 m. Figure 7c indicates that more than 60 % of the measured and simulated LEFs have WI ranging from 0.3 to 1.4 m. The measured DE data showed an overall mean of 1.36 ± 1.20 m, and coefficient of variation of 88.34 %, while the simulated data showed an overall mean of 1.97 ± 1.31 m and coefficient of variation of 66.82 % For the DI variable, the class with the highest probability of occurrence is the one with the mean value of 0.2 m. The measured data showed an overall mean of 1.07 ± 0.96 m and a coefficient of variation of 89.66 %, while the respective statistical indicators for the simulated data were: 0.8 ± 0.55 m; 70 %. According to the data in figure 7d, 70 % of the studied erosion incisions have DI ranging from 0.2 to 0.9 m. During the measurement of the cross-sections, the shapes of the gullies and rills were also observed, and there were 51.7 % of measured U-shaped gullies and hills with steep walls (which were more or less vertical) and a flat bottom. However, at other points, the same incision was V-shaped (48.3 %), open at the top and narrow at the bottom, indicating that the materials of the subsoil or deeper horizons are more resistant in comparison to the surface horizon. According to the Monte Carlo simulation results, there were 50.4 % of U-shaped gullies and hills. In summary, based on the small difference in percentage, it can be understood that almost half of the linear features along the slopes are V-shaped, while the other half are U-shaped. It should be considered that the observed incisions extend up to the slope, with tiny widths; therefore, they were not visualized in the images and were not mapped in the present study. Also, there is a significant number of rills spread over the area, connected to gullies and directly linked to fluvial courses or in small independent incisions along the slopes. Due to the limits of the spatial resolution of the satellite images, these rills were not considered in this study. From the Monte Carlo simulations were then performed based on the data presented in figure 7, and on the shapes of the LEFs. To calculate the soil volume displaced from the slopes, we have applied equation one from using the data obtained in the Monte Carlo simulations. Aiming to compare the results obtained from Monte Carlo simulations with the data collected in field activities, we have performed Monte Carlo simulations for determining the geometrical predictions for 201 cross-sections. From simulations, we obtained that approximately 11,500 m 3 had been deposited along the Salitre river bed due to the LEFs. The calculation of the soil volume displaced from the slopes to the river obtained from the field activity data is approximately 12,300 m 3 . Therefore, the agreement between the Monte Carlo simulations and field measurements was approximately 93 %. This finding motivated us to analyze and extrapolate the Monte Carlo simulation to study the volume of soil displaced from the slopes along the entire mapped area. The simulated results allow us to estimate that a total of approximately 515,000 m 3 of soil had been displaced, by linear erosion, from the slopes along the Southern portion of the middle Salitre River. Since the extension of the Salitre River in this section is 23,900 m, we can conclude that approximately 22 m 3 of soil was deposited in each linear meter along the river bed and its flood plain. One of the most immediate consequences of the intense erosion processes on the slopes is the accumulation of sediments along the drainage line of the Salitre River. The map in figure 6 also presents the spatial intermittency of surface water along the Salitre River during the period of the highest rainfall in the region (May 2015). The blue marks represent small stationary water bodies, which is not consistent with a fluvial behavior even for intermittent rivers of the semiarid region. Figure 8 shows a bridge over the studied River, where it can be observed that there is a considerable accumulation of sediment in its bed, evidenced by the filling of the space that existed between the base of the bridge and the river bed. In this process, it is important to consider that the natural conditions of prolonged drought periods and long intermittence of the Salitre River do not provide enough water volume in the riverbed to transport a part of the sediments that arrive through the fluvial channel, thus contributing to the accumulation (siltation). DISCUSSION Several studies have proved the efficiency of the high-resolution satellite image used in the detection of linear erosion features (LEFs). Forinstance, Shruthi et al. (2011) mapped gullies using Ikonos and GeoEye-1 images in Morocco, highlighting the importance of work as a basis for land use planning and data collection on sediment production. Desprats et al. (2013) used QuickBird images with 0.6 m of spatial resolution to map the erosive phenomenon in the North of Tunisia. Nevertheless, in these works, the efficiency of this technique is verified only in the measurement of length and direction of the LEFs, in such a way that the width and depth of LEFs are not well determined. To avoid such kind of indetermination, in this work, we have used images from WorldView-2, with a resolution of 0.5 m. Moreover, field activities allowed us to obtain pieces of information on the width and depth of the incisive erosions, and the statistical analysis showed the occurrence probability of the measured widths and depths along the mapped LEFs. Finally, Monte Carlo simulations, a method that has been already used to study erosive process (Nicholls and Stephenson, 1995), was able to simulate these variables from extrapolating the data obtained in field activities to all area, assuming that the soil properties have small variation along the studied area. All these data were joined to estimate, with reasonable confidence, the soil volume that was removed from the slopes, by linear erosion. It is important to highlight that the used methodological procedures can be replicated in other semiarid regions, with low vegetation cover and high soil exposure, aiming to specialize in the LEFs and estimate the soil removed from the slopes. The quantity and dimensions of LEF's in the study area showed that the area is under an intense morphodynamics instability, represented by the large amount of soil that is being removed from the slopes and deposited in the Salitre River bed (area with declivity greater than 8 %). Even though a discussion on the causes of the erosive process is out of the scope of this work, the analysis of the map presented in figure 6 evidences that most of the observed area under erosion is located around the village/community of Salgadinho, Taboa, Lagoa Branca de Cima, and Lagoa Branca. These areas present share the characteristic to have excessive soil use in agriculture activities, pointing to the anthropic contribution in the trigger and acceleration of the linear erosive phenomenon. Nevertheless, LEFs were not observed in the plane areas (with declivity lower than 3 %), even with an accentuated soil use. In this context, environmental conditions (relief and soil) has a crucial influence on the instability of the soil and in the trigger of the erosive process. The sediment carried to the Salitre River changes the form of the fluvial channel, increasing its width and raising the topographic level of the valley floor (local base level). Such topography changes in the bottom of the riverbed have a direct impact on the dynamics and functioning of the Salitre River, reducing the number of days of the fluvial flow of its waters. It should be noticed that there is a close relationship between the groundwaters and the availability of water on the riverbed (Karmann, 2000). Therefore, the massive accumulation of sediments causes the groundwater level to be at a greater depth concerning the bottom of the fluvial channel. This difference in the water level compromises the maintenance of water on the riverbed, even during rainy periods. In most dry sections on the riverbed, there is a significant accumulation of sediments, which is a result of the excessive erosion on the slopes. Since most of the 275 points of connection (confluence) between the LEFs and the bed of the Salitre River are found in the dry sections, it can be inferred that the rills and gullies contribute to the increment of sediments at these points. The remobilization of sediment is so intense that river channels cannot carry them, causing extensive filling in the saltpeter river. It is noteworthy that for more than a decade, the Salitre River has not presented water flow in its bed. Although the causes of this water intermittence along the Salitre River are related to a series of environmental factors, as interventions along the river, constructions, extinction of river sources, and reduction of the water table (Santos, 2016), intense silting must be pointed out as one of them. This silting deposition process results from the erosion processes on nearby slopes (potentially due to the association between human and natural causes) that coincide with the limits of the area presented in this study. The numbers, dimensions, and spatial density of the linear erosive systems (rills and gullies) identified along the middle course of the Salitre River call attention due to the possibility to form a desertification hotspot. Global warming associated with the extension of grazed and cropped areas should put more regions at high risk of gully erosion , which is an important indicator of desertification (Avni, 2005;Zweig et al., 2018). In discussions on the theme of desertification, soil erosion always appears as one of the most critical indicators (Symeonakis et al., 2014;Sarparast et al., 2018). Indeed, it is considered the most severe problem of land degradation and one of the main conditions of desertification in the Brazilian semiarid (Sampaio et al., 2005;Sá et al., 2010). The intense erosion of soils in semiarid environments can point to the existence of an advanced stage of desertification, in which erosion reduces the soil fertility and biological potential of the area and undermines the river channels and water quality. These changes in the natural environment potentiate economic losses with social impacts. In this context, this work contributes to deepening studies on the process of desertification in the Salitre basin, by providing information on the dimensions of the ongoing erosion process and the possible consequences on the dynamics of the Salitre river. Finally, it is fundamental to consider that the pedological and climatic conditions in the studied area are very different from those in the rest of Brazil. In the middle course of Salitre River, the soils are shallow and poorly developed, and the rainfall is around 400 mm; despite that, the area has a high density of active gullies, certainly above the patterns observed in the semiarid region of Brazilian Northeast. From using integrated geoenvironmental analysis and evaluation by biophysical and socioeconomic indicators, Santos (2016) stated that the Salitre middle course is an area affected by desertification hotspot. Kulik et al. (2013) point out that these hotspots are places with advanced levels of land degradation. At the same time, Singh and Ajai (2019) argue that these hotspots are extremely vulnerable to degradation/desertification, requiring urgent interventions aiming at their rehabilitation. Therefore, this research pointed out that the Salitre middle course presents landscapes in an advanced stage of soil degradation by linear erosion, reinforcing the significant risk of desertification, demanding an urgent construction of recovery plans for essential ecosystem services in the area. It is worth highlighting that the desertification is a multivariate phenomenon, and its study must include other indicators. However, the data obtained allow us to state that this area is a possible desertification hotspot, and it is essential to analyze other variables to confirm our statement, in addition to monitoring the increase (over the years) in the number and dimensions of the LEFs. CONCLUSIONS The use of high-resolution satellite images allowed us to point out the spatial distribution and significant density of Linear Erosive Features in an area in the Salitre Sub-basin, located in Bahia, Brazil. Based on the integration of remote sensing and Monte Carlo simulation, we have estimated the volume of soil removed from the slopes by the linear erosive process. The estimate of soil loss on the slopes serves to measure the impact of the linear erosive process acting on the slopes, as well as to detect the significant impacts on the Salitre river clogging and local water dynamics. The information provided confirms that the studied area is under advanced stage of land degradation, and constitutes an element capable of contributing, as an important indicator, in recognition of an ongoing desertification process.	2020-05-28T09:18:54.424Z	2020-05-26T00:00:00.000	{ "year": 2020, "sha1": "2b2dceffc91d2232155d2ff13357b26af9b5aa90", "oa_license": "CCBY", "oa_url": "https://www.rbcsjournal.org/wp-content/uploads/articles_xml/1806-9657-rbcs-44-e0190159/1806-9657-rbcs-44-e0190159.pdf", "oa_status": "GOLD", "pdf_src": "Anansi", "pdf_hash": "305e540b227a38390ea133f2cbdce156e9a0f453", "s2fieldsofstudy": [ "Agricultural And Food Sciences" ], "extfieldsofstudy": [ "Environmental Science" ] }
252896751	pes2o/s2orc	v3-fos-license	Bayesian hierarchical quantile regression with application to characterizing the immune architecture of lung cancer The successful development and implementation of precision immuno‐oncology therapies requires a deeper understanding of the immune architecture at a patient level. T‐cell receptor (TCR) repertoire sequencing is a relatively new technology that enables monitoring of T‐cells, a subset of immune cells that play a central role in modulating immune response. These immunologic relationships are complex and are governed by various distributional aspects of an individual patient's tumor profile. We propose Bayesian QUANTIle regression for hierarchical COvariates (QUANTICO) that allows simultaneous modeling of hierarchical relationships between multilevel covariates, conducts explicit variable selection, estimates quantile and patient‐specific coefficient effects, to induce individualized inference. We show QUANTICO outperforms existing approaches in multiple simulation scenarios. We demonstrate the utility of QUANTICO to investigate the effect of TCR variables on immune response in a cohort of lung cancer patients. At population level, our analyses reveal the mechanistic role of T‐cell proportion on the immune cell abundance, with tumor mutation burden as an important factor modulating this relationship. At a patient level, we find several outlier patients based on their quantile‐specific coefficient functions, who have higher mutational rates and different smoking history. INTRODUCTION Immunotherapy is a class of cancer treatments that fosters the patient's own immune system to fight cancer (Waldman et al., 2020). Although immunotherapies represent a major breakthrough in cancer treatment, offering immense clinical benefit to some patients with a lower toxicity burden than chemotherapy, resistance to immunotherapy remains a major challenge (Walsh & Soo, 2020). Tumors use various strategies to protect themselves from antitumor immunity which might vary across patients. Antitumor immune responses might also be mediated by several different mechanisms, driven by patient-specific immune architecture. Cancer immunotherapy therefore needs to be personalized to recognize patient-specific rate-limiting steps and employ strategies for overcoming these hurdles (Kakimi et al., 2017). Identifying patient-specific influences on immune response requires integrating measures of immune activity and characterizing their dependence on upstream factors. Here, we consider a setting where the response variable is a continuous measure of immune activation, such as the abundance of CD8+ T-cells, which play a key role in directly killing cancer cells. The abundance of these cells is driven by "Level 1" covariates 1 , … , , which represent measures of upstream immune activity. Here we take these Level 1 covariates to be features derived from T-cell receptor (TCR) sequencing, which capture activity of the adaptive immune system. The TCR repertoire depends, in turn on "Level 2" influences further upstream, such as antigens that the T-cells have been exposed to. This includes tumor-specific antigens, which are abnormal proteins produced by tumor cells due to DNA mutations in the cell (see Figure 1A). We consider mutational variables 1 , … , as (hierarchical) Level 2 covariates in our model. Our goal is to develop a hierarchical modeling approach providing insights into the mechanistic relationships among these variables wherein the measures of immune activity may depend on the upstream factors in a complex nonlinear fashion. We now briefly review prior work addressing the challenge of flexible regression modeling, which lays the foundation for the proposed approach. In order to estimate the subject-specific effect of variables on the outcome variable, Hastie and Tibshirani (1993) proposed the varying coefficient model (VCM). Since then, several variations of the VCM have been proposed (Fan & Zhang, 1999;Park et al., 2013), including approaches that incorporate shrinkage in estimation (Wang & Xia, 2009). However, existing VCM methods do not enable explicit multilevel variable selection, which is crucial in settings such as ours where there are a large number of explanatory variables within a hierarchy. Although VCMs allow more flexible coefficients than traditional regression models, they are still focused on estimation of the mean of the response variable. If one is interested in obtaining a comprehensive picture of the effect of the predictors on the response variable, mean regression might be insufficient. For example, if the dependent variable is multimodal or skewed, estimating the mean effect might be misleading. In such scenarios, median or more generally, quantile regression may be more appropriate (Koenkar & Bassett, 1978). Specifically, if the interest lies specifically at higher or lower quantiles of the response variable, quantile regression is more suitable. Over the last couple of decades, methodological developments in quantile regression have been proposed in both the classical and Bayesian frameworks (Das & Ghosal, 2018;Kottas & Gelfand, 2001;Reich, 2012;Yu & Moyeed, 2001). Several articles emerged in the literature integrating quantile regression with VCMs (Honda, 2004;Kim, 2007;Tang & Wang, 2005). In Bayesian settings, the literature on VCMs is relatively sparse (Biller & Fahrmeir, 2001), and these proposals do not address quantile regression. Recently, Ni et al. (2018) proposed a VCM incorporating variable selection in the Bayesian framework, which was applied to characterize the relationship of cancer patient outcomes to proteomics and genomics data. Although there are approaches for Bayesian variable selection in multilevel models that assume linear covariate effects (Koslovsky et al., 2020;Stingo et al., 2013), to the best of our knowledge, no prior work has considered variable selection in VCM for quantile regression. In order to understand the subject-specific effect of hierarchically structured covariates on the outcome variable we propose Bayesian QUANTIle regression for hierarchical COvariates (QUANTICO), where the regression coefficients are allowed to differ across patients for any given quantile level of the outcome variable. Among the hierarchically structured (multilevel) covariates, we consider the Level 1 covariates have a direct effect on the response variable, modulated by Level 2 covariates. Since we expect the covariate effects to be heterogeneous across patients, we want to perform individualized inference as well as allowing variable selection for both Level 1 and Level 2 covariates. As we also expect the effects to be different at different parts of the distribution of the response variable, we model across the quantiles, rather than only considering fixed moments (e.g., mean). This provides a richer, broader, and more flexible exploration of the relationship structure. We show an illustration of the proposed model in Figure 1B with subjects, Level 1 and Level 2 covariates. As shown in the figure, the selection of Level 1 and 2 covariates is allowed to vary across quantiles. Due F I G U R E 1 (A) Illustration of how mutation within tumor cells and T-cell receptor (TCR) repertoire impact immune cells. (B) Illustration of the proposed model for a scenario with patients, mutation variables 1 , … , , and TCR variables 1 , … , . The response variable is denoted 1 , … , . Different colored lines describe the estimated dependency structure at different quantile levels denoted by ( 1 , 2 , … ). (C) Directed acyclic graph (DAG) of the QUANTICO model. Parameters are shown in circles and the observed data are shown in boxes to the presence of two levels of covariates and due to the fact that the effect of Level 2 covariates is induced on the output variable via its effect on Level 1 covariates, we call it a hierarchical model. The rest of the paper is organized as follows: Section 2 describes the QUANTICO modeling framework, with a discussion of the priors in Section 3. We then describe the computational algorithm for performing posterior inference (Section 4), and provide a simulation study comparing the performance of the proposed method with existing alternatives (Section 5). In Section 6, we apply QUANTICO to characterize the relationship of the CD8 immune marker with TCR and mutation variables for lung cancer patients. We conclude with a brief discussion and possible extensions of our methodology in Section 7. QUANTICO MODEL In Section 2.1, we introduce quantile regression in a VCM setting with two levels of covariates. In Section 2.2, we describe variable selection procedures on Level 1 and Level 2 covariates. Section 2.3 summarizes the likelihood constructions. Varying sparsity quantile regression model Suppose there are subjects whose response variables (immune marker values) are denoted by = ( 1 , … , ) and Level 1 covariates (TCR variables) = ( 1 , … , ). For the th subject, the set of Level 1 covariates are denoted as = ( 1 , … , ) for = 1, … , . We assume a linear relationship between the response variable and the Level 1 covariates . Now, for the th Level 1 covariate , consider a set of Level 2 covariates (mutation variables) = ( 1 , … , ) for = 1, … , . A Level 2 covariate induces its effect on the response variable through its effect on the th Level 1 covariate, that is, for = 1, … , . Note that the same Level 2 covariate may influence multiple Level 1 covariates, and that the number of Level 2 covariates can be different for each Level 1 covariate. However, in the particular case study considered in this paper, the same set of Level 2 covariates (mutation variables) are considered for all Level 1 covariates (TCR variables). For the th subject, the set of Level 2 covariates is denoted by = ( 1 , … , ) for = 1, … , , = 1, … , . The total number of possible effects for the variables which can be identified as a part of the coefficient of the variables is given by ∑ =1 . Instead of estimating the effect of and on the mean of the response variable, our interest lies in estimating their relationship for different quantile levels. To obtain the quantile-specific estimates, we assume the effect of the covariates on to be dependent on the quantile level (0 < < 1). Let denote the th conditional quantile (0 < < 1) of the response variable at = , = . The relation between the response and the covariates at the th quantile is given by where 0 denotes the values of all distinct Level 2 covariates for the th subject, and denotes the values of the Level 2 covariates corresponding to the th Level 1 covariate for th subject. This equation shows the dependence structure of the th quantile of the dependent variable for the th patient on the corresponding Level 1 ( ) and Level 2 ( ) covariates. Note that for each Level 1 covariate we have many distinct variables, but since two different Level 1 covariates may share the same variables, the ∑ =1 many variables may not be distinct. Also note that both the intercept and slope terms are patient-specific (indexed by ). In order to incorporate the effect of the variables on the dependent variable, at any given quantile level , all the slope terms (of ) are semiparametric functions of the variables and likewise the intercept is a semiparametric function of the distinct variables. The explicit structure of ( , ) is discussed in Section 2.2. Varying-sparsity coefficient modeling and selection For any given quantile level , we consider ( , ) as a smooth function of . The main motivation for taking ( , ) as a smooth function of is so that the coefficients of any Level 1 covariate ( ) for two different subjects are similar if the values of the corresponding Level 2 covariates ( ⋅ ) are similar as well. Since for any Level 1 covariate, "neighboring" patients (with respect to the variables) are expected to have a similar slope coefficient, this assumption enables borrowing of strength, increasing our power to estimate the subject-specific slope coefficients. For modeling the slope and intercept terms, we use spline functions due to their flexible construction, interpretation, and the ease of incorporating penalization. At any given quantile (0 < < 1), the slope and the intercept terms are estimated as the sum of spline functions given by where 0 ( , 0 ) denotes the (global) intercept term, and ( , ) for ≥ 1 denotes the slope term. The spline components ( ) ( ) = where denotes the cubic B-spline bases for and denotes the corresponding spline coefficient. Note that we consider the intercept term to be a function of all Level 2 covariates. The number of knots for the B-spline bases is taken to be sufficiently large to capture the nonlinear features. We do not perform knot selection; rather equally spaced quantile knots on each of the Level 2 covariates are considered. Smoothing is induced via regularization and overfitting is controlled through a roughness penalty on the spline coefficients, details of which are provided in Section A of the Supporting Information. In Equation (3), ( , ⋅) is modeled as a sum of a set of smooth spline functions. As discussed in Section 3, we construct a prior on the spline coefficients, so that, for any given variable, the linear and nonlinear effects corresponding to all associated variables can be identified. Under these assumptions, the estimated coefficients of the variables would be zero if both the linear and nonlinear effects of all associated variables are zero. However, for a larger number of variables, it is unlikely that both the linear and nonlinear effects of all variables corresponding to any variable are zero. If the number of Level 1 covariates ( variables) is also large, it becomes crucial to perform variable selection on them to enforce sparsity and interpretability. In order to incorporate sparsity on the Level 1 covariates, one naïve approach could be to use discrete mixture priors such as spike-and-slab (George & McCulloch, 1993). But, since we consider the selection of the covariates to be patient-specific, to apply spike-and-slab prior, we would need to assign a latent indicator for each coefficient for every patient. This approach would therefore substantially increase the number of parameters in the model. In addition, as discussed below, the spike-and-slab prior is not well-suited for functional regression coefficients. Instead, we rely on a Bayesian hard-thresholding approach where we truncate the coefficients with smaller absolute values to zeros so that only the important Level 1 variables are selected. For truncation, we take the Bayesian hardthresholding function ℎ( , ) = (\| \| > ) to threshold the slope coefficients of the variables. We modify the values of the slope coefficients given in Equation (3) as where is a "minimum effect size" for the effect of to be considered as nonzero. Our use of the hard-thresholding prior instead of the more commonly adopted Bayesian spike-and-slab prior is due to the functional nature of the regression coefficients (⋅). The hard-thresholding prior allows selection at any given input whereas spikeand-slab is not able to handle an infinite-dimensional object. Another advantage of using the Bayesian hardthresholding is that this "minimum effect size" can be set at any reasonable value by the user based on intuition or prior experience, or estimated by assigning a prior (as we discuss in the next section). Note that no hard-thresholding is performed (or needed) on the intercept term as our main interest is to perform variable selection among Level 1 covariates only. PRIOR FORMULATIONS Note that the intercept and the slope terms ( ( ) 0 ( ) and ( ) ( ), respectively) are functions of the spline coefficients . In this section, we describe the prior on the spline coefficients , the prior used to induce selection of Level 2 covariates, and the prior assigned to the thresholding parameter to select Level 1 covariates. Prior on spline coefficients We propose the use of a penalized spline in order to have a flexible but smooth fit. Specifically, we choose a large number of knots placed at equally spaced quantiles of the covariates so that the local features can be captured, and penalize the roughness of (⋅) through an improper Gaussian random walk prior on given by ∼ ( , − ). The penalty matrix is constructed from the second-order differences of adjacent spline coefficients, which essentially penalizes the second derivatives of (⋅). Larger value of leads to a smoother fit, while smaller value of leads to an irregular fit (Ni et al., 2015). Note that Equation (3) is not identifiable since adding some quantity to any term of the summation and deducting the same quantity from any other term would yield the same summation value. In addition, is singular and therefore no penalty is imposed on the linear and constant trend of (⋅) (i.e., the null space of ). In order to alleviate these two issues, we consider a similar approach as in Scheipl et al. (2012) and transform the spline bases into orthonormal bases. Let = ( 1 , … , ). For a given quantile , consider the spectral decomposition of the covariance of where is the orthonormal matrix of the eigenvectors corresponding to the positive eigenvalues in the diagonal matrix . Let * = 1 2 . Now if we assume an independent proper prior * ∼ ( , 2 ), then the nonlinear part of (⋅) can be parameterized by * * which has a proper distribution proportional to the distribution of the original improper prior . Suppose we denote the effect size of the th covariate for the th subject before applying the hard-thresholding by * ( , ). Thus the full reparameterization of * ( , ) is now given by * ( , where is the global constant term absorbing all constant terms from splines. This parameterization adds the flexibility to separately shrink and estimate the linear, nonlinear, and constant effects of the Level 2 ( ) variables on the coefficients of the Level 1 ( ) variables. In order to make the proposed method computationally more efficient, we only consider the first several eigenvectors which explain at least 99.5% of the variation, a similar idea as in principal component analysis. In order to induce sparsity on the linear, nonlinear, and the constant effects, we consider a parameter-expanded normal mixture of inverse gamma (peNMIG) prior on each * , 0 , and separately. We opt for the peNMIG prior since it is known to provide a more efficient Markov chain Monte Carlo (MCMC) algorithm compared to traditional spike-and-slab priors given the multivariate nature of the spline coefficients (Scheipl et al., 2012). peNMIG multiplicatively expands * as * = , where is a scalar parameter and is a vector of the same length as * . Each of 0 and is also expanded in a similar fashion. A brief discussion on the choice of such a prior is provided in Section A of the Supporting Information. Priors for selection of Level 2 covariates Since the Level 2 coefficients are at population level, we can induce explicit selection using a spike-and-slab prior on , where ∼ ( ) and 0 is a fixed very small quantity close to zero. The selection of as a nonzero effect is indicated by the binary variable , and thus indicates the selection of * vector due to the multiplicative construction. In terms of interpretation, the binary variable = 1 indicates that has nonzero nonlinear effect on . Similarly, in the expansion of 0 and , = 1 indicates the presence of linear and constant effects of on , respectively. Thus, based on the estimated values of in the expansion of * and 0 , we can identify the presence of nonlinear and linear effects of Level 2 covariates, respectively. In essence, this construction allows the flexibility of considering only linear or only nonlinear or joint effects simultaneously. We choose conjugate hyperpriors for and , ∼ ( , ) and ∼ ( , ). As the number of Level 2 covariates increases, this Beta-Bernoulli prior automatically corrects for multiplicity by making the posterior distribution of concentrated at small values near 0 (Scott & Berger, 2010). We assign a mixture normal prior on each element of the vector = ( ( ) ), (1, 1) + 1 2 (−1, 1). The structure of this assumed prior discourages small effects. In a similar fashion, peNMIG priors are assumed for 0 and as well. Prior on hard-thresholding for selection of Level 1 covariates As mentioned before, to select the Level 1 covariates, we adopt a hard-thresholding approach where we truncate the nonzero coefficient of the th Level 1 covariate with absolute value less than to 0. Since sparsity is induced while estimating the linear, nonlinear, and constant effects of the Level 2 covariates on the Level 1 covariates, it is possible that (⋅) = 0 even before hard-thresholding (when * , 0 , and are zero for all = 1, … , ). In that case, can take any value without affecting the resulting estimates. To resolve this identifiability issue, we take = for = 1, … , . In the presence of at least one nonzero (⋅), is now well-defined. We put a gamma prior on , ∼ ( , ). The values of the shape and scale parameters ( , ) can be taken so that the mean of the gamma distribution (i.e., ) is equal to the desired cutoff (based on intuition or prior experience). A brief discussion on how to choose the parameters of the gamma prior is provided in Supporting Information (Section A). Note that a more conventional variable selection prior, the spike-and-slab, is often used for finite-dimensional parametric models. We choose to use the random hard-thresholding because it is more suitable for infinite-dimensional functional selection, as discussed in Section 2.2. A schematic illustration of the proposed model and its key parameters is given in Figure 1C. MCMC AND POSTERIOR INFERENCE The posterior distributions of the model parameters are not analytically tractable. Therefore, an MCMC sampling algorithm is required to generate samples from the posterior distribution. We use a Gibbs sampling scheme for updating parameters using their full conditional distributions ( , , ) and Metropolis sampling scheme for the parameters without closed-form conditional distributions ( , , ). Details on the full conditional distributions are provided in Section B of the Supporting Information. Inferential summaries When using this sampling algorithm, the quantile level of interest is fixed at the desired level. In order to estimate multiple quantiles simultaneously, the algorithm can be run in parallel for faster computation. The selection of Level 1 and Level 2 covariates can be performed in several ways based on the marginal posterior probability of inclusion. For the selection of Level 1 covariates, we consider the cutoff of the posterior probability of inclusion to be 0.5. A similar approach of thresholding the posterior inclusion probabilities can be taken for the selection of Level 2 covariates. In the presence of a higher number of Level 2 covariates, a false discovery rate (FDR) controlling approach can also be considered (Baladandayuthapani et al., 2010); see Section C of the Supporting Information. The proposed model allows the identification of both linear and nonlinear effect components of within the effect size of for all subjects. From the posterior samples, the posterior probabilities of both linear and nonlinear effect of Level 2 covariates within the effect size of Level 1 covariate can be calculated explicitly. Due to the induction of sparsity on the effect-size components of Level 2 covariates, the effect size of each Level 2 covariate can be categorized as one of four possible cases: linear, nonlinear, both, or none. Based on the posterior mean estimate of the effect sizes of the Level 1 covariates over a grid of quantiles, patientspecific posterior credible intervals over the quantiles can be obtained using QUANTICO. To calculate the posterior 95% credible interval of the quantile function for the th patient, we calculate the posterior mean of ( \| ⋅ , ⋅ ) over the quantile grid = 0.1 for = 1, … , 9. Then, for quantile level , we calculate the 95th percentile of the values \| ( ) ( \| ⋅ , ⋅ ) −ˆ( \| ⋅ , ⋅ )\| from the posterior sample, where ( ) ( \| ⋅ , ⋅ ) denotes the value of ( \| ⋅ , ⋅ ) obtained in the th posterior sample. Thus, we derive the width of the posterior 95% credible interval at all 's. Furthermore, by taking a dense grid of quantiles over [0,1], patient-specific uniform posterior credible intervals can be calculated. A detailed description on the calculation of posterior pointwise and uniform credible intervals is provided in Supporting Information Section D. SIMULATION STUDIES In this section, we evaluate the variable selection performance of QUANTICO across both Level 1 and Level 2 covariates and illustrate the subject-specific estimation at different quantile levels using simulation studies. The performance of QUANTICO is compared with varying coefficient quantile regression model (VCQRM) and variable selection in quantile regression using the lasso penalty (LASSO-QR) (Wu & Liu, 2009) as well as its Bayesian alternatives. All of these methods are variants of quantile-based VCMs and we provide explicit details of each method in the Section C of the Supporting Information. Variable selection performance To compare the variable selection performance across both Level 1 ( ) and Level 2 ( ) covariates, we calculate the true positive rate (TPR), false positive rate (FPR), and area under receiver operating characteristic (ROC) curve (AUC) separately for the and variables. A detailed description of the simulation design and computation details of the metrics is provided in Supporting Information Section C. Simulation structure To compare the performance of QUANTICO, VCQRM, and LASSO-QR, we consider two scenarios with sample sizes of = 100 and = 200. In order to understand how the selection of the variables is carried out in the presence of a higher number of covariates, we consider the cases = 5, 10 for sample size scenario = 100 and the cases = 5, 10, 20 for the sample size scenario = 200. For the cases where > 5, the true model remains as given by Equation (1) of the Supporting Information. So, the additional covariates considered in > 5 scenarios have no true effect on . The simulation is repeated 25 times, and each time new data are generated from the quantile function given in Equation (1) of the Supporting Information. The mean and the standard deviation of the TPR, FPR, and AUC for the and variables are computed separately at = 0.1, 0.5, 0.9 for all the methods. In addition, for a few high-dimensional scenarios the performance of QUANTICO is evaluated in Section F of the Supporting Information. We observe that QUANTICO maintains a high TPR and AUC for larger values of and as well, along with low FPR rates. We also evaluate the performance of QUANTICO compared to two other existing quantile regression methods, implemented in the R packages rqPen (Sherwood & Maidman, 2019) and Brq (Alhamzawi & Ali, 2020) which is also provided in Section F of the Supporting Information. It is observed that QUANTICO outperforms rqPen and Brq, in general. Comparative performance evaluation The comparative performance of the three methods is reported in Table 1. In general, QUANTICO and VCQRM perform better than LASSO-QR. QUANTICO results in better selection of the (Level 1) variables, and a large improvement in terms of FPR and AUC over VCQRM. In VCQRM, we do not incorporate any thresholding of the slope terms. Although it is possible to have a zero slope or intercept term without thresholding if all estimated linear, TA B L E 1 The result for the method(s) with the best performance is marked in bold, comparative performance study of QUANTICO, varying coefficient quantile regression model (VCQRM) and LASSO quantile regression (LASSO-QR) methods at quantile levels = 0.1, 0.5, 0.9 for different sample size and number of variable scenarios. lvl1TPR, lvl1FPR, and lvl1AUC denote the true positive rate, false positive rate, and area under receiver operating characteristic curve (AUC) of Level 1 covariates; and lvl2TPR, lvl2FPR, and lvl2AUC denote the same for Level 2 covariates ( , ) Measures = . = . nonlinear, and constant effects of all Level 2 covariates are zero, in the simulation results, the FPR of Level 1 covariates came out to be 1. We do not observe any strong pattern of differences in the comparative performance of QUANTICO and VCQRM in terms of TPR, FPR, and AUC for (Level 2) variables, since they follow same mechanism for selection of variables. In terms of performance, in Table 1, we report that QUANTICO yields very low TPR as well as FPR for Level 2 covariates for sample size scenario = 100 despite a high AUC, which implies that QUANTICO is very conservative in its estimation. In QUANTICO, the selection of Level 2 covariates is performed using an FDR-based approach with cutoff = 0.2. The performance of QUANTICO improves as increases (as noted in Table 1). QUANTICO We further report the selection and estimation performance of QUANTICO evaluated at quantile levels = 0.1, 0.2, … , 0.9. Figure 2A illustrates the selection of variables. QUANTICO selected the correct variables across all quantiles. Specifically, it selected variable 3 for the quantiles greater than 0.5 as it is in the true model. In Figure 2B,C, the corresponding true variables have the highest posterior probability of selection for 1 and 2 , respectively. In Figure 2D-F, the true and estimated quantile functions are plotted for three randomly selected subjects along with the corresponding estimated 95% credible bounds. We also compute the uniform posterior 95% credible intervals of the quantile functions. To improve coverage of the uniform credible intervals, we increase the width using an inflation factor. Although for point estimates, the observed coverage might come close to the percentage of the computed credible interval, for functions, in practice, it is a common phenomenon that the observed coverage may be less than the actual percentage of the credible interval. This undercoverage of the uniform posterior credible band of smooth functions is a well-known property and has been addressed in several articles (Cox, 1993;Das & Ghosal, 2017;Knapik et al., 2011;Szabo et al., 2015). In order to improve coverage, either undersmoothing or inflation of the obtained credible interval is required (Yoo & Ghosal, 2016). To improve coverage of the posterior uniform credible intervals, we increase the width using an inflation factor. As mentioned in Remark 5.4 and Theorem 5.3 in Yoo and Ghosal (2016), the inflation factor of the uniform credible interval should be taken such that it slowly increases to infinity as a function of sample size. Through experimentation, we observe that the inflation factor ( ) = 1.5 √ log( ) (which has a similar form to that considered in Das and Ghosal (2017) in the context of uniform credible bound over quantiles) works well in simulation under various settings for the sample size and number of Level 1 covariates. A detailed discussion regarding the computation of the uniform posterior 95% credible intervals of the quantile functions along with an extensive study on the computational time is provided in Supporting Information Section D. Scientific problem and data description Lung cancer is the second most common cancer and is one of the leading causes of cancer death. Though early stage patients can be treated with surgery, late stage lung cancer requires the use of systemic therapies (Ettinger et al., 2017). In recent years, immunotherapies have emerged as a successful treatment in a subset of late stage lung cancers, largely through their ability to boost the activity of T-cells, the subset of immune cells that target infected or malignant cells based on the detection of specific antigens. However, a large proportion of lung cancer patients still do not respond to immunotherapy (Doroshow et al., 2019). The nature of the antigens specifically recognized by Tcells has been the object of intense focus, and recent work has suggested that somatic mutations harbored by the tumor can be presented to T-cells as neoantigens (Lee et al., 2018). Analysis of specific somatic mutations and overall tumor mutational burden (TMB) has shown a positive association with response to immunotherapy in patients with NSCLC. This supports the role of these mutations in aiding T-cell responses by increasing tumor immunogenicity. Recent technologies have emerged to sequence the TCR to gain insight into T-cell responses, and studies have confirmed that TCR sequencing can be used to monitor immune response in various types of cancer (Page et al., 2016). In order to develop patient-specific immunotherapeutic treatment strategies for NSCLC, it is of critical importance to understand the interplay between somatic mutations, TCR variables, and the immune microenvironment, illustrated schematically in Figure 1A. We consider a cohort of 215 NSCLC patients recruited at UT MD Anderson Cancer Center. The tumors of these subjects were analyzed to obtain immune profiling, TCR sequencing, and mutational status of immune-related genes. We focus here on understanding the impact of mutation and TCR variables on the CD8 marker, which is the outcome variable in our analysis and is discussed in more detail below. In order to assess the effect of the explanatory variables on the outcome variable across different patients, we would like to estimate the patientspecific effect at different quantiles of the response. As our Level 1 covariates, we take standard summary measures of TCR sequencing data, including T-cell clonality, T-cell entropy, T-cell productive proportion of cells, and T-cell richness. T-cell clonality is a measure of heterogeneity among T-cells and has been linked to patient outcomes . T-cell entropy is Shannon's entropy, and highly diverse samples have comparatively higher entropy. The T-cell productive proportion of cells denotes the proportion of the tumor that consists of T-cells. TCR richness is another measure of TCR heterogeneity, defined as the number of different unique sequences in the sample. As our Level 2 covariates, we consider mutation counts for the top six most frequently mutated genes across the cohort, namely, CSMD3, MUC16, RYR2, TP53, TTN, and USH2A, along with total TMB, which is the total number of mutations observed per sample. As the mutation counts for individual genes have sparse values (i.e., zeros for a large proportion of patients), we only consider the linear and constant effects for those six variables. For TMB, linear, nonlinear, and constant effects are allowed. As our response variable, we focus on CD8 abundance as a key measure of immune activity. CD8 is a protein found on the surface of cytotoxic T-cells. CD8+ T-cells have the ability to mount a response against pathogens and defend against tumors by killing transformed tumor cells (Berg & Forman, 2006), and are therefore a vital part of cancer immunity. In precision oncology, understanding the patient-specific effect of T-cell architecture on CD8+ T-cell abundance is therefore crucial. Out of the cohort of 215 NSCLC patients, we focus on the 87 patients for which all three data types (immune, TCR, and mutational profiling) are available. Since the set of TCR variables considered have differing magnitudes, it is crucial to transform them to a similar range of values to make any comparison of their effect sizes meaningful. The same applies for the mutation variables. Therefore, before applying QUANTICO, we transform the TCR variables and the CD8 immune markers to unit intervals using log-normal cumulative distribution function (cdf) transformation (see Supporting Information Section G). The mutation variables are transformed into the unit interval using a linear transformation. Population level findings Using the proposed method, the coefficients of the TCR variables are estimated for each patient at nine quantile levels = 0.1, 0.2, … , 0.9. We use the same values of the hyperparameters as in the simulation study except the total number of iterations and burn-in, which are taken to be 50,000 and 10,000, respectively. The average posterior probability of selection of all TCR variables is plotted in Figure 3A. The T-cell productive proportion variable has an average posterior probability greater than 0.5 at all quantile levels except at = 0.9, and T-cell entropy has an average posterior probability marginally higher than 0.5 at = 0.2. In general, the average posterior probability of the TCR variables having a nonzero effect on the CD8+ immune cell abundance decreases at higher quantile levels. This implies that for patients with a lower abundance of CD8+ cells, the number of CD8+ cells has a stronger dependence on the TCR variable measures. However, in patients with a higher density of CD8+ cells, this dependence is less prominent. To summarize our Level 2 findings, we assessed the effect of the mutation variables on the coefficient of the Tcell productive proportion, which was identified to be the most important among the TCR variables ( Figure 3B). The nonzero effects of the mutation variables in the coefficient of the T-cell productive proportion are not strong for lower quantile levels, while TMB is shown to have a marginally higher posterior probability of having a nonzero effect at higher quantiles. Thus, at higher quantiles, a large proportion of the effect of the T-cell productive proportion on CD8+ immune cells is due to TMB. This is consistent with previous studies that have shown a positive correlation between TMB and CD8+ in melanoma , where immunotherapy using checkpoint inhibitors have shown to be successful. Hence, our findings suggest that TMB could be used for predicting the response to anti-PD-1/PD-L1 therapies in NSCLC. Patient-level findings Our results also provide insights into patient-specific immune profiles offering the potential to guide the development of precision immuno-oncology treatment strategies based on patient-specific information such as patient smoking history and mutational profiles. As an illustration of this, in Figure 3C, we show the estimated coefficients of the T-cell productive proportions for each subject across quantiles in the rows of a heatmap, along with patientlevel covariates. In this figure, we rely on hierarchical clustering over the estimated coefficients at quantile levels = 0.1, … , 0.8 to group patients with similar coefficients. Focusing on the last few rows of the heatmap, it is apparent that patients with overall lower values of the coefficient of the T-cell productive proportion have higher recurrence rates of NSCLC. Layering this analysis with clinical covariates reveals that the effect size of T-cell productive proportion is in general lower for nonsmokers compared to recent and former smokers ( Figure S3). Reuben et al. (2020) found higher T-cell clonality in current and former smokers compared to never smokers. In Figure 4A, we show a boxplot of the coefficients of the T-cell productive proportion for all patients, and we detect outlier patients (patient numbers 4, 45, 51, 56, and 76) at quantile level 0.1 and an overlapping set of outliers (patient numbers 40, 45, 56, and 76) at quantile level 0.3. Since this coefficient is modeled as a function of the mutation variables, we further compare the number of mutations for these six distinct patients ( Figures 4B-D). In general, these outlier patients have a higher number of mutations compared to the average value in the patient cohort; specifically for TMB, TP53, and MUC16 genes. CONCLUSIONS AND FUTURE WORK In this paper, we propose QUANTICO with multilevel covariates where, at any specific quantile level, selection over the direct (Level 1) covariates is performed for each subject. A novel feature of the proposed model is the development of a quantile-specific varying sparsity coefficient estimation approach which allows us to explicitly delineate how at different quantiles, the response variable depends on different covariates for each subject. The proposed method also enables selection of the indirect (Level 2) covariates, and can be used for obtaining patientspecific posterior credible bands over the quantile levels. The proposed method is used to analyze how the CD8 immune marker depends on TCR and mutation variables at different quantiles. We find that T-cell productive proportion is the most important TCR variable, and influences CD8 immune cells for most quantile levels. Out of all mutation variables considered, total TMB is found to be most important. Based on the structure of the relationship between the TCR and immune variable, we identify outlier patients, who turn out to have a higher number of mutations across several critical genes of known clinical relevance in cancer. This information is potentially useful to devise effective immunologic therapies for such patients(s) based on their unique immune architectures. There are several potential refinements that could be made to our modeling framework. We can extend our approach to simultaneous quantile regression, where instead of estimating the model parameters at specific quantiles, the entire quantile function and associated parameters are estimated simultaneously (Das & Ghosal, 2018;Yang & Tokdar, 2017). In terms of theoretical excursions, one could investigate for such hierarchical VCQRMs, building on some of the theoretical results proposed for Bayesian quantile regression by ourselves and others (Das & Ghosal, 2017;Yang & Tokdar, 2017) regarding posterior consistency, rates of convergence and posterior contraction. Given the nontrivial nature of these explorations, we leave them as future work. A C K N O W L E D G M E N T S VB was supported by NIH grants R01-CA160736, R01CA244845-01A1, R21-CA220299, and P30 CA46592, NSF Grant 1463233, and start-up funds from the U-M Rogel Cancer Center and School of Public Health. CBP was supported by NIH/NCI CCSG P30CA016672 and CPRIT Grant RP150521. KAD was partially supported by a CCSG NCI Grant P30 CA016672, NIH Grants UL1TR003167, 5R01GM122775, the prostate cancer SPORE P50 CA140388, CPRIT Grant RP160693, and the Moon Shots funding at MD Anderson Cancer Center. YN was partially supported by the NSF DMS-1918851 and NSF DMS-2112943. O P E N R E S E A R C H B A D G E S This article has earned an Open Materials badge for making publicly available the components of the research methodology needed to reproduce the reported procedure and analysis. All materials are available at http://re3data. org/. D ATA AVA I L A B I L I T Y S TAT E M E N T Code and data: Code for the simulations and real data analysis, along with a synthetic data set designed to closely resemble our example T-cell receptor (TCR) data, are available online at https://github.com/bayesrx/QUANTICO.	2022-10-15T06:17:45.008Z	2022-10-14T00:00:00.000	{ "year": 2022, "sha1": "602bfbba1962f10ef3aac7c2f802d63570ac6863", "oa_license": "CCBY", "oa_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/biom.13774", "oa_status": "HYBRID", "pdf_src": "Wiley", "pdf_hash": "d3d779cd02946b5c16277832853583db8680eb7d", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
56227465	pes2o/s2orc	v3-fos-license	Mechanism and Prevention of Titanium Particle-Induced Inflammation and Osteolysis The worldwide number of dental implants and orthopedic prostheses is steadily increasing. Orthopedic implant loosening, in the absence of infection, is mostly attributable to the generation of wear debris. Dental peri-implantitis is characterized by a multifactorial etiology and is the main cause of implant failure. It consists of a peri-implant inflammatory lesion that often results in loss of supporting bone. Disease management includes cleaning the surrounding flora by hand instruments, ultrasonic tips, lasers, or chemical agents. We recently published a paper indicating that US scaling of titanium (Ti) implants releases particles that provoke an inflammatory response and osteolysis. Here we show that a strong inflammatory response occurs; however, very few of the titanium particles are phagocytosed by the macrophages. We then measured a dramatic Ti particle-induced stimulation of IL1β, IL6, and TNFα secretion by these macrophages using multiplex immunoassay. The particle-induced expression profile, examined by FACS, also indicated an M1 macrophage polarization. To assess how the secreted cytokines contributed to the paracrine exacerbation of the inflammatory response and to osteoclastogenesis, we treated macrophage/preosteoclast cultures with neutralizing antibodies against IL1β, IL6, or TNFα. We found that anti-TNFα antibodies attenuated the overall expression of both the inflammatory cytokines and osteoclastogenesis. On the other hand, anti-IL1β antibodies affected osteoclastogenesis but not the paracrine expression of inflammatory cytokines, whereas anti-IL6 antibodies did the opposite. We then tested these neutralizing antibodies in vivo using our mouse calvarial model of Ti particle-induced osteolysis and microCT analysis. Here, all neutralizing antibodies, administered by intraperitoneal injection, completely abrogated the particle-induced osteolysis. This suggests that blockage of paracrine inflammatory stimulation and osteoclastogenesis are similarly effective in preventing bone resorption induced by Ti particles. Blocking both the inflammation and osteoclastogenesis by anti-TNFα antibodies, incorporated locally into a slow-release membrane, also significantly prevented osteolysis. The osteolytic inflammatory response, fueled by ultrasonic scaling of Ti implants, results from an inflammatory positive feedback loop and osteoclastogenic stimulation. Our findings suggest that blocking IL1β, IL6, and/or TNFα systemically or locally around titanium implants is a promising therapeutic approach for the clinical management of peri-implant bone loss. The worldwide number of dental implants and orthopedic prostheses is steadily increasing. Orthopedic implant loosening, in the absence of infection, is mostly attributable to the generation of wear debris. Dental peri-implantitis is characterized by a multifactorial etiology and is the main cause of implant failure. It consists of a peri-implant inflammatory lesion that often results in loss of supporting bone. Disease management includes cleaning the surrounding flora by hand instruments, ultrasonic tips, lasers, or chemical agents. We recently published a paper indicating that US scaling of titanium (Ti) implants releases particles that provoke an inflammatory response and osteolysis. Here we show that a strong inflammatory response occurs; however, very few of the titanium particles are phagocytosed by the macrophages. We then measured a dramatic Ti particle-induced stimulation of IL1β, IL6, and TNFα secretion by these macrophages using multiplex immunoassay. The particle-induced expression profile, examined by FACS, also indicated an M1 macrophage polarization. To assess how the secreted cytokines contributed to the paracrine exacerbation of the inflammatory response and to osteoclastogenesis, we treated macrophage/preosteoclast cultures with neutralizing antibodies against IL1β, IL6, or TNFα. We found that anti-TNFα antibodies attenuated the overall expression of both the inflammatory cytokines and osteoclastogenesis. On the other hand, anti-IL1β antibodies affected osteoclastogenesis but not the paracrine expression of inflammatory cytokines, whereas anti-IL6 antibodies did the opposite. We then tested these neutralizing antibodies in vivo using our mouse calvarial model of Ti particle-induced osteolysis and microCT analysis. Here, all neutralizing antibodies, administered by intraperitoneal injection, completely abrogated the particle-induced osteolysis. This suggests that blockage of paracrine inflammatory stimulation and osteoclastogenesis are similarly effective in preventing bone resorption induced by Ti particles. Blocking both the inflammation and osteoclastogenesis by anti-TNFα antibodies, incorporated locally into a slow-release membrane, also significantly prevented osteolysis. The osteolytic inflammatory response, fueled by ultrasonic scaling INTRODUCTION The prevalence of orthopedic and dental titanium (Ti) implants has increased steadily worldwide. Despite high success rates during the first 10 years (1, 2), orthopedic loosening and oral peri-implantitis remain a major problem for clinicians and constitute a major health problem for the profession. As mentioned, dental peri-implantitis is the main cause of implant failure. The risk of peri-prosthetic complications after 10 years ranges from 20 to 56% (3,4); however, currently there are no acceptable, standardized protocols for treatment and consequently, recurrence rates remain high (5). Increasing the life expectancy of Ti prostheses is thus a major challenge in orthopedics and oral rehabilitation. Oral peri-implantitis is believed to have a microbial etiology. However, a strong body of evidence links implant failure to aseptic inflammation around implants due to shedding of debris, Ti ions, and particles (1). Similarly, aseptic loosening of orthopedic implants has been attributed to Ti debris and particles (6). Although Ti prostheses are very biocompatible, these Ti byproducts are far from being bio-inert. Soft tissue biopsies around failing orthopedic (7,8) and oral (9-11) implants revealed severe inflammatory reactions around aggregates of Ti particles. We have previously shown that Ti particles released from ultrasonic (US) scaling around dental implants induce a marked inflammatory response in macrophages, with increased expression of pro-inflammatory cytokines, mainly IL1β, IL6, and TNFα. These Ti particles activate osteoclastogenesis in vitro and trigger inflammatory bone resorption in vivo (12). Our previous results led us to further investigate the mechanism by which Ti particles entrain bone resorption and to investigate the therapeutic potential of neutralizing antibodies against IL1β, IL6, or TNFα in preventing Ti particle-induced osteolysis. MATERIALS AND METHODS All procedures involving animals were carried out in accordance with the guidelines of Tel Aviv University and were approved by the Institutional Animal Care and Use Committee (permit number M-015-047). Cell Culture Primary bone marrow-derived macrophages (BMDMs) were isolated from the femora and tibiae of adult C57BL/6J mice (Envigo, Israel), as previously described (13). Briefly, cells were cultured overnight in 6-well dishes at 37 • C in a humidified atmosphere with 5% CO 2 in our "standard medium" consisting of alpha-modified Eagle's medium (αMEM, Life Science Technology, NY, USA) and 10% fetal bovine serum (FBS, Rhenium, Ltd, Modi'in, Israel). After 24 h, the nonadherent fraction was cultured in 10-cm non-culture-treated dishes containing standard medium and 100 ng/ml macrophage colony stimulating factor (M-CSF), prepared as previously described (14). The resulting adherent BMDMs were collected after 3 days for the specific assays described below. Particle Generation To obtain Ti particles that correspond to the particles shedding from oral implants during routine scaling, we subjected Ti discs that were made from Ti6Al4V (AlphaBio Tec., Petah-Tikva, Israel) to ultrasonic scaling (Newtron Led, Satelec, Acteon, Marignac, France), adjusted to a frequency of 32 kHz. Particles were obtained from discs with a machined (M), sand-blasted and acid-etched (SLA) or sand-blasted (SB) surface topography as described previously (12). When not specified, SLA-derived particles were used. All particles were generated in a sterile environment. Each disc was subjected to US scaling for 60 s in distilled water (ddH2O), then cleaned twice with ethanol, and finally resuspended in distilled water. We previously showed that each 6 mm diameter disc generates ∼2.54 million particles on average. In all our in vitro assays and for the preparation of the fibrinogen-thrombin membranes (see below) we used a particle density of 1,293 particles/mm 2 . Environmental Scanning Electron Microscopy (E-SEM) To examine the cellular response of macrophages to Ti particles, BMDM were seeded on glass slides in a 10-cm plate (10 6 cells per well) and cultured for 24 h in the presence of Ti particles released by the US scaling of SLA-treated discs. Cultures were then visualized by E-SEM. Each field was taken either in backscattered electrons mode (BSE) or secondary electrons (SE) mode to distinguish between extracellular vs. total Ti particles in the culture, respectively. Protein and Nitric Oxide (NO) Measurements in Conditioned Medium, RNA Isolation, and RT-qPCR Following a 24-48 h incubation with Ti particles (or LPS or vehicle only), the supernatant was collected and referred to as conditioned medium (CM). Secreted protein amounts of pro-and anti-inflammatory cytokines in CM were measured using a multiplex assay and expressed in MFI units (Multiplex Fluorescent Immunoassay, ProcartaPlex Multiplex Immunoassay, eBioscience, San Diego, CA, USA). NO content was measured using the Griess Reagent System kit (Promega, Madison, WI, USA). After collecting the supernatant, macrophages were washed with sterile PBS, and RNA was extracted using Tri-RNA Reagent (Favorgen Biotech Corp, Kaohsiung, Taiwan). The 260/280 absorbance ratio was measured to verify the RNA purity and concentration. cDNA was produced using a high-capacity cDNA reverse transcription kit (Invitrogen, Grand Island, NY, USA), and real-time PCR was performed using Kapa SYBR Fast qPCR (Kapa Biosystems, Wilmington, MA, USA) on a StepOne real-time PCR machine (Applied Biosystems, Grand Island, NY, USA). We examined the expression of IL1β, IL6, and TNFα, which are established markers of macrophage inflammation. The primer sets were as follows: F-GAA ATG CCA CCT TTT GAC AGTG and R-TGG ATG CTC TCA TCA GGA CAG for mouse IL1β; F-TAG TCC TTC CTA CCC CAA TTT CC and R-TTG GTC CTT AGC CAC TCC TTC for mouse IL6; and F-TCT TCT CAT TCC TGC TTG TGG and R-GGT CTG GGC CAT AGA ACT GA for mouse TNFα. The reaction was subjected to 40 cycles of amplification using the following program: 95 • C for 20 s, 60 • C for 20 s, and 72 • C for 25 s. The relative mRNA expression levels of the selected genes were normalized to the level of GPDH, which was amplified using the following primers: F-ACC CAG AAG ACT GTG GATG G and R-CAC ATT GGG GGTA GG AACAC. Animal Model and Micro-Computed Tomography (µCT) We used our calvarial model, which was described before (12). Briefly, US-released Ti particles (from M/SLA/SB-treated discs) were incorporated into a fibrinogen-thrombin degradable membrane used as a scaffold to localize the Ti particles. Membranes with no particles or only LPS were prepared as positive and negative controls. As indicated, neutralizing antibodies against TNFα (or saline as control) were incorporated into the membrane together with the Ti particles. After anesthesia, the skin of 10-week-old C57Bl/6J female mice was shaved and disinfected. The parietal bones of the mice were exposed via a 10-mm incision in the nape area, and the periosteum was removed using a periosteal elevator. The fibrinogen-thrombin membranes were inserted to cover both parietal bones. The surgical incision was then closed using nylon monofilament surgical sutures (5/0). In the sham controls, no membranes were inserted and the incisions were closed. Another control group consisted of inserting an empty fibrinogen-thrombin membrane (with no Ti particles). To test the therapeutic potential of systemic administration of neutralizing antibodies, we injected antibodies against IL1β, IL6, or TNFα (or saline as control) intraperitoneally 1 day before and once a week (2 mg/kg IL6 and 10 mg/kg TNFα) or daily (10 mg/kg IL1β) after membrane insertion, in accordance with the established clinical procedures (20)(21)(22)(23). All groups comprised a minimum of 6 animals. Animals were euthanized at the indicated post-operative time, and the skull of each mouse was removed, fixed for 24 h in 4% phosphatebuffered formalin, followed by 70% ethanol. All specimens were scanned and analyzed using a µCT system (µCT 50, Scanco Medical AG, Switzerland). Scans were performed at a 10-µm resolution in all three spatial dimensions, with 90 kV energy, 88 µA intensity, and 1,000 projections at a 1,000 m s integration time. The region of interest (ROI) was defined as two 3.7mm circles in the center of the parietal bones. A custom-made algorithm, based on Image-Processing Language (IPL, Scanco Medical), was developed to isolate the resorption pits, defined as unmineralized pits that were 10 to 40-µm deep on the bone surface (12). Morphometric parameters were determined at the 3D level and included the total volume of bone pits (Pit Resorption Volume, PRV, mm 3 ) and the bone tissue volume inside ROI (TV, mm 3 ), which was used to determine the PRV/TV (%). FACS Analyses Using our calvarial model, we also determined the effect of Ti particles on macrophage polarization, in vivo. The mice received either control membranes (fibrinogen-thrombin only) or membranes including LPS/Ti particles. Mice were euthanized 3 weeks post-op. This time point was chosen to be after the acute inflammatory response induced by the surgery but before the potential resolution of the inflammatory response to the Ti particles. The soft tissue covering the parietal bones was collected and processed using collagenase. Cells were labeled using specific markers (CD11b, Ly6-C, Ly6-G, F4/80, and MHC-II). This panel was used to define macrophages and to differentiate between M1 (CD11b+, Ly6C+, Ly6G+, F4/80+, MHC-II+) and M2 (CD11b+, Ly6C-, Ly6G+, F4/80+, and MHC-II+) polarization (24,25). Cells were fixed with 1% paraformaldehyde and analyzed on a Gallios flow cytometer (Beckman Coulter, Brea, CA, USA). Environmental Scanning Electron Microscopy (E-SEM) To determine whether particles are internalized by the cells, a culture of BMDM and Ti particles was examined by E-SEM (Figure 1). Each field was taken either in BSE or SE mode. BSE provides a topographic view of the upper-most layer, which means that the Ti particles seen in this mode were not internalized by the cells but rather, were laid on top of the cell membrane (Figures 1A,B). The SE mode is a deep penetrating mode, able to detect all metals in the sample, within or outside the cells (Figures 1C,D). Based on the similarity between the two scanning modes (the total number of Ti particles and particles in the upper-most layer), it can be concluded that BMDM failed to internalize most of the US-released Ti particles. Effect of Ti Particles on Macrophage Polarization The effect of Ti particles on macrophage polarization was examined in vivo, using a mouse calvarial model (12). Either Ti particles (from M/SLA/SB-treated discs) or LPS was incorporated into a fibrinogen-thrombin degradable membrane. Mice were euthanized 3 weeks post-op to examine the macrophages inside the soft tissue covering the parietal bones. Our FACS analysis first gated a CD11b+/F480+/MHC-II+ monocyte/macrophage population, which was further divided into Ly6G-/Ly6C-(Gr1-) non-inflammatory monocytes/macrophages (26), Ly6G+/Ly6C+ M1 and Ly6G+/Ly6C-M2 macrophages (24,25). Our results show no appreciable difference in the M2 population in either group (Figure 2). However, the M1 population was significantly higher in animals exposed to either machined or SLA surface-derived Ti particles. Surprisingly, milder differences were recorded for the SB particles. LPS alone also did not cause an appreciable increase in the M1 macrophage population (Figure 2). It is reasonable to assume that the fibrinogen membrane by itself caused some inflammatory changes that masked the differences between LPS and the empty membrane (control). Cytokine Secretion Profile of Macrophages Exposed to Ti Particles To further reveal the mechanism underlying the particleinduced inflammatory response, the proteomic profile of macrophages exposed to Ti particles was assessed using the multiplex cytokine kit, after 24 h exposure to either LPS or SLA-derived particles. For most cytokines, the presence of Ti particles induced a 2-to 10-fold elevation in the proinflammatory cytokine levels (Figure 3). A modest (∼2fold) elevation in the anti-inflammatory IL1α and IL10 levels was also found, as well as in IFNγ. With respect to IL1β, IL6, and TNFα, our multiplex analysis showed a marked increase in their levels in the supernatant of macrophages cultured with Ti particles, in line with our previously published RT-qPCR analysis (12). Overall, this cytokine profile is characteristic of an M1 polarization, supporting our findings in vivo (Figure 2). We also observed a significant increase in NO secretion, which is a hallmark of M1 macrophage polarization (27). Anti-inflammatory and Anti-osteoclastogenic Effects of Neutralizing Antibodies in BMDM and Preosteoclast Cultures The concomitant increase in all inflammatory cytokines (Figure 3) raised a question regarding a positive paracrine feedback loop that fuels the inflammatory response of the macrophages. To determine whether blocking the paracrine effect of the cytokines secreted in response to Ti particles attenuates the overall inflammatory response of the macrophages, BMDM were cultured in the presence of Ti particles and treated with neutralizing antibodies for IL1β, IL6, or TNFα. Cells were collected after 24 h and changes in the levels of pro-inflammatory cytokines were measured using RT-qPCR. The secretion of IL1β was not affected by either of the neutralizing antibodies in the culture (Figure 4). On the other hand, IL6 expression in macrophages reduced significantly in the presence of each of the neutralizing antibodies. In contrast, TNFα expression in the culture was only affected by anti-TNFα antibodies. These findings suggest that the overall inflammatory response to Ti particles not only results from direct contacts between macrophages and Ti particles, but also results from positive feedback induced by paracrine signals. Next, we tested the potential antagonizing effect of the same neutralizing antibodies on the osteoclastogenic signals emitted by macrophages in the presence of Ti particles. To this end, preosteoclasts cultured under osteoclastogenic conditions for 48 h were supplemented with the CM of BMDM cultured with or without Ti particles. Concomitantly, neutralizing antibodies against IL1β, IL6, or TNFα were added as indicated (Figure 5). Cultures were stopped on the 4th day, and the total area covered by multinucleated (>3 nuclei) TRAP-positive osteoclasts was calculated. Treatment with anti-IL6 antibodies did not affect osteoclastogenesis in our culture of isolated preosteoclasts. However, neutralizing antibodies against either IL1β or TNFα suppressed osteoclastogenesis in the presence of CM from BMDM exposed to Ti particles ( Figure 5). Dynamics of Ti Particle-Induced Osteolysis To account for the dynamics of the inflammatory response, we first established the time course of particle-induced osteolysis in our calvarial model. Ti particles released by the US scaling of SLA-treated discs were incorporated into the inserted membranes. A sham group consisted of mice undergoing the same surgical procedure without membrane insertion, which were euthanized after 4 weeks. Mice carrying a Ti particle-loaded membrane were euthanized every 2 weeks to monitor the extent of calvarial resorption using µCT. The results indicate that a dramatic resorption occurred during the first 5 weeks, along with a recovery from the 6th week onward. Of note, the membrane was totally degraded within 4 to 5 weeks; however, the presence of Ti particles was still detected even after 10 weeks (Figure 6C). This suggests that the osteolytic response is self-contained and somewhat recovers within 8 weeks after the last insult of Ti particles. Importantly, despite this recovery, the PR volume after 8 and 10 weeks FIGURE 3 \| Secreted protein and NO levels in the supernatant of macrophages exposed to LPS/Ti particles. BMDM were cultured in the presence of either LPS or SLA-derived Ti particles. After 24 h, the supernatant was collected and analyzed using multiplex cytokine assay for secreted protein levels, and Griess Reagent System kit for NO. Results are presented as mean Melt Flow Index (MFI) units ±SD, and µM (for NO). n = 4. Experiment was repeated once with similar results. p < 0.05 vs. control; non-parametric ANOVA. remained significantly higher in the Ti than in the sham group. Based on these results, further treatments were given for 4 weeks. After 4 weeks, the calvariae were collected and processed for µCT. Therapeutic Potential of Neutralizing Antibodies on Ti Particle-Induced Osteolysis We repeated the same calvaria model with membranes loaded with SLA-derived Ti-particles. Neutralizing antibodies for IL1β, IL6, or TNFα (or saline as control) were injected intraperitoneally 1 day before, once a week (IL6 and TNFα), or daily (IL1β) after the membrane insertion. As shown in Figure 7, each of the 3 neutralizing antibodies had a dramatic, statistically significant effect in preventing particle-induced osteolysis. Next, we tested the therapeutic potential of topically administered neutralizing antibodies. As a proof-of-concept, we elected to test anti-TNFα neutralizing antibodies incorporated into the membrane together with the Ti particles (Figure 8). Importantly, we found that topically-administered neutralizing antibodies significantly suppressed the particle-induced osteolysis (Figure 8). DISCUSSION In this study, we examined the mechanism by which Ti particles lead to inflammation and bone resorption and tested the therapeutic potential of using specific neutralizing antibodies in preventing particle-induced osteolysis. Our multiplex analysis revealed an interesting secretion profile of macrophages cultured in the presence of Ti particles. There was a dramatic elevation in the levels of most proinflammatory cytokines relative to the modest rise in the anti-inflammatory IL1α and IL10 levels, which may be a compensatory restraining response (28,29). IFNγ also displayed FIGURE 4 \| The effect of neutralizing Ab on Ti particle-induced inflammatory response: BMDM were cultured for 24 h in the presence of SLA-derived Ti particles released by US with neutralizing Ab for IL1β, IL6, or TNFα. IL1β, IL6, and TNFα expression levels were measured using RT-qPCR, normalized to GAPDH, and expressed as a fold change relative to the control (no particles). Data are shown as the mean ± SD. p < 0.05 vs. Ti, non-parametric ANOVA, n = 3. a significant rise; however, its contribution to the inflammatory response remains unclear, since both pro-and anti-inflammatory actions have been reported for this cytokine (30)(31)(32). MIP1α and MCP1 are chemokine proteins (CCL3 and CCL2, respectively). MIP1α is involved in the acute inflammatory state and is responsible for recruitment of polymorphonuclear cells, whereas MCP1 was found in the vicinity of bone resorption sites (33). The latter chemokine may drive osteoclast differentiation in the absence of RANKL (34). With respect to IL1β, IL6, and TNFα, our multiplex analysis revealed a trend very similar to our RT-qPCR analysis reported recently (12), with significantly elevated levels in the supernatant of macrophages cultured with Ti particles. In general, Ti particles induce in macrophages a response similar to that of LPS. The resulting inflammatory response drives the bone tissue damage mediated by osteoclasts. Together with the FACS, gene expression and secretome profiling on macrophages in vitro and in vivo, these changes indicate that macrophages undergo an "M1-like" polarization in response to Ti particles. It should be noted however that all our assays were conducted at the early stages of the inflammatory response. The resolving inflammation manifested in the partial tissue repair observed in vivo after 6-8 weeks (Figure 6), suggests a more dynamic and complex spatio-temporal distribution of M1 and M2 macrophages (27). Next, we examined how neutralizing antibodies against IL1β, IL6, and TNFα affected Ti particle-induced inflammatory response in macrophages. Importantly, we found that (i) the secretion of IL1β was not affected by either of the neutralizing antibodies, (ii) IL6 expression was significantly decreased by the presence of each of the neutralizing antibodies, and (iii) TNFα expression in culture was only affected by anti-TNFα antibodies. This is in accordance with the notion that both TNFα and IL1β are upstream factors in the inflammatory cascade (35). Indeed, our findings suggest that IL1β expression is independent of the FIGURE 9 \| Putative model of the hierarchical roles of IL1β, IL6, and TNFα in particle-induced osteoclastogenesis. This model describes the paracrine and autocrine induction of cytokine secretion by macrophages exposed to SLA-derived Ti particles. The thickness of the green arrows reflects the importance of one pathway over the others. secretion of IL1β, IL6, and TNFα by neighboring macrophages. In contrast, most IL6 expression by macrophages depends on paracrine signals, including IL1β, IL6, and TNFα. TNFα expression is only partly dependent on these signals, since only blockade of TNFα partly suppressed its own expression. Importantly, we also observed that in vivo, blocking any of these 3 cytokines significantly attenuated bone resorption. Based on these observations we propose a model to describe this additive and synergistic interrelationship between these 3 cytokines and the resulting stimulation of osteoclasts, thus inducing bone resorption (Figure 9). In this model, blocking IL1β does not affect TNFα expression and vice versa, and blocking IL6 does not affect either IL1β or TNFα expression. Moreover, both IL1β and TNFα directly stimulate osteoclasts while IL6 does so indirectly via the osteoblasts. The latter is suggested by the significant blockade of osteolysis by anti-IL6 antibodies administered in vivo, but not in vitro in the absence of stromal cells in the cultures. This conclusion is also in line with another study demonstrating that IL6 stimulates osteoclastogenesis via osteoblasts (36). Overall, our model suggests that Ti particles stimulate osteoclastogenesis via 3 pathways, the predominant one being the synergistically increased expression of IL6 by IL1β and TNFα, which in turn stimulates osteoblast-mediated osteoclastogenesis. It is reasonable to assume that both in vitro and in vivo, not all macrophages are in direct contact with Ti particles. The paracrine effect, depicted here, portrays a chain reaction that could provide the inflammatory signals with an extended range. We conducted here a time-lapse experiment and evaluated the long-term effects of Ti particles on bone resorption in vivo (Figure 6). Our data show that the osteolytic response is selfcontained and somewhat recovers, although not entirely. After 8 and 10 weeks, the extent of bone residual defects remained significantly higher than in the sham group. Moreover, in clinical settings, a Ti prosthesis is likely to continuously release ions, debris, and particles, thus fueling the inflammatory response, and further aggravating osteolysis. Previous articles studied different approaches to block the progression of bone resorption. These approaches included promoting apoptosis of osteoclasts, genetic intervention (37,38) or using bisphosphonates, which caused pathological fractures (39). In our previous publication, we characterized the size, number, physical and chemical properties of the titanium particles shedding from the surface of rough titanium implant commonly used in contemporary dentistry following ultrasonic scaling (12). Regarding the size, we used an automated cell counter to measure the number and size of the particles shedding from one 6-mm diameter titanium disc with an SLA surface. We found that ultrasonic scaling of each SLA disc generates 2.54 million titanium particles of an average size of 6 to 12 µm. A previous study established that the range of wear debris generated from orthopedic prostheses is between 1 and 30 µm (40). It is thus likely to assume that the cellular and inflammatory responses described in the current study with ultrasonic-generated particles are similar to those observed with wear debris surrounding orthopedic prostheses. Three main biological strategies were tested to block the chronic inflammatory reaction to orthopedic wear particles, including (i) interference with systemic macrophage trafficking to the local implant site, (ii) modulation of macrophages from an M1 to an M2 phenotype in periprosthetic tissues, and (iii) local inhibition of the transcription factor nuclear factor-kappa B, thereby interfering with the production of pro-inflammatory mediators (38). All three approaches showed promising results in preclinical studies but have not yet been evaluated clinically. Here we tested three clinically-approved neutralizing antibodies that are prescribed for the management of autoimmune and inflammatory diseases. Anti-IL1β antibodies (Anakinra) are prescribed to rheumatoid arthritis and neonatalonset multisystem inflammatory disease, anti-IL6 receptor antibodies (Tocilizumab) are administered to treat arthritis and anti-TNFα antibodies (Adalimumab) are effective against rheumatoid arthritis, psoriasis and Crohn's disease (41)(42)(43). Our in vivo and in vitro results showed a clear association between each of these cytokines and the macrophage response to Ti particles. Using our mouse calvarial model, we demonstrated that blocking each of these cytokines prevents Ti particleinduced osteolysis. We further showed that both systemic and local administration are conceptually possible approaches. This preclinical study therefore advocates that neutralizing antibodies be further tested against IL1β, IL6, or TNFα in clinical settings for managing the aseptic loosening of orthopedic prostheses and oral peri-implantitis. AUTHOR CONTRIBUTIONS ME, SH-B, TL, NS, YC, DK, and YG: contributed to the design of the research; ME, SH-B, and TL: performed research and contributed to the obtained results; YG and DK: supervised the project; ME and YG: wrote the paper and prepared it for publication. All authors read and approved the final manuscript. FUNDING This work was supported by Israel Science Foundation (ISF) Grants No. 1822/12, 1367/12, and 1086/17 to YG, and a Rothstein-Foundation grant to DK, NS, and YG.	2018-12-18T14:05:24.007Z	2018-12-18T00:00:00.000	{ "year": 2018, "sha1": "d519ec1a950f8f83bb21f33562fdea1fc6b8116f", "oa_license": "CCBY", "oa_url": "https://doi.org/10.3389/fimmu.2018.02963", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "d519ec1a950f8f83bb21f33562fdea1fc6b8116f", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
210887337	pes2o/s2orc	v3-fos-license	Quantification of Discontinuities in Welded Joints Using Gamma Tomography Gamma tomography was used in this work to evaluate the recurrent defects in welding processes in naval steel sheets. It was used a first-generation equipment consisting of a source of Cesium-137, with activity of the order of 200 mCi, coupled to a Thallium-doped Sodium Iodide detector, NaI (Tl). For the study, specimens were produced in ASTM A131-AH36 steel sheets with 13.7 mm thickness; all welded by Metal Active Gas process. One reference sample was fabricated, with no macroscopically measurable defects and another welded under wind conditions producing a weld bead with different types of discontinuities. The microstructural characterization of welded joints made possible a qualitative evaluation between defective joints, in relation to the joint without defects. With the data of the tomography, 3D graphics were drawn that enabled the statistical survey and analysis of clusters of the results that allowed the localization and the dimensioning of the discontinuities that appeared as counting peaks in these graphics. It was revealed that the defective welding showed porosity of up to 39.28% of its volume, and the acceptable size crack must be less than 1 mm based on ASME-B31.3, showing the viability of the tomography for this type of nondestructive analysis. Introduction The joining of metallic materials has been usually accomplished by an electric arc welding process which uses heat source to melt the base metals and addition, forming a melt pool which must be protected by an inert or active gas 1 . In order to assure the quality of the welded structures it is necessary to investigate the integrity of the joint that can be made several methods of Non-Destructive Testing (NDT). The computerized tomography (CT) gamma radiation is one of these methods that lends itself to analyze the attenuation suffered by a beam of radiation in relation to the volume of the object under analysis, allowing the acquisition of data according to the spatial density distribution 2 . Recently, Oliveira et al. 3 used gamma tomography to analyze weld defects in tubes of API 5L X80 steel. In this study, computed tomography was used to calculate the porosity in metals, with similar methodology to that used by Pires et al. 4 to analyze porosity in rocks. After the adequate adaptations for irradiation in flat plates the discontinuities of a welded joint of naval plates could be investigated. Microstructural characterization (destructive test) was associated for ratification of gamma-tomography (non-destructive test) results. Due to the nature of the gamma transmission when crossing a material, Figure 1, the intensity of the radiation undergoes an exponential fall described by the Beer-Lambert relation 5 , Equation 1: In equation (1), I is the final intensity of the radiation which arrived in detector, I 0 the initial intensity of the radiation, μ the linear attenuation coefficient and x is the thickness of the material. The linear attenuation coefficient (cm -1 ) is specific to each material. Often, the term μ is replaced by μ m ρ, where μ m is the mass attenuation coefficient (cm 2 /g) and ρ is the density (g/cm 3 ). The mass attenuation coefficient is independent of physical state 6 . Materials and Methods The study material was ASTM A-131 AH 36 Steel plates with 13.7 mm thick. From this material, test sheets measuring 100 x 300 mm were machined, two by two, with a bevel angle of 25 degrees in each of them. Then, the plates were welded by the MAG process, using AWS ER70S-6 wire with φ = 1,2 mm and the Ar-25% CO 2 protection gas as consumables. Table 1 shows the parameters used for welding. The welding process was semi-automatically performed with constant speed in all passes, where a motorized system was used to move the torch with specified speed. The plates were made differently, so that a comparative study on the quality of the weld could be made. A test plate was welded while respecting all the cited parameters, so as not to produce any discontinuity macroscopically measurable. In the other test plates, the discontinuities were purposely induced by failures in the shielding gas caused by the use of winds blown transversely to the weld bead. Microstructure characterization For microstructural characterization, the test plates were cut into small specimens to be subjected to traditional metallography: cutting, sanding, polishing and etching. A microstructural evaluation of these specimens was made to allow a comparison between the reference plate and the defective plate, considering the discontinuities found in it. Gamma tomography essays Gamma tomography (source of Cesium-137) was performed with the plate fixed in the holder and submitted to an activity of the order of 200 mCi. Gamma rays emitted energy of 0.662 MeV which, after being collimated (5.5 mm aperture) and attenuated by the material, reached a Thalliumdoped Sodium Iodide NaI (TI) detector. The source / detector distance was 33.9 cm. Scanning of the welded region was done in 99 steps of 1 mm each, with 10 s counting for each step. After each horizontal scan, the table height was raised 3 mm for further scanning. Current (A) 140 Gas flow rate (l/min) 18 Vel. Wire (m/min) 6 Vel. Torch (mm/min) 6 Stick out (mm) 8 For the calculation of radiometric density ρ m , Equation 2 7 based on the Beer-Lambert relationship was used 8 . This equation (2) is a function of the values of the mass attenuation coefficient μ m of the base metal and weld metal, the diameter or thickness of the plate D and the logarithmic ratio between the empty tube intensities I v and of the whole with flow I f , respectively. This equation can be used by replacing I v by I 0 and I f by I referring to Equation 1. (2) Using equation (2), the mass linear attenuation coefficient for the base metal (μ m = 0.0742) and for the weld metal (μ m = 0.0733) were obtained. These results are in accordance with the NIST XCOM 9 . The gravimetric density was calculated by Equation 3: In equation (3), ρ is the gravimetric density, m the mass and V the volume. Samples with parallelepiped-shaped were produced by machining with edges aligned by a milling machine to ensure the scanning the edges of the volume. For sample volume measurements, a micrometer with a sensitivity of 0.001 mm was used. The volumes were then weighed using an analytical precision balance, with a capacity of 220 g and a precision of 0.1 mg. Using these values, the gravimetric density ρ of each sample was calculated. With the CT scanner was made reading the I o , which is the radiation level measured without the specimen. For this, ten measurements with a duration of 10 seconds each were performed and then an average of the values taken. Based on the results of density values and attenuation coefficients, the irradiated volume were calculed. The volume of the radiation beam that passes through the material and reaches the detector is cylindrical in shape with a diameter of 5.5 mm. In this way, the volume V of the material traversed by the beam of radius r in a thickness l was calculated by S X A defect (discontinuity) represents an empty space resulting from the lack of material that is not filled by the weld. Considering the smaller attenuation of the beam, it can be considered that the irradiated effective volume would be equivalently smaller than the physical volume of the sample without defects as shown in Figure 2. This method closely approximates the work described by Schon 10 to calculate the percentage of porosity in rocks. The percentage of porosity φ, given as a function of the relationship between pore volume v p and total volume v t , can be calculated by equation (6). (6) However, by the method using gamma transmission, it is possible to estimate the porosity by relating the density value of the sample 4 , according to Equation (7): (7) In equation (7), ρ m is the radiometric density of the gamma transmission of the material, obtained by Equation (2), ρ the density of the material used in the experiment provided by the literature, such as NIST XCOM 7 . The gravimetric density ρ can also be calculated by Equation (3). Reference sheet The high quality of the reference weld can be seen from the microscopic images shown in figure 3. No pores, cracks or any discontinuities were found in its structure. It is demonstrated that this weld served perfectly as reference in this work. In Figure 3A it is observed a pass of the weld bead indicated by the red arrow and no existing defect. Figure 3B shows the root pass of the weld, where the effect of epitaxy on the columnar growth of the grains in the molten region is seen, without any apparent discontinuity. Figure 3C shows the three zones of the weld joint: base metal (BM), heat-affected zone (HAZ) and weld metal (WM). Due to the microstructure of the three adjacent regions, the weld quality is evidenced. As expected, the microscopic images of welded joints with failures in the shielding gas showed large pores, cracks and lack of fusion. The absence of the shielding gas besides destabilizing the electric arc, produces a great oxidation of the joint, contributing to the formation of cracks and large pores. Studying the effects of wind on welding with tubular wire in shipyards located in the brazilian northeast, Lessa et al. 11 obtained equivalent results.. Figure 4 shows cracks with diameters ranging from 22.01 μm to 34.46 μm in thickness and pores with diameters ranging from 1608.54 μm to 2114.50 μm. In Figure 5 it is possible to observe pores with 25,94 μm and up to 561,83 μm of diameter. A long crack with an average thickness of 4.92 μm was also found. Based on the data obtained with the gamma tomography of the plates, 3D graphics were created, as shown in figures 6 and 7. These graphs show the existence or not of defects in a welded joint. These defects are characterized by high counting peaks. This type of signal (high peaks) indicates a lot of porosity and other types of discontinuities that caused less attenuation in the gamma beam due to lack of material in the weld. A greater number of defects implies a lower attenuation of the gamma beam, increasing the gamma particle count that reaches the detector. The 3D graph, obtained from the computed tomography of the reference weld, revealed that there are no measurable discontinuities capable of altering the attenuation of the gamma beam, both on the weld bead and on the base metal ( Figure 6). All the irradiated surface, base metal and weld metal, presented peaks of equivalent height, characterizing the good quality of the welding, as already evidenced by metallography, where there are no detectable defects in this sample, ensuring compliance with ASME-B31.3 standard for tomography. The 3D graph obtained from the computed tomography in the weld with fault of the protection gas (Figure 7) revealed peaks that reached 8738 counts in a region on the weld bead, the regions between 0 to 2 and between 6 to 8 corresponding the HAZ. These high points characteristic of the presence of discontinuities (absence of material) revealed an effective reduction in weld bead thickness. Based on the attenuation of the gamma beam, the effective thickness of the defective weld bead was measured by three scans straight S1, S2 and S3, as shown schematically in Figure 8. The average thickness measured along the S1 scan was 1.066 cm ± 0.049 cm, in the S2 scan was 1.018 cm ± 0.063 cm and in the S3 scan was 1.301 cm ± 0.037 cm. These results showed an average error of 17.58%, in relation to the nominal thickness of the sheet, which was 1.37 cm (Figure 8). Thickness below 1.29 cm is an indication of the presence of defects. On defective weld bead, the effective thickness reached 0.817 cm. Based on this result, we can admit the existence of a large number of discontinuities (cracks, bubbles and pores) throughout this weld bead, even without the visualization of the internal part that was only ratified after the metallographic tests. This variation of thickness associated with the presence of discontinuities (defects) can also be observed on surfaces with sharp curvature, as reported by Moura et al. 5 when applying gamma radiation in small tubes The results of the porosity φ were obtained by calculation according to Pires 4 through Equation 7. For the base metal, the result of the porosity φ was 0.053% considering the reference density and, considering the gravimetric density (Equation 3), the porosity φ was 0.068%. For the region of the reference solder, the porosity φ calculated by the method used by Pires 4 was 0.177%, considering the reference density and 0.192%, considering the gravimetric density. In the test plate with defective bead, where regions with material thickness of 0.817 cm were found, the porosity φ measured by the method of Pires 4 using the reference density was 39.28%. In this same sample, considering the gravimetric density, the porosity φ calculated with Equation 3 was 37.21%. Conclusion The results showed the potential of computerized tomography for inspection of defects in welded joints. From the results obtained, it was possible to observe a satisfactory response of the tomographic signal, comparing a reference solder with a solder with discontinuities. The presence of discontinuity in a welded joint decreases its effective volume, reflected in a lower attenuation of the gamma beam, when crossing the sample. In this way, the detector is hit by a larger volume of gamma particles, generating high peaks in the 3D diagrams. In the tomograph, the volume reduction is measured from the effective thickness reduction. In the reference weld the average value of the effective thickness of weld beam was 1.128 cm. In the defective weld, the effective thickness of weld beam reached 0.817 cm. Considering that the thickness of the sheet was 1.37 cm, the reductions of the effective thickness in the weld without and with defect were of 17.66% and 40.36%, respectively. The porosity measured on the basis of the reference and gravimetric densities were respectively: 0.051% and 0.066% for the base metal; 0.177% and 0.192% for the weld bead without defect and 39.28% and 37.21% for the defective weld bead.	2020-01-02T21:18:34.998Z	2019-01-01T00:00:00.000	{ "year": 2019, "sha1": "7e3dbb0a17e2e667fc4bb4f22ce3e4cb382180a4", "oa_license": "CCBY", "oa_url": "http://www.scielo.br/pdf/mr/v22n6/1980-5373-mr-22-06-e20190278.pdf", "oa_status": "GOLD", "pdf_src": "Anansi", "pdf_hash": "e0f5ebfa1787175a9e83c902c6fd742c4e6c0bb7", "s2fieldsofstudy": [ "Engineering", "Materials Science" ], "extfieldsofstudy": [ "Materials Science" ] }
29260734	pes2o/s2orc	v3-fos-license	Urinary Biomarkers for Kidney Disease in ATTR Amyloidosis 1Department of Nephrology, St. Anthony Hospital, Central Hospital of Porto, Porto, Portugal 2Department of Clinical Chemistry, St. Anthony Hospital, Central Hospital of Porto, Porto, Portugal 3Unit Clinic Paramiloidose, St. Anthony Hospital, Central Hospital of Porto, Porto, Portugal 4UMIB, Institute of Biomedical Sciences Abel Salazar ICBAS, University of Porto, Porto, Portugal 5Department of Pathology, St. Anthony Hospital, Central Hospital of Porto, Porto, Portugal Introduction The Amyloidoses Associated with Transthyretin (ATTR) are autosomal-dominant diseases related to at least 100 different Transthyretin (TTR) mutations. The single amino-acid substitution of methionine for valine at position 30 is the most common [1]. Although this disorder was initially thought to follow a benign evolution concerning the kidney, it was later recognized that progression to End-Stage Renal Disease (ESRD) occurs in up to 10 percent of patients as natural course of the disease [2]. The detection and prognosis of ATTR nephropathy depend on the presence of albuminuria and an elevated serum creatinine concentration. These are correlated with the amount of amyloid in the glomeruli, arterioles, and medium vessels. When amyloid is confined to the tubulointerstitium or vasculature, proteinuria is minimal and reduced Glomerular Filtration Rate (GFR) is the principal clinical manifestation. In some patients, proximal tubular epithelial cells contained reabsorption-like droplets TTR positive and Congo-red stain negative, but clinical expression of tubular dysfunction has not been described until now [2]. Conventional measurements of renal function, such as creatinine and BUN levels, are limited by several non-renal factors, including body weight and nutritional status, which are particularly relevant in this population. Of special note, increased concentrations of albuminuria among patients with a GFR >60 ml/min (an area of weakness for serum creatinine and GFR), may define a higher risk patients to develop clinical nephropathy. Specific urinary biomarkers for tubular and interstitial pathologic abnormalities are needed for early detection and timely treatment. Ideally, there should be early markers of nephropathy in initial stages of ATTR, even before neurological manifestations. Although Orthotopic Liver Transplantation (OLT) is performed as a potential curative treatment, new strategies have been developed to treat Familial Amyloidotic Polyneuropathy (FAP) [3]. Tafamidis was approved for the treatment of ATTR in adult patients with stage 1 symptomatic polyneuropathy to delay peripheral neurologic impairment [4]. Several trials, some already completed and others recruiting participants, are evaluating new drugs [5]. Until now, trials did not clarify whether kidney disease is a criterion for excluding or adopting the use of a given drug. Most trials accept patients with evidence of neuropathy, some with cardiomyopathy but none of them have admitted patients with nephropathy as an isolated feature. It is questionable whether a patient with proteinuria, renal amyloid deposition identified as TTR and without other manifestations of disease would be a candidate for any future therapy. In the past decades, several urinary proteins have been identified as early prognostic markers in different kidney diseases [6]. Beta-2-Microglobulin (B2M) and Alpha-1-Microglobulin (A1M) are both low molecular weight proteins that are freely filtered by glomerulus, efficiently reabsorbed and catabolized by proximal tubule. No active tubular secretion or significant extra renal elimination occurs. Therefore, in the presence of renal dysfunction, B2M and A1M serum levels are increased when compared to those patients with normal renal function [7]. Alpha-2 Macroglobulin (A2M) is a tetrameric glycoprotein, produced in human plasma that has a molecular mass of 725 kDa [8]. Urinary levels of A2M increase when glomerular leakage occurs and, consequently, can act as a marker of such impairment. The aim of this study was to evaluate urinary A1M, B2M and A2M as early markers of ATTR nephropathy and predictors of outcome of renal disease. Patients and Methods We evaluated a cohort of thirty patients and eleven asymptomatic gene carriers with the TTR V30M mutation. Information about demographics and clinical characteristics was collected. Age-at-onset of FAP was defined as the age of initial neurologic features. Amyloid deposition was confirmed histologically in 24 patients; sections from formalin-fixed, paraffin-embedded biopsy specimens were stained with Congo red and viewed under cross-polarized light. Patients who attended our clinic were followed by the same nephrological team. Diabetic patients and patients submitted to OLT, therapy with tafamidis or under any therapeutic clinical trial were excluded. Markers Measurements On the day of the urine sample collection, blood samples were taken. Serum and urinary creatinine levels (mg/dL) were measured by a rate-blanked compensated Jaffé method on a Modular P analyzer (Roche Diagnostics, Mannheim, Germany). Urinary albumin levels (mg) were measured with an automated immunoturbidimetric assay using the Cobas Integra 800 analyser (Roche Diagnostics, Mannheim, Germany). The amount of albuminuria was estimated by the albumin to creatinine ratio (mg/g). Albuminuria <30 mg/g was considered normal. The inflammatory state was evaluated by determination of C-reactive protein (CRP) (reference value <5 mg/L) and ferritin values (reference range 12.5-454 ng/mL). Pro-B-type natriuretic peptide (pro-BNP) concentration was measured in all subjects (reference value < 227 pg/mL). Statistical Analysis Correlations were assessed by Spearman correlation coefficient test. The level of significance was considered to be P<0.05. Values are expressed as mean ± standard deviation. The laboratory data is summarized in Tables 1, 2 and 3. Eleven patients showed overt renal failure, with 5 of them progressing to dialysis. None of the patients showed glycosuria. B2M was detected in all asymptomatic gene carriers with a mean value of 62.7 ng/mL (range 16.1 to 121 ng/mL). Also, A1M was present in 4 subjects with a mean value of 9.7 mg/L (range 8.5-12 mg/L) and A2M was only detected in one individual with a value of 3.34 mg/L. All values were on the reference range. Pathological urinary A1M, B2M and A2M levels were detected in 17, 6 and 5 patients respectively. However, none of the asymptomatic gene carriers showed such abnormal excretion. Six patients had albuminuria <30 mg/g, 4 between 30 and 300 mg/g and 20 >300 mg/g. Among normoalbuminuric patients we found one with urinary pathological levels of A2M and A1M and another with abnormal B2M levels. Among 14 patients who evolved to ESRD, 5 presented simultaneous detection of A1M and A2M. A1M and B2M, were positively correlated with albuminuria, serum creatinine and cystatin C (Table 4) in all patients. A2M was almost exclusively found in the presence of albuminuria >30 mg/g, although their levels do not correlate with the severity of albuminuria. There were no significant correlations between urinary levels of A1M, B2M and A2M and pro-BNP, CRP and ferritin levels. Discussion In current clinical practice, definitive diagnosis of ATTR nephropathy is based on renal biopsy findings. In our experience, however, the diagnosis can be reliably made in patients with albuminuria in the unequivocal presence of neuropathy. Conversely, we face two constraints. The first is that albuminuria is not a marker of tubulointerstitial damage. The second refers to the fact that this marker is absent in 10 percent of the patients who progress to ESRD [10]. Thus, improved methods for detect onset of kidney amyloid deposits, even before clinical disease, are needed to allow earlier treatment. This study is the first description of the contribution of urinary proteins, other than albumin, as non-invasive and costeffective markers to anticipate renal TTR amyloidosis. A low content of protein in the urine, readily determinable, offers advantage over current biofluids widely used such as serum and plasma. The urine proteome represents the integrated product of glomerular filtration of plasma and protein shedding by cells of the proximal renal tubule, suggestive of both systemic and local contributions [11]. In our study, concentrations of all 3 urinary biomarkers increased progressively with decreasing GFR. One third of our patients presented a clear glomerular proteinuria, although significant tubular component was also observed. Urinary excretion of low-molecular proteins A1M and B2M, which are reliable indicators of tubular impairment, was present in 60 percent of patients. The occurrence of low molecular weight proteinuria, despite the absence of typical tubular syndrome, can be explained by the fact that megalin, a multiligand receptor expressed on the renal proximal tubules, functions as a specialized chaperone protein for internalization and degradation of a number of proteins, including A1M e B2M [12]. We speculate that the mechanism for low molecular weight proteinuria is not tubular damage but rather a saturation of the megalin-mediated endocytosis. So, the protein overload present in the lumen of the proximal tubule results in a combined low and high molecular weight proteinuria. We presume that urinary excretion of tubular proteins is related with severity of kidney injury and it is not a precocious marker. The progressive dysfunction of the glomerular barrier leads to nonselective waste of high molecular weight proteins. Excretion of A2M is considered to be related with the severity of albuminuria. Unexpectedly, this correlation was not found. Tencer et al. reported the proteinuria selectivity index as useful to describe changes of the glomerular permeability for macromolecules in glomerular diseases [13]. Proteins the size of A2M cannot normally pass the glomerular barrier. Based on a comparison of the clearance of high-molecular-weight proteins to that of albumin, the pattern of glomerular proteinuria may be described as either selective or nonselective. Early in the course of diabetic glomerular disease, selective proteinuria in both micro-and macroalbuminuric stages is observed. As the disease progresses, proteinuria becomes more nonselective and those with selective proteinuria tend to have a better outcome. We consider that probably the low number of patients with pathological elimination of A2M did not allow a significant correlation. However, the combined excretion of low and high molecular weight proteins was exclusively found in patients who progressed to ESRD. Vyssouli et al. in a study involving 1445 nondiabetic patients revealed that urinary A1M is independently associated with circulating acute phase proteins in patients with newly diagnosed hypertension. They concluded that urinary A1M may reflect the overall inflammatory status in patients with newly diagnosed hypertension, beyond its value as a marker of renal function [14]. In order to exclude the variation associated with inflammatory markers we evaluated CRP and ferritin. None of the individuals presented abnormal values. We did not find a correlation between A1M and these inflammatory proteins. Cardiomyopathy is another well-known complication in FAP. Considering that pro-BNP appears to be a sensitive marker for heart complications and proved valuable for follow-up purposes [15], we decided to search for a correlation between cardiac and kidney biomarkers. Nonetheless, no significant relation was found for this sample. It must be highlighted that our study design has some limitations, such as the small size of the sample and the lack of a gold standard method to evaluate GFR. Additionally, we only had single measurements of B2M, A1M, A2M, CysC, and creatinine, and these measurements are known to vary within participants. However, it should be noted the fact that several proteins were evaluated together in the same population. It is likely that a combination of biomarkers will be required for assessing disease detection and future response to a treatment. In conclusion, the use of urinary low and high molecular weight proteins, like A1M, B2M and A2M, represented useful markers to estimate injury severity and monitoring the progression of renal lesion in ATTR V30M. The design of clinical urinary proteomics studies may increase our understanding of renal involvement in ATTR in the near future.	2019-03-11T13:11:43.853Z	2014-09-18T00:00:00.000	{ "year": 2014, "sha1": "d640d64b2c6f7ced32b67c59c06a5b97a421e72f", "oa_license": "CCBY", "oa_url": "https://www.omicsonline.org/open-access/urinary-biomarkers-for-kidney-disease-in-attr-amyloidosis-2161-0959.1000181.pdf", "oa_status": "HYBRID", "pdf_src": "MergedPDFExtraction", "pdf_hash": "585202e41d0bb13319d96283542750cd24b7f672", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
255600550	pes2o/s2orc	v3-fos-license	Living in ancestrally diverse states of the United States is associated with greater vagal tone Historically, exposure to dissimilar others (“strangers”) was a physiologically arousing event—resulting in avoidance, distrust, and even conflict. Despite this, contemporary migration patterns are increasing intergroup contact. What gives rise to an individual’s ability to regulate their arousal such that social engagement with outgroup members is possible? We propose that cultural practices that evolve in ancestrally diverse, compared to ancestrally homogeneous, societies provide more opportunities for society members to engage in emotion regulation. This regulatory exercise, in turn, promotes higher vagal tone—a physiological indicator of one’s ability to effectively manage arousal in social interaction. In a secondary analysis of data from the MIDUS 2 Biomarker Project, we find that the ancestral diversity of the states of the United States significantly predicts the average vagal tone of the state’s citizens. The findings suggest that social context is associated with predictable and significant adaptations of human physiology over individual lifetimes. Introduction The emergence of a globalized economy as well as increases in international migration has altered the world's demographic landscape such that contact with people from different cultural and linguistic backgrounds has become a common feature of social life. In some regions of the world, defined at multiple geo-political levels (e.g., countries and states of countries), this trend is historically recent. In other regions, the intermingling of people from different cultures and linguistic groups has occurred continuously over a long history. With respect to evolved human predispositions, the intermingling of different peoplesintergroup contact-is a social challenge because humans largely prefer homogeneity. That is, people are generally more attracted to and comfortable around people similar to themselves (Moreland and Beach, 1992;McPherson et al., 2001). Contact with dissimilar people ("strangers") typically elicits higher sympathetic nervous system arousal (e.g., stress Frontiers in Psychology 02 frontiersin.org and anxiety) relative to contact with familiar people (Blascovich et al., 2001;Mendes et al., 2002). Considering the trend toward cultural intermingling, a question urgently in need of an answer is what gives rise to the ability to successfully tolerate and socially engage with diverse others? Here, we develop and test the idea that living in ancestrally diverse societies is associated with adaptations of human physiology, specifically of the vagal system, which in theory supports the ability to regulate sympathetic arousal such that social engagement is tolerable. We propose that the cultural practices and values that represent solutions to the challenges of living in ancestrally diverse populations over generations constitute opportunities to regulate arousal through the activation of the parasympathetic nervous system. Repeated exercise of parasympathetic control-via behavioral and emotional regulation during social interactions-should give rise to greater average vagal tone, an indicator of individual differences in regulatory control exerted by the parasympathetic nervous system. In the reported study, we test these ideas by examining the relationship between the ancestral diversity of the states of the United States and the average vagal tone of the states' citizens, controlling for other factors known to explain significant variance in vagal tone. Beginning roughly 500 years ago, innovations in cartography, shipbuilding, and navigation supported unprecedented waves of migration around the globe (Diamond, 1997). The distribution of the migration was, however, uneven. Some regions (e.g., present-day Brazil and the United Kingdom) became host to migrants from many different countries and, as a result, members of their populations had frequent contact with cultural outgroup members over history. Conversely, other regions (e.g., present-day Japan and Norway) were recipients of far fewer migrants (Putterman and Weil, 2010). Members of their populations therefore had very little contact with cultural outgroup members, or their norms and practices. Differences in the migratory history of regions are captured by a socio-ecological variable, ancestral diversity (Putterman and Weil, 2010), which can be quantified by the number of individual source countries that contributed to at least 0.5% of the current population of a given country since 1500 AD (Wood et al., 2016). Importantly, life in ancestrally diverse societies is very different from life in less ancestrally diverse societies (Niedenthal et al., 2017). First, the former societies present the unique challenge of regularly navigating interactions with people very different from the self. Throughout human history, strangers and dissimilar people (i.e., outgroup members); on the rare occasions they were encountered, were viewed with suspicion, and were attacked or avoided (Neuberg and Schaller, 2016). Recent metaanalyses have noted that the prospect of interacting with outgroup members still produces negative affect and anxiety (Toosi et al., 2012), and the encounters themselves can trigger automatic threat-like responses indicated by sympathetic nervous system arousal, amygdala activation, and excessive cardiac reactivity (Phelps et al., 2000;Blascovich et al., 2001;Mendes et al., 2002). Second, ancestrally diverse societies are more unpredictable, meaning that it is harder to anticipate a social partner's behavior, feelings, or goals (Niedenthal et al., 2019). Indeed, indicators of the ancestral diversity of countries of the world are negatively correlated with the tightness of prevalent social norms and positively correlated with present-day relational mobility, suggesting that norms and social ties in ancestrally diverse societies tend to be less rigid (Niedenthal et al., 2019). As noted by Niedenthal and colleagues, social unpredictability can be reduced if people make their internal states explicit through eye contact and nonverbal communication. Consistent with this claim, recent studies demonstrate that ancestral diversity positively predicts display rules for the overt expression (versus suppression) of emotion, cross-cultural recognizability of facial expressions, and the frequency of use of smiles of reward and affiliation (Rychlowska et al., 2015;Niedenthal et al., 2018Niedenthal et al., , 2019. Importantly, the exchange of non-verbal signals such as eye contact and facial expressions, are themselves social challenges in that they are arousing and require emotion regulation. It must also be noted that the development of such norms and practices requires long periods of time over which they can be transmitted generationally and evolve (Mesoudi, 2016). One potential consequence of this is that the alleviation of social tensions that often accompany initial exposure to diverse others (e.g., increases in outgroup prejudice and conflict; Esteban et al., 2012), may require long-term exposure to outgroup members in order for adaptive social strategies to evolve within a culture. Indeed, this idea is supported by recent work demonstrating that, while short term-exposure to diversity is associated with increases in outgroup prejudice and discrimination, long-term exposure to diversity is associated with decreases in prejudice and discrimination toward outgroup members (Ramos et al., 2019) In the context of ancestral diversity, this idea is supported by evidence that the effects of ancestral diversity on social cues and emotional expressivity are distinct from the impact of present-day diversity-quantified as the number of individual source countries which have contributed to at least 0.5% of a region's population from 2000 to 2010 years (Niedenthal et al., 2018). To summarize, societies defined by high vs. low ancestral diversity are very different social contexts for human development. The practice and norm characteristics of the former entail frequent experiences (e.g., eye contact, explicit displays of emotion, and numerous encounters with outgroup members) that require effective control over sympathetic arousal to successfully engage with individuals different from the self. Moreover, there is evidence to support the claim that frequent experiences of social challenges only result in socially beneficial practices and norms when afforded enough time for cultural evolution to occur. Unlike species with phylogenetically older vagal systems, in humans, the vagal system governs two distinct strategies when responding to environmental challenges, particularly social ones (Porges, 1995(Porges, , 2003(Porges, , 2007. When threat levels are perceived to be high, the vagus' influence is suspended or reduced, sympathetic arousal is uninhibited, and bodily Frontiers in Psychology 03 frontiersin.org resources needed to take appropriate action (e.g., fight or flight) in response to the threat are mobilized. When the environment is perceived as non-threatening, vagal activation can serve to rapidly "brake" the sympathetic nervous system's influence on the heart to reduce activity of the hypothalamic-pituitaryadrenal axis and generate a calm state that enables successful social engagement (Porges, 2007). In contexts in which behavioral practices and norms involve activation of sympathetic arousal, the application of the vagal "brake" must occur frequently and accurately. The efficacy and accuracy of the vagal system's ability to "step on the brakes" at the level of the heart are referred to as vagal tone (Porges et al., 1996). Vagal tone varies across individuals and higher levels are associated with positive social outcomes including the tendency to engage in prosocial behavior and the ability to respond flexibly to environmental conditions (Kok and Fredrickson, 2010;Geisler et al., 2013). Children with greater vagal tone are less aggressive and less likely to exhibit behavioral control problems than children with lower vagal tone (Porges et al., 1996;Beauchaine et al., 2007). And adults with greater vagal tone have lower social anxiety (Porges et al., 1996), less defensiveness (Alvares et al., 2013), better emotion recognition (Movius and Allen, 2005), and stronger feelings of social connection when interacting with others (Colzato et al., 2017). Critically for the present account, experiences of emotion regulation can reliably lead to temporary increases in an individual's average, baseline vagal tone (Butler et al., 2006;Colzato et al., 2017), and, in theory, occur over medium timescales (e.g., ontogeny) if repeated. We therefore reasoned that if the norms and practices which evolved in societies with high ancestral diversity require frequent regulation of emotions and arousal, members of such societies should have, on average, greater vagal tone than members of societies low in ancestral diversity. For the present study, we accessed measures of high-frequency heart rate variability (HF-HRV) from the Midlife in the United States (MIDUS 2) Biomarker Project as an indicator of vagal tone. We then tested the potential relationship between the average HF-HRV of citizens of the U.S. and a measure of the ancestral diversity of the state in which they reside. We expected to observe a significant positive relationship, even when controlling for individual and state-level health and economic status. Our account also holds that the adaptive norms and practices designed to successfully navigate the social challenges of diverse living may take multiple generations to fully develop, and benefit a population. Put another way, in contexts in which cultural practices have had insufficient time to evolve, initial interactions with diverse others may be linked to poor arousal regulationindexed via vagal tone. We assessed this claim by also testing the potential relationship between the average HF-HRV of citizens of the U.S. and the present-day diversity of the state in which they reside while controlling for that state's ancestral diversity as well as the other correlates of vagal tone. We expected to find no relationship or a negative relationship between present-day diversity and vagal tone-highlighting an important distinction between the effects of historical exposure to diversity and shortterm exposure to diversity on the adaptation of one's physiology. HF-HRV and demographics We conducted analyses using data from three existing datasets. Baseline HF-HRV, demographic, and health characteristic data were originally collected during the Bioindicators project of the National Survey of Midlife Development in the United States (MIDUS 1) follow-up study (MIDUS 2). MIDUS 1 is a nationally representative sample of 7,108 individuals across 50,000 households selected via random telephone dialing (Brim et al., 2018). MIDUS 2 followed up with 5,895 of the original MIDUS 1 participants with a 30-min telephone survey (Ryff et al., 2019). Of these 5,895 participants, 1,255 were eligible, and subsequently selected, to participate in the Bioindicators project (Ryff et al., Frontiers in Psychology 04 frontiersin.org 2017). Participants (n = 1,255) who comprise the Bioindicators project had a mean age of 54.5 (SD = 11.7), were mostly White (78.3%), and 56.8% were female. Participants attended a 2-day hospital clinic visit during which they underwent cardiovascular, immune, neuroendocrine, and basic physical assessments. Continuous ECG and respiration recordings were taken during the entire 2-day visit and used to calculate a measurement of both low and HF-HRV for each participant. Complete details about how ECG and respiration data were collected, cleaned, used to calculate HF-HRV, and transformed can be found in the "Bioindicators in the MIDUS National Study Protocol" (Dienberg Love et al., 2010). The present analyses utilized the HF-HRV data collected during a 6-min resting baseline period. Each participant's baseline HF-HRV data were residualized on their sex, age, and BMI. These health-adjusted HF-HRV scores were then used to calculate cluster-means at the level of the state. The lack of individual-level data from Alaska, Hawaii, and Wyoming, as well as the inclusion of data from Washington DC (counted as a state here for the sake of uniformity in terms), resulted in 48 total "state-level" HF-HRV scores. All analyses that included HF-HRV as a variable were performed using these demographic-adjusted, state-level HF-HRV scores. Scores varied between −0.764 and 1. State health State health scores were retrieved from the 2019 Opportunity Index (Opportunity Index, 2019). The Opportunity Index is an annual report which collects demographic, social, and economic data from individuals across all 50 US states. These data are used to determine a single Opportunity Index score for each state. The Opportunity Index calculates state health scores from three health indicators: The percentage of infants born weighing less than 5.5 pounds, the percentage of a state's population (under the age of 65) without health insurance, and the number of deaths per 100,000 attributed to alcohol or drug poisoning, or suicide. Scores were then standardized, and reverse coded so that higher scores indicated greater health opportunities. State Health scores varied between 36.7 and 71.2 (M = 54.6, SD = 8.7). Data cleaning and outlier removal All analyses reported in the present manuscript, unless otherwise stated; use state-level data from 46 of the 50 US states. HF-HRV and ancestral diversity data from Alaska, Hawaii, and Wyoming were not available and, thus, not included in the analyses. An evaluation of model assumptions as well as outlier analyses was conducted to check for observations with strong influence on our final models. Following Ancestral diversity of the United States. Average number of foreign countries which have contributed to at least 0.5% of a state's population since 1850. Darker colors indicate a larger number of source countries (operationalized as high ancestral diversity). Scores are based on decennial census data from 1850 to 2010. Gray indicates missing data. Frontiers in Psychology 05 frontiersin.org field standards (Belsley, 1980), a DFBETA cutoff of 2/sqrt(n) (0.29) was set. All observations which exceeded this cutoff were classified of model outliers and not included in our final analyses. Both Rhode Island and Washington DC exceeded this cutoff (DFBETA >0.29) and were excluded from analyses. Our main findings are derived from analyses utilizing the outlier-adjusted dataset (i.e., excluding Rhode Island and Washington). However, we have also provided the output from our analyses when using the full dataset (i.e., a dataset including the aforementioned outlier states). Ancestral diversity and HF-HRV The potential relationship between ancestral diversity and HF-HRV was assessed with a weighted least squares, multiple regression model in which health-adjusted HF-HRV (i.e., HF-HRV while statistically controlling for the effects of individual health, sex, age, and BMI) was regressed on ancestral diversity while statistically controlling for present-day diversity and state health. Model weights were based upon the number of observations per state (M = 23.74, SD = 33.36). As predicted, there was a significant linear relationship between ancestral diversity and HF-HRV, b = 0.029, sd_b = 0.435, t (42) = 2.291, p = 0.027, η p 2 = 0.11, CILO = 0.003, CIHI = 0.055 (Figure 2). Residents of states high in ancestral diversity (e.g., Illinois, New Jersey, and Michigan) had greater baseline HF-HRV than did residents of ancestrally homogeneous states (e.g., Alabama, Georgia, and Tennessee). The full model accounted for 13% of the total variance (R 2 = 0.134) in HF-HRV scores. On the x axis, the average number of foreign countries which have contributed to at least 0.5% of a state's population since 2010. On the y axis, a mean-centered average of baseline, HF-HRV residualized on health, age, and sex. This figure is generated from a data set consisting of 46 data points. Each data point represents one of the U.S. states (excluding Alaska, Hawaii, Rhode Island, and Wyoming). The size of each data point represents the number of observations per state from which the HF-HRV scores were calculated. Present-day diversity and HF-HRV The potential relationship between present-day diversity and HF-HRV was also assessed through the analysis of the previously introduced statistical model. As predicted, analysis revealed a significant, negative relationship between present-day diversity and HF-HRV, b = −0.018, sd_b = −0.277, t (42) = −2.299, p = 0.0265, η p 2 = 0.11, CILO = −0.034, CIHI = −0.002 (Figure 3). On average, individuals living in states in which the present-day population has been informed by a greater number of foreign countries since 2010 (i.e., more heterogenous and present-day state populations) have lower baseline HF-HRV as compared to individuals residing in states in which the present-day population has been informed by fewer countries since 2000-2010 (i.e., more homogeneous and present-day state populations). This relationship held above and beyond the effects of ancestral diversity, state health (i.e., health insurance coverage, low birth weight, and substance-related deaths), and individual health (i.e., BMI, age, and Sex). Control variables and outlier impact In light of the present study's limitations in power (see our limitations section for more information), the selection of control variables was done conservatively and in observation of the potential impact that the careless inclusion of control variables can have on the conclusions drawn from our analyses (Spector and Brannick, 2011). As such, our choice of control variables was primarily informed by existing guidelines for the analysis of HF-HRV data (Quintana and Heathers, 2014). This included both state health and individual health (accounted for by residualizing HF-HRV scores on individual health variables such as BMI, sex, Frontiers in Psychology 07 frontiersin.org and age). Despite its inclusion as a control variable, the effect of state health on HF-HRV, while statistically accounting for all other predictors in the model, was not significant (p = 0.064). Additionally, preliminary analyses revealed that state-level demographic variables such as state education and economic health were not significantly related to HF-HRV (p > 0.10). This fact, in combination with the previously mentioned conservative approach to the implementation of statistical controls, resulted in the decision to not include these variables in our final model. Analysis of the abovementioned statistical model while using the full dataset (i.e., data including the previously identified outlier states) yielded similar results. There was a significant, positive, linear relationship between HF-HRV and ancestral diversity, b = 0.029, sd_b = 0.291, t (44) = 2.291, p = 0.032, η p 2 = 0.10, CILO = 0.002 CIHI = 0.055, as well as a significant, negative, linear relationship between HF-HRV and present-day diversity, b = −0.018, sd_b = −0.185, t(44) = −2.226, p = 0.0312, η p 2 = 0.10, CILO = −0.035 CIHI = −0.001. Discussion Despite historically fearing and avoiding individuals considered to be outgroup members (e.g., of different ethnic, racial, or religious groups), humans now find themselves in societies that grow more diverse by the day. Previous work has shown that the intermingling of heterogeneous populations, over long history, gives rise to practices and values that support successful communication and cooperation (Niedenthal et al., 2019). Results of the present study suggest further that these practices and values constitute a social context associated with specific adaptations of human physiology. We find that citizens who reside in states of the U.S. that are high in ancestral diversity have greater average vagal tone, an indicator of effective emotion regulation, than citizens who reside in states lower in ancestral diversity. This effect persists when controlling for factors such as individual and state health-both of which can have a substantial relationship to vagal tone (Karason et al., 1999;Aziz et al., 2012;Hill et al., 2015;Koenig and Thayer, 2016). The negative relationship between present-day diversity and vagal tone-while controlling for ancestral diversity-suggests an important distinction between the effects of long-term (ancestral) and short-term (present-day) exposure to diversity. We posit that the cultural norms and practices which evolved in ancestrally diverse populations-greater emotional expressivity, behavioral regulation, and emotional regulation-are specific to social interactions with diverse others for whom those norms were necessitated. For example, norms and practices initially designed to facilitate healthy interactions with large number of immigrants from Germany and Norway in Wisconsin circa 1850 may not extend to, or be effective during, social interactions with immigrants from India or Mexico in present-day Wisconsin. Indeed, past research has shown that large-scale changes in cultural norms require transmission across multiple generations (Mesoudi, 2016). Moreover, when interacting with diverse others for whom adaptive cultural norms have had insufficient time to evolve, it is possible that the norms and cultural practices for interacting with these newer out-group members lead to increased conflict-an outcome that has been linked to initial increases in diversity (Esteban et al., 2012). Future work will test the theory that the negative link between present-day diversity-despite the positive relationship with ancestral diversity-may be explained by differences between the countries which historically immigrated to and contributed to a state's population and the countries which are presently immigrating and contributing to a state's population. Limitations Although the present theory is novel and the findings are robust, claims about the cause of the observed relationship must be made with caution. It is possible that individuals with greater vagal tone were more likely to migrate to states that became ancestrally diverse. It is also possible that a third variable is responsible for our findings. For example, immigrants from many different countries with diets high in fish oil-the consumption of fish is linked to greater vagal tone (Mozaffarian et al., 2008)-may have migrated to the Great Lake States or other coastal regions of the U.S. due to a desire to replicate their diet in the New World. Thus, these populations would have been both diverse and high in average vagal tone without one variable causing the other. The issue is a complex one, to which future experimental, as compared to cross-sectional, studies may offer greater insight. Additionally, the majoritively White populations (78.3%) from the MIDUS 2 Bioindicators project limits the generalizability of the present findings. The social contexts in which Americans lived from 1850 to 2010, and the practices required to successfully navigate those contexts have been-and still are-irrefutably different for people of color, as compared to Whites. Consequentially, it cannot be assumed that the theoretical motivation behind the present analyses is equally compelling for populations which have dealt and are dealing with the added social-and physiological (Williams and Mohammed, 2013)threat of discrimination and systemic racism. Nor can it be assumed that the present findings hold true populations of color. Indeed, a 2015 meta-analysis revealed that, on average, Black Americans, as compared to White Americans, possess greater baseline HRV (Hill et al., 2015). It has since been posited that this finding-dubbed the "African American Vagal Advantage"-may be linked to the experience and anticipation of racial discrimination (Snijders, 2005). Future work aims to address this limitation by examining the relationship between ancestral diversity and HRV in majoritively Black populations and by examining the potential moderating effect of participant race and state-level indicators of historic and present-day discrimination. It should also be noted that the number of observations (n = 1,255) is not evenly distributed across the 48 sampled states. This uneven distribution is addressed and accounted for in our Frontiers in Psychology 08 frontiersin.org statistical models through the use of a weighted analysis. Similarly, it must be noted that statistical power for a level-2 analysis-such as the analyses conducted in the present study-is derived from the number of level-2 observations (Snijders, 2005). Because the maximum possible sample size is 51 (i.e., the number of U.S. states and Washington D.C.), the study's potential power is hard capped at 0.66. With a sample size of 46 (following the removal of the two outlier states, Rhode Island and Washington), the present study has a statistical power of 0.63. Certainly, this limit in power hinders the extent to which firm conclusions can be confidently drawn from the present findings. As such, these findings should be considered skeptically. Future studies are aimed at replicating the present study's findings on an international scale. Sampling from countries, rather than states, would remove the cap of 51 possible observations and, in turn, the limitations in power. Finally, the present findings only provide information about state-level effects and may not generalize to the individual-level. As such, more precise claims about variation at the individuallevel or variation within the states cannot be made from these findings but instead stand as valuable questions to be explored by future research and may also lend insight into questions of directionality. Conclusion The present work serves as a foundational step for multiple lines of research. The findings imply that short-term exposure to greater diversity is not sufficient in promoting adaptive physiological changes and, in turn, greater social tolerance. Rather, such physiological changes may require continued, long-history exposure to the practices and values codified in the cultures that arise over time in ancestrally diverse contexts. This nuance raises an interesting question of how to foster lasting social tolerance in the face of diversity despite its negative short-term impact on vagal tone. Further research will not only further our knowledge of how cultural differences arise but will also serve as a useful guide as outgroup interactions inevitably become more commonplace. Within cultural psychology and the study of cultural differences, the present work provides a newfound physiological mechanism for the well-documented cultural differences in emotional expression, emotional regulation, stress regulation, and outgroup prejudice. Finally, within the field of physiology, the noted long-history changes and regional differences in vagal tone will pave the way for research dedicated to understanding well-documented, but seemingly paradoxical, group-level phenomena such as the African American vagal advantage or the Female vagal advantage (Hill et al., 2017;Shaffer and Ginsberg, 2017). The value of the present work lies in its role as the grounds for a broader theoretical relationship between cultural differences in social behavior and human physiology. It not only demonstrates that where we live and what cultures we are a part of shape our physical and social processes, but ultimately sets the stage for research dedicated to understanding how to cultivate and maintain greater social tolerance-knowledge which is crucial in our ever-shrinking world. Data availability statement The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author. Ethics statement The studies involving human participants were reviewed and approved by the University of Wisconsin-Madison IRB. Written informed consent for participation was not required for this study in accordance with the national legislation and institutional requirements. Author contributions EH: first authorship, formal analysis, and visualization. EH, IS, JM, and PN: conceptualization. EH and JM: methodology. JM and IS: data curation. PN: supervision. IS: writing-original draft. EH and PN: writing-review and editing. All authors contributed to the article and approved the submitted version. Funding This work was supported by Cattell Sabbatical Award MSN206694 (PN). Frontiers in Psychology 09 frontiersin.org 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.	2023-01-12T15:18:41.434Z	2023-01-12T00:00:00.000	{ "year": 2023, "sha1": "312324d23532eb2b57e3c0a20f8038a248848687", "oa_license": null, "oa_url": null, "oa_status": null, "pdf_src": "Frontier", "pdf_hash": "312324d23532eb2b57e3c0a20f8038a248848687", "s2fieldsofstudy": [ "Psychology" ], "extfieldsofstudy": [] }
242808050	pes2o/s2orc	v3-fos-license	Detecting and describing stability and change in COVID-19 vaccine receptibility in the United Kingdom and Ireland COVID-19 continues to pose a threat to global public health. Multiple safe and effective vaccines against COVID-19 are available with one-third of the global population now vaccinated. Achieving a sufficient level of vaccine coverage to suppress COVID-19 requires, in part, sufficient acceptance among the public. However, relatively high rates of hesitance and resistance to COVID-19 vaccination persists, threating public health efforts to achieve vaccine-induced population protection. In this study, we examined longitudinal changes in COVID-19 vaccine acceptance, hesitance, and resistance in two nations (the United Kingdom and the Republic of Ireland) during the first nine months of the pandemic, and identified individual and psychological factors associated with consistent non-acceptance of COVID-19 vaccination. Using nationally representative, longitudinal data from the United Kingdom (UK; N = 2025) and Ireland (N = 1041), we found that (1) COVID-19 vaccine acceptance declined in the UK and remained unchanged in Ireland following the emergence of approved vaccines; (2) multiple subgroups existed reflecting people who were consistently willing to be vaccinated (‘Accepters’: 68% in the UK and 61% in Ireland), consistently unwilling to be vaccinated (‘Deniers’: 12% in the UK and 16% in Ireland), and who fluctuated over time (‘Moveable Middle’: 20% in the UK and 23% in Ireland); and (3) the ‘deniers’ and ‘moveable middle’ were distinguishable from the ‘accepters’ on a range of individual (e.g., younger, low income, living alone) and psychological (e.g., distrust of scientists and doctors, conspiracy mindedness) factors. The use of two high-income, Western European nations limits the generalizability of these findings. Nevertheless, understanding how receptibility to COVID-19 vaccination changes as the pandemic unfolds, and the factors that distinguish and characterise those that are hesitant and resistant to vaccination is helpful for public health efforts to achieve vaccine-induced population protection against COVID-19. Introduction The rapid development of safe and effective vaccines against Coronavirus Disease (COVID- 19) represents one of the greatest collaborative scientific achievements of our lifetime. As of August 2021, four vaccines have been authorised by the European Medicines Agency, three have been authorized for emergency use by the United States Food and Drug Administration, and 99 are undergoing clinical trials on humans [1]. Just under five billion vaccines doses have been administered, globally, meaning that 31% of the world's population have been vaccinated and it is estimated that 75% of the world's population will be vaccinated by February 2021 [2]. Sufficient uptake of COVID-19 vaccines not only requires the coordinated action of governments, communities, and individuals alike to ensure adequate vaccine delivery (e.g., via production, logistics, procurement, financing, and service delivery components of the health system), but also to ensure vaccine receptibility. COVID-19 vaccine acceptance rates across the world range from lows of 24% in Kuwait and 44% in Lebanon to highs of 88% in China and 91% in India [3][4][5]. Concurrently, rising rates of vaccine hesitancy, whereby an individual delays or refuses vaccination despite the availability of inoculation services [6], remains one of the greatest global health threats listed by the World Health Organization [7]. As the term implies, however, vaccine hesitancy is not immutable, and individual attitudes towards a specific vaccine can change over time as a function of a wide-range of interdependent individual, social, and vaccination-specific factors [8,9] including, but not limited to, perceptions of susceptibility to pathogen exposure [10], severity of illness [11], perceived vaccine safety and efficacy [11,12], and recency of vaccine development [13]. Accordingly, some have suggested that vaccine hesitancy is better conceptualised as existing on a continuum and bookended by 'decliners' and 'accepters', or those who completely reject or accept all vaccines, respectively [14]. Levels of COVID-19 vaccine acceptability have fluctuated considerably throughout the pandemic. Most recent data from the global survey of knowledge, attitudes, and practices around COVID-19 (KAP COVID-19)which has reached over 1.7 million people in 67 countries across as many as 19 waves of data collection in some contexts-indicates that only 63% of individuals would accept a COVID-19 vaccine as of the 31 st of January 2021 [15]. Encouragingly, however, these same data suggest that willingness to be vaccinated has increased in nations that have successfully launched COVID-19 vaccination programmes (e.g., the United Kingdom [UK]). Previous work carried out by our group, the COVID-19 Psychological Research Consortium, found that resistance to COVID-19 vaccination in the UK and the Republic of Ireland is associated with distrust of experts and authority figures (i.e., scientists, health care professionals, and government), stronger religious, conspiratorial, and paranoid beliefs, a higher internal locus of control, preference for hierarchically structured and authoritarian societies, antimigrant views, lower levels of agreeableness, conscientiousness, and emotional stability [16]. Similarly, the 'attitude roots' model of science rejection proposes that conspiratorial beliefs, disgust sensitivity, trait reactance-as a motivational state that arises when people feel that their behavioural freedom has been threatened or taken away [17]-and hierarchical worldviews are central to understanding individual differences in vaccine resistant attitudes [18][19][20][21]. Thus, understanding the individual factors, including psychological dispositions, that predict whether vaccine hesitant individuals change their minds about COVID-19 vaccination, as well as the factors that might predict a move towards acceptance or resistance over time is paramount, albeit currently less well understood [9]. In light of these existing gaps, the current study was planned with three primary objectives. The first was to examine changes in COVID-19 vaccine acceptance, hesitance, and resistance in the Irish and UK adult populations across four time periods (Waves) during the first nine months of the global pandemic. We have previously reported on the changes in these populations across the first three waves of the survey (i.e., March-April, April-May, and July-August 2020) [22]; however, as these data were obtained prior to the development of safe and effective vaccines for COVID-19, our focus in this study is on changes from Wave 3 (July/August 2020) to Wave 4 (November/December 2020) when populations transitioned from having to contemplate a hypothetical vaccine to considering an actual, available vaccine. Understanding that people's willingness to accept a COVID-19 vaccine may fluctuate over time, our second objective was to determine if there were multiple groups in each sample with distinct probabilities of accepting a COVID-19 vaccine over time. We hypothesised that there would be two stable groups in each sample: one representing people with consistently high probabilities of accepting a COVID-19 vaccine ('Accepters'), and the other representing people with consistently low probabilities of accepting a COVID-19 vaccine ('Deniers'). Additionally, we expected to identify a group (or groups) in each sample with fluctuating probabilities of accepting a COVID-19 vaccine; a group that have often been termed the movable middle. Finally, we sought to identity key sociodemographic and psychological factors that were associated with belonging to any group that was not consistent in their acceptance of a COVID-19 vaccine. Our intention with the second and third objectives was to develop a comprehensive understanding of the people who were not consistent in their willingness to accept a COVID-19 vaccine so that targeted and effective public health strategies could be developed to reach those who can still change their minds. Participants and procedures This study is based on data from the Irish and UK strands of the COVID-19 Psychological Research Consortium (C19PRC) study. The C19PRC study was established to track the social, political, economic, and mental health effects of the COVID-19 pandemic on society. Data for this study were collected at four assessment points during the first nine months of the COVID-19 pandemic. Wave 1 data were collected in the UK between March 23 rd and 28 th , 2020, and in Ireland between March 30 th and April 5 th , 2020. These dates coincided with the initial public health lockdown measures in the respective countries. Wave 2 data were collected in the UK from April 22 nd to May 1st, 2020, and in Ireland from April 30 th to May 19 th , 2020. Wave 3 data were collected in the UK from July 9 th to July 23 rd , 2020, and in Ireland from July 16 th to August 8 th , 2020. Finally, Wave 4 data were collected in the UK from November 25 th to December 22 nd , 2020, and in Ireland from December 2 nd to December 22 nd , 2020. The UK and Irish samples were collected using a non-probability Internet panel survey design. The survey research company Qualtrics was employed to recruit participants from traditional, actively managed, double-opt-in research panels via email, SMS, or in-app notifications. Inclusion criteria for both samples were that respondents were aged 18 years or older, residing in the UK or Ireland, respectively, and capable of completing the survey in English. Ethical approval was granted by the research ethics committees at the University of Sheffield (Reference number: 033759), Ulster University (Reference number: 230320), and Maynooth University (Reference number: SRESC-2020-2402202). Participants were remunerated by Qualtrics, and informed electronic consent was obtained from all participants. Quota sampling methods were used at Wave 1 to generate samples that represented the general adult populations of both nations. In the UK, the sample was recruited to match known population quotas for sex, age, and income distributions. In Ireland, the sample was recruited to match known population quotas for sex, age, and regional distribution. Further details regarding the UK and Irish samples, including evidence of their representativeness, are presented elsewhere [23 -25]. As described in an earlier study [16], power analyses to determine optimal sample sizes were calculated to detect common mental health disorders such as Major Depressive Disorder and Posttraumatic Stress Disorder. Sample size calculations were performed to detect a disorder with a 4% prevalence rate, with a precision of 1%, and 95% confidence levels. This resulted in a required sample size of 1,476. As Qualtrics could only guarantee a sample size of 1,000 participants in Ireland, this was set as the target sample size in Ireland. Holding all other parameters in the sample size calculation equal, this sample size resulted in a precision of 1.21%. Given the substantially larger population of the UK and thus the availability of a larger pool of potential participants, we set a target sample size of 2,000 people. At Wave 1, the sample size in the UK was 2,025 and 1,041 in Ireland. The sociodemographic characteristics for both samples at Wave 1 are reported in Table 1. In the UK, the recontact rate was 69% (n = 1406) at Wave 2, 58% (n = 1166) at Wave 3, and 63% (n = 1271) at Wave 4. Those who responded at each wave significantly differed (p < .05) from non-responders on a range of sociodemographic variables including being older, male, living with fewer adults, higher income earners, born in the UK, not living in a city, having a post-secondary education, and not having a suspected or confirmed COVID-19 infection. In Ireland, the recontact rate was 49% (n = 506) at Wave 2, 51% (n = 534) at Wave 3, and 40% (n = 416) at Wave 4. Respondents significantly differed (p < .05) from non-responders by being older, more likely to have been born in Ireland, not living in a city, to have a pre-existing health condition, and not having a suspected or confirmed COVID-19 infection. Management of missing data is outlined in the data analysis section. COVID-19 vaccination status. In the UK and Irish samples, participants were asked the following question at Waves 1, 2, and 3: 'If a new vaccine were to be developed that could prevent COVID-19, would you accept it for yourself?' At Wave 4, participants in both samples were asked: 'Multiple vaccines for COVID-19 have now been developed. Will you take a vaccine for COVID-19 when it becomes available to you?' The response options at all times were 'Yes', 'Maybe', and 'No'. Those who answered 'Yes' were classified as 'vaccine accepting', those who responded 'Maybe' were classified as 'vaccine hesitant', and those who responded 'No' were classified as 'vaccine resistant'. Sociodemographic, political, and health indicators (Measured at Wave 1). The sociodemographic, political, and health indicator variables used in this study were identical to those utilized in our previous study [16], and all are listed in Table 1. For analytical purposes, several of these variables were recoded. Living location was recoded to represent city dwelling vs. noncity dwelling; education status was recoded to represent post-secondary education vs. nonpost-secondary education; employment status was recoded to represent unemployed vs. all other options; and religion was recoded to represent any religious identification vs. atheist or agnostic. Additionally, due to limited numbers in various subgroups, ethnicity was recoded to represent self-identified Irish ethnicity vs. non-Irish ethnicity in the Irish sample. Psychological indicators (Measured at Wave 1). Personality traits. The Big-Five Inventory (BFI-10) [26] measures the traits of openness to experience, conscientiousness, extraversion, agreeableness, and neuroticism. Each trait is measured by two items using a five-point Likert scale that ranges from 'strongly disagree' (1) to 'strongly agree' (5). Higher scores reflect higher levels of each personality trait, and Rammstedt and John [26] reported good reliability and validity for the BFI-10 scale scores. Internal reliability coefficients are not provided as this scale measures each trait using only two items, and it is well documented that coefficient alpha is inappropriate and meaningless for two-item scales [27]. Locus of control. The Locus of Control Scale (LoC) [28] measures internal (e.g., 'My life is determined by my own actions') and external locus of control. The latter has two components, 'Chance' (e.g., 'To a great extent, my life is controlled by accidental happenings') and 'Powerful Others' (e.g., 'Getting what I want requires pleasing those people above me'). Each subscale was measured using three questions and a seven-point Likert scale that ranges from 'strongly disagree' (1) to 'strongly agree' (7). Higher scores reflect higher levels of each construct. The internal reliabilities of the Internal and Chance subscale scores in the Irish sample were slightly lower than desirable (α = .67 &.63, respectively) but somewhat stronger for the UK sample (α = .71 &.70, respectively), while those for the Powerful Others subscale scores were good in both samples (Ireland: α = .78; UK: α = .85). Analytical/reflective reasoning. The Cognitive Reflection Task (CRT) [29] is a three-item measure of analytical reasoning where respondents are asked to solve logical problems designed to hint at intuitively appealing but incorrect responses. The response format was multiple choice with three foil answers (including the hinted incorrect answer), as recommended by Sirota and Juanchich [30]. The internal reliabilities of the CRT scores in the Irish and UK samples were α = .67 and α = .69, respectively. Altruism. The Identification with all Humanity scale (IWAH) [31] is a nine-item scale. Respondents are asked to respond to three statements with reference to three groups; people in my community, people from Ireland/ the UK, and all humans everywhere. The three statements were presented to respondents separately for each of the three groups, as follows: (1) How much do you identify with (feel a part of, feel love toward, have concern for) . . .? (2) How much would you say you care (feel upset, want to help) when bad things happen to . . Conspiracy beliefs. The Conspiracy Mentality Scale (CMS) [32] measures conspiracy mindedness using five items with each scored on an 11-point scale (1 = 'Certainly not 0%' to 11 = 'Certainly 100%'). Items include, 'I think that many very important things happen in the world, which the public is never informed about', and 'I think that there are secret organizations that greatly influence political decisions'. The internal reliability of the CMS in both the Irish and UK samples was good (α = .84 &.85, respectively). Paranoia. The five-item persecution subscale from the Persecution and Deservedness Scale was used [33]. Participants rate their agreement with statements such as "I'm often suspicious of other people's intentions towards me" and "You should only trust yourself." Response options ranged from 'strongly disagree' (1) to 'strongly agree' (5) with higher scores reflecting higher levels of paranoia. The psychometric properties of the scale scores have been previously supported [34], and the internal reliability in both the Irish and UK samples was good (α = .83 &.86, respectively). Trust. Respondents were asked to indicate the level of trust they have in political parties, Parliament, the government, the police, the legal system, scientists, and doctors and other health professionals. Responses were scored on a five-point Likert scale ranging from 'do not trust at all' (1) to 'completely trust' (5). For this study, responses to the first five institutions were summed to generate a total score for 'trust in the state'. Responses to the final two questions were summed to generate a total score for 'trust in scientists and doctors/health professionals'. Authoritarianism. The Very Short Authoritarianism Scale [35] includes six items assessing agreement with statements such as: 'It's great that many young people today are prepared to defy authority' and 'What our country needs most is discipline, with everyone following our leaders in unity'. All items were scored on a five-point Likert scale ranging from 'strongly disagree' (1) to 'strongly agree' (5), with higher scores reflecting higher levels of authoritarianism. The internal reliability of the scale scores in the Irish sample was lower than desirable (α = .58) but somewhat stronger for the UK sample (a = .65). Social dominance. Respondents' levels of social dominance orientation were assessed using the eight-item Social Dominance Scale [36]. Respondents were asked the extent to which they opposed/favoured statements such as: 'An ideal society requires some groups to be on top and others to be on the bottom'; 'Some groups of people are simply inferior to other groups'; and 'We should do what we can to equalize conditions for different groups'. Responses were scored using a 5-point Likert scale ranging from 1 'Strongly oppose' to 5 'Strongly Favour'. Ho and colleagues demonstrated that the scale had good criterion and construct validity [36]. The internal reliability of the scale scores in both the Irish and UK samples was good (α = .79 &.82, respectively). Attitude towards migrants. Two items assessing respondents' attitudes towards migrants were taken from the British Social Attitudes Survey 2015 [37]. These were, (1) 'Would you say it is generally bad or good for the UK's economy that migrants come to the UK from other countries?' (scored on a 10-point scale ranging from 1 'extremely bad' to 10 'extremely good'), and (2) 'Would you say that the UK's cultural life is generally undermined or enriched by migrants coming to live here from other countries?' (scored on a 10-point scale ranging from 1 'undermined' to 10 'enriched'). These items were phrased appropriately for use with the Irish sample. Data analysis The first objective was assessed by means of structural equation modelling (SEM). A SEM approach was used so that missing data could be most effectively managed using full information robust maximum likelihood estimation (MLR) [38]. This approach is helpful because it means that all available information at Wave 1 is used to estimate missingness at future waves, thus ensuring minimal loss of statistical power or sample representativeness. This method of estimation can also handle non-normally distributed variables [39]. This analytic process involved three steps. First, a 'null' model was specified where the proportions (e.g., in vaccine acceptance, hesitance, and resistance-all are estimated individually) at Waves 1-4 were constrained to be equal. Second, an 'alternative' model was specified where the proportions were freely estimated at each wave. These models differed by three degrees of freedom and significant differences in model fit were tested using a loglikelihood ratio test (LRT), which follows a chisquare (χ 2 ) distribution. Third, post-hoc pairwise comparisons were tested using a Wald χ 2 test. The second objective was assessed using latent class analysis (LCA). Responses to the question about willingness to accept a COVID-19 vaccine (0 = Yes, 1 = Maybe, 2 = No) at Waves 1-4 were used as the observed indicators in the model. To understand the probability of consistent acceptance of a COVID-19 vaccine across time, we focused our interpretations on the probability of the first response (i.e., 'Yes') within each class. Models with one to six classes were estimated in the Irish and UK samples using MLR. To avoid solutions based on local maxima, 500 random starting values and 50 final stage optimizations were used. The relative fit of these models was compared using three information theory based fit statistics: the Akaike Information Criterion (AIC) [40], the Bayesian Information Criterion (BIC) [41] and the sample size adjusted Bayesian Information Criterion (ssaBIC) [42]. The solution with the lowest value of these statistics is deemed superior, or if no minimum is found then the 'diminishing gains in model fit' for additional classes can be examined [43]. Simulation studies suggest that the BIC is optimal for identifying the correct number of classes [44]. Additionally, the Lo-Mendell-Rubin adjusted likelihood ratio test (LMR-A) [45] was used to compare models with increasing numbers of latent classes. When a non-significant value occurs, the model with one fewer class should be accepted. Model convergence, replication of the log-likelihood, entropy values, the plausibility of the model estimates, and the interpretability of the model solutions were also used to determine the optimal solution. The third objective was assessed by adding the demographic and psychological predictor variables to the best fitting LCA models in the Irish and UK samples, respectively. A 3-step approach was used so that the inclusion of the predictor variables did not influence the formation of the classes [46]. Objective 2: Changing probabilities of vaccine acceptance over time The full set of latent class analysis (LCA) results for the Irish and UK samples are presented in Table 3. In both samples, iterative models with one to four classes terminated normally, and the loglikelihood values were replicated. Models with more than four classes failed to converge or terminate normally in both samples suggesting that models with more than four classes were not viable representations of the sample data. Overall, the results were similar in the two samples in that the Bayesian Information Criteria (BIC) and sample size adjusted BIC (ssaBIC) values were lowest for the three-class models. The Lo-Mendell-Rubin adjusted likelihood-ratio test (LMR-A) values become non-significant at five classes, which suggests that a four-class model may be optimal; however, the p-values for the four-class model were also elevated (Ireland: p = .022; UK: p = .027), suggesting a better fit for the three-class model. Comparing the profiles of the three-and four-class models, a relatively large group of people with high probabilities of accepting a COVID-19 vaccine in the three-class model was differentiated in the four-class model to represent groups with high and moderate-to-high probabilities of vaccine acceptance. Thus, the addition of another class in the four-class model was not qualitatively different from the classes identified in the more parsimonious three-class model. Consequently, based on parsimony, model interpretability, and recognition that BIC is an optimal index for model selection, the three-class model was selected as the best fitting model of the Irish and UK sample data. The probabilities of accepting a COVID-19 vaccine over time in the Irish and UK samples are represented in Figs 2 and 3, respectively. In the Irish sample, class 1 included 16% of people and was characterised by extremely low probabilities of accepting a COVID-19 vaccine over time. Notably, there was a drop-off from an already low probability at Wave 1 (.15) to near zero probabilities of accepting a vaccine through Waves 2-4. This class was labelled 'Deniers'. Class 2 included 61% of the sample and was characterised by high probabilities of accepting a COVID-19 vaccine over time. Yet, it is noteworthy that the probability of acceptance steadily declined from Wave 2 (.93) to Wave 4 (.82), despite remaining high. This class was labelled 'Accepters'. Finally, class 3 included 23% of the sample and was characterised by fluctuating probabilities of accepting a COVID-19 vaccine. This class had a low-to-moderate probability of vaccine acceptance at Wave 1 (.34) that declined markedly by Wave 3 (.05) before increasing again at Wave 4 (.26). This class was labelled 'Movable Middle'. In the UK sample, class 1 included 12% of people and was characterised by declining probabilities of accepting a COVID-19 vaccine over time. This class had a low-to-moderate probability of vaccine acceptance at Wave 1 (.32) that declined through Wave 2 (.17) and Wave 3 (.09) and remained low at Wave 4 (.10), even after the introduction of an approved vaccine. This class was labelled 'Deniers'. Class 2 included 68% of the sample and was characterised by consistently high probabilities of vaccine acceptance. Notably, the probability of vaccine acceptance rose steadily from Wave 1 (.86) to Wave 3 (.97) before decreasing at Wave 4 (.88). This class was labelled 'Accepters'. Finally, class 3 included 20% of the sample and, like class 1, demonstrated declining probabilities of vaccine acceptance from Wave 1 (.33) to Wave 2 (.12) but then diverged from class 1 as the probability of vaccine acceptance increased steadily through Wave 3 (. 19) and Wave 4 (.24). This class was labelled 'Movable Middle'. Objective 3: Correlates of class membership Based on our desire to understand why individuals were not consistent in their willingness to accept a COVID-19 vaccine, the class of 'Accepters' in the Irish and UK samples were set as the reference categories for analyses to determine the correlates of membership in the 'Deniers' and 'Movable Middle' classes. These findings for the Irish and UK samples are presented in Tables 4 and 5 Discussion Three important findings emerged from the analyses. First, the arrival of vaccines against COVID-19 coincided with a significant change in vaccine receptibility, but in only one of the two countries sampled. Second, within both samples, vaccine receptibility over time was most parsimoniously represented by three distinct groups. In Ireland and the UK, the majority of respondents belonged to a group characterised by stable acceptance that accounted for 61% and 68% of each sample, respectively. Conversely, the fewest respondents in both samples belonged to a group characterised by stable non-acceptance (Ireland: 16%) or decreasing acceptance (UK: 12%). A final group characterised by fluctuating probabilities of accepting a COVID-19 vaccine over time was also identified within each sample (Ireland: 23%; UK: 20%). Third, compared to those characterised by stable acceptance over time, individuals characterised by changing or decreasing acceptance of a COVID-19 vaccine were distinguishable, and also comparable, in relation to several individual, socio-economic, and psychological variables. The significance of these findings is described in turn below. Compared to data that had been collected at a time when vaccine receptibility could only be considered in relation to a hypothetical vaccine (i.e., July/August 2020), data from a period when approved vaccines for COVID-19 had been introduced in both countries (December 2020) showed a significant increase in vaccine resistance in the UK, and a significant decrease in vaccine acceptance. No change in vaccine acceptance, resistance, or hesitance was identified in the Irish sample between these timepoints. The proportion of UK respondents in November/December 2020 who indicated that they would be receptive to one of the approved vaccines for COVID-19 when it became available to them (65.5%) was slightly lower than the proportion of the sample who indicated acceptance of a hypothetical vaccine in July of the same year (71.1%). Moreover, the proportion of the sample in November/December 2020 who indicated that they would be resistant to accepting one of the approved vaccines when made available to them (15.6%) was markedly higher than the proportion who indicated resistance to a hypothetical vaccine in July 2020 (10.6%). While this trend may have been attributable to factors other than the arrival of approved vaccines (i.e., our analyses clearly indicated that fluctuation in vaccine receptibility has been at play in both countries for some time), recency of vaccine development and distribution has been identified as one of many factors that can influence vaccine hesitancy [13]. In relation to the COVID-19 pandemic specifically, a study of 1,941 Israeli healthcare workers and members of the general Israeli population has shown that the vast majority of responders' concerns were due to the assumed speed of vaccine development and related concerns surrounding quality controls [10]. It is notable that while the extant literature covers vaccine efficacy and safety extensively, and the rigorous quality controls that precede, dictate, and follow approvals [47,48], members of the general population still identify speed, safety, efficacy, and quality control as key reasons for hesitation/concern about receiving a vaccine. It is imperative therefore that public health authorities do more to educate, inform, and intervene to challenge vaccine hesitancy on these grounds. The current study revealed important vaccine receptibility subgroups and trends in both countries. Mixture modelling of our longitudinal data afforded a valuable opportunity to investigate (i) the proportion of each population that displayed a sustained high probability of vaccine acceptance throughout the pandemic, (ii) the proportion that displayed a sustained low probability of vaccine acceptance, and importantly, (iii) whether a 'moveable middle' group-or groups-existed, and what their receptibility profiles looked like. Overall, 61% and 68% of the Irish and UK samples, respectively, exhibited stable vaccine acceptance with acceptance probabilities in both samples above 80% across all four timepoints. However, somewhat concerningly, the trajectories for both groups ended in a downward trend. It will be important, therefore, to monitor these stable acceptance groups at later survey waves to determine what effect, if any, national vaccination programmes and communication strategies are having on acceptance levels for those who seem committed to vaccination. Notably, however, the size of these groups also reveals significant differences between countries regarding rates of acceptance and highlights the importance of country-specific approaches to understanding and tackling vaccine hesitancy and promoting vaccine receptibility. While we expected to identify distinct subgroups in both populations characterised by low probabilities of vaccine acceptance over time, the profiles for these groups differed in important ways. While the Irish sample included a group characterised by sustained low-tonear zero probabilities of acceptance at each survey wave (16%), the UK's most resistant group (12%) began with a 32% probability of acceptance that steadily declined to 10% by Wave 4. The Irish non-acceptance group, therefore, reflected more extreme and stable resistance compared to those who were most resistant in the UK. Several studies have shown that upwards of approximately 10% of study populations appear to be opposed to vaccinations in whatever form they take [49,50]; therefore, these findings were not entirely surprising. It was notable, however, that resistance was lowest in both countries at the beginning of the pandemic (~6-10% in March/April 2020), and that this resistance steadily rose (significantly between some survey waves) to~16-18% by Waves 3 (July/August 2020) and 4 (November/December 2020). Resistance to actual approved vaccines in December 2020, therefore, was concerningly high. If resistance remains at this level or continues to rise, public health officials will likely need to consider how to reach and persuade a now substantial subpopulation that has traditionally been shown to be extremely resistant to vaccine promotional campaigns and public health messaging regarding inoculation generally [51,52]. A third group was also identified in both countries. This group was considered to represent a 'moveable middle' or 'changing' group that may hold important significance for future public health initiatives that seek to achieve herd-protection against SARS-CoV-2. In the Irish sample, this group was characterised by a 26% probability of accepting a vaccine in December 2020 when approved vaccines had been developed. However, in the months preceding vaccine development (July/August 2020), this same group of respondents exhibited only a 5% probability of acceptance, while at the beginning of the pandemic, acceptance probability was at its highest (34%). Comparatively, the 'moveable middle' group in the UK sample exhibited a similar probability of acceptance in November/December 2020 (24%), and at the beginning of the pandemic (33%) but had its lowest level of acceptance in April/May 2020 (12%). These groups have fluctuated in their positions over the duration of the pandemic, and while there may be cause for optimism in the upward trends identified at the most recent data collection timepoints, it must be noted that neither of these groups displayed a probability of acceptance above 34% at any time since the beginning of the pandemic. While the extant research literature details many distinct socio-demographic and psychological indicators of vaccine hesitancy generally [6,53,54], and a burgeoning literature has begun to list those common to COVID-19 vaccines specifically [10,16,55], studies describing characteristics associated with stability or change in vaccine receptibility over time are lacking. Our findings revealed important similarities and distinctions in vaccine receptibility between those in the 'movable middle' and those characterised by stable resistance in both countries. First, those who fluctuated in their receptiveness to a COVID-19 vaccine in Ireland and the UK were more likely to be female and to lack trust in scientists and health care professionals. Evidence suggests that, in relation to COVID-19 vaccination specifically, females may have concerns surrounding issues such as fertility and pregnancy [56,57]. As has been highlighted earlier, trust in scientists and health care professionals (particularly regarding the speed of vaccine development and distribution) seems also to be of particular concern for many who are hesitant about a COVID-19 vaccine specifically [10]. Public health messaging, therefore, tailored specifically to allay concerns and/or fears that may be specific to women, and/or to educate and reassure the public about quality controls and standards relating to the development, distribution, administration, and review of COVID-19 vaccines may prove useful. Notable distinctions were also evident for the moveable middle groups across samples. In Ireland, those who fluctuated over time were more likely than accepters to believe that powerful others were responsible for their experiences and to hold conspiratorial beliefs, while those in the UK were more likely than accepters to be younger, of Chinese/Asian ethnicity, have a lower level of income, have voted 'other' in the last general election, be lower in extraversion, and higher in openness. These distinct country specific characteristics may help to further inform and refine public health messaging in ways that are contextually sensitive to each population. Second, those who remained resistant over time in Ireland and the UK tended not to live with any other adults, to hold conspiratorial beliefs, and to lack trust in scientists and health care professionals. While those who remained resistant over time may be more challenging to reach or persuade than those who fluctuate in their receptibility, these common indicators of resistance may prove useful in informing our understanding of who these people are and why they are susceptible and committed to the beliefs they hold. Individuals living alone have been shown to lack important opportunities to explore/discuss their concerns or to reality test their assumptions about the world in which they live [58,59], while those who are open/receptive to conspiratorial interpretations of world events often dismiss information sourced from or disseminated by traditional, scientific and/or authoritative sources [60,61]. Notably, as was also evident for the change groups, stable resisters in both countries also differed in specific ways. In Ireland, these individuals were uniquely characterised by low income and negative views towards migrants, while in the UK, those most resistant to a COVID-19 vaccine were more likely to have children, not to have voted in the last general election, and to be lower in the personality traits of agreeableness and neuroticism. Each of these indicators has previously been shown to be associated with vaccine hesitancy/resistance [6,62]. That they do not predict resistance in the same way within different populations and in relation to common vaccines likely reflects the context specific complexity of vaccine hesitancy as a phenomenon and the challenging terrain that must be navigated by those seeking to tackle it. These findings should be interpreted considering several limitations. First, non-probability quota-based sampling methods were used to recruit samples via the Internet. This opt-in mode of recruitment employed by the survey company who facilitated the data collection (Qualtrics), albeit being a cost-effective method for gaining fast access to a large and diverse sample (and the most feasible method of recruitment during the pandemic), inevitably meant that it was not possible to know if participants in these panels differed in important ways from members of the public that do not belong to the panels. Second, the current study was also limited to two western, European countries whose populations had many social, cultural, economic, and political similarities. However, while these populations may have been similar in many respects, our findings highlight notable differences between countries in relation to (i) the proportions of each population that were receptive, hesitant, and resistant over time, (ii) the profiles and trajectories of these groups, and (iii) the specific indicators that predicted fluctuation and stable resistance over time. Now that vaccination programmes are underway in many countries, our findings highlight the importance of population-specific analyses of vaccine hesitancy and the continued monitoring of this phenomenon as vaccination programmes advance. Relatedly, the extent to which these results will generalise to other nations is unknown. It is essential that other (low, middle, and high income) countries obtain estimates of change in hesitancy/resistance to COVID-19 vaccination in their general populations, given that vaccination efforts will only succeed if sufficiently undertaken globally. Third, while the use of nationally representative samples from two countries is a key strength, these samples are representative of general adult populations and do not include members of the public that are institutionalised (e.g., hospital care, prisons, refugee centres) or difficult to reach (e.g., those not online, the homeless, etc.). The inability to survey these members of society also limits the generalisability of our results. Conclusion Our findings suggest that approximately two-thirds of adults in the general populations of the UK and Ireland had consistently high probabilities of accepting a COVID-19 vaccine during the first nine months of the global pandemic. To achieve wider vaccine coverage, it will be important to reach the 20-25% of people in society who belong to the so-called 'moveable middle'. In both samples, these individuals were more likely to be women, and to have lower levels of trust in scientists, doctors, and other healthcare professionals. Furthermore, contextspecific identifiers were also evident such as younger age, Asian ethnicity, and lower income in the UK, and conspiracy mindedness and external locus of control in Ireland. These findings can be used to aid public health efforts in both countries to reach those in society whose minds can be changed with regards to COVID-19 vaccination.	2021-11-05T05:10:28.126Z	2021-11-03T00:00:00.000	{ "year": 2021, "sha1": "b7f0ef296697f533128115514b6e0a165ae7e9d2", "oa_license": "CCBY", "oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0258871&type=printable", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "b7f0ef296697f533128115514b6e0a165ae7e9d2", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
4635971	pes2o/s2orc	v3-fos-license	Four-state memory based on a giant and non-volatile converse magnetoelectric effect in FeAl/PIN-PMN-PT structure We report a stable, tunable and non-volatile converse magnetoelectric effect (ME) in a new type of FeAl/PIN-PMN-PT heterostructure at room temperature, with a giant electrical modulation of magnetization for which the maximum relative magnetization change (ΔM/M) is up to 66%. The 109° ferroelastic domain switching in the PIN-PMN-PT and coupling with the ferromagnetic (FM) film via uniaxial anisotropy originating from the PIN-PMN-PT (011) surface are the key roles in converse ME effect. We also propose here a new, four-state memory through which it is possible to modify the remanent magnetism state by adjusting the electric field. This work represents a helpful approach to securing electric-writing magnetic-reading with low energy consumption for future high-density information storage applications. Scientific RepoRts \| 6:30002 \| DOI: 10.1038/srep30002 et al. reported a non-volatile, three-state resistance switching technique via electric impulses in converse ME multiferroic heterostructures but reached a maximum relative magnetization change (Δ M/M) of only 4.3% 26 . FeAl is a magnetostriction material 27 , which shows perfect soft ferromagnetism. Moreover, when considering economic reasons and commercialization, low cost FeAl alloys are the most promising FM candidate. By contrast, yielding ultra-high piezoelectric behaviors, single crystals of lead magnesium niobate-lead titanate (PMN-PT) are generally used for FE 15,25 . More recently, due to the higher thermal and electrical stability of the rhombohedral ferroelectric phase compared to binary PMN-PT systems, ternary lead indium niobate (PIN)-PMN-PT has been studied, as it has an approximately 40 °C higher rhombohedral-tetragonal phase transition temperature than that of PMN-PT and maintains the same excellent piezoelectric properties as PMN-PT [28][29][30][31][32] . Therefore, to achieve a giant, stable, tunable and non-volatile ME effect, the combination of a FeAl thin film and a PIN-PMN-PT(011) single crystal is a perfect candidate. In this paper, we report a giant, stable, tunable and non-volatile converse ME effect in a new type of FeAl/ PIN-PMN-PT FM/FE heterostructure at room temperature with electrical magnetization modulation; the maximum relative magnetization change (Δ M/M) of the proposed system is up to 66%, i.e., quite high compared to those reported in similar studies 14,16,26 . We also propose a new four-state memory which can be modified with the remanent magnetism state by adjusting the electric field. We established an illustrative example using the ASCII codes for electric field-based magnetism control in information technology. To our knowledge, this is the first report on controlling multiple states of remanent magnetism in this type of FeAl/PIN-PMN-PT FM/FE heterostructure. Results To construct the FeAl/PIN-PMN-PT heterostructure (Fig. 1a), 10-nm-thick amorphous Fe 81 Al 19 film were deposited on (011)-oriented PIN-PMN-PT single-crystal substrates via molecular beam epitaxy (MBE). Before observing the converse magnetoelectric effect in the FeAl/PIN-PMN-PT structure, the reflection high-energy electron diffraction (RHEED) patterns and atomic force microscopy (AFM) images of the PIN-PMN-PT substrate and Fe 81 Al 19 film after deposition was complete were first examined. Figure 1b shows the RHEED pattern of the (011)-oriented PIN-PMN-PT single-crystal substrate. After growing Fe 81 Al 19 , the streak patterns gradually changed to diffusion; finally, no diffraction pattern was observed upon completion of the deposition, as shown in Fig. 1c. This indicates that the amorphous phases of the Fe 81 Al 19 film are formed on the PIN-PMN-PT substrate. Figure 1d shows an AFM micrograph of the PIN-PMN-PT substrate, with a root-mean-square (rms) roughness of 1.1 nm, which is mostly smooth for deposition. When deposition of the Fe 81 Al 19 film was complete, the surface exhibited a small rms roughness of 0.5 nm, as shown in Fig. 1e, which is perfect for the next stage of the study. To demonstrate the magnetic anisotropy of the FeAl film, angle-remanent curves were measured in different electric fields of 0 kV/cm, 10 kV/cm and 12 kV/cm using a vibrating sample magnetometer (VSM). Before obtaining experimental data, a saturation field of 2000 Oe was applied and then set to zero to achieve a remanence state. Figure 2a shows the angular dependence of remanent magnetization (M r ) in different electric fields; here, an angle of 0° is the starting point in our experiment. The 0 kV/cm curve shows a uniaxial anisotropy of ultrathin FeAl film that originates from epitaxial growth on the PIN-PMN-PT(011) surfaces. At 0° and 90°, the M r reaches its maximum and minimum value, respectively, which indicate the easy magnetizing axis and hard magnetizing axis, respectively, and correspond to the PIN-PMN-PT [100] and [01-1] directions. When applying a dc electric field of 10 kV/cm along the [0-1-1] direction normal to the substrate, an obvious 55° shift in the easy axis direction is observed, which indicates electric field-based control of the magnetism. However, after exceeding 10 kV/cm, almost no changes can be observed, as shown in the 12 kV/cm curve. According to the angle-remanent curves, the magnetic hysteresis loops were measured along the [100] direction at 0° and the [01-1] direction at 90 o , with electric fields of 0 kV/cm and 10 kV/cm using VSM, respectively. The magnetization process of the sample along the [01-1] direction becomes easy and the M r increases when the electric field increases, as shown in Fig. 2b. Meanwhile, the magnetization process along the [100] direction is the reverse, with M r decreasing in the electric field of 10 kV/cm. How these data related to the anisotropic strain of the (011)-oriented PIN-PMN-PT under electric fields will be discussed in the discussion section. To study the origin of the variation of magnetization that is due to the electric field, we measured the polarization and piezoelectric strain of (011)-oriented PIN-PMN-PT single crystals as a function of the electric field using a ferroelectric test system and a strain gauge, respectively. As expected, the strain-electric field (S-E) curves in Fig. 3a show a special loop-like shape, which may originate from the 109° ferroelastic domain switching. To clarify the correlation between the variation of magnetization and electric field directly, the variation of magnetization along the [01-1] direction with the electric field was measured at 8 Oe, at which the converse ME effect is more notable. The numbers marked on the curves in Fig. 3b show the variation sequence of magnetization to the electric field, and yield a loop-like M-E curve that matches the S-E curve shape and the intense magnetization process changes that occur near the coercive electric field of PIN-PMN-PT. Obviously, the M-E curve for our FeAl/PIN-PMN-PT FM-FE structure is distinct from previous reports, which exhibited butterfly-like M-E curves and volatile changes of magnetization [14][15][16][17][18] . Although in Ga 1−x Mn x As/FE and La 0.8 Sr 0.2 MnO 3 /FE structures, researchers found loop-like M-E curves, the M-E effect in these samples occurred under extreme conditions below − 170 °C 33,34 . However, in the FeAl/PIN-PMN-PT FM-FE structure, we obtained a giant, stable, tunable and non-volatile converse ME effect at room temperature. Specifically, in the CoFeB/PMN-PT structure reported in ref. 8., the authors attributed the large electric-field-controlled magnetization to the in-plane magnetic isotropy of the FM layer in addition to the strain mediation of the piezoelectric substrates. However, for our FeAl/ PIN-PMN-PT structure, we grew the FM layer with an obvious uniaxial anisotropy and even obtained a large electric-field-controlled magnetization Δ M/M of up to 66%. Note that the magnetoelectric coefficient, defined as α = μ 0 dM/dE 2 , can be obtained by differentiating the M-E curve in Fig. 3b. The magnetoelectric coefficient α , as a function of the electric field, is shown in Fig. 3c; the maximum value of α is 1.6 × 10 −6 s/m, which is 2 orders of magnitude greater than those from the former reports 2,14,16 . Thus, the tunable converse ME effect reported here is particularly significant in terms of giant and non-volatile magnetization variation. Considering the requirements of applications, M r under different pulsed electric fields was measured. Intermittent positive electric fields of 4 kV/cm, 6 kV/cm and 8 kV/cm and a negative electric field of 8 kV/cm as an "erased electric pulse", were applied to the FeAl/PIN-PMN-PT FM-FE structure. Here, the electric field of − 8 kV/cm also plays the role of the reset voltage. Routinely, a saturation field of 2000 Oe was applied which was then reduced to zero. Figure 4a shows the electric field-induced non-volatile magnetization switching under the various pulsed electric fields along the [01-1] direction, with a giant electrical modulation of magnetization for which the maximum relative magnetization change (Δ M/M) is up to 66%, which is much larger than those found in the previous reports 14, 16,26 . Therefore, four notable and stable magnetization states can be achieved by simply modulating the electric field. The magnetization under the various pulsed electric fields, labeled as "0", "1", "2", and "3", exhibit obvious multistate properties. In the strategy of utilizing electric-writing magnetic-reading, one of the key components is the converse ME effect device, which is used to convert the electric signals into magnetic signals. For instance, the four items of two-digit information of "00", "01", "10", and "11" can be demodulated and stored from remanent magnetism with the corresponding M r /M s device as 0.15, 0.3, 0.4, and 0.5 in response to the electric fields of − 8 kV/cm (E D ), 4 kV/cm (Ec), 6 kV/cm (E B ) and 8 kV/cm (E A ), respectively, as shown in Fig. 4b. Moreover, according to the ASCII codes and the standard eight-bit codes of "01001100", the letter "L" can be demodulated and stored in the FeAl/PIN-PMN-PT FM-FE structure. This non-volatile multistate memory through electric-writing magnetic-reading can be written efficiently with lower energy consumption and can store information with higher density compared to the traditional memory methods. The other letters can also be demodulated in the same way. Therefore, the word "LZUV", which is an abbreviation for Lanzhou University, can be demodulated from the electric field and stored in the FeAl/PIN-PMN-PT FM-FE structure, as shown in Fig. 4c. Discussion To illustrate the strain-mediated ME effect, we propose the model shown in Fig. 5. In this work, (011)-oriented PIN-PMN-PT single-crystal substrates are used for FE and the spontaneous polarizations of PIN-PMN-PT with rhombohedral phase are along the < 111> directions (Fig. 5a). Upon being exposed to an external positive electric field along the [0-1-1] direction normal to the substrate, all of the eight possible polarization directions are switched downward, except p2 − and p3 − (Fig. 5a). Larger external electric fields further switch the polarizations p2 − /p3 − toward the [0-1-1] direction, inducing a tensile stress strain along the [01-1] direction and a compressive strain along the [100] direction (Fig. 5b). This causes the magnetization process of the sample along the [01-1] direction to become easy and along the [100] direction to become hard. As shown in Fig. 5a, polarization changes with p1 + /p4 − to p2 − or p4 + /p1 − to p3 − are related to 109° ferroelastic domain switching, which can remain and result in loop-like behavior for the strain-electric field curve and is the origin of the non-volatile converse ME effect 12 . When an external negative electric field is applied to this heterostructure, once beyond the coercive electric field of PIN-PMN-PT, all of the polarizations return to the initial state and the heterostructure recovers its original condition. This model can explain the loop-like behavior of the S-E curve and the origin of the nonvolatility. We should note that the results cannot be explained by the terms related to the electric-field induced charge accumulation at the interface, because its contribution is too small to account for the achieved large modulation in magnetization which should come from the entire film 33,35 . To further discuss the correlations between polarization states and different electric fields in (011)-oriented PIN-PMN-PT, we investigated the substrates with piezoresponse force microscopy (PFM). In non-poled PIN-PMN-PT, the polarization vectors appeared randomly along the eight body diagonals as shown in Fig. 6a. This created the PFM phase images out-of-plane and in-plane shown in Fig. 6c,d, respectively, with the cantilever along the [100] direction. In out-of-plane measurement, the tip was sensitive to the vertical piezo response and the PFM phase images had two kinds of color contrast. while in-plane measurements, the tip was sensitive to component polarization perpendicular to the cantilever besides out-of-plane polarization and, so there were three kinds of color contrast in the PFM phase images. After applying a voltage of + 10 V on the tip along the [0-1-1] direction normal to the substrate (green box) with a 5 μ m by 5 μ m square (Fig. 6e), polarization switching with a two-step switching sequence composed of a 71° switch (e.g., P2 + changes to P1 + ) followed by a 109° (e.g., P1 + changes to P2 − ) switch took place and finally changed to P2 − and P3 − , thus the color of the out-of-plane image became black and the color of the in-plane image bacame brown. Accordingly, the in-plane polarization had no component along the [0-11] direction and thus there was no contrast within the green box in Fig. 6f. When a voltage of − 10 V was applied along the [0-1-1] direction normal to the substrate (red box) within a 3 μ m by 3 μ m square, the polarization directions were switched conversely upward apart from p2 + and p3 + , thus the out-of-plane image became white (Fig. 6e). The in-plane polarization again had no component along the [0-11] directions, so there is no contrast changes within the red box as shown in Fig. 6f. The stable and reversible in-plane and out-of-plane polarization states, including 109° ferroelastic domain switching, enable non-volatile converse ME effect. Finally, we calculated the angular dependence of M r to understand the experimental shift of anisotropy. The total energy E t is written according to the Stoner-Wohlfarth model 36 where E k is the magnetic anisotropy energy, E σ is the stress energy and E h is the Zeeman energy. Here, we define E u as the total anisotropy energy, which consists of E k and E σ . Figure 7a shows the coordinate system for the analysis; θ 0 is the angle between the directions of the applied magnetic field H and the as-deposited anisotropy field H k ; ϕ is the angle between the directions of the total effective field H t after applying a dc electric field and H k ; θ is the angle between the directions of H t and M s . In the actual measurement of VSM, H k is not always along the x-axis and θ is always 0. In this case, the total energy can be expressed as where K t is the in-plane total effective uniaxial anisotropy constant. M r , achieved in the angle-remanent curves, satisfies The equilibrium conditions can be written as Because M r is always positive and a differential constant C, we can obtain Based on formula (5), fitted curves of M r vs θ 0 are shown in Fig. 7b, which agree well with the experimental data. When applying an electric field, the easy axis is driven to a new stable direction due to the total effective field H t . Namely, the angle ϕ, fitted from Fig. 2a, in formula (5) determines the easy magnetizing direction under various electric fields. This indicates that the total effective anisotropy field H t has switched by nearly 55 o from the as-deposited H k after applying the 10 kV/cm electric field; the calculation result agrees well with the experimental shift in anisotropy. In summary, a stable, tunable and non-volatile converse ME effect is obtained in a new type of FeAl/ PIN-PMN-PT FM/FE heterostructure at room temperature with a giant electrical modulation of magnetization. Test results show that the 109 o ferroelastic domain switching in PIN-PMN-PT and coupling with FM film with uniaxial anisotropy originated from the surface of PIN-PMN-PT (011) were the key factors in the converse ME effect observed. Owing to the giant electrical modulation of magnetization, four-state memory through electric-writing magnetic-reading can meet the application requirements of high-density information storage, which allows demodulation from the electric field and can be stored in the FeAl/PIN-PMN-PT FM-FE structure. This work is helpful for exploring the basic principles of electric-field modulation of magnetism and applications, especially electric-writing magnetic-reading, to achieve low energy consumption for future high-density information storage. Methods 10-nm-thick amorphous Fe 81 Al 19 film were first grown at RT on PIN-PMN-PT(011) substrates using MBE equipped with solid-source effusion cells for Fe and Al, at a rate of 0.13 nm/min and 0.33 nm/min for Fe and Al respectively. Then, a capping layer of 3 nm of Al was grown on the top of the sample to prevent the FeAl film from oxidizing and also it could be used as a top electrode. Prior to growth, the substrate was chemically cleaned using trichloromethane, acetone, and methanol. After the substrate was loaded into the growth chamber, the substrate was further thermally cleansed at an annealing temperature of 100 °C. Nucleation and growth were monitored in situ using RHEED. Pt layers were finally sputtered as the bottom electrode via magnetron sputtering at room temperature. The surface morphologies of substrates and samples were examined using AFM. In the magnetoelectric coupling test, Cu varnished wires were connected to the electrode by adhesive tape and the electric field applied between the top and bottom electrode was controlled by the dc power supply Keithley 6517B. The polarization and piezoelectric strain of (011)-oriented PIN-PMN-PT single crystals, as a function of the electric field, were measured using a ferroelectric test system (Radiant, Precision Premier II, USA) and a strain gauge (KFG-1-120-D17, Kyowa), respectively. The ferroelastic domain switching in the PIN-PMN-PT under electric field was observed using PFM.	2018-04-03T03:43:21.178Z	2016-07-15T00:00:00.000	{ "year": 2016, "sha1": "6cbb15322665375ff7085ae4c4f04b34c79c6592", "oa_license": "CCBY", "oa_url": "https://doi.org/10.1038/srep30002", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "6cbb15322665375ff7085ae4c4f04b34c79c6592", "s2fieldsofstudy": [ "Materials Science", "Physics" ], "extfieldsofstudy": [ "Materials Science", "Medicine" ] }
20874237	pes2o/s2orc	v3-fos-license	Atypical Subacute Recurrence of Catastrophic Antiphospholipid Syndrome in a Japanese Female Patient. Catastrophic antiphospholipid syndrome (CAPS) survivors rarely relapse. We herein report a case of a second CAPS episode with an unusual subacute course and no microangiopathic hemolytic anemia (MAHA), a common CAPS symptom. During the first episode, the 69-year-old woman responded well to high-dose glucocorticoids and plasma exchange. On relapse, these treatments plus rituximab were ineffective and she died of multi-organ failure and bacterial cholangitis. The absence of MAHA and a subacute course do not exclude a CAPS recurrence. Introduction Catastrophic antiphospholipid syndrome (CAPS), a rare subtype of antiphospholipid syndrome (APS), is characterized by multiple organ dysfunctions due to thrombosis of the small vessels occurring over a short period of time, typically less than 1 week. The mortality of CAPS is about 30% even when treated with high-dose glucocorticoids (GCs) and plasma exchange (PE) (1,2), but patients who survive the initial event rarely have recurrent events (3). We herein report a case of CAPS relapse in a Japanese woman with an atypical slow clinical course and the absence of microangiopathic hemolytic anemia (MAHA). Case Report In July 2013, a 69-year-old Japanese woman with a 7year history of recurrent cerebral infarction due to APS, presented with symptoms of fever at a local hospital. She was diagnosed with aspiration pneumonia and successfully treated with antibiotics. She also exhibited dysphasia due to a cerebral infarction, and was transferred to our hospital at the behest of her family for rehabilitation. Three years prior to this admission, she was referred to our hospital because of fever and a disturbance of consciousness, and was subsequently diagnosed with CAPS based on the presence of multiple organ impairment, including the brain, liver and kidney, and the presence of an antiphospholipid antibody, which had been detected previously. The patient had hemolytic anemia with peripheral schistocytes, thus suggesting MAHA. She was successfully treated by anticoagulation, high-dose GC, and PE, and has since been on 6-mg prednisolone (PSL) maintenance therapy. At the current admission, the patient was afebrile and well oriented. She had been bed-bound for years and needed help with her daily activities. A physical examination revealed a systolic ejection murmur in the apex and mild to moderate muscle weakness of the extremities. She did not have any skin manifestations. The complete blood count, aspartate transaminase (AST), alanine transaminase (ALT), creatinine levels, and a urinalysis including urine sediment were normal, except for slight anemia. The activated partial thromboplastin time was prolonged (79.5 s; normal range <39.0 s) as was the prothrombin time due to anticoagulation with warfarin (international normalization ratio 1.69). The immunological tests were positive for β2 glycoprotein 1 dependent anticardiolipin antibody (>125 U/mL; normal range <3.4 U/mL), whereas the anti-nuclear antibody titer was low (1:40, homogenous/speckled pattern) and the anti-dsDNA was negative. Chest radiograph showed a slight infiltration in the right lower lung field, which was attributed to the preceding pneumonia. A central venous catheter was inserted into the right femoral vein for parenteral nutrition. Three days after the catheter insertion, however, the patient developed fever, anemia, and thrombocytopenia. Computed tomography (CT) revealed diffuse lung infiltration and venous thrombosis at the insertion site. The catheter was removed and anticoagulation with unfractionated heparin and antibiotics was started. Over the next 4 weeks, the patient developed a diffuse gastric and bladder mucosal hemorrhage as revealed by gastroscopy and cystoscopy, the elevation of AST and ALT, proteinuria, hypertension, and the elevation of the creatinine level (0.87 mg/dL on admission, 1.18 mg/dL on hospital day 38). A fever was persistent and refractory to antibiotics, including meropenem, cefepime, and clindamycin. The procalcitonin levels were constantly positive (1.49 ng/mL on hospital day 23, 1.07 ng/mL on hospital day 50). Three weeks after hospitalization, the patient gradually developed palpable purpura in the bilateral palms (Fig. 1a). A skin biopsy revealed multiple fibrin thrombi in dermal small vessels (Fig. 1b). On the basis of multiple organ dysfunction and microthrombosis in small vessels, a relapse of CAPS was diagnosed on hospital day 38. The laboratory data on admission and on hospital day 37 are summarized in Table. We started methylprednisolone (mPSL) at a pulse dose of 1,000 mg/day for 3 days followed by 40 mg daily, which resulted in a dramatic improvement of fever, lung infiltration, partial improvement of thrombocytopenia, and liver enzyme elevation. One month later, however, the patient again exhibited fever, thrombocytopenia, and liver enzyme elevation. Intravenous immunoglobulin (IVIg) at 400 mg/kg/day for 4 days had no effect. Plasma exchange every other day was effective during treatment only, and symptoms and laboratory parameters worsened during the 2-day interval. Renal dysfunction and proteinuria were refractory to these treatments. As the patient had primary CAPS (CAPS without systemic lupus erythematosus), we decided to administer rituximab as further treatment rather than cyclophosphamide. However, 4 administrations of rituximab at 375 mg/m 2 had no effect. Despite increasing mPSL to 60 mg, she developed petechiae in the lower palpebral conjunctiva and on the toes around hospital day 110 (Fig. 2). The patient died on hospital day 120. The entire treatment course is summarized in Fig. 3. An autopsy revealed the presence of multiple thrombi in small vessels in multiple organs. In the liver, organized thrombi were detected in the interlobular portal veins and arteries (Fig. 4a). Organized thrombi were also detected in the arterioles of the gastric mucosa and pancreas. A kidney specimen showed a thickened and swollen intima of the afferent arterioles and fibrin thrombi in the glomerular tufts, compatible with renal thrombotic microangiopathy (Fig. 4b, c). Fibrin thrombi were detected in the small arteries of the left ventricle and Libman-Sacks vegetation was seen on the tricuspid valve. In addition to these findings which are consistent with APS, the common bile duct was obstructed by stones. Neutrophil and histiocyte infiltration, as well as bacterial colonies, were seen at the site of the bile duct obstruction, which is consistent with bacterial cholangitis. Discussion The subacute clinical course and lack of MAHA in the present case were atypical for CAPS. The major differential diagnoses were infective endocarditis, thrombotic thrombocytopenic purpura (TTP), and heparin-induced thrombocytopenia (HIT). Blood cultures were taken at least once every week and the findings were negative. Transthoracic echocardiography showed no vegetation. Fragmented red blood cells were not detected in blood smears, which is atypical for TTP. Changing the anticoagulant first to dalteparin and then to fondaparinux, which are less likely to cause HIT, did not improve the patient's status or thrombocytopenia. She was refractory to intensive immunosuppressive therapy, including rituximab. Multiple organ failure gradually worsened her general status and she eventually died. The autopsy findings indicated that the cause of death was bacterial cholangitis and sepsis due to a common bile duct obstruction. The obstruction likely occurred a few days prior to death because it was not evident on the abdominal enhanced CT 1 month prior and the total bilirubin level had been <2 mg/dL. Moreover, a blood culture taken just 2 days before death was negative. There are few reports on the prognosis of CAPS patients who survive the initial event. Erkan et al. reported that none of the 58 CAPS patients who survived the initial event developed further catastrophic events during an average follow-up of 67.2 months (3). Moreover, since the first CAPS relapse case was reported in 1999 (4), only 10 additional cases have been documented (5)(6)(7)(8)(9)). An international registry of patients with CAPS was created in 2000 (2, 10-12). Espinosa et al. reviewed 9 cases from the CAPS registry (9). With the exception of 1 case in which detailed data were not available, 13 of 18 catastrophic episodes in the 8 cases were complicated by MAHA and the presence of schistocytes in the peripheral blood smear. Furthermore, in 2 relapsed cases reported in 2012 and 2014, all 5 episodes (2 in one and 3 in the other) were complicated by MAHA (7,8). Thus, MAHA has been detected in 18 of 23 episodes (78%) in the 10 reported relapsing CAPS patients. In our case, however, schistocytes were not detected in the peripheral smear and MAHA was not clinically diagnosed, although it was present during the first event 3 years prior. The absence of MAHA, therefore, does not seem to preclude a CAPS recurrence. Another unusual aspect of the present case was that our patient developed a relapse with an atypical subacute course. She gradually developed fever, lung infiltration, thrombocy-topenia, anemia, mucosal bleeding in the stomach and bladder, elevation of liver enzymes, acute kidney injury with hypertension and proteinuria, and palpable purpura over 1 month, while preliminary classification criteria (1,13) state that manifestations of CAPS usually develop simultaneously or in <1 week. Indeed, data from the CAPS registry reveal that 175 of 176 patients (99%) fulfill this acute course criterion (1). A subacute course even over 1 month does not seem to preclude a second catastrophic event in APS patients if other possible diseases are excluded. Treatment is always challenging if a CAPS patient is refractory to anticoagulation and corticosteroids (first-line therapies) and to the addition of plasma exchange and/or IVIg (second-line therapies) (12,14). Although rituximab is recommended rather than cyclophosphamide for the treatment of refractory CAPS without systemic lupus erythematous (12,15), our patient was also refractory to rituximab and died. Eculizumab is another agent reported to be effective for CAPS treatment (7), but the cost is high and it is approved only for paroxysmal nocturnal hemoglobinuria. The administration of glucocorticoids and the plasma exchange were at least partially effective for fever, lung involvement, and liver enzyme elevation, so we speculate that anti-cytokine therapy such as anti-tumor necrosis alpha or interleukin-6 might have been effective for our patient, although there have been no reports using these agents for CAPS. In conclusion, the lack of MAHA and a subacute course over a month are therefore insufficient to exclude a subsequent catastrophic event in APS or CAPS patients. The most effective third-line therapy for CAPS is still unresolved. The authors state that they have no Conflict of Interest (COI).	2018-04-03T06:14:42.391Z	2015-01-01T00:00:00.000	{ "year": 2015, "sha1": "396f517e88132568a682ab5ac95ee52d7ebbc01c", "oa_license": null, "oa_url": "https://www.jstage.jst.go.jp/article/internalmedicine/54/22/54_54.5150/_pdf", "oa_status": "GOLD", "pdf_src": "Adhoc", "pdf_hash": "c326f465f43bfb9a9e5c1d8ce78a2115520c838a", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
226538235	pes2o/s2orc	v3-fos-license	Postural Control of People in treatment for substance use disorder - Aims: One of the common sequelae of chronic abuse of alcohol and/or illicit drugs is the impairment of body balance control, caused by long-term neurological damage. This study aimed to investigate the postural control of individuals hospitalized for the treatment of substance use disorder (initial phase) and to compare the results obtained by a control group. Method: For this, One-hundred fourteen individuals hospitalized for the treatment of substance use disorder and eighty-eight healthy controls, all males, were analyzed. Body mass, height, and waist circumference were measured. Were performed a balance test in a static upright position with feet side by side using a plantar pressure platform. Results: Individuals with substance use disorder have shown significantly lower results on body balance compared with controls. Individuals with chemical dependence showed balance results open-eyes, similar/lower than control subjects with close-eyes. Still, those who used only alcohol or alcohol combined with other illicit drugs presented worse results. Finally, impairments in body balance variables showed a significant correlation with age and substance time of use. Conclusion: Even in the early stage of treatment, substance use disorder considerably impairs the balance of the investigated men. The worst results were found in alcohol-dependents (alcohol alone or alcohol combined with other drugs). Introduction Chronic and excessive consumption of alcohol and illicit drugs can impair several central nervous system functions, including cognitive and motor aspects of individuals who misuse these substances. The acute effect of alcohol and other drugs on postural control is well established in the literature 1,2 . Deficits in postural control and balance, as well as adverse sedative effects, such as impaired cognitive functioning and muscle relaxation, are reversible after cessation (withdrawal) of the abuse of these psychotropic drugs 1,3,4 . According to Carlini et al. 5 , data from the last household survey on the use of psychotropic drugs in Brazil showed that 22.8% of the population surveyed has already used illicit drugs. Lifetime marijuana use appears first among illicit drugs, with 8.8% of respondents, and cocaine use was reported by 2.9% of the individuals interviewed. Alcohol use was reported by 74.6% of the interviewees, and the prevalence of individuals classified as a dependent was 12.3%. The treatment procedures for substance use disorders (SUD) involve psychological and/or pharmacological support and an extended follow-up period 4 . Psychological approaches used in the treatment of alcohol and drug dependence may include mutual help groups, group therapy, and psychotherapy, as well as short-or long-term hospitalization in specialized institutions for the treatment of chemical dependence 6 . Chemical dependence can lead to several impairments to the individual's mobility, including impairments in their postural control 7,8 . Postural control can be defined as the act of maintaining, reaching, or restoring a state of balance during any posture or activity. Adequate postural control is essential for daily activities and requires the integration of visual, vestibular, and proprioceptive information 9,10 . Individuals with chronic SUD with varying periods of abstinence presented lower levels of balance and coordination than control groups in different sensory conditions analyzed, both in static and dynamic situations 2,10,11 . Few studies have been conducted to investigate postural control in individuals initiating the recovery process and the harm caused by the long-term interaction of alcohol and other illicit drug abuse 12 . The main cause of this scarcity is the poor adherence iD iD iD DOI: http://dx.doi.org/10.1590/S1980-6574202000020010 to treatment and subsequent maintenance of the recovery process, making it difficult to follow the patient for long periods and gain a better understanding of the mechanisms involved in the neural plasticity that occurs during the sobriety process after years of excessive alcohol and other drug consumption 7 . Most studies on body balance in individuals with SUD have been conducted with individuals with alcoholism, leaving a gap on the effects of other drugs on this variable. Alcohol dependence has been associated with a decrease in postural control in more than 60% of abstinent patients, suggesting that this deficiency is directly related to alcohol consumption throughout life 2 . Thus, it can be assumed that the negative impacts on postural control in these patients may have been caused, mainly, by the toxic effect related to the dose of alcohol consumed, that is, the greater the amount and time of exposure to addiction, the greater the losses in postural control. The study conducted by Schmidt et al. 13 using dynamic posturography also concluded that alcohol consumption has a deleterious effect on the body balance of abstinent alcoholic individuals and that these individuals present significant alterations in postural balance when compared to non-alcoholics. Moreira et al. 14 in their study using posturography, highlight that time of use of illicit drugs demonstrated a significant and positive correlation with oscillation velocity in the closed eyes condition. Thus, the longer the use of illicit drugs, the greater the individual's difficulty in maintaining body balance when visual and proprioceptive information was absent or distorted. Some variables that intervene in SUD, such as subjects' age, time of consumption, and type of substance consumed (alcohol or other illicit drugs) may interfere with postural control. In addition, deficits in postural control mechanisms in individuals who are dependent on different drugs are not well known. Thus, this study aims to investigate postural control of men in treatment for SUD (initial phase) and to compare these results with those of healthy control subjects. In addition, individuals who use alcohol and other illicit drugs will also be compared. Sample The study was approved by the local Ethics Committee (number 2.125.747/2017). We evaluated one-hundred fourteeen adult males diagnosed with Substance Use Disorders (SUD), aged from 18-59 years (34.9 ± 12.2 years), hospitalized for treatment in a private clinic in south Brazil. The clinic had a multidisciplinary approach to the treatment of chemical dependency, offering non-pharmacological and pharmacological therapy when necessary. Only individuals who were in the initial phase of treatment (after the first two weeks of hospitalization) and who presented at least one year of dependence diagnosis were included in the sample. Since withdrawal symptoms may occur in the first two weeks, the subjects were evaluated on the 15th day of hospitalization. Exclusion criteria were having a motor or cognitive impairment that hamper participation in motor tests and understanding of assessment tools (e.g. excessive tremble or Mini-Mental State Examination < 21). Individuals with acute withdrawal symptoms were excluded from the study. The institution physician consulted all the patients before the test sessions and only those considered fit to participate in the research were included. In the control group (CG), eighty-eight healthy adult male participants with a mean age of 35.3 ± 12 years and without a chemical dependence diagnosis were evaluated, selected from the local university community. These individuals were in the same age group and presented similar anthropometric characteristics to the clinical sample. Individuals were excluded if they reported the use of illicit psychotropic substances (illicit drugs) and alcohol abuse -4 (four) or more times a week, or consumption of 5 (five) or more doses on a single occasion during the 30 days preceding the study 15 . A sample size of 82 individuals for each group was estimated by the G * Power 3.1.9 software for power (1 − β) of 80%, a one-tailed significance level (α) of .05, and high effect size (d = .8). All the participants signed an Informed Consent, agreeing to participate in the research Procedure Initially, all the individuals answered an anamnesis with information about age, schooling, type of substance consumed, time of use/abuse, medication use, and regular participation in a physical exercise program in the month preceding hospitalization. Measurements of anthropometric variables, weight, height, and waist circumference were collected, and BMI was calculated. Analysis of postural control was conducted through a balance test in a static upright position with feet side by side using the plantar pressure platform myoPressure Noraxon®. During the test individuals remained in the bipodal position for 30 seconds in two conditions -eyes open and eyes closed -and data were recorded, stored, and then analyzed using Noraxon® myoRE-SEARCH 3.10 software. To evaluate the body balance of the participants, COP (center of pressure) path length (the distance traveled by the COP in millimeters during the analyzed time) and the ellipse area (the area covered by the center of pressure in the mediolateral and anteroposterior axes) were verified. The lower the values recorded, the better the balance. Statistics Descriptive statistics were calculated, one-way ANOVA was used to compare body balance variables between groups, and a generalized linear model (GLM) was used to adjust differences according to age and physical exercise practice in the month preceding hospitalization (YES or NO). Effect size between groups (Cohen's d) was calculated using the equation: Cohen's d = Mean1 -Mean2/Standard Deviation pooled . The effect size was interpreted as follows: d<0.20 very small, d= 0.20-0.59 small, d= 0.60-1.19 medium, d= 1.20-1.99 large, d≥ 2.00 very large, according to the recommendations of Hopkins et al. 16 . Pearson's correlation test was applied to verify the relationships between body balance variables, age, and time of substance use. Significance was set at p< 0.05. Results Among participants with SUD, 31 subjects (27.2%) reported regular use of alcohol, 21 (18.4%) use of cocaine/crack, and 62 (54.4%) reported polysubstance use (alcohol + illicit drugs). The mean time of consumption of the substance was 13.97 ±1.13 years. In this variable, the group of alcohol dependents presented significantly longer consumption time (22.87 ±14.4 years) than cocaine/crack (10.14 ±7.3) or polysubstance (10.82 ±7.4) users. Despite the great variability in the data, 68 subjects from the SUD group (59.9%) stated that they had been using the substance for at least 10 years. Regarding the level of schooling, 62 subjects (54.4%) with SUD reported elementary education, while only 14 (12.3%) reported higher education. When questioned about physical exercise, 42 subjects (47.7%) from the CG and 25 (21.9%) from the SUD group stated that they had participated in a regular and supervised program in the previous month. There were no significant differences between age and BMI values of the participants from both groups (p> .05). Among individuals with SUD, alcohol users had a higher average age than drug or alcohol + drug (polysubstance) users. Participants' descriptive data are summarized in table 1. It was verified that the group of individuals with alcohol-dependence was significantly older when compared with cocaine/crack and polysubstance users (F = 22.51; p< .001). No differences were observed in BMI. Comparisons between groups (SUD and CG) related to the variables of body balance are summarized in Figure 1. When the SUD group was analyzed according to the type of substance consumed, different behaviors were observed in the balance variables, as shown in Figures 3a and 3b. In both open and closed eyes conditions, only the alcohol and polysubstance groups had higher values than the CG for COP Path and Ellipse. In addition, the alcohol group presented higher values than the cocaine/crack and polysubstance groups, except for the Ellipse in the closed eyes condition. Finally, significant correlations were observed in the open eyes situation between COP Path and age (r =.19; p = .045), ellipse area and age (r =.19; p = .040), COP Path and time of substance consumption (r =.34; p =.000), and ellipse area and time of substance consumption (r = .25; p =.008). Discussion The study aimed to compare data related to body balance between individuals in treatment for SUD and control individuals with no history of substance abuse. In the descriptive analysis of the participants with SUD, it was possible to verify a high prevalence of individuals dependent on polysubstance, with at least 10 years of consumption, and a low level of education. According to the World Health Organization 17 , the onset of dependence is early, which usually leads to dropping out of school at younger ages, helping to explain the low level of schooling found 12 . Another point observed was the low prevalence of individuals with SUD who had practiced regular physical exercise in the month preceding hospitalization when compared to the CG. The data should be interpreted with caution since some sequelae of chronic SUD, such as weakness, and motor coordination and balance problems can reduce participation in physical exercise programs [18][19][20] . In addition, the practice or not of physical exercise in the present study was only reported by the individual. Despite this, other studies in the literature have demonstrated a higher prevalence of sedentary behaviors among individuals with SUD, especially among those with a longer time of use 21,22 . When both groups were compared, the CG presented higher results in the COP Path and Ellipse area, denoting worse body balance control. These worse results of the SUD group were seen in both open and closed eye conditions. In addition, as highlighted in figure 1, the SUD group with eyes open presented results similar to those of the CG with eyes closed for the variable COP Path. It is known that the consumption of alcohol and/or illicit drugs can negatively affect body balance 8,12,14 , which may be due to the increase in the occurrence of tremors and excessive body oscillation 1,23,24 , damage to structures of the central and peripheral nervous system 13 , or, in the specific case of alcohol consumption, cerebellar damage characterized by degeneration and loss of volume of Purkinje cells 25 . Although the practice of physical exercise was considered as a covariant for the comparison between groups, it is known that individuals with SUD, in addition to neurological damage, may present low levels of muscular strength when compared to people without dependence, a factor which may also negatively influence the results 7,19 . When comparing the users by the type of illicit drug consumed, it was verified that those with alcohol dependence were significantly older and had a long time of dependence. This fact may mean that alcohol dependents take longer to seek treatment since alcohol is a drug considered "acceptable" and licit by society 24 . Likewise, the subgroup of alcohol-dependent individuals, when compared to cocaine/crack or polysubstance (alcohol + illicit drugs) dependents, presented greater impairments in body balance. Furthermore, the polysubstance dependent group demonstrated significantly worse values in the body balance variables when compared to the CG, which was not observed for the cocaine/crack group. These differences reinforce the potential impairment that abusive alcohol consumption can bring to the control of body balance. It should be noted, however, that the alcohol-dependent group presented a significantly higher mean age and substance consumption time than the other groups. The results found in the present study diverge somewhat from those seen by Fein, Smith, and Greenstein 12 , who compared alcohol-dependent, alcohol + illicit drug dependent, and control individuals and showed that the alcohol + illicit drug group presented the worst results in body balance. However, in this study, individuals in the alcohol + drug group also presented longer substance use time, which may have interfered with the results. Another study by Moreira, Ganança, and Caovila 14 comparing alcohol-dependent, illicit drug-dependent, and control individuals found that the first two groups had worse results in body balance. However, as the specific types of drug consumed, as well as the consumption time, was not specified in this study, comparisons with our results are hampered. It is known that the consumption time of the substance is a factor that negatively affects balance control 26 , especially in the case of alcohol. Wöber et al. 2 analyzed 82 alcohol-dependent individuals who had abstained less than one month previously and found that they demonstrated worse results in balance tests when compared to controls. The authors also found a positive relationship between the damage to balance and the time of alcohol consumption, reinforcing that chronic dependence can cause severe impairment over the years. In the specific case of the present study, a positive relationship was also observed between impaired balance, age, and duration of substance use, indicating that the older the individual and the longer the consumption time, the higher the values measured in the COP path and ellipse area. It would be possible to speculate that alcohol-dependent individuals presented the worst results only because they were older, however, the polysubstance group (alcohol + illicit drugs) also presented lower values in the body balance variables when compared to the control, despite having a lower mean age. In this way, it seems that the use of alcohol, combined or not with other illicit drugs, is an independent factor for damage to body balance and that this impairment is greater the longer the consumption time of the substance. Regarding the withdrawal time, in our study this variable was not analyzed since all the participants had been in rehabilitation for two weeks, that is, they had a very short period of abstinence. Nevertheless, it is known that prolonged periods of abstinence can reverse some of the damage caused by abusive consumption of substances to body balance. Fein and Greenstein 10 followed individuals with chronic alcoholism who were abstainers for 6 weeks and one year, comparing their body balance to a CG. However, even after one year of abstinence, the authors did not find improvements in body balance, noting that it would probably take longer without substance use to observe any improvement. Rosenbloon et al. 3 , following the body balance of individuals with alcoholism over two years of abstinence, observed that after this period there were significant improvements in the analyzed variables. These findings indicate that long periods without the use of the substance are necessary so that beneficial effects can be verified in body balance. Our study presents some limitations, such as the unequal distribution in the subgroups by type of substance consumed, as well as the type of instruments applied to analyze body balance, Postural control and substance use disorder. which evaluates only the static upright position and to assess the physical exercise practice. Nevertheless, it is possible to highlight that the verified data can provide relevant subsidies for professionals working in health services that provide care for individuals with SUD, to emphasize the importance of intervention programs aimed at improving the balance of these individuals, supporting their recovery process and reducing the risk of motor impairments. Conclusions From the data analyzed, it is observed that the balance of individuals with SUD was significantly impaired when compared to control individuals. In addition, those who used only alcohol or alcohol combined with other illicit drugs presented worse results compared to cocaine/crack users. Finally, impairments in body balance variables presented a significant correlation with age and duration of substance use. In this way, the importance of interventions as early as possible is reinforced, in order to avoid prolonged use of substances, in addition to intervention programs aimed at improving or maintaining body balance. The increasing number of people in SUD situations in Brazil imposes an imminent need for more in-depth information about their health status so that professionals from different areas can carry out interventions in a specific and grounded way. Thus, the importance of new studies with this population is evidenced, with interventions involving protocols of physical exercises, analyses of other variables such as muscle strength, as well as larger and more homogeneous samples, in order to obtain reliable data about the balance of these individuals.	2020-09-03T09:14:37.367Z	2020-01-01T00:00:00.000	{ "year": 2020, "sha1": "7fea206ace659d386aa3eb660bf9ba4878cff8ec", "oa_license": "CCBY", "oa_url": "http://www.scielo.br/j/motriz/a/mqnBjqCqVt5WCtsmgMQTqvx/?format=pdf&lang=en", "oa_status": "GOLD", "pdf_src": "Adhoc", "pdf_hash": "8563955860b593f30d2b62f7128a1cf0e4d1175d", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
245862338	pes2o/s2orc	v3-fos-license	How Far Is Far Enough? The Social Constitution of Geothermal Energy through Spacing Regulations : This article argues that the sociotechnical context in which near-surface geothermal energy is embedded draws out its characteristic of being temporarily depletable. Thereby, the minimization of unavoidable side effects, such as cold plumes, which result from the social constitution of geothermal energy, is a crucial area of consideration. Using the situation in Germany as a touchstone, we discuss how cold plumes and interferences from neighboring ground source heat pumps test the limits of the existing regulatory framework, requiring negotiations between different knowledge sets stemming from areas as diverse as planning law, geology, cultural habits, and engineering. This makes the operation of geothermal energy highly uncertain and continuous negotiations on sustainable modes of extractions a pressing issue. Introduction: The Transition to Low Carbon Heating in Germany Growing awareness of the negative consequences of fossil fuel-based energy production has led Germany to plan taking all coal-fired power generation off-grid by no later than 2038 and to meet its energy demand with mainly non-fossil energy sources. At the same time, the country aims to reduce energy consumption and increase energy efficiency [1]. The Federal Ministry for Economic Affairs and Energy ("Bundesministerium für Wirtschaft und Energie"-BMWi) sees great potential both in reducing energy consumption and in switching to non-fossil alternatives in the heating sector. Thus, the Renewable Energies Heat Act ("Erneuerbare-Energien-Wärmegesetz"-EEWärmeG), which came into force on 1 January 2009, obliges citizens and companies in Germany to incorporate renewable energy sources, such as solar thermal, biomass, or geothermal energy, into the heating supply for new buildings if their floor space exceeds 50 m 2 [2,3]. The use of renewable energy for heating and cooling purposes is also subsidized by the federal government under the Market Incentive Program (MAP) in accordance with §13 of the Renewable Energies Heat Act [3]. This program funds technologies that use renewable energy for heating or cooling, including heat pumps (air source heat pumps, ground source heat pumps, and exhaust air heat pumps), as well as deep geothermal systems and heat storage systems. These government incentives have contributed to the rising popularity of near-surface geothermal energy, and by 2020 there were over 440,000 ground source heat pumps (GSHPs) installed in Germany [4]. In the same year, heat pumps (both ground and air source) made up about 10 percent of renewable heating (17.5 billion kWh). As Figure 1 illustrates, the installed capacity of GSHPs considerably varies between German federal states. This is not only caused by differences in geological conditions or population sizes, but also by varying regulations and incentive programs [5]. Heating transitions are currently taking place in many industrialized countries all over the world and they can be understood as transformations of social, technological, and material configurations in the heating sector. As outlined above, the political goal in Germany, as well as elsewhere, is to transform the sociotechnical system of heating in such a way that it becomes less CO2-intensive. Thus, heating systems which rely on fossil fuels need to be phased out, while renewable heating systems, such as heat pumps, are expected to spread throughout society. However, the integration of new elements into a larger sociotechnical system almost always goes hand-in-hand with unforeseen and unintended side effects [6]. Unexpected issues come to the fore, brought about by the interplay of social, technological, and material elements of the respective sociotechnical systems. With regard to geothermal energy, one such unexpected issue is the technology's temporary depletability when it is used in specific ways. This temporary depletability becomes observable in the form of so-called cold plumes, which represent a cooling of the subsoil due to the extraction of heat by GSHPs and will be explained in depth below. With reference to science and technology studies (STS) as a conceptual perspective, and by drawing on issues of GSHP spacing regulation in Germany, we argue that the Heating transitions are currently taking place in many industrialized countries all over the world and they can be understood as transformations of social, technological, and material configurations in the heating sector. As outlined above, the political goal in Germany, as well as elsewhere, is to transform the sociotechnical system of heating in such a way that it becomes less CO 2 -intensive. Thus, heating systems which rely on fossil fuels need to be phased out, while renewable heating systems, such as heat pumps, are expected to spread throughout society. However, the integration of new elements into a larger sociotechnical system almost always goes hand-in-hand with unforeseen and unintended side effects [6]. Unexpected issues come to the fore, brought about by the interplay of social, technological, and material elements of the respective sociotechnical systems. With regard to geothermal energy, one such unexpected issue is the technology's temporary depletability when it is used in specific ways. This temporary depletability becomes observable in the form of so-called cold plumes, which represent a cooling of the subsoil due to the extraction of heat by GSHPs and will be explained in depth below. With reference to science and technology studies (STS) as a conceptual perspective, and by drawing on issues of GSHP spacing regulation in Germany, we argue that the renewability of geothermal energy depends on the sociotechnical context in which geothermal energy is embedded and where it is used as a resource. The embedding in specific sociotechnical contexts brings certain characteristics of geothermal energy to the fore in the form of unintended side effects which had previously been unknown, ignored, or regarded as irrelevant. This may also result in social conflicts. Near-surface geothermal energy development in Germany offers a case study for this. In the following section we briefly outline a science and technology studies-(STS) inspired perspective on the social constitution of geothermal energy as a resource. We then turn to our case study of spacing regulations and cold plumes in Germany, digging deeper into the legal and scientific issues which serve as a crucial knowledge base for the regulation of geothermal energy. After this, we utilize our STS perspective in order to discuss how the renewability of geothermal energy becomes contested within specific sociotechnical contexts. We then end with a short summary and outlook. Conceptual Considerations: How Geothermal Energy Obtains Its Characteristics From an STS perspective, a crucial question regarding energy sources is why and how they become resources in the first place [7]. Thereby, resources can be understood as something that serves the realization of human interests, and thus are considered useful and valuable [8] (p. 1219). Humans draw on biophysical materialities in order to satisfy and pursue individual or collective needs, aspirations, or projects. For example, natural materials are indispensable as sources of food, building materials, and for the production of everyday utensils. Similarly, geothermal energy, like any energy source, is a resource that we harness for a specific purpose-in this case, for heat generation-and to which we thus assign a certain significance and value. This process can be understood as a form of social constitution that presupposes various forms of knowledge and involves integration with larger sociotechnical systems so that the resource can be exploited at scale and in a useful way. According to Thomas P. Hughes, the concept of sociotechnical systems denotes a set of complex components: "Among the components in technological systems are physical artifacts, such as the turbogenerators, transformers, and transmission lines in electric light and power systems. Technological systems also include organizations, such as manufacturing firms, utility companies, and investment banks, and they incorporate components usually labeled scientific, such as books, articles, and university teaching and research programs. Legislative artifacts, such as regulatory laws, can also be part of technological systems" [9] (p. 51). However, through the integration with a specific sociotechnical system, certain characteristics of a resource come to the fore. In the context of an energy system that is fed with significant amounts of volatile renewable energies, for example, the base load capability attributed to geothermal energy gains particular relevance: since heat pumps can generate a stable supply of heat, the fluctuation of other renewables used for heat generation (e.g., solar thermal collectors) can be counterbalanced. By contrast, in a fossil fuel-based energy system, the base load capability of a heat source becomes largely irrelevant, and thus "invisible." Moreover, resources that are integrated into a sociotechnical system can also change that system's configuration. This is the case, for example, since resources are an element of the co-production of energy. They are part of a process involving different actors, technologies, and materialities that produces certain outputs, such as heat. The heating produced by heat pumps is the outcome of a complex process of co-production between homeowners, heating system installers, drilling companies, geologists, environmental administrations, energy production technologies, underground heat, and others [10]. In addition, the integration of heat pumps (which require electricity to operate) increases both the load on the power grid and its peak demand, which, in turn, may require further reinforcement of the grid and other redesigns of the system on a wider scale, such as the integration of energy storage solutions to mitigate peak load [11] (p. 13). As a consequence, heat pumps also alter the configuration of the energy system. The crucial point here is that the CO 2 balance of a heating system, questions of base load capability, and the issue of peak loads are only of technical and political relevance in the sociotechnical context of an energy system that increasingly relies on fluctuating renewables, as well as in the face of aspirations to decarbonize the energy system. Thus, the characteristics of resources such as geothermal energy arise in part from their embeddedness in larger sociotechnical systems featuring specific social, technological, and material configurations. One important "way of knowing" the characteristics of geothermal energy is through their representation by science; in particular, through measurements, quantifications, and modeling. Sheila Jasanoff coined the notion of "ways of knowing" [12] in order to describe the methods used to produce those things which are considered to be facts. In this stream of thought, technoscientific knowledge is central for understanding and thinking about resources. Technoscientific ways of knowing are both normative and functional: "For example, ideas of risk and opportunity are pervasive in the energy sector, and these calculations often reflect the biases of particular groups. The desirable path forward often looks very different depending on whether one is a policy-maker, an energy entrepreneur, or a local citizen" [13] (p. 142). Thus, we are always dealing with a kind of situated knowledge that is dependent on the circumstances under which it is produced and which, therefore, varies from context to context. Models, for example, produce different outputs depending on the often implicit assumptions that are built into them. Specific "ways of knowing" are, therefore, vehicles for the production of facts and always entail predefined (normative and descriptive) ideas of the object under study. These "ways of knowing," nevertheless, also shape our ideas of the characteristics of resources. Ensuring Renewability through Spacing As we will elaborate in the following case study, the design of boreholes for GSHPs seems to affect geothermal energy's status as a renewable energy resource, with the science, laws, rules, and regulations which govern the integration of geothermal energy in the sociotechnical system of energy also playing a crucial role. Our considerations emerged partly out of eight expert interviews of about one hour duration each with researchers, urban planners, and officials from Germany, Austria, and Switzerland dealing with near-surface geothermal energy. The interviews were conducted over the course of half a year in 2021 as part of the German Research Foundation-(DFG) funded project on "Wind harvest and heat theft as indicators of new ownership structures" (2021-2024) (for further information see here: https://sfb294-eigentum.de/en/subprojects/windernte-und-warmeklau-alsindikatoren-neuer-eigentumsordnungen/ (accessed on 29 December 2021)). The interview material was subjected to a thematic analysis [14] with the help of the computer program MAXQDA. Since the interviews are part of a larger research endeavor, only parts of them were relevant for the scope of this paper. Additionally, we have reviewed official documents, such as legal codes and spatial planning schemes, in order to supplement and deepen the insights gained from the expert interviews. Geothermal Energy as Depletable Resource? One of the main aspects that distinguishes renewables from fossil fuels is their ability to renew themselves, ideally at the same rate as they are being consumed. From a natural science perspective, an energy resource can be considered renewable when three crucial parameters are in an equilibrium: the inflow of energy, the outflow of energy, and the stability of the energy reservoir. The inflow of energy has to be equal to the outflow of energy, so that a specific energy reservoir can remain in balance [15] (p. 1). In this sense, "geothermal resources can be considered renewable on the time-scales of technological/societal systems and do not require the geological times of fossil fuel reserves such as coal, oil, and gas" [16] (p. 469). Near-surface geothermal energy systems utilize geothermal heat at depths of up to 400 m and temperatures of up to 25 • C. At this depth, heat sources can be renewed by the natural vertical flow of energy from below, by lateral flows from the side, and from groundwater. However, if more energy is extracted than can be renewed in a particular time frame, the energy reservoir gradually depletes [17] (p. 1). The decline of geysers in New Zealand serves as a prime example of such geothermal resource depletion. In this particular case, extensive extraction of heat caused an irreversible depletion of geothermal resources in certain areas of New Zealand [18] (p. 803). So-called cold plumes are more common and less severe phenomena which can develop as a negative side effect of the application of GSHPs [19]. Even though deep geothermal heat extraction can also cause the heat reservoir to cool down [20], for this paper, the focus lies solely on the near-surface geothermal energy use, which, unlike deep geothermals, is an integral part of the German heating transition. Cold plumes emerge because GSHPs extract underground energy for utilization within heating purposes. In order to do this, they require groundwater (in open systems such as wells) or brine (in closed systems such as probes) to absorb the geothermal heat. GSHPs transfer the heat via heat exchangers and return the cooled water or brine back underground. The liquids can then reheat again due to the stable heat stored in the subsoil. This cyclic extraction-and, in particular, cold liquid or water reinfusion-can cool down the heat reservoir, thus potentially causing cold plumes to emerge [19] (p. 495). These cold plumes develop differently depending on the geothermal extraction system used and the conditions under the ground. If the heat transfer takes place via conduction, then the cooling spreads around the GSHP probe in a radial pattern. If the heat transfer is convective in nature, then a plume-like spreading in the direction of groundwater flow can be observed [21] (p. 70). The speed at which the cooling spreads depends on several factors. These include "soil characteristics, moisture content, building load, initial ground temperature, borehole spacing, etc." [22] (p. 222). Plumes can accumulate if boreholes are located too close to each other. Meng and colleagues [23] investigated cold plumes in a neighborhood in Cologne (Germany) which is characterized by a high density of GSHPs. The adverse placing of real estate and the spacing of GSHPs in partial alignment with the groundwater caused the emergence of cold plumes affecting several households. Whereas property lines clearly denote the boundaries of individuals' property, no physical barrier exists under the ground in order to enforce these. Thus, thermal anomalies can traverse property lines with ease. As a result, cold plumes can cause several issues. As the study by Meng and colleagues shows [23] (p. 10), the most obvious of these is the decreasing efficiency of GSHPs when placed close together, requiring higher electricity input into the heat exchanger. Cold plumes can also result in social costs, such as litigation, if home owners feel negatively affected by neighboring GSHPs [24] (p. 482). These disputes are already part of public debates on geothermal energy extraction, often framed in terms of "heat theft" [25] (p. 37). Furthermore, heat extraction constitutes an artificial interference with the temperature regime that can potentially affect the groundwater ecosystem. Microorganisms, such as bacteria, provide important ecosystem services through filtering or cleaning the groundwater, but they are likely to be affected by changes in temperature [26]. To avoid such interferences, as well as unintended side effects in the form of cold plumes, the proper spacing of heat pumps is crucial. Keeping GSHPs at a Distance With the growing, politically subsidized diffusion of GSHPs, the density of GSHPs is also increasing in Germany, particularly in urban areas, leading to shorter distances between installations. This makes the cold plume phenomenon more relevant and interference more likely. If geothermal energy as a renewable resource is supposed to contribute to the German heat transition in an optimal way, the phenomenon of cold plumes could become an issue of growing relevance with regard to the planning of GSHPs. For this reason, the different German federal states have implemented distancing guidelines for near-surface geothermal energy systems. Depending on the state and/or municipality, these vary between 5 and 10 m [27] (p. 123). However, it remains unclear on which knowledge base these distances have been established. Furthermore, in an international survey of national legislation, temperature limits, and distances between near-surface geothermal energy systems, Hähnlein and colleagues found that regulations vary widely from country to country and are not solely derived from scientific considerations and evidence, but also from supposed best practice examples and rules of thumb [28]. Furthermore, plumes not only depend on spacing, but also on the characteristics of the underground environment. For example, in gravel aquifers, cold plumes can exceed 10 m within a single heating period at average withdrawal rates, as water can flow through the ground quicker [29]. Conditions under the ground vary regionally, making general distancing guidelines partially useless. Since distancing guidelines do not necessarily meet the actual requirements for the efficient use of geothermal energy, they are a subject of controversial debate among experts: "The existing legal frameworks all over the world have failed to some degree to provide a scientific-based solution to this problem and aimed to use simple approaches, often with their roots in groundwater resource management, that have ended in disperse incoherent legal enforcements" [29] (p. 10). To avoid "significant errors," simple models can help to determine suitable spacing for boreholes based on "all geometrical parameters" [19] (p. 497). However, despite modeling, problems can arise, especially when geothermal probes do not run directly downwards, but run at an angle, extending beyond the boundaries of individual property [30] (p. 2). Moreover, the results produced by computer models are always and inevitably affected by uncertainties, since they are simplifications of a complex reality. It follows from the above discussion that sound spacing of GSHPs seems fundamental to guaranteeing the sustainable extraction of the resource without causing interference with adjacent GSHPs. However, as we have laid out, thresholds and distancing regulations are contested and both seem to be somewhat arbitrary and, in general, to be constituted through contingent social processes. This not only holds true for distancing recommendations, but also, in part, for formal regulations for geothermal energy extraction more broadly as they are implemented in German law. Governing Geothermal Energy Extraction by Law In Germany, the extraction of geothermal energy is subject to several legal preconditions. In general, the German Civil Law Book ("Bürgerliches Gesetzbuch"-BGB), which allows landowners to make use of their land, includes in this permission both the sky above and the land underneath a piece of property ( §905 I 1 BGB). However, this does not hold true on all accounts. Even though the book suggests that natural resources in the ground can be exploited by landowners, the Federal Mining Act ("Bundesberggesetz"-BBergG) further defines under which conditions resources may be extracted. The Federal Mining Act distinguishes between two types of extractable resources under the ground. On the one hand are the "grundeigene Bodenschätze," that is, mineral resources that belong to landowners ("grundeigen") and can be extracted without any formal permission ( §3 II 1 BBergG). On the other hand are "bergfreie Bodenschätze" that do not belong to the landowner, and thus require permission to extract ( §3 II 2 BBergG). The adjective "bergfrei" in connection to mineral resources is a reference to the geological subsoil free for mining. Moreover, the Federal Mining Act explicitly states that land rights do not extend under the ground in the case of "bergfreie Bodenschätze" [24] (p. 477). In general, private installations by homeowners do not require mining permission in Germany, unless the boreholes exceed a depth of 100 m or the scope of the permitted field extends beyond the boundaries of the property [25] (p. 37). Then they require a permission in accordance with §127 Federal Mining Act ("Bundesberggesetz"-BBergG). Thus, up to a depth of 100 m, formal permission for the extraction of geothermal energy is only required if the geothermal energy system extends beyond the property boundaries or if the neighboring property is thermally or materially affected [21] (p. 71). There is no formal reason or scientifically grounded justification for the 100-m threshold. Nevertheless, it has a direct impact on the installation, efficiency, and functioning of GSHPs. When it comes to the active regeneration of underground heat reservoirs, existing cold plumes could be mitigated more easily with boreholes exceeding 100 m. Thus, boreholes from different neighboring GSHPs could more easily be placed at different depths. However, this would be associated with higher financial costs (as the cost of drilling increases with depth), even though it is generally cheaper than actively regenerating underground reserves in the long term [31] (6f). Deeper drilling is also more time consuming and involves higher bureaucratic hurdles since the relevant permissions must first be obtained from the State Mining Office ("Landesbergamt"). Nevertheless, considering the level of heat under the ground which increases with every meter, deeper boreholes-depending on underground characteristics-could improve efficiency and reduce thermal interference between adjacent heat pumps [32] (p. 35). Accordingly, in Germany, the Federal Association of Heat Pumps ("Bundesverband Wärmepumpe") is already demanding a reconsideration of the 100-m threshold. The association suggests expanding the limit to a depth of 400 m, the maximum depth for near-surface geothermal energy systems (VDI-The Association of German Engineers 4640, Blatt 1, 2010, as referenced in [26] (p. 26)). Interestingly, in the neighboring countries of Austria and Switzerland, no comparable regulations in terms of depth of drilling have been imposed. Rather, drilling is carried out until a suitable geothermal heat reservoir has been reached. This can be at a depth of 50 m or 200 m. Of course, the embedding of GSHPs into a regulatory framework is necessary for the prevention of conflicts and to steer the further development of geothermal energy usage. However, as we have seen, regulatory frameworks do not necessarily serve to promote the expansion of GSHPs in an efficient way. They can also serve as obstacles or contribute to the formation of negative side effects, such as cold plumes. In Spain, for example, García-Gil and colleagues have concluded that the present regulatory framework is slowing down the development of geothermal energy. They describe a similar ambiguity and arbitrariness with regard to the depth of boreholes as we have found to be the case in Germany [33]. Discussion We have suggested above that geothermal energy can only be regarded as a fully renewable resource if the spatial layout of GSHPs is appropriately regulated. This means that, when geothermal heat is extracted from under the ground, it must be ensured that at least the same amount of heat can be returned or regained through natural or technical processes. This is due to the physical characteristics of geothermal energy as a renewable energy source. If this is not guaranteed, the ground will gradually cool down, making the operation of affected GSHPs inefficient. Thus, proper distancing between individual GSHPs is necessary in order to avoid interferences. In the context of the German energy transition, the demand for GSHPs is growing and thus spacing regulations are becoming increasingly important. As has been illustrated above, the distance that should be maintained between individual GSHPs is based on an understanding of cold plume formation and propagation that is barely adequate in the context of contemporary scales of geothermal energy usage. That existing spacing regulations appear to have arbitrary origins, and yet they continue to be followed, indicates that they have become part of general "knowledge" about geothermal energy to a point where they are hardly ever questioned. German distancing regulations were established before geothermal energy was considered to be an important part of the energy transition, with other kinds of resources in mind, and at a time when underground resources remained relatively unexplored. Further, different knowledges based on specific "ways of knowing" geothermal energy compete with each other for being incorporated in spatial planning practices and regulations [34]. These knowledges become incorporated in these practices and regulations, to different extents. Present regulations seem to be mainly derived from practical knowledge, conventional wisdom, and past experiences. However, these knowledges seem to blend out or underestimate geothermal energy's ability to produce cold plumes when being used by GSHPs. At the time when these knowledges have become embedded in practices and regulations of distancing, they may have been sufficient to ensure an appropriate management of geothermal energy as a resource. However, under conditions of density between geothermal probes, these knowledges seem to fail the test of practical applicability. Most interestingly, other countries, such as Switzerland and Austria, have adopted these guidelines despite different soil conditions and use patterns in either country. Switzerland's largest city, Zurich, for example, has a vast amount of aquifers and, therefore, relies heavily on groundwater and open geothermal systems, whereas Germany installs mostly geothermal probes [35]. Still, the German distance regulation of 10 m has been adopted. This transplantation of regulations [36] has also resulted in a transplantation of the congealed knowledges incorporated in the regulations and, thus, in a transfer of the boundaries of these knowledges. Modeling and measurements as ways of knowing the underground may appear to produce more robust knowledge since they are able to capture the cold plume phenomenon. However, each way of knowing produces certain blind spots and is affected by uncertainties. Furthermore, as Brian Wynne has famously pointed out [37], technoscientific knowledges are not necessarily superior to knowledges grounded in practical experience or conventional wisdom. It is not only different ways of knowing which constitute certain characteristics of geothermal energy. As we have seen, the embedding of heat pumps within the specific configurations and expectations of the present energy system brings the (temporary) depletability of geothermal energy to the fore. Thereby, the material world exerts its own influence on the constitution of energy system characteristics. As Bruno Latour [38] has elaborated in detail, biophysical materialities follow their own trajectories, and thereby intervene in the flow of events in the world. That is, the characteristics of resources are not arbitrarily malleable, but exhibit moments of unruliness. As shown above, these moments of unruliness are revealed by the complex and partially arbitrary interplay between different knowledges formed by different knowledge-producing institutions and the biophysical entity of geothermal energy [39]. In our case, the main knowledge-producing institutions are law and science. As we have seen, law is a crucial knowledge-producing institution when it comes to the installation of GSHPs, while science and its inherent ways of knowing [12] (p. 249) are crucial for understanding the biophysical characteristics of geothermal energy. This understanding, however, also changes over time. Nevertheless, law and science are central to our understanding of geothermal energy and its characteristics. Miller et al. [13] thus speak of "energy epistemics" in order to denote the cognitive representation of energies on the part of different knowledge-producing institutions. These energy epistemics, together with the material world, co-produce [40] the cold plume phenomenon. These cold plumes, however, only come to the fore and become problematized through the reconfiguration of the sociotechnical system of energy and the growing diffusion of heat pumps in order to decarbonize heating. All in all, it is the constellation of different knowledge sources (which have partly sunk into oblivion), the resource's inherent unruliness, and the political project of decarbonization that brings cold plumes into being, which makes them a debated issue, and which reveals geothermal energy's characteristic of temporary depletability. This is what we understand to be the social constitution of geothermal energy. Conclusions Drawing on the example of cold plumes and regulatory issues of spacing in Germany, we have elaborated the unruly nature of geothermal energy, which reveals novel characteristics when exploited at scale. The scope of our research is mainly limited to Germany. However, the phenomenon of cold plumes and accompanying regulatory issues are not Germany-specific. If geothermal energy is intended to become an important component of a decarbonized energy mix in Germany and other countries around the world, careful management of the resource will become increasingly necessary. Thereby, regulatory frameworks and distancing regulations which draw on various knowledges grounded in distinct ways of knowing play a crucial role [41,42]. As we have seen, the sociotechnical context in which geothermal energy is currently embedded in Germany draws out its characteristic of being (at least temporarily) depletable. Unintended side effects are a common issue in complex sociotechnical transformations, such as the German energy transition. Thus, when it comes to dealing with such side effects, it is not so much a question of whether these unintended side effects could have been foreseen (since they cannot be avoided) as of how these effects can be minimized after they became perceivable. In the case of cold plumes and interference between neighboring GSHPs, a carefully balanced regulatory framework for locating GSHPs seems to be the obvious solution. However, this regulatory framework must be based on extant sets of knowledge stemming from the areas of planning, law, science, and engineering. Even when drawing together these different areas, such knowledge will never be complete, and the generation of new knowledge often includes the awareness of new knowledge gaps that foster new uncertainties, risks, and ignorance [43][44][45]. Whenever knowledge grows, so does the perception of ignorance. Uncertainties in geothermal energy operations will thus appear as soon as the sociotechnical context in which the use of geothermal energy is embedded changes. Ultimately, this article throws into relief the social constitution of geothermal energy. It is intended to illustrate the relevance of the issue and whet appetite for further research to gain a deeper understanding of which actors-ranging from natural scientists and energy suppliers to NGOs and governments-contribute in what way to the epistemics of energy surrounding geothermal heating. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Data Availability Statement: The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy restrictions.	2022-01-12T16:14:24.206Z	2022-01-04T00:00:00.000	{ "year": 2022, "sha1": "4da705b05cbf2daa8287b480779e55b0da290034", "oa_license": "CCBY", "oa_url": "https://www.mdpi.com/2071-1050/14/1/496/pdf?version=1641280102", "oa_status": "GOLD", "pdf_src": "Adhoc", "pdf_hash": "5f75910db28b6746142deb4f942ce15e9ced49a2", "s2fieldsofstudy": [ "Geology" ], "extfieldsofstudy": [] }
19118760	pes2o/s2orc	v3-fos-license	Dry eye, sleep quality, and mood status in glaucoma patients receiving prostaglandin monotherapy were comparable with those in non-glaucoma subjects Purpose Prior studies suggested that glaucoma patients suffer worse dry eye and mood and sleep disorders than non-glaucoma subjects. Prostaglandin analogues are first-line therapy for glaucoma, inducing few instillation problems and sufficient pressure-reduction effects. This study compared dry eye, sleep quality, and mood status between glaucoma patients receiving prostaglandin monotherapy and non-glaucoma subjects. Methods This cross-sectional study evaluated 1520 patients (579 males and 941 females) for glaucoma status and dry eye-related symptoms (dryness, eye fatigue, photophobia, pain, blurring) and signs (Schirmer test, tear break-up time, corneal staining scores). Of the total cohort, 93 patients were also evaluated by Pittsburgh sleep quality index (PSQI) and hospital anxiety and depression score (HADS). Inclusion criteria were consecutive patients ≥ 51 years of age and best-corrected visual acuity ≥ 20/25. Glaucoma patients included those treated with prostaglandin or a fixed combination including prostaglandin. Exclusion criteria were history of ocular surgery within one month. Data were analyzed using the chi-square or Mann-Whitney U tests, at 5% significance. Results There were no significant differences in dry eye-related signs and symptoms between the control (n = 1431, mean age of 66.9 years) and glaucoma groups (n = 89, 67.9 years). The psychiatric sub-analysis of the control (n = 61, 66.2 years) and glaucoma groups (n = 32, 67.3 years) revealed mean scores of 5.02 ± 3.10 and 5.16 ± 3.46 for PSQI (normal range ≤ 5), 9.47 ± 5.61 and 9.42 ± 7.36 for HADS (normal range ≤ 10), 4.84 ± 3.22 and 4.71 ± 3.45 for anxiety (normal range ≤ 5), and 4.63 ± 3.05 and 4.71 ± 4.40 for depression (normal range ≤ 5), respectively, without statistical significance. Conclusions Our results were comparable between glaucoma patients on prostaglandin monotherapy and non-glaucoma subjects for dry eye-related clinical manifestations, sleep quality, and mood status. Introduction Topical therapy and laser applications have become major strategies in modern management for glaucoma that strives for less invasive intervention. Following recent advances in topical therapies for glaucoma and its diagnosis, the number of patients using glaucoma eye drops has increased [1] and the use of prostaglandin (PG) analogues and fixed combination therapies in glaucoma treatment is markedly more prominent. All such advances have combined to effectively reduce intraocular pressure, which is beneficial for both patients and ophthalmologists. Ocular surface side effects are mostly from preservatives contained in glaucoma eye drops [2,3]. Patients may feel some discomfort during the use of such eye drops including irritation, burning, stinging, and a foreign body sensation. [4] Benzalkoniun chloride (BAK) is the primary preservative used in glaucoma eye drops [5,6], while other preservatives are also used to reduce toxicity including SofZia TM (Alcon Laboratories, Tokyo, Japan) and PolyQuad TM (Alcon Laboratories, Tokyo, Japan).Numerous reports detail BAK cytotoxicity [7,8,9], which generally depends on BAK concentrations. Recent studies suggested that glaucoma patients tend to have sleep disturbance [10][11][12] and mental disorder, while other groups reported glaucoma patients were depressive in comparison to non-glaucoma subjects [13][14][15][16], although the severity of glaucoma and visual field defects may contribute to the depression. Damage of intrinsically photosensitive retinal ganglion cells (ipRGCs) was proposed as underlying condition in sleep disorder of glaucoma patients [11] and the presence of circadian rhythm disorder has also been suggested to associate with glaucoma, [17], while the burden of eyedrop instillation might also contribute to sleep disturbance [12]. PGs have several benefits for pressure reduction and prevention of side effects since currently introduced PGs contain only a small amount of BAK and require a single instillation per day. Compared to conventional therapies with multiple medication and instillation, a single instillation is free from cumulative toxicity to ocular surface [18] and the burden of daily medication management [12]. Consequently, patient's adherence was significantly improved with accompanying substantial reductions in intraocular pressure [19][20][21]. Taken together, previous findings indicated that glaucoma patients receiving PG monotherapy are expected to have less negative effects with regard to ocular and psychiatric health. However, to the best of our knowledge, there are few investigations comparing sleep status between non-glaucoma subjects and glaucoma patients taking PGs. The aim of this study was to compare ocular surface complications, sleep quality, and mood status between glaucoma patients on prostaglandin monotherapy and non-glaucoma subjects. We conducted a multi-center study in a large cohort under the standardized diagnostic criteria for glaucoma and dry eye disease (DED) to minimize selection bias. Study institutions and Institutional Review Board approval The Institutional Review Boards and Ethics Committees of Keio University School of Medicine, Shinseikai Toyama Hospital, Shinkokai Medical Group, and Komoro Kosei Hospital approved this study, and the study was performed in accordance with the principles of the Declaration of Helsinki. Informed consent was obtained from all participants. Ophthalmological examinations and treatments Board-certified ophthalmologists examined all patients, with diagnostic criteria for glaucoma in the present study comprising glaucomatous visual field loss less than -4.0 dB MD in the worse eye, an ophthalmoscopic NFL defect, a c/d ratio > 0.6, or elevated IOP (> 21 mmHg) requiring topical medication for more than 6 months. Exclusion criteria were coexisting cataract with significant lens opacity disturbing the optical axis that accounted for subjective visual disturbance or decreased visual function, retinal pathology, retinal surgery, or photocoagulation affecting the visual field. Topical glaucoma medication included Xalatan 1 (Pfizer, Tokyo, Japan), Tapros 1 (Santen Pharmaceutical Co. Ltd., Osaka, Japan), Travatanz 1 (Alcon Laboratories, Tokyo, Japan), Lumigan 1 (Senjyu Pharmaceutical Co. Ltd., Osaka, Japan), and Xalacom 1 (Pfizer, Tokyo, Japan). Table 1 (S1 Table) describes the distribution of prescribed PGs. The concentration of BAK in the medications was 0.001% to 0.02%. The low-toxicity preservatives, PolyQuald 1 is included in Travatanz 1 . Patient interviews for DED-related symptoms Outpatients were interviewed regarding symptoms related to DED. Questions determined the presence or absence of six common DED-related symptoms, namely dryness, irritation, pain, eye fatigue, blurring, and photophobia. The questions put to patients were selected from items in the Dry Eye Questionnaire Score (DEQS) [23], which is based on the six most prevalent symptoms of DED patients visiting the Dry Eye Clinic of the Department of Ophthalmology, Keio University Hospital, in 2012. Questionnaire-based survey for sleep and mood disorders Outpatients were invited to complete the Pittsburgh Sleep Quality Index (PSQI) [24] and the Hospital Anxiety and Depression Scale (HADS) [25] questionnaires. Scores for each scale were calculated according to separate algorithms and then subjected to analysis. The cut-off point for PSQI was a total score of 5 or greater, being indicative of poor sleep quality. The cut-off points for the HADS were 0-10 as normal and 11-21 as abnormal. The PSQI consists of seven subscales, whereas the HADS consists of depression (HADS-D) and anxiety (HADS-A) subscales. These questionnaires have been widely used for hospital-based surveys and are easy to answer, even by eye clinic visitors, because they do not contain difficult questions concerning severe psychiatric disease (e.g. suicide and hallucinations). Patient evaluations for systemic comorbidities The presence of systemic comorbidities was evaluated on the basis of brief interviews, chart review, and results from annual health check records, if available. Assessments of the use of sleep medications and antidepressants were included in the PSQI score. Inclusion and exclusion criteria for people with eye disease and systemic comorbidities The present study recruited consecutive patients older than 50 years with best corrected visual acuity equal to or better than 20/25 in both eyes, from participating clinics during the study period. In addition, patients with ocular surgery within one month, diagnosed psychiatric disease, and taking specific psychiatric medications were excluded from the study. DED is often accompanied with sleep and mood disorders, and therefore those with DED were excluded from the sleep study [26] so as to not overestimate sleep and mood status. Statistical analysis Subjects were divided into two groups: glaucoma patients treated with PG-containing eyedrops and non-glaucoma subjects. Results for DED-related signs and symptoms as well as sleep and mood disorders were compared between groups, based on results of the Schirmer test and the tear BUT in the right eye. Where appropriate, data are given as the mean ± SD. Chi-square and Mann-Whitney U tests were used as appropriate for all analyses, which were performed using StatFlex (Atech, Osaka, Japan) with P < 0.05 considered significant. Results We evaluated 1520 patients (579 males and 941 females) by comprehensive glaucoma and dry eye examinations. A total of 93 patients in this series were also evaluated for sleep quality and mood status using two questionnaires. Tables 2 and 3 (S1 and S2 Tables), and indicated no significant difference in either DED-related symptoms and signs or sleep and mood indices between the glaucoma and control groups. Dry eye related eyedrops and other topical medications were not different between two groups ( Table 2). Comparisons of representative data between the groups are shown in Fig 1A, 1B and 1C which reveals no significant difference between the groups in corneal staining score, PSQI global score, and HADS score. Discussion Our results demonstrated that PG monotherapy had no negative impacts on ocular surface health, sleep, and mood in glaucoma patients. Despite that DED patients often report discrepancies between symptoms and signs [27], and no examined parameters showed inconsistency as to whether glaucoma patients receiving PG monotherapy really enjoy reduced ocular discomfort. This benefit should contribute to better adherence and avoidance of additional care for ocular surface management. Recent formulation of glaucoma eye drops contains less or no BAK and it may be the cause of the improved tolerance resulting in decreased side effects. There was no difference in the sleep and mood indices between the glaucoma and control patients, while previous reports showed that glaucoma patients tended to have sleep and mood disorders [10][11][12][13][14][15][16]. The reason for this apparent discrepancy is that the present study may not include very severe cases since all patients were controlled with PG monotherapy only and the burden of medication was minimal. Previous investigations with detailed ophthalmological evaluation included many cases with severe visual field loss and multiple medications and they indicated the psychiatric disorders were associated with the severity of glaucoma [10][11][12][13]. Damage to the ipRGC is an attractive hypothesis for sleep disturbance in glaucoma although clinical evidence to support this proposal are still to be determined. Gracitelli et al 1. Comparison of representative parameters for dry eye disease, sleep, and mood between an affected and control groups. There was no significant difference between the control patients those with glaucoma in corneal staining score (Fig 1A), PSQI global score (Fig 1B), and HADS score (Fig 1C). The data are given as mean ± SD. PSQI, Pittsburgh Sleep Quality Index; HADS, Hospital Anxiety and Depression Scale; HADS-A, Anxiety subscale; HADS-D, Depression subscale. https://doi.org/10.1371/journal.pone.0188534.g001 Dry eye, sleep quality, and mood status in glaucoma patients receiving prostaglandin monotherapy demonstrated decreased sleep quality in glaucoma patients with electroencephalography and it was correlated to reduced ipRGC function measured with a pupillometer under blue-light stimulus [11]. Actually, numerous factors affect sleep quality in aged glaucoma patients and we speculate many glaucoma patients may be depressive due to the progressive nature of the disease and the burden of life-long medication use [12], thus it is likely that depression may be a major cause of sleep disorders in such patients since it is strongly correlated with sleep disturbance [28,29]. Social awareness of glaucoma and advancement of optical coherence tomography have led to easy and early diagnosis of glaucoma [30], and many patients are now treated appropriately. A recent review of PG therapy describes no significant side effects for the ocular surface [31]. We therefore speculate that cumulative epidemiological evidence, diagnostic and pharmaceutical advancements, and reduced preservative in PGs may contribute to both effective treatment and prevention of dry eye and psychiatric disorders in glaucoma patients receiving PG monotherapy. Limitations The present study has limitations. The gender ratio between control and glaucoma groups in dry eye study were significantly different, however, we believe it did not affect the results since the difference may not be clinically significant and topical medications were similar in both groups. For more informative results, sleep and mood in such patients should be further evaluated with polysomonography (electroencephalogram) and psychiatric consultation. In addition, further investigations should be conducted in patients with severe glaucoma being treated with beta-blocker eyedrops with a single instillation formulation. Finally, the number of subjects who completed the Schirmer test was small, although still acceptable because it was not among the essential new diagnostic criteria for dry eye [32]. Conclusions In conclusion, our results demonstrated no significant difference between control and glaucoma groups in DED, sleep, and mood, suggesting that adverse ocular and psychiatric effects with prostaglandin monotherapy might be minimal.	2018-04-03T02:25:19.370Z	2017-11-27T00:00:00.000	{ "year": 2017, "sha1": "8e94bafb56e6d5b3fc057f5ff88945ee76d6b21b", "oa_license": "CCBY", "oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0188534&type=printable", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "8e94bafb56e6d5b3fc057f5ff88945ee76d6b21b", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
256111983	pes2o/s2orc	v3-fos-license	Failed reinnervation in aging skeletal muscle Skeletal muscle displays a marked accumulation of denervated myofibers at advanced age, which coincides with an acceleration of muscle atrophy. In this study, we evaluated the hypothesis that the accumulation of denervated myofibers in advanced age is due to failed reinnervation by examining muscle from young adult (YA) and very old (VO) rats and from a murine model of sporadic denervation secondary to neurotrypsin over-expression (Sarco mouse). Both aging rat muscle and Sarco mouse muscle exhibited marked fiber-type grouping, consistent with repeating cycles of denervation and reinnervation, yet in VO muscle, rapsyn at the endplate increased and was associated with only a 10 % decline in acetylcholine receptor (AChR) intensity, whereas in Sarco mice, there was a decline in rapsyn and a 25 % decrease in AChR intensity. Transcripts of muscle-specific kinase (21-fold), acetylcholine receptor subunits α (68-fold), ε (threefold) and γ (47-fold), neural cell adhesion molecule (66-fold), and runt-related transcription factor 1 (33-fold) were upregulated in VO muscle of the rat, consistent with the marked persistent denervation evidenced by a large proportion of very small fibers (>20 %). In the Sarco mice, there were much smaller increases in denervation transcripts (0–3.5-fold) and accumulation of very small fibers (2–6 %) compared to the VO rat, suggesting a reduced capacity for reinnervation in aging muscle. Despite the marked persistent denervation in the VO rat muscle, transcripts of neurotrophins involved in promoting axonal sprouting following denervation exhibited no increase, and several miRNAs predicted to suppress neurotrophins were elevated in VO rat. Our results support the hypothesis that the accumulation of denervated fibers with aging is due to failed reinnervation and that this may be affected by a limited neurotrophin response that mediates axonal sprouting following denervation. Background Skeletal muscle undergoes repeating cycles of denervation and reinnervation for much of adult life [1]. However, in very advanced age, when muscle atrophy becomes severe and has clinical impact in terms of impairing mobility [2], there is a marked accumulation of persistently denervated muscle fibers [3,4]. Notably, we have shown in very old (VO) rat muscle that denervated myofibers are on average 35-50 % smaller than innervated fibers and innervated fibers of VO muscle are only 7 % smaller than in young adult [3]. The significance of this is that denervation is the primary cause of atrophy in the period of the lifespan when the clinical impact of aging muscle is most relevant. As an accumulation of persistently denervated fibers marks the acceleration of atrophy in both slow and fast twitch muscles with aging [5], understanding the mechanisms contributing to their accumulation is essential to provide effective targets for combating age-related muscle atrophy. Although there have been a number of recent studies examining the mechanisms that may contribute to destabilization of the aging neuromuscular junction (NMJ) and subsequently lead to denervation of muscle endplates [6][7][8], there is scant information about the potential mechanisms contributing to impaired reinnervation of denervated fibers in very advanced age. Importantly, the progressive increase in fiber type grouping seen with increasing age [9] is clear evidence that following a sporadic denervation event, such fibers are usually successfully reinnervated for much of the lifespan. For this reason, a likely cause for the accumulation of persistently denervated fibers in very advanced age is a failure of reinnervation. In this context, previous studies have noted an impaired reinnervation response following acute nerve injury in aged animals that is due to impaired axonal regrowth [10,11]. Despite this evidence, no prior studies have considered whether this impaired capacity for axonal regrowth (identified following acute nerve injury in aged animals) might also contribute to the accumulation of persistently denervated muscle fibers following denervation events that are part of the normal process of aging. In addressing factors that might cause a compromised reinnervation response with aging, it is relevant to note that following surgical denervation in healthy young animals, there is a robust increase in neurotrophins, including brain-derived neurotrophic factor (BDNF) [12]. Furthermore, forced over-expression of the neurotrophins BDNF, glial-derived neurotrophic factor (GDNF) and nerve growth factor (NGF) in muscle facilitate axonal sprouting and reinnervation following surgical denervation [13]. Similarly, the exercise-induced enhancement of reinnervation following nerve crush injury is dependent upon muscle BDNF signaling [14]. As such, examination of neurotrophins involved in reinnervation and microRNAs (miRNAs) predicted to suppress neurotrophins could provide insights to the potential cause of failed reinnervation in very advanced age or at the least identify whether this pathway might represent a therapeutic target. To help address these issues, we examined muscle morphology (fiber type and size distribution, fiber type grouping) and denervation-related transcripts (e.g., muscle specific kinase [MuSK], AChR subunits, cytoplasmic phospholipase A2 [cPLA2], neural cell adhesion molecule [NCAM], and runt-related transcription factor 1 [RUNX1]) in vastus lateralis (VL) muscle from young adult (YA; 8 months old) and very old (VO; 35 months old) male Fisher 344 × Brown Norway F1-hybrid (F344BN) rats and compared this to the same measures made in soleus and gastrocnemius muscles from male wild-type (WT) and neurotrypsin-over-expressing mice (Sarco mouse; a model of sporadic denervation). A key element of the study design is that the age of the Sarco mice (8 months of age) allows us to evaluate the capacity for reinnervation of young adult muscle subjected to sporadic denervation. We hypothesized that both VO rats and Sarco mice would exhibit marked fiber type grouping (indicating repeating cycles of denervation and reinnervation in both models) but compared to Sarco mice, VO rat muscle would exhibit a greater accumulation of small fibers indicative of persistent sporadic denervation and a commensurately greater expression of denervation-related transcripts. As both aging [9] and neurotrypsin over-expression [7] cause repeating cycles of denervation and reinnervation, a greater accumulation of sporadically denervated muscle fibers and up-regulation of denervation-related transcripts would implicate failed reinnervation as a major contributor to the accumulation of persistently denervated muscle fibers in advanced age. To provide insights to potential mechanisms of failed reinnervation in VO muscle, we also examined neurotrophin transcripts and miRNAs predicted to blunt neurotrophin levels. Results Fiber type proportions and fiber type grouping in aged rat and Sarco mice Figure 1a shows representative images of the rat vastus lateralis and Sarco mouse soleus muscle cross-sections labeled using primary antibodies against myosin heavy chain (MHC) isoforms to identify fiber type and laminin to distinguish fiber borders. Proportions of pure fibers (types I, IIa, IIb, and IIx) were lower in the VO rat muscle compared to YA and lower in Sarco mice than wild-type mice. This decline in pure fibers was accounted for by an increase in abundance of fibers co-expressing two or more MHCs (MHC co-expressing fibers) in the VO rat and Sarco mice (Fig. 1b), a phenotype that we have shown previously to be largely secondary to denervation in the VO rat muscle [3]. To obtain an index of the occurrence of denervation-reinnervation events, we examined the abundance of fibers totally enclosed by fibers of the same type (fiber type grouping) in rat and mouse muscle. This analysis revealed that there was a large increase in fiber type grouping in both VO rat and Sarco mouse muscle (Fig. 1c). MuSK-rapsyn at the neuromuscular junction To evaluate changes in one of the primary signaling axes for regulating the integrity of the post-synaptic AChR plaque, we examined phospho-MuSK, rapsyn, and AChR intensity levels in the endplate region of the muscle fiber in both the aging and Sarco mice (Fig. 2). Interestingly, whereas immunofluorescent images show relatively preserved phospho-MuSK and semi-quantitative analysis in muscle cross-sections revealed an increase in rapsyn and only a small (10 %) decline in AChR intensity in VO rat muscle, phospho-MuSK appeared considerably reduced in Sarco Soleus muscle and this was accompanied by a decline in rapsyn and relatively large (25 %) decline in AChR intensity. Fiber size in aged rat and Sarco mice To gain insight into the frequency of failed reinnervation events in the VO rat, we examined the abundance of muscle fibers exhibiting a size of ≤1000 μm 2 , as these represent a population of muscle fibers shown by our group in VO rat gastrocnemius muscle to exhibit a high association with positive labeling for a marker of denervation (>90 % of these fibers are positive for the sodium channel isoform, Nav 1.5 ) and to represent <1 % of fibers from young adult muscle [3]. We then compared this to a corresponding fiber size in Sarco mice (650 μm 2 in gastrocnemius, 575 μm 2 in soleus), based upon these representing a similar percentile of fiber size in the WT mice as in the YA rat. Interestingly, whereas VO rat exhibited >20 % of these very small fibers, Sarco mice exhibited approximately 2-6 % of these very small fibers (higher in soleus than gastrocnemius) (Fig. 3a). Taken in the context of the high frequency of denervationreinnervation events evidenced by the large increase in fiber type grouping in both VO rat VL and Sarco mouse soleus muscle noted above (Fig. 1c), our analysis of very small fiber abundance reveals a marked impairment in reinnervation in VO rat muscle. On the other hand, a degree of compensatory hypertrophy was visually evident in photomicrographs of Sarco soleus muscle (Fig. 1a), consistent with a previous report [7]. Transcriptional profile of denervation in aged rat and Sarco mice The expression of denervation-related transcripts was dramatically increased in VL muscle of VO rat samples when compared to YA, and these increases were markedly greater than seen in Sarco mouse soleus or gastrocnemius muscles (Fig. 3b). Specifically, transcript levels of MuSK, a key signaling molecule involved in regulation of AChR clustering at the neuromuscular junction [15,16], increased 21-fold in VO muscle (20-50 % change in Sarco mouse muscles). Similarly, transcript levels of the AChR subunit α (68-fold in VO vs YA) and γ (47-fold in VO vs YA) increased to a much greater degree in VO rat muscle than in Sarco mouse soleus muscle where AChR α had a borderline significant increase and AChRγ modestly increased (threefold) (no change in gastrocnemius muscle) vs wild-type mice. Interestingly, whereas AChR ε increased approximately threefold in VO rat VL muscle, it declined by nearly 60 % in Sarco soleus muscle. Transcript levels of NCAM (50-fold higher in VO vs YA), which encodes a protein that increases in aging muscle [17] and is involved in restoration of nerve contact to the muscle [18], and RUNX1 (40-fold higher in VO vs YA), which encodes a protein that also increases in aging muscle [19] and counters muscle atrophy [20], were markedly elevated in VO rat muscle but not in Sarco mouse soleus or gastrocnemius muscle. Transcript levels of cPLA 2 , an enzyme catalyzing release of mitochondrial lipids as fatty acid hydroperoxides following denervation [21], increased in VO rat VL muscle and in Sarco mouse Soleus muscle (both approximately twofold relative to their respective controls). Collectively, these results identify a markedly greater denervation-related transcriptional profile in VO rat muscle than the Sarco mouse, despite both models being subjected to recurring cycles of sporadic denervation. Transcriptional profile of neurotrophins and related receptors in aged muscle The gene expression of neurotrophins ( Fig. 4a left panel), including those implicated in promoting axonal growth following nerve injury [12,13], were not significantly elevated in VO muscle, despite the marked accumulation of very small muscle fibers and large increases in expression of denervation-related transcripts. This was also the case in both muscles examined in Sarco mice (Fig. 4b, c left panels), although the denervation stress here is clearly less than with aging and there is likely less requirement for axonal sprouting due to the exclusively post-synaptic nature of the NMJ instability in this model. When considering transcripts for the neurotrophin receptors (Fig. 4b), only neurotrophic tyrosine kinase receptor type 3 (NTRK3; encodes the NT3 receptor) and ciliary neurotrophic factor receptor (CNTFR; encodes the CNTF receptor) were increased in rat VO muscle. The other neurotrophin receptors, p75NTR (encoding the low affinity nerve growth factor receptor for several neurotrophins), neurotrophic tyrosine kinase receptor type 1 (NTRK1; encoding the high-affinity receptor for NGF), neurotrophic tyrosine kinase receptor type 2 (NTRK2; encoding the high-affinity receptor for BDNF), and GDNF family receptor alpha-1 (GFRA1; encoding the GDNF receptor) exhibited no change in rat VO muscle compared to YA muscle (Fig. 3a, c). Similar to aging muscle, there were minimal changes in neurotrophin TBP was used as endogenous control. Data are shown as means with standard error. T tests were performed. P < 0.05 was considered as statistically significant receptor transcripts in the Sarco mice, with only NTRK1 exhibiting a small decline in expression in Sarco mouse gastrocnemius muscle Fig. 4b, c left panels. Although this analysis does not reveal a decline in neurotrophin signaling in the VO muscle, taken in the context of the large increase in very small fibers and denervation-related transcripts, a reasonable conclusion is that the neurotrophin response is insufficient to facilitate successful reinnervation. Discussion There is a need for identification of viable therapeutic targets to treat aging muscle in advanced age where the clinical impact of muscle atrophy becomes meaningful. A key step in this endeavor is a clear understanding of the processes that lead to muscle atrophy in advanced age. To this end, the purpose of this study was to evaluate the hypothesis that the accumulation of persistently denervated muscle fibers in advanced age is due to failed reinnervation by comparing changes with aging to changes seen in a mouse model of sporadic denervation, the Sarco mouse. Our results show that VO rat VL muscle is characterized by marked accumulation of very small fibers and MHC co-expression, two hallmark features of sporadic denervation in aging muscle [3]. Although there are many reasons fiber atrophy could occur, its sporadic nature (atrophied fibers interspersed amongst relatively T tests were performed. P < 0.05 was considered as statistically significant normal sized fibers) in VO muscle is also seen in neuromuscular diseases that exhibit denervation, such as amyotrophic lateral sclerosis and spinal muscular atrophy [22]. Further, to this point, although MHC coexpression can also occur in the context of sedentary lifestyle with aging and is reduced in response to exercise training [23], a distinguishing feature in VO rat muscle is that MHC co-expression most often colocalizes with other features of denervation, including small and angular fiber size and Nav 1.5 expression [3]. Furthermore, VO rat muscle exhibited marked fiber type grouping (indicative of many repeating cycles of denervation and reinnervation [9]) and large induction of transcripts known to respond to denervation. In contrast, despite significant fiber type grouping in Sarco mouse muscle and a high abundance of MHC co-expressing fibers (suggesting that, similar to normal aging, Sarco mice also undergo many repeating cycles of denervation and reinnervation), the abundance of small fibers and induction of denervation-related transcripts compared to WT mice was much smaller than seen with aging. This analysis suggests a robust reinnervation response in young muscle (evidenced by the Sarco mouse) and in turn a poor reinnervation response of aging muscle. In identifying a possible mechanism for failed reinnervation, our analysis suggests that an insufficient neurotrophin response in aging muscle may contribute to the persistence of denervated muscle fibers and that this may be attributable to increases in miRNAs that suppress neurotrophin expression in aging muscle. The translational significance of these observations is that they suggest that up-regulation of neurotrophin signaling could be an effective therapeutic approach for attenuating muscle atrophy by facilitating reinnervation in advanced age where muscle impact is clinically meaningful. The evolving causes of atrophy in aging muscle Whereas at ages ≤75 years old muscle fibers atrophy relatively uniformly [24], in more advanced ages where the severity of muscle atrophy is more likely to have clinical impact [2], the muscle morphology is punctuated by a marked accumulation of small angular fibers that are interspersed amongst relatively normal-looking muscle fibers [4]. Identification of processes that can lead to sporadic fiber atrophy in aging muscle is thus key to understanding the causes of muscle atrophy in advanced age. Although segmental accumulation of mitochondrial defects has been considered a cause of muscle fiber atrophy with aging [25], the fraction of fibers harboring mitochondrial defects that exhibit atrophy is very small (5 %) [5,26] and thus unlikely to represent a key target. On the other hand, we have shown that there is a marked accumulation of small angular fibers that coincides with the acceleration of muscle atrophy in both slow and fast twitch muscles in advanced aging [5] and that >90 % of these fibers label positively for a sodium channel isoform, Nav 1.5 [3], that is upregulated in response to denervation [27]. It is for this reason that we have focused the current study on identifying mechanisms that could account for the accumulation of persistently denervated muscle fibers in aging muscle. Sarco mouse as a model of sporadic denervation In our studies, we used a genetic model of sporadic denervation, the Sarco mouse [28]. This model was engineered to over-express the endogenous protease, neurotrypsin, in neurons which in turn causes increased agrin cleavage at the neuromuscular junction and reduces activation of MuSK [7]. The result is less recruitment of rapsyn to the subsynaptic microdomain, a resulting loss of cytoskeletal tethering of the postjunctional AChRs, and thus, an increased susceptibility to sporadic denervation. Although we present data from both gastrocnemius and soleus muscles, our analysis shows that for the parameters examined, the changes are generally greater in the soleus muscle. For example, there was a greater increase in fraction of very small muscle fibers and larger changes in denervation-related transcripts in the relatively slow twitch postural soleus muscle compared to the largely fast twitch gastrocnemius muscle in the Sarco mice. Thus, whilst fiber type and muscle recruitment are clearly different between rat VO and mouse Soleus muscles, the Sarco mouse soleus is more affected and might on this basis provide a better indicator of the degree to which reinnervation is compromised with aging. Consistent with prior studies using this model [7,28], the Sarco mice exhibited smaller muscle mass and a pronounced fiber type shift punctuated by a large accumulation of MHC co-expressing fibers. Furthermore, phospho-MuSK levels at the NMJ were seen to decrease (Fig. 2a), as would be expected in response to reduced agrin activation of MuSK [7], and this was accompanied by a reduced rapsyn and AChR intensity in the post-synaptic endplate region. In addition, we observed marked fiber type grouping (only examined in soleus muscle) and moderate increases in some denervation-related transcripts (MuSK in both muscles, and AChRγ and cPLA2 in soleus muscle). We also noted a nearly 60 % decrease in AChR ε subunit in Sarco soleus muscle, which contrasts with the increase in AChR ε that was seen in VO rat muscle. In this respect, previous studies have shown that following denervation AChR ε expression undergoes a multiphasic response: (i) an initial decrease in the first 10 day post-denervation [29]; (ii) a sharp increase observed approximately 4 weeks following denervation; (iii) a decline over the next 4 weeks; and (iv) a gradual secondary increase when denervation persists beyond 2 months [30,31]. Thus, the differences between the VO rat and Sarco mouse likely reflect differences in the duration of denervation, with the higher AChR ε in VO rat reflecting more prolonged denervation and the reduced AChR ε in Sarco mouse reflecting the acute denervation response. A caveat to comparing the Sarco mouse to aging muscle concerns the frequency of sporadic denervation events. The large increase in frequency of enclosed fibers (Fig. 1c) shows both are exposed to many recurring cycles of denervation and reinnervation. Accepting that it is not possible to obtain a quantitative comparison of the exact frequency of denervation (e.g., on a daily or weekly basis) events in the two models, the value of the Sarco mouse herein is that it provides an indicator of the normal capacity for reinnervation of young adult muscle in the context of significant neuromuscular junction instability, and as such, suggests a compromised reinnervation capacity of VO muscle. On the other hand, we cannot rule out that the aging muscle might undergo a higher frequency of denervation events, which is itself the problem (rather than failed reinnervation per se). As noted above, however, the AChR ε response in the VO rat is more consistent with prolonged persistent denervation rather than a greater burden of denervation per se. Causes of denervation and failed reinnervation in aging muscle Skeletal muscle undergoes repeating cycles of denervation and reinnervation throughout much of the adult life, resulting in remodeling of the aging motor unit that in turn causes significant grouping of fibers of the same type [9]. The causes of individual denervation events with aging may involve gradual changes in the signaling pathways necessary for NMJ maintenance and/or a sudden change such as muscle injury or motor neuron loss [32]. In our aging rat model, we have previously shown in the same animals studied here that there is a 27 % loss of motor neuron cell bodies in the spinal cord [3], indicating a significant pre-synaptic element causing denervation in VO rat muscle. This is in contrast to the Sarco mouse where the defect is downstream of the motor neuron and where motor neuron number in the spinal cord is maintained [7]. Thus, although both models exhibit denervation, the causes may be quite different. This notion is supported by the fact that in our study, phospho-MuSK was maintained in aging muscle (decreased in Sarco mouse), rapsyn was increased (reduced in Sarco mouse), and this resulted in only a small loss of AChR intensity in the post-synaptic endplate region of aging muscle (much larger reduction in AChR intensity in Sarco mouse). Consistent with other studies that have examined transcriptional responses in aging muscle [19,33], in VO rat muscle, we observed increases in several transcripts that change with denervation. The large magnitude of these changes underscores the very significant denervation stress present in the VO muscle, and this contrasts markedly with the much smaller changes seen in Sarco mouse muscles. Notably, previous studies have shown that the neurotrophins BDNF, NGF, and GDNF are important to promote motor axon sprouting and recovery of innervation following surgical denervation [12,13,34]. Furthermore, interventions that increase signaling through BDNF and/or its receptor TrkB markedly improve recovery of diaphragm function following cervical spinal cord injury [35][36][37]. Yet, we observed largely non-significant increases in the neurotrophin and neurotrophin receptor transcripts in VO rat muscle which could indicate a failed response to the marked presence of sporadically denervated fibers in aging muscle. This notion is consistent with previous studies showing impaired BNDF signaling in aging limb muscle [38] and diaphragm [39]. In contrast to aging muscle, the fact that motor neurons are spared in neurotrypsin-overexpressing mice [7] would obligate less extensive axonal sprouting to facilitate reinnervation in this context. Thus, the negligible changes in neurotrophins and their receptors in Sarco mouse muscles are not unexpected. It follows that a significant limitation to our interpretation of the neurotrophin response in VO muscle is that we do not have a model to indicate the "normal" adult neurotrophin response to sporadic motor neuron loss. Such information could be obtained from either experimental models involving a less severe form of denervation involving partial nerve crush or a genetic model involving motor neuron loss. Regardless of this limitation, on the basis of the aforementioned evidence showing the importance of neurotrophins in promoting axonal sprouting and reinnervation following experimental denervation, therapeutically targeting the neurotrophins in advanced age is worthy of evaluation. Potential involvement of miRNA-induced suppression of neurotrophins There has been remarkable progress in recent years in identification of miRNAs, opening up new avenues of research into mechanisms of disease [40,41], as well as mechanisms that can contribute to muscle impairment in various disorders [42,43]. Although it is well established that miRNAs can suppress gene expression by pairing with target mRNAs to direct their degradation or influence the mRNA translation capacity [44], very few studies have examined miRNAs in the context of aging muscle [45,46], and none specifically in the context of the causes of persistent denervation. Thus, in seeking mechanisms that might account for a blunted neurotrophin response to denervation in aging muscle, we first used TargetScan to identify potential miRNAs that could suppress neurotrophins. This analysis identified 10 miRNAs that are conserved in humans and which target three neurotrophins involved in promoting axonal sprouting following nerve injury. For BDNF, our analysis identified increases in miR-206-3p, miR-195-5p, and miR-497-5p in VO rat muscle. Similarly, for NT3, our analysis identified significant increases in miR-21-5p and miR-221-3P in VO rat muscle. Finally, for NGF, our analysis revealed an increase in both let-7b-5p and miR-98-5p in VO rat muscle. Of these, miR-206 has previously been shown to increase in aging mouse muscle [46], consistent with our analysis in VO rat muscle. Similarly, miR-195 (predicted to target BDNF), and let-7b and miR-98 (both predicted to target NGF) have been shown previously in human muscle to increase with aging [45], consistent with our analysis in VO rat muscle. To our knowledge, none of the other miRNAs that we identified as potentially suppressing neurotrophins have been examined previously in the context of aging muscle. In contrast to the aforementioned studies in aging muscle, most of the miRNAs studied herein have been examined in the context of experimental denervation, including miR-206 (increases after reinnervation), miR-10a-5p (increases four-to seven-fold with denervation), miR-1 (increases up to 10-fold following denervation and remains elevated after reinnervation), miR-195 (increases up to 10-fold with denervation), miR-21 (increases with denervation), miR-221 (no consistent change), miR-222 (no consistent change), and miR-98 (increases up to 10-fold with denervation) [43,47,48]. Note that we also saw an increase in miR-98-5p in Sarco mouse muscle, which likely reflects a normal response to denervation. Thus, the normal response to experimental denervation, and in the current study, a model of NMJ instability (Sarco mouse), is often an increase in miRNAs that are predicted to target neurotrophin genes. The miRNAs that exhibited a different response in VO rat muscle from what is seen with experimental denervation are miR-1, which in experimental denervation increases [48] but in VO rat muscle was observed to decline, and miR-221, which does not change with denervation [47] but was observed to increase in VO rat muscle. Since miRNAs frequently target more than a single mRNA species, a simple interpretation of the biological significance of the changes observed is challenging. Thus, whether miRNA sponges [41] or other approaches to silencing miRNA signaling would be effective in promoting reinnervation in aging muscle by amplifying the neurotrophin response, will need experimental validation. Conclusions The purpose of our study was to evaluate whether the marked accumulation of persistent denervated fibers in muscle at advanced age is due to failed reinnervation. Consistent with our hypothesis, VO rat muscle exhibited a much greater accumulation of very small muscle fibers and denervation-related transcripts than Sarco mouse muscle, even though both models evidenced marked fiber type grouping indicative of recurring cycles of denervation and reinnervation. In spite of this significant denervation stimulus in VO rat muscle, the neurotrophin pathways known to drive axonal sprouting following acute nerve injury exhibited a minimal induction (with most changes not reaching statistical significance). We conclude that failed reinnervation is an important mechanism driving the acceleration of muscle atrophy in advanced age and that therapeutically targeting the neurotrophin response to stimulate axonal sprouting would be worthy of exploration. Experimental procedures Animals and tissue collection We studied 8-month-old (young adult; YA; n = 8) and 35-36-month-old (very old; VO; n = 10) male Fischer 344 X Brown Norway F1-hybrid (F344BN) rats obtained from the National Institute on Aging (USA) colony. Upon arrival at the University of Calgary Biological Sciences vivarium, the animals were housed individually in cages fitted with filter bonnets and maintained in a 12:12-h light/dark cycle at 23°C for at least 48 h before experiments. On the day of experiments, the animals were anesthetized with 55-65 mg kg −1 of sodium pentobarbital and the left vastus lateralis muscle was removed, trimmed of adipose and connective tissue, and weighed. A slice through the entire midbelly of the vastus lateralis muscle was mounted transversely on pieces of cork using optimal cutting temperature compound and frozen in isopentane pre-cooled in liquid nitrogen before being stored at −80°C. Remaining muscle portions were frozen in liquid nitrogen for RNA and microRNA measurements and stored at −80°C. All procedures with F344BN rats were conducted with approval from the Animal Care Committee at the University of Calgary (ID B109R-11). Neurotrypsin over-expressing C57BL/6 mice were provided by Neurotune, Switzerland, and bred at the Research Institute of the McGill University Health Centre vivarium. All of the mice were maintained in a heterozygous state and were backcrossed with wild-type C57BL/6 mice. In all experiments, 8-month-old mice were used to study the impact of sporadic denervation and the fidelity of reinnervation in young adult muscle. On the day of the experiment, the animals were sacrificed with CO 2 asphyxiation followed by cervical dislocation. The right soleus and gastrocnemius muscles were dissected free of adipose and connective tissue and weighed, and a portion of each was mounted in tragacanth gum and frozen in isopentane precooled in liquid nitrogen. The remaining portion of each muscle was frozen in liquid nitrogen for mRNA and miRNA (gastrocnemius muscle only in Sarco mouse) analyses. All experimental procedures with mice were conducted with approval from the Animal Care Committee at McGill University Health Centre. Muscle sectioning Ten-micrometer-thick cross-sections were cut using a Leica CM-3050-S cryostat (−20°C) and mounted onto glass slides. The sections were left to air-dry for 2 h before being stored at −80°C until use. Prior to labeling protocols (see below), the sections were defrosted in slide boxes for 30 min and allowed to air dry. Fiber type, size, and proportion assessments Muscle cross-sections were immunolabeled for myosin heavy chains (MHCs) I, IIa, IIx, and IIb by a previously described method [24]. The sections were rehydrated with PBS (pH 7.2). These sections were blocked using goat serum (10 % Due to the presence of four MHC isoforms in rodent muscle, labeling was performed on two serial sections (section 1: three MHC antibodies + laminin, section 2: one MHC antibody + laminin). Muscle cross-sections were then washed three times in PBS, and coverslips were applied to slides using Prolong Gold (ThermoFisher Scientific; P36930) as mounting medium. The slides were imaged with a Zeiss fluorescence microscope (Axio Imager, Zeiss, Oberkochen, Germany). Frames for analysis were systematically sampled across each muscle section. The frequency of very small fibers was assessed as an index of the degree of persistent denervation in both aging rat and Sarco mice, based upon previous data from our group showing that fibers of this size in VO rat gastrocnemius muscle exhibit >90 % positive labeling for a marker of denervation, the sodium channel isoform Nav 1.5 [3]. Specifically, we used the fiber size corresponding to the first percentile in YA (for rat) or WT (for mice) animals to define very small fiber size. The first percentile was determined by calculating the size distribution of all fibers for all animals in either YA rats or WT mice. In VO rat VL muscle, this corresponded to fibers with a size of ≤1000 μm 2 , whereas in Sarco mouse, this corresponded to fibers with a size of 650 μm 2 in gastrocnemius muscle and ≤575 μm 2 in soleus muscle. Neuromuscular junction labeling We first labeled sections to identify levels of activated MuSK based upon phospho-MuSK staining intensity. Tissue cross-sections of rat VL and mouse gastrocnemius muscle were rehydrated with PBS and then fixed in 2 % paraformaldehyde containing Phosphatase Inhibitor Cocktail tablet for 10 min at room temperature and then washed (4 × 5 min PBS, 1× PBS) before being incubated in permeabilization solution (0.2 % Tween 20 in PBS) at room temperature for 10 min. The sections were washed again and incubated in blocking solution (10 % NGS) for 60 min. Incubation with anti-MuSK (phospho Y755) antibody (1:250 in blocking buffer; Rabbit polyclonal, Abcam; ab192583) was performed overnight at 4°C in a humidified chamber. The slides were washed and then incubated with AF594 IgG goat anti-rabbit (Thermo-Fisher Scientific, A11037; 1:100 in blocking buffer) antibody for 1 h at room temperature. The sections were washed a final time and then mounted with ProLong Gold Antifade Mountant (ThermoFischer Scientific). Images were captured on an Axio Imager M2 (Zeiss) microscope using ×40 objective. We next examined the downstream impact of alterations in MuSK activation by examining muscle endplate levels of the AChR tethering protein, rapsyn, and AChR intensity levels. Tissue cross-sections of rat VL and mouse gastrocnemius muscle were fixed in acetone for 15 min at 4°C then washed (2 × 5 min PBST, 1× PBS) before being incubated in permeabilization solution (0.1 % Triton in PBS) at room temperature for 15 min. The sections were washed again and incubated in blocking solution (1 % NGS 5 % BSA) for 30 min. Incubation with αbungarotoxin 488 conjugate (ThermoFisher Scientific; B13422) and rapsyn (mouse monoclonal, Abcam; Ab11423) was performed at room temperature for 2 h in a humidified chamber, where all dilutions were made up in blocking solution. The slides were washed, and AF350 IgG1 goat anti-mouse (ThermoFisher Scientific 1:100) was applied for 1 h at room temperature. The sections were washed a final time and then mounted with ProLong Gold Antifade Mountant (ThermoFischer Scientific). Images were captured on an Axio Imager M2 (Zeiss) and then deconvoluted on Autoquant X3 using the "blind deconvolution" setting and image parameters provided by the Axio imager M2 metadata. Analysis of NMJ labeling was performed using ImageJ (National Institutes of Health, Bethesda), with NMJs being traced semi-automatically and the investigator blinded to the group status of the animal. Gene expression Total RNA was extracted from YA and VO rat vastus lateralis muscle samples using the RNeasy Lipid Tissue Mini Kit (Qiagen; 74804). Quantification and purity of total RNA was assessed with a nanodrop by examining the A260/A280 absorption ratio. RNA (1 μg) was reverse transcribed to cDNA using qScript™ cDNA Synthesis Kit (Quanta biosciences; 95047-025), according to the manufacturer's instructions. Real-time PCR was performed using a StepOnePlus™ Real-Time PCR system (ThermoFisher Scientific) to quantify the different denervation-related genes (Musk, AChRα, AChRγ, cPLA2, NCAM, and RUNX1) (Additional file 1: Table S1) and neurotrophin-related genes (BDNF, NGF, GDNF, CNTF, p75NTR, NTRK1, NTRK2, NTRK3, GFRA1, CNTFR) (Additional file 1: Table S2). TATA box binding protein (TBP) (endogenous control) was obtained from RealTime-Primers.com. Primers were designed with a freely available software, Primer 3 plus. Power SYBR® Green PCR Master Mix (ThermoFisher Scientific; 4367659) was used to quantify the mRNA. The thermal profile was 95°C for 10 min and 40 cycles of 95°C for 15 s and 55-60°C for 60 s. All real-time PCR experiments were performed in triplicate, and melt curve analysis for each PCR experiment was performed to assess primer dimer formation or contamination. The comparative threshold cycle (CT) method was used to calculate fold changes in expression. Relative foldchanges in gene expression were presented as 2−ΔΔCT and normalized to YA (F344BN rats) or WT (Sarco mice) values. MicroRNA analysis Freely available target scan software (www.targetscan.org) was used for predicting microRNAs targeting BDNF, NT3, and NGF genes. Subsequently, we used the software miRBase (www.mirbase.org) to obtain the sequence of the different miRNAs predicted to target these neurotrophin genes. Total RNA was extracted from both YA and VO rat VL muscle using mirVana™ miRNA Isolation Kit, with phenol (Thermo Fisher Scientific Inc; AM1560), according to the manufacturers' protocols. MiRNAs were detected using an NCodeTM miRNA qRT-PCR Kit (ThermoFisher Scientific; 45-6612) and real-time PCR (StepOnePlus™ Real-Time PCR system, Thermal Cycling Block; ThermoFisher Scientific) with specific primers (Additional file 1: Table S3), Platinum® SYBR® Green qPCR SuperMix-UDG (ThermoFisher Scientific; 11733-038). All experiments were performed in triplicate. Relative miRNA expression was determined using the CT method, where CT values of individual miRNA data were normalized to CT values of miR-191a-5p. (Fig. 5) Statistical analysis GRAPHPAD PRISM (GraphPad Software, Inc. La Jolla, CA, USA) software was used for statistical analysis. All statistical analyses involved an unpaired two sided T test when comparing two groups. Fiber type distribution was analyzed by two-way ANOVA (fiber type x group), with a Sidak's post hoc test. The outliers were detected by using GraphPad prism Grubbs' test. P < 0.05 was considered as statistically significant. Additional file Additional file 1: Table S1. DNA sequences and primers for qPCR analyses of neurotrophins and neurotrophin receptors in rat skeletal muscle. Table S2. DNA sequences of primers used for qPCR analyses of neurotrophins and neurotrophin receptors in mouse muscle. Table S3. microRNA sequences of primers used to perform QPCR experiments to detect miR levels in rat vastus lateralis and mouse gastrocnemius muscle. (PPTX 1239 kb)	2023-01-24T15:29:11.229Z	2016-09-01T00:00:00.000	{ "year": 2016, "sha1": "9f9f3aa29428927ff4d4b395cf186c59d6f0d3cf", "oa_license": "CCBY", "oa_url": "https://skeletalmusclejournal.biomedcentral.com/track/pdf/10.1186/s13395-016-0101-y", "oa_status": "GOLD", "pdf_src": "SpringerNature", "pdf_hash": "9f9f3aa29428927ff4d4b395cf186c59d6f0d3cf", "s2fieldsofstudy": [ "Biology" ], "extfieldsofstudy": [] }
202177408	pes2o/s2orc	v3-fos-license	Japan's Position in the Context of Agricultural Trade Issues The Doha Round has long been on the rocks. The biggest factor is that the conflicts between developed and developing countries over various themes cannot be resolved. The multilateral negotiation, named formally the Doha Development Agenda, places importance on considerations for developing countries. It is somewhat ironic that the stalemate cannot be broken because developing countries have increasingly powerful voices. The Ministerial Declaration states, “The majority of WTO members are developing countries. We seek to place their needs and interests at the heart of the Work Programme adopted in this Declaration”. Such a stance by the WTO still has not lost its importance in considering future trade negotiations including agricultural products. This paper discusses how research into agricultural economics can contribute to the creation of international rules for future agricultural trade taking into consideration the current framework of the WTO, which has in its scope the issues of developing countries. The author’s past research mainly focused on Japanese agriculture and food, with occasional reference to EU and Japanese agricultural policy. Hence, the paper firstly gives an overview of the position of Japanese agriculture and food from a rather long-term viewpoint. Then it briefly looks back on the postwar history of how the Japanese government dealt with issues of agricultural trade. Based on these, the author gives his view on the appropriate form of agricultural trade negotiations with basic theory of trade in mind. The point is to explore the possibility of applying the framework of the role of a market economy and that of the government to the international community. The history of Japanese agriculture, food and trade negotiations may contain informative points for other countries, especially for Monsoon Asia, on how they deal with agricultural trade. Among them, those countries that are at the initial stage of economic growth may find useful elements in the experiences of Japan. It is needless to say that there should also be experiences that are negative lessons. Introduction The Doha Round has long been on the rocks. The biggest factor is that the conflicts between developed and developing countries over various themes cannot be resolved. The multilateral negotiation, named formally the Doha Development Agenda, places importance on considerations for developing countries. It is somewhat ironic that the stalemate cannot be broken because developing countries have increasingly powerful voices. The Ministerial Declaration states, "The majority of WTO members are developing countries. We seek to place their needs and interests at the heart of the Work Programme adopted in this Declaration". Such a stance by the WTO still has not lost its importance in considering future trade negotiations including agricultural products. This paper discusses how research into agricultural economics can contribute to the creation of international rules for future agricultural trade taking into consideration the current framework of the WTO, which has in its scope the issues of developing countries. The author's past research mainly focused on Japanese agriculture and food, with occasional reference to EU and Japanese agricultural policy. Hence, the paper firstly gives an overview of the position of Japanese agriculture and food from a rather long-term viewpoint. Then it briefly looks back on the postwar history of how the Japanese government dealt with issues of agricultural trade. Based on these, the author gives his view on the appropriate form of agricultural trade negotiations with basic theory of trade in mind. The point is to explore the possibility of applying the framework of the role of a market economy and that of the government to the international community. The history of Japanese agriculture, food and trade negotiations may contain informative points for other countries, especially for Monsoon Asia, on how they deal with agricultural trade. Among them, those countries that are at the initial stage of economic growth may find useful elements in the experiences of Japan. It is needless to say that there should also be experiences that are negative lessons. 1) Japan as a newly developed country Let us clarify the position of Japan as a developed country from an international bird's-eye view. Since actual situations of agriculture and food vary widely among developed countries, they could broadly be classified into two groups, namely, developed countries among EU members 1) and developed countries of European origin on new continents, that is, the United States, Canada, Australia and New Zealand. The first group developed earlier and the second later. The history reflects on their differences in farm land endowment, which basically determines their agricultural competitiveness. Japan is a newly developed country, neither of first or second group. So far it has been the leading runner of economic development in Monsoon Asia. At the same time, there is a trend toward the birth of developed countries in this area that are different from the first and second group. This is the third group, and South Korea and Taiwan are already developed countries in the OECD context. 2) The author mentioned in the introduction that the experiences as a newly developed country may serve as good references for the agricultural trade issues of countries in Monsoon Asia. From this viewpoint, the next section summarizes Japan's experiences since the postwar economic growth era. If we go back to Japanese agricultural history of the Meiji Era, it may provide more lessons for least-developed countries that also exist in Asia. However, the author is not knowledgeable enough about prewar agricultural history to spread the wings of argument so widely. 2) Japanese agriculture, food and society To understand Japan's position in the context of agricultural trade, it is appropriate to apply a two-step approach, that is, to confirm the situations of Japanese agriculture and food as the base and then uncover the characteristics of how Japan dealt with agricultural trade issues, especially, market-opening issues. As for Japanese agriculture and food, we will just confirm the main points which contain not only the characteristics common to Monsoon Asia but also characteristics unique to Japan. 3) Japanese agriculture cannot be discussed uniformly. On one hand, there is paddy farming in which cultivation by part-time farmers is the majority and aging of farmers is pronounced. On the other hand, the sectors like livestock farming and greenhouse horticulture have steadily expanded their size and farm management has been upgraded through incorporation and the introduction of employees. Land using agriculture in Hokkaido, including paddy farming, is dominated by full-time farmers and is different from the situation in the rest of Japan. For the regions with highly aged paddy farmers, it is possible to regard the conditions for structural reform as being ready, after the issue having been postponed for half a century. Regarding Japanese food, consumption of animal protein, fat and oil has increased remarkably with postwar economic growth. The ingredient composition of Japanese diet dramatically changed, partly supported by mass imports of feed grains for livestock and soybeans for oil. Contrasted to this is the consumption of rice as the staple food. It has already decreased to less than half of its peak of 118 kg per capita in 1962. At the same time, most items whose consumption had increased remarkably have already hit the ceiling and are in the declining phase. As for this transition, we should take into consideration factors like recent aging. However, based on Japan's postwar experiences, it may be possible to estimate future saturation level of food consumption of the regions where Asians are the major population. Not only has ingredient composition changed, but also Japanese ways to purchase and eat food; that is, purchase of processed food and eating out have increased. According to the estimation by the Ministry of Agriculture, Forestry and Fisheries based on the 2011 Input Output Table, out of a total expenditure on food and drink of 76 billion yen, fresh food, processed food and eating out account for 16%, 51% and 33%, respectively. From a different point of view, the food industry, which is composed of food processing, food distribution and eating out service, has expanded significantly. However, the dependence on processed food and eating out differs according to whether or not the households are single. Besides, in recent years, food expenditure level has varied significantly according to household income level. 4) As for Japanese eating habits, we could say that there is an increasing tendency of polarization both by household composition and income level. While eating habits and agriculture have changed greatly with the general increase in the income level, there could also be changes in the themes which Japa-1) Among 28 present member countries of EU, 6 countries are not included OECD. Definition of developed country varies among international organizations and OECD member countries might not be developed countries. 2) Taiwan is not OECD member country, but they participate in its activities as an observer. Taiwan's recent per capita income is 70% of that in Japan. 3) It may be too rough to discuss Monsoon Asia as one uniform area. There are differences among countries and, within a country, differences of agricultural areas with old history and newly developed ones. For this point, Hara (2013) is highly suggestive. 4) For the realities of Japan's eating habits from the viewpoint of age groups and income levels, Ishibashi (2006) and Kusakari's series of study (Kusakari, 2011, etc.) carried out interesting analysis. nese society is interested in. One of the changes is a spread of interest or concern over income disparity. Recently, the use of the phrase 'income gap' has clearly increased. Another change is a rising interest in byproducts and side effects that are transferred outside the market, that is, external economy and diseconomy. Regarding agriculture, multi-functional roles of agriculture have been referred to frequently. This is external economy which also has a characteristic of public goods with joint consumption and non-excludability. Using a theoretical model of interpretation, increase in income level has lowered marginal utility of goods and services purchased in the market and has raised marginal utility of byproducts and side effects transferred outside the market. 3) Framework of agricultural trade negotiations The transition of issues postwar Japanese agriculture faced related to trade liberalization corresponded to the changing trends in global trade liberalization. Japan joined GATT in 1955, the year of the turning point from postwar recovery period to high economic growth era. After that, starting from the liberalization of 121 items such as rye and coffee beans in 1960, that is, removal of quantitative import restrictions, a considerable number of items had been liberalized by early 1970s. 5) Not only niche agricultural products, but also products that are familiar in the ordinary Japanese diet, like soybeans, onions, chicken, bananas, grapefruit and pork, were included. However, trade liberalization then didn't provoke a big political issue. The whole situation changed after the United States requested the market opening of beef and oranges. Through the first negotiation in 1977 to the third one in 1988, expansion or removal of import quotas, tariffication and reduction of tariffs were agreed on. The issue of beef and oranges was apparently related to the trade friction over the sharp increase in Japan's export of automobiles and other manufactured goods, mass media such as newspaper and TV reported widely. Nevertheless, the form of negotiation persisted as a bilateral one on specific agricultural items. The Uruguay Round, started in 1986, fundamentally changed the structure of agricultural trade as it was a multilateral full-scaled negotiation covering both agricultural and non-agricultural products. 1) Inter-industry adjustment Economic views on international trade are diversi-fied even within those deemed classic. Two leading theorists of classical economics, Ricardo and Malthus, also had different views. While Ricardo supported a free trade theory that was based on comparative production cost, Malthus stayed on the opposite side regarding the elimination of the Corn Laws. As a criticism of simple free trade theory, there was an argument by Friedrich List based on a long-term view. He was a well-known historian in Germany, which was then a less-developed country. Although their arguments were different, what they had in common was that they did not focus on specific goods alone. Since today's trade negotiations cover various goods and services, pursuing multilateral agreements, it might be meaningful to go back to classical theory on trade. One of the points of the classic proposition of free trade is the relation between industries with comparative advantage and those with comparative disadvantage. Under free trade, production factors move from disadvantaged industries to advantaged industries, and comparison between with and without trade shows that trade is beneficial for all countries concerned. In other words, it brings about Pareto improvement. However, most of the models in textbooks do not consider the cost of moving production factors between industries, that is, inter-industry adjustment cost. Or explanation is given on the implicit assumption that the increase in the total net profit is well above the adjustment cost caused by free trade. In the real world, there should be many cases in which the inter-industry adjustment cost cannot be ignored. However, this point has not been sufficiently examined in postwar market opening issues concerning Japanese agricultural products. It is appropriate to understand that the option of scaling down agriculture as a disadvantaged industry was not explicitly proposed, at least in official places. It could be said that successive Japanese governments had the policy, at least as their announced principle, to minimize trade liberalization of agricultural products and lessen the damage on domestic agriculture by adopting various measures against worsened competitive conditions due to the change in their border situation. Therefore, they have allocated policy resources for the areas of corresponding products, often with financial expenditure. 2) Distribution adjustment between benefit and loss In regard to such situations, it would be useful to discuss relying on a more general framework of profit 5) 'Annual reports during last 50 years' attached to '2010 White paper on food, agriculture and rural areas' shows a brief summary of history of postwar import liberalization of agricultural products. re-distribution. More specifically, an examination using the framework of compensation principle by Kardor-Hicks is recommended. This is the standard of judgment whether to accept an institutional change when there are both people who gain profits and those who suffer losses. The change is acceptable if the compensation for the losses by re-distribution of additional profit brings about Pareto improvement. It is a very normal situation that an institutional change generates both areas that gain profit and those that suffer losses. However, the viewing angle required for the analysis of agricultural products is not so simple. Although there is a basic structure of re-distribution between the farmers of the agricultural product of concern and the companies or regions that benefit from the free trade, the actual method of re-distribution would be through measures adopted by the government. Moreover, input of resources could be merely compensation in response to the loss of farmers or could be a way of strengthening measures for their long-term competitiveness. In the case of the latter, it is possible that the resources once allocated to the agricultural sector would be returned to consumers through price decline of the products with lower cost. In addition, today's eating habits make the issue more complicated. While the food industry has become influential in the areas between agriculture and consumers' dining tables, price change in agricultural products cannot be reflected in retail food prices directly. In addition, as mentioned earlier, there is an increasing polarization of eating habits and therefore consumers cannot be treated in the same way. Moreover, as for influence on food processing industries, we should not discuss this in a unique way. There are cases that are located in rural areas and use domestic materials and cases that are located next to a large seaport and use imported materials. As shown in the domestic situation regarding the Uruguay Round and TPP, recent discussions on Japanese agricultural trade have been heated prior to actual negotiations. The topic was frequently covered by general media like newspapers and TV, not to mention by agriculture-related media which took a stand against free trade. On the other hand, there have not been many ex-post-facto attempts to review the actual impacts of market opening. 6) The ex-post-facto assessment of the government's policy adopted in response to changes of border measures has not been highlighted, with a few exceptions of partial comments. Such a situation suggests the important subject of research which will look back on the past. Surely it is not an easy task. In impact assessment there is difficulty related to the common method of estimation, especially that of separation of changes by free trade-related events from those caused by the general domestic economic situation. Japanese Society Facing Agricultural Trade Issues 1) Third group of developed countries The author grouped developed countries into three types at the beginning of the second section. This grouping reflects their characteristics of import and export of agricultural products, especially basic food represented by cereals, namely, developed countries of new continents of European origin, which are the export side, making use of abundant land (second group), developed countries of the EU which maintain relatively high domestic production supported by longterm protection policy under the Common Agricultural Policy (first group), and newly developed Asian countries such as Japan and Korea, which import large quantities of food (third group). Japan, so far the leading runner of newly developed Asian countries, experienced drastic changes in eating habits supported by imported food and livestock feed. Japan also has been requested to further open the market for agricultural products by developed countries of new continents, especially by the United States. The same structure can be found in Korea and Taiwan, whose self-sufficiency rate of cereals in 2013 were 25% and 20%, respectively, which is below the 28% in Japan ('Self-sufficiency rate of cereals of foreign countries (2013) (trial calculation)' in 2016 Food Balance Sheet by MAFF). Japan's import counterparts of agricultural products are not only developed countries like the United States. Japan imports from such countries as China, Thailand and Brazil, too. In a rather generalized way, it can be said that developed countries of the third group also face the issue of how to deal with requests for market opening raised by developing countries. The awareness of this issue was shared among many countries at the beginning of the Doha Round as it placed importance on issues of developing countries. However, viewpoints to criticize developed countries' protection policy for agriculture in connection with import of agricul-6) As an ex-post examination for trade liberalization, there is an interesting document 'Impact assessment on past import liberalization' submitted on February 2007 by Ministry of Agriculture, Forestry and Fisheries to EPA agricultural working group at the Council on Economic and Fiscal Policy. From a long-term viewpoint, with domestic measures introduced in mind, it attempted to assess impacts on both consumption and production of four items: beef, citrus, apple and cherry. tural products by developed countries had already become an internationally common understanding just before the end of the Uruguay Round. Agenda 21 adopted in 'The United Nations Conference on Environment and Development' in 1992 included the following sentence: In particular, the achievement of this objective requires that there be substantial and progressive reduction in the support and protection of agriculture -covering internal regimes, market access and export subsidies -as well as of industry and other sectors, in order to avoid inflicting large losses on the more efficient producers, especially in developing countries. 2) Food security As the author pointed out, Japan has maintained a cautious attitude towards market opening of agricultural products. While there is a factor like the political power of agricultural organizations, there are also discussions on the positive meaning of maintaining domestic agricultural production from the viewpoint of Japanese society as a whole. Such discussions include some points which put brakes on simple free trade proposals based on the theory of comparative advantage. In other words, the points are not included in the framework of classic trade theory. While the contents of those points have been reflected in the arguments of successive Japanese governments, it is also important to once again examine their validity in the international context. One of the points is the food security issue. Food is a highly selective good in a society like Japan, yet it is still an absolutely necessity, as human beings cannot maintain a healthy life without the minimum amount of it. Since food has two such aspects, it is the government's responsibility of the highest priority to provide systems under which people can secure the necessary amount of food under any circumstances. The author assumes here unexpected situations like international friction and serious natural disasters. One of the issues is to select carefully the specific contents necessary for food security in both quantity and quality and the application of a scheme of agricultural resources for this purpose. We know that these will be quite different from those in daily life where people enjoy highly selective and rich diets. In the context of international validity, we should clarify the difference between the concept of food security which is discussed here and the concept of food security in general. This recognition is quite important for sending out a clear message abroad. Food security exists when all people, at all times, have physical, social and economic access to sufficient, safe and nutritious food to meet their dietary needs and food preferences for an active and healthy life. This is the definition of food security by the FAO, based on the agreement made in the World Food Summit (agreed at the World Food Summit in 1996 with an addition made at the summit in 2009). As plainly indicated in the world's distribution of undernourished people, food security is basically an issue of the daily diets of the poor in developing countries. At the same time, the definition by the FAO includes food security with which necessary food is available under unexpected contingencies. Food security in abnormal situations can be said to be the subset of food security as a general concept. Furthermore, not only is such a simple categorical clarification necessary, but it is also important to clarify concrete images of food insecurity and measures against it. 7) It may be one of the subjects of agricultural economic research. As for food security in developed countries, due to the nature of the matter, it is important to acquire consensus from the international community, especially from neighbor countries. Developed countries in the third group are close to each other and there are commonalities that people depend on other countries for a large part of their food supply. It is necessary to mutually understand that food security is the basis that supports people's cool judgment and stable behavior. It is important to send out a message worldwide on the importance of food security of developed countries based on such a mutual understanding in neighboring countries. In addition, if agricultural policy that preserves minimum necessary agricultural production or agricultural resources for food security is accepted internationally, it may be linked to the restriction on agricultural protection policy implemented beyond that level. 3) External economy of agricultural production Another point of argument raised as a brake on simple free trade theory for agricultural products is the existence of an external economy. The term used in policy acknowledges the multi-functional roles of agriculture. In Japan the system of direct payment for hilly and mountainous areas was introduced in 2000 to preserve multi-functional roles. For agricultural byproducts with a public good character, we share some common understanding with early developed European 7) This may be one of Doha Round's points of dispute. Regarding procurement of official food stocks which India and other developing countries insist, careful examination is necessary as to how they assume impairment of food security. countries. It may also be possible to have positive discussions on the value of agricultural byproducts with Monsoon Asia, which has also long history of rural areas. At the same time, in the sense of market failure, it is necessary to keep in view the burdens on the environment, that is, the external diseconomy involved in agriculture. How has the presence of external economy and diseconomy been dealt with in agricultural trade negotiations so far? As described in the WTO Agricultural Agreement at the Uruguay Round, issues of externalities have not been dealt with in a way that is directly reflected in the structure of border measures. However, the biggest characteristic of the Uruguay Round related to agriculture is that it imposed regulations on domestic policies (amber box). The purpose of the regulations is to control production-stimulating policy, which was recognized as the background of trade friction over agricultural products. Contrarily, the Agreement also listed policies which can be implemented by the decision of individual countries (green box). For instance, policy for handicapped areas and that for environmental preservation are left to member countries' discretion. Japan's direct payment for hilly and mountainous areas is also green box policy. 8) As for this theme, it is important to compare the WTO with the FTA and EPA. While cases of the FTA and EPA vary widely, it is common with the WTO that an agreement covers both agricultural and nonagricultural products. In EPA negotiations, a change in domestic regulations concerning business activities often becomes the focal point. However, so far, the case of a concrete agreement based on agricultural externality has never become their topic. This issue appears to have two aspects. One is that it may be regarded to have little meaning to set additional regulations as the WTO agricultural agreement is still effective for a wide range of countries. The other aspect is that even if the EPA agrees on some regulation that differs from the WTO agreement, its effectiveness is limited to specific countries under the treaty. There might also be producers who trade with countries outside the treaty in concern. The question is whether it is possible or not to impose rules on agricultural policy which uniformly regulate producers in the country. If possible, what form would it take? This question is worth brainstorming. Conclusion Economists whose object of analysis is market economy have a perception that correction for market failure is the role of government. While we have such an understanding for the society of a country, the appropriate institution to correct market failure in the international community is under question. The WTO agricultural agreement, though by an indirect method, had stepped into this area. On the other hand, we have to pay attention to the movement of international organizations other than the WTO. For the issue of external economy or diseconomy, there is the United Nations Framework Convention on Climate Change. Today's developed countries, while at different levels, are concerned about the expansion of income gap and adopt policies in view of income re-distribution. Internationally, various forms of international cooperation support improvement for developing countries. The ILO, which aims to improve people's working conditions, will soon celebrate its 100th anniversary. However, the international community does not have a direct means to correct income gap between countries or regions. On this point, expansion of the FTA or EPA has aspects to be concerned about. We cannot deny the possibility that trade diversion effects between countries under treaty and those out of treaty will lead to an increase in the gap. We also have to consider the situation where there are countries or regions left out of the FTA and EPA. Although an ex-post-facto correction system of income disparity is difficult in the international community, the viewpoint is becoming increasingly important to provide a brake on trade circumstances that lead to a widening of the gap. Market economy is not almighty. However, it is a very useful system. Such understanding is shared by many societies in developed countries. In other words, market economy is a system which we should utilize wisely. It may be expected to spread such a stance from a single country's economy to the economy of the global society.	2019-09-10T20:24:08.851Z	2019-03-31T00:00:00.000	{ "year": 2019, "sha1": "2504e46d1d67bad5de5ab0ee0528bfe29e5b7892", "oa_license": null, "oa_url": "https://www.jstage.jst.go.jp/article/jjae/21/0/21_56/_pdf", "oa_status": "GOLD", "pdf_src": "Adhoc", "pdf_hash": "c9e1b6f4432ec104499326a1cb124ddff819f2ef", "s2fieldsofstudy": [ "Economics" ], "extfieldsofstudy": [ "Business" ] }
244881469	pes2o/s2orc	v3-fos-license	Construction of a Triple Gene-Deleted Live Vaccine Candidate against Pseudorabies Virus Using CRISPR/Cas9 and Cre/LoxP System-Based Strategy Received: Revised: Accepted: Published online: April 03, 2021 June 09, 2021 June 25, 2021 July 24, 2021 The immune protective effects of pseudorabies virus (PRV) vaccines in the market have gradually reduced and they have failed to provide complete protection against the new PRV variant. In this study, a triple gene-deleted live PRV strain–rZDΔTKgE-gI was successfully constructed by simultaneously knocking out the three major virulence genes (gE/gI and TK) with CRISPR/Cas9 and Cre/LoxP gene editing system and low melting point agarose purification method. After challenge with the virulent PRV variant, all the 3-week-old piglets vaccinated with the rZDΔTK-gE-gI PRV vaccine candidate survived without any clinical symptoms, whereas all the unvaccinated piglets exhibited PRV respiratory and neurological signs with 100% mortality rate within 7 days post infection. High levels of anti-gB antibodies were induced in the vaccinated piglets after vaccination with the rZDΔTK-gE-gI vaccine candidate, which elicited a better immune protective effect than the classical strain Bartha-K61. Therefore, the triple gene-deleted live PRV vaccine candidate is expected to control the current outbreak of pseudorabies caused by the PRV variants. ©2021 PVJ. All rights reserved INTRODUCTION Pseudorabies (PR) is an acute infectious disease caused by pseudorabies virus (PRV) in many domestic and wild animals (Fan et al., 2016;Papageorgiou et al., 2016;Yu et al., 2016;Masot et al., 2017;Sun et al., 2018;Wu et al., 2018). PRV belongs to the Family Herpesviridae, subfamily Alphaherpesvirus and genus Varicellovirus. The main clinical symptoms are fever, itching, encephalomyelitis, respiratory and nervous system disorders. Inactivated or attenuated live vaccines can be used to control pseudorabies in pigs. Classical attenuated PRV vaccines developed from field virulent PRV strains by deleting one or more genes have been a new tool for disease control. Firstly, the gene engineered live virus (labeled) vaccine with missing virulence determining gene is constructed, and then wild virus infection and labeled vaccine vaccination can be differentiated by the serological identification using the corresponding serological diagnostic methods. The foundation of this process is the molecular characteristics of PRV and the identification of nonessential envelope glycoproteins, such as glycoprotein gE, which can be deleted or removed from the virus without affecting the replication or immunogenicity of the virus to achieve the purpose of differential diagnosis (Ren et al., 2018) . The practice of vaccination to eradicate PR in pig farms has been successful in several countries, including Germany and the United States. Over the years, the new epidemic situation of pseudorabies has occurred and prevailed in many areas of China; it can be more severe in some provinces, with a large number of infected pig farms and huge loss. The antigenicity of the new epidemic strain has changed, and the pathogenicity of the new virus strain has increased. The existing attenuated vaccine like Bartha-K61 cannot fully prevent pigs from the infection of the epidemic variants (Hu et al., 2015). The PRV variants were found to be more virulent than the classical PRVs in piglets of different ages (Tong et al., 2015;Yang et al., 2016). Therefore, it is necessary to develop a new vaccine against PRV variant strain. With the continuous progress and advance of molecular biology and gene engineering technology, the molecular biological characteristics and gene functions of RESEARCH ARTICLE PRV have been completely analyzed. Lack of one or more non-essential virus replication genes can greatly reduce the virulence of PRV, but it does not affect its proliferation and immunogenicity (Wang et al., 2015;Cong et al., 2016) . CRISPR/CAS system, an innovative and powerful genome editing tool, has been widely used in many research fields. The endonuclease Cas9 protein identifies specific genomic sites through guide RNA (gRNA) and cleaves double stranded DNA (Ma et al., 2014) . Cells then repair the cleaved sites by nonhomologous end joining (NHEJ) or homologous recombination (HR) to achieve DNA level gene knockout or precise editing. CRISPR/CAS system uses the interaction between RNA and DNA to locate gene in the genome. Only a small RNA sequence is needed to anchor the target gene position, which greatly reduces the workload of gene editing and has high specificity. In addition, Cre LoxP system is a recombinant enzyme system, which can control the occurrence of site-specific recombination in the genomic DNA (Van Duyne, 2015) . It is widely used in gene knockout, insertion, turnover and translocation of specific sites, and can achieve the goal of targeted genetic transformation of organisms at the gene level. It is mainly composed of Cre and LoxP. Cre recombinase can effectively knock out the sequence between two LoxP loci when they are located in the same direction on a DNA strand. The purpose of this study was to construct a live gene deleted vaccine against pseudorabies virus with a newly isolated PRV using CRISPR/Cas9 and Cre/Lox systems. MATERIALS AND METHODS Ethics statement: Animal experiments were approved by Animal Ethics Committee of Harbin Veterinary Research Institute of the Chinese Academy of Agricultural Sciences and the Animal Ethics Committee approval number was Heilongjiang-SYXK 2015-005. Isolation and identification of virulent PRV strain: A PRV variant was isolated and identified from a PRV Bartha-K61 vaccinated pig farm in Heilongjiang Province in China in 2017. The PRV isolate was cultured in PK15 cells and named Strain PRV-ZD. PRV-SC strain is the standard classical PRV virulent strain. PRV Bartha-K61 is the classical PRV attenuated vaccine. Constructions of recombinant plasmid, preparation of virus genome and PCR amplification: Based on the PRV sequence (GenBank accession number: KJ789182.1), six pairs of primers were designed to amplify target genes and construct the vector. The guide RNA of TK/gE/gI genes of PRV was cloned into sgRNA / Cas9n expression vector (Addgene, Beijing, China) homologous arms of TK/gE/gI gene were amplified in vitro and LoxP, and Lox2272 were added, respectively. The expression frame containing eGFP, RFP, TK/gE/gI homologous arms was constructed by overlapping PCR using TK l-arm and TK r-arm (green), and gI/gE l-arm and gI/gE r-arm (red, two genes). The whole genome of PRV was extracted by TIANamp Virus DNA/RNA Kit (TIAGEN Biotech, Beijing, China). All the primer sequences are shown in Table 1. PRV gene knockout flowchart is presented in Fig. 1. Cell culture, vector transfection and virus infection: As shown in Fig. 1, PK15 cells were co-transfected by Lipofectamine®2000 Kit (Thermo Fisher, Carsbad, CA) with recombinant plasmids, eGFP DNA fragment containing homologous arms of TK gene and RFP DNA fragment containing homologous arms of gI/gE gene. Transfected cells were inoculated with PRV-ZD at 8 hours after transfection. The infected cells were cultured in the DMEM solution containing 2% FBS. Identification and purification of gene deletion virus: Cells containing two kinds of fluorescent proteins of PRV were selected by fluorescent microscopy and continuous plaque purification. The two kinds of fluorescent protein genes were knocked out by Cre/Lox system, and the cells containing PRV without fluorescence were obtained by low melting point agarose plaque purification to produce a triple gene-deleted live PRV strain, rZDΔTK-gE-gI. The genome was then extracted with TIANamp Virus DNA/RNA Kit (TIAGEN Biotech, Beijing, China), amplified by PCR and sequenced to identify the knockout effects of the three genes TK/gE/gI. The homology arms containing TK gene, eGFP gene fragments on the LoxP site (green), homology arms containing the gE/gI gene, the RFP gene fragment at Lox2272 site (red) were constructed by the overlapping PCR method. These gene fragments and corresponding gene knock-out Cas9 recombinant plasmids were co-transfected into PK15 cells. Pseudorabies gene deletion strain was constructed by Cas9 gene editing system and homologous recombination technology, and fluorescent protein genes were introduced into the recombinant plasmid as a reporting system. (Reed, 1938) . Animal experiments: Thirty-five 3-week-old piglets were randomly divided into 7 groups (5 in each experimental group and 5 in the control group), including Bartha-K61/SC, Bartha-K61/PRV-ZD, rZDΔTK-gE-gI/SC, rZDΔTK-gE-gI/PRV-ZD, SC (mock/SC), PRV-ZD (mock/PRV-ZD) and the control group. Each piglet in the immunization groups was intramuscularly inoculated with 10 5 TCID50 vaccine virus whereas each piglet in the control group was intramuscularly inoculated with PBS. Each piglet in Bartha-K61/PRV-ZD, rZDΔTK-gE-gI /PRV-ZD and mock/PRV-ZD was challenged with 2 ml of PRV-ZD containing 1.0 × 10 6 TCID50 on the 14th day after immunization. In Bartha-K61/SC, rZDΔTK-gE-gI/SC and mock/SC piglets, 2 mL of PRV SC containing 1.0 × 10 5 TCID50 was applied for challenge on the 14th day, and each one in the control group was inoculated with 2 ml of PBS. Serum samples and nasal and rectal swabs of piglets were collected for testing after challenge. Serological tests and detection of virus shedding: All serum samples were detected by ELISA Kits for PRV gBspecific and gE-specific antibodies following the manufacturer′s instructions (IDEXX, Maine). The nasal and rectal swabs were treated and inoculated into PK15 cells to detect the virus titer. Statistical analysis: The significant differences of the animal experiments were analyzed using GraphPad Prism (version 5.01) (San Diego, CA). Differences were considered statistically significant when P<0.05. RESULTS Isolation, culture and genome sequence analysis of PRV strain: There was an epidemic of pseudorabies in a PRV Bartha-K61 vaccinated pig farm in Heilongjiang Province in China in 2017, where samples were collected from the diseased animals. The PCR amplification results are shown in Fig. 2. A fragment of the same size as that of the positive control was amplified using the brain tissue, which was 865 bp (Fig 2A). The full-length PCR sequencing analysis of gE gene showed that the isolated strain was the similar to the pseudorabies variant after 2011 ( Fig 2B). The material that was tested PCR positive was inoculated into PK15 cells and cultured by successive passages. The isolate was named strain PRV-ZD. Construction of a PRV rZDΔTK-gE-gI strain by knocking out three virulence related genes using CRISPR/Cas9 and Lox/Cre system: In order to knock out the virulence related genes (TK-gE-gI) of PRV efficiently and swiftly, CRISPR/Cas9 combined with Lox/Cre system was used to knock out three virulence genes at the same time. Meanwhile, two fluorescent proteins and continuous plaque purification were used to obtain the PRV strain with stable deletion of three genes (Fig. 3). In terms of avoiding the false positive signal caused by the expression of fluorescent gene in the homologous recombinant donor template, the expression of GFP and RFP are driven by the endogenous virus promoter, which only occurs after the precise homologous recombination of DNA. In this study, GFP and RFP genes were designed to be located on both sides of LoxP and Lox2272 pairs, separately, to facilitate the removal of fluorescent marker genes. TK, gE and gI are the main virulence-related genes of PRV and the main targets of PRV gene deficient vaccine. The simultaneous deletion of TK, gE and gI genes was realized by the homologous recombination mediated by CRISPR/Cas9 system. Firstly, sgRNA targeting TK, gE and gI genes (designed by web-based tools http://crispr.mit.edu/) were co-transfected into PK15 cells with DNA donors containing fluorescent selection genes and homologous arms. At 8 hours after transfection, the cells were infected with PRV-ZD with different MOI (infection complex) and incubated until the recombinant virus expressing red and green fluorescence was observed (Fig. 3A, 3B and 3C). It was noted that no green/red overlapping cells were observed in the control group (Fig. 3D). Plaques containing two kinds of fluorescence were selected for plaque purification to obtain pure recombinant virus (P2). The purity of the recombinant virus was verified by PCR amplification. As shown in Fig. 4, except for the control, positive control PCR bands were observed in all samples, indicating the integrity of virus DNA, while TK, gE and gI gene amplifications in the control group were still complete (Fig. 3E). Further experiments showed that the triple gene deletion virus was stable for six passages. Observation of clinical symptoms after challenge: In order to evaluate the cross protection of the live vaccine against classical and variant PRV strains, a commercial PRV live vaccine (Bartha-K61) and PRV rZDΔTK-gE-gI were used to inoculate piglets. PRV-SC and PRV-ZD were then used to challenge piglets and the rectal temperature, clinical signs, weight gain and virus titers in the blood of vaccinated piglets 14 days after vaccination were checked. As shown in Fig. 4A, the body temperature of mock/SC group and mock/PRV-ZD group increased to 40.3-41.8℃ on the second day after challenge. Compared with the mock/SC group, the average temperature of mock/PRV-ZD group was higher. It also revealed obvious clinical symptoms of PR, mental depression, respiratory distress, loss of appetite, vomiting, tremor and ataxia, and finally all died within 7 days. All the piglets challenged with PRV-ZD died between 4 and 7 days. All the nonimmunized piglets in the control group challenged with PRV-SC died between 6 and 8 days (Fig. 4B). These results showed that PRV-ZD had higher virulence in 4-week-old piglets. Yet, all piglets in rZDΔTK-gE-gI /PRV-ZD, rZDΔTK-gE-gI/SC and Bartha-K61/SC groups remained healthy like those in the control group and no PR symptoms were detected during the 2-week observation period. Vaccination can significantly shorten the growth stagnation period and decrease the economic loss caused by PRV infection. Thus, we evaluated the effect of the live vaccine on the weight gain of Bartha-K61-and rZDΔTK-gE-gI-immunized piglets after challenge with virulent virus. As shown in Fig. 4C, the weight gains in all the groups were significantly lower than that of the control group (P<0.05). Pigs in rZDΔTK-gE-gI/SC and Bartha-K61/SC groups gained similar weight throughout the experiment (P>0.05). However, the body weight of rZDΔTK-gE-gI /PRV-ZD group was significantly higher than that of Bartha-K61/PRV-ZD group (P<0.05). Virus detection: From day 0 to day 14, nasal and anal swabs were collected from piglets in each group to determine the virus shedding of the experimental animals. The results showed that virus shedding was detected in all piglets, including the controls and the challenged animals. In the four immunization groups, virus shedding started on the second day and lasted for only 4 days, which was shorter than that of the non-immunized piglets, as shown in Fig. 4D. Vaccination of piglets with Bartha-K61 vaccine reduced the viral load after challenge with PRV-ZD. From the second day to the seventh day, the virus shedding content of rZDΔTK-gE-gI/PRV-ZD group was significantly lower than that of Bartha-K61/PRV-ZD group (P<0.05). Consistent results were observed between PRV-ZD/SC and Bartha-K61/SC groups. Antibody response to PRV infection: In order to monitor the levels of PRV-gB specific antibody produced by PRV-ZD and SC strains, serum samples were collected after vaccination and ELISA used to detect the antibodies. The results showed that the piglets immunized with rZDΔTK-gE-gI or Bartha-K61 had the gB-specific antibody on the 3rd day, and the antibody levels of gB increased steadily in all the immunized pigs, and at the 15th day all the infected piglets produced the high level of anti-gB antibody. During the 14-day after immunization, no specific antibody of gB was detected in the control group (Fig. 4E). These results showed that rZDΔTK-gE-gI vaccine candidate could protect pigs from classical PRV and PRV variants. DISCUSSION PR has always been regarded as one of the most imperative infectious diseases in pig industry . Vaccinations with large doses of Bartha-K61 vaccine were an effective method to control the prevalence and spread of PR (Zhou et al., 2017) . This method was very effective, and only a few cases of pseudorabies had been reported. Since 2011, the PRV epidemic successively broke out in a large number of pig farms in many provinces in China, including the largescale pig farms where pigs were immunized with attenuated PRV vaccine, causing huge economic losses. The main manifestations of the infected animals were that the positive rate of gE antibody in the pig population significantly increased, the sows had abortion, weak piglets and stillborn fetuses, and newborn piglets had obvious neurological symptoms and died in a large number. All the piglets in these pig farms were vaccinated with the attenuated PRV gE gene deletion vaccine. PRV could be isolated from dead piglets and aborted fetuses. The preliminary results showed that compared with the previous PRV strains, the new PRV strains had significant variations, and the pathogenicity of the new PRV strains was significantly increased. The existing vaccine cannot effectively protect against the current epidemic strains. Therefore, it is an urgent task to develop a new PR vaccine against the variant. With the advance of molecular biology and gene engineering technology, the molecular biological characteristics and gene functions of PRV have been clearly recognized. The absence of one or more nonessential genes for PRV replication can greatly reduce the virulence of PRV, but at the same time does not affect its proliferation and immunogenicity. Many research results show that PRV gene deletion vaccines have high safety and good immunogenicity and have been widely used in practical production. A TK/gE/gI triple gene-deleted PRV derived from a PRV variant was generated by using bacterial artificial chromosome techniques, which elicited high levels of protective gB-specific antibodies and provided full protection to the viral challenge after vaccination (Zhang et el., 2015). A TK/gE/gI-deleted PRV rZJ01ΔTK/gE/gI was constructed from the gE/gIdeleted PRV, and rZJ01ΔTK/gE/gI provided full protection against a lethal challenge with the PRV ZJ01 variant in pigs and had lower pathogenicity compared with the gE/gI gene-deleted virus strain rZJ01ΔgE/gI (Dong et al., 2017).	2021-12-05T16:19:35.843Z	2021-10-01T00:00:00.000	{ "year": 2021, "sha1": "05146fbdad701e6993a940224941f709f122f62a", "oa_license": "CCBYNC", "oa_url": "https://doi.org/10.29261/pakvetj/2021.053", "oa_status": "GOLD", "pdf_src": "MergedPDFExtraction", "pdf_hash": "89da574eeadf3d38e9fd0b93a118b7e93bb57a2f", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [] }
237197666	pes2o/s2orc	v3-fos-license	The ‘Surprise question’ in heart failure: a prospective cohort study Objective The Surprise Question (SQ) is a prognostic screening tool used to identify patients with limited life expectancy. We assessed the SQ’s performance predicting 1-year mortality among patients in ambulatory heart failure (HF) clinics. We determined that the SQ’s performance changes according to sex and other demographic (age) and clinical characteristics, mainly left ventricular ejection fraction (LVEF) and the New York Heart Association (NYHA) functional classifications. Methods We conducted a prospective cohort study in two HF clinics. To assess the performance of the SQ in predicting 1-year mortality, we calculated the sensitivity, specificity, positive and negative likelihood ratios, and the positive and negative predictive values. To illustrate if the results of the SQ changes the probability that a patient dies within 1 year, we created Fagan’s nomograms. We report the results from the overall sample and for subgroups according to sex, age, LVEF and NYHA functional class. Results We observed that the SQ showed a sensitivity of 85% identifying ambulatory patients with HF who are in the last year of life. We determined that the SQ’s performance predicting 1-year mortality was similar among women and men. The SQ performed better for patients aged under 70 years, for patients with reduced or mildly reduced ejection fraction, and for patients NYHA class III/IV. Conclusions We consider the tool an easy and fast first step to identify patients with HF who might benefit from an advance care planning discussion or a referral to palliative care due to limited life expectancy. INTRODUCTION People living with heart failure (HF), especially those in advanced stages of the disease, might present with uncontrolled symptoms such as shortness of breath, pain, sleep disorders and fatigue 1 ; they also often suffer from comorbidities, such as depression and anxiety. 2 Advance care planning (ACP) has proven to improve the quality of life and patient satisfaction with end-of-life care for patients with HF by promoting their autonomy concerning medical decisions. 3Identifying those with HF who are in the last year of their life is paramount to guide discussions and determine other strategies that are part of ACP, including possible referrals to specialised palliative care (PC). The Surprise Question (SQ) is a prognostic screening tool used to identify patients with limited life expectancy.The Key messages What was already known?⇒ The performance of the Surprise Question (SQ) screening tool to predict 1-year mortality had been assessed among inpatient populations with heart failure (HF) and populations with HF in an emergency setting. What are the new findings? ⇒ The SQ's performance predicting 1-year mortality in an ambulatory setting, a HF clinic, where patients with HF are more stable.⇒ The SQ's performance predicting 1year mortality for a population with HF according to patients' sex, age, New York Heart Association functional class and left ventricular ejection fraction. What is their significance? Clinical: ⇒ The SQ's new psychometric profile allows for determining the appropriateness of its use in clinical practice at HF clinics and its predictive value for subgroup populations. Research: ⇒ This study contributes to filling gaps in the knowledge of the SQ's performance in patients with HF in the ambulatory setting. Original research SQ is a reflective question a physician or other healthcare providers ask themselves about a patient's prognosis: 'would I be surprised if this patient dies within the next 12 months?'An SQ is positive (+SQ) if the healthcare provider's answer is 'no, I would not be surprised'.There are also versions of the SQ within the context of 3 or 6 months.However, 1 year is the most common version.The SQ's performance has been evaluated for patients with both oncological [4][5][6][7] and non-oncological diseases, including chronic kidney disease 8 and chronic obstructive pulmonary disease. 9][10] Despite the heterogeneity among performance measures, two systematic reviews suggest that, in general, the performance of the SQ predicting 1-year mortality is better for patients with oncological disease than for patients with nononcological disease. 10 11Recently, the SQ's performance was assessed for an inpatient population with HF 12 and for a population with HF in an emergency setting. 13he results were promising with a sensitivity of 85% for hospitalised patients and 79% for the emergency setting.However, the SQ's performance in ambulatory settings, such as HF clinics where patients are more stable, is unknown.Furthermore, understanding the SQ's performance for men and women is important to ensure equity in the delivery of ACP and PC.To our knowledge, the SQ's performance for patients with HF, stratified according to the patient's sex, has not yet been reported.Finally, given that HF clinic admission criteria and samples vary, it is important to understand the SQ's performance according to other demographic and clinical characteristics to increase the generalisability of the results.Performance of the SQ is likely to differ across subgroups due to previous knowledge of the staff answering the question.For example, evidence has shown a trend of increasing mortality rates with increasing New York Heart Association (NYHA) stage, 14 increasing age 15 and decreasing left ventricular ejection fraction (LVEF). 16 17herefore, our objectives included (1) assessing the performance of the SQ predicting 1-year mortality among patients in ambulatory HF clinics, and (2) assessing whether performance changes according to sex, age, NYHA classification and LVEF category. METHODS This study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology guidelines. 18 Study design and setting The prospective cohort included 174 ambulatory patients with HF who were recruited from two HF clinics in Medellín, Colombia between November 2017 and November 2018.One-year vital status was determined by consulting the national mortality register in Colombia.Both clinics are part of tertiary care institutions that are referral centres for patients with cardiovascular disease.They offer comprehensive, multidisciplinary care that includes clinical follow-up by HF cardiologists, nursing education and telephone follow-up, cardiac rehabilitation, and a psychoeducational programme for patients and their families. Participants Patients were potentially eligible for the study if they were 18 years or older and existing patients at the HF clinic who had at least two prior consultations.Since the first two consultations provide the cardiologist an opportunity to optimise treatment if necessary and to get to know patients under optimal treatment circumstances according to clinical guidelines, we did not enrol newer HF clinic patients.We enrolled consecutive eligible patients in the study.There were no exclusion criteria. Ethical aspects We conducted the study in accordance with the ethical guidelines from the Declaration of Helsinki 19 ; our study was approved by the research ethics committees of the institutions involved in the study.Informed consent was collected before participants enrolled in the study. Data sources and measurements We obtained sociodemographic characteristics (age, sex and marital status) from electronic medical records, along with values of the clinical variables: LVEF, number of hospitalisations in the last year, presence of cardiac implantable devices, NYHA functional class, comorbidities and current medications.Comorbidities included clinical depression, atrial fibrillation, type 2 diabetes mellitus, kidney disease, lung disease, coronary artery disease, obstructive sleep apnoea and hypothyroidism.Current medications included ACE inhibitors, beta-blockers and angiotensin receptor blockers. When a patient met the eligibility requirements for study inclusion, the treating cardiologist answered the SQ for that patient.For patients that the cardiologist would not be surprised if they died within the next year, we coded as +SQ.For patients that the cardiologist would be surprised if they died within the next year, we coded as a negative SQ (−SQ). Statistical methods To describe the study sample, we used mean and SD to summarise continuous variables in case of normal distribution.In cases of skewed distribution, we used median and IQR.We assessed normality using Q-Q plots.We summarised categorical variables as frequencies and percentages. Original research To assess the SQ's performance predicting 1-year mortality, we calculated the sensitivity and specificity, as well as the SQ's positive (+LR) and negative (-LR) likelihood ratios and the positive (PPV) and negative (NPV) predictive values (online supplemental tables 1 and 2).We interpreted the effect of the +LR on the likelihood of dying within 1 year based on the following classifications: no change if +LR=1; minimal increase if +LR between 1 and 2; small increase if +LR between 2 and 5; moderate increase if +LR between 5 and 10; and substantial increase if +LR >10. 20o illustrate how the result of the SQ changes the probability that a patient dies within 1 year, we created Fagan's nomograms. 21We also conducted a univariable regression to assess the relation between a +SQ and 1-year mortality. Participants Of the 184 patients who met the inclusion criteria and were potentially eligible, 178 consented to participate in the study.Among these 178 participants, 4 were excluded because their 1-year vital status was unknown (figure 1). Table 1 shows baseline demographic and clinical characteristics of the study's 174 participants.The sample had a median age of 70 (58-77), was predominantly male, had reduced LVEF and NYHA class II. The prevalence of a +SQ was 48%.After 1 year, 20 patients had died, giving an overall mortality rate of 12%. Performance of the SQ predicting 1-year mortality After 1 year, mortality among those with a +SQ was 21%; mortality among those with a -SQ was 3% (p<0.001).Participants with a +SQ had 7.5 times higher odds of death at 1 year compared with those with a -SQ (OR 7.6, 95% CI 2.1 to 26.9). The +LR is the probability that patients with +SQ will die within 1 year divided by the probability that patients with +SQ will be alive in 1 year. 20The +LR of the SQ was 1.98.The +SQ was nearly twice as likely for patients who died within 1 year than it was for patients who were alive in 1 year (figure 2).According to the classification of the +LR's effect on the likelihood of dying within 1 year, it is a minor increase.The 1-year mortality rate for our study was 12%.With the pretest probability of a patient dying within 1 year at 12% and a +LR of 1.98, the post-test probability of dying within 1 year is 20% (figure 3).A +SQ increased the probability that an ambulant patient with HF died within 1 year by 8 percentage points.The -LR was 0.26.The -SQ was nearly four times more likely assigned to participants who were alive in 1 year than it was for patients who died within 1 year. For our sample, the probability that the SQ correctly identified an individual who would die in the course of a year, the sensitivity was 85% (95% CI 69% to 100%).The probability that the SQ would correctly identify an individual who would survive over a year, the specificity was 57% (95% CI 49% to 65%) (figure 2). Subgroup analyses We compared the SQ's performance between women and men.Whereas other parameters were similar, the SQ's sensitivity was 5 percentage points higher for women.With higher sensitivity, higher specificity and higher +LR, the SQ performed better for participants younger than 70 years (figure 2).Regarding LVEF's classification, sensitivity, NPV and -LR were better for patients with reduced LVEF.Specificity, PPV and +LR were better for patients with mildly reduced LVEF.We observed the SQ's worst performance for patients with preserved LVEF (figure 2).Among patients classified as NYHA III/IV, the SQ's sensitivity was perfect (100%).This subgroup also had the best values of PPV and NPV.However, the SQ's specificity was very low (31%) for patients at NYHA III/IV classification (figure 2). Based on Fagan's nomograms, we accounted for the clinical application of the +LR.We observed the biggest probability changes for patients with +SQ dying within 1 year for those aged under 70 years and for patients with mildly reduced LVEF.Having a +SQ increased the probability of dying within 1 year by 12 and 14 percentage points, respectively (online supplemental figures 1-8).According to the +LR's effect on the likelihood of dying within 1 year, the +LRs showed a small increase in the likelihood among women; those aged under 70 years and patients with mildly reduced LVEF.In the remaining subgroups, +LR showed a minimal increase in the likelihood of dying within 1 year. Key results Our primary objective was to assess the SQ's performance predicting 1-year mortality among patients in ambulatory HF clinics.With a sensitivity of 85%, the SQ is a good tool to screen ambulatory HF clinic patients who might be in the last year of life.The SQ's performance predicting 1-year mortality was similar among women and men.We also assessed the SQ's performance according to sex and other demographic (age) and clinical characteristics.The SQ performed better for patients aged under 70 years; for those with reduced or mildly reduced ejection fraction; and for patients at NYHA III/IV classification.For the whole sample and the different subgroups, +LRs showed a minor or a small increase in the likelihood of a patient dying within 1 year, which is not good enough to consider that a patient has a life expectancy limited to 1 year. Performance of the SQ by subgroup The risk factors for developing HF differ between men and women, as do responses to treatment, symptom burden, and comorbidities due to both biological and cultural factors. 23Because of this, sex-specific results should be presented in research. 24Accounting for sex, determining survival rates for patients with HF has been inconclusive.Initially, the Framingham study showed better survival rates after HF diagnosis for women than for men. 25 Later, other studies suggested worse survival rates for women, 15 which was supported Original research by women presenting with HF when they are older and have more comorbidities. 23However, the most recent evidence suggests that age-adjusted mortality is similar between sexes. 26mong the subgroups, the SQ's best performance was for those aged under 70 years.The main difference between its performance for those older and younger than 70 years was specificity.An age over 70 years likely contributes to a +SQ response from cardiologists, leading to an increase in the proportion of false positives; thus, a reduction in specificity. The best sensitivity was for the group of patients with reduced LVEF.With more evidence of effective therapies to reduce morbidity and mortality, this type of HF is the most studied and best understood. 27In addition, evidence has shown that for patients with LVEF regardless of age, the lower the LVEF, the higher the mortality. 28This might explain why the performance of an intuitive prognostic tool is better for this type of HF.The worst performance was among patients with preserved LVEF.This type of HF is not well understood, and there is no evidence of pharmacological therapy that decreases mortality for patients with preserved LVEF. 27 29An intuitive prediction of mortality for this group of patients is especially complex because the relationship between LVEF and mortality is U shaped. 16 28Above certain LVEF values, age-adjusted mortality increases, which is comparable with patients with LVEF between 30% and 35%. 16 28s for NYHA classifications, all patients classified at NYHA III/IV who died had a +SQ, which led to a sensitivity of 100%.However, for patients at NYHA III/IV functional class who survived 1 year, the majority also had a +SQ.Perhaps due to previous knowledge that mortality increases with increasing NYHA functional class, cardiologists are more likely to assign +SQ to patients with HF in more advanced stages of NYHA, 14 which is similar to what happens with older patients with HF. Comparison with previous studies Previously, the SQ's performance predicting 1-year mortality had been evaluated for patients with HF in emergency 13 and inpatient settings. 12In the emergency department, the SQ has a sensitivity of 79% and a specificity of 57% when answered by emergency physicians. 13For hospitalised patients, the SQ has a sensitivity of 85% and a specificity of 59% when answered by cardiologists. 12We found that the SQ's sensitivity (85%) and specificity (57%) for outpatient settings are equal to inpatient settings.Since the SQ's sensitivity and specificity were the same for decompensated (inpatients) and stable patients (outpatients), it suggests that the SQ's performance predicting 1-year mortality for patients with HF does not vary significantly.However, to compare the SQ's performance across settings and its interpretation in clinical practice for individual patients, we would have to compare LRs.No studies assessing the SQ's performance for populations with HF reported LRs.However, since LRs are calculated using sensitivity and specificity, we estimated them using other studies' reported test sensitivities and specificities.The +LR and -LR were similar across settings and represent minimal increases in the likelihood of dying within 1 year. Interpretations for clinical practice When there is high test sensitivity, fewer false negatives occur, which increases the chance that patients with HF in need of ACP or PC will receive these services.Since determining sensitivity was the main criterion to evaluate the SQ's performance, we consider it is an acceptable prognostic screening tool for patients with HF.Yet, there remains a significant percentage of patients with HF within the last year of their life who might be left out (15% of the whole sample) and might benefit from ACP or PC.Both in this population and in the literature, the SQ predicts 1-year mortality well among patients with HF. 12 13 However, there are several aspects of the tool that users should be aware of before applying it in clinical practice. First, the SQ's performance depends on the clinical expertise and experience of the staff using it, as well as their knowledge of the patient with HF.For example, Straw et al conducted a study that evaluated the SQ's performance among an inpatient population with HF. 12 They showed that the SQ's sensitivity decreased from 85% when cardiologists used the tool to 75% when physicians in training used it; and from 90% when HF nurses used the SQ to 66% among non-specialist nurses.Second, screening strategies such as the SQ do not afford assessment of the complexity of patients' needs or the level of training that professionals should have to address ACP discussions or provide PC . Finally, a major limitation of this screening tool is that it potentially excludes patients who will survive longer than 1 year but still would benefit from ACP or PC.Even in scenarios when the SQ predicts mortality well, using life expectancy as the sole criterion for assessing the need for ACP or PC is limiting.We consider the SQ can be used as a screening tool to initiate ACP or refer to PC for patients with a life expectancy of less than 1 year.However, we also consider the parallel use of needs assessment tools for patients with life expectancy of more than 1 year.For example, two recent systematic reviews of available tools to assess PC needs in patients with HF concluded that the Needs Assessment Tool: Progressive Disease-Heart Failure (NAT: PD-HF) was the most appropriate tool to determine the unmet needs of patients with HF. 30 31 The NAT: PD-HF offers an alternative solution to several previously discussed points: it is not based on survival prognosis or severity factors, but it comprehensively evaluates different spheres. The NAT: PD-HF is made of questions to determine a patient's physical and psychological symptoms, daily life activity limitations, spiritual concerns, financial or legal concerns, and health-related information needs.Although the NAT: PD-HF does not focus on PC needs or referrals for specialised PC services, it does assess patients' unmet needs and matches those needs with appropriate services, including specialised PC and other services.Finally, the NAT: PD-HF assesses the patients' and the caregivers' needs, including the caregiver's ability to take care of the patient. 32However, in clinical situations where there is not enough time to gather answers to the NAT: PD-HF's comprehensive question sets, there is enough time for the clinician to ask themselves the singular SQ, which is better than no needs screening at all. Strengths and limitations of this study The low proportion of patients classified as NYHA III/ IV might be a limit of the generalisability of our results.The risk of mortality increases with a higher NYHA classification. 33Among our population, 80% of the patients were classified as NYHA I or II.As expected, the mortality rate was low compared with what has been reported in other HF clinics where mortality is around 30%. 12 13 However, as we conducted different subgroup analyses, including analysis according to NYHA functional class, we provide different analyses Original research that can be assessed according to each HF clinic population. As most encounters between healthcare personnel and patients occur in ambulatory settings, a strength of this study is the contribution of the SQ's psychometric profile for ambulatory patients with HF.In addition, when screening for limited life expectancy and the need to initiate ACP, since the outpatient population are more stable patients, ACP needs may be overlooked.The systematic use of a tool such as the SQ could help identify patients with HF eligible for end-of-life care without being limited by having ACP discussion or making decisions within unstable medical contexts.Furthermore, to support better generalisation, this study provides substantial data regarding subcategories. CONCLUSION The SQ showed a good sensitivity predicting 1-year mortality for patients in ambulatory HF clinics.However, the likelihood of dying within 1 year increases little when having a +SQ.We suggest that the SQ can be used as a starting point to identify patients who might benefit from having an ACP discussion or a referral to PC due to limited life expectancy toward a patient's end of life. Figure 1 Figure 1 Flow chart of the patients included in the study. Figure 2 Figure 2 Performance of the SQ in predicting 1-year mortality among patients in ambulatory HF clinics.HFmrEF, heart failure with mildly reduced ejection fraction; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; LR, likelihood ratio; NPV, negative predictive value; NYHA, New York Heart Association; PPV, positive predictive value; SQ, Surprise Question. Figure 3 Figure3 Fagan's nomogram for the overall sample.Based on a pretest probability of dying within 1 year of 12%, the blue line shows a post-test probability of a patient with a positive Surprise Question (+SQ) dying within a year of 20% (95% CI 17% to 25%) according to the positive likelihood ratio (+LR) of 1.98.A +SQ increases the probability of a patient dying within 1 year by 8 percentage points.The red line shows a post-test probability of a patient with a negative SQ (-SQ) dying within a year of 3% (95% CI 1% to 9%) according to the negative LR (-LR) of 0.22.A -SQ decreases the probability of a patient dying within 1 year by 9 percentage points. Table 1 Clinical and demographic characteristics of study participants	2021-08-19T06:16:38.504Z	2021-08-17T00:00:00.000	{ "year": 2021, "sha1": "8fe959955ff8a13db00200fceeccaa2de02ff0d5", "oa_license": "CCBYNC", "oa_url": "https://spcare.bmj.com/content/bmjspcare/early/2021/08/16/bmjspcare-2021-003143.full.pdf", "oa_status": "HYBRID", "pdf_src": "PubMedCentral", "pdf_hash": "170e232c7d1c3421f87c28ee77b0b8303298811c", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
250444147	pes2o/s2orc	v3-fos-license	Let’s all dance: Enhancing amateur dance motions Professional dance is characterized by high impulsiveness, elegance, and aesthetic beauty. In order to reach the desired professionalism, it requires years of long and exhausting practice, good physical condition, musicality, but also, a good understanding of choreography. Capturing dance motions and transferring them to digital avatars is commonly used in the film and entertainment industries. However, so far, access to high-quality dance data is very limited, mainly due to the many practical difficulties in capturing the movements of dancers, making it prohibitive for large-scale data acquisition. In this paper, we present a model that enhances the professionalism of amateur dance movements, allowing movement quality to be improved in both spatial and temporal domains. Our model consists of a dance-to-music alignment stage responsible for learning the optimal temporal alignment path between dance and music, and a dance-enhancement stage that injects features of professionalism in both spatial and temporal domains. To learn a homogeneous distribution and credible mapping between the heterogeneous professional and amateur datasets, we generate amateur data from professional dances taken from the AIST++ dataset. We demonstrate the effectiveness of our method by comparing it with two baseline motion transfer methods via thorough qualitative visual controls, quantitative metrics, and a perceptual study. We also provide temporal and spatial module analysis to examine the mechanisms and necessity of key components of our framework. Introduction Dance is a performing art form that consists of purposeful, rhythmical, and well-patterned sequences of body movement; it has aesthetic and often symbolic value [1]. Capturing dance motions and transferring them to avatars not only facilitates expressive film or animation production process, but also contributes to the conservation of cultural heritage and dance education. However, so far, access to high-quality dance data has been limited. Most currently available motion capture repositories typically contain basic human movements, while only a limited number of dance-specific databases comprise prime dance movements performed by professionals [2,3]. This is because professional dance is characterized by dynamic body language, high impulsiveness, elegance, smoothness, fluidity, and aesthetic beauty that usually require the performer to have long-term dance experience and skills, followed by extensive practice sessions, excellent physical condition, and acquaintance with years of dance studies. This poses a practical challenge when capturing realistic and high-quality dance motions, which is restrictive for large-scale acquisitions, or regular acquisition [4]. To perform a professional dance, the performer should be familiar with the content and rhythm of the choreography, and achieve the specific physical amplitude of the choreography with the appropriate energy and balance [5]. On top of that, in order to achieve a satisfactory dance quality during the motion capture process, dancers have to repeat the performance many times to avoid mistakes. In this paper, we present a technique that enhances professionalism of dance moves, allowing the movement quality to be improved in both the spatial and temporal domains, meeting the following key constraints: (i) production of flowing and smooth dance moves, (ii) expansion of the anatomical and physical amplitude of human movements, to meet the demanding restrictions of the choreography, and (iii) good synchronization of movements to the rhythm of the music. In this way, our method reduces the need to hire professional dancers, facilitates the process of obtaining high-quality dance movements even from amateurs, enriches existing databases with professional data to enable better training of deep networks, and finally aligns dance motion data to a given audio file. One obvious approach to deal with the challenge of enhancing professionalism of dance movements is to leverage a deep style-transfer framework [6][7][8][9][10], by considering amateur dances as the source style and professional dances as the reference style. However, while style transfer algorithms provide a possible way to handle this problem, they do not address exactly the same problem. Professionalism is not a specific style, but closer to an evaluation metric. Dances with different styles might be seen as professional. Professional dance preparation, whatever the style, not only has specific anatomical and physical demands, but also requires artistic qualities, such as musicality, expression, and distinct communication skills. On top of that, existing style-transfer methods face two technical challenges. Firstly, they mainly focus on motions with welldefined styles, while different styles of motions have explicit changes over the whole sequence. In contrast, dances often contain highly-dynamic and heterogeneous movements, and the professional and non-professional dances may share a large number of similar poses but with a limited number of local changes: see Fig. 1. Therefore, it is difficult to learn mappings between unpaired professional and nonprofessional dances, as state-of-the-art motion style transfer methods do [8,9]. Secondly, they mainly focus on music-free motions, with no explicit and deterministic control over the correlations between motion content and other external factors, such as musical rhythm. Even though existing methods may cause timing changes in motion based on the style differences hidden in the data, such changes are uniformly distributed over the temporal domain. In this work, we propose a two-stage dance enhancement model that adds professionalism to existing dance motions, and release a new dataset with paired professional and amateur dances that enables the training on the model. We define Fig. 1 Our approach enhances the professionalism of dances performed by non-professional dancers. Top: input amateur dance sequence. Second row: our enhanced dance motion. Third row: corresponding ground-truth professional dance. Note that our results have similar temporal and spatial features to the ground-truth dance sequence. dance professionalism term, and describe how it can be evaluated through various attributes, e.g., flow, amplitude, and rhythm of a dance. In particular, we improve the quality of dance motions in both spatial and temporal domains, focusing on the following three professional properties: (i) the production of fluent and smooth movements; (ii) the physical amplitude of intense movements that is restricted by the poor physical condition of the amateur dancer; and (iii) the temporal alignment of the dance movements to a given musical rhythm. Firstly, our model estimates temporal correlations between dance motions and musical rhythm, followed by a temporal alignment and spatial motion enhancement process, under the guidance of the proposed professionalism metrics. The dance-to-music alignment stage consists of a network that learns the affinity matrix between dance and music with attention mechanisms, and a classic dynamic time-warping module to infer an optimal temporal alignment path matrix. Secondly, the dance-enhancement-stage enables adjustment of the dance motion in both temporal and spatial domains, under the guidance of the optimal alignment path, a reconstruction loss, and a consistency loss. The reconstruction loss constraints the network to preserve the original motion content of the amateur dance, while adjusting it to be similar to the corresponding professional dance. The consistency loss preserves the temporal continuity of the enhanced dance motion and decreases temporal noise. One of the most critical challenges we faced in this project was the lack of data for training our network. Professional and corresponding amateur dances may differ in various combinations of the above professionalism properties. Since the professionalism of a dance is independent of its choreography or style, the dances in a professional or amateur dance dataset may contain highly dynamic and heterogeneous movements. This makes it difficult to learn a homogeneous distribution for the professional or amateur dataset using existing methods, let alone a credible mapping between the two heterogeneous datasets. In addition, the mappings between professional dances and amateur ones are not deterministic. Therefore, before designing our network, we first introduce a key-pose based data augmentation scheme to generate amateur data from professional dances, taken from the AIST++ dataset [3]. The data augmentation scheme modifies the movements in all three professionalism metrics, and the constructed dataset contains many-to-one paired amateur and professional dances. We demonstrate the effectiveness of our method by comparing it to two state-of-the-art motion transfer methods [6,8] via thorough qualitative visual controls, quantitative metrics, and a perceptual study. Apart from using our synthesized amateur data, we additionally captured several dance sequences performed by amateur dancers, to further examine the generalizability of our method. User responses indicate that our method enhances amateur motion so that it cannot be easily distinguished from actual professional dance. In addition, we provide temporal and spatial module analysis via an ablation study to evaluate the mechanisms and necessity of key components of our framework. The main scientific contributions of the paper are: • The concept of enhancing professionalism in dance movements: we give a first definition of what dance professionalism is, and how a professional dance can be distinguished from an amateur one. • A novel two-stage deep learning framework that extracts meaningful features from motion inputs, in terms of the newly defined professionalism criteria, to improve the quality of dance motions. It integrates a reconstruction loss, to preserve the original content of the dance, and a consistency loss, to maintain the temporal coherency of the reconstructed motion. • A novel model designed to synchronize 3D dance motions with reference audio in the face of nonuniform and irregular misalignment. • Thorough evaluation, and an ablation study to examine the efficiency and necessity of our methods. Dance evaluation Dance is an expressive form of performing art that consists of aesthetic movements of the body in a rhythmic way, usually to music, for the purpose of expressing an idea or emotion, releasing energy, or simply taking delight in the movement itself [5]. To professionally perform dance, performers regularly attend long routine training, and have extensive experience in dance studies, choreography, and musicality, along with excellent physical condition, which enables them to perform complex movements with extreme physical amplitude demands in some instances [11,12]. Only a few works in the dance research community have identified qualitative indicators of professional dances. For example, Neave et al. [13] and Torrents et al. [14] have reported qualitative experiments showing that kinematic parameters related to the amplitude of movement are highly associated with perceived dance beauty and aesthetics, while Park [15] investigated the correlation between dance professionalism and motion smoothness (using jerk-based quantitative measures). However, no explicit quantitative metrics have been proposed so far to completely evaluate dance professionalism. In computer graphics, several interactive dance systems have been proposed to enhance dance learning and teaching [16][17][18]. Basically, these methods export dance movement features to enable comparisons between dances performed by professional dancers (teachers) and amateurs (students). For example, Chan et al. [19] implemented a self-learning dance system by visually comparing motion accuracy through Euclidean distance between professional and amateur motions. Aristidou et al. [20] leveraged the well-known Laban movement analysis (LMA) theories [21] to introduce quantitative feature components that measure quality characteristics relating two dance motions. Although these movement measurements can be used to understand motion qualities and to compare similarities between dance motions, no metrics have been developed so far that explicitly measure the professionalism of dance motions. Motion style transfer One obvious way to add professionalism to existing motion sequences is to use methods based on the concept of motion style transfer. These methods aim to transform the style of a reference motion to a source motion, while simultaneously preserving the original source motion content. Several approaches [22,23] have been proposed in the literature to infer styles of motions using hand-crafted features. For example, Tenenbaum and Freeman [22] explicitly separated style and content using asymmetric bilinear models. Aristidou et al. [23] built statistical correlations between LMA features and emotions, and used such correlations to support interactive emotion-based motion transfer. However, those methods explicitly construct common mappings between hand-crafted features and motions, which are hard to generalize to heterogeneous or large-scale datasets. Machine learning based techniques. To avoid the disadvantages of selecting hand-crafted features, researchers started to extract style information from large-scale paired data using machine learning techniques [24][25][26]. Brand and Hertzmann [24] introduced a style hidden Markov model (HMM) and minimized information entropies to separate structure, style, and accidental properties. Following their work, Hsu et al. [25] built dense correspondences between different motions with an iterative motion warping algorithm, and then proposed a linear timeinvariant model to translate motion styles, while Xia et al. [26] proposed to learn local regression models. However, machine learning-based methods require explicit or implicit motion registration between the input and output motions, and are therefore limited to styles and content that exist in the training dataset; as a result, they do not generalize well to new styles of motion. Deep learning based techniques. In recent years, deep learning techniques have been widely adopted to transfer motion styles [6][7][8][9][10][27][28][29], enabling more efficient and effective results for complex and even unpaired motions. For instance, Holden et al. [6] leveraged convolutional autoencoders [30] to learn hidden motion representations with paired input and output motions. The same authors in Ref. [7] further improved this model with an additional feed-forward neural network, and transformed motion style in the hidden motion space under the constraint of a Gram matrix [31]. Later, Aberman et al. [8] proposed a neural network to disentangle latent style and content codes, where the latent style code is used to modify the decoded motion content through an AdaIN [32] operator. By using a multi-style discriminator, this method can handle unpaired motions. Following their work, Wen et al. [10] recently proposed an unpaired and unsupervised motion style transfer method using a generative flow model. Despite their great progress, existing deep-learning-based methods mainly focus on locomotions with a limited number of motion structures, and have no explicit control over musical correlations. Unlike locomotions, dance performed by professional dancers may contain heterogeneous motions with various choreographies (e.g., different motion poses and ordering of poses) and are wellsynchronized to temporal rhythmic patterns. Our method deals with these challenges by simultaneously learning the intrinsic motion attributes and the motion-rhythm correlations that commonly appear in professional dance. Music-driven motion synthesis Many scholars have worked on methods for musicdriven dance synthesis. Typical solutions leverage a graph-based framework [33][34][35][36]. In pioneering work, Kim et al. [34] constructed a movement transition graph based on extracted motion beats and synthesized new motions under kinematic and rhythmic constraints. More recently, the use of machine learning to synthesize music-driven dance motions has witnessed impressive progress [4,[37][38][39][40][41]. For instance, Lee et al. [38] proposed a decomposition-to-composition framework to generate 2D movements conditioned on a given piece of music, under the guidance of learned correlations between musical beat and dance units. Chen et al. [4] proposed a choreomusical embedding module to learn stylistic and rhythmic music-dance correspondences, and incorporated the embedding distances into the traditional graph-based dance synthesis framework. More recently, Aristidou et al. [41] introduced a music-driven neural framework that generates rich and diverse dance motions that respect the overall choreographic structure of a dance genre. However, music-driven dance synthesis learning methods heavily rely on high-quality dance motion data synchronized to given audio for adequate training. Since access to dance data made by professionals is not always possible, our method can be used to enrich databases using data from amateurs that have been artificially enhanced to look more visually appealing; our work simultaneously learns music-to-dance correspondences and leverages them to learn dance-to-dance correlations. Audio alignment To enhance non-professional dances, an essential goal is to align dance motions to reference audio. Motion-audio synchronization aims to temporally align human motion dynamics to audio rhythms, which is fundamental to synthesis of rhythmic human motions. Traditional motion-audio synchronization methods leverage hand-crafted 2D motion and rhythmic features, determine their correspondences, and warp motions under the guidance of motionrhythm matches [35,[42][43][44]. Over the years, more attention has been devoted to the video-audio timing alignment problem. A common basic idea is to find optimal video-to-audio correspondences and use them to guide warping between visual and audio features, either using hand-crafted features [45,46] or deep multi-modal features [47,48]. Among such methods, Wang et al. [49] introduced two attention modules before the feature extraction stage to highlight important spatial and temporal regions. In contrast, instead of emphasizing specific features, we introduce an integrated attention module to map correspondences between spatial and temporal motion features, and audio rhythms, without hand-crafted elements or post-processing. Our model is the first to synchronize 3D dance motions with reference audio under irregular and non-uniform misalignment. Data augmentation An important challenge that needs to be addressed in this work is the lack of training data. The limited existing dance motion datasets [2,3,39] typically contain high-quality professional dances, but lack corresponding non-professional ones. As a result, it is difficult to build correlations between paired professional and non-professional movements. Capturing realistic non-professional dances requires amateur dancers to learn the original professional choreography, which can be challenging and requires practice, and it may be difficult to cover the large variability of the movements of professional dancers. Instead, we propose a data augmentation scheme that artificially synthesizes random non-professional dances, by altering professional ones taken from the AIST++ dataset [3], both in the spatial and temporal domains. Definition of dance professionalism Before we present our data augmentation schema, it is important to define first the criteria that distinguish a professional dance movement from an amateur one. In that matter, we consulted expert choreographers, experienced dancers, and dance teachers, who pointed out the following key criteria: • Sense of rhythm. Professional dancers can perfectly follow the beat of the music, while amateur dancers often lose synchronization and have difficulty in following the rhythm of the music. • Physical amplitude. Professional dancers have excellent physical condition, which allows them to perform complex and dynamic movements, in some cases reaching the limits of the body. In contrast, non-professional dancers usually have difficulties in completing certain dance moves as they have limitations due to their poor physical condition (to extend their body to the limits, to perform the splits, etc.). • Motion quality. The movements of a professional dancer are elegant, smooth, and the movement cycle is nicely completed. In contrast, the movements of an amateur dancer are often not in balance, they abruptly start and end movements without fully completing the movement cycle (sharp movements), and may be shaky, lacking smoothness. All these result in amateurs requiring more effort than professionals because they do not control their movements as experts do. • Concentration and consistency. Amateur dancers usually focus on one part of the body (e.g., legs or arms) and may neglect the consistency of movements of other parts of the body (such as the head, and overall style). Note that our method does not take this feature into account. • Choreography. Professionals have a richer choreography in terms of the diversity of movements, compared to amateurs who usually repeat the same movements multiple times. Again note that changing dance choreography is outside the scope of this paper. Generating amateur dance movements Amateur dancers have, in general, difficulties in synchronizing their movements to the musical beat, to achieve certain physical amplitudes, and to perform controlled and smooth movements. Therefore, to enrich our database, we introduce a method that artificially alters professional dance movements, through random disturbances, to generate corresponding amateur counterparts. It is important to note that our approach needs to meet the following three conditions: (a) to keep the choreography of the professional dances unchanged, (b) temporal disturbance: to alter the temporal alignment between motion, and music and rhythm; (c) spatial disturbance: to change the physical amplitudes of motions. In this manner, we propose a key-pose-based scheme that first extracts key poses based on the motion beat; then, it randomly generates spatial disturbance factors to limit or exaggerate the physical amplitudes of movements, and temporal disturbance factors to disrupt the music-motion synchronization. Finally, it computes the spatial and temporal disturbances in between those key-poses using piecewise linear interpolation [50]. These are later used to modify the professional dances. Figure 2 overviews our data augmentation method. Motion representation. We represent a dance motion M as a sequence of T skeleton poses. Each skeleton pose P is represented by J = 21 joint rotations that are organized in a hierarchical order and depicted by unit local positions [51] between parent-child joints, denoted as P ∈ R (J−1)×3 . Therefore, a dance motion can be represented by M ∈ R T ×(J−1)×3 , where T = 426 to 2878 frames without being trimmed into short clips. Poses are then translated back to rotations via a Jacobianbased inverse kinematics solver [52]. Note that the root rotations and translations are discarded in the motion representation to avoid significant changes to the choreography. Key-pose extraction. When learning to dance, it is usually easier for students to identify prominent changes of movements (such as pausing and turning). Based on this observation, we define the representative poses to be those with changes of velocity direction [46]. To facilitate key-pose extraction, we first uniformly sample several poses over a certain time duration t. The time duration t is set to three seconds in our implementation. We then search for the nearest motion beats as the corresponding keyposes, where the motion beat is found using the minimum of all joint speeds in a certain frame. Since some neighboring key-poses may bring about rapid direction changes, we filter out key-poses that have neighboring motion beats within 1 s. The key-poses are spatially modified based on the spatial disturbance curves. We highlight one curve for a specific joint with the spatial factors on key-poses (indicated by yellow dots). (c) The temporally modified key-poses and accompanying temporal disturbance curves. Spatial disturbance. The spatial disturbance aims to disrupt physical amplitudes of the movements, limiting or exaggerating their intensity. Thus, we define the spatial factor S ∈ R N ×J for the N selected key-poses to control the spatial disturbance of all skeleton joints, and randomly generate corresponding values through an approximately inverse normal distribution: where s n is the randomly generated spatial disturbance value for key-pose n. α and β are used to control the shape of the inverse normal distribution; in our implementation they equal 1.1 and 1.3 respectively. d is a randomly generated binary parameter that enables (d = 1) or disables (d = 0) exaggeration of the pose. All joints in a specific frame share the same d value. S n is then propagated to each frame of the entire sequence as S ∈ R T ×J (T > N ) via linear interpolation. A straightforward way to apply the aforementioned spatial factor for motion disturbance is to directly multiply the rotations or positions of each joint by it. However, this may produce infeasible poses that violate physical or bone constraints. Instead, we interpolate the new pose (local position) between the current and a standard standing pose, guided by the spatial factor, as Eq. (2): where p t,j denotes the local position of the j-th joint in the t-th frame, and u j is the pre-defined direction for joint j of the standard standing pose. To simplify the process, we define three key interpolation directions for the standing pose: up (u j = (0, 0, 1)) for joints on the spine, no modification for the shoulder and waist joints, and down (u j = (0, 0, −1)) for the other joints. Temporal disturbance. The temporal disturbance aims to disrupt the mappings between the dance motion and the corresponding musical rhythm. We define the temporal factor Q ∈ R N to control the temporal disturbance at the N key-poses, and randomly generate values to warp the original dance motion sequence according to Eq. (1). The parameters α and β are set to 50 and 0 respectively. We then move the key-poses to new positions by shifting them Q frames, where negative Q n means shifting backward and positive Q n indicates shifting forward. Note that in this process we need to check time crossings between key-poses and preserve motion monotonicity. Finally we calculate the movements of the intermediate poses between adjacent key-poses by linear interpolation. Augmented data. Using the above approach, we constructed a large, highly variable, non-professional dataset of dances paired to professional ones. We repeated the temporal and spatial disturbance four times, creating many-to-one paired amateur and professional dances, in a dataset 4 times the size of the original AIST++ database. Dance professionalism framework Our framework, by taking as input a dance motion sequence performed by an amateur dancer, and its corresponding audio file, aims to enhance professionalism by considering the following three conditions: retention of the original choreographic content, generation of fluid movements, and amplification of physical amplitudes. Enhancement is made in both temporal and spatial domains. In the temporal domain, our framework aligns the amateur dance motion to music to achieve fluid and consistent motions, while in the spatial domain, it increases the physical amplitudes of the input motions to match those of a professional dancer, while preserving the original content of the dance's choreography. To do so, we have designed a two-stage deep framework: a dance-to-music alignment stage, and a dance enhancement stage. The dance-to-music alignment stage estimates the temporal mapping of the amateur dance required to match the input music; these estimates are later integrated into the dance enhancement stage to enable temporal warping of the encoded dance content features, which are later decoded to reconstruct a professional dance with the same choreographic content. Figure 3 shows the twostage architecture of our framework, whose details will be described in the following sections. Dance-to-music alignment stage The main goal of the dance-to-music alignment stage is to find the optimal alignment between the input music and dance sequences. Taking into account the highly complex correspondences between a dance motion and music, we use auto-encoders to learn the cross-modal frame-to-frame mapping between the high-level motion and the music features extracted from the raw data. High-level feature extractor. For an input music signal with T frames, we compute the melscaled spectrogram using the well-known librosa [53] audio analysis library, depicted as G ∈ R T ×B where B is the number of frequency bins. For the input T -frame dance motion given by position offset vectors M ∈ R T ×(J−1)×3 , we first calculate the corresponding joint positions, and then estimate the velocities and accelerations of each joint in x, y, z directions per frame, denoted K ∈ R T ×C ; C = J × (3 + 3) where J = 21 joints in the human skeleton. Dance-to-music alignment network. The dance-to-music alignment network is composed of two encoders, MusicEnc and MotionEnc, to map the music feature G and dance motion feature K to the corresponding latent feature sequences f G and f K , respectively. The two encoders share the same network architecture but have different weights. Following the two encoders, we compute the Fig. 3 Our two-stage dance professionalism architecture. The dance-to-music alignment stage learns the temporal alignment of the input dance motion to the corresponding music, through a dynamic-time-warping operation on the encoded deep features of dance motion and music. In the dance enhancement stage, we first extract the hidden dance motion features to express the original motion content, which are then modified under the guidance of the temporal alignment matrix, and further decoded into the enhanced dance motion under the constraints of a reconstruction and consistency loss. Euclidean distance between the frames of the two latent feature sequences to form a T × T affinity matrix, defined as where i is the index of the music frame, and j is that of the motion frame. Figure 4 shows the structure of our dance-to-music alignment network. Specifically, for each encoder, the input sequence is first processed by three temporal 1D-convolution layers sequentially, and each is followed by batch normalization and a ReLU layer. Considering the complex correlations between dance choreographies and their musical correspondences, we use the attention mechanisms in a transformer network to learn contextualized dance-to-music information, providing adaptive local neighbors for both the dance and music encoders. In particular, we add a shallow transformer with two multi-head self-attention and feed-forward layers on the basis of the 1D-convolution layers, and thus obtain latent feature sequences f G or f K encoding temporal context information. Note that the self-attention layers in the transformer are biased towards the local neighbors of each frame by setting the attention mask matrix B a as Eq. (4): where δ is a parameter to control the neighborhood size and is set to 50 in our implementation. Dynamic time warping. The target now is to find the optimal alignment between the input dance and music, so that each dance frame can be matched to the music frame with minimal alignment distance. Under the guidance of the affinity matrix deduced from the dance-to-music alignment network, we perform dynamic programming [54,55] an optimal alignment path matrix W between the latent dance and music features. Dance enhancement network The enhancement stage aims to modify amateur dances so as to look more professional in terms of physical amplitudes and dance-to-music synchronization. To achieve this goal, we leverage an auto-encoder network to modify the non-professional dances in latent feature space. Specifically, the non-professional dance sequence given by unit local positions is used to extract corresponding latent features via an encoder, DanceEnc. The latent dance features are then temporally warped under the guidance of the optimal dance-to-music alignment path, followed by a decoder to output the corresponding professional dance sequence. DanceEnc has similar implementation details to the MusicEnc and MotionEnc networks. We warp the encoded feature sequence f D by calculating the dotproduct between f D and the alignment path matrix W obtained from the dynamic time warping module. The decoder is implemented as a three-layer MLP network to project the feature sequence to the final enhanced dance. Training and loss The two stages in our framework are trained separately with different loss functions. In particular, the dance-to-music alignment network is trained using an alignment loss, while the dance enhancement network is trained with a reconstruction and a consistency loss. In this process, we leverage the optimal alignment path as a condition to modify the latent dance features; we use the ground-truth alignment path as an initial warping condition, and then fine-tune the network with the estimated alignment path. Note that the lengths of the input dance motions during training and testing can be arbitrary. Alignment loss. We assume that the dance sequences and their corresponding music sequence are well-synchronized: a motion frame is well matched with its paired music frame. Therefore, we design the temporal alignment loss on the affinity matrix in a contrastive learning manner. To be more specific, for each music frame, we select the corresponding dance frame as the positive sample and a randomly selected frame as the negative one. Then, we compute the triplet loss on the latent features of the three frames as the alignment loss: where f G , f K are the music feature and dance feature respectively, r is a randomly sampled frame index, andφ(t) is the index of the corresponding dance frame for the music frame t. Reconstruction loss. To improve the physical amplitudes of the movements in amateur dances so that they look more professional, we trained our network using paired amateur and professional data; our target is to force the enhanced amateur movements to be as close as possible to the corresponding professional ones. Therefore, we define a reconstruction loss that minimizes the local position error between the enhanced motion and the ground truth, given by Eq. (6): where p t,j is the local position of joint j at frame t in the enhanced dance motion, andp t,j is the corresponding local position in the ground-truth professional dance motion. Consistency loss. To enforce temporal coherence in the enhanced dance, we introduce a consistency loss by measuring the error between the velocity of the enhanced dance and that of the corresponding ground-truth. Our consistency loss is where v andv are the velocities of the enhanced and ground-truth dance motion, respectively. Results and discussion In this section, we present the dataset used for training and testing our method, implementation details, and evaluation metrics. We also demonstrate the efficiency of our framework in several experiments, a perceptual survey that evaluates its performance in terms of professionalism, realism, and dance-to-music synchronization, and an ablation study. Figure 1 shows a gallery of selected frames extracted from the input amateur motion (yellow), our result (red), and the ground-truth professional dance (blue). It can be observed that our method enhances professionalism so that the input's temporal and spatial features better match those of the ground-truth dance sequence. The quality of our enhanced dance animations may be examined in the supplementary video in the Electronic Supplementary Material (ESM). Dataset. The original AIST++ dataset [3] contains 1408 sequences of 3D human dance motion represented as joint rotations along with root trajectories. Each sequence of dance motion is accompanied by corresponding music well-synchronized to the animation. Overall, the dataset includes 10 dance genres with hundreds of different choreographies, providing rich and varied dance content. We follow the music-choreography data splits used in the original paper [3] for network training and testing/validation. For each professional music-dance pair in the AIST++ dataset, we produced multiple amateur dance counterparts using our key-pose based dance synthesis algorithm (see Section 3), by controlling their temporal and spatial disturbance factors. In total, we generated 3680 nonprofessional dances for training, 80 for testing, and 80 for validation. Implementation details. We implemented our framework in PyTorch and tested it on a 6-core PC with a 3.7 GHz Intel i7 CPU, 16 GB RAM, and an NVIDIA Tesla P100 GPU. All networks in our framework were trained with a batch size of 64 and learning rate of 10 −4 , and optimized by the Adam optimizer [56]. In total, it took about 12 training hours for the dance-to-music alignment network and 6 training hours for the dance enhancement network, on 4 NVIDIA Tesla P100 GPUs. Evaluation metrics. To the best of our knowledge, no quantitative metrics currently exist to evaluate the professionalism of dances. Therefore, we used the temporal alignment error (time error), the pose error (PE), and the Fréchet inception distance (FID), as evaluation metrics, and observed the temporal and spatial differences between the input motions, our enhanced dance motions, and the corresponding ground-truth professional dances. The three evaluation metrics are defined as follows: • Temporal alignment error is the average distance between the indices of motion poses per music frame, in the optimal dance-to-music alignment path and the ground-truth alignment path. • Pose error measures the average Euclidean distance between joint positions for specific poses in two motions sequences to be compared. • Fréchet inception distance measures how far the distribution of the enhanced dance is to that of the ground-truth professional one [4,57]. We calculate FID based on the extracted kinematic features [3] of the enhanced and ground-truth professional dances. Evaluation In this section, we evaluate the performance of our method with two baseline methods -the Holden et al.'s [7] and the Aberman et al.'s [8] methodsusing the three aforementioned evaluation metrics. In addition, we conducted three perceptual studies to qualitatively evaluate: (a) the quality and realism of our artificially generated amateur dance motions; the quality and realism of our experimental results in enhancing professionalism on amateur movements using (b) our synthetically generated amateur dataset, and (c) real, motion-captured amateur dances. More details about our perceptual study can be found in our supplementary material in the ESM. Comparisons Baseline methods. As far as we know, there are no other methods in the literature that deal with the dance enhancement problem. Thus, we compare the results of our approach with two state-of-the-art motion style transfer methods due to Holden et al. [7] and Aberman et al. [8], which also use auto-encoders as a backbone network. Unlike our problem, these methods take a content motion and a target style motion as input, and then generate an output motion by preserving the same but desired style of input content with the target motion. Note that, these methods do not take music into consideration. To adapt the two baseline methods to our problem, we made the following modifications. (1) Since they require motions to have the same length, we downsampled our dataset to the same length (400 frames). (2) We used our synthesized amateur dance together with the accompanying music as the input, and randomly selected another professional dance of the same genre as the target motion style for their network. (3) As Aberman et al.'s [8] network is trained using unpaired motion data with a consistency loss, minimizing a reconstruction error between the input and output content when the input content sequence and the style sequence have the same style. To make their method applicable to paired motion data, we use the consistency loss to calculate the reconstruction error between the output of the network and the ground-truth professional dance. Holden et al.'s [7] method was trained with its original loss functions. Fig. 5 Qualitative comparison of our method to baseline methods [7,8]. Each row shows a set of frames selected from the same music beat. It can be observed that our results are closer to the ground-truth than the two alternatives, in terms of dance-to-music alignment and pose reconstruction. For an animated version, please refer to our supplementary video in the ESM. Qualitative comparison. Figure 5 illustrates selected poses from the input dance motion sequence (yellow), our method (red), Aberman et al.'s method (green), Holden et al.'s method (gray), and the ground truth (blue). The music beat [53] is marked with a gray dotted line to indicate the temporal coherence. It can be observed that our method successfully produces good correspondences to the professional dance sequences, with satisfactory temporal alignment and spatial amplitudes. In contrast to our method, the two alternatives are not synchronized to the beat (since they are not designed for aligning dance-to-music), and their reconstructed poses are further than ours from the ground-truth movement. Quantitative comparison. Table 1 quantitatively reports the pose error and the FID metric. These metrics confirm our observations; the two baseline methods produce worse results than our method, having larger pose error and FID score. Since they are not designed to compute an explicit temporal alignment between dance and music, we do not consider the time error metric in this evaluation. In addition, we use a professional dance sequence as input to the network to further evaluate the naturalness and realism of the output motion. As Fig. 6 shows, the results confirm the ability of our method to generate natural movements, returning a . 6 In this example, we tested our network by inputting a dance sequence performed by a professional dancer. It can be observed that the output motion remains natural and realistic, and similar to the input. movement that is realistic and well aligned to the music beat. Perceptual study Evaluation of our synthetic amateur dance dataset. We first conducted a perceptual study to evaluate the quality and realism of our synthetic amateur motions, and whether they resemble true amateur dances. For this task, we recruited, in total, 20 participants, 11 female and 9 male. Each participant watched 28 pairs of side-by-side dance motions; on the right side, we showed amateur motions, which were either selected from our synthetic dataset (16 samples), or captured by amateur dancers who imitated professional dance moves (12 samples); on the left side of the video, we showed the corresponding ground-truth dance expert motions, so that the participants could use the professional motion as a reference to examine the quality of the amateur and synthetic motions. The participants were asked to rate on a Likert scale whether the presented motion on the right side was captured from an amateur dancer, or generated by a computer algorithm. The scale was 0: the motion was not performed by an amateur, there is too much computer-generated noise; 2: it is hard to decide; 4: the participant is strongly confident that the motion was performed by an amateur dancer. The scores were statistically analyzed to compare our synthetic motions and the true amateur motions. Figure 7 shows box-plots of the average score for the synthetic and true amateur motions. Both cases have an average score between two and three, which indicates that it is hard for participants to discriminate whether the motions are computergenerated or not. However, it is important to mention here that our synthetic dataset may have some differences compared to the true, motion-captured data. Our synthetic amateur motions are generated by randomly setting disturbances in spatial and temporal spaces to imitate the amplitude and music synchronization of amateur dances, so some motions may exist with too exaggerated or limited movements. In addition, dances performed by amateurs may have lower consistency of the body parts and contain different choreographies compared to those performed by experts; these differences have not been considered in our synthetic data generation process. Therefore, as expected, our synthetic amateur motions got a Fig. 7 Average scores evaluating whether dance motions were captured from amateur dancers or algorithmically synthesized. Red: score for all synthetic amateur motions. Green: score for real, motioncaptured, amateur motions. slightly lower score than the true amateur motions. Professionalism evaluation on synthetic data. We conducted a perceptual survey to evaluate the quality of our results when the synthetic dataset was used. We compared the results of our method with two baselines [7,8], the input and the ground truth, considering the following three aspects of professionalism: (i) the smoothness and fluency of the dance motions; (ii) the naturalism of the dance physical amplitudes; and (iii) the dance-to-music synchronization. For this evaluation, we randomly selected seven motions, each from a different dance genre. We recruited 20 participants, 15 amateur dancers with less than one year of dance experience and 5 expert dancers with more than eight years of experience. Each participant was shown, in total, 28 pairs of dance motions; each pair included one generated by our approach, and the other from the ground-truth dataset or generated using one of the two baselines. For each pair, in three independent questions, the participants were asked to select the dance motion that: (i) is smoother and more fluid, (ii) has more natural physical amplitudes, and (iii) is better synchronized to the music. All experimental dance motions were randomly ordered to avoid learning effects. The answers were gathered to quantify the overall professionalism of the dance motions. Results of the perceptual study are shown in Fig. 8, which lists the average percentage of participants who preferred our results over the results of the two baseline methods, the input and the ground-truth. It can be observed that our method received higher scores than the two baselines and the input for all three aspects of professionalism, in the votes of both amateur and expert participants. Apart from smoother and more natural motion, we believe that the better dance- to-music alignment plays an important role in these results. As expected, both the amateur and dance expert participants gave higher scores to the groundtruth motions than to ours. Professionalism evaluation on real motions. Finally, we used the 12 motion-captured dance sequences performed by amateur dancers to further evaluate our method on real amateur data. The motion-captured dances were performed by amateur dancers, who imitated 12 ground-truth professional dances chosen from our testing dataset. In this survey, we recruited 20 participants, 15 amateur dancers and 5 expert dancers. As in the previous study, each participant was shown pairs in random order of true amateur motions and our enhanced results, and asked to select the motion that: (i) is smoother and more fluid, (ii) has more natural physical amplitudes, and (iii) has better synchronization to the given music. Figure 9 presents the results of this study. Compared to the input amateur dances, our enhanced results were preferred by most amateur and expert participants; our results scored significantly higher with respect to motion fluency and music synchronization. As expected, our method performs worse on the true amateur dataset than the synthetic dataset, because our network has been explicitly trained using synthetic amateur data. An interesting future problem is to enrich our training dataset with true amateur dances or to better simulate synthetic data so that they can better execute real amateur dance motions. Ablation study To evaluate the contribution of the dance-to-music alignment stage and the necessity of each of its components, we conducted an ablation study which evaluated several variations of the proposed network, by removing or replacing key components with other alternatives. In detail, we assessed our network: (i) without integrating the dynamic time warping module (W/O DTW); and (ii) without combining the dance-to-music alignment stage (W/O Alignment). Table 2 reports the results of the ablation study; for visual comparisons, please refer to our supplementary material in the ESM. Temporal-alignment stage. If the temporal alignment stage is not integrated into our framework (W/O Alignment), we concatenate the input music and amateur dance, and provide them as input to the dance encoder during the dance enhancement stage. The encoded latent features are directly fed into the decoder without warping. Table 2 lists results using this setup; we can easily observe the necessity of having the temporal alignment stage: the network cannot implicitly learn the temporal warping from the convolutional layers. Dynamic-time-warping component. To evaluate the effect of the dynamic-time-warping component, we use temporal attention mechanisms as an alternative to our learning alignment method. Without the dynamic-time-warping component, we built the optimal temporal alignment path for each music frame by selecting the motion frame with the maximum attention value in the affinity matrix. More details of how we have built and trained the affinity matrix can be found in the supplementary material in the ESM. The results in Table 2 confirm that the performance of the attentionbased implementation (W/O DTW) is worse than the original implementation. To better demonstrate the results, we show alignment results of our method and the attention-based implementation in Fig. 10. In each case, the blue background gives the T × T temporal affinity matrix (rows are motion frames, and columns are music frames), and it is overdrawn by the optimal alignment path (yellow curve). The white color indicates low alignment correspondence between the motion and music frame. It can be observed that the optimal alignment path produced by the W/O DTW approach is scattered: aligned poses between neighboring frames may have large changes, causing motion jitters. Instead, our optimal alignment path is continuous and monotonic. This validates the necessity for a separate dynamic time warping component to give temporal alignment. Dance enhancement stage. To examine the impact of our deep neural architectures, we reimplemented the dance enhancement network with two baseline structures, ConvNet and Transformer. The ConvNet encoder is composed of three Conv-BN-ReLU blocks. The encoder of Transformer is implemented as a shallow network with two attentionforward blocks, while our method (ConvNet with a Transformer) concatenates three Conv-BN-ReLU blocks and two attention-forward blocks. The decoder for all three structures is implemented as the threelayer MLP. In this experiment we kept the same parameters, for the dance-to-music alignment stage, for all three structures. The last three rows in Table 2 show that our network's structure performs better than Transformer. Compared to ConvNet, our performance is slightly better in pose error, but a little worse in FID. Since pose error measures pose similarity to the ground-truth per frame while FID measures overall kinematic feature distribution, we believe that the pose error metric is more important when evaluating visual effects. As Fig. 11 shows, the enhanced poses produced by our structure are visually closer to the ground-truth than when using ConvNet or Transformer alone. This evaluation indicates that the convolution layers are essential for encoding dance features with temporal contextual information. Application: Dance-to-music synchronization One of the main features of our method is that it aligns 3D motion data with audio files in the presence of non-uniform and irregular misalignment. This feature enables some very interesting applications, where the same dance can be reused with audio files with different beats. Figure 12 shows an example of such dance-to-music synchronization. It can be Fig. 11 Ablation study: visual comparison between our final configuration, and when using only ConvNet or Transformer. Our final structure produces dance poses closest to the ground-truth. observed that the input and output dance sequences share similar poses, but are temporally misaligned since they are artificially synchronized to music files played at different beats. To the best of our knowledge, there are no other works in the literature that do dance-to-music synchronization in 3D motion. Our approach enables data reuse, and puts the foundations to facilitate future development in this important application area. Limitations and future work Our study mainly focuses on two attributes of dance professionalism, extending specific physical amplitudes via spatial amplitude enhancement, and making dances fluid by temporal motion-rhythm synchronization. However, dance professionalism is also correlated with other semantic attributes, such as smoothness, energy, balance, and aesthetics. In future work, it would be interesting to investigate these attributes, and design algorithms that emphasize semantics in dances, e.g., to enhance the aesthetics of the input dance. Furthermore, our framework modifies input amateur dances based on their original content. No additional constraints have been considered for adding or deleting poses in the original amateur dances. Therefore, when choreographic errors exist in the input amateur dances, or their choreography is not rich or diverse enough, our method cannot improve it. A possible future direction is to use motion motifs [58] to learn fine-grained mappings between professional and non-professional dances, and to build a knowledge code-book for dance enhancement, similar to the concept of Ref. [41]. Last but not least, our framework is built on a paired professional and synthetic non-professional dance motion dataset. When the input amateur dance contains poses far away from the distribution of dances in the dataset, our method may produce unsatisfactory results. A future improvement would be to enrich the synthetic non-professional dataset with real captured amateur dance motions, or to design an unpaired dance enhancement approach by leveraging characteristics of different dance genres. We would also like to experiment with other pose representations, e.g., see Refs. [59,60], to avoid the use of inverse kinematics to restore joint rotations, and avert potential rotation discontinuities caused by the network. Finally, our method can be used to improve e-learning applications, e.g., for XR systems when users try to learn dance with a virtual avatar. Conclusions In this paper, we have presented a deep learning framework that enhances professionalism of amateur dances, satisfying three main professionalism properties: fluid dance movements, physical amplitudes, and temporal alignment of dance and music, without changing the content of the original choreography. The framework consists of a dance-to-music alignment stage and a dance-enhancement-stage, the first learning an optimal temporal alignment path between the input dance and the accompanying music, and the second enhancing the dance motion in both spatial and temporal domains. We have also presented a key-pose based dance augmentation scheme that artificially generates non-professional dance data from the AIST++ [3] dataset. We demonstrate the effectiveness of our framework by comparing it to two baseline style transfer methods [7,8] via a qualitative visual survey, quantitative metrics, and a perceptual study. We have also presented a useful application that reuses existing dance motion files by synchronizing them with audio files with a different rhythm.	2022-07-11T21:38:15.996Z	2023-03-31T00:00:00.000	{ "year": 2023, "sha1": "ff089b9c2dee7ddd9f12c3bf567d14ce15f594ec", "oa_license": "CCBY", "oa_url": "https://link.springer.com/content/pdf/10.1007/s41095-022-0292-6.pdf", "oa_status": "GOLD", "pdf_src": "Springer", "pdf_hash": "68f1fc461869d7da022f1544a09692079596ddaf", "s2fieldsofstudy": [ "Computer Science" ], "extfieldsofstudy": [] }
257253108	pes2o/s2orc	v3-fos-license	SIV Infection Regulates Compartmentalization of Circulating Blood Plasma miRNAs within Extracellular Vesicles (EVs) and Extracellular Condensates (ECs) and Decreases EV-Associated miRNA-128 Background: This is Manuscript 1 of a two-part Manuscript of the same series. Here, we present findings from our first set of studies on the abundance and compartmentalization of blood plasma extracellular microRNAs (exmiRNAs) into extracellular particles, including blood plasma extracellular vesicles (EVs) and extracellular condensates (ECs) in the setting of untreated HIV/SIV infection. The goals of the study presented in this Manuscript 1 are to (i) assess the abundance and compartmentalization of exmiRNAs in EVs versus ECs in the healthy uninfected state, and (ii) evaluate how SIV infection may affect exmiRNA abundance and compartmentalization in these particles. Considerable effort has been devoted to studying the epigenetic control of viral infection, particularly in understanding the role of exmiRNAs as key regulators of viral pathogenesis. MicroRNA (miRNAs) are small (~20–22 nts) non-coding RNAs that regulate cellular processes through targeted mRNA degradation and/or repression of protein translation. Originally associated with the cellular microenvironment, circulating miRNAs are now known to be present in various extracellular environments, including blood serum and plasma. While in circulation, miRNAs are protected from degradation by ribonucleases through their association with lipid and protein carriers, such as lipoproteins and other extracellular particles—EVs and ECs. Functionally, miRNAs play important roles in diverse biological processes and diseases (cell proliferation, differentiation, apoptosis, stress responses, inflammation, cardiovascular diseases, cancer, aging, neurological diseases, and HIV/SIV pathogenesis). While lipoproteins and EV-associated exmiRNAs have been characterized and linked to various disease processes, the association of exmiRNAs with ECs is yet to be made. Likewise, the effect of SIV infection on the abundance and compartmentalization of exmiRNAs within extracellular particles is unclear. Literature in the EV field has suggested that most circulating miRNAs may not be associated with EVs. However, a systematic analysis of the carriers of exmiRNAs has not been conducted due to the inefficient separation of EVs from other extracellular particles, including ECs. Methods: Paired EVs and ECs were separated from EDTA blood plasma of SIV-uninfected male Indian rhesus macaques (RMs, n = 15). Additionally, paired EVs and ECs were isolated from EDTA blood plasma of combination anti-retroviral therapy (cART) naïve SIV-infected (SIV+, n = 3) RMs at two time points (1- and 5-months post infection, 1 MPI and 5 MPI). Separation of EVs and ECs was achieved with PPLC, a state-of-the-art, innovative technology equipped with gradient agarose bead sizes and a fast fraction collector that allows high-resolution separation and retrieval of preparative quantities of sub-populations of extracellular particles. Global miRNA profiles of the paired EVs and ECs were determined with RealSeq Biosciences (Santa Cruz, CA) custom sequencing platform by conducting small RNA (sRNA)-seq. The sRNA-seq data were analyzed using various bioinformatic tools. Validation of key exmiRNAs was performed using specific TaqMan microRNA stem-loop RT-qPCR assays. Results: We showed that exmiRNAs in blood plasma are not restricted to any type of extracellular particles but are associated with lipid-based carriers—EVs and non-lipid-based carriers—ECs, with a significant (~30%) proportion of the exmiRNAs being associated with ECs. In the blood plasma of uninfected RMs, a total of 315 miRNAs were associated with EVs, while 410 miRNAs were associated with ECs. A comparison of detectable miRNAs within paired EVs and ECs revealed 19 and 114 common miRNAs, respectively, detected in all 15 RMs. Let-7a-5p, Let-7c-5p, miR-26a-5p, miR-191-5p, and let-7f-5p were among the top 5 detectable miRNAs associated with EVs in that order. In ECs, miR-16-5p, miR-451, miR-191-5p, miR-27a-3p, and miR-27b-3p, in that order, were the top detectable miRNAs in ECs. miRNA-target enrichment analysis of the top 10 detected common EV and EC miRNAs identified MYC and TNPO1 as top target genes, respectively. Functional enrichment analysis of top EV- and EC-associated miRNAs identified common and distinct gene-network signatures associated with various biological and disease processes. Top EV-associated miRNAs were implicated in cytokine–cytokine receptor interactions, Th17 cell differentiation, IL-17 signaling, inflammatory bowel disease, and glioma. On the other hand, top EC-associated miRNAs were implicated in lipid and atherosclerosis, Th1 and Th2 cell differentiation, Th17 cell differentiation, and glioma. Interestingly, infection of RMs with SIV revealed that the brain-enriched miR-128-3p was longitudinally and significantly downregulated in EVs, but not ECs. This SIV-mediated decrease in miR-128-3p counts was validated by specific TaqMan microRNA stem-loop RT-qPCR assay. Remarkably, the observed SIV-mediated decrease in miR-128-3p levels in EVs from RMs agrees with publicly available EV miRNAome data by Kaddour et al., 2021, which showed that miR-128-3p levels were significantly lower in semen-derived EVs from HIV-infected men who used or did not use cocaine compared to HIV-uninfected individuals. These findings confirmed our previously reported finding and suggested that miR-128 may be a target of HIV/SIV. Conclusions: In the present study, we used sRNA sequencing to provide a holistic understanding of the repertoire of circulating exmiRNAs and their association with extracellular particles, such as EVs and ECs. Our data also showed that SIV infection altered the profile of the miRNAome of EVs and revealed that miR-128-3p may be a potential target of HIV/SIV. The significant decrease in miR-128-3p in HIV-infected humans and in SIV-infected RMs may indicate disease progression. Our study has important implications for the development of biomarker approaches for various types of cancer, cardiovascular diseases, organ injury, and HIV based on the capture and analysis of circulating exmiRNAs. Introduction A novel mechanism of gene regulation in host health and disease emerged with the discovery of non-coding RNAs (ncRNAs), including microRNAs (miRNAs), which are about~20-22 nt long. miRNAs bound to specific messenger RNAs (mRNAs) are directed to miRNA-induced silencing complex to downregulate their expression by triggering mRNA degradation or repression of translation [1,2]. More than 60% of mRNAs in the mammalian genome can be targeted by a single miRNA [3]. The human transcriptome is predicted to be regulated by miRNAs [4] with roles in proliferation, apoptosis, cellular development, cellular signaling, development of cancer, substance use disorder, and HIV pathogenesis [5][6][7][8]. After the discovery of miRNAs as posttranscriptional regulators of gene expression, it was shown that miRNAs were transported from one cell to another, where they could regulate gene expression despite not being synthesized by the target cells, but their carriers were not known. Originally associated with the cellular microenvironments, circulating or extracellular miRNAs (exmiRNAs) have been shown to be present in various extracellular environments, including saliva, tears, urine, breast milk, colostrum, peritoneal fluid, cerebrospinal fluid, bronchial lavage, seminal fluid, as well as blood serum and plasma [9,10]. The RMs used in this study are listed in Figure 1A and Table 1, while the workflow for the isolation of EVs, ECs, and their characterization are shown in Figure 1B. We used PPLC [13] to isolate EVs and ECs from the blood plasma of RMs and separate them from other extracellular particles in the blood. A representative elution spectrum showing the locations of EVs and ECs is shown in Figure 1C. The spectra depict regions of different analytes, with the first peak (blue box) from fractions 66 to 80 marking the EV-enriched region, and the last peak (green box) from fractions 233 to 264 marking EC-enriched regions [13]. The elution profiles from individual RMs (n = 15) showed no significant variation among Viruses 2023, 15, 622 4 of 28 animals or groups, as expected. The EVs (blue box) and ECs (green box) were collected, aliquoted, and stored at −80 • C until used. To assess the morphological differences between EVs and ECs, pooled (n = 15) samples were subjected to negative stain transmission electron microscopy (TEM). The TEM data show the successful separation of EVs from ECs ( Figure 1D). As revealed by TEM, the EC fraction is enriched in membrane-less particles that are <20 nm in size (green arrow, Figure 1D). Immunogold labeling showed a positive signature of CD9 on the surface of EVs (blue arrow, Figure 1D). Nanoparticle tracking analysis (NTA) of EV size and concentration was conducted. On average, EVs from healthy RMs have a size of 141.83 nm ( Figure 1E), concentration of 2.92 × 10 8 particles/mL ( Figure 1F), and surface charge, measured as zeta-potential (ζ-potential) of −39.06 mV ( Figure 1G). While no significant differences were observed in EV size, concentration, and ζ-potential, there were individual variabilities in EV size, concentration, and ζ-potential. EC size, concentration, and ζ-potential were not measured as the average EC particle primary size was lower than the 20 nm detection limit of the Zetaview PMX 110 (https://www. excilone.com/client/document/particle-metrix--zetaview-brochure-0319_en_540.pdf, accessed on 20 November 2022) with the standard laser configuration, therefore preventing the acquisition of accurate and reproducible data. Complexity and Enrichment of Circulating miRNA in EVs and ECs The miRNAome of EVs and ECs were evaluated using state-of-the-art low-bias sRNA-Seq technology-RealSeq ® [51]. Briefly, total RNA was extracted from paired EVs and ECs isolated from 100 µL of plasma per sample. EV RNA yields ranged between 119 and 1116 ng, with an A260/A280 ratio of 1.37 to 1.62. EC RNA yields ranged between 199.5 and 457.5 ng, with an A260/A280 ratio of 1.3 to 1.61. The number of detectable miRNAs significantly increased (30.2%) in ECs relative to EVs (Figure 2A). Detectable miRNAs in EVs ranged from 52 to 222, with an average of 138 mRNAs per sample (Figure 2A), while detectable miRNAs in ECs ranged from 166 to 310, with an average of 257 ( Figure 2A) per sample. As shown in Figure 2B, 34.6% of all miRNAs detected in paired ECs were not present in EVs, while 14.9% of miRNAs detected in paired EVs were unique to EVs. The 15 RMs were randomly assigned into five groups of n = 3 to identify common miRNAs enriched in EVs and ECs using a five-way Venn diagram analysis. For miRNA to be included in the five-way Venn analysis, we used a cutoff miRNA distribution count of ≥1 miRNA for each group. The miRNAs were inputted into InteractiVenn to identify common miRNAs within a group (n = 3) and repeated for all five groups. The identified common miRNAs for all five groups were then inputted into InteractiVenn to identify common miRNAs. We identified 19 common miRNAs in EVs ( Figure 2C) and 114 miRNAs in ECs ( Figure 2D). Analysis of the 19 common EV-associated miRNAs by distribution count identified various Let-7 families of miRNAs, including Let-7a-5p, Let-7c-5p, Let-7f-5p, Let-7g-5p, Let-7d, Let-7b-5p, as well as miR-26-a5p, miR-191-5p, miR-16-5p, and miR-23a-3p as the highest (top 10) detectable miRNAs ( Figure 2E). Among the 114 common EC-associated miRNAs, we identified miR-16-5p, miR-27a-3p/miR27b-3p, miR-191-5p, miR-376c-3p, miR-451, miR-23a-3p, miR-23b-3p, miR-221-3p, and miR-21-5p as the highest (top 10) detectable miRNAs ( Figure 2F). Previously, Arroyo et al. reported that let-7a-5p is exclusively associated with EVs [52]. In our dataset, although the majority of the let-7 family of miRNAs were amongst the top 10 miRNAs in EVs, they were also present in ECs at very low counts compared to EVs ( Table 2). The raw miRNA counts for all uninfected samples are provided in Supplemental Table S1. miRNAs in EVs, they were also present in ECs at very low counts compared to EVs 2). The raw miRNA counts for all uninfected samples are provided in Supplementa S1. Predictive Functional and Pathways Categories of Differentially Enriched miRNAs in EVs and ECs Based on predicted targets of the top ten miRNAs ( Figure 2E,F) using MIENTUR-NET miRTarbase, which uses data from experimentally validated miRNA-target interactions [53], the functional categories of miRNAomes in EVs and ECs were identified. In EVs, miRNA-target enrichment analysis identified 839 target genes (p < 0.05 and FDR < 0.01) for miRNAs in Figure 2E. The top ten most significant target genes included MYC, TXLNG, TMTC3, GGA3, CDV3, ONECUT2, NAT8L, ECHDC1, DIABLO, and BAZ1B ( Figure 3A). Let-7d was excluded from the miRNA-target enrichment analysis because, at the time of analysis, Let-7d was not found in the miRTarbase database, used by MIENTURNET and DIANA. In ECs, miRNA-target enrichment analysis identified 737 target genes associated with EC miRNAs (p < 0.05 and FDR < 0.01), shown in Figure 2F. The top ten most significant target genes identified were TNPO1, PNRC2, SOCS6, RNF38, PSAP, TOPBP1, FLNA, DYNC2LI1, UTP14A, and PDCD6IP. ( Figure 3B). Similar to Let-7d, miR-451 was excluded from the miRNA-target enrichment analysis, as it was not found in the miRTarbase database. Visualization of the EVs ( Figure 3C) and ECs ( Figure 3D) miRNA-target interaction networks depicts the connections between the top 10 miRNAs and the predicted genes they are associated with. The blue circles represent the miRNAs, while yellow circles represent the target genes. Pathways are represented by blue lines. Let-7b-5p in EVs and miR-16-5p in ECs appeared as major hubs in the networks, as they had several interactions with other elements in the network, and thus, appeared likely to have more influence on the networks ( Figure 3C,D). Based on the miRNA-target interactome, the genes predicted to be targeted by the greatest number of miRNAs (n = 8) was MYC for EV ( Figure 3A), and the panel of miRNAs is shown in Figure 3C. Likewise, TNPO1 is the gene predicted to be targeted by the greatest number of miRNAs (n = 6) for ECs ( Figure 3B), with the network of miRNA shown in Figure 3D. KEGG pathway analysis identified the target genes of multiple miRNAs in EVs to be associated with: microRNAs in cancer, human cytomegalovirus infection, Hepatitis C, bladder cancer, non-small-cell lung cancer, proteoglycans in cancer, cellular senescence, melanoma, glioma, and chronic myeloid leukemia, JAK-STAT signaling, p53 signaling, EGFR tyrosine kinase inhibitor resistance, IL-17 signaling pathway, inflammatory bowel disease, Th17 cell differentiation, and cytokine-cytokine receptor interaction ( Figure 3E). In ECs, the KEGG pathway analysis suggested the target genes of multiple miRNAs to be associated with: PI3K-Akt signaling, microRNAs in cancer, pancreatic cancer, prostate cancer, hepatitis C, melanoma, glioma, measles, breast cancer, cellular senescence, p53 signaling, lipid and atherosclerosis, Th1 and Th2 cell differentiation, Th17 cell differentiation, and adherens junctions ( Figure 3F). A closer look into the KEGG-predicted functions/pathways revealed that 10 and 8 KEGG pathways were unique to EVs and ECs, respectively, while seven functions/pathways were common to both EVs and ECs ( Figure 3G). Disease ontology predicted that EV-associated miRNAs may target the function of diseases/disorders such as ovarian cancer, malignant ovarian surface epithelial-stromal neoplasm, ovary epithelial cancer, bone marrow cancer, stomach carcinoma, neuroblastoma, lymphoblastic leukemia, malignant glioma, and acquired immunodeficiency syndrome, while EC-associated miRNAs were predicted to target diseases such as prostate cancer, male reproductive organ cancer non-small-cell lung carcinoma, gastrointestinal system benign neoplasm, lymphoblastic leukemia, multiple myeloma, neuroblastoma, Tcell leukemia, malignant glioma, chronic lymphocytic leukemia, and acquired immunodeficiency syndrome. Furthermore, WikiPathways analysis suggested that the target genes of multiple miR-NAs in EVs were associated with non-small cell lung cancer, hepatitis C and hepatocellular carcinoma, bladder cancer, RAC1/PAK1/p38/MMP2 pathway, aryl hydrocarbon receptor, signaling pathways in glioblastoma, pancreatic adenocarcinoma, breast cancer pathway, DNA damage response, miRNA regulation of DNA damage response, as well as IL-6 signaling and cytokines and inflammatory response. In ECs, the WikiPathways analysis suggested that the target genes of multiple miRNAs were associated with signaling pathways in glioblastoma, PI3K-Akt signaling pathway, pancreatic adenocarcinoma pathway, focal adhesion, integrated breast cancer pathway as well as interleukin-11 signaling pathway, extracellular vesicle-mediated signaling in recipient cells, VEGFA-VEGFR2, hypothesized pathways in the pathogenesis of cardiovascular disease, signaling pathway and brain-derived neurotrophic factor (BDNF) signaling pathway. Additionally, Reactome analysis suggested that the target genes of multiple miRNAs in EVs were associated with cellular senescence, MAPK family signaling cascades, constitutive signaling by EGFRvIII, signaling by EGFRvIII in cancer, GRB2 events in ERBB2 signaling, regulation of RUNX1 expression and activity, cellular responses to stress, oncogene-induced senescence, mitotic G1-G1/S phases, MAPK6/MAPK4 signaling, as well as VEGFA-VEGFR2 pathway, signaling by VEGF, and interleukin-4 and 13 signaling. In ECs, the Reactome analysis suggests that the target genes of multiple miRNAs are associated with transcriptional regulation by RUNX1, regulation of RUNX1 expression and activity, mitotic G1-G1/S phase, cyclin D-associated events in G1, G1 phase, interleukin-4 and 13 signaling, transcriptional regulation by RUNX3, cellular senescence, disease of signal transduction, constitutive signaling by AKT1 E17k in cancer, as well as interleukin-7 signaling. Common and Unique EV/EC-Associated miRNA-Linked Biological Networks and Functional Pathways The focus of the analysis in this section was the 19 and 114 miRNAs ( Figure 2C,D) common to EVs and ECs isolated from the 15 study subjects. Two-way Venn diagram analysis revealed that 18 of the miRNAs are present in both EVs and ECs, with 1 miRNA unique to EVs, while 96 miRNAs are unique to ECs ( Figure 4A). The distribution of the unique miRNAs associated with EVs ( Figure 4B) and ECs ( Figure 4C) revealed that the counts of EC-associated miRNAs are higher than EV-associated miRNAs. EVs contained one unique miRNA, miR-6132 ( Figure 4B). ECs contained 96 unique miRNAs, with the top ten by distribution count, including miR-27a-3p/miR-27b-3p, miR-451, miR-21-5p, miR-185-5p, miR-376b-3p, miR-376a-3p, miR-17-5p, miR-126, and miR-423-5p ( Figure 4C). To gain insight into functional categories and for an assessment of how these miRNA-mRNA regulatory networks in EVs and ECs may be affected, we examined miRNA-linked biological networks and canonical pathways of the unique EVs and ECs. miRNA-target enrichment analysis of miR-6132 revealed TRIM29, THBD, POU4F1, PHF21A, PCBD1, NECAP1, LAMA5, HSPB6, GALNS, FOXI2, CRABP2, CPLX1, CACFD1 as the top significant target genes ( Figure 4D), and the visualization network showed that miR-6132 is significantly linked to 109 target genes ( Figure 4E). Similarly, miRNA-target enrichment analysis of the EC miRNAs identified APEX1, UBR5, SOCS6, GPAM, TMSB10, RNF139, MGEA5, DERL1, DCUN1D4, and ATN1 as top significant target genes ( Figure 4F). The visualization network showed that miR-21-5p has the most connections ( Figure 4G). KEGG pathway analysis suggested that the top ten detectable EC miRNAs are implicated in bladder cancer, pancreatic cancer, and chronic myeloid leukemia, among others ( Figure 4H). Disease ontology revealed EC miRNAs to be linked to sensory system cancer, ocular cancer, retinal cancer, renal cell carcinoma, urinary system cancer, connective tissue cancer, pancreatic cancer, musculoskeletal system cancer, as well as neuroblastoma, malignant glioma, and neuroendocrine tumor. In ECs, the WikiPathways analysis suggests that the target genes of multiple miRNAs are associated with the integrated breast cancer pathway, pancreatic adenocarcinoma pathway, G1 to S cell cycle control, bladder cancer, DNA damage response (only ATM dependent), retinoblastoma gene in cancer, hepatitis C and hepatocellular carci-noma, androgen receptor signaling pathway, signaling pathways in glioblastoma, as well as viral acute myocarditis, extracellular vesicle-mediated signaling in recipient cells, and microRNAs in cardiomyocyte hypertrophy. Moreover, the Reactome analysis suggests that the target genes of multiple miRNAs are associated with interleukin-4 and 13 signaling, cyclin D-associated events in G1, G1 phase, mitotic-G1/S phases, oncogene-induced senescence, regulation of gene expression by hypoxia-inducible factor, signaling by TGFbeta receptor complex, as well as cell surface interactions at the vascular wall. While differential expression analysis did not reveal any significant differences in miRNA distribution counts of the 18 common miRNAs in EVs and ECs (Table 3), principal component analysis (PCA) showed a distinct separation of miRNAs associated with EVs and ECs ( Figure 4I). Intergroup variation was assessed by hierarchical clustering heatmap. There was a complete hierarchical sample clustering between the 18 common EV and EC miRNAs ( Figure 4J). Together, these data suggest that the miRNAs that are common to EVs and ECs have different enrichment levels in blood plasma EVs (BEVs) and blood plasma ECs (BECs). For insight into the potential functions of the common miRNAs, we performed a functional-enrichment analysis. The top ten most significant target genes are shown in Figure 4K, which include MYC, VCL, TMTC3, GGA3, CDV3, FAM222B, DIABLO, COX1, SNRPC, and CCNB2. Visualization of the miRNA-target interaction network revealed a complex network showing that multiple miRNAs may target the same genes. For example, MYC is a target of 10 miRNAs, including let-7a-5p, let-7g-5p, let-7c-5p, miR-26a-5p, let-7f-5p, miR-320b, miR-320a, let-7b-5p, miR-23a-3p, and miR-16-5p ( Figure 4L). KEGG pathway enrichment analysis of the common BEV and BEC miRNAs implicated them in microRNAs in cancer, human cytomegalovirus infection, and Hepatitis C. Interestingly, these miRNAs are also significantly associated with p53 signaling, glioma, lipid and atherosclerosis, and Th17 cell differentiation. Disease ontology revealed that the common BEV and BEC miR-NAs are linked to ovarian cancer, musculoskeletal system cancer, malignant ovarian surface epithelial-stromal neoplasm, ovary epithelial cancer, ovarian carcinoma, connective tissue cancer, female reproductive organ cancer, non-small-cell lung carcinoma, as well as neuroblastoma, lymphoblastic leukemia, T-cell leukemia, and lymphoid leukemia. Furthermore, WikiPathways analysis suggests that the target genes of multiple miRNAs are associated with non-small-cell lung cancer, DNA damage response, hepatitis C, and hepatocellular carcinoma, as well as extracellular vesicles in the crosstalk of cardiac cells, focal adhesions, VEGFA-VEGFR2 signaling pathway, TLR4 signaling/ tolerance, and heart development. Moreover, the Reactome analysis suggested that the target genes of multiple miRNAs are associated with cellular senescence, MAPK family signaling cascades, oncogene-induced senescence, constitutive signaling by EGFRvIII, signaling by EGFRvIII in cancer, as well as VEGFA-VEGFR2 pathway, interleukin-4 and 13 signaling, interleukin-6 family signaling, TCR signaling, and TNFR1-induced NF-kB signaling pathway. Effect of SIV Infection on EV and EC miRNAome It is well established that humans infected with HIV have changes in their miRNA profiles [54], with HIV altering the levels of several miRNAs at various time points after infection [55,56]. We previously showed that HIV infection and cocaine use regulated the repertoire and functions of miRNAs associated with EVs, with specific emphasis on the modulation of miR-128 [57]. To determine the effect of SIV on BEV and BEC miRNAome, three RMs were infected with SIV. Blood plasma was collected from all three RMs 1-month post-infection (MPI) and 5 MPI, as shown in Figure 5A. EVs and ECs were isolated from blood plasma via PPLC [13], as described in Figure 1B. SIV infection at both 1 MPI and 5 MPI had no significant effect on the isolation spectra of EVs and ECs. Similarly, NTA analysis showed that SIV had no significant effect on BEV size, concentration, and ζ-potential at both 1 MPI and 5 MPI compared to their respective pre-infection timepoint. Interestingly, while there were no differences in total EV and EC protein content at the pre-infection time point, in VEH/SIV RMs, total EV protein was significantly lower than total EC protein at both 1 MPI and 5 MPI. Materials and Methods In this two-part Manuscript of the same series, the Materials and Methods used for Manuscript 1 are similar to that of the follow-up study presented in Manuscript 2. Minor differences (if present) are detailed in the various Materials and Methods subsections. Macaques and Viruses (Used for This Study and the Follow-Up Study Presented in Manuscript 2) Pre-infection and pre-treatment blood samples were collected from a total of 15 age and weight-matched Mamu-A0*1 − /B08 − /B17 − specific-pathogen-free (free of SIV, D retrovirus, STLV, and Herpes B) male Indian rhesus macaques (Table 1 of Manuscript 1). The animals were then randomly assigned to five experimental groups. Rhesus macaques in groups 1 to 4 were infected intravenously with 100 TCID 50 dose of the CCR5 tropic SIVmac251 (TNPRC virus isolation and production core). Group 1: (JD66, IN24, JH47, Table 1 [58]. Group 5 (HI78, HN79, HN39) macaques received twice daily injections of THC, similar to groups 2 and 4 but remained SIV uninfected and served as THC-only controls (Table 1 of Manuscript 2). THC (NIDA/NIH) was prepared as an emulsion using alcohol, emulphor, and saline (1:1:18) as vehicle before use. Chronic administration of VEH (groups 1, 3) or ∆ 9 -THC (groups 2, 4, and 5) was initiated four weeks before SIV infection at 0.18 mg/kg, as used in previous studies [58][59][60][61]. This dose of ∆ 9 -THC was found to eliminate responding in a complex operant behavioral task in almost all animals [61]. Beginning the day of SIV infection, the THC dose was increased for each subject to 0.32 mg/kg over a period of approximately two weeks when responding was no longer affected by 0.18 mg/kg on a daily basis (i.e., tolerance developed), and maintained for the duration of the study. The optimization of the THC dosing in RMs accounts for the development of tolerance during the initial period of administration. Because previously published studies [60,61] on this dose of THC showed protection, the same dose was used in this study. Rhesus macaques in groups 3 (VEH/SIV/ART) and 4 (THC/SIV/ART) began combination anti-retroviral (ART) treatments (PMPA 20 mg/kg, FTC or Emtricitabine 30 mg/kg and Dolutegravir 2.5 mg/kg) at 2 weeks post-SIV infection daily by subcutaneous route until 6 MPI. We want to mention that macaques in groups 3 and 4 received two injections of (50 mg/kg of anti-alpha4beta7) (LN60 and LM85) or control IgG (LM56) beginning 4 MPI at three-week intervals before the plasma sample at 5 MPI was collected. Nevertheless, we did not see any differences in EV characteristics. Moreover, viral replication was significantly suppressed by ART at this stage. Note that anti-alpha4beta7 functions by blocking alpha4beta7 positive T cells from trafficking to the intestine and did not have an effect on viral rebound after 6-8 treatments at 3-week intervals ( Table 2 of Manuscript 2). Therefore, it will not have any effects on EV composition, and we did not observe any effect. For studies in Manuscript 1, only pre-infection and pre-treatment blood samples, as well as 1 MPI and 5 MPI blood samples collected from Group 1 (JD66, IN24, JH47) animals, were used. Isolation of EVs and ECsViruses (Used for This Study and the Follow-Up Study in Manuscript 2) The EVs and ECs were isolated from EDTA blood plasma samples using the PPLCbased size exclusion chromatography (SEC), as previously described [36]. Briefly, samples were liquefied at room temperature for 30 min, centrifuged at 2000× g for 10 min, and 10,000× g for 30 min to remove cellular debris and large vesicles. EVs and ECs were purified using a gravity-packed 7-bead gradient (G-10, G-15, G-25, G-50, G-75, G-100, 2% BCL agarose bead standard) into a 100 cm × 1 cm Econo-Column. Elution was achieved by gravity using 0.1× Phosphate Buffered Saline (PBS, Corning, NY, USA). Fractions of 250 µL were collected, and elution profiles were determined by absorbance measurements at 280 nm. The first peak, which contained EVs, and the last peak, which contained ECs, were independently collected and stored in aliquots at −80 • C until further analysis. Transmission Electron Microscopy (TEM) TEM analysis was performed on pooled samples (n = 15) for both EVs and ECs, as previously described [36]. Briefly, 10 µL of pooled BEC sample was applied to a carbon-coated grid and allowed to sit for 30 s. Excess samples were removed with filter paper. The grids were washed with distilled deionized water (ddH2O) twice, stained with 0.7% Uranyl Formate solution for 20 s, and then allowed to air dry. Images were viewed and collected using an FEI Tecnai12 BioTwinG 2 electron microscope. The samples were captured with an AMT XR-60 CCD Digital Camera system. For EVs, 10 µL of pooled sample was incubated with anti-CD9 (Iowa Hybridoma Bank, University of Iowa, Iowa City, IA, 1:50 dilution with 1% BSA-PBS) at 4 • C overnight. The next day, EVs were washed with 0.1% BSA-PBS 5 times and incubated with 10 nm gold-conjugated anti-mouse IgG (Electron Microscopy Sciences, Hatfield, PA, USA, 1:20 dilution with PBS, #25129) for 1 h. Specimens were then washed with deionized water 5 times, followed by a post-stain with uranyl acetate (1%). Specimens were characterized using TEM (JEM1400, JEOL) at the accelerating voltage of 80 kV. Nanoparticle Tracking Analysis (NTA) (Used for This Study and the Follow-Up Study in Manuscript 2) BEV size, concentration, and zeta-potential (ζ-potential) were measured using Ze-taView PMX 110 (Particle Metrix, Mebane, NC, USA) and the corresponding software ZetaView v8.04.02, as previously described [57]. Briefly, the system was calibrated and aligned with 102 nm polystyrene standard beads before the experiment. BEV samples were left at room temperature for 30 min to acclimatize before measurement. Although samples were isolated using 0.1X PBS, during NTA, samples were diluted to appropriate concentrations (1:20,000 to 1:320,000) in filtered ultra-pure water because salts in buffered solutions, such as PBS, may skew sample measurements. All samples were analyzed under the same conditions (room temperature-20 • C to 25 • C, pH 5.8, sensitivity 92, shutter speed 70, and frame rate 30 fps). Triplicate measurements were taken for size and concentration, and each replicate included eleven positions with two cycles of reading at each position. For ζ-potential measurement, data were acquired at least in quintuplicate, and each replicate corresponds to two cycles of reading. EC size, concentration, and zeta-potential were not measured as the average EC particle size was lower than the 20 nm detection limit of the Zetaview PMX 110 with the standard laser configuration, therefore preventing the acquisition of accurate and reproducible data (https://www.excilone.com/client/document/ particle-metrix--zetaview-brochure-0319_en_540.pdf, accessed on 20 November 2022). Total RNA Isolation (Used for This Study and the Follow-Up Study in Manuscript 2) Total RNA were isolated from 100 µL of BEV and BEC samples from each study subject using miRNeasy plasma kit (Qiagen), with the optional on-column DNase-I digestion step. RNA was eluted in 25 µL RNase-free water once and re-eluted with 25 µL RNase-free water. RNA quality control was assessed by Nanodrop 1000 prior to sequencing. Library Preparation and sRNA Sequencing (Used for This Study and the Follow-Up Study in Manuscript 2) Library preparation and sRNA sequencing were performed, as previously described [57], for each subject. Briefly, libraries were amplified by 20 cycles of PCR. Libraries were sequenced in one NextSeq 550 run with the NextSeq 500/550 High Output Kit v2.5 (75 cycles); sequencing was done with Single End 75 nt reads and dual 6 nt indexes. Libraries were loaded at 1.5 pM and sequenced with a RealSeq Biosciences (Santa Cruz, CA, USA) custom sequencing primer for read one; 5% PhiX control was used. FastQ files were trimmed of adapter sequences by using Cutadapt [62] with the following parameters: cutadapt -u 1 -a TGGAATTCTCGGGTGCCAAGG -m 15. Identification of Common miRNAs (Used for This Study and the Follow-Up Study in Manuscript 2) InteractiVenn web tool [63] was utilized to identify common miRNAs for EVs and ECs between the five groups. Briefly, for a miRNA to be included in this analysis, a cutoff of miRNA distribution count of ≥1 was used. miRNAs for each single group were inputted into InteractiVenn to identify common miRNAs by group (n = 3) for all groups. Common miRNAs for all five groups were then inputted into InteractiVenn to identify common miR-NAs for both EVs and ECs. PCA Plot and Heatmap Generation (Used for This Study and the Follow-Up Study in Manuscript 2) ClustVis web tool [64] was utilized to generate both PCA plots and heatmaps using the same input data. Briefly, log-10 normalized miRNA read counts were inputted into ClustVis for each EV and EC, and default data pre-processing options were applied. For PCA plot generation: PC1 was assigned to the X-axis and PC2 to the Y-axis, with a margin ratio of 0.05. For heatmap generation: rows were centered; unit variance scaling was applied to rows. Imputation was used for missing value estimation. Both rows and columns were clustered using correlation distance and average linkage. Identification of Differentially-Enriched miRNAs (Used for This Study and the Follow-Up Study in Manuscript 2) Identification of differentially-enriched miRNA was performed, as previously described [57], using a cutoff for the sum of reads defined as equal or larger than the number of samples being compared per group times two. Thus, for a miRNA to be included in the analysis, the sum of reads should be ≥12. Subsequently, the read counts were log-10 normalized, and two-way ANOVA comparisons between different groups were performed in Prism software. The false discovery rate (FDR) was controlled using the method of Benjamini and Hochberg and was set to <0.05. miRNA-Target Enrichment Analysis (Used for This Study and the Follow-Up Study in Manuscript 2) MIENTURNET web tool [53] was utilized for all miRNA-target enrichment analyses. Briefly, miRTarBase was used for miRNA-target enrichment analyses, with a threshold for the minimum number of miRNA-target interactions of 2, and a threshold for adjusted FDR of 1. Network analyses were performed with the same filter settings, with the selection of strong evidence only. miRTarBase was again used for subsequent functional enrichment analyses (KEGG, REACTOME, WikiPathways, and disease ontology). Validation of miRNA by Real-Time Quantitative PCR (RT-qPCR) RNA was extracted from pooled EVs (n = 5) and ECs (n = 5) and used for cDNA synthesis. RT-qPCR was performed using an ABI 7500 FAST machine and the Thermo Fisher Scientific hsa-miR-128a (Assay ID 002216) TaqMan PCR-specific assays, following manufacturer's instructions. Statistical Analyses (Used for This Study and the Follow-Up Study in Manuscript 2) Statistical tests were performed using the GraphPad Prism (Version 9.3.1) software and are detailed in the figure legends. For two-group comparison, an unpaired t-test with Welch's correction was used to determine the differences between the groups. Ordinary one-way ANOVA (Brown-Forsythe and Bartlett tests, with Sidak's multiple comparison tests) was used to determine the differences between multiple groups. Discussion In the present study, we found that in healthy RMs, the circulating EV-miRNA profile is significantly different from those identified in ECs. Notably, a greater number of circulating plasma miRNAs were associated with ECs than EVs, and over time, SIV infection altered the level of miRNAs in EVs and ECs ( Figure 6). Indeed, studies on miR-NAs have burgeoned since their discovery as mediators of cell-cell communication, and in particular, their presence in the extracellular space. There is no doubt that circulating miRNAs are wrapped within EVs and other lipid-containing membranous vesicles, as well as protein complexes [52,65,66]. In our study, we found that in healthy RM blood plasma, miR-6132 may be specifically associated with EVs, while miR-27a-3p, miR-27b-3p, miR-451, miR-21-5p, miR-185-5p, miR-376b-3p, miR-376a-3p, miR-17-5p, miR-126, miR-423-5p were uniquely associated with ECs ( Figure 4A). Additionally, there were miRNAs shared between EVs and ECs with variable intensities. These include the let-7 family of miRNAs, one of which (Let-7a) was previously reported to be exclusively associated with EVs, while miR-16 and miR-92a are outside of known membrane-containing particles [52]. In our study, the majority of the let-7 family of proteins were present in both carriers, although let-7a-5p, let-7c-5p, let-7f-5p, let-7g-5p, let-7d, and let-7b-5p were amongst the top 10 EV-associated miRNAs, their counts in ECs were significantly lower compared to their counts in EVs ( Table 2), suggesting that EVs may be the main carriers of let-7 family of miRNAs. The significance of the enrichment of the let-7 family of miRNAs in EVs is yet to be determined. However, let-7 miRNAs regulate CNS inflammation and neurological outcomes through the suppression of neuroinflammation [67] and the production of proinflammatory cytokines, such as TNFα expression [68,69]. Overexpression of Let-7g was shown to preserve-brain barrier integrity, reduce proinflammatory cytokine release, and prevent immune cell infiltration into the infarcted region, leading to improved behavioral outcomes [70][71][72]. With respect to SIV infections, EV-associated miR-128a-3p, miR-378d, and miR-99a-5p and EC-associated miR-656-3p and miR-671-5p were longitudinally decreased and increased, respectively, by SIV infection. The longitudinally decreased and increased miRNAs in EVs and ECs, respectively, in VEH-treated SIV-infected RMs may indicate disease progression. It is interesting that SIV infection leads EV-associated miRNAs to decrease, while EC-associated miRNAs tend to increase ( Figure 5C-F). The significance or implication of this regulatory pattern is yet to be determined. Of the miRNAs that were longitudinally regulated in EVs, the decrease in miRNA-128a-3p level agrees with our published observation in people living with HIV (PLWH), where HIV infection with or without concomitant cocaine use was shown to decrease the levels of miR-128 in human semen-derived EVs [57] (Table 5). These observations are remarkable because miRNA pathways have been shown to regulate HIV replication and contribute to latency [73]. Several miRNAs, including miRNAs-28, -29, -125b, -150, -223, and -382, negatively affect HIV by directly targeting the viral RNA genome and/or by repressing virus-dependent cellular cofactors [74][75][76]. For example, the Let-7 family of miRNAs was shown to be With respect to SIV infections, EV-associated miR-128a-3p, miR-378d, and miR-99a-5p and EC-associated miR-656-3p and miR-671-5p were longitudinally decreased and increased, respectively, by SIV infection. The longitudinally decreased and increased miR-NAs in EVs and ECs, respectively, in VEH-treated SIV-infected RMs may indicate disease progression. It is interesting that SIV infection leads EV-associated miRNAs to decrease, while EC-associated miRNAs tend to increase ( Figure 5C-F). The significance or implication of this regulatory pattern is yet to be determined. Of the miRNAs that were longitudinally regulated in EVs, the decrease in miRNA-128a-3p level agrees with our published observation in people living with HIV (PLWH), where HIV infection with or without concomitant cocaine use was shown to decrease the levels of miR-128 in human semen-derived EVs [57] (Table 5). These observations are remarkable because miRNA pathways have been shown to regulate HIV replication and contribute to latency [73]. Several miRNAs, including miRNAs-28, -29, -125b, -150, -223, and -382, negatively affect HIV by directly targeting the viral RNA genome and/or by repressing virus-dependent cellular cofactors [74][75][76]. For example, the Let-7 family of miRNAs was shown to be significantly lower in patients with chronic HIV infection compared to healthy controls [77] and several let-7 family members (let-7d-5p, let-7a, let-7c) decreased over time in progressive liver fibrosis in hepatitis C-infected patients [78]. One implication of our study is that EV-and EC-associated miRNAs could potentially serve as biomarkers in various diseases [104]. However, whether or not circulating miR-128-3p associated with EVs or ECs in blood plasma have any role in health and disease is yet to be determined. We do not wish to speculate on the predicted regulatory roles miR-128-3p may play in the pathogenesis of HIV, as the topic deserves further research. However, our group previously showed that the miR-128 network mediates strategic monocyte haptotaxis [57]. It is, therefore, plausible that the decrease in miR-128-3p in HIV-infected humans [57] and the longitudinal decrease in SIV-infected RMs, may indicate disease progression. The origin of circulating EV-and EC-associated miRNAs in plasma remains to be identified. The nonvesicular miRNAs, such as EC-associated miRNAs (Figure 2), constitute a significant fraction of the circulating miRNAome [105]. However, it is worth noting that an observed EV or EC enrichment of a given miRNA may not necessarily mean selective release, packaging, enrichment, or function of the miRNA. There may be biological reasons for the association of specific miRNAs with the cell and extracellular carriers. Like cellular miRNAs, circulating miRNAs are effective mediators of intercellular communication. Unlike cellular miRNAs, circulating miRNAs are protected by various extracellular miRNA carriers (EVs, ECs, other RNPs) and are shuttled to proximal or distal sites to exert their functions. Indeed, these different miRNA carriers may have characteristic functions in the extracellular space and in target cells. However, EV-and EC-associated miRNAs may convey messages intracellularly, even if they do not have the 'intention' to do so. As such, EV-and EC-associated miRNAs may serve as novel intercellular communication pathways or as minimally invasive biomarkers [106,107]. Beyond their potential roles as signaling mediators and biomarkers, circulating miR-NAs associated with EVs and ECs may serve to maintain steady-state levels of miRNAs under homeostatic conditions by balancing miRNA synthesis, degradation, excretion, and reuptake by cells. This process may facilitate miRNA-mediated gene silencing in a cell without the cell making the miRNA. In addition, it has been shown that the association of miRNAs with extracellular carriers enhances their stability in austere conditions, including RNase digestion in the bloodstream and extreme pH values and temperatures during handling and storage. The stability of exmiRNA may regulate their degradation and help prevent uncontrolled cytoplasmic RNase activation that may be cytotoxic [108] or promote cellular energetics [109]. Despite the significant findings in this study, the following questions need to be addressed. Studies are needed to assess whether miRNAs are selectively packaged within EVs or ECs and what the significance of such selective association is. It has been suggested that in addition to the active secretion of miRNAs through cell-EVs [110,111] and protein complex lipoproteins, such as high-density lipoprotein-HDL [66] andAgo2 [52]-miRNAs may also be secreted from cells due to injury, chronic inflammation, apoptosis, or necrosis, or from cells, such as platelets with short half-lives [11,112]; hence, studies to identify the sources of the miRNAs associated with EVs and ECs and the mechanisms that control their release. Studies are also needed to clarify if changes in the miRNAome (e.g., during viral infection, licit or illicit drug use) are reflected in EVs and ECs. This last point will be addressed in Manuscript 2 of this series. Conclusions and Translational Relevance To our knowledge, this study is the first of its kind. We used rigorous experimental approaches to profile exmiRNAs, characterize their association with two distinct EPs (EVs, ECs), and elucidate the effect of SIV infection on exmiRNA association with EVs and ECs. The clinical importance of the longitudinally altered EV-and EC-associated miRNAs in the SIV-infected group may indicate disease progression, in which case, such miRNAs may be used as a biomarker to assess the severity of HIV/SIV infection.	2023-03-01T16:13:08.355Z	2023-02-24T00:00:00.000	{ "year": 2023, "sha1": "c44c100d2e480b67399875f677074a5a33fa833b", "oa_license": "CCBY", "oa_url": "https://www.mdpi.com/1999-4915/15/3/622/pdf?version=1677239268", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "9e2a3fb251204669990ebcb53113824e33f0ec50", "s2fieldsofstudy": [ "Biology" ], "extfieldsofstudy": [] }
55993841	pes2o/s2orc	v3-fos-license	Exploring the Coordinated Management of Meaning of Sex: The Social Construction of Male College Student Logical Forces This exploratory qualitative research study examined male college students’ narratives describing memorable sexual experiences, and how those encounters inform current sexual practices. Drawing from a larger collaborative research project, this study explores the narratives of 130 men who were attending college at one of three diverse US campuses in 2012. Utilizing a coordinated management of meaning theoretical frame we examine how sexual experiences are informed by logical forces of looped narratives that exhibit both charmed and strange loops. Findings demonstrate how men’s management of meaning regarding their sexual experiences is informed by larger expectations steeped in rigid masculine values, a major force in socially constructing sexuality. . While most women and men report falling in love at least once, men begin dating at earlier ages than young women do (Carpenter, 2005).In a review of early studies, DeLamater (1987) found that males have a more positive emotional reaction to first intercourse than females do, in part because men are ten times more likely than women to have an orgasm during their first experience (see also Sprecher et al., 1995).In comparison to women, men report feeling more pleasure and less guilt (Darling, Davidson & Passarello, 1992).As illustrated through these studies, the dominant theme of past research investigating gendered perspectives is different, despite the fact that researchers are likely to find as many similarities between women and men's experiences as differences (Canary & Dindia, 1998;Sprecher et al., 1995). More recent research has reinforced the differences between women's and men's sexual experiences through studies that have focused on a variety of issues.For instance, Mongeau and colleges (Mongeau, Serewicz, Morr, & Therrien, 2004;Moor & Mongeau, 2004) found that men and women have different goals during first dates, with men having higher sexual expectations-especially when alcohol was consumed, something that is typical in sexual situations (see also, Abbey, Zawacki, Buck, & Clinto, 2001).Sexual intercourse has been established as a means to relieve stress for both women and men (Ein-Dor & Hirschberger, 2012), however, men's aggressiveness is more dominant than women's, something linked to rigid gender socialization (Mulac, Jansma, & Linz, 2002) and alcohol consumption (Wilson, Calhoun, & McNair, 2002).Gender socialization regarding emotional intimacy expectations also has been studied in terms of sexual experiences.This research has explored the psychological meaning of sexual behavior (Peplau, Rubin, & Hill, 1977), and women and men's emotional support.Disclosing positive feelings for one's partners after sexual activity is positively associated with relational satisfaction, trust, and psychological immediacy (Denes, 2012).Yet, research has demonstrated that these forms of communication are primarily the behaviors that women, and not men, engage in (Burleson, Homstrom, & Gilstrap, 2005).Yet, through an exploration of gendered communicative styles, Sidelinger, Frisby, and McMullen (2009) found that both feminine women and androgynous men were likely to perform emotionally supportive behaviors. While a traditional belief is that heterosexual women are more likely to feel a lack of control involving virginity loss than young men, recent studies have suggested that an increasing number of adolescent men experience similar emotions (Wight et al., 2000).Carpenter's (2005) qualitative research project is especially relevant here.Based on in-depth interviews with 61 US Americans from diverse backgrounds, she explored how perceptions of virginity loss were informed by the preconceived values that they brought to the experiences.Specifically, she discovered three metaphors that reflect different perspectives: virginity as a gift, virginity as a stigma, and virginity loss as a rite of passage1 .According to Carpenter, these metaphors inform larger frameworks through which people understand virginity and influence their sexual practices.For instance, she found that "gifters" were much more likely to practice safer sex compared to individuals who viewed their virginity as a stigma.In terms of sex, most women and men describe their first sexual experiences as satisfying-and even enjoyable (Carpenter, 2005). As demonstrated within this brief introduction, a significant body of literature on sex, gender, and sexual experiences exists.The vast majority of this research positions women's and men's sexual experiences in contrast to one another, with a particular focus on how experiences relate to relational expectations, emotional intimacy, aggressiveness/coercion, alcohol consumption, and other factors (e.g., Palmer, McMahon, Rounsaville, & Ball, 2010).Yet, there is little contemporary scholarly research that focuses on how individuals make sense of their sexual experiences in general, and memorable sexual experiences in particular.Sexuality is a social construction (Berger & Luckmann, 1967) yet researchers have largely failed to study it as such.In an era marked by increased sexual activity, high rates of sexually transmitted infections (STIs) in young adult populations, and billons of dollars of direct costs to treat them (Chesson, Blandford, Gift, Tao, & Irwin, 2004), explorations on this topic can provide a unique vantage point for communication scholars and practitioners.As documented next, existing literature on the coordinated management of meaning (CMM) provides a valuable conceptual framework to explore how men socially construct meanings of their most memorable sexual experiences.Following an examination on existing literature of CMM, we provide the methodology of the study, followed by an articulation of our findings.We then close with discussion of our findings followed by concluding remarks. CMM: Meaning and Action through Storytelling Introduced in the mid-1970s (Pearce, 1976), the coordinated management of meaning (CMM) theory is grounded in a basic premise: Communication is the process by which people "cocreate, maintain, and alter social order, personal relationships, and individual identities" (Cronen, Pearce, & Harris, 1982: p. 64).The theory is conceptualized in broad terms, drawing from many communicative aspects as it explores how reality is constructed in social interaction (Littlejohn, 2009); not surprisingly it has been used in a variety of interpersonal and intercultural contexts (e.g., Bruss et al., 2005;Pearce & Pearce, 2000).CMM theory contends that people coordinate their lives by managing the ways in which messages have meaning for, and through, larger patterns of meaning (Heath & Bryant, 2000).Pearce and Cronen (1980) describe communication as "a process in which each person interprets and responds to the acts of another, monitors the sequence, and compares it to his or her desires and expectations" (p.68).Consequently, a coordinated management of meaning depends on particular interaction rules, the content of messages, and the ways in which various kinds of interaction are structured (Heath & Bryant, 2000). According to CMM, people learn behaviors that are appropriate for specific contexts (Pearce, 1976).Specifically, early conceptualizations focused on the interpretative process through which individuals drew from six levels of understanding to create particular meanings (Pearce & Cronen, 1980;Philipsen, 1995).These six levels are: 1) content (the words used to communicate), 2) speech acts (how we perform the content), 3) contracts (a system of formal and/or informal rules that guide two or more individuals' communication), 4) episodes (communication routines that consist of a describable sequence of speech acts), 5) identity or life scripts (individual's self-perception that shapes, and is shaped by, communication), and 6) cultural archetypes (understandings of speech acts, contracts, episodes, and life scripts that are shared by a particular social group).Collectively, these levels created a hierarchy of meaning whereby understandings at lower levels "were said to be embedded in, and derive their meaning from" (Pearce, 2005: p. 41) Over the past decade, CMM scholars have acknowledged the fluid and dynamic nature of these contexts; accordingly they have adopted models (e.g., LUUUTT model, Daisy Model, SHEDD model) that emphasis the reflexive relationship among contextual levels of understanding (Littlejohn, 2009;Pearce, 2005;Pearce & Pearce, 2001) that serve as the fundamental logic that people use to frame, or define, experience (Heath & Bryant, 2000). CMM is grounded in the premise that communicators engage in two different things in every interaction: "They interpret, or ascribe meaning, and they act-two functions closely tied to one another: Meaning leads to action, and action forms meaning" (italics in original) (Littlejohn, 2009(Littlejohn, : p. 2000)).CMM scholars study how individuals coordinate their meanings and actions over time, often in larger contexts that contain conflicting messages that compete for greater salience (Bruss et al., 2005).According to the theory, this negotiation is best understood through the recognition of two different types of rules: those that are constitutive and regulative (Pearce & Cronen, 1980).Constitutive rules are guidelines that reveal meaning, they help answer the fundamental question of "What does this mean?"Regulative rules, in contrast, are rules of action, they indicate what actions should be taken in any given communicative context.In short, people interpret messages and know what actions constitute appropriate responses because they can follow rules that guide what they do and say (regulative rules) in the context of how they interpret what transpires (constitutive rules) (Heath & Bryant, 2000). Over the course of one's live, rules of meaning and action work to create a logical force that reinforces what is perceived as "logically right" in any given context (Littlejohn, 2009: p. 201).Yet, each individual lives simultaneously in many different social worlds, a reality that means that numerous forms of logic are available to them over the course of their lives (Pearce, 2008).These potentially contradictory sets of regulative and constitutive rules can result in coherence or mystery (Pearce, 2005).When meanings and actions are consistent and reinforcing of one another, a charmed loop is said to exist (coherence).Alternatively, a strange loop is experienced when meanings and actions are contradictory and inconsistent with a person's logical force, something that triggers a quest for new understanding (mystery) (Pearce, 2008).According to CMM, the narratives that an individual shares create a particular social world that includes rules of meaning and action that guide communication with others (Pearce, 2008).Given this, CMM appears as a productive lens through which to explore men's memorable sexual experiences, especially in terms of how individuals negotiate competing messages occurring at different levels of meaning (Bruss et al., 2005).Accordingly, we utilize it as an interpretive lens to gain insight into the following research questions: RQ1: How is meaning situated in male college student's narrations of memorable sexual experiences?and RQ2: What logical forces exist as they create social worlds regarding sex, self, and society? Methods The data analyzed within this manuscript were collected as a part of a large national study.From April-June, 2012 we conducted a research project focusing on college student sexual experiences and knowledge of sexually transmitted infections.Within this section, we describe the participants, survey, and analytic process of the study. Participants and Procedures Participants for this study were recruited from three different college campuses from across the US: 1) a small private urban college located in the Northeast, 2) a large state land grant university located in the upper Midwest, and 3) a mid-sized historically black university in the South.Initially, each of the faculty researchers provided extra credit opportunities to students in their communication classes who volunteered to complete the 10-minute survey on-line.In order to widen and diversify the participant pool, other students across campus beyond this initial scope were also encouraged to participate.Through the process, 476 surveys were collected from the three different campuses: 147 from the Northeast campus, 215 from the Midwest campus, and 114 from the Southern campus.Of the 476 surveys, 130 were completed by men.This smaller data subset, similar in size to other qualitative studies exploring sexual messages (Morgan & Zurbriggen, 2007), is the focus of our current analysis. In order to counter existing criticism on the ways that traditional researchers collect and analyze demographic markers such as race, age, and gender (Houston, 2002), demographic information was collected from participants via an open-ended prompt that asked them to provide a self-description of their identity.The approach mirrors the advice of Martin, Krizek, Nakayama and Bradford (1996) who assert that individuals should be asked to provide their own labels, presumably ones that have meaning for them.Because of the lack of consistency in these self-descriptions, reporting participant demographics with any confidence is difficult2 . Using an on-line survey, we asked participants a few binary questions (e.g., Have you ever had sex?) and several Likert-style questions that asked about levels of knowledge and awareness of Human Papillomavirus (HPV) and comfort in discussing sex with family members and romantic partners.We also included one openended question that asked participants to provide "a brief but detailed story that describes something from your past-a story, memory, experience, and/or message from another source-that has most impacted your CURRENT SEXUAL PRACTICES" 3 .This prompt provided a rich source of detailed data concerning a variety of details regarding narratives about memorable sexual messages.This self-report methodological strategy (Lauckner et al., 2012) was consistent with our desire to have participants "narrate their own experiences within these interactions as opposed to asking participants to respond to topics chosen by the researchers" (Morgan & Zurbriggen, 2007: p. 519).Within this study, participant stories ranged from those that were 1 -2 sentences long to those that were several paragraphs. Thematic Analytical Process In our thematic analysis of narratives, we drew from the work of Owen (1984) and used three criteria (repetition, recurrence, and forcefulness) to assist in the emergence of primary themes (see also Apker, Propp, & Ford, 2005).Specifically, the thematic analytic process included four steps.First, the 130 responses related to memorable sexual experiences were extracted from the larger data set and organized into one Microsoft Excel sheet.Second, we reviewed this data subset to locate frequent appearances of specific words and phrases (repetition) across participant narratives.Third, we recorded how similar meanings were articulated through various articulations (recurrence) from different participants.Fourth, we also took note of the power behind certain excerpts that were emphasized through different codes (e.g., ALL CAPS), punctuation (!!!! or ???), or format (bold or italics).The initial thematization process generated approximately 10 potential themes.At this juncture, we re-engaged the data using these preliminary themes as conceptualized through a coordinated management of meaning theoretical lens.Through this rigorous process, we were able to facilitate another level of analysis whereby several key over-arching memorable experience-based themes could be identified. Thematic Results Our analysis revealed memorable narratives about sex that largely were limited to descriptions of speech acts and episodes4 .These memorable narratives cultivated from what the traditional interpretive approach would place at the most basic levels of understanding, in which content descriptions were offered in a certain manner to reflect a particular speech act, or a series of speech acts leading to an episode.Despite their brevity, at the core of these narratives were various issues such as decision making, negative consequences associated with sex, and positive outcomes of practicing safe sex.For instance, one 35-year-old heterosexual male from Michigan explained, "I developed critical thinking skills", which placed decision-making and agency at the foundation of this memorable narrative.One 19-year-old straight white male simply attributed "Watching porn, I guess", to his understanding of sex.Other participants described memorable sexual experiences where meaning was negotiated at the speech act or episode level of understanding.These primarily centered on the negative consequences of obtaining sexually transmitted infections or the result of pregnancy whereas others interpreted messages of the positive results in being sexually active. One Christian male offered a narrative centered on pregnancy: "Hearing that even with birth control can get a girl pregn[a]nt!So me and my girlfriend now [wear] condoms every time we have sexual intercourse".Similarly, one 19-year-old white single male expressed the fear of pregnancy in his narrative, "My friends having kids at the age of 16 scared the hell out of me".Whereas narratives similar to these demonstrate speech acts centered on unwanted pregnancies, others revealed obtaining sexually transmitted infections (STI's) at the core of their narratives."I really don't know anyone that has gotten a STD or HIV but being a responsible person a condom is a must to pro[t]ect from unplanned birth and STDs or HPVs", described a 20-year-old heterosexual White Catholic male from Michigan. Other narratives unraveled a more optimistic understanding of sex, describing the positive consequences associated sex education and performing sex.One 19-year-old heterosexual white male from Massachusetts explained, "I just feel like education has made it an easy decision to have safe sex", attributing positive outcomes resulting from sexual education.Similarly one 21-year-old straight White male from Michigan stated, "In elementary school a presentation about teen pregnancies has always made me aware of safe sex".For these individuals, certain speech acts or episodes created a positive understanding of sex and the outcomes of "safe sex".Other participants offered narrative that highlighted the pleasure associated with sex."I remember the first time I had sex.It was great.I ejaculated within two minutes", described a 20-year-old Black male.These narratives were primarily descriptive with little context. The various memorable narratives above depicted meaning situated and limited to the speech act or the episode.Narratives expanded beyond the levels of speech act and episode depicted how levels of understanding can shift and become more elaborate with larger frames of meaning-making.For example, an episode could be best understood in the context of the relationship, or the self could be best understood in the context of the speech act.Relationships between meaning and action rules constitute a reflexive relationship between various contextual frames reflective of more elaborate structures. Reflexive Relationships Various narratives constitute a reflexive relationship between meaning and action.A loop metaphor emphasizes the fluidity of contexts and depicts the analytical nature of meaning and action rules.In the traditional hierarchical version of CMM, "the act is understood in terms of the relationship, the relationship in terms of the episode, the episode in terms of the self, and the self in terms of the archetype" (Littlejohn, 2009: p. 200).Utilizing the loop metaphor, however, demonstrates how communicators can shift amongst contexts within given different situations.By providing CMM models that organize each level of understanding for selected narratives, this section will preview how we have provided insight into the meaning making of a few select participant narratives-each of which will demonstrate how meaning was situated in loops of CMM contextual frames. The value of CMM's loop model is clear in participant narratives that offer descriptions that reflect mean-ing-making across contextual frames.This is evident in all of the memorable sexual narratives that we analyze in this section, including the one featured below.For the participant below, a narrative where identity was understood in the context of the archetype was disclosed.As described by this 21-year-old heterosexual African American male: After growing up in a catholic grade school from kindergarten to seventh grade I transferred to a public junior high school.After my first [day] I noticed right away the differences between a catholic and public school.One image that changed my sexual life for the rest of my life before it even beg[a]n was seeing a young girl around my age easily six months pregnant and I realized right away I did not want that to be [associated with] me. For this participant constitutive rules were associated at the archetype level in that the archetype established meaning for how sexual values differ, within his example, from private school to public school (see Figure 1).Transferring from a private school to a public school provided a different culture with different expectations and norms regarding sex.This led to regulative rules that explicitly work to avoid pregnancy while still in school.This narrative is unique in that it does not involve a personal sexual experience, but reflects on an observation of the consequences of someone else's sexual experience. Several narratives associated within the level of episodes portrayed reflexive relationships among various levels of CMM.While issues of male identity and masculinity were implicit in earlier analyzed narratives, the issue was more explicit in other narratives.This was illustrated by one 21-year-old heterosexual man not in a relationship, who described one episode that was associated with identity or life script: Yes, I have had sexual encounter before.I will tell the first time I had sex was my freshmen year in high school.I was kind of nervous because I knew it was soon for me to lose my virginity one night after a party I had went to.My girlfriend was a senior in high school obviously with way more experience than me.I went over her cousins place, we kicked it in the living room for a hot second then started kissing.Soon after we took it to the next room.I came prepared that night with like 10 condoms!I have older brothers and they always taught me to always keep condoms with me.As we were kissing she took her pants off and told me to sit down because she was going to get on top of me.Knowing the big ego I had and for my first experience with this whole ordeal, I told her to lay down and I would start everything off.I took my pants off already aroused, put my condom on slightly nervous and started to have sex.That night I felt good about myself lol. For this participant, constitutive rules were informed by deeper contextual frames but enacted at the episode level in that the episode established meaning for how he should behave (see Figure 2).The participant described how he was not ready for sex, but came prepared to a party with 10 condoms.The episode taking place at his cousin's house after the party created rules of meaning associated with when performing the act of sex should take place.This episode led to regulative rules influenced by his "ego" which coincides with the life script of being a traditional, masculine man.When kissing and told to sit down because she was going to get on top of him, this male participant described how he told her to lay down as he would "start everything off", describing how his "big ego" influenced his actions during his first experience. Other narratives demonstrated masculinity as essential to understanding sex.For instance, one twenty-something straight Caucasian Christian male explained: I'm a male in my twenties.Growing up I was always interested in pornography and would watch it on premiere TV whenever it was on.It introduced me to different things involving sex such as the act and the typical procedure involved with a sexual situation.Once I became sexually active I would refer back to what I had seen growing up and try to get my partners comfortable with what I wanted to do or try.I always wanted to please my partners by giving oral sex and turning them on.Then I would usually go into intercourse.One time I experienced sex with a girl that orgasmed so much that she squirted.I had heard of this before and seen it in porn but had never experienced it personally.I must say, it was very excit[ing] and turned me on greatly!Knowing that I had given my partner such an intense orgasm was such a turn on.From that point on I always strived to give my partners the same experience, not just for their satisfaction but, for mine as well. This narrative reveals understanding of this interaction was reached through five hierarchical levels of meaning, with rules of meaning and rules of action situated in different levels of understanding, creating a loop of meaning making (see Figure 3). For this participant constitutive rules were associated at the archetype level in that a cultural artifact-pornography-established meaning for how he should behave.Once sexually active, the participant would refer back to pornography to get new ideas to try with his partner.This lead to regulative rules of control influenced by his masculinity, part of his identity or life script.Pleasing his partners in the way depicted in pornography is described as this participant's sexual objective, and the ability to control his partners' bodily functions served as the ultimate sexual experience for this male participant. Logical Forces of Looped Narratives Within this final analysis section we demonstrate the ways in which the participants created cognitive connections between rules of meaning and rules of action, a concept known as a logical force.At the heart of this concept is the idea that individuals create different rules of meaning and action depending on the context of the situation that feel "logically correct".Given that every individual lives within various social spheres, many different "logics" of meaning and action are available.At times individuals can be in limbo over competing rules of meaning and rules of actions. Logical forces.For the participants of this study, different themes were found at the foundation of their narratives that formed logical forces used to make sense of memorable sexual experiences.These themes include STI protection and pregnancy avoidance, responsibility, and relationship status. Several participants offered narratives that centered on STI protection and pregnancy avoidance.Some offered how observations of others were used to form logical forces about STI prevention and pregnancy avoidance.This was the case with one 22-year-old white male from Michigan who stated, "I've heard of kids who had unsafe sex and got STDs also pregnant.So [I] always wear a condom!I don't remember the people's names".Others described education as the main factor contributing to their logical forces about STI prevention and pregnancy avoidance.As described by one 21-year-old white male from Michigan, his logical force about pregnancy was cultivated during a school assembly, stating that, "In elementary school a presentation about teen pregnancies has always made me aware of safe sex".Other participants described personal experiences involved with STIs and pregnancy that formed logical forces.One 20-year-old white male provided a narrative about a previous experience with a condom breakingduring sex and how that has impacted his sexual practices, disclosing, "During sexual intercourse when younger the condom broke.I realized how serious the repercussions could be and made sure to be extra careful.Now I am extremely cautious and my partner is now also taking birth control". Logical forces centering on the theme of STI prevention and pregnancy avoidance such as above represent a coherent logical force in that the rules of meaning about unsafe sex leading to obtaining STI's and pregnancy prompted consistent rules of action to practice safe sex.A second clear theme that organically developed in our analysis centered on responsibility. Several participants provided narratives that placed a large importance on responsibility.This second theme differs from the first theme in that these participants provided understandings about sex evolving around being a responsible person, whereas those contributing to the previous theme were specifically concerned with pregnancy and STIs.For example, one 20-year-old male described the teachings he has received from his family about not having sex until marriage: I come from a Christian family and I have been taught since young to not have sex until marriage.I've been taught that sex leaves a lasting impression on both parties and having sex with multiple partners can destroy one spiritually.I decided to not have sex till marriage because I respect my parent's teachings. Similarly, another 20-year-old male from Texas attributes his understanding of sex to the sex education that he received from his current partner, saying, "I've learned a lot from my current partner who is very passionate about sex education.It has influenced me to be more responsible and encouraged me to help friends be more responsible". Narratives such as these focused on the importance of being a responsible person and prompted logical forces between rules of meaning and rules of action.More specifically, these narratives represent contextual forcesways a communicator feels they must act in a given situation (Littlejohn, 2009).Rules of meaning about being sexually responsible through teachings and experiences are consistent with rules of action to practice sexual responsibility. Charmed and Strange Loops The loop metaphor provides insight in how understanding can cultivate from the various levels of understanding of CMM.The reflexive relationship between constitutive rules and regulative rules emphasize the systematic nature of shifting contexts.Taking this concept one step further demonstrates the two different types of loops that form to create understanding.As described earlier, a charmed loop is present when meanings and actions a parallel to one another and (re)produce one another.In contrast, a strange loop is present when meanings and actions seemingly contradict one another.In this section, we will provide examples of charmed loops and strange loops to gain a better understanding of how various loop(ed) narratives had either consistent and (re)productive rules of meaning and action or inconsistent and contradictory rules of meaning and action. Charmed loops.Several participants provided accounts of memorable sexual experiences that demonstrated charmed loops.Themes emerging from charmed loop narratives included various forms of responsibility, pleasure, and masculine values.As one 20-year-old heterosexual male from Houston, TX offered, "I've learned a lot from my current partner who is very passionate about sex education.It has influenced me to be more responsible and encouraged me to help friends be more responsible".For this narrative, constitutive rules associated in the contract of learning about sex education through his partner has in turn led to regulative rules associated in his life script of identity level of practicing safe sex and helping friends be more responsible with sex. Other narratives provided more in-depth descriptions that led to the formation of charmed loops.For example, a 20-year-old homosexual white male from Connecticut disclosed how his fear of getting AIDS has led to rules of meaning and rules of action to prevent obtaining AIDS: Ever since I came out it has been ingrained in my head to be careful about getting AIDS.Because of this, I didn't have my first sexual encounter until I was 17 years old.(Two years after I came out).I also didn't have anal sex until I was 19, because I was so fearful of contracting AIDS.To this day I still remain extremely fearful of AIDS and do not take anal sex lightly. For this participant, constitutive rules were interpreted within the speech act of coming out.In addition, being told to take caution regarding AIDS become part of his life script as a young gay man.The messages within these contextual frames led to several regulative rules: wait to have his first sexual encounter, remain extremely fearful of AIDS, and not take anal sex lightly.These rules of action can be best understood in the participant's life script or identity level of understanding.A charmed loop exists between rules of meaning, as a young gay man and rules of action in that he waited to have sex and remains quite cautious when engaging in sex as a way to prevent obtaining AIDS. Strange loops.In contrast to charmed loops that show consistency between rules of meaning and rules of action, strange loops demonstrate the contradiction and inconsistencies that can exist within meaning-making processes.Although still demonstrating a loop in that they emphasize the fluidity of contexts and depict the analytical nature of meaning and action, these differ from charmed loops in that rules of meaning and rules of action do not align.For example, as one Baptist 21-year-old racially mixed black male described an experience involving a "promiscuous" woman and how it "turned him off": I had an experience with a girl I was introduced to about 4 years ago.She was very attractive but I was told that she was promiscuous so I went along with it.We were watching a movie at her house and I noticed how often she was itching her private areas.I totally was turned-off but I still got some oral from her.Ever since that experience I became really cautious about who I have sex with, I have to at least have known you for a while before jumping into bed with someone. Within this experience, the narrative provides rules of meaning and rules of action that seemingly contradict.The constitutive rules established involved being a promiscuous woman itching her private areas; something that was a sexual turnoff.Given this meaning-making, a charmed loop would suppose that regulative rules would facilitate avoidance in terms of any sexual contact.Yet, this male participant explains that he still "got some oral [sex] from her".The rules of action cultivated from the experience involved sexual caution and knowing someone for a while before jumping into bed with them.The inconsistency between constitutive and regulative rules reflects a strange loop; however, it also leaves open the possibility that the participant's regulative rules involve a perception that oral sex is safer than sexual intercourse. Whereas the previous narrative centered on caution and knowing someone before engaging in sexual intercourse, others were grounded with a need to meet some sort of masculine expectation.For example, a 22-yearold African American male from Texas described a conversation between him and his father about sex which led to him having sex as a way to avoid feeling like "an outcast": As with the previous narrative, this narrative depicts a strange loop between the rules of meaning at the contract level and the rules of action at the speech act level, in that they seemingly contradict.Constitutive rules about being sexual active were understood at the contract level between the man and his father.What it meant to be sexually active or not was negotiated between the man and his father.It appears that what the participant took away from the conversation with his father is that engaging in sex reflects a norm in masculine culture.After the discussion, regulative rules about having sex to not feel like an outcast were then understood.Although he did not have any desires to become sexually active at 17 years old, this participant had sex at age 17 as a way to be accepted into some masculine norms regarding sex and sexual activity. Discussion Our study was designed to explore how male college students socially constructed memorable sexual experiences generally, and how the meanings associated with these experiences influence current sexual practices.Our data was comprised of self-generated narratives, those in which male participants described memorable sexual experiences.Through a coordinated management of meaning theoretical framework, we explored how meaning was constructed in both unidimensional (i.e., via speech acts and episodes) and multidimensional (i.e., reflexive relationships, charmed/strange loops) contextual frames.Utilizing CMM's loop model, we found how men's memorable sexual experiences were implicitly and explicitly informed through larger frames of cultural mascu-linity (i.e., through archetypes and life scripts).In addition our thematic analysis of participant narratives reveals logical forces of both charmed and strange loops.Within this context, three themes emerged: Protection against STIs and unwanted pregnancy, responsibility, and relationship status.While we acknowledge the polyvocality of participant narratives (Lyotard, 1984), our findings provide valuable insight into the complexity of how men socially construct meaning around sexual experiences. A significant outcome of our study lies within the ways it demonstrates the ways that masculine archetypes influence and inform the social construction of male sexual experiences.Through several CMM loop analyses, narratives illustrated how their decisions about sex, as well as their role during sexual encounters, reflect negotiations of larger values regarding masculinity.We focused on one memorable experience narrative, however, future studies might find greater insight through analyses of multiple narratives describing memorable experiences.As such, future research can add considerable depth to the existing literature by examining how current behaviors are informed by clusters of memorable narratives that are oftentimes complementary and oppositional.In addition to CMM, scholars interested in this line of research might draw from the concept of sensemaking as articulated by Weick (1995).Grounded in both individual and social activities, sensemaking "involves turning circumstances into a situation that is comprehended explicitly in words and that serves as a springboard into action" (Weick, Sutcliffe, & Obstfeld, 2005: p. 409).Sensemaking occurs within a larger societal frame, and incorporating gender role theories-like those established by Eagly (1987) and Bem (1981)-would prove invaluable to understanding how individuals negotiate societal expectations and personal preferences (McCabe et al., 2010).Specific research projects that explore how individuals make sense of multiple, sometimes competing, memorable experiences about sex appear as potentially insightful and heuristically rich.Future research is needed to further explore the salient factors, in addition to sex and gender, that inform memorable experiences regarding sexual intimacy5 . Our study also lays the foundation for future research that reflects both theoretical, methodological, and practical value.A major contribution of our exploratory study of male sexual experiences to existing literature is the productive ways in which we investigated the narratives of memorable experiences given the powerful ways in which they serve as a template for understanding human life (Browning, 2009).We have used CMM to focus on how men make meaning of their own sexual experiences, in the context of complimentary and competing contexts of understanding (Bruss et al., 2005).However, CMM is quick to point out that "communicators must manage their own meanings and actions, while responding to the meanings and actions of others" (Littlejohn, 2009: p. 200).Our study focuses on men's narratives of memorable sexual experiences, and provides valuable insight into the perspectives that are part of their social worlds (Pearce & Pearce, 2001).While important in this regard, the research is limiting in that it does not incorporate relational contexts through which meaning is given to sexual experiences.Future research might consider collecting narratives from relational partners to explore how the negotiation of meaning exists within social constructions that include self, other, and relationship.This approach is consistent with a social construction approach given that people jointly construct their understandings of the world (Berger & Luckmann, 1967). Using a CMM loop model (Pearce, 2005) provided an opportunity to focus in on the functions that narratives play in the creation of meaning and action.In this conceptualization, CMM acknowledges how the stories that are told construct a world of meanings and actions; yet it also recognizes how narratives can never tell the entire story at any given time (Littlejohn, 2009).Utilizing CMM as a practical theory (Pearce & Pearce, 2000), then, can involve "an attempt to change the mode of storytelling to one that has more opportunities for good things to happen" (Pearce, 2005: p. 48).We believe, as Littlejohn (2009) does, that "patterns of communication can be changed and expanded and coordination achieved by telling and different stories" (p.203).In the context of men's memorable sexual narratives, we wonder what might this look like-especially if practitioners could create contexts where men make meaning of their sexual experiences with others who are negotiating masculine values in ways that resist traditional conceptualizations (Anderson, 2010;Burleson, Holmstrom, & Gilstrap, 2005).What stories might they tell in an emotionally supportive environment?How might the nature and meanings of these stories shift, and in turn alter their social constructions of sex, gender, and sexuality? In regards to methodology, our study highlights the productive ways in which on-line surveys can engage a large, diverse participant pool and collect anonymous, confidential data on a sensitive (if not taboo) topic such as sex.While much of the data was frank, honest, and uncompromising, collecting via an open-ended survey question also was limiting in that we did not have any opportunities to directly engage with participants, something that hindered our ability to ask follow-up clarifying questions6 .Future research on sexual memorable experiences, like that which we've outlined in this discussion section, would be wise to facilitate data collection strategies that allow for extended exchanges that can provide greater depth and contextualization.Conducting such research on-line (e.g., via personal or group chats) exists as a distinct possibility (see, for example, Grabner-Kanter & Kaluscha, 2003).Engaging participants through communicative channels that foster trust and confidentiality is crucial to maximizing the depth and richness of data. Conclusion We conclude with some brief comments regarding the practical implications of this line of research.CMM is "a valuable resource for understanding, describing, and facilitating the development of new forms of communication called for by the challenges of contemporary society" (Pearce, 2005: p. 36).As a theory with practical applications, CMM analyses demonstrate how narratives can represent a crucial tool to assist in understanding how individuals make sense of their sexual experiences.This understanding-of both self and others' experiencesis a salient factor in creating a healthy attitude toward human sexuality, something advocated by the World Health Organization: Sexual health is a state of physical, emotional, mental and social well-being related to sexuality; it is not merely the absence of disease, dysfunction or infirmity.Sexual health requires a positive and respectful approach to sexuality and sexual relationships, as well as the possibility of having pleasurable and safe sexual experiences, free of coercion, discrimination and violence (World Health Organizations, 2002, as cited in Carpenter, 2005: p. 194). Promoting a healthy attitude toward human sexuality must be a core objective within any contemporary sex education efforts.While some sex education curriculum has proven success in delaying teen sex, a recent study reported that outdated curriculum-including antiquated stereotypical thinking in terms of gender-are hindering their ultimate success (Wong, 2012).This shortcoming is reflective of younger generations negotiation, or blatant rejection, of traditional masculine and feminine roles. Within this research endeavor we sought out the narrations of college males to better understand how meanings of sex are socially constructed.Using coordinated management of meaning (CMM) as a theoretical lens we discovered multiple themes centered on rigid masculine values that participants referenced to construct meanings of sex and gender.Understanding how individuals in the 21 st century make meaning of their own sexuality must acknowledge shifting attitudes regarding gender identity and expression.Without this type of recognition, efforts to promote a healthy attitude toward sex will continue to fall short. Figure 1 . Figure 1.CMM Model [Catholic and public school sexual differences]. Figure 2 . Figure 2. CMM Model [Male ego and first time having sex]. Figure 3 . Figure 3. CMM Model [Giving partners intense orgasms is the ultimate satisfaction]. I remember having sex at the age of 17 not because [I] wanted [too] but because of a conversation [I] had with my dad about w[h]ether or not [I] was active or not[.][A]fter the disscussion [I] went and had sex because [I] didn't want to feel like an outcast but mostly [I] did it because [I] wanted to [I] was extremely nervous.	2018-12-13T06:59:36.615Z	2014-08-28T00:00:00.000	{ "year": 2014, "sha1": "2c54e27b036e6c38de97dca85387df90465a6e7d", "oa_license": "CCBY", "oa_url": "http://www.scirp.org/journal/PaperDownload.aspx?paperID=49293", "oa_status": "GOLD", "pdf_src": "MergedPDFExtraction", "pdf_hash": "2c54e27b036e6c38de97dca85387df90465a6e7d", "s2fieldsofstudy": [ "Sociology" ], "extfieldsofstudy": [ "Psychology" ] }
52121528	pes2o/s2orc	v3-fos-license	Iran supports a great share of biodiversity and floristic endemism for Fritillaria spp. (Liliaceae): A review Iran supports a great share of exotic and/or endemic plant genera and species. The genus Fritillaria (Liliaceae) is a precious part of this botanical richness with 19 species, of which 10 are endemic to the country. However, signs are mounting that the country is truly at a crossroads when it comes to preservation of this national wealth. In this regard, an effective conservation strategy should thoroughly consider the classification of Fritillaria, as conservation practices are compromised by knowledge gaps in systematics and taxonomy. As published studies on Fritillaria in Iran have been sporadic and limited in scope, the aim of this review is to provide information necessary to help bridge these information gaps. Our objective is to facilitate increased understanding of the geographic, taxonomic, cytogenetic and phylogenetic status of Iranian Fritillaria, which is vital to meeting the goal of sustainable conservation of the genus in Iran and neighboring areas. Iran is a land of extremes; altitudes range from 5604 m on Mount Damavand's summit (Kharazipour, 2009) to 28 m below sea-level on the shores of the Caspian Sea (Akhani and Ghorbanli, 1993). The cold climates of the Zagros range, running from the northwest to the south, are replaced by the hot desert climates to the center and east. Mean January temperatures range from 20 C along the Persian Gulf in Hormozgan province to about À11 C in Chaharmahal & Bakhtiari (C & B) province. Dry lands of the interior areas change suddenly to the wet and moderate coastal climates of the Caspian coastal areas; the lush Caspian forest may receive an annual rainfall of 1950 mm, whereas lifeless sand-dunes of the 'Lut' remain arid (~30 mm annual rainfall). Iran supports a great share of plant species and countless natural habitats characterized by many unique plants and centers of local endemism (Khourang et al., 2014). Up to 8200 plant species are recognized throughout the country, of which almost 1900 are endemic (Kharazipour, 2009). The genus Fritillaria, with a variety of species that grow naturally across different areas, is a major component of this botanical richness. However, signs are mounting that the country is truly at a crossroads when it comes to preservation of this national wealth. Taxonomy and species conservation are often assumed to be interdependent activities; a shortage of taxonomic information and skills causes problems for conservation practitioners (Mace, 2004). From this point of view, dealing with the genus Fritillaria is a tricky effort, as the accepted units of diversity in the genus are hard to define and the exact habitat range of the species is hard to estimate. The existence of intermediate forms and continuous ranges of variation within this group of plants, which are phenomena generally caused by frequent hybridization at the species and subspecies level, accounts for the extreme difficulty in identifying many doubtful species. Furthermore, many of the species are native to remote uninhabited areas, where access to the habitats is extremely difficult. To make this even more complicated, members of the genus in Iran exhibit a wide range of variation in their morphological features (Sharifi-Tehrani and and physiological responses to the environment (Khourang et al., 2014). Thus, it is not surprising that among the large number of Fritillaria species known to occur in Iran, the majority of species are little known. Despite the necessity of an accessible comprehensive source of data to set practical conservation strategies, studies on Fritillaria in Iran have been quite sporadic and limited in scope. Following a comparative approach, here, we present a review of detailed basic information necessary for understanding the geographic, taxonomic, cytogenetic and phylogenetic status of Iranian Fritillaria, which seem to be vital to meeting the goal of sustainable conservation of Fritillaria in Iran and neighboring areas. Genus Fritillaria (L.) The name Fritillaria is said to come from either the Latin term fritillus (the chequered Roman dice tower) (Kiani, 2015) or more likely the Latin root frittillo (the chess-board) (Gerard, 1995); both terms, however, denote the checkerboard pattern of the petals in Fritillaria meleagris L., the type species of the genus. Fritillaria is also the scientific name for butterflies in the family Nymphalidae, again referring to their patterned wings (Dunford and Sims, 2008). Fritillaria (L.) is one of the most complex genera recognized in the family Liliaceae (Li et al., 2009). According to Rix (2001), the genus embraces about 140 species (165 taxa) of geophytic perennials, indigenous to the temperate climatic regions of the Northern Hemisphere (Rønsted et al., 2005) between latitudes 32 and 62 , especially throughout Europe, the Northern part of Africa bordering the Mediterranean Sea, the Mediterranean basin, the Middle East, Northern Asia and as far south as Iran, Afghanistan, the Himalayas, China, Korea and Japan. In North America Fritillaria occurs in Canada and in a narrow coastal strip from the Aleutian Islands to northern Mexico, extending inland to 16 states of the United States, especially California and Oregon (Beetle, 1944;Tekş en and Aytaç, 2011;Tomovi c et al., 2007). The genus has undergone marked species radiations in the Middle East, Southwest Asia and the eastern Mediterranean regions (Zaharof, 1988), particularly in Turkey and Iran. While Turkey, with 35 species and 6 subspecies (Tekş en and Aytaç, 2011), supports the greatest share of the world's Fritillaria resources, the significance of Iran rather lies in having a greater representation of different subgenera than any other country. Thus, the main center of the genetic diversity of the genus probably is Iran, where subgenera from Caucasus, central Asia and the Mediterranean meet (Rix, 1977). To date, 19 species (including at least 1 variety and 3 subspecies) are reported in Iran, of which at least 10 are endemic to the country (Table 1) Bakhshi Khaniki, 1997a, 1997bRechinger, 1990;Rix, 1977Rix, , 2001. Interestingly, the pattern of distribution of Iranian fritillaries largely corresponds to those parts of the country belonging to the Irano-Anatolian biodiversity hotspot (Fig. 1, Right). Fritillares are quite hardy and occur over a wide range of climates and habitats (Beetle, 1944). Most of the Iranian species are well-adapted to the uplands, foothills and associated plains of the Zagros range (Badfar-Chaleshtori et al., 2012), characterized by very rocky and sloping areas, wetlands, riparian zones, and different types of rangelands. The remaining species typically occur either on the northern slopes of the Mountain Elburz (also spelled Alborz) along the Caspian Sea in North Iran (Khaniki and Persson, 1997) or in flat and arid areas of the central parts of the country (Rix and Zarrei, 2007b). Iranian fritillaries mainly inhabit montane to alpine zones (1000e3600 m), but some species e.g. Fritillaria raddeana, at least in some parts, inhabit lowlands too (Khourang et al., 2014). With the exception of Fritillaria imperialis and F. persica, which are more southerly species, from the northern borders of the Zagros chain downwards, the geographical distribution of most of the Iranian species gradually decreases . The Zagros highlands at the border of Kordestan and Kermanshah provinces support the highest level of species richness in Iran (Wallis and Wallis, 2009), while the number of species in the southerly C & B province decreases to only four . The typical bulb of Fritillaria species consists of a few fleshy, tightly packed scales and a thin, translucent tunic, which usually disappears as the bulb increases in size. Some Fritillaria species have naked bulbs consisting of many scales, which resemble those of Lilium; these species also have numerous bulbils, loosely attached (Rønsted et al., 2005). In the genus Fritillaria, flowers are usually pendant, solitary (sometimes in umbels or many-flowered racemes) and have a campanulate and typical tulip-like perianth (Zych and Stpiczy nska, 2012). Fritillaries generally bear actinomorphic flowers (Khaniki and Persson, 1997), except for members of subgenus Rhinopetalum, in which flowers are usually zygomorphic (Tamura, 1998). Nectaries are often flattish, usually linear, lanceolate or ovate, sometimes circular, and pollen grains are usually faveolate to reticulate (Khaniki, 2003). Species range from <10 to over 120 cm in height. Although fritillaries can be seen in nearly every color, white-yellow, orange-red and purplish-brown are most common among the Iranian species (Kiani, 2015). F. imperialis is well known for its unpleasant foxy odor (Helsper et al., 2006), emanated by all parts of the plant, which is thought to be a way for repelling herbivores. The majority of the species with green or brown, broadly campanulate, flowers have what is often called a "spermatic scent", associated only with the flowers and presumably related to the attraction of pollinators. Fritillaries generally produce bisexual flowers (Shimizu et al., 1998); however, a few other mating systems like androdioecy (in F. persica, F. involucrate All., F. messanensis Raf. and F. montana Hoppe ex W.D.J.Koch) and andromonoecy (in F. montana) have also been documented (Peruzzi, 2012). The breeding systems vary from self-compatible to self-incompatible, demonstrating various degrees of out-crossing. Fritillaria species are probably pollinated by insects, especially Hymenoptera (mostly various species of bees and wasps), Diptera, Lepidoptera and Coleoptera (Zych and Stpiczy nska, 2012), and even birds, as in the case of F. imperialis (Burquez, 1989;Peters et al., 1995). Up to now, 19 species belonging to four subgenera are recorded in Iran; subgenus I) Petilium consists of four morphologically similar species, F. imperialis, F. raddeana, F. chitralensis (auct.) B. Mathew (also treated as F. imperialis var. chitralensis auct.) and F. eduardii A.Regel ex Regel (Wietsma et al., 2015); the first two species have been found in Iran. F. persica is the only species recognized in the monotypic subgenus II) Theresia (Türktaş et al., 2012). Fritillaria gibbosa, F. ariana, F. karelinii, F. stenanthera Rgl. and F. bucharica Rgl. are five morphologically similar species recognized in the subgenus III) Rhinopetalum (Rix and Zarrei, 2007b), of which the occurrence of the first two has been confirmed in Iran; historically, a separate genus closely related to Fritillaria, called Rhinopetalum Fisch. ex Alexand. was described in 1830 (Mathew, 2005). Several authors used the genus Rhinopetalum subsequently, but others such as Boissier (1846) and Baker (1874) opted to place these species in a distinct subgenus in the genus Fritillaria. Recent molecular studies of Fritillaria (Rønsted et al., 2005;Day et al., 2014) have confirmed that members of Rhinopetalum are nested in Fritillaria; so the incorporation of Rhinopetalum into Fritillaria is justified. The largest subgenus of the genus, IV) Fritillaria, is morphologically classified into six complexes, namely Fritillaria crassifolia, Fritillaria kotschyana, Fritillaria caucasica, Fritillaria graeca, F. meleagris and Fritillaria cirrhosa groups (Rix, 2001), of which members of the first three complexes are reported in Iran. No member of the F. meleagris group has yet been recorded in Iran, but a representative of the group (Fritillaria latifolia) is found in neighboring areas of Turkey and the Caucasus, and is very likely to be found in northwest Iran (Rix, 1977). Following the classification proposed by Rix (2001), data in Fig. 2 confirm the monophyly of all subgenera within the genus, except subgenus Fritillaria, which resolves into two phylogenetically distant clades. On the basis of molecular evidence, members of the group Fritillaria A (a large predominantly European, Middle Eastern and North African clade) are distantly related to the group Fritillaria B (a small clade of mainly Chinese and Central Asian species). This result supports recognition of the clade Fritillaria B as a new subgenus, yet to be formally named. In Iran, phylogenetic research of Fritillaria is limited to few studies (e.g. Khourang et al., 2014;. Although the findings of Khourang et al. (2014) reconfirmed most results of the previous phylogenetic studies of Fritillaria (e.g. Rønsted et al., 2005;Day et al., 2014), they suggested a new phylogenetic arrangement within the genus in some cases. In comparison to Day et al. (2014), the most comprehensive phylogenetic study of the genus, Khourang et al. (2014) used nine species (F. imperialis, F. imperialis var. lutea [the yellow form of the species], F. persica, F. crassifolia, F. straussii, F. zagrica, F. kotschyana, F. gibbosa, F. reuteri and F. raddeana), on which no information has been released on molecular phylogeny of F. imperialis var. lutea and F. straussii in the literature. Therefore, based on molecular evidence, the status of these taxa within the genus Fritillaria has remained undefined. Their results confirmed that F. straussii falls into the Middle Eastern group of subgenus Fritillaria (Fritillaria A). Furthermore, they claimed that red-and yellow-colored crown imperials are closely related, and both belong to subgenus Petilium. Karyology and cytogenetic characteristics The genus Fritillaria has been used as a cytogenetic model for the study of chromosomal structure and cell division (meiosis & mitosis) processes, due to its exceptionally huge genome size (Darlington and La Cour, 1941); the tetraploid Fritillaria assyriaca is known to be the largest plant genome so far reported with 254.8 pg, nearly 800 times the size of the Arabidopsis thaliana genome (Zonneveld, 2010). Therefore, species of the genus have long held a special place in plant biology (Ambro zov a et al., 2010). Chromosome numbers have been reported for over 50 species of Fritillaria; with a few exceptions (Kiran and Gedik, 2014;Matsuura, 1934;Marchant and Macfarlane, 1980;Noda, 1975;La Cour, 1978;Li and Shang, 1989), most of the species within the genus have a basic chromosome number x ¼ 12 (Table 2). In addition, Fritillaria is a genus with frequent variations in the chromosome structure, ascribed to polyploidy as well as fusion and fragmentation of some chromosomes. Although tetraploidy in fritillaries has been documented in few cases (Table 2), polyploidy within the genus is almost limited to triploid plants, which are frequently found (Ambro zov a et al., 2010). Karyo-ecotypes are reported among Fritillaria species, as some taxa growing in different geographical-environmental conditions show various levels of ploidy, which directly affect their reproductive behavior, ecological tolerance, and distribution range (Marchant and Macfarlane, 1980). Cytogenetic and karyotypic characteristics are available for all of the Iranian species, except Fritillaria avromanica which has just recently been reported (Khaniki, 2002e, 2002f, 2002g;Advay et al., 2015). All of the Iranian species are diploid (2n ¼ 2x ¼ 24), showing a variety of karyotype formulas; in the majority of the species, the presence of 1e4 pairs of satellites has been confirmed Ahmadi-Roshan et al., 2016;Bakhshi, 2000;Bakhshi and Persson, 2002;Bakhshi Khaniki, 1997a, 1997bJafari et al., 2014;Khaniki, 2002e, 2002f, 2002g), mostly located on the long arms (Ahmadi-Roshan et al., 2016;Jafari et al., 2014); this may be evaluated as an indicator for chromosomal evolution or possible ways for structural differentiation. Khaniki (2002b) reported the presence of accessory B chromosomes (also known as supernumerary chromosomes) for F. zagrica, whereas Ahmadi-Roshan et al. (2016) and Jafari et al. (2014) found no accessory chromosomes in their assessments. From an evolutionary point of view, this would be considered quite normal, since it seems unlikely that accessories would persist in a species unless there was some significant adaptive advantage. Different chromosome types have been reported in Fritillaria. In accordance with Peruzzi et al. (2009), Jafari et al. (2014 and Ahmadi-Roshan et al. (2016), four chromosome types have been confirmed to include 'm', 'sm', 'st' and 'T' in nine species (F. imperialis, F. persica, F. raddeana, F. crassifolia ssp. kurdica, F. zagrica, F. kotschyana ssp. kotschyana, F. gibbosa, F. reuteri and F. straussii), with the 'st' and 'sm' types appearing at the highest and the lowest frequencies. The 'T' chromosome type has been reported for the first time in F. imperialis, F. imperialis var. lutea, F. raddeana, F. persica, F. crassifolia, F. zagrica and F. gibbosa (Jafari et al., 2014;Ahmadi-Roshan et al., 2016); the 3x 4x presence of the 'T' type chromosome in the karyotype of species directly correlates with the chromosomal asymmetry inferring the level of the species evolution. According to Stebbins's classification (1971), karyotypes of different species of Iranian Fritillaria can be classified as relatively asymmetric and/or evolved chromosomal structures (Jafari et al., 2014). Fritillaries are quite different in the geometry of the chromosome structure, which directly affects their cross-compatibility and the fertility of the offspring. Compatible taxa generally show high degrees of chromosomal homology, making them potentially suitable for fertile interbreeding. Accordingly, a set of crosscompatible species, including F. imperialis Â F. imperialis var. lutea, F. persica Â F. straussii, F. kotschyana Â F. crassifolia, F. imperialis Â F. crassifolia, F. imperialis Â F. gibbosa, F. gibbosa Â F. raddeana, F. reuteri Â F. zagrica, have been proposed by Ahmadi-Roshan et al. (2016) and Jafari et al. (2014), each of which are expected to make normal chromosome pairing and probably produce the most fertile hybrids when crossed to each other. The specific objectives of Fritillaria cytogenetics may be essentially diverse, but such knowledge is extremely useful, particularly in commercial improvement of cultivars through hybridization and/or identification of doubtful specimens via guestimating potential parents when several taxa inhabit in a given area. Morphology and geographic distribution In the following sections, detailed morphological characters and distribution ranges of the genus Fritillaria in Iran (including 19 species and 3 subspecies) are discussed. Keys to these taxa are also provided. Subgenus Petilium comprises a small group of larger and sturdier species, distributed in Turkey, Iraq, Turkmenistan, Iran, Pakistan, Afghanistan, and the western Himalayas (Rønsted et al., 2005). They differ from the other subgenera of Fritillaria in the top half of the robust stem being leafless with a clear tuft of bract leaves above an umbel of 4e8 hanging flowers. The bulbs are much larger than those of most Fritillaria species are, and consist of a few, large, erect, imbricate, fleshy scales (Rønsted et al., 2005). The bulbs typically have a big hole in the top center where last year's stem grew, causing a considerable decrease in the density of the bulb. This subgenus is also characterized by having white to yellowish, broadly triangularovate to circular nectaries, which are positioned very close to the base of the perianth segments. The style is 3-fid and the capsules are winged (Rønsted et al., 2005). Two well-known Iranian species, F. imperialis and F. raddeana, are discussed here. Crown imperial (Fritillaria imperialis L.) is known to be among the most widely grown species of the genus (Fig. 3). The predominant habitats of crown imperials in Iran are high elevations of the alpine Zagros region (Badfar-Chaleshtori et al., 2012) and neighboring areas (Table 3). Distribution: Iraq, Turkey, across the plateau of Iran to Afghanistan, Pakistan and the western Himalayan foothills. Flowers in the crown imperial (in Persian Taaj-e Khosro) are at the top of the stem, topped by a crown of small bract leaves, hence the name. In Iran, crown imperial is also commonly known under the local name of 'Ashk-e Maryam' meaning 'Tears of Mary' (Badfar-Chaleshtori et al., 2012), which refers to the drops of nectar (200e350 ml) that regularly appear at the petal base through the flowering stage, replaceable in less than a day upon extraction. It is also called 'Gel-e Begeriv' (Weepy flower) amongst local 'Lor' people of C & B and Kohgiluyeh & Boyerahmad (K & B) provinces, two provinces with a remarkable share of the distribution. Mountainous oak woodlands; single and/or small colonies can be seen at the borders of roads, plains, hills and valleys, edge of fields, rocky slopes, stony places (1500e2000 m). Although various colors ranging from nearly a true scarlet through orangeeyellow are found in crown imperials, they are regularly classified as red-and yellow-colored forms (Alp and Koyuncu, 2009;Kiani, 2015). In comparison to the red-colored form, the yellow-colored (known as F. imperialis var. lutea) individuals are rare in Iran and are considered a critically endangered variety. These are two different forms of the species that naturally co-exist in the same habitats (Khourang et al., 2014). It is worth noting that the yellow form is not a hybrid (Alp and Koyuncu, 2009). Despite the color difference, they show a high level of morphological similarity. The color of the main upright stem is variable based on the flower color; the stem of the red-colored form normally ranges black to dark brownish/purplish, while bright green is the commonest stem color for the yellow-colored individuals (Kiani et al., 2015b). Crown imperial is mostly an out-crossing species, morphologically best suited to be pollinated by birds; although it may be visited by a variety of insects (Burquez, 1989). Crown imperials usually release a distinctly foxy odor unpleasant to some people, caused by a single sulphurous terpene component identified as 3methyl-2-butene-1-thiol (Helsper et al., 2006). The odoriferous nature of the bulbs, leaves, and flowers may have evolved as a way to repel herbivores. Like other members of Liliaceae, F. imperialis is susceptible to depredation by the lily beetle Lilioceris Reitter spp. (Salahi Ardakani, 2014). Crown imperial is among a few species in the genus with the ability to form extraordinarily huge colonies. At least three huge crown imperial colonies are recognized in Iran; 1) Aligoodarz (Dalani) crown imperial plain (~2900 ha), 2) Khansar (Golestan-Kuh) crown imperial plain (~15 ha), and 3) Koohrang (Chelgerd) crown imperial plain (Fig. 12, left), among which the latter, occupying an area of over 3600 ha, is of high national/international importance to tourism. Fritillaria raddeana Rgl. Gorgan lily (Fritillaria raddeana Rgl.) (Fig. 4, Right) is an endemic species native to the semi-humid climate condition and lower elevations of north and northeast Iran. The species is found scattered primarily in Golestan province (Table 3). Distribution: Iran, Turkmenistan. F. raddeana closely resembles the yellow-colored form of crown imperials. F. raddeana is easily distinguished from them by the larger number of flowers in the umbel, the more narrowly campanulate flowers, and much smaller and more angular nectaries (Rix, 1977). As with crown imperial, F. raddeana is a species with the ability to form huge populations in its normal range. Such colonies are primarily centered in the Khombi and Sara Mountains in Garmeh region (Northern Khorassan province), where at least three plains covered by F. raddeana occupy a total area of over 18 ha (Kiani, 2015). Subgenus Theresia (K. Koch) Baker Fritillaria persica is the sole member of the monotypic subgenus Theresia. The species has a bulb consisting of only one massive fleshy scale, second in size only to F. imperialis, and numerous flowers on a tall stem. The nectary color largely depends on that of the flower, so that in some populations the nectaries are not very noticeable. Given its striking appearance and geometry, the species can be easily distinguished from other species. Fritillaria persica L. The distribution of Persian bells (Fritillaria persica L.) (Fig. 4, Left) across Iran is restricted to northwest through southwest areas (Table 3), relatively similar to those of F. imperialis, but at slightly lower elevations (1500e2000 m). In general, F. persica is a species of wider distribution, but of much lower number of plants in comparison to crown imperials (Kiani, 2015). F. persica is very variable, especially in flower color, leaf width and in presence or absence of bract leaves in the inflorescence. A variety of flower colors from white to deep purple is known; indeed, pale flowers are in great abundance in Iran (Kiani, 2015). Three forms, F. eggeri, F. libanotica and F. arabica, have been described as separate species, but there is much overlap between them and none seem to merit recognition at even subspecific level (Rix, 1977). Androdioecy is a rare mating system in plants and F. persica is one of a few plant taxa in which androdioecious reproduction is reported (Mancuso and Peruzzi, 2010). Stem diameter and number of flowers are believed to be higher in the hermaphrodite plants compared with the males (Mancuso and Peruzzi, 2010). Each flowering stem bears numerous basal leaves as well as a terminal pyramidal raceme composed by many small nodding flowers of the same sexual morphology; either male or hermaphrodite. Apparently, flowers open gradually from the bottom of the raceme upwards. Likewise, those species in the Central Asian clade (Fig. 2), have the flowers arranged in a raceme, and open their flowers from the bottom upwards (e.g. F. verticillata, F. olgae). Subgenus Fritillaria (L.) Subgenus Fritillaria is the largest subgenus of the genus Fritillaria, comprising more than half of the species, including the type species, F. meleagris. The species in this subgenus are widely distributed from Western Europe and the Mediterranean region to eastern Asia. These are characterized by having the typical Fritillaria bulb; the bulb usually consists of 2 (sometimes 3e4 if one or both of the scales from the previous year have persisted into a second season) subglobose and fleshy scales with/or without a tunica, formed by the scarious remains of the scales of the previous year or years. The stem bears usually 1, sometimes 2e3 flowers, rarely up to 8 flowers as in F. olgae Vved. and F. camschatcensis or up to 12 flowers as in F. pluriflora Torr. ex Benth. (Khaniki and Persson, 1997). Subgenus Fritillaria is normally subdivided into two sections based on the style; species with a clearly 3-fid style are included in section Fritillaria, whereas species with an undivided style or a style that is only shortly trilobulate at the apex are placed in section Olostyleae (Rønsted et al., 2005). Rix (2001) subdivided section Olostyleae into 6 series and section Fritillaria into 10. F. crassifolia group This group consists of a complex of species and subspecies close to Fritillaria crassifolia Boiss. & Reut., with a distribution centered in eastern Anatolia and northern Zagros. The main species in this group is F. crassifolia comprising three dwarf subspecies (Rix, 2000b), of which only ssp. kurdica has been recorded in Iran (Bakhshi Khaniki, 2001a;Wallis and Wallis, 2009). The most widespread member of the complex is F. crassifolia ssp. crassifolia, which occurs in Turkey. Subspecies kurdica is frequent in alpine steppe throughout northwest Iran (Azarbaijan), east Turkey and northeastern Iraq; ssp. hakkarensis is scattered in Turkey and northeastern Iraq. In addition to this typical species, a) F. reuteri, b) F. straussii (Kiani et al., 2015a), and c) F. poluninii (Wallis and Wallis, 2009) are three morphologically distinct species of this group which occur in Iran. They all have usually broadly campanulate flowers pendant at maturity, and rather large linear nectaries, 5e10 Â 1.5e2 mm (half or more as long as the perianth segments, placed 3e5 mm above the base), which are usually black-purplish at the base. All members of this complex, except F. poluninii and F. reuteri, have groups of warts or similar processes scattered on the surface of the tepal (Khaniki and Persson, 1997 (Fig. 5, Left) is a common plant from Zagros Mountains mainly reported along a relatively narrow border strip, bordering Turkey and Iraq in West and Northwest Iran (Rix, 1977). Bulb: up to 3 cm in diameter, sometimes with bulbils. Stem: up to 10 cm high. Leaves: 4 (5e7), glaucous, lanceolate or linear lanceolate, all alternate; the lowest is wider than the rest, 3e5 cm long, 6e15 mm wide. Flowers: 1e2 (À4), campanulate, pendant at maturity, yellowish or green, stippled and tessellated inside and out, sometimes on a greenish or yellowish background, with a green stripe along the center of each perianth segment; perianth segments, obovate to oblanceolate, obtuse. Nectaries: linear, about 8e10 Â 1.5e2 mm, usually with a raised ridges on each side, more blackish at the base. Capsule: 4e4.5 cm long, obvoid, narrowly tapering at base, truncate at apex, not winged. Flowers April to July. Distribution: Iran, Turkey, Iraq. The species is rather diminutive, very variable especially in overall size and flower color, ranging from greenish, brownish to yellowish (or combinations thereof), often heavily tessellated, brown or purple inside and outside, often with green fascia. Subspecies kurdica is distinguishable from ssp. crassifolia in relatively very narrower basal leaves that are also higher in number; subspecies crassifolia usually has only 4 leaves, whereas ssp. kurdica bears 5e6 (Rix, 1977). It also differs in its more clearly defined fascia and the raised ridge on the inner side of the perianth segment, which forms the nectary. On the other hand, F. poluninii differs from F. crassifolia ssp. kurdica mainly in narrower perianth segments, smaller flowers and absence of the swollen nectary ridge. (Fig. 5, Right) is the rarest and smallest of four closely related species or subspecies, being recognized by its short stem (Rix, 2006). Bulb: large for the size of the top growth, up to 2 cm in diameter, without bulbils and stolons. Stem: 4e8 cm, smooth. Leaves: 6e8, usually 6, the lowest narrowly lanceolate 4e7 Â 8e10 mm, usually 5e7 times as long as wide, all alternate or subopposite, shiny green. Flowers: usually 1 or 2, rarely 3, nodding or horizontal, broadly campanulate or cup-shaped, pale purple to greyish, not or vaguely tessellated but lined with wavy green or brownish-purple veins, without fascia; tepals smooth, entirely lacking warty cells or other processes, usually spreading and not overlapping, acute, 11.5e14 mm long, the outer 4 mm wide narrowly ovate, the inner slightly wider, obovate. Nectaries: green, linear, 3.5e5 mm long, 1e1.5 mm wide, 3e5 mm above the tepal base. Filaments: ca. 6 mm, sparsely papillose. Style: 4e5 mm, the branches 2e3 mm, reflexed, slender, smooth. Capsule: ca. 35 mm long, cylindrical, tapering towards the base, not winged. Flowers April to May. Distribution: West Iran and adjacent part of northeastern Iraq in Sulaimaniyah. Fritillaria poluninii was previously treated as a subspecies of F. crassifolia, but later Khaniki and Persson (1997) raised this to the rank of species, because of several distinctive characteristics. It is differentiated from the other members of the F. crassifolia complex in the narrower leaves, smooth tepal surface without warty cells, much shorter/narrower nectaries, the absence of the swollen nectary ridge and the absence of a dark spot at the base of the nectaries. As opposed to the other subspecies of F. crassifolia that have more oblanceolate tepals, in F. poluninii tepals are usually acute. The flowers are not tessellated (sometimes vaguely tessellated), but lined with green or brownish-purple veins; the veins are distinctly wavy, and can be confused with tessellation. Given these striking features, elevating F. crassifolia ssp. poluninii to a specific level is fully justified. Distribution: Endemic to Iran. F. straussii is unique in that the flowers open green, and then mature to dark brown until before fading (Khaniki, 2002d). This is also unlike other species in its combination of a linear nectary and leaves opposite or in whorls (Rix, 1977). Distribution: Endemic to Iran. A pure yellow-flowered form has been recorded on several occasions, but otherwise the species is relatively invariable (Rix, 1977). Fritillaria kotschyana group In Fritillaria kotschyana group flowers are nodding and broadly campanulate (Khaniki, 2002b). The nectary is ovate or lanceolate, and always shorter than half the length of the limb of the perianth segment, located at the inflection of the bell (Khaniki and Persson, 1997;Rix, 1977). Two species of this group are restricted to small areas of high elevations; 1) F. olivieri Baker is found in the northern Zagros Range, while 2) F. whittallii Baker is scattered in southwest Anatolia; 3) F. kotschyana is very variable in color (ranging from pure green with little tessellation to brown) and flower size. Fritillaria kotschyana ssp. kotschyana has been found mainly along upper slopes of the Elburz and Talysh Mountains as well as at the southern end of the Talysh, as far south as Khalkhal, where it inhabits a variety of habitats including sparse meadowland, unstable screes, rock crevices and among roots of small shrubs (Wallis and Wallis, 2009). Khourang et al. (2014) also reported this subspecies in meadows and sloping areas near Mianeh (East Azarbaijan); F. kotschyana ssp. grandiflora is only collected from the northern Talysh; 4) F. hermonis is only found in S. Syria and Lebanon (Lebanon, Chouf and Antilebanon ranges), while 5) F. amana, which was previously treated as a subspecies of F. hermonis, is limited to Turkey and NW Syria (Amanus and Nusairiah Ranges). Distribution: Endemic to Iran. Distribution: Endemic to Iran. Distribution: Endemic to Iran. Fritillaria caucasica group Fritillaria caucasica is the most diverse group within subgenus Fritillaria in Iran. All Iranian species belonging to F. caucasica group i.e., F. caucasica, F. chlorantha, F. zagrica, F. uva-vulpis, F. assyriaca, F. chlororhabdota, Fritillaria atrolineata and F. avromanica, are distributed along the Zagros chain (Bakhshi Khaniki, 1997a, 1997bRechinger, 1990;Rix, 1977). Within this group, the species have comparatively small and narrowly Deserted areas, on mobile sand dunes and sandy hills and plains, with a mixture of small annuals, which usually flower and set seed very quickly before the arrival of the dry summer (500e1000 m). campanulate, narrowly lanceolate flowers, without tessellation. The small nectaries (not or only slightly depressed) are placed 0.5e1 mm above the base of tepals (Khaniki and Persson, 1997). Some species of this group e.g., F. atrolineata, F. armena, F. zagrica and F. chlorantha, have warts or warty cells scattered on the tepal surface. Distribution: Iran, Turkey, Iraq. F. uva-vulpis is morphologically distinct; it has usually four flat, lanceolate, shiny green leaves spaced evenly up the stem; the radical leaves are narrowly lanceolate. F. assyriaca is widespread and very variable, especially in width of the leaf; it has 4e6 glaucous, narrow and canaliculate leaves, mainly placed on the upper half of the stem; the radical leaves are broadly ovate. F. assyriaca also has narrowly tubular-campanulate flowers with a more greenish tinge and a slightly flaring mouth, while the flowers of F. uva-vulpis are usually more rounded. Distribution: Endemic to Iran. Tekş en and Aytaç (2011), in a revision of the genus in Turkey, lumped Fritillaria zagrica with F. pinardii, but they are quite distinct. F. zagrica normally has the little yellow tips to the segments, while F. pinardii sometimes has a yellow distal margin, but if present, this always is a continuous margin on the distal edge. On the other hand, the F. zagrica that Tekş en and Aytaç used as a comparator is not the type specimen, but one from Turkey that is a significant outlier from the normal range of F. zagrica. 6.3.3.4. Fritillaria atrolineata Bakhshi Khaniki. Fritillaria atrolineata Bakhshi Khaniki (Fig. 9, Right) has been reported only from one locality in West Azarbaijan province (Table 4) by Bakhshi Khaniki (1997a). Distribution: Endemic to Iran. In the type description, Bakhshi Khaniki (1997a) referred to the black nectaries of Fritillaria atrolineata, hence the name. Bakhshi Khaniki classified the species in subgenus Fritillaria (F. caucasica group) based on its comparatively small overall size, geometry, position and color of the nectaries as well as narrowly campanulate flowers. Fritillaria chlorantha seems to be related to F. atrolineata, primarily on account of the flower color. There are, however, striking differences separating them: F. atrolineata has glaucous, long and narrowly oblanceolate leaves with the lowest ones sometimes two or three in subopposite or subverticillate (subternate) position; a comparatively long stem; small, narrowly campanulate flowers; dark, elongate linear nectaries, and entire styles. The foleolate exine of pollen grains, and slightly trilobulate styles are also typical of the species. By contrast, F. chlorantha has short stems; short, wide, shiny green leaves, all alternate; comparatively big tubular flowers; green, lanceolate nectaries, and styles 3-fid at apex (Bakhshi Khaniki, 1997a;Khaniki, 2002a). (Fig. 10, Right) has lately been reported only from Hawraman (also spelled Avroman) area in west Iran (Table 4) . It forms small, fewnumbered populations of less than twenty individuals. Distribution: Endemic to Iran. Advay et al. (2015) argued that Fritillaria avromanica and F. assyriaca are closely related based on floral traits, but these species clearly show very little morphological resemblance since the leaves are a completely different shape. Robert B. Wallis believes F. avromanica is much more similar to F. chlorantha, particularly in its ovate lanceolate lower leaves, and the former may just be a variant of the latter species (Personal communication). However, they are distinct in flower color and to some degree in leaf texture/arrangement. F. chlorantha is relatively invariable, but some collections have perianth segments marked with purple, which may be a sign of hybridization with F. zagrica or F. assyriaca, both of which occur in the same area (Rix, 1977). On the other hand, the patterned coloration of the petals in F. avromanica do not seem to be a sign of variation or random hybridization in F. chlorantha. As in the case of taxonomic status of Rhinopetalum (See Section 4), molecular phylogenetic studies should help shed light on this challenging issue. Distribution: Endemic to Iran. Among the species belonging to the Caucasica group, Fritillaria chlororhabdota seems to be closest to F. caucasica, by phyllotaxy (position and number of the leaves) as well as shape and its external purplish color of the tepals. However, the species can be distinguished from F. caucasica by several characters, e.g., folded leaves at fruiting time, narrowly campanulate flowers, obtuse tepals with a median green stripe on the inside, wide lanceolate nectaries, shorter filaments, rough pollen exine with big luminae and an entire stout style. In contrast, F. caucasica has flat leaves during the whole of vegetative period, acute tepals (sometimes inner ones obtuse) that are paler purple inside (sometimes greenish) and without a green stripe along the whole of their length, linear-lanceolate nectaries, longer filaments, smooth pollen exine (knobs are absent) with very small luminae (foveolate sculpturing type), and longer tribulate styles. Subgenus Rhinopetalum (Fisch. ex Alexand.) Baker The five members of subgenus Rhinopetalum (F. gibbosa, F. karelinii (Fisch. ex D.Don) Baker, F. stenanthera Rgl., F. bucharica Rgl., F. ariana) all have rather flat, pink and distinctive starry flowers different in shape compared with other subgenera and species of the genus. They are inhabitants of the mountains and steppes of central and western Asia, from the Caspian sea and Iran, to Afghanistan and the dry hills of western Pakistan; two members of this subgenus, F. gibbosa and F. ariana, occur in Iran (Rix and Strange, 2017). The filament movement is only found in members of Rhinopetalum and Theresia (F. persica); the stamens bend back against the tepals before dehiscence of the anthers, and then bend forward at dehiscence. The nectary structure in members of this subgenus clearly differ from all other species; the nectaries are deeply impressed and the nectary orifice is slit-like which is bordered by two lobes, and appears densely hairy, at least in the lower part Zarrei, 2007a, 2007b). As against other subgenera, F. gibbosa and F. ariana are unique in that the flowers are zygomorphic, as the nectary lobes are rather fringed and broad, and the nectary on the uppermost tepal is more depressed than the others. In the other subgenera, the nectaries are less depressed (often flattish), and usually linear to lanceolate or ovate, except in subgenus Petilium having usually circular nectaries. Following a morphological approach, the unique structure of the nectaries and other distinctive floral characters in all species of this subgenus imply that subgenus Rhinopetalum does not seem to be more than distantly related to the genus Fritillaria, and may probably merit formal recognition as a distinct genus (Rhinopetalum Fisch. ex Alexand). We have discussed this uncertainty earlier through Section 4 based on molecular evidence. Fritillaria gibbosa Boiss Gibbous lily (Fritillaria gibbosa Boiss.) (Fig. 11, Center) has been claimed (Rix, 1977) to be geographically the most widespread species of the genus in Iran (Table 4). Normally, F. gibbosa is a species adapted to arid climates. Flowering begins very early, from March through May, on arid steppes, often among Artemisia L. spp., Rosa persica Michx. ex Juss., Lactuca orientalis (Boiss.) Boiss. and Poa bulbosa L. In east and southeast Iran where the summers are dry, it is very easy to grow this species. The species is highly variable, especially in the number and color of the flowers (Kiani, 2015). Flower color is usually pink, but ranges from white, orange-yellowish or purplish. The nectaries of F. gibbosa are very complex; the nectary on the topmost outer tepal is the largest (slightly however), and the inner one opposite it, is the smallest. All have two papillose ridges running down their middle, pressed together and concealing the nectar. On the outer part of the ridges, the papillae become flattened yellow lamellae, giving the appearance of an irregular crest. As with F. imperialis, the nectaries may exude drops of nectar in early mornings. The area around the nectaries is yellowish-green, heavily marked with blackish-purple. It differs from F. karelinii in that all the tepals, which are relatively broader, have nectaries, though some are larger than others; in F. karelinii only the uppermost nectary is enlarged and the rest seem to be vestigial, and do not produce nectar (Rix and Zarrei, 2007b (Fig. 11, Right) in Iran occurs nowhere in great abundance, but rather in very small populations mainly in northeastern Iran (Table 4) (Rix, 1977). It requires generally the same climatic conditions as were described for F. gibbosa. The geographical distribution of these two species is quite different; F. gibbosa is widespread in center and, northeast through northwest, all the way down along Zagros chain (Rix, 1977), but F. ariana is very rare in Iran and is categorized as a critically endangered species. On the other hand, F. ariana grows primarily in sandy deserts, while F. gibbosa prefers clay steppes or rocky slopes (Rix and Strange, 2017). Bulb: up to 3 cm diameter, without bulbils or stolons. Stem: 10e30 cm (À50 cm in fruit), glabrous or papillose only below the lowest leaves and at the nodes. Leaves: 8e12, long, the lowest pair considerably longer up to 120 Â 12 mm, linear-lanceolate, opposite, the rest linear, scattered; bract leaves 20e40 Â 2 mm, linear, 2 at the base of each pedicle. Flowers: 2e14, usually 6, flat, horizontal at maturity; perianth segments pink-reddish, unspotted or sometimes slightly mottled, marked with brown around the nectaries; the outer 25 Â 10 mm, obovate, acute; the inner somewhat wider. Nectaries: about 4 mm long, deeply impressed, the uppermost largest, the outer larger than the inner. Filaments: 6e7 mm, slender, papillose or glabrous only near the base; anthers about 5 mm before, 2 mm after dehiscence. Style: 5e7 mm, entire, slender, glabrous. Capsule: 15 mm, not winged, but toothed at the upper corners. Flowers March to early April. Fritillaria ariana is very similar to F. gibbosa, but they are separated by several characters; F. ariana is usually a much taller plant with narrower basal leaves, and more numerous flowers nearly always unspotted. The distinction between the broad basal leaves and narrower stem leaves of F. gibbosa, and the narrow basal and stem leaves of F. ariana is usually obvious. The stems of F. gibbosa are usually papillose all over, while those of F. ariana are glabrous or papillose only at the leaf bases (at the nodes) and below the lowest leaves (Rix, 1977;Rix and Zarrei, 2007b;Rix and Strange, 2017). Eight-Frit Mountain The Iraqi border is truly a paradise for those in search of fritillaries; within this region lies a range of mountains which has been christened "Eight-Frit Mountain" (Wallis and Wallis, 2009), as it is home to at least eight different species of the genus, each of which is growing in its own ecological niche. This location is south of Marivan in Kordestan province southwards on the road to Nowsud and on the border with Kermanshah Province. It is very close to the border with Iraq and includes a pass at about 2500 m (Fig. 12, Right). The place is a habitat of a strange-looking form of F. uva-vulpis growing together with Zagrosia persica (Hausskn.) Speta. and Bellevalia paradoxa (Fisch. & C.A.Mey.) Boiss. There are also populations of F. straussii on the north side of the slope; F. imperialis on the top of the ridge and a very large F. crassifolia ssp. kurdica on the south-facing side of the slope; F. avromanica grows in the spiny bushes on the west-facing slope below the road on the south side of the cleft; F poluninii grows in rock crevices in several parts of the pass; F. assyriaca can be found in several places on the slopes on both sides of the pass but a little lower down; F. persica populations are spread on the slopes while descending from the pass on the road towards Nowsud. Such a rich habitat, where members of the four subgenera known to occur in Iran coexist, is believed to be unique in the world (Wallis and Wallis, 2009). Systematics serves conservation In general, the large number of species (over 160 taxa), extensive polyploidy and highly polymorphic nature of several species of the genus Fritillaria indicate a genus near its peak of speciation (Beetle, 1944). As stated earlier, the occurrence of four different subgenera of the genus (including 19 species and 3 subspecies) has been confirmed in Iran. This makes the country an important area for diversification and evolution of new species, as species diversity itself could be a driver of further species diversification (Emerson and Kolm, 2005). Increasing species diversity may lead to greater genetic makeup complexity, which has been suggested as an evolutionary force driving speciation. This happens in a variety of ways, among which hybridization has played an especially important role in the evolution of the genus. Frequent hybridization can result in morphologically intermediate specimens and increased confusion in identification of Fritillaria species (Beck, 1947), which in turn increases the difficulty of making decisions related to conservation. Setting conservation priorities in taxonomically complex taxa is an essential but especially complicated process. Incomplete understanding of taxonomy can have disastrous impacts on conservation, since a given threatened species may not be readily distinguishable from a more frequent species. Although the current classification of the genus proposed by Rix (2001) is supported by molecular phylogenetic studies at the subgeneric level (Rønsted et al., 2005;Day et al., 2014), the whole genus still raises important evolutionary and systematic questions, and many taxa recognized within different species belonging to subgenera Fritillaria, Rhinopetalum and Petilium present taxonomic problems. In the following sections, we provide a brief overview of the possible uncertainties that may easily mislead conservation activities. Subgenus Fritillaria Taxonomic uncertainties are most frequent in taxa belonging to subgenus Fritillaria. One such problem has been already encountered in the case of F. uva-vulpis and F. assyriaca, which are known as two species of high similarity in general morphological characters. Although the name F. assyriaca has often been used for F. uva-vulpis, detailed examinations reveal their distinctiveness (Rix, 2000b). F. assyriaca is rather more widespread and is a common species in Iran, whereas F. uva-vulpis is considered as an endangered species. To make the case even more complicated, these two are among the most variable species in the genus and have quite diverse morphologies in Iran. In the most confusing case, Wallis and Wallis (2009) recorded a strange-looking F. uva-vulpis in the Eight-Frit Mountain area that is quite different from the description of the species. On the other hand, F. kotschyana subsp. kotschyana has been confused with F. crassifolia, all the subspecies of which are more dwarfed and have linear nectaries (Rix, 1977). In another instance, the name F. kotschyana has been used for dwarf F. latifolia from northern Turkey, a very different species that has never been recorded in Iran. At least two unknown taxa have been recorded in Kordestan, very close to the Iraq border, which seemed to be F. crassifolia, but did not fit any of the known subspecies (Wallis and Wallis, 2009). F. pinardii and F. zagrica are two taxa frequently mixed up in the literature (see Section 6.3.3.3), once again, for extreme resemblance in their morphological characters. For this reason, although, it has been recently a single report of F. pinardii in West Azarbaijan province (Sharifi-Tehrani and , the inclusion of this species within Iranian flora awaits further studies. In short, there is a variety of similar cases mainly due to crossing compatibility of different taxa and/or morphological similarities (e.g. Beck, 1947;Rix, 1974Rix, , 1975Bakhshi Khaniki, 1997b;Wallis and Wallis, 2009; Tekş en and Aytaç, 2011). Subgenus Rhinopetalum As with subgenus Fritillaria, identification errors are frequent and easily occur in subgenus Rhinopetalum mainly due to morphological similarities. For example, the red-listed F. ariana can be easily confused with F. gibbosa, which is widely distributed across Iran (Rix and Zarrei, 2007a). Moreover, F. pterocarpa Stocks, which is native to Baluchistan (Pakistan), is probably identical to F. gibbosa (Rix and Zarrei, 2007a). Although F. karelinii Fisch. ex D.Don has been claimed to occur in Iran on 31st May 1964 by Paul Furse (Rix and Zarrei, 2007b), we strongly believe this is an identification error; when Furse reported this record, the whole genus was poorly known. In addition, F. karelinii normally flowers very early in March; at this late date (31st May), it is very likely to have been in fruit and not in flower, making identification nearly impossible. Furse reported his record in Northern Khorassan (80 km west of Bojnurd, arid plains, yellow fine soil with stones and scattered Artemisia, scrub, 1800 m), where F. gibbosa is recorded 20 km from here in Artemisia steppe at 1300 m. Considering the normal range of F. karelinii, described from steppe in Southern Urals north of the Caspian Sea, as far south as Koppe Dagh, this is likely not F. karelinii. Subgenus Petilium In a similar case from subgenus Petilium, the abundance of F. imperialis var. imperialis, the most frequent species in Iran, is quite different from its very close relative, F. imperialis var. lutea, which is seriously on the brink of extinction (Khourang et al., 2014). Although the endemic F. raddeana is also reported in Kashmir (Ali, 2007), Pakistan is not within the normal range of the species. This record is doubtful and presumably is correctly assigned to either F. chitralensis, native to the Chitral District of northern Pakistan, or to F. imperialis var. lutea. Subgenus Theresia General morphological characters of F. persica, the sole member of subgenus Theresia, are clear-cut with no indication of introgressions or intermediates in Iran. This is the only species of the genus in Iran with numerous flowers arranged in a raceme, so that it can be easily distinguished from other species. Thus, in contrast to the above-mentioned cases, there is no identification concern about this species in Iran. Conclusion Implementation of conservation frameworks is negatively affected by a lack of basic information on taxonomy and biology of fritillaries in Iran, as taxonomy provides the foundation for conservation practices and sustainable management of the world's remaining resources. Studies carried out on the genus Fritillaria through the last four decades in Iran have been technically sound efforts, but quite sporadic and limited in scope. These studies give a fairly good overview of different aspects of Iranian Fritillaria, yet provide little information relevant to conservation, which remains relatively poorly understood. By describing geographical and ecological features of Iranian Fritillaria along with an intensive look into biodiversity of the genus with respect to its taxonomic status, a much clearer understanding of the essential elements of conservation strategies can be provided for all involved practitioners. Information described in this review may serve as a stepping-stone for setting conservation priorities, hopefully in the near future, in Iran and neighboring centers of biodiversity of the genus Fritillaria.	2018-09-15T22:01:10.942Z	2017-09-13T00:00:00.000	{ "year": 2017, "sha1": "f05e2e3412cae27bd634f183ab2c4313fd22c9ff", "oa_license": "CCBYNCND", "oa_url": "https://doi.org/10.1016/j.pld.2017.09.002", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "f05e2e3412cae27bd634f183ab2c4313fd22c9ff", "s2fieldsofstudy": [ "Environmental Science", "Biology" ], "extfieldsofstudy": [ "Medicine", "Biology" ] }
52989093	pes2o/s2orc	v3-fos-license	Effect of angiotensin converting enzyme gene I/D polymorphism in South Indian children with nephrotic syndrome Nephrotic syndrome is one of the most common childhood kidney diseases. It is mostly found in the age group of 2 to 8 years. Around 10%–15% of nephrotic syndrome cases are non-responders of steroid treatment (SRNS). Angiotensin converting enzyme (ACE) (I/D) gene association studies are important for detecting kidney disease and herein we assessed the association of ACE (I/D) polymorphism with nephrotic syndrome in South Indian children. We recruited 260 nephrotic syndrome (162 boys and 98 girls) and 218 (140 boys and 78 girls) control subjects. ACE I/D polymorphism was analyzed by PCR using genotype allele specific primers. In ACE (I/D), we did not find significant association for the ungrouped data of nephrotic syndrome children and the control subjects. Kidney biopsies were done in 86 nephrotic syndrome cases (minimal change disease, n=51; focal segmental glomerulosclerosis, n=27; diffuse mesangial proliferation, n=8). We segregated them into the minimal change disease / focal segmental glomerulosclerosis groups and observed that the ACE ‘D’ allele was identified with borderline significance in cases of focal segmental glomerulosclerosis and the ‘I’ allele was assessed as having very weak association in cases of minimal change disease. ‘II’ genotype was weakly associated with minimal change disease. Gender specific analysis revealed weak association of ‘ID’ genotype with female nephrotic syndrome in females. Dominant expression of DD genotype was observed in males with nephrotic syndrome. Our finding indicated that ACE (I/D) has moderate association with focal segmental glomerulosclerosis. However, due to the limited number of biopsy proven focal segmental glomerulosclerosis subjects enrolled, further studies are required to confirm these results. Introduction Nephrotic syndrome (NS) is a condition where damages are prominent in filtering units of kidney. NS is more predominant in children and commonly exists in the ages of 2 and 8 years. It seems to affect boys more often than girls [1][2] . The annual incidence of NS in most countries is estimated to range from 2 to 7 new cases per 100,000 children. The annual incidence of NS in US, European, Africo-Caribbean and Asian children were 2, 2.6, 3.4 and 16.9 per 100,000 children respectively and the cumulative prevalence was about 16 per 100,000 [1][2][3][4] . A geographically based epidemiological study of NS suggested that Asian children have higher incidence; mainly, it was higher in lower socioeconomic groups [5][6] . Ethnic origin may play a major role of the histological modification and response to immunosuppressive treatment. Based on immunosuppressive treatment, NS is mainly divided into steroid sensitive, steroid dependent and steroid resistant. The histological evaluation of renal tissues in nephrotic cases has three main categories: minimal change disease (MCD), focal segmental glomerulosclerosis (FSGS) and diffuse mesangial proliferation (DMP). MCD is idiopathic, with no change in the number of podocytes and it mostly responds to steroid treatment (remission). FSGS is a severe form which decreases the number of podocytes and does not respond to steroid treatment. Finally, DMP is also a severe form of nephrotic which is rarely seen. Genetic factors have been suggested to be important in determining the progression of NS. Recently, well documented evidences indicate that the renin-angiotensin system (RAS) is involved/implicated in pathogenesis of kidney disease [7][8] . Angiotensin converting enzyme (ACE) is an important enzyme of RAS. The stability of plasma ACE levels, combined with marked inter individual differences and familial clustering of plasma ACE levels suggested its regulation to be under major gene control. Genes that are involved in the RAS system are functional candidates/close link with these phenotypes, it would be the prognostic factors of renal damage such as diabetic nephropathy, Immunoglobulin A (IgA) nephropathy [7] , systemic lupus erythematosus (SLE) [9] and/or lupus nephritis and finally lead to development of end stage renal disease. Some of the existing reports from ACE polymorphism identified the presence of 'D' allele is indicated for the risk of cardiovascular [10] and kidney damage [11] . ACE DD genotype is responsible for hypertension [8] , diabetes mellitus [12] , IgA nephropathy [7] , renal artery stenosis [7] , cardiomyopathies [13] and carotid atherosclerosis [14] . However, some previous studies evaluating/identify and proven relationship between steroid sensitive/FSGS and ACE gene polymorphism of NS in different ethnic populations [15][16] . ACE-II genotype was more frequent in steroid sensitive nephrotic syndrome (SSNS) patients as compared to controls in North India [17] , Malaysia [18] . DD genotype with NS has been reported from Taiwan Province (China) [19] , Egypt [20] . However, the DD genotype was associated only with FSGS-SRNS in the Kuwaiti Arab children [21] . In contrast to the previous studies, no such association of ACE gene polymorphism was found in Swiss children [22] . Hence, we analysed the distribution of ACE (I/D) gene polymorphism of NS [SSNS/steroid resistant nephrotic syndrome (SRNS)] in South Indian children between the age of 2 and 12 years old. Study subjects In the present study, we recruited 260 (mean age of 7.17AE3.58) NS cases from a single center; 218 control subjects (mean age of 10.5AE1.07) were recruited. Our institutional ethical committee approved the study and written informed consent was obtained from subject's parents. All enrolled children were born in South India, were of ethnically South Indian ancestry and belonged to the lower and middle classes. Nephrotic cases were sporadic and it was noted that 85 enrolled (nephrotic 58; control 27) subjects were born from second degree consanguineous parents. The enrolled subjects were between the ages of 2 and 12 years old. The inclusion criterion was the clinical presentation of NS. The diagnosis of NS was based on the presence of edema, urinary protein excretion ≥ 40 mg/(m 2 $hour), spot albumin to creatinine ratio > 2 mg, and hypoalbuminemia < 2.5 g/dL. All the nephrotic cases received the standard steroid therapy and were classified into two categories on the basis of their clinical responses towards steroids: SSNS group and SRNS group. SSNS is defined as (remission) stratified into proteinuria negative to trace for three consecutive days or urine protein excretion < 4 mg/(m 2 $hour). SRNS is defined as failure to achieve remission after 4 week of daily oral prednisolone at a dose of 2 mg/(kg$day). Children with NS and a history of positive HIV, HbsAg were excluded from the study. The control group consists of unrelated healthy individuals with no history of kidney disease/ hypertension. DNA isolation and ACE (I/D) genotyping Genomic DNA was isolated from 2 mL of blood using the standard salting-out method. The 16 th intron of the polymorphic ACE (I/D) gene was amplified by PCR. The primers (forward/reverse) 5'-TGGAGAC-CACTCCCATCCTTTCT-3' and 5'-GATGTGGCCAT-CACATTCGTCACGAT-3' were used to amplify the region of intron 16 which produced a 287-bp insertion/ deletion polymorphism. PCR reaction was performed in a final volume of 12 mL containing 7.64 mL of milli Q water, 3 mL of genomic DNA (200 ng/mL), 0.2 mL of 5U Taq polymerase (Genet Bio, Korea), 1.2 mL of 10 Â PCR buffer, 0.24 mL of 10 mmol/L dNTP (Cinna Gen, Iran), 0.36 mL each of forward and reverse primer (10 mmol/L). PCR amplification was carried with an initial denaturation at 94°C for 5 minutes, followed by 30 cycles of denaturation at 94°C for 2 minutes, annealing at 58°C for 1 minute and extension at 72°C for 1 minute and a final extension at 72°C for 5 (Agilent, USA; model no: G8800A). Amplified products were detected on a 2% agarose gel containing 0.5 mg/mL of Ethidium Bromide. The amplicon containing the insertion allele (I) was visible as a band at approximately 490 bp; while deletion (D) at 190 bp; and 190-bp and 490-bp PCR products for ID heterozygotes. Due to preferential amplification of the D allele, it was possible that ID heterozygotes might be mistyped as DD homozygotes. Therefore, in order to increase the specificity of DD genotyping, all samples identified as DD after initial amplification were reconfirmed with the second PCR containing an insertion specific primers. Statistical analysis The statistical analysis was performed by STATA 11.1 (College Station, TX, USA). The continuous variables of age, serum creatinine, protein and albumin and cholesterol level were expressed as mean and standard deviation. ACE (I/D) genotype and allele frequencies were compared between the cases and controls. Odds ratio with 95% confidence interval expressed as genotype and allele frequencies. Genotype and allele frequencies distribution were expressed as frequency and percentage. Hardy-Weinberg equilibrium (HWE) for ACE genotype was tested by Chi square test in each group. Distribution of ACE (I/D) polymorphism dominance and recessive model were expressed. Statistically significant accepted value was P < 0.05. Results The clinical and demographic data of the nephrotic cases are presented in Table 1. There were no statistically significant differences in the distribution of ACE (I/D) gene polymorphism between the NS cases and controls ( Table 2). The data were stratified based on steroid treatment response in NS cases. The differences in the frequencies for genotypes or alleles for SSNS and SRNS were found to be statistically insignificant. We segregated NS biopsy cases into the MCD and FSGS groups. We found II genotype frequency was observed in 29.41% (15/51) of the MCD cases and in 11.11% (3/ 27) FSGS cases. This association was estimated as [OR (95%CI)= 3.33(0.80-19.61); P = 0.06] ( Table 3). The frequency of DD genotype was 25.9% (7/27) in FSGS, but 11.7% (6/51) in MCD. The 'I' allele and 'D' allele were found to be weakly associated (both had borderline significance) with MCD and FSGS, respectively. Additionally, in 8 DMP cases, II and ID genotype frequencies were 50% (4/8) and 50% (4/8), respectively, no cases had DD genotype. However, when we segregated the groups into male and female populations of nephrotic cases and controls, we could not found a significant association for II, ID and DD genotype in males. In the female groups, ID genotype was slightly more prevalent (59.1%) in the nephrotic group than the control (51.2%) which showed a weak association [OR (95%CI) = 1.74 (0.93-3.24); P = 0.06] ( Table 4). (Table 5). Further, we calculated the degree of dominance (h) test to find out the deviation of heterozygous state from the risk of disease. The analysis revealed that the degree of dominance was < 1 from the ACE (I/D) variant and NS in South Indian children. Discussion ACE is a glycoprotein and its stability in plasma, combined with marked inter-individual differences and the familial clustering of plasma ACE levels, suggests its regulation is under the control of a major gene. ACE I/D polymorphism was originally thought to account for varying degrees of phenotypic expression in circulation. The generation of Ang II depends on ACE. All nephrotic patients ultimately received the suggested treatment, with the synergic combination of ACE inhibitor and Ang II receptor blockers, as they decrease the proteinuria. Thus, it reduces the glomerular capillary pressure in addition to altering the glomerular permeability. Recently, Type 2 diabetes patients with nephropathy (proteinuria > 150 mg/24 hours) treated with standard ACEI therapy found progressive reduction of the proteinuria during 3-6 months study period [21] . Effects of Ang II increased systemic and glomerular blood pressure, leading to tubule-interstitial fibrosis and glomerulosclerosis; finally progressed to loss of kidney function. DD homozygous or D allele is associated with elevated circulating and tissue ACE activity compared to I allele. Several studies have found that D allele is an independent risk factor for hypertension, diabetic, IgA nephropathy, congenital renal malformations and NS [16,20,24] . In the present study, we observed that ACE (I/D) allelic distribution was not elevated in the control group when compared to the different subsets of nephrotic (SSNS/SRNS) cases, which is in agreement with previous studies from different ethnic populations such as Swedish and Egyptian [22,[25][26][27] . On the contrary, the ACE-II genotype was found more frequent in SSNS in North Indian and Pakistani children [17,28] . However, DD genotype was higher in north Indian population [15] which also coincided with the Egyptein [25] , Turkish [29] and Taiwanese [19] nephrotic children. Therefore, the ACE (I/D) polymorphism and its linkage with NS in different ethnic populations are contentious. Podocyte loss is the main reason behind the development of FSGS. According to the histopathological condition, NS is distinguishable notably as MCD and FSGS. The pathogenesis of MCD has not been clear so far; but there are evidences of immune dysfunction [30] . Most MCD cases were in remission, but a very few of them was advanced to CKD. In our center, some of the SRNS cases had a slow progression, while most have fast progressions to end-stage renal disease. It seems that genetic markers/polymorphism plays a susceptible/protective role in the disease mechanism [31] . Moreover, existing studies have identified the genetic markers and their linkage to the progression of kidney disease. Firstly, a study on PcKO mice, the Atg5 gene (functional block of autophagy) was associated with slow progression of podocyte degeneration and then to glomerulosclerosis [32] . Secondly, the ACE II-A9570G, ACE-ID/DD and/or AGT-M235T polymorphism were associated with an increased risk factor for arterial hypertension in FSGS with fast progression to chronic kidney disease (CKD) [16,[31][32][33][34] . In the present study, FSGS cases have 'D' allele association, which showed substantial persistence of the 'I' allele. However, the incidence of II genotype cases (3/27) was less dominant in FSGS. All three 'II' genotype FSGS cases had late onset of nephrotic below 10 years of age had slow progression to end-stage renal disease. In MCD cases, the 'I' allele showed strong association and the II genotype a weak association with very few of them having DD genotype (11.7%). In this manner MCD-DD genotype may play a role in the slow progression of MCD to FSGS. However, a second biopsy is required to confirm the conversion of MCD to FSGS. In one study from Kuwaiti, in Arab-Jewish nephrotic children the DD genotype was associated with FSGS cases [21] . In contrast, one meta-analysis study showed DD genotype/ 'D' allele is associated with MCD and II genotype played a protective role [35] . In our study, the frequency of ID genotype was slightly increased in SSNS 32.71% (35/107) compared to SRNS 26.79% (41/153) but it was not statistically significant. Moreover, a large number of biopsy proven MCD/FSGS cases may be recommended along with routine follow-up for better interpretation of the role of the ACE ID/DD genotype. In childhood, a preponderance of nephrotic in males is well-established. In an experimental rat model, an inactivating mutation in ACE resulted in lower ACE protein, compared with wild type rats. In female rats, a low level of ACE protein did not affect the blood pressure, but male rats were protected from hypertension [36] . In gender based study, ACE DD genotype was associated with hypertension in males [37] . Further, a large study (the Suita Study, Japan) which enrolled 14,200 individuals suggested a unique sex-specific effect of ACE on hypertension. In addition, a metaanalysis study involving Asian males with hypertension revealed significant association between ACE (I/D) polymorphism and CKD risk [37] . On the other hand SLE in females (analysis of 644 families) was proven with ACE association [9] . In various studies, established ACE (I/D) polymorphism and gender based disease association have been documented in different ethnic populations. Lin et al. stated that ACE (I/D) polymorphism and gender dependent effect have been observed very commonly among different populations. So, ACE locus is a sex-specific candidate gene/association for different diseases. In the present study, the subjects were segregated, gender-wise. The gender-wise comparison results revealed that female nephrotic children had weak association of 'ID' [OR (95%CI) = 1.74 (0.93-3.24); P = 0.06]. ACE ID genotype present in females with NS was 59.1% (58/98) and the control was 51.2% (40/78). In females, we observed ID genotype (subset of nephrotic) MCD, FSGS, DMP and pooled were 66.6% (12/18), 63.6% (7/11), 60% (3/5) and 64.7% (22/34) respectively. We conclude the ID genotype frequency was significantly increased in the NS and subset, compared to control females. Finally, we found the moderate significant association between the ACE 'D' allele and FSGS cases. The study results are positive and suggest that future studies should recruit a larger number of biopsy proven FSGS cases from different geographical regions for a better understanding of NS.	2018-11-01T20:39:02.761Z	2016-03-13T00:00:00.000	{ "year": 2016, "sha1": "de7ba96500df4f1c9b3a01eec237a8481ca05027", "oa_license": "CCBY", "oa_url": "https://doi.org/10.7555/jbr.32.20150095", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "8e1312797632a6db7f1663232a4d7a2f085daf2d", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
7716529	pes2o/s2orc	v3-fos-license	Copper-Mediated Radiofluorination of Arylstannanes with [18F]KF A copper-mediated nucleophilic radiofluorination of aryl- and vinylstannanes with [18F]KF is described. This method is fast, uses commercially available reagents, and is compatible with both electron-rich and electron-deficient arene substrates. This method has been applied to the manual synthesis of a variety of clinically relevant radiotracers including protected [18F]F-phenylalanine and [18F]F-DOPA. In addition, an automated synthesis of [18F]MPPF is demonstrated that delivers a clinically validated dose of 200 ± 20 mCi with a high specific activity of 2400 ± 900 Ci/mmol. Step 1: Iodide S1 was prepared via a modification of a literature procedure. 8 To an oven-dried flask, Boc-Phe(4-I)-OH (1016.6 mg, 2.6 mmol, 1.0 equiv) was dissolved in dichloromethane (22 mL) at room temperature. DMAP (34.3 mg, 0.3 mmol, 0.1 equiv) and ethanol (0.3 mL, 5.1 mmol, 2.0 equiv) were added to the solution, and the reaction was placed under nitrogen and cooled to 0 °C. DCC (862.5 mg, 4.2 mmol, 1.6 equiv) was added slowly. The mixture was allowed to warm up to room temperature and react overnight at room temperature under nitrogen. The white precipitate formed was filtered off and the organic filtrate was washed with brine (1 x 20 mL), dried using magnesium sulfate, and concentrated. The crude residue was purified by column chromatography (silica gel, 20% ethyl acetate in hexane) affording the product (S1) as a white solid ( Step 2: The general procedure for stannane synthesis was followed using S1 (404.8 mg, 1.0 mmol) and heating for 2 h. Purification by flash column chromatography eluting with 20% ethyl acetate in hexanes afforded 14-SnBu 3 as a colorless oil (73.4 mg, 17% yield, R f = 0.5 in 20% ethyl acetate in hexanes). NBoc 2 -Phe(4-SnMe 3 ) ethyl ester (15-SnMe 3 ) was prepared by the following 3 step procedure. Step 1: Aryl bromide S7 was prepared via a literature procedure. 11,12 The product was purified by flash column chromatography on silica gel (15% ethyl acetate in hexanes), which afforded S7 as a yellow solid (687.0 mg, 49% yield, R f = 0.32 in 30% ethyl acetate in hexanes, mp = 108 -109 °C). The 1 H and 13 C NMR spectra matched those reported previously in the literature. 12 Step 2: The general procedure was followed using S7 (263.9 mg, 0.93 mmol) and heating for 7 h. Purification by flash column chromatography eluting with 30% ethyl acetate in hexanes afforded 18-SnBu 3 as a yellow oil (245.8 mg, 54% yield, R f = 0.43 in 20% ethyl acetate in hexanes). The 1 H and 13 C NMR spectra matched those reported previously in the literature. 13 Synthesis and Characterization of Fluorinated Standards (E)-Fluorostyrene (S8) In a nitrogen atmosphere glovebox, tributyl(phenylethenyl)tin (415.9 mg, 1.1 mmol, 1 equiv), Cu(OTf) 2 (766.2 mg, 2.1 mmol, 2 equiv), KF (245.3 mg, 4.2 mmol, 4 equiv), and pyridine (1.3 mL, 16.1 mmol, 15 equiv) were weighed into a flask equipped with a magnetic stir bar. DMA (100 mL) was added. The reaction mixture was allowed to stir at 140 ºC for 4 h. The resulting solution was cooled to room temperature, diluted with diethyl ether and washed with water (2 x 200 mL), and brine (200 mL). The organic extracts were dried and concentrated. Purification by flash column chromatography eluting with 20% diethyl ether in pentane afforded S8 as a colorless oil. The 19 F NMR spectroscopic data was identical to that reported. 14 NHBoc-Phe(4-F) ethyl ester (S9) Authentic standard S9 was prepared by the following procedure via a modification of a literature procedure. 8 To an oven-dried flask, NHBoc-Phe(4-F)-OH (509.4 mg, 1.8 mmol, 1.0 equiv) was dissolved in dichloromethane (15 mL, 0.12 M) at room temperature. DMAP (22.2 mg, 0.18 mmol, 0.1 equiv) and ethanol (0.21 mL, 3.6 mmol, 2.0 equiv) were added to the solution and the reaction was placed under nitrogen and cooled to 0 °C. DCC (595.0 mg, 2.9 mmol, 1.6 equiv) was added slowly. The mixture was allowed to warm to room temperature and was then stirred overnight at room temperature under nitrogen. Over this time, a white precipitate formed and was removed by filtration. The filtrate was washed with brine (1 x 20 mL), dried over magnesium sulfate, and concentrated. The crude residue was purified by column chromatography (silica gel, 20% ethyl acetate in hexane) affording the product (S9) as a colorless oil (453.1 mg, 81% yield, R f = 0.44 in 20% ethyl acetate in hexanes). The 1 H and 13 C NMR spectra matched those reported previously in the literature. 15 NBoc 2 -Phe(4-F) ethyl ester (S10) Authentic standard S10 was prepared via a modification of a literature procedure. 9 To a solution of S9 (334.5 mg, 1.07 mmol, 1 equiv) in dry acetonitrile (20 mL) under nitrogen was added 4-dimethylaminopyridine (DMAP) (54.0 mg, 0.44 mmol, 0.4 equiv) in dry acetonitrile (5 mL) and di-tert-butyl dicarbonate (0.5 mL, 2.2 mmol, 2.0 equiv) in dry acetonitrile (5 mL). The mixture was stirred at room temperature overnight and then concentrated under nitrogen. Purification by flash column chromatography eluting with 20% ethyl acetate in hexanes afforded S10 as a colorless oil (81.5 mg, 18% yield, R f = 0.5 in 20% ethyl acetate in hexanes). F-PEB (S12) Aryl fluoride S12 was prepared via a literature procedure. 12 The product was purified by flash chromatography on silica gel (15% ethyl acetate in hexanes), which afforded S12 as a brown solid (142.9 mg, 63% yield R f = 0.15 in 20% ethyl acetate in hexanes, mp = 75 -77 °C). The 1 H and 13 C NMR spectra matched those reported previously in the literature. 12 Experimental Details Copper Screen for Cu-Mediated Fluorination of 1-SnBu 3 with KF In a nitrogen atmosphere glovebox, substrate 1-SnBu 3 (0.1 mL, 0.25 M, 0.025 mmol, 1 equiv), copper salt (0.1 mmol, 4 equiv) and KF (5.8 mg, 0.1 mmol, 4 equiv) were weighed into a 4 mL vial. MeCN (0.3 mL total volume) was added, and the vial was sealed with a Teflon-lined cap. The reaction mixture was allowed to stir at 60 ºC for 18 h. The resulting solution was diluted with MeCN (1 mL). 1,2-Difluorobenzene was added as an internal standard, and the crude reaction was analyzed by 19 F NMR spectroscopy. The yields of 1 are listed in Table S1. In a nitrogen atmosphere glovebox, substrate 1 (0.025 mmol, 1 equiv), Cu(OTf) 2 (36 mg, 0.1 mmol, 4 equiv), KF (5.8 mg, 0.1 mmol, 4 equiv), and 18-crown-6 (26 mg, 0.1 mmol, 4 equiv) were weighed into a 4 mL vial. MeCN (0.3 mL total volume) was added, and the vial was sealed with a Teflon-lined cap. The reaction mixture was allowed to stir at 60 ºC for the appropriate time. The resulting solution was diluted with MeCN (1 mL). 1,2-Difluorobenzene was added as an internal standard, and the crude reaction was analyzed by 19 F NMR spectroscopy. The yields of 1 are listed in Table S7. Table S8. Trimethyltin versus Tributyltin Substrates In a nitrogen atmosphere glovebox, arylstannane substrate (0.1 mL, 0.25 M, 0.025 mmol, 1 equiv), Cu(OTf) 2 (36 mg, 0.1 mmol, 4 equiv), KF (5.8 mg, 0.1 mmol, 4 equiv), and 18-crown-6 (26 mg, 0.1 mmol, 4 equiv) were weighed into a 4 mL vial. MeCN (0.3 mL total volume) was added, and the vial was sealed with a Teflon-lined cap. The reaction mixture was allowed to stir at 60 ºC for 2 h. The resulting solution was diluted with MeCN (1 mL). 1,2-Difluorobenzene was added as an internal standard, and the crude reaction was analyzed by 19 F NMR spectroscopy. The yields of aryl fluoride product are listed in Table S10. Control reactions for the fluorination of substrate 1-SnBu 3 were conducted without copper salt and without KF. In both cases, the fluorinated product 1 was not observed. The reactions were conducted using to the general procedure (vida infra), on a 0.025 mmol scale. Materials and Methods Unless otherwise stated, reagents and solvents were commercially available and used without further purification. HPLC grade acetonitrile, anhydrous N,N-dimethylformamide, anhydrous N,N-dimethylacetamide, potassium trifluoromethanesulfonate, and potassium carbonate were purchased from Fisher Scientific. Sterile product vials were purchased from Hollister-Stier. QMA-light Sep-Paks were purchased from Waters Corporation. QMA-light Sep-Paks were flushed with 10 mL of ethanol, followed by 10 mL of 90 mg/mL potassium trifluoromethanesulfonate solution, and finally 10 mL of sterile water prior to use. Azeotropic drying/evaporation was achieved by heating the reaction vessel to 100 ºC and drawing vacuum for 6 min. The reaction vessel was then subjected to an argon stream and simultaneous vacuum draw for an additional 6 min. N,N-Dimethylacetamide (8 mL) was added to the dried reagent, and the sample was cooled to 40 ºC and was transferred to a S10 sterile vial for subsequent use in reactions. As an example, approximately 80 mCi of prepared [ 18 F]KF in 8 mL DMA is isolated with a 5 min cyclotron beam. It should be noted that percent recovery data is only relevant for manual reactions, not automated one-pot syntheses. Synthesis of 18 F-Labeled Molecules (Manual Synthesis) Unless otherwise noted, this procedure was used for the synthesis of the [ 18 aluminum block without stirring at 140 ºC for 30 min. After 30 min, the reaction was allowed to cool to room temperature. Radio-TLC analysis was conducted to determine radiochemical conversion (% RCC). The crude reaction mixture was spotted onto a standard silica-coated glass plate and run using 1:1 hexane/ethyl acetate in a glass TLC chamber. The RCC was then determined by dividing the integrated area under the fluorinated product spot by the total integrated area of the fluorine-18 on the TLC plate. To prepare samples for HPLC analysis: 0.1 mL of the reaction mixture or for the coinjection analysis 0.1 mL of the reaction mixture spiked with 0.1 mL of 1 mg/mL fluorinated standard solution were transferred to an HPLC autosampler vial. Eluent systems and columns used for HPLC analysis are described below. RCC = integration of 18 F product peak / sum of integration of all 18 F peaks General HPLC Conditions Two general sets of HPLC conditions were used. A gradient method (Conditions A) or an isocratic method (Conditions B) was used for all manual reactions. Additional Optimization Results (manual synthesis conditions) Unless otherwise stated, 4-(tributylstannane)-1,1'-biphenyl (3-SnBu 3 ) was used for all optimization screens. The reaction scheme and accompanying tables in each subsection describe the reaction conditions employed, with bold typeface in the reaction scheme denoting the variable tested in each case. All reactant values are expressed in equivalents relative to arylstannane for brevity and simplicity. Red typeface denotes the 18 F source used, typically 0.1 mL of a 8 mL DMA solution containing [ 18 F]KF, 10 mg KOTf and 50 μg K 2 CO 3 . All RCC are n = 2 or greater. The product was trapped on a C18 extraction disk, washed with 10 mL of sterile water, eluted with 1 mL of EtOH, and then rinsed with 9 mL of saline solution. The resulting 10 mL solution was passed through a sterile filter and submitted to standard quality control tests (tests and results are described in detail below). [ 18 F]MPPF was produced in a 13 ± 1% (n=4) yield (200 mCi ± 20, n=4). Specific Activity Calculation An aliquot of the sample was injected onto an analytical HPLC using Conditions A. The UV peak corresponding to the radiofluorinated product was determined by overlaying the UV and RAD traces (with a 0.2 min offset as described in the HPLC section). The UV area was then used to calculate the concentration of the product based on linear regression analysis of appropriate fluoroarene standards. A standard curve was generated from the standard solutions, each run in duplicate (0.0001 mg/mL to 1.0 mg/mL). This provided the concentration of the product in mmol/mL. Dividing the activity concentration (Ci/mL) by the HPLC-derived concentration of product (mmol/mL) provided the specific activity in Ci/mmol. This reflects an end of synthesis (EoS) specific activity. [ 18 F]MPPF was produced with a specific activity 2,400 ± 900 Ci/mmol (n=4). In the section below, two HPLC traces are presented for each substrate contained in Figure 2 and Figure 3. The trace shows the RAD and UV trace (254 nm or 280 nm) from the crude reaction mixture spiked with an authentic standard of the fluorinated product. This was used to confirm the identity of the radiofluorinated product. The wavelength shown is the wavelength where the analyte compound exhibited greatest absorptivity. Because of the physical separation of the two detectors, a horizontal offset of 0.	2018-04-03T02:15:51.794Z	2016-10-10T00:00:00.000	{ "year": 2016, "sha1": "8157f45cc4c08ff10c5224f3c5d7b83c7a2876d0", "oa_license": "acs-specific: authorchoice/editors choice usage agreement", "oa_url": "https://doi.org/10.1021/acs.orglett.6b02911", "oa_status": "HYBRID", "pdf_src": "PubMedCentral", "pdf_hash": "8157f45cc4c08ff10c5224f3c5d7b83c7a2876d0", "s2fieldsofstudy": [ "Chemistry", "Biology" ], "extfieldsofstudy": [ "Chemistry", "Medicine" ] }
212670135	pes2o/s2orc	v3-fos-license	Association between sleep quality and time with energy metabolism in sedentary adults The aim of the present study was to investigate the relationship of sleep quality and time with basal metabolic rate (BMR) and fuel oxidation in basal conditions and during exercise in sedentary middle-aged adults. We also studied the mediation role of dietary intake and adherence to the traditional Mediterranean Diet in the relationship between sleep parameters and energy metabolism parameters.A secondary analysis of the FIT-AGEING study was undertaken. 70 middle-aged sedentary adults (40–65 years old) participated in the present study. Sleep quality was assessed using the Pittsburgh Sleep Quality Index (PSQI) and wrist accelerometers (ActiSleep, Actigraph, Pensacola, Florida, USA) for 7 consecutive days. BMR was measured with indirect calorimetry and fuel oxidation was estimated through stoichiometric equations. Maximal fat oxidation was determined by a walking graded exercise test and dietary intake with 24 h recalls. Adherence to the traditional Mediterranean diet was assessed through the PREDIMED questionnaire. PSQI global score (poor sleep quality) was associated with lower basal fat oxidation (BFox), both expressed in g/min and as a percentage of BMR, independently of confounders. We did not find any association between other sleep and energy metabolism parameters. No mediating role of the dietary intake or PREDIMED global score was observed in the association of PSQI and BFox. In conclusion, our study showed that a subjective poor sleep quality was associated with lower BFox, which is not mediated by dietary intake in sedentary adults. Materials and Methods Participants and design. A total of 70 (36 women) middle-aged sedentary adults (40-65 years old) participated in this cross-sectional study. The participants were enrolled in the FIT-AGEING study 20 , an exercise-based randomized controlled trial (clinicaltrial.gov: ID: NCT03334357). Data for these subjects were collected at baseline data collection in the FIT-AGEING study. All of them declared: (i) to be non-physically active (<20 min of moderate-intensity physical activity on 3 days/week), (ii) to have a stable weight (weight changes <5 kg) over the last 5 months, (iii) to be healthy and (iv) to be free of medication (medication for thyroid, betablockers, benzodiazepines, glucose lowering medication, or cardiovascular medication) during the last 5 months. All participants gave their oral and written informed consent before the beginning of the intervention. The study was approved by the Ethics Committee on Human Research of the University of Granada and the Andalusian Health Service (SAS) (CEI-Granada) (0838-N-2017). The study protocols and experimental design were applied according to the last revised ethical guidelines of the Declaration of Helsinki. All assessments were made at the Sport and Health University Research Institute (iMUDS, Granada, Spain) during September and October 2016 and September and October 2017. Anthropometric and body composition measurements. Anthropometric variables were measured by a certified anthropometrist [the International Society for the Advancement of Kinanthropometry (ISAK)] following the ISAK guidelines 21 . Both body weight and height were assessed using an electronic scale and stadiometer (Seca model 799, Hamburg, Germany), and the body mass index (BMI) was calculated as (weight [kg]/ height 2 [m]). A dual-energy X-ray absorptiometry scanner (Discovery Wi, Hologic, Inc., Bedford, MA, USA) was used to measure body composition following the manufacturer's recommendations. We conducted the quality controls, the positioning of the participants and the analyses of the results following the manufacturer's recommendations. An automatic delineation of the anatomic regions was performed by the software APEX 4.0.2. We acquired spine phantom quality control scans on each study day. The results displayed lean mass and fat mass and the lean mass index (LMI) and the fat mass index (FMI), which were calculated as (lean mass [ Sleep quality and time assessment. The Pittsburgh Sleep Quality Index (PSQI) is a self-report tool which consists of 19-item scale that provides 7 component scores (ranges 0-3): (i) subjective sleep quality (very good to very bad), (ii) sleep latency (≤15 minutes to >60 minutes), (iii) sleep duration (≥7 hours to <5 hours), (iv) sleep efficiency (≥85% to <65% hours sleep/hours in bed), (v) sleep disturbances (not during the past month to ≥ 3 times per week), (vi) use of sleeping medications (none to ≥ 3 times a week), and (vii) daytime dysfunction (not a problem to a very big problem), with a total global score ranging from 0 to 21 22 . A PSQI global score higher than 5 indicates poor sleep quality 22 . Objective characteristics of sleep-wake cycles were monitored with a wrist-worn accelerometer (ActiSleep, Actigraph Pensacola, Florida, USA) for 7 consecutive days (24 hours/day) 20 . Participants received detailed information on how to wear the accelerometer and were asked to remove it only for water activities. It was also recorded the times in which the participants went to bed every night, woke up every morning and removed the device every day. The accelerometers used an epoch length of 5 seconds and a frequency rate of 100 Hz to store raw accelerations 23 . The raw accelerations were exported in ".csv" format using ActiLife v. 6.13.3 software (ActiGraph, Pensacola, FL, US) and processed using the GGIR package (v. 1.6-0, https://cran.r-project.org/web/ packages/GGIR/index.html) 24 in R (v. 3.1.2, https://www.cran.r-project.org/). We derived the Euclidean Norm Minus One G (ENMO) as √(x^2 + y^2 + z^2) −1G (where 1 G ~ 9.8 m/s2) with the accelerometer's z angle to describe sleep patterns. We used a previously published algorithm combining data from the accelerometers and diary reports to detect sleep period time 25,26 . According to this algorithm, sleep was defined as any period of sustained inactivity, in which there were minimal changes in the arm angle (i.e., as much 5 degrees for 5 minutes periods) during a period recorded as sleep by the participant in their diary reports 25 . The following variables were analyzed: total sleep time (minutes slept between bedtime and wake time), sleep efficiency (percentage of time asleep while in bed) and wake after sleep onset (minutes awake between sleep onset and wake time). It is to note that only the participants wearing the accelerometers for ≥16 hours/day for at least 4 days (including at least 1 weekend day) were included in the final analyses (i.e. a total of 4 participants did not meet these conditions) 23 . The mean accelerometer wear-time for the 70 participants included in the final analyses were 6.7 days (4.3% of non-wear time). www.nature.com/scientificreports www.nature.com/scientificreports/ Basal metabolic rate and fuel oxidation in basal conditions. Subjects were told to arrive at the laboratory in fasting condition of at least 8 h in a motor vehicle and to avoid any moderate/vigorous physical activity in the previous 24 h/48 h respectively; all were required to confirm that they had met these conditions. The evening meal consumed by the subjects prior to fasting was previously standardized: an egg omelet with fried tomato and boiled rice. BMR and fuel oxidation in basal conditions were measured through indirect calorimetry (IC) following the current scientific consensus 27 . All tests were conducted in the same quiet room with controlled room temperature (i.e. 22-24 °C) and humidity (i.e. 35-45%). IC measurements were performed during 30-minute periods with a CPX Ultima CardiO2 system (Medical Graphics Corp, St Paul, MN, USA) employing a neoprene face-mask with no external ventilation 27 . The Ultima CardiO2 metabolic cart device assessed oxygen consumption (VO2) using a galvanic fuel cell, and carbon dioxide production (VCO2) via non-dispersive infrared analysis using a breath-by-breath system 28 . A gas calibration using 2 standard gas concentrations and a flow calibration using a 3-L calibration syringe were performed following the manufacturer's recommendations. Prior to the start of the BMR assessment, the subjects reclined on a bed for ~30 min in a comfortable supine position covered by a sheet 29,30 . During the assessment, participants laid on a bed in a supine position and were instructed to breathe normally and not to talk, fidget or sleep. The first 5 minutes of each measurement were discarded and the most stable 5-min period that met steady state criteria (i.e. coefficient of variation <10% in VO2, CO2, minute ventilation, and coefficient of variation <10% in respiratory exchange ratio) was considered for further analyses following previous studies [29][30][31][32] . The Weir's abbreviated equation 33 was used to estimate the BMR expressed in kcal/day and also calculated with respect to the lean mass (BMR LM ). The Frayn's equation was used to estimate BFox and basal carbohydrate oxidation (BCHox) expressed in g/min 34 . The BFox and BCHox were also expressed as a percentage of the BMR. Maximal fat oxidation during exercise assessment. MFO and the intensity that elicit MFO (FATmax) were assessed in a different day of the BMR and BFox/BCHox test (i.e. interval 3 to 15 days). Participants were asked to arrive at the laboratory in a fasted state of 6 hours and to avoid any physical activity both moderate (24 h) and vigorous intensity (48 h) before the measurement. A walking graded test on a treadmill (H/P/cosmos pulsar, H/P/cosmos sports & medical GmbH, Nussdorf-Traunstein, Germany) was performed to calculate MFO and FATmax following a previously validated methodology 35 . Briefly, the protocol started assessing the maximal walking speed of each participant [35][36][37] . After ~3 minutes resting, the walking graded test started with a 3-minute warm up at 3.5 km/h. Subsequently, the treadmill speed was increased 1 km/h every 3 minutes until the maximal walking speed was reached. Thereafter, the treadmill gradient was increased 2% every 3 minutes until the respiratory exchange ratio was above 1.0. An automated gas analysis system (CPX Ultima CardiO2; Medical Graphics Corp, St Paul, MN) was used to record breath-by-breath gas exchange measurements. Participants wore an oronasal mask (model 7400, Hans Rudolph Inc, Kansas City, MO, USA) equipped with a prevent ™ metabolic flow sensor (Medgraphics Corp, Minnesota, USA). Gas analysis systems were calibrated following the manufacturer's recommendations. VO2 and VCO2 were averaged over and the last 60 seconds of each graded exercise protocol stage. Frayn's equation was used to estimate fat oxidation rates 34 . These fat oxidation values were plotted against the relative-exercise intensity, expressed as the percentage of maximum oxygen uptake (VO2max); a third-degree polynomial curve was built to determine MFO and FATmax 38 . MFO was also expressed as MFO LM in order to relativize it to the lean mass. Maximal carbohydrate oxidation was not included in the analyses since it is not a key factor of energy metabolism during exercise 39 . Indeed, our recent systematic review has analyzed a total of 112 studies which included data about fuel oxidation during exercise 37 . None of those studies reported maximal carbohydrate oxidation during exercise. Cardiorespiratory fitness assessment. VO2max was determined using a maximum treadmill (H/P/ Cosmos Pulsar treadmill, H/P/Cosmos Sport & Medical GMBH, Germany) exercise test following the modified Balke protocol, which has been extensively validated 40 . In short, the warm up consisted in walking at 3 km/h for 1 minute followed by 2 minutes at 4 km/h. The incremental protocol started at a speed of 5.3 km/h (0% grade), which was kept constant with the gradient increasing by 1% every minute until the participants reached their volitional exhaustion. We used the same indirect calorimetry and software as in the MFO assessment. Dietary intake assessment. Diet was assessed using three 24-hour recalls carried out on 3 separate days (2 weekdays and 1 weekend day) by a qualified and trained research dietitian. Dietary recalls were done on different days than the MFO and VO 2 assessments. In the face-to-face interviews, the participants were asked to recall all food consumed during the previous day. The interviews involved a detailed assessment and description of the food consumption using colored photographs of different-size portions of food to improve the participants' accuracy of food quantification 41 . These data were introduced by two independent qualified and trained dietitians in the EvalFINUT ® software. Energy, macronutrient, fiber, lipid profile and ethanol intake data were obtained by EvalFINUT ® , which is based on the USDA (United States Department of Agriculture) and BEDCA ("Base de Datos Española de Composición de Alimentos") databases. Dietary energy density was calculated by dividing the energy contained in food and beverages (excluding water) by the total weight of daily food and beverages (expressed as kcal/g) 42 . Energy and weight data of daily food and beverages were obtained from the 24-hours recalls. The traditional Mediterranean diet is associated with a lower prevalence of chronic diseases (i.e. obesity, metabolic syndrome, cardiovascular diseases, cancer) and mortality 43 . The adherence to the traditional Mediterranean Diet (MedDiet) was estimated by using the 14-point questionnaire of adherence to the MedDiet used and validated in the PREDIMED trial 44 . The PREDIMED questionnaire includes 12 questions related to frequency intake of key foods and 2 questions related to specific dietary habits of the MedDiet. Each question scores 0 or 1 point. The global score ranges from 0 to 14, being 0 points null adherence and 14 points complete adherence to the MedDiet. The PREDIMED questionnaire proved to be very useful in a large Spanish cohort for a quick adherence estimation to the traditional MedDiet 44 . Statistical analysis. The sample size and power calculations were made based on the data of a pilot study of the FIT-AGEING study 20 . This study aimed to compare the influence of different exercise programs on BMR, BFox and MFO in sedentary middle-aged adults. We based the sample size calculations on a minimum predicted change in MFO of 0.05 g/min between the intervention groups and the control group, and an SD for this change of 0.05 g/min. A sample size of 17 participants was predicted to provide a statistical power of 80%, considering a type I error of 0.05. Assuming a maximum loss of 25% at follow-up, we decided to recruit at least 20 participants for each group (N = ~80 individuals). The present study is based on a secondary analysis using baseline data from the FIT-AGEING study, and therefore a specific sample size calculation was not conducted. We used the Shapiro-Wilk test, visual check of histograms, Q-Q and box plots to verify all variable distributions. The descriptive parameters were reported as mean and standard deviation. Given that we did not observe a sex interaction, we conducted the analysis including men and women together. Simple linear regressions were performed to examine the association between sleep time and quality (PSQI global score, total sleep time, sleep efficiency and wake after sleep onset) with BMR, BMR LM , BFox, BCHox, MFO, MFO LM and FATmax. We also conducted multiple linear regression models to test these associations after adjusting for sex (Model 1), sex and age (Model 2) and sex, age and FMI (Model 3). Pearson correlation was performed to assess the association between sleep parameters and dietary outcomes. Effect modification analyses were conducted to test the joint effects of dietary intake (dietary intake outcome*sleep outcome) and sleep quality on energy metabolism. To quantify the mediating role of dietary intake (i.e. energy, macronutrient, fiber, ethanol and lipid profile intake, and PREDIMED total score) in the relationship between sleep parameters and BMR and fuel oxidation, we conducted mediation analyses 45 . We used the PROCESS macro version 3.3, model 4 with 5,000 bias-corrected bootstrap samples and 95% confidence intervals. Bootstrapping is a nonparametric resampling procedure that does not require the assumption of normality of the sampling distribution 46 The mediation was estimated using the indirect effect, which indicates the change in the effect of the independent variable on the outcome that can be endorsed to the proposed mediator. Indirect effects (a × b paths) with confidence intervals not including zero are interpreted as statistically significant 47 which could occur regardless of the significance of the total effect (c path, effect of the independent variable on the dependent variable) and the direct effect (c' path, effect on the dependent variable when both the independent and the mediator variables are included as independent variables) 45 . To quantify how much of the total effect was due to the mediation, we calculated the percentage of mediation ([indirect effect / total effect] × 100) provided when the total effect was larger than the indirect effect with the same direction 45 . All analyses were conducted using the Statistical Package for Social Sciences (SPSS, v. 25.0, IBM SPSS Statistics, IBM Corporation) and the level of significance was set at <0.05. Graphical presentations were prepared using GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA). Ethical standards. Ethical approval for the study was given by the Ethics Committee on Human Research at the University of Granada and Servicio Andaluz de Salud (CEI-Granada) (0838-N-2017). Written informed consent was obtained from all subjects. This study was in accordance with the last revised ethical guidelines of the Declaration of Helsinki. Results The characteristics of the study sample are shown in Table 1. We observed an inverse association between total sleep time and BMR (P < 0.001; Table 2), which disappeared after including sex, age and FMI in the model (all P > 0.105; Table 2). No association was found between the remaining subjective or objective sleep parameters with BMR and BMR LM (all P > 0.071; Table 2), neither when we accounted for confounders. An inverse association was detected between PSQI global score with BFox (expressed in g/min, and as %BMR) (all P < 0.001; Table 3), which remained significant after including sex, age and FMI in the model (all P < 0.002; Table 3). We did not find any significant association between any objective sleep parameter with BFox (all P > 0.265; Table 3). PSQI was also positively associated with BCHox (expressed in g/min, and as %BMR) even after controlling for sex, age and FMI (all P < 0.002; Table S1). We showed an inverse association between total sleep time with MFO (expressed in g/min; P = 0.008; Table 4), which disappeared when the model includes sex, age and FMI (all P > 0.651; Table 4). No association was found between the remaining sleep parameters with MFO and MFO LM (all P > 0.171; Table 4). We repeated all previous associations controlling for menopausal status (pre-or post-menopausal) in order to avoid the possible cofounder of female hormones, and the results did not change (data not shown). We observed only a negative association of fiber intake and PSQI global score and cholesterol intake negatively and positively associated with total sleep time and wake after sleep onset respectively (Table S5). However, we observed a modification effect of different dietary factors (i.e. fiber and ethanol intake; Table S5). Despite this modification effect of dietary factors and the several associations between dietary factors and sleep parameters, we did not find a mediating effect of energy, dietary energy density, fat, protein, carbohydrate, fiber intake, lipid profile intake, ethanol intake and PREDIMED total score on the association of the PSQI global score and BFox both expressed in g/min and in %BMR (Figs. S1-S4). Discussion The main finding of the present study is that a poor subjective sleep quality was associated with lower BFox independently of sex, age and body composition outcomes. No consistent association was observed between any sleep quality and time parameters with BMR, MFO and FATmax. Moreover, our results indicated that the association of PSQI global score with BFox was not mediated by dietary intake and MedDiet adherence. We observed an inverse association between total sleep time and BMR which disappeared after controlling for confounders. The energy expenditure is lowest during sleep, therefore a high total sleep time is related with a prolonged period of the lowest energy expenditure 48 . Sleep deprivation could increase energy expenditure since energy expenditure is reduced during sleep 48 . Sharma et al. proposed that these reduction in energy expenditure could be influenced by circadian rhythm, body temperature and muscle temperature 48 . However, the results should be interpreted with caution because this association disappeared after controlling for sex, age and FMI. Several physiological mechanisms could explain the relationship between sleep quality and BFox. Sleep restriction is associated with insulin resistance characterized by a decreased insulin-mediated glucose uptake 49 www.nature.com/scientificreports www.nature.com/scientificreports/ arealso related to low leptin levels or leptin resistance 51 which are associated with an impaired fatty acid oxidation 52 . Sleep disruption (discontinuity of sleep) can lead to the disruption of circadian rhythms 13 , which orchestrate crucial physiological and behavioral functions, being one of these the regulation of carbohydrate and fatty acid metabolism 12 . Higher sleep duration and quality are associated with a healthier gut microbiome 53 , which could suppress insulin signaling, increase β-oxidation and inhibit fat oxidation derived from the production of short-chain fatty acids 54 . Furthermore, sleep disruption (discontinuity of sleep) could decrease melatonin production 13 , which has important metabolic functions, such as lipolysis, regulating the energy flow 55 . An increase in the production of pro-inflammatory cytokines and reactive oxygen species is observed in impaired sleep patterns 13 . Both inflammation and oxidative stress could modulate metabolic flexibility, specifically fat oxidation 56,57 . Therefore, based on the above-mentioned mechanisms, a healthy sleep pattern could improve metabolic health via the increment of BFox and viceversa. In addition, an impaired sleep pattern, determined by a low sleep duration could increase energy intake through several potential mechanisms: increment of time and opportunities for eating, psychological distress, sensitivity to food reward, energy needed to sustain wakefulness, hunger hormones and decrease dietary restraint 16 . A lack of sleep or low sleep quality could increase the intake of high energy-dense foods, high fat and sugary snacks, which are low in fiber 16 . In this sense, although we did not find any association between energy and macronutrient intake, we observed that fiber intake was negatively associated with PSQI global score. Fiber intake could have different metabolic effects (i.e. insulin sensitivity and glycemia improvement) 58 , that could have a potential role in the regulation of fat oxidation. However, we did not find any mediating role of dietary intake (i.e. fiber intake) between the association of PSQI with BFox. The lack of a mediating role may be due to specific issues: (i) since dietary outcomes were assessed in a specific time point, it could be that the dietary intake was insufficiently maintained over time to modify BFox; (ii) the possible lower and upper threshold for when dietary intake (i.e. fat intake) could modify fat oxidation 59 ; (iii) the inter-individual variability, body composition and metabolic status influence on fat oxidation 8 ; (iv) a sleep patterns insufficiently maintained over time. The lack of association between any sleep outcomes with BMR and MFO could be explained by different factors. Sleep is a complex phenomenon influenced by behavioral and physiological mechanisms (i.e. homeostatic, circadian and metabolic control) under the participant's natural sleep environment that we have not investigated 60 . These factors could influence the relationship between sleep parameters and BMR and MFO. We also observed an inverse association between total sleep time with MFO. A previous study of Konishi et al. observed that a night of sleep deprivation did not affect MFO in healthy young men 15 . It has been reported several detrimental effects of long sleep for optimal health 61 . In addition, long sleep could increase fatigue, physiological deprivation, which could influence insulin resistance and hormonal imbalance 62 . Although the mechanisms are not clear, the above-mentioned mechanisms could have influenced this relationship. However, the results should be interpreted with caution because this association disappeared after controlling for sex, age and FMI. Surprisingly, different results were observed when the association between sleep quality and energy metabolism was performed considering subjective instead of objective measures of sleep quality. It has been previously reported that PSQI and accelerometer records measure different attributes of sleep, highlighting the bias of accelerometry to register wakefulness, thus lying in bed awake but motionless is likely to be coded as sleep 63 . Therefore, it is recommended to use both methods to obtain complementary information additionally to the body movements 64 . These differences in measurement of sleep attributes could explain the different results of the associations between sleep quality and energy metabolism. Despite accelerometer records and subjective measurements are a valid and extensively used measure of sleep quality 26,65 they cannot differentiate between rapid eye movement sleep (REM) and non-rapid eye movement sleep (NREM), restricting the detailed assessment of the real biologic process of sleep. REM and NREM phases are metabolically different 66 . In REM sleep glucose uptake is increased, leading to anaerobic glucose metabolism 67,68 , therefore sleep quality in each phase could be differently associated with energy metabolism. Future studies that examine the relationship between REM and NREM sleep using polysomnography records with BMR and fuel oxidation in basal conditions and during exercise are needed. The present study should be interpreted with caution; the study has a cross-sectional design that does not allow to establish causal relationship. Therefore, experimental studies should manipulate BMR and fuel oxidation and/or sleep (e.g. sleep deprivation) under well-controlled lab conditions in order to establish causal relationship. Furthermore, sleep and dietary parameters were assessed only in a specific timepoint, which do not allow us to extrapolate our results to chronic sleep or dietary patterns. Our study only included middle-aged sedentary adults and, consequently, we cannot extrapolate our results to older, younger, and/or physically active individuals. The difficulty of an accurate dietary evaluation with possible underreporting or misclassification should be considered, as in all cross-sectional studies. Lastly, the narrow PSQI global score range should be taken into account when interpreting our results. conclusions In conclusion, our study showed that a subjective poor sleep quality was associated with lower BFox. No association was found between the remaining sleep parameters with BMR and fuel oxidation in basal conditions and during exercise. Moreover, our findings indicated that the association of PSQI global score with BFox was not mediated by dietary intake and MedDiet adherence. Further studies are needed to better understand the physiological mechanisms of sleep regulation and how it could influence the BMR and fuel oxidation in basal conditions and during exercise.	2020-03-12T15:05:36.361Z	2020-03-12T00:00:00.000	{ "year": 2020, "sha1": "0b6c151abfa898a5af7496501091ab454b309f49", "oa_license": "CCBY", "oa_url": "https://www.nature.com/articles/s41598-020-61493-2.pdf", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "896730e56d78ce9bcb4215043fb692117cdc0226", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
77395051	pes2o/s2orc	v3-fos-license	Persistence of chloroquine resistance alleles in malaria endemic countries: a systematic review of burden and risk factors Background Chloroquine, a previous highly efficacious, easy to use and affordable anti-malarial agent was withdrawn from malaria endemic regions due to high levels of resistance. This review collated evidence from published-reviewed articles to establish prevalence of Pfcrt 76T and Pfmdr-1 86Y alleles in malaria affected countries following official discontinuation of chloroquine use. Methods A review protocol was developed, registered in PROSPERO (#CRD42018083957) and published in a peer-reviewed journal. Article search was done in PubMed, Scopus, Lilacs/Vhl and Embase databases by two experienced librarians (AK, RS) for the period 1990-to-Febuary 2018. Mesh terms and Boolean operators (AND, OR) were used. Data extraction form was designed in Excel spread sheet 2007. Data extraction was done by three reviewers (NL, BB and MO), discrepancies were resolved by discussion. Random effects analysis was done in Open Meta Analyst software. Heterogeneity was established using I2-statistic. Results A total of 4721 citations were retrieved from article search (Pubmed = 361, Lilac/vhl = 28, Science Direct = 944, Scopus = 3388). Additional targeted search resulted in three (03) eligible articles. After removal of duplicates (n = 523) and screening, 38 articles were included in the final review. Average genotyping success rate was 63.6% (18,343/28,820) for Pfcrt K76T and 93.5% (16,232/17,365) for Pfmdr-1 86Y mutations. Prevalence of Pfcrt 76T was as follows; East Africa 48.9% (2528/5242), Southern Africa 18.6% (373/2163), West Africa 58.3% (3321/6608), Asia 80.2% (1951/2436). Prevalence of Pfmdr-1 86Y was; East Africa 32.4% (1447/5722), Southern Africa 36.1% (544/1640), West Africa 52.2% (1986/4200), Asia 46.4% (1276/2217). Over half, 52.6% (20/38) of included studies reported continued unofficial chloroquine use following policy change. Studies done in Madagascar and Kenya reported re-emergence of chloroquine sensitive parasites (IC50 < 30.9 nM). The average time (years) since discontinuation of chloroquine use to data collection was 8.7 ± 7.4. There was high heterogeneity (I2 > 95%). Conclusion The prevalence of chloroquine resistance alleles among Plasmodium falciparum parasites have steadily declined since discontinuation of chloroquine use. However, Pfcrt K76T and Pfmdr-1 N86Y mutations still persist at moderate frequencies in most malaria affected countries. Background Chloroquine was once an important medicine used in malaria treatment especially due to its affordability, ease of use and high anti-malarial efficacy. However, due to high level of resistance among Plasmodium falciparum parasites, chloroquine was withdrawn from malaria treatment in most malaria endemic countries [1]. It is estimated that the loss of chloroquine to resistance was responsible for more than doubling of malaria-associated mortality in sub-Saharan Africa, a region which bears over 90% of malaria burden [2,3]. Chloroquine resistance reached fixation levels across malaria endemic countries by late 1990s [4]. As a result artemisinin agents and their derivatives were introduced in malaria treatment and have since been the first-line anti-malarial agents [3]. The use of artemisinin-based combination therapy (ACT) in malaria treatment, however, is limited by the high cost, pill burden and currently emerging risk of decreased P. falciparum parasite sensitivity [5][6][7]. Due to lack of current effective alternative agents to ACT, malaria treatment faces uncertain future which could expose populations most affected by the disease to the risk of increased malaria-associated mortality as previously seen with chloroquine [2]. Recent studies have indicated emergence of P. falciparum parasites susceptible to chloroquine following cessation of its use [4,8]. However, considerations to re-introduce chloroquine in malaria treatment is faced with the challenge of inadequacy of information on current extent of chloroquine sensitivity and the uncertainty over how this might affect future resistance. The current review was intended to collate evidence and provide current evidence on genotypic and phenotypic chloroquine resistance among P. falciparum parasites in malaria affected countries. Protocol development A systematic review protocol was developed following STREGA [9] and PRISMA-P [10] guidelines. The protocol was registered in International Prospective Register of Systematic Reviews, PROSPERO (#CRD42018083957, http://www.crd.york.ac.uk/prosp ero) and published in a peer-reviewed journal [11]. Review question The review sought to establish the prevalence of Pfcrt K76T and Pfmdr1 N86Y alleles among Plasmodium falciparum parasites in malaria affected countries since official discontinuation of chloroquine use in malaria treatment. Search strategy Electronic search Electronic search for Pubmed data base is reported in the published protocol [11]. The search terms were combined using Boolean logic 'OR' for synonymous terms and ' AND' across elements of PECOS (Population, Exposure, Comparison, Outcome and Study design). Additional searches The reference lists of included articles were screened and for any reference that could potentially be eligible for inclusion in the review, a full text article was retrieved. In addition, authors of included articles were contacted for any relevant publications on the review topic but did obtain any response. Data management All article citations retrieved from database searches were exported into EndNote software version X7 (Thomson Reuters, 2015) and duplicates removed. The articles were grouped into relevant categories as indicated in the PRISMA flow diagram (Fig. 1). Eligibility criteria The titles and abstracts were screened using a priori criteria [11]. Articles that reported on at least one of the chloroquine resistance alleles, Pfcrt K76T or Pfmdr1-N86Y were considered for inclusion. Articles that reported prevalence of the above chloroquine resistance alleles and genotyped more than 300 P. falciparum DNA samples were included. Studies that reported multiple parasite resistance genotype infections among patients were included in the review. Studies that assessed prevalence of resistance alleles (Pfcrt K76T or Pfmdr1-N86Y) following cessation of chloroquine use in malaria treatment were included. The review included studies that assessed prevalence of chloroquine resistance alleles both before and after official discontinuation of chloroquine use in malaria treatment [11]. Exclusion criteria for ineligible studies For exclusion, the following articles were considered for exclusion; focused on other drugs/insecticides, other malaria parasites, reviews, non-molecular studies, nonhuman studies, method development, non-Pfcrt/Pfmdr-1 genotypes, epidemiological studies, In vitro efficacy studies only without genotypic analysis, genetic studies, those that did not report any single nucleotide polymorphisms (SNPs). The articles were also considered for exclusion if they were letters to the editor/policy/communication, year of data collection is not indicated, chloroquine still being officially used after change in malaria treatment policy, data collected from countries that have not officially stopped the use of chloroquine in malaria treatment, and studies that screened few P. falciparum DNA (< 300 samples). Citations whose full text articles could not be retrieved (02) were considered for exclusion ( Fig. 1). Minimizing bias in article identification, selection and data extraction A second librarian (RS), validated electronic search in PubMed by performing an independent and duplicate search. A second reviewer (EAO) screened all full text articles excluded by the first reviewer (MO). Any discrepancies among the reviewers were resolved by discussion and consensus. Two reviewers (NL and BB) performed duplicate and independent data extraction. Any disagreement between the reviewers was resolved by discussion and consensus. Any further disagreement between the two reviewers was referred to the third reviewer for a final decision (MO). Data extraction Data extraction form was developed in Excel spread sheet 2007, pre-tested on 5 articles and adjusted as appropriate. The following data was extracted from included articles, author, citation, country where study was done, study design, method of sample collection, years covered by data collection, year when chloroquine use was officially stopped, duration (years) since discontinuation of chloroquine use, whether chloroquine is still being used in the study area, In vitro assay (IC 50 ), genotyping success rate, DNA extraction method, laboratory where genotyping was done, prevalence of Pfcrt 76T before and after cessation in chloroquine use, prevalence of Pfmdr-1 86Y before and after discontinuation of chloroquine use, trends in prevalence of chloroquine resistance alleles (Pfcrt 76T, Pfmdr-1 86Y), nature of malaria transmission, prevalence of mixed genotype infections, prevalence of other mutations (184F, 1246D, 1042D), and factors associated with chloroquine malaria treatment outcomes [11]. Data synthesis of included studies Descriptive data synthesis was conducted with findings from a single primary study being the unit of analysis. Sub-groups were created, region (West Africa, East Africa, Southern Africa, and Asia), study-design (cross sectional, cohort, RCTs and non-randomized clinical trials). Extracted data was analyzed using Open Meta Analyst software [12]. Descriptive summaries of study outcomes were generated including; year of official cessation of chloroquine use, chloroquine in vitro IC 50 , duration from discontinuation of chloroquine use to data collection, chloroquine resistance alleles, factors associated with chloroquine resistance, genotyping success rate, allele calling algorithms, chloroquine use, and trends in genotypic chloroquine resistance. DerSimonian-Laird (DL) random effects analysis was performed to establish a summary estimate of prevalence of P. falciparum Pfcrt 76T and Pfmdr1 86Y resistance alleles in malaria affected regions using Open Meta Analyst software. Sub-group analysis was performed, region (East Africa, Southern Africa, West Africa, Asia) and study design (cross sectional, Non-randomized clinical trials and RCTs). Heterogeneity in included articles was inferred from the summary estimates of I 2 -statistic. The I 2 statistic was used to indicate percentage (%) heterogeneity that could be attributed to between-study variance. Interpretation: I 2 = 25% (small heterogeneity), I 2 = 50% (moderate heterogeneity), I 2 = 75% (large heterogeneity) [13]. Due to high level of heterogeneity in the included studies the authors were unable to perform quantitative data analysis. Missing data and risk of bias assessment The variables that were missing from included articles were recorded as not reported. No statistical test was applied in handling missing data. However, available information was used in recalculating some variables in addition to contacting authors. The risk of publication bias was assessed using indirect assessment of rank correlation between effect size and sample size (Kendall's tau) method [14]. In this method correlation of the articles is interpreted from the analysis output in Open Meta (Analyst) software where 1, represent perfect correlation and 0 no correlation. Description of included studies Article search yielded a total of 4721 citations (Pubmed = 361, Lilacs/Vhl = 28, Science Direct = 944, Scopus = 3388). Additional searches resulted in 03 articles. After removing duplicates, 4198 articles remained and their titles and abstracts were screened for inclusion using a priori criteria. A total of 3, 834 articles were excluded after title and abstract screening. Full text of the remaining 364 articles were obtained and screened using a set criteria. Two (02) citations were excluded as their full text could not be obtained. A total of 326 articles were excluded as they did not meet the a priori inclusion criteria. Review data was extracted from a total of 38 articles that met the a priori inclusion criteria ( Fig. 1). Genotyping errors and parasite DNA extraction methods For Pfcrt 76T allele, genotyping was done in a total of 28,820 DNA samples with 18,343 being successfully genotyped, 63.6% (18,343/28,820). While genotyping was done for Pfmdr-1 86Y mutation in a total of 17,365 DNA samples with 16,232 being successfully genotyped, 93.5% (16,232/17,365). One study by Baraka et al. [15] done in Bukina-Faso did not report DNA extraction method used. The other 37 articles reported using either of the following DNA extraction methods, Chelex-100, Quiagen blood mini kit, Change in malaria treatment policy from chloroquine to sulfadoxine-pyrimethamine Majority of countries officially stopped chloroquine use in malaria treatment in 2004 (Range 1982-2008). India was the first country to stop chloroquine use in 1982. Guinea-Bissau in West Africa continued using chloroquine in malaria treatment until 2008. The average duration of time since official change in malaria treatment policy to data collection of the individual included studies is 8.7 ± 7.4 (95% CI 6.1-11.3) ( Table 1). Five studies, 13.2% (5/38) done in Zambia, Gabon, Rwanda, and India reported elimination of chloroquine use after official discontinuation. In Zambia [38], there was no detectable chloroquine resistance alleles 7 years after cessation of chloroquine use. In Gabon, there was no detectable chloroquine use 1 year after official discontinuation of chloroquine use and prevalence of Pfcrt 76T was 95.2% and Pfmdr-1 86Y, 66.5% [39]. Fourteen years after change in malaria treatment policy in Rwanda chloroquine use was stopped and prevalence of resistance alleles was, Pfcrt 76T, 39.2% and Pfmdr-1 86Y, 3.7% [40]. In India, there was no chloroquine use for over two decades after change in policy and resistance allele frequency was Pfcrt 76T, 77.6% and Pfmdr-1 86Y, 59.5% [28,41] (Table 1). Multiple chloroquine resistance surveys were done in eight countries (Kenya, Tanzania, Mozambique, Burkina Faso, Ghana, India, Gabon and Senegal). In all the eight countries chloroquine resistance allele prevalence decreased with increasing duration of time since discontinuation of chloroquine use in malaria treatment. None of the eight countries recorded zero prevalence of Pfcrt 76T allele after discontinuation of chloroquine use. Only Mozambique reported zero prevalence of Pfmdr1-86Y allele [17] 4 years after cessation of chloroquine use. The lowest reported Pfcrt 76T resistance allele prevalence, 2.3% was observed in Mozambique [42] 13 years after discontinuation of chloroquine use in malaria treatment. In India, a study by Mittra [43] reported the highest prevalence, 84.2% of Pfcrt 76T resistance allele after over a decade after discontinuation of chloroquine use in malaria treatment. The Phenotypic chloroquine resistance among P. falciparum after cessation of chloroquine use Five studies reported phenotypic parasite resistance to chloroquine after official discontinuation of use in malaria treatment. In Madagascar, one (01) year after change of policy, chloroquine IC 50 was 18.7 nM (95% CI 14.7-23.7 nM) [37]. India that stopped chloroquine use over two decades (26 years) [22]. A study in Southern Thailand [33], 14 years after change in policy reported high average chloroquine IC 50 value, 129.2 ± 45.2 nM. A study by Mwai et al. [21] in Kenya showed that P. falciparum parasites carrying Pfcrt 76T mutations had chloroquine IC 50 Hemming-Schroeder et al. [20] of 63 ± 90 nM while those carrying Pfmdr1 86Y mutation had IC 50 of 68 ± 87 nM (Table 3). Factors associated with chloroquine treatment outcomes in malaria patients A study in Madagascar [37] showed that patients with P. falciparum parasites having a Pfmdr-1 86Y mutation (OR = 4.6, 95% CI 2.3 to 8.9), and age (OR = 1.2, 95% CI 1.1 to 1.3) were predictors of chloroquine treatment failure among malaria patients. Patients with P. falciparum parasites having Pfcrt 76T are more likely to fail on treatment Malaysia, [36]. A study in Ghana [26] showed that being infected with parasites carrying Pfcrt 76T mutation was associated with staying in a place where CQ is sold and used for malaria treatment (P < 0.0001). In India (Purulia), chloroquine treatment failure was strongly associated with P. falciparum parasites carrying Pfmdr1 86Y+ 1246Y mutation in 2008-2009. Lucchi et al. indicated that P. falciparum parasites with Pfcrt 76T and Pfmdr-1 86Y mutation had significantly higher IC 50 values compared to wild-type in Kenya [22], and in Senegal [30]. A study by Afoakwah et al. [27] (Iran) showed that patients infected with P. falciparum parasites having Pfcrt 76T mutation had higher parasite density with mean density of 73,529 parasites/µl of blood compared to wild type infections. Duration since official cessation of chloroquine use and prevalence of resistance alleles Thirty-two (32) of the 38 included studies (84.2%) reported time duration (years) since discontinuation of chloroquine use to when data collection was done. Of which the average time (years) since cessation of chloroquine use in malaria treatment up to data collection date was 8.7 ± 7.4 (95% CI 6.1-11.3) years. India stopped use of chloroquine in malaria treatment for more than two decades (26 years) prior to data collection in all included studies. Fifteen studies were conducted in regions/countries less than 5 years after discontinuation of chloroquine use in malaria treatment and reported average prevalence, 63.5% of Pfcrt 76T resistance allele. Twelve studies done in the same regions and time-period reported average prevalence, 52.2% of Pfmdr1 86Y. Twelve studies were conducted in countries between 6 and 10 years after cessation of chloroquine use and reported average prevalence, 49.1% of Pfcrt 76T resistance alleles. Seven studies conducted in the same countries and time-period reported average prevalence, 37.4% of Pfmdr1-86Y. Eight studies were conducted in countries after more than 10 years since cessation in use of chloroquine in malaria treatment and reported average prevalence of resistance alleles, Pfcrt 76T (56.5%) and Pfmdr1 86Y (38.8%). A study by Mungthin et al. [33] conducted in Southern Thailand 14 years since cessation of chloroquine use reported 100% prevalence of Pfcrt 76T resistance alleles. In Gabon, Frank et al. [29] reported high prevalence, 99.2% of Pfcrt 76T resistance alleles 8 years after change in chloroquine policy in malaria treatment. In Benin, Ogouyemi-Hounto et al. [23] reported 93.9% prevalence of Pfcrt 76T 7 years after cessation of chloroquine use. In Liberia, 93.5% prevalence of Pfcrt 76T resistance alleles was reported in a study done 5 years after cessation of chloroquine use in malaria treatment [47]. Discussion Chloroquine resistance was first detected in Thailand in Southeast Asia [48] and spread to other malaria affected regions [49,50]. Due to high levels of resistance, chloroquine was withdrawn from malaria treatment globally except in Central America where parasites remain susceptible [1]. The review observed progressive decline in prevalence of Pfcrt 76T and Pfmdr-1 86Y resistance alleles following official discontinuation of chloroquine use across malaria endemic countries. This finding is similar to previous review which indicated a reduction in prevalence of chloroquine resistance alleles, Pfcrt 76T and Pfmdr-1 86Y since discontinuation of chloroquine use [51]. The current review further highlights presence of phenotypic P. falciparum chloroquine sensitivity. In Madagascar, a study by Andriantsoanirina et al. [37] reported chloroquine IC 50 , 18.7 nM (sensitive), India, [28], 146.5 nM, 162.25 nM (highly resistant), Kenya, < 25 nM (sensitive), Southern Thailand, [33], 129.2 nM (highly resistant) (highly resistant (IC 50 > 101 nM), moderately resistant (IC 50 30.9 nM < IC 50 < 100.9 nM) and sensitive (IC 50 < 30.9 nM) [52]. The observed re-emergence of chloroquine susceptibility is timely especially due to recent emergence of P. falciparum parasites with decreased artemisinin susceptibility in Southeast Asia [6,7]. Chloroquine sensitive P. falciparum parasites that have recently emerged especially in Africa following discontinuation of chloroquine use have evolved independently from African parasites [53]. Immune individuals in high malaria transmission settings serve as reservoirs [54] and are likely to be the source of current observed chloroquine susceptible parasite strains that have emerged following cessation of chloroquine use [53]. A previous review by Venkatesan et al. [51] showed continued prevalence of Pfcrt 76T alleles in malaria endemic regions, East Africa (67.6%), and West Africa (73.3%) following discontinuation of chloroquine use. The current review observed a further decline in Pfcrt 76T prevalence rates from those reported by Venkatesan et al. [51], East Africa (48.9%, 2528/5242), Southern Africa (18.6%, 373/2163), West Africa (58.3%, 3321/6608) and Asia (80.2%, 1951/2436). A similar trend in Pfmdr1 86Y mutation was seen by Ventaketsan et al. [51] in different regions of Africa, East Africa (61.1%), and West Africa (48.7%). A recent review by Okell et al. [55] also indicated a reduction in prevalence of Pfmdr1 86Y gene following change in malaria treatment policy. The current review observed lower average prevalence rate of Pfmdr-1 86Y mutation, East Africa (32.4%, 1447/5722), Southern Africa (12.7%, 544/1640), West Africa (52.2%, 1986/4200), and Asia (22.3%, 1276/2217). Despite an overall decline in frequency, chloroquine resistance alleles have persisted in malaria endemic countries following official change in policy more than a decade ago. Persistence of chloroquine resistance alleles in the population could be due to development of compensatory mutations in P. falciparum parasites that restore parasite fitness in the absence of drug pressure [56]. This incomplete reversal of resistance is a drawback to efforts considering potential re-emerging role of chloroquine in malaria control and elimination efforts. Chloroquine resistance alleles Pfcrt 76T and Pfmdr-1 86Y are selected for by the current artemisinin-based combinations used in malaria treatment [57,58]. The wide spread use of artemisinin agents in malaria treatment across malaria endemic regions could be contributing to the observed persistence of chloroquine resistance alleles in the population. Duraisingh et al. [59] indicated that increased prevalence of Pfmdr-1 86 N alleles was associated with decreased P. falciparum parasite sensitivity to lumefantrine and mefloquine. The persistence of P. falciparum parasites carrying Pfcrt 76T and Pfmdr-1 86Y mutations could indicate the high efficacy of artemisinin and partner drug (lumefantrine) especially among African parasites [51,60]. The review observed varying extents of decline in prevalence of Pfcrt 76T and Pfmdr-1 86Y chloroquine resistance alleles across malaria affected countries following discontinuation of chloroquine use. In addition, chloroquine resistance allele frequencies prior to policy change varied across malaria endemic countries. The drivers for observed variations in rates of re-emergence of chloroquine sensitive P. falciparum genotype are not clear. However, continued chloroquine use after official discontinuation, parasite biology, fixation of chloroquine resistance alleles prior to change in malaria policy could be having a role. A study by Laufer et al. [53] indicated that the re-emerging susceptible P. falciparum parasite strains are an expansion of susceptible parasite population that survived during the period when chloroquine was still being used for malaria treatment. Therefore, the observed variation in prevalence of chloroquine resistance alleles could also be due to differences in baseline levels of Pfcrt 76T and Pfmdr-1 86Y alleles in the population at the time of discontinuation of chloroquine use across malaria endemic regions. Drug-susceptible organisms may regain predominance as long as there is a population that survives due to exposure to non-lethal drug concentration despite prolonged drug pressure. The review showed that the rate of decline of prevalence of Pfcrt 76T and Pfmdr-1 86Y resistance alleles do not correspond to the duration of time since cessation in chloroquine use in most malaria endemic countries. Previous studies have shown that the extent of resistance corresponds to the prevailing drug pressure [61]. Continued use of chloroquine was observed in majority of malaria affected regions. This may be indicative of the challenges in implementation of malaria treatment policies across malaria affected countries [5]. This is especially the case in sub-Saharan Africa where chloroquine remained accessible to the general population through unofficial channels. In some countries, chloroquine continued to be imported for use in other indications such as P. vivax infections, and inflammatory conditions such as arthritis [5]. The ease of access of medicines over the counter in most low and middle income countries provides avenue for continued access and use of chloroquine in malaria treatment despite official discontinuation of its use [62,63]. The low price coupled with ease of taking chloroquine compared to current artemisinin-based combinations, further influences continued non-prescribed use of chloroquine in malaria treatment [63]. Resistance selection in an area is an inevitable consequence of antimicrobial use especially when resistant organisms to the same medicine have already emerged in other areas [51,61,64]. It follows that reduction in extent of antimicrobial use would result in emergence of more susceptible micro-organisms [61]. This is exemplified by chloroquine in which withdrawal from malaria treatment resulted in the return of susceptible parasites [4,65]. This pattern of re-emergence of chloroquine susceptible P. falciparum parasites across malaria affected regions has been observed in the current review. However, complete reversal to chloroquine sensitive parasites was not observed. This is likely due to continued drug pressure as chloroquine and related drugs are still accessed and used in the population [53]. The continued use of amodiaquine, a partner drug in artemisinin-based combinations for example has been shown to select for resistance to chloroquine [66]. Ursing et al. showed that even in presence of resistance, a change in dosing regimen restores chloroquine efficacy among P. falciparum parasites [67]. The study indicated that in vivo chloroquine failure rate can be decreased by giving the drug twice per day instead of once daily. Furthermore, doubling of chloroquine dose achieved 95% cure despite pre-existing parasite resistance with no observed increase in adverse events. In Guinea-Bissau, Ursing et al. [68] showed that doubling chloroquine dose helped stabilize spread of chloroquine resistance allele Pfcrt 76T. Pharmacodynamics modelling revealed that higher doses of chloroquine can eliminate resistant P. falciparum parasites [69]. Ginsburg [70] suggested that there is a finite concentration of chloroquine above which resistance will not develop. The hypothesized explanation for this suggestion is that the energy consumption required by the parasite to survive in presence of high chloroquine concentrations causes decrease in general fitness such that the parasites will not prevail [71]. A study by Nosten et al. [72] in Thailand also showed that if parasites develop resistance to a drug, combining that drug with an appropriate partner agent may in some cases effectively reduce drug pressure and allow for re-emergence of susceptible forms of the parasite. This helps inform current efforts considering reintroduction of chloroquine in malaria treatment. From recent evidence chloroquine can only be considered for re-introduction in malaria treatment in combination with other agents [73]. While a partner drug is sought to be combined with chloroquine, considerations should be made on the half-lives of potential agents with aim of having agents with similar half-lives to chloroquine [74,75]. These observations are important in guiding efforts re-evaluating the potential emerging role of chloroquine in malaria treatment. The review had some limitations, some of the articles did not report on all the review variables. Some variables were re-calculated from the data provided in the articles in addition to contacting authors to obtain more information and what was not established was reported as missing. Majority of the articles were from sub-Saharan African region however, since a wide article search criteria that was set a priori was applied, this is unlikely to affect the outcomes of the review. Conclusion There is a general uneven distribution of decline without complete disappearance of chloroquine resistance alleles Pfcrt 76T and Pfmdr-1 86Y across malaria endemic regions following official discontinuation of chloroquine use. Implementation of complete withdrawal of chloroquine following policy change was not achieved in most countries thus its continued access and use could have contributed to persistence of resistance alleles in the population. Evidence from the current review of prevalence of chloroquine resistance coupled with presence of highly efficacious artemisinin agents does not support re-introduction of chloroquine in malaria treatment in areas where chloroquine was previously used. Abbreviations PRISMA: Preferred Reporting Items for Systematic reviews and Meta-Analysis; STROBE: Strengthening Reporting of Observational studies in Epidemiology;	2019-03-15T14:12:14.899Z	2019-03-12T00:00:00.000	{ "year": 2019, "sha1": "1aeef8433473b07965f6218db445635045e4aea0", "oa_license": "CCBY", "oa_url": "https://malariajournal.biomedcentral.com/track/pdf/10.1186/s12936-019-2716-z", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "430da4ac0a2f9411f67450e2964865cd189327bf", "s2fieldsofstudy": [ "Environmental Science", "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
6794135	pes2o/s2orc	v3-fos-license	Ghrelin and eating behavior: evidence and insights from genetically-modified mouse models Ghrelin is an octanoylated peptide hormone, produced by endocrine cells of the stomach, which acts in the brain to increase food intake and body weight. Our understanding of the mechanisms underlying ghrelin's effects on eating behaviors has been greatly improved by the generation and study of several genetically manipulated mouse models. These models include mice overexpressing ghrelin and also mice with genetic deletion of ghrelin, the ghrelin receptor [the growth hormone secretagogue receptor (GHSR)] or the enzyme that post-translationally modifies ghrelin [ghrelin O-acyltransferase (GOAT)]. In addition, a GHSR-null mouse model in which GHSR transcription is globally blocked but can be cell-specifically reactivated in a Cre recombinase-mediated fashion has been generated. Here, we summarize findings obtained with these genetically manipulated mice, with the aim to highlight the significance of the ghrelin system in the regulation of both homeostatic and hedonic eating, including that occurring in the setting of chronic psychosocial stress. INTRODUCTION Ghrelin is a 28-amino acid octanoylated peptide hormone secreted predominantly from a distinct population of endocrine "ghrelin" cells located within the gastric oxyntic mucosa, previously referred to as P/D1-type cells in humans and A liketype cells in rodents (Kojima and Kangawa, 2010). Ghrelin is the only known naturally occurring peptide to be modified by Ser 3 O-octanoylation, a reaction catalyzed by the enzyme ghrelin O-acyl transferase (GOAT) (Yang et al., 2008;Kojima and Kangawa, 2010). This post-translational modification is essential for the binding of ghrelin to and subsequently activation of its specific receptor, the growth hormone secretagogue receptor type 1a (GHSR). GHSR is a 7 transmembrane protein coupled to the Gq subfamily of heterotrimeric GTP-binding proteins that activates phospholipase C and in turn, induces release of calcium from intracellular stores (Howard et al., 1996;Damian et al., 2012). Stimulation of other signal transduction cascades also have been demonstrated upon activation of GHSR, although much of that data was characterized using transfected cell systems (Cruz and Smith, 2008). GHSR is expressed in several regions of the central nervous system and some peripheral organs (Cruz and Smith, 2008). Ghrelin's actions are diverse. Ghrelin is a potent growth hormone (GH) secretagogue and orexigenic peptide (Kojima and Kangawa, 2010). Ghrelin also prevents falls in blood glucose upon extreme caloric restriction and minimizes depressive-like behavior upon chronic psychosocial stress (Lutter et al., 2008;Goldstein et al., 2011). Moreover, ghrelin stimulates motor activity in the gastrointestinal tract and accelerates gastric emptying (Peeters, 2003). Although desacyl-ghrelin, the non-octanoylated form, does not bind to GHSR, some studies report that it can modify certain physiological responses, including food intake and glucose homeostasis (Delhanty et al., 2012). Genetically-modified mouse models have been instrumental in dissecting out the physiological roles of ghrelin, as well as the specific sites of ghrelin action. Genetic modifications in mice can be grouped in two classes: (1) Standard transgenic models, in which a foreign DNA construct is randomly integrated into the mouse genome (commonly used for overexpression models) and (2) Gene-targeted transgenic models, in which a gene disruption or manipulation is introduced in a single gene via homologous recombination between the foreign DNA and the endogenous gene (commonly used for "knockout" and "knock-in" models) (Davey and Maclean, 2006). In this review, we discuss ghrelin's action on eating behaviors as determined using mice with genetic manipulations in the ghrelin signaling system (Tables 1 and 2). We first focus on the role of ghrelin in homeostatic eating and subsequently discuss the role of ghrelin in hedonic eating. GHRELIN AND HOMEOSTATIC EATING Homeostatic eating is defined as food intake driven by necessity, due to energy deficiency as perceived by the brain and body. Ghrelin is the only known circulating peptide hormone Table 1 \| Studies of homeostatic eating behaviors in mouse models with genetic alterations in ghrelin-signaling system. Model Summary References Ghrelin overexpression (driven by various promoters) Elevated desacyl-ghrelin levels with conflicting food intake and body weight results Ariyasu et al., 2005;Asakawa et al., 2005;Iwakura et al., 2005;Wei et al., 2006;Zhang et al., 2008 Elevated ghrelin levels with no change in food intake (Reed et al.) or increase in food intake (Bewick et al.) Reed et al., 2008;Bewick et al., 2009 Ghrelin overexpression (SV40 T-antigen) Increased ghrelin levels, with food intake and body weights indistinguishable from wild-type littermates; Older transgenic mice lose weight, likely due to large gastric tumors Iwakura et al., 2009;Zhao et al., 2010b Ghrelin analog overexpression Increase of ghrelin-like activity, no change in food intake, body weight, energy expenditure or fat mass Matsumoto et al., 2001;Yamada et al., 2010;Ariyasu et al., 2012 Human ghrelin and GOAT overexpression Elevated human ghrelin only when fed diet containing medium-chain triglycerides. No change in food intake but increase in body weight Kirchner et al., 2009 Ghrelin knockout No changes in eating behaviors under CHD. Inconsistent results in HFD feeding studies Sun et al., 2003;Wortley et al., 2004Wortley et al., , 2005De Smet et al., 2006;Dezaki et al., 2006;Pfluger et al., 2008;Sato et al., 2008 GHSR knockout Attenuated anticipatory locomotor response. Similar food intake, but subtle decrease in body weight under CHD feeding Sun et al., 2004Sun et al., , 2008Abizaid et al., 2006;Blum et al., 2009;Lin et al., 2011;Ma et al., 2011 GHSR null (removable transcriptional blocking cassette) Female mice on CHD weigh less than wild-type controls. Consume less food and are resistant to HFD-induced body weight gain upon early HFD exposure Zigman et al., 2005; Ghrelin and GHSR knockout Lower body weight, but no change in food intake Pfluger et al., 2008 Site-selective GHSR Expression (GHSR null crossed with cell-specific Cre) TH-promoter driven Cre: partially restores ghrelin-stimulated food intake Phox2b-promoter driven Cre: lack ghrelin-induced food intake Chuang et al., 2011;Scott et al., 2012 GHSR overexpression in GHRH neurons Increase in post-weaning growth rate. Increase in growth rate not maintained in adulthood. No change in body weight on HFD, although fat pad mass trended lower. Possible increase in HFD intake Lall et al., 2004 GOAT deficiency Normal body weight and fat mass under CHD feeding. Increase in food intake but lower body weights under medium-chain triglyceride rich diet Gutierrez et al., 2008;Kirchner et al., 2009;Zhao et al., 2010a Ghrelin knockout (Ob/Ob background) Similar food intake and a modest increase in body weight as compared to ob/ob mice Sun et al., 2006 GHSR knockout (Ob/Ob background) Similar food intake and body weights to ob/ob mice Ma et al., 2012 Table 2 \| Studies of hedonic eating behaviors in mouse models with genetic alterations in ghrelin-signaling system. Model References GHSR knockout Suppressed intake of rewarding food in a free choice (CHD/rewarding food) paradigm Blum et al., 2009;Egecioglu et al., 2010;Disse et al., 2011;Verhagen et al., 2011;Lamont et al., 2012 GHSR null Lack cue potentiated feeding behavior. Do demonstrate CPP for HFD post-CSDS. Fail to increase intake of CHD in response to chronic stress Perello et al., 2010;Chuang et al., 2011;Walker et al., 2012 Site-selective GHSR expression (GHSR null crossed with cell-specific Cre) TH-promoter driven Cre: sufficient to restore CSDS-induced CPP for HFD Chuang et al., 2011 GOAT deficiency Attenuated motivation for food in an operant responding model. Decreased hedonic feeding response as examined in a "dessert effect" protocol Davis et al., 2012 that stimulates food intake (Cummings, 2006). Ghrelin triggers eating even at times of minimal spontaneous food intake, and its orexigenic actions are rapid, short-lived and potent (Nakazato et al., 2001;Wren et al., 2001;Cummings, 2006). It has been suggested that ghrelin is a physiologic meal initiator, since its levels increase preprandially and decrease post-prandially (Cummings et al., 2001). Ghrelin levels, in individuals under restricted feeding schedules, increase approximately 1 h prior to food presentation. In animals, this preprandial rise in ghrelin occurs simultaneously or possibly precedes the increased locomotion in anticipation of a scheduled meal (Drazen et al., 2006). Indeed, ghrelin administration increases locomotor activity in rats and foraging-like activities in hamsters (Keen-Rhinehart and Bartness, 2005;Jerlhag et al., 2007). Thus, Blum et al. hypothesized that ghrelin is secreted in anticipation of a meal and correlates with anticipatory locomotor activity (Blum et al., 2009). The mechanism for ghrelin's action on stimulating food intake was initially thought to be through homeostatic hypothalamic circuits originating in the arcuate nucleus (ARC) (Willesen et al., 1999;Nakazato et al., 2001;Kageyama et al., 2010). The expression of GHSR, the only known receptor for ghrelin, has been confirmed within the ARC by many techniques, including in situ hybridization histochemistry, RT-PCR, immunohistochemistry and Western blot analysis (Howard et al., 1996;Guan et al., 1997;Willesen et al., 1999;Zigman et al., 2006). Several studies have demonstrated that ghrelin has the capacity to bind its receptor on arcuate NPY/AgRP neurons, resulting in their activation and ensuing orexigenic behaviors (Willesen et al., 1999;Nakazato et al., 2001;Wren et al., 2001;Kageyama et al., 2010;Schaeffer et al., 2013). Interaction of ghrelin with several other neuronal subtypes and brain sites also has been shown to increase intake of freely-available food. These include, among others, orexin-producing neurons of the lateral hypothalamus (Olszewski et al., 2003;Toshinai et al., 2003). Additionally, some evidence indicates that vagus nerve integrity is required for ghrelin-induced food intake as well (Date et al., 2002;Date, 2012). HOMEOSTATIC EATING IN MICE OVER EXPRESSING GHRELIN In order to confirm the effects of ghrelin on homeostatic eating, several groups have attempted to produce transgenic mice with chronically elevated ghrelin. These initial transgenes contained the ghrelin gene driven by various promoters-including those for CAG (Ariyasu et al., 2005;Asakawa et al., 2005), rat insulin II (Iwakura et al., 2005), rat glucagon (Iwakura et al., 2005), CMV (Wei et al., 2006) and FABP4 (Zhang et al., 2008)resulting in elevated ghrelin gene expression mainly in cell types in which ghrelin gene expression does not usually occur. These mouse models resulted in elevated desacyl-ghrelin rather than elevations of ghrelin, presumably because this overexpression of the ghrelin gene was targeted mostly to tissues lacking the ghrelinacylating enzyme, GOAT. Indeed, GOAT was not discovered until 2008. Most mouse models overproducing desacyl-ghrelin exhibited similar food intake and body weight as compared to wildtype mice (Iwakura et al., 2005;Wei et al., 2006); nonetheless, some studies found reduced body weight, fat mass, nose-to-anus length and/or food intake in their transgenic mice overexpressing desacyl-ghrelin (Ariyasu et al., 2005;Asakawa et al., 2005;Zhang et al., 2008). These findings may suggest that desacyl-ghrelin has metabolic effects opposite to those of ghrelin. Two groups successfully produced standard transgenic mice with ghrelin overexpression. Reed et al. generated a transgenic mouse model overexpressing ghrelin in neurons by using a neuron-specific enolase promoter. These mice showed increased ghrelin expression in brain and increased circulating ghrelin (∼5fold), although they did not differ from wild-type controls in body weight, food intake, energy expenditure or fat mass (Reed et al., 2008). Likewise, Bewick et al. generated a transgenic mouse model overexpressing ghrelin, using a bacterial artificial chromosome containing the ghrelin gene and its promoter (Bewick et al., 2009). These mice exhibited increased circulating ghrelin (∼1.5fold) and increased food intake without long-term body weight gain, perhaps due to a paradoxical increase in energy expenditure (Bewick et al., 2009). Bewick et al. proposed that the differences between their model and the Reed et al. mouse model could be due to differential peripheral vs. central nervous system ghrelin concentrations (Bewick et al., 2009). Of note, results regarding the existence, sites of production and physiological significance of endogenous central ghrelin production in non-geneticallymanipulated models have been inconsistent (Furness et al., 2011), although ghrelin from the periphery does make its way to the central nervous system. After GOAT was identified, Kirchner et al. generated a doubletransgenic mouse model overexpressing both human ghrelin and MBOAT4 (the membrane-bound O-acyltransferase domain containing 4 or human variant of GOAT) genes in the liver via the human apo-lipoprotein E promoter (Kirchner et al., 2009). These transgenic mice, which exhibited elevated concentrations of human ghrelin only when fed on a diet containing mediumchain triglycerides, showed decreased energy expenditure and increased rate of body weight gain without changes in food intake (Kirchner et al., 2009). Other unique techniques have been employed to increase ghrelin expression. Yamada and colleagues generated transgenic mice overexpressing a ghrelin analog, which replaced Ser 3 with Trp 3 , under the control of human serum amyloid-P promoter (Yamada et al., 2010). Trp 3 -ghrelin possesses lower levels of ghrelin-like activity without the need for post-transcriptional modification (Matsumoto et al., 2001). Although mice overexpressing Trp 3 -ghrelin displayed a ∼6-fold increase of ghrelin-like activity, neither body weight, food intake, energy expenditure or fat mass differed from wild-type controls (Ariyasu et al., 2012). Our group generated transgenic mice that express the immortalizing SV40 large T-antigen in ghrelin cells (Zhao et al., 2010b). These mice develop ghrelin-secreting tumors, which result in ∼25-fold higher plasma ghrelin concentrations and ∼37-fold higher plasma desacyl-ghrelin levels by the time they reach 20 weeks of age (Zhao et al., 2010b). Despite the sustained increase in plasma ghrelin levels, food intake and body weights in these transgenic mice were indistinguishable from those in wild-type mice; older transgenic mice lost weight, presumably due to the development of large gastric tumors (Zhao et al., 2010b). Iwakura et al. also created ghrelin promoter SV40 large T-antigen transgenic mice (Iwakura et al., 2009). These mice showed no difference in body weights as compared to wild-type mice before 12 weeks of age, after which body weight and food intake declined, likely due to tumor growth (Iwakura et al., 2009). HOMEOSTATIC EATING IN MICE WITH GHRELIN DEFICIENCY To our knowledge, four different ghrelin knockout mouse models have been reported (Sun et al., 2003;Wortley et al., 2004;De Smet et al., 2006;Dezaki et al., 2006). Regardless of the complete absence of ghrelin in circulation in these models, under ad libitum standard rodent chow diet (CHD) feeding, food intake and body weights in these ghrelin knockout mice were indistinguishable from those in wild-type mice. In addition, no differences were observed when other parameters of food intake were measured, including post-fasting hyperphagia (Sun et al., 2003;Wortley et al., 2004;Pfluger et al., 2008), forced dark cycle-induced eating (De Smet et al., 2006) or eating memory (Sato et al., 2008). These findings suggest that genetic ghrelin deficiencies fail to cause significant alterations in homeostatic aspects of eating behaviors. Furthermore, three of the four ghrelin knockout models have been studied under chronic high fat diet (HFD) feeding, but yielded inconsistent results (Wortley et al., 2005;Dezaki et al., 2006;Sun et al., 2008). Surprisingly, none of the studies exhibited the expected reduction of HFD intake. Wortley et al. did demonstrate that ghrelin deficiency resulted in reduced body weight and fat mass, but these results were not recapitulated in the other mouse ghrelin knockout mouse models. Differences in the age at which the mice were first exposed to HFD diet are thought to account for the inconsistencies in the body weight phenotype among the different ghrelin knockout models. In particular, resistance to the development of diet-induced obesity may manifest in ghrelin knockout mice upon early exposure to HFD, but not when HFD challenge is initiated later in life (Wortley et al., 2004(Wortley et al., , 2005Sun et al., 2008). HOMEOSTATIC EATING IN MICE WITH GHSR DEFICIENCY To our knowledge, three different mouse models with GHSR deficiency have been reported in the literature (Sun et al., 2004;Zigman et al., 2005;Abizaid et al., 2006). Two are traditional GHSR knockout models, in which the GHSR gene has been removed (Sun et al., 2004;Abizaid et al., 2006), and the other is a GHSR-null model in which the GHSR locus has been modified by the insertion of a loxP-flanked transcriptional blocking cassette (Zigman et al., 2005). More specifically, this transcriptional blocking cassette sequence disrupts GHSR gene expression and, as a consequence, the GHSR-null mice (harboring 2 copies of the recombinant GHSR allele) have no detectable functional GHSR expression. Cre recombinase mediates reactivation of GHSR gene expression by removal of the transcriptional blocking cassette sequence. GHSR-deficient mice show a subtle but significant decrease in body weight. Sun et al. reported that GHSR knockout mice fed on CHD show a lower body weight than wild-type mice from 16 to 24 weeks of age (Sun et al., 2004). GHSR-deficient mice do not differ from wild-type controls in food intake under CHD (Sun et al., 2004;Zigman et al., 2005). Sun et al. found that GHSR knockout mice have a similar susceptibility to dietinduced obesity compared to wild-type mice upon exposure to HFD as adults (Sun et al., 2008); however, GHSR knockout mice manifest reduced age-associated obesity mainly due to reduced adiposity and increased thermogenesis Ma et al., 2011). Using a restrictive eating paradigm, Abizaid et al. showed that GHSR knockout mice did not increase their food intake in response to repeated fasts as seen in wild-type mice. This was also true in ghrelin knockout mice (Abizaid et al., 2006). In the GHSR-null mice from our group, female GHSR-null mice on CHD were significantly lighter than wild-type controls starting at 12 weeks of age (Zigman et al., 2005). Also, in two independent studies, GHSR-null mice were resistant to HFD-induced body weight gain upon early HFD exposure (Zigman et al., 2005;. As previously mentioned, in animals, a preprandial rise in ghrelin occurs simultaneously or possibly precedes the increased locomotion exhibited in anticipation of a scheduled meal (Drazen et al., 2006). In preliminary studies, Gooley et al. examined food-entrainable circadian rhythms in a very small cohort of female GHSR-null mice (Gooley et al., 2006). Similar to wild-type littermates, the GHSR-null mice continued to show a preprandial rise in locomotor activity and body temperature, suggesting that intact ghrelin signaling was not required for the expression of food-entrainable circadian rhythms that are likely related to finding food (Gooley et al., 2006). Furthermore, Blum et al. also demonstrated increased activity prior to scheduled mealtime in both wild-type and GHSR knockout mice, although in their study, the response in GHSR knockout mice was attenuated (Blum et al., 2009). Others have also similarly reported that GHSR knockout mice under a restricted feeding paradigm demonstrate attenuated anticipatory locomotor responses and reduced expression of the marker of cellular activation c-fos in the mesolimbic dopamine pathway (Blum et al., 2009;Lamont et al., 2012). Using a different paradigm, Verhagen et al. have shown that GHSR knockout mice did not anticipate food when exposed to an activity-based anorexia model, in which mice are given free access to a running wheel and fed once per day for 2 h (Verhagen et al., 2011). Thus, most of these studies suggest that a disruption in the ghrelin signaling results in the lack of food anticipation behavior. HOMEOSTATIC EATING IN MICE WITH BOTH GHRELIN AND GHSR DEFICIENCY In another study, significant differences in body weight (decreased), energy expenditure (increased) and locomotor activity (increased) were observed when both ghrelin and GHSR were deficient in mice (double knockout mice), whereas neither ghrelin deficiency nor GHSR deficiency independently affected energy balance under CHD feeding conditions . Thus, the authors speculated that additional components of the ghrelin signaling system might be present. HOMEOSTATIC EATING IN MICE WITH TISSUE-SPECIFIC GHSR EXPRESSION The GHSR-null mouse model with Cre recombinase-dependent reactivation of GHSR has been valuable in establishing the physiological roles of certain of ghrelin's brain targets (Zigman et al., 2005). In one study, we have used GHSR-null mice to assess the role of direct action of ghrelin on the ventral tegmental area (VTA) and other catecholaminergic neurons. We achieved this by crossing GHSR-null mice with mice containing Cre recombinase expression driven by the tyrosine hydroxylase (TH) promoter, resulting in mice expressing GHSR selectively in TH-containing cells normally programmed to express this receptor (Chuang et al., 2011). This selective GHSR expression did not affect body weight, but partially restored ghrelin-stimulated food intake to levels achievable in wild-type littermates (Chuang et al., 2011). In another study, we crossed GHSR-null mice with mice containing expression driven by the paired-like homeobox 2b (Phox2b) promoter, resulting in mice expressing GHSR selectively in specific hindbrain nuclei, including the nucleus of the solitary tract, dorsomotor nucleus of the vagus, area postrema, nucleus ambiguus and facial motor nucleus . In contrast to mice with catecholaminergic neuron-selective GHSR expression, mice with hindbrain-selective GHSR expression lacked ghrelininduced acute food intake . Thus, as opposed to GHSR-expressing catecholaminergic neurons, GHSR-expressing hindbrain neurons are not sufficient to mediate the acute orexigenic effects of ghrelin. HOMEOSTATIC EATING IN MICE WITH GHSR OVER EXPRESSION IN GROWTH HORMONE RELEASING HORMONE NEURONS Lall et al. generated a mouse model overexpressing GHSR in GH releasing hormone (GHRH) neurons using a rat GHRH cosmid promoter (GHRH-GHSR) (Lall et al., 2004). GHRH-GHSR mice had an increased post-weaning growth rate compared to wildtype littermates. When placed on HFD at 2 months of age, weight gain of GHRH-GHSR mice was similar to that of their wild-type counterparts, although fat pad mass trended lower. Food intake was not measured for individually-housed mice, however groups of GHRH-GHSR mice ate significantly more calories when fed HFD than when fed CHD, while groups of wild-type mice ate similar caloric amounts of HFD and CHD (Lall et al., 2004). HOMEOSTATIC EATING IN MICE WITH GOAT DEFICIENCY Two GOAT knockout mouse lines have been reported (Gutierrez et al., 2008;Zhao et al., 2010a). Using one of these lines, GOAT knockout mice demonstrated higher levels of desacyl-ghrelin compared to their wild-type littermates, and under CHD feeding had normal body weight and fat mass (Kirchner et al., 2009). However, when fed HFD, these GOAT knockout mice displayed significantly lower body weights compared to wild-type animals, despite not having significant changes in body composition. Challenging GOAT knockout mice with HFD rich in mediumchain triglycerides significantly reduced body weight and fat mass compared to similarly-treated wild-type littermates (Kirchner et al., 2009). Such was likely due to increased energy expenditure, since food intake in the GOAT knockout mice was higher (Kirchner et al., 2009). In contrast, the GOAT knockout mice generated by Zhao et al. had similar body weight, body composition and food intake as wild-type littermates, both on CHD and HFD (Zhao et al., 2010a). They also demonstrated similar weight loss curves when subjected to a severe caloric restriction protocol (Zhao et al., 2010a). HOMEOSTATIC EATING IN MICE WITH DEFICIENCIES IN GHRELIN SIGNALING AND LEPTIN Ghrelin and leptin are both regulators of feeding, but have opposite effects on food intake. Unlike ghrelin, leptin increases with positive energy balance and inhibits food intake. Furthermore, leptin-and ghrelin-responsive central targets partially overlap (Zigman and Elmquist, 2003;. To investigate the physiological roles of ghrelin and leptin signaling, and their potential interaction in regulating energy balance, Sun et al. crossed ghrelin knockout mice with leptin deficient (ob/ob) mice to generate mice lacking both ghrelin and leptin (ghrelin knockout ob/ob). Although they predicted that the ghrelin knockout ob/ob mice would have a leaner and relatively hypophagic phenotype, the double knockout mice instead displayed similar food intake and a modest increase in body weight as compared to ob/ob mice (Sun et al., 2006). Ma and colleagues generated mice with GHSR deletion on the ob/ob background (GHSR knockout ob/ob) (Ma et al., 2012). They found that food intake and body weights of GHSR knockout ob/ob mice were similar to those of ob/ob mice, possibly suggesting that the obesity in ob/ob mice is unrelated to unopposed ghrelin action. GHRELIN AND HEDONIC EATING Hedonic eating is defined as food intake motivated primarily by pleasure. Hedonic eating also encompasses behaviors aimed at facilitating access to pleasurable food. Ghrelin is thought to increase the rewarding value of palatable foods. Recent studies show that ghrelin is able to regulate different aspects of hedonic eating. The two-food choice test is a conventional behavioral method for determining preference. In this test, rodents are offered a pair of different types of food simultaneously, and the consumed amount is measured for a defined time. Using this test, it has been shown that ghrelin administration shifts food preference toward diets rich in fat (Shimbara et al., 2004). Ghrelin administration also increases intake of palatable saccharin solution and preference for saccharin-flavored foods in mice (Disse et al., 2010). Similarly, rats treated with a GHSR antagonist consume less peanut butter and Ensure®, but do not change intake of CHD in a free choice protocol (Egecioglu et al., 2010). As will be discussed below, ghrelin administration also affects other models of hedonic eating, including the conditioned place preference (CPP) test and operant responding. The central distribution of GHSRs within mesolimbic centers supports the likelihood of ghrelin playing a major role in hedonic eating. In particular, GHSR is highly expressed in dopaminergic VTA neurons, which project to the nucleus accumbens (NAc) and strongly drive reward behaviors (Perello and Zigman, 2012). Mesolimbic circuitries also may be indirectly regulated by cholinergic neurons located in the laterodorsal tegmental area (LDTg), which express GHSR and innervate the VTA (Dickson et al., 2010). In addition, we have recently shown that ghrelin action on food reward requires intact orexin signaling, although the neuronal circuits linking orexin neurons and ghrelin action remain unclear (Perello et al., 2010). HEDONIC EATING IN MICE WITH GENETIC DELETION OF GHSR The use of mouse models deficient in GHSR have been instrumental in demonstrating an obligatory role for intact ghrelin signaling in various hedonic aspects of eating. Egecioglu et al. showed that GHSR knockout mice have a suppressed intake of the rewarding food in a free choice (CHD/rewarding food) paradigm and a reduced accumbal dopamine release induced by rewarding foods (Egecioglu et al., 2010). We recently subjected GHSR-null mice to a cue-potentiated feeding protocol, in which cue-potentiated feeding was demonstrated by enhanced intake of grain-based pellets selectively upon presentation of a positive conditioned stimulus (Walker et al., 2012). While wild-type mice demonstrated cue-potentiated feeding behavior, GHSR-null mice lacked this behavior. More specifically, GHSR-null mice displayed enhanced food intake in response to both positive and negative conditioned stimuli (Walker et al., 2012). This study supports pharmacologic studies suggesting a key role for intact ghrelin signaling pathways in establishing a specific positive cue-food association (Kanoski et al., 2013). Altered eating behaviors are observed under stress as well as depression. In order to study the role of ghrelin in stressassociated hedonic eating, our group subjected both GHSR-null and wild-type littermates to a mouse behavioral model of chronic psychosocial stress and major depression, known as chronic social defeat stress (CSDS). Mice were challenged with ten daily bouts of social defeat by aggressive CD1 male mice, a protocol which results in an increase in plasma ghrelin (Lutter et al., 2008). Despite an increase in plasma ghrelin levels, GHSR-null mice failed to increase intake of CHD in response to chronic stress, unlike wild-type controls which increased their food intake. This suggests that activation of GHSR signaling is required for a stress-induced increase in caloric intake. In addition, stressed GHSR-null and wild-type mice were also tested in a food CPP test, which models a more complex, reward-related type of eating behavior. In the food CPP test, animals are conditioned to associate one chamber of the CPP apparatus with CHD and a second, visually and texturally distinct chamber with an equal calorie amount of a more pleasurable food, such as HFD. After conditioning, animals are given free access to both chambers in the absence of the food, and CPP is demonstrated when more time is spent in the chamber previously paired to the food reward. The CSDS-CPP protocol seems to model stress-based comfort food eating. Following exposure to the CSDS protocol, GHSRnull mice did not demonstrate CPP for HFD, whereas wild-type mice did (Chuang et al., 2011). The above findings in GHSR-null mice confirm pharmacological studies; as administration of ghrelin and other means of achieving increases in plasma ghrelin (caloric restriction) both enable acquisition of food CPP (Perello et al., 2010;Disse et al., 2011), while administration of GHSR antagonists blocks food CPP (Egecioglu et al., 2010;Perello et al., 2010). Likewise, in our studies using mice subjected to a caloric restriction in which mice were limited to 80% of their usual daily food intake, GHSRnull mice did not demonstrate CPP for HFD, whereas wild-type mice did (Perello et al., 2010). Interestingly, when the study was repeated in mice exposed to a more intense calorie restriction protocol, in which mice were limited to ad libitum access to CHD for 4 h per day, no differences between genotypes were detected in the food CPP test. Thus, endogenous increases in GHSR signaling following stress as modeled with the CSDS protocol or under mild food restrictive conditions enable acquisition of CPP for HFD (Perello et al., 2010). HEDONIC EATING IN MICE WITH TISSUE-SPECIFIC REACTIVATION OF GHSR We also assessed food CPP performance following CSDS or as induced by administered ghrelin in mice with catecholaminergic selective GHSR expression, using the mice described above in section Homeostatic Eating in Mice with Tissue Specific GHSR Expression (Chuang et al., 2011). Such selective GHSR expression in TH-expressing cells was sufficient to mediate the ability of administered ghrelin or chronic stress to induce CPP for HFD. HEDONIC EATING IN MICE WITH GOAT DEFICIENCY Recently, it was determined that GOAT-mediated acylation of ghrelin is a critical modulator of food reward. Following a 24-h food deprivation period, GOAT-deficient mice displayed an attenuated motivation for food in an operant responding model (Davis et al., 2012). This supports several pharmacologic studies in which ghrelin administration increased operant lever pressing for sucrose, peanut butter-flavored sucrose or HFD pellets in rodents, and in which GHSR antagonist reduced operant responding for sucrose solution (Perello et al., 2010;Finger et al., 2012;Landgren et al., 2012;Skibicka et al., 2012). GOAT knockout mice also showed a decreased hedonic feeding response as examined in a "dessert effect" protocol, wherein the intake of a palatable HFD "dessert" was assessed in calorically-sated mice (Davis et al., 2012). Furthermore, GOAT knockout mice displayed atypical transcriptional changes in orexin 1 receptor gene expression in the NAc, potentially implying the involvement of orexin signaling in these effects of ghrelin (Davis et al., 2012). Taken together, these results suggest that the motivation to obtain food is regulated by endogenous GOAT activity. DISCUSSION In this review, we have summarized the findings of ghrelin action on eating behavior as determined or confirmed using different mouse models with genetic manipulations of the ghrelin signaling system. Of note, the various transgenic mouse models more consistently demonstrated alterations in hedonic eating, while changes to homeostatic eating and body weight changes were more variable and subtle. An important consideration when evaluating the eatingrelated phenotypes of the various transgenic models relates to differences between manipulations of ghrelin vs. GHSR vs. GOAT. For example, variable degrees of altered homeostatic eating, hedonic eating and body weight have been observed in ghrelindeficient, GHSR-deficient, and GOAT-deficient models. Although the exact reasons for these differences are unclear, they could result from differences inherent to each of those molecules. For instance, GHSR possesses strong ligand-independent constitutive activity when studied in vitro , and thus phenotypic changes in GHSR-null mice could result from the combined loss of ghrelin binding and of baseline GHSR "tone." Additionally, it has been proposed that ghrelin has other receptors besides GHSR, which could explain why simultaneous deletion of both ghrelin and GHSR result in a more severe phenotype than deletion of either alone Pfluger et al., 2008). Furthermore, GOAT may have other substrates besides ghrelin, and thus its loss could directly affect more than just ghrelin bioactivity. Another concern when evaluating these mouse models, as with any transgenic mouse models, is the possibility of functional redundancy or developmental compensation by another related system, which may mask the true effect of manipulation of the ghrelin system. While this has not yet been demonstrated specifically for the ghrelin system, it has been demonstrated for AgRP neurons that are thought to be crucial for ghrelin's effects on eating. Importantly, unlike the lack of significant body weight or food intake phenotypes in mice with genetic AgRP ablation achieved since inception or in neonates, rapid starvation ensues when AgRP expression is deleted in adult mice (Qian et al., 2002;Gropp et al., 2005;Luquet et al., 2005). Such a time-dependent effect of genetic AgRP deletion may be influenced by a time-sensitive capacity for the development of compensatory neurocircuits (Bouret et al., 2004;Luquet et al., 2005). Indeed, leptin's neurotrophic effects on hypothalamic development are restricted to a critical neonatal period (Bouret et al., 2004). Similar time-sensitive neurotrophic actions have been predicted for ghrelin (Steculorum and Bouret, 2011). Further considerations when evaluating the effects of genetic engineering-in particular, those changes induced in standard transgenic models in which the transgene is randomly inserted into the genome-include the potential for the site of integration to affect tissue specificity, levels of transgene expression as well as expression of other genes present in the nearby genomic vicinity. For all these reasons, the use of the multiple mouse models generated using different genetic strategies and comparison of these mice to mice treated pharmacologically with agents targeting the ghrelin system is important for best determining the importance of the ghrelin system in eating behavior It should also be noted that these various geneticallyengineered mouse models targeting the ghrelin system additionally have been used to assess the importance of ghrelin action in domains other than those related to eating and body weight. For instance, GHSR-null mice have been used to demonstrate a key role for ghrelin action in minimizing depressive-like behavior following chronic stress (Lutter et al., 2008). Also, ghrelinoverexpressing, ghrelin-knockout, GHSR-knockout, GHSR-null and GOAT-knockout mice all demonstrate alterations in glucose homeostasis, with the most extreme phenotype-namely marked hypoglycemia and near death-appearing upon severe caloric restriction of mice with genetic deletion of either GOAT or ghrelin (Zhao et al., 2010a;Li et al., 2012). These studies predict eating-and body weight-related phenotypes for people if they were to carry a genetic mutation engendering either ghrelin and/or GOAT overexpression or ghrelin, GHSR or GOAT deletion. Indeed, the extreme hyperphagia, early-onset obesity and debilitating food seeking behaviors associated with Prader-Willi Syndrome, in which ghrelin levels are found to elevated and abnormally regulated, represent much more extreme phenotypes than the mouse models of ghrelin over expression would predict (Feigerlova et al., 2008;Hinton et al., 2010). As of yet, no definitive examples of ghrelin, GHSR or GOAT deletion have been identified in people. However, other types of mutations-in the ghrelin gene and the GHSR gene-have been identified. One, the Leu72Met ghrelin polymorphism has been associated with binge eating disorder in small cohorts (Monteleone et al., 2007). Several GHSR polymorphisms have been found in individuals with short stature, which makes sense given ghrelin's ability to potently stimulate GH secretion. Using in vitro methods, these mutations result in loss of endogenous constitutive activity of GHSR, altered binding of ghrelin, decreased ghrelin-stimulated signal transduction, and/or decrease cell surface expression of GHSR Pantel et al., 2006;Inoue et al., 2011). Interestingly, carriers of some of these GHSR mutations demonstrate an incomplete penetrance of overweight and obesity, although the relationship of the mutation and the body weight phenotypes have not been confirmed . Given the crucial insight regarding ghrelin action on feeding and other domains provided by these mouse models, we would encourage further investigations using not only the currently available mouse models, but also novel genetically modified mouse models. For instance, a GHSR-conditional knockout mouse in which GHSR is eliminated exclusively from a single cell type or a single brain region by using the Cre-loxP technology is an exciting alternative. Additionally, the Cre-loxP system could be used to investigate the role of GHSR at specific stages of development by breeding GHSR floxed mouse lines with transgenic mice bearing a tamoxifen-dependent Cre recombinase expressed under the control of specific promoters, thus allowing the manipulation of GHSR to be controlled in a spatiotemporal-specific manner. In addition, mouse models containing some of the above-described ghrelin and GHSR polymorphisms described in individuals with binge eating disorder and families with short stature ± overweight/obesity would help clarify and causative role for the mutations in those conditions.	2016-05-04T20:20:58.661Z	2013-05-15T00:00:00.000	{ "year": 2013, "sha1": "4859567496b05f426efc1aa2c1b55fa28308d14a", "oa_license": "CCBY", "oa_url": "https://www.frontiersin.org/articles/10.3389/fnins.2013.00121/pdf", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "4859567496b05f426efc1aa2c1b55fa28308d14a", "s2fieldsofstudy": [ "Biology", "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
258947412	pes2o/s2orc	v3-fos-license	RankCSE: Unsupervised Sentence Representations Learning via Learning to Rank Unsupervised sentence representation learning is one of the fundamental problems in natural language processing with various downstream applications. Recently, contrastive learning has been widely adopted which derives high-quality sentence representations by pulling similar semantics closer and pushing dissimilar ones away. However, these methods fail to capture the fine-grained ranking information among the sentences, where each sentence is only treated as either positive or negative. In many real-world scenarios, one needs to distinguish and rank the sentences based on their similarities to a query sentence, e.g., very relevant, moderate relevant, less relevant, irrelevant, etc. In this paper, we propose a novel approach, RankCSE, for unsupervised sentence representation learning, which incorporates ranking consistency and ranking distillation with contrastive learning into a unified framework. In particular, we learn semantically discriminative sentence representations by simultaneously ensuring ranking consistency between two representations with different dropout masks, and distilling listwise ranking knowledge from the teacher. An extensive set of experiments are conducted on both semantic textual similarity (STS) and transfer (TR) tasks. Experimental results demonstrate the superior performance of our approach over several state-of-the-art baselines. Introduction Sentence representation learning refers to the task of encoding sentences into fixed-dimensional em- * Work done during internship at Meituan. † † beddings. The sentence embeddings can be leveraged in various applications, including information retrieval (Le and Mikolov, 2014), text clustering (Ma et al., 2016) and semantic textual similarity comparison (Agirre et al., 2012). With the recent success of pre-trained language models (PLMs), such as BERT/RoBERTa (Devlin et al., 2019;Liu et al., 2019), a straightforward way to generate sentence representations is to directly use the [CLS] token embedding or the average token embeddings from the last layer of PLMs (Reimers and Gurevych, 2019). However, several studies (Ethayarajh, 2019;Li et al., 2020) have found that the native sentence representations derived by PLMs It can be seen that RankCSE captures more fine-grained ranking information than SimCSE (Gao et al., 2021) and DiffCSE (Chuang et al., 2022). occupy a narrow cone in the vector space, and thus severely limits their representation capabilities, which is known as the anisotropy problem. Supervised methods like SBERT (Reimers and Gurevych, 2019) usually generate better sentence representations, but require fine-tuning on a large amount of labeled data. Recent unsupervised models (Carlsson et al., 2021;Giorgi et al., 2021) adopt contrastive learning framework without any labels, which pulls similar semantics closer and pushes dissimilar ones away. These methods usually design different augmentation algorithms for generating positive examples, such as back-translation , dropout (Gao et al., 2021) and token shuffling or cutoff . In-batch negatives are further combined with the positives. Despite achieving promising results, they treat positives/negatives equally without capturing the fine-grained semantic ranking information, resulting in less effective sentence representations which fail to distinguish between very similar and less similar sentences. For example, Table 1 shows two examples of a query sentence and several target sentences from semantic textual similarity datasets. It is clear that the similarity scores produced by the contrastive learning method SimCSE are not optimized, where the sentence rankings are not preserved in the learned representations. On the other hand, our RankCSE generates effective sentence representations with consistent rankings to the ground-truth labels. Figure 1 further shows the advantage of RankCSE in terms of two ranking metrics. The fine-grained ranking information is crucial in various real-world applications including search and recommendation. The ability to differentiate between subtle distinctions in sentence meaning can help these systems provide more relevant and accurate results, leading to a better user experience. Therefore, it is an important problem to learn ranking preserving sentence representations from unsupervised data. To obtain semantically discriminative sentence representations, we propose a novel approach, RankCSE, which incorporates ranking consistency and ranking distillation with contrastive learning into a unified framework. Specifically, our model ensures ranking consistency between two representations with different dropout masks and minimizes the Jensen-Shannon (JS) divergence as the learning objective. In the meanwhile, our model also distills listwise ranking knowledge from the teacher model to the learned sentence representations. In our work, we explore two listwise ranking methods, ListNet (Cao et al., 2007) and ListMLE (Xia et al., 2008), and utilize the pre-trained SimCSE (Gao et al., 2021) models with coarse-grained semantic ranking information as the teachers to provide pseudo ranking labels. Our RankCSE is able to generalize fine-grained ranking information from the weak ranking knowledge learned by SimCSE. We conduct an extensive set of experiments on semantic textual similarity (STS) and transfer (TR) tasks. Experimental results show that RankCSE outperforms the existing state-of-the-art baselines. Related Work Unsupervised Sentence Representation Learning Early works typically augment the idea of word2vec (Mikolov et al., 2013) to learn sentence representations, including Skip- Thought (Kiros et al., 2015), FastSent (Hill et al., 2016) and Quick-Thought (Logeswaran and Lee, 2018). With the great success of PLMs, various attempts focus on generating sentence representations by leveraging the embedding of [CLS] token or applying mean pooling on the last layer of BERT (Reimers and Gurevych, 2019). However, Ethayarajh (2019) identifies the anisotropy problem in language representations, which means the native learned embeddings from PLMs occupy a narrow cone in the vector space. BERT-flow (Li et al., 2020) and BERTwhitening (Su et al., 2021) propose to resolve the anisotropy problem through post-processing. Recently, contrastive learning has been adopted to learn sentence representations by designing dif- : Encoder S2 S1 S3 S2 S1 softmax Figure 2: The framework of RankCSE which consists of three components: (1) contrastive learning object; (2) ranking consistency loss which ensures ranking consistency between two representations with different dropout masks; (3) ranking distillation loss which distills listwise ranking knowledge from the teacher. ferent augmentation methods (Zhang et al., 2020;Carlsson et al., 2021;Giorgi et al., 2021;Yan et al., 2021;Kim et al., 2021;Gao et al., 2021). A typical example SimCSE uses dropout as a data augmentation strategy and is also the foundation of many following works. ArcCSE enhances the pairwise discriminative power and models the entailment relation among triplet sentences. DCLR (Zhou et al., 2022) alleviates the influence of improper negatives. DiffCSE (Chuang et al., 2022) introduces equivariant contrastive learning to Sim-CSE. PCL (Wu et al., 2022a) proposes contrastive representation learning with diverse augmentation strategies for an inherent anti-bias ability. InfoCSE (Wu et al., 2022b) learns sentence representations with the ability to reconstruct the original sentence fragments. Generative learning techniques Wu and Zhao, 2022) have also been proposed to enhance the linguistic interpretability of sentence representations. Although achieving promising results, these methods fail to capture the fine-grained ranking knowledge among the sentences. Learning to Rank Given a query example, learning to rank aims to rank a list of examples according to their similarities with the query. Learning to rank methods can be divided into three categories: pointwise (Li et al., 2007), pairwise (Burges et al., 2005(Burges et al., , 2006 and listwise (Cao et al., 2007;Xia et al., 2008;Volkovs and Zemel, 2009;Pobrotyn and Bialobrzeski, 2021). Pointwise methods optimize the similarity between the query and each example, while pairwise approaches learn to cor-rectly model the preference between two examples. Listwise methods directly evaluate the ranking of a list of examples based on the ground truth. In our framework, we leverage listwise ranking objectives for learning effective sentence representations, which have shown better performance compared to pointwise and pairwise methods. Preliminary We provide some conceptual explanations and definitions in learning to rank. Top One Probability Given the scores of all objects S = {s i } n i=1 , the top one probability of an object is the probability of its being ranked at top-1: s i = e s i /τ n j=1 e s j /τ where τ is a temperature hyperparameter, usually utilized to smooth the distribution. We simply denote the formulation for calculating the top one distribution based on the scores S as: denote a permutation of the object indexes, which represents that the π(i)-th sample is ranked i-th. The probability of a specific permutation π is given as: P (π\|S, τ ) = n i=1 e s π(i) /τ n j=i e s π(j) /τ . Problem Formulation Our goal is to learn sentence representations such that semantic similar sentences stay close while dissimilar ones should be far away in an unsupervised manner. Specifically, We aim to find an optimal function f that maps a sentence s ∈ p s to a ddimensional vector f (s) ∈ p e ⊆ R d , where p s and p e denote the distributions of sentences and sentence representations, respectively. Supposing s 1 and s 2 are more semantic similar than s 1 and s 3 (s 1 , s 2 , s 3 ∈ p s ), a good mapping function f should satisfy that the distance between f (s 1 ) and f (s 2 ) is smaller than that between f (s 1 ) and f (s 3 ), i.e., d(f (s 1 ), f (s 2 )) < d(f (s 1 ), f (s 3 )), where d is the distance metric such as Euclidean distance and cosine distance. In this way, the similarities among the sentences are preserved in the learned sentence representations. The general idea of RankCSE is to learn semantically discriminative sentence representations by capturing the ranking information among the sentences. As shown in Figure 2, our model consists of three components: (1) contrastive learning objective (Section 4.2); (2) ranking consistency loss which ensures ranking consistency between two representations with different dropout masks (Section 4.3); (3) ranking distillation loss which distills listwise ranking knowledge from the teacher (Section 4.4). Contrastive Learning Contrastive learning aims to learn effective representations by pulling similar semantics closer and pushing away dissimilar ones. SimCSE (Gao et al., 2021) creates positive examples by applying different dropout masks and takes a cross-entropy object with in-batch negatives (Chen et al., 2017). More specifically, for any sentence x i in a min-batch, we send it to the encoder f (·) twice and obtain two representations with different dropout masks f (x i ), f (x i ) ′ . SimCSE use the InfoNCE loss (van den Oord et al., 2018) as the training objective: where N is the batch size, τ 1 is a temperature hyperparameter and ϕ(f is the cosine similarity used in this work. Essentially, the contrastive learning objective is equivalent to maximizing the top one probability of the positive sample. Although contrastive learning is effective in separating positive sentences with negative ones, it ignores the continuity modeling of the similarity. In other words, it is not effective in distinguishing highly similar sentences with moderate similar ones. To address this issue, we propose to directly model the ranking information among the sentences, which could enhance the discrimination of semantic similarity in the learned sentence representations. Ranking Consistency The main drawback of contrastive learning is that the distinction between the in-batch negatives is not modeled, resulting in less effective sentence representations in capturing the fine-grained sentence similarity. Therefore, instead of treating the negatives equivalently, we propose to explicitly model the ranking information within the sentences by ensuring the ranking consistency between the two similarity sets (circled by the solid and dashed curves respectively in the right part of Figure 2). Concretely, by taking a close look at the contrastive modeling in Section 4.2, there are two sets of sentence representations, f (x i ) and f (x i ) ′ , derived from different dropout masks. For each sentence x i , two lists of similarities with other sentences can be naturally obtained from the two representations, i.e., S( We then enforce the ranking consistency between these two similarity lists in our modeling. Intuitively, all corresponding elements in S(x i ) and S(x i ) ′ should have the same ranking positions. Given two similarity lists S(x i ) and S(x i ) ′ , we can obtain their top one probability distributions The ranking consistency can be ensured by minimizing the Jensen-Shannon (JS) divergence between the two top one probability distributions: where P i and Q i represents S τ 1 (x i ) and S τ 1 (x i ) ′ respectively. The reason we choose JS divergence instead of Kullback-Leibler (KL) divergence is that the two distributions are symmetric rather than one side being the ground truth. Ranking Distillation Contrastive learning based methods like SimCSE learn effective sentence representations with coarsegrained semantic ranking information (shown in Appendix F and G), which have demonstrated their effectiveness in various downstream tasks. Orthogonal to ranking consistency, we further introduce ranking distillation by distilling the ranking knowledge from pre-trained teacher models into our learned sentence representations, to generalize effective ranking information from the weak ranking knowledge learned in the teachers. More specifically, for each sentence in a min-batch, we obtain the similarity score list from the teacher model, which is then served as pseudo ranking labels in the ranking distillation. The intuitive idea is to transfer the ranking knowledge from the teacher to the student as guidance for learning ranking preserved sentence representations. In the ranking distillation, ListNet (Cao et al., 2007) and ListMLE (Xia et al., 2008) methods are utilized. Formally they are defined as: where S(x i ) and S T (x i ) are the similarity score lists obtained from the student and the teacher, respectively, rank(·, ·) is the listwise method. ListNet The original ListNet minimizes the cross entropy between the permutation probability distribution and the ground truth as the training objective. However, the computations will be intractable when the number of examples n is large, since the number of permutations is n!. To reduce the computation complexity, the top one probability distribution is usually adopted as a substitute: where τ 2 and τ 3 are temperature hyperparameters. 1 ListMLE Different from ListNet, ListMLE aims to maximize the likelihood of the ground truth permutation π T i which represents the sorted indexes of the similarity scores calculated by the teacher model. The objective of ListMLE is defined as: In this work, we propose to use a multi-teacher from which more listwise ranking knowledge can be transferred and preserved. In our experiments, we utilize the weighted average similarity scores of two teachers as pseudo ranking labels: where α is a hyperparameter to balance the weight of the teachers. The contrastive learning loss L infoNCE pushes apart the representations of different sentences to maximize the representation space, while the ranking consistency loss L consistency and the ranking distillation loss L rank pull similar negatives closer, thus capturing fine-grained semantic ranking information. Combining the above three loss functions, we can obtain the overall objective: where β and γ are hyperparameters to balance different losses. For fair comparison, we use the same 10 6 randomly sampled sentences from English Wikipedia provided by SimCSE. Following previous works, we start from pre-trained checkpoints of BERT (Devlin et al., 2019) Table 2: Sentence representations performance on STS tasks (Spearman's correlation). We directly import the results from the original papers and mark the best (bold) and second-best (underlined) results among models with the same PLMs. Results are statistically significant with respect to all baselines on each PLM (all p-value < 0.005). BERT large , SimCSE-RoBERTa base and SimCSE-RoBERTa large . We utilize the first two as a multiteacher for RankCSE-BERT base and RankCSE-BERT large , while the last two for RankCSE-RoBERTa base and RankCSE-RoBERTa large . We evaluate our model every 125 training steps on the dev set of STS-B and keep the best checkpoint for the evaluation on test sets of all STS and TR tasks. More training details can be found in Appendix A. We compare RankCSE with several unsupervised sentence representation learning methods, including average GloVe embeddings (Pennington et al., 2014), USE (Cer et al., 2018) and Skipthought (Kiros et al., 2015), average BERT embeddings from the last layer, post-processing methods such as BERT-flow (Li et al., 2020) and BERTwhitening (Su et al., 2021), and contrastive learning methods such as IS-BERT (Zhang et al., 2020) and ConSERT . We also include recent strong unsupervised sentence representation baselines, including SimCSE (Gao et al., 2021), DCLR (Zhou et al., 2022), ArcCSE , DiffCSE (Chuang et al., 2022), PaSER (Wu and Zhao, 2022) and PCL (Wu et al., 2022a). Since RankCSE and the teacher model SimCSE are using the same unsupervised training data, the comparison between RankCSE and baselines is fair. representations by incorporating ranking consistency and ranking distillation. We also observe that the performances of RankCSE listNet and RankCSE listMLE are very consistent across all datasets, which demonstrates the effectiveness of both listwise ranking methods. Results on STS Tasks As shown in Results on TR Tasks It can be seen in Table 3 that RankCSE achieves the best performance among all the compared baselines on all PLMs. Note that for DiffCSE, we obtain the results from the publicly available code and checkpoints, because DiffCSE uses different dev sets to find the best hyperparameters for TR tasks than other baselines. More detailed explanation and comprehensive comparison are provided in Appendix B. Another observation is that the performance of the RankCSE listNet is slightly better than that of the RankCSE listMLE . Our hypothesis is that the inaccurate pseudo ranking labels introduce more errors in the calculation of the permutation probability than the top one probability. Nevertheless, both listwise methods achieve better results than the baselines, which is consistent with the results in Table 2. Figure 3: Effect of the temperatures τ 2 and τ 3 . Results are average STS performance, and RankCSE listNet is based on BERT base while RankCSE listMLE is based on different PLMs. We do not demonstrate results below 78 to make the variation obvious. Analysis and Discussion Ablation Study To investigate the impact of different losses in our approach, we conduct a set of ablation studies by removing L infoNCE , L consistency and L rank from Eq.(6). The average results on STS and TR tasks are reported in Table 4. There are several observations from the results. First, when L rank is removed, the performance significantly drops in both STS and TR tasks, which indicates the effectiveness of L rank in our modeling. Second, it is also clear that without L infoNCE or L consistency , the model performance also decreases, especially on TR tasks. Thirdly, it is worth mentioning that RankCSE with only L rank can also outperform the teachers on STS tasks. The reason is that RankCSE is able to preserve ranking knowledge from multiple teachers, and generalize fine-grained ranking information from multiple coarse-grained representations. Fourthly, since L consistency does not explicitly distinguish the positives from negatives, RankCSE with only L consistency will preserve inaccurate rankings leading to significant performance drop. Finally, RankCSE with all components achieves the best performance on both STS and TR tasks. Comparisons of Different Teachers We conduct experiments to explore the impact of dif-ferent teachers on the performance of RankCSE. As shown in Table 5, RankCSE outperforms the teacher model which indicates that incorporating ranking consistency and ranking distillation leads to more semantically discriminative sentence representations. Comparing the performance of RankCSE using different teachers, we observe that better teacher leads to better RankCSE, which is consistent with our expectation since accurate ranking labels yield more effective ranking knowledge transfer. Another observation is that the performance of RankCSE with a multi-teacher is better than that with a single teacher, which verifies that RankCSE is able to preserve listwise ranking knowledge from more than one teacher. It is also interesting to see that using DiffCSE-BERT base and SimCSE-BERT large as multi-teacher leads to even higher performance than the results in Table 2. We plan to conduct more investigation along this direction to explore the upper bound of improvements. Effect of Hyperparameters To study the effect of temperature hyperparameters, we conduct experiments by setting different τ 2 and τ 3 . As shown in Figure 3a, we find that large discrepancy between τ 2 and τ 3 leads to significant drop in the performance of RankCSE ListNet . The best temperature setting for RankCSE ListNet is τ 2 : τ 3 = 2 : 1. The performance of RankCSE ListMLE has similar trends based on different PLMs, as shown in Figure 3b. For both RankCSE ListNet and RankCSE ListMLE , the temperature should be set moderate. Robustness of RankCSE We conduct 5 runs of model training with the hyperparameter settings which can be referred to Appendix A with different random seeds, and then calculate the mean and standard deviation values. The results provided in Table 6 demonstrate both the superior performance and the robustness of our model. It can also be seen that RankCSE listMLE achieves similar performance but more stable results compared with RankCSE listNet . Alignment and Uniformity Following previous works (Wang and Isola, 2020), we use alignment and uniformity to measure the quality of representation space. Alignment measures the distance between similar instances, while uniformity measures how well the representations are uniformly distributed (detailed in Appendix H). For both measures, the smaller value indicates the better result. We plot the distribution of ℓ align -ℓ uniform for different models using BERT base which are measured on the STS-B dev set. As shown in Figure 4, RankCSE effectively improves both alignment and uniformity compared with average BERT embeddings, while SimCSE and DiffCSE only improve uniformity and alignment respectively. Since RankCSE pulls similar negatives closer during incorporating ranking consistency and ranking distillation, RankCSE has smaller alignment and bigger uniformity than SimCSE. We consider that RankCSE achieves a better trade-off than Sim-CSE. When compared with DiffCSE, RankCSE has smaller uniformity whereas similar alignment. We can also observe that RankCSE outperforms PCL on both metrics. Conclusion In this work, we propose RankCSE, an unsupervised approach to learn more semantically discriminative sentence representations. The core idea of RankCSE is incorporating ranking consistency and ranking distillation with contrastive learning into a unified framework. When simultaneously ensuring ranking consistency and distilling listwise ranking knowledge from the teacher, RankCSE can learn how to make fine-grained distinctions in semantics, leading to more semantically discriminative sentence representations. Experimental results on STS and TR tasks demonstrate that RankCSE outperforms previous state-of-the-art methods. We also conduct thorough ablation study and analysis to demonstrate the effectiveness of each component and justify the inner workings of our approach. We leave what is the upper bound of improvements of the teacher for future work. Limitations In this section, we discuss the limitations of our work as follows. First, despite achieving promising results, our model needs to calculate pseudo ranking labels of the teacher which requires additional training time per epoch than the teacher. The training efficiency of RankCSE and SimCSE can be seen in Appendix D. Second, we directly use SimCSE base and SimCSE large as a multi-teacher in our implementation and experiments. However, how to choose the best combination of the teacher models is worth further exploration. It could help researchers to better understand the upper bound of improvements. We plan to investigate more along this direction in the future. A Training Details We implement all experiments with the deep learning framework PyTorch on a single NVIDIA Tesla A100 GPU (40GB memory). We carry out grid-search of learning rate ∈ {2e-5, 3e-5} and temperatures τ 2 , τ 3 ∈ {0.0125, 0.025, 0.05}, while setting batch size to 128, temperature τ 1 to 0.05, α to 1/3, β to 1, γ to 1 and the rate of linear scheduling warm-up to 0.05 for all the experiments. We train our models for 4 epochs, and evaluate the model every 125 steps on the dev set of STS-B and keep the best checkpoint for the final evaluation on test sets of all STS and TR tasks. The hyperparameter settings we adopt are shown in Table 9. Following SimCSE, we utilize the embedding corresponding to [CLS] token as the representation of the input sentence. We utilize SimCSE-BERT base and SimCSE-BERT large as a multi-teacher for RankCSE-BERT base and RankCSE-BERT large , while SimCSE-RoBERTa base and SimCSE-RoBERTa large as a multi-teacher for RankCSE-RoBERTa base and RankCSE-RoBERTa large . B DiffCSE Settings for Transfer Tasks DiffCSE uses different dev sets to find the best hyperparameters for the two tasks (STS-B dev set for STS tasks, dev sets of 7 TR tasks for TR tasks), while other methods only use the STS-B dev set for both tasks, which is not fair. Therefore we obtain the results in Table 3 from its publicly available code and checkpoints for STS tasks 2 instead of directly importing the results from its original paper. For a more comprehensive comparison with Dif-fCSE on TR tasks, we also use dev sets of 7 TR tasks to find the best hyperparameters and checkpoints. As shown in Table 10, RankCSE still outperforms DiffCSE in this setting. C Data Statistics The complete listings of train/dev/test stats of STS and TR datasets can be found in Table 7 D Training Efficiency We compare the training efficiency of SimCSE and RankCSE , which are tested on a single NVIDIA Tesla A100 GPU (40GB memory). We set batch size to 128 for both SimCSE and RankCSE, and training epoch to their original settings (1 for SimCSE, 4 for RankCSE). RankCSE utilizes SimCSE base and SimCSE large as a multi-teacher to provide pseudo ranking labels. As shown in Table 11, RankCSE base and RankCSE large can be trained within 2 hours and 3.7 hours respectively. Since RankCSE needs to calculate pseudo ranking labels of the teacher, it requires additional training time per epoch than SimCSE. E Cosine Similarity Distribution We demonstrate the distribution of cosine similarities for sentence pairs of STS-B dev set in Figure 5. We can observe that cosine similarity distributions from all models are consistent with human ratings. However, the cosine similarities of RankCSE are slightly higher than that of SimCSE under the same human rating, as RankCSE pulls similar negatives closer during incorporating ranking consistency and ranking distillation, and shows lower variance. Compared with DiffCSE, RankCSE shows a more scattered distribution. This observation further validates that RankCSE can achieve a better alignmentuniformity balance. F Case Study We present another two examples of a query sentence and several target sentences from the STS datasets, with their similarity scores and rankings in Table 12. It is obvious that the similarity scores produced by RankCSE are more effective than Sim-CSE, with consistent rankings to the ground-truth labels. It further demonstrates that SimCSE only captures coarse-grained semantic ranking information via contrastive learning, while RankCSE can capture fine-grained semantic ranking information. For example, SimCSE can distinguish between similar and dissimilar sentences, however, it can not distinguish between very similar and less similar sentences as RankCSE. G Ranking Tasks We build the ranking task based on each STS dataset to verify that RankCSE can capture finegrained semantic ranking information. For one sentence x i , if there are more than three sentence pairs (x i , x j i ) containing x i with similarity score label y j i in the dataset, we view as a sample of the ranking task, as shown in Table 12. We adopt KCC (Kendall's correlation coefficient (Abdi, 2007)) and NDCG (normalized discounted cumulative gain (Järvelin and Kekäläinen, 2002)) as evaluation metrics for ranking tasks, and demonstrate the results in Table 13. RankCSE outperforms SimCSE and DiffCSE on both KCC and NDCG, which validates that RankCSE can capture fine-grained semantic ranking information by incorporating ranking consistency and ranking distillation. Another observation is that SimCSE and Dif-fCSE also achieve moderate results, which shows they can distinguish coarse-grained semantic differences via contrastive learning. H Alignment and Uniformity Wang and Isola (2020) use two properties related to contrastive learning, alignment and uniformity, to measure the quality of representation space. Alignment calculates expected distance between normalized representations of positive pairs p pos : while uniformity measures how well the normalized representations are uniformly distributed: where p data denotes the distribution of sentence pairs. Smaller alignment means positive instances have been pulled closer, while smaller uniformity means random instances scatter on the hypersphere. These two measures are smaller the better, and well aligned with the object of contrastive learning. Table 13: Sentence representations performance on ranking tasks (KCC and NDCG) using BERT base . The results of SimCSE and DiffCSE are obtained from their publicly available codes and checkpoints. We mark the best (bold) and second-best (underlined) results. ACL 2023 Responsible NLP Checklist A For every submission: A1. Did you describe the limitations of your work? Limitations A2. Did you discuss any potential risks of your work? Sections 1 and 6 A3. Do the abstract and introduction summarize the paper's main claims? Abstract and Section 1 A4. Have you used AI writing assistants when working on this paper? Left blank. B Did you use or create scientific artifacts? Section 5 B1. Did you cite the creators of artifacts you used? Section 5 B2. Did you discuss the license or terms for use and / or distribution of any artifacts? We were unable to find the license for the dataset we used. B3. Did you discuss if your use of existing artifact(s) was consistent with their intended use, provided that it was specified? For the artifacts you create, do you specify intended use and whether that is compatible with the original access conditions (in particular, derivatives of data accessed for research purposes should not be used outside of research contexts)? Section 5 B4. Did you discuss the steps taken to check whether the data that was collected / used contains any information that names or uniquely identifies individual people or offensive content, and the steps taken to protect / anonymize it? The datasets we use are the commonly-used benchmarks. B5. Did you provide documentation of the artifacts, e.g., coverage of domains, languages, and Appendix E C1. Did you report the number of parameters in the models used, the total computational budget (e.g., GPU hours), and computing infrastructure used? Section 5 and Appendix E The Responsible NLP Checklist used at ACL 2023 is adopted from NAACL 2022, with the addition of a question on AI writing assistance. C2. Did you discuss the experimental setup, including hyperparameter search and best-found hyperparameter values? Section 5 and Appendix A C3. Did you report descriptive statistics about your results (e.g., error bars around results, summary statistics from sets of experiments), and is it transparent whether you are reporting the max, mean, etc. or just a single run? Section 5 C4. If you used existing packages (e.g., for preprocessing, for normalization, or for evaluation), did you report the implementation, model, and parameter settings used (e.g., NLTK, Spacy, ROUGE, etc.)? Section 5 D Did you use human annotators (e.g., crowdworkers) or research with human participants? Left blank. D1. Did you report the full text of instructions given to participants, including e.g., screenshots, disclaimers of any risks to participants or annotators, etc.? No response. D2. Did you report information about how you recruited (e.g., crowdsourcing platform, students) and paid participants, and discuss if such payment is adequate given the participants' demographic (e.g., country of residence)? No response. D3. Did you discuss whether and how consent was obtained from people whose data you're using/curating? For example, if you collected data via crowdsourcing, did your instructions to crowdworkers explain how the data would be used? No response. D4. Was the data collection protocol approved (or determined exempt) by an ethics review board? No response. D5. Did you report the basic demographic and geographic characteristics of the annotator population that is the source of the data? No response.	2023-05-29T01:22:23.507Z	2023-05-26T00:00:00.000	{ "year": 2023, "sha1": "a6f7a120f3bdbabe97f8563ce25bfef3fa652400", "oa_license": null, "oa_url": null, "oa_status": null, "pdf_src": "ACL", "pdf_hash": "6762b3421f9c16f7e3a051667ad180482086c785", "s2fieldsofstudy": [ "Computer Science" ], "extfieldsofstudy": [ "Computer Science" ] }
30142697	pes2o/s2orc	v3-fos-license	Superfragile glassy dynamics of onecomponent system with isotropic potential: competition of diffusion and frustration We investigate glassy dynamical properties of one component three-dimensional system of particles interacting via pair repulsive potential by the molecular dynamic simulation in the wide region of densities. The glass state is superfragile and it has high glassforming ability. The glass transition temperature Tg has pronounced minimum at densities where the frustration is maximal. The ubiquitous glass formation and jamming still puzzle physicists [1,2]. The microscopic mechanism of the drastic slowing down of the structural relaxation of a liquid on cooling is one of the central issues of the physics of the liquid-glass transition. The question "why some liquids form a glass easily but others do not," is still the matter of debates. There is a paradigm that one component liquid with isotropic potential typically spontaneously crystallizes being supercooled in (quasi)equilibrium conditions [3][4][5][6]. It is a formidable challenge to avoid, e.g., spontaneous crystallization in quasiequilibrium cooling of the onecomponent Lennard-Jones liquid. Yet it was discovered not long ago that there are some exceptions from the paradigm. The common fitch of these exceptions is the pronounced attractive well of the pair potential, see, e.g., [7,8]. Here we show using the molecular dynamics (MD) that the one component simple liquid with pure repulsive potential shows glassy behaviour in quasiequilibrium cooling. Frustration, when one cannot minimize the energy of the system by merely minimizing all local interactions, is one of the basic factors that stipulates the glass-forming ability [1,9]. E.g., it can be related with the long-range alternating interactions (e.g., in spin glasses [10]) or with geometrical reasons [11]. The potential used in Ref. [7,8] was optimized to produce icosahedral local order and so geometrical frustration. Our potential has the soft step and the simple liquid with this interaction has two characteristic scales. The competition between these scales makes the system effectively quasibinary [12]. And this is the origin of frustration in our system. Intuitively increasing frustration one should favor the formation of the glass. Here we show that on the contrary, the glass transition temperature T g may have minimum at parameters where the frustration is maximal. The second important concept of glassy physics is fragility. According to Angell-classification [2,13] the glassforming liquids effectively divide into two classes: "strong" and "fragile", where the viscosity of the liquid shows either nearly Arrhenius behaviour with temperature or much faster one. We found the fragility index and concluded that our system is superfragile [its fragility index exceeds that of the "Decalin" -one the most fragile liquid [14]], see Fig. 1. In the remainder of this paper we use the dimensionless quantities:r ≡ r/σ,Ũ = U/ε, temperatureT = T / , densityρ ≡ N σ 3 /V , and timet = t/[σ m/ε], where m and V are the molecular mass and system volume correspondingly. As we will only use these reduced variables, we omit the tildes. So we take here n = 14, k 0 = 10, and σ 1 = 1.35. These parameters values reveal complex system behavior such as phase diagram with polymorphous transitions and disordered gap, see (Color online) We show the glass transition temperature on the sketch of the phase diagram obtained in Ref. [12]. The red circles and blue triangles correspond to Tg extracted from D(T ) using the Vogel-Fulcher formula (VF) [29] and the power law from mode-coupling theory (MCT) [31] correspondingly. In the inset we show the accuracy of D(T ) approximation by VF and MCT for ρ = 0.6. The circles show the result of MD simulations. For MD simulations we have used the system of N = 5000 particles that were simulated under periodic boundary conditions in Nose-Hoover NVT ensemble. We have checked that N = 5000 is enough amount of particles to eliminate the finite size effects that agrees with Ref. 21. The MD time step was δt = 0.01. It is nearly the maximum possible time step that satisfies the energy conservation condition. The system was studied in the density region of ρ ∈ (0.35 − 0.75). At all densities of this region the system was cooled in a stepwise manner from high temperature state and completely equilibrated at each step until convergence of time dependence of mean square displacement (up to 5 × 10 7 time steps). The time dependencies of temperature, pressure and configuration energy were additionally analyzed to control equilibration. Data were subsequently collected during the time t samp that was chosen to be large enough for correct calculation of diffusion coefficients by Einstein relation. So this time was approximately equal to t samp ∼ 3τ , where τ is the time necessary to reach diffusive regime after ballistic and plateau ones. For more details see Supplementary Material. Avoiding spontaneous crystallization during equilibration process is the principal difficulty of MD simulations of glassy dynamics of liquids. For model glass-forming systems this problem is usually solved by either using of non-isotropic potentials [5,22], or considering multicomponent systems [23,24], or using nonequilibrium cooling [18,25]. For collapsing soft spheres system, it is possible to avoid crystallization in the one-component system with isotropic potential due to the quasibinary behavior that develops itself in certain density interval. In our case this range is ρ ∈ (0.51 − 0.74). In the inset in Fig. 3b we show the splitting of the first peak in the radial distribution functions (RDF) that illustrates the quasibinary behaviour of our system in hand. Outside this interval, it is hardly possible to study the glassy dynam- Intermediate scattering function for the density ρ = 0.6. Inset shows the snapshot of the typical distribution of particles for T = 0.1 and ρ = 0.6. We mark by red (blue) color particles that have a neighbor at distance of the order of the first (second) maximum in rdf [hard-core and soft-core diameters]. ics because system spontaneously crystalizes whereupon supercooling below the melting line. Conversely, inside the region mentioned, one can equilibrate supercooled liquid without crystallization down to temperatures at which relaxation time becomes too large for simulation. In our case, these minimal temperatures were chosen so that diffusion coefficients were on the order of 10 −5 . At that temperatures, the total calculation time required for equilibration and data collecting is ∼ 3 × 10 7 MD steps [three days of calculation on 32 processors in parallel]. In order to control stability of glassy state, we performed calculation at lower temperatures also (down to T g ). At that temperatures, the system cannot be equilibrated completely because of large relaxation times and so we did not use these data for T g calculations. But what we have observed is the absence of any crystallization up to 10 8 MD steps and so we conclude that the glass state is stable (at least at simulation time scales). To acquire information about the glass state we focus on the temperature range T T g where the "glassforming fluctuations" [27] slow down the system dynamics with temperature decreasing. The conventional correla-tion function tools have been used: the mean square displacement ∆r 2 (t) and the intermediate scattering function F s (q, t). The time dependencies of these functions for ρ = 0.6 and different temperatures are shown in Fig. 4. One can see the typical picture glassformers demonstrate at low temperature [28]. Namely, the "plateau" reflecting the cage effect (when the particle is trapped in the "cage" of the nearest neighbours) appears on ∆r 2 (t) and F s (q, t) at sufficiently low temperatures that indicates the onset of glassy regime in system dynamics. Meanwhile, the system remains in disordered state as can be seen from radial distribution functions, see inset of Fig. 3b. We note the splitting of the first and second peaks of the radial distribution function (Fig. 3b). The splitting of the first peak reflects the quasibinarity of the system caused by the form of the potential (see discussion below). While the splitting of the second peak is apparently the (system independent) attribute of glassy state [7,18,26]. In order to estimate the glass transition temperature, we calculated diffusion coefficient D at each of the temperature and density investigated. The temperature dependencies D(T ) were approximated by both the Vogel-Fulcher formula (VF) [29,30] and the power law from mode-coupling theory (MCT) [30,31] The parameters D , γ were obtained using method of nonlinear least squares. Since the expressions (2), (3) are correct only in a vicinity of glass transition temperature, the temperature interval T ∈ (T min , T max ) for least squares approximation was chosen so that D(T min ) 10 −5 , D(T max ) 10 −3 that approximately corresponds T /T g ∈ (1.15, 1.6). The typical temperature dependence D(T ) of diffusion coefficient obtained from simulation and its VF and MCT approximations are shown in the inset of Fig. 2. One can see that both formulas provide good fitting of simulation data. Having the parameters of (2), (3) one can get the glass transition temperature. According to generally accepted definition, glass transition occurs at D 0 /D = 10 n , where 13 n 17. Using (2), (3) we obtain The graph of T g (ρ) is shown in Fig. 2 on the phase diagram (previously obtained in [12]). It follows that T g (ρ) dependencies obtained using VF and MCT approaches are in good agreement with each other. Particularly, the The equation (7) particularly shows that the fragility index has minimal value m min = n corresponding to the limit T (vf) 0 /A → 0 that turns (2) to Arrhenius law. Fig. 1 shows the density dependence of the reduced fragility, m/m min , of our system in comparison to several glassformers. One can see that the dynamics of our system is extremely fragile since the mean value of its fragility exceeds the one for decalin -one of the most fragile system [14]. Note that m(ρ) is nonmonotonic and reveals clear minimum at ρ = 0.6 as well as for T g (ρ), see Fig. 2. Recently it was shown that the increase of the interaction softness can lead to the increase of the fragility [34,35]. The "softness" of our pair potential is large at r corresponding to the peaks of the radial distribution function, see Fig. 5. Away from ρ * = 0.6 one of the RDF peaks dominates, see inset in Fig. 3b, that effectively makes softer the effective interparticle interaction and helps to interpret the nonmonotonic behaviour of m(ρ). Advances over the last decade have linked non-Arrhenius behavior of fragile glass formers with the presence of locally heterogeneous dynamics: i.e. the presence of distinct (if transient) slow and fast regions within the material [2,36]. The quasibinary character of our particle system allows to make the selection among the particles: we mark by red (blue) color particles that have a neighbor at distance of the order of the first (second) maximum in rdf [hard-core and soft-core diameters], see Fig.5. The snapshot of the spatial particle distribution over the simulation volume (see inset in Fig. 4 and Fig. 2 in the Supplementary) shows the high degree of the heterogeneity in our system that favors the existence of the locally heterogeneous dynamics within the system because of different free volumes (and so diffusion coefficients) of particles with different effective radii. It has been mentioned above that the system demonstrates quasibinary behavior due to the repulsive shoulder of the pair potential, see inset in Fig. 3b. It reflects the competition between hard-core and soft-core scales and, as a result, the frustration in the system. As the quasibinarity index we choose where ρ 1 = and r 2 are minima of the first and the second RDF peaks. So ρ 1 , ρ 2 are local densities in the vicinity of the peaks and F is their inverse symmetrized ratio. The origin of the frustration in our system is the same as for binary soft-sphere like systems where the frustration is due to the large, local rearrangement of atoms required for the formation of a crystal from a fluid or glassy configuration [37]. This situation is caused by presence of second component and so the frustration of such type increases with increasing concentration. Thus the quasibinary index F can serve as frustration measure in our system. It is clear that the value F 0 corresponds to the situation than only one of the scales dominates and so there is no frustration in the system. On the contrary, if F 1 the competition between hard-core and softcore scales and so the frustration are maximal. It should be noted that F depends on temperature as well as on density. In order to study the influence of frustration on glass-forming ability of the system we calculate the density dependence of the frustration index at sufficiently low temperatures. In Fig. 3b we show F (ρ) at T = 0.08 (of the order of T g for all the densities investigated). It follows from Fig. 3 that F (ρ) and T g (ρ) have opposite behaviour: the bigger F the smaller T g and the maximum of the former curve is located at the same density as the minimum of the later one. At the density ρ * where T g (ρ) has minimum the system has potentially a number of different lattice constants as we mentioned above. If we imagined the long range order formation at ρ * then the crystal would be strongly distorted by defects due to the competition of the different lattice constants. This situation should favour the diffusion and frustration at the same time, see Fig. 3. T g is determined both by the diffusion and by the frustration: diffusion tries to decrease T g while the frustration does the opposite. However the diffusion defeats frustration in our system, so T g has minimum at ρ * . Finally we discuss the boundaries of the density domain, ρ ∈ (0.35 − 0.75). Beyond these boundaries the glass forming ability quickly decreases. This one can judge from the quick rise of the diffusion coefficient, D(ρ), at the melting line, see inset in Fig. 3c, and the destruction of the quasibinary behaviour, see inset in Fig. 3b. In conclusion, we show using the molecular dynamics that the one component simple liquid with pure repulsive potential shows glassy behaviour in quasiequilibrium cooling. We explain the nonmonotonic density dependence of T g , frustration, the diffusion coefficient and fragility by the evolution of the quasibinary properties of our system. We observe that our system belongs to the short list of the most fragile systems. Our system can be used as the simple toy model for investigation of the quasibinary frustrated systems. In the search for (super)fragility in the one component repulsive simple liquids one should test them for quasibinarity. SUPPLEMENTARY MATERIAL Appendix A: Details of simulation and results handling MD simulation For MD simulations, we have used DL POLY 2.20 Molecular Simulation Package [38] developed at Daresbury Laboratory. Initially, the system was equilibrated from lattice configuration at high temperature for 10 6 MD time steps and then it was consecutively equilibrated at lower temperatures from final configurations of previous runs. At relatively high temperatures (T 1.2T g ), the equilibration time was chosen to be in order of 10 4 − 2 × 10 6 time steps. The time dependencies of temperature, pressure and configuration energy were analyzed to control equilibration. At low (T < 1.2T g ) temperatures addition control is needed so systems was consecutively equilibrated until convergence of time dependence of mean square displacement. Data were subsequently collected during the time t samp that was chosen to be large enough for correct calculation of diffusion coefficients by Einstein relation. So this time was approximately equal to t samp ∼ 3τ , where τ is the relaxation time, see Fig. 6(a). The reason is simple: one τ is needed to reach diffusive regime after ballistic and plateau ones, another τ is the main time interval for calculation by Einstein formula and one more τ is to improve averaging statistics for MSD calculations in the main interval. We obtain T g by quasiequilibrium method that is just approximation of equilibrium D(T ) curve for super- cooled liquid by appropriate fitting formula (for example, Vogel-Fulcher law) and using the relation D(T g )/D 0 = 10 n , where n = 13 − 17. This method is natural for (quasi)equilibrium simulation we have used but it differs from the dynamic method. According to the dynamic method, glass transition temperature is obtained as temperature at which the system falls out of equilibrium (i.e. when the cooling rate exceeds the relaxation one). This temperature can be determined by sharp peaks on heat capacity C p (T ) or thermal expansion coefficient, α p (T ). The temperature, T g , obtained by the dynamic method, depends on the cooling rate v, i.e. T g = T g (v). The natural question appears: how do the glass transition temperatures obtained by static (quasiequilibrium) and dynamic methods correlate with each other? The answer is not conclusive because the rigorous theory is still absent but some arguments may be given. There are a number of experimental and theoretical works (see, e.g., Refs. [2,4,25,[39][40][41][42][43]) where T g (v) dependence has been investigated. It was shown that this dependence can be represented in the form, where f (v) is some function of the cooling rate v = dT dt , where T and t are temperature and time correspondingly. The function f (v) is often approximated as follows: where B and v 0 are constants [39,41]. Expression (A2) is able to give a satisfactory fit to the experimental data when v is varied over 3 decades [39]. The shape of Eq. (A2) can be justified if we use the Vogel-Fulcher approximation for the relaxation time, and if we assume [see, e.g., Refs. [39,41,43]] that the system falls out of equilibrium at temperature T g (v) where the relaxation time is of the order of the inverse of the cooling rate, i.e. It follows from the derivation that Eq. (A2) is applicable while v v 0 , see, e.g., Ref. [43]. Note that there are other suggestions for f (v) that also agree not so bad with experiments [39][40][41]43]. In any case, at sufficiently small cooling rate T g becomes nearly independent on the cooling rate v. This is the limit where we perform our simulations. Then pressure and temperature are well defined in Nozier-Hover thermostat that we have used. The potential of the collapsing soft spheres model The potential of the collapsing soft spheres model is the pure repulsive potential with two characteristic scales corresponding to hard-and soft-cores, see Fig. 6(b). This potential was introduced in Refs. [12,15,16] in the following form: Using the identity where the Fermi-function, n F (x) = 1/[1 + exp(x)], we rewrote the potential of the collapsing soft spheres model: Angell plot Angell plot for temperature dependence of diffusion coefficient demonstrating essential deviation of D(T ) from Arrhenius law and so the fragile character of system dynamics, see Fig. 6(c). The construction of such plot is the traditional way to distinguish between the strong and fragile glassformer [13]. Appendix B: Discussion What is the basis for the paradigm that no glass can be expected for purely repulsive potential for quasiequilibrium cooling? First of all this is "experimental" fact. Namely, it is definitely established last decades from numerical molecular-dynamic experiments that purely repulsive potential in one-component isotropic system does not allow to freeze the system into a glass while cooling is quasiequilibrium (as it was defined above): the system spontaneously crystalizes (there is lack of frustration built in this system), see also discussion in Refs. [4][5][6]. However, if the cooling rate is large enough, larger than the typical relaxation rate, then it is possible to freeze such system into a glass with T g essentially depending on the cooling rate. But in this dynamical cooling regime the region above T g , we focus on in our paper, is not accessible where glassy slowing down is already present but the system is still in the liquid state. The dynamical cooling is in fact the method to "cook" the glass and find T g (v). Both regimes, qusiequilibrium and dynamical, are accessible experimentally. Therefore, if we want to study the mechanism of glassy slow down in supercooled state we have to use quasiequilibrium method. So our model system should not spontaneously crystallize being supercooled in quasiequilibrium conditions due to some frustrations. If we will speak on the intuitive level it becomes difficult to imagine frustration in simple monatomic pure repulsive system: potential is isotropic and what is most important only one characteristic scale is built in. So we propose the repulsive potential with additional source of frustration. Our potential has the soft step and the simple liquid with this interaction has two characteristic scales. The competition between these scales makes the system effectively quasibinary. And this is the origin of frustration in our system. The physical origin of the non-Arrhenius behavior of fragile glass formers is an area of active investigation in glass physics. Advances over the last decade have linked this phenomenon with the presence of locally heterogeneous dynamics in fragile glass formers; i.e. the presence of distinct (if transient) slow and fast regions within the material [2,36]. The quasibinary character of our particle system allows to make the selection among the particles: we mark by red (blue) color particles that have a neighbor at distance of the order of the first (second) maximum in rdf [hard-core and soft-core diameters], see Fig. 5 in the main paper. To demonstrate the high degree of the heterogeneity in our system we place the snapshot of the spatial particle distribution over the simulation volume, see Fig. 7. The high spatial heterogeneity of our system favors the existence of the locally heterogeneous dynamics within the system because of different free volumes (and so diffusion coefficients) of particles with different effective radii.	2013-01-09T08:31:47.000Z	2013-01-07T00:00:00.000	{ "year": 2013, "sha1": "d12ef0a63833de11aaef8f208e9ff46b344a002a", "oa_license": null, "oa_url": "http://arxiv.org/pdf/1301.2162", "oa_status": "GREEN", "pdf_src": "Arxiv", "pdf_hash": "d12ef0a63833de11aaef8f208e9ff46b344a002a", "s2fieldsofstudy": [ "Materials Science" ], "extfieldsofstudy": [ "Materials Science", "Medicine", "Physics" ] }
232237849	pes2o/s2orc	v3-fos-license	THE INFLUENCE OF MINERALOGICAL COMPOSITION ON THE ADSORPTION CAPACITY OF HEAVY METALS SOLUTION BY JAVA NATURAL CLAY, INDONESIA The adsorption capacity of four clay samples (Boyolali-BYL, Sleman-SLM, Gunungkidul-GK, and Pacitan-PCT) from Java, Indonesia, for copper (Cu) lead (Pb), zinc (Zn) and cadmium (Cd) solution was investigated by using batch equilibrium test. The adsorption data were presented using an isotherm curve, and adjusted to the Langmuir equation model, which produced the maximum adsorption capacity of the clay samples. The physical, chemical, and mineralogical analysis showed that the BYL and PCT samples have higher montmorillonite content, cation exchange capacity (CEC), and specific surface area (SSA) compared to SLM and GK clay samples. The batch equilibrium test revealed that the clay samples with higher montmorillonite content produced higher heavy metal adsorption capacity due to the higher cation exchange capacity (CEC), and specific surface area (SSA). The batch equilibrium test also show that the adsorption order for the metals cations followed the selectivity order Cu > Pb > Zn >Cd. The Langmuir model resulted in the adsorption processes, offering maximum adsorption capacities from 196.08 to 769.23 mg/g for Cu, 217.39 to 416.67 mg/g for Pb, 106.38 to 114.94 mg/g for Zn and 104.17 to 113.24 mg/g for Cd of four clay samples studied. The highest adsorption capacity was achieved for the BYL sample. The lowest was the GK sample, in which the order of the maximum adsorption capacity of clay samples was BYL > PCT > SLM > GK sample. This research indicated that the proportion of montmorillonite content in the clay samples affected the maximum adsorption capacity of the heavy metal in the solution. Introduction Nowadays, the presence of lead (Pb), copper (Cu), zinc (Zn), and cadmium (Cd), which are a heavy metals contaminant group that has toxic properties to humans and are usually found in waste [1,2]. Natural clay has physical and chemical properties to be able to adsorb heavy metals [3]. Several studies on natural clays' ability to adsorb heavy metals have been carried out using the batch equilibrium test, which is the simple but effective method [4,5,6,7]. The batch equilibrium test is a practical and quick test to predict the maximum amount of contaminants that can be adsorbed by a material and a representation of the experiment that can destroy all material structures; for example, the surface of the clay, which is an adsorbent medium. The batch equilibrium test is also carried out to measure the distribution coefficient (K). A measurement of this K value is essential because it is an indication of contaminant retention by solids. Several studies have been conducted only for geological of natural clay deposit and mineralogy in Java, Indonesia [8,9,10,11]. However, the comprehensive research for the mineralogical composition of these natural clay influenced the heavy metals' adsorption need to be investigated. The study of heavy metal adsorption in natural clay is an important aspect, as stated by several previous studies [12,13,3]. This understanding of the adsorption will deliver knowledge for the use of natural clay for environmental remediation purposes. This research aims to study characteristics and adsorption capacity on clay samples obtained from several locations in Java, Indonesia, and quantify the value of heavy metal adsorption. In this research, the batch equilibrium test was performed to obtain adsorption parameters, including maximum adsorption capacity (Qmax) and adsorption constant, by calculating the Langmuir equation model determination. The Langmuir model is widely used because of its mathematical simplicity and also easily interpretable constant which related to the highest possible adsorption [14] Clay Samples The clay samples in this study were obtained from four locations in Java, Indonesia, with different geological backgrounds. They were the clays of Boyolali (BYL), Gunungkidul (GK), Sleman (SLM), and Pacitan (PCT). The sampling location can be seen in Figure 1, and the geological information on the clay samples can be seen in Table 1 based on other studies [15,16,17,18]. Punung Formation Volcanic glass alteration Some of the main properties of clay samples were analyzed using the following procedure. The grain size was analyzed using the wet sieving and hydrometer method. The clay sample's pH was measured using a glass electrode, with a ratio of 1:5 for clay and water samples. The cation exchange capacity (CEC) of clay samples was analyzed by the BaCl2 method [19]. The specific surface area (SSA) of clay samples was measured by Brunauer-Emmett-Teller (BET) method. The mineralogical content and types of clay minerals were analyzed using X-Ray Diffraction (XRD) technique. The XRD was operated using a Philips PW1710 diffractometer with Ni-filtered CuKα radiation at randomly oriented samples. The quantitative assessment value of the mineral phases' presence was obtained from the XRD data, adopting the intensity of specific reflections and standard external variety of minerals [20]. Batch Equilibrium Test The heavy metal adsorption experiment on clay samples was carried out using the batch equilibrium test. In this experiment, the adsorption performance of Pb, Cu, Zn, and Cd was investigated in the clay samples used in this study that was done by preparing 10 grams of clay sample, which was placed in a 500-ml measuring cup. The desired eight heavy metal concentration solutions 5.0, 2.5, 1.25, 0.63, 0.32, 0.16, 0.08 and 0.04 mmol/l were prepared in this experiment at constant temperatures (25 • C). Then, 250 ml of a solution containing Cu(NO3)2, Cd(NO3)4, Pb(NO3)2, and Zn(NO3)2 has been added into the cup. The clay sample was dried, and then the batch test is carried out. The clay samples and heavy metal solution were stirred for 12 hours at room temperature and then were left without stirring for 24 hours. The agitation rates tested were 200 rpm, and the pH of the solution was kept on 6.5. The mixture was filtered with a 0.45 µm filter. The filtered solution was then analyzed for Pb, Cu, Zn, and Cd concentration with the measurement of inductively coupled plasma atomic emission spectroscopy (ICP AES). All analyzes were carried out by making duplicates three times. The maximum adsorption capacity (Qmax) of the Langmuir isotherm equation for adsorption was calculated using the Equation [21]: where, Q = adsorbed amount (mg/g) Qmax = maximum adsorption capacity (mg/g) K = adsorption constant (l/mg) C = initial metal concentration (mg/l) Physical, Chemical, and Mineralogy of Clay Samples Several properties of clay samples are given in Table 2. The BYL and PCT samples had higher CEC and SSA values compared to the SLM and GK samples. This phenomenon occurred may be attributed to the montmorillonite-typed mineral had three layers, which were a repetition of one layer of alumina and two layers of silica. Generally, CEC and SSA values of four clay samples show that relatively intermediate compared with other natural clay that has been studied by another researcher [22,23,24]. Another study state that the CEC and SSA value of pure montmorillonite mineral up to 150 meq/100 g and >100 m 2 /g [22] [25]. This difference may be due to differences in the percentage of clay minerals, and the existence of other mineral impurities like quartz. Also, this may cause for reducing the value of CEC and SSA, as found by other studies [26,27,28] Figure 2 shows the result of the XRD analysis of the clay samples. The diffractogram of XRD indicates the presence of clay minerals such as montmorillonite, illite, kaolinite, but another impurity mineral such as quartz also present. These results have been confirmed by other researchers, which show that the existence of clay in the study area was heterogeneous of clay mineralogy contents [8,9,10,11]. The semi-quantitative mineralogical composition is shown in Table 3. XRD analysis was found dominant for montmorillonite-type clay for BYL and PCT sample with percentages of montmorillonite 36.1% to 38.1%, respectively. Other samples are found dominant for kaolinite-type clay for SLM and GK sample with portions of kaolinite 35.1% to 36.1%, respectively. Other minerals such as quartz, plagioclase, feldspar, and other nonclay mineral were generally present in natural clay [29]. Result of Batch Equilibrium Test The result of batch equilibrium is shown in Figures 3 to 6 and Table 3. This test is conducted to obtain the Qmax value of the clay by using eight solutions. Figures 3 to 6 show the linear relationship between the values 1/Q and 1/C used to get the Langmuir constant value (K) and the Qmax value. The adsorption data in this experiment was obtained from 1/Q plot value in the 1/C function as the Langmuir model representation. Table 4 shows the amount of the Qmax and K parameters, and the value of the linear regression, R 2 . Adsorption isotherm models were recommended to determine adsorption capacity, as stated by several studies [30,31,32]. In this experiment, the Langmuir equation model defines the tendency of adsorption data with values of > 0.70 for R 2 , as shown in Table 4. This adsorption experiment indicates that the Langmuir isotherm model represents the adsorption progress with a high coefficient of determination. The coverage of metal cations on the clay samples' surface was reflected by the high R 2 value obtained. In the Langmuir isotherm model, the equilibrium is limited to the molecular layer's determination and correlate to electrostatic chemistry in the outer sphere complex [33]. The selectivity of four metals adsorption was exhibited by Qmax values (Table 4), and the value was not similar for metals investigated. Cu and Pb, the most greatly adsorbed metal in this experiment, were less influenced by competition than Zn and Cd for clay samples. This competition may be attributed to the fact that Cu and Pb were strongly adsorbed by the clay fraction and hence attempted effectively with the less adsorbed Cd and Zn and agree with another study [34]. Cu and Pb can enter clay lattices and form an insoluble group (e.g., Cu, Pb(OH)2, as the alternative hydrolysis product of Cu, Pb). These are irreversible processes in a laboratory experiment range. They could cause an important constituent of Pb and Cu's addition to be unavailable for cation exchange reactions. Another parameter influence the selectivity was related to the hydrated radius of the cation, as described by other researchers [35,36]. The hydration radius of the cations for Pb 2+ (4.01Å), Cu 2+ (4.19Å), Zn 2+ (4.30Å) and Cd 2+ (4.26 Å) respectively. Thus Pb and Cu seem to be less influenced by competition than other metals that are less adsorbed (Zn and Cd). The result revealed that Pb occupied a higher affinity for ion exchange on clay samples. This result of the selectivity is agreed with another researcher [37]. The adsorption of heavy metals ions on clay samples also depends on the ionic potential of the metals. The cations in the solution are connected with molecules of water in complex hydration. Generally, this hydration has a radius that was inversely corresponding to the cation radius. Hence, the smaller cations, the greater the hydration radius and cannot be easily eliminated from the clay surface. This finding also advises the low adsorption capacity of the Zn and Cd against the Cu and Pb component. In general, the adsorption order for the metals cations followed the selectivity order Cu 2+ > Pb 2+ > Zn 2+ > Cd 2+ . Generally, the montmorillonite type-mineral has a higher adsorption capacity than the kaolinite type-mineral. The adsorption capacity values may reach as three times or higher [27]. Adsorption of heavy metals on clay samples appearance to follow Langmuir isotherm with a maximum adsorption capacity, i.e., 769.23 mg/g for Cu of BYL clay sample. This maximum adsorption capacity was decreased to 196.08 for GK clay samples (see Table 4). Metals adsorption on clay sample is sensitive to the montmorillonite content. In this study, the BYL and PCT clay samples have higher montmorillonite content than SLM and GK samples. Other studies revealed that Pb's adsorption by montmorillonite clay from Pb solution up to 20 mg/g [38]. It was determined that adsorption capacity is higher on dominant montmorillonite on BYL and PCT samples (Cu: 769.23 mg/g, Cd: 588.24 mg/g) than on kaolinite on SLM and GK samples (Cu: 416.67 mg/g, Cu: 196.08 mg/g). In another study, the better removal efficiency was obtained for heavy metals removal using montmorillonite instead of another type of clay [39]. The higher maximum adsorption values obtained in the BYL and PCT samples, and the lower values obtained in the SLM and GK samples, are shown in Figure 7. This adsorption capacity was consistent with the results of another researcher who stated that clay mineral type such as, for example, higher montmorillonite content, influenced the maximum adsorption capacity value [26]. Other research suggests that the high content of montmorillonite will provide high binding energy for metal absorption. The results obtained showed that clay samples had different behavior in adsorbing Cu, Pb, Zn, and Cd. This behavior might be related to the adsorption mechanism caused by the mineralogical content of the four types of clay samples studied, confirmed by other research [40]. These results indicated that the difference in the proportion of mineralogical contents in clay samples would affect the adsorption capacity of metal ions. This result shows that the order of maximum adsorption capacity of clay samples was BYL > PCT > SLM > GK sample is obtained by analyzing the values of maximum sorption capacities calculated from the Langmuir equations, The graphical correlation between maximum adsorption capacity (Qmax), CEC, SSA, and % montmorillonite values are shown in Figure 8-10. The graph in Figure 8 to 10 indicates a good correlation between maximum adsorption and CEC, SSA, and % montmorillonite values with R 2 values of more than 0.6. This result is agreed with other research results, which mentioned that clays with higher CEC, SSA value, and % montmorillonite have higher adsorption values to adsorb metal cations [7,41]. Parametric Pearson correlation analyses were applied to the four parameters tested in pairs to assess the degree of association between the parameters examined. The Qmax of Cu, Pb, Zn, and Cd was strongly correlated with CEC, SSA, and % montmorillonite with Pearson's correlation coefficient of 0.01 and 0.05 level, as shown in Table 5. The results showed that CEC, SSA, and % montmorillonite had a strong correlation with Qmax of the four heavy metals. The trend shown by the Qmax values in the Langmuir model also suggested that adsorption capacity increased with CEC, SSA, and % montmorillonite. The results indicated that the correlations between Qmax and % montmorillonite value more than 0.75 were the most significant, for Qmax of Cu. This highest correlation value may be attributed to the highest adsorption capacity (Figure 7). On the other hand, the correlations between Qmax and % montmorillonite value was less than 0.75. This correlation value showed a lower degree of significance for Qmax Pb, Zn, and Cd. This lower correlation value probably due to lower adsorption capacity (Figure 7). The statistical correlation of 0.01 and 0.05 level identified for % montmorillonite strongly influenced on Cu, Pb, Zn and Cd adsorption. Conclusions This work investigated the adsorption characteristics of four clay samples obtained from Java, Indonesia, and measured the adsorption of heavy metals by batch equilibrium test to evaluate the adsorption parameters. The four clay samples were composed of dominant kaolinite contents for SLM and GK samples, while the BYL and PCT clay samples were mainly montmorillonite mineral. The adsorption capacity from batch equilibrium tests indicates that BYL and PCT have higher adsorption capacity on heavy metals than samples SLM and GK. The study also found a positive relationship between the adsorption capacity and the clay mineralogy content of the clay samples. The adsorption capacity of natural clay containing dominant montmorillonite-clay mineral is much higher than of kaolinite-clay mineral for all the four metal ions used in the experiment. The study also suggests that physicochemical characterization and batch equilibrium tests can be beneficial in assessing clay for an adsorbent material. Both clay samples containing dominant kaolinite-clay mineral and montmorillonite-clay mineral are capable of removing Cu, Pb, Zn, and Cd from aqueous solution. This finding also shows that the proportion of different mineralogical content in clay samples is affected by heavy metal cations removal from wastewater. Future work for investigation of the study, such as the influence of pH, should be considered in the batch test. Also, comparison by using another isotherm model such Freundlich isotherm need to be done to get a more comprehensive understanding of the adsorption study. Finally, tentative finding obtained in this study shows that clay samples in this study is suitable and has potential value for the adsorption of heavy metal contamination from wastewater, considering it is a low-cost, abundant, and locally available material.	2021-03-16T08:46:14.726Z	2021-03-16T00:00:00.000	{ "year": 2021, "sha1": "c9119144531e211568efcd8555473de0359ce8d0", "oa_license": null, "oa_url": "https://journals.utm.my/aej/article/download/16679/7637", "oa_status": "GOLD", "pdf_src": "MergedPDFExtraction", "pdf_hash": "c9119144531e211568efcd8555473de0359ce8d0", "s2fieldsofstudy": [ "Agricultural And Food Sciences" ], "extfieldsofstudy": [ "Chemistry" ] }
247100410	pes2o/s2orc	v3-fos-license	Low and No-Contact Euthanasia: Associated Ethical Challenges Experienced by Veterinary Team Members during the Early Months of the COVID-19 Pandemic Simple Summary During the COVID-19 pandemic, many veterinary practices have been required to move to a low or no-contact consultation model to minimise the risk of SARS-CoV-2. Utilising data from a global survey, we explored the experiences of veterinary team members performing low and no-contact euthanasia during the early months of the COVID-19 pandemic. We found that low and no-contact euthanasia were encountered as common and/or stressful ethical challenges in the pandemic. In order to minimise the potential negative impacts of low and no-contact euthanasia on veterinary team members, clients and animal patients, there is a need for a toolkit of protocols to assist veterinary team members in provision of low-contact euthanasia, and avoidance of no-contact euthanasia wherever possible. Abstract Background: During the ongoing COVID-19 pandemic, many veterinary practices around the world have shifted to a low or no-contact consultation model to ensure the safety of their team members and clients, and comply with public health orders, while continuing to provide veterinary care. Methods: We performed reflexive thematic analysis on a subset of data collected using a mixed-methods survey of veterinary team members globally. Results: There were 540 valid responses available for analysis. Low and no-contact euthanasia we raised as a common and/or stressful ethical challenge for 22.8% of respondents. We identified five key themes: no-contact euthanasia as a unique ethical challenge; balancing veterinary team safety with the emotional needs of clients; low and no-contact protocols may cause or exacerbate fear, anxiety and distress in veterinary patients; physical distancing was more challenging during euthanasia consultations; and biosecurity measures complicated communication around euthanasia and end-of-life decision making. Recommendations: In light of concerns highlighted by respondents, we recommend the development of a toolkit of protocols that will assist veterinary team members in performing low-contact euthanasia in a range of circumstances, in alignment with their values and professional ethical codes. Professional bodies may be involved in developing, updating and disseminating this information, and ensuring a continuous supply chain of PPE. Introduction The World Health Organisation (WHO) declared a global pandemic on 11 March 2020 [1]. The COVID-19 pandemic led to major changes in veterinary practice to minimise the risk of transmission of SARS-CoV-2 in veterinary settings and in some instances comply with public health orders, including the increased use of personal protective equipment (PPE), and low and no-contact consultations to facilitate physical distancing. For the purposes of this discussion, "low-contact" refers to strategies aimed at minimising physical contact with clients, such as minimising the number of clients in the consultation room, physical distancing, and requesting that clients wear PPE during the consultation. "No-contact" refers to strategies aimed at eliminating physical contact with clients. In such instances, clients may have had verbal contact with veterinary team members, for example via telephone or internet, to communicate their concerns, provide a patient history or give consent, but were required to remain outside of the premises at all times, for example in the case of "drop-off" or "curbside" consultations, or telemedicine. Due to a surge in demand, disruption of global supply chains, shortage of raw materials for production, competition between countries for PPE and in some cases interception of PPE imports, there was a global shortage of PPE [2]. In particular, masks, goggles, face shields, gowns and N95 respirators were in very short supply [3]. Shortages of PPE in human healthcare settings left those caring for patients with COVID-19 extremely vulnerable to COVID-19-associated morbidity and mortality [4,5]. Healthcare workers were also deemed an important source of SARS-CoV-2 transmission [5]. These factors made protection of healthcare workers from infection a priority for disease control. Vaccinations were not approved for use until December 2020 in the UK, and in many countries not until much later [6,7]. With no approved vaccinations or effective treatment, physical distancing, PPE, and minimising the duration of proximity to others where physical distancing was not possible, were key elements of prevention. For this reason, veterinary professional organisations and associations such as the American Veterinary Medical Association, promoted conserving PPE by postponing elective procedures, extending the use of disposable PPE or even reusing disposable PPE in some circumstances [8]. To minimise the risk of transmission of SARS-CoV-2 between clients and veterinary team members in veterinary clinical settings, most amended their practice to align with local public health orders or recommendations; to limit service provision only to "essential" services; to reduce the number of veterinary team members on site at any one time, and minimise client contact by limiting the number of clients entering veterinary premises, or to exclude clients from veterinary premises entirely [9][10][11]. In 2020 we surveyed veterinary team members around the world about the types of ethically challenging situations they had faced since the beginning of the pandemic. The results are published elsewhere [12,13]. We identified "no-contact consultations" in general, particularly "no-contact euthanasia consultations", as distinct, novel types of ethically challenging situations faced by veterinary team members. Euthanasia is commonly performed in veterinary contexts [14][15][16][17][18][19][20]. Derived from the Greek "eu" for good and "thanatos", pertaining to death, "euthanasia" describes the killing of an animal in such a way that minimises pain and distress to the animal patient, and emotional distress of those present, including animal owners [19]. While historically it was commonplace to separate animals from clients at the time of euthanasia, best practice is now keeping bonded humans and animals together, or at least providing that option for clients [21]. Previous surveys have identified euthanasia of animals as a source of moral stress and moral distress for veterinarians in particular [22][23][24][25]. Whether veterinary team members experience euthanasia as an ethical challenge may depend on the indication or reasons for a euthanasia request [26]. In preventing veterinary team members from keeping bonded humans and animals together, low and no-contact euthanasia may violate their expectations/values/beliefs around what constitutes a good death. In order to better prepare current and prospective veterinary team members working in the context of this and future pandemics, we sought to better understand the ethical challenges posed by low and no-contact euthanasia during the early months of the COVID-19 pandemic. Materials and Methods The methodology for this project has been described in detail elsewhere [12]. Briefly, we developed and administered an online, mixed-methods survey to explore the frequency and stressfulness of ethically challenging situations encountered in the early months of the COVID-19 pandemic. The anonymous survey, hosted on the secure web application Research Electronic Data Capture (REDCap), consisted of 29 questions across three sections. Participants were invited to provide free-text responses to three questions: "Since the advent of COVID-19, describe the most COMMON ethically challenging situation you have encountered as a veterinary team member?"; "Since the advent of COVID-19, describe the most STRESSFUL ethically challenging situation you have encountered as a veterinary team member? (If the response is the same as above, enter "same")"; and "Is there anything else you would like to add about your experience with ethically challenging situations since the advent of COVID-19?" For the first two questions, participants were instructed that the ethically challenging situation identified did not have to be specific to the COVID-19 pandemic. For all questions participants were advised not to include potential identifying information such as names of individuals or workplaces in their responses. In this study, we pooled and analysed the free-text responses to these three questions. De-identified data were downloaded into Microsoft ® Excel for Microsoft Office 365 MSO (16.0.13328.20262). Responses were sorted into categories for the purposes of descriptive statistics. Summary statistics were calculated for the demographic variables using IBM SPSS version 24. Responses were screened to exclude identifying information, then uploaded onto NVivo ® 12 Plus software (QSR International) to facilitate thematic analysis. For this paper, free-text responses referring to the practice of low-and no-contact consultations relating to critically ill patients, or where euthanasia was discussed or performed, were compiled in order to perform a reflexive thematic analysis on this subset of data. Where respondents had written "same" in response to the second free-text question, to indicate that the most common ethically challenging situation was also the most stressful, this second comment was excluded from analysis. When performed rigorously, qualitative research is explicitly acknowledged to be "context-bound, positioned and situated" [27], with analysis of data reliant on interpretation of the situated researcher. Researcher subjectivity is recognised as a resource rather than a barrier to knowledge production [27]. Thematic analysis is "an interpretive activity undertaken by a researcher who is situated in various ways, and who reads data through the lenses of their particular social, cultural, historical, disciplinary, political and ideological positionings" (original emphasis) [28]. It is therefore considered best practice for those performing reflexive thematic analysis to describe their own perspectives, including their "personal and social standpoint, and positioning" [28]. The first author is a companion animal veterinarian, practicing as a primary accession veterinarian and a lecturer in the Sydney School of Veterinary Science. She is also a lifelong companion animal owner. She has been a practicing veterinarian since 2005, well before the COVID-19 pandemic was declared, and has continued to practice since then, modifying her practices in line with public health orders and protocols at practices where she works. Therefore, during the study period, she performed low and no-contact consultations, as well as low-contact euthanasia consultations. The second author is a veterinarian, researcher and lecturer in animal welfare and veterinary ethics at University College, Dublin. She has a long-standing interest in animal welfare science, ethics and law, starting as a student and continuing through practice and into teaching. The third author is a veterinarian, lecturer in epidemiology and public health, and a researcher in the Sydney School of Veterinary Science. His veterinary clinical experience is derived exclusively from government practice as a field veterinarian. He has a strong interest in infectious and transboundary diseases and has conducted original research on the COVID-19 pandemic. Since the declaration of the global pandemic, all authors have engaged either wholly or mostly, in virtual (no-contact) teaching of DVM students. Data analysis involved six stages. Firstly, the first author read all comments at least three times. Secondly, initial codes were generated. Each comment was coded inductively for semantic themes, employing a realist approach without a pre-existing theoretical framework. An iterative approach was used. A single comment could be coded multiple times. Where a comment could not be assigned an existing code, a new code was generated. Thirdly, initial themes were generated. Codes were examined to identify clusters of codes and complex codes which were grouped together as themes thought to best represent the data. Themes were reviewed for both internal coherence and distinctiveness from other themes. This involved regularly re-reading all coded extracts from each theme. Where extracts did not fit a theme, these were either reallocated to a more appropriate theme or allocated to a new theme. The fourth and fifth stages-refining themes and developing a thematic map, and defining and naming themes, were performed concurrently, and involved further discussion between all authors. The sixth and final stage involved selection of examples illustrative of each theme. Results There was a total of 540 valid responses. There were 141 comments, provided by 123 respondents (22.8%). Key demographic frequencies of both the overall respondent population, and the subset who commented on low or no-contact euthanasia in the free-text comments, are summarized in Tables 1 and 2 We identified five major themes relating to euthanasia: no-contact euthanasia as a unique ethical challenge; balancing veterinary team safety with the emotional needs of clients; low and no-contact protocols may cause or exacerbate fear, anxiety and distress in veterinary patients; physical distancing is more challenging during euthanasia consultations; and biosecurity measures complicated communication around euthanasia and endof-life decision making ( Figure 1). No-Contact Euthanasia as a Unique Ethical Challenge A number of respondents raised no-contact euthanasia as the most common, th stressful or both the most common and most stressful ethically challenging situatio encountered since the beginning of the pandemic, due largely to the absolute exclu owners: No-Contact Euthanasia as a Unique Ethical Challenge A number of respondents raised no-contact euthanasia as the most common, the most stressful or both the most common and most stressful ethically challenging situation (ECS) encountered since the beginning of the pandemic, due largely to the absolute exclusion of owners: "Disallowing witnessing euthanasia" veterinarian, Singapore "Putting animals to sleep without owners allowed to be present" veterinary nurse, UK "Having to make owners stay outside while we take their pet inside" veterinary nurse, Australia The third comment in particular suggests externally imposed rules or protocols followed by veterinary team members that disallow owner presence. Balancing Veterinary Team Safety with the Emotional Needs of Clients Some respondents described the challenge of managing the conflict between ensuring the safety of the veterinary team, through strict physical distancing, while meeting the needs of clients to be present during euthanasia of an animal. A veterinary nurse from the USA described the difficulty in weighing up the costs (to the veterinary team) of allowing clients to be present, against the emotional costs (to the client) of not allowing them to be present during euthanasia: "Being forced to choose between allowing clients into the facility for a euthanasia or maintain "no client access" policies instituted to reduce potential exposure and allow for social distancing. By allowing clients to be present during a euthanasia there is a risk of exposure to both our staff members and the clients in question. It uses scarce and valuable PPE and adds further stress to the team in an already emotionally taxing situation. However, denying clients the opportunity to be present during the euthanasia compounds the grief and loss of an already deeply traumatic situation and denies them a sense of closure and control." Veterinary nurse, USA Other respondents confidently prioritised the safety of the veterinary team, justifying it as " . . . the ethical decision to protect our staff." (Veterinarian, USA). Some were prepared to take a calculated risk in breaching workplace protocols excluding owners from attending euthanasia consultations: "Not being allowed to have owners present or even visit their pet again prior to euthanasia. I found it to be excessive and unnecessary. While I am worried about COVID just as much as the next person and I want to take precautions, I don't see why we can't offer the client to be present outside the building on a bench with masks and long extension set etc. I was reprimanded by management for doing just that." Veterinarian, Canada No-contact euthanasia was experienced for some respondents as an ethical challenge even when the owner was self-isolating or diagnosed with COVID-19. An Australian veterinarian described the most stressful ECS they encountered as " . . . inability of owner to be present for euthanasia when known COVID positive," underscoring the view that the costs to the owner of not being present were significant, even in the face of high likelihood of exposure of the veterinary team member(s) to the virus in such situations. Some respondents mentioned not being able to provide a home euthanasia service, due to factors such as reduced staff numbers and increased workplace biosecurity restrictions, as a source of distress, despite justifying such measures as a means of protecting both veterinary team members and clients: " . . . having to decline house calls for elderly clients for both our, and their, protection." Veterinarian, Australia For clients such as the elderly, persons living with disabilities and those without transport, house call consultations may have been their only means of accessing veterinary care. For these clients, loss of the house call service may have equated to loss of access to veterinary care altogether. Some respondents found that not allowing owners into the clinic or hospital to visit critically unwell or dying patients, or even leave a familiar-scented item to comfort an animal due to concerns about fomite transmission of SARS-CoV-2, particularly challenging: "It is difficult to receive a patient, specifically one that is critical or in pain, and tell the owner they have to wait in their car and/or are not allowed to come in with their pet. Similarly, when owners want their hospitalized pet to have a blanket or shirt with them, it is hard to tell them no." Veterinary nurse, USA For others, not allowing owners to visit critical or dying animals transgressed their values about acceptable care of animals and their owners: "Not being able to provide clients contact with their seriously ill hospitalized pet . . . it is not what I would consider acceptable for my own pets." Veterinary nurse, Australia One respondent referred to a client-free hospital as the "ideal", but described weighing up the needs of both the veterinary team and the client, suggesting a possibly flexible approach: "Balancing the needs of clients to see/visit their critically ill pet with the needs of our staff/hospital to maintain a socially distant and ideally client free hospital." Veterinarian, Australia Low and No-Contact Protocols May Cause or Exacerbate Fear, Anxiety and Distress in Veterinary Patients A number of respondents felt that no-contact euthanasia in particular not only negatively impacted clients, but animal patients themselves, with the key stressor identified as separation from their owner in an unfamiliar environment: "Many dogs are stressed away from their owner. Also, in the case of very sick animals/emergencies/euthanasias owners are distressed about not being able to be with their animal. Do you cave and let them be there knowing that if you get covid19[sic] the entire clinic team and possibly other clients could get infected, or stick to the policy knowing you are causing emotional distress to the owner and animal?" Veterinarian, Australia "Anxious animals being away from their owners creating a more negative environment for the animal to be in." Veterinary nurse, Australia " . . . distress of pets and owners when separated from [each] other to allow social distancing during exam." Veterinarian, Australia One respondent described being able to implement work-arounds to avoid separation of owners and animals, though did not elaborate on the nature of these: "People want to be with their animals, and some need to be . . . sometimes the dog or cat needs them. We finds[sic] ways to accommodate that." Veterinarian, Canada There were concerns that persons wearing PPE may add to fear, anxiety or distress in veterinary patients: "I thought it was very over the top that in Australia some clinics were either not allowing clients to be present for euthanasia of their pet or required the client to be gowned up in a hazmat suit to be present (and scaring the poor dog with the outfit)." Veterinarian, Australia Physical Distancing Is More Challenging during Euthanasia Consultations Where low-contact euthanasia was performed, respondents described different strategies to maintain physical distancing, including minimising the time in which owners and veterinary team members were in close proximity, performing euthanasia outdoors, the use of intravenous catheters and lines to allow remote injection, and/or minimising the number of people present. One respondent described these extra measures as presenting the most stressful ECS during the pandemic, highlighting the need to separate the client from the animal during the process of euthanasia: "Not being able to allow clients to be present the whole euthanasia procedure i.e., taking the animal off them in the car park, placing IVC [intravenous catheter], then bringing clients around the back to outside where they must remain for the procedure, never allowing them in the clinic." Veterinary nurse, Australia Some respondents noted concerns about potential increased risk for COVID-19 transmission in euthanasia consultations: "Being in close proximity to grieving owners (with increased secretions) is stressful on the staff." Animal health technician, USA "Owners crying without masks during euthanasia." Veterinarian, USA Restricting the number of persons present in the euthanasia consultation was the most stressful ECS for some team members, due to concerns about the impact on clients excluded from the procedure: " . . . only allowing 1 person to be present when saying goodbye to their pet. It causes moral conflict because it feels wrong asking other family members to leave in a hard time when they are also grieving and would like closure." Veterinary nurse, Australia Some respondents were concerned about the impact of excluding others on individuals forced to attend euthanasia without the support of others: "Family's [sic] not allowed to be present during euthanasia. Only one family member outside the building. Seeing the sadness/distress of the one family member shouldering the burden alone." Veterinary nurse, Australia One respondent described physical distancing requirements as a deterrent to work-up of cases for some clients, perhaps leading to premature euthanasia decisions: "Clients are quicker to elect euthanasia as apposed [sic] to diagnostics as it's more difficult to bring them into the practice." Veterinarian, Canada Biosecurity Measures Complicated Communication around Euthanasia and End-of-Life Decision Making Physical distancing complicated communication around euthanasia and end-of-lifedecision making. Respondents described the challenge of being unable to demonstrate the clinical status of an animal to the client as they may have done previously: "Trying to convince an owner that it's the right time to euthanise their pet when the owner is unable to see their pet's clinical status and what is happening in the hospital." Veterinarian, Canada In some instances, the prospect of no-contact euthanasia impacted end-of-life decision making. Some respondents reported that the most stressful ECS they encountered was client refusal to euthanise an animal if they could not be present: "Euthanasia being refused by clients as they cannot be present with their animal." Veterinarian, UK "Clients not wanting to put their pets to sleep as they are unable to attend euthanasia." Veterinary nurse, UK Some respondents noted the lack of contact between themselves and grieving owners as a common or stressful ECS, due to inability to express compassion in a way they were accustomed to: "It has been difficult to have to refrain from any human touch or closeness during such a personal procedure which requires empathy." Veterinary nurse, Australia " . . . not hugging the client or spending time with them which we normally do." Veterinarian, Australia Discussion Low and no-contact euthanasia of veterinary patients were experienced as stressful by veterinary team members during the COVID-19 pandemic. Traditionally, the veterinary ethical literature has focused on the client, the animal and the veterinarian as key stakeholders [29]. Comments from respondents suggest that veterinary team members were conscious of the needs of a much broader range of stakeholders. Because of the infectious nature of SARS-CoV-2, decisions around whether and how to perform low or no-contact euthanasia also had the potential to impact household members and contacts of clients and veterinary team members, as well as the wider community [12], all with varying risks of viral exposure. Additional stakeholders mentioned included human healthcare workers and the human healthcare system, due to scarcity of resources such as PPE, as well as professional associations, registration boards, charities and non-Government organisations [12]. In addition to complying with professional codes of conduct, veterinary professionals in many jurisdictions were required to comply with public health orders. The focus of our survey was the frequency and type of ECS encountered during the pandemic and did not attempt to discern the predominant ethical framework(s) utilised by veterinary team members, or whether the ethical approach of veterinary team members shifted with the advent of the COVID-19 pandemic. Nonetheless, responses tended to be most aligned with deontological, utilitarian or virtue ethics approaches. Deontology holds that an action is good or right if it conforms to a rule or a moral norm, and prioritises the intentions of the decision-maker [30]. The theme "No-contact euthanasia as a unique ethical challenge" comprised comments about following rules, for example, "disallowing" clients from being present during euthanasia, as well as comments indicating distress about the inflexibility of such rules, and the consequences, particularly for grieving clients. The emphasis on disallowing owners, or "having to" exclude them from the process, suggests that these veterinary team members were following protocols, or felt compelled by circumstances to act, in conflict with their values. While it may be unavoidable due to workplace policies or public health orders, acting in a way that transgresses one's deeply held moral beliefs causes moral distress [31]. This can impact the welfare of veterinary team members [32]. In human healthcare, bans on visitors of hospitalised patients, particularly those in ICU and those dying from COVID-19, were instituted around the world [33,34]. These caused distress not just to family members of those patients, but also to healthcare workers [35,36]. The idea of dying alone contravenes beliefs about what is considered a "good death" in many cultures [37]. Selman and others note that "a key clinical debate is whether, and how, to facilitate family members and close friends to be present when someone dies in hospital, hospice or care home during a pandemic" [37]. Family members who are not able to visit dying relatives to say goodbye are at higher risk of developing complicated grief and post-traumatic stress-like disorders [35]. Being unable to see a family member right before, during or immediately after death made it hard for some to accept that the person had died [38]. Given the attachment that many owners have to their animals, it is likely that veterinary clients who were unable to be present during euthanasia may be susceptible to similar negative sequelae. In our study, many respondents appeared to take a utilitarian approach to decision making around owner presence at the time of euthanasia. Broadly speaking, utilitarians seek to achieve the greatest positive consequences (or the least worst) for the greatest number of stakeholders [30]. This is captured in the theme "Balancing veterinary team safety with the emotional needs of clients". Consider the respondent who posed the question about whether one allows a client to be present "knowing that if you get covid19 [sic] the entire clinic team and possibly other clients could get infected", vs. not allowing the client to be present, leading to distress for both the animal and the client. However, weighing costs and harms did not necessarily yield a satisfactory approach. Utilitarians evaluate decisions according to their consequences-but the respondent could not have predicted with certainty whether they would acquire COVID-19, infect other team members and clients, or indeed how severe such infections would be. Nor could they measure with any certainty the degree of harm to the client or the animal. According to a utilitarian framework, steps taken to mitigate or eliminate the risk of harm are ultimately evaluated according to their consequences, which cannot be known until after those steps are taken. In the context of a pandemic, uncertainty is increased. Interestingly, some commentators attributed moral distress among healthcare workers during periods of extreme resource constraint during the pandemic to a shift in the predominant medical ethic toward utilitarianism [36,39]. Some human healthcare workers (such as veterinarians) breached no-contact protocols in order to "minimise the negative psychological effects caused by not being able to say goodbye and possible ongoing complications of mourning" [38]. Increasingly, the professional identity of veterinary team members has come to be centered around primary concern for animal welfare. Indeed, "protecting and promoting animal welfare" is described as the veterinarian's "raison d'etre" [40], and is embedded in codes of professional conduct for veterinarians, animal health technicians and veterinary nurses [41][42][43][44][45][46]. This focus on animal welfare has been accompanied by a recognition of the potential iatrogenic harms of veterinary care [47], and concerted efforts to minimise fear, anxiety and distress in veterinary patients [48][49][50][51][52]. For example, the European Veterinary Code of Conduct states that "euthanasia must be practiced with as little pain, distress and fear as possible" (1.2, Recommendation 4) [44]. Yet the theme "Low-and no-contact protocols may cause or exacerbate fear, anxiety and distress in veterinary patients" suggests that public health considerations (also embedded in professional codes of conduct), came into conflict with this iatrogenic harm minimisation ethos. A randomized crossover trial of 44 client-owned dogs examined in the consultation room in the presence of their owner, and the common treatment area ("out the back") without the owner present, reported higher levels of fear, anxiety and stress in more dogs examined in the common treatment area, without their owners [53]. Similarly, a randomized crossover trial of 21 client-owned cats found that separation from owners and examination in the common treatment area were associated with clinically significant increases in perceived stress in cats [54]. These findings suggest that, where possible, examinations and minor procedures should be performed in the consultation room, with the owner present [49]. At the time of euthanasia in particular, it is recommended to keep the client and patient together throughout the euthanasia appointment "to reduce anxiety for both" [21]. While implemented for the safety of veterinary team members and clients, a potential unintended consequence of no-contact consultations is an increased risk of injury. Some respondents highlighted concerns around safety associated with separation of animals from their owners. Anxious and fearful animals are more likely to scratch, bite or otherwise injure veterinary team members, and may be more refractory to sedation [49,50]. Another common approach to ethics is virtue ethics, which prioritises cultivation of morally relevant, persistent character traits such as compassion, honesty, trustworthiness, integrity and discernment [30]. Virtues are linked to one's role(s), which may vary. For example, a respondent may have roles as a veterinary team member, a parent, a carer, and a community member. Low and no-contact euthanasia may have led to moral distress for veterinary team members because they were unable to perform their role in alignment with their core values (for example, compassion), or because their professional role as a veterinary team member caring for animals and clients came into conflict with their other roles (for example as a family member or carer seeking to protect those they live with). Indeed, we found conflict between the wellbeing of family/household members and professional role was reported to be among the most common (reported by 46.3% of respondents) and most stressful (33.6%) ethically challenging situations encountered by veterinary team members during the early months of the pandemic [12]. One challenge with virtue ethics is how to manage conflict between different virtues. The finding that, for at least some respondents, physical distancing was more challenging during euthanasia consultations, may reflect a conflict between the expectation for veterinary team members to be discerning, to follow reasonable public health orders and to minimise biosecurity risk, and the expectation that veterinary team members are compassionate in the face of the grief of clients and their family members. It can be difficult to navigate conflict between different roles and virtues [30]. Euthanasia, in particular, presents a challenge when physical distancing, as it is a time when veterinary team members must be in close physical proximity to an animal to prepare for and perform euthanasia. It is also commonly a time when owners wish to be close to the animal, bringing them into close proximity with veterinary team members. Prior to the pandemic, the presence of multiple family members, friends, support persons and even other animals prior to, during and after euthanasia of animals was common. Extended appointments for euthanasia were routine, and it was common for multiple persons to attend. This may reflect the reality that "euthanasia appointments are as close to a funeral as some clients will have for their pets" [21]. But extended appointments conflicted with advice to minimise duration of client contact. As stated by several respondents, it is not uncommon for clients, and sometimes veterinary team members, to cry during euthanasia consultations. Tears, along with respiratory droplets, are a potential source of SARS-CoV-2 infection [55], as alluded to by some of the respondents. Few published protocols for low-contact euthanasia were available at the time. In the experience of the first author (AQ), most veterinary teams in Australia developed their own approaches to low-contract euthanasia on an ad hoc basis, or, where possible, referred clients to home euthanasia services. Indeed, the USA-based Companion Animal Euthanasia Training Academy (CAETA) reported an increase in referrals to home euthanasia services during the pandemic, as well as an increased number of outdoor euthanasias [56], as some hospitals sought to avoid admitting clients onto the premises. Available guidelines focused on minimising contact time between veterinary team members and clients and reducing the risk of fomite transmission. For example, an early edition of "COVID-19: A guide to reopening veterinary medicine in Ontario" recommended the following: "Euthanasia appointments should be structured so that time in close proximity to the client is minimized. For example, contactless or quick transfer of the patient, distanced escort of an owner to a room, insertion of a catheter in a separate room, keeping personnel distant from the owner until the time of injection, having the owner stand distant or, if they will hold the animal, have personnel wear PPE to protect themselves (mask and eye protection); Documentation of verbal consent rather than requiring signatures; Using contactless electronic payment wherever possible" [57] Some continuing education providers shared strategies for performing low-contact euthanasia. For example, CAETA recommended that house call veterinary team members reduce their exposure by reducing overall appointment volume and minimising the number of people present at euthanasia, screening clients ahead of the appointment for signs of illness, explaining the procedure and collect payment over the phone; dispensing pre-visit pharmaceuticals that clients could administer to animals prior to the appointment to promote sedation and anxiolysis, wearing PPE, requesting that clients present wear PPE, ceasing physical contact with clients (avoid handshakes, hugs), encouraging virtual presence at euthanasia, performing euthanasia outdoors where possible, minimising potential fomite transmission by documenting verbal or electronic instead of written consent, reducing handling of animal bodies and using disposable pads rather than towels beneath animals [58,59]. Euthanasia protocols for anxious or aggressive animals are designed to minimise contact between veterinary team members and the conscious patient, often incorporating oral premedication or sedation [60]. Anecdotally, some teams began using these protocols routinely during the pandemic to minimise contact between veterinary team members and clients. For example, where it was safe to do so, some veterinarians utilized a threestep euthanasia process in canine patients involving (a) oral transmucosal application of detomidine hydrochloride gel (an oral transmucosal preparation typically used to sedate and restrain equine patients, but known to cause reversible sedation in dogs [61,62]) by the owner under the direct supervision of the veterinarian; (b) subcutaneous or intramuscular injection of a sedative agent, and (c) placement of an intravenous catheter in a hindlimb, attached to a long extension set to facilitate pentobarbitone sodium injection at a distance from the patient and clients, or intrahepatic injection of pentobarbitone sodium (J. Campbell, personal communication, December 2021). Non-veterinary team members present would be asked to step away from the dog while injections were given or intravenous catheters placed in steps (b) and (c) but could resume physical contact with the animal once veterinary team members moved away from the patient. Biosecurity Measures Complicated Communication around Euthanasia and End-of-Life Decision Making Biosecurity measures, including low-and no-contact consultations, and the use of PPE-in particular, masks-complicated communication between veterinary team members and clients in general [63], so it is not unexpected that they also complicated communication and end-of-life decision making. Communication that might normally occur in the consultation room may have occurred over the phone or via telemedicine, reducing the ability of both veterinary team members and clients to read non-verbal cues [63]. According to Ware and colleagues, briefer appointments and those where the client is separated from the animal can complicate decision making around treatments, monitoring of outcomes and establishing humane endpoints [64]. In human healthcare settings, virtual communication presented a challenge for family members of some patients, including difficulty hearing and unreliable WiFi-connection [65], and could be a source of stress for some family members if not managed appropriately [66]. Video calls could be a source of comfort to some family members [66][67][68], though some bereaved family members and friends displayed an ambivalent attitude to the use of devices to facilitate virtual farewells [37]. Telephone communication was associated with a perceived decrease in communication quality, information and support [69]. Masks reduced the ability to read facial expressions, eliminated lip-reading, and may have reduced audibility of verbal communication [65,70,71]. Veterinary team members typically play an important role in supporting pet owners during end-of-life discussions, euthanasia and the immediate aftercare of the animal's body [72][73][74]. In a study of 2043 dog and cat owners in the USA, more than half reported that the veterinarian was their primary support in relation to pet dying and death [74]. A systematic review of 19 qualitative papers from 17 studies found that when clients reported positive interactions and high levels of support from veterinarians, they were better able to trust and collaborate, felt more reassured, felt better able to grieve and experienced reduced trauma [75]. Discussions around euthanasia, including the sharing of bad news, quality of life assessment, end-of-life decision making, and comforting grieving clients take time [76][77][78], yet the predominant advice given to veterinary team members was to reduce direct contact time with clients. Due to the circumstances of the pandemic, clients may have wished for more time with veterinary team members. It is possible that the human-animal bond intensified due to changes brought about by the pandemic, including spending more time with companion animals due to working from home, or loss of employment [79]. Isolation may have intensified grief over loss of an animal, particularly owners for whom that animal was their only source of comfort or companionship [79]. Respondents reported that, in some cases, the prospect of no-contact euthanasia was a reason for clients to refuse euthanasia. This may have led to situations where euthanasia was delayed, or animals suffered a bad death (dysthanasia). Where veterinary team members had no alternative to no-contact euthanasia (for example, the ability to perform low-contact euthanasia or refer to a service provider who could do so), this likely caused moral distress. It is possible that in such situations, veterinarians continued to treat animals despite poor welfare, or what they felt was futility of treatment. Previous studies have reported that situations in which a client wished to continue treatment despite a patient's poor quality of life are experienced as ethically challenging by veterinarians [22,24]. Strengths and Limitations Limitations of the larger study from which the data discussed in this paper have been discussed at length elsewhere [12,13]. For the purposes of the current discussion, a key limitation was the anonymity of the survey, precluding the opportunity to clarify responses, and explore the social, cultural and contextual factors influencing whether respondents experienced low and/or no-contact euthanasia as ethically challenging. Of the subset of respondents who did report experiencing low and/or no-contact euthanasia as an ethical challenge in the early months of the pandemic, we did not have the opportunity to interview them regarding their experiences, what might have helped them in navigating low and/or no-contact euthanasia, and what they learned from the experience. The focus of our study was ethically challenging situations in general, not specifically low and nocontact euthanasia, which may have limited the extent to which respondents elaborated on this particular topic. However, anonymity may have facilitated more open, honest responses, removing social desirability bias. The voluntary nature of the survey predisposes it to self-selection bias, whereby those with stronger views on ethically challenging situations may have been more likely to respond. It is possible that those who had more negative views about or experiences with low and no-contact euthanasia were more likely to respond to the survey. Alternatively, those distressed by their experiences may have avoided responding due to concerns about recalling distressing ethical challenges. While every effort was made to distribute the survey globally, responses came from veterinary team members based in 22 countries. Results are biased towards wealthy, Western countries where the majority of veterinary teams work with companion animals [80][81][82][83][84]. This study may not reflect the experiences of veterinary team members working in other countries. Nonetheless, this study captures the experiences of veterinary team members from multiple countries during the early months of the global pandemic. It provides a snapshot of ethical challenges around low and no-contact euthanasia at a unique time in history. By its design, it does not document the evolution of ethical challenges faced by veterinary team members during the pandemic. Data were collected in the context of a shortage of PPE, prior to the identification of variants including Delta and Omicron, the availability of vaccinations, and the availability of rapid antigen tests, any and all of which have the potential to modify the likelihood of infection and therefore impact the way veterinary team members and clients behave and interact, including in euthanasia consultations. This study did not capture the experiences of veterinary clients. While our study provides evidence that low and no-contact euthanasia was a source of stress for veterinary team members, other studies show that low and no-contact veterinary consultations in general were anticipated to be or experienced as stressful by animal guardians/owners. In a study of low-income pet guardian's experiences at private veterinary clinics and hospitals during the pandemic, interviewees highlighted their inability to accompany the animal during the visit as a stressor both for themselves and for the animal [85]. Interviewees also reported challenges communicating with veterinary team members over the phone. A survey of 2254 pet owners in the US in a similar time period to this study (April to July 2020) reported that 13% of owners had concerns about accessing veterinary care during the pandemic [79]. Such concerns included protocols that precluded pet owners from accompanying animals during appointments, particularly euthanasia appointments. A qualitative study of Canadian pet owners with (dis)abilities found that for some, their (dis)ability (e.g., sensory, cognitive or motor) posed a barrier to virtual or telephone consultations or commuting to veterinary clinics during the pandemic [86]. A number reported that the inability to accompany their animals into the veterinary hospital led to distress, and reduced their willingness to access veterinary care [86]. Prior to the pandemic, viral posts on social media platforms Twitter and Facebook implored owners to stay in the room when their companion animals were euthanised, otherwise their pet's final moments may entail "frantically looking around for their owners" [87]. These posts assume that owners have a choice as to whether to be present during euthanasia but may serve to exacerbate owner distress in situations where this choice is removed, such as in a pandemic. In light of negative experiences and profound psychological harms suffered by bereaved family members, numerous authors emphatically recommend development of protocols, policies and guidelines to preserve of the ability of family members to visit dying loved ones in healthcare settings and/or be present at the time of death during this and future pandemics [33,37,38,65,66,69]. Recommendations Our study highlights a strong need to prepare veterinary team members to navigate ethical challenges presented by low and no-contact euthanasia. As the way euthanasia was discussed and ultimately performed was the main source of concern, we believe that providing further information and training on low-contact euthanasia may help veterinary team members preserve the ability of clients to accompany animals during euthanasia should they wish to do so. This would, in turn, enable veterinary team members to perform euthanasia in alignment with their values, thereby reducing moral distress. Specific guidelines as to how to assess risk, communicate about and perform lowand no-contact euthanasia in different circumstances, including for example pre-visit pharmaceutical and sedation protocols and checklists for preparing clients, could be included in future editions of guidelines such as the American Veterinary Medical Association's Guidelines for the Euthanasia of Animals [19], and/or updated biosecurity guidelines distributed by veterinary professional organisations [88]. At various stages in the pandemic, and in some cases throughout the pandemic, veterinary teams have been understaffed and under-resourced, with little time to digest and implement extensive guidelines [12,89,90]. Information contained in biosecurity guidelines and protocols needs to be as accessible as possible. There is an opportunity for professional organisations and continuing professional development providers to train veterinary team members in implementing such guidelines. This includes undertaking risk management with regard to euthanasia consultations, and communication and euthanasia techniques for situations where contact with clients must be minimised or eliminated. Low and nocontact euthanasia guidelines and risk assessment tools should be incorporated in hospital emergency plans. It would be beneficial if accessible information for clients, for example around the wearing of PPE in the euthanasia consultation, could be developed alongside guidelines and protocols for low and no-contact euthanasia. This information may help reduce miscommunication around practical matters, such as instructing clients how to wear PPE during a consultation and explaining expectations around physical distancing. It may also enable clients to better prepare for euthanasia, particularly if shared ahead of the consultation where possible. Additionally, customisable templates providing contact details of local support services could be made available to veterinary teams to provide to their clients, ensuring that these details are consistently and accurately communicated. Veterinary team members performing or assisting in low-contact euthanasia will require a reliable supply of PPE for themselves, and potentially any clients present. Veterinary professional organisations may have a role in helping to secure a continuous supply chain of appropriate PPE. Conclusions The identification of low and no-contact euthanasia as ethical challenges by over one fifth of respondents underscores that it isn't just the indications for euthanasia, but the practical aspects of how it is performed, that may be ethically challenging and potentially lead to moral distress for veterinary team members. Wherever possible, no-contact euthanasia should be avoided. Veterinary team members should be better prepared and equipped to perform low-contact euthanasia in the context of this and future pandemics.	2022-02-26T00:04:00.895Z	2022-02-23T00:00:00.000	{ "year": 2022, "sha1": "7a668c13592fc46dd18c126b5d31ff1f08c42ab1", "oa_license": "CCBY", "oa_url": "https://www.mdpi.com/2076-2615/12/5/560/pdf", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "80c7bfa579527a65354cb7bf4f5a9d8d0c96d39f", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
2962988	pes2o/s2orc	v3-fos-license	A revision of the genus Kaszabister Mazur (Histeridae, Histerinae, Exosternini) Abstract We revise the four species of Kaszabister Mazur, 1972, one of which, Kaszabister barrigai sp. n., is described as new. The other species in the genus are Kaszabister rubellus (Erichson, 1834), Kaszabister ferrugineus (Kirsch, 1873) and Kaszabister carinatus (Lewis, 1888). The species are principally known from the subtropics of South America, with one in Central America. Lectotypes are designated for Kaszabister rubellus and Kaszabister ferrugineus, and a key is provided for all the species. Ants of the genus Solenopsis Westwood, mainly Solenopsis invicta Buren and Solenopsis saevissima (Smith), are documented as hosts of three of the four species. Introduction The genus Kaszabister Mazur, 1972, was initially described for a single species, now known as K. rubellus (Erichson, 1834), and assigned to the myrmecophilous and termitophilus subfamily Haeteriinae. Later, two other species originally described in other genera (Epierus ferrugineus Kirsch, 1873, and Phelister carinatus Lewis, 1888) were moved here, and the entire genus was moved to Histerinae: Exosternini (Mazur 1997). This dynamic taxonomic history underscores the enigmatic nature of this genus, and its affinities within the tribe remain unclear. Similarly, little is known about the biology of these species, although it has become clear that they live as guests in the nests of fire ants of the genus Solenopsis Westwood. These ants are a major nuisance as a result of invasion into many tropical and subtropical climates (Hölldobler and Wilson 1990). It is therefore important to better understand the commensal predators that may play a role in controlling ant populations. This paper represents the first installment of an ongoing revision of all the species of New World Exosternini. The fauna is large and complex, with the limits of most genera poorly understood. Kaszabister is among the more straightforward and clearly monophyletic groups, and preliminary analyses (Caterino et al. in prep) indicate that it lies outside any other named genus. Thus we are confident in retaining its status as a genus. Materials and methods The morphological terminology used is that defined by Wenzel and Dybas (1941), supplemented by Helava et al. (1985), Ôhara (1994) and Lawrence et al. (2011). Following histerid conventions, total body length is measured from the anterior margin of the pronotum to the posterior margin of the elytra (to exclude preservation variability in head and pygidial extension), while width is taken at the widest point, generally near the elytral humeri. Type material of all valid species was examined by one or more of the authors. Photographic imaging was done using a Visionary Digital's 'Passport' portable imaging system, which incorporates a Canon D7 with MP-E 65mm 1-5X macro zoom lens. Images were stacked using Helicon Focus software. SEM imaging was done on a Zeiss EVO 40 scope, with most specimens sputter coated with gold. Photographs of all type specimens are available through the Encyclopedia of Life (www.eol.org). Specimens from the following institutions were utilized: Description. Body length 1.7-2.3mm, width 1.3-1.7mm, oval or oblong, more or less convex, reddish brown, glabrous. Head: Frons bordered by a prominent, moderately to strongly carinate frontal stria; antennae inserted under the rim of the frons in front of eyes; antennal scape slightly setose; antennal club oval, tomentose, lacking sutures or annuli, with small oval subapical sensoria on upper and lower surfaces; epistoma flat to convex, bordered by striae or carinae; labrum short, broad, rounded at sides and emarginate at middle; mandibles with strong furrows along lower outer margins and very weak subapical teeth on incisor edge; gena setose and weakly depressed; gular sutures impressed; submentum with numerous fringed setae, projecting slightly between maxillar bases; mentum about one-fourth broader than long, sides rounded, tapering apically, margin faintly emarginate; palpi relatively short, with truncate apices. Pronotum: pronotum widest at base, sides rounded, anterior angles acute; prescutellar impression absent; gland openings annulate, situated about one-third from anterior margin, behind inner edge of eye on each side; with 3 pores along each side; marginal stria complete, continuous with anterior marginal stria; lateral stria absent. Elytra: Dorsal striae of elytra simple or carinate, variously abbreviated; dorsal stria 1, subhumeral striae, and epipleural stria carinate and convergent apically. Prosternum: Antennal cavities of the prosternum visible in ventral view, located in the anterior angles of pronotum; prosternal lobe short, broad, reaching hypomeron laterally, with marginal stria at least medially; base of prosternal keel weakly emarginate; with complete carinate striae diverging anterad and posterad, not joined. Mesoventrite: Disk flat, weakly projecting at middle, with complete marginal stria; mesometaventral stria present, angulate forward onto mesoventral disk. Metaventrite: Metaventral disk with postcoxal and lateral striae, both extending laterad toward metepisternum. Abdomen: Propygidium short, moderately convex, with annulate gland openings in anterolateral corners; pygidium lacking apical stria; abdominal ventrite 1 with one or two lateral striae; ventrites 2-5 with or without posterior marginal striae. Legs: Protrochanter with seta; protibia lacking teeth, but with 8-10 stout marginal spines; protibial spurs short, strong; protarsus with fine ventral spines, pretarsal claws simple and equal; meso-and metatrochanters lacking setae; mesoand metatibiae nearly parallel-sided, the former with few weak marginal spines; mesoand metatarsomeres with single pair of apicoventral setae. Male: Eighth tergite with accessory sclerites, shallowly narrowly incised at subtruncate apex, with basal membrane attachment distad basal emargination; ventral apodemes of 8 th tergite broadly rounded, not meeting at midline; 8 th sternite approximately parallel-sided, halves not joined along midline, with apical guides gradually more strongly elevated toward apices; 9th tergite with median emargination deep, ventral apodemes situated just behind midpoint, strongly toothed; spiculum gastrale (9 th sternite) narrowed in distal two-thirds, with thin apical arms and short median apical flanges; halves of 10 th tergite well developed, separated along midline. Female: Eighth tergite forming a single, apically emarginate plate; 8 th sternite divided into two lateral plates, with thin, separate basal baculi which are articulated with the disk of S8; 9 th sternite present, elongate, connected to apex of S8; tenth tergite present, without basal apodemes; valvifers elongate, enlarged basally; coxites with two strong and one weak inner tooth; gonostyle present, free, setose; bursa copulatrix small; spermatheca short, sclerotized, forming a ventral concave disk over oviduct; spermathecal gland attached at base of spermatheca, elongate, gradually expanded to apex. Distribution. The distribution of the species of this genus is interestingly discontinuous, with three species concentrated in subtropical South America, and a single species from Central America, with few records from the northern half of South America. Diagnostic description. Length 1.9-2.3mm; width 1.5-1.7mm; Frontal stria descending onto epistoma as a strong carina, epistoma strongly depressed, depression broader than in K. rubellus; fourth dorsal elytral stria present in apical half to twothirds; inner subhumeral elytral stria present in apical two-thirds; fifth elytral stria generally absent; sutural stria present in apical half; elytral ground punctures denser and more uniformly distributed; mesometaventral stria arched forward to between one-half to one-third from anterior mesoventral margin; lateral metaventral stria strongly abbreviated mediad, ending about one-third from mesometaventral margin; inner postmetacoxal striae forming a complete, narrow arc across anterior margin of abdominal ventrite 1; abdominal ventrites 2-4 with apical marginal stria. Male: Aedeagus narrow, elongate, swollen toward base, flattened apically in lateral view. Distribution. Known principally from the Brazilian states of Mato Grosso and São Paulo, with one record each from Argentina and Paraguay, and a single specimen found in an Argentinian grape shipment at Kennedy Airport, New York, USA. Biology. While the majority of the type specimens say merely that they were collected from 'ant nest', there is little doubt that the hosts were Solenopsis Westwood, probably Solenopsis invicta Buren, 1972, the focus of the work of collectors Lennartz and Whitcomb . Other records specify 'fire ants', Solenopsis sp., and 'Solenopsis saevissima group' as hosts. Etymology. This species is named for the notable Chilean collector Juan Enrique Barriga-Tuñón, who provided us with one of the first known examples of the species. Remarks. Specimens labelled as 'São Paulo, Varzea Grande' most probably came from Vargem Grande do Sul in the same state. We were not able to find any place called Varzea Grande in São Paulo. On the other hand, there is a town of this name that is a part of greater Cuiabá, the capital of Mato Grosso, where the fire ant researchers who collected specimens from Mato Grosso and São Paulo were apparently based. But given that we have seen specimens collected on the same day, 1.xii.1972, in 'São Paulo, Vargem Grande' (3) and 'São Paulo, Varzea Grande' (1), we suspect confusion and mislabeling for the latter locality. Distribution. This species is restricted to Central America, from southern Mexico to Costa Rica. Biology. As for K. barrigai, many records of K. ferrugineus from 'ant nests' almost certainly refer to Solenopsis invicta as host, while a few other labels specify S. saevis-sima (Smith). One specimen from Bahia in the Lewis collection bears a label "with an Aphaenogaster" (Formicidae). However, no host specimen is present for verification. Remarks. Mazur (1997) cites "Phelister marginicollis Bruch, 1914" as "nom. nud. -syn. nov., Wenzel in litt.". Bruch (1914 cites this as "Phelister marginicollis Lew. in litteris". This appears to have been a species that Lewis intended to describe based on Bruch's specimens, and he indicated such to Bruch. Specimens labeled "P. marginicollis Lewis" are present in the BMNH and NMNH. However, for whatever reason, the species was never formally published, so it is indeed a nomen nudum. In Bruch's later works (e.g., 1935) he apparently realized that the species was not properly described and omitted it from his catalog. Remarks. The synonymy of K. mahunkai Mazur with Epierus rubellus Erichson was originally designated by Mazur (1984) (citing "Wenzel, in litt."). We have not studied the type of K. mahunkai first-hand, but Ottó Merkl of the HNHM very kindly compared the type specimen with our descriptions, keys and figures, and had no doubt that the synonymy is valid. His study also confirmed that Wenzel had compared types of K. mahunkai and E. rubellus side-by-side in coming to his original conclusion that the two were conspecific. Discussion Collecting records indicate that most of the species are strongly or exclusively associated with fire ants in the Solenopsis saevissima species group (excepting K. carinatus for which no ecological data is available). Most specimens appear to have been washed out of host mounds, and there have been no reported observations of any behavioral interactions of the Kaszabister beetles and their hosts. As all known histerids are predaceous (Kovarik and Caterino 2005), there can be little doubt that the beetles prey on their hosts, probably the larvae and pupae. From some focused collecting it appears that the density of these beetles can be relatively high, and it's conceivable that they provide a significant level of natural control of fire ant populations. Significant interest and resources have focused on introduction of parasitic flies as natural enemies of invasive fire ants in the United States (Callcott et al. 2011). None, to our knowledge, has considered predaceous myrmecophilous beetles. It should also be noted here that, although not as common a host as Eciton Latreille, Solenopsis does host other neotropical histerid genera, including Hippeutister Reichensperger and Procolonides Reichensperger (Helava et al. 1985, Kovarik and Caterino 2005, Caterino and Tishechkin 2008. Myrmecophily is a common phenomenon in Histeridae, with two entire large subfamilies (Haeteriinae and Chlamydopsinae) composed almost entirely of myrmecophiles (Kovarik and Caterino 2005). Outside these major groups, however, there have been numerous independent acquisitions of myrmecophilous habits in nearly all other major taxa. In Exosternini myrmecophily has arisen in both Old and New World genera, including Paratropus Gerstaecker, Coelocraera Marseul, Phelister Marseul, Pseudister Bickhardt and Tribalister Horn. Specialized myrmecophily is often associated with a suite of morphological specializations (Helava et al. 1985, Kovarik andCaterino 2005). Of these, only exaggerated striae/carinae (frontal, pronotal, elytral) present themselves in Kaszabister. Kaszabister exhibits a largely disjunct distribution, with very few records from a large area of northern South America. At this point it seems most likely that this relates to sampling effort. The vast majority of existing specimens have resulted from direct sampling of Solenopsis colonies, which has generally been focused on those areas with species that have become invasive elsewhere, primarily in southern Brazil. At the same time, fairly intensive passive trapping (flight intercept trapping and pitfall trapping) in some of these areas has resulted in no specimens (Vaz-de-Mello unpublished data, Flechtmann unpublished data). So although other parts of South America have seen significant trapping, by ourselves and others, there has been relatively little effort to collect in the appropriate ant colonies. So it remains to be seen whether Kaszabister is really more widespread but undetected in South America. This situation is very similar to that of Hippeutister (Histeridae: Haeteriinae), with species known from Central America and southern Brazil (Caterino and Tishechkin 2008). Dedicated collecting efforts in Solenopsis nests elsewhere may well turn up additional species in both these groups. An extensive collecting of Kaszabister in host ant colonies in 1972 in Mato Grosso has revealed some interesting facts about sympatry and syntopy of its species. Out of nine localities where K. barrigai was collected, at six K. ferrugineus was also found. Moreover, collecting colony codes reveal that both species have been collected in the same nests at Cáceres, Cuiabá, Mato Grosso Co., and Poconé. At the six localitites where both species were found K. barrigai was much more common, outnumbering K. ferrugineus by 100 to eight specimens. The situation at another locality, Arenapolis, was completely reversed: 23 specimens of K. ferrugineus were found in three nests there but no K. barrigai specimens are known from that site. Although nothing else known about this remarkable syntopy, the abundance patterns are suggestive of either interspecific competition, or very fine habitat and/or host preferences. Kaszabister ferrugineus and K. rubellus are also known to co-occur in at least one location; both were collected at Rosas, Argentina, represented by seven and six specimens, respectively. However, due to poor labeling, it is unclear whether they coexist microsympatrically or simultaneously, e.g. in the same colony or at the same time. This would be interesting to explore further. The phylogenetic position of Kaszabister within Exosternini has never been addressed. Ongoing analyses suggest a relationship with another inquilinous and enigmatic genus, Mecistostethus Marseul, together as sister or near sister to the large genus Operclipygus Marseul. However, some of the morphological characters these share may be convergences related to myrmecophily. Final analyses remain ongoing, and a confident result with respect to Kaszabister will probably have to await the availability of molecular sequence data.	2016-05-13T22:46:33.677Z	2012-04-06T00:00:00.000	{ "year": 2012, "sha1": "48fcacf87308c50bfe7412aeb6a7e881b32ec804", "oa_license": "CCBY", "oa_url": "https://zookeys.pensoft.net/lib/ajax_srv/article_elements_srv.php?action=download_pdf&item_id=2838", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "c70ca360be7d13f27763ebed646ff1970446b086", "s2fieldsofstudy": [ "Biology" ], "extfieldsofstudy": [ "Biology", "Medicine" ] }
220325924	pes2o/s2orc	v3-fos-license	Preconditioning beef cattle for long-duration transportation stress with rumen-protected methionine supplementation: A nutrigenetics study In general, beef cattle long-distance transportation from cow-calf operations to feedlots or from feedlots to abattoirs is a common situation in the beef industry. The aim of this study was to determine the effect of rumen-protected methionine (RPM) supplementation on a proposed gene network for muscle fatigue, creatine synthesis (CKM), and reactive oxygen species (ROS) metabolism after a transportation simulation in a test track. Angus × Simmental heifers (n = 18) were stratified by body weight (408 ± 64 kg; BW) and randomly assigned to dietary treatments: 1) control diet (CTRL) or 2) control diet + 8 gr/hd/day of top-dressed rumen-protected methionine (RPM). After an adaptation period to Calan gates, animals received the mentioned dietary treatment consisting of Bermuda hay ad libitum and a soy hulls and corn gluten feed based supplement. After 45 days of supplementation, animals were loaded onto a trailer and transported for 22 hours (long-term transportation). Longissimus muscle biopsies, BW and blood samples were obtained on day 0 (Baseline), 43 (Pre-transport; PRET), and 46 (Post-transport; POST). Heifers’ average daily gain did not differ between baseline and PRET. Control heifer’s shrink was 10% of BW while RPM heifers shrink was 8%. Serum cortisol decreased, and glucose and creatine kinase levels increased after transportation, but no differences were observed between treatments. Messenger RNA was extracted from skeletal muscle tissue and gene expression analysis was performed by RT-qPCR. Results showed that AHCY and DNMT3A (DNA methylation), SSPN (Sarcoglycan complex), and SOD2 (Oxidative Stress-ROS) were upregulated in CTRL between baseline and PRET and, decreased between pre and POST while they remained constant for RPM. Furthermore, CKM was not affected by treatments. In conclusion, RPM supplementation may affect ROS production and enhance DNA hypermethylation, after a long-term transportation. Introduction Long-distance road transportation (> 400 km) is a stressful event that most cattle experience at least once in their lifetime [1]. It is important to consider that road transportation includes immediate precursor of creatine. In skeletal muscle, the dystroglycan complex works as a transmembrane linkage between the extracellular matrix and the cytoskeleton. α-dystroglycan is extracellular and binds to laminin in the basement membrane, while β-dystroglycan is a transmembrane protein and binds to dystrophin. Dystrophin attaches to intracellular actin cables. In this way, the dystroglycan complex, which links the extracellular matrix to the intracellular actin cables, is thought to provide structural integrity in muscle tissues. The neuronal form of nitric oxide (nNOS) inhibits GAMT and the enzyme creatine kinase present in muscle (CKM), responsible for the conversion of creatine to phosphocreatine which releases the energy (ATP) required to cope against muscle fatigue (Fig 1). The objective of this study was to analyze the effect of administration of RPM on a proposed muscle fatigue gene network during pre-and post-exposure to a long-duration transportation simulation. Furthermore, we aimed to determine if RPM induces changes in expression of genes related to oxidative stress and DNA methylation. Materials and methods All the procedures for this study were conducted in accordance with a protocol approved by the Institutional Animal Care and Use Committee of Auburn University (IACUC Protocol #2017-3129). Eighteen Angus × Simmental heifers were divided in two groups of 9 each, with the same average body weight (BW; 408 ± 64 kg) and age (375 ± 44 days old) per treatment. Animals belonged to the E. V. Smith Research Center Beef Unit located in Shorter, AL (36075), and they were relocated to the Beef Cattle Evaluation Center, Auburn University, Auburn, AL (36849) for the duration of the study. After a successful 21-day adaptation period to a Calan Gate System (Northwood, NH), both groups received two pounds per day of a diet consisting of soyhulls (50%) and corn gluten feed (50%) with bermudagrass hay ad libitum. Additionally, during the following 45 days, nine heifers received supplement containing 8 gr/ hd/day as a fixed rate of top-dressed RPM (Smartamine M, Adisseo Inc., Antony, France), whereas the remaining nine heifers received supplement without the addition of RPM (CTRL). After 45 days on treatment, both groups were loaded onto a 32' × 7' steel gooseneck trailer (Circle W trailers, McKenzie, AL 36456) at a stocking density of 1.15 m 2 /hd. Subsequently, animals were transported via pickup truck and trailer from the Beef Cattle Evaluation Center located at 405 Shug Jordan Parkway, Auburn, AL 36832 to the National Center for Asphalt Technology (NCAT) Test Track located at 1600 Lee Road 151, Opelika, AL 36804 (~18 miles apart). The transportation simulation took place on a 1.7-mile oval test track at NCAT, where animals remained loaded and transported for 22 hours, at a constant speed of 45 mph. A team of three truck drivers followed a modified version of the NCAT driver safety schedule which consisted of stops once per hour (5 minutes per stop) to rest, and every three hours for fuel (~10 minutes per fuel stop). Every time the truck stopped, visual evaluations of the animals were performed to check their well-being (i.e. panting, respiration rate, proportion of heifers resting or standing). Animals were weighed at the beginning of the adaptation period to Calan Gate System (Day-21), at the beginning of the study (Day 0), 2 days before transportation (Day 43, pretransportation, PRET), and immediately after transportation (Day 46, post-transportation, POST). The first skeletal muscle biopsy was performed at the beginning of the administration of RPM (Day 0 or baseline-BASE). The second biopsy was performed 2 days prior to transportation simulation (PRET), and a final biopsy was performed right after unloading the trailer following transportation simulation (POST). These biopsies were taken in order to assess muscle fatigue-related gene expression changes due to RPM supplementation throughout the preconditioning period, during and after transportation. Animals' IDs were recorded so that the same animals were biopsied at each time point to provide repeated measures sampling. Skeletal muscle biopsies of Longissimus dorsi muscle were performed for epigenetic regulation (DNA methylation) and gene expression analysis using real-time quantitative PCR (RT-qPCR). Each muscle sample was obtained from the left side and moving up the side toward the head (longitudinally) with each sequential biopsy. Anesthesia injections of 5 mL of Lidocaine HCl 2% (VetOne1, Boise, ID) were placed before perform each biopsy incision. A muscle biopsy core (500-600 mg) was removed using a biopsy needle (12-gauge × 16 cm) inserted in a biopsy gun (Bard Magnum MG#1522, Tempe, AZ). The first biopsy area coincided with the Longissimus dorsi muscle at the level of the last rib. The second biopsy was performed 10 cm towards the head from the first biopsy area, and the final biopsy was 10 cm towards the head from the second biopsy area. The incision sites for all biopsies were located 5 cm down the vertebral column with the aim to take samples from the center of the Longissimus dorsi muscle (Fig 2). Ten mL blood samples were collected via jugular venipuncture at BASE, PRET and POST to determine hormonal and metabolic changes. Cortisol concentrations were measured at the Auburn University Endocrine Diagnostics Laboratory with the IMMULITE 2000 (Siemens Healthcare Diagnostics, Deerfield, IL, USA), which uses a solid-phase competitive enzyme-amplified chemiluminescent immunoassay. Creatine kinase and glucose analysis were performed at Auburn University Clinical Pathology Laboratory using methods previously described [12,13]. Statistical analysis Real time quantitative PCR data were analyzed using the MIXED procedure of SAS (SAS 9.4 Institute, Cary, NC, USA). Before statistical analysis, RT-qPCR data was normalized using quantities values of expression of each gene under study, divided by the geometric mean of 3 housekeeping genes (i.e., UXT, MTG1 and RPS15A). Then, normalized data was transformed to fold-change relative to day 0 (i.e. BASE). To estimate standard errors at day 0 and prevent biases in statistical analysis, normalized RT-qPCR data were transformed to obtain a perfect mean of 1.0 at day 0, leaving the proportional difference between the biological replicate. The same proportional change was calculated at all other time points to obtain a fold-change relative to day 0. Fixed effects in the statistical model for each variable analyzed (i.e. genes, blood metabolite) included treatment, time on experiment and treatment × time on experiment interactions when appropriate (e.g. mRNA expression over time). Gene expression data analysis included a repeated-measures statement with an autoregressive covariate structure. Animal performance and blood metabolite levels were also analyzed using the MIXED procedure of SAS, and treatment was the fixed effect in the statistical model. The random effect in all models was heifer within treatment. Average daily gain data was analyzed using PROC GLM procedure of SAS. The statistical model used was: Yijl = μ + Ci + Tj + Sl + (C × T)ij + εijl; where, Y ijl is the background-adjusted normalized fold change or blood data value; μ is the overall mean; C i is the fixed effect of time (3 levels); T j is the fixed effect of treatment (2 levels); S l is the random effect of heifer nested within treatment; C × T is the interactions of time by treatment and ε ijl is the random error (0, σ e 2 ) associated with Y ijl . Animal performance Body weight did not present a treatment × time interaction (P = 0.44), although it showed a decrease after transportation in CTRL and RPM heifers (P < 0.01) (Fig 3). Shrink loss was calculated by difference in body weight before and after transportation. Control heifers lost 43 kg (10% shrink), while RPM heifers lost 35 kg (8% shrink). The P values for shrink loss in kilograms and by percent were not significant (P = 0.29 and P = 0.28, respectively). Furthermore, average daily gain (ADG) did not differ between treatments (P = 0.41) during the preconditioning period (i.e., between BASE and PRET). Heifers in RPM group had an ADG of 0.98 kg/ day, whereas heifers in CTRL an ADG of 1.12 kg/day(S1 Fig). Serum cortisol, glucose, and creatine kinase Glucose levels showed a treatment × time interaction (P < 0.05) and a time effect (P < 0.05). There was not a significant difference (P = 0.14) in glucose concentration in both treatments between BASE and PRET Heifers in RPM group had a significant increase (P < 0.05) in glucose levels between PRET (83.9 mg/dL) and POST (98.8 mg/dL); and heifers in CTRL did not show a significant difference between PRET (91.7 mg/dL), and POST (96.8 mg/dL) (Fig 4). Circulating creatine kinase levels did not show a treatment × time interaction (P = 0.16). However, creatine kinase concentration showed a time effect (P < 0.05). Both groups showed a significant increase between PRET and POST (Fig 4). Finally, serum cortisol concentrations did not present a treatment × time interaction (P = 0.68). However, there was an important reduction in cortisol levels after transportation in CTRL and RPM heifers (Fig 4). (P < 0.05) between BASE and PRET, while SLC6A8 decreased between PRET and POST for CTRL but did not change for RPM. Finally, SLC6A8 had a tendency for a time effect (P = 0.07) (Fig 7). DNA methylation. There was a treatment × time interaction (P = 0.05), and a time effect (P < 0.01) for DNMT1 mRNA expression. Furthermore, DNMT3A showed a treatment × time interaction (P < 0.01), a treatment effect (P = 0.05), and a time effect (P < 0.01). Furthermore, there was an increase of DNMT1 mRNA expression in RPM between PRET and POST. For DNMT3A, CTRL had a greater expression (P < 0.05) at PRET than RPM. Finally, DNMT3A mRNA expression decreased (P < 0.05) in RPM between BASE and PRET. Similarly, DNMT3A expression was downregulated (P < 0.05) in CTRL between PRET and POST. There was a significant treatment × time interaction (P < 0.01), treatment effect (P < 0.05), and time effect (P < 0.01) for AHCY mRNA expression. In addition, AHCY mRNA expression Expression of genes related to the dystrophin-glycoprotein complex in Longissimus dorsi muscle of beef heifers that received rumen-protected methionine (RPM) for 45 days before transportation and heifers control (CTRL) that did not received RPM. Statistically significant differences were declared at P < 0.05 and tendencies at P > 0.05 and < 0.1. Treatment × time interaction ( �� ), time effect ( �� ) and treatment effect ( � ). Tendencies are denoted if symbols ( � , �� or �� ) are underlined. Symbols (▲) on lines denote significant differences (P < 0.05) between two time points for the same treatment, symbols (•) denote significant differences (P < 0.05) between treatment at the same time point. Because the data presented in this chart are fold-change values, the baseline reference value is 1 for RPM and CTRL, therefore, there is overlap between symbols with error bars at baseline. in CTRL was greater (P < 0.05) than RPM at PRET. There was a significant downregulation of AHCY (P < 0.05) in RPM between BASE and PRET; however, it was significantly upregulated (P < 0.05) between PRET and POST (Fig 8). Oxidative stress. There was a significant treatment × time interaction (P < 0.05), and a time effect (P < 0.01), for SOD1, NQO1, and NOS3 mRNA expression. Furthermore, SOD2 showed a significant treatment × time interaction (P < 0.01), and treatment effect (P < 0.05). There was not a significant treatment × time interaction (P > 0.05) for PGC1α and NFKB1. At PRET, CTRL had a greater expression (P < 0.05) for SOD2, and NQO1 expressions in comparison to RPM. There was a significant increase (P < 0.05) in SOD1, NOS3, NFKB1, and PGC1α mRNA expressions in CTRL between BASE and PRET. In CTRL heifers, SOD2 mRNA expression presented upregulation (P < 0.05) between BASE and PRET; however, its expression was Expression of genes located at the sarcoplasmic reticulum in Longissimus dorsi muscle of beef heifers that received rumen-protected methionine (RPM) for 45 days before transportation and heifers control (CTRL) that did not received RPM. Statistically significant differences were declared at P < 0.05 and tendencies at P > 0.05 and < 0.1. Treatment × time interaction ( �� ), time effect ( �� ) and treatment effect ( � ). Tendencies are denoted if symbols ( � , �� or �� ) are underlined. Symbols (▲) on lines denote significant differences (P < 0.05) between two time points for the same treatment, symbols (•) denote significant differences (P < 0.05) between treatment at the same time point. Because the data presented in this chart are fold-change values, the baseline reference value is 1 for RPM and CTRL, therefore, there is overlap between symbols with error bars at baseline. https://doi.org/10.1371/journal.pone.0235481.g006 downregulated between PRET and POST. In the case of SOD2 expression in RPM, a significant decrease (P < 0.05) in expression was observed between BASE and PRET, which remained steady between PRET and POST. Finally, NQO1 mRNA expression increased significantly (P < 0.05) in CTRL between BASE and PRET; however, a significant decrease (P < 0.05) was observed between PRET and POST, while it remained steady for RPM group (Fig 9). Animal performance With respect to animal performance, CTRL group showed no difference in ADG compared with RPM group (1.12 kg/d vs. 0.98 kg/d, respectively). There are no previous studies that Expression of genes related to the creatine synthesis pathway in Longissimus dorsi muscle of beef heifers that received rumen-protected methionine (RPM) for 45 days before transportation and heifers control (CTRL) that did not received RPM. Statistically significant differences were declared at P < 0.05 and tendencies at P > 0.05 and < 0.1. Treatment × time interaction ( �� ), time effect ( �� ) and treatment effect ( � ). Tendencies are denoted if symbols ( � , �� or �� ) are underlined. Symbols (▲) on lines denote significant differences (P < 0.05) between two time points for the same treatment, symbols (•) denote significant differences (P < 0.05) between treatment at the same time point. Because the data presented in this chart are fold-change values, the baseline reference value is 1 for RPM and CTRL, therefore, there is overlap between symbols with error bars at baseline. https://doi.org/10.1371/journal.pone.0235481.g007 report the performance of beef heifers under RPM supplementation; however, our results support Chen et al. (2011), who evaluated RPM supplementation (Smartamine1) in dairy cows, and reported no difference in ADG with a control group [15]. In addition, Torrentera et al. (2017) tested the performance of Holstein steers (127 ± 4.9 kg) under the inclusion of 0.032%, 0.064%, 0.096%, or 0.128% of RPM. They obtained superior results with RPM supplementation when Smartamine1 was included at a level of 0.096% of dietary dry matter (DM) [16]. In our study, the RPM supplementation was 0.07% of DM; inclusion up to 0.09% might possibly have results that are more notable. The most important parameter to assess, from the beef production standpoint, is the body weight loss or shrink associated with transportation. Numerous factors affect shrink during Expression of genes related to DNA methylation in Longissimus dorsi muscle of beef heifers that received rumen-protected methionine (RPM) for 45 days before transportation and heifers control (CTRL) that did not received RPM. Statistically significant differences were declared at P < 0.05 and tendencies at P > 0.05 and < 0.1. Treatment × time interaction ( �� ), time effect ( �� ) and treatment effect ( � ). Tendencies are denoted if symbols ( � , �� or �� ) are underlined. Symbols (▲) on lines denote significant differences (P < 0.05) between two time points for the same treatment, symbols (•) denote significant differences (P < 0.05) between treatment at the same time point. Because the data presented in this chart are fold-change values, the baseline reference value is 1 for RPM and CTRL, therefore, there is overlap between symbols with error bars at baseline. https://doi.org/10.1371/journal.pone.0235481.g008 Expression of genes related to oxidative stress in Longissimus dorsi muscle of beef heifers that received rumen-protected methionine (RPM) for 45 days before transportation and heifers control (CTRL) that did not received RPM. Statistically significant differences were declared at P < 0.05 and tendencies at P > 0.05 and < 0.1. Treatment × time interaction ( �� ), time effect ( �� ) and treatment effect ( � ). Tendencies are denoted if symbols ( � , �� or �� ) are underlined. Symbols (▲) on lines denote significant differences (P < 0.05) between two time point for the same treatment, symbols (•) denote significant differences (P < 0.05) between treatment at livestock transportation. Transportation duration and distance travelled are especially critical. In our study, animals were loaded on the trailer for 22 hours; however, because of the stops to rest, the net transportation time was 20 hours. Interestingly, Lambooy & Hulsegge (1988) studied the effect of 24 hours transportation in pregnant heifers, and reported body weight losses of 8% on average [17]. This result is similar to the shrink loss we observed for RPM heifers. Environmental conditions during the transportation simulation day was retrieved from AWIS (Agricultural Weather Information Service, Inc.) at the Auburn Mesonet weather station (Auburn_CR10_92-18, Lee County) which is located 5.2 miles from NCAT. The average temperature was 5˚C, relative humidity was 48.5%, and no rain was recorded during the 22 hours of transportation nor on the previous or subsequent days. According to Gonzalez et al. (2012), 20-25 hours transportation with an ambient temperature of~10˚C will produce a~8% of BW losses on feeder cattle, and~10% of BW losses on cull cattle [18]. Thus, the environmental conditions during our trial were not detrimental for body weight losses. However, RPM supplementation may affect shrink during warmer seasons in the Southeast. Serum cortisol, glucose, and creatine kinase Transportation is a stressful event that causes physiological and endocrine changes in animals' metabolism. One indicator of these changes is cortisol, a glucocorticoid hormone released from the adrenal gland in response to stress [19]. Several studies have shown an increase in cortisol levels due to transportation in cattle, especially in the first hours of the trip [19, 20]. The release of blood cortisol increases when the transportation starts, reaching its highest point at~4 hours, and then it starts to decline until reaching similar levels to the beginning of transportation. This level is maintained for the next 24 hours post-transportation [21]. However, these studies assessed the effect of shorter transportation time (i.e., 9 hours maximum), in comparison to a longer transportation period as in this study. Nevertheless, evidence of the reduction of blood cortisol level due to a long transportation could be associated to a familiarization of the animal with the environment inside the truck [22]. In total, our truck was stopped for 160 minutes during the 22-hour trip, either to rest or to refuel, and visual evaluations of the animals' status took place during these stops. The small amount of time that the truck was stopped, could be beneficial for both, driver and animals. Frequent rest stops with loading and unloading may be more stressful than continuously being in the vehicle, consequently, this practice would likely be detrimental for the well-being of the animals [8]. Blood glucose level is used as a metabolic indicator of energy status in cattle. Our results suggest that glucose levels in blood during the preconditioning period are comparable to previous studies that reported 80 (2010) reported an increase in circulating CK levels after 24-hours of transportation in Charolais bulls with BW of 367 ± 35 kg. In that study, before transportation, average CK level was 162.3 U/L, which is lower than in our study at PRET (CTRL, 227.4 U/L; RPM, 266.7 U/L). However, circulating CK after 24-hours transportation was 496 U/L, which is similar to heifers' POST CK levels measured in this study (CTRL, 406 U/L; RPM, 424.8 U/L) [25]. In conclusion, beyond the rise in glucose and CK levels after transportation, no significant effect was observed due to the addition of RPM in heifers' diet pre-transportation. Counter to our hypothesis, there was not a greater CK concentration, which could mitigate the effect of muscle fatigue due to greater amounts of creatine phosphate synthetized from RPM administration. Gene expression Dystrophin-glycoprotein complex. Dystrophin-glycoprotein complex (DGC) is a key element for the integrity of the sarcolemma, composed primarily of dystrophin and sarcoglycan complexes. The function of DGC is associated with the anchoring of sarcolemma to the actin cytoskeleton [30]. The DGC is formed by the dystroglycan complex, composed by dystrophin, the syntrophins, α-and β-dystroglycan; and sarcoglycan complex (SG), which is formed by four different transmembrane glycoproteins called α-, ß-, γ-, and δ -sarcoglycan Sarcospan (SSPN), a transmembrane protein, is a DGC member [35] with a role in stabilizing the sarcolemma. Additionally, SSPN may serve as a chaperone in the DGC assembly and may help enhance localization of other adhesion complex proteins [36]. To the best of our knowledge, there is no literature reporting the effects of fatigue on SSPN expression. Therefore, it would be possible that CTRL group had more stabilization in the sarcolemma by SSPN upregulation between BASE and PRET. However, in a previous study, SSPN deficient mice present normal creatine kinase levels, sarcolemma integrity, and force generation aptitude [36]. The gene expression regulation of Syntrophin isoforms is an important factor for the sarcolemma integrity and binding signaling proteins. If one isoform is missed or deleted, another can replace it with no functional effects observed [37]. The absence α-Syntrophin 1 (SNTA1) gene in mutant mice did not show reduced force recovery in a fatigue stimulation trial [38]. Mice lacking α-and ß-syntrophins (SNTA1 and SNTB1, respectively) produce a decrease in skeletal muscle function [37]. It is possible to infer that sarcolemma integrity might not be affected by RPM supplementation nor fatigue due to transportation, since both SNTA1 and SNTB1 expression did not differ between treatments. Similarly, ß-sarcoglycan (SGCB), a gene that encodes for a transmembrane protein in the sarcoglycan complex, showed a similar pattern of expression in comparison to SNTA1 and SSPN, indicating its possible correlation with the whole Dystrophin-Glycoprotein complex. Because SSPN was the only gene that showed a treatment effect x time interaction in the Dystrophin-Glycoprotein complex, it would not be accurate to declare that the sarcolemma maintains its stability in CTRL group during transportation due to the unique upregulation of SSPN at PRET. Consequently, administration of RPM may not alter skeletal muscle membrane integrity in growing heifers. Further analysis of more genes related to Dystrophin-Glycoprotein complex and their phenotypic effects will help to fully understand the mechanism behind changes on skeletal muscle membrane's integrity under this stressful condition. Sarcoplasmic reticulum. The regulation of Ca 2+ ions in muscle cells has been widely studied in different species; however, little is known about the effect of fatigue on beef cattle. In a previous study, a downregulation of CASQ1 protein expression in mice after an endurance trial [39] was reported. These results are consistent with our study for the CTRL group, because there was a downregulation of CASQ1 between PRET and POST. Similarly, reduction in activity of ATP2A1 in fatigued skeletal muscle was reported in rats [40] and humans [41]. Even though, we did not measure Ca 2+ levels in skeletal muscle, the downregulation of CASQ1, ATP2A1, and SYPL2 in CTRL during transportation may be related to Ca 2+ depletion after a repetitive use of muscle [42]. These responses were not observed in heifers that received RPM between BASE and PRET. It has been shown that, after a methylation process of sarcolemma throughout SAM as a methyl donor agent, the Na + -Ca + exchange is inhibited in the cardiac muscle. In skeletal muscle, phospholipid methyltransferase activity is highly localized in sarcoplasmic reticulum and present to a lesser extent in sarcolemma. Methylation of phosphatidylethanolamine modulates sarcoplasmic reticulum phospholipid composition, which in turn alters the energetic efficiency of SERCA (Ca 2+ uptake / ATP hydrolysis) ion pump, decreasing the skeletal muscle metabolic rate [43][44][45]. Our results can be explained by a possible inhibition of Ca 2+ uptake under an excess of methylation in the sarcoplasmic reticulum of RPM. The administration of RPM downregulated only one of the genes that we selected to study the sarcoplasmic reticulum complex, SYPL2. It is possible that the methyl groups provided by RPM inhibited SYPL2 expression. Nagaraj et al (2000) reported that wild-type mice and mice lacking MG29, the protein that SYPL2 encodes, exhibited lower force generation after fatigue in the diaphragm, a skeletal mixed-fiber muscle like the Longissimus Dorsi [46]. In addition, mice lacking SYPL2 did not show decreased force generation after fatigue compared with wild-type mice, indicating the effect of MG29 can be compensated in a muscle fatigue scenario for that species. In RPM, ATP2A1 and CASQ1 remained unchanged among time points, suggesting that the greater availability of methyl groups does not modify Ca 2+ regulation in skeletal muscle or inhibits the expression of genes related to sarcoplasmic reticulum. Skeletal muscle fatigue impairs Ca 2+ uptake as previously described [47]. Down-regulation in the expression of all the genes analyzed related to sarcoplasmic reticulum between PRET and POST in CTRL may suggest that Ca 2+ metabolism could be impaired under fatigue conditions. In contrast, RPM did not show a significant difference between PRET and POST. Thus, the stabilization of the expression of genes related to sarcoplasmic reticulum could potentially explain a better Ca 2+ uptake by the skeletal muscle [47,48]. Creatine synthesis pathway. To the best of our knowledge, there are no published nutrigenomics studies of muscle fatigue during transportation stress in cattle, and because of this, we utilized information from other speciesCreatine biosynthesis occurs mainly in the kidney, pancreas and liver, and involves two steps [49]. First, glycine amidinotransferase (GATM or AGAT) catalyzes the transamidination of guanidine group from arginine to glycine, producing ornithine and guanidinoacetate (GAA) [50]. Second, guanidinoacetate methyl transferase (GAMT) catalyzes the addition of a methyl group from S-adenosyl methionine (SAM) to GAA, producing creatine and S-adenosyl homocysteine (SAH) [51]. Once creatine is formed, it reaches different tissues via its transporter called solute carrier family 6 member 8 (SLC6A8) [52]. In our study, the difference in GATM expression can be explained by the maintenance of the expression in RPM at POST, in comparison to CTRL, which was downregulated. In mice, GATM upregulation in skeletal muscle does not compensate deficiency in SLC6A8 [53]. Thus, it is possible that the GATM expression in CTRL does not make up for decreased expression of CKM and SLC6A8 at POST. In addition, GAMT deficiency has a negative effect in force maintenance in both, highintensity electrical stimulation, and lower electrical stimulation frequency in mice [54]. The downregulation of GAMT expression in CTRL between PRET and POST would explain a decrease in creatine synthesis due to muscle fatigue since CKM expression was downregulated in CTRL for the same period of time. In skeletal muscle of Thoroughbreds horses, no difference in CKM expression was found under moderate and high intensity exercise on a treadmill [55]. Samples were taken before the exercise, immediately after exercise, and 4 hours later. It has been shown that muscle force generation vary depending on the activity of the muscle. For example, muscle generates larger forces at moments of high speeds (i.e. running) compared with lower speed of muscle use (i.e. walking) [56]. Consequently, the velocity of muscle contraction has an important influence in energy use. For example, energy expenditure during an active movement (i.e. walking) is greater compared with less motion activities (i.e. standing) [57]. In our study, animals were not doing an intense exercise but they were certainly contracting theirs muscles in order to maintain equilibrium inside the truck for a long period of time, it is possible to compare these results with ours. However, even though RPM treatment remained stable among time points for CKM expression, it is not possible to explain the upregulation in CTRL heifers between BASE and PRET. Interestingly, SLC6A8 shows the same pattern as CKM expression for both treatments among all time points. Since SLC6A8 encodes for the major creatine transporter, the decrease of creatine levels produces an inhibition of SLC6A8 expression, similarly to results reported in previous studies with creatine transporter deficiency in mice [52]. Oxidative stress. The generation of free radical elements such as Nitric Oxide (NO -) and superoxide (O 2 -) is a result of normal cellular metabolism [58]. Superoxide suffers reductions generating reactive oxygen species (ROS) [i.e., O 2 -, hydrogen peroxide (H 2 O 2 ) and, hydroxyl radicals (OH -)] during resting conditions. Nitric oxide can be inactivated by coupling with O 2 -, producing the oxidant proxynitrite (ONOO -) as a result [59]. Nitric oxide creates reactive nitrogen species (RNS) like peroxynitrous acid (HNO 3 -), ONOOand other nitrogen-derived oxidants [60]. These products are released in low concentration during resting, and they are necessary for normal cellular function [59]. For example, these compounds are involved in muscle regeneration, acting with other elements like growth factors and chemokines. Additionally, ROS promote mitochondriogenesis during exercise, which generates higher ATP levels for cell utilization [61]. However, skeletal muscle generates oxidants at high rates during muscle contraction, and they either became blockers of myogenic differentiation, or cause injuries by oxidative damage [62]. A single bout of exhaustive exercise leads to strong increases in ROS, which cannot be buffered by endogenous antioxidants, particularly in untrained individuals [63]. Therefore, high levels of ROS results in contractile dysfunction and fatigue [64]. In contrast, in our study, a prolonged period of water deprivation and utilization of energy reserves, which occur during long-term transportation of livestock, ultimately result in dehydration and a switch to a heightened gluconeogenic state while sparing muscle protein degradation [65]. The generated state of energy deprivation produces an increase in oxidative stress [66]. Exercise and fatigue increase muscle ONOOand O 2 producing an excess of oxidants in the cellular milieu. Nevertheless, cellular defense mechanisms against oxidative damage can effectively degrade these products [67]. Superoxide dismutases (SODs) are oxidoreductases generated by cells as a protection against oxidants. They are capable of spontaneous dismutation of O 2 into H 2 O 2. Consequently, they also help reducing the production of ONOO -. Mammals present three isoforms of SOD, and they are differentiated by their composition and site of action. The cytosolic SOD1 (Cu/ZnSOD) is the main intracellular dismutase, which has enzymatic activity dependent on the presence of copper and zinc [68]. The mitochondrial SOD2 (MnSOD) has enzymatic activity dependent on the presence of manganese. Manganese plays a catalytic role of the dismutation of O 2 to H 2 O 2 , similar to SOD1 [59]. The major site of O 2 generation is the mitochondrial respiratory chain; consequently, SOD2 is a very important regulator of ROS equilibrium in the cell [69]. Finally, an extracellular dismutase called SOD3 (ecSOD) catalyzes ROS in the extracellular matrix. It is mostly expressed in internal organs and blood vessels. Skeletal muscle does not present high expression of SOD3, thus, it was not subject to analysis in our study [59,70]. In our study, expression of SOD1 and SOD2 was significantly activated in CTRL heifers during the pre-conditioning period compared to RPM heifers. The lack of response in RPM heifers could be interpreted as a better controlled oxidant-antioxidant balance within skeletal muscle. We base our hypothesis on the observation that regular exercise appears to gradually increase the level of adaptation by the repeated activation of antioxidant genes and proteins [71]. We believe that after a continuous contraction of muscles during the 22 hours on the trailer, RPM heifers responded better than CTRL heifers in the adaption to a stressful situation and, they were able to reach homeostasis in their ROS production by the end of the study. Thus, activation of SOD1 and SOD2 was not in demand. As reported in our study, BW losses of 8% in RPM, and 10% in CTRL were caused by the transportation, and up to 50% of that BW loss may be tissue partitioning [18]. During the 45-day preconditioning period with RPM supplementation, most ROS-related genes had an upregulation pattern in CTRL heifers. Methionine is a precursor of S-adenosylmethionine and hydrogen sulfide, which alleviate oxidant stress and protect the tissue from the damage [72]. Therefore, RPM supplementation might prevent an oxidative stress response in the RPM group before and after transportation contrary to CTRL, which showed an upregulation in ROS-related genes during the same period of time. The upregulation in CTRL compared with RPM before transportation may be related to a lower antioxidant capacity due to a minor methionine residues content in skeletal muscle cells. Methionine residues are known as ROS scavengers [73]. The greater expression of superoxide agents in CTRL as compared to RPM may be related to the oxidative stress imbalance due to tissue partitioning in heifers' adipose and skeletal muscle tissues, due to the long-duration transportation without access to water and feed [74,75].Furthermore, the occurrence of oxidative stress due to starvation has been reported in different species [76][77][78]. Fasting rats had a higher hepatic oxidative stress imbalance due to different processes of tissue degradation such as lipid peroxidation, protein misfolding, and glycolysis [79]. In addition, ROS generation increased in skeletal muscle after 24 hours of fasting in mice [80]. Metabolic stress inherent to endurance exercise has been demonstrated to stimulate mitochondrial biogenesis [81]. Mitochondrial synthesis is stimulated by the PGC-1α-NRF1-TFAM pathway [63]. Therefore, PGC-1α is the first stimulator of mitochondrial biogenesis [82], and its expression increases with exercise training [83]. PGC-1α is a key player in the adaptation of muscle cells to exercise [63]. Oxidative stress increases expression of PGC-1α [84], and PGC-1α controls cellular antioxidant homeostasis by stimulating SOD2, catalase, glutathione peroxidase 1 (GPx1), and uncoupling protein (UCP) [85]. In our study, neither administration of RPM nor transportation stress affected PGC-1α expression, although the peak in PGC-1α expression at PRET in CTRL heifers leads us to suggest that RPM could be producing an inhibition in the expression of this gene. Furthermore, the presence of additional methionine could be producing a more balanced environment in the RPM heifer's skeletal muscle, where the activation of the oxidative stress-related gene network was not required. In contrast to our results, PGC-1α is upregulated in liver of fasting mice [86]. However, PGC-1α stabilization in RPM during transportation could mean that tissue partitioning caused by fasting metabolism was not related to the expression of this gen. Furthermore, the effect of mice lacking PGC-1α on substrate utilization from fed to fasted state was not relevant [87]. This could explain the lack of significant differences in PGC-1α. The generation of nitric oxide (NO) in endothelial cells is produced by nitric oxide synthetase Type III or eNOS, which is encoded by NOS3. [89]. Finally, NQO1 is a member of an antioxidant defense system. An acute bout of exercise at sufficient intensity activates this antioxidant enzyme, among others [90]. In contrast, in patients with chronic fatigue syndrome, NQO1 was downregulated [91]. Similar to other antioxidant genes assessed in this study, NQO1 expression remained unchanged among time points in RPM heifers. The higher concentration of available methyl groups provided by the addition of RPM could decrease ROS production. Consequently, the activation of NQO1 would not be necessary for the muscle cells after the transportation simulation was performed. DNA methylation. DNA methylation is the major regulator of epigenetics in the genome of mammalians by the covalent addition of a methyl group (-CH 3 ) to cytosine in combination of CpG dinucleotide called "CpG" island [92]. Enzymes capable of adding methyl groups to the hemimethylated and unmethylated DNA are called DNA methyltransferase (DNMT), like DNMT1 and DNMT3A, respectively [93]. Thus, DNMT1 and DNMT3A gene expression in skeletal muscle would potentially explain epigenetic changes due to muscle fatigue. DNMT3A presents a catalytic activity of de novo DNA methyltransferases while, DNMT1 works maintaining methylation of the CpG sites [94]. Furthermore, a cooperation between DNMT1 and DNMT3B was observed in humans; when disrupted, DNA methylation was decrease by 95% in colon rectal cancer cell [95]. The addition of RPM to the diet increases methionine availability [96]. Consequently, it can be transferred to S-adenosylmethionine (SAM), which is a molecule considered as the universal methyl donor [97]. Similar to our results, Osorio et al. (2014) did not obtain a significant difference in hepatic DNMT1 expression in periparturient dairy cows under Smartamine1 supplementation. Furthermore, in contrast to our results, hepatic DNMT3A showed a greater expression after 11 days under Smartamine1 supplementation during the pre-partum period [98]. However, the difference was mitigated after calving, and no difference was obtained from 11 to 31 days of supplementation. In addition, there was a similar pattern of expression of DNMT1 and DNMT3A in CTRL across time points, similar to results from Rhee et al (2002). Although, this pattern was not observed in RPM [95]. The AHCY gene provides instructions for producing the enzyme S-adenosylhomocysteine hydrolase, an important inhibitor of the majority of the methytransferases, affecting DNA methylation, RNA, and protein methylation in humans [99]. Specifically, S-adenosylhomocysteine hydrolase controls the chemical reaction that converts S-adenosylhomocysteine to adenosine and homocysteine [100]. This chemical reaction also plays an important role in regulating the addition of methyl groups to other compounds (i.e., methylation) [101]. Therefore, the principal role of AHCY in metabolism is to hydrolyse and efficiently remove S-adenosylhomocysteine, the by-product of transmethylation reactions and one of the most potent methyltransferase inhibitors [102]. S-adenosylhomocysteine hydrolase (AHCY) is the only mammalian enzyme able to hydrolyze S-adenosyl-l-homocysteine, and it binds to DNMT1 during DNA replication [101]. Hypermethylation of the genome can occur due to DNMT1 overexpression caused by AHCY regulation in vitro [101]. Furthermore, in isolated guinea pig heart muscles, adrenaline secretion inhibits S-adenosylhomocysteine hydrolase activity by a calcium mediated mechanism [103]. Stress can also directly influence the transcriptional regulation at the epigenetic level. In a previous study, prenatal transportation stress in Brahman bull calves was assessed at the methylome level. Results showed at least 10% more methylation in stressed calves as compared to controls [104]. In our study, RPM produce the inhibition of AHCY during preconditioning. However, under the stress of transportation, AHCY had activation in RPM heifers, coinciding with DNMT1 upregulation. Furthermore, in a previous study, the transcriptional repression of DNMT1 by glucocorticoid exposure is considered as a proxy for stress response [105]. Hypermethylation plays a role in controlling the physiological response to stressors by regulating the release of glucocorticoids in response to challenges [106]. Our results indicate that preconditioning with RPM supplementation might produce an activation of the process of DNA methylation during transportation stress. The animal's benefit due to this metabolic response (i.e., hypermethylation) to long-term transportation stress still needs to be elucidated. Conclusion There was not a significant reduction in shrink with RPM; however, the utilization of more animals in further studies, could make this difference significant. In addition, serum glucose, creatine kinase and cortisol levels were not affected by the administration of RPM, suggesting that the catalytic processes generated due to high stress level could not be diminished by a greater methionine availability. Contrary to our hypothesis, rumen-protected methionine during a preconditioning period of 45 days does not have an effect in CKM neither genes related to creatine synthesis in beef cattle transported for a period of 22 hours. Expression of genes related to muscle integrity and fatigue, creatine synthesis, sarcoplasmic reticulum and Dystrophin-glycoprotein complex were not affected by the administration of RPM in beef heifer's skeletal muscle. However, our results indicate that administration of 8 gr/hd/ day of RPM could promote a better response to transportation stress in heifer's skeletal muscle due to a possibly more stable redox balance after long-duration transportation. Perhaps, evaluating the Semitendinosus muscle, which might suffer more tension during transportation, would make these results more clear. However, we wanted to focus on the ribeye area (i.e., Longissimus dorsi) where the most expensive beef cuts are located. In conclusion, preconditioning growing Angus × Simmental heifers with RPM pre-transportation may have an effect on oxidative stress (SOD2), DNA methylation (DNMT3A) and hypermethylation (AHCY).	2020-07-04T13:05:42.021Z	2020-07-02T00:00:00.000	{ "year": 2020, "sha1": "291e242481286a39191754177d07c6ebf1cc2937", "oa_license": "CCBY", "oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0235481&type=printable", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "245eba0080c8ac508ee2e4d5c8630d8e56c8ef78", "s2fieldsofstudy": [ "Agricultural And Food Sciences" ], "extfieldsofstudy": [ "Biology", "Medicine" ] }
214639749	pes2o/s2orc	v3-fos-license	Lessons from a costs analysis in potato production *Corresponding author: Gilson Rogerio Marcomini, Professor, Federal Institute of Education Science and Technology of São Paulo (IFSP), São João da Boa Vista Campus, Lauro Sargaco Street, Jardim Santa Helena. CEP 13874057. São João da Boa Vista, SP. Brazil. E-mail: gilson.professor@yahoo.com.br Received: 01 July 2019; Accepted: 30 October 2019 R E S E A R C H A R T I C L E INTRODUCTION As the fourth largest crop by global production volume, the potato is one of the most cultivated horticultural crops worldwide. World potato production in 2016 was over 377 million tons, harvested from almost 19 million hectares, an average yield of around 20 tons per hectare (FAOSTAT, 2018). This global potato production was only smaller than whey (750 million tons in 2016), rice (740 million tons in 2016), and maize (1.06 billion tons in 2016) (FAOSTAT, 2018). It is a crop that is widely consumed worldwide, due to its ease of preparation, characteristics, and the innumerable possible ways in which potatoes can be consumed. Whether traditionally cooked, fried, or roasted, or as a component of other recipes, it is a product that all people can afford. The largest Brazilian producer state is Minas Gerais, with production of 1.26 million tons in 2016 in an area of 39,431 hectares; this represents about 33 percent of the national (Brazilian) production. The second largest Brazilian producer state is Parana, producing 775 thousand tons in an area of 30,249 hectares, with Sao Paulo in third place, producing 665 thousand tons in 21,651 hectares. These three states together produce 75 percent of Brazil's total production, which corresponds to about 2.7 million tons (IBGE, 2018). On the other hand, potato production in the US is higher than that in Brazil, and it ranks as the fifth largest producer in the world. The potato is produced in the US in three harvests throughout the year: in the autumn, spring, and summer. The fall harvest is the most important; about 376 thousand hectares are cultivated and produce about 18 million tons of potatoes (USDA, 2018). The main producing states in the US are Idaho, Washington, Wisconsin, North Dakota, and Colorado, which together account for 84 percent of the country's total potato production (USDA, 2018). The US potato market has the same characteristics as the Brazilian market, since prices fluctuate depending on the crop and the period. The potato production costs indicate that the average total costs in southwestern Idaho in 2010 was US$ 6,503.79 per hectare, while in 2017 the average total costs reached US$ 8,573.86 per hectare. Likewise, the total production costs in the Vargem Grande do Sul region in 2010 were US$ 11,581.59 per hectare, reaching US$ 13,209.44 per hectare in 2012, US$ 9,314.04 in 2015, and US $10,831.22 in 2017 Analyzing the structure of costs in the two areas shows that in southwestern Idaho, the highest costs are related to land leasing, fertilizers, pesticides, seeds, and machinery. Likewise, for the Vargem Grande do Sul region, the highest cost items are seeds, post-harvest and sorting, fertilizer, labor, and pesticides. Thus, it is possible to deduce that potato production in the US has advantages over Brazilian production, allowing it to serve as a benchmark for Brazilian producers, with the goal of obtaining better results and becoming more competitive. Faced with this, a comparative economic analysis of potato production in Brazil (Vargem Grande do Sul region), and in the United States (the southwestern region of Idaho) was performed, which allows for better understanding of the production costs, prices received by producers, and gross and net operating revenue in each region, as well as investigation of each region's cost structure, allowing identification of the operational and financial advantages of each region. In this context, this study's main objective is to analyze and compare the economic, financial aspects of potato production in the Vargem Grande do Sul region, SP (Brazil), and southwestern Idaho (US), seeking to identify which aspects of American production can serve as reference for technical and financial improvement of the Brazilian production. MATERIALS AND METHODS Data from official sources was used to describe the economic aspects of potato production in the regions in Brazil and the US. The analysis was restricted to southwestern Idaho, the largest potato producer in the United States, which is subdivided into three potato producing regions and the region of Vargem Grande do Sul, in the state of São Paulo, Brazil, which accounts for 60 percent of Brazilian winter potato production and approximately 20 percent of Brazil's total production. This information was used to conduct comparative economic and financial analyses of the two regions, taking into account the values of gross revenues, total costs, and profit. To verify the relationships, principal components analysis (PCA) was used for each of the two regions, considering the same eleven cost component variables (seed, fertilizer, pesticides, irrigation, machinery, labor, post-harvesting and sorting, fees and assessment, tractor depreciation, land leasing, and overhead) for the period between 2009 and 2017. Principal component analysis (PCA) is a statistical tool that is used to reduce data sets that are large and complex, whereby this reduction is made by transforming the data into new variables (principal components). These new variables house most of the dataset, becoming a few components, which facilitate the analysis and interpretation of the original data (Todorov, H., Fournier, D., Gerber, S., 2018) RESULTS AND DISCUSSION The analysis of production costs demonstrates which region has an advantage in the individual cost components to obtain benchmarks for the other region. It is expected that certain cost structure aspects of potato production in the US will be more advantageous, serving as suggestions for production in the state of São Paulo. Fig. 1 shows that there is great fluctuation in the prices received by producers in the Vargem Grande do Sul region (SP), mainly due to the region's level of potato production. In some years, such as 2009, 2012, and 2015, the volume produced was very high, which resulted in low prices. On the other hand, production was low in 2010, 2011, 2013, and 2017 and prices were higher. Analyzing these variations in the volume produced and the prices received by producers, in 2010, compared to 2009, there was a decrease of 12 percent in potato production in the Vargem Grande do Sul region and an increase of 403 percent in the prices received by producers. Similarly, there was an 11 percent decrease in production in 2011 and an increase of 188 percent in prices, while in 2017, production decreased by 30 percent and prices increased by 41 percent. The results show there was less price instability in the United States than in Brazil. In 2009, the average price received by US producers was $165.34, which fell to $154.32 in 2010, returned to $165.34 in 2011, fell again in 2014 to $159.83, and closed in 2017 at $170.86. In 2010, therefore, production decreased by 15 percent and prices fell by 7 percent, while production fell by 8 percent in 2011 with no change in prices, and by 2017, there was a 6 percent reduction in production and a 7 percent increase in prices. The average production yield in the southwestern region of Idaho in 2009 and 2017 was 59.38 and 56 tons per hectare, respectively, while it was only 28.12 and 30.78 tons per hectare in Vargem Grande do Sul, showing that the US state achieved a higher yield than the Brazilian state. It also indicates that the yield in the two regions remained stable in the period between 2009 and 201; however, the most striking fact is the yield difference between the two regions. The average yield in the southwestern Idaho region was 57 tons per hectare, while it was only 28 tons per hectare in the Brazilian region. This higher yield in the southwestern region of Idaho can be explained by edaphoclimatic factors, seed technology, and increased fertilizer use. Using the yield information and prices received by producers, it is possible to compute the average gross operating revenue obtained each year in each region, which is shown in Fig. 2. The average gross operating revenues in the southwestern region of Idaho are stable, with values each year close to around US $9,500 per hectare. In the Vargem Grande do Sul region, however, the fluctuations each year are larger; in some years, average gross operating revenues per hectare have been US $ 5,041 (2014), while in others, they have reached US $ 15,175 (2016) per hectare. The profitability of potato production in the two regions can be estimated by calculating the gross operating profit, obtained as the difference between gross operating revenues and total production costs. Disregarding inflationary adjustments, the potato producers in southwestern Idaho experienced negative results in 2011 and 2012; in all other years, the results were positive, resulting in total financial gains of US$ 4,720.85 per hectare between 2009 and 2017. In the Vargem Grande do Sul region, negative results were found in 2010, 2011, 2012, 2014, and 2017, and in the other years (2009, 2013, 2015, and 2016) the results were positive, giving rise to a negative value of US$ 18,116.95 per hectare for the entire period. Comparing the two regions, it can be seen that the prices paid to producers in southwestern Idaho are lower than those in the Vargem Grande do Sul region, but the average production yield is higher, and the costs of production are lower, producing better financial results for southwestern Idaho. Table 1 shows that in southwestern Idaho, the highest costs are related to land leasing, fertilizers, pesticides, seeds, and machinery, and for the Vargem Grande do Sul region, the highest cost items are seeds, post-harvest and sorting, fertilizer, labor, and pesticides. An examination of the seed costs of the two regions ( Fig. 3) reveals that in southwestern Idaho, the average annual costs were around US $893.00 per hectare, while in the Vargem Grande do Sul region, the annual average was US $2,082 per hectare. This cost component is extremely complex in the Brazilian region, due to the multiplication processes that are necessary for Brazilian plantations. Unlike in southwestern Idaho, where seeds are produced in the region and have less impact on the total cost of potato production, the parent seeds for Brazil are imported from European countries and then multiplied in the region. Similarly, when analyzing fertilizer costs in the two regions, the annual costs in southwestern Idaho average US $1,241 per hectare, and those in the Vargem Grande do Sul region average US $1,491 per hectare, approximately 20 percent higher in the latter than in the former. Likewise, the average annual labor costs in southwestern Idaho are around US $555 per hectare, whereas the total annual labor cost in the Vargem Grande do Sul region averages US $1,311 per hectare, almost 237 percent higher than in southwestern Idaho. This higher cost in the Vargem Grande do Sul region can be explained by the intense use of labor in harvesting operations, since almost all properties employ a semi-mechanized harvesting system. A possible solution for reducing this cost would be to completely mechanize the harvest operations, as is done in southwestern Idaho (Fig. 4). We also analyze the costs of post-harvesting and sorting in the two regions; the results show that in southwestern Idaho, post-harvesting and sorting generate an average annual cost of around US $188 per hectare. On the other hand, in the Vargem Grande do Sul region, the average annual cost is US $2,036 per hectare-1,082 percent higher than in southwestern Idaho. The high cost in this region of Brazil can explained by the fact that almost all potatoes harvested in the region undergo a washing and classification process, which is required by the Brazilian market. The most aggravating fact in this context is that almost all potato producers in the region outsource this operation, which increases the cost. Another constraint in the Vargem Grande do Sul region is the fact that marketing is carried out through middlemen agents, who link producers with wholesalers and retailers. (2014) These agents are very important for marketing the potatoes produced in the region because they maintain contact with major wholesalers and retailers and deliver the entire production to consumer markets. However, the agents charge costs that reduce the viability of the business for producers, particularly in years when prices are low. In developing the relationships between the two main components for producers in the Vargem Grande do Sul region, principal components analysis showed that the first four major components (CP) have eigenvalues greater than one and explain about 89 percent of the total variation, while each contributes about 40, 26, 14, and 9 percent, respectively. The first major component (F1) is strongly correlated with 4 of the 11 variables, showing correlation values above 0.3, including correlation with the variable "fertilizers" of 0.4, as shown in Fig. 5. F1 is also correlated with the variables related to total costs of production, such as fertilizers, land leasing, sorting, seeds, labor, and pesticides. The second main component (F2) presents a higher correlation with irrigation variables, administrative costs, equipment depreciation, and, negatively, the overhead variable. The principal components analysis for southwestern Idaho (US) showed that the first four main components (CP) have eigenvalues greater than one (1) and explain about 98 percent of the total variation, while each one contributes about 50, 29, 14, and 5 percent, respectively. The first major component (F1) is strongly correlated with 2 of the 11 variables, with correlation values above 0.4. F1 is more correlated with variables related to pesticides, overhead, and land leasing, and this association between the variables F1 and F2 is highlighted in Fig. 6. The second main component (F2) shows a stronger correlation with variable machinery. Thus, the cost components that most impact potato production in the Vargem Grande do Sul region are fertilizers, seeds, land, labor, and sorting, corroborating that the profitability gaps in production are strongly related to agricultural production factors. In contrast, in southwestern Idaho, the cost components most closely related to the profitability of potato production are pesticides, land, overhead, and mechanization. Given this context, what lessons can Brazilian producers extract from potato production in Southwestern Idaho? With respect to reducing the costs of production factors, in potato production in the Vargem Grande do Sul, fertilizer 0 150 300 450 600 750 900 1,050 1,200 1,350 1,500 1,650 1,800 1,950 2,100 2,250 2,400 2,550 2,700 2, use accounts for 13 percent of average total costs, while other production factors that also impact production costs are seeds (19 percent), defensive (10 percent), work (12 percent), processing (19 percent), and land leasing (8 percent). Carvalho et al., (2010) show that the high cost of seeds is due to their importation and the cold chamber storage system. Some potential actions would reduce these costs, such as new growing technologies, seed multiplication and storage, and reducing the region's dependence on imports, including the use of varieties developed in country. Although this is not an easy task, there is research that addresses this scope and can, in the short and medium term, reduce dependence on imported seeds. The adoption of modern and technological pesticide products with smaller applications per area and increased efficiency, which are already available in the market, becomes essential for cost reduction. In addition, if the culture is healthy, with good nutritional levels, pest and disease attacks will be diminished. Producers need to be attentive to new technologies, since the reduction process must be constant. Purchases of pesticides by producer associations and cooperatives are also approaches that can reduce costs, since purchasing greater quantities increases trading power. In the context of pest and disease control, the adoption of biological control can also help reduce production costs. The labor factor is perhaps one of the most important aspects in this analysis, since it accounts for 12 percent of total costs. Looking at the lessons of production in southwestern Idaho, we found that the solution would be the complete mechanization of the crop. This is already done in some regions of Brazil by large producers, but the number of producers who implement such a process is small. It is necessary to analyze the investment required, which is still high, but studies show that this investment is financially viable, mainly due to the reduced amount of labor compared to the semi-mechanized system. Land leasing is a common practice, as it becomes more advantageous for producers to absorb the cost of renting land rather than to invest in land acquisition. One solution would be negotiating with landowners to reduce the amounts paid, or planting other crops in the same area, such as corn, soybeans, and beans, as part of a system of crop diversification and rotation. In addition to mechanizing the harvest, the post-harvest processing, which is mostly carried out by third parties, is responsible for 19 percent of total production costs. Similarly, the beneficiation process requires large investments when performed by the farm itself and can be made more financially viable by substituting outsourcing. A palliative measure would be the joint realization of the investments necessary for mechanizing the harvest and beneficiation through an association of producers, cooperatives, or producer pool for this purpose, including being able to provide services to other producers, generating extraordinary revenues. The potato marketing process is totally disadvantageous for producers, especially in low-price years, since practically all producers in the region use market agents (brokers) to sell their potatoes. These agents are the link between the rural producers and the wholesale and retail sectors and perform the planning and scheduling tasks for product delivery. Brokers receive the wholesaler and retailer demands and pass these to the producers, organizing the product's delivery process. In the same way, it is the agent who makes payments to the producers and determines other marketing actions. Thus, the agent ends up having great power in the market (Carvalho et al., 2010), providing an essential link for rural producers to commercialize their production. Moreover, this link provides an important contact between the wholesalers and retailers to producers in several Brazilian regions to supply potatoes throughout the year, meeting the demands of these sectors downstream of production. However, rural producers do not have any form of guarantee of financial receipts or prices and assume all market risks, due to the absence of formal marketing contracts and the informality of the process. This corroborates Carvalho et al., (2010) study. It is common for producers to face default by brokers or receive insignificant prices, well below the break-even values. One possible solution would be for producers to organize and seek new forms of marketing, perhaps by working directly with wholesalers and retailers. They could also opt for investments in production storage, using new technologies that exist in European countries, which provide good operating results and maintain product quality for longer periods of time at low cost per unit of production. Still in the scope of commercialization, another solution would be to diversify the forms of sale, dividing the production between the market in nature and the chips and frozen potato chips industries. Finally, prices are a more complex aspect, as they depend on variables such as production, time of year, market power, etc. However, producers have some opportunities to increase prices, such as attempting to have their product on the market when there is less supply, that is, to plan the crop cycle so that harvest occurs before or after the region's peak production. Since peak production occurs between May and October, producers should try to schedule their harvest before May or after October, especially in low-price years. Another possibility would be to store the production for later commercialization by making investments in storage, notably in cold rooms. These are commonly used in European countries for this purpose, but are more prominently used in Brazil for seed storage. Likewise, production cycles could be planned based on historical production and price series; when two or three years of low prices are observed, producers could reduce the area to be planted, minimizing negative results, and increase planting areas in years of higher price trends. CONCLUSIONS Lessons learned from potato production in southwestern Idaho expose the need for reduced costs of production factors, (such as seeds, fertilizer, labor and processing/ commercialization), whereas average total seed costs in the Vargem Grande do Sul region are 134% higher than in the Southwestern Idaho region and the total average costs of fertilizers and processing/commercialization are 22% and 983% higher than Southwestern Idaho, respectively. Labor factor is 137% higher in Vargem Grande do Sul than Southwestern Idaho, but the machinery costs in Southwestern Idaho are 35% higher than Vargem Grande do Sul, exposing the greater use of mechanization in potato production in Idaho. Potato producer in Vargem Grande do Sul need learning how to reduce the costs these components to have better financial results in your potato production, and there is the need for new research in these topics for helping them to improve these actions.	2020-03-12T10:17:03.734Z	2019-11-26T00:00:00.000	{ "year": 2019, "sha1": "97f3abdc64d353b2e055b7998d7ccef0b8724feb", "oa_license": "CCBYNC", "oa_url": "https://ejfa.me/index.php/journal/article/download/2028/1292", "oa_status": "GOLD", "pdf_src": "Anansi", "pdf_hash": "c118a5824a2d78824f77807b7e19dd56534e0095", "s2fieldsofstudy": [ "Education" ], "extfieldsofstudy": [ "Business" ] }
210299760	pes2o/s2orc	v3-fos-license	Modelling the impact of ground temperature and ground insulation on cooling energy use in a tropical house constructed to the Passivhaus Standard In tropical climates ventilation introduces high levels of moisture into dwellings, leading to raised values of indoor relative humidity and air temperatures, creating occupant thermal discomfort. The Passivhaus standard, by creating a super insulated and air-tight envelope, reduces heat losses and provides low energy comfort. This approach in tropical housing might be effective but could potentially increase mechanical cooling demand. This research investigated the effect of ground temperatures on thermal comfort and energy-saving in a tropical dwelling. Terraced houses in Jakarta, Indonesia were chosen as this a popular house type. IES VE software was used to study building energy use, and the Passivhaus software PHPP examined the application of Passivhaus criteria. Field measurements of air temperature and relative humidity were used to validate the software model. Analysis revealed that the building’s predicted air temperatures were affected by the assumed ground temperature models in IES. Good agreement between the measured and modelled values was achieved only when a particular ground model was chosen. Having validated the IES model, dwelling insulation and air-tightness levels were incrementally changed until they met the Passivhaus standard. Finally, the feasibility of meeting the Passivhaus energy standard for cooling in the modified tropical house was tested. Introduction Low-density housing demand in Indonesia has contributed significantly to uncontrolled urban sprawl around the edges of Indonesia's major cities [1]. This unplanned development is characteristic of contemporary Asian Urbanization [2]. Many of these dwellings are badly constructed, being poorly insulated and not well-sealed. The air conditioning (AC) that is employed to achieve thermal comfort for residents is used inefficiently and expensively. Indonesia's hot and humid climate, coupled with rapid urbanisation and economic growth, means the demand for cooling energy in residential buildings will increase sharply in the coming decades [3]. In 2017, the three largest energy user sectors in Indonesia were transport (47%), industry (30%) and housing (16%) [4]. It is forecast that, by 2020, building energy consumption in South East Asia countries will exceed that of developed countries, with 56% of total building energy demand being used for air-conditioning [5]. Therefore, it is desirable to investigate low energy cooling systems that could help reduce the need for electric air conditioning. Radiant cooling is one such strategy to cool buildings. Concrete Core Temperature Control (CCTC), for example, pumps cool water around pipes cast inside the floor [6]. In tropical climates, ground soil temperatures range from around 15qC to 25qC, and so can be much cooler than the ambient air temperature [7]. Floor slabs in contact with the ground can, therefore, be effective in cooling a building. The use of ventilation to try and cool buildings can transfer high amounts of moisture in to rooms, resulting in poor internal thermal and health conditions in which the optimum humidity levels of 40% to 60% [5] are exceeded. It is, therefore, necessary to create design approaches to building design that can keep indoor air humidity low while still reducing cooling energy consumption. One radical alternative solution is to apply the German Passivhaus standard [8], with its very high levels of insulation and air tightness, to Indonesian dwellings. This paper presents the validation and testing of a house model developed using the commercial dynamic thermal simulation software IES. Monitored air temperature and relative humidity data from a real dwelling were used to validate the IES software model before Passivhaus concepts were applied to the house model. This paper describes the development of a Passivhaus standard for housing built in Indonesia that works for both thermal comfort and energy efficiency. The energy advantages from cool ground temperatures and no floor insulation were observed in the development of the house model. The Passivhaus standard The requirements for the Passivhaus standard, which originated in Germany, are shown in Table 1 [8]. These requirements are principally performance-based, using passive measures, minimal thermal bridging and a whole house mechanical ventilation system with highly efficient heat recovery [9]. Even though the initial developments were made within the relatively mild climate of Central and Northern Europe, studies have suggested that Passivhaus could be a feasible option in other climate types [10]. The Passive-On study forecast a number of issues related to Passivhaus criteria for warmer climate, including the introduction of a limit for summer cooling energy demand, a higher infiltration rate and an indoor comfort temperature that coincided with adaptive thermal comfort standards [11]. The application of the Passivhaus standard must also properly consider moisture balances and latent loads for buildings in a hot and humid climate. Table 1. Passive House Criteria. Thermal Comfort Thermal comfort assessment is an important tool to assess energy usage and measurement thermal satisfaction achievement. The term 'thermal envelope' used here refers to the shell of the dwelling as a barrier to unwanted heat or mass transfer between the interior of the building and the outside conditions [12]. In hot climates, the effect of air movement and humidity are particularly important because heat lost by evaporation dominates [13]. The impact of the external temperatures on perceived comfort has to some extent also been incorporated in to design standards such as ASHRAE Standard 55. The finding on the ASHRAE adaptive comfort standard study was the difficulty of creating generalization for areas that have mean outdoor temperatures above 23°C. The research indicated that buildings in this zone were unable to maintain thermal comfort for many hours of the day [14]. This is consistent with other research that found some climatic differences between cities in the lowland and highland in Indonesia, 3 which could lead difference in people's comfort temperature due to physical adaptation [15]. However, the Indonesia National Standardization Agency (SNI) still issues standards whereby the comfort temperature is set at 25°C ± 1 °C and comfort relative humidity at 60% ± 10% [16]. The result from a study of comfort temperatures for people in the Depok area (a Jakarta satellite town) indicated that the comfort temperature was higher than this national standard, at 27.6°C [17]. Research methodology This paper investigated the impact of ground temperature model on a simulated Passivhaus building. A row (terraced) house of the type that is commonly found in the Jakarta Metropolitan Region was chosen for the study. IES VE 2018 software was used to predict the dwelling's indoor temperatures, relative humidity and energy use. The building model in IES VE was built by inserting building information such as building materials, cooling systems, lights and appliances, and a presupposed occupancy schedule. The selected monitored house had a floor area between 50m 2 to 69m 2 floor area in which form most of the existing urban housing stock for housing in the Jakarta Metropolitan Region [18]. Empirical validation compared the computer simulation results with monitored data from the house. Analysis was done by investigating the typical housing characteristic, building performance and energy consumption. Case study monitoring The dwelling was in the major housing development area of Bogor, Jakarta Metropolitan Region. The dwelling measured 6m x 10m in plan, with a total floor area of 55m 2 , and a floor-to-ceiling height of 2.85m. The building's main axis orientation was northerly; it was not insulated, being constructed from a single layer of brick, and with single glazing windows ( Figure 1 and Figure 2). The monitoring of the case study building was undertaken during a period of consistent occupation over two selected periods, one in January -February for the rainy season, and the other in September -November for the hot season. Air temperature and relative humidity values were collected at 30 minute intervals using Tinytag and Rotronic data loggers placed at a height of 1.5m in the living room, master bedroom and a sheltered outdoor area. Because of space limitations, only the living room results are presented in this paper. Validation of the IES VE 2018 software A three-dimensional dwelling model in IES VE software was built using the data obtained from field measurements. The IES VE simulation result was compared with measured period data, in this paper in mid-September, which has comparatively high weekly air temperature and a larger gap of temperature differences between maximum and minimum temperature compared to all the measured data. The dwelling model in this software was using plan provided by the home owner as a guidance, building materials information was based on the contractor specification, and the occupant activity schedule was gained from field observations. The building elements used to build the base model in IES VE 2018 can be seen in Table 2. Figure 3 shows the detailed analysis of the air temperature and relative humidity measured data compared with simulation results for the living room. A large divergence between measured and modelled values was evident for daytime results, with simulation data showing internal air temperatures similar to the external temperatures. Undertaking a series of parametric studies of the IES model indicated that the assumed ground temperatures were creating the divergence. The default setting in IES VE sets the ground temperature equal to the external air temperature, but this may not be appropriate for tropical weather and soil conditions. To find out the setting for the ground temperature, the research continued by looking at ground temperatures in tropical countries, particularly Jakarta, Indonesia. Soil properties are difficult to generalize because of local heterogeneity and the lack of broad based data; also, soil condition is not a standard variable collected at most weather stations [20]. Available literature indicates that underground soil temperatures can vary from 8qC to 27qC in some parts of the world, especially in cold-dominated areas (North America), and between 15qC and 25qC in tropical climates [7]. The average monthly ground temperature in Jakarta, based on the weather analysis software Climate Consultant, is 26qC [20]. Because soil temperature responds to the net effect of the daily surface energy balance, it can be estimated by computing a running average of air temperature, with progressively longer integration times as soil depth increases [21]. The ground temperature model used in the simulation By using the literature information about ground temperature above, the setting on the IES VE simulation was set using outside air with an offset temperature profile. The analysis method for ground construction setting in IES VE software is based on EN-ISO 13370 [22]. The software analysis indicated that an offset temperature profile set at -5qC from the external temperature was appropriate for this simulation. The graph in Figure 4 is ground temperature that was estimated by offsetting measured external temperature by -5qC following analysis above. The graph indicated that there was more than 75% of the time when the ground temperatures were below the adaptive comfort level. Figure 5 shows the air temperature and relative humidity for the living room after applying the adjusted ground temperature model. The results indicate that the simulations were satisfactory in displaying the same trends as the measured data. Living room results in Figure 5 suggest that the differences between the measured and modelled air temperature values were around 1 to 2qC, and the differences for relative humidity were around 10%. To ensure that the IES dwelling model was statistically valid, an analysis was conducted in accordance with ASHRAE Standard 14-2002 [23]. As an indication of the mean ratio of relative error between two values (measured and modelled), the mean bias error MBE and Coefficient of Variation of the Root Mean Square Error CVRMSE were calculated. ASHRAE Standard 14 defines the acceptable limits for calibration to hourly data as within ±10% MBE (hourly) and ≤30% CVRMSE (hourly) measured at a utilities level [23]. The evaluation results for this study showed relative humidity and temperature MBE values for the living room of -5.9% and +0.3% respectively and CVRMSE values of +7.5% and +3.1% respectively. Therefore, it can be concluded that the IES simulation can model satisfactorily the studied house in the tropical climate and produce relatively similar results compared with the field measurement data. The impact of ground temperature impact on a Passivhaus building in a tropical climate To study the effect of ground temperature on a tropical Passivhaus building, the validated IES building model had Passivhaus elements applied to it. Table 3 indicates the materials that were used in the building model to follow the Passivhaus concept. This performance of the Passivhaus model will be compared with a Passivhaus model that had the floor insulation removed to let the ground floor act as a thermal sink and potentially provide radiant cooling. The house with its original building elements was also included in the comparison to look at the performance of the building with the Passivhaus standard to study the energy saving to reach thermal comfort. All three scenarios used the same HVAC system (AC and dehumidifier) that was located in the living room, master bedroom, and children's bedroom. For the simulation, the AC temperature set point in IES was 26qC, and relative humidity controller was set at 60%. The loggers used in this study recorded air temperature. Operative temperature is often considered a better indicator of thermal comfort as it combines both air and mean radiant temperatures in a space but is not easy to measure in real building situations. The IES software was run to test for any significant differences between predicted air and operative temperatures in the house. However, as shown in Figure 5, there were only minor difference between simulated air temperatures with simulated operative temperature and so it was feasible to use air temperature for the validation and the comfort analysis. Result and discussion With the application of the Passivhaus standard to the IES dwelling model, simulation results for the living room showed that the Passivhaus building model provided stable air temperatures (Figure 6), although the air temperatures were mostly above 27.6qC (the adaptive comfort level). For the original building elements scenario, the air temperature was still fluctuating and following the external air temperature, with a few hours being above 27.6qC. Meanwhile, the Passivhaus application without floor insulation always maintained room air temperatures below 27.6qC. Relative humidity for the original house scenario tracked outside relative humidity and, for most of the time, was above 60%. On the other hand, the Passivhaus model and Passivhaus without floor insulation indicated low relative humidities, with relative humidity staying below 60% by using the AC + Dehumidifier system. With the same building HVAC system and occupancy schedule, the IES simulation predicted an annual cooling energy use of 11.41 MWh for the original layout, 10.89 MWh for the house with Passivhaus application, and 8.61 MWh for the Passivhaus without floor insulation (Figure 7). From this study, the application of Passivhaus into a tropical dwelling provided stable interior air temperatures and relative humidities. The building airtightness reduced cooling energy slightly, and by removing floor insulation significant energy savings were made. The floor can be seen to act as a radiant Conclusions and discussion This paper explored the effect of ground temperature on the performance of a Passivhaus-enhanced dwelling in a tropical climate. The original dwelling was monitored to gain air temperature, relative humidity and activity schedule data that were used to create a validated IES building model. The validation process identified ground temperature as an important variable in the model. The impact of coupling the building to the ground without using insulation was also examined and found to reduce cooling energy demand whilst maintaining good comfort levels. The application of the Passivhaus concept in tropical climates still needs further analysis. Besides the additional construction cost for the building, the skill of construction workers needs to be improved, especially residential construction workers, because the use of insulation, double glazing, and constructing air-tight building is not common practice in Indonesia.	2019-10-17T09:00:44.112Z	2019-10-11T00:00:00.000	{ "year": 2019, "sha1": "5f936c8125411b19c2495a5b1264c47f25ed2648", "oa_license": null, "oa_url": "https://doi.org/10.1088/1755-1315/329/1/012010", "oa_status": "GOLD", "pdf_src": "IOP", "pdf_hash": "b98f3d891b65bb9187c9ada78238becd667364cd", "s2fieldsofstudy": [ "Environmental Science", "Engineering" ], "extfieldsofstudy": [ "Environmental Science", "Physics" ] }
118939133	pes2o/s2orc	v3-fos-license	On the 1/f spectrum in the solar wind and its connection with magnetic compressibility We discuss properties of Alfv\'enic fluctuations with large amplitude in plasmas characterised by low magnetic field compression. We note that in such systems power laws can not develop with arbitrarily steep slopes at large scales, i.e. when $\|\delta \bf{B}\|$ becomes of the order of the background field $\|\bf{B}\|$. In such systems there is a scale $l_0$ at which the spectrum has to break due to the condition of weak compressibility. A very good example of this dynamics is offered by solar wind fluctuations in Alfv\'enic fast streams, characterised by the property of constant field magnitude. We show here that the distribution of $\delta B=\|\delta \bf{B}\|$ in the fast wind displays a strong cut-off at $\delta B/\|{\bf B}\|\lesssim2$, as expected for fluctuations bounded on a sphere of radius $B=\|{\bf B}\|$. This is also associated with a saturation of the rms of the fluctuations at large scales and introduces a specific length $l_0$ above which the amplitude of the fluctuations becomes independent on the scale $l$. Consistent with that, the power spectrum at $l>l_0$ is characterised by a -1 spectral slope, as expected for fluctuations that are scale-independent. Moreover, we show that the spectral break between the 1/f and inertial range in solar wind spectra indeed corresponds to the scale $l_0$ at which $\left<\delta B/B\right>\sim1$. Such a simple model provides a possible alternative explanation of magnetic spectra observed in interplanetary space, also pointing out the inconsistency for a plasma to simultaneously maintain $\|\bf{B}\|\sim$const. at arbitrarily large scales and satisfy a Kolmogorov scaling. INTRODUCTION The magnetic field spectrum of the fast solar wind is characterised by a double power law at intermediate and large scales, with power indices -5/3 and -1 respectively (Bavassano et al. 1982;Denskat & Neubauer 1982;Burlaga & Goldstein 1984). The former corresponds to the MHD turbulence inertial range (observed typically for 10 −3 Hz f 10 −1 Hz, where frequencies in the spacecraft frame correspond to Doppler-shifted spatial kvectors in the plasma frame), while the latter, so-called 1/f range (typically for f 10 −3 Hz at 1 AU), is considered the energy reservoir feeding the turbulent cascade, although the origin of this range is not well understood and still under debate (Matthaeus & Goldstein 1986;Velli et al. 1990;Verdini et al. 2012;Chandran 2018). Note that a spectrum with index -1 is indicative of scaleindependent underlying fluctuations and a long memory in the system (e.g. Keshner 1982). In this work, we investigate the possible link of spectral properties with another well-established property observed during fast streams with large amplitude Alfvénic fluctuations, namely the almost constancy of magnetic field intensity. This property is surprising because while the total amplitude of the fluctuations δB = \|δB\| is of the order of the field strength B = \|B\|, the associated variations in the magnetic field intensity δ\|B\| remain small at all scales: δ\|B\| δB ∼ B (Belcher & Davis 1 LESIA, Observatoire de Paris, Université PSL, CNRS, Sorbonne Université, Univ. Paris Diderot, Sorbonne Paris Cité, 5 place Jules Janssen, 92195 Meudon, France 2 Department of Physics, Imperial College London, SW7 2AZ London, UK 3 School of Physics and Astronomy, Queen Mary University of London, London E1 4NS, UK 1971). Note that the nearly constant magnetic pressure has an impact also on the gas pressure, leading to equally small perturbation of the plasma density during Alfvénic fast streams. Geometrically, it means that the tip of the magnetic field vector is approximatively constrained on a sphere of constant radius \|B\| (e.g. Bruno et al. 2004). We discuss here how this behaviour has a direct impact on the scaling of δB and on the shape of its distribution (PDF). In particular, we suggest that if the plasma is characterised by a regime of small magnetic field intensity variations, then a conflict with the expected MHD turbulence spectrum at large scales can arise. Indeed, large amplitude fluctuations (with δB ∼ B) cannot simultaneously be organised with an arbitrary power law and maintain a low magnetic compressibility at all scales. On the contrary, if relaxing the former, the latter condition imposes a saturation of δB for scales l larger than the reference scale l 0 at which the average level of the fluctuations reaches the mean field amplitude B. As a consequence, this saturation introduces a break in the spectrum at l 0 and a slope of index -1 (or shallower) for l > l 0 , as for fluctuations that are independent of the scale. This simple argument, based on the phenomenological constraint δ\|B\| δB motivated by in situ observations, leads to a straightforward justification of the spectral shape that is usually observed at large scales in space plasmas. We show in this work that the existence of a 1/f spectrum in Alfvénic fast streams is indeed associated with the presence of an observational cutoff in the distribution of the fluctuations and the saturation of their mean amplitude. Moreover, the spectral break between inertial and 1/f ranges as observed at various radial distances from the Sun always corresponds to the scale l 0 at which the average level of fluctuations reaches δB/B ∼ 1. DATA ANALYSIS We analyse Ulysses magnetic field observations over the pole during solar minimum, when the spacecraft was permanently embedded in fast wind. We selected 150 continuous days starting from day 100 of year 1995, with time resolution of 1s, corresponding to radial distances 1.4-2.2 AU and a heliographic latitudinal variation from 30 to 80 degrees (Wicks et al. 2010;Chen et al. 2012). We also use selected intervals of measurements near the ecliptic by Ulysses and by Helios at perihelion (0.3 AU). Time increments δB(t, ∆t) = B(t) − B(t + ∆t) are calculated using 17 logarithmically spaced lags ∆t, ranging from 1 to 2 · 10 5 s. In terms of physical scales, this covers the full MHD inertial range of the turbulence and the largest scales (∆t 10 4 s) correspond to the 1/f range (Bruno & Carbone 2013). In the following we define δB as \|δB\| and δB/B as \|δB\|/B, where B is the average intensity \|B\| over the interval ∆t. Figure 1 shows an example of typical magnetic field fluctuations in the fast solar wind taken during a period of 6hrs (top) and then within shorter intervals of 1 hour (middle) and 10 minutes (bottom). The amplitude of the fluctuations in the components decreases with the scale, going from δB ∼ B at the largest scale to δB B at smaller scales. Remarkably, B (black line) is very steady and close to constant all the time (δ\|B\|/B 10%), and its value is the same at all scales. DISTRIBUTION OF δB/B This property characterises Alfvénic fluctuations at all distances and has the geometrical consequence that Middle: Distributions of δB/B over different scales, from 1s (red) to 2 · 10 5 s (black). The PDF of ∆t ∼ 5 · 10 3 s, roughly identifying the 1/f break scale, is highlighted by squares. The insert shows the PDFs from the 1/f range (10 4 − 2 · 10 5 s). Bottom: PDFs of δB 2 /2B 2 which is directly related to the cosine of the rotation angle θ. The dashed line shows the exponential dependence of Eq. (1). the tip of the magnetic field vector moves on a sphere of approximatively constant radius (e.g. Bruno et al. 2004;Matteini et al. 2015). Although this is well known (Belcher & Davis 1971), there is an important consequence of this state that has not yet been noted. The regime of low-magnetic compressibility sets a well defined limit for the amplitude of the fluctuations: the maximum amplitude δB of the difference between two magnetic field measurements is twice the radius of the sphere, i.e. δB ≤ 2B. We can then expect that this limit is visible in the distribution of δB. This is confirmed by the top panel of Figure 2, which shows the distribution of δB/B at a scale of 500s, well inside the inertial range. A very clear cutoff is visible in the distribution, located at δB/B ∼ 2. Only fluctuations changing significantly the modulus of B (more than 10-20%) populate the far righthand side of the PDF; these are mostly field depressions associated with non-Alfvénic, isolated magnetic holes (Fränz et al. 2000), for which δB/B is particularly enhanced. On the contrary the larger part of the PDF on the lefthand side (δB/B 2) constitutes the main incompressible component of the turbulence, for which δ\|B\| δB. The middle panel of Figure 2 shows that the same cutoff is visible in all PDFs at large scales (green to blue), while it gradually disappears approaching kinetic scales (green to red) as expected since δB B at small scales. The PDFs at kinetic scales become narrow, corresponding to small rotations of the magnetic field vector (Chen et al. 2015). The PDF of ∆t ∼ 5·10 3 s approximately corresponding to the break scale separating the inertial range from the 1/f is highlighted by black squares. Interestingly, at the largest scales (cyan to black) the distributions do not evolve further and tend to a roughly symmetric shape between 0 < δB/B < 2, peaked at δB/B ∼ 1, as highlighted in the insert panel. Moreover, at these scales the PDFs lie approximately on top of each other, as expected for a 1/f range in which the amplitude of the fluctuations becomes independent of scale. The last panel (bottom) shows the distribution of δB 2 /2B 2 which is related to the rotation angle θ between the two magnetic field vectors. For pure rotations δB 2 = 2B 2 − 2B 2 cos θ, so in the range [0,2] we can consider δB 2 /2B 2 ∼ 1 − cos(θ) (Zhdankin et al. 2012). Moving towards large scales the distribution of cos(θ) becomes shallower, indicating that fluctuations tend to spread over the full sphere as δB ∼ B. However, even at the largest scales, corresponding to the 1/f range, the distribution does not become flat, meaning that the fluctuations do not uniformly cover the sphere and maintain a memory about the underlying mean field direction. The PDF saturates to an approximately exponential shape: where x = 1 − cos(θ) and α is an empirical constant close to unity (α = 0.8). We recall that for fluctuations uniformly distributed over the sphere α = 0. 3.1. Connection to power spectra It is now instructive to plot the average value of δB/B for each of the PDFs in Figure 2 as a function of the scale. This is shown in the top panel of Figure 3 which for each scale ∆t displays the mean value of δB/B (red line). This increases linearly for scales ranging from subminutes to ∼ 1hr, while at larger scales, t 2hrs, the curve starts to flatten and saturates at δB/B ∼ 1. As Figure 1 shows, B is locally independent of the scale, and the distribution displayed here in practice corresponds to the distribution of the scale-dependent δB normalised to the scale-independent reference B. Note that this is then analogous to simply normalising the mean value of the PDF of each scale to a constant value (the average B measured at the largest scale). As a consequence, the quantity shown in Figure 3 corresponds to the (normalised) first order structure function and has a direct connection to the power density spectrum of the fluctuations, thanks to the well known relation that connects δB 2 at a scale l to the k-space power spectrum P (k): δB 2 l = P (k) · k, where l = 1/k. In particular the slope of the power spectrum P (k) ∝ k −α is related to the exponent of the second order structure function δB 2 ∝ l β as: α = β + 1 (e.g. Monin & Yaglom 1975), for 1 < α < 3. There is then a straight correspondence between the behaviour in the top panel of Figure 3 and the spectral slopes commonly observed in the solar wind. The increasing part with δB ∝ l 1/3 corresponds to the inertial range of spectral index -5/3, while the flat part is the 1/f with spectral index -1. This is obviously well known; what is however new here is that the region of constant amplitude in the structure function and corresponding to the 1/f range in spectra, saturates at a normalised value δB/B ∼ 1. This is consistent with the condition of small magnetic field compressibility discussed above. As a consequence, the spectrum breaks at the scale l 0 (vertical dotted line) at which the level δB/B = 1 is reached. By contrast, the blue curve in the same panel shows the behaviour observed during a period of non-Alfvénic slow wind (when Ulysses was on the ecliptic, days 337-349 of year 1990). In this case the structure function does not saturate as the spectrum is not characterised by a 1/f range. The bottom panel of the figure shows the level of magnetic compressibility, δ\|B\|/δB, in the two cases: this is small in the Alfvénic fast wind (with a minimum value ∼ 0.1 in the 1/f range), while is substantially larger in the slow wind where the condition \|B\| ∼ const is not observed (Grappin et al. 1991). The above picture is consistent with the idea that the constraint of low magnetic field compression sets a limiting value of δB/B. This implies that spectra with arbitrary δB/B at large scales cannot be realised within a condition of low magnetic field compressibility. If the latter is well satisfied, as in fast solar wind observations, then it implies that there is a scale l 0 such that for l > l 0 a steep power spectrum of fluctuations cannot be maintained. In other words, this means that there is an inconsistency between extending a -5/3 inertial range to arbitrarily large scales and the fluctuations being nearly incompressible in B. If the level of fluctuations saturates at δB/B ∼ 1 for l > l 0 , so that the amplitude becomes independent on the scale l, then a spectrum with index -1 is the steepest possible realisation. RADIAL EVOLUTION To confirm the behaviour described above, we test the results of Figure 3 against radial variations. The top left panel of Figure 4 shows the scale dependent average amplitude of the fluctuations for different subintervals in the Ulysses dataset, corresponding to different radial distances. The amplitude δB decreases with distance. The small radial variation (from 1.4 to 2.2 AU) does not introduce a major change in the break scale, although one can see that the flat region (1/f range in the spectrum) is shifted to a slightly longer time scale as known (e.g. Bruno & Carbone 2013). The right top panel shows the same curves where the amplitude is normalised to the local magnetic field strength B; when normalised, all spectra collapse on the same curve as in Figure 3, indicating that, within the radial excursion explored by Ulysses, fluctuations populating the 1/f region are maintained at the saturation level δB/B ∼ 1 all along the expansion. Moreover, the break between the inertial and 1/f ranges is always identified by the scale l 0 at which the rms of δB/B approaches 1. To test further our analysis, we compare the Ulysses results at R > 1AU with data of the Helios spacecraft, which approached the Sun as close as 0.3AU, providing thus a much larger radial excursion. We have selected a 5-day period of 1976 (Bruno et al. 1985), when Helios continuously observed a highly fast stream at 0.3AU. Data from both spacecraft are shown in the bottom panels of Figure 4 (Helios in blue, Ulysses in red). The bottom-left panel highlights the large variation in the mean amplitude values due to the large radial separation between the two measurements. In this case, we can clearly distinguish the change in the break characteristic scale, which moves to longer periods as the wind expands (Bavassano et al. 1982;Horbury et al. 1996). Also in this case, when the amplitude rms is normalised to the mean field the two curves are brought closer together, almost overlapping and both saturating approximatively at the expected value δB/B ∼ 1. DISCUSSION A possible interpretation of the dynamics discussed here is sketched in Figure 5. If the amplitude of the fluctuations δB at each scale l is always lower than the critical threshold δB/B ∼ 1 (top left panel), then a standard Kolmogorov-like spectrum P (k) of slope −5/3 can be formed at all scales (top-right). However, if part of the fluctuations exceeds the maximum amplitude (dotted red section in the left bottom panel), the system can maintain a small magnetic field compression only somehow removing the amplitude excess and imposing a constant saturated amplitude over all the scales reaching the δB/B ∼ 1 condition (blue section of the curve). This situation then leads to a spectrum with two spectral slopes (bottom right), with -1 for the largest scales. Obviously, this qualitative picture does not clarify how the plasma imposes and maintains a low magnetic compressibility, nor the physical processes that are responsible of the removal of the most compressible part of the fluctuations; these are certainly interesting and challeng- Figure 5. Cartoon of the distribution of the fluctuations amplitude δB over scales l (left) and the corresponding power density spectra P (k) (right). A case with δB B is shown in the top panels, while a case with δB/B ∼ 1 is shown in the bottom. The blue sections in the bottom panels encode the part of the spectrum which is modified by the constraint of low magnetic field compression, leading to the 1/f range. ing questions for future theoretical studies. However, such a simple scheme has possibly a direct application to the solar wind. Once accelerated, the absolute level of fluctuations in the solar wind is predicted to be maximum around the Alfvénic point; however, if compared to the mean field δB/B ∼ 0.1-0.2 (e.g., Cranmer & van Ballegooijen 2005;Verdini & Velli 2007). One can then expect that the turbulence can fully develop as in the top panels of Figure 5 and without relevant effects related to the low compressibility limit. As the wind expands, the relative amplitude of the fluctuations δB/B increases and reaches unity at 0.3 AU (Helios), leading to the scenario of the bottom panels. Consequently, if the existence of a 1/f region is really related to the presence of a cutoff in the amplitude of the fluctuations and a saturation of their mean amplitude as suggested by Figure 2, it is then possible that this range forms during solar wind expansion somewhere in between the Alfvénic point and 0.3 AU. From 0.3 AU onward the plasma lies continuously on the condition δB/B ∼ 1 at large scales (e.g. Mariani et al. 1978;Behannon 1978), supporting further the idea of a saturation of the fluctuation amplitude. Outside 0.3 AU the width of the 1/f is observed to decrease with radial distance, as expected for a WKB expansion at large scales and a faster decay for the inertial range (e.g. Tu & Marsch 1995); the break scale then moves to lower frequencies, perhaps also due to some coupling between Alfvénic and compressive fluctuations (Bavassano et al. 2000;Malara et al. 2001;Del Zanna et al. 2015). However, the evolution could be different inside 0.3 AU, before the amplitude saturation; according to our model, moving towards a lower δB/B level closer to the Sun would also imply a reduction of the extent of the 1/f range. CONCLUSION In this work we have proposed that the 1/f region of the fast solar wind magnetic field spectrum could be generated by a saturation of the fluctuation amplitude at large scales imposed by the constraint \|B\| = const., which is well verified in Alfvénic streams. There are other models in the literature for the origin of the 1/f range (e.g. Matthaeus & Goldstein 1986;Velli et al. 1989;Ruzmaikin et al. 1996;Verdini et al. 2012;Chandran 2018). The one described in this work constitutes an alternative scenario; further testing against experimental data is needed in order to discriminate among different models. However, intriguingly, the simple mechanism discussed above has the advantage of explaining several observational properties listed below, without the need of further assumptions. First, and mainly, it explains the difference between the high speed Alfvénic streams where a 1/f regime is ubiquitously observed and the typical slow wind where such a feature is absent (see e.g. Bruno & Carbone 2013). The slow solar wind has commonly a lower level of power in the fluctuations, and due to its more irregular and compressible nature with respect to the fast wind, is not characterised by a \|B\| = const. condition, see Figure 3. As a further confirmation, there are periods of slow solar wind which are particularly Alfvénic (D'Amicis & Bruno 2015). During those periods the plasma shows properties very similar to fast Alfvénic streams, including a higher level of fluctuations (δB/B ∼ 1) and a low magnetic field compression. It is then not surprising, following what has been discussed in this work, that the Alfvénic slow wind also displays a spectrum with a 1/f range, breaking at a scale l 0 similar to that of fast streams (D'Amicis et al. 2018). The 1/f part of the magnetic field spectrum has no counterpart in all fluctuating fields (Tu et al. 1989;Bruno et al. 1996;Wicks et al. 2013). In particular, when decomposing the turbulent fluctuations using the Elsasser variables, only the dominant outward component (δz + ) shows a spectrum with 1/f range, while the inward (δz − ) does not. This can be easily explained according to the model proposed in this work. As only the outward component of the fluctuations really reaches a high enough level to satisfy δB/B ∼ 1, only its spectrum has a break at l 0 . Viceversa, the lower power in the inward component keeps its spectrum below the threshold for the formation of a 1/f range and displays thus a more extended -5/3 inertial range. A consequence of this model is that, unless the 1/f range is formed in the Corona and just advected in interplanetary space preserving its shape, it should gradually disappear moving closer to the Sun where δB/B < 1. It will be possible to test this prediction soon with Parker Solar Probe measurements as close as the Alfvén radius. Finally we note that a similar saturation of Alfvénic fluctuations could be at work also in other space and astrophysical plasmas with δB/B ∼ 1, leading to spectra shallower than Kolmogorov at large scales (e.g. Hadid et al. 2015). Authors acknowledge valuable discussions with S. Landi, R. Bruno, A. Verdini, and M. Velli. This work was supported by the Programme National PNST of CNRS/INSU co-funded by CNES. TH is supported by STFC grant ST/N000692/1, DS is supported by STFC studentship ST/N504336/1, CHKC is supported by STFC Ernest Rutherford Fellowship ST/N003748/2.	2018-12-13T22:17:38.000Z	2018-12-13T00:00:00.000	{ "year": 2018, "sha1": "3b56db0a378cc76002dfa77cf3ab4f0eb833209d", "oa_license": null, "oa_url": "http://spiral.imperial.ac.uk/bitstream/10044/1/74149/7/On-the-1f-Spectrum-in-the-Solar-Wind.pdf", "oa_status": "GREEN", "pdf_src": "Arxiv", "pdf_hash": "849a0a1893b8ccd922ea89aa56a56fc1da63772b", "s2fieldsofstudy": [ "Physics" ], "extfieldsofstudy": [ "Physics" ] }
118439353	pes2o/s2orc	v3-fos-license	Numerical modeling of carbon dioxide sequestration on the rate of pressure solution creep in limestone: Preliminary results When carbon dioxide (CO2) is injected into an aquifer or a depleted geological reservoir, its dissolution into solution results in acidification of the pore waters. As a consequence, the pore waters become more reactive, which leads to enhanced dissolution-precipitation processes and a modification of the mechanical and hydrological properties of the rock. This effect is especially important for limestones given that the solubility and reactivity of carbonates is strongly dependent on pH and the partial pressure of CO2. The main mechanism that couples dissolution, precipitation and rock matrix deformation is commonly referred to as intergranular pressure solution creep (IPS) or pervasive pressure solution creep (PSC). This process involves dissolution at intergranular grain contacts subject to elevated stress, diffusion of dissolved material in an intergranular fluid, and precipitation in pore spaces subject to lower stress. This leads to an overall and pervasive reduction in porosity due to both grain indentation and precipitation in pore spaces. The percolation of CO2-rich fluids may influence on-going compaction due to pressure solution and can therefore potentially affect the reservoir and its long-term CO2 storage capacity. We aim at quantifying this effect by using a 2D numerical model to study the coupling between dissolution-precipitation processes, local mass transfer, and deformation of the rock over long time scales. We show that high partial pressures of dissolved CO2 (up to 30 MPa) significantly increase the rates of compaction by a factor of ~ 50 to ~ 75, and also result in a concomitant decrease in the viscosity of the rock matrix. Introduction The subsurface sequestration of CO 2 in geological repositories is frequently cited as a promising solution for reducing the amount of anthropogenically-produced CO 2 in the atmosphere. Some of the important issues involved in the long-term sequestration of CO 2 in such sites are discussed in an overview by Wawersik et al. (2001), for example. Therein it has been judged essential that models need to predict CO 2 sequestration behavior over time periods of several thousand years, which is the same order of magnitude as some climatic cycles. Therefore, in order to advance our knowledge of processes involved in the geological sequestration of CO 2 , one of the most important objectives confronting geoscientists is understanding and quantifying all of the mechanochemical processes, at both short and long time scales, that are relevant to CO 2 storage in geological formations. When considering O H CO 2 2 − injection into a geological repository, the following general chemical reactions (Stumm and Morgan, 1996) (2) Since we examine CO 2 sequestration within the context of pervasive pressure solution creep (PSC) in limestone, Eq. (2) is thus of particular relevance to this study. PSC is a mechano-chemical process characterized by ductile deformation and local mass transfer affecting water saturated porous rocks (e.g. Weyl, 1959;Rutter, 1976;Gratier and Guiguet, 1986;Dewers and Ortoleva, 1990;Spiers and Brzesowsky, 1993;Gundersen et al., 2002;Yasuhara et al., 2003). This ductile deformation mechanism occurs in the upper crust and plays an important role in the compaction of sedimentary rocks during diagenesis (Ortoleva, 1994;Tuncay et al., 2000). PSC is driven by differences in chemical potential induced by differential stress along grain surfaces in the rock matrix. PSC can be modeled as a serial process involving four successive steps: -stress-enhanced dissolution at grain-grain interfaces subject to elevated normal stress; -diffusion of dissolved material (solutes) through intergranular fluid films; -precipitation of dissolved material in adjacent pore spaces (grain surfaces in contact with pore fluid); -transport of dissolved material to distant pores, which can induce local mass transfer (Gundersen et al., 2002). Since it is assumed that PSC operates as a serial process, the slowest step imposes the overall rate for deformation (Rutter, 1976;Gratz, 1991). The first three individual steps of PSC (dissolution, diffusion, precipitation) are in turn influenced by local parameters such as temperature, stress state of the rock matrix, fluid pressure, and fluid chemistry (Rutter, 1976). The PSC mechanism of deformation is slow and operates over long geological time scales. Because of this, it is even possible that the slow step can change from one process to another over time during compaction of the rock matrix. Because PSC operates over geological time scales, correctly predicting the longterm stability of a CO 2 repository requires accurate modeling. PSC models are based for the most part on kinetic and equilibrium parameters derived from laboratory dissolution and precipitation experiments, as well as from pressure solution experiments that typically run for only a small fraction of the time scales associated with natural PSC deformation. Injection of CO 2 causes chemical and flow regime perturbations that affect the PSC process, causing the porosity, permeability, and mechanical stability of the porous rock matrix to evolve over time. There are multiple reasons that are responsible for this. The acidification of pore fluids due to the dissolution of CO 2 generally increases rates of fluid-rock interactions. This is particularly important in limestones where a decrease in pH increases both the rate of calcite dissolution and calcite solubility. In addition, higher concentrations of dissolved calcium carbonate can result in increased rates of precipitation. The porosity of the rock matrix is reduced during PSC, due primarily to grain indentation and precipitation in pore spaces. Taken together, the dissolution, diffusion, and precipitation processes have the potential for modifying the long-term porosity and permeability of the repository rock, as well as its mechanical stability. The deformation of chalk, both dry and in the presence of fluids, has been widely investigated experimentally. Most of the published work is based on chalk deformation data that have been interpreted solely in terms of mechanical processes (e.g., Botter, 1985;Da Silva et al., 1985;Jones and Leddra, 1989;Monjoie et al., 1990;Shao et al., 1994;Piau and Maury, 1995;Schroeder and Shao, 1996;Homand et al., 1998;Risnes and Flaageng, 1999). Recently, however, a few studies have examined chalk-fluid deformation within the context of chemical processes associated with PSC (Hellmann et al., 2002a, b;Heggheim et al., in press). In addition, PSC rates have also been measured for calcite aggregates (Zhang et al., 2002) and single calcite grains (Zubtsov et al., 2004). In this study, we examine the effect of elevated concentrations of dissolved CO 2 on the overall PSC rate of limestone dominated aquifers or reservoirs at burial depths relevant to CO 2 storage (1000-3000 m). The model treats the post-injection phase of sequestration, where the aqueous pore fluids have been homogeneously acidified by the presence of CO 2 . The partial pressure of CO 2 is fixed and remains constant for each simulation. The model also makes the approximation that the 2 CO p equals the pore pressure in the reservoir. Using a 2D numerical model, we examine how various parameters such as grain size, burial depth, rock texture, and the partial pressure of CO 2 modify the rate of matrix compaction by pervasive PSC. Only the effect of CO 2 dissolved in water is considered (i.e. a single subcritical aqueous phase); we do not consider the injection of a supercritical CO 2 phase since the reactivity (i.e. solubility) of calcite therein is predicted to be minimal. Below, we first present a brief review of the relevant thermodynamics and kinetics of the calcite-H 2 O-CO 2 system that is used to model the dissolution-precipitation processes. This is followed by a description of the basics of our PSC model. Lastly, we present several results from the 2D simulations. Conventions Before we address the thermodynamics of the calcite-H 2 O-CO 2 system, it is important to note the conventions we use. First, even though equilibrium constants are by definition based on activities of product and reactant species and phases, we make a simplifying assumption of equating activities of aqueous species with concentrations (in mol m -3 ); that is we assume that their activity coefficients are equal to 1. This approximation is based on the relatively low ionic strength of the solutions (the maximum ionic strength (I) = 0.115 molal, calculated using EQ3NR, Wolery, 1992, and the activity coefficient is greater than 0.85, see Fig. 1 of Kervevan et al., 2005). Solid phases are always assigned an activity equal to 1. We also assume that the molarities are equal to the molalities and the density correction is neglected. Lastly, the use of the term 'fluid' has the same meaning as 'solution', no supercriticality is implied. Overview of equilibrium thermodynamic relations at 25 °C and 0.1MPa In this study we consider the system CaCO 3 -H 2 O-CO 2 with the following equilibrium constants (Nordstrom et al., 1990) 6.5, Langmuir, 1997). The only exception to this is the species − 2 3 CO whose concentration, albeit low, is necessary for the calculation of the equilibrium constant sp K . In addition to the above equilibria, the following aqueous charge CO , and OHare of minor importance and can be neglected in this charge balance. Using the chemical equilibria and charge balance relations given in Eqs. 3a-i, the concentrations of all chemical species pertinent to the calcite-H 2 O-CO 2 system at STP can be calculated. However, in order to be useful for PSC, the code recalculates these equilibrium concentrations to conform to the relevant temperature and pressure conditions of the fluids present either in the contact zone or in the pore space (as discussed in the following sections). Note, however, that the charge balance relation in Eq. 3i is not pressure or temperature dependent. In order to simplify the numerical code, only the + 2 Ca concentration is allowed to vary as a function of time. All of the other species' concentrations remain fixed at their initial equilibrium values calculated for t = 0. Thus, at this stage in the development of the code, we have made the simplifying assumption that PSC depends only on one chemical species, such that only + 2 Ca dissolution, precipitation, and local transport are rate-determining. 1.3. Equilibrium thermodynamic relations at elevated temperatures and pressures 1.3.1. Dependence on temperature An empirical expression for the solubility of calcite (Eq. 3f) as a function of temperature T and salinity S has been derived by Mucci (1983) and is given by where T is in Kelvin. As defined by Mucci (1983), this equation is valid for 0-40°C and S = 5-35 g/kg. For the purposes of this study, we restrict ourselves to the first four terms of Eq. 4, which is the expression originally determined by Plummer and Busenberg (1982) and is applicable at 0-90°C. The values of a, b, c, d, and e in Eq. (5) are given in Table 2. 1.3.2. Dependence on pressure of the equilibrium constant of calcite dissociation The expression for the pressure dependence of K sp (Eq. 3f) that we use is the following (Lown et al., 1968;Millero, 1982) ln K sp Here, ΔV r 0 and Δκ r are the molal volume and compressibility changes for the dissolution reaction, P is the applied pressure in MPa, and R = 8.32 cm 3 MPa mol -1 K -1 . Thus, using 0 Ksp V Δ = -58.3 cm 3 mol -1 at 25°C (Langmuir, 1997) and Δκ r = -1.5·10 -3 cm 3 MPa -1 mol -1 at 25°C for the solubility of calcite in pure water (Owen and Brinkley, 1941), we can then calculate K sp at elevated pressures. As an example, a change in system pressure from P 0 =0.1 to P=10 MPa results in the following increase in the solubility product: 26 . In the present version of the code, the pressure dependencies of the other equilbrium constants (Eqs. 3a-e, g-h) have not been considered. Overview of kinetic rate laws for calcite dissolution The rates of calcite dissolution and precipitation in aqueous solutions have been actively investigated over many decades (e.g. see a thorough review by Morse and Arvidson, 2002; and references therein). One of the most widely used empiricallyderived kinetic rate laws is based on the work of Busenberg, Plummer, and co-workers (Plummer et al., 1978;Busenberg and Plummer, 1986), in which the rate of reaction (dissolution or precipitation) of calcite can be expressed as a function of three parallel forward rates and one backward rate where R is the overall reaction rate, k 1 , k 2 , k 3 , k 4 are kinetic rate constants, and [i] represents the bulk solution concentrations (in mol/m 3 of water) of aqueous species i. When considering only the three forward reactions (k 1 , k 2 , k 3 ) in Eq. (7) space there is also a region where all three terms must be considered (see Fig. 12, Plummer et al., 1978). In circum-neutral pH solutions, and close to calcite equilibrium ( Ω > 0.6), the dissolution reaction can be represented by a simpler alternative kinetic rate law for dissolution that has the following form (Berner and Morse, 1974;Busenberg and Plummer, 1986;Wollast, 1990;Hales and Emerson, 1997) n sp diss n diss Here k diss is the rate constant for calcite dissolution (denoted by k 5 in Busenberg and Plummer, 1986), is the chemical reaction affinity term (i.e. degree of solution undersaturation or oversaturation), n is the reaction order, and Ω is defined as the ratio of the ion activity product (see Section 1.1) to the equilibrium constant K sp (Eq. 3f). For calcite, n = 1 and log k diss = -9.93 at 25 °C and 0.1 MPa (Plummer et al., 1978, units of k diss in mol cm -2 s -1 ). A linear relation between the rate of dissolution and chemical affinity (i.e. n ≈ 1) at conditions close to equilibrium has been reported in other studies, as well (e.g. Cubillas et al., 2004). The rate law given in Eq. (8) has in fact been described as, "the most commonly used equation in geosciences to describe the rate of carbonate mineral dissolution" (Morse and Arvidson, 2002). The kinetic expression in Eq. (8), which describes the net rate of reaction (dissolution and precipitation), has both a mechanistic and empirical origin (Wollast, 1990), based on the However, the simple, quasi-linear relationship between rate and chemical affinity shown in Eq. 8 is perhaps still debatable. As an example, Svensson and Dreybrodt (1992) report that for natural (i.e. impure) calcite, the dissolution rate-affinity relation deviates from a linear relation chose to equilibrium, yielding a value of n ≈ 3-4 in Eq. 8. In the following treatment, however, we neglect the uncertainty associated with this potential non-linearity, as it is much smaller than the uncertainty associated with the measurement of the value of k diss that we adopt. Effect of temperature and pressure on the rate constant k diss Elevated T and P conditions have two effects on the overall rate of calcite dissolution: they modify the rate constant k diss , and also the chemical affinity term . The dependence of k diss on T can be expressed using the classical Arrhenius where A is a pre-exponential frequency factor, and E a is the overall thermal activation energy. Plummer et al. (1978) determined an E a = 33.05 kJ mol -1 for the rate constant k 3 (Eq. 7); this corresponds to A = 11.7. An E a value of 35 kJ mol -1 was also reported by Sjöberg (1978). Morse and Arvidson (2002) proposed that activation energy values in this range are consistent with surface reaction control kinetics, but the value of 35 kJ/mol suggests a mixed control: activation energies in the range 0-25 kJ/mol are typical for diffusion in a fluid media, whereas 50 kJ/mol and greater is typical for surface reaction control (Lasaga, 1998). In the study of Busenberg and Plummer (1986), the values of k 3 (Eq. 7 above) and k diss (Eq. 8 above, equiv. to k 5 ) are very close in value, and therefore we base the temperature dependence of k diss on the following empirical expression for k 3 (25-48°C) given in Plummer et al. (1978) ( ) The effect of fluid pressure (neglecting the effect of 2 CO p -see next section) on the rate constant k diss is not considered, given the lack of pertinent data in the literature. (Plummer et al., 1978;Busenberg and Plummer,1986;Arakaki and Mucci, 1995 In addition, these authors have also shown that the ionic strength has no effect on the dissolution rate constant. In our model, we use the results of Pokrovsky et al. (2004) for defining the change in k diss as a function of 2 CO p . Coupling between dissolution-diffusion-precipitation processes and matrix deformation Using our model we compute how the petrophysical properties of the rock matrix (permeability, porosity) and the volumetric strain (compaction) evolve with time due to PSC. The rock aggregate is a monomineralic limestone (i.e., pure calcite), with either a homogeneous or variable grain size (i.e. random spatial variation based on a uniform grain size distribution). The model treats the coupling between chemical reactions of calcite dissolution and precipitation, diffusion, and PSC deformation in two steps: textural equations at the grain scale allow for grain deformation due to chemical reactions occurring at contact and pore surfaces (including diffusion in the contact fluid); -a mass conservation relationship at a global scale which takes into account the transfer of dissolved material by diffusion in the interconnected pore spaces. Therefore, microscopic grain scale processes and macroscopic mass transport within the model volume are fully coupled (for a complete review, see Gundersen et al., 2002). The basis of our model for treating the effect of elevated normal stress on individual grains comprising a rock matrix is the thermodynamic relationship which relates normal stress to chemical potential (Gibbs, 1878;Kamb, 1961;Paterson, 1973;Lehner, 1995;Renard et al., 1999). The higher chemical potential of calcite surfaces in the contact zone results in a higher solubility with respect to calcite surfaces of the pore space. Ultimately, this difference in chemical potentials drives two processes: stressenhanced dissolution of the contact zone surfaces and diffusion of dissolved material from the contact zone out to the pore fluid. The diffusion of dissolved material occurs within a trapped water film separating the grains (Rutter, 1976;de Meer et al., 2002;Dysthe et al., 2002b). Finally, grain-scale dissolution-precipitation processes in pore spaces are coupled to bulk diffusion of solute within the entire porous medium. The conservation relationships are derived by a global mass balance of the solute phase in the pore volume and a local mass balance at each grain contact (Gundersen et al., 2002). These relationships are then coupled to equations which express the deformation of the grains and the evolution of the rock texture. Deformation of individual grains and rock matrix In our model the rock matrix is a 20x20 cm domain, located at a specified depth. The geological conditions (stress at the boundaries, pore pressure, temperature) are assumed to stay constant and are chosen to represent relevant conditions for CO 2 sequestration (depths between 1 and 3 km, see Table 3). The rock matrix is modeled as a network of solid calcite grains with well-defined grain-grain contacts and interconnected pore spaces between grains (Fig. 1). The aggregate grain framework is represented by a dense cubic packing of truncated spheres that represent the individual grains. In Fig. 1 it is also important to distinguish the two different types of grain-fluid interfaces that are treated in the model: -grain surfaces at intergranular (i.e. grain-grain) contacts that are separated by a trapped fluid film, -'free' grain surfaces that are in contact with pore fluids. The intergranular grain boundary is considered as a flat interface (Hickman and Evans, 1995) with a mean roughness of several nanometers averaged over the contact. This interfacial contact zone contains a trapped fluid film that is postulated to have the unusual characteristic of being able to support a shear stress transmitted by the normal stress imposed on the grains. In addition to nanoscale topography, there is also ample evidence that grain-grain contact zones have a structure of channels and islands at a micrometer scale (Hickman and Evans, 1992;Schutjens and Spiers, 1999;Renard et al., 1999;Dysthe et al., 2002a). The exact relation between these nano-and microscale features is not yet well understood. The initial radius of the spherical grains, L f , can vary between the different elements. The other lengths, describing the truncations of the spheres, L x and L z , are given . The choice of this value allows initial porosities close to 30%. Each simulation lasts until the porosity in the whole simulation domain is less than a threshold value equal to 5%. Below this value, transport by diffusion between the pores ceases and pressure solution only continues locally inside this element, as a closed system. The decrease in porosity from an initial 30% to a final value of 5% represents a finite volumetric strain of approximately 20%. The time associated with this porosity reduction defines the average rate of deformation for PSC and forms the basis for comparing the various cases examined (variable depth, 2 CO p ) The entire matrix of grains with cubic packing is subjected to a constant normal stress component (σ n ). We then define a contact normal stress component (σ c ) as the mean stress at each grain contact surface. This stress depends on the relationship between the surface area and the diameter of the sphere. This relationship is independent of the initial size of the individual grains (Dewers and Ortoleva, 1990), and the stress at the contact is proportional to some (positive) power of the porosity. The evolution of the rock texture, which leads to a compaction of the aggregate, is computed as a result of the coupling between dissolution, diffusion, precipitation and mass transport in the fluid. This textural model allows us to take into account the sequential coupling between chemical processes (dissolution at grain contacts, diffusion, and precipitation on free pore surfaces) and the mechanical evolution (deformation of the grains) of the rock matrix. The resulting model is then a set of highly coupled non-linear equations that can only be solved using numerical methods. All of the parameters used in the following equations and their units are given in Table 1. Matrix-fluid chemical equilibrium at t = 0 The initial state of the model is based on two separate equilibrium chemical states, such that at t = 0, the rock matrix is in equilibrium with, respectively, the contact and Thus, an initial chemical imbalance exists between the contact zones and the pore spaces. The evolution of the chemistry of the contact fluid and the pore fluid is treated in a different manner once deformation starts, that is for t > 0. The initial chemical imbalance between the contact zone fluid and the pore space fluid drives diffusion that initiates the PSC deformation of the rock matrix. This diffusion process, which is constrained by the code to depend only on the negative [ Conservation equations in the pore fluid Defining c p, Ca to represent the concentration of where the lengths L x , L y , L z are geometric grain parameters defined in Fig. 1, D Ca on the pore walls (i.e. pore free faces) and the solute moles derived from diffusion from the grain contacts to the pore fluid, respectively (see Table 1 for symbols). The first term on the right side in Eq. (11) describes the diffusive transport of solutes in the pore space by a standard Fickian diffusion relation. The second term on the right hand side in Eq. (11) represents mass loss due to precipitation on the grain surfaces and is given as: where k p is the rate constant for calcite precipitation, A p is the grain surface area in contact with the pore fluid (reactive area), is the ion activity product in the pore fluid, and K sp,p is the calcite solubility product in the pore fluid. Notice that as the geometry of the grains changes during compaction, A p varies as well. At conditions close to equilibrium, the rate expression used in the model to calculate the overall reactivity of calcite does not differentiate between dissolution and precipitation (i.e. k diss = k prec ). This appears to be a reasonable approximation based on available experimental evidence at conditions very close to equilibrium (e.g. Fig. 8, Busenberg and Plummer, 1986). Conservation equations at the grain contacts and grain shape evolution The last term in Eq. (11) describes mass production due to diffusional mass transport out of the intergranular contact zone and reads: where c c,i,Ca is the concentration of + 2 Ca in the grain contact perpendicular to axis i (i=x, y, z). The driving force for this transport is a concentration gradient related to the difference in stress concentration at intergranular and free grain contacts. In the model, the length of the intergranular diffusion path relates directly to the rate of compaction (i.e. smaller grain matrices compact faster than larger grains). The transfer of dissolved + 2 Ca within the interfacial fluid from grain-grain contacts to the pore fluid occurs via a diffusion process. Thus, the chemistry of the interfacial fluid evolves not only by the addition of material via dissolution, but also by the removal of dissolved material by diffusion. In general, the rate of ion diffusion (D c ) in thin, confined interfacial fluid films is not directly measurable, but recent studies (e.g. Dysthe et al., 2002b;Alcantar et al., 2003) conclude that the rate is no more than two orders of magnitude slower than the corresponding bulk fluid value (at a given T and P). The adopted value for the diffusion coefficient D c of + 2 Ca along the grain contact used in the model is 0.01·D p (D p is the diffusion coefficient in the pore fluid) at the T and P of the contact fluid. The coefficient of diffusion within the water film, D c , follows an Arrhenius law with an activation energy of 15 kJ/mol (Dysthe et al., 2002b). Considering the uncertainty in D c , we can estimate that the true value probably differs by at least one order of magnitude from the value we have chosen. In Eq. (13) the thickness of the water film at the grain contacts, Δ, is divided by two since it is shared between two contact surfaces. We assume that this water film has a constant thickness of 3 nm (Dysthe et al., 2002b). The global mass balance equation (Eq. 11) is then coupled to Eq. (14) which represents the local mass balance for each contact on a truncated spherical grain. The concentration of dissolved material within the intergranular fluid phase is given as the difference between the mass produced by dissolution and the mass lost by diffusion: In the expression above, R c,i is the radius of the contact surface in the i-direction, which is a function of the truncations of the spherical grains. The first term in Eq. (14) represents the local production of dissolved material by dissolution at the grain contacts. The model defines the chemical reaction (i.e. dissolution) of grain-grain contacts in terms of the rate law already presented in Eq. 8, where R is the overall rate (mol m -2 s -1 ), k diss is the rate constant, Q c,i is the ion activity product in the contact zone, and K sp,c,i is the calcite solubility product in the intergranular contact zone perpendicular to the i axis (i=x, y, z): Using this law, one can write the flux corresponding to the dissolution process as: is not limited to an increase in the rate of dissolution (via k diss ), however, since this also results in a concomitant increase in calcite solubility (i.e. K sp,c and K sp,p ), which will also affect the rate of PSC. The grain shape evolution is given by the equations where V is the molar volume of calcite (3.69·10 -5 m 3 mol -1 ) and the calcite precipitation rate k prec is equal to the dissolution rate constant k diss . With these equations the code continually updates the texture of the grains and thus couples chemical reactions to grain deformation. It is important to note that even though the two kinetic processes, dissolution (Eq. 17) and precipitation (Eq. 18), respectively control the grain shape evolution in the contact zones and the pore spaces, these processes are coupled to diffusion in the contact zones, and therefore are not independent. This coupling results in a modification of the chemical affinity terms in both rate equations above (bracketed terms), thereby changing the overall rate of grain shape evolution. As an example, taking the rate of dissolution in The pore space, however, is a special case since dissolved material enters by diffusion from adjacent contact zones, but it can also diffuse to other pore spaces further away. Stress at grain contacts The simulation domain is a square rock matrix located at depths of 1, 2 or 3 km and is exposed to constant lithostatic stress and temperature. The normal stress component on an i-contact, σ c, i , is related to the lithostatic stress σ n on the boundaries Γ i by the three relationships which take into account the stress concentration effects due to the porosity of the rock (Dewers and Ortoleva, 1990). These three equations are obtained by circular permutation of the indices in the following relationship: where i, j, k represent the three space directions x, y, and z. When the lithostatic stress is kept constant, the contact surface area A i increases as the grains become more and more truncated (Fig. 1). As a consequence the stress normal to this contact decreases with time, and therefore, the driving force for pressure solution also decreases with time. All of these phenomena are due to the continual removal of material at grain-grain contacts by dissolution and the outward diffusion of this dissolved material into the pore fluid. The final equation updates the porosity, which is defined as the difference between the volume of a cubic box, with lengths L x , L y , L z , and the volume of the truncated grains (Dewers and Ortoleva, 1990). This relationship arises from the textural model of truncated spheres ordered in a centered cubic network (Fig. 1). Numerical methods and boundary conditions The equations presented in sections 2.4-2.6 are highly coupled. The deformation of the grains is coupled to transport in the pore fluid through the diffusion step at the contacts and the Fickian diffusion in the pores. Therefore we choose two different numerical modelling techniques. We couple a discrete treatment of the grains (Eqs. 17- 18) with a continuous determination of the dissolved Ca in the pore fluid (Eq. 11). The space domain is a 2D rock sample in the x-z plane and is modeled as a cubic packing of truncated spheres. It has a thickness of one layer of grains in the y direction (Fig. 1) The geometrical evolution of the grains is modeled as a discrete process where the grains in each grid element are treated separately (see . In order to visualize the deformation of the rock matrix as the grains deform during PSC, an adaptive grid is developed. The nodes change position after each time step as the grains deform. In addition, the contact surface area of each element is updated according to the lengths and radii of the truncated spheres, and is subsequently used to calculate the porosity loss in the xz-plane. The total change in volume of each element is then obtained by multiplying the volume variation of each grain by the number of grains in the element. The total deformation of the rock sample is determined by a summation over all the elements. The mass conservation equation (Eqn. 11) is derived from a continuum model, assuming a continuous solute phase. This equation is solved using a Galerkin finite element method, with linear basis functions on the spatial domain and a Crank-Nicolson scheme in the time domain. All the numerical schemes were programmed in C++ and make use of the numerical library Diffpack (Langtangen, 1999;Gundersen et al., 2002). The program was extensively tested for numerical stability. In all of the simulations, we verified that the total mass of solid was conserved as it should be. The time step is adjusted such that the maximum deformation is less than 0.1% in all the elements between each time increment. Therefore, the time step automatically adapts in the regions with the most rapid deformation. Results In our model, three different geometries of rock matrix are studied: -a single homogeneous layer with an initially constant grain size; -a sedimentary layered rock with an initial local variation of grain size; -a gouge filling a fracture. In the results presented here, we aim to compare the evolution of the rock with and without CO 2 injection. In order to compare the various simulations, we have chosen a reference case that provides a normalized time scale for the measured PSC rates: the reference (see Figs. 2a and 6) is based on the time needed to achieve a porosity of 5% for a rock at 1 km depth and 40 °C, with an initial grain size of 2 mm (L f = 1 mm), and a pore fluid that is CO 2 -free (i.e. 2 CO p = 10 -4.5 MPa). In all the simulations, the times scales of deformation are normalized to this reference case (total time for reference deformation equals 740,000 years-see Table 3). In this way we measure the degree to which the sequestration of CO 2 perturbs the system and enhances deformation and compaction by PSC. The absolute time scales for PSC deformation for all of the simulations are given in Table 3. Because of uncertainties with respect to several parameters (k diss , k prec , D c ), the absolute time scales related to PSC deformation are difficult to quantify and should be used with caution (see Discussion). In all cases, both CO 2 -free and at elevated 2 CO p , the PSC process is controlled by the diffusion process in the contact zones. This indicates that the intergranular diffusion process is slower than either dissolution in the contact zones or precipitation in the pores. Diffusion in the contact zones, controlled by the + 2 Ca concentration gradient between the intergranular fluid and the pore fluid, therefore, plays the dominant role in defining the rate of diffusion, the rate of grain indentation, and most importantly, the overall rate of PSC deformation of the rock matrix. This result is predictable, given the rapid kinetics of calcite dissolution and precipitation. Figure 2c, which shows that the + 2 Ca concentration in the contacts is always greater than that in the pore fluid, graphically demonstrates that the rate of diffusion is the limiting process of PSC for these simulation conditions. CO 2 -free simulations The rock matrix, located at 1 km depth, consists of grains with a homogeneous size of 2 mm, which corresponds to a coarse oolitic limestone. We consider here a CO 2free pore fluid (i.e. atmospheric 2 CO p = 10 -4.5 MPa). In our model, a typical compaction process without CO 2 can last between 10 4 to 10 6 years, depending on the choice of the parameters D c , k prec, k diss . As already stated, this configuration represents a reference case for comparison with other simulations at elevated 2 CO p . As PSC proceeds, the porosity decreases progressively with time both by grain indentation and by precipitation in the pores (Fig. 2a). As shown in Fig. 2b, L f increases whereas L z decreases. With time the grains become truncated and the stress on the contacts decreases. As a consequence, the gradient in [ Ca ] between the grain contacts and the pore fluid also decreases (Fig. 2c), leading to a decrease in the rate of overall PSC deformation For a heterogeneous medium, such as a layered rock, the grain size may vary spatially (Fig. 3). In this case, the PSC rate is faster in the layers with small grains than in layers with coarse grains. Because of this initial grain size difference, the different domains compact at different rates. Layers with smaller grains reach the limit of 5% porosity faster than coarser grain layers. This result is an example of the well-known inverse dependence of pressure solution creep rate on grain size (Rutter, 1976;Gratz 1991). Similar behavior can be observed for a fracture filled with a gouge containing smaller grains (Fig. 4). Both examples show that compaction is homogeneous in the domains with a constant grain size. We do not observe mass transfer between layers as we did for the simulation of PSC in sandstones (Gundersen et al., 2002). This is related to the faster compaction rate in limestones, and the smaller time scales involved for long distance diffusion in the pores. Effect of elevated partial pressures of CO 2 When CO 2 is present at hydrostatic pressure (i.e. 2 CO p = P p ), PSC is significantly faster compared to the CO 2 -free case. This situation is clearly visualized in Figure 5 which shows the compaction process for two fractured rocks at 2 km depth, one with 2 CO p =10 -4.5 MPa (i.e. CO 2 -free), the other where 2 CO p = P p (see Table 3). The pattern of the porosity reduction is similar in both cases, however the characteristic time scales are very different; at 2 CO p = P p , the process is roughly 65 times faster. We also compare the compaction process at different depths with and without injection of CO 2 . With no injection of CO 2 the rock matrix compacts naturally as PSC progresses. As depth increases, the temperature and the stress increase as well. On the one hand, the effect of stress enhances PSC, mainly by increasing the solubility and the rate of dissolution. On the other hand, increasing the temperature decreases the solubility of calcite (i.e. calcite has retrograde solubility), and this therefore slows down the PSC rate. These two competing factors explain why the PSC rates increase by less than a factor of 3 between 1 and 3 km burial depth (Figure 6a). In a similar manner, for elevated 2 CO p , the relative increases in the PSC rates with increasing depth are modest (Table 3), again reflecting the competing effects of stress, solubility and kinetics. When CO 2 is present at elevated concentrations, where 2 CO p equals the pore fluid pressure, the rates of PSC are greater by factors of ~50-75 (Fig. 6b), as compared to the CO 2 -free simulations (Fig. 6a), the exact amount depending on depth (Table 3). This translates to a reduction in porosity significantly more rapid than compaction at the same depth in the absence of CO 2 . This can be explained by two factors. First, at elevated levels of CO 2 , the pH decreases dramatically ( (Pokrovsky et al., 2004) and precipitation. However, an increase in the rate of calcite dissolution in the intergranular zone should not affect the rate of deformation since it appears that the diffusion process is the rate-limiting step in the overall PSC process. On the other hand, an increase in the rate of precipitation in the pores has a positive effect on the rate of diffusion from the contact zone to the pores. What, if any, effect increased levels of CO 2 has on the rates of diffusion is unknown. Discussion and conclusions The injection of CO 2 into a geological reservoir and the resulting increase in the 2 CO p cause acidification of the pore fluids. As the pH of the pore fluids decreases, they become more reactive with respect to the rock matrix; this effect is especially pronounced for carbonates since both the solubility and the reaction kinetics increase dramatically with decreasing pH. This is in accord with the general finding that, to a first approximation, PSC rates depend linearly on the solubility of minerals (Rutter, 1976). Our model shows that for a given depth and temperature, elevated concentrations of dissolved CO 2 in the pore fluids lead to rates of PSC deformation that are up to 75 times faster than reference simulations based on CO 2 -free pore fluids. Thus, our results convincingly demonstrate the direct relation between elevated 2 CO p levels, calcite solubility, pore fluid pH, and increased rates of PSC deformation. Of particular importance to predicting the behavior of future CO 2 geological repositories is the accuracy of theoretical PSC deformation rates. Taking just one example from our model, the absolute time scale associated with porosity reduction from 30 to 5% for a carbonate rock matrix with CO 2 -free pore fluids at 1 km depth (40 °C, 2mm grain size) is on the order of 700,000 years (Table 3). This time frame is most probably unrealistically rapid, given that in nature PSC occurs over much longer time scales. As an example, it is observed that Mesozoic chalks and limestones retain some porosity after several tens of million years, which represents a time period several orders of magnitude greater. There are undoubtedly a myriad number of reasons for this, for which a detailed analysis is beyond the scope of this article. Nonetheless, we briefly discuss how some important parameters influence theoretical PSC deformation rates derived from our model, and how they may contribute to discrepancies with natural PSC rates. The present study reports 'preliminary results' that are derived from a model that incorporates many simplifications that are both chemical and physical in nature. H into a consistent model. Even though a fully coupled model should produce more realistic PSC rates, we estimate that the deformation patterns would not differ too much from those generated in the present study. The thermodynamic, kinetic, and diffusion parameters that are incorporated in the code play perhaps the most important role in determining the accuracy of theoretical PSC rates of deformation. The rates that we have chosen for the kinetic rate parameters k diss and k prec , the value of n (see Eq. 8), and the overall form of the kinetic rate equation(s), determine the rates of dissolution and precipitation in the contact zone and pore space, respectively, and therefore potentially influence the overall rate of PSC deformation. The kinetic parameters we use are not unique and therefore future modeling will include sensitivity analyses based on a variation of these kinetic parameters. In addition, the accuracy of theoretical PSC rates depends on the availability and quality of experimental data, both kinetic and thermodynamic, for conditions relevant to CO 2 sequestration. Further refinements in PSC models will depend on additional experimental work to unravel more precisely the effect of elevated 2 CO p on k diss and k prec of calcite (and other minerals, as well) at conditions close to equilibrium at elevated T and P. If we consider CO 2 sequestration and PSC deformation within the context of conditions in a geological repository, the chemical complexity of natural pore fluids may dictate the need for additional kinetic parameters to be used in the kinetic rate laws. As an example, the presence of certains ions (e.g. PO ; Svensson and Dreybrodt, 1992;Zuddas and Mucci, 1998;Zhang et al., 2002) or humic acids (e.g. Zuddas et al., 2003) have been shown to inhibit calcite dissolution/precipitation rates. The inhibition of calcite kinetic rates has an important implication for carbonate PSC deformation since the limiting process may switch from diffusion limited to reaction limited. This points out that more experimental work is needed to investigate the effect of fluid chemistry on rates of PSC deformation. The mineralogy of the solid phase is also important. At present, our code only considers a monomineralic (pure calcite) limestone. The effect of impurities, such as clays, potentially has a large effect on PSC deformation. For example, it has been experimentally shown that the presence of clays can dramatically increase rates of PSC (Hickman and Evans, 1995;Renard et al., 2001). Even though the thermodynamics of the calcite-H 2 O-CO 2 system at STP are well known and quite accurate, future versions of our code need to further develop the temperature and pressure dependencies of most of the equilibria given in Eqs. 3a-h. Accounting for non-ideal behavior at high 2 CO p is also necessary with respect to Henry's law constant in Eq. 3b. Moreover, future versions of the code will need to consider equilibria relations in terms of activities and not concentrations; this will be especially important for solutions with elevated ionic strengths. Two other parameters of considerable importance, especially for the case of diffusion-limited PSC deformation, are the rate of diffusion and the thickness (and continuity) of the fluid film in intergranular contacts. Measurement of diffusion coefficients in thin films is experimentally challenging, and therefore the diffusion coefficient and film thickness will remain uncertain quantities that probably contribute significantly to the discrepancy between natural and theoretical PSC deformation rates. At this stage in the development of our model, many uncertainties exist with respect to the thermodynamic, kinetic, and physical parameters that have been incorporated in the code. Future versions of the code will hopefully be able to address many of these issues. We believe that the main value of the results generated in the present study lies not in the absolute rates, but rather in the relative rates of PSC deformation. Intercomparison of these relative PSC rates will help in better understanding and evaluating the long-term effect of increased rates of PSC on the porosity and viscosity of rock matrices in contact with high 2 CO p fluids. Another important benefit of PSC models such as ours is that they may help in focusing laboratory-based PSC experiments. The relation between solubility, pH and PSC rates has important ramifications with respect to future experimental laboratory protocols for investigating the sequestration of CO 2 . As an example, the injection of subcritical CO 2 -H 2 O mixtures into a rock sample could be simplified by the simple injection of an acidic aqueous fluid, where the pH is adjusted to be equivalent to the pH due to CO 2 acidification alone. Nonetheless, such an approach has its limitations and would not be applicable under all circumstances, especially in the case of injection of supercritical CO 2 fluids.	2017-09-14T08:37:22.138Z	2005-03-01T00:00:00.000	{ "year": 2008, "sha1": "ea99a6591b5b8460e12bb826a5e93fab50437f30", "oa_license": "CCBY", "oa_url": "https://ogst.ifpenergiesnouvelles.fr/articles/ogst/pdf/2005/02/renard_vol60n2.pdf", "oa_status": "GOLD", "pdf_src": "Arxiv", "pdf_hash": "ea99a6591b5b8460e12bb826a5e93fab50437f30", "s2fieldsofstudy": [ "Environmental Science", "Engineering", "Geology" ], "extfieldsofstudy": [ "Geology", "Materials Science", "Physics" ] }
106404037	pes2o/s2orc	v3-fos-license	Detection of Boulders in Side Scan Sonar Mosaics by a Neural Network Boulders provide ecologically important hard grounds in shelf seas, and form protected habitats under the European Habitats Directive. Boulders on the seafloor can usually be recognized in backscatter mosaics due to a characteristic pattern of high backscatter intensity followed by an acoustic shadow. The manual identification of boulders on mosaics is tedious and subjective, and thus could benefit from automation. In this study, we train an object detection framework, RetinaNet, based on a neural network backbone, ResNet, to detect boulders in backscatter mosaics derived from a sidescan-sonar operating at 384 kHz. A training dataset comprising 4617 boulders and 2005 negative examples similar to boulders was used to train RetinaNet. The trained model was applied to a test area located in the Kriegers Flak area (Baltic Sea), and the results compared to mosaic interpretation by expert analysis. Some misclassification of water column noise and boundaries of artificial plough marks occurs, but the results of the trained model are comparable to the human interpretation. While the trained model correctly identified a higher number of boulders, the human interpreter had an advantage at recognizing smaller objects comprising a bounding box of less than 7 × 7 pixels. Almost identical performance between the best model and expert analysis was found when classifying boulder density into three classes (0, 1–5, more than 5) over 10,000 m2 areas, with the best performing model reaching an agreement with the human interpretation of 90%. Introduction Hard substrata are ecologically important as they provide habitat structures for highly specialized communities that in turn significantly contribute to benthic production, and provide irreplaceable ecosystem services [1]. Often, hard bottom habitats feature a complex topography. The three-dimensional structures enhance small-scale spatial habitat variability and thereby lead to high biodiversity. They provide refuge or serve as nursery areas for a variety of species throughout the marine food-webs and the inhabiting communities play a major role in benthic-pelagic coupling and nutrient cycling. Consequently, marine hard-substrata are not only areas of special interest to study ecosystem functioning, but also for marine nature conservation and their detection should be a key goal of geological seafloor mapping. In the Baltic Sea, hardgrounds are mostly fields of cobbles (6.4 cm to 25.6 cm), boulders (25.6 cm to 4.1 m), and blocks (>4.1 m) [2] that eroded out of glacial till deposits [3][4][5]. Hard substrates as the southern Baltic boulder fields are protected habitats under European Habitats Directive (HD 92/43/EEC annex 1 1170-reefs). Boulder distribution and the area they provide for settlement of marine life is not quantitatively known. A reason for this is that cobbles and boulders (or features of similar acoustic appearance such as certain biogenic reefs [6] or underwater mines [7]) currently cannot reliably be detected automatically especially on heterogenous seafloor. In current practice, boulder detection often requires time-consuming manual interpretation of sidescan-sonar or multibeam echo sounder backscatter mosaics and/or confirmation by underwater video footage [8]. In principle, it has to be differentiated between the detection of buried boulders, and boulders exposed at the seafloor. Buried boulders may be detected by seismic methods [8]. While the detection of slightly buried boulders is important, especially considering dynamic areas such as parts of the North Sea, where boulders may be frequently exposed or covered based on environmental conditions [9], the focus of this study is on the detection of stable surficial boulders widespread in the Baltic Sea. Ongoing, full-coverage hydroacoustic surveys in the Baltic Sea [10] utilize sidescan-sonar technology. Lacking the morphological information provided by multibeam echo sounders frequently used for habitat classification in recent years [11], towed sidescan-sonar systems offer the advantage of lower grazing angles and thus increased presence of acoustic shadows due to the elevated boulder surface [12,13]. Since backscatter mosaics are essentially georeferenced acoustic imagery of the seafloor following data processing, techniques for the detection of objects in images can be applied to the problem. Techniques that have been developed for the identification of man-made objects rely on feature extraction and selection of classification algorithms by human experts [14][15][16]. These methods rely on spatial domain filtering to detect highlight-shadow sequences and often operate on raw sonar images to transparently provide high detection performance in real-time. In contrast, during training on labeled data, neural networks independently learn patterns in input data-in this study, comprising processed backscatter mosaics and no raw sonar data-that can be used for classification. Especially in the field of computer vision, deep convolutional neural networks [17] were found to be highly capable tools for the recognition and classification of objects within images. While neural networks have been previously applied to backscatter mosaics in the field of marine habitat mapping [18,19], these were not convolutional networks optimized for image analysis. Convolutional neural networks operate by convolving an input image with different kernel functions, thus generating a large number of filtered images that differ due to the different weights of the kernels. The basic assumption of convolutional neural networks is that no connection between all pixels of an image is required for classification, but that local pixel patterns have a higher significance. During training on labelled images, the weights of the randomly initialized kernels are continuously adjusted to be sensitive to characteristic features of the sample data. A stack of several trained convolutional layers, each layer operating on the output of the previous layer and combined with activation functions that allow the network to learn nonlinear relationships as well as regular reductions in image size to assess large scale patterns, detects increasingly abstract features of images and can be used for image classification tasks. Well known implementations of such convolutional neural networks for image classification are, e.g., AlexNet [20] or different instances of the VGG network [21]. Deeper convolutional networks become increasingly difficult to train due to different learning rates in the individual layers of the network, which is an issue commonly termed the vanishing gradient problem [22,23]. To resolve this issue, deep residual networks (e.g., ResNet) [24] were developed, which add shortcuts between selected layers of a convolutional network. This causes networks to learn on the residuals of the filtered input image, allowing the construction of very deep networks. For the purpose of object detection, the network structures introduced above are embedded into frameworks [25] that merge the task of image classification with the detection of boundary boxes or masks of different objects within one image. These frameworks are directly applicable to the interpretation of backscatter mosaics. Thus, the objective of this study is to train an object detection framework (RetinaNet) on boulders apparent in backscatter mosaics in the Baltic Sea ( Figure 1) and determine boulder densities as currently suggested for the German Mapping Guidelines for North Sea and Baltic Sea. Preparation of Train, Validation, and Test Datasets The Department of Hydrographic Surveying and Marine Geodesy of the German Federal Maritime and Hydrographic Agency (BSH) provided sidescan-sonar raw data. Data were recorded using an EG&G DF-1000 side scan sonar in July 2013. The 384 kHz channel was used to create the mosaics. The altitude of the sidescan-sonar in the investigation area varied between 15 m and 37 m above the seafloor, due to extensive seafloor morphology and a constant cable length. The recorded slant range was 84 m. Sidescan-sonar mosaics were produced within an administrative agreement between the Leibniz Institute for Baltic Sea Research Warnemünde (IOW) and the BSH. In this study, higher backscatter intensities are displayed in darker colors. To obtain a training dataset, three sidescan-sonar mosaics of ~2.5 km² in total ( Subsequently, the mosaics of the training area were exported as square GeoTiff tiles with a spatial extension of 25 m² and 225 m². The individual tiles include between 0 and more than 160 manually detected boulders. To avoid boundary effects, where bounding boxes of manually interpreted boulders are cut by the tile boundary, tiles were exported with an overlap of 1.0 m. This ensures that the large majority of observed boulders are completely visible on at least one image. Due to the overlap, identical boulders can be present on different images. Therefore, for the purpose of this study, subsequent detections that are within 0.5 m of an already present detection are ignored, which may cause an underestimation of boulders in areas with high boulder density. The coordinates of the manually interpreted bounding boxes in the database were compared with the geographical extension of each tile. If the bounding box of boulder is situated completely within a tile, pixel coordinates of the boulder were calculated and a training/validation sample was created. This check is necessary to avoid the boundary of an image tile being trained as a boulder. Preparation of Train, Validation, and Test Datasets The Department of Hydrographic Surveying and Marine Geodesy of the German Federal Maritime and Hydrographic Agency (BSH) provided sidescan-sonar raw data. Data were recorded using an EG&G DF-1000 side scan sonar in July 2013. The 384 kHz channel was used to create the mosaics. The altitude of the sidescan-sonar in the investigation area varied between 15 m and 37 m above the seafloor, due to extensive seafloor morphology and a constant cable length. The recorded slant range was 84 m. Sidescan-sonar mosaics were produced within an administrative agreement between the Leibniz Institute for Baltic Sea Research Warnemünde (IOW) and the BSH. In this study, higher backscatter intensities are displayed in darker colors. To obtain a training dataset, three sidescan-sonar mosaics of~2.5 km 2 in total ( Subsequently, the mosaics of the training area were exported as square GeoTiff tiles with a spatial extension of 25 m 2 and 225 m 2 . The individual tiles include between 0 and more than 160 manually detected boulders. To avoid boundary effects, where bounding boxes of manually interpreted boulders are cut by the tile boundary, tiles were exported with an overlap of 1.0 m. This ensures that the large majority of observed boulders are completely visible on at least one image. Due to the overlap, identical boulders can be present on different images. Therefore, for the purpose of this study, subsequent detections that are within 0.5 m of an already present detection are ignored, which may cause an underestimation of boulders in areas with high boulder density. The coordinates of the manually interpreted bounding boxes in the database were compared with the geographical extension of each tile. If the bounding box of boulder is situated completely within a tile, pixel coordinates of the boulder were calculated and a training/validation sample was created. This check is necessary to avoid the boundary of an image tile being trained as a boulder. Following this procedure, the results were randomly shuffled to create separated training (80% of samples) and validation (20% of samples) datasets using the scikit-learn library of Python [26]. For testing of the trained models, a fourth mosaic (Figure 1) was tiled with the same spatial extension, resolution and overlap. To assess the performance of the trained model on widespread 1 m resolution mosaics, the test backscatter mosaic was also downsampled to a resolution of 1 m. Geosciences 2019, 9, 159 4 of 16 Following this procedure, the results were randomly shuffled to create separated training (80% of samples) and validation (20% of samples) datasets using the scikit-learn library of Python [26]. For testing of the trained models, a fourth mosaic (Figure 1) was tiled with the same spatial extension, resolution and overlap. To assess the performance of the trained model on widespread 1 m resolution mosaics, the test backscatter mosaic was also downsampled to a resolution of 1 m. For classification and object detection, we use an open source RetinaNet [27] implementation in Python, available on GitHub (https://github.com/fizyr/keras-retinanet, last accessed on 6 February 2019). The architecture of RetinaNet includes a backbone convolutional neural network connected to a feature pyramid [28], which detects potential objects in the backscatter mosaics. The detected objects are forwarded to two subnetworks, one for classifying the objects within the bounding box (in this case, the two available classes are boulders and background), and one for the regression of the bounding box coordinates. As a backbone for classification we used ResNet50, with pretrained weights from the Microsoft Coco dataset [29] to speed up training. Initial tests not reported here were carried out using VGG16 as a backbone as well, but ResNet50 consistently provided better results. To update the weights of the convolutional layers, RetinaNet implements a focal loss mechanism [27] for backpropagation of errors during training, which puts an emphasis on the loss contributed by difficult training samples. RetinaNet expects by default an input image side length of 800 pixels, corresponding to 200 m or 800 m depending on the resolution of the image tiles (0.25 m or 1 m). During training, the extent of the input image is reduced to 1/16th of the original image. This deletes most of the cobbles and fine boulder information from the data if large scale mosaics were fed directly into the network. Hence, the smaller tiles exported from the backscatter mosaics were upscaled to values between 300 and 1200 pixels, which is the simplest approach to facilitate small object detection [30,31]. The size of anchor boxes used to determine the bounding box of objects were left at their standard settings of 32, 64, 128, For classification and object detection, we use an open source RetinaNet [27] implementation in Python, available on GitHub (https://github.com/fizyr/keras-retinanet, last accessed on 6 February 2019). The architecture of RetinaNet includes a backbone convolutional neural network connected to a feature pyramid [28], which detects potential objects in the backscatter mosaics. The detected objects are forwarded to two subnetworks, one for classifying the objects within the bounding box (in this case, the two available classes are boulders and background), and one for the regression of the bounding box coordinates. As a backbone for classification we used ResNet50, with pretrained weights from the Microsoft Coco dataset [29] to speed up training. Initial tests not reported here were carried out using VGG16 as a backbone as well, but ResNet50 consistently provided better results. To update the weights of the convolutional layers, RetinaNet implements a focal loss mechanism [27] for backpropagation of errors during training, which puts an emphasis on the loss contributed by difficult training samples. RetinaNet expects by default an input image side length of 800 pixels, corresponding to 200 m or 800 m depending on the resolution of the image tiles (0.25 m or 1 m). During training, the extent of the input image is reduced to 1/16th of the original image. This deletes most of the cobbles and fine boulder information from the data if large scale mosaics were fed directly into the network. Hence, the smaller tiles exported from the backscatter mosaics were upscaled to values between 300 and 1200 pixels, which is the simplest approach to facilitate small object detection [30,31]. The size of anchor boxes used to determine the bounding box of objects were left at their standard settings of 32, 64, 128, 256, and 512 pixels. An areal overlap of the bounding box with a manually interpreted bounding box of 0.5 was required to count as a positive detection. To counter overfitting, and the constant orientation of acoustic shadows in the North-South direction due to the East-West tow direction of the acoustic sonar, which cannot be assumed in general, a strong random transformation was applied to the training data. We applied a random rotation, a translation by −0.3 to 0.3 times the image size, a shearing of the picked boulders between −114 • and 114 • , random scaling down to 0.7 and up to 1.4 and a chance of 0.5 to mirror the image in the North-South and East-West directions. Training of RetinaNet For training of RetinaNet, we set the batch size to 2 and the steps per training epoch between 50 and 800. The number of epochs was kept constant at 30, since the deep network is rapidly overfitting on the limited training dataset after few epochs. Independent models were calculated for the tiles of 25 m 2 and 225 m 2 , and termed the 25 m 2 and 225 m 2 model. For all image sizes, losses and mean average precision (mAP) [32] are peaking rapidly following 5-15 training epochs. The mAP describes the classification accuracy of the detected boulders, the overlap of the detected bounding boxes with the training bounding boxes, and the percentage of objects identified in every tile. A higher mAP corresponds to a better identification. The data were evaluated on the validation dataset ( Figure 3). The following description relates to training on mosaics with 0.25 m resolution. It can be observed that the mAP on the validation dataset is controlled by the size of the input image. The mAP for tiles of 25 m 2 exceed 0.6. It drops to slightly below 0.4 for image sizes of 225 m 2 . This decrease is only partially offset by increasing the upscaled image extension presented to the neural network. Here, a higher upscaling causes a change in mAP of less than 0.1. As larger images consume increasingly more memory and computing capacity, it was not possible to test the relationship routinely upwards of 800 pixels, and limited example runs indicate no mAP increase for upscaling factors exceeding~15×. The number of steps per epoch has an impact on final mAP of 0.1 to 0.15 chiefly for small image tiles of 25 m 2 , where it was found that a mean number of steps of~500 delivered the best results. For larger images, the impact of steps per epoch decreases, and is offset by a larger number of training epochs before overfitting occurs. For mosaics downscaled to 1 m resolution, mAP on the validation dataset were generally less than 0.1. 256, and 512 pixels. An areal overlap of the bounding box with a manually interpreted bounding box of 0.5 was required to count as a positive detection. To counter overfitting, and the constant orientation of acoustic shadows in the North-South direction due to the East-West tow direction of the acoustic sonar, which cannot be assumed in general, a strong random transformation was applied to the training data. We applied a random rotation, a translation by −0.3 to 0.3 times the image size, a shearing of the picked boulders between −114° and 114°, random scaling down to 0.7 and up to 1.4 and a chance of 0.5 to mirror the image in the North-South and East-West directions. Training of RetinaNet For training of RetinaNet, we set the batch size to 2 and the steps per training epoch between 50 and 800. The number of epochs was kept constant at 30, since the deep network is rapidly overfitting on the limited training dataset after few epochs. Independent models were calculated for the tiles of 25 m² and 225 m², and termed the 25 m² and 225 m² model. For all image sizes, losses and mean average precision (mAP) [32] are peaking rapidly following 5-15 training epochs. The mAP describes the classification accuracy of the detected boulders, the overlap of the detected bounding boxes with the training bounding boxes, and the percentage of objects identified in every tile. A higher mAP corresponds to a better identification. The data were evaluated on the validation dataset ( Figure 3). The following description relates to training on mosaics with 0.25 m resolution. It can be observed that the mAP on the validation dataset is controlled by the size of the input image. The mAP for tiles of 25 m² exceed 0.6. It drops to slightly below 0.4 for image sizes of 225 m². This decrease is only partially offset by increasing the upscaled image extension presented to the neural network. Here, a higher upscaling causes a change in mAP of less than 0.1. As larger images consume increasingly more memory and computing capacity, it was not possible to test the relationship routinely upwards of 800 pixels, and limited example runs indicate no mAP increase for upscaling factors exceeding ~15×. The number of steps per epoch has an impact on final mAP of 0.1 to 0.15 chiefly for small image tiles of 25 m², where it was found that a mean number of steps of ~500 delivered the best results. For larger images, the impact of steps per epoch decreases, and is offset by a larger number of training epochs before overfitting occurs. For mosaics downscaled to 1 m resolution, mAP on the validation dataset were generally less than 0.1. Based on these observations, a model trained with 225 m 2 mosaic tiles, upscaled to 800 pixels, and a model trained with 25 m 2 image tiles, upscaled to 300 pixels, were used for further testing on 0.25 m resolution mosaics. The upscaling causes a magnification of boulders by a factor of 13.3 to 15. A model with 225 m 2 tiles, upscaled to 300 pixels, was used for classification of mosaics with 1 m resolution, corresponding to an upscaling factor of 20×. The threshold score (probability) for including an object as a boulder was set to 0.5 for all models trained on 0.25 m mosaics and 0.25 for models trained on 1 m mosaics. Detections with lower scores were omitted. Determination of Boulder Densities Boulder densities over 100 × 100 m boxes in the test area ( Figure 1) were grouped into three classes, following recent agreements of national German expert groups for the initial assessment of boulder density in large-scale surveys in the German Baltic Sea and North Sea. Classes comprise 0 boulders, 1-5 boulders, and 6 and more boulders. The human interpretation was taken as a baseline. To estimate model performance in comparison to the human interpretation, a confusion matrix was calculated, and the accuracy (ACC), precision (P), and recall (R) were determined by utilizing Pythons sklearn.metrics package (version 0.20.3) [26]. Accuracy describes the percentage of boxes where the interpretation of human interpreter and trained model agree. Precision describes the tendency of the model to not produce false positives, where the best value is 1 and the worst 0. Recall describes the ability of the model to find the same number of boxes for each boulder density class as the human interpreter, where the best value is 1 and the worst 0. Description of Backscatter Mosaics Two natural features-glacial till covered by lag deposits and sand areas-are distinguished in the backscatter mosaics and occur both in the training and testing area (Figure 1), corresponding to published maps of seafloor composition [33]. Displayed in darker colors that represent higher backscatter intensities, extended areas of glacial till covered with lag sediment are visible. The majority of surficial boulders in the investigation area are located within lag deposits. The boulders are characterized by increased backscatter on the side that faces the sonar, and decreased backscatter (acoustic shadow) on the other side. Acoustic shadows are either facing towards north or south in all our backscatter mosaics due to east-west ship routes. Grayish colors representing intermediate backscatter intensities, typically corresponding to sandy sediments, are located adjacent to the glacial till covered with lag sediments. Pockets of sandy material are situated within areas covered by lag sediment. While visually decreasing in abundance, individual boulders occur on top of the sandy seafloor, indicating the presence of lag deposits or glacial till closely beneath the surface. In total, 4617 boulders have been recognized on the training mosaics, while a manual counting at a scale of 1:250 yielded 3192 boulders in the test area ( Figure 4). The size distribution of the visually interpreted objects in the training mosaics ( Figure 5) shows an increasing probability of occurrence for objects with bounding boxes approx. 4.5 m 2 in area, with rapidly decreasing probability of occurrence for larger and smaller objects. Crossing both glacial till and sand, elongated plough marks created as a side effect of offshore windpark construction work are present in all backscatter mosaics. The boundaries of the plough marks are typically marked by one or two lines of elongated high backscatter intensity ( Figure 6). Within the cable plough marks, boulders or boulder-like features occur frequently, as well as areas of low backscatter intensity between the elongated high backscatter lines. The sedimentological composition of the areas of low backscatter intensities is unknown due to the lack of dedicated ground truthing. East-west directed areas of low backscatter contrast present in all mosaics correspond to the nadir lines of the used sidescan-sonar (Figure 6), where no usable backscatter information is present. In addition, the impact of internal water column stratification is frequently observed. It is caused by small irregularities of the stratification scattering the acoustic signal and is recognized by a curved, chaotic change of backscatter intensities especially in the outer parts of the sonar swath ( Figure 6). Application of RetinaNet to the Test Backscatter Mosaic The models trained on 25 m 2 and 225 m 2 m image subsets were applied to the test mosaic. The following description relates to the application on the 0.25 m resolution mosaic. For the 25 m 2 model, the total number of detected boulders in the test area is 5330 (Figure 4). This model detects a higher number of boulder candidates within the sandy seafloor area characterized by intermediate backscatter intensities. In comparison to the visual interpretation, a higher number of objects are detected near the boundaries of the high backscatter plough marks. Some misclassification occurs due to artefacts caused by scatter on internal water column stratification ( Figure 6). Within the glacial till boundaries, the 25 m 2 model detects a high number of cobbles or small boulders, only a few pixels in size and partly showing only a minimal acoustic shadow 1 pixel in width. However, a visual comparison with the manually interpreted mosaics indicates that the human expert is able to interpret smaller features (Figure 4). A histogram showing the bounding box size of the detected objects ( Figure 5) confirms this observation, with the percentage of objects with a bounding box size smaller than 5 m 2 considerably higher compared to the 225 m 2 model, but still below manual interpretation. Boulders with extensive shadows exceeding 1 m in size can be detected several times due to the overlapping image tiles. For the 25 m 2 model, it was tested whether model and human agree if a boulder is present in a tile for 6933 randomly selected samples chosen from the test area. The results are displayed as a binary Receiver Operating Characteristic (ROC) curve available in the supplemental material ( Figure S1). The area under the ROC (AUC) is 0.97. The number of boulders detected by the 225 m 2 model is 666 (Figure 4). Few classifications near the boundaries of the cable plough marks occur. The model is not sensitive to artifacts from water column stratification. The number of boulders detected in sandy areas and inside the lag deposits is considerably lower compared to the 25 m 2 model (Figure 4). The 225 m 2 model captures individual boulders but is not detecting fine boulders or boulders that show only a minimal acoustic shadow. This is reflected in a histogram of the bounding box area ( Figure 5), which demonstrates a decrease in detected boulders of small size compared to the 25 m 2 model. Application of RetinaNet to the Test Backscatter Mosaic The models trained on 25 m² and 225 m² m image subsets were applied to the test mosaic. The following description relates to the application on the 0.25 m resolution mosaic. For the 25 m² model, the total number of detected boulders in the test area is 5330 (Figure 4). This model detects a higher number of boulder candidates within the sandy seafloor area characterized by intermediate backscatter intensities. In comparison to the visual interpretation, a higher number of objects are detected near the boundaries of the high backscatter plough marks. Some misclassification occurs due to artefacts caused by scatter on internal water column stratification ( Figure 6). Within the glacial till boundaries, the 25 m² model detects a high number of cobbles or small boulders, only a few pixels in size and partly showing only a minimal acoustic shadow 1 pixel in width. However, a visual comparison with the manually interpreted mosaics indicates that the human expert is able to interpret smaller features ( Figure 4). A histogram showing the bounding box size of the detected objects ( Figure 5) confirms this observation, with the percentage of objects with a bounding box size smaller than 5 m² considerably higher compared to the 225 m² model, but still below manual interpretation. Boulders with extensive shadows exceeding 1 m in size can be detected several times due to the overlapping image tiles. For the 25 m² model, it was tested whether model and human agree if a boulder is present in a tile for 6933 randomly selected samples chosen from the test area. The results are displayed as a binary Receiver Operating Characteristic (ROC) curve available in the supplemental material ( Figure S1). The area under the ROC (AUC) is 0.97. The number of boulders detected by the 225 m² model is 666 (Figure 4). Few classifications near the boundaries of the cable plough marks occur. The model is not sensitive to artifacts from water column stratification. The number of boulders detected in sandy areas and inside the lag deposits is considerably lower compared to the 25 m² model (Figure 4). The 225 m² model captures individual boulders but is not detecting fine boulders or boulders that show only a minimal acoustic shadow. This is reflected in a histogram of the bounding box area ( Figure 5), which demonstrates a decrease in detected boulders of small size compared to the 25 m² model. For the mosaic of 1 m resolution, the 225 m² model found a total of 572 boulders (Figure 4). A direct comparison with the interpretation of 0.25 m resolution mosaics shows that both smaller and larger boulders are not detected when using a coarser resolution, and several small boulders may be interpreted as a single large object in lower resolution. The first mode of the bounding box size distribution is located at 10 m² ( Figure 5). Boulder Densities For assessing the applicability of the network in practical work, we compared the classification of boulder density per 10,000 m² boxes in three groups (no boulder, 1-5 boulders, > 5 boulders). The visual interpretation of boulder densities into three classes (Figure 7) confirms high boulder densities on areas covered by lag sediment, while low boulder abundances or the absence of boulders characterize the sand seafloor facies. Accuracy for the 25 m² model is 0.9 over all classes. The model detects higher boulder densities in low density areas compared to expert classification, with 27% of the boxes that were visually empty counted to contain at least one object by the model. Therefore, the number of empty boxes is underestimated, resulting in a reduced recall for class 0, and reduced precision for class 1-5 (Table 1). Almost no misclassification occurs for boxes visually interpreted to include 1-5 or more than 5 boulders, with recall values of 0.88 and 0.98, respectively. In contrast, the 225 m² model tends to underestimate boulder density compared to the expert interpretation, with most of the boxes with 1-5 boulders classified as empty, and 28% of the boxes with > 5 boulders classified as medium density (Figure 8). This outcome is reflected in low precision and recall rates for the 1-5 class, and reduced recall for the > 5 class. The automatic classification of boulder densities based on mosaics with 1 m resolution shows an increase of empty and intermediate boxes compared both to the human interpretation and the higher resolution models, with overall accuracy dropping to 0.54, and precision and recall being less than 0.7 for class 0, and less than 0.5 for class 1-5. For the mosaic of 1 m resolution, the 225 m 2 model found a total of 572 boulders (Figure 4). A direct comparison with the interpretation of 0.25 m resolution mosaics shows that both smaller and larger boulders are not detected when using a coarser resolution, and several small boulders may be interpreted as a single large object in lower resolution. The first mode of the bounding box size distribution is located at 10 m 2 ( Figure 5). Boulder Densities For assessing the applicability of the network in practical work, we compared the classification of boulder density per 10,000 m 2 boxes in three groups (no boulder, 1-5 boulders, >5 boulders). The visual interpretation of boulder densities into three classes (Figure 7) confirms high boulder densities on areas covered by lag sediment, while low boulder abundances or the absence of boulders characterize the sand seafloor facies. Accuracy for the 25 m 2 model is 0.9 over all classes. The model detects higher boulder densities in low density areas compared to expert classification, with 27% of the boxes that were visually empty counted to contain at least one object by the model. Therefore, the number of empty boxes is underestimated, resulting in a reduced recall for class 0, and reduced precision for class 1-5 (Table 1). Almost no misclassification occurs for boxes visually interpreted to include 1-5 or more than 5 boulders, with recall values of 0.88 and 0.98, respectively. In contrast, the 225 m 2 model tends to underestimate boulder density compared to the expert interpretation, with most of the boxes with 1-5 boulders classified as empty, and 28% of the boxes with >5 boulders classified as medium density (Figure 8). This outcome is reflected in low precision and recall rates for the 1-5 class, and reduced recall for the >5 class. The automatic classification of boulder densities based on mosaics with 1 m resolution shows an increase of empty and intermediate boxes compared both to the human interpretation and the higher resolution models, with overall accuracy dropping to 0.54, and precision and recall being less than 0.7 for class 0, and less than 0.5 for class 1-5. Table 1. Constraining the Minimum Size of Detected Boulders In marine conservation, all hard substrata with a diameter >64 mm can potentially form a geogenic reef (DG ENV). Therefore, it is mandatory to detect objects of the smallest possible size and to consider the minimum object size detectable by the trained models. The minimum size of objects Table 1. Table 1. Accuracy (ACC), precision (P), and recall (R) of the boulder density confusion matrix. Constraining the Minimum Size of Detected Boulders In marine conservation, all hard substrata with a diameter >64 mm can potentially form a geogenic reef (DG ENV). Therefore, it is mandatory to detect objects of the smallest possible size and to consider the minimum object size detectable by the trained models. The minimum size of objects Table 1. Constraining the Minimum Size of Detected Boulders In marine conservation, all hard substrata with a diameter >64 mm can potentially form a geogenic reef (DG ENV). Therefore, it is mandatory to detect objects of the smallest possible size and to consider the minimum object size detectable by the trained models. The minimum size of objects whose detection can be trained by RetinaNet depends on a) the resolution of the input backscatter mosaic and b) the minimum anchor box of the network measured in pixels multiplied by the threshold of areal overlap of 0.5 required for a positive training [30,31]. For a minimum anchor box of 32 pixels, this results in a theoretical minimum threshold for positive training of 23 × 23 pixels. With the used strong upscaling factor of 13×-15×, this would in principle allow objects of 1 × 3 pixel extension to be trained upon, corresponding to a minimum bounding box area of approx. 0.2 m 2 at a backscatter mosaic resolution of 0.25 m, and 3 m 2 at a backscatter mosaic resolution of 1 m. However, the detection of objects three pixels in size prior to upscaling would likely result in substantial misclassification due to noise and artefacts inherent in the sidescan-sonar data or due to upscaling artefacts, since no real internal features could be learned by a neural network. In addition, the creation of training datasets at these scales is highly challenging, since the differentiation between real features and noise or artifacts becomes difficult for the human expert interpreter as well, and the pixel-accurate digitization of bounding boxes is unfeasible. Therefore, almost no objects of this size are included in the training dataset ( Figure 5). For objects with a 4 × 4 pixel extension, equal to 1 m 2 at 0.25 m mosaic resolution, a detection appears possible, since they result in 52 × 52 to 60 × 60 pixels following upscaling and a higher number of corresponding objects is included in the training dataset. However, the bounding box threshold criterion of 0.5 may still be missed for these size classes in case of different alignment of bounding boxes. In addition, it needs to be considered that manually digitized bounding boxes are necessarily inaccurate at the pixel-level and may not be reproduced by the model, and errors on bounding boxes effect smaller objects more than larger objects [30]. For 7 × 7 pixel bounding boxes (3.1 m 2 at 0.25 m resolution), a detection appears feasible following the upscaling, as can be observed by the increased detection frequency by the 25 m 2 model rivaling the human interpreter ( Figure 5). Naturally, the size of the boulders is not equal to the size of the bounding box, which includes the shadow and some surrounding background. Sidescan-sonar backscatter mosaics do not preserve information on acoustic incidence angles and towfish height that would be needed for an accurate geometrical reconstruction of object size [34], but rough approximations may be made nonetheless. To constrain object size, the area of several boulders facing towards the sidescan-sonar and an eventual transition zone representing the boulder top were measured and divided by the total area of the bounding box (examples shown in Figure 9). Albeit highly variable, the actual boulder encompasses approximately 20% of a bounding box area on average for mosaics of 0.25 m resolution. For bounding boxes of 7 × 7 pixels (3.1 m 2 ) that are picked up with high frequency by the model and the human interpreter, this would equal a boulder-covered area of 3 × 3 pixels (approx. 0.6 m 2 ) at 0.25 m mosaic resolution. In the following, it is assumed that most boulders have a round shape. If the assumption of round shape is violated, e.g., for artificial objects, their detection by the model depends on the orientation of the object to the sidescan-sonar. Ideally, this orientation would be made available as additional data during model training. For round boulders, the diameter of a boulder fitting within a 3 × 3 pixel square will be between 0.5 m at minimum and 0.75 m at maximum. Based on the resolution of the input data, the composition of the training dataset and training parameters chosen in this study, boulders of smaller diameter would likely not be picked up reliably by either the trained model or a human interpreter. For mosaics of 1 m resolution, the minimum size of detected objects decreases correspondingly. For these mosaics, the ratio between bounding box area and estimated stone surface is closer to 0.3 (one example shown in Figure 9). With the most frequent bounding box at 1 m mosaic resolution located at 10 m 2 , corresponding to approx. 3 × 3 pixels ( Figure 5), boulders below diameters of 2 m to 3 m will not be detected reliably. With the spatial resolution available in large-scale practical applications today, the models trained by neural networks may be useful for the secure identification of medium and large boulders as well as blocks, but not fine boulders or cobbles. A reliable direct classification of cobbles and fine boulders will need higher resolution input data, which can be difficult for large areas of interest. If covering a large percentage of seafloor, the presence of cobbles and fine boulders may be detected indirectly, e.g., by utilizing angular response curves [35,36] and multifrequency backscatter approaches currently under active development [37][38][39][40]. Eventually, the density of medium and large boulders may be used as a proxy for the overall density of hard substrata available for biota and consequently for the identification of reefs sensu habitats directive (code 1170). However, future work on the statistical distribution of cobbles and boulders in lag deposits of different geological origin is recommended. Model Performance on the Validation Dataset The accuracy of the different models on the validation dataset were highly variable. In contrast to the models trained on 0.25 m resolution mosaics, the mAP of models validated on the 1.0 m backscatter mosaic is significantly worse (Figure 3). This is caused by the use of a training dataset digitized on the 0.25 m resolution mosaics for training. On the 1 m mosaic, a high number of boulders cannot be differentiated due to the decreased resolution, and are not found during model validation. It was attempted to create a training dataset on 1 m mosaics, but due to the coarse resolution (examples in Figures 4 and 9), it is not possible for a human interpreter to reliably distinguish potential boulders from noise and create accurate bounding boxes. As an effect, the score assigned by the 1 m trained model to detections in the test area decreased, which was offset by cutting the score threshold required for a positive detection from 0.5 to 0.25. For the 0.25 m resolution mosaics, the main difference in performance between the 25 m² and 225 m² model is the higher sensitivity of the 25 m² model, picking more boulders on the lag deposits and the sandy seafloor, but also being more sensitive to boundaries of the cable plough marks. It is unlikely that the different upscaling of boulders in the original mosaic is responsible for the sensitivity increase, as it is very similar for both models (15× for the 25 m² model and 13.3× for the 225 m² model), and both should be able to detect boulders of 7 × 7 pixels in extension as discussed With the spatial resolution available in large-scale practical applications today, the models trained by neural networks may be useful for the secure identification of medium and large boulders as well as blocks, but not fine boulders or cobbles. A reliable direct classification of cobbles and fine boulders will need higher resolution input data, which can be difficult for large areas of interest. If covering a large percentage of seafloor, the presence of cobbles and fine boulders may be detected indirectly, e.g., by utilizing angular response curves [35,36] and multifrequency backscatter approaches currently under active development [37][38][39][40]. Eventually, the density of medium and large boulders may be used as a proxy for the overall density of hard substrata available for biota and consequently for the identification of reefs sensu habitats directive (code 1170). However, future work on the statistical distribution of cobbles and boulders in lag deposits of different geological origin is recommended. Model Performance on the Validation Dataset The accuracy of the different models on the validation dataset were highly variable. In contrast to the models trained on 0.25 m resolution mosaics, the mAP of models validated on the 1.0 m backscatter mosaic is significantly worse (Figure 3). This is caused by the use of a training dataset digitized on the 0.25 m resolution mosaics for training. On the 1 m mosaic, a high number of boulders cannot be differentiated due to the decreased resolution, and are not found during model validation. It was attempted to create a training dataset on 1 m mosaics, but due to the coarse resolution (examples in Figures 4 and 9), it is not possible for a human interpreter to reliably distinguish potential boulders from noise and create accurate bounding boxes. As an effect, the score assigned by the 1 m trained model to detections in the test area decreased, which was offset by cutting the score threshold required for a positive detection from 0.5 to 0.25. For the 0.25 m resolution mosaics, the main difference in performance between the 25 m 2 and 225 m 2 model is the higher sensitivity of the 25 m 2 model, picking more boulders on the lag deposits and the sandy seafloor, but also being more sensitive to boundaries of the cable plough marks. It is unlikely that the different upscaling of boulders in the original mosaic is responsible for the sensitivity increase, as it is very similar for both models (15× for the 25 m 2 model and 13.3× for the 225 m 2 model), and both should be able to detect boulders of 7 × 7 pixels in extension as discussed above. It may rather be assumed that the reason is the different amount of background, which is implicitly trained in ResNet. While to our knowledge no boulder size distributions of the different Baltic Sea lag deposits area exist, it may be a reasonable assumption that the number of small boulders is increasing nonlinearly: The source of objects exposed on Kriegers Flak are glacial till deposits. A correlation of grain size distribution on the type of till, original bedrock, and topography exists [41], but lodgement tills widespread in the Baltic show grain size maxima in the silt-clay and gravel domain [42]. Given that significant reemergence of cobbles and boulders from glacial till takes place over decades [43], thus quickly regenerating following submergence of the former coastlines, effects of Holocene water level fluctuations [44] in the Baltic Sea do not need to be considered. The grain size of the armor layer that forms on top of glacial is typically located in the gravel fraction in the Baltic Sea [45], not resolved by the acoustic backscatter mosaics and below the 3 m 2 to 5 m 2 bounding box related to boulders of~0.7 m diameter. Hence, the decrease in objects with bounding boxes smaller than approx. 3 m 2 is presumably caused by increasing difficulty of interpreting small objects from the background, ultimately limited by mosaic resolution, and does not represent the true geological conditions. Therefore, a lot of small boulders or large cobbles are overlooked by the human interpreter (Figure 2), forming false negatives during training. These objects are trained as background during the training runs especially for the 225 m 2 model, reducing its sensitivity to small objects. For further optimization of the neural network, efforts to improve the training dataset with more extensive ground truthing including small objects are required. Sources of Error and Comparison with Expert Interpretation A main criteria for the practical applicability of automated objects detections is to reduce the number of false positive detections. An especially difficult source of error during the detection of objects of small size-both for humans and the trained models-is caused by stratification of the water column, which scatters acoustic waves at oblique incidence angles and is very prominent in sidescan-sonar backscatter data especially in the Baltic Sea (an example shown in Figure 6). A second source of misclassification are the boundaries of the plough marks. The higher backscatter intensities at their boundary facing towards the sidescan-sonar due the morphological depression formed by the central trough, are in some cases wrongly picked up especially by the 25 m 2 model ( Figure 6). However, due to the glacial till situated closely beneath the sand surface, boulders may have been exposed within the plough marks. Therefore, a judgement whether an object was picked correctly or erroneously by the model was not always possible. It is expected that the use of snippet-derived multibeam echo sounder backscatter data [46], would mitigate both sources of misclassification. Since snippet derived backscatter centers the backscatter time series on the detected seafloor position, effects of noise situated in the water column are minimized. In addition, the availability of a local bathymetric data would help to distinguish the boundaries of plough marks from discrete objects. Still, for a random selection of samples, the AUC of 0.97 between the best model and human interpreter ( Figure S1) indicates a generally very high agreement between model and human. The remaining misclassifications in backscatter mosaics can be mitigated using a raster-based approach (Figure 7) that also allows to distinguish boulder densities into the three classes mentioned above. The reasoning behind this classification was that a human interpreter can quickly recognize 5 items in a 100 × 100 m box at a glance, but is required to count objects exceeding four or five [47]. With an overall average accuracy of 0.9, the performance of the trained 25 m 2 model is very close to the results of a human interpreter. All models achieve a precision of 1.0 on boulder densities class >5 (Table 1), indicating the quality of the found boxes needs no further quality check by the human interpreter. However, the recall (and corresponding overall accuracy) is continuously decreasing with tile size and mosaic resolution, indicating an increasing amount of seafloor with high density coverage of boulders is not recognized. A higher disagreement is observed for boxes interpreted as empty, or including 1-5 boulders, where values for either precision or recall drop below 0.6, potentially requiring additional quality assurance by human interpreters. However, it has to be taken into account that the human interpretation is certainly not always correct, given the inevitable lack of ground truthing, and absolute performance of the models may be better than indicated by these values. In summary, it is suggested that neural networks could take over most of the tedious process of digitizing boulder sized objects from backscatter data into maps of boulder density. Those maps can in turn serve as a baseline for the identification of reefs (code 1170) protected under European legislation, but to obtain complete maps of individual boulders, an improvement of the training datasets to increase the sensitivity of the models will be required. Conclusions Neural networks are capable of detecting boulder sized objects in sidescan-sonar backscatter mosaics. It is found that a reliable detection requires boulders to encompass at least 3 × 3 pixels in the backscatter mosaics of 0.25 m and 1 m resolution, thus dictating the minimum boulder diameter that is reliably detected. To further improve the trained models, an improved training dataset with more positive and negative examples including other environments, such as shallow waters, will be required. In addition, the models should be transferred to include multibeam echo sounders or interferometric sidescan-sonar derived mosaics, because coregistered bathymetric data would be available for detection and effects of water column noise minimized. With these modifications in place, the proposed approach may serve as a tool to identify reefs that are protected under European legislation in offshore areas where remote sensing by drones, airplanes, or satellites is not applicable and may also replace the time consuming manual interpretation in reporting for marine constructions.	2019-04-07T21:00:17.543Z	2019-04-03T00:00:00.000	{ "year": 2019, "sha1": "6046f7fee425efe612490636adf393f6c4bc3b69", "oa_license": "CCBY", "oa_url": "https://www.mdpi.com/2076-3263/9/4/159/pdf", "oa_status": "GOLD", "pdf_src": "MergedPDFExtraction", "pdf_hash": "9e5f5b971796720f9aa34f9fbc81a1be4f6f6fad", "s2fieldsofstudy": [ "Environmental Science" ], "extfieldsofstudy": [ "Geology" ] }
225411207	pes2o/s2orc	v3-fos-license	Enhancing Effects of NaHSO3 on Corrosion of T91 Steel In the paper, corrosion behavior of T91 steel in different concentrations of NaHSO3 solution was studied in combination with scanning electron microscope (SEM) and electrochemical measurements. The results showed that the steel exhibited active anodic dissolution characteristics in the solution, and NaHSO3 concentration affected both cathodic and anodic behaviors. The steel surface was covered by intact corrosion products in the solutions, but the compactness and mechanical properties of the corrosion products degraded with the increase of NaHSO3 concentration. In low-concentration NaHSO3 solution the steel tended to undergo uniform corrosion with slight corrosion pits, but its corrosion mode gradually transited to localized corrosion as the NaHSO3 concentration increased. The mechanical property degradation of the corrosion products caused by sulfur compounds and the pH decrease of the solution are the important factors to accelerating its corrosion process. Introduction In recent years, acid rain has become a global environmental issue. Europe, North America and China are the three largest acid rain regions. The acid rain in China is mainly distributed in south of the Yangtze River. Its main type is sulfuric acid rain, and SO 2 emission from industrial manufacture is its main reason [1]. Acid rain is one of the important reasons of metal corrosion in atmospheric environment. The economic loss caused by acid rain corrosion throughout the world reached 20 billion US dollars in 1999, and the according loss in China also reached 3 billion RMB [2]. Many scholars have conducted a lot of researches on acid rain corrosion of metals [3]. It has become a common belief that acid rain promotes the decomposition of natural stones [4], causing damages to cultural heritage (especially outdoor marble and bronze sculptures) [5]. Acid rain can enhance the corrosion rate of rebars in concrete, leading to a significant decrease of yield strength and elastic modulus [6,7]. Shi et al. [8] found that the cathodic reaction of 2024-T3 aluminum alloy in acid rain environment is mainly the reduction of hydrogen ions, and its corrosion rate increases in low-pH solution. The high activity and migration rate of hydrogen ions in acidic solutions are the main reasons for the accelerated corrosion [8]. AZ91 magnesium alloy experience uniform corrosion in acid rain environments and the corrosion mainly occurs in its α phase, while its β phase shows a passive characteristic [9]. The corrosion process of AZ31 magnesium alloy in acid rain environment is jointly controlled by anodic dissolution and hydrogen ion reduction, and pitting corrosion occurs around the AlMn phase near grain boundary [10]. Varvara et al. [11] indicated that Horse-chestnut ethanolic extract has a good inhibition role in acid rain corrosion of copper, and the inhibition role stems from the stable physical adsorption layer forming on the copper surface. Rotaru et al. [12] confirmed that some antibiotics own the similar inhibition effect of the Horse-chestnut ethanolic extract, and it is related to the parameters of the energy and energy gap of organic molecules. T91 steel has been selected as one of the main heatresistant materials and replacement materials of steam pipe [13], and widely used in power stations. The steel is subjected to atmospheric corrosion during the storage and service before its service, and some scholars have studied its atmospheric corrosion behavior [14][15][16]. Zhang et al. [17] verified that seasonal acid rain and severe atmospheric contamination due to heavy industries induced localized corrosion on internal and external surfaces of T91 steel after stored for only one year. Yu et al. [18] found that H + and Clcan attack the passive film of T92 steel and promote its corrosion process. However, there is a lack of the studies on corrosion behavior of T91 steel in acid rain environment. In this paper, the electrochemical measurements and scanning electron microscope (SEM) were jointly used to study the corrosion behavior of T91 steel in different concentrations of NaHSO 3 solution (to simulating acid rain), and the pH effect on corrosion behavior was discussed. Materials The material used in this work is T91 steel. Its chemical compositions are as follows (wt.%): C 0.08, Mn 0.30, Si 0.20, S < 0.010, P < 0.020, Cr 8.00~9.50, Ni < 0.40, Mo 0.85, Nb 0.06~0.10, V 0.18, N 0.03, Fe balance. The material was processed into a size of 10 mm × 10 mm × 3 mm, and then sealed with epoxy resin to expose an area of 10 mm × 10 mm for electrochemical measurements. According to the national standard GB5776-86, the working surfaces of the specimens were polished into a mirror using a series of waterproof abrasive papers, and cleaned with acetone and alcohol in sequence. After that, the specimens were stored in a desiccator for use. The experimental solution used in this work is a series of NaHSO 3 solutions with different concentrations, with concentrations and corresponding pH values showing in Table 1. It can be seen that the pH of the solutions are between 4.1 and 5.1, which are suitable for simulating the moderate acid rains. Electrochemical measurements In the experiment, electrochemical impedance spectroscopy (EIS) and Tafel polarization were used to study the electrochemical performance of T91 steel in the solutions. All electrochemical measurements were performed in a classic three-electrode system, with a saturated calomel electrode (SCE) as a reference electrode, a platinum electrode as an auxiliary electrode, and a T91 specimen as a working electrode. All measurements were performed on a CS350 electrochemical test system produced by Wuhan Coster Equipment Co., Ltd. The frequency range of EIS measurement was from 100 kHz to 10 mHz, and the excitation signal was a sine wave voltage signal with an amplitude of 10 mV. The EIS results were modeled and fitted by a ZSimpWin software. The scanning range of Tafel polarization curve was ± 250 mV vs. open circuit potential (OCP), and its scan rate was 0.25 mV s -1 . The experimental results were simulated by Origin software package. The EIS and Tafel polarization tests in this experiment were repeated once, and the average value or typical value was written into the paper. The potentials reported in the paper were all relative to SCE. Analysis of corrosion products After immersed for 504 hours, the specimens were taken out from the solutions, washed with absolute ethanol for 10 minutes, dried with cold air, and stored in a vacuum desiccator. The morphologies and compositions of the corrosion products on the steel surfaces were analyzed by a SEM (JSM-6360LV) with an energy disperse spectroscopy (EDS, NECA 250). After that, the surface corrosion products were removed by the rust remover to expose the corrosion morphologies, and the steel surface were observed again by the SEM. The rust remover used is composed of 500 ml hydrochloric acid (37 wt.%) + 500 mL deionized water + 3.6 g hexamethylenetetramine. Fig. 1 shows SEM micrographs of corrosion products on the steel surface after 504 h immersion in the NaHSO 3 solutions, and the corresponding EDS results are listed in Table 2. After immersed in 0.01% NaHSO 3 solution for 504 h, the steel surface has been completely covered with dense corrosion products ( Fig. 1a), and there are obvious cracks in the inner layer of the corrosion products (Fig. 1b), which may stem from the dehydration process during the posttreatment of the specimens. The color of the corrosion products in this layer is deep, containing elements such as O, Fe, Cr and S (Table 2), and its main components are oxides and sulfides of iron and chromium. The outer layer of the corrosion product is composed of some light-color materials, with the similar main components as the inner layer, but its oxidation content is slightly higher than the later. As for in 0.1% NaHSO 3 solution, the corrosion products have completely covered the steel surface and cracked (Fig. 1c). Compared with the corrosion products in 0.01% NaHSO 3 solution, the number and width of the cracks in the corrosion products are increased (Fig. 1c), indicating that the mechanical properties of the corrosion products are poorer than the former. A layer of mantle like materials is distributed on the corrosion products (Fig. 1d), and it is composed of oxides and sulfides of iron and chromium (Table 2). When the NaHSO 3 concentration increased to 1.0%, massive cracks appears in the corrosion products, and some corrosion products fall off from the steel surface, which indicates that the mechanical properties of the corrosion products are the worst in the 1.0% NaHSO 3 solution (Fig. 1e). Intact granular corrosion products exist in the cracks and the fall-off region (Fig. 1f). The composition of granular corrosion products is close to other corrosion products on the surface, with high contents of chromium and sulfur ( Table 2). The above results show that after immersed in NaH-SO 3 solutions for 504 h, complete corrosion products with high contents of chromium and sulfur generate on the steel surface. The compactness and mechanical properties of corrosion products decreases with the concentration increase of NaHSO 3 in the solutions. Fig. 2 shows the corrosion attack morphologies of the steel after removing the corrosion product. In the 0.01% and 0.1% NaHSO 3 solutions, corrosion morphologies of the steel surface are relatively uniform (Figs. 2a and 2c), with many shallow corrosion spots inlaying in it. The corrosion spots look like bird's nests, and their diameters are about 50 µm (Figs. 2b and 2d). In the above solutions, the steel tends to take place uniform corrosion only with slight corrosion spots. SEM analysis of corrosion morphology When NaHSO 3 concentration increased to 1.0%, localized corrosion dominates the steel surface (Fig. 2e). The corrosion spots connect with each other, forming the contiguous localized corrosion spots. Besides, out of the localized corrosion spots, some small pits seem to be forming (Fig. 2f). From the above, it can be concluded that the T91 steel tends to occur uniform corrosion with slight corrosion pits in the low-concentration NaHSO 3 solutions, while the Electrochemical impedance spectroscopy (EIS) Nyquist plots of the steel immersed in the NaH-SO 3 solutions with time are shown in Fig. 3. At the beginning of the experiment, the impedance of the steel firstly increases with the increase of NaHSO 3 concentration, while it decreases after the NaHSO 3 concentration reaching 0.1% (Fig. 3a). What's more, the impedances of the steel in all solutions remain within one order of magnitude. After 72 h immer- sion, the impedance of the steel is still very high in the solutions with NaHSO 3 concentration higher than 0.1%, but it becomes very low in the low-concentration solutions. At 144 h, the impedance increases slightly in the low-concentration solutions, decreases sharply in the 0.1% NaHSO 3 solution, and continues to increase in the high-concentration solutions. At 216 h, the impedance of the steel in low concentration solution continues to rise, while that in the highconcentration solution slightly decreased. At 336 h, the impedance in the low-concentration solution is higher than that in the high-concentration solution. At the end of the experiment, the impedance of the steel shows a negative correlation with the NaHSO 3 concentration in solutions. From the above discussion, it is reasonable to infer that the impedance of the steel is relatively high once the steel is immersed in a NaHSO 3 solution, and it will experience a drop at a certain time during the immersion process. The concentration of NaHSO 3 mainly affects the time of the impedance drop and the subsequent trend of the impedance. The lower the concentration of NaHSO 3 is, the earlier the impedance drop will occur. For example, the impedance drop occurs at 72 nd h in 0.05% and 0.1% NaHSO 3 solutions. In the low-concentration solutions, the impedance increases immediately after the drop, which may be related to the better mechanical properties of the corrosion products forming on the steel surface in the solution. In order to further study the electrochemical behavior of the steel in NaHSO 3 solution, equivalent circuits were used to simulate the EIS data. Twoconstant model is adopted for fitting as shown in Fig. 4. R s is the solution resistance. R f and Q f are the resistance and capacitance of surface corrosion products, respectively. R ct is the charge transfer resistance, and C dl is the double electric layer capacitance. W is the Weber impedance, and the corresponding fitting results are shown in Tables 3-7. In the following, the evolutions of R f and R ct of the steel are used to discuss the experimental results. R f is the resistance of the surface corrosion products, which can represent the integrity of the products and its blocking effect to corrosive ions. R f evolution of the steel with time in the NaHSO 3 solutions is shown in Fig. 5. R f variations can be divided into two types. When the NaHSO 3 concentration is less than or equal to 0.1%, R f of the steel gradually decreases in the early stage of the experiment, and then slightly increases with time. During the whole immersion process, R f always decreases with the increase of NaHSO 3 concentration. However, when the NaHSO 3 Table 3. Fitting results of the EIS data of the steel in 0.01 wt.% NaHSO 3 solution. concentration is 0.5% or 1.0%, R f of the steel increases slightly in the early stage of the experiment, and then it decreases gradually. What's more, R f of the steel in the 1.0% NaHSO 3 solution is higher than that in the 0.5% NaHSO 3 solution in the early stage, while it decreases sharply from 144 th h and is lower 8.971×10 -4 Table 6. Fitting results of the EIS data of the steel in 0.5 wt.% NaHSO 13.27 1.233×10 than the later in the end stage of the experiment. Generally, R ct is inversely related to corrosion rate of metals. Thus, R ct evolution with time is represented as the R ct -1 change with time, as shown in Fig. 6. When NaHSO 3 concentration is lower than 0.1%, R ct -1 of the steel sharply increases, and gradually decreases from the 72 nd or 144 th h to the end of the experiment. However, R ct -1 of the steel firstly decreases and then rapidly increases from 144 th h in the 0.50% and 1.0% NaHSO 3 solution. Obviously, the corrosion behavior of the steel present almost the opposite modes in low-and high-concentration NaH-SO 3 solutions. However, at the end of the experiment, there is a positive correlation between the corrosion rate and the NaHSO 3 concentration. Comparing the evolutions of R f and R ct , they display the same evolution pattern, which indicates that the change of corrosion products in the corrosion system plays an important role in corrosion rate. Tafel polarization curve Tafel curves of the steel immersed in the NaHSO 3 solutions for 504 h are shown in Fig. 7, and the corresponding fitting results are shown in Table 8. The steel shows the characteristics of active anodic dissolution in all solutions, and the cathodic and anodic polarization behaviors are all directly affected by the NaHSO 3 concentration. As shown in Table 8, the corrosion potential E corr of the steel decreases gradually but the corrosion current density i corr increases gradually with the increase of NaHSO 3 concentration, indicating that NaHSO 3 accelerates the corrosion process of the steel during the immersion. The cathodic Tafel slope β c of the steel is in the range of 260~320 mV/ dec in the solutions, which is a limited difference. However, the anodic Tafel slope β a decreases from 390 mV/dec to 110 mV/dec with the increase of NaHSO 3 concentration, and then increases sharply. The polarization resistance R p of the system can be calculated by combining i corr and Tafel slopes: The calculated R p is also shown in Table 8. The polarization resistance of the system decreases and the corrosion rate increases with the increase of NaH-SO 3 concentration at the end stage of the experiment. In order to compare R p and R ct , the evolutions of R p and R ct as a function of NaHSO 3 concentration after 504 h immersion are presented in Fig. 8, and the results are fitted as shown in a black solid line. It can be seen from the figure that the results of R p and R ct are similar, and they are scattered near the model results. The facts indicate that the result of Tafel curve is consistent with that of EIS. Discussion From the result section, the increase of NaHSO 3 concentration in the solutions results into E corr decrease and i corr increase of the T91 steel, with the increase of chromium and sulfur contents in corrosion products. Uniform corrosion dominates the corrosion process of the steel in the low-concentration NaHSO 3 solution, but high-concentration NaHSO 3 tends to accelerate the corrosion process of the steel and increase its localized corrosion sensitivity. The corrosion mechanism of the steel in this solution can be summarized as follows: When the steel is immersed in the low-concentration NaHSO 3 solutions (0.01%, 0.05% and 0.1%), the steel surface may quickly corrode, and the corrosion products forming in the corrosion process will cover on the steel surface quickly, which prevents the direct contact between the corrosion medium and the base metal, thus reducing the corrosion rate. The whole process can be completed within 24 h [19], so the impedance of the steel is high and the corrosion rate is very low at 12h (Figs. 3a and 6). The corrosion products formed in the low-concentration NaHSO 3 solutions are thin, and thus the corrosion medium can gradually diffuse into the interface between the steel and the corrosion product. At the same time, the high-valence iron oxide generated at the early stage may accelerate the steel corrosion through solid redox reaction. Therefore, the impedance of the steel decreases and the corrosion rate increases accordingly at 72~144h (Figs. 3b, 3c and 6). On the other hand, the diffusion of the corrosive medium and the solid-state reaction may be local, or proceed through some certain channels in the corrosion product, inducing many localized corrosion pits on the steel surface (Figs. 2b and 2d). Afterwards, newly-formed corrosion products will fill the channels in the corrosion products, making them denser. The corrosion products will also exist on the steel/product interface, which acts as a physical shielding film for the diffusion of corrosive medium. Then the steel corrosion rate gradually decreases (Fig. 6). When the steel is immersed in the high-concentration NaHSO 3 (0.5% and 1.0%) solutions, corrosion products will also cover on the steel surface and inhibit the corrosion process in the early stage of the experiment. What's more, the corrosion rate and R f of the steel decreases and increases respectively in the solutions in the early stage. The fact shows that the corrosion products may be thicker and the inhibition effects will last longer. However, the NaHSO 3 concentration is relatively higher in the bold solution, thus more corrosive medium may diffuse into the steel/product interface, leading to an extremely rapid corrosion rate (Figs. 3c~3f) and inducing severe localized corrosion (Figs. 2e and 2f). These localized corrosion spots may act as the anode of the corrosion reaction, and the corrosion product layer that has not been penetrated by the corrosive medium may function as the cathode of the corrosion reaction, thereby forming a corrosion couple with large cathode/small anode and accelerating the corrosion process of the steel. Therefore, in the high-concentration NaHSO 3 solutions, the corrosion rate is very high in the later period of the experiment (Fig. 6), and the enhancing effects of NaHSO 3 is obvious. And, the contiguous localized corrosion will be the main corrosion type for the steel (Figs. 2e and 2f), where the anodic reaction may take place on the region around the contiguous localized corrosion. pH of the solutions is distributed in range from 5.1 to 4.1 ( Table 1). The corrosion process of the steel in the solutions is an electrochemical reaction in the acidic solutions. This may be the first reason why the steel shows the characteristics of active anodic dissolution in all solutions (Fig. 7). Besides, pH of the solutions decreases with the increase of NaHSO 3 concentration (Table 1), which may promote the dissolution of the steel substrate and the reduction of the hydrogen ion in the solution. This may be another way for NaHSO 3 concentration to affect the anodic and cathodic polarizations of the steel. What's more, as shown in EDS results (Table 2), the chromium content in corrosion products on the steel surface is in a range from 12 to 20 at.%. The content is much higher than that in the steel matrix (8.00~9.50 wt%). The fact manifests that chromium deposits on the steel surface during the corrosion process, participates in the formation process of corrosion products, and plays a role in protecting the steel. However, high sulfur content also is detected in the corrosion products, and the sulfur content in corrosion products increases with the increase of NaHSO 3 concentration. It has been reported in the literature [20,21] that the presence of S element can reduce the mechanical properties of corrosion products, which is the reason why the micro-cracks in corrosion products increase and the mechanical properties decrease with the increase of NaHSO 3 concentration (Fig. 1). Conclusions The corrosion behavior of a T91 steel in different concentrations of NaHSO 3 solutions was studied, and the following conclusions can be drawn: 1. After immersing in NaHSO 3 solution for 504 h, intact corrosion products can be generated on the steel surface. With the increase of NaHSO 3 concentration, the sulfur content in corrosion products increases, but the compactness and mechanical properties of the corrosion products decrease accordingly. 2. In the low-concentration NaHSO 3 solutions, the steel tends to undergo uniform corrosion with slight corrosion spots. But, as the NaHSO 3 concentration increases, the localized corrosion sensitivity of the steel increases, and the corrosion model gradually changes into localized corrosion. 3. In low-and high-concentration NaHSO 3 solutions, R f of the steel exhibits different evolution patterns, and the corrosion behavior of the steel is almost opposite. 4. The steel in the NaHSO 3 solutions exhibits an obvious characteristic of active anodic dissolution. The cathodic and anodic polarization behaviors are all directly affected by the NaHSO 3 concentration, and the polarization resistance of the system gradually decreases with the increase of NaHSO 3 concentration in solution.	2020-08-06T09:03:47.817Z	2020-08-04T00:00:00.000	{ "year": 2020, "sha1": "98c60c0003be39e40bed1cfdffb43288001e62fb", "oa_license": "CCBYNC", "oa_url": "https://www.jecst.org/upload/pdf/jecst-2020-00997.pdf", "oa_status": "GOLD", "pdf_src": "Anansi", "pdf_hash": "7c323d97e285e6e811c36365c3ad8d9cc8d3533e", "s2fieldsofstudy": [ "Materials Science" ], "extfieldsofstudy": [ "Materials Science" ] }
103983427	pes2o/s2orc	v3-fos-license	Green synthesis of amphiphilic carbon dots from organic solvents: application in fluorescent polymer composites and bio-imaging Carbon dots (CDs) have sparked tremendous attention due to their unique properties and vast potential in diverse fields. Herein, we report a green and cost-effective hydrothermal route for the synthesis of a series of CDs from readily available organics solvents. Since the organics were completely recyclable after the separation of CDs, this method holds immense potential for the large-scale synthesis of CDs. We found the DMF-CDs and DMAc-CDs possessed amphiphilicity and the diameter of amphiphilic DMF-CDs was ca. 3.5 nm with a narrow distribution. Moreover, these amphiphilic CDs emitted blue light under UV irradiation (365 nm) and the quantum yield could reach more than 30%. Due to their good solubility in organic solvent, DMF-CDs were successfully imbedded into polymers (i.e., PS and PMMA), which revealed their potential in painting, coating, and optical devices. In addition, benefiting from high quantum yield and low cytotoxicity, the DMF-CDs in aqueous media were used as fluorescent probes in living cells, which demonstrated their great potential in bio-imaging. Introduction Carbon dots (CDs) or carbon quantum dots, with a typical diameter below 10 nm, have emerged as potent metal-free uorescent nanomaterials in the past decade. Due to their electronic transfer property, 1 bandgap transitions, 2 and surface defects, 3 CDs exhibit an emission dependence on the excitation wavelength or surface passivation. 4 Compared to the conventional semiconductor quantum dots, 5 CDs possess the distinctive advantages of chem high biocompatibility, 6 good watersolubility, low cytotoxicity, 7 functional diversity, 8 convenient synthesis, and low cost. 9 A great deal of research endeavors have been devoted to the synthesis of CDs. A myriad of top-down approaches have been developed for the synthesis of CDs. However, these methods typically require toxic chemical reagents and special equipment (e.g., microwave irradiation, 10 laser ablation, 11 etc.) in the exfoliation process. Instead, bottomup approaches, such as sovlothermal carbonization of glucose, chitosan, glycol, and citric acids, have spurred signicant attention recently. 12 Despite signicant advances have been made in this area, it is of continuous and considerable interest to develop uorescent CDs in an efficient, facile and green manner. On the other hand, CDs have been widely exploited in biological applications such as cell imaging, 13 biosensors, 14 ions testing, 15 and drug delivery, 16 because of their good biocompatibility, robust chemical stability, and high resistance to photobleaching. In these applications CDs were designed to be soluble in aqueous media. Besides aforementioned biological applications by watersoluble CDs, amphiphilic CDs, which receive relative less attention, could open up new applications in hydrophobic environments as well. For example, considering the excellent photoluminescence, low cost and biocompatibility, CDs are excellent alternative of current organic printing dyes in printing industries. However, the current methods for the synthesis of amphiphilic CDs are always associated with low yield and high cost, which severely restrains their further utilization in printing, uorescent composites synthesis and biology related applications. In this context, we report the synthesis of a series of CDs through a facile, one-step hydrothermal treatment of N,N 0dimethylformamide (DMF), N,N 0 -dimethylacetamide (DMAc) and other organic solvents. With this approach, amphiphilic DMF-CDs were achieved with simultaneous surface passivation and tunable uorescence emission. This synthetic route offers multiple advantages. First, it represents a rare example of simple synthesis of amphiphilic CDs from the readily available organic solvents for the rst time, which is an appealing approach from the perspective of energy and cost. Remarkably, the organic solvents were completely recyclable aer the separation of CDs and by-products were not observed, which could achieve high atomic efficiency, one of the core focuses in green chemistry. Last, the surface passivation of DMF-CDs took place simultaneously in the hydrothermal process, rendering intrinsic uorescence emission. Due to the good solubility in polar organic solvents (i.e., DMF), the as-synthesized DMF-CDs could be easily doped into polymer materials and the resultant CDs-polymers composites exhibit high uorescent properties. Furthermore, the DMF-CDs demonstrate excellent biocompatibility and there is no appreciable cytotoxicity observed. Synthesis of the amphiphilic CDs Typically, DMF (20 mL) was added to a polyphenol autoclave (50 mL) and heated at 260 C for 12 hours. Aer solvothermal reaction, the autoclave was cooled to room temperature and the solution was rotary evaporated to obtain CDs. The distillate was collected to recycle the reagents for the successive runs. The other CDs were synthesized by analogous procedures in which DMF was replaced by DMAc, xylene, n-hexane and cyclohexane. Characterization of amphiphilic CDs Transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM) images were recorded on JEOL JEM-1011 and JEOL JSM-2100, respectively. The CDs were drop-cast onto 400-mesh carbon-coated Cu grids followed by drying in open air at room temperature. The uorescence spectra of the CDs were recorded on a FluoroMax-4 spectrouorometer (Horiba Scientic, Japan). X-ray diffraction (XRD) pattern was performed by using a D8 Advance X-ray diffractometer (Bruker) with Cu Ka radiation. Raman spectra were acquired on a Lab RAM Aramis spectrometer (Horiba Scientic, Japan). Fourier transform infrared (FTIR) spectra were measured on Vector-22 spectrometer (Bruker) ranging from 500 to 4000 cm À1 , and the samples were dispersed in KBr pellets. Ultraviolet-visible (UV-vis) absorption of the CDs solution was obtained using an UV-1800(PC) UV-vis spectrophotometer (Mapada). X-ray photoelectron spectroscopy (XPS) measurements were performed on a PHI 5000 VersaProbe spectrometer (UlVAC-PHI, Japan). The uorescence confocal images were obtained using a Leica TCS SP5 confocal scanning microscope (Leica Microsystems, Heidelberg GmbH, Mannheim, Germany). The specimens were excited at 488 nm, and the emission was detected from 500 to 530 nm. Preparation of DMF-CDs-polymer composite 0, 0.04, 0.16 and 0.2 mg of DMF-CDs were separately added into 5 mL of DMF together with 1 g of PMMA or PS. Aer the polymer was dissolved, the solution was transferred into a watch glass and dried at 60 C. Cell imaging of DMF-CDs MCF-7 cells were incubated in the plates with glucose and 1 mL Dulbecco's modied Eagle's medium (DMEM) at 37 C for 24 h. 0.25 mL of DMF-CDs aqueous solution (3 mg mL À1 ) was added into the medium and the nal concentration of DMF-CDs were 0.6 mg mL À1 . The cells were cultured with the CDs at 37 C for 2 h, and washed with the physiological extracellular buffer (ECB: 135 mM NaCl, 5 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , 10 mM HEPES, and 10 mM glucose, pH 7.4) for three times prior to taking the uorescence image. The uorescence intensity of the cells was measured using the uorescence confocal microscope. Cell toxicity of DMF-CDs MCF-7 cells (10 5 cells per mL) in DMEM with glucose were cultured in a 96-well microplate (100 mL per well) for 6 h at 37 C with 5% CO 2 . Then 100 mL of the DMF-CDs solution with different concentrations was introduced into the microplate to acquire the nal concentrations of 1, 10, 100, 1000 mg mL À1 for the culturing of cells for 24 h. The cells were washed with phosphate-buffered saline (PBS, pH 7.4). 20 mL of 5 mg mL À1 MTT and 100 mL of DMEM were added to cell wells for 4 h. Aer removing the culture medium with MTT, 150 mL of DMSO was added. The obtained mixture was shaken for 10 min at room temperature. The optical density (OD) of the mixture was measured at 490 nm in the En-Spire multimode plate reader. The cell viability was calculated according to the given equation: where OD s and OD c refer to OD of samples and control groups, respectively. Quantum yield measurements The quantum yield of amphiphilic CDs was measured by the equation below: where s and r refer to sample and reference. The reference quinine sulfate was dissolved in 0.1 M H 2 SO 4 , of which the quantum yield was 0.54 at 360 nm. 17 I is the integrated emission intensity of uorescent spectra from 380 to 700 nm at the excited wavelength of 360 nm. A represents UV-vis absorbance at 360 nm, which is from 0.1 to 0.01 in the 1 cm quartz absorbance cell. h refers to the refractive index which is listed in Table S1 of the ESI †. Then the uorescent emission spectra of the above samples was tested and compared. Results and discussion Following the procedure depicted in Scheme 1, we prepared a variety of CDs from different organic solvents (Table S2 †), implying the generality of this method. We speculate that the formation of CDs involves three steps, including decomposition, oxidation/carbonization, and surface passivation. Taking DMF-CDs as a representative example, DMF was decomposed into carbon monoxide and dimethylamine in the rst step. 18 Since the autoclave was airtight and the temperature was high, the decomposition products of DMF molecules was oxidized, carbonized and turned into CDs directly. 19 Finally, the external surface of CDs was passivated via covalent bonding of aminecontaining agents. 20 The effect of hydrothermal temperature on the formation of CDs was explored. As shown in Fig. S1 † (ESI), the lower temperature led to the decrease of uorescence, while no uorescence was observed when the reaction temperature was 120 C. This result indicated that the temperature played a key role in the formation of CDs. We found that the synthesized DMF-CDs were small and monodisperse, as revealed by transmission electron microscopy (TEM) image in Fig. 1a. No lattice structure was observed in the HRTEM image (Fig. 1b), indicating amorphous dots were formed. Fig. 1c showed the average diameter of the CDs is 3.5 AE 1.0 nm by measuring 100 particles. The XRD pattern of DMF-CDs (Fig. S2, ESI †) showed a broad diffraction peak at 27.3 , which revealed the amorphous crystal structure of DMF-CDs and conformed to the result of HRTEM. 21 Raman spectra exhibited weak D band at around 1314 cm À1 and strong G band at 1600 cm À1 (Fig. S3, ESI †), which were related to sp 3 carbon and the stretching vibration of C]C bonds, respectively. 22 This result indicated the hybrid form of the carbon atoms in DMF-CDs was mainly disordered sp 2 . 23 The uorescence spectra of the solution were measured to investigate the optical property of DMF-CDs. Under ultraviolet (UV) illumination at 365 nm, DMF-CDs showed an intense blue color (Fig. S4b, ESI †), indicating DMF-CDs had a good uorescent property. The uorescence spectra of the solution displayed wavelength-dependent excitation and broad emissions (Fig. S4c, ESI †). 4 With the excitation wavelength increasing from 320 to 440 nm, the emission wavelength increased from 420 to 500 nm. Such excitation-dependent photo luminescent behavior was consistent with previously reported CDs, which was probably due to the defect sites from the surface passivation. 4a The quantum yields of the CDs were summarized in Table S2. † The DMF-CDs and DMAc-CDs exhibited the high yields of 33.9% and 31.8%, which were superior to CDs derived from organic solvents in the previous reports (Table S3, ESI †). On the contrary, the quantum yields of CDs prepared from other nitrogen-free solvents (i.e., xylene, n-hexane and cyclohexane) were much lower (<13% , Table S2, ESI †). Given the fact that quantum yield of CDs are readily inuenced by the composition, 24 we infer that the intrinsic nitrogen in DMF and DMAc could account for the high quantum yields of DMF-CDs and DMAc-CDs. 25 In an effort to explore the chemical structure and functional groups on DMF-CDs, UV-vis and FT-IR analyses were performed. The UV-vis absorption spectrum of DMF-CDs had a strong peak centered at 320 nm (Fig. 2a), typically corresponding to the n-p* transition of the C]O bond. 26 The FTIR spectrum of DMF-CDs (Fig. 2b) shows an intense peak at 1600 cm À1 , indicating the stretch vibration of carbonyl groups. Additionally, the peaks of other functional groups can be observed in DMF-CDs spectrum, evidenced by their characteristic vibrations centered at 3420 cm À1 (O-H stretching), 2920 cm À1 (sp 3 C-H stretch), 1380 cm À1 (sp 3 C-H bending) and 1270 cm À1 (C-O stretching). To gain a further insight of the surface composition of CDs, XPS analyses were conducted and the DMF-CDs displayed much lower nitrogen content than DMF in the overall XPS spectra (Fig. 2c). The atomic ratio of N/C in DMF is 1/3 and it became very low in DMF-CDs. The XPS of C1s spectrum (Fig. 2d) was deconvoluted into three bands with the binding energy of 284.6, 286, 286.6 eV, corresponding to C]C/ C-C, C-N, and C]O, respectively. The N1s spectrum (Fig. S5a, ESI †) exhibited two peaks at 400.2 and 398.6 eV, which were assigned to N-H and C-N-C groups, respectively. For the O1s Scheme 1 Schematic illustration of the preparation of DMF-CDs and the cyclic process of DMF and DMF-CDs. In order to investigate the chemical stability and the effect of metal ions on the emission of the DMF-CDs, we measured their uorescence under different pH or in the presence of Fe 3+ /Cu 2+ ions. As shown in Fig. 3, there is no signicant loss (less than 5%) in the uorescence intensity of DMF-CDs. The result showed that the DMF-CDs possessed high ion and pH stability and could emit stable uorescence in different solutions. Since the DMF-CDs were synthesized in/from DMF, the DMF-CDs are readily soluble in DMF. Considering DMF is a good solvent for dissolving polymers, we resort to dope CDs into polymers, such as polymethylmethacrylate (PMMA) and polystyrene (PS), with the aim to explore the potential of using amphiphilic CDs as uorescent inks. As shown in Fig. 4, PS and PMMA lms emit blue light aer loading CDs. Increasing the concentration of DMF-CDs lead to enhanced intensity of emitting light from the polymers. No uorescence light was observed from the bare polymers, indicating the DMF-CDs are the only uorescence source. Impressively, these CDs-polymer composites are highly stable and emitted blue uorescent light without obvious decay even aer one year. The addition of hydrochloric acid, sodium hydroxide, ferric nitrate and copper sulfate solution, had negligible effect on the intensity of uorescence. On the basis of the above results, the DMF-CDspolymer composites were quite stable in different solutions. All these results uncover the important potential of DMF-CDs as promising ink for printing. By virtue of their strong uorescence and tunable excitation and emission wavelength, CDs emerged as promising candidates in optical bio-imaging. We sought to evaluate the cytotoxicity of the DMF-CDs, which have the highest quantum yield in the present study. The DMF-CDs were dispersed into water with ultrasonic treatment for 10 min to obtain the aqueous CD solution. Notably, the photoluminescence emission spectra and particle size of DMF-CDs did not show obvious changes in H 2 O compared to that in DMF ( Fig. S6 and S7, ESI †). DMF-CDs were introduced into the cells and the cells emitted green light upon 490 nm excitation (Fig. 5). The observation of uorescence at cells conrmed that the DMF-CDs were water-soluble, and could be uptaken into the cells. The cytotoxicity of the carbon dots was tested (Fig. S8, ESI †) and the DMF-CDs rendered extremely low toxicity below 100 mg mL À1 . In addition, there was no obvious morphological cell change observed aer incubation with the CDs, further demonstrating their prominent biocompatibility. The abovementioned results unambiguously imply that the DMF-CDs, featuring low toxicity and high quantum yield, hold great potential in biological applications, such as bio-imaging and protein analysis by deposit. Conclusions In conclusion, we have successfully prepared a series of CDs via a green one-step hydrothermal synthesis route without any catalyst, additives or by-products in a recycle manner. Among the CDs series, DMF-and DMAc-CDs possess amphiphilic feature and enhanced quantum yields, presumably due to the presence of nitrogen. The prepared DMF-CDs have small size with a narrow size distribution. Moreover, the oxygen/nitrogen doped structure derived from simultaneous surface passivation endows them with strong and excitation-dependent uorescence. Due to the amphiphilic feature, remarkable luminescence stability, and good biocompatibility, DMF-CDs could be readily doped into the polymers and the robust CDspolymers composites exhibited high uorescence. Additionally, the DMF-CDs were appropriately uptaken by cells with low cytotoxicity, thus demonstrating their potential for uorescent cellular imaging. This green method opens up fascinating opportunities for the large-scale synthesis of CDs with the immense potential in uorescent composites synthesis and biological imaging. Conflicts of interest There are no conicts of interest to declare.	2019-04-09T13:09:07.821Z	2018-04-03T00:00:00.000	{ "year": 2018, "sha1": "778447115079db780ce1724d92b0050f1c6de30c", "oa_license": "CCBYNC", "oa_url": "https://pubs.rsc.org/en/content/articlepdf/2018/ra/c8ra01085d", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "82769a772a33f25a869176a78b746aa49310b628", "s2fieldsofstudy": [ "Chemistry", "Materials Science" ], "extfieldsofstudy": [ "Medicine", "Materials Science" ] }
23400633	pes2o/s2orc	v3-fos-license	Opportunities for topical antimicrobial therapy: permeation of canine skin by fusidic acid Background Staphylococcal infection of the canine epidermis and hair follicle is amongst the commonest reasons for antimicrobial prescribing in small animal veterinary practice. Topical therapy with fusidic acid (FA) is an attractive alternative to systemic therapy based on low minimum inhibitory concentrations (MICs, commonly <0.03 mg/l) documented in canine pathogenic staphylococci, including strains of MRSA and MRSP (methicillin-resistant Staphylococcus aureus and S. pseudintermedius). However, permeation of canine skin by FA has not been evaluated in detail. This study aimed to define the degree and extent of FA permeation in canine skin in vitro from two sites with different hair follicle density following application of a licensed ophthalmic formulation that shares the same vehicle as an FA-betamethasone combination product approved for dermal application in dogs. Topical FA application was modelled using skin held in Franz-type diffusion cells. Concentrations of FA in surface swabs, receptor fluid, and transverse skin sections of defined anatomical depth were determined using high-performance liquid chromatography and ultraviolet (HPLC-UV) analysis. Results The majority of FA was recovered by surface swabs after 24 h, as expected (mean ± SEM: 76.0 ± 17.0%). FA was detected within 424/470 (90%) groups of serial sections of transversely cryotomed skin containing follicular infundibula, but never in 48/48 (100%) groups of sections containing only deeper follicular structures, nor in receptor fluid, suggesting that FA does not permeate beyond the infundibulum. The FA concentration (mean ± SEM) in the most superficial 240 μm of skin was 2000 ± 815 μg/g. Conclusions Topically applied FA can greatly exceed MICs for canine pathogenic staphylococci at the most common sites of infection. Topical FA therapy should now be evaluated using available formulations in vivo as an alternative to systemic therapy for canine superficial bacterial folliculitis. Electronic supplementary material The online version of this article (10.1186/s12917-017-1270-6) contains supplementary material, which is available to authorized users. Background Antibiotic resistance is a major threat to global health and modern medicine [1]. Canine pyoderma caused by Staphylococcus pseudintermedius is amongst the commonest reasons for prescribing antimicrobial drugs in small animal veterinary practice [2]. Traditionally in canine practice, surface infections (confined to the interfollicular epidermis) are treated topically, whereas superficial infections such as bacterial folliculitis (that extend to the follicular infundibulum without extension into the surrounding dermis) are treated with oral antibiotics. The recent emergence of methicillin-resistant S. pseudintermedius (MRSP) [3] that are routinely resistant to licensed oral antibiotics has renewed interest in the direct application of topical antibiotics and antiseptics for superficial pyoderma [4,5]. Fusidic acid (FA) is an antibiotic which has a steroidlike structure, with proven activity in vitro against coagulase-positive staphylococci including MRSP [6,7]. The physico-chemical properties of this large, lipophilic molecule (molecular weight of 517 kDa, octanol/water partition co-efficient >4, 6 free hydrogen bonding groups) predict limited diffusivity through stratum corneum and restricted partitioning to the more hydrophilic living epidermis, after topical application [8,9]. These features correlate with clinical efficacy of licensed FAcontaining topical veterinary products in surface infections such as canine acute moist [pyotraumatic] dermatitis [10]. Utility in canine superficial pyoderma, however, is dependent upon adequate permeation into hair follicles, but this has received little attention. Studies of clinical efficacy of topical FA in canine superficial pyoderma / bacterial folliculitis are lacking [4,11]. Stuttgen and Bauer established that in sparsely-haired human skin, FA is limited to the stratum corneum and epidermis after topical gel application, and does not penetrate into the deep dermis or subcutaneous fat [12]. By contrast, Degim et al. reported that 1.3% of FA in a betamethasone-containing gel formulation penetrated full-thickness haired canine skin in diffusion cell studies [13]. Skin integrity was not assessed prior to gel application, and FA was quantified in only receptor fluid and not on or within skin itself [13]. In view of these prior conflicting and incomplete data, we developed an in vitro model of topical FA application using canine skin held in Franz-type diffusion cells and high-performance liquid chromatography and ultraviolet (HPLC-UV) analysis of FA concentrations to define the degree and extent of drug permeation in skin from sites with varying hair follicle density. We describe for the first time how the depth of drug permeation into dermal layers can be defined by concurrent observation, in representative paired transverse histological sections [14,15], of the variations in hair follicle anatomy that mark the infundibulum, isthmus and inferior portions of hair follicles. In addition, conventional analyses of drug recovery in receptor fluid and swabs from surface of dosed skin complemented evaluation of dermal drug concentrations. These data were used to inform likely clinical utility in canine superficial and deep pyoderma. HPLC-UV detection of fusidic acid. Validation Fusidic acid sodium salt (≥98%, Sigma-Aldrich, Irvine, UK) was diluted in absolute ethanol to produce standard (0.5-49 μg/ml) and quality control (0.5, 1.0, 6.5 and 40 μg/ml) solutions (see Additional file 1 for chemicals used). High Performance Liquid Chromatographyultraviolet analysis (HPLC-UV) was performed using an Ultimate 3000 (Thermo Scientific, Paisley, UK) system comprising quaternary pump, autosampler, column oven and diode array detector. The column was from Kinetex (C18 2.1 mm × 50 mm, 1.7 μm particle size; Phenomenex, Macclesfield, UK) held at 35°C. Mobile phase A comprised methanol; mobile phase B 0.1 M acetic acid. Mobile phase A/B was ramped from 30/70 to 78/22 v/v ratio over 4 min and then held for 5 min. Mobile phase A/B then returned to 30/70 over 30 s and reequilibrated for 7 min. The flow rate was maintained at 0.35 ml/min. The retention time of FA was 9.1 min, with UV detection at 240 nm. Samples were analysed using a validated method developed at the University of Hertfordshire in accordance with OECD guidelines for studies of skin absorption in vitro [16]. Injection volume was 2 μl. The linear dynamic range for FA, based on peak areas with 1/x 2 weighted regression was 1.24-249 ng on column (R 2 = 0.9998), with limit of detection (LOD) of 0.50 ng on column (signal/ noise = 3). System precision, determined using replicate injections (n = 10) at 1.24 and 187 ng on column, was 6.6 and 0.76%, respectively. Receptor fluid, cotton wool swab and canine cryosection extracts (from 8 different animals) demonstrated no matrix interference at the retention time of FA. The calibration range was 0.5-49 μg/ml. Satisfactory intra and inter-run accuracy (87-107%) and precision (±15%) was obtained at low (1.0 μg/ ml), mid (6.5 μg/ml) and high (40 μg/ml) concentrations of the calibration curve and at the limit of quantification (LOQ, 0.5 μg/ml). Extraction efficacies of FA from spiked canine skin cryosections (at low, mid, and high QC levels, 8 replicates of each) using ethanol confirmed recoveries to be 98.7-101.3%. Stabilities of spiking solutions and spiked matrices were shown to be at least 2 weeks when refrigerated. All experimental samples were analysed within 14 days of refrigerated storage. Analysis of samples Standard solutions covering the calibration range were run at the start of each batch with low, mid and high QCs bracketing no more than 15 test samples. Batch sample data were accepted when the accuracy and precision of these met validation criteria. In order to obtain concentrations of FA found in skin samples, the amount of FA found in the skin was adjusted for the sample weight from which the sample was obtained. Canine skin collection Full thickness canine skin was obtained from healthy Beagle dogs (three male, three female, aged 6-12 months, 8-14 kg) immediately after euthanasia for reasons unrelated to this study (approved by the Royal Veterinary College's Clinical Research Ethical Review Board 2016 1651-2-R). Hair was clipped (Moser Arco 1854, Wahl, Sterling, IL, USA) to within 3 mm of skin surface taking care not to damage skin integrity, and skin was then excised from the dorsum and groin of each dog, immediately wrapped in tin foil and chilled by frozen ice blocks prior to storage at -20°C within 6 h of collection. Harvested skin was used within 5 months [17]. Skin measurements Thickness of whole skin specimens The thickness of the centre of each 3 × 3 cm portion of skin was measured using callipers immediately before assembly into the diffusion cells, as described below. Thickness of stratum corneum Vertical cryosections through full thickness skin were prepared from undosed skin from each treatment group (undamaged, shampoo-treated and tape-stripped skin) for both dorsum and groin for all six animals (total n = 36). Eight cryosections were taken from each piece of skin, sectioning both from panniculus to epidermis (n = 4) and from epidermis to panniculus (n = 4), and stained with haematoxylin and eosin. Thickness of stratum corneum was measured at three points per section where stratum corneum was at its most compact / intact and not obviously folded [18] using a light microscope (×40 magnification) and Image-Pro Plus v5.0.1.11 software (Media Cybernetics, Duxford, UK). Hair follicle density Hair follicle density (compound follicles per mm 2 ) and infundibular area (as a percentage of skin area) at the level of the common infundibulum were compared in replicate control untreated dorsal (n = 6) and groin (n = 6) skin samples by microscopy of transverse haematoxylin and eosin-stained paraffin sections of skin from one male and one female Beagle dog. The hair follicle count and area were measured at three randomly selected areas per section using microscope settings and software as for stratum corneum measurements described above. Electrical resistance Electrical resistance between saline treated epidermal skin surface and receptor fluid was used to assess skin barrier integrity in each assembled diffusion cell using an ohmmeter (Iso-Tech LCR-821 Meter, Iso-Tech, Southport, UK) [19]. Diffusion cell experiment Four dermal absorption experiments with full thickness canine skin were conducted using a 10 mg/g FA suspension (Isathal®, Dechra Veterinary Products (DVP), Shropshire, UK) which contains the same vehicle as a licensed topical skin product for dogs (Isaderm®, DVP). Dorsal and groin skin sourced from six animals (three dogs per experiment) was defrosted and defatted by blunt dissection before division into three evenly sized pieces (70 cm 2 ), one for each treatment group and assembled into diffusion cells containing receptor fluid within 3 h. For each experiment, 21 static Franz diffusion cells (Permgear Inc. Hellertown, PA USA) were assembled with portions of either dorsal or groin skin (3 cm × 3 cm) which had either been left untreated (n = 6), repeatedly tape-stripped (n = 6) to mimic damage to the stratum corneum, or shampooed with a 2% chlorhexidine and 2% miconazole shampoo (Malaseb®, DVP) to mimic clinical use (n = 6). One diffusion cell containing untreated skin from each animal was assembled but left undosed (negative controls). Skin allocated to be tape-stripped was quickly [20] and repeatedly (n = 30) [20] stripped with D-Squame discs (22 mm; Cuderm, Dallas, TX, USA) prior to assembly into the cells. A uniform pressure was applied to each disc for 2 s using a 225 g/cm 2 applicator before disc removal with forceps. The epidermal surface of relevant portions of excised skin (approximately 70 cm 2 ) were moistened and shampooed (0.02 ml/cm 2 ) for 2 min by hand, then left for 10 min, as per label instructions for clinical use, prior to rinsing with water (2 × 5 ml). Treated skin was blotted dry with paper towels and cut into pieces for assembly into the cells. Each piece of skin was placed between a glass receptor chamber, containing measured volumes (approximately 14 ml) of ethanol / pH 5.0 phosphate buffered saline, 25/75 v/v [13], and a magnetic stirrer, and glass donor chamber and secured by pinch clamp, exposing 1.77 cm 2 . Diffusion cells were randomly assigned positions in stirrer blocks and plumbed into a heated water circulator system in order to maintain a constant skin surface temperature of 32.0 ± 1.0°C confirmed by infrared camera (P620, FLIR, West Malling UK). Skin barrier integrity was established in each cell by measurement of electrical resistance as described above. After equilibration overnight, 18 of the 21 assembled cells were dosed with 100 μl of the FA suspension using a calibrated positive displacement pipette. Saline (0.9%, 100 μl) was added by pipette to all cells in order to liquefy the gel [13]. A glass rod was used to gently spread the gel across the entire exposed surface of skin. The total amount of FA applied to each cell was determined by weighing the filled and emptied pipette tip and the glass rod before and after use. The donor chambers were then promptly occluded using plastic paraffin film. Receptor fluid samples (250 μl) were collected from each cell before (pre-dose) and at 12 and 24 h after dosing, with replacement of equal volumes of fresh receptor fluid at each time point. Additional sampling intervals were deemed unnecessary in anticipation of negligible penetration into the receptor fluid. After 24 h, the donor chamber was removed and the surface skin and inside of the donor chamber were both swabbed with cotton wool. Swabs were transferred to glass vials and soaked in ethanol (10 ml) for a minimum of 24 h at 4°C to extract the FA. Aliquots of the swab extracts and receptor fluid samples were transferred to autosampler vials (2 ml) and crimped capped. Skin specimens were gently removed from the receptor chamber using forceps, taking care not to touch the exposed area of skin, and were stored in foil at −70°C prior to cryosectioning (to minimise drug lability in the skin). FA concentrations in ethanolic swab extracts and receptor fluid samples were determined by direct injection of aliquots, transferred to autosampler vials, using the HPLC-UV method described above. Skin cryosectioning OCT-embedded frozen specimens of skin were cryosectioned transversely starting from the deep dermis proceeding towards epidermis in groups of seven sections, to avoid cross contamination of sections with the blade. Each group comprised a) five sequential 20 μm sections that were placed in individual glass vials and extracted in ethanol (5 ml) for 24 h, for subsequent FA HPLCanalysis, and b) a further two 10 μm sections which were mounted on Polysine™ slides (ThermoFisher Scientific, Paisley, UK) prior to staining with haematoxylin and eosin. The stained slides were examined by light microscopical observation of hair follicle anatomy and presence of other skin structures by a blinded assessor (RB) to determine the anatomical depth within the skin (Figure 1). As the highest proportion of hair follicle infundibula were present in the uppermost two vials (equivalent to a depth of approximately 240 μm) these were selected for determination of the maximum FA concentration achievable in the superficial skin layers in these experiments. Statistical analyses Statistical analyses utilised IBM SPSS Statistics Package 21 (IBM, Portsmouth, UK) with P values of ≤0.05 considered significant. Normality was assessed through Shapiro-Wilk test prior to use of either the Kruskal-Wallis test with Dunn's post-hoc test, or one-way ANOVA with Bonferroni post-hoc correction as appropriate. Chi-squared tests evaluate contingency table data. FA recovery The amount of FA applied to canine skin in each diffusion cell ranged from 762 to 1087 μg (mean ± SEM 946 ± 9 μg). All QCs and standards running alongside samples met validation criteria. FA was never detected in any sample from un-dosed control cells, nor detected within quantifiable limits in any receptor fluid sample 24 h after application (Table 1). HPLC-UV analyses indicated that total FA recovery was 90.2 ± 9.0% (range 65-107%) after 24 h; no significant difference was found between skin sites or treatment groups (Table 1). From skin surface swabs, overall recovery of FA was 76.0 ± 17.7% independent of skin site or treatment group. A significantly (P = 0.002) a d c b Fig. 1 Composite image of the histology of the canine compound hair follicle. a Traditional vertical section through the long axis of a compound follicle from epidermis (right) to panniculus adiposus (left). Lines indicate planes of section for corresponding transverse images that define depth of section. b Transverse section at common infundibulum: follicle is lined by stratified squamous keratinising epithelium that recapitulates that of the interfollicular epidermis and contains multiple naked hair fibres. c Transverse section at isthmus: compound follicle comprises a cranial primary hair and a group of (commonly 14-18) secondary hair follicles; each hair shaft is surrounded by root sheaths whose anatomy varies with stage of hair growth. d Transverse section at inferior portion of follicles: presence indicates anagen phase represented by hair fibre surrounded by inner root sheath and glycogen-rich outer root sheath higher percentage of the applied dose was found within cryosections from the dorsum (17.7 ± 2.4%) in comparison to the groin (10.7 ± 1.0%; Figure 2). No significant inter-dog variability in FA distribution was seen between skin from the six donor dogs within skin group (undamaged, shampooed or tape-stripped) or site (Additional file 2). FA was detected in 80% of vials (376 of 470 vials, equivalent to 72.6% of all 518 vials; Table 2) containing cryosections where follicular infundibula or more superficial structures (surface hairs, living epidermis, keratin and free hairs) were observed in representative (paired) histological specimens. FA was never detected in the absence of these structures, i.e. in sections containing only isthmi, inferior portions of the hair follicle or subcutaneous fat (n = 48 vials; Table 2; P < 0.0005, Chi-squared test). A mean (± SEM) FA concentration of 395.4 ± 30.9 μg/ g was found in each treated portion of skin 24 h after topical application (Table 3). No significant differences (P > 0.05) were seen between the concentrations achieved in any site or treatment group. The concentration of FA (mean ± SEM) in the uppermost (from epidermal aspect) two vials of cryosections of each skin specimen (approximately 240 μm) was 2000 ± 815 μg/g. Discussion The combination of transverse histological sectioning and HPLC-UV assessment of FA concentrations in serial sections was pivotal in ascertaining in detail the depth of FA permeation through canine skin. Coupled with the more conventional processes of drug recovery in post-treatment skin, surface swab and receptor fluid samples, these data confirmed FA permeation to the level of the follicular infundibulum and thus, the presence of drug at the level of infection in canine superficial bacterial folliculitis following topical application. Whilst formulation with different vehicles commonly influences skin permeation, we used a commercially available ophthalmological product that contains the same vehicle as a licensed steroid-containing topical skin product for dogs (Isaderm®, DVP) to maximise the clinical relevance of the results of our in vitro study to veterinary practitioners. This first report of the proportion of FA remaining on canine skin surface after application in vitro (76.0 ± 17.7% at 24 h) was remarkably similar to the 80% figure reported in an analogous study of human skin [12]. The higher hair follicle density dorsally likely accounts for the increased amounts of FA present within skin from this site when compared with groin, particularly since the presence of drug in skin sections was significantly associated with histological observation of infundibulae and other superficial structures. The failure to detect FA in receptor fluid in this canine study was in accordance with a previous in vitro penetration study of human skin [12], and not un-expected from the physicochemical properties of the molecule [8,9]. The limit of detection of this HPLC method was well below the predicted concentration of FA in receptor fluid had we reproduced the 1.3% bioavailability described by Degim et al [13]. The full thickness penetration of FA reported in that study might reflect technical or procedural differences such as apparent absence of barrier integrity testing prior to dosing; [13] ensuring that the epidermal barrier layer maintains its integrity is an essential factor to the successful performance of diffusion cell experiments [16]. In this study, electrical resistance was used for barrier integrity testing, but this does not appear to have been described previously for dogs. Values obtained here fell between those reported for rat (3 kΩ) and pig (4 kΩ) skin [19], in parallel with relative stratum corneum thickness in these species (rat 6.0-13.3 μm < dog 9.4-15.1 μm < pig 13.1-18.1 μm) [21]. Comparable electrical resistance and thus barrier integrity in undamaged, shampooed and tape-stripped skin correlated with the equivalent FA penetration across the three groups. Values reported here should be of use for future skin integrity testing for canine in vitro diffusion experiments. Our model indicates that topical therapy with FA in canine skin is likely to achieve concentrations that markedly exceed MICs of staphylococcal strains deemed both 'susceptible' and 'resistant' using existing interpretative criteria. By extrapolating the mean FA concentration achieved in the top 240 μm of skin (2000 ± 815 μg/g) using a skin density value of 1.09 [22], the overall concentration of FA in this region can be estimated as 2180 ± 634 mg/l. This markedly exceeds previously reported MIC 90 of both methicillin-resistant and susceptible S. aureus and S. pseudintermedius [6,23], and EUCAST systemic therapy breakpoint for 'resistance' (1 mg/l) [24] and compares favourably with MIC 100 values for FAresistant MRSA (1024 mg/l) [7]. Development of interpretive criteria for topical rather than just systemic use of antimicrobial therapy is urgently required. The stratum corneum thickness of undamaged canine skin in this study was closely comparable to those of previous reports [21,25]. Tape-strip removal of stratum Mean (± SEM) concentration of FA measured in transverse cryosections obtained from full thickness dorsum or groin skin from healthy Beagle dogs (n = 6) after topical application of a 10 mg/g FA suspension (Isathal®) for 24 h in four diffusion cell experiments. Skin samples were undamaged, shampooed or tape-stripped (n = 6 per group) prior to dosing. After exposure, the skin was swabbed prior to transverse cryosectioning. No significant differences were seen between the concentrations achieved in any site or treatment group (ANOVA, P > 0.05) corneum cells and lipid is commonly used to degrade the barrier and enhance drug permeability, although poststripping measurements of thickness in cryosections (which best preserve stratum corneum architecture in haired skin) [21,26] are very rarely reported [27][28][29]. We speculate that the failure of tape stripping to significantly reduce canine interfollicular stratum corneum thickness reflects the combined effects of a dense mat of short stubbly hairs reducing D-squame tape access to the interfollicular epidermis (close clipping was avoided to prevent stratum corneum disruption) [29], uneven skin surface [30], thicker corneum at follicular ostia [31], and or preferential removal of loose corneocytes that may be lost or otherwise not included in measurements of residual compact layers. Further studies that optimise parameters, such as applicator pressures, numbers of repeat strips and clipping methods [20] for the thin but compact corneal layers of canine haired skin, are indicated. Conclusions These data suggest that topical FA should be useful in the treatment of canine surface and superficial pyoderma (intact follicles) caused by bacteria susceptible to fusidic acid, in countries where it is available, but not deep pyoderma (where infection extends to surrounding dermis). Clinical studies are now required to confirm this. Similar studies should now be performed for other topically applied antibiotics to inform evidence-based antibiotic treatment guidelines. Although prevalence of antimicrobial resistance should be monitored prospectively, FA provides an opportunity for topical antibiotic therapy in the treatment of staphylococcal folliculitis in dogs and an option to reduce selection pressure for antimicrobial resistance on these zoonotic canine pathogens from conventional systemic antibiotic use.	2017-11-21T17:42:19.260Z	2017-11-21T00:00:00.000	{ "year": 2017, "sha1": "1f55c9ad126977d53db182887b96e554bd64dc57", "oa_license": "CCBY", "oa_url": "https://bmcvetres.biomedcentral.com/track/pdf/10.1186/s12917-017-1270-6", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "1f55c9ad126977d53db182887b96e554bd64dc57", "s2fieldsofstudy": [ "Environmental Science", "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
234594466	pes2o/s2orc	v3-fos-license	Study of Adaptive-Reuse Application in Rumah Atsiri (Ex. Citronella Factory), Karanganyar Abandoned and unused old buildings are often seen as a waste of land. One of the ways that old buildings can be reused is to use the adaptive-reuse concept, which is an attempt to reuse an old, unused building into a different function from the original function while maintaining the original building values. The purpose of this research is to understand how the Adaptive Reuse concept is applied in Rumah Atsiri Indonesia. Firstly, the data has been collected through literature survey and field survey. Then, analyzed the elements of Rumah Atsiri Indonesia that can be categorized as adaptive reuse examples and have been investigated in the light of the defined factors. At the end, according to the result we can find what kind of work that has been done in Rumah Atsiri Indonesia that makes it as a sample of adaptive reuse projects. Rumah Atsiri that we see today is the result of the revitalization of an old factory building with a touch of modern architecture. Revitalization is carried out while maintaining the authenticity of the design of the building. The contrass of materials makes visitors able to distinguish which of the old buildings and which buildings and new elements are added by looking at the materials used. Introduction Abandoned and unused old buildings are often seen as a waste of land. One of the ways that old buildings can be reused is to use the adaptive-reuse concept, which is an attempt to reuse an old, unused building into a different function from the original function while maintaining the original building values. This concept is mainly used in old, historical buildings that are unused and located in the city center or locations that are visited by many tourists. Adaptive-reuse opens up opportunities for buildings to function more efficiently and sustainably as well as opportunities to preserve historical legacies for future generations. With adaptive-reuse, new job opportunities will also open up for the community (Gao et al, 2020). Recently, the adaptive-reuse concept has been widely applied as an effort to conserve historical old buildings. One of those that applies this concept is the building of Rumah Atsiri (eng: House of essentials), ex.The Citronella Factory which later became an integrated tourist destination with facilities for aromatic gardens, laboratories, production house, training centers, museums, restaurants, shops, and MICE facilities. The purpose of this research is to understand how the Adaptive Reuse concept is applied in Rumah Atsiri Indonesia. Firstly, the data has been collected through literature survey and field survey. Then, analyzed the elements of Rumah Atsiri Indonesia that can be categorized as adaptive reuse examples and have been investigated in the light of the defined factors. At the end, according to the result we can find what kind of work that has been done in Rumah Atsiri Indonesia that makes it as a sample of adaptive reuse projects. Adaptive-Reuse Adaptive-reuse is an attempt to reuse an old unused building into new functions that are relevant to current needs while maintaining the original values of the building. 1 st International Conference on Engineering (ICONE 2020) Architectural Support of Heritage, Culture, and Sustainable Development Faculty of Engineering Tunas Pembangunan Surakarta University 50 According to Ennis Davis, there are several advantages of adaptive-reuse, including supporting the concept of sustainability, both for the building itself, environmentally, socially, and economically. According to Rodrigues andFreire (2017, in Gao et al. 2020) adaptive-reuse is the process of strengthening old buildings for new functions. In general, adaptive-reuse application to old buildings is intended for tourism. Therefore, visitor response to adaptive-reuse objects is very important in the success of an adaptive-reuse project. According to Shao et al (2018, in Susanti et al., 2020, adaptive-reuse is a process of revitalizing or reusing existing structures that already exist, but adapted to new use functions, and adaptive-reuse is also a process of transforming buildings that has become obsolete and ineffective into something new that can be reused for a different purpose. Adaptivereuse applications that work with historical building structures are more sustainable from an environmental aspect and also minimize costs incurred for development compared to building a new building with new construction. Adaptive-Reuse in Historical Building Plevoets and Cleempoel (2013( , in Susanti, 2020 introduced 7 conversion concepts based on the Palimsest metaphor from Machado, including (1) building in it, (2) building on it, (3) building around it, (4) building along the sides, (5) ) adapt to the new function, (6) build by maintaining the style, (7) recycling the remaining material. In the same study, Plevoets and Cleempoel also provide different design strategies in adaptive-reuse applications in historical buildings, namely: (1) physical intervention (2) insertion (3) installation. According to them, the most important and significant factor in adaptive-reuse is the original building. Rumah Atsiri Built on an area of 2.3 hectares, according to their website, Rumah Atsiri is an integrated tourist destination containing aromatic gardens, laboratories, essential oil's production houses, essential oil's training centers, citronella factory museums, restaurants, shops, and MICE (Meeting, Incentive, Convention, and Exhibition) facilities. Rumah Atsiri is located in Plumbon Village, Karanganyar Regency, Central Java, about 35 kilometers east of Surakarta City. The old building of Rumah Atsiri was built in 1963-1967 as a form of cooperation between the Indonesian government and the Bulgarian government. After going through a development period of approximately four years, in 1968 the building was opened for the first time and functioned as a Citronella Factory which produces perfume essential oils and is planned to be the largest distillation factory in Asia. The Citronella factory was closed in 1986 due to domestic political conditions at that time. After that, the factory building changed ownership 1 st International Conference on Engineering (ICONE 2020) Architectural Support of Heritage, Culture, and Sustainable Development Faculty of Engineering Tunas Pembangunan Surakarta University 51 several times until finally in 2015 it was owned by PT Rumah Atsiri Indonesia. When it was "rediscovered" in 2015, the condition of the ex.Citronella factory was like any other stalled building. Some corners of the building are damaged with production machines that are not maintained and are missing in several parts. A number of building improvements were made to make the building functional again, however the revitalization or repairs were made to avoid changing the original style of the building which was characterized by a modern architecture. Parts of the building that showed its original function as an essential oil processing plant are also still being preserved, such as the bricks and roster that had been shipped directly from Bulgaria. Figure 2: The original brick and roster from Bulgaria is still used in new building concept Credit: personal documentation The new buildings, which mostly use light steel structures, are made to be connected both in a tourist concept and a visual concept by accentuating the old building with a concrete structure. Adaptive Reuse in Rumah Atsiri Rumah Atsiri that we see today is the result of the revitalization of an old factory building with a touch of modern architecture. Revitalization is carried out while maintaining the authenticity of the design of the building. What remains of the Citronella Factory building are the physical buildings of the three refineries, without machines and large tanks. Until now these buildings still stand firmly with the original construction, the addition of new constructions is more about fulfilling new activity functions. The contrast of materials makes visitors able to distinguish which of the old buildings and which buildings and new elements are added by looking at the materials used. Most of the old buildings use iron and concrete, while the new ones use steel, wood and glass. Figure 5: Contrast between old materials and new materials inside Rumah Atsiri Courtesy of Rumah Atsiri Indonesia Marigold Plaza's arrangement plants, (4)Frame in restaurant's chair Cortesy of Rumah Atsiri Indonesia & Personal Documentation Based on their website, the revitalization of Rumah Atsiri is done using the juxtaposition method (contextual juxtaposition). The original building-that has become the identity of the place-was maintained, but new elements were included without damaging the old building. Today, this area functions as a shop that sells various signature products and souvenirs of Rumah Atsiri Indonesia, while the lower floor is used as a classroom for children to learn about essential oils. The addition of interior and sclupture elements as a sign that the function of space has changed, but still leaves a memory of its old history. Figure 9: Area for (now) The Shops and classroom was then made for kettle and condenser Courtesy of Rumah Atsiri Indonesia Paulus Mintarga as the owner of Rumah Atsiri, in his interview with Kompas.com said that the Rumah Atsiri design was based on an old building that was not demolished but was trying to be connected so that it could accommodate new functions. The vision of Rumah Atsiri still refers to its original goal of developing the essential industry in Indonesia through educational tourism. Rumah Atsiri, seen from its modeling, illustrates how this location shows elements of sustainable tourism. Valerina Daniel, Chair of the Team for the Acceleration of Sustainable Tourism Development, in her report on Kompas.com once explained that sustainable tourism carries 3P elements, namely: People, Planet and Prosperity. Rumah Atsiri shows how this place tries to defend the building from its heretical side. Its existence also seeks to always work together with the village community. According to the manager, around 90 people in the surrounding community become employees at Rumah Atsiri. Some local people were trained to make various signature products and souvenirs that are sold in Rumah Atsiri's Shop. Conclusion The concept of adaptive reuse is to embrace the original character, history, and energy of a building and make it relevant to today's needs. If done well, adaptive reuse in historical buildings can provide significant environmental, social, cultural and economic results. Although it did not explicitly mention the application of the adaptive-reuse concept, we can easily find the characteristics of its application, both tangible and intangible. In Rumah Atsiri Indonesia, the old and the new are juxtaposed to strengthen one another and to show that development can be done without erasing the history of a place	2021-05-16T00:04:05.608Z	2020-11-28T00:00:00.000	{ "year": 2020, "sha1": "52adb70ca4675977e478cea35220fe442c4e6b6c", "oa_license": "CCBYNCSA", "oa_url": "http://ejournal.utp.ac.id/index.php/ICONE/article/download/1270/520521054", "oa_status": "HYBRID", "pdf_src": "Anansi", "pdf_hash": "90bab4ac4bfa2b473098afc9e8417c7f70a43a82", "s2fieldsofstudy": [ "Environmental Science", "Engineering" ], "extfieldsofstudy": [ "Computer Science" ] }
267197483	pes2o/s2orc	v3-fos-license	Validation of the acoustic change complex (ACC) prediction model to predict speech perception in noise in adult patients with hearing loss: a study protocol Background Speech perception tests are essential to measure the functional use of hearing and to determine the effectiveness of hearing aids and implantable auditory devices. However, these language-based tests require active participation and are influenced by linguistic and neurocognitive skills limiting their use in patients with insufficient language proficiency, cognitive impairment, or in children. We recently developed a non-attentive and objective speech perception prediction model: the Acoustic Change Complex (ACC) prediction model. The ACC prediction model uses electroencephalography to measure alterations in cortical auditory activity caused by frequency changes. The aim is to validate this model in a large-scale external validation study in adult patients with varying degrees of sensorineural hearing loss (SNHL) to confirm the high predictive value of the ACC model and to assess its test–retest reliability. Methods A total of 80 participants, aged 18–65 years, will be enrolled in the study. The categories of severity of hearing loss will be used as a blocking factor to establish an equal distribution of patients with various degrees of sensorineural hearing loss. During the first visit, pure tone audiometry, speech in noise tests, a phoneme discrimination test, and the first ACC measurement will be performed. During the second visit (after 1–4 weeks), the same ACC measurement will be performed to assess the test–retest reliability. The acoustic change stimuli for ACC measurements consist of a reference tone with a base frequency of 1000, 2000, or 4000 Hz with a duration of 3000 ms, gliding to a 300-ms target tone with a frequency that is 12% higher than the base frequency. The primary outcome measures are (1) the level of agreement between the predicted speech reception threshold (SRT) and the behavioral SRT, and (2) the level of agreement between the SRT calculated by the first ACC measurement and the SRT of the second ACC measurement. Level of agreement will be assessed with Bland–Altman plots. Discussion Previous studies by our group have shown the high predictive value of the ACC model. The successful validation of this model as an effective and reliable biomarker of speech perception will directly benefit the general population, as it will increase the accuracy of hearing evaluations and improve access to adequate hearing rehabilitation. Background Hearing impairment is the most frequent sensory deficit in humans, affecting almost 20% of the population worldwide.It has been listed by the World Health Organization as a priority disease for research into therapeutic interventions [1,2].Hearing loss not only affects communication and quality of life but also causes social distress and anxiety and has a negative impact on cognitive functioning [3][4][5][6].Currently, no treatment is available to prevent or halt the progression of sensorineural hearing loss (SNHL).Management mainly consists of hearing rehabilitation with hearing aids and/or cochlear implants depending on the grade of hearing loss.Speech perception tests are essential to measure the functional use of hearing, to determine the effectiveness of hearing aid fittings, and to evaluate cochlear implant candidacy [7].However, these language-based tests are influenced by linguistic and neurocognitive skills [8,9].In Belgium, and in particular in the Flemish region, validated speech perception tests are only available in the Dutch and French languages.For patients with insufficient proficiency in these languages, audiologists are unable to obtain a reliable evaluation of their functional hearing impairment.The lack of language-independent tests to assess auditory function and speech perception further restricts access to healthcare and adequate hearing revalidation with hearing aids or cochlear implants for adults as well as for children [10].The consequences of untreated hearing loss in these vulnerable populations are far-reaching, worsening their social isolation and impeding their process of integration into society.The same holds for adults and children with intellectual disabilities in whom no conventional speech discrimination tests can be performed due to their limited cognitive abilities.In this population, genetic causes of hearing loss (syndromic and non-syndromic) are frequent, and adequate hearing rehabilitation is challenging due to their cognition, attention span, and cooperation.To overcome these challenges we want to introduce a new biomarker to predict speech perception in noise. Speech perception is strongly associated with spectral shape discrimination [11,12].Unfortunately, frequency discrimination tests require active participation and can be challenging for those with hearing impairment.Therefore, we have developed a non-attentive model using the Acoustic Change Complex (ACC) to predict speech perception.The ACC is an auditory evoked potential revealing the cortical response evoked by changes within an ongoing stimulus [13]. For our prediction model, ACCs are evoked in response to 12% frequency increases (i.e., 1000-1120 Hz; 2000-2240 Hz; 4000-4480 Hz) at three base frequencies (1, 2, and 4 kHz) [14].Multiple regression analysis for prediction of speech perception in noise (SRT; dB SNR) revealed that the strongest prediction model was obtained by averaging the obtained ACC latencies and average hearing loss measured with pure-tone audiometry at those three frequencies [14].This model was able to explain 87% of the total variance, indicating that subjects with longer ACC latencies have worse speech perception in noise than subjects with comparable hearing thresholds and shorter ACC latencies.If HL was removed from this model, the combination of ACC amplitude and latency over those three frequencies still explained 74% of the total variance in speech perception in noise (r 2 = 0.74, p < 0.001) [14].The major advantage of the ACC over the currently available audiometric tests is that it is languageindependent, and it does not require active participation of the listeners.Therefore, it can be used in patients with insufficient language proficiency and-potentiallycognitive impairment, or in cases where conventional audiometry is unreliable, or malingering is suspected.Moreover, the ACC would be most beneficial for evaluating the auditory abilities of the growing population of patients who struggle with healthcare accessibility due to language barriers. In the current study, we aim to validate this ACC prediction model in a large-scale study.This project will be the first study to externally validate a highly promising prediction model for speech perception in noise.The results of this study are essential to confirm and validate the predictive potential of this new objective measurement, developed at UMC Utrecht [14] and is a prerequisite for clinical implementation. Study design Patients visiting the Antwerp University Hospital (UZA) Ear, Nose and Throat (ENT) outpatient clinic will be screened thoroughly for potential eligibility.If they meet the inclusion criteria and agree to participate, patients will be enrolled in the study (Fig. 1).The baseline evaluation will consist of pure-tone audiometry, speech-innoise tests, a phoneme discrimination test, and a total of three ACC recordings.The entire test procedure will take about 2 h.Additionally, participants will be asked to fill in a questionnaire asking about their hand dominance, language proficiency, and musical experience.Recent literature has shown that professional musical training is beneficial for improving frequency discrimination and the ability to detect changes in frequency, leading to alterations in the ACC amplitudes [15,16].At the end of the first session, an appointment for the second session will be scheduled after 1 to 4 weeks. During the second visit, the same three ACC recordings will be obtained for each patient.This allows us to assess the test-retest reliability of the ACC prediction model.Test-retest reliability is a valuable method for assessing the stability of the ACC prediction model over time.Pure tone audiometry, speech-in-noise testing, and a phoneme discrimination test will not be performed again during the second visit.In the brief timeframe spanning 1 to 4 weeks, significant alterations in puretone thresholds, speech-in-noise scores, or phoneme discrimination are not anticipated.If patients perceive a subjective change in hearing during the test interval, pure-tone audiometry will be repeated.In case of significant deviations in thresholds, patients will be excluded from the study. To ensure that all audiometric tests and ACC recordings are performed in an identical manner for all participants, the investigators are trained prior to the start of the study and will adhere to a strict study protocol.All tests will be performed in the same order for all participants.One researcher will perform audiometric testing while another researcher will perform the ACC recordings.For each patient, the investigator performing the audiometric tests will be blinded from the ACC recordings and vice versa to prevent potential bias. Participants Study participants will be consecutively sampled from patients visiting the Antwerp University Hospital (UZA) ENT outpatient clinic for evaluation of their hearing.The study population is therefore a direct sample from the target population of adults presenting with subjective hearing loss to an outpatient ENT/audiology clinic and requesting evaluation of their hearing.Potential subjects must meet specific criteria during screening at the ENT clinic of the UZA before inclusion in the study.During the study, the subject has the right to drop out of the study at any time.A subject can be included in the study if the following criteria are met: If potential subjects present themselves with any cerebral condition (e.g., CVA), neurodegenerative diseases (e.g., multiple sclerosis, Parkinson's disease), or insufficient language proficiency and/or cognition, they will be excluded from participation in the study.This will be verified based on the patient's medical history.Patients with known middle ear pathology, Menière's disease, or observed pathology during baseline examination will also be excluded. If patients report alterations of their hearing at the time of the second visit, pure tone audiometry will be repeated and if there is deviation of ± 1 dB SD for these tests compared to the first visit, the patient will also be excluded. ACC measurement ACCs will be recorded using the procedure described by Vonck et al. [17].The acoustic change stimuli consist of a reference tone with a base frequency of 1000, 2000, or 4000 Hz with a duration of 3000 ms gliding to a 300 ms target tone with a frequency that is 12% higher than the base frequency (i.e., 1000-1120 Hz; 2000-2240 Hz; 4000-4480 Hz).The same triplet of ACC stimuli will be used during the first and second visits.The three ACC stimuli will be presented at three base frequencies of 1000, 2000, and 4000 Hz.Sound stimuli will be presented monaurally, to the best hearing ear, through RadioEar DD45 supra-aural headphones at a level of 75 dB SPL in normal-hearing subjects or at maximum comfortable loudness (MCL) level in subjects with SNHL in order to attempt to correct for differences in loudness.This will result in stimulus presentation levels ranging from 75 dB SPL to MCL level with 100 dB SPL as the upper limit for those with SNHL. Participants will be seated in a comfortable reclining chair in an electrically shielded, sound-attenuated booth and are allowed to watch a silent, captioned movie.Electrophysiological responses will be recorded by electrodes placed according to the 10-20 system using the Synergy Nicolet EDX evoked potential system.The active electrode will be placed at the vertex of the skull (Cz), the contralateral mastoid (A1/A2) will be used as the reference electrode and the ground electrode will be placed on the forehead.Eye movements and blinks will be monitored using electrodes above and below the eye, contralateral of the stimulated ear.Blink artefact rejection will be applied during the recordings.Responses will be recorded using a sampling frequency of 50 kHz and filtered from 0.01 to 100 Hz.For each recording, 100 accepted sweeps will be averaged to obtain individual grand averages.The total duration of three ACC recordings required for the prediction model is around 30 min.Only the best hearing ear will be assessed.For waveform analysis, the N1 latency, N1 amplitude, P2 latency, P2 amplitude, and N1-P2 amplitude of the ACC response will be determined. Speech reception in noise testing (SRT) Speech in noise scores (in dB SNR) will be obtained according to two clinically validated procedures; the Leuven Intelligibility Sentences Test (LIST) [18] and the Nederlandse Vereniging voor Audiologie (NVA) Consonant-Vowel-Consonant (CVC) words [19]. Phoneme discrimination test (A §E discrimination test) The Auditory Speech Sounds Evaluation (A §E ® ) is a software package containing a phoneme discrimination test used to determine the ability to discriminate two different phonemes [20].This auditory test will solely be used as a secondary outcome measurement. Questionnaire All participants will be questioned about their hand dominance, language proficiency, and musical experience.Participants will be asked if they practice music and if so, how many hours they play per week and for how many years.In accordance with Vonck et al. (2021) and Van Heteren et al. ( 2022), a 'musical experience score' will be calculated by multiplying the average amount of musical experience in hours per week by the years of active engagement.A score of > 15 reflects significant musical engagement [17,21]. Primary outcome measurements The level of agreement between the predicted SRT by the original ACC prediction model and the actual SRT as determined using the LIST and NVA list will be assessed.The maximum limit of acceptable difference is set at ± 2 dB. The level of agreement between the SRT calculated from the first ACC measurement (1st visit) and the SRT calculated from the second ACC measurement (2nd visit) will be assessed with Bland-Altman plots.The maximum limit of acceptable difference is set at ± 2 dB (2 × 1 dB SD). Secondary outcome measurements First, the intraclass correlation coefficient (ICC) and Pearson r will be determined between the ACC-predicted SRT and the measured SRT.Secondly, the correlation between the ACC N1 peak latencies and amplitudes and the A §E discrimination test will be determined.Lastly, the ICC and Pearson r between the SRT calculated by the first ACC prediction model (1st visit) and the SRT calculated by the second ACC prediction model (2nd visit) will be assessed. Sample size calculation Based on the prediction model: SRT = − 6.4 + 0.071HL + 0.083(ACC latency -100), with ACC latency being the average of ACC recordings at three different base frequencies and HL being the PTA (average hearing loss for 1000, 2000, and 4000 Hz), we will have 6 variables in the model with SRT and HL in dB and ACC latency in ms [14].With a large effect size of f 2 = 0.6667, an α = 0.01 and β = 0.99, we would need a total sample size of n = 62 (GPower 3.1, F-test linear multiple regression, R 2 deviation from zero).This is in line with the rule of thumb for sample size calculations for regression analyses of n = 10 per variable.We expect a potential patient dropout of 20% due to study visit duration.Therefore, a total of 78 subjects will be recruited. The categories of severity of SNHL will be used as a blocking factor to establish an equal distribution of patients with various degrees of SNHL.The following 5 categories of SNHL degree, according to the international standards established by the World Health Organization, will be included: normal, mild, moderate, moderatesevere, and severe [22]. Discussion Each year, the UZA ENT department has over 10,000 hearing-related consultations.Based on this large number of patient visits, we expect the risk of low enrollment to be minimal.If study enrollment is lower than expected, the research group will reach out to the advisory board partners, which include patient organizations and hearing clinics.Study participants will be asked to visit twice for a study session of approximately 2 h for the first session and 1 h for the second.Since patients do not directly benefit from participation in this study, we expect a dropout of 20%.If the dropout is higher, the following fallback strategies will be considered.First, the duration of the first visit can be reduced by omitting the A §E phoneme discrimination test.This will reduce the duration of the study visit by approximately 10-15 min.Secondly, a subject can be excluded from the test-retest validity study from the study protocol.This will reduce the time investment of the participant by half, as the second visit is not required or mandatory anymore.With an effect size of 0.6667, an α = 0.01, and β = 0.80, we would need at least a remaining sample size of n = 30 (GPower 3.1, t-test means, difference between two dependent means (matched pairs)).If we expect a drop-out of 20% during the second visit, we need a total sample size of n = 36 in order to evaluate the test-retest reliability. So far, attempts have been made to find a good model to predict speech in noise with different brainstem and cortical response paradigms, like the cortical auditory evoked potential (CAEP) and P300.However, due to a lack of clinically useful and consistent correlations, none of these objective measures have made it beyond research applications [23][24][25][26][27].The prognostic ACC model is the first objective measure to predict speech in noise perception with high accuracy [14].Since this prediction model was based on a study population of 37 adult subjects with only 13 subjects with SNHL, it is essential to validate this prediction model in an independent and large study population.This project is highly innovative because the ACC prediction model can fill a notorious gap in the evaluation of hearing impairment.ENT departments and audiology clinics worldwide struggle with patients who have insufficient language proficiency or cognitive abilities to fulfill the conventional speech perception tests, children who are difficult to test, or cases where malingering is suspected.If this project proves that the ACC model provides a reliable prediction of speech in noise perception, it will provide the opportunity to perform adequate hearing evaluations in the aforementioned populations in whom no reliable audiometric assessments can be performed.The ACC model can improve hearing evaluation in the general population and patients will directly benefit from this diagnostic advancement, as it can provide access to better hearing rehabilitation. - All subjects must be 18-65 years old and must have signed an informed consent form.-For inclusion in the normal hearing group, subjects must have a hearing threshold of ≤ 15 dB hearing level (HL) on pure tone average (PTA) at 500, 1000, 2000, and 4000 Hz, or ≤ 20 dB HL at one or more frequencies between 125 and 8000 Hz.The air-bone gap must be < 15 dB at 500, 1000, 2000, and 4000 Hz. -For inclusion in the SNHL group, subjects must have a hearing threshold of > 15 dB HL on PTA at 500, 1000, 2000, and 4000 Hz, or > 20 dB HL at one or more frequencies between 125 and 8000 Hz.Patients exceeding 70 dB HL will be excluded.	2024-01-25T06:17:20.474Z	2024-01-23T00:00:00.000	{ "year": 2024, "sha1": "9263df0d14c61c49972bc83c5a00c4e84095e8b8", "oa_license": "CCBY", "oa_url": "https://diagnprognres.biomedcentral.com/counter/pdf/10.1186/s41512-024-00164-6", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "2dcd50c79da4685c354758f47054d00e34130e1b", "s2fieldsofstudy": [ "Medicine", "Engineering" ], "extfieldsofstudy": [ "Medicine" ] }
235733746	pes2o/s2orc	v3-fos-license	Insula to mPFC reciprocal connectivity differentially underlies novel taste neophobic response and learning in mice To survive in an ever-changing environment, animals must detect and learn salient information. The anterior insular cortex (aIC) and medial prefrontal cortex (mPFC) are heavily implicated in salience and novelty processing, and specifically, the processing of taste sensory information. Here, we examined the role of aIC-mPFC reciprocal connectivity in novel taste neophobia and memory formation, in mice. Using pERK and neuronal intrinsic properties as markers for neuronal activation, and retrograde AAV (rAAV) constructs for connectivity, we demonstrate a correlation between aIC-mPFC activity and novel taste experience. Furthermore, by expressing inhibitory chemogenetic receptors in these projections, we show that aIC-to-mPFC activity is necessary for both taste neophobia and its attenuation. However, activity within mPFC-to-aIC projections is essential only for the neophobic reaction but not for the learning process. These results provide an insight into the cortical circuitry needed to detect, react to- and learn salient stimuli, a process critically involved in psychiatric disorders. Introduction Determining salient stimuli from the continuously perceived sensory input is a key component for effective learning that has been shaped throughout evolution. The term 'salient' is used in psychology and neuroscience to denote a particularly distinct or important stimulus, or an aspect of it (Yantis and Hillstrom, 1994). Attributing salience to a given cue or stimulus, can be affected by genetics, previous experiences, current psychological and motivational state, stress response, goals and motivation (Puglisi-Allegra and Ventura, 2012; Goldberg et al., 2006). Numerous studies have identified cortical and subcortical brain structures that are involved in the detection of salient stimuli, with the insular cortex (IC) being a key node of the salience network (Gogolla, 2017;Livneh et al., 2017;Uddin, 2015). In particular, the IC is believed to have a critical role in the bottom-up detection of salient events and attention allocation (Menon and Uddin, 2010). Thus, when a salient stimulus is detected, the insula assists in targeting other brain structures that enable access to attention and working memory resources. Salient experiences are comprised of several dimensions, which include novelty amongst others (Modinos et al., 2015;Schultz, 2013;Winton-Brown et al., 2014). Experiencing a new taste therefore represents a salient experience of novel sensory information, and the memory for this event is dependent on the functionality of the IC (Gal-Ben-Ari and Rosenblum, 2011;Yiannakas and Rosenblum, 2017). Taste recognition, upon which novelty and salience detection inversely depend, relies on a combination of instinctive responses (genetic elements) and brain processes involved in recalling past experiences to help animals cautiously avoid poisonous food, while taking advantage of food that is deemed safe. Therefore, when animals encounter a taste that is completely new to them, they approach it with cautiousness, a phenomenon called taste neophobia (Lin et al., 2015). This neophobic response is the first line of defense, and therefore an important and evolutionarily conserved behavior, manifesting in reluctance to consume the novel taste, for fear of potentially adverse outcomes. Following this tentative consumption of the novel taste, and in the absence of any proceeding aversive visceral consequences, a memory for a safe taste is formed (Osorio-Gó mez et al., 2018). Thus, the consumption of this taste gradually increases in the following encounters, a phenomenon called attenuation of neophobia (Barot and Bernstein, 2005;Green and Parker, 1975;Moró n and Gallo, 2007). This is an example in which a salient, novel stimulus (1) increases attention which affects the preliminary response, and then (2) initiates incidental encoding for a memory of the new experience as to subsequently prime future adaptive behavior (Borsook et al., 2007;Anderson et al., 2006;McGaugh, 2006). Thus, acting when the sensory information is new, and then learning the new stimulus (familiarization), are both crucial but different components in the process of behavior to a salient, novel stimulus. The process by which a novel taste becomes a familiar one is partially subserved by the IC (Bermudez-Rattoni, 2014), and the gustatory cortex, located within the medial portion of the anterior IC (aIC) (Gehrlach et al., 2020;Shura et al., 2014;Rolls, 2016;Yiannakas and Rosenblum, 2017). The aIC plays a critical role in the formation and retrieval of novel taste memory (Kayyal et al., 2019;Lavi et al., 2018;Yiannakas and Rosenblum, 2017). Inhibition of protein synthesis or lesions to the IC cause impairment in novel taste learning (Rosenblum et al., 1993). Several molecular changes occur in the IC following novel taste learning (Gal-Ben-Ari and Rosenblum, 2011;Gould et al., 2020), including increased phosphorylation-activation of the extracellular regulated kinase (ERK). Phosphorylated ERK (pERK) levels are increased in the IC following novel taste consumption (Berman et al., 1998) and inhibition of MEK, an upstream kinase of ERK in the IC, impairs novel taste learning (Berman et al., 1998). Although an accumulating body of work has been directed toward understanding the aIC and the molecular mechanisms by which novel taste memory is formed and retrieved (Adaikkan and Rosenblum, 2015;Merhav et al., 2006;Yefet et al., 2006), the cellular and circuit mechanisms are yet to be deciphered. Additionally to the aIC, the mPFC is another region that plays a prominent component of the novelty network that is engaged following an encounter with a novel taste stimulus (Morici et al., 2015;Takehara-Nishiuchi et al., 2020). Lesions of the mPFC result in impaired learning of novel tasks (Barker et al., 2007;Devito and Eichenbaum, 2011), including taste learning (Gonzalez et al., 2015). The mPFC is reciprocally connected with the IC, however, whether and how their connectivity contributes to novel taste learning has not been comprehensively examined (Gabbott et al., 2003). Hence, we set out to assess the involvement of aIC-mPFC functional connectivity in the neophobic response evoked during novel taste exposure, as well as novel taste learning. To do so, we measured the correlation between the number of activated, pERK + neurons within the general aIC and mPFC populations, and specifically within aIC-to-mPFC or mPFC-to-aIC projecting neurons. We found that experiencing a novel taste causes a significant increase in reciprocal aIC-mPFC neuronal activity as measured by pERK + neurons (Jones et al., 1999). From an electrophysiological perspective, we identify a correlation between taste novelty and increased excitability in aIC-to-mPFC projecting neurons. In order to ascertain behavioral causality of the molecular and electrophysiological correlations we identified, we inhibited the activity of each pathway during the reaction to-or the learning of-a novel taste. We found that chemogenic inhibition of the aIC-to-mPFC pathway resulted both in impaired neophobic response, as well as impaired taste learning and memory. In contrast, inhibition of the opposite, reciprocal, mPFC-to-aIC pathway, caused impairment to the neophobic response only, but not to the learning process. Our results show that activation of the aIC-mPFC circuit is correlative and necessary for both the neophobic response to novel tastants, as well as memory formation, with functionally discrete reciprocity. This provides part of a functional map of the broader salience system that is yet to be described, as well as providing a guideline for setting the existing molecular knowledge to reveal the unexplored circuit framework. Materials and methods Animals Animals used were 8-to 12-week-old wild type (WT) adult (C57BL/6) male mice. Mice were kept in the local animal resource unit at the University of Haifa in an environment that is temperature-controlled and under a 12 hr dark/light cycle. Water and chow pellet were available ad libitum. All experiments and procedures conducted were approved by the University of Haifa Animal Care and Use committee under Ethical license 554/18 and were in accordance with the National Institutes of Health guidelines for ethical treatment of animals. Surgery and viral injections For all surgeries, naive adult male mice were used as detailed. Mice received i.p. injections of norocarp (0.5 mg/kg), 30 min before surgery and 24 hr later, to minimize pain. Animals were anesthetized by i.p. injection of ketamine and Dormitor (0.5 mg/kg) and then placed in a stereotaxic device (a Model 963 Kopf Instruments stereotaxic injection system). After exposing the scalp, and using bregma and lambda as relevant alignment points, holes were drilled (0.4 nm) in both hemispheres. Mice were injected with AAV construct at the IC (A-P: +0.86; D-V-4; M-L: ±3.4) and/or the mPFC (A-P: 2.34; D-V: À2; M-L:±0.3) or the basolateral amygdala (BLA) (A-P: À1.6; D-V-4.8; M-L: ±3.375), as stated in each experiment. All AAV constructs used in this study were obtained from the viral vector facility of the University of Zurich (http://www.vvf.uzh.ch). All mice used in our studies were injected with 0.25 ul/site/hemisphere of said AAV constructs (physical titer: 4.5 x 10E12 vg/ml) as defined in each experiment. Viral constructs were injected at a rate on 0.1 ml/min, while the syringe was kept in the injection site for 5 min prior and 10 min following delivery, to ensure efficient distribution of the virus. The skin was closed using Vetbond glue, and mice were given 4 weeks of recovery to guarantee viral expression. After recovery, mice were split into individual cages for behavioral experimental purposes. Immunohistochemistry and quantification For perfusion, mice that underwent behavioral experimentation were deeply anesthetized using isoflurane. When the mice were completely anesthetized, they were perfused transcardially using 4% paraformaldehyde (PFA) in 0.1M PBS solution (PBS, MFCD00131855, Sigma-Aldrich). After perfusion, brains were placed in falcon tubes containing 4% formaldehyde solution overnight at 4˚C. On the following day, the solution was replaced with 30% sucrose in 0.1 M phosphate buffered saline (PBS) and brains were incubated for another 48 hr. Afterwards, brains were stored at À80˚C. Coronal sections (40 mm thickness) were then collected from the IC (Bregma: 1.10 mm to 0.26 mm) or the mPFC (Bregma 1.94 mm to 1.54 mm) utilizing a Leica cryostat CM 1950. Following sectioning, slices were washed using PBS for three times, and then blocked for 1 hr using a 0.3% bovine serum albumin (10775835001, Sigma-Aldrich), 0.3% triton X-100 (MFCD00128254, Sigma-Aldrich) and 10% fetal bovine (MFCD00132239, Sigma-Aldrich) solution in PBS. After washing three times with PBS, slices were incubated at 4˚C with primary antibodies against rabbit phospho-ERK (1:200 D13.14.4E Cell Signaling) and mouse NeuN (1:500 MAB377 MERCK) in blocking solution. On the next day, the primary antibody solution was removed, and slices were washed three times with PBS. Then, slices were kept in the dark in blocking solution containing donkey anti-rabbit Alexa Fluor 647 (1:500 ab150075 Invitrogen) and Goat anti-mouse Alexa Fluor 488 (1:500 A11001 ThermoFisher) secondary antibodies, for 1.5 hr. Slices were then washed three times using PBS, and mounted on glass slides and Vectashield mounting solution containing DAPI (H-1200) was added prior to adding coverslips. Slices were visualized and images were acquired using a vertical light microscope (Olympus Cell-Sens Dimension ) using Â10 or Â20 magnification. pERK data was subsequently quantified and analyzed in terms of the bilateral number of positive cells/slice using two-way ANOVA (Graphpad Prism ), to examine the respective responses in the general population. The number of rAAV + cells was similarly quantified and analyzed in the above slices. The images of IC and/or mPFC slices were processed with Image-Pro Plus V-7, Media Cybernetics , pERK + and rAAV + neurons were quantified manually in the relevant subregions and layers. To assess the recruitment of the IC-to-mPFC or mPFC-to-IC projecting neurons in the pERK population, the number of cells co-localized with rAAV + cells was normalized to the total rAAV + population [(rAAV + pERK + / rAAV + )100%] in the respective subregions and layers in slices of the IC or the mPFC. Quantification was done using randomly assigned IDs for individual animals, regardless of treatment. Following quantification, we confirmed treatments, and data were analyzed (Graphpad Prism ). Behavioral experiments Mice were randomly allocated to experimental groups. Group size range estimation was based on previously published results using similar methods, as well power calculations (https://www.stat.ubc. ca/~rollin/stats/ssize/n2.html). Immunohistochemical studies Mice aged 8-12 weeks were injected with rAAV-hSyn1-chI-mCherry-WPRE-SV40p(A), rAAV construct, at the mPFC (to label IC-to-mPFC projecting neurons) or at the IC (to label mPFC-to-IC projecting neurons). One month later, animals were individually separated and given 5 acclimation days. Next, mice were water deprived for 24 hr, and were given tap water (for novel saccharin group) or saccharin (0.5% dissolved in tap water, for familiar saccharin group) pipettes for 20 min sessions each day for 6 consecutive days. On the 7th day, all animals were presented with 1 ml of saccharin, and perfused 20 min later, for immunohistochemical analysis. To target mPFC-to-IC projecting neurons for inhibition, the same constructs were used to injections in a different set of mice, with the retrograde-Cre AAV construct being injected in the IC, and the Cre-dependent DREADD injected at the mPFC. Animals were then given 4 weeks of recovery. Subsequently, animals were individually separated and given 5 days for acclimation. Mice were then water deprived for 24 hr, and given 20 min of pipettes containing tap water for 20 min session for 3 days. Next, animals were i.p. injected with CNO (diluted in saline; 0.5 mg/kg; Enzo) or Saline for control (1% body weight) 1 hr prior to a choice test. Choice test of two pipettes was provided to the mice, with one pipette containing 0.5% saccharin (their first exposure to saccharin) and the other containing water. Drinking volumes for each solution were measured and aversion index scores were calculated (aversion index = volume of water consumed/volume of (water + saccharin) consumed%). During the subsequent 4 days, mice were presented with unreinforced choice tests of saccharin and water, for 20 min session each day, without any intervention. On the 6th day, animals were only given water in pipettes. Toward examining the importance of the circuitry during familiar taste retrieval, mice were injected on the 7th day with CNO or saline, 1 hr prior to an additional choice test. Circuit inhibition during aversive taste memory retrieval In order to test the importance of IC-to-mPFC projecting neurons in aversive taste memory retrieval, animals were injected with AAV8_hEF1a-dlox-hM4D(Gi)_mCherry(rev)-dlox-WPRE-hGHp(A), a Credependent AAV construct expressing the inhibitory DREADD receptor at the IC, and rAAV-hSyn1-chI-EGFP_2A_iCre-WPRE-SV40p(A), a retrograde AAV Cre construct at the mPFC. 4 weeks after the surgery, mice were individually housed and allowed 5 days for acclimation. Mice were then water deprived for 24 hr prior to being allowed access to water pipettes in 20 min sessions for 3 days. On the following day, mice were trained in CTA by exposing them to saccharin for the first time. Forty min after saccharin consumption, mice were i.p. injected with a 2% body weight dose of LiCl (0.14 M), a malaise inducing agent. Mice were water-restricted for the subsequent 2 days. On the 4th day, mice were injected with CNO or saline, 1 hr before saccharin/water choice test. After 20 min of choice test, drinking volumes were measured and aversion index scores were calculated as above. Circuit inhibition during innately aversive novel taste exposure In order to target IC-to-mPFC projecting neurons for inhibition, mice aged 8-12 weeks were bilaterally injected with AAV8_hEF1a-dlox-hM4D(Gi)_mCherry(rev)-dlox-WPRE-hGHp(A), a Cre-dependent AAV construct expressing the inhibitory DREADD receptor in neurons of the IC, and bilateral injection of rAAV-hSyn1-chI-EGFP_2A_iCre-WPRE-SV40p(A), a retrograde-Cre AAV construct, at the mPFC. Animals were individually separated a month after the injection and given 5 days of acclimation. Mice were then water deprived for 24 hr and were given water pipettes for 20 min each day for 3 days. On the test day, mice were injected with CNO or saline one hour prior to a choice test between water and Quinine (0.04%). The mice were presented with the choice test for 20 min and drinking volumes of each solution was measured. Aversion was calculated accordingly. IC-to-BLA circuit inhibition during expression of neophobia In order to target the IC-to-BLA circuit for inhibition, mice aged 8-12 weeks were bilaterally injected with AAV8_hEF1a-dlox-hM4D(Gi)_mCherry(rev)-dlox-WPRE-hGHp(A), a Cre-dependent AAV construct expressing the inhibitory DREADD receptor in neurons of the IC. Bilateral injection in the BLA with rAAV-hSyn1-chI-EGFP_2A_iCre-WPRE-SV40p(A), a retrograde-Cre AAV construct, was also conducted, as to target IC-to-BLA projecting neurons for inhibition. Animals were then given 4 weeks of recovery, prior to being individually housed and given 5 days for acclimation. Then, animals were water deprived for 24 hr, and given 20 min of pipettes containing tap water for 20 min session for 3 consecutive days. Next, animals were i.p. injected with CNO (0.5 mg/kg; Enzo) or Saline for control (1% body weight) 1 hr prior to a choice test. Choice test of two pipettes was provided to the mice, with one pipette containing 0.5% saccharin (their first exposure to saccharin) and the other containing water. Drinking volumes were measured and aversion was calculated accordingly. During the subsequent 4 days, mice were presented with unreinforced choice tests of saccharin and water, in 20 min sessions each day, without any additional intervention. IC-to-mPFC/ mPFC-to-IC circuit inhibition during an open field test Mice aged 8-12 weeks were bilaterally injected with AAV construct expressing the inhibitory DREADD receptor in IC-to-mPFC or mPFC-to-IC projecting neurons as previously described. As a control study, another set of mice were injected with a control Cre-dependent mCherry-expressing AAV vector at the aIC and a retrograde-Cre AAV construct at the mPFC. All three groups were i.p. injected with CNO, and 1 hr later, they were placed in a 50Â50 cm open-field arena (Noldus Information Technology), for 10 min. Electrophysiology To label neurons in the IC that project to the mPFC, mice were injected with rAAV-hSyn1-chI-mCherry-WPRE-SV40p(A) construct, at the mPFC. One month after the surgery, mice were split to individual cages, and given 5 days of acclimation period. Mice were water deprived for 24 hr, and then presented with water (for novel saccharin group) or saccharin (for familiar saccharin group) for six consecutive days. On the 7th day, mice were presented with 1 ml of saccharin and sacrificed 1 hr later. A cage control group, injected with the rAAV-hSyn1-chI-mCherry-WPRE-SV40p(A) construct at the mPFC, were also sacrificed without any separation nor water restriction procedure. To obtain brain slices containing the aIC, mice were deeply anesthetized with 5%isoflurane and transcardially perfused with 40 ml of ice-cold oxygenated cutting solution containing the following (in mM): 25 NaHCO3, 105 Choline-Chloride, 2.5 KCl, 7 MgCl2, 0.5 CaCl2, 1.25 NaH2PO4, 25 D-glucose, 1 Na-Ascorbate and 3 Na-Pyruvate. All reagents were commercially obtained from Sigma-Aldrich Israel, except where stated. The 300-mm-thick coronal brain slices were cut with a Campden-1000 Vibrotome using the same cutting solution. The slices were allowed to recover for 30 min at 37˚C in artificial CSF (ACSF) containing the following (in mM): 125 NaCl, 2.5 KCl, 1.25 NaH2PO4, 25 NaHCO 3 , 25 D-glucose, 2 CaCl 2 , and 1 MgCl 2 , followed by additional recovery for at least 30 min in ACSF at room temperature until electrophysiological recording. The solutions were constantly gassed with carbogen (95% O 2 5% CO 2 ). Intracellular whole-cell recording After the recovery period, slices were placed in the recording chamber and maintained at 32-34˚C with continuous perfusion of carbogenated ACSF (2 ml/min). Brain slices containing the aIC were illuminated with infrared light and pyramidal cells were visualized under a differential interference contrast microscope with 10X or 40X water-immersion objectives mounted on a fixed-stage microscope (BX51-WI; Olympus). The image was displayed on a video monitor using a charge-coupled device (CCD) camera (Dage MTI). aIC-to-mPFC projecting neurons were identified by mCherry, and whole cell recordings from aIC to mPFC-projecting neurons were performed using an Axopatch 200B amplifier and digitized by Digidata 1440 (Molecular Devices). The recording electrode was pulled from a borosilicate glass pipette (3-5 MW) using an electrode puller (P-1000; Sutter Instruments) and filled with a K-gluconate-based internal solution (in mM): 130 K-gluconate, 5 KCl, 10 HEPES, 2.5 MgCl2, 0.6 EGTA, 4 Mg-ATP, 0.4 Na3GTP and 10 phosphocreatine (Na salt). The osmolarity was 290 mOsm, and pH was 7.3. The recording glass pipettes were patched onto the soma region of rAAV + pyramidal neurons. The recordings were made from the soma of aIC pyramidal cells, particularly from layer 2/3 and Layer 5/6. Liquid junction potential (10 mV) was not corrected online. All current clamp recordings were low-pass filtered at 10 kHz and sampled at 50 kHz. Series resistance was compensated and only series resistance <20 MW was included in the dataset. Pipette capacitance was~80% compensated. The method for measuring active intrinsic properties was based on a modified version of previous protocols (Kaphzan et al., 2013;Sharma et al., 2018). Recording parameters Resting membrane potential (RMP) was measured 10 s immediately after the beginning of whole cell recording (rupture of the membrane under the recording pipette). The dependence of firing rate on the injected current was obtained by injection of current steps (of 500 ms duration from +50 to 400 pA in 50 pA increments). Input resistance (Rin) was calculated from the voltage response to a hyperpolarizing current pulse (À150 pA). Sag ratio was calculated from voltage response À150 pA. The sag ratio during the hyperpolarizing steps calculated as [(1-DV SS / DV max ) x 100%] as previously reported (Song et al., 2015). Membrane time constant was determined using a single exponential fit first 100 ms of raising phase of cell response to 1 s À150 pA hyperpolarization step. For measurements of a single action potential (AP), after initial assessment of the current required to induce an AP at 15 ms from the start of the current injection with large steps (50 pA), a series of brief depolarizing currents were injected for 10 ms in steps of 10 pA increments. The first AP that appeared on the 5 ms time point was analyzed. A curve of dV/dt was created for that trace and the 30 V/s point in the rising slope of the AP was considered as threshold (Sharma et al., 2018). AP amplitude was measured from the equipotential point of the threshold to the spike peak, whereas AP duration was measured at the point of half-amplitude of the spike. The medium after-hyperpolarization (mAHP) was measured using prolonged (3 s), high-amplitude (3 nA) somatic current injections to initiate time-locked AP trains of 50 Hz frequency and duration (10-50 Hz, 1 or 3 s) in pyramidal cells. These AP trains generated prolonged (20 s) AHP, the amplitudes and integrals of which increased with the number of APs in the spike train. AHP was measured from the equipotential point of the threshold to the anti-peak of the same spike (Gulledge et al., 2013). Series resistance, Rin, and membrane capacitance were monitored during the entire experiment. Changes 30% in these parameters were criteria for exclusion of data. Results Novel taste experience increases the number of pERK + cells in mPFCprojecting neurons from inner layers of the aIC Activity in the IC is correlated and necessary for the formation of taste memory traces (Gal-Ben-Ari and Rosenblum, 2011). The IC displays high connectivity with other brain regions, and these reciprocal interactions facilitate integration of sensory information, valence and reward (Yiannakas and Rosenblum, 2017). Previously, we identified a correlation between an aversive taste valence and activation of aIC-to-BLA projecting neurons (Lavi et al., 2018). We also proved that activity of these projections is necessary for the acquisition and retrieval of an aversive taste memory, but not for attenuating a neophobic response to a novel, non-aversive taste (Kayyal et al., 2019). Given that mPFC activity is crucial for novel taste memory formation (Gonzalez et al., 2015;Uematsu et al., 2015;Mickley et al., 2007) and is correlated with learning novel experiences (Euston et al., 2012), we examined the role of aIC-mPFC reciprocal projections in novel taste learning. Toward that end, WT mice were injected with a retrograde virus containing an mCherry construct in the mPFC, labeling neurons in the IC that project to the mPFC ( Figure 1i). As ERK activation is necessary for taste learning (Elkobi et al., 2008), we used pERK, as an indicator of activated neurons (Sweatt, 2001;Adaikkan and Rosenblum, 2012). We quantified the number of pERK + neurons in 24 aIC slices (bregma 1.10 mm-0.26 mm) of these mice following novel (one exposure, n=3) or familiar (eight exposures, n=5) taste experiences involving the same taste, saccharin (see Materials and methods, Figure 1a,h). Quantification of pERK + cells across the aIC demonstrated novel saccharin consumption to be associated with an increased number of activated, pERK + neurons compared to familiar saccharin, in agreement with the literature (Berman et al., 1998) (unpaired t-test: p=0.0002, t=3.883, DF=114; Figure 1b). The aIC itself is comprised of three major subregions: the agranular (AIC), dysgranular (DIC), and granular IC (GIC) with the 4 th cortical layer persisting only in the dorsal portion of the GIC (Figure 7). Within these subregions, experiencing novel saccharin increased the number of pERK + neurons in the AIC (p=0.04, t=2.094, DF=67; Figure 1c) and the DIC (unpaired t-test: t=3.749, p=0.0003, DF=94; Figure 1d), but not the GIC (unpaired t-test: t=1.340, p=0.184, DF=92; Figure 1e) compared to familiar saccharin. Different cortical layers play different roles in sensory information processing (Dikecligil et al., 2020;Harris and Mrsic-Flogel, 2013). Hence, we examined the expression of pERK in the outer (1,2,3 combined) and inner (5,6 combined) layers. We found that both in outer (unpaired t-test: p=0.0002, t=3.897, DF=70; Figure 1f) and inner layers (unpaired t-test; p=0.002, t=3.121, DF=89; Figure 1g) of the entire aIC, there is a significant increase in pERK + neurons following novel saccharin experience, compared to the familiar saccharin group. Extensive reciprocal connectivity exists between the IC (mainly the AIC and DIC) and the mPFC (Gabbott et al., 2003), and this may be analogous to the salience network in humans (Uddin, 2015). We therefore hypothesized that aIC-to-mPFC projections have a vital role in novel taste behavior. We quantified pERK + neurons in the AIC and DIC that colocalized with retrograde-virus labeling, originating from the mPFC. We found a significant increases in the percentage of mPFC-projecting aIC neurons that are pERK + following consumption of a novel-compared to familiar taste (unpaired t-test: t=3.301, p=0.0015, DF=69; Figure 1j). This effect was evident at the AIC (unpaired t-test: t=3.513, p=0.0008, DF=69; Figure 1k) but not the DIC (unpaired t-test: t=1.353, p=0.181, DF=70; Figure 1l). Looking at the cortical layer distribution, we found that pERK induction in the aIC is observed in both the inner layers (unpaired t-test: t=3.050, p=0.003, DF=92; Figure 1n) and outer layers of the aIC (Mann Whitney test: p=0.0069, U=352; Figure 1m). To test for variability of labeling between novel/familiar saccharin groups, we perfromed an analysis of number of the aIC rAAV + neurons in its different subregions/layers and found no significant difference in number of rAAV + neurons between the groups (Figure 1-figure supplement 1a,b). Novel taste experience increases excitability in mPFC-projecting neurons of the inner layers of the aIC Following the correlation between experiencing a novel taste and increased activation of aIC-mPFC projecting neurons as measured by pERK (Figure 1), considering previous evidence regarding the role of pERK on neuronal intrinsic properties (Cohen-Matsliah et al., 2007), we tested the hypothesis that following the experience of a novel taste, aIC-to-mPFC projecting neurons will become more excitable. Toward that end, we once more injected a rAAV mCherry construct in the mPFC (Figure 2b), allowing the labeling of mPFC-projecting neurons of the IC. A month later, different groups of mice were exposed to novel (n=four mice) or familiar saccharin (n=four mice) and were sacrificed one hour later (see materials and methods, Figure 2a) while cage control mice were sacrificed without water restriction regime (n=4). Whole-cell patch-clamp recordings were then performed in slices from aIC-to-mPFC-projecting neurons (Figure 2b). In accordance with our hypothesis, novel taste consumption increased the excitability of mPFC-projections in the inner Table 1). Taken together, these results demonstrate that following the salient experience of novel taste consumption, deep layer IC-to-mPFC neurons are activated and display increased excitability, thus identifying part of a circuit involved in this form of learning. Novel taste experience increases activation of aIC-projecting neurons of the mPFC The mPFC is involved in novelty behavior and the connectivity between the IC and the mPFC is reciprocal (Gabbott et al., 2003;Gehrlach et al., 2020). Considering the present results, we sought to investigate the role of mPFC inputs to the aIC in novel taste learning. Hence, a separate group of mice were injected with the rAAV mCherry-expressing construct in the aIC, labeling neurons in the mPFC that project to the IC (Figure 3h,i). Mice were exposed to novel (first exposure, n=3, Figure 3a) and familiar (eight exposures, n=3) saccharin. As previously done in the IC, the general population of pERK + neurons were evaluated in the different mPFC subregions -prelimbic (PrL), infralimbic (InfrL), and cingulate cortex (Cg) from eight slices (Bregma 1.94 mm-1.54 mm). The whole mPFC (all subregions pooled together) showed an increase in activated, pERK + neurons following novel taste consumption compared to familiar taste (unpaired t-test: t=5.359, DF=14, p=0.0001; Figure 3b). Analysis of the different layers yielded significant effects both in the outer (t=5.835, DF=14, p<0.0001; Figure 3c) and inner (t=4.512, DF=14, p=0.0005; Figure 3d) layers of the mPFC. The changes in the total mPFC are derived from both the Cg1 (t=3.111, DF=14, p=0.0077; Figure 3e), and the PrL (t=4.306, DF=14, p=0.007; Figure 3f), but not the InfrL (t=0.6712, DF=14, p=0.513; Figure 3g). A significant increase in activated mPFC-to-aIC projecting neurons following (f) Number of pERK + neurons was significantly increased in the outer layers of the IC following novel (76.58±4.103) compared to familiar (60.75±2.016) saccharin consumption. (g) Number of pERK + neurons was significantly increased in the inner layers of the IC following novel (63.87±3.009) compared to familiar (52.76±1.868) saccharin consumption. (h) Representative coronal IC sections immunostained for pERK (green) and DAPI (blue) from mice injected with retroAAV at the mPFC (red) following novel (upper) and familiar (lower) saccharin. Scale bar, 50 mm, 20x. (i) Stereotaxic injection of rAAV-mCherry construct (red) at the mPFC and its subsequent labeling in the IC. Representative schematic overlays of the Cre-dependent expression of the chemogenetic receptors using the rAAV systems is shown, demonstrating the expressionto be restricted in the aIC and mPFC. Number of double-labeled (pERK + , rAAV + ) neurons was calculated as a percentage of all rAAV + neurons. (j) Percentage of double-labeled neurons of the IC was significantly higher following novel (9.517±1.337) compared to familiar (6.065 ± 0.313%) saccharin consumption. (k) Percentage of double-labeled neurons of the AIC was significantly higher following novel (11.36 ± 2.050%) compared to familiar (6.036±0.303) saccharin consumption. (l) Percentage of double-labeled neurons of the DIC was similar following novel (8.126 ± 2%) and familiar (5.956 ± 0.546%) saccharin consumption. (m) Percentage of double-labeled neurons of the outer layers of the IC was similar following novel (10.08 ± 2.079%) and familiar (5.02 ± 0.381%) saccharin consumption. (n) Percentage of double-labeled neurons of the inner layers of the IC was similar following novel (10.05 ± 1.089%) and familiar (6.562 ± 0.411%) saccharin consumption. Data are shown as mean ± SEM. p<0.05, p< 0.01, p<0.001, **p<0.0001. The online version of this article includes the following source data and figure supplement(s) for figure 1: Source data 1. Novel taste experience increases number of pERK cells in mPFC-projecting aIC neurons. novel compared to familiar taste was also identified (t=9.549, DF=14, p<0.0001; Figure 3j). The increase in activated aIC-projecting neurons of the mPFC is derived from both outer (t=5.504, DF=14, p<0.0001; Figure 3k), and inner layers (t=9.923, DF=14, p<0.0001; Figure 3i Animals were water restricted in the first day and administered with saccharin (familiar saccharin group) or water (novel saccharin group) in the following 6 days. Mice were given novel or familiar saccharin and sacrificed 1 hr later. Intrinsic properties were measured in the different layers of the aIC. (b) Whole-cell current-clamp recordings of aIC neurons projecting to mPFC (red). (c, d) The dependence of firing rate on current step magnitude is significantly higher in layers V/VI insular neurons projecting to mPFC following novel compared to familiar taste or cage controls (two-way repeated measurements ANOVA, n=13-14 cells per group; p=0.0009; F (2, 37) = 8.465). Error bars represent SEM; p<0.05, p< 0.01, p<0.001, **p<0.0001. The online version of this article includes the following source data and figure supplement(s) for figure 2: Source data 1. Novel taste exposure increases excitability of mPFC projecting neurons of the inner insular layers. supplement 1c,d). These results show that the reciprocal aIC-to-mPFC circuit is activated following novel taste consumption. Activation of aIC-to-mPFC projecting neurons is necessary for both novel taste neophobic responses and memory formation In many correlative experimental set-ups, it is difficult to dissociate between the cellular activity needed for expressing the behavior from that which is needed to acquire the information for a learning process. Based on the correlation yielded above (Figures 1-3), between novel taste consumption and the reciprocal activation of aIC to/from mPFC, we sought to identify whether activation of the aIC-to-mPFC part of the circuit is necessary for novel taste neophobic responses and/or familiarization learning. We therefore injected mice with retrograde-Cre virus in the mPFC, and Cre-dependent inhibitory DREADDs (hM4Di) in the aIC (Materials and methods, Figure 4a,b). Although we used a low concentration of CNO, in order to account for any nonspecific effects of CNO administration (Gogolla, 2017), we injected another group of animals with AAV constructs at the aIC and mPFC, resulting in the expression of mCherry, without DREADD receptors in aIC-to-mPFC projections (see Materials and methods). Animals expressing hM4Di in the aIC-to-mPFC projecting neurons were administered either CNO (n=11) or saline (n=10), 1 hr prior to a choice test of novel saccharin (first exposure) and water (Figure 4c). Choice tests between water and saccharin were then carried out over the next 4 days. We found that inhibition of aIC-to-mPFC projecting neurons not only impairs the neophobic response in the first day (unpaired t-test; p=0.0095, t=2.883, DF=19; Figure 4d), but also causes a measurable increase in the neophobic response on the next day (twoway ANOVA: F(1,175)=7.558, p=0.0075; Figure 4e and Figure 4-figure supplement 1a). In order to control for possible effects of CNO, additional group of mice that were injected with control virus that does not express hM4Di in aIC-to-mPFC projections, were treated with CNO (n=7) 1 hr prior to the same experimental design. Aversion toward the novel taste and its attenuation were not different from the saline treated group (Figure 4-figure supplement 1a), which excludes a possible effect of CNO itself. Our results indicate a necessity for the circuit in denoting the taste as novel, although it does not rule out its effect on other facets such as taste recognition, retrieval or aversion. 1-3 4-6 InfrL PrL PrL PrL Figure 3. Novel taste experience increases number of pERK + cells in aIC-projecting neurons of the mPFC. (a) Schematic representation of novel or familiar taste learning; animals were water restricted in the first day and administered with saccharin (familiar saccharin group) or water (novel saccharin group) in the following 6 days. In the last day, mice were presented with 1 ml of Saccharin 20 min prior to perfusion.number of pERK + neurons was quantified in the different subregions and layers of the mPFC. (b) Number of pERK + neurons of the mPFC was significantly increased following novel Figure 3 continued on next page CNO (n=11) or saline (n=11; unpaired t-test: t=0.4114, DF=20, p=0.6852; Figure 4-figure supplement 1d,e). Furthermore, we inhibited the aIC-to-mPFC part of the circuit prior to a choice test between innately aversive quinine and water (CNO=7, Saline=7), and found no significant difference in aversion between the two groups (unpaired t-test: t=1.102, DF=12, p=0.2919; Figure 4-figure supplement 1f,g). These results therefore indicate that the aIC-to-mPFC part of the circuit is not necessary for the retrieval of a familiar taste, taste aversion or taste recognition, but is vital for both the initial neophobic response, as well as incidental memory formation. We have previously shown that activity within aIC-to-BLA projecting neurons is not necessary for attenuation of neophobia (Kayyal et al., 2019). However, since activity within the BLA is necessary for the expression of a neophobic response (Shinohara and Yasoshima, 2019;Lin et al., 2018) and in order to better understand the broader circuit involved in the expression of a neophobic reaction, we sought to inhibit aIC neurons projecting to the BLA during the first encounter with novel saccharin. Mice were injected with retrograde-Cre AAV at the BLA, and Cre-dependent inhibitory DREADDs at the IC (see Materials and methods, Figure 5a-b). A month later, mice received an i.p. injection of CNO (n=8) or saline (n=6) one hour prior to a choice test of novel saccharin and water ( Figure 5c). Inhibition of aIC-to-BLA projecting neurons caused a reduction in neophobia (unpaired t-test; p=0.007, t=3.243, DF=12) compared to the control group (Figure 5d). However, the attenuation curve did not differ between the two treatments in accordance with our previous report (two way ANOVA: F(1,43) = 0.07789, p=0.7815; Figure 5e), indicating no change in novelty learning. Together, the results show that aIC-to-mPFC projections are necessary both for learning and expressing taste neophobia, while aIC-to-BLA projections are necessary only for the expression of the neophobic response, but not for its attenuation. Activity of mPFC-to-aIC projecting neurons is necessary for novel taste neophobic response but not memory formation To better understand the interplay of information between the mPFC and IC and the role the mPFC plays in novelty behavior, we sought to inhibit the directionally opposite element of the pathway, the mPFC-to-aIC projection. Mice were therefore injected with rAAV-Cre at the aIC, and Cre-dependent inhibitory DREADDs at the mPFC (see materials and methods, Figure 6a-b). We then injected mice with CNO (n=8) or Saline (n=5) one hour prior to a choice test between novel saccharin (first exposure) and water (Figure 6c). Inhibition of the mPFC-to-aIC circuit caused a reduction in aversion index during the first choice test (unpaired t-test: t=2.995, DF=11, p=0.0131; Figure 6d). However, Figure 3 continued (548.1±36.24) compared to familiar (329±18.89) saccharin consumption. (c) Number of pERK + neurons of the outer layers of the mPFC was significantly increased following novel (198±16.15) compared to familiar (95.75±6.920) saccharin consumption. (d) Number of pERK + neurons of the inner layers of the mPFC was significantly increased following novel (349±21.67) in comparison to familiar (233.4±14.040) saccharin consumption. (e) Number of pERK + neurons in the Cg1 was significantly higher following novel (143±17.71) compared to familiar (86.63±4.988) saccharin consumption. (f) Number of pERK + neurons in the PrL was significantly higher following novel (314±19.68) in comparison to familiar (203.6±16.44) saccharin consumption. (g) Number of pERK + neurons in the InfrL was similar following novel (90.23±6.423) and familiar (102.6±17.32) saccharin consumption. (h) Representative coronal mPFC sections immunostained for pERK (Green) and DAPI (blue) from mice injected with rAAV at the IC (Red) following novel (upper) and familiar (lower) saccharin. Scale bar, 50 mm, 20x. (i) Stereotaxic injection of rAAV-mCherry construct (red) at the IC and its subsequent labeling in the mPFC. Representative schematic overlays of the Cre-dependent expression of the chemogenetic receptors using the rAAV systems is shown, demonstrating the expressionto be restricted in the aIC and mPFC. Percentage of double labeled (pERK + ,rAAV + ) neurons was calculated as a percentage of all rAAV + neurons. (j) Percentage of double-labeled neurons of the mPFC was significantly increased following novel (10.44 ± 0.434%) compared to familiar (4.819 ± 0.397%) saccharin consumption. (k) Percentage of double-labeled neurons of the outer layers of the mPFC was significantly increased following novel (8.49 ± 0.230%) compared to familiar (4.23 ± 0.739%) saccharin consumption. (l) Percentage of double-labeled neurons of the inner layers of the mPFC was significantly increased following novel (11.81 ± 0.561%) compared to familiar (5.296 ± 0.340%) saccharin consumption. (m) Percentage of double-labeled neurons of the Cg1 was significantly increased following novel (8.543 ± 0.633%) compared to familiar (4.897±0.617) saccharin consumption. (n) Percentage of double-labeled neurons of the Prl was significantly increased following novel (12.23 ± 0.665%) in comparison to familiar (4.893 ± 0.358%) saccharin consumption. (o) Percentage of double-labeled neurons of the InfrL was significantly increased following novel (7.744 ± 0.554%) in comparison to familiar (3.261 ± 0.673%) saccharin consumption. Data are shown as mean ± SEM. p<0.05, p< 0.01, p<0.001, **p<0.0001. The online version of this article includes the following source data for figure 3: Figure 4. aIC-to-mPFC projecting neurons are necessary for both novel taste neophobic response and learning. WT male mice were injected with retrograde-Cre AAV at the mPFC (green) and the Cre-dependent DREADDs at the IC (red), resulting in expression of inhibitory hMD4Gi in aIC-to-mPFC neurons. (a) Representative images of retrograde-Cre AAV injection in the mPFC and the consequent labeling in the IC (red). (b) Representative images of neurons targeted for inhibition in the IC: Cre-dependent expression of the chemogenetic receptors (red), retrograde-Cre AAV (green) and DAPI (blue). (c) Schematic representation of behavioral test conducted. Mice injected with IC-to-mPC inhibitory DREADDs received intraperitoneal injection of CNO or Saline 1 hr prior to novel saccharin (first exposure) and water choice test, while aversionwas assessed. In the following days, mice were given a choice test without intervention. (d) CNO-injected mice exhibited significantly lower aversion (51.21 ± 8.255%) levels than do Saline-injected controls (78.36 ± 4.33%) in the first choice day (expression of neophobia). (e) CNO-injected mice were significantly more averse to saccharin (87.45 ± 4.336%, compared to saline (54.57 ± 5.708%)) injected controls in the second day. CNO-injected and saline-injected mice showed similar aversion in the third (CNO: 60.70 ± 8.599%, Saline: 54.93 ± 12.68%) fourth (CNO: 51.13 ± 7.804%, Saline: 27.29 ± 14.34%) and fifth (CNO: 45.77 ± 7.852%, Saline: 29.63 ± 6.974%) day of choice tests. Data are shown as mean ± SEM. p<0.05, p< 0.01, p<0.001, **p<0.0001. The online version of this article includes the following source data and figure supplement(s) for figure 4: Source data 1. aIC-to-mPFC projecting neurons are necessary for both novel taste neophobic response and learning. Figure 5. Activity within aIC-to-BLA projecting neurons is necessary for novel taste response but not for its attenuation. WT male mice were injected with retrograde-Cre AAV at the BLA (green) and Cre-dependent DREADDs at the IC (red), resulting in expression of inhibitory hMD4Gi in aIC-to-BLA projecting neurons. (a) Representative images of retrograde-Cre AAV injection in the BLA (green) and the consequent labeling in the IC. (b) Representative images of IC-to-BLA projecting neurons that are targeted for inhibition using Cre-dependent expression of the chemogenetic receptors (red), retrograde-Cre AAV (green) and DAPI (blue). (c) Schematic representation of behavioral test conducted. Mice injected with IC-to-BLA inhibitory DREADDs received intraperitoneal injection of CNO or Saline 1 hr prior to novel saccharin (first exposure) and water choice test, while aversion was assessed. In the following days, mice were given a choice test without intervention. (d) CNO-injected mice exhibited lower aversion (41.11 ± 9.309%) levels than do Saline-injected controls (79.60 ± 5.604%) in the first choice day t=3.243,DF=12). (e) CNO-injected and salineinjected mice showed similar aversion on the second (CNO: 36.26 ± 11.04%, Saline: 40.0 ± 12.98%), third (CNO: 16.06 ± 5.01%, Saline: 18.25 ± 8.920%), fourth (CNO: 26.13 ± 10.52%, Saline: 18.67 ± 12.48%), and fifth (CNO: 6.50 ± 4.153%, Saline: 0.00 ± 0.00%) day of expression of neophobia. Data are shown as mean ± SEM. p<0.05, p< 0.01, p<0.001, **p<0.0001. The online version of this article includes the following source data for figure 5: Source data 1. Activity within aIC-to-BLA projecting neurons is necessary for novel taste response but not for its attenuation. Figure 6. Activity of mPFC-to-aIC projecting neurons is necessary for novel taste neophobic reaction but not for the learning. WT male mice were injected with Cre-dependent DREADDs at the mPFC and rAAV-Cre at the IC, resulting in expression of inhibitory hMD4Gi in mPFC-to-aIC neurons. (a) Representative images of retrograde-Cre AAV injection in the IC (red) and the consequent labeling in the mPFC. (b) Representative images of mPFCto-aIC projecting neurons that are targeted for inhibition: Cre-dependent expression of the chemogenetic receptors (red), retrograde-Cre AAV (green) and DAPI (blue). (c) Schematic representation of behavioral test conducted. Mice injected with mPFC-to-IC inhibitory DREADDs received intraperitoneal injection of CNO or Saline 1 hr prior to novel saccharin (first exposure) and water choice test, while aversion was assessed. In the following days, mice were given a choice test without intervention. (d) CNO-injected mice (47.81 ± 11.58%) exhibited significantly lower aversion levels than do salineinjected controls (92.08 ± 2.111%) in the first choice day (expression of neophobia). (e) CNO-injected and saline-injected mice showed similar aversion on the second (CNO: 54.06 ± 13.71%, Saline: 66.84 ± 11.60%), third (CNO: 34.98 ± 11.44%, Saline: 54.76 ± 5.693%), fourth (CNO: 16.85 ± 5.986%, Saline: 34.54 ± 7.856%), and fifth (CNO: 23.80 ± 8.681%, Saline: 10.74 ± 4.712%) day of attenuation of neophobia. Data are shown as mean ± SEM. p<0.05, p< 0.01, p<0.001, ***p<0.0001. The online version of this article includes the following source data and figure supplement(s) for figure 6: Source data 1. Activity of mPFC-to-aIC projecting neurons is necessary for novel taste neophobic reaction but not for the learning. Figure supplement 1. mPFC-to-aIC projecting neurons are necessary for novel taste response but not for familiar taste taste retrieval. Figure supplement 1-source data 1. mPFC-to-aIC projecting neurons are necessary for novel taste response but not for familiar taste retrieval. it did not affect attenuation of neophobia in the following days (two way ANOVA: F(1,44) = 1.608, p=0.2114; Figure 6e, Figure 6-figure supplement 1a). In order to test whether this was due to an effect on taste identification, we performed another experiment in which we injected animals with CNO (n=4) or saline (n=5) prior to a choice test, conducted after mice were familiarized with saccharin (six exposures; Figure 6-figure supplement 1b). Inhibition of mPFC-to-aIC projecting neurons did not affect aversion after taste familiarization ( Figure 6-figure supplement 1c; DF=7,p=0.7716). This indicates that the activation of mPFC-to-aIC projecting neurons is necessary for the neophobic response to a novel taste, but not for the formation of a memory trace, as is additionally seen in the reciprocal aIC-to-mPFC part of the pathway. To further investigate whether aIC-mPFC inhibition affects anxiety levels, that may be associated with and/or affecting novel taste experiences, we performed an open-field test (Seibenhener and Wooten, 2015). Groups expressing hM4Di in aIC-to-mPFC (n=8) or mPFC-to-aIC projecting neurons (n=9), as well as their respective control virus injected groups (n=8), were all treated with i.p. injections of CNO. One hour later, mice were placed in an open arena ( Figure 6-figure supplement 2a). Time spent in the center of the arena and frequency of crossing the center were analyzed. No significant differences was found in both parameters, indicating no effect on anxiety as a consequence of inhibition of aIC-mPFC reciprocal connectivity ( Figure 6-figure supplement 2b,c). Aiming to clarify further whether the circuit manipulations affect the detection of the novel taste or anxiety associated with presentation of this taste, we performed another set of experiments, in which we tested the latency toward a novel saccharin and water presentation. Groups expressing hM4Di in aIC-to-mPFC (n=8) or mPFC-to-aIC projecting neurons (n=9) as well as their respective control virus injected groups (n=9) were all injected i.p with CNO. Mice were then presented with a pipette of novel saccharin and time to approach the pipette was recorded. Following 5 s of consumption, the saccharin pipette was removed and 10 s later a water pipette was presented. This routine of saccharin/water presentation was performed twice, consecutively (Figure 6-figure supplement 2d). No significant difference was found between the groups in the time it took the mice to start drinking, either during the first or second presentation of the novel saccharin or water ( Figure 6-figure supplement 2e), further indicating no effect on anxiety levels caused by our intervention. Discussion Detecting salient, new information from the environment is a crucial aspect of human and animal behavior. The detection and proper reaction to a novel stimulus, and the ability to form a stable memory of it, are two distinct facets of novelty behavior. The molecular and cellular mechanisms underlying these centrally important processes, which are critical for animal survival, have been the subject of much research (Mishkin and Murray, 1994;Schomaker and Meeter, 2015). As a result, numerous studies have explored the role of distinct brain areas and molecular or cellular processes within them that are crucial for salience and novelty detection and learning (Peters et al., 2016;van Kesteren et al., 2012). Although studies conducted in humans show a correlation between several brain structures and novelty detection, little is yet known of the detailed novelty network and directionality in the brain, resulting from causative research (Li et al., 2017). It is therefore important to understand how functionally relevant, distal brain regions act as a circuit to define salience and encode unfamiliar sensory information within a wider salience system. In the present study, we causally demonstrate that reciprocal activity between the aIC and mPFC, or the BLA, conveys novelty related gustatory information, in one such novelty circuit. Specifically, we show that general neuronal activation in the aIC, as denoted by pERK (Thiels and Klann, 2001), is increased following novel taste experience. This is in line with previous studies, showing that ERK phosphorylation in the aIC allows the formation of a memory trace of the safe taste, and is correlated with taste novelty (Sweatt, 2001;Gutkind, 1998;Kelleher et al., 2004;Thomas and Huganir, 2004;Yiannakas and Rosenblum, 2017). Importantly, our results indicate pERK is increased specifically in the AIC and the DIC subregions of the aIC, which have previously been implicated in processing chemosensory, somatosensory visceral and limbic functions (Jones et al., 2006;Katz et al., 2001;Yokota et al., 2011), emphasizing their relevance in novel taste learning. Extensive reciprocal connections exist between the IC and the mPFC, that are thought to be an essential component in salience processing both in humans and rodents (Uddin, 2015). Here too, our results show a correlation between the number of activated, pERK + neurons within reciprocal aIC-mPFC projections following novel taste experience, suggesting plasticity related changes in aIC-mPFC neurons are initiated following taste learning (Impey et al., 1999;Philips et al., 2013;Purcell et al., 2003;Rosenblum et al., 2002). Interestingly, in the aIC this was particularly apparent in the AIC, while in the mPFC the spread of pERK + neurons was more generalized, being detected across the PrL, InfrL and Cg. The projections of the aIC to the mPFC mainly originate from the AIC (compared to the DIC or GIC), which could be a possible explanation for the specific subregioneffect. Recent reports indicate the capability of PrL neurons to form ensemble codes for novel stimulus associations within minutes (Takehara-Nishiuchi et al., 2020). This ability seems to underlie the necessity of the mPFC to rapidly detect and selectively encode novel experiences, although our data suggest a role for all mPFC subregions and layers in novel taste processing and learning. Whether and how mPFC subregions, as well as the emerging local circuitry of the aIC are involved in novel taste learning via cell-and subregion-specific mechanisms, on a molecular and cellular level across different phases of learning, will be the subject of future research. We found that ERK activation within aIC-to-PFC projections correlated with taste novelty, and given ERK involvement in maintaining long-term memory-relevant excitability changes (Cohen-Matsliah et al., 2007), we assessed the intrinsic properties of these projections subsequent to novel taste consumption. In line with the pERK results, we found that aIC-to-mPFC neurons display distinct electrophysiological properties following novel taste consumption. This is in agreement with previous studies in other modalities, showing that novel stimuli and learning of associative experience differentially affect intrinsic properties of neurons in cortical and subcortical regions (Kaczorowski et al., 2012;Sehgal et al., 2013;Sehgal et al., 2014;Song et al., 2015;Song et al., 2015). The increased excitability of aIC-to-mPFC neurons following novel taste experience was observed only in the deep layer of the aIC. Limbic and gustatory information converge in the deep layers of the entire IC, and the main output source of the IC to other cortical and subcortical regions lies within these layers (Kobayashi, 2011;Krushel and van Der Kooy, 1988). In addition, the percentage of taste-coding neurons is higher in the deep layers of the aIC in comparison with the superficial layers (Dikecligil et al., 2020). Our results further strengthen the involvement of the deep layers of the aIC in processing and transmitting taste-related information to other cortical regions, suggesting that these specific molecular and intrinsic changes of aIC-to-mPFC neurons of the deep layers of the aIC allow the formation of a familiar and safe taste memory trace. Importantly, even though the correlated increase in pERK and in excitability of aIC-mPFC projections following a novel taste demonstrate the involvement of this circuit in novel taste processing, it does not differentiate between the two components of novel taste behavior: novel taste response, or novel taste learning. We thus evaluated if indeed there is a functional aIC-mPFC circuit, by using chemogenetic inhibition to causally examine the reciprocal role of both components, the aIC and the mPFC, in novel taste expression or learning. We found that inhibition of aIC-to-mPFC projecting neurons during novel taste exposure resulted in impaired neophobic responses and novel taste learning. In contrast, inhibiting mPFC-to-aIC projections impaired the expression of neophobia, but not taste memory formation. This indicates that the mPFC plays a major role in novelty gating, while the detection of novelty and storage of familiar taste memories are both mediated by the aIC. We show that aIC-to-mPFC projections need to be activated during novel taste exposure in order to form a safe taste memory trace, while aIC activation from the mPFC is necessary in order to evoke appropriate behavioral responses to novelty. We thus causally demonstrate that specific connectivity within the aIC and the mPFC allows the identification, the learning and reaction to novel taste stimuli with the memory being formed in the aIC. However, this does not exclude the involvement of other additional stimuli or other circuits that were not the focus of the current study. The entire IC is an integrative hub for saliency processing. We recently demonstrated that the reciprocal connectivity with the BLA underlies aversive valence processing (Kayyal et al., 2019;Lavi et al., 2018). The BLA is an essential structure for encoding salient stimuli ) and reward learning , and its signals to the aIC transmit palatability related information (Piette et al., 2012), while facilitating the acquisition and recall of sensory information related to body states, as well as the processing of valence (Gogolla, 2017;Haley and Maffei, 2018;Tye, 2018). Although we have previously shown that activity within aIC-to-BLA projecting neurons is not necessary for novel taste learning, we did not test if it is necessary for the expression of taste neophobia. Here, we show that activity within aIC-to-BLA projecting neurons is essential for the expression of neophobic responses. This goes along with previous data showing that inactivation of the BLA impairs the neophobic response but does not affect the process of attenuation of neophobia, and therefore learning (Lin et al., 2018;Shinohara and Yasoshima, 2019). Together, our findings suggest that aIC connectivity to different brain areas differentially conveys the novelty state (aIC-to-mPFC/BLA), while following learning-induced plasticity, a now modified circuit state underlies the encoding of the familiarized taste (aIC-to-mPFC but not to BLA) (Figure 7). A functional reciprocal connectivity exists between the BLA and the mPFC (Yizhar and Klavir, 2018), that is of great importance in relation to drinking behavior and water intake (Zhao et al., 2020). In addition, activity within BLA-mPFC projections is necessary for valence-specific behaviors (Yizhar and Klavir, 2018) including innate fear responses (Jhang et al., 2018). Hence, our results do not exclude the possibility of mPFC-BLA activity during novel taste experience and/or learning, but rather suggest a triage of IC, BLA and mPFC participation in novel taste behavior. Detecting, responding and remembering salient information is a vital property of animal behavior. We show here, for the first time, that the interplay between the aIC and the mPFC in the mammalian brain grasps within it novelty related data, as part of a salience network. Both aIC-mPFC and aIC-to-BLA reciprocal connections are necessary for exhibiting a neophobic response; however, the aIC-mPFC is necessary also to form a memory for a familiar safe taste. In addition, the acquisition and retrieval of learned associations between novel taste and aversive visceral information requires the activity of aIC-to-BLA projecting neurons (Kayyal et al., 2019;Lavi et al., 2018). Together, it places the insula as an integrator of two facets of taste saliency; the novel/familiar axis and aversive/appetitive axis. In addition to the necessity of aIC-mPFC interaction in taste novelty behavior and learning, a recent study reported a prominent role of mPFC-to-AIC in learning of novel olfactory tasks (Zhu et al., 2020). However, it is yet to be defined if and how the IC-mPFC circuitry encodes saliency for other modalities. There is a growing body of evidence showing that impairment in entire IC-mPFC circuitry is found in schizophrenia, addiction, and depression (Cisler et al., 2013;Penner et al., 2016;Samara et al., 2018), raising the importance for further investigation of this circuit and its dysfunction in patients. Defining the molecular and cellular mechanisms by which salience of sensory or innate information is processed and defined remains a key question in basic and clinical neuroscience. Our results suggest that while the mPFC plays a critical role in reacting to a novel cue, the aIC is essential both in reacting and in learning of the novel stimulus. This suggests that proposed non-invasive treatments for schizophrenia or addiction (such as transcranial magnetic stimulation therapy) could benefit from regiments that target the IC and/or other salience network components, while the integrity of these networks could be of particular pathognomonic value in the developing and adult brain.	2021-07-06T06:16:26.448Z	2021-07-05T00:00:00.000	{ "year": 2021, "sha1": "f4dafdbdabf7718fdb1497ba471f14b5f75ad582", "oa_license": "CCBY", "oa_url": "https://doi.org/10.7554/elife.66686", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "3751db93762abaa345c96043b5e10790cd6778da", "s2fieldsofstudy": [ "Biology", "Psychology" ], "extfieldsofstudy": [ "Medicine" ] }
228833006	pes2o/s2orc	v3-fos-license	An efficient method to produce segregating populations in quinoa ( Chenopodium quinoa ) Rapid global population growth together with the lack of availability of arable land and accessible water is imposing a challenge especially in many poor countries, to feed the population with sufficient and nutritious food. The pseudocereal quinoa, a member of the Amaranthaceae family, has gained increasing interest as an alternative staple food particularly in marginal lands, due to its high nutritional value and strong tolerance to abiotic stresses like drought, salinity, frost and heat (Jacobsen et al., 2003). Quinoa originated and has been cultivated along the Andes region of South America for the last 7,000 years (Williams & Brenner, 1995). Quinoa seeds have a high protein level (10 to 18%) with a perfect balance of amino acids specifically the essential amino acids (Vilche et al., 2003). They also provide a valuable combination of beneficial micronutrients like potassium, copper, zinc, iron and calcium along with fibre, lipids, carbohydrates and vitamins (Vega-Gálvez et al., 2010). Moreover quinoa is low in gluten, which offers a perfect substitute for wheat for people suffering from celiac disease. Despite of its exceptional characteristics, the seed yield of quinoa is generally low (around 1–2 t/ha), which makes breeding activities inevitable. Activities to breed varieties with higher seed yield Received: 2 June 2020 \| Revised: 10 July 2020 \| Accepted: 28 September 2020 DOI: 10.1111/pbr.12873 \| 1191 EMRANI Et Al. and good nutritional quality were driven from a small number of accessions in the Altiplano, which constitutes a very narrow genetic base for quinoa breeding programmes (Jacobsen & Mujica, 2002). Therefore, along with the conservation of landraces, which is essential for the preservation of the genetic material, efforts should be concentrated on the introduction of new germplasm into breeding programmes to increase genetic diversity. Quinoa is mainly an autogamous species with very small flowers, which particularly complicates crosses. Therefore, development of an efficient crossing method should serve as the first step in quinoa breeding programmes. Quinoa produces panicle inflorescences consisting of mostly hermaphrodite, but also pistillate flowers. The hermaphrodite flowers consist of five sepals and five stamens surrounding the ovary, and two stigmas (Abdelbar, 2018;Figure 1a). The hand emasculation of quinoa flowers has shown to be very difficult due to the floral morphology, the size of the flowers and the rapid progress of flowering within the inflorescence (Jacobsen & Stølen, 1993;Peterson et al., 2015). In many species, heat treatment and vacuum emasculation have been routinely applied as alternative methods for physical emasculation of the seed parents (Almeida et al., 2017;Mukasa et al., 2007;Otsuka et al., 2010;Sha, 2013). Warm water emasculation relies on the fact that pollen is generally more sensitive to higher temperatures than the ovary and the stigma. Therefore, selecting the appropriate temperature, at which the pollen will not be viable anymore, but the ovary and stigma are still active (usually around 45°C), is crucial for the application of this method (Sha, 2013). Almeida et al. (2017) reported the hyperthermotherapy in a water bath of 46°C for a panicle immersion time of 2.5 min to be more effective for the production of sterile plants compared to vacuum emasculation in rice. However, warm water emasculation has been reported as ineffective in quinoa, because it damages the inflorescence (Fleming & Galwey, 1995). Stetter et al. (2016) considered three different crossing methods (open pollination, warm water emasculation and hand emasculation) for creating inter-and intra-specific hybrids in three grain species of the genus Amaranth. Their results indicated hand emasculation and open pollination to be the most and the least efficient methods for creating hybrids, respectively. As an alternative to mechanical emasculation, male sterility systems can be considered as the method of choice, particularly for the production of hybrid varieties in a commercial level. Different sources for male sterility have been identified in quinoa germplasm (Simmonds, 1971;Ward & Johnson, 1994), including a cytoplasmic male sterility system, for which restorer genes can be found in many quinoa accessions (Ward, 1998). However, introducing this CMS system into parents to facilitate crossings is time-consuming. Moreover using the CMS system in quinoa breeding programmes does not seem to be feasible in the near future, since heterosis has not been reported yet and therefore commercial hybrid production seems elusive at the moment. Breeding inbred line varieties is still the method of choice as the floral morphology of quinoa rather promotes line development. Apart from the availability of efficient methods for crossing, strategies for the identification of F 1 plants are also required to produce a large array of F 2 populations. Two strategies for the identification of F 1 plants have been proposed using morphological and/ or molecular markers. In case of the former, parents should differ by easy to detect qualitative morphological traits with dominance/ recessive inheritance. In quinoa, seed colour, inflorescence color, axillary pigmentation, and plant colour could be used as morphological markers, if the pollinator is homozygous for the dominant alleles (Peterson et al., 2015). This approach would however limit F I G U R E 1 Floral morphology, crossing methods and selection of hybrid plants. (a) A hermaphrodite flower during pollination (stage BBCH60). The stamen consists of five anthers and the filaments form a ring around the ovary of the carpel. (b) Warm water emasculation crossing method. (c) Growing seed parent and pollinator side by side under a plastic bag ('no emasculation method'). (d) Axil pigmentation as morphological marker for selection of F 1 plants. Seed parents with green axil pigmentation were crossed with red-axil pigmented pollinators. The F 1 plants showed red axil pigmentation. Photos were taken at BBCH59 the number of crosses that can be performed, because only parents with contrasting phenotypes can be used. On the contrary, selection with molecular markers would allow crosses between any two parents, even if they show exactly the same morphological traits. Theoretically, one marker that is polymorphic between the two parents would be enough for the selection of true F 1 plants. In the current study, we compared three different crossing methods to produce F 1 seeds. We found that one polymorphic marker locus unequivocally distinguishes F 1 hybrid plants from selfing offspring. In combination with phenotypic selection, we propose a two-step selection procedure for the identification of true F 1 plants. Moreover we produced 30 different F 2 populations, which can serve as starting material in quinoa breeding programmes. \| Plant material and growth conditions To test different crossing methods and their efficiency in quinoa (Chenopodium quinoa Willd.), we considered 12 red-axil and 11 green-axil quinoa accessions from five different countries of origin for crossing (Table 1). Accessions with the dominant morphological trait (red axil) were considered as male parent and accessions with the recessive morphological trait (green axil) were considered as female parents for the crossing experiment (Table 1). We planted two plants per accession in 13 cm 2 pots in a cold greenhouse in May 2018 under natural long-day conditions in Kiel, Germany. To synchronize the flowering time of both crossing parents, we sowed the late-flowering accessions in the first week and then considered five different sowing dates in five consecutive weeks for the early-flowering accessions, based on the flowering time data of all the accessions from previous experiments (data not shown). We harvested the seeds on the seed parent and sowed 280 seeds per cross in 35x-multi trays in September 2018 in the cold greenhouse to identify the hybrid progenies. After leaf sampling for DNA isolation, we transferred the putative F 1 plants to 13 cm 2 pots for efficient seed production. All the true F 1 plants were bag-isolated to produce F 2 seeds. \| Crossing methods We considered three different crossing methods for the production of F 1 seeds (Figure 1b,c and Figure 2). \| Hand emasculation of the seed parent For this method, we followed the procedure suggested by Peterson et al. (2015) with slight modifications. To reduce the number of flower clusters to a manageable number, the flower bud of the seed parent was removed with scissors, once it was visible on the top of the plant (approximately 1.5 cm in size). We kept only 3-4 flower clusters on the seed parent. Then we opened the flowers under a magnifying lens (with 10x magnification) and removed all five anthers with a tweezer, when they were still green or yellowish-green ( Figure 2). \| Emasculation of the seed parent with warm water The inflorescence of the seed parents was dipped into a water bath of 45°C for 10 min (Figure 1b). This procedure was previously reported to be successful for the emasculation of the seed parents in grain amaranth (Stetter et al., 2016). \| No emasculation Here we did not perform any treatments for the emasculation of the seed parent and simply placed the seed parent and the pollinator together under an isolation bag (Figure 1c). \| DNA extraction We collected leaf samples from all the putative F 1 plants and lyophilized them before DNA isolation. We used the NucleoSpin® Plant II DNA isolation kit (Macherey-Nagel) or CTAB method with slight modifications (Saghai-Maroof et al., 1984) to extract the DNA from the dried leaf samples. \| Confirmation of crosses We conducted a two-stage selection for the identification of F 1 plants. First, we selected all the putative F 1 plants based on axil pigmentation. We phenotyped the seedlings for axil pigmentation four weeks after sowing. The putative F 1 plants are the ones that show red axil pigmentation, although their seeds have been collected on a green-axil seed parent ( Figure 1d). We expect that these plants are hybrid plants, since they show the dominant phenotype they have inherited from the pollinator. Additionally, we considered at least one green-axil plant from each cross as control for genotyping. As a second step, we used the publicly available Insertion-Deletion (InDel) markers described in Zhang et al. (2017) to confirm the genotype of the F 1 plants (Table S1). The parental lines of each cross were first screened with the InDel markers to find the polymorphic markers for each cross combination. These markers were then used for the identification of true F 1 plants from the putative ones ( Figure 3). Every PCR reaction with the InDel markers had a total volume of 20 µl with the following amplification conditions: 94°C for 5 min as initial denaturation, 35 cycles of: 30 s at 94°C, 30 s at primer pair annealing temperature and 30 s at 72°C and a final extension step of 10 min at 72°C. \| Success rate evaluation and statistical analysis For the calculation of success rates and selection accuracy and for their statistical analysis, we used the statistical software R (Team, R. C, 2019). We used appropriate generalized linear models (McCullagh & Nelder, 1989) with the parental accessions combination and crossing method as influence factors. No interaction effects were assumed. The residuals were assumed to follow a binomial distribution. For success rates, the distribution is based on the number of confirmed F 1 plants and the total number of the plants phenotyped for each cross. For selection accuracy, the distribution is based on the number of confirmed F 1 plants using molecular markers and the total number of putative F 1 plants for a given cross, which were selected using the phenotypic selection method. In this context and in order to enhance the model estimates, we used a Bayesian generalized linear model described by Gelman et al. (2008) for data analysis. Based on the model for the success rates, an analysis of variances was conducted, followed by multiple contrast tests (Bretz et al., 2011) in order to (a) identify the superior parental combinations and (b) compare the crossing methods. \| RE SULTS We produced in total 214 F 1 hybrids from 30 crosses between 23 parents using three crossing methods. All the hybrid plants were bag-isolated and selfed to produce 30 F 2 populations. The crossing parents were selected based on their axil pigmentation which had been determined before (Table 1). We considered three crossing methods (no emasculation, hand emasculation and warm water emasculation) and the parental combinations as sources of variation. Our results showed that both parental combinations and crossing methods have a significant effect on the success rate of the crossing experiment (ρ = <2.20E-16 and 2.32E-10, respectively). We used a two-step method for the identification of F 1 plants from all the crosses. As axil pigmentation has been reported to have a dominant monogenic inheritance (Simmonds, 1971), we expect all the F 1 plants from a cross between a green-axil seed parent and a red-axil pollinator to have red axils ( Figure 1d). Therefore, we only considered genotyping with the molecular markers for the red-axil plants together with a green-axil plant as control. For genotyping, we used six publicly available InDel markers (Zhang et al., 2017), which were polymorphic between the parents ( Not<del author="Nazgol Emrani" command="Delete" timestamp="1602617894146" title="Deleted by Nazgol Emrani on 10/13/2020, 9:38:14 PM" class="reU3">e</del>: For the identification of hybrid plants, genotyping with the published InDel markers was considered (Zhang et al., 2017). The size of the amplicons in base pairs is written in brackets for seed parent and pollinator, respectively. Success rates were calculated based on the number of confirmed F 1 plants and the total number of the plants phenotyped for each cross, while selection accuracy rate was calculated based on the number of confirmed F 1 plants using molecular markers and the total number of putative F 1 plants for a given cross, which were selected using axil pigmentation. comparable to the accuracy rate we reported for axil pigmentation, considering the low number of investigated crosses for hypocotyl colour. The comparison of different crossing methods revealed significant differences between all three methods ( produced the least number of F 1 plants with an average success rate of 0.06% (Table 2). To select the most successful parental combinations for the production of hybrids, we identified the crosses, which showed a higher success rate over all three crossing methods compared to the overall mean of the whole experiment. Five crossing combinations produced hybrid seeds with a significantly higher success rate compared to the mean success rate of the experiment (Table 4). We wanted to know if the phylogenetic relationship of the parents plays a role in the success rate of their cross. Therefore, we used 1.6 million SNPs derived from the whole genome re-sequencing of 20 out of the 23 investigated accessions in this study for phylogenetic analysis. The individual's dissimilarity and coancestry coefficient grouped the accessions in two main clusters (group I and group II; Figure 4). Nine out of the ten (90%) accessions in group I were originated from Peru or Bolivia, while all but one of the Chilean accessions investigated in this study were grouped in group II. Group I represents quinoa highland accessions in our experiment, while accessions in group II belong to the coastal quinoa population (Jarvis et al., 2017). We found crosses between more related accessions Note: For the statistical analysis, the logit estimates of the success rates were calculated. , , **: parental genotype combinations that show a higher success rate compared to the mean of all the crosses at α = 0.1, α = 0.05, α = 0.01 and α = 0.001, respectively. The ρ-values are based on appropriate multiple contrast tests (Bretz et al., 2011). it is crucial to develop a simple method for crossing and hybrid seed detection. Manual emasculation has been suggested as a promising method for production of hybrid seeds. However, hand emasculation is reported to be very difficult in quinoa due to its compact inflorescence and its tiny florets (Jacobsen & Stølen, 1993), while warm water emasculation has been considered ineffective due to the damage to the inflorescence (Fleming & Galwey, 1995). Despite of the challenges of the hand emasculation of quinoa flowers, attempts have been made to develop F 2 populations after hand crossing in quinoa. A detailed instruction for hand emasculation of quinoa flowers under controlled conditions is available (Peterson et al., 2015). Nonetheless, this protocol only describes the crossing procedure and does not report on further selection of F 1 plants using molecular markers. Here, we present a two-step selection strategy for the selection of true F 1 plants based on morphological and molecular markers. In this way, we created 30 segregating populations from crosses between closely related as well as distantly related quinoa accessions. \| D ISCUSS I ON We reasoned that two components (parental combination and crossing method) would account for the crossing success. Quinoa is an autogamous species with highly variable outcrossing rates (0.5 to 17%) across different accessions (Murphy et al., 2016). In a crossing experiment, parental accessions with a high outcrossing rate can be Quinoa accessions, which produce less cleistogamous and more chasmogamous flowers with exposed anthers and abundant pollen over a longer period of time, would be ideal candidates as pollinators in crossing programmes. Additionally, accessions with higher number of pistillate flowers would be more suitable as seed parents in crossing programmes. We would suggest phenotyping these traits in different quinoa accessions to identify putative candidates as crossing parents for breeding programmes in quinoa. Apart from mating type and floral morphology, pollinators with dominant qualitative morphological traits can facilitate the identification of F 1 plants. In the current experiment, we selected different parents from highland and coastal origin displaying high genetic diversity (Jarvis et al., 2017). To reduce the molecular marker screening workload, we considered axil pigmentation for the first selection step of the F 1 plants. It is important to mention, that the identification of red axil plants is not always easy. In some cases, axils of putative F 1 plants display a pink colour, which makes it hard to distinguish them from the green axils of selfing progenies. Moreover axil pigmentation may appear sporadically, which makes it difficult to detect (Peterson et al., 2015). Therefore, the selection accuracy due to axil pigmentation was highly variable and not suitable as a sole criterium to detect hybrids. There are other morphological traits like stem colour, inflorescence colour, saponin content or seed colour, which can be considered for the selection of F 1 plants. In our study, the parental combination had a significant effect on the success rate of the cross. We identified five parental combinations, which produced significantly higher numbers of hybrid plants. By comparing the most successful crosses, it is evident that the accessions 171,024 and 170,916 were present in more than one significantly successful cross as seed parent and pollinator, respectively. Among the five most successful crosses, two crosses were made between highland ecotypes, while one cross was made between the coastal accessions and two crosses were made between different ecotype groups. Therefore, considering the material explored in the current study, we did not see any associations between the phylogenetic relationship of the accessions and the success rate of the crosses. This suggests that crossing success in quinoa will be independent of the parental origin. It is noteworthy that hand emasculation has been successfully used to generate interspecific hybrids between C. quinoa and C. berlandieri, which opens new perspectives for breeding quinoa for disease resistances (Bonifacio, 2003). Hand emasculation was the most successful crossing method in the current study, followed by warm water emasculation and no emasculation, which was in line with the previous report in grain amaranth (Stetter et al., 2016). However, we recorded a lower success rate for our most efficient method (maximum success rate of 55.94% for hand emasculation) compared to the reported success rates in amaranth (74% for hand emasculation; Stetter et al., 2016). We believe, this difference is due to slight differences in the execution of the crossing methods. In the previous study in amaranth, the emasculation was repeated after seven days of the first emasculation and any flowers developed after the emasculation were removed. In our study due to technical reasons and the high number of crosses that we performed, repeated emasculation of the seed parents and removal of extra flowers were not technically feasible. Therefore, we expected a higher percentage of selfing seeds in the crossing progenies, which was a reason to consider 280 seeds per crossing event for identification of F 1 plants. In future experiments, we will consider multiple emasculation events and elimination of the intact flowers on the seed parents, to increase the success rate of the crosses. Nevertheless, it is evident that the hand emasculation significantly increases the success rate of the crosses and should be considered to assure the production of hybrid seeds. could significantly shorten the growth cycle of amaranth plants using the same strategy. A speed breeding protocol for shortening the growth cycle of quinoa is already available (Ghosh et al., 2018). Differences in However, this protocol is only designed for day-neutral/long-day quinoa accessions. Using the measures mentioned above, we expect to be able to shorten the growth cycle of short-day quinoa accessions, which encompass the majority of quinoa germplasm. We produced 30 segregating populations from crosses between genetically diverse quinoa accessions, which can serve as starting material for breeding programmes in quinoa. The improved progenies of the crosses can be selected and investigated in consecutive generations using the established breeding methods like pedigree or single seed descent. Moreover the F 2 populations and their progenies can be used for QTL mapping and identification of candidate genes for agronomically important traits in quinoa. ACK N OWLED G EM ENTS The authors thank Monika Bruisch for her support in conducting the experiment and performing the crosses in the greenhouse. The authors also acknowledge Verena Kowalewski and Brigitte Neidhardt-Olf for their support in the lab. Additionally, the authors also thank Prof. Mark Tester (King Abdullah University of Science and Technology, Saudi Arabia) for providing seeds for some of the parental accessions. CO N FLI C T O F I NTE R E S T The authors declare that there is no competing interest. AUTH O R CO NTR I B UTI O N S NE designed the experiment, performed the crosses and wrote the manuscript. MH performed the statistical analysis. DS calculated the phylogenetic relationship between the parental accessions. NM performed genotyping of the progenies using the molecular markers. ER produced the SNP file for the parental accessions. CJ supervised the study and critically revised the manuscript. All authors read and approved the manuscript.	2020-11-05T09:07:54.588Z	2020-11-03T00:00:00.000	{ "year": 2020, "sha1": "9e7693325f2d90cc1edb668ff3cd90d2656464fb", "oa_license": "CCBY", "oa_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/pbr.12873", "oa_status": "HYBRID", "pdf_src": "Adhoc", "pdf_hash": "feefee2af764a82ffa7110a8257a0d65529474f2", "s2fieldsofstudy": [ "Agricultural and Food Sciences" ], "extfieldsofstudy": [ "Biology" ] }
91983252	pes2o/s2orc	v3-fos-license	Protection of Grid Connected Photovoltaic Systems ( GCPVS ) Installations of photovoltaic systems connected or not to the electrical network have become increasingly popular, but it is often carried out by unqualified people using low quality components. The present study aims to describe the methodology adopted for the protection of grid connected photovoltaic systems (GCPVS) and the implications caused by their lack. Photovoltaic modules are typically installed in galvanized, painted, stainless aluminum or steel structures, which must be connected to a common ground, as these structures and any other components of the system could be energized by the photovoltaic array and may cause serious injuries or even death during routine maintenance, these fatalities can still result in ramifications for the entire industry, with millions of dollars in liability (for death or injury), negative publicity for GCPVS, and increased complications in obtaining licenses. Also, its components must be of good quality and the systems of protection must be well dimensioned and correctly installed, in order to avoid fire and electric damages, with reduction of possible damages in cases of short circuits. INTRODUCTION The installations of photovoltaic systems connected or not to the electrical network have become increasingly popular, but it is often carried out by unqualified people using low quality components.The present study aims to describe the methodology adopted for the protection of photovoltaic systems connected to the grid (GCPVS) and the implications caused by its lack.Photovoltaic modules are typically installed in galvanized, painted, stainless steel or aluminum structures, which must be connected to a common ground, since these structures and any other components of the system, which could be energized by the photovoltaic matrix, can cause serious injuries or even death during routine maintenance, these fatalities can still result in ramifications for the entire industry, with millions of dollars in compensation (for death or injury), negative publicity for GCPVS, and increased complications in obtaining licenses.Also its components must be of good quality and the systems of protection must be well dimensioned and correctly installed, in order to avoid fire and electrical damages, with reduction of possible damages in cases of short circuits.Therefore, the objective of this study is to address the main technologies used to protect SFVCR, based on the methods adopted in the United States, since in Brazil there are no current norms or regulations that address this issue 1 . GRID-CONNECTED PHOTOVOLTAIC SYSTEMS (GCPVS) A GCPVS consists basically of the photovoltaic panel, which converts the energy of the sun into electrical energy in direct current by an inverter, which converts the direct current to alternating current and, in addition to making it available for installation, also injects the surplus of that energy in the power distribution company´s network by the bidirectional meter, which accounts for the energy injected into the distributor´s grid. 2' 3 Figure 1 shows these components.In a photovoltaic system connected to the electricity grid (GCPVS), besides the protection system, the quality of the inverters must be analyzed, since the inverter must be disconnected from the grid in the event of a lack of electricity by the power distribution company, as well as having a sinusoidal wave of good quality, so as to avoid possible disturbances in the systems of protection of the distribution company. Grounding of GCPVS A photovoltaic system is considered grounded when one of the conductors of the DC circuit, positive or negative, is connected to the grounding system, which in turn is connected to ground. 7 ' 5 Photovoltaic modules are typically installed in galvanized, painted or stainless steel or aluminum structures.These structures and any other components, which could be energized by the photovoltaic matrix and which may be accessible during routine maintenance, should be connected to a common ground as shown in Table .2. The National Electric Code (NEC) does not provide guidance on how metal parts should be joined together to provide reliable grounding and it is beyond the scope of the Underwriters Laboratories (UL) 1703 standard for flat plate PV modules and panels.There have not been any industry guides on how to design, install, or maintain a reliable external electrical connection between metal parts for more than 20 years, leaving the installer with no option but to improvise solutions to this problem, resulting in a surprisingly large number of field systems that violate UL standards by failing to comply with the method prescribed by the module's grounding installation manual, often leaving the module structures ungrounded, or creating an electrical connection, with improvised methods, without testing , field validation, or certification to verify that the newly created connection would last the system´s lifespan.Such practices pose a safety risk to maintenance staff, leading to serious injury or death, and such event would also have ramifications for the entire industry, with millions of dollars in compensation (for death or injury), negative publicity for the photovoltaic system connected to the electricity grid (GCPVS), and increased complications in obtaining licenses. 5 Standard UL-1741 In an GCPVS system, one of the conductors of the DC circuit, positive or negative, must be connected to the grounding system, which in turn must be connected to ground.These grounding requirements require that each inverter has at least three terminals available and that they are connected together with the inverter housing, all on an appropriate bus in the inverter itself. 5rounding in a PV arrangement is usually done through the Ground Fault Protection Device (GFPD) inside the inverters connected to the network, so that this device plays the role of the system connection jumper.The presence of this connection jumper requires, in accordance with UL Standard 1741, that inverters have clearly marked terminals marked for the conductor connection of the grounding electrode, as shown below in Fig. 3.The UL 1741 also requires that the input DC and the output on the inverters have terminals, wiring, or other arrangements for accepting equipment grounding conductors (EGC).These grounding requirements require that each inverter has at least three terminals available and that they are electrically connected to each other and to the inverter housing, all on an appropriate bus in the inverter itself, as shown in Fig. 4. Other countries with standards and grounding standards different from the US, accept inverters without a grounding electrode terminal. Protection Devices The protective devices are available to be added to the isolated photovoltaic systems as well as connected to the mains.Some inverters already come equipped with these protection devices, which serve as junction and in the presence of fault current, they can open this junction.If these devices do not perform their junction function, they simply act by disconnecting the current (negative and positive) conductors from the circuit 7 .Incorrect or faulty installed device such as AC isolator switches being used erroneously in DC circuits, may result in a buildup of heat inside the switch compartment leading to fire as shown in Fig. 5. 8 Other incidents resulted from the use of defective inverters or absence of DC switches, and any switching faults or connection on the DC side of a system can result in the generation of a high temperature arc that can start fire, as shown in Fig. 6.DC arcs can be difficult to extinguish and pose a risk to firefighters. 8 CONCLUSION Photovoltaic systems are exposed to the effects of lightning and are short-circuited, resulting in high currents and consequently, increase the risk of fire and damage to electrical appliances.In order to reduce these risks, protective devices are installed in these systems, which are capable of detecting excessive short-circuit current or a lack of grounding, and are capable of interrupting it.The over current protection components, located on the DC conductor, constitute a current sensor that measures the current difference between the positive and negative conductors.In any photovoltaic arrangement, leakage currents are expected and they increase with the aging of systems, especially in rainy environments.In order for the system not to get unprotected unnecessarily, the over current limit of these devices, which varies according to the size of the system, should allow leakage currents up to certain values.NEC requires that protective devices be capable of identifying and disrupting ground fault currents, however, it says nothing about the necessary procedures to achieve them.Compared to Brazil, the solar energy industry in the United States has been used longer and, consequently, has increasingly improved standards.It can be concluded that for photovoltaic systems to work correctly and without risk of accidents or fire, it is necessary that in Brazil there is a regulation not only for the components of the system, but also for the professionals who will install and maintain the GCPVS.	2019-04-03T13:08:06.005Z	2018-10-29T00:00:00.000	{ "year": 2018, "sha1": "036c627ff5700aab67830e53adcdbc88b1dea55b", "oa_license": "CCBYNC", "oa_url": "http://www.scielo.br/j/babt/a/8sJPb3cDFkZCCksVts7tG8m/?format=pdf&lang=en", "oa_status": "GOLD", "pdf_src": "Anansi", "pdf_hash": "036c627ff5700aab67830e53adcdbc88b1dea55b", "s2fieldsofstudy": [ "Environmental Science", "Engineering" ], "extfieldsofstudy": [ "Computer Science" ] }
216111192	pes2o/s2orc	v3-fos-license	Proton Pump Inhibitor-Related Gastric Mucosal Changes Proton pump inhibitors (PPIs) are used worldwide to treat of acid-related disorders such as peptic ulcer and gastroesophageal reflux disease and to prevent gastroduodenal injuries due to nonsteroidal anti-inflammatory drugs. PPIs are the most potent inhibitors of gastric acid secretion currently available, and they are one of the most commonly prescribed classes of drugs because of their high efficacy and low toxicity. However, long-term PPI use causes histopathological changes such as parietal cell protrusion into the gland lumen, cystic dilation of gastric fundic glands, and foveolar epithelial hyperplasia. These changes can manifest on endoscopic examination as fundic gland polyps, hyperplastic polyps, multiple white and flat elevated lesions, cobblestone-like mucosa, or black spots. Clinicians must be aware of PPI-induced endoscopic features in patients with chronic long-term PPI use. Conversely, identifying patients with long-term PPI use based on their endoscopic findings is important. Recently, potassium-competitive acid blockers (P-CABs), a new class of acid suppressants that inhibit gastric acid secretion more strongly than PPIs, have recently been introduced clinically. Further long-term prospective studies on these gastric mucosal lesions in patients with either PPI or P-CAB use are required to investigate their association with histopathological changes and to establish the clinical significance of these findings. (Gut Liver 2021;15:-652) INTRODUCTION Proton pump inhibitors (PPIs), since their introduction in the late 1980s, have been used worldwide for the treatment of acid-related disorders such as gastroesophageal reflux disease (GERD) and peptic ulcer. They have also been administered prophylactically to prevent gastroduodenal mucosal damage due to nonsteroidal anti-inflammatory drugs (NSAIDs). PPIs are the most potent inhibitors of gastric acid secretion currently available. High efficacy and low toxicity of PPIs, combined with the high prevalence of GERD and NSAID use, has resulted in these becoming one of the most commonly prescribed agents. 1 GERD is a chronic condition, and the majority of affected patients experience a symptomatic relapse if PPI therapy is discontinued. Therefore, many GERD patients require continuous maintenance PPI therapy. 2 An important issue with long-term PPI usage is that these drugs raise the level of the peptide hormone gastrin; 3 as a result of the homoeostatic response of antral G cells, to the reduced acidity of gastric juice. Gastrin exerts trophic effects on the entire gastrointestinal tract tissue, including on both parietal and enterochromaffin-like cells distributed throughout the oxyntic mucosa. 4 It has been reported that long-term PPI use also induces histopathological changes such as protrusion of parietal cells into the gland lumen and cystic dilation of gastric fundic glands. [5][6][7][8] PPIinduced endoscopic features such as formation of fundic gland polyps, hyperplastic polyps, multiple white and flat elevated lesions, cobblestone-like mucosa, and black spots have been reported in recent studies (Table 1). [9][10][11][12] In the clinical setting, it is important to be aware of PPI-related gastric endoscopic features exhibited by long-term users. Conversely, it is also important to identify long-term PPI users based on their endoscopic findings. In this review of the latest reports on the topic, I have summarized the PPI-RELATED HISTOPHATOLOGICAL GASTRIC MUCOSAL CHANGES Chronic PPI use is associated with changes in the morphology of gastric parietal cells, including an increase in cell size and number and is associated with enterochromaffin-like cell proliferation in the gastric mucosa. 13 Histopathologically, parietal cell enlargement can easily be identified as a convex bulging and swelling of the apical membrane into the lumen of oxyntic glands. This so-called parietal cell protrusion (PCP) appears as a serrated contour of the internal glands (Fig. 1A). 7,13 Gastric hydrochloric acid accumulates in parietal cells due to inhibition of its active release from the secretory canaliculi by PPIs, which results in PCP. 13 Development of this aberrant cellular morphology shows a positive correlation with hypergastrinemia. 13 PCP is also observed in other conditions causing hypergastrinemia, such as Helicobacter pylori gastritis, peptic ulcer, or Zollinger-Ellison syndrome. 14 Therefore, PCP is not a finding unique to acid-suppressive therapy and might be associated with the increase of serum gastrin concentration, to some extent. 13 PCP further leads to cystic dilation of fundic glands and cytoplasmic vacuolation (Fig. 1B). 15 Cystic dilation of fundic glands may result from outflow-obstruction. This is attributable to glandular isthmus blockage due to protrusion of parietal cells and plugging by mucus secreted from proliferating foveolar cells, secondary to PPI-induced hypergastrinemia. 16,17 The resulting cystic dilation has the potential to further enlarge and progress to fundic gland polyp formation. 18 Previous studies have reported that fundic gland dilatation occurred between 8 months and 60 months after starting PPIs, 7,19 while a recent meta-analysis concluded that long-term use of PPIs (≥1 year) was associated with an increased risk of formation of fundic gland polyps. 18 Several studies have shown that fundic gland polyps are observed particularly in patients without accompanying H. pylori infection. 13,20 Enzymatic degradation of gastric mucus by H. pylori protease may facilitate glandular out- flow, thus protecting gastric glands against acid-retention and cystic dilation. 21 While β-catenin gene mutations were found in 64% to 91% of patients with sporadic fundic gland polyps, these mutations were not found in patients developing these polyps following PPI treatment. 22 In a histological analysis, PCP was found in 9.2% and 64.4% of fundic gland polyps in patients not taking PPIs and in those taking the drugs, respectively. The maximum diameter of fundic gland cysts within the fundic gland polyps was larger in patients on PPI treatment than in those not receiving the drugs. 22 A retrospective study revealed three types of PPI-associated gastric polyps in chronic users, (1) fundic gland type, (2) hyperplastic type, and (3) inflammatory type. 23 However, whether these polyps differ from those occurring sporadically in the general population, is still unclear. PPI-induced hypergastrinemia possibly also causes foveolar epithelial hyperplasia (Fig. 1C), which generally presents as hyperplastic polyps. 24 This may also present as multiple white and flat elevated lesions. 12,25 Therefore, hyperplastic polyps and multiple white and flat elevated lesions are considered as endoscopic features of foveolar epithelial hyperplasia, caused by chronic PPI consumption. Fundic gland polyps Fundic gland polyps are usually found in the gastric corpus and fundus, and they are generally multiple, small (<1 cm), sessile, and whitish-pink in color, with a smooth, partially translucent surface ( Fig. 2A). In 1992, Graham 26 first reported that fundic gland polyps arose during omeprazole treatment. In a prospective study of 191 longterm PPI users, fundic gland polyps were identified in 13.6% of patients, with the incidence in H. pylori-positive patients being significantly lower than in H. pylori-negative patients (hazard ratio, 0.29; 95% confidence interval [CI], 0.11 to 0.76); the majority of fundic gland polyps arose within 2 years of initiating treatment with PPIs. 24 Other studies on long-term PPI therapy have reported a 9% to 36% incidence of new, fundic gland polyp formation. 9,27 A recent meta-analysis of 12 studies also revealed PPI use to be significantly associated with an increased prevalence of fundic gland polyps (odds ratio [OR], 2.4; 95% CI, 1.42 to 4.27). 28 This association was observed to positively correlate with a longer time-interval of PPI exposure (for ≥6 months: OR, 4.71; 95% CI, 2.22 to 9.99 and for 12 months: OR, 5.32; 95% CI, 2.58 to 10.99). 28 In addition to the appearance of new fundic gland polyps, the size of existing polyps was often found to increase with PPI intake. Conversely, the fundic gland polyps were also observed to decrease in size or regress, following discontinuation of PPIs. There is no need to resect fundic gland polyps, as they have virtually no malignant potential in the general population. 29,30 However, a case-series report has documented a very rare, fundic gland type gastric adenocarcinoma. 31 A recent cross-sectional study showed that the occurrence of fundic gland polyps correlated with female sex (OR, 1.77; 95% CI, 1.31 to 2.39) and atrophic gastritis (OR, 0.12; 95% CI, 0.08 to 0.18). 12 Hyperplastic polyps Cases of patients with newly formed hyperplastic polyps due to long-term PPI use have been reported, especially in conditions associated with hypergastrinemia and concurrent H. pylori infection (Fig. 2B). 11,23,24 In a prospective study, hyperplastic polyps were identified in 8.9% of longterm PPI users. Development of new hyperplastic polyps was observed in 6% and 22% of patients with existing fundic gland polyps and hyperplastic polyps, respectively, over a 2-year period of PPI usage. 24 Furthermore, the incidence of hyperplastic polyps tended to be higher in H. pyloripositive patients than in H. pylori-negative patients. 24 Although the exact pathophysiological mechanism of hyperplastic polyp-formation is unclear, the gastrin receptor was reportedly expressed on the foveolar epithelium of hyperplastic polyps. 32 Gastrin induces proliferative effects on the gastric mucosa, which in turn enhances the effects of growth factors such as epidermal growth factor and tumor growth factor-α families, thereby promoting proliferation of crypt epithelial cells. 21 Therefore, PPI-induced hypergastrinemia is considered to cause hyperplasia of the gastric foveolar epithelium, which leads to the formation of hyperplastic polyps. Multiple white and flat elevated lesions Multiple white and flat elevated lesions are defined as apparently circumscribed and sharply demarcated white-colored, round, and slightly elevated mucosa with a smooth surface (Fig. 2C). 12,25 These lesions are usually found in the upper gastric corpus and fundus, and close observation allows identification of tubular structures on their surface (Fig. 2D). Dilated vessels usually observed in fundic gland polyps, are absent. 25 The rate of detection of these characteristic lesions can be increased by endoscopic examination or by using image-enhanced endoscopy such as narrow-band imaging. These lesions have been reported in 14.3% to 26.3% of PPI users. 10,33 In observational studies, multiple white and flat elevated lesions were found to be significantly associated with PPI use (OR, 3.58; 95% CI, 1.94 to 6.61), H. pylori eradication therapy (OR, 2.11; 95% CI, 1.08 to 4.11), female sex (OR, 1.92; 95% CI, 1.19 to 3.12), aging (OR, 1.05; 95% CI, 1.02 to 1.08), and moderate to severe atrophic gastritis (OR, 2.63; 95% CI, 1.52 to 5.63). 12,[33][34][35] These lesions are also related with administration of histamine receptor antagonists and H. pylori eradication, and are therefore not uniquely associated with chronic PPI use. Furthermore, these lesions can be misdiagnosed as intestinal metaplasia. Points of differentiation are as follows: (1) multiple white and flat elevated lesions appear in the upper corpus and fundus (intestinal metaplasia in the antrum often appears as whitish elevated patches, while intestinal metaplasia in the corpus is usually flat), and (2) multiple white and flat elevated lesions have a ridged or papillary surface structure, lacking the light-blue crest sign, on magnifying endoscopy with narrow-band imaging. 36 Cobblestone-like mucosa Cobblestone-like mucosa is defined as numerous, approximately 3 to 5 mm-sized, uneven, elevated, mucosal lesions in the gastric corpus (Fig. 2E). 12,25 Cobblestone-like mucosa has coloration similar to that of the surrounding mucosa, is usually observed in-between gastric folds, and has a fluffy appearance. The magnifying endoscopy with narrow-band imaging shows dilation of the oval crypt opening and the intervening part. 11 Several observational, case-control studies have shown that this cobblestone-like mucosa is significantly associated with PPI intake with a frequency of 9.1% to 35.1% in chronic users. 10,37,38 While chronic PPI users with cobblestone-like mucosa did not differ from those without cobblestone-like mucosa with respect to factors including age, sex, period of use, and H. pylori eradication, the characteristic mucosal appearance was observed more frequently in patients without atrophic gastritis. 38 In a cross-sectional study, multivariate analyses showed that the cobblestone-like mucosa was associated with PPI use (OR, 4.57; 95% CI, 2.34 to 9.96) and diabetes mellitus (OR, 3.41; 95% CI, 1.47 to 7.91). 12 A prospective, multicenter study also showed that the cobblestone-like mucosa was associated with PPI use (OR, 13.72; 95%, CI, 5.49 to 35.28), sex (female: OR, 0.32; 95% CI, 0.09 to 0.89), and age of the patients (≥50 years: OR, 4.46, 95% CI, 1.15 to 29.49). 10 Cobblestone-like mucosa has been histopathologically characterized by PCP and cystic dilatation of fundic glands, and these changes are especially prominent in nonatrophic gastric regions with abundant fundic glands. 39 When cobblestone-like mucosa is observed in regions with atrophic gastritis, the remaining gastric mucosal area with conserved fundic glands shows a characteristic, reddish-colored, elevated morphology, against a background of discolored, atrophic mucosa. Black spots Black spots are defined as black-colored pigmentation in gastric mucosa on conventional endoscopy, and are localized to the gastric body and fundus, which contain fundic glands. 40 Superficially, they look like blood clots, but the overlying mucosa is flat in appearance (Fig. 2F). They are usually multiple (approximately in 65% of patients), 40 and are sometimes observed inside fundic gland polyps ( Fig. 2A). Nonetheless, it is difficult to detect black spots on routine endoscopy without detailed examination. Observation studies have reported a frequency of black spots to be 0.2% to 6.2%. 10,12,40 On multivariate analyses, black spots were associated with PPI use (OR, 2.94; 95% CI, 1.66 to 5.21), H. pylori eradication therapy (OR, 3.01; 95% CI, 1.73 to 5.24), and a lower body mass index (OR, 0.89 per kg/m 2 ; 95% CI, 0.82 to 0.96). 12 The frequency of black spots was reported as 14% in the H. pylori eradication group or in PPI users, while it was <5% each in the group without H. pylori eradication and in PPI nonusers. 12 Histopathologically, black spots appear to comprise brownish substances and/or eosinophilic exudates within fundic gland cysts, which are thought to develop due to the accumulation of secretions from the lining cells of fundic gland cysts. 12 However, the composition of the contents of black spots and their mechanism of formation have yet to be elucidated. CONCLUSIONS In long-term PPI users, characteristic mucosal changes, such as fundic gland polyps, hyperplastic polyps, multiple white and flat elevated lesions, cobblestone-like mucosa, and black spots, can be observed on endoscopy. Histopathologically, while the development of fundic gland polyps, cobblestone-like mucosa, and black spots are considered to be caused by PCP and the cystic dilatation of fundic glands, the formation of hyperplastic polyps and multiple white and flat elevated lesions are attributed to foveolar epithelial hyperplasia (Fig. 3). With an ever-increasing number of PPI users, endoscopic identification of these mucosal changes is important. Recently, potassium competitive acid blockers (P-CABs), a new class of acid suppressants, which inhibit gastric H + , K + -ATPase activity via reversible and K + -competitive ionic binding to the enzyme and have stronger inhibitory effect on gastric acid secretion than PPIs, has recently been introduced in the clinical setting. Although an association between the aforementioned histopathological changes and P-CAB use has not yet been established, there is a possibility that long-term P-CAB users may also demonstrate similar gastric mucosal changes as observed in long-term PPI users. Further long-term, prospective studies examining these gastric mucosal lesions in both PPI and P-CAB users are required to investigate their association with histopathological changes and to establish the clinical significance of these changes. CONFLICTS OF INTEREST G.H.K. is an editorial board member of the journal but was not involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflicts of interest relevant to this article were reported.	2020-04-25T13:05:40.810Z	2020-04-27T00:00:00.000	{ "year": 2020, "sha1": "9a31e38efe6bd9a7b5768343e4a9ccefbd5d1aa0", "oa_license": "CCBYNC", "oa_url": "http://www.gutnliver.org/journal/download_pdf.php?doi=10.5009/gnl20036", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "6c2dce46ba0fb9e1dc84548698fd9870e838ac2a", "s2fieldsofstudy": [ "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
9043002	pes2o/s2orc	v3-fos-license	The Knee Joint Loose Body as a Source of Viable Autologous Human Chondrocytes Loose bodies are fragments of cartilage or bone present in the synovial fluid. In the present study we assessed if loose bodies could be used as a source of autologous human chondrocytes for experimental purposes. Histochemical examination of loose bodies and differential enzymatic digestions were undertaken, the isolated cells were cultured in alginate bead microspheres and immunolocalisations were undertaken for chondrogenic markers such as aggrecan, and type II collagen. Isolated loose body cells had high viability (≥90% viable), expressed chondrogenic markers (aggrecan, type II collagen) but no type I collagen. Loose bodies may be a useful source of autologous chondrocytes of high viability. Introduction Loose bodies (also known as rice bodies and knee joint mice) are fragments of cartilage or bone that float freely within the synovial fluid component of the knee joint. They may occur in single or multiple forms 1 but generally only affect a single knee. Loose bodies are classified as either stable or unstable. The former are located in fixed positions in the knee and are generally well tolerated and asymptomatic while the latter are free to move around the knee joint and may be the cause of pain, knee joint swelling, joint weakness and may cause the knee to lock abruptly. Loose bodies have a traumatic origin such as dislocation of the patella 2 or a complication arising from an orthopaedic procedure [3][4][5] and their occurrence is more likely in patients affected by osteoarthritis (OA) or rheumatoid arthritis (RA). Smooth bodies are classified as fibrinous, cartilaginous and osteocartilaginous. Fibrinous loose bodies result from intra-articular bleeding or by death of synovial tissue associated with tuberculosis, OA or RA. Cartilaginous loose bodies are caused by traumatic injury to the OA joint. Osteocartilaginous loose bodies are caused by fractures, osteochondritis dissecans, 6,7 inflammation, synovial chondromatosis 7 or tumours (osteochondromas). 8,9 Loose bodies are normally small, 9 but cases of giant loose bodies have also been reported. 8 In the present study histological examination of loose bodies showed to our surprise that they were highly cellular containing large numbers of viable chondrocytes and suggested that they may be a potential source of autologous human chondrocytes for repair strategies. We subsequently went on to isolate this chondrocyte population, they displayed high viability and were highly proliferative compared to chondrocytes isolated from residual tibial and femoral articular cartilage from the same knee joint replacements. The loose body chondrocytes expressed type II collagen, aggrecan and 5D4 KS in 3D alginate bead culture but no detectable type I collagen. Materials and Methods All chemicals and supplier details are as previously indicated described. 10,11 Monoclonal antibodies to aggrecan (clone 4D11 2A9), type I collagen (clone I-8H5), type II collagen (clone II-4CII), biotinylated anti-mouse IgG and antirabbit IgG secondary antibody, horseradish peroxidase conjugated streptavidin have been described previously. 10,12 Tissues and cells Loose bodies harvested from total knee replacement surgical discards from 6 patients (median age 56 years, 3 male, 3 female) of the Orthopaedics Clinic at North Shore Private Hospital, St. Leonards, were obtained with informed consent under ethical approval of the human ethics and care committee of the Royal North Shore Hospital who approved all procedures. Isolation and culture of chondrocytes from the loose bodies Chondrocytes were enzymatically isolated from loose bodies by sequential digestion with: i) 0.1% (w/v) pronase in DMEM-F12 media containing 10% (v/v) FCS for 2 h at 37°C; ii) 0.05% (w/v) clostridial collagenase in media for 4 h; and iii) overnight digestion with collagenase. 11 The cells were spun down (10 min x 800 g) and cell viability and numbers determined on a haemocytometer using trypan blue exclusion. Examination of the front and side scatter characteristics of these cells by flowcytometry was similar to results obtained earlier with only one major cell population evident. 11 The cells were established in monolayer culture in 75 cm 2 canted neck flasks at a density of 100,000 cells/mL in DMEM-F12 + 10% FCS + antibiotics under an atmosphere of 5% CO 2 and 98% humidity at 37°C. 11 After cellular attachment overnight, the flasks were rinsed in PBS to remove non-adherent cells and cultured in DMEM-F12 media + 10% FCS + antibiotics, with media changes every 3 days. The cells became confluent on day 5-6, the cells were sub-cultured up to passage 9. Preparation and culture of loose body cells in calcium alginate beads The loose body cells from one 75 cm 2 canted neck flask were detached using trypsin-EDTA and pelleted (800 g x 10 min) then washed in sterile DMEM-Hams Fl2 + 10% v/v FCS and resuspended in 2 mL of DMEM-Hams Fl2-FCS. Cell numbers and viability assays were measured on an aliquot using a haemocytometer and trypan blue exclusion. A known number of cells (95% viability) were spun down again and dispersed at a density of 3 x 10 6 cells/mL alginate in sterile isotonic saline. This mixture was loaded into a 2 mL syringe and extruded drop-wise through a 23 guage needle into an agitated solution of sterile CaCl 2 (102 mM) in a laminar flow hood to maintain sterility. After 10 min curing time the calcium alginate beads (~10 mL/30,000 cells) were established in culture in small petri dishes (100 beads/plate) in DMEM-F12 + 10% FCS + 50 µg/mL ascorbic acid (5 mL media/plate). The plates were incubated at 37°C, in an atmosphere of 5% C0 2 in air, with a humidity of 98% and the medium replenished every 48 h. The loose body cells were cultured up to 4 weeks and samples of beads collected after 2, 3 and 4 weeks of culture as indicated earlier. 11,13,14 The remaining beads were rinsed in isotonic saline and solubilized in 55 mM trisodium citrate in 150 mM NaCl (2 mL/50 beads) and the cells spun down at 800g x 10 min. The cells were then either cryopreserved in liquid nitrogen or were reestablished in monolayer culture. Histological processing of alginate beads and embedded in paraffin then sectioned at 4 µm thickness and attached to SuperFrost ultraPlus positively charged microscope slides. The slides were de-paraffinised in xylene (2 changes x 5 min), re-hydrated through graded ethanol washes (100-70% v/v) to water. Immunolocalisation of type I and type II collagen, aggrecan and keratan sulphate Endogenous peroxidase was blocked in bead samples with 0.3% (v/v) H 2 O 2 for 10 min then blocking undertaken in DAKO non-protein blocking agent. Aggrecan immunolocalisation were pre-digested with chondroitinase ABC (0.1 U/mL) in 50 mM Tris HCl pH 7.2 + 2% (w/v) BSA for 1 h, type I and II collagen immunolocalisations were pre-digested with proteinase K for 6 min and bovine testicular hyaluronidase (1000 U/ml) for 1 h at 37°C in phosphate buffer pH 5.0. The bead sections were incubated with antiaggrecan (1/10,000 dilution), anti-type I collagen (1/300 dilution), anti type II collagen (1/200 dilution) and MAb 5-D-4, anti-KS (1/1000 dilution) in TBS + 2% (w/v) BSA overnight at 4°C then biotinylated anti-mouse and anti-rabbit IgG antibodies and horse-radish peroxidase conjugated streptavidin were added using Nova RED substrate for visualisation. Negative control sections were run omitting primary Ab or using an irrelevant primary antibody. Both yielded negative results. The stained specimens were examined using a Leica photomicroscope linked to a DFC 480 digital camera using brightfield illumination. Results Loose bodies were observed in 12 of 18 total knee joint replacements, and typically 2-5 mm in size, smooth and glistening. Histological examination demonstrated a high cell density and abundant deposition of proteoglycan in the loose bodies ( Figure 1). The loose body contained a central necrotic core where the cells were arranged in clumps. Closer inspection within the loose body sections revealed a tran-Brief Report sition in cell morphology and tissue organisation from the surface zones into the transitional zone through to the cartilaginous and necrotic core (Figure 1 a,d). The surface zone region contained a population of flattened fusiform cells in a fibrous matrix overlying a mixed population of cells of a rounded chondrocyte like morphology (Figure 2a). This led into the transitional zone where larger chondrocyte-like cells were also evident ( Figure 2b). The cartilaginous zone contained closely packed large chondrocyte-like cells ( Figure 2c). The central core contained clumps of cells, many were dead or necrotic and many of the lacunae were unoccupied in the tissue sections examined (Figure 2d). Examination of surface zone 1 revealed stratification of the cells into three discernable zones, two surface zone regions containing cells of a flattened morphology but little proteoglycan (Figure 2f), and a cartilaginous zone underlying this containing a dense population of rounded chondrocyte like cells surrounded by metachromatically stained proteoglycan (Figure 2g). Surface zone 2 displayed similar traits to surface zone 1 with a surface fibrous region containing flattened cells overlying a cartilaginous zone containing larger chondrocytes in a proteoglycan rich matrix (Figure 2 h,i). Examination of the front and side scatter characteristics of the chondrocytes released from the loose bodies by flow-cytometry showed only one cell population (Figure 3a). A small proportion of dead/non-viable cells was also evident however these did not attach during the expansion of cell numbers by monolayer culture. Morphometric image analysis of bead sections immunolocalised for type I and II collagen and aggrecan demonstrated a steady increase of these components in the alginate beads over 2-4 weeks in culture (Figure 4 b,c). Type I collagen was not expressed by rounded chondrocyte-like loose body cells in alginate bead culture (Figure 4a) however it was produced by a few (<5%) cells of a flattened morphology at the bead surface after 4 weeks of bead culture. These expressed low levels of type I collagen (Figure 4e Discussion The present study arose from a series of failed attempts to isolate articular chondrocytes from the articular remnants of discarded surgical material from total knee arthroplasty patients. This was not an unexpected finding given that these specimens had little residual cartilage, displayed severe eburnation, loss of > 50% of the menisci and marginal osteophytic features of advanced knee-joint degeneration. The chondrocytes that were isolated typically were 50-60% viable, had poor replicative capability, and were obtained in insufficient numbers to warrant further investigation. Incidental observations on a number of these knee replacement tissues drew our attention to small glistening loose bodies present in a significant number of cases (12/18). On closer inspection histologically we were surprised to see that the loose bodies were highly cellular and contained abundant proteoglycan, the same could not be said of the residual articular cartilage of these joints. A necrotic core was a common finding particularly in the loose bodies ≥4 mm in size however this was to be expected given the advanced degenerative features of these knee specimens. We subsequently developed a protocol to isolate the loose body chondrocytes at ≥95% viability. The therapeutic use of autologous chondrocytes in isolation 15 or in combination with mesenchymal stem cells 16 in the development of cartilage repair strategies are of considerable interest in repair medicine. 17 The present study describes a simple convenient procedure for the isolation of these cells from loose bod-Brief Report cells (f, g, h, i). ies in the knee joint and could be of further application. We initially examined the cells isolated by flow cytometry, front and side scatter data demonstrated a single cell population with only a small proportion of dead or nonviable cells. Loose body cell numbers were initially expanded in monolayer culture which eliminated these dead cells and demonstrated a homogeneous cell population of cells with a typical cobblestone morphology and of high replicative capability and viability. Further culture of these cells encapsulated in alginate microspheres allowed us to demonstrate the synthesis of type II collagen, aggrecan, and KS but not of type I collagen by the loose body cells. Chondrocytes do not divide in this culture system but produce cartilage specific ECM components. The lack of type I collagen synthesis was further evidence of the chondrocytic pedigree of the isolated cell population. Furthermore morphometric image analysis of the immunolocalised bead sections over a 2-8 week culture period demonstrated a steady increase in chondrocyte ECM products plateauing at 4 weeks of culture. Our laboratory has formerly used this culture system with intervertebral disc cells, meniscal cells and articular chondrocytes. 11,13,14,[18][19][20] The microcarrier cell culture system was first introduced in 1967 by van Wezel. 21 These spheres were typically 125-250 μm in size and made from DEAE-dextran, bioglass, polystyrene, acrylamide, collagen or alginate and were available commercially as dextran beads (Cytodex, GE Healthcare), alginate (GEM, Global Cell Solutions), collagen (Cultispher, Percell) and polystyrene (Solohill Engineering). The surface chemistries of these microspheres were conducive to cell attachment and they were robust enough for use in spinner cultures. Cancer cells were one cell type which was cultured on and within these spheres in an effort to simulate a spheroid cell mass and promote cellular cross-talk to maintain cell viability and provide a culture system suitable for the assessment of anti-cancer agents in vitro. 22,23 In 1992 alginate microspheres were developed as an encapsulated system for the culture of cells of a chondrogenic background. 24,25 The high negative charge of the mannuronic and guluronic acid alginate copolymer was envisaged to reproduce the high fixed charge density 3D microenvironment chondrocytes experience in cartilaginous matrices. Chondrocytes rarely undergo division when surrounded by a mature 3D ECM however with maturity terminal differentiation can lead to hypertrophy. Neither of these features were displayed by the loose body chondrocytes in the alginate bead cultures examined in the present study (Figures 3 and 4) however these features were evident in histological sections of the loose bodies and typical of the columnar arrangement of growth plate chondrocytes (Figures 1 and 2) suggesting that the loose bodies may also be useful as models of the growth plate. The deer antler has been suggested as a growth plate model, mainly stemming from its impressive rate of growth however deer antler has a tissue organisation quite dissimilar to that of the long bone growth plates 26,27 whereas, as seen in the present study, loose bodies have similar cellular arrangements to these found in growth plate cartilage. Loose bodies can be stored in standard tissue culture media containing foetal calf serum and antibiotics to maintain cell viability for at least 3 weeks making them a potentially useful research tool and they could be considered as explants without cut edges. In the present study, the distribution of chondrocytes in alginate beads and absence of cell division were well illustrated in the negative control bead sections (Figure 4 h). Long term culture (8 months) of chondrocytes in alginate beads have been reported to lead to morphological changes in the outer bead cells with the appearance of cells of a flattened morphology similar to annular fibrochondrocytes in the intervertebral disc, whereas cells located more centrally within the bead have well Brief Report defined rounded chondrocytic morphologies. 28 A re-evaluation of the microsphere culture system has occurred in the last few years where cancer cells are encapsulated in an environment permissive to cell-cell cross-talk. Various matrix components can also be introduced into the bead in an effort to develop a more appropriate cell culture micro-environment similar to that found in vivo. 22,[29][30][31][32][33][34] This culture system has been used to culture macrophages and fibroblasts in breast cancer, 35 invasive hepatocellular cells, 32 and epithelial-stromal cells in prostate cancer. 36 Porous chitosan-alginate microspheres have also been developed to examine prostate cancer 37 and glioma. 38 Hepatocarcinoma spheroids have also been prepared using gelatin microspheres. 39 Pullulan 40 and controlled release rhBMP2 in 3D printed porous hydroxyapatite, 41 injectable nanofibrous microspheres, 41 and arginine-chitosan BMP-2 nanoparticle cell delivery vehicles for bone repair have also been developed. 32,33 These promising initial findings with the loose body chondrocytes warrants further studies to determine the gene expression profiles of the loose body chondrocytes compared to articular and growth plate chondrocytes. Human chondrocytes are difficult to source and it is only relatively recently that knee chondrocytes have become available commercially with most suppliers previously using hip cartilage as a tissue source. Pascual-Garrido et al. 42 found the viability of loose bodies from paediatric patients were 88% vs 92% for healthy articular cartilage. Others have also suggested that the loose body cells represent a valuable resource for autologous cell transplantation. 43 Thus, the loose bodies examined in the present study should be considered a valuable cell resource rather than as a surgical discard.	2017-10-14T18:27:12.032Z	2016-04-11T00:00:00.000	{ "year": 2016, "sha1": "9f9cc3369bfbea3b4bc6e26c302fb33f46d3b012", "oa_license": "CCBYNC", "oa_url": "https://www.ejh.it/index.php/ejh/article/download/2645/2577", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "333726c461884497092182fc1a0d4a60519385c8", "s2fieldsofstudy": [ "Medicine", "Biology" ], "extfieldsofstudy": [ "Biology", "Medicine" ] }
225082880	pes2o/s2orc	v3-fos-license	Identification of NPB, NPW and Their Receptor in the Rat Heart Members of neuropeptide B/W signaling system have been predominantly detected and mapped within the CNS. In the rat, this system includes neuropeptide B (NPB), neuropeptide W (NPW) and their specific receptor NPBWR1. This signaling system has a wide spectrum of functions including a role in modulation of inflammatory pain and neuroendocrine functions. Expression of NPB, NPW and NPBWR1 in separate heart compartments, dorsal root ganglia (DRG) and stellate ganglia was proven by RT-qPCR, Western blot (WB) and immunofluorescence. Presence of mRNA for all tested genes was detected within all heart compartments and ganglia. The presence of proteins preproNPB, preproNPW and NPBWR1 was confirmed in all the chambers of heart by WB. Expression of preproNPW and preproNPB was proven in cardiac ganglionic cells obtained by laser capture microdissection. In immunofluorescence analysis, NPB immunoreactivity was detected in nerve fibers, some nerve cell bodies and smooth muscle within heart and both ganglia. NPW immunoreactivity was present in the nerve cell bodies and nerve fibers of heart ganglia. Weak nonhomogenous staining of cardiomyocytes was present within heart ventricles. NPBWR1 immunoreactivity was detected on cardiomyocytes and some nerve fibers. We confirmed the presence of NPB/W signaling system in heart, DRG and stellate ganglia by proteomic and genomic analyses. Introduction Neuropeptide B (NPB) and neuropeptide W (NPW) are structurally and functionally related endogenous neuropeptides. NPB peptide is produced by proteolytic cleavage from a preproNPB molecule, and two NPB isoforms, NPB23 and NPB29 (consisting of 23 and 29 amino acids, respectively) can be produced in humans. However, only NPB29 has been described in nonhuman species. Similarly, two isoforms of NPW have been observed, NPW23 and NPW30, which are produced from a common precursor, preproNPW. Abundance of both NPW isoforms has been demonstrated in several species, including human, rat, mouse, pig and chicken. Biological activities of NPB and NPW are mediated by orphan G-protein-coupled receptors, GPR7 (also called NPBWR1) in humans, rats and mice and GPR8 (NPBWR2) in humans only. NPB and NPW together with these receptors constitute NPB/W signaling system [1]. Detection of mRNAs of the Members of NPB/W Signaling System by RT-qPCR Presence of mRNA for preproNPB, preproNPW and NPBWR1 was detected in all heart compartments, i.e., left atrium (LA), right atrium (RA), left ventricle (LV) and right ventricle (RV), as well as in DRG and stellate ganglia. However, very high Cq values oscillating in the range of 29-34 for all tested genes indicate their low level of expression. Statistical analysis of Cq values of studied genes within each separate heart compartment, DRG and stellate ganglia revealed that in the LA and RV, NPB and NPBWR1 mRNAs were expressed less than the gene for NPW. In RA and LV, NPBWR1 gene was expressed in a lesser extent than genes for NPB and NPW ( Figure 1A). In the DRG and stellate ganglion, the level of mRNA expression of genes for NPB and NPW did not differ significantly, but both were expressed more than mRNA for NPBWR1 ( Figure 1A). Furthermore, for better understanding, expression in LA, RA and LV was compared with expression in RV by assigning expression level of RV to 1 and setting expression in other heart chambers against it. Expression of mRNA for NPB was highest in RA and lowest in RV; mRNA expression for NPW was highest in RV and lowest in LA, mRNA expression for NPBWR1 was higher in the ventricles than in the atria ( Figure 1B). Generally, the differences in the expression of studied genes between separated heart chambers were relatively small, not even reaching statistical significance in some cases. Identification of Peptides/Proteins of NPB/W Signaling System by WB Good protein profiling of rat heart lysate was visible on SDS-PAGE. Similar amounts of proteins were transferred to nitrocellulose membrane and used for detection of NPB/W signaling system through Western blot (WB) analysis. Single immunogenic bands were detected for NPBWR1 and NPB at 50 and 25 kDa, respectively. Two immunogenic bands were observed for NPW at 35 and 26 kDa ( Figure 2). Immunogenic bands of NPBWR1, preproNPB and preproNPW were identified from different heart chambers such as LA (n = 3), RA (n = 4), LV (n = 3) and RV (n = 3) and ganglia such as DRG (pool of five) and stellate ganglia (pool of five). Immunogenic bands of NPBWR1 and preproNPB were present in all the chambers of rat heart (Table 1. An immunogenic band of preproNPW at 35 kDa was present in all the chambers of rat heart except LA. On the contrary, the presence of an immunogenic band of preproNPW at 26 kDa was observed in one out of four and one out of three rats in RA and LV, respectively. Additionally, a strong immunogenic band of NPBWR1 was detected Figure 1. Quantitative RT-PCR analyses of neuropeptide B (NPB), neuropeptide W (NPW), and NPBWR1 mRNAs in heart chambers of rats. (A) Comparison of expression of studied genes within each heart compartment, dorsal root ganglia (DRG) and stellate ganglia. Data are presented as ΔCq values (compared to β-actin) to indicate differences in expression between different targets. Hence, low values represent high expression. Statistically significant differences are shown graphically (Mann-Whitney test; n = 5-8 in each group). (B) Comparison of separate heart compartments: left atrium (LA), right atrium (RA), left ventricle (LV) and right ventricle (RV). Data are presented as relative expression ± SEM. Mean values from RV were used as a calibrator and were settled as 1. Statistically significant differences between the heart compartments are shown graphically (Mann-Whitney test; n = 5-8 in each group). # p < 0.05, + p < 0.01, § p < 0.005, * p < 0.001. Identification of Peptides/Proteins of NPB/W Signaling System by WB Good protein profiling of rat heart lysate was visible on SDS-PAGE. Similar amounts of proteins were transferred to nitrocellulose membrane and used for detection of NPB/W signaling system through Western blot (WB) analysis. Single immunogenic bands were detected for NPBWR1 and NPB at 50 and 25 kDa, respectively. Two immunogenic bands were observed for NPW at 35 and 26 kDa ( Figure 2). Immunogenic bands of NPBWR1, preproNPB and preproNPW were identified from different heart chambers such as LA (n = 3), RA (n = 4), LV (n = 3) and RV (n = 3) and ganglia such as DRG (pool of five) and stellate ganglia (pool of five). Immunogenic bands of NPBWR1 and preproNPB were present in all the chambers of rat heart (Table1. An immunogenic band of preproNPW at 35 kDa was present in all the chambers of rat heart except LA. On the contrary, the presence of an immunogenic band of preproNPW at 26 kDa was observed in one out of four and one out of three rats in RA and LV, respectively. Additionally, a strong immunogenic band of NPBWR1 was detected for DRG and stellate ganglia ( Figure 3); however, a very weak band for preproNPB and no band for preproNPW was observed in the same tissues (Table 1). Figure 1. Quantitative RT-PCR analyses of neuropeptide B (NPB), neuropeptide W (NPW), and NPBWR1 mRNAs in heart chambers of rats. (A) Comparison of expression of studied genes within each heart compartment, dorsal root ganglia (DRG) and stellate ganglia. Data are presented as ∆Cq values (compared to β-actin) to indicate differences in expression between different targets. Hence, low values represent high expression. Statistically significant differences are shown graphically (Mann-Whitney test; n = 5-8 in each group). (B) Comparison of separate heart compartments: left atrium (LA), right atrium (RA), left ventricle (LV) and right ventricle (RV). Data are presented as relative expression ± SEM. Mean values from RV were used as a calibrator and were settled as 1. Statistically significant differences between the heart compartments are shown graphically (Mann-Whitney test; n = 5-8 in each group). # p < 0.05, + p < 0.01, § p < 0.005, * p < 0.001. To validate the commercial antibodies, we performed a pre-absorbance assay. Highly immunogenic peptide sequences were selected and synthesized for pre-absorption assay. We used identical peptide sequences (species-specific) that were used as an immunogen to raise the respective antibodies. Moreover, we also observed 93% sequence homology in selected NPB peptide mouse versus rat, 97% sequence homology in selected NPW peptide human versus rat and 79% sequence homology in selected NPBWR1 human versus NPBWR1 rat ( Table 2). All the antibodies used in the experiment were cross-reactive to rat species. Therefore, we synthesized the peptide sequence of NPW and NPBWR1 from human and NPB from mouse, not from rat. Immunoreactive proteins identified on WB were further examined by using the pre-adsorption assay. Overnight pre-adsorption of the primary antibody with the corresponding synthetic peptide prevented the recognition of NPBWR1 protein in WB, but not completely ( Figure 4A). Due to the lower reactivity/signal ratio of anti-NPB towards heart lysate and high background issued with 1:200, 50 µg of protein was used with anti-NPB (1:400). Overnight pre-absorption of the primary antibody with the corresponding synthetic peptide prevented the recognition of preproNPB protein in WB ( Figure 4B). Pre-adsorption Assay using the Synthetic Peptides To validate the commercial antibodies, we performed a pre-absorbance assay. Highly immunogenic peptide sequences were selected and synthesized for pre-absorption assay. We used identical peptide sequences (species-specific) that were used as an immunogen to raise the respective antibodies. Moreover, we also observed 93% sequence homology in selected NPB peptide mouse versus rat, 97% sequence homology in selected NPW peptide human versus rat and 79% sequence homology in selected NPBWR1 human versus NPBWR1 rat ( Table 2). All the antibodies used in the experiment were cross-reactive to rat species. Therefore, we synthesized the peptide sequence of NPW and NPBWR1 from human and NPB from mouse, not from rat. Immunoreactive proteins identified on WB were further examined by using the pre-adsorption assay. Overnight pre-adsorption of the primary antibody with the corresponding synthetic peptide prevented the recognition of NPBWR1 protein in WB, but not completely ( Figure 4A). Due to the lower reactivity/signal ratio of anti-NPB towards heart lysate and high background issued with 1:200, 50 µg of protein was used with anti-NPB (1:400). Overnight pre-absorption of the primary antibody with the corresponding synthetic peptide prevented the recognition of preproNPB protein in WB ( Figure 4B). Overnight pre-adsorption of the primary antibody with the corresponding synthetic peptide prevented the recognition of preproNPW protein at 35 and 26 kDa. The band observed at 26 kDa showed the best depletion among all the peptides used. However, the band at 35 kDa was only partially depleted ( Figure 4C). The percentage of inhibition was calculated by using the following formula (1): Immunofluorescence Analysis of NPB, NPW and NPBWR1 Localization Specific NPB immunoreactivity (IR) was detected within the heart, mainly in nerve fibers. These NPB immunoreactive nerve fibers were more numerous within the atria, where they were localized in between atrial cardiomyocytes, in the vicinity of nerve cell bodies within heart ganglia and within nerve fiber bundles (Figure 5a-c). Additionally, some but not all nerve cell bodies within heart ganglia exerted specific NPB-IR (Figure 5c). Smooth muscle cells of some coronary arteries exerted weak reactivity with NPB antibody, while cardiomyocytes showed no specific NPB-IR (Figure 5d). Some but not all nerve cells bodies and nerve fibers showed NPB-IR within DRG (Figure 5e). In the stellate ganglia, only few nerve fibers within nerve bundles exerted NPB-IR and few nerve cell bodies showed weak IR (Figure 5f). Within heart atria, NPW-IR was visible in the nerve cell bodies as well as in the nerve fibers of heart ganglia ( Figure. Overnight pre-adsorption of the primary antibody with the corresponding synthetic peptide prevented the recognition of preproNPW protein at 35 and 26 kDa. The band observed at 26 kDa showed the best depletion among all the peptides used. However, the band at 35 kDa was only partially depleted ( Figure 4C). The percentage of inhibition was calculated by using the following Formula (1): % Inhibition = WB signal without peptide − WB signal with peptide WB signal without peptide × 100 (1) More than 90% inhibition of anti-NPW at 26 kDa, 31.6% inhibition of anti-NPW at 35 kDa, 40.1% inhibition of anti-NPB at 25 kDa and 58.7% inhibition of anti-NPBWR1 at 50 kDa were observed (Supplementary Figure S1). Immunofluorescence Analysis of NPB, NPW and NPBWR1 Localization Specific NPB immunoreactivity (IR) was detected within the heart, mainly in nerve fibers. These NPB immunoreactive nerve fibers were more numerous within the atria, where they were localized in between atrial cardiomyocytes, in the vicinity of nerve cell bodies within heart ganglia and within nerve fiber bundles (Figure 5a-c). Additionally, some but not all nerve cell bodies within heart ganglia exerted specific NPB-IR (Figure 5c). Smooth muscle cells of some coronary arteries exerted weak reactivity with NPB antibody, while cardiomyocytes showed no specific NPB-IR (Figure 5d). Some but not all nerve cells bodies and nerve fibers showed NPB-IR within DRG (Figure 5e). In the stellate ganglia, only few nerve fibers within nerve bundles exerted NPB-IR and few nerve cell bodies showed weak IR (Figure 5f). Specific NPBWR1-IR was visible on the surface of atrial as well as ventricular cardiomyocytes (Figure 5m,n), although the intensity of immunofluorescence was not homogenous throughout the heart. Additionally, some nerve fibers within heart atria showed specific reaction with NPBWR1 antisera (Figure 5i). No specific labeling was observed in smooth muscle cells of coronary circulation (Figure 5n). No specific reaction with NPBWR1 antisera was detected in DRG (Figure 5o), and only NPBWR1-IR nerve fibers were detected in stellate ganglia (Figure 5p). and their receptor 1 (NPBWR1; l-p) in the rat heart compartments, dorsal root ganglia (DRG) and stellate ganglia. Antiserum against NPB exerts specific immunoreactivity (IR) in nerve fibers (some of them marked by arrows), cardiac ganglionic cells within heart atria (arrows) and smooth muscle cells of coronary vessels (arrow). Some neurons within DRG exert specific NPB immunoreactivity (IR; arrow), while some of them do not (arrowheads). Neurons within stellate ganglion do not show any specific reaction with NPB antiserum (arrowheads). Within the heart, NPW-IR is visible in some nerve fibers and bodies of intracardiac neurons (arrows). Neurons of DRG and stellate ganglion do not exert specific IR with NPW antiserum (arrowheads). NPBWR1 antiserum shows IR in cardiomyocyte cell membranes (some marked by arrows), some (arrow) but not all (arrowheads) neurons within stellate ganglia and some nerve fibers within stellate ganglia (arrows). No specific IR is present within the DRG. n = 3 of each tissue. LCM Analysis of NPB and NPW mRNA Expression within the Heart Expression of mRNAs for NPB and NPW was detected in some individual cell types of heart tissue, nerve cell bodies of intracardiac nervous system (GCs), cardiomyocytes (CMs) and smooth muscle cells (SMCs). These cells were isolated by means of laser capture microdissection (LCM) and acquired samples were analyzed by RT-qPCR. In total, 19 GC, 4 CM and 7 SMC samples were prepared and analyzed. Expression of NPB mRNA was detected in 8 GC, 1 SMC and 0 CM samples. NPW mRNA expression was detected only in 1 GC sample and in no samples of other tested cell types. Figure 5. Immunofluorescence for neuropeptide B (NPB; a-f), neuropeptide W (NPW; g-k) and their receptor 1 (NPBWR1; l-p) in the rat heart compartments, dorsal root ganglia (DRG) and stellate ganglia. Antiserum against NPB exerts specific immunoreactivity (IR) in nerve fibers (some of them marked by arrows), cardiac ganglionic cells within heart atria (arrows) and smooth muscle cells of coronary vessels (arrow). Some neurons within DRG exert specific NPB immunoreactivity (IR; arrow), while some of them do not (arrowheads). Neurons within stellate ganglion do not show any specific reaction with NPB antiserum (arrowheads). Within the heart, NPW-IR is visible in some nerve fibers and bodies of intracardiac neurons (arrows). Neurons of DRG and stellate ganglion do not exert specific IR with NPW antiserum (arrowheads). NPBWR1 antiserum shows IR in cardiomyocyte cell membranes (some marked by arrows), some (arrow) but not all (arrowheads) neurons within stellate ganglia and some nerve fibers within stellate ganglia (arrows). No specific IR is present within the DRG. n = 3 of each tissue. Within heart atria, NPW-IR was visible in the nerve cell bodies as well as in the nerve fibers of heart ganglia (Figure 5g,h). Very weak nonhomogenous staining of cardiomyocytes was present within heart ventricles (Figure 5i). No specific reaction with NPW antisera was detected in DRG ( Figure 5j) and stellate ganglia (Figure 5k). Specific NPBWR1-IR was visible on the surface of atrial as well as ventricular cardiomyocytes (Figure 5m,n), although the intensity of immunofluorescence was not homogenous throughout the heart. Additionally, some nerve fibers within heart atria showed specific reaction with NPBWR1 antisera (Figure 5i). No specific labeling was observed in smooth muscle cells of coronary circulation (Figure 5n). No specific reaction with NPBWR1 antisera was detected in DRG (Figure 5o), and only NPBWR1-IR nerve fibers were detected in stellate ganglia (Figure 5p). LCM Analysis of NPB and NPW mRNA Expression within the Heart Expression of mRNAs for NPB and NPW was detected in some individual cell types of heart tissue, nerve cell bodies of intracardiac nervous system (GCs), cardiomyocytes (CMs) and smooth muscle cells (SMCs). These cells were isolated by means of laser capture microdissection (LCM) and acquired samples were analyzed by RT-qPCR. In total, 19 GC, 4 CM and 7 SMC samples were prepared and analyzed. Expression of NPB mRNA was detected in 8 GC, 1 SMC and 0 CM samples. NPW mRNA expression was detected only in 1 GC sample and in no samples of other tested cell types. Discussion According to our knowledge, this is the first study demonstrating the presence and tissue distribution of members of NPB/W signaling system within separate compartments of rat heart and ganglia innervating it at both genomic and proteomic levels. The presence and tissue distribution of individual members of the NPB/W signaling system were studied by several working groups in the central nervous system, and detailed information is available in the literature [1]. It has been suggested that the NPB/W signaling system is involved in several central regulations including regulation of neuroendocrine functions; regulation of feeding activity; autonomic regulation; and activation of stress axis, pain sensation and energy homeostasis. Additionally, it plays a role in emotions, anxiety and fear [12]. Despite the action throughout common receptor, NPB and NPW exert some differences in their biological activities [1]. Relatively little is known about the localization of this signaling system within the peripheral nervous system and/or other tissues. Expression of members of this signaling system at mRNA level has been more or less comprehensively studied in human, rat, pig and chicken peripheral tissues, confirming the presence of mRNA for individual members of the NPB/W signaling system, for example in kidneys and organs of respiratory system, urogenital system and gastrointestinal tract. The presence of mRNA for NPB was also demonstrated in whole-heart RNA in human, chicken and rat; for NPW in chicken and pig; and for NPBWR1 in pig only [2,10,11,13]. In this study, the heart was divided into individual chambers in order to obtain more precise information about expression and distribution of members of NPB/W signaling system, as cardiac tissue is not homogenous in terms of the content of different cell types in the individual cardiac compartments, especially with respect to the localization of the intracardiac nervous system. The results of the qPCR analysis show an overall very low level of expression of individual genes of the investigated signaling system. The level of NPB mRNA expression was higher in atria than in ventricles, suggesting predominant neuronal localization, as nerve cell bodies of the intracardiac nervous system are localized only within the atria in rats [14]. This assumption was confirmed by immunofluorescence analysis, which showed specific NPB and NPW-IR in the bodies of intracardiac ganglion neurons. This finding indicates the functioning of these peptides in the heart as neurotransmitters. However, compared to the expression of the most abundant neuropeptide transmitter of intracardiac neurons, NPY, the expression of these neuropeptides is about a thousand times lower [15]. By means of WB analysis, the presence of NPB and NPW peptides was proven in each heart chamber. Results of qPCR showed expression of NPB as well as NPW genes in the heart ventricles, which also indicates non-neuronal expression of these genes. However, the total amounts of NPB mRNA and protein detected in ventricles were low, just at the limit of detectability. The differences in NPB and NPW gene expression between the right and left ventricles cannot be explained by the different proportions of myocytes and fibroblasts, as the percentages of these cell types were found to be similar in both ventricles [16]. However, differences in the proportion of endothelial cells or smooth muscle cells could be considered, given that the results of our immunofluorescence analysis show the presence of NPB in smooth muscle cells and suggest that smooth muscle cells are probably the main non-neuronal source of NPB in the heart. The expression of NPW mRNA and peptide was also demonstrated in the heart ventricles, but the source is not entirely clear. Immunofluorescence results indicate weak expression in cardiomyocytes, but due to the overall very low expression of this gene in the ventricles, its presence cannot be ruled out in endothelial cells too, where such a low amount would not be detectable by immunofluorescence. The case of expression and detection of NPY in the heart is similar. The expression of NPY in the rat heart was also demonstrated in non-neuronal cells, specifically in cardiomyocytes and endothelial cells [17,18]. However, the presence of preproNPY in endothelial cells cannot be reliably detected by immunofluorescence [15,19]. This is probably due to the NPY content in these cells being lower than the immunofluorescence detection limit. RT-qPCR analysis of RNA from samples obtained by LCM was done in order to prove the expression of NPB and NPW in individual cell types. We have shown that some intracardiac neurons express mRNA for NPB and/or NPW. We also demonstrated the expression of NPB mRNA in vascular smooth muscle cells. Unfortunately, we were unable to demonstrate the expression of these genes in cardiomyocyte cells. This could be due either to the absence of expression of these genes in cardiomyocytes or, possibly, to their very low expression level, which cannot be detected in such a small sample. A limitation in the determination of the expression of the monitored genes in the sample of cardiomyocytes obtained by LCM is the impossibility of excluding the presence of endothelial cells and fibroblasts. In fact, in a tissue section stained with hematoxylin, it is not possible to distinguish between these cells and cardiomyocytes and remove them from the sample. Absence of WB signal in some protein samples could be also caused by low amount of studied proteins. Another source of NPB in the heart could be nerve fibers coming from the DRG, as the majority of DRG neurons expressing NPB are visible from immunofluorescence experiments. On the contrary, neurons within the stellate ganglion contain very small amounts of NPB, which leads to assumption that sympathetic fibers innervating the heart are not the source of NPB. NPW probably does not enter the heart via sympathetic or sensory nerves because we have failed to demonstrate the presence of this peptide in the ganglia innervating the heart. The regulation of NPBW signaling system is still only partially disclosed. Within the central nervous system, NPW neurons could be directly or indirectly stimulated by osmotic stimuli [20], stress [21] and/or nerve growth factor and brain derived neurotrophic factor [22]. Additionally, NPBWR1 mRNA expression is regulated by leptin, NPW and ghrelin in the neurons of nodose ganglia [23]. NPB is involved in sleep regulation [1]. So far, we can only speculate about the function of this signaling system in the heart, because very little information is currently available in the literature. It has been demonstrated that NPB/W signaling system is involved in the regulation of blood pressure both at the central and peripheral levels [1]. Our finding demonstrating the presence of NPB within smooth muscle cells of coronary circulation supports the idea that NPB could be involved in regulation of coronary circulation. In addition, both neuropeptides described in this study are present in the bodies of some intracardiac neurons, suggesting their possible involvement in the regulation of cardiac function. However, this hypothesis must be confirmed by a functional study. Finally, this report confirms the presence of neuropeptide W/B signaling system in all the chambers of rat heart as well as in DRG and stellate ganglia. However, the function of this signaling system in the heart is a matter for further investigation. (for immunofluorescence and/or for laser capture microdissection) or directly frozen in liquid nitrogen (for RNA and/or protein isolation). A similar procedure was used to prepare DRG, stellate ganglia and thoracic spinal cord (SC) samples. Samples were kept at −80 • C until use. Eight animals were used for RNA isolation and subsequent RT-qPCR, 5 animals for protein isolation with subsequent Western blotting and 3 animals for immunofluorescence and LCM. Extraction of Protein Tissues (n = 3-5 of each) were weighed and minced into small pieces. PBS was added in a total amount of 1 µL per 1 µg of tissue. Samples were homogenized for 3 min with rest for 1 min in ice per cycle. In total, three cycles were used. Homogenate was centrifuged at 10,000× g for 15 min and debris was removed. Supernatant was collected and protein content was estimated by Bradford dye. For WB analysis, 25 µg of protein was used. Selection of Neuropeptides NPB, NPW and NPBWR1 UniProtKB/Swiss-Prot protein database was used to select the peptide sequences. Accession numbers for NPB (Mouse), NPBWR1 (Human) and NPW (Human) were Q8K4P1, P48145 and Q8N729, respectively. Peptide sequences of mouse NPB , human NPBWR1 (220-250) and human NPW (33-62) were selected (Table 3). Basic Local Alignment Search Tool (BLAST) was used for comparing primary biological sequence information in order to find alignment of the sequences of different species. Moreover, availability of antibodies was also confirmed, where selected peptides were used as immunogens for respective antibody production with cross-reactivity with rat species. Selection of Neuropeptide Sequences for Peptide Synthesis A bioinformatics approach was used to identify the highly immunogenic sequence of NPB, NPW and NPBWR1. BLAST (Basic Local Alignment Search Tool) was used to find alignment of the sequence of different species in the protein database. Three sequences were selected: mouse NPB , human NPBWR1 (220-250) and human preproNPW (33-62). The selection of peptide sequences to be synthesized was done according to the availability of the corresponding antibodies on the market, while selected sequences should be used to produce the antibody used for experiments. Peptide Synthesis and Purification Highly immunogenic peptides, i.e., NPB , NPW (220-250) and NPBWR1 (33-62) were synthesized in solid phase using the automated microwave-assisted peptide synthesizer Liberty Blue (CEM Corporation, Matthews, NC, USA) following the standard protocols for Fmoc/tBu strategy. The peptides were purified by semipreparative RP-HPLC on a Waters instrument (Separation Module 2695, detector diode array 2996) using a Phenomenex (Torrance, CA, USA) Jupiter column C18 (10 µm, 250 × 10 mm) at 4 mL/min with solvent systems A (0.1% TFA in H 2 O) and B (0.1% TFA in CH 3 CN). The purity of the peptides was analyzed by analytical HPLC using a Waters ACQUITY HPLC coupled to a single-quadrupole ESI-MS (Waters 3100 Mass Detector) supplied with a BEH C18 (1.7 µm 2.1 × 50 mm) column at 35 • C at 0.6 mL/min with solvent systems A (0.1% TFA in H 2 O) and B (0.1% TFA in CH 3 CN). Data were acquired and processed using MassLynx software (Waters, Milford, MA, USA). The analytical data are reported in detail in Table 4. Laser Capture Microdissection (LCM) The whole procedure was done according to a protocol published earlier [24]. Briefly, sections of heart atria were stained with alum hematoxylin solution, washed with water and dehydrated with ethanol. Intracardiac neurons, vascular smooth muscle cells and cardiomyocytes were collected from the sections. Subsequently, RNA was isolated from obtained samples using RNeasy Micro Kits (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Real-time PCR was performed in the iCycler (Bio-Rad, Prague, Czech Republic). Final assay volumes were 15 µL and contained 7.5 µL iQ SYBR Green Supermix (Bio-Rad, Prague, Czech Republic), 0.15 µL of each primer (20 nmol/L), 3 µL of diluted cDNA and 4.2 µL of ultrapure water. The quantitative PCR reactions were performed as follows: denaturation at 95 • C for 10 min followed by 45 cycles of amplification (95 • C for 20 s, 60 • C for 20 s and 72 • C for 20 s). Each run was completed with a melting curve analysis in order to confirm the specificity of amplification and lack of primer dimers. Each pair of primers yielded a single peak in the melting curve and a single band of the expected size in agarose gel. Reactions for all samples were performed in triplicate. Standard curves were generated for each pair of primers using 10-fold serial dilution of standards. Blank controls with the omitted template were used. NPB, NPW and NPBWR1 mRNAs were determined in all heart chambers by subtracting their quantitative cycle values (Cq) to Cq of reference gene. As it exerts good stability within the heart, β-actin was used as a reference gene. The relative expression ratios were calculated using the 2 −∆∆Cq method. Immunofluorescence Shock-frozen hearts, stellate ganglia and/or DRG (n = 3 of each) were cut into 10-µm-thick sections using a Leica CM1850 cryostat (Leica, Bensheim, Germany) and fixed for 10 min in cold acetone. Samples were exposed at room temperature for 1 h to the medium blocking the nonspecific binding sites (1% BSA, 0.1% Triton X-100 and 2% normal goat serum) and then incubated Statistical Analysis All data are expressed as the mean ± standard error of the mean (SEM). All the data obtained from the individual experimental groups were first subjected to Shapiro-Wilk normality test. As the results of this test showed that the values in the individual groups did not show a normal distribution, a nonparametric Mann-Whitney test was subsequently used for statistical evaluation. Values of p < 0.05 were considered statistically significant. The analysis was performed using the software package STATISTICA Cz, version 7 (StatSoft CR, Prague, Czech Republic).	2020-10-28T13:05:59.506Z	2020-10-22T00:00:00.000	{ "year": 2020, "sha1": "490e91019c3c4a14731831acf2051e39c0e6b233", "oa_license": "CCBY", "oa_url": "https://www.mdpi.com/1422-0067/21/21/7827/pdf", "oa_status": "GOLD", "pdf_src": "Adhoc", "pdf_hash": "4fb51035fe34f5ea3bd00edd45a96eac88a48074", "s2fieldsofstudy": [ "Biology", "Medicine" ], "extfieldsofstudy": [ "Medicine", "Chemistry" ] }
11753689	pes2o/s2orc	v3-fos-license	The Human N-Alpha-Acetyltransferase 40 (hNaa40p/hNatD) Is Conserved from Yeast and N-Terminally Acetylates Histones H2A and H4 Protein Nα-terminal acetylation (Nt-acetylation) is considered one of the most common protein modification in eukaryotes, and 80-90% of all soluble human proteins are modified in this way, with functional implications ranging from altered protein function and stability to translocation potency amongst others. Nt-acetylation is catalyzed by N-terminal acetyltransferases (NATs), and in yeast five NAT types are identified and denoted NatA-NatE. Higher eukaryotes additionally express NatF. Except for NatD, human orthologues for all yeast NATs are identified. yNatD is defined as the catalytic unit Naa40p (Nat4) which co-translationally Nt-acetylates histones H2A and H4. In this study we identified and characterized hNaa40p/hNatD, the human orthologue of the yeast Naa40p. An in vitro proteome-derived peptide library Nt-acetylation assay indicated that recombinant hNaa40p acetylates N-termini starting with the consensus sequence Ser-Gly-Gly-Gly-Lys-, strongly resembling the N-termini of the human histones H2A and H4. This was confirmed as recombinant hNaa40p Nt-acetylated the oligopeptides derived from the N-termini of both histones. In contrast, a synthetically Nt-acetylated H4 N-terminal peptide with all lysines being non-acetylated, was not significantly acetylated by hNaa40p, indicating that hNaa40p catalyzed H4 Nα-acetylation and not H4 lysine Nε-acetylation. Also, immunoprecipitated hNaa40p specifically Nt-acetylated H4 in vitro. Heterologous expression of hNaa40p in a yeast naa40-Δ strain restored Nt-acetylation of yeast histone H4, but not H2A in vivo, probably reflecting the fact that the N-terminal sequences of human H2A and H4 are highly similar to each other and to yeast H4 while the N-terminal sequence of yeast H2A differs. Thus, Naa40p seems to have co-evolved with the human H2A sequence. Finally, a partial co-sedimentation with ribosomes indicates that hNaa40p co-translationally acetylates H2A and H4. Combined, our results strongly suggest that human Naa40p/NatD is conserved from yeast. Thus, the NATs of all classes of N-terminally acetylated proteins in humans now appear to be accounted for. Introduction N a -terminal acetylation (Nt-acetylation) occurs on 80-90% of soluble human proteins and 50-70% of soluble yeast proteins [1][2][3], making it one of the most widespread protein modifications in eukaryotes. This modification was previously largely believed to provide stability to proteins by preventing their degradation [4], which is contrary to the recent finding of Hwang et al. indicating that Nt-acetylation of specific N-termini in yeast acts as a degradation signal for the Doa10 ubiquitin ligase, thereby potentially targeting Nt-acetylated proteins for ubiquitin mediated degradation [5]. Studies over the last years have also linked Ntacetylation to several other important protein functions, such as sub-cellular protein targeting [6][7][8][9][10]. Nt-acetylation of specific proteins is also essential for their correct activity or for their interaction with other proteins [11][12][13]. More recently, an inverse correlation between Nt-acetylation and the ability of a protein to post-translationally enter the secretory pathway was revealed [14]. Taken together, these studies implicate Nt-acetylation as an essential modulator of diverse protein properties [15]. Finally, a defective NAT was recently found to be the cause of a lethal syndrome affecting male infants, indicating that proper Ntacetylation is functionally essential in humans [16]. Nt-acetylation is catalyzed by N-terminal acetyltransferases (NATs) and involves the transfer of the acetyl group from acetylcoenzyme A to the primary a-amino group of the very N-terminal amino acid residue of a protein. The amino acid sequence at the N-terminus is the major factor determining whether or not a given protein is Nt-acetylated and by which NAT. In yeast, five different NAT types are identified, designated NatA-NatE [17][18][19][20][21]. All NATs are composed of a catalytical subunit responsible for catalyzing the acetylation reaction, and up to two auxiliary subunits. The NATs are associated with ribosomes where the acetylation reaction occurs cotranslationally as the peptide extrudes from the ribosome, typically after 20-50 residues have left the exit tunnel [22][23][24]. All NATs display by and large their own unique substrate specificity; NatA acetylates Alanine, Serine, Threonine, Glycine and Valine N-termini after initiator methionine (iMet) removal by methionine aminopeptidases [17,18]. NatB acetylates proteins with Met-Asn-, Met-Asp-and Met-Glu-Ntermini [25], while NatC acetylates the iMet when it is followed by a hydrophobic residue [26]. For NatE no substrates have been identified in yeast, and the yNatE deletion strain shows no obvious phenotype [23]. However, studies on homologues from higher eukaryotes show that this acetyltransferase has the potency to Ntacetylate Met-Leu-and similarly starting peptide N-termini in vitro and is important for proper sister chromatid cohesion in vivo [27][28][29][30][31]. However, no in vivo substrates explaining this phenotype are identified. In addition to NatE, NatA, NatB, and NatC are also characterized in higher eukaryotes and shown to be evolutionarily conserved from yeast with respect to both complex composition and substrate specificity [1,10,32,33]. Recently, an additional NAT exclusively present in higher eukaryotes was identified and designated Naa60/NatF. The acetyltransferase activity of hNatF partially explains the increased levels of Nt-acetylation seen in humans as compared to yeast [3]. The activity of Naa40p/NatD (Nat4) is so far characterized in yeast only. yNaa40p was identified in the search for the acetyltransferase responsible for Nt-acetylation of histone H4. In the same study, histone H2A was also identified as a substrate for yNaa40p [20]. Later, the activity of yNaa40p was denoted yNatD, and the enzyme was predicted to act co-translationally since a fraction of yNaa40p associated with ribosomes [34]. The Nterminal residue of histone H4 is a serine, hence a predicted NatA substrate. In contrast however to other NatA-type N-termini, Ntacetylation of H4 was not affected in strains with mutant NatA subunits (ard1or nat1 -) suggesting that H4 is acetylated by another NAT [17]. The basic histone tail makes these proteins unusual substrates for a NAT since such residues are predicted to be inhibitory for acetylation [35]. In addition to perhaps acetylating only two selected substrates, H2A and H4, yNaa40p is also unusual in its requirement for as many as 20-50 available residues to effectively acetylate its substrates in vivo [34]. A search for additional substrates has proven unsuccessful, and this is also the case for the identification of putative auxiliary subunits which potentially remain to be identified. Song et al. failed to detect any phenotypes in a ynaa40-deletion strain [20]. However, Polevoda et al. [34] later found that deletion of ynaa40 induced some minor growth defects when this deletion strain was grown on media containing an inhibitor of transcription (3-amino-1,2,4-triazole), anti-mitotic microtubule-destabilizing drugs (benomyl and thiabendazole) or a nonspecific membrane inhibitor (dinitrobenzene). A double mutant encoding histone H4 with K5R, K8R and K12R replacements, as to create an N e -unacetylatable mutant, in addition to a disrupted ynaa40 gene, displayed an even greater growth defect on media containing some of the aforementioned compounds than either of the mutant strains alone, indicating that the lack of Nt-acetylation of histone H4 acts synergistically with the lack of lysine acetylation in the H4 tail, making Nt-acetylation of histone H4 part of a charge patch, thereby explaining the observed phenotypes based on a sub-optimal function of the histone tail [34]. The objective in the present study was to identify and investigate the human Naa40p/NatD. By the use of different methods, we found that the substrate specificity of hNaa40p is highly conserved from yeast, and recombinant hNaa40p was indeed capable of acetylating synthetic H4 and H2A peptides in vitro. Furthermore, heterologous hNaa40p also Nt-acetylated yeast H4 in vivo. Endogenous hNaa40p was additionally found to be associated with ribosomes in agreement with a co-translational activity, but interestingly a fraction of hNaa40p was also found to localize to the nucleus. With the recent characterization of several human NATs and the current identification of the human NatD, it is likely that all human NATs acting co-translationally have been identified since all classes of substrates found to be Nt-acetylated are covered. Results and Discussion A human homologue of the yeast N-alpha acetyltransferase Naa40p (Nat4) Sternglanz and coworkers previously identified yeast Naa40p (Nat4) to be an N a -terminal acetyltransferase, responsible for acetylating the N-termini of histones H2A and H4 [20]. In addition, they predicted yNaa40p to have homologues in higher eukaryotes, including humans. Later this putative human homologue (accession number NP_079047) was denoted hNaa40p, the catalytic subunit of the NatD type activity in accordance with the revised NAT nomenclature [19]. An alignment of the predicted Naa40 protein from different species is shown in Figure 1A. Of importance is the conservation of the acetyl-coenzyme A (Ac-CoA) binding motif, RxxGxG/A (where x can be any amino acid), from yeast to humans. This motif is found in all members of the Gcn5-related N-acetyltransferase (GNAT) superfamily and is probably required for Ac-CoA binding to the enzyme and hence its catalytic activity [36][37][38]. The conservation of this binding motif, in addition to an overall sequence similarity (23% amino acid sequence identity between yNaa40p and hNaa40p), is also seen in the alignment of the putative human Naa40p and the catalytical subunits of all characterized human NATs ( Figure 1B), which also revealed that some stretches of amino acids are unique for Naa40p. These stretches are marked with red lines above the alignment. The conservation of the Ac-CoA binding motif is strongly indicative of an N-terminal acetyltransferase activity, and the alignment point to hNaa40p as being the yNaa40p orthologue. hNaa40p is a NAT selectively acetylating Ser-Gly or Gly-Gly starting peptides Although the amino acid sequence of hNaa40p clearly supports its candidacy as a novel human NAT of the NatD type, there is to our knowledge no experimental evidence yet in support of this. In order to assess whether hNaa40p is truly a NAT, the predicted hNAA40 gene was cloned into a prokaryotic expression vector for recombinant MBP-hNaa40p fusion protein expression. A recently developed in vitro proteome-derived peptide library assay [30] was performed for an largely unbiased characterization of the potential enzymatic activity and substrate specificity of hNaa40p. As such, 177 unique substrate peptides were identified as being Nterminally acetylated by hNaa40p, providing evidence that hNaa40p is a genuine NAT. The resulting substrate specificity profile of hNaa40p, displayed in Figure 2, illustrates that hNaa40p preferentially acetylates Ser-Gly-Gly-Gly-Lys-starting peptides. The similarity between the sequence presented in the specificity profile and the N-terminal sequences of the human histones H2A and H4 being Ser-Gly-Arg-Gly-Lys-is striking, only deviating at position 3 with a Gly replaced by an Arg. It should be noted that since the assay is based on an N e -modified proteome-derived peptide library generated by the use of trypsin, internal Arginines will be generally missed and consequently will be absent among preferred N-terminal residues. The 177 substrate peptides acetylated by hNaa40p in this screen range from 7-19 amino acid residues in length, a result in contrast with the finding of Polevoda et al. postulating that acetylation by yNaa40p required at least 30-50 amino acid residues at the N-terminus of histone H2A and H4 [34]. The apparent difference in requirement for more residues in the yeast experiments might implicate differences between species, or solely reflect different requirements between an in vitro setting with purified enzyme and peptides, and the in vivo setting in yeast where a ribosome-bound yNaa40p targets the emerging nascent polypeptide. For the in vitro assay described here, seven amino acids are clearly enough to ensure specificity and acetylation by hNaa40p. It should be noted that the N-termini in the peptide library assay do not represent true N-termini since they are generated by proteolytic cleavage. However, the significant match between the hNaa40p substrates from the specificity profile and the N-termini of histones H2A and H4 strongly suggests a conserved function from yeast to humans with respect to Ntacetylation of histones H2A and H4. hNaa40p N-terminally acetylates histones H2A and H4 in vitro In order to obtain further insight into the substrate specificity of hNaa40p, the quantitative aspect of substrate Nt-acetylation, and particularly the role of hNaa40p in histone acetylation thus evaluating its candidacy as a true orthologue of yeast Naa40p/ NatD, an in vitro Nt-acetylation assay with purified recombinant hNaa40p was conducted. hNaa40p was challenged by synthetic peptides representing the N-termini of selected proteins, and interestingly, hNaa40p strongly preferred histones H2A and H4 ( Figure 3). A histone H4 peptide with a synthetically pre-acetylated N-terminus was no longer acetylated efficiently by hNaa40p, consistent with hNaa40p acetylating the a-amino group of H4 and not significantly any of the lysine residues known to be N eacetylated by lysine acetyltransferases (KATs) [39]. Furthermore, only minor acetylation activity towards the HMG-1 (SESS), hnRNPF (MLGP), or histone H3 (ARTQ) peptides was observed. These three peptides are representatives of a classical NatA substrate [1], a NatE substrate [28], and a non-acetylated Nterminus [20], respectively. The lack of acetylation of the SESS peptide could be explained by the fact that a Glu residue in position P29 is generally inhibitory for hNaa40p acetylation according to the specificity profile as deduced from the peptide library assays ( Figure 2). To determine whether hNaa40p expressed in human cells possesses specific acetyltransferase activity, as shown when using recombinant hNaa40p purified from E. coli, hNaa40p-V5 was expressed in HEK293 cells and immunoprecipitated. Immunoprecipitated hNaa40p-V5 was used in an in vitro acetylation assay with synthetic peptides as described in Material and Methods. Histone H4 was significantly more acetylated as compared to the N-terminally blocked peptide ( Figure 4). Taken together, the results from the in vitro experiments strongly suggest that hNaa40p is a NAT conserved from yeast which specifically Nt-acetylates histones H2A and H4 and that such specific acetylation may also take place in cultured cells. Ectopically expressed hNaa40p in yeast acetylates yeast histone H4 but not H2A in vivo In order to assess whether hNaa40p represents a true orthologue of yNaa40p in vivo, we generated a yeast naa40-D strain expressing hNaa40p ( Figure 5). When comparing Ntacetylation of histones in control yeast (yeast Ctrl 1 or 2) to yeast lacking NAA40 (ynaa40-D) and to the yeast strain lacking NAA40 but expressing hNaa40p (y[hNAA40]), hNaa40p expression restored the diminished histone H4 acetylation ( Figure 6). In fact, the expression of hNaa40p increased the total level of H4acetylation (i.e. the combined Nt-acetylation and lysine acetylation observed). This could indicate that Nt-acetylation stimulates lysine acetylation in vivo, that hNaa40p is also capable of catalyzing lysine acetylation of histone H4, or perhaps that the degree of Ntacetylation is higher in yeast expressing hNaa40p as compared to yNaa40p. Lysine acetylation is normally catalyzed by the lysine acetyltransferases (KATs), while Nt-acetylation is catalyzed by NATs. However, a dual NAT/KAT-role for the NAT-enzymes , NatE and NatF. The conserved acetyl coenzyme A (Ac-CoA) binding motif RxxGxG/A, where x designates any amino acid, is indicated in both alignments. The red lines above the alignment represent examples of stretches of amino acid residues unique to Naa40p. Purple background indicates acidic residues, red indicates basic residues, orange indicates glycine, yellow indicates proline, blue indicates hydrophobic residues, green indicates polar residues, and turquoise indicates histidine and tyrosine residues. The alignments were created using Jalview multiple alignment editor [47,48]. doi:10.1371/journal.pone.0024713.g001 has been suggested for hNaa50p [28,40] as well as for hNaa10p [41,42]. The in vitro assays described above (Figures 3 and 4) indicate that hNaa40p mostly acts as a NAT towards histone H4, but a minor KAT activity might occur. Further detailed experiments are required to conclude on this point. As expected, no alterations in histone acetylation were observed for H2B and H3 for any of the yeast strains used (data not shown). Surprisingly, histone H2A, which was less acetylated in the yeast naa40-D strain as compared to the controls, was not acetylated by hNaa40p ( Figure 6). This probably reflects the fact that the first five Nterminal amino acid residues of histone H2A are altered from yeast (Ser-Gly-Gly-Lys-Gly-) to humans (Ser-Gly-Arg-Gly-Lys-) to become identical to the Nt-sequences of yeast and human histone H4 (Ser-Gly-Arg-Gly-Lys-). From the in vitro specificity profile (Figure 2), it is clear that Lys4 and Gly5 are not optimal for hNaa40p activity. Naa40p might have evolved to become even more specific in humans, acetylating the N-termini of the very similar human H2A and H4, while the yeast homologue is able to Nt-acetylate the more divergent N-termini of yeast H2A and H4. Despite the lack of yeast histone H2A acetylation, the in vitro and in vivo data combined, strongly suggest that hNaa40p is the true orthologue of yNaa40p. Subcellular localization of hNaa40p Subcellular fractionation was performed to elucidate on the localization pattern of hNaa40p. As shown in Figure 7 endogenous hNaa40p localizes to both the cytoplasm and the nuclei of A431 cells. The same localization was also observed in HEK293 cells (data not shown). In addition, subcellular localization of exogenously expressed hNaa40p-V5 was determined using immunofluorescence microscopy ( Figure 8). In both cell lines investigated, HeLa and A431, hNaa40p-V5 was found to have a distributed cytoplasmic localization with some degree of nuclear localization. Surprisingly, two main populations of localization patterns were observed, one where hNaa40p-V5 was found mainly in the cytoplasm and to a lesser degree in the nucleus (Figure 8, panel 1 (white arrow), and 2), and another where hNaa40p-V5 was found enriched in the nucleus (Figure 8, panel 1 (red arrow), and 3). The latter could possibly be due to an artificial nuclear accumulation of the overexpressed protein. However, only low to medium intensity cells were taken into account. In addition, cells with a low amount of hNaa40p-V5 in the nucleus compared to the cytoplasm, were also observed among highly overexpressing cells (figure S1). Furthermore, there also existed a smaller population of cells in which the fluorescence intensity level was equal in the two compartments (data not shown), potentially representing an intermediate or transition stage. hNaa40p associates with ribosomes NATs are involved in co-translational acetylation of nascent polypeptides, and are therefore physically associated with ribosomes. The yeast Naa40p was also demonstrated to associate with ribosomes [34]. In order to investigate whether hNaa40p shares this feature, ribosomes from HEK293 cells were isolated and the binding of endogenous hNaa40p was analyzed by Western blotting. Indeed, a small but reproducible fraction of hNaa40p was detected in the polysomal pellet while the majority of hNaa40p was present in the supernatant post ultracentrifugation ( Figure 9). As demonstrated here for the catalytic subunit of the NatA complex, hNaa10p, this pattern is similar to what was observed previously for all the other human NAT subunits [10,30,33,43]. yNaa10p has previously been shown to associate with ribosomes through an interaction with the auxiliary subunit yNaa15p [23]. No such auxiliary subunit has been identified for Naa40p neither in yeast nor in humans, suggesting that hNaa40p itself might interact with the ribosome directly. However, the lack of identified additional NatD subunit(s) does not exclude their presence. The large fraction of hNaa40p not in complex with ribosomes indicates that most of hNaa40p is in a free form in the cytoplasm or in the nucleus based on the fractionation and immunofluorescence experiments presented above. Concluding remarks In conclusion, we have here identified the human Naa40p (NatD) acting as a co-translational NAT towards the N-termini of the histones H2A and H4. Overall, the set of functionally characterized human NATs now comprises NatA-NatF which are likely to be responsible for all co-translational Nt-acetylation events in human cells. A potential role for hNaa40p in the nucleus as well as a potential function as a KAT awaits further investigation. Construction of plasmids The plasmid (phNAA40-V5) encoding V5-tagged hNaa40p was constructed from cDNA made from total RNA isolated from human HEK293 cells. Superscript TM II Reverse Transcriptase (Invitrogen) was used to make cDNA. PCR products were inserted into the TOPO TA vector pcDNA 3.1/V5-Hisß TOPOH (Invitrogen) according to the instruction manual. Prokaryotic expression vector for hNAA40 was made by subcloning the amplified hNAA40 gene from phNAA40-V5 to the pETM41 expression vector using XmaI and EcoRI restriction sites. pETM41 was generously provided by G. Stier, EMBL, Heidelberg. E. coli hNaa40p expression and purification E. coli BL21 Star TM (DE3) (Invitrogen) was transformed by heat shock with the pETM41-hNAA40 plasmid encoding a MBP- hNaa40p fusion protein, and grown overnight on LB agar containing kanamycin. A well separated colony was picked and inoculated at 37uC in 200 ml LB medium supplemented with kanamycin to an OD 600 of 0.6. Protein expression was induced by IPTG (1 mM) at 20uC for 18 hours. The culture was harvested by centrifugation at 3000 x g for 20 minutes. The cells were lysed by sonication in lysis buffer (50 mM Tris-HCl (pH 8.3), 300 mM NaCl, 1 mM DTT, 1 tablet EDTA-free protease Inhibitor cocktail per 50 ml (Roche)) and cleared by centrifugation at 10000 x g for 20 minutes. Cleared lysate was applied on a metal affinity FPLC column (HisTrap HP, GE Healthcare, Uppsala, Sweden), and MBP-hNaa40p was eluted with 300 mM imidazole in 50 mM Tris (pH 8.3), 300 mM NaCl, 1 mM DTT. MBP-hNaa40p was further purified on a size exclusion chromatography column (Superdex TM 75, GE Healthcare). In vitro proteome-derived peptide library Nt-acetylation assay Proteome-derived N a -free peptide libraries were generated from human K-562 cells as described previously [30]. 100 nmol of the desalted N a -free peptide pool was reconstituted in acetylation buffer (50 mM Tris-HCl (pH 8.3), 1 mM DTT, 800 mM EDTA, 10% glycerol) together with equimolar amounts of a stable isotope encoded variant of acetyl-CoA, 13 C 2 -acetyl CoA, (99% 13 C 2 -acetyl CoA, ISOTEC-Sigma (lithium salt)) and 500 pmol of enzyme (i.e. recombinant MBP-hNaa40p) was added to a final reaction volume of 500 ml. The reaction was allowed to proceed for 2 h at 37uC and stopped by addition of acetic acid to a 5% final concentration. NAT oligopeptide-substrate recovery and LC-MS/MS analyses was performed as decribed previously [30]. LC-MS/MS analysis LC-MS/MS analysis was performed as described previously [30]. The generated MS/MS peak lists were then searched with Mascot using the Mascot Daemon interface (version 2.2.0, Matrix Science). Searches were performed in the Swiss-Prot database with taxonomy set to human (Homo sapiens) (UniProtKB/Swiss-Prot database). Trideutero(d 3 )-acetylation at lysines, carbamidomethylation of cysteine and methionine oxidation to methioninesulfoxide were set as fixed modifications. Variable modifications included 13 C 2 -acetylation, d 3 -acetylation, acetylation and pyroglutamate formation of N-terminal glutamine. Endoproteinase Arg-C/P (Arg-C specificity with arginine-proline cleavage allowed) was set as enzyme allowing no missed cleavages. The mass tolerance on the precursor ion was set to 10 ppm and on fragment ions to 0.5 Da. Peptide identifications were considered significant when their Mascot ion score was above the identity threshold set at a 99% confidence interval and when they were top-ranked. In vitro acetylation assay using synthetic oligopeptides as substrates For the in vitro acetylation assay using custom-made synthetic peptides, purified MBP-hNaa40p (400 nM) was mixed with selected oligopeptide substrates (200 mM) and 200 mM of [1-14 C] acetyl-CoA in a total volume of 170 ml acetylation buffer (50 mM Tris-HCl (pH 8.3), 800 mM EDTA, 10% glycerol, 1 mM DTT) and incubated for 1 hour at 37uC. Following incubation, 150 ml SP Sepharose beads (50% slurry with 0.5 M acetic acid) were added to the mix and incubated at room temperature for 5 minutes. After centrifugation 6000 x g for 2 min, the beads were washed three times with 0.5 M acetic acid. Following the last wash, the supernatant was removed and the beads washed with 300 ml ice cold methanol, after centrifugation the supernatant was removed and the beads were transferred to 5 ml counting fluid and radioactivity was measured as disintegrations per minute (DPM) using a scintillation counter. For the immunoprecipitation experiments, HEK293 cells were transfected with a plasmid encoding hNaa40p-V5 using FugeneH 6 Transfection reagent. Cells were harvested 48 hours post-transfection and resuspended in lysis buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 5 mM EDTA, 0.5% NP-40, Pefabloc). After lysis on ice for five minutes, the lysate was cleared by centrifugation 3000 x g for 5 minutes. hNaa40p-V5 was immunoprecipitated from the lysate using specific antibodies against the V5-tag (Invitrogen) and Protein A/G Agarose beads (Santa Cruz). Immunoprecipitation using antibodies against the Xpress-tag (Invitrogen) was used as a negative control. Following washing (3 times) with lysis buffer and acetylation buffer the immunoprecipitates were used in an in vitro acetylation assay as described above. Yeast strain generation and cultivation for histone isolation Yeast was grown and manipulated according to standard procedures [44]. Expression vector for hNAA40 was made by subcloning the amplified hNAA40 gene from phNAA40-V5 to pBEVY-U bidirectional expression vector [45] using the BsmBI restriction site. The S. cerevisiae MAT alpha strain Y10000 was transformed with pBEVY-U and acted as control strains (yeast Ctrl). Strain Y16202 Figure 7. Endogenous hNaa40p localizes to the cytosol and nucleus. A431 cells were harvested and fractionated. Total cell lysate (T), cytosolic fraction (C), and nuclear fraction (N) were analyzed by SDS-PAGE and western blotting. Beta-tubulin was used as a cytosolic marker, while histone H4 was used as a nuclear marker. hNaa40p was detected in all fractions using an anti-hNaa40p antibody. Results shown are representative of three independent experiments. doi:10.1371/journal.pone.0024713.g007 Figure 6. Ectopically expressed hNaa40p in yeast acetylates histone H4 but not H2A in vivo. (A) The mature N-terminus of yeast histone H2A ( 2 SGGKGGKAGSAAKASQSR 19 ) (left panels) and H4 ( 2 SGRGKGGKGLGKGGAKR 18 ) (right panels) was identified as being partially in vivo N-acetylated. Various N-acetylated forms of these histone N-termini were detected in all samples analyzed. A decrease of 5 Da, corresponds with the gain of an extra in vivo N-acetyl group (i.e. all free amines were in vitro N-acetylated using 13 C 2 D 3 -NHS acetate; i.e. adding a 5 Da heavier acetyl group to free amines). Colored squares indicate the number of in vivo (orange) and in vitro (gray) acetyl group present per detected isotopic envelope. Clearly deletion of yNAA40 reduces the overall degree of N-acetylation (middle panels) for both histone N-termini while the ectopic expression of hNaa40p restores and even augments the degree of N-aectylation of the H2A N-terminus, while not having any observable effect on the overall N-acetylation status of the H4 N-terminus (B) The percentage of all isobaric N-terminal variants as shown in panel A were calculated and their contribution to the overall percentage of the respective histone N-terminus plotted. doi:10.1371/journal.pone.0024713.g006 (naa40-D, Euroscarf) was transformed with pBEVY-U or pBEVY-U-hNAA40 to generate the strains ynaa40-D and y[hNAA40], respectively. The three strains generated were selected on plates lacking uracil. Disruption of NAA40 was verified by isolation of genomic DNA, and PCR using the primers yNAA40-86F: 59-CCACATATGTATGCGCGGTGC-39 and yNAA40 941R: 59-GGAGTCAAGGATTCGAGTCGC-39. Expression of hNaa40p was verified by SDS-PAGE and Western blotting analysis of yeast lysates using anti-hNaa40p (Biosite). For histone isolation, the strains were cultivated in 10 ml synthetic medium lacking uracil (Sigma) to an OD600 nm of ,3. Cells were collected by centrifugation for 5 min at 1500 x g and resuspended in 500 ml lysis buffer (0.2% TRITON-X-100, 10 mM HEPES (pH 7.6) 1.5 mM MgCl 2 , 10 mM KCl, 1 mM Pefabloc, 1 mM Na 3 VO 4 and 5 mM TSA). Glass beads were added before several rounds of vortex/ice. After lysis, the glass beads were removed and the supernatant was centrifuged for 5 minutes at 3000 x g. Pellets were washed in lysis buffer followed by centrifugation for 5 min at 300 x g. The pellets were incubated in 0.4 M HCl for 1 hour and centrifuged for 15 min at 13000 x g. 6 volumes of ice-cold aceton were added to the supernatant and acid-extracted proteins precipitated overnight at 220uC. After centrifugation for 15 min at 13000 x g, and 3 rounds of washing with ice-cold aceton the pellets were dried for 5 min in a vacuum concentrator. SDS-PAGE of histone samples and determination of the N-acetylation status of yeast histone N-termini using in-gel stable-isotope labeling (ISIL) followed by in-gel digestion All dried histone pellets were dissolved in 50 ml sample buffer (8 M urea, 5% b-mercaptoethanol and 10 mM Tris-HCl (pH 7.0)). 20 ml of each sample was added SDS-sample buffer (Biorad XT sample buffer) and the samples were analyzed on a 4%-16% gradient XT precast Criterion gel using XT-MOPS buffer (Biorad). In-gel-stable isotope labelling was performed as described previously [43] with the following modifications; primary amines were ( 13 C 2 d 3 )-acetylated by the addition of 1 mg N-hydroxysuccinimide ( 13 C 2 d 3 )-acetate and trypsin digestion proceeded for 309 at 37uC, representative of a sub-optimal incubation time, to increase the occurrence of missed cleavages (i.e. the yeast histone H4 N- terminus harbours an Arg-residue at the third position (i.e. the Nterminus w/o the iMet)) thereby allowing for the identification of the yeast H4 histone N-terminus (next to the H2A N-terminus). The resulting peptide mixtures were acidified (0.1% formic acid) and analyzed by LC-MS/MS analysis as described previously [43]. The generated MS/MS peak lists were then searched with Mascot using the Mascot Daemon interface (version 2.2.0, Matrix Science). Searches were performed in the Swiss-Prot database with taxonomy set to human (Homo sapiens) (UniProtKB/Swiss-Prot database). 13 C 2 d 3 -or normal acetylation at lysines, 13 C 2 d 3 -or normal acetylation of the N-terminus, pyro-glutamate formation of N-terminal glutamine and methionine oxidation to methioninesulfoxide were set as variable modifications. Endoproteinase Arg-C/P (Arg-C specificity with arginine-proline cleavage allowed) was set as enzyme allowing 2 missed cleavages. The mass tolerance on the precursor ion was set to 10 ppm and on fragment ions to 0.5 Da. Peptide identifications were considered significant when their Mascot ion score was above the identity threshold set at a 99% confidence interval and when they were top-ranked. The extent of histone N-acetylation was calculated from the peptide ion signals observed in the MS spectra as described in [3]. Immunofluorescence For fluorescent microscopic analysis, cells were transfected with plasmid encoding hNaa40p-V5. Transfection was carried out using FugeneH 6 Transfection reagent (Roche) and 0.5 mg plasmid per ml medium. 24 hours post transfection, cells were incubated for 10 min with Hoechst 33342 (Invitrogen) (0.4 ml per ml medium). Cells were then washed in PBS and fixed for 20 min in 4% paraformaldehyde in PBS. Fixed cells were washed three times in PBS, permeabilized using 0.2% Triton X-100 for 10 min and then washed once in PBS before incubation in blocking buffer (10% BSA) for 1 hour. Primary antibody, Mouse IgG2a anti-V5 (Invitrogen), diluted 1:200 in 2% BSA was incubated for 1 h. Cells were washed one time in 0.1% Triton X-100 for 5 min, followed by two washes in PBS. The fluorochrome-conjugated secondary antibody, Alexa FluorH 594 goat anti-mouse IgG 2a (Molecular Probes, Eugene, Oregon, USA), diluted 1:100 in 2% BSA was incubated for 40 min. Cells were again washed once in 0.1% Triton X-100 for 5 min followed by two PBS washes. Finally, coverslips were dipped in ddH 2 O before mounting on glass slides using SlowFade Gold antifade reagent (Invitrogen) after which coverslips were sealed. Imaging was performed using a Leica DMRXA microscope equipped with a 26 magnification lens, a Leica DC500 camera, and a Leica HCX PL APO 1006oil immersion objective. Hoechst stain was visualized using a 340-380 nm bandpass filter with longpass suppression filter (425 nm); Alexa-594 was visualized using a 560/40 nm bandpass filter with bandpass suppression filter (645/75 nm). Untransfected cells and secondary antibody staining only were considered as a control for unspecific binding of the antibodies. Subcellular fractionation Subcellular fractionation was used to separate nuclear from cytosolic proteins. Based on the protocol by Suzuki et al. [46], cells were washed in PBS and harvested by ''pop spin'' for 10 seconds (3000 x g). The resulting cell pellet was resuspended in 900 ml of ice-cold lysis buffer (0.1% NP40 in PBS) and a sample was taken as total lysate fraction. Following a second pop-spin for 10 seconds, a sample from the supernatant was removed representative of the cytosolic fraction. The nuclear pellet was washed by resuspension in 1 ml lysis buffer followed by centrifugation for 10 seconds. The supernatant was discarded and the pellet was resuspended in sample buffer and designated as the nuclear fraction. The nuclear fraction and the total lysate fraction were sonicated two times for 10 seconds. The subcellular fractions were analysed by SDS-PAGE and Western blotting. Ribosome isolation Approximately 2610 7 HEK293 cells were used per experiment. Prior to harvesting, cells were treated with 10 mg/ml cycloheximide (CHX) for 5 minutes at 37uC. Cells were harvested, lysed with KCl-containing ribosome lysis buffer (1.1% (w/v) KCl, 0.15% (w/v) triethanolamine, 0.1% (w/v) magnesium acetate, 8.6% (w/v) sucrose, 0.05% (w/v) sodium deoxycholate, 0.5% (v/v) Triton-X100, 0.25% (v/v) Pefabloc), and incubated on ice for 15 minutes. After removing the nuclear and membrane containing fraction by centrifugation at 400 6 g for 10 min, 700 ml cell lysate was ultracentrifuged at 436,000 6 g for 25 minutes on a 0.4 ml cushion of 25% sucrose in KCl ribosome lysis buffer using a MLA-130 rotor (Beckman, Geneva, Switzerland). Pellets were resuspended in ribosome lysis buffer with the indicated KCl concentrations, followed by ultracentrifugation as described above. Pellets were re-suspended in KCl ribosome lysis buffer, and prepared for analysis by SDS-PAGE and Western blotting. Supporting Information Figure S1 Subcellular localization of hNaa40p-V5 in highly overexpressing cells. A431 and HeLa cells were prepared as described in Figure 8. The same expression pattern as shown in Figure 8 (panel 1 (white arrows), and 2) was also observed in highly overexpressing cells. Scale bars are 5 mm.	2014-10-01T00:00:00.000Z	2011-09-15T00:00:00.000	{ "year": 2011, "sha1": "3a2e6c0828419f589043d442d1f09d5a21d26fc7", "oa_license": "CCBY", "oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0024713&type=printable", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "2762c2b232e16b3e5d5f46d878ef061e408a3693", "s2fieldsofstudy": [ "Biology" ], "extfieldsofstudy": [ "Biology", "Medicine" ] }
55735590	pes2o/s2orc	v3-fos-license	A LOFAR view on the duty cycle of young radio sources Compact Steep Spectrum, Gigahertz Peaked Spectrum and High Frequency Peak (CSS, GPS, HFP) sources are considered to be young radio sources but the details of their duty cycle are not well understood. In some cases they are thought to develop in large radio galaxies, while in other cases their jets may experience intermittent activity or die prematurely and remain confined within the host galaxy. By studying in a systematic way the presence and the properties of any extended emission surrounding these compact sources we can provide firmer constraints on their evolutionary history and on the timescales of activity of the radio source. Remnant emission from previous outbursts is supposed to have very low surface brightness and to be brighter at low frequency. Taking advantage of the unprecedented sensitivity and resolution provided by the Low Frequency Array (LOFAR) we have started a systematic search of new CSS, GPS and HFP sources with extended emission, as well as a more detailed study of some well-known of these sources. Here we present the key points of our search in the LOFAR fields and a more in-depth analysis on the source B2 0258+35, a CSS source surrounded by a pair of large, diffuse radio lobes. Introduction To date, Gigahertz Peak Spectrum, Compact Steep Spectrum and High Frequency Peak (GPS, CSS, HFP) sources are thought to be the best representatives of young radio galaxies because of their compact double morphology which resembles the FRI/FRII extended structures. Both dynamical and radiative arguments suggest that the typical ages of these sources lie in the range 10 2 − 10 5 yr (O'Dea (1998), Owsianik & Conway (1998), Murgia et al. (1999)). It is expected that in some cases the compact sources will expand to hundreds of kpc sizes on timescales of 1-100 Myr (Parma et al. (1999)). However, a significant fraction may never evolve into large sources (O'Dea & Baum (1997), Readhead et al. (1996), An & Baan (2012)). The best way to broaden our knowledge on the duty cycle of these objects is to investigate the presence of extended emission surrounding the compact source. Stanghellini et al. (1990) and Stanghellini et al. (2005) find respectively that about 25% (4/15) and 20% (6/33) of the GPS sources in their samples show associated diffuse emission on the arcsecond-scale, which translates into structures of hundreds of kpc up to few Mpc. The luminosity of this extended emission is often negligible with respect to the compact source. However, its presence can be relevant in terms of the evolutionary history of the radio source. The most common interpretation of this emission is connected to previous epochs of activity. To date many observations probe that radio galaxies are episodic in nature (Saikia & Jamrozy (2009)) and statistical studies suggest that this Corresponding author: e-mail: brienza@astron.nl duty cycle strongly depends on the power of the radio source (Best et al. (2005)). Sources which exhibit at the same time remnant emitting plasma from previous outbursts and newborn jets are termed restarted radio galaxies and can be effectively used to investigate the timescales of activity of the radio AGN. One of the most well-known case of GPS with extended emission is the source 0108+388. This source shows a double compact morphology on milli-arcsecond scales as well as some diffuse emission at a distance of 20 from the core connected by a bridge. Multi-frequency studies have been performed by Baum et al. (1990) in order to clarify the nature of the source. However, whether the extended plasma is fuelled by the central activity remains unclear. The clearest evidence of restarted activity in CSS and GPS are the sources 3C236 (Willis et al. (1974)), B1144+35 (Schoenmakers et al. (1999)) and B1245+676 (Marecki et al. (2003), Saikia et al. (2007)). All these sources have Mpc-scale morphologies recalling the structure of a Double-Double Radio Galaxy (DDRG, Schoenmakers et al. (1999)) which is suggested to originate from multiple AGN outbursts. Alternative explanations to the restarting scenario have also been proposed. For example, Baum et al. (1990) suggest that GPS/CSS sources could arise within a large-scale radio galaxy if the jet propagation was obstructed or impeded on scales of tens of parsecs to few kiloparsecs. In that case, the extended structure would remain visible along with the confined parsec scale source that would have the appearance of a GPS source. This may occur in coincidence with gas-rich galaxy-galaxy interactions or, more in general, because of instabilities of the accretion process. At present, the statistics on young sources with associated extended emission is not sufficient to provide meaningful constraints on their incidence within the general radio galaxy population, and on the implications for the triggering and evolution of radio loud active galaxies. Taking advantage of the high sensitivity and high angular resolution at low frequencies provided by the Low Frequency Array (LOFAR, van Haarlem et al. (2013)) we have started a systematic search of these objects aiming at broaden their statistics. The need for the low frequency is connected to the spectral evolution of the radio emission as the source ages. Indeed, as the injection of fresh particles stops, an exponential break appears in the power-law radio spectrum (S ∝ ν −α ) at high frequency due to preferential radiative cooling of high-energy electrons (Kardashev (1962), Pacholczyk (1970)). Therefore, it is at low frequencies where the source is visible for longer time and where we plan to address our searches. In order to explore the variety of source characteristics and prepare the selection criteria to be used in this search we are also performing single object detailed studies such as the CSS source B2 0258+35, which is surrounded by a pair of large, diffuse radio lobes Shulevski et al. (2012). A combination with higher frequencies will also allow to perform spectral studies to infer the physics of these objects and finally give constraints on their origin and evolution. In Section 2 we discuss how and why LOFAR promises to be an important instrument in this field, and in Section 3 we present LOFAR observations of B2 0258+35 and discuss possible interpretations for the source. The cosmology adopted in this work assumes a flat universe and the following parameters: H 0 = 70 km s −1 Mpc −1 , Ω Λ = 0.7, Ω M = 0.3. Searching for AGN remnants with LOFAR Because of its unprecedented capabilities at low frequency LOFAR (van Haarlem et al. (2013)) is now playing a primary role in this field of research. The LOFAR Dutch array is currently composed of 24 stations located within a radius of 2 km referred to as the core and 14 remote stations arranged in an approximation to a logarithmic spiral distribution extending out to a radius of 90 km. This configuration provides a maximum resolution of ∼5 in the high-band (HBA, and ∼10 in the low-band (LBA, 30-90 MHz) as well as a good instantaneous uv-coverage on the short baselines. Moreover, the 9 currently operational international stations distributed over 4 different European countries generate baselines up to 1000 km allowing to achieve sub-arcsecond resolutions (e.g. 0.33 at 150 MHz). In the context of the search and study of AGN remnant radio plasma there are a number of characteristics that make LOFAR a unique instrument in the field. The first one is its high sensitivity with noise values of 0.2 mJy at 150 MHz and 3.9 mJy at 60 MHz for 10 hours observations. This is essential for detecting low-surface brightness sources like AGN remnant radio plasma. The second one is the variety of resolutions that are provided by a single observation. This unique array configuration, which combines a dense core of short baselines together with longer baselines, allows at the same time a good imaging of diffuse emission as well as compact, small-scale structures providing a complete morphological characterization of the sources. Moreover, LO-FAR highest resolutions match for the first time those at higher frequency and open the way to systematic resolved spectral studies. Finally, the wide field of view produced by LOFAR observations (e.g. 6 deg at 150 MHz and 12 deg at 60 MHz) allows significant statistical studies and source searches directly in the surroundings of the target source. A few detections of remnant and restarted radio galaxy in the LOFAR fields have already been presented in literature (NGC 5590, van Weeren et al. (2014); 4C35.06, Shulevski et al. (2015); J1828+49, Brienza et al. (2015)) providing a taste of the instrument capabilities. These sources have been morphologically identified based on their amorphous shapes and low-surface brightness of a few mJy arcmin −2 . This class of objects is expected to be numerous in the LO-FAR fields and therefore new systematic searches are being planned. One of the Key Science Project of LOFAR is the continuum surveys which follow a three-tier approach with different observational setups (Röttgering et al. (2003), van Haarlem et al. (2013)). The Tier-1 high-frequency survey will cover the entire northern sky with a resolution of 5 and an rms of 0.07 mJy at frequencies 120-180 MHz. As to June 2015 the high-frequency Tier-1 survey has completed about 1000 deg 2 observations. Typical maps have 20 resolution and noise equal to 0.5 mJy beam −1 with the only limitation to the image quality being the ionospheric calibration. A low-frequency Tier-1 survey is also scheduled to start in the next future. Using the fields already observed and reduced for the surveys we have started a search of remnant and restarted radio galaxies that will be further expanded to the whole northern sky in the future. Beyond blind searches we aim to systematically investigate the surroundings of all the wellknown CSS/GPS/HFP sources in order to provide a better statistic of the presence of associated extended emission and investigate its nature in closer detail. An initial search in this direction has already been performed by Shulevski (PhD thesis, 2015) using the LOFAR Multifrequency Snapshot Sky Survey (MSSS, Heald et al. (2015)). However, the sensitivity and spatial resolution of this survey turned out to be not ideally suited for this search and no new detections were made even though spectral index studies suggested that some CSS sources might show an excess of emission. Despite that, MSSS nicely detects two of the well-known giant radio galaxies with GPS or CSS central cores mentioned in Sec. 1 (i.e. 3C236 and B1245+676). These issues are overcome by the high resolution and sensitivity provided by the LOFAR Tier-1 survey which can probe extended emission on 5-10 arcsecond scales. Further details about the selection criteria we intend to use for identifying remnant radio plasma are discussed in Brienza et al. (2015). Higher frequency data will be then required to complete the radio spectral coverage allowing radiative ageing analysis of the sources. Furthermore, complementary data in other frequency regimes will probe the properties of the host galaxy and its environment. In this way we aim at creating new representative samples of remnant and restarted radio galaxies to start a systematic investigation of their properties as a function of energetics, host galaxy and environment. B2 0258+35: a restarting radio galaxy? In the context of restarting radio galaxies the CSS source B2 0258+35 has revealed a very intriguing story. The radio source has a power equal to 10 24.37 W Hz −1 at 408 MHz (Giovannini et al. (2001)) and it is associated to the field early-type galaxy NGC 1167 (z=0.0165) which is optically classified as a Seyfert 2 galaxy (Ho et al. (1997)). Its radio morphology was first studied by Sanghera et al. (1995) using the VLA and MERLIN+EVN at 5 and 1.6 GHz respectively. They classified it as a CSS source even though the high frequency spectrum is not very steep α = 0.54. Giroletti et al. (2005) provide better imaging of the source using the VLA at 8 and 22 GHz. At sub-arcsecond resolution the source shows a two-lobe morphology of 3.2 size (1.1 kpc) without any clear indication of hot-spots and the southern lobe exhibits an extension towards north-east (see inset in Fig. 1 in the top panel). At milli-arcsecond resolution Giovannini et al. (2001) further detect a compact component with jet-like structure. Assuming an equipartition magnetic field of B = 9×10 −5 Gauss and the detected break frequency at 4.6 GHz, Giroletti et al. (2005) compute an average spectral age of 9 × 10 5 yr. It is worth noting though that the outermost regions of the lobes show a quite steep spectral index in the range 1.0-1.5. Giroletti et al. (2005) speculate that the source will not grow into an FRI/II radio galaxy because i) it doesn't show hot-spots, ii) the bending of the southern lobe indicates a dense surrounding ISM, iii) its size is small for the estimated spectral ages. A new significant ingredient for reconstructing the evolutionary history of the source is the discovery by Shulevski et al. (2012) of an extended emission with double-lobe morphology of 240 kpc size surrounding the CSS (see Fig.1 top panel). The very low surface brightness emission at 1.4 GHz with average value of 1.4 mJy arcmin −2 remained undetected in all of the previous studies. The low surface brightness and lack of compact features such as hotspots or jets likely indicate that the emission could be remnant plasma from a previous cycle of activity (Shulevski et al. (2012)). Assuming the lobes to be buoyant bubbles expanding in the intergalactic medium Shulevski et al. (2012) LOFAR observations of B2 0258+35 In order to further investigate the nature of the source we performed LOFAR observations in the HBA. The source was observed for 8 hours using the Dutch array with a bandwidth of 40 MHz. 3C48 has been used as flux density calibrator and the flux scale has been set according to Scaife & Heald (2012). The amplitude corrected visibilities have been phase self-calibrated and the final image was produced using a robust weighting of -0.3 and has final beam size of 98 × 80 arcsec. The RMS of the map is 3 mJy beam −1 . The integrated spectrum of the central CSS source is presented in Fig. 2 with the LOFAR measurements at 120 MHz and 160 MHz nicely following the literature points. The linear relation between size and turnover frequency found by O'Dea & Baum (1997) predicts a turnover in the CSS spectrum at an observed frequency of 570 MHz. The observed spectral shape suggests that the turnover frequency is at ν ≤ 300 MHz. It is also worth noting that at lower frequency there may be contamination to the CSS flux density from the surrounding extended emission. Future high resolution LOFAR LBA observations will trace the spectrum at low frequency in more detail giving constraints on the spectral peak. Fig. 2 The integrated radio spectrum of the B2 0258+35 CSS source is shown. Red triangles represent the new LO-FAR flux density measurements at 120 MHz and 160 MHz while blue circles are taken from literature (Giroletti et al. (2005), NVSS, WENSS, NED) The morphology of the extended emission as seen by LOFAR at 145 MHz reflects what was previously observed at 1.4 GHz by Shulevski et al. (2012) (see Fig. 1). The linear extension of the lobes at low frequency is in agreement with the high frequencies and the surface brightness has an average value of 4.7 mJy arcmin −2 . The two broad enhancements in surface brightness observed in the Northern and Southern direction are clearly visible at both high and low frequencies but we cannot interpret univocally their origin. The extended emission that we see has likely been originated by large-scale jets that are currently switched off or smothered. As discussed in Sec. 2 the spectrum of an ageing plasma is expected to steepen because of radiative cooling to typical spectral indices α ≥ 1.2. Surprisingly, a preliminary spectral index study of the extended lobes of B2 0258+35 reveal a non-steep spectrum in the range 145-1400 MHz with values α 145 1400 = 0.6 − 0.8. Two scenarios can be invoked in order to explain our findings of a relatively "normal" spectral index for the extended lobes, when a much steeper spectrum typical of aged plasma may have been expected. On one hand the outer lobes of B2 0258+35 could be completely detached from the current central activity but they could manifest their spectral steepening only at frequencies higher than 1.4 GHz, therefore beyond our available spectral window. This occurrence has already been verified in a few remnant radio galaxies e.g. B2 1610+29 (Murgia et al. (2011)) and J1828+49 (Brienza et al. (2015)) and can be explained by a combination of remnant plasma energetics and possibly local environment. These sources show low magnetic field values equal to 3.2 µG and 1 µG respectively and do not suffer significant expansion energy losses so that the spectral shape at low frequency is still comparable to those observed in active sources. In this way, they remain detectable with spectral steepening occurring only at high frequency, for a long time after the jets switch off. On the other hand, the outer lobes of B2 0258+35 may still be fuelled by the current AGN activity and consequently the spectrum has not developed a high energy cutoff. Because we do not recognize any defined large-scale jets this could be happening via decollimated jets and likely at a low rate. The best example of this phenomenon is observed Centaurus A. This radio galaxy with clear signs of intermittent jet activity ) likely due to a merger event, shows a relevant difference between the northern and southern outer lobe. The first one indeed exhibits an intermediate-scale northern middle lobe ) which is interpreted as an "open channel" between the inner and the outer lobe through which fresh injected particles can reach the outskirts. This idea is supported by the fact the the entire northern lobe does not present any signature of spectral steepening up to 90 GHz (Hardcastle et al. (2009)). Although this last scenario might explain what we observe in B2 0258+35, the reason why the jets would have been smothered remains unclear. Struve et al. (2010) show indeed that the last merger the host galaxy has experienced goes back up to 1 Gyr, hence incompatible with the radio activity timescales. Conclusions and future perspective B2 0258+35 is a rare example of a CSS source with large low surface brightness radio lobes surrounding it. The extended lobes may be the remnants from a previous cycle of jet activity that arise due to intermittency of the central engine. LOFAR nicely detects this low surface brightness structure at 145 MHz and allows a first investigation of its spectral properties in combination with 1.4 GHz WSRT data. Contrary to expectations, we find that the extended lobes do not show very steep spectral index as predicted for old ageing plasma. If the lobes are actually remnants of a previous cycle of activity, the spectral steepening must occur at frequencies higher than 1.4 GHz. Otherwise they could still being fueled from the central engine via an open channel between the inner and outer lobes through which freshly injected particles can reach the outskirts. Upcoming 320 MHz and 5 GHz observations will provide new constraints on the radio spectrum of the extended emission enabling a more thorough spectral ageing study, from which firmer conclusions may be drawn. This study confirms the idea that large scale diffuse emission surrounding compact sources can show a variety of physical properties and may have experienced a variety of evolutionary paths. A clearer understanding of the spectral properties of the remnant plasma surrounding CSS, GPS and HFP sources is crucial for the identification and selection of this class of objects from the LOFAR survey. For this reason it is important to investigate in more details wellknown sources in preparation for all-sky selections. Upcoming complete samples of CSS, GPS and HFP with extended emission will finally allow a more systematic study of the young radio sources duty cycle.	2015-11-06T09:18:25.000Z	2015-11-06T00:00:00.000	{ "year": 2016, "sha1": "7a83ebba1c5e10072489867e6d20d1e8705c4080", "oa_license": null, "oa_url": "http://arxiv.org/pdf/1511.02011", "oa_status": "GREEN", "pdf_src": "Arxiv", "pdf_hash": "7a83ebba1c5e10072489867e6d20d1e8705c4080", "s2fieldsofstudy": [ "Physics" ], "extfieldsofstudy": [ "Physics" ] }
231857896	pes2o/s2orc	v3-fos-license	Environmental challenges induced by extensive use of face masks during COVID-19: A review and potential solutions The ongoing COVID-19 disease significantly affects not only human health, it also affects the wealth of country’ economy and everyday routine of human life. To control the spread of the virus, face mask is used as primary personal protective equipment (PPE). Thus, the production and usage of face masks significantly increase as the COVID-19 pandemic still escalating. Further, most of these masks contain plastics or other derivatives of plastics. Therefore, this extensive usage of face masks generates million tons of plastic wastes to the environments in a short span of time. This study aims to investigate the environmental impact induced by face mask wastes and sustainable solution to reduce this waste. An online survey was carried out to identify the types of face mask and number of masks used per week by an individual from 1033 people. Based on this survey and available literature, this study quantifies the amount of plastics waste generated by face masks. However, this survey was limited with certain ages, country and durations (July–August 2020). Thus, the prediction of plastic waste generation, only provide fundamental knowledge about the mask wastes. Results revealed that there is a huge plastic waste remained in land and marine environment in the form of mask waste, which will contribute to micro-plastic pollution. Therefore, this paper also highlights the sustainable approach to the mask production by integrating the use of natural plant fiber in the woven face mask technology to reduce the plastic waste induced by masks. Further, upcycling the mask waste and producing construction materials also discussed. Introduction The ongoing COVID-19 pandemic significantly induced uncertain environments for every human, business, education, job, and economy of each country. There is no viable meditation to prevent the spread of this deadly coronavirus disease ( Fadare and Okoffo, 2020 ). The use of personal protective equipment (PPE), social distance, travel restrictions and lockdown were currently employed to reduce this spreading level of coronavirus ( Rubio-Romero et al., 2020 ;Sun et al., 2020 ). This ongoing pandemic situation created that wearing mask is must for every human life. There are various types of masks such as surgical , N95 , and commercial fabric/cloth masks used to tackle the ongoing pandemic situation ( Fig. 1 ). According to the World Health Organization (WHO) study, in USA about 89 million medical masks are anticipated to be required to respond the COVID-19 as this crisis is likely to persist for some time ( Xiang et al., 2020 ). Further, the plastic innovation hub has identified Governments and publics have already begun to explore the alternative solutions including the reuse, reprocessing and disinfection of approved disposable masks, and producing biodegradable masks and homemade or non-certified masks ( Rubio-Romero et al., 2020 ). The effectiveness and impact of these alternatives are not yet understood and practiced by every people. Thus, this study conducts the online survey to identify the mask types, duration and disposal method being employed currently. Furthermore, a comprehensive review was conducted to find the material contents in the mask, impact on the mask wastes and suggest a sustainable upcycling solutions to the mask waste. Data collections Online survey and literature review were employed to achieve the objectives of this research. Following sub-sections describe the survey and review methods. Online survey The data is collected confidentially by conducting an online survey among different age groups (i.e. children (12-15), teenagers (16-25) and adults (26-65)) in many countries, particularly Australia, America, UK, Singapore, Sri Lanka and India. This survey was performed on a total of 1033 individuals for a period of one month (5th July 2020 -6th August 2020) at COVID-19 pandemic outbreak. In this survey, the questions were focusing following parameters: the types of mask (i.e. surgical mask, N95 mask, cloth mask or both); the amount of masks usage per week; and methods of mask disposal (i.e. wash and dispose, flush in toilets, put in appropriate bin, burning and throw away). The aim of survey is identifying the mask waste generation and provide the fundamental information about effect of the mask waste to the environment. Due to this, the survey did not focus mask waste generation by gender basis (male/female). This limitation could be creating the variation in the amount of mask waste generation. However, finding of this study provide the fundamental knowledge of mask waste generation, which will help to develop the waste control and management policies. Literature review The data collection approach of the current study was focused on the review of journal articles related to rapidly growing COVID-19. The data and knowledge on this particular topic are outstanding, showing the severity of the pandemic crisis. Such information is collected not only from scientific literature, but also from numerous reliable online resources, journals, and policy and media reports. The literature was gathered from 2nd February 2020 (at the peak of the first wave of COVID-19 pandemic) until the completion of this review (30th November 2020). The data collection concentrated on a following categories of: (a) precautionary measures used to monitor the forms of COVID-19; (b) the components of the mask; (c) problems relating to the disposal of the mask; (d) and sustainable solutions to address the effect of waste mask. Two reliable and detailed databases such as Scopus (scopus.com) and Web of Science (webofscience.com) were used to search for the relevant topics. As they provide high coverage of similar papers to the subject of this review and minimize the risk of missing any relevant document. To start with the searching process, a keyword of "COVID-19 or COVID 19 " from 2014-2020 was used to develop the final search criteria. The initial results of this search through Scopus and Web of Science (WoS) databases were 26920 and 48406, respectively. Additional limits were given for filtering the quantity of the publications to exclude the irrelevant sources. In that way, the findings decreased to 415 for Scopus and 19 for WoS using key terms such as "COVID-19 " AND "Masks " AND "Wastes". In addition, the results were restricted to 131 and 7 journal articles in Scopus and WoS, respectively by using the keywords "COVID-19 " AND "Masks " AND "wastes " AND "Environmental " as shown in Fig. 2 . Survey results and discussion Survey results and current issue from the mask waste are presented in the following sections. Survey results The data obtained from the online survey was confidentially analyzed in age-wise. Fig. 3 shows the percentage of people who participated in the online survey with different age group. For this survey people with age of 26-45 show more interested to participate . Fig. 4 shows that about 80% of people always wear masks and about 16% people occasionally wear the mask. This indicates that about 96% of individuals understand the importance of wearing the mask during the pandemic. However, 3% of them rarely used the mask, which may be due to lack of awareness and given less importance of the situation. About 1% of people have never used the mask due to their pre-existing medical condition. The type of masks used by people who participate in the survey was shown in Fig. 5 . This figure indicates that highest population (i.e. 40%) of people are used surgical mask for their personal safety. Whilst, the cloth mask is the second highest for use (34%) as it's cheaper than N95 mask and also be self-made. The N95 mask is highly recommended by WHO, only about 9% people use it as it is expensive than other mask types. Fig. 6 shows the quantity of waste mask generated by a person per week. This quantification may varies depending on the duration of mask usage, type of mask, degree of people's hygiene, place visited, etc. However, the results from this survey illustrate the fundamental quantification value of mask waste. The survey reveals that more than 25% of people generate 5 masks waste per week. Consequently, at least one mask waste was created by an individual per day. Mask wastes generated from each country (i.e. Sri Lanka, India, Australia, Singapore, UK and USA) was derived and shown in Fig. 7 . This was calculated based on the minimum (i.e. 1) and maximum (i.e. 5) number of masks used by individual ( Fig. 6 ) with the population of people at age between 16 and 65, which were obtained from census data ( Clark, 2020 ;Erin, 2020 ;Hirschmann, 2020 ;Plecher, 2020a , b , c ). Fig. 7 shows the fundamental idea of mask waste generation per week from Sri Lanka, India, Australia, Singapore, UK and USA. During this peak of the first wave of COVID-19 crisis most of these countries are went under lock down. Hence, the transportation in public and private vehicles, walking was very less, which may have affected the lower percentage of mask usage per week. Therefore, if the usage of mask is con- sidered worldwide with the whole population, it is expected to reach the peak level, which will eventually create a significant amount of plastic waste to the environment. A study by Akber et al. (2020) highlighted that the amount of polypropylene in surgical mask and N95 mask are 4.5 g and 9 g, respectively. Total maximum and minimum amount of polypropylene generated per week from different countries were calculated based on the mask wastes generation per week ( Fig. 7 ). Fig. 8 shows that minimum about 2.5 kt, 0.6 kt, 0.04 kt plastic (i.e. polypropylene) waste were generated, respectively from India, USA, Australia per week. This indicates that this Covid-19 pandemic will impact not only the health and economy, but it will also impact the sustainability of environment. In order to identify the mask disposal method, the survey question was generated with six general disposal methods: 1) disposed into the mixed waste bin; 2) disposed into the hazardous waste bin; 3) burning; 4) flush in toilets; 5) wash and disposed into mixed waste bin; and 6) throw away. Fig. 9 shows that about 34% and 11% individuals dispose the mask as the mixed waste and hazardous waste, respectively. This indicates that about 45% of them appear to be extremely responsible for the solid waste that they produce. Whereas 19% individuals recklessly throw away the face masks in the street and 12% of people are wash and dispose the mask. This has the potential to create the environmental issues due to its indecomposable nature. Consequently, has high possibilities in creating a future global warming issue. About 3% of people flush the mask in toilets while 10% people burn the mask. This 10% of individuals are responsible for the contamination of the air by the releasing adverse compounds into the atmosphere. Ultimately, this phenomenon could create chronic respiratory illness ( Stafford and Jones, 2019 ). Further, if the prevailing pandemic situation is to be continued, the people should adapt to live with the safety precaution, which can give them a healthy life. Thus, the mask production will be significant and increase the mask waste in near future. The mask is made of different types of plastic, which is not decomposable, and it induce the negative impact on the environments. Therefore, without disrupting the economy, any innovative steps can be taken with respect to the environment and the social well-being of people. On the other hand, instead of using different types of plastic raw material, steps should be taken to replace some decomposable raw material in production of face masks. In addition, the waste mask can be recycled by different methodologies and used as a supplementary material in any innovative products such as construction materials. Material contents in the masks The survey results highlighted that higher population of people are using surgical mask to protect their self from the COVID-19 virus. This surgical face masks ( Fig. 10 a) are made of non-woven fabric that has better filtration of bacteria and air permeability, while remaining less slippery than woven fabric. Based on the barrier properties and breathing resistance, the ASTM F2100 ( ASTM, 2007 ) and EN ISO 15223-1 (BS EN ISO 15223-1, 2016 ) classified this mask into different types (i.e. low barrier, moderate barrier and high barrier) ( Chellamani et al., 2020 ). This surgical mask mainly consists of three layers ( Fig. 10 b) such as outer hydrophobic non-woven layer (translucent), middle melt-blown layer (generally in white colour), and an inner soft absorbent non-woven layer (green, blue, or white colour). The polypropylene (i.e. known as plastic) is used as a major material to produce this surgical mask. However, other polymers like polystyrene, polycarbonate, polyethylene or polyester are also used to produce this mask ( Akber et al., 2020 ). The highly recommended mask to prevent the virous spreading is N95, which consists of four layers of material: an outer layer of spunbond polypropylene; a second layer of cellulose/polyester; a third layer of melt-blown polypropylene filter material; and an inner (fourth) layer of spun-bound polypropylene. The typical raw materials used to produce N95 mask is polypropylene similar as the surgical mask ( Barycka et al., 2020 ). Further, the weight of the surgical and N95 face masks is 3.5 g and 18.14 g, respectively. The ear loops of both face masks were made of natural and synthetic polyisoprene (i.e. latex-free) rubber. While, the ear loops in the cloth mask was made by cloth. There are mostly two kinds of cloth masks, "commercial cloth masks" and "homemade cloth masks" ( Santarsiero et al., 2020a ;Santarsiero et al., 2020b ). These cloth masks are made of multi layered of cloth like old t-shirts and sewing material ( Ayse et al., 2020 ). Cloth mask are also made of cotton and nylon, which have water resistant property. Furthermore, the fabric of cloth mask comes under the non-medical mask category, whilst the surgical and N95 masks are known as the medical mask. The filtration performance of these cloth masks are depends on many variables, such as thread count, number of layers, fabric type, and resistance to water ( Chughtai et al., 2020 ). The filtration efficiency of cloth mask is varying as some fabrics filter better than others. However, the filtration efficiency of cloth mask is lower than that of from medical mask. Impact of mask waste on environment The mask wastes are increased across the world as the people are not following the appropriate disposal methods for the used mask. Thus, it creates a new environmental challenge. Further, there are no appropriate mask or plastic waste collecting method specified in whole countries or part of the region in Sri Lanka, India, Pakistan and China ( Sangkham, 2020 ). This is adding a vast amount of plastic and plastic particle waste in the environment, which may end up in the streets and landfills. Besides, it gets into the waterways and reach the fresh water and marine water. This adds the presence of the plastics into the aquatic medium. The health and environmental effects of plastic and plastic particles due to the inappropriate disposal of facemasks were also highlighted by number of literatures ( de Sousa, 2020 ; Parashar and Hait, 2021 ;Sangkham, 2020 ). Furthermore, the production of the face masks also contributes the emission of CO 2 , which will potentially contribute to the global warming ( Liebsch, 2020 ). The processes of propylene, small aluminum strips and polypropylene in the production of N95 and surgical mask contributes the significant amount of CO 2 emission to the environments. Furthermore, production of fabric, sewing and weaving process of cloth mask fabrication also contributes the CO 2 emission to the environments ( Liebsch, 2020 ). The N95 mask production release 50 g CO 2 -eq per single mask, excluding the transportation process ( Kleme š et al., 2020a ). Surgical mask is embodied with 59 g CO 2 -eq per single and the highest share is from the transportation process ( Kleme š et al., 2020a ). Whilst, the cloth mask production contributes about 60 g CO 2 -eq greenhouse gas emission per single mask ( Kleme š et al., 2020a ). However, this would create a massive impact to the atmosphere because, millions of face masks are produced all over the world to control the pandemic situation. The face masks used by medical examiners in hospitals are carefully collected as its hazardous waste. A study was conducted in the UK and analyzed that if each individual uses one disposable surgical mask every day for a year, this would create over 124,000 tons of unrecyclable plastic waste 66,000 tons of contaminated waste and 57,000 tons of plastic packaging ( Ayse et al., 2020 ). However, there is currently no specific waste stream for these products if it used by the public. Mostly, it is thrown recklessly in the streets or collected as a mixed waste. In the handling of urban solid waste and hazardous medical waste, the pandemic has led to a significant challenge. The collected hospital face masks and other mixed waste are sent to the incineration and landfill. However, due to the existence of the plastics in the mask, such methodologies often have the potential to cause adverse environmental effects. Most plastics are chemically stable, resistant to corrosion and, difficult to degrade by microorganisms ( Webb et al., 2013 ). Yet they prefer to remain in the soil and pose environmental threats. The solution that allows the chemical energy content of plastics to be recovered for useful purposes is the incineration of medical waste coupled with waste heat recovery. For medical waste incineration, the WHO has suggested 900 °C and 1200 °C to guarantee safe destruction, but most of them are unaware of the temperature range ( Kleme š et al., 2020a ). However, with heat recovery, there are limitations to the widespread use of incineration. Public worries about dioxin and furan trace emissions can become troublesome. The transportation of those waste to relevant disposal site also consume energy and release greenhouse gases to the environment. Recent study by Kumar et al. (2020) stated that 10 ton of PPE waste including face masks travelled 10 km for the relevant disposal site resulted in total global warming potential (GWP) impact of 2.76 kg CO 2 -eq. The mask littered in the soil can impact the fauna in which it causes entanglement and can cause death ( Fig. 11 ). For instance, it is reported in Columbia that a bird was tangled in a discarded coronavirus facemask in a tree. Then died after the mask is wrapped around its body and beak ( Boyle, 2020 ). Further, when the mask is mistaken for food by an Fig. 11. Threat to birds due to usage of mask ( Boyle, 2020 ). animal, which is unfortunately a common occurrence, the plastic can fill stomachs, decrease food intake, cause animals to starve and die. The mask wastes also conveyed into the rivers and enters the fresh water and sea water. This creates the plastic pollutions to the aquatic medium ( Fig. 12 ). Marine plastic adsorbs toxins and organic contaminants, which ensures the pollutant particles bind as a toxic film to the surface of the plastic ( Williams-Wynn and Naidoo, 2020 ). As a result, it is also possible to poison the marine animals, which ingest plastic. It may destroy them directly, or weaken them, rendering them more vulnerable to other threats ( Ferraro and Failler, 2020 ). Ingested plastic can interfere with impair reproduction, growth and development of young ( Kleme š et al., 2020b ;Yang et al., 2020 ). Further, it can also cause entanglement, which lead to death in aquatic fauna like birds and other under water animals. Fragmentation of the macro plastic in the mask could occur due to various abiotic factors such as photodegradation, weathering, corrosion, and aquatic immersion forming the secondary micro plastics . Hence, bio accumulation of such microplastic occurs in the major food web to human existence and cause accumulation of toxins. This causes not only detrimental environmental effects, but also economic and social effects as shown in Table 1 . Solutions to reduce the mask waste Plastic is a significant continuing debate, which is not biodegradable material and induces further climate change pollution by affecting land and groundwater ( Thompson et al., 2009 ). To overcome this issue, different management and assessment approaches have been used such as incineration and landfilling ( Vanapalli et al., 2021 ). However, these are no longer preferred options to create the circular economy. Reduction of the plastic usage is completely not possible as it is an inevitable part of human behaviour, hence the search for another option to manage plastic waste is necessary. Therefore, governments are employing numerous international agreements to regulate plastic pollution. They include the Basel convention and its 2019 amendment, the United nations convention on the law of the sea, the international convention on the prevention of ship pollution, and the United nations global marine litter partnership ( Patrício Silva et al., 2020 ). • The clean-up activities, lifesaving activities of the aquatic bodies are expensive. • Tourism industry experiences significant loss. Reusing and recycling are viable options for plastic waste management, but prior to that identification of the condition of the plastic is required and cleaning or repairing steps are necessary according to its source ( Liang et al., 2021 ). The collected plastic wastes are shredded initially and sorted using the techniques like spectroscopy, X-ray fluorescence, flotation, magnetic or density separation. Then the optical sorter is used to differentiate the colour, and the separated plastics are melted and extruded into pellets for reuse. The recycled plastics are sold to local plastic manufacturing firms that can ultimately manufacture useful products such as engine oil, textile, footwears, and concrete additives. These process is financially not feasible as it is expensive, hence as an alternative way, the automated separation of various plastics could be adopted prior to the shredding ( Williams-Wynn and Naidoo, 2020 ). Furthermore, medical waste also contribute significantly to the global plastic waste ( Huang and Huang, 2007 ). These wastes should be given a special consideration for disposal as it might be hazardous or radioactive. As a first step, the classification of hospital waste at its origin is done for the management of waste, which has the possibilities to avoid the spread of infection to the waste handlers. The waste is collected in separate bins or bags with a clear identification and then the waste containing bags must be disinfected and sealed with doublelayered plastic bags (yellow color) prior to transportation from the origin. According to WHO, the biomedical waste includes 10% of infectious hazardous waste, and 5% of toxic or chemical waste (Ilyas et al., 2020). The commonly used disinfection techniques for the management of hospital wastes are pyrolysis technique and microwave technique. The temperature is sustained at around 540°C-830°C in high-temperature pyrolysis, which involves pyrolysis-oxidation, plasma pyrolysis, inductionbased pyrolysis, and laser-based pyrolysis ( Ilyas et al., 2020 ). This effectively offers advantages such as low emission rate, reduction of inert residual volume by up to 95% and reduction of mass by up to 90% ( Ilyas et al., 2020 ). Another technique used for the medical waste is the medium microwave technique, which works within the temperature range of 177-540 °C. The key benefits of the microwave technique are comparatively lower energy and action temperature, limited heat loss, and less environmental burden without harmful residue during the disinfection phase ( Haque et al., 2021 ). There is a dramatic rise in the generation of plastic waste during the COVID-19 due to the increased use of face masks and change in customer preference to single use mask due to the hygienic problem ( Patrício Silva et al., 2020 ). Therefore, many advanced waste management methods with minor variations in the general treatment of medical waste have been developed ( Sangkham, 2020 ). Shortly after the first outbreak of COVID-19 in South Korea, 'Special Safe Waste Management Measures against COVID-19 was enforced on 28 January 2020 ( Sangkham, 2020 ). COVID-19 waste cannot be kept for more than 24 h as per its guidelines and must be incinerated on the same day of collection. Therefore, most of the COVID-19 waste is sent for incineration at a temperature above 1100 °C ( Sangkham, 2020 ). The microwave technique is often used in conjunction with autoclaving in the case of COVID-19 waste disinfection, where steam is used for sterilization (in the temperature range of 93-177 °C). The chemical disinfection technique is commonly used in combination with prior mechanical shredding for the pre-treatment of COVID-19 waste. The polypropylene (i.e. plastic) used to produce the mask is according to the technical standards ( Huang and Huang, 2007 ). Currently, this mask is discarded after the usage, as they are sensitive to heat and cannot tolerate high heat. However, due to the huge demand of the face masks caused by the COVID-19 crisis, the lifespan of the usage of face masks is investigated for the reuse ( Chua et al., 2020 ). Reuse can be done using a strategy called mask rotation. In this strategy, rotation of the mask used each day can be carried out, allowing them to dry for long enough periods so that the virus is no longer viable ( > 72 h) ( SAGES Webmaster, 2020 ). Proper storage for this technique requires either hanging the face masks to dry, or keeping them in a clean, breathable container like a paper bag between uses ( Vanapalli et al., 2021 ;Williams-Wynn and Naidoo, 2020 ). Re-processing/decontamination is carried out by various methods like moist heat, dry heat, UV treatment, hydrogen peroxide(H 2 O 2 ) vaporization. The H 2 O 2 , dry heat techniques have the opportunity for reprocessing of personal protectives (N95 face masks) and their re-use ( Ilyas et al., 2020 ). The recycling the mask by the appropriate processes is one of the alternatives used to reduce the plastic pollution generated by mask waste. Broadly, there are two ways for recycling such as primary recycling and secondary/chemical recycling. Primary recycling is the reusing of the product in their original structure. In the secondary recycling, the mask consisting of thermoplastic can be reused ( Lackner, 2015 ). As they can be re-melted and re-processed into various end products; either to ( Chellamani et al., 2020 ); and (b) plant fibre ( Misnon et al., 2014 ). produce the same product or composite products. Hence, this method can be employed to recycle the mask in a useful manner. However, considering the cost of a brand-new mask, this recycled mask is expensive for processing, also the filtering efficiency and the quality of the recycled mask is less than that of a new mask ( Chua et al., 2020 ). Therefore, it is important to further examine for other alternatives to reduce the mask wastes. Biodegradable mask is one of the modern sustainable alternative options to reduce the mask that induces plastic waste. The polypropylene in the mask can be replaced with other substitute organic and biodegradable materials with similar mechanical, physical and chemical properties such as light weight, high tensile strength, ecological safety, low cost and high biodegradable potential ( Glukhikh et al., 2020 ;Samper et al., 2018 ;Siracusa and Blanco, 2020 ). The bioplastic and biodegradable polymers can be an option to replace the polypropylene. Bioplastic is a type of biodegradable plastic derived from biological substances rather than conventional plastic made of petroleum. The production of the mask using bioplastics must follow the standard requirements such as elasticity, water resistance and filtering properties. In general, due to the existence of polypropylene, all these properties can be found in bioplastics ( Duc et al., 2011 ). This biodegradable plastic reduces the CO 2 emissions by 30%-70% compared with the conventional one ( Lackner, 2015 ). The biodegradable polymers can be obtained from different families like biomass production from agro resources such as polysaccharides (starches, lignocellulose), proteins, lipids and micro-organisms. Natural fibers such as cactus, banana, avocado, lotus, sisal, straw, hemp, maize bamboo, hemp, coffee and sugar cane have capability to meet the standard requirements to produce the mask ( Luhar et al., 2020 ;Ramesh et al., 2017 ;Yan et al., 2016 ). Furthermore, Fig. 13 a shows the woven technology used in the filter of the surgical mask. Similar filters can be made with biodegradable plant fibers ( Fig. 13 b). Polysaccharides such as bamboo, hemp, coffee and sugar fibers are used to develop a bio mask and it prevents pollution caused by mask disposal and they can be processed and recycled ( Ho, 2020 ;Layt, 2020 ;Staff, 2020 ). Tea leaf waste also can be used to produce the filter part of the mask as it contains polypropylene and enhances the properties of poly lactic acid (PLA) in bioplastics ( Ferraro and Failler, 2020 ). This indicates there is a possibility to use natural fibers waste and raw natural fibers to produce a cost-effective sustainable biodegradable green mask. Sugar cane waste mask ( Layt, 2020 ), Coffee-Based Face Masks ( Ho, 2020 ) and Hemp Fiber mask ( Staff, 2020 ) are currently available biodegradable mask. These types of biodegradable masks are manufactured with 99.99% dual antibacterial technology and have the high filtration capacity. Currently many research studies are going on developing the required specification for biodegradable mask to bring it to the standard of medical mask ( Choi et al., 2021 ;Santarsiero et al., 2020a ;Santarsiero et al., 2020b ). Although biodegradable face masks are made, individuals who are unaware of these benefits may not appear to alter their behavioral patterns from single-use plastics. Such masks are unpopular because they are only made in limited countries. Whilst the biodegradable face mask made from bioplastics claims to be degradable, there is still an uncertainty about the complete degradable nature of the biodegradable mask whereas a number of limitations still restrict it. For example, some plastics such as PLA require industrial scale composter to biodegrade ( Vanapalli et al., 2021 ). Consequently, it is also mandatory and costly to build the appropriate infrastructure for biowaste management. Therefore, more investigations and research studies are needed, to assess the effectiveness and lifespan of sustainable biodegradable mask. The use of recycled waste face masks as a composite material in various applications from all other alternatives to solve the global plastic crisis (i.e. mask pollution) would be a better approach in terms of circular economy. This is not only promotes the removal of collected waste from landfill sites, but also helps to reduce the production of waste in the first place. Substantially, it is restricting the use of natural resources and limiting the effects on the environment. A novel approach to reduce the plastic waste is convert this into construction materials. Different types of recycling methods such as mechanical recycling, chemical recycling and incineration have been used for the extraction of polypropylene from plastic wastes. In mechanical recycling, the polypropylene is separated from other forms of resin, washed to remove dirt and contaminants. Then grinded and compressed to minimize the particle size of the plastics, accompanied by heat extrusion and reprocessing into new plastic products. Chemical recycling allows plastic waste to be turned into initial monomers or another useful chemical ( Ahmad et al., 2015 ;Canopoli et al., 2020 ;Matei et al., 2017 ;Poulakis and Papaspyrides, 1997 ;Rahimi and García, 2017 ). The extraction of the pure polypropylene from the plastic waste containing various types of polymers could be achieved by pyrolysis method because it does not require pure plastics. The polypropylene pyrolysis was carried out using a two-stage continuous process equipped with auger and fluidized bed reactors linked in sequence ( Park et al., 2020 ;Park et al., 2019 ). There have been different forms of reactors, such as stirred, screw kiln, circulating sphere, fluidized bed, spouted bed, plasma, and microwave-assisted reactors for pyrolysis ( Park et al., 2019 ). It is stated that -scission, end-chain -scission, and intermolecular and intramolecular hydrogen transfer reactions are essential reaction mechanisms of pyrolysis ( Park et al., 2020 ). A study by Ali et al., 2011 performed polypropylene pyrolysis blended with petroleum residue and coal to obtain a transportation fuel. Most of propylene are developed by steam cracking of naphtha or low-molecular hydrocarbons ( Park et al., 2020 ). These extracted recycled polyethylene and polypropylene derivatives were used in road construction as a partial replacement for asphalt/bitumen ( Williams-Wynn and Naidoo, 2020 ). There has been a great deal of research on the replacement of bitumen with plastics, use of plastics in asphalt-concrete mixtures ( Appiah et al., 2017 ;Awoyera and Adesina, 2020 ;Bahij et al., 2020 ;Biswas et al., 2020 ;del Rey Castillo et al., 2020 ). A study by Hama and Hilal, 2017 concluded that the addition of fine, coarse, and mixed plastic wastes to self-consolidating concrete (SCC) enhanced its fresh properties, such as passing capability and filling capacity, at 12.5% replacement level of plastic by weight of fine aggregate. Furthermore, Khalid et al., 2018 investigated the performance of ring-shaped polyethylene terephthalate (PET) wastes in fibrereinforced concrete beams. Their study showed that there was no significant impact on the failure mode of plastic fibers applied to concrete, but it improved the mechanical properties of the beams in terms of load at first crack and strength. A study by Arulrajah et al., 2017 also investigated the feasibility of using recycled plastic waste granules (i.e. linear low density polyethylene filled with calcium carbonate (LDCAL), high density polyethylene (HDPE) and low density polyethylene (LDPE)) blends with crushed brick and reclaimed asphalt pavement for road construction. Results showed that the introduction of plastic waste above 5% for road construction material, decreases the blends' stiffness, bearing capacity and robust module. However, the required performance was still achieved, and it has been concluded that the use of plastic waste as aggregate in asphalt improves the pavement's skid and crack resistance ( Asi, 2007 ). A study by Kumi-Larbi et al. (2018) investigated the potential of using low-density polyethylene (LDPE) sheets plastic waste in sand blocks. Their study showed that strong and durable sand blocks can be produced without the need of water. Furthermore, recent study by Aciu et al. (2018) used recycled plastic waste to make sustainable mortars with 75% of plastics as partial sand replacement to obtain M20 grade masonry mortar. The latest invention from mask waste has proved strongly that the face mask can be used for sustainable brick manufacturing ( Adlakha, 2020 ). The sustainable brick was produced with 52% mask waste, 45% paper waste and 3% binder. This brick is alternative to the traditional masonry bricks, which have potential to reduce the plastic waste and cost. Existing technologies used to extract the plastic derivatives from waste such as mechanical recycling, chemical recycling, incineration and pyrolysis have potential to separate the plastic particle in the mask wastes. The separated plastic derivatives from the mask have the potential to partially replace the cement and aggregates in the construction materials. This is strong, durable, waterproof, light weight, simple to mould and recyclable. Therefore, it can be used in the construction sectors. Further, it can be utilized as air and moisture barriers films and sheets used in insulating the building wraps, industrial adhesive tapes, plastic parts including pipes, masonry bricks. However, there is limited research available as it is a new waste. Therefore, more research needed to ensure the performance, efficiency and economic of construction materials or other products produced from this mask generated plastic wastes. Summary and conclusions This study analyses the number of different types of face masks usage in Australia, America, UK, Singapore, Sri Lanka, and India, through a public survey analysis. The study reveals that the number of face masks used during the COVID-19 pandemic. Results will help to understand the fundamental inside knowledge of mask waste generation and type of mask. These additional enhanced face masks containing plastic contributed to micro-plastic pollution in the aqua environment and also significantly impact the soil. A detailed study was therefore carried out to identify the difficulties of using more face masks and preventive measures. This paper highlighted the sustainable approach by integrating the use of natural plant fiber in the woven face mask technology to reduce the plastic waste induced by face masks. Further, upcycling this mask waste and producing construction materials such as artificial aggregates, light weight plastic block and production of ecological mortar can be a viable solution in the near future to reduce the plastic waste and environmental and health impact. 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 article.	2021-02-10T14:08:18.129Z	2021-02-10T00:00:00.000	{ "year": 2021, "sha1": "b18563e89aa776b7ee370f267aec662a1dd8957b", "oa_license": "CCBYNCND", "oa_url": "https://doi.org/10.1016/j.envc.2021.100039", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "6c6e31b8e3dee4a9d9a727895d2de227ae5fc71d", "s2fieldsofstudy": [ "Engineering" ], "extfieldsofstudy": [ "Medicine" ] }
259373932	pes2o/s2orc	v3-fos-license	Three‐dimensional genome structure and function Abstract Linear DNA undergoes a series of compression and folding events, forming various three‐dimensional (3D) structural units in mammalian cells, including chromosomal territory, compartment, topologically associating domain, and chromatin loop. These structures play crucial roles in regulating gene expression, cell differentiation, and disease progression. Deciphering the principles underlying 3D genome folding and the molecular mechanisms governing cell fate determination remains a challenge. With advancements in high‐throughput sequencing and imaging techniques, the hierarchical organization and functional roles of higher‐order chromatin structures have been gradually illuminated. This review systematically discussed the structural hierarchy of the 3D genome, the effects and mechanisms of cis‐regulatory elements interaction in the 3D genome for regulating spatiotemporally specific gene expression, the roles and mechanisms of dynamic changes in 3D chromatin conformation during embryonic development, and the pathological mechanisms of diseases such as congenital developmental abnormalities and cancer, which are attributed to alterations in 3D genome organization and aberrations in key structural proteins. Finally, prospects were made for the research about 3D genome structure, function, and genetic intervention, and the roles in disease development, prevention, and treatment, which may offer some clues for precise diagnosis and treatment of related diseases. INTRODUCTION The three-dimensional (3D) genome structure plays a crucial role in determining cell fate and maintaining normal cellular functions. 1 In the late 19th century, the existence of chromatin territory and euchromatin/heterochromatin in the cell nucleus was proposed based on findings from optical microscopes and chromatin dyes, 2 and Hans Winker introduced the term "genome" in 1920. 3 With the development of fluorescent in situ hybridization, the individual chromosomal territories could be directly observed. 4,5 Chromosome conformation capture was first reported in 2002, which employs formaldehyde-induced chromatin cross-linking and restriction enzyme digestion to connect spatially proximate DNA fragments. 6 The interaction frequency between two specific loci in the genome is detected in a "one versus one" format, using polymerase chain reaction or next-generation sequencing technology. 6 Circular chromosome conformation capture facilitates the detection of specific DNA fragments' interactions with other genomic regions. Designing primers for the detected fragment achieves a "one versus all" format, allowing for the detection of all genome regions. 7,8 In 2009, genome-wide chromosome conformation capture (Hi-C) was reported and can detect all chromatin interactions throughout the genome, referred to as "all versus all". 9 Furthermore, remarkable advancements in optical and electron microscopy technologies have enabled direct observation of chromatin structures. The development of super-resolution fluorescence imaging techniques, such as stimulated emission depletion and structured illumination microscopy, has been especially notable. 10 By combining these techniques with specific DNA probes, the chromatin conformation of particular genomic regions within individual cells can be observed, providing new insights for exploring gene expression regulation and other intricate biological processes within the cell nucleus. 11,12 DNA sequence undergoes a series of complex compressions and folding to ultimately form chromosomes. 13 A chromosome maintains a specific location within the nucleus, denoted as a chromosome territory (CT). 14 These chromosomes are organized in a structured manner within the nucleus and can be classified into euchromatin and heterochromatin based on gene transcriptional activity. 15,16 At the subchromosomal level, A/B compartments, with an average size of 3−5 Mb, correspond to euchromatin and heterochromatin, respectively. 17,18 Compartments at the megabase level can be further divided into different topologically associating domains (TADs). 19 TAD represents the fundamental units of 3D genome structure and function. [20][21][22] High-frequency interactions between enhancers and promoters within TAD regulate cell-specific expression of developmentrelated genes. 22 The CCCTC-binding factor (CTCF) at TAD boundaries isolates aberrant regulatory interference, ensuring efficient transcription. 23 At a smaller scale (<2 Mb), chromatin loops often link enhancers and promoters, playing a crucial role in the regulation of gene transcription. 15 Structural variations or structural protein abnormalities can lead to disease by disrupting TAD structure and affecting cis-regulatory element function. [24][25][26] These may be critical pathological mechanisms for many diseases, such as congenital diseases and cancers. 27,28 Here, we tried to make a systematic review of the 3D genome structure, function, and relationship with diseases. First, the structural characteristics of mammalian 3D chromatin organization were summarized. Then, the effect and mechanisms of cis-regulatory elements' interaction in the3D genome for regulating spatiotemporally specific gene expression, the role and mechanisms of dynamical changes of 3D chromatin conformation during embryonic development were discussed. Subsequently, the pathological mechanisms of some diseases, such as congenital developmental abnormalities and cancer, which attribute to alterations in 3D genome organization and aberrations in key structural proteins, were systematically reviewed. Finally, prospects were made for the research about 3D genome structure, function, and genetic intervention, and the roles in disease development, prevention, and treatment. F I G U R E 1 Chromosome 3D structure. In mammalian cells, chromosomes nonrandomly occupy specific regions within the interphase nucleus, known as chromosome territory (CT). Chromosomes can be classified into euchromatin and heterochromatin based on their transcriptional activity and regulatory roles. Euchromatin contains numerous transcriptionally active genes and is loosely arranged, while heterochromatin is tightly organized and primarily comprises transcriptionally silent genes. Heterochromatin at the nuclear periphery interacts with nuclear lamins, giving rise to Lamin-associated domains (LADs). At the megabase level, the A/B compartment is further subdivided into TAD. Covering approximately 90% of chromatin structure, TAD represents the basic unit of 3D genome function and structure. Within TAD, high-frequency interactions between enhancers and promoters regulate cell-specific expression of developmental genes. CTCF, located at TAD boundaries, effectively isolates aberrant regulatory information interference and ensures smooth transcription. 3D GENOME STRUCTURE The hierarchical structure of the 3D genome, from high to low levels, includes CT, A/B compartment, topologically associated domain, and chromatin loop ( Figure 1). Chromatin and chromosome territory In mammalian cells, chromosomes are formed from approximately 2 m of DNA sequences through a series of compressions and folding events. [29][30][31] Each nucleosome encompasses around 146 base pairs of DNA, giving rise to chromatin fibers that subsequently condense and fold to create chromosomes. 21,32 These chromosomes nonrandomly occupy specific regions within the interphase nucleus, known as CT. 13,22 Chromosomes with similar characteristics tend to cluster and are influenced by size, gene density, and transcriptional activity. [33][34][35] The distribution of chromosome territories can vary among different cell types. [36][37][38] For instance, differences in nuclear shapes between human fibroblasts and lymphocytes result in distinct radial distributions of chromosome territories. 39 Moreover, interactions between CT boundaries may be associated with disease-related chromosomal translocations. 14 Chromosomes can be classified into euchromatin and heterochromatin based on their transcriptional activity and regulatory roles. 9 Euchromatin, comprising numerous transcriptionally active genes, exhibits a loosely organized structure, whereas heterochromatin, characterized by transcriptionally silent genes, displays a tightly packed arrangement. 40 Heterochromatin at the nuclear periphery interacts with nuclear lamins (lamin A/C, B1, and B2), giving rise to Lamin-associated domains (LADs). [41][42][43][44] Over 1000 LADs exist in humans and mice, with a median size of approximately 0.5 Mb, comprising 30−40% of the entire genome. 45,46 LADs are characterized by heterochromatin histone modifications H3K9me2 and H3K9me3 and predominantly exhibit low or silent gene expression. 47 LADs can be further categorized into constitutive LADs (cLADs) and facultative LADs (fLADs), based on their conservation across cell types. 48 cLADs retain a relatively stable position in the genome and display high conservation among cell types. 46 In contrast, fLADs demonstrate variability across cell types, associate with cell type-specific gene expression regulation and functions, and experience dynamic changes during the cell cycle, cellular differentiation, and circadian rhythms. 49,50 A/B compartment At the sub-chromosomal level, A/B compartments correspond to transcriptionally active and silent chromatin. 51 These compartmental structures have an average size of 3−5 Mb. 17 The A compartment is situated near nuclear speckles regions and modified by active histones, while the B compartment locates at the nuclear periphery and is modified by inactive histones. 15 Sites within the same compartment exhibit more frequent interactions. 15 Highly spatially plastic, compartments undergo extensive A-B switching during embryonic stem cell differentiation. Genes associated with A-B compartment switching demonstrate increased transcriptional activity throughout development. 16 Topologically associating domain At the megabase level, A/B compartments are further divided into different TADs. 52 In mammalian genomes, TADs tend to be conserved across species and different cell types, 15,53,54 covering approximately 90% of chromatin structure and serving as the fundamental unit of 3D genome function and structure. 55 In most cases, TAD formation precedes gene expression. 56 In the human genome, TADs emerge during the eight-cell stage when the interactions between regulatory elements within TAD are low and gradually increase throughout embryo development. TADs span approximately 0.1-1 Mb, and up to a maximum of 2 Mb, containing one or more genes. 52 Smaller sub-TAD structures reside within TADs and are associated with tissue-specific gene expression. [57][58][59] Although earlier studies have suggested that TADs function as a unit at the cell population level, recent studies indicate that TADs are chromatin structures in individual cells, exhibiting globular conformations and sharp structural domain boundaries. 60,61 TADs boundaries are co-localized with housekeeping genes, cohesin, and CTCF. 53,54 CTCF is located at TAD boundaries and effectively compartmentalizes external regulatory information. 53,54 As a highly conserved zinc finger protein, CTCF plays crucial roles in transcriptional activation/repression, insulation, and the formation of higher-order chromatin structures. 62,63 Due to the DNA sequences that bind CTCF being asymmetric, CTCF located at TADs boundaries has directionality. 64,65 In over 90% of TADs, CTCF sites are situated at boundaries in a convergent orientation. 66 CRISPR-edited inversions of CTCF binding sequences disrupt TAD structure and cause abnormal gene expression, suggesting that CTCF orientation is important for TAD. 67 Multiple CTCF binding sites at TAD boundaries ensure precise gene expression through synergistic and redundant effects. [68][69][70] Recent studies showed that six CTCF binding sites exist between Epha4 and Pax3 TADs, and are located adjacent to each other. Deletion of a single CTCF binding site can alter Pax3 expression level without causing observable phenotypes.However, deletion of multiple CTCF binding sites leads to TADs fusion and abnormal finger development in mice. 70 TAD boundaries are highly conserved and remain relatively stable during organ development and tissue differentiation. 16,23,71 However, human genome analysis suggests that although the vast majority of CTCF loci are conserved, thousands of CTCF loci may be associated with tissue-specific gene expression. 72 For instance, CTCF function and insulating activity are positively correlated with the level of enhancer-promoter interactions within TAD. 73 CTCF at TAD boundaries is more conserved than CTCF located within TAD, with stronger binding, more open chromatin, and lower DNA methylation levels. 74 Furthermore, TAD boundaries containing a single gene are more conserved than TAD boundaries that contain multiple genes. 75 Chromatin loop Chromatin loop is relatively stable, ring-shaped structure formed by chromatin within 3D space, often containing regulatory elements such as enhancers and promoters. 76 In 2014, Rao and colleagues discovered approximately 10,000 chromatin loops through the analysis of in situ Hi-C conducted on human lymphoblastoid cells. 15 Most of these loops are conservation across various species and cell types. 15 Genes that form chromatin loops exhibit higher expression levels compared with those did not form chromatin loops, indicating the involvement of chromatin loops in regulating gene expression. 15 Furthermore, an analysis of 24 cell lines derived from three human germ layers revealed that approximately 28% of chromatin loops varied among different cells. 77 These tissue-specific chromatin loops are strongly associated with the tissue-specific gene expression. 77 The "loop extrusion" model, mediated by cohesin-CTCF cooperation, aptly describes the formation of chromatin loops and TADs 78 ( Figure 2). In this process, CTCF and cohesin exhibit distinct roles. 78 CTCF acts as a barrier to prevent overextension of cohesin. 79 Cohesin is a ring-shaped DNA-entrapping adenosine triphosphatase (ATPase) complex that works as a molecular motor. 80 Human cohesin comprises three subunits: SMC1, SMC2, F I G U R E 2 Mechanism of loop extrusion. Cohesin acts as a molecular motor in the "loop extrusion" model: After binding to chromosomes, cohesin moves in two opposite directions along chromatin fibers and extrudes DNA loops until it contacts the target CTCF site in the converging direction. During TAD formation, cohesin undergoes functional changes by interacting with various regulatory factors. The cohesin loading factor, NIPBL, loads cohesin at specific DNA sites and facilitates cohesin's translocation on chromosomal fibers. Cohesin's release from chromosomes requires the PDS5 subunit to recruit the cohesin-releasing factor, WAPL. The unloading efficiency of WAPL is stronger than that of PDS5, and both participate in the cohesin release process. Finally, cohesin ensures its smooth arrival at the target CTCF site through direct interaction with CTCF. and Scc. 81 Cohesin moves along the chromatin fiber with a speed of 1.5 kb/s until it interacts with convergently oriented CTCF sites. 82 This process lasts 10−30 min. 83 Scc1 plays an essential role in loop extrusion by binding to several regulators, such as NIPBL and PDS5. 84 The cohesin loading factor NIPBL loads cohesin at specific DNA sites, promoting cohesin translocation across chromosome fibers. [85][86][87] In addition, NIPBL can activate ATPase to supply energy for the high-speed movement of cohesin on chromatin fibers. 88 Cohesin release from chromosomes requires PDS5 to recruit WAPL, a cohesinreleasing factor. 89 The unloading efficiency of WAPL is stronger than that of PDS5, and both participate in the process of cohesin release. [89][90][91] Cohesin ensures its smooth arrival at the CTCF site by directly interacting with CTCF. 92 The N-terminal 77-aa region of CTCF can directly interact with the cohesin Scc1-SA2, providing a structural and functional basis for the precisely anchoring of cohesin to CTCF binding sites. 93 The functional changes of the aforementioned cohesin subunits or complexes affect the formation of TAD. For example, deletion of the Scc1 subunit leads to the disruption of TAD, 75,94 while depletion of WAPL leads to a more than 20-fold prolongation of the duration of cohesin action, resulting in the formation of a larger chromatin loop, which increases the amount of TAD and/or sub-TAD. 90,95-97 MECHANISMS OF 3D GENOME REGULATING GENE EXPRESSION: CIS-REGULATORY ELEMENTS INTERACTION IN 3D GENOME In the human genome, protein-coding sequences constitute merely 2% of the DNA sequence, 98 and gene expression is subject to regulation by various noncoding structures, such as promoters and enhancers within F I G U R E 3 Mechanisms of enhancer-promoter interactions. Within TADs, cis-regulatory elements play crucial roles in gene expression. Cohesin compresses DNA to form chromatin loops, thereby shortening the distance between enhancers and promoters. The promoter is the transcription start site for mammalian gene expression, determining the location and direction of transcription. The core promoter is a DNA sequence situated approximately 50 bp upstream and downstream of the transcription start site, containing multiple general transcription factor binding sites, and represents the minimal sequence required for initiating gene expression. Enhancers are significant non-coding elements, typically located in nucleosome-depleted regions, and contain multiple transcription factor binding sites. the 3D genome, and does not operate in isolation ( Figure 3). 99,100 Promoter The promoter serves as the starting point for mammalian gene expression and determines the location and direction of transcription. 101 The core promoter is a DNA sequence located approximately 50 base pairs upstream and downstream of the transcription start site, containing multiple universal transcription factor start sites, and is the minimal sequence that initiates gene expression. 102,103 The promoter has some transcriptional activity, but it is relatively weak. 104,105 In specific cases, promoters can act as enhancers to regulate gene expression by interacting with other promoters. 106,107 The transcription preinitiation complex(PIC) assembles at the promoter and binds to the promoter sequence, a prerequisite for gene transcription. 108,109 The structure and function of the PIC vary between stages, and the PIC consists of universal transcription factors (TFIIA, TFIIB, TFIID, TFIIE, TFIIF, TFIIH), RNA polymerase II (RNAP II), and other factors in most cases. 110 These general transcription factors play distinct roles in regulating gene expression, with TFIID recognizing and binding to promoter DNA sequences and TFIIB directing other transcription factors and RNAPII to bind to promoter sequences. 111 RNAPII undergoes three stages of transcription initiation, pause, and pause release at the promoter, 112 with RNAPII pausing after about 60 bases of transcription and restarting transcription. 113 The exact mechanism of this process is unclear, and current studies suggest that RNAPII transcription is influenced by various factors. For example, p300/CBP promotes RNAPII transcription initiation and pause release, whereas TFIID is associated with RNAPII pause. 114,115 Enhancer Enhancers are important noncoding elements that regulate the tissue-specific expression of genes. 116 Active enhancers typically reside in nucleosome-deficient regions and contain multiple transcription factor binding sites. 117,118 The regulatory function of enhancers is generally direction-independent, with enhancers separated from their cognate promoters by approximately 10 4 −10 6 base pairs. 119 Cohesin facilitates enhancer-promoter interactions by shortening the distance between enhancers and promoters through chromatin loop formation. 120 A single enhancer can act directly with its cognate promoter to regulate target gene expression, 121 while multiple enhancers can act on the same target gene and regulate its transcription through synergistic or cumulative effects. [122][123][124] In addition, superenhancers span longer DNA sequences, often containing multiple common enhancers within them, and can drive higher expression levels of target genes with increased activity. 125 Shadow enhancers are a class of enhancers that partially or completely overlap in expression patterns. 126 During mammalian development, especially under conditions of physiological or genetic stress, these enhancers with similar expression patterns are able to resist genetic variation and keep relatively stable gene expression levels. 127 These seemingly "redundant" enhancers provide a guarantee of accurate gene expression. 128 Mechanisms of enhancer-promoter interactions Chromatin loops can shorten the distance between enhancers and promoters, but the interaction between them also requires the involvement of transcription factors and multiple cofactors. 129 Enhancer sequences contain multiple transcription factor binding sites and are able to specifically bind to a variety of transcription factors. Pioneer transcription factors are able to efficiently degrade nucleosomes by binding to DNA sequences through an alpha helix structure, 130 resulting in a chromatin-open state. 131,132 Transcription factors recruit p300/CBP and mediator complexes. 133 P300/CBP, a transcriptional activator with acetyltransferase activity, directly activates enhancers through histone acetylation. 134 The mediator complex serves as a functional hub between the enhancer and the promoter and consists of four parts: head, middle, tail, and kinase module (CDK8 module). 135,136 This complex is capable of transmitting genetic information from the enhancer to the promoter. 137 First, the tail of the mediator interacts with transcription factors at the enhancer, whereas its head and middle of the mediator are involved in RNAP II recruitment and assembly of the transcriptional preinitiation complex. 138 The kinase module (CDK8) phosphorylates the carboxy terminus of RNAP II to facilitate transcription initiation. 139 In conjunction with transcription factors and cofactors, enhancers deliver genetic information to the promoter for precise expression of target genes. MECHANISMS OF DYNAMICAL CHANGE OF 3D GENOME AMONG DIFFERENT FUNCTIONAL STATES The 3D genome undergoes dynamic changes during cell differentiation and organ development to regulate interactions between cis-regulatory elements, exhibiting diverse functional states. In this section, the mechanisms of these dynamic changes in the 3D genome are summarized. Chromatin compartment switches During mammalian cell differentiation, chromatin compartments exhibit dynamic changes at various stages and significantly influence the spatiotemporal-specific expression of development-related genes. 140 Throughout the differentiation of human embryonic stem cells (hESCs), substantial transitions transpire between compartments A and B. [141][142][143] The number of genomes participating in A-B compartment switches varies among cell types, and the number of TADs involved declines as differentiation progresses. 144,145 For instance, when ESCs differentiate into skeletal muscle progenitors, approximately 20% of the genome participates in the A-B compartment switches. In contrast, only 6.5% of genomes are involved in compartmentalization when skeletal muscle progenitor cells further differentiate. 144,146 Among the 10 marker genes associated with preadipocyte differentiation, nine are observed to convert from compartment B to A during differentiation. 147 Similarly, Prdn1 and Atf4, related to plasma cell fate determination, are transferred to euchromatin, whereas Ebf1, an inhibitor of their differentiation, transitions from euchromatin to heterochromatin. 148 Dynamical change of TAD During mammalian development, TAD facilitates precise gene expression by providing a suitable environment while minimizing interference from external regulatory information. 149 TAD-D (∼250 kb) and TAD-E (∼500 kb) are adjacent TADs within the mouse X chromosome inactivation center, containing mutually repressed genes Tsix and Xist, respectively. 54 Xist is associated with X chromosome inactivation, while Tsix represses the expression of Xist. 150 In male mice, embryonic stem cell differentiation (mESCs), Tsix is transcriptionally active, but Xist is barely expressing. 151 In female mESCs, although Tsix is also transcriptionally active, Xist transcript levels gradually increase as mESCs further differentiate, eventually leading to X chromosome inactivation. 152 In male mESCs, a 40 kb inversion across the TAD-D and TAD-E boundaries places the promoters of Tsix and Xist inside each other's TAD, respectively. This alteration in the regulatory environment enables enhancers to interact with nontarget promoters, causing promoters that should regulate Tsix gene expression to be ectopically activated within TAD-E, ultimately leading to aberrant Xist expression in male mESCs. 56 The HoxD gene clusters locate between two adjacent TADs: the telomeric side (T-DOM) and the centromeric side (C-DOM). During mammalian limb development, HoxD genes are sequentially regulated by spatiotemporalspecific enhancers within these two TADs. [153][154][155] The T-DOM TAD contains multiple enhancers that regulate proximal limb development by controlling the expression of Hoxd9, Hoxd10, and Hoxd11 genes. 156 In contrast, during subsequent distal limb development, enhancers within T-DOM are silenced, whereas enhancers within C-DOM are activated. 157 Enhancer II1, one of the enhancers in C-DOM responsible for distal limb development, plays a crucial role in tetrapod distal limb development by binding to the transcription factor Hox13. When the enhancer II1 sequence is inserted into T-DOM, its activity is almost entirely suppressed, even though the transcription factor Hox13 remains explicitly bound to the sequence within enhancer II1. 158,159 The regulatory activity of enhancer II1 in the distal limb is restored only after most of the sequence at the mitotic end of T-DOM (Mtx2-II1-T-DOM) is deleted. 160 Thus, promoter and enhancer activities relate to the chromatin environment during gene expression. 157 TADs, as higher-order chromatin structures, provide an appropriate regulatory environment for enhancer-promoter interactions, ensuring that development-related genes are expressed in a cellor tissue-specific manner. During cell differentiation, sub-TADs gradually form within TADs, promoting cis-regulatory elements to precisely regulate target gene expression. 161,162 Runx1, a transcription factor, is associated with hematopoietic development and is situated within a 1.1 Mb TAD, which forms prior to Runx1 gene expression. 163,164 Analysis of mESCs reveals the emergence of two sub-TADs within the Runx1 TAD, promoting tissue-specific enhancer-promoter interactions. 165 Similarly, observation of α-globin protein motifs within erythrocytes at different stages of differentiation in mice shows the appearance of sub-TADs during cell differentiation. 166,167 Furthermore, a progressive increase in chromatin accessibility, accompanied by a corresponding rise in the frequency of enhancer-promoter interactions within TADs, is observed during the differentiation of various cell types. 166 For instance, due to changes in enhancer activity, frequent chromatin loop reconnections are observed during the differentiation from preadipocytes to adipocytes. 168 ROLE OF DYNAMICAL CHANGE OF 3D GENOME IN GAMETOGENESIS AND EARLY EMBRYOGENESIS Meiosis plays a crucial role in germ cell development, facilitating the maturation of sperm and oocytes. 169,170 Throughout this process, germ cells undergo morphological transformations, and their chromosomes experience intricate 3D structural alterations, ultimately leading to highly differentiated germ cells. 171 In this section, the dynamic changes of the 3D genome in gametogenesis and early embryogenesis are systematically reviewed. Role of dynamical change of 3D genome in spermatogenesis Meiosis I and II enable the development of spermatogonia into spermatozoa, 172 involving homologous chromosome separation and sister chromatid separation to form round spermatids. 169,173 Subsequently, these round spermatids undergo a series of complex, dynamic changes to generate mature spermatozoa. 174 During meiotic prophase, although A/B compartments are still present, TAD disappears due to alterations in cohesin activity in mouse spermatocytes. [175][176][177] In postmeiotic stages, the level of genomic compartmentalization in round spermatocytes gradually increases. Mature spermatocytes exhibit chromatin higher-order structures such as compartments and TADs, demonstrating frequent remote interactions between TADs. 178,179 Comparing Hi-C data from mature sperm cells, fibroblasts, and ESCs reveals considerable similarity in their 3D genomes. 180 As protamine replaces over 85% of histone loci in mature spermatozoa, chromatin becomes highly compressed, leading to a more than 10-fold reduction in sperm cell nuclear volume compared with fibroblasts. 180 In contrast, retained histones are highly enriched at the promoters of crucial developmental genes. For instance, nucleosomes modified by active histones were observed at the transcription start sites of genes such as EVX1/2 and ID1. 174,181 Moreover, studies on mouse sperm reveal that mature sperm contain over 10,000 enhancers and over 500 superenhancers. 180,182 The vast majority of these are identical to those in mESCs, suggesting that regulatory elements are ready for tissue differentiation and cell fate decisions as early as in the sperm. 177 Role of dynamical change of 3D genome in oogenesis In contrast to sperm development, oocyte formation begins during the embryonic stage. 183,184 Primordial germ cells, originating from the proximal epiblast, differentiate into oogenic cells within the embryonic gonads. These oogenic cells undergo a series of differentiations, eventually forming primary oocytes that enter meiosis I prophase. Upon reaching the diplotene stage of meiosis I, the primary oocytes constitute primordial follicles, which are subsequently stored in the ovary. During puberty, the primordial follicles gradually transition hormonally into germinal vesicle oocytes. 185 These germinal vesicle oocytes continue to undergo hormone-stimulated meiosis and remain suspended in the metaphase of meiosis II until fertilization. 185 During oogenesis, the 3D genome structure undergoes a series of dynamic changes, characterized by the gradual disappearance of compartment, TAD, and loop. 186,187 Meiosis I oocytes and germinal vesicle oocytes exhibit similar numbers of TADs, whereas metaphase II (MII) oocytes display an almost complete absence of TADs and A/B compartments, resulting in a uniform chromatin folding pattern. 188 Role of dynamical change of 3D genome in zygote After fertilization, two transcriptionally quiescent gametes fuse to form a zygote, characterized by a more relaxed chromatin structure. [189][190][191] Transcription remains inactive in zygotes and early embryos. 189,192 Concurrent with zygotic genome activation (ZGA), gene expression is activated, and an extensive reorganization of the 3D genome structure occurs. [193][194][195] Mouse embryo studies reveal a progressive increase in TAD intensity and the extent of chromatin remote interactions at the two-cell stage. 196,197 In contrast, human embryos exhibit compartments and TADs as early as the eight-cell stage, with continuous enhancement throughout development. 181,198 Although the onset of ZGA often coincides with TAD formation in many species, recent research indicates that TAD formation associates with DNA replication and that inhibiting ZGA does not impede TAD formation. 195 ROLE OF DYNAMICAL CHANGE OF 3D GENOME IN ORGAN DEVELOPMENT In addition to playing a crucial role in gametogenesis, dynamic changes in the 3D genome are also important for regulating organ development. 6.1 Role of dynamical change of 3D genome in brain development The mammalian brain, a highly complex organ, is primarily composed of neurons and neuroglial cells. 199 Glial cells include astrocytes, oligodendrocytes, and microglia. 200 Neurons can be classified as either excitatory or inhibitory. 201,202 Apart from microglia, all other cells are derived from neural progenitor cells (NPCs), which strictly regulate the spatiotemporal distribution of gene expression to modulate physiological functions such as memory, cognition, and emotion. 203,204 During brain development, ESCs first differentiate into NPCs, which subsequently differentiate into neurons and further into various cell subtypes. 199 Throughout this process, chromatin experiences extensive reorganization and compaction. 205 Initially, the chromatin of ESCs is globally accessible, but as differentiation progresses, it becomes increasingly condensed, reducing accessibility. 206 Concurrently, at the compartment level, the A compartment size decreases by approximately 5% compared with ESCs as neural cell differentiation continues. 16,71 Despite this change, TAD boundaries do not exhibit significant alterations; the TAD landscape remains stable in NPCs, neurons, and glial cells. 207 As differentiation ensues, TAD sizes undergo a slight increase, interaction levels between TADs in the A compartment decrease, and those in the B compartment increase. 71 Additionally, human-specific brain structures, such as the formation of the subplate layer, are associated with the development of human-specific TADs and chromatin loops. 208,209 These human-specific TADs are generally smaller than conserved TADs between species due to chromatin structure alterations, and their boundary CTCF binding strengths are marginally weaker. 208 Human-specific enhancers play a crucial role in brain development by effectively regulating target gene tissuespecific expression. [210][211][212] For instance, the human-specific enhancer EOMES controls FGFR2 expression, which is associated with NPCs proliferation. 206 EPHA7 is regulated by an upstream human-specific enhancer, and its inactivation impacts the development of human cortical pyramidal cells, 208 ARHGAP11B collaborates with a human-specific enhancer located 500 kb away, contributing to the expansion of the human neocortex. 213 6.2 Role of dynamical change of 3D genome in cardiac development The heart, originating from the mesoderm, has chromatin regulatory mechanisms in its development process that have long captivated researchers. Primarily composed of cardiomyocytes, the heart derives these cells from cardiac progenitor cells, which in turn arise from distinct subgroups of cardiac mesodermal cells. 214 Cardiomyocyte-specific compartments form early during heart development and differentiation. 215 Studies reveal that in adult mouse cardiomyocytes, approximately 66.7% of cardiac-specific genes reside in compartment A. 216 During cardiomyocyte differentiation, around 20% of the genome undergoes compartment conversion, with genes transitioning from compartment B to A demonstrating greater cardiac specificity. 217-219 Moreover, as hESCs or human induced pluripotent stem cells (hiPSCs) differentiate into cardiomyocytes, a decrease in the number of TADs, loss of boundaries, and reduction in intensity are observed. 218 Roughly 70% of TADs remain stable throughout development, suggesting their importance in heart development through the promotion of enhancerpromoter interactions and regulation of developmental gene expression. For instance, Handsdown (Hdn) and Hand2, which are located within the same TAD and closely related to heart development, play a significant role in cardiomyocyte differentiation and heart development. 220,221 Hdn encodes a lncRNA and modulates Hand2 transcription by inhibiting its upstream enhancer activity. Hdn deficiency results in an abnormal increase in Hand2 expression, leading to the thickening of the right ventricular wall. 222 Furthermore, structural protein abnormalities associated with TAD formation may cause heart development abnormalities of varying degrees. 214,223,224 For example, STAG2, one of the cohesin subunits, has its deficiency affecting the proliferation and migration of secondary heart field progenitor cells, leading to delayed heart development and morphological defects. 225 A heterozygous deletion of the adhesion loading factor NIPBL may induce atrial septal defects. 226,227 6.3 Role of dynamical change of 3D genome in blood system development The development and differentiation of hematopoietic stem cells (HSCs) provide a valuable model for 3D genomics research. During embryonic development, HSCs initially appear within the principal arteries and subsequently migrate to the liver. 228 Before birth, HSCs relocate to the bone marrow, where they remain for an extended period. In adult mouse bone marrow, more than 70% of HSCs exist in a quiescent differentiation state. 229 Studies have demonstrated that as ESCs differentiate into HSCs, the spatial arrangement of TADs experiences dynamic shifts coinciding with the transition of A-B compartments. 230,231 Within TADs, chromatin accessibility progressively increases, and sub-TADs form gradually. 165,166 Furthermore, research on mouse embryos and adult mouse HSCs has revealed that compartments and TADs are predominantly conserved, with a mere 5% of compartments and 12% of TADs undergoing alterations. 232 In adult mouse HSCs, both compartmentalization and TAD boundary strength augment, whereas chromatin accessibility declines. Approximately 52% of enhancerpromoter interaction levels vary, correlating with differential gene expression. 232 Erythropoiesis refers to the biological process wherein hematopoietic stem cells and progenitor cells differentiate into mature red blood cells. 233 Erythroid progenitor cells proliferate and undergo a sequence of morphological transformations, ultimately yielding enucleated reticulocytes. 234 Throughout erythroid cell differentiation, widespread chromatin condensation occurs, and both chromatin accessibility and transcriptional activity diminish. 235,236 Heterochromatin significantly compacts, and H3K9me3 is repositioned, leading to numerous longrange interactions. 237 Approximately 58% of TADs are disrupted, and TAD boundaries weaken considerably during the final stages of erythroid development. TADs with active chromatin modifications are partially conserved, and GATA1 is thought to participate in the preservation of TADs. 237 6.4 Role of dynamical change of 3D genome in immune system development B cells and T cells both derive from hematopoietic stem cells in the bone marrow. 238 Within this environment, common lymphoid cells either differentiate into B cells and innate lymphoid-like cells or migrate to the thymus, where T cell differentiation is initiated. The development of B and T cells is intimately associated with the recombination of antigen gene loci. Throughout B cell development, the dimensions, quantity, and positions of the majority of TADs remain relatively constant. 239 However, significant gene loci associated with B cell fate determination exhibit dynamic changes. The Ebf1 gene locus encodes an essential B cell transport protein and serves a crucial regulatory function in B cell differentiation. 240 At the pluripotent stage, the Ebf1 locus resides in the transcriptionally repressed B compartment. 241,242 As cells transition from pre-pro-B cells to pro-B cells, Ebf1 relocates to the transcriptionally active A compartment, playing a role in B cell fate determination. 239 During the differentiation of B cells into plasma cells, Ebf1 is repositioned to the heterochromatin region. 148 T cell development relies on Bcl11b expression. In the pluripotent stage, Bcl11b is situated within the B compartment, with its activation closely tied to the noncoding RNA ThymoD. 243,244 ThymoD facilitates local DNA demethylation through transcription, promoting CTCF-DNA binding and cohesin protein recruitment. This process contributes to Bcl11b-TAD formation and strengthens the interactions between enhancers and promoters. 245 3D GENOME STRUCTURAL VARIATIONS AND DISEASE The 3D genome structure is pivotal in gene expression regulation. Pathological conditions can lead to extensive reorganization of the higher-order chromatin structure, which affects the expression of functional genes. The switching of aberrant A/B compartments at the subchromosomal level is closely linked to the onset and progression of the disease. Chromatin structural variations (SVs) at the megabase scale modify the TAD structure and interactions between regulatory elements, resulting in abnormal gene expression levels. Additionally, the formation of disease-specific chromatin loops alters enhancer-promoter interactions, exacerbating gene expression dysfunction and contributing to disease development. This section systematically reviews the dynamic reorganization of higher-order chromatin structures at different levels during the onset and progression of diseases. This may provide valuable insights into the complex relationship between chromatin conformation and diseases (Table 1). Compartment and disease The abnormal conversion of A/B compartments under pathological conditions may result in dysregulated gene expression, playing a crucial role in the onset and progression of diseases. 246 Nonalcoholic fatty liver disease (NAFLD) is a chronic liver disease primarily observed in obese individuals, characterized by hepatic fat accumulation. 247 NAFLD increases the risk of liver cirrhosis and hepatoma. 248 Studies on NAFLD-induced mice have revealed that around 33% of the genome undergoes A/B compartment conversion. Sugct and Fgfr2 are examples of genes transferred to the A compartment, linked to lipid metabolism disorders and tumor development potential, respectively. [249][250][251] This finding enhances our understanding of NAFLD pathogenesis, disease progression, and the pathological mechanisms that make patients susceptible to liver cancer. 251 The metastatic potential and invasiveness of prostate cancer are also intimately connected to dynamic genome alterations. Recent in vitro investigations on prostate cancer metastasis have revealed that during cancer progression, there is a widespread genome compartment conversion, which results in significant changes in the nuclear chromatin activation environment, leading to increased mixing and interaction in the A compartment. 252 Forty-eight genes have been transferred from transcriptionally repressed B compartments to A compartments and activated, including genes associated with prostate cancer progression and poor prognoses, such as WNT5A, CDK44, and TMPRS22. These research findings reveal potential key factors in the pathogenesis and progression of prostate cancer. 252 Additionally, viruses can reshape the host's 3D genome structure, which affects gene expression and is closely linked to disease progression. The Hepatitis B virus (HBV) is a leading cause of hepatocellular carcinoma. 253 Integration of HBV DNA into the human genome results in the remodeling of the 3D genome structure. 254,255 During this process, there are dynamic changes in the A/B compartments, mainly in the genomic regions on chromosomes 9, 13, and 21. Most genes are moved from transcriptionally repressed B compartments to A compartments, where they are subsequently activated. 256 These regions have many transposable elements that promote viral DNA replication. A smaller subset of regions moves from the A compartment to the B compartment, including enhancers and genes associated with the regulation of inflammatory responses. 256 Research on human cells infected with SARS-CoV-2 shows that about 30% of the genome undergoes compartment switching, which is accompanied by a decrease in chromatin interaction levels within the A compartment. This may be linked to the high incidence of acute sequelae resulting from SARS-CoV-2 infection. 257 TAD and disease Approximately 5% of the human genome exhibits structural variation. 258 SVs include both balanced and unbalanced rearrangements, which lead to altered expression levels of functional genes by impacting the TAD structure. This results in aberrant enhancer-promoter interactions, contributing to the development of numerous human diseases (Table 1). 259 Copy number variation comprises chromatin deletions and duplications, whereas inversions and translocations represent balanced rearrangements. 25 This process is distinct from chromatin remodeling during cell cycle progression, which occurs in the nucleus The proto-oncogene MIR31HG-TAD is adjacent to CDKN2A-TAD, and sequence deletions involving TAD boundaries can result in the fusion of the two TADs. The protooncogene MIR31HG promoter is aberrantly regulated by the enhancer, resulting in an increase in its expression level. 299 Topologically associating domain Sox9 Kcnj2 Cooks Syndrome Duplications spanning the Sox9-TAD and Kcnj-TAD give rise to "neo-TAD"., and result in an enhancer otherwise regulating Sox9 expression incorrectly interacting with the Kcnj2 promoter in the "neo-TAD". 304 (Continues) Disease-specific chromatin loop formation has been observed in AML, involving oncogenes. The specific interaction between the MYCN promoter and enhancers situated 650 kb downstream is related to AML onset. 379 without concomitant changes. 173,260 Copy number variants within the same TAD can alter gene expression by influencing the amount of regulatory elements or the frequency of interactions. 261 Structural variations between TADs can affect gene expression by disrupting enhancer-promoter interactions or creating ectopic interactions ( Figure 4). Additionally, recent studies have shown that structural variation is a significant mechanism contributing to cancer. 262,263 Genomic rearrangements cause transcriptional dysregulation of proto-oncogenes and oncogenes by changing interactions between cis-regulatory elements in somatic cells, ultimately leading to abnormal proliferation and differentiation of somatic cells. [264][265][266] 7.2.1 Deletions and duplications within TADs lead to diseases Copy number variations within TADs significantly modulate gene expression by changing the quantity of regulatory elements or the frequency of interactions between them. Such changes can cause the dysregulation of gene expression, contributing to various human diseases and developmental disorders ( Figure 4A). [267][268][269][270][271][272] Sox9 is located within a 2 Mb TAD, 273,274 containing multiple tissue-specific enhancers that participate in mammalian sex determination, craniomaxillofacial development, and chondrogenesis. 268,275-277 Located 1.45 and 1.25 Mb upstream of the Sox9 gene, remote enhancer clusters Ec1.45 and Ec1.25 regulate its tissue-specific expression in the mandibular process and the first branchial arch region. 267 Deletion of Ec1.45 and/or Ec1.25 in the Sox9-TAD leads to reduced Sox9 expression levels, causing Pierre-Robin sequence, 267 a group of craniomaxillofacial developmental malformations such as mandibular dysplasia, cleft palate, and tongue recession. 278,279 Specifically, deletion of Ec1.45 in the human genome can cause a decrease of over 50% in the expression level of Sox9, 267 whereas simultaneous deletion of both appears to have a more substantial impact on Sox9 expression. 280 The Ihh gene participates in skeletal development, and its expression is regulated by at least nine enhancers (i1-i9) located within Ihh-TAD. 281 Deletion of Ihh in mouse models causes joint fusion and skeletal shortening, 282 whereas Ihh duplication is associated with finger deformities and premature closure of cranial sutures. 283,284 Deleting enhancers i2-i9 within Ihh-TAD in mice results in a 98% reduction of Ihh mRNA expression levels, leading to abnormal cranial ossification, reduced bone cortex, and shortened extremities. 281 In contrast, duplication of i2-i9 enhances Ihh-TAD endointeraction, increasing the expression of Ihh to five-fold in the head and limbs, which results in premature closure of the cranial suture and syndactyly. 281 It is worth noting that not all intra-TAD fragment duplications result in increased enhancer-promoter interactions and target gene expression levels within the TAD. Alterations in enhancer-promoter distances can influence interaction frequency and, as a result, impact gene expression. 285 2q31 syndrome, characterized by facial malformations, mental retardation, and limb deformities, 286 can result from sequence duplication or deletion within HoxD-TAD. 287,288 This occurs because sequence duplication increases the distance between the enhancer and F I G U R E 4 Structural variation located between TAD. (A) Sequence deletions or duplications within TADs primarily affect regulatory functions by altering the number of enhancers and causing abnormal expression of target genes. (B) Sequence deletions at the boundaries result in the merging of adjacent TADs into one, forming a new TAD, referred to as "TAD fusion." Duplication of DNA sequences containing boundary elements can create a "neo-TAD." The "neo-TAD" is situated between the original TADs, wherein the interacting enhancers and promoters originate from different initial TADs, and their interactions do not interfere with the enhancer-promoter interactions within the original TADs. Chromosomal inversions occurring between adjacent TADs alter the position and/or orientation of DNA segments, placing genes and/or regulatory elements in different chromosomal contexts, and causing pathological effects known as "TAD shuffling". gene, which weakens the interaction between the HoxD promoter and its enhancer, resulting in a 50% downregulation of the HoxD gene cluster transcript levels. In this case, sequence duplication within HoxD-TAD leads to similar changes in gene expression levels as deletion. 288 7.2.2 Deletions between TADs lead to diseases Copy number variations across TAD boundaries can disrupt chromosomal rearrangements of boundary elements. Deletions at the boundaries can merge adjacent TADs, resulting in a "TAD fusion", 258 and can lead to abnormal gene expression levels as enhancers from different TADs interact with promoters ( Figure 4B). 289 Structural variation has been identified as a crucial mechanism underlying cancer development. 262,263 Genomic rearrangements can cause transcriptional dysregulation of proto-oncogenes and oncogenes by altering interactions between cisregulatory elements in somatic cells, leading to abnormal proliferation and differentiation. [264][265][266] Congenital diseases An example of chromosomal rearrangement that impacts gene expression is the Wnt6/Ihh/Epha4/Pax3 locus located on chromosome 2q35-36. A heterozygous deletion of 1.75-1.9 Mb in the 2q35 region results in short-fingered malformation in humans and mice. This deletion disrupts the TAD boundary between Epha4 and Pax3, leading to TAD fusion and producing an 800 kb fused TAD. 258 Within this fused TAD, the enhancer that initially regulated Epha4 interacts with the Pax3 promoter, causing an increased expression level of Pax3 and a decreased expression level of Epha4, ultimately leading to the development of short-fingered malformations. Not all deletions across TAD boundaries lead to TAD fusion. In the mouse genome, adjacent motifs Sox9-Kcnj show that deleting only the CTCF locus at the boundary does not result in TAD fusion. TAD fusion occurs only after deleting all four CTCF loci within the TADs. Only deleting all four CTCF loci within the TADs leads to TAD fusion, but it does not significantly affect gene expression. 290 The limited impact on gene expression resulting from small deletions may be due to the redundancy of CTCF sites in the TADs. This redundancy mechanism helps maintain the structural and functional stability of TADs and ensures precise gene expression. Cancers The oncogene MYC is a critical downstream target gene of Notch1 signaling. 238,291 Its expression levels are frequently elevated in patients with T-cell acute lymphoblastic leukemia (T-ALL). 292 Recent studies have shown that deletions of TAD boundaries in T-ALL patients result in the fusion of MYC-TAD with adjacent TADs, leading to abnormal interactions between the enhancer BDME/BENC and the MYC promoter within the fused TAD. This results in increased expression of MYC and the development of T-ALL. 293 CFTR and WNT2 genes are located in adjacent TADs on chromosome 7. Research on patients with intestinal neoplasia pedigrees has shown that a 121.1 kb heterozygous deletion on 7q31.2 disrupts the border between CFTR and WNT2 and deletes the CFTR promoter sequence. 294 Consequently, the enhancer located in introns 1, 10, and 11 that originally regulates the tissue-specific expression of CFTR interacts with the WNT2 promoter, resulting in increased levels of WNT2 expression and decreased levels of CFTR expression in this pedigree. [294][295][296] This dysregulation is associated with the development of intestinal adenocarcinoma and small intestinal neuroendocrine tumors. 295,297 The majority of patients with pancreatic ductal carcinoma present a homozygous deletion of CDKN2A. 298 Recent studies have shown that MIR31HG-TAD is adjacent to CDKN2A-TAD in various pancreatic ductal carcinoma cell lines. Deletions in the boundary regions of the two TADs result in their fusion. In the fused TADs, the MIR31HG promoter is abnormally regulated by the enhancer, leading to increased expression levels. 299 7.2.3 Duplications between TADs lead to diseases Duplication of DNA sequences containing boundary elements can result in the formation of "neo-TADs", [300][301][302] which are located between the original TADs. The enhancers and promoters within these "neo-TADs" originate from different initial TADs, but their interactions do not interfere with the enhancer-promoter interactions in the original TADs ( Figure 4B). Congenital diseases Cooks syndrome is linked to chromosomal duplications and is characterized by nail hypoplasia and short-fingered malformations. 303 The duplication of the Sox9-TAD and Kcnj-TAD creates a "neo-TAD", as evidenced by RNA-seq results from mouse limbs at various developmental stages that show no changes in the expression levels of Sox9 and Kcnj16, but an increase in Kcnj2 expression. 304 This rise in Kcnj2 expression is a result of misinteraction between the enhancer that originally regulated Sox9 expression and the Kcnj2 promoter within the "neo-TAD". Retinitis pigmentosa is a common inherited retinal disease characterized by progressive peripheral vision loss and night blindness, which can lead to blindness in severe cases. 305 The disease is associated with a chromosome 17q22 duplication spanning the YPEL2 and GDPD1 genes. YPEL2 is highly expressed in the brain and retina, 306 whereas GDPD1 encodes glycerophosphodiesterase and is predominantly expressed in the brain and testis. 307 This duplication leads to the formation of new interaction domains, where enhancers originally located in YPEL2-TAD regulate GDPD1 expression ectopically. 307 Elevated GDPD1 expression levels can cause dysregulation of lipid metabolism, a pathogenic factor contributing to photoreceptor cell inactivation and the development of retinitis pigmentosa. [308][309][310] Cancer EGFR and LINC01446 are situated in adjacent TADs on chromosome 7p11.2. In cell lines derived from glioblastoma patients, a sequence duplication was observed between EGFR and LINC01446, generating a "neo-TAD" between the two adjacent TADs. Within the "neo-TAD," the enhancer initially intended to regulate LINC01446 expression aberrantly interacts with the promoter responsible for regulating EGFR expression, leading to an increase in EGFR expression levels. This is one of the pathological mechanisms that contribute to the development of glioblastoma. 311 7.2.4 Inversion between TADs leads to diseases Balanced chromosomal rearrangements, including inversions and translocations, occur between spatially adjacent or separated TADs, changing the position and/or orientation of DNA segments. 27 Balanced rearrangements cause pathological effects by placing genes and/or regulatory elements in different chromosomal environments, a phenomenon known as "TAD shuffling". 312 TADs that experience balancing rearrangements result in abnormal gene expression due to disrupted enhancer-promoter interactions, enhancer hijacking effects, or positional effects ( Figure 4B). 27 Congenital diseases The mechanisms described above can explain the etiology of chromosomal inversions causing congenital diseases such as branchiooculofacial syndrome and Liebenberg syndrome. Branchiooculofacial syndrome is a rare developmental defect caused by heterozygous deletions or mutations in the TFAP2A gene. 313,314 In patients with branchiooculofacial syndrome, an 89Mb inversion was found with a breakpoint located in the TAD, which disconnects the TFAP2A gene from enhancers such as Enh100 and Enh105, leading to haploinsufficient expression of TFAP2A in human neural crest cells. 315 Liebenberg syndrome is characterized by a heterozygous leg-arm transformation, where the upper limb exhibits morphological features typically seen in the lower limb. 316,317 Pitx1 is expressed in the hindlimbs, pituitary gland, and first-gill arch, and gene rearrangements involving Pitx1 are associated with the disorder. 318,319 During limb development, Pitx1 is regulated by enhancers RA4 and RA5, in both the anterior and posterior limbs. 320 However, enhancer Pen specifically regulates Pitx1 expression in the hindlimbs, and there is no interaction with Pitx1 during forelimb development. 318,320 In Liebenberg syndrome, an inversion of a 113-kb fragment containing enhancers Pen and RA4 results in the placement of Pen in the position of RA4. This misplacement causes Pitx1 to be activated by Pen during forelimb development, leading to the development of phenotypes such as radial curvature, patellar heterotaxy, and shortened ulnar hawk in adult mice. 320 Cancers Medulloblastoma is a highly malignant tumor of the central nervous system that commonly occurs in children. [321][322][323] Recent studies suggest that chromosomal inversions may significantly contribute to the development of this disease. 324 The PRDM6 gene, encoding a histone transferase, is located approximately 600 kb downstream from SNCAIP at 5q23. Chromosomal inversion leads to the interaction between the super-enhancer responsible for regulating SNCAIP and PRDM6, resulting in a significant increase of approximately 20-fold in PRDM6 expression levels. 262,324 7.2.5 Translocations between TADs lead to diseases Congenital diseases Chromosomal translocations can also cause developmental defects through pathological mechanisms akin to inversions. MEF2C is considered one of the crucial pathogenic genes in the 5q14.3 microdeletion syndrome, 325,326 which is associated with developmental brain malformations, epilepsy, and intellectual deficits. 327,328 A 1 Mb complex translocation between 5q14.3 and 3q24 disrupts the TAD structure, disconnecting MEF2C from the enhancer that regulates its expression. 329 Chromosomal translocations reduce MEF2C expression to 50% compared with controls, 329 a level comparable to that observed in humans with MEF2C heterozygous deletions. 330 Cancers A chromosomal translocation involving 11q13 and 14q32 has been identified in multiple myeloma and mantle cell lymphoma, and is associated with the CCND1 and IGH gene loci. 331 The proto-oncogene CCND1, situated on chromosome 11, encodes a cell cycle protein, while the IGH gene is located on chromosome 14. 332 This translocation allows the superenhancer, which initially regulates IGH, to aberrantly activate CCND1 expression, leading to an up to 500-fold increase in its expression levels. 333 Acute myeloid leukemia (AML) is linked to aberrant expression of the stem cell factor EVI1. Chromosomal translocations involving 3q21 and 3q26.2 have been observed in AML patients. 334 These rearrangements permit the upstream enhancer of the GATA2 gene to relocate and aberrantly interact with the EVI1 promoter, leading to increased EVI1 expression and GATA2 haploinsufficiency. As a result, hematopoietic stem cell growth and differentiation are impeded, contributing to the development of AML. 335 Angiocentric glioma, a low-grade malignant glioma, is associated with the MYB-QKI rearrangement. 336 The proto-oncogene MYB and the oncogene QKI reside at 6q23.3 and 6q25.3, respectively. Chromosomal translocations lead to the enhancers Q3SE1 and Q3SE2, which initially regulate QKI transcription, aberrantly activating the proto-oncogene MYB expression. Consequently, the fusion protein MYB-QKI, a proto-oncoprotein, contributes to the development of angiocentric glioma. 337 Chromatin remodeling can contribute to cancer development by altering the local 3D genomic structure. Human papillomavirus (HPV) genes can integrate into the human genome, facilitating cervical cancer development through TAD structure remodeling. 338,339 PEG3 and CCDC16, genes co-located within the same TAD on chromosome 19, are impacted by this alteration. 340,341 HPV reshapes the TAD structure, dividing it into two unequal TADs. As a result, enhancers initially regulating PEG3 aberrantly activate the proto-oncogene CCDC16 expression, contributing to cervical cancer development. 338 7.2.6 Abnormalities in structural proteins lead to diseases The "loop extrusion" model, mediated by CTCF, cohesin, and their regulators, effectively elucidates TAD formation and remote enhancer-promoter interactions within TAD. 342 Previous research has demonstrated distinct functions of cohesin and CTCF in TAD formation. 78,79 Abnormalities in cohesin or CTCF function can result in aberrant gene expression due to alterations in higher chromatin structure. Mutations in genes encoding cohesin subunits or regulatory factors can disrupt TAD formation and chromatin interactions, affecting normal gene expression. 343 These mutations, which impact cohesin activity and function, are termed "cohesinopathies". 344 Cornelia de Lange syndrome (CdLS) is characterized by intellectual disability, microcephaly, growth retardation, and upper limb deformities. 345 CdLS pathogenesis is linked to cohesin dysfunction. 344,345 NIPBL mutations are found in 65% of CdLS patients. 346,347 Normally, the cohesin loader NIPBL introduces cohesin into the promoter of highly expressed genes, 84,348,349 facilitating its movement along chromatin fibers. 87,350 In primary fibroblasts derived from CdLS patients, cohesin is still loaded at specific sites, but its chromatin fiber binding stability is reduced. 351 The NIPBL mutation decreases cohesin mobility and ultimately affects the DNA loop extrusion process and TAD formation. 351 CTCF is involved in forming chromosomal higher-order structures and plays a crucial role in cell differentiation and apoptosis. [352][353][354] Located at TAD boundaries, CTCF avoids ectopic contact with enhancers while promoting high-frequency enhancer-promoter interactions within the TAD. However, CTCF is highly sensitive to methylation levels, 355,356 and abnormally elevated methylation levels at CTCF-DNA binding sites cause disruption of the TAD boundaries, 357 leading to a decrease in enhancer-promoter interactions within the TAD and an increase in inter-TAD interactions. 78 This mechanism was first identified in the study of isocitrate dehydrogenase (IDH) mutant gliomas. IDH mutant glioma patients have increased levels of DNA methylation, 357,358 resulting in ectopic interaction between the enhancer located 50 kb upstream of FIP1L1 and the proto-oncogene PDGFRA promoter, leading to a threefold increase in the expression level of proto-oncogene PDGFRA and promoting glioma cell proliferation. 359 Similarly, in regions of pathogenic short tandem repeats, local DNA methylation levels elevate and affect local CTCF binding sites. 360,361 Short tandem repeats comprise repeats of three or more base pairs in a DNA sequence. They make up approximately 1% of the human genome and are generally non-pathogenic. 362 Recent Hi-C data suggested that in various congenital disorders, such as fragile X syndrome, Friedreich's ataxia, and Huntington's disease, pathogenic short tandem repeat sequences are located at the TAD boundary. [363][364][365][366] These pathogenic short tandem repeat regions have abnormally elevated local methylation levels and alter the TAD boundaries, contributing to many congenital disorders. 367,368 For example, FMR1 has been identified as the pathogenic gene in fragile X syndrome. 369 Disruption of the FMR1-TAD boundary due to increased local methylation in the genome of Fragile X syndrome patients results in disrupted FMR1 interaction with telomere orientation cognate enhancers and decreased FMR1 expression. 370 Chromatin loop and disease During disease development, chromatin loops experience reprogramming in a cell-specific manner, regulating gene expression. Systemic lupus erythematosus (SLE) is an autoimmune disease frequently involving multiple organs. 371 In CD4 + T cells derived from SLE patients, 391 disease-specific chromatin loops are present, encompassing crucial inflammation-related and immunity-related genes, including DDX60L and CXCL13. 372,373 Notably, the DDX60L locus contains two disease-specific chromatin loops, whose formation is associated with histone modifications in the promoter region mediated by transcription factor SPI1. 372,373 Dilated cardiomyopathy (DCM) is a leading cause of heart failure. 374,375 Recent studies have revealed that during DCM development, enhancer-promoter loop dynamic remodeling occurs extensively across the genome in response to rapid transcription under cardiac stress conditions. 376 Chromatin loops reprogramming, directly driven by transcription factor HAND1, results in elevated expression levels of DCM-associated pathogenic genes. 376 For instance, the NPPA-AS1 promoter possesses enhancer functions, interacting with NPPA and NPPB promoters during DCM development and leading to the co-transcription of NPPA and NPPB. 376 Distant metastasis in pancreatic cancer significantly contributes to its poor prognosis. 377 Recent investigations suggest that pancreatic cancer distant metastasis correlates with epigenetic alterations. 378 In metastatic pancreatic cancer cells, the number of chromatin loops increases, along with the emergence of cell-specific chromatin loops. LIPC, a gene promoting pancreatic cancer metastasis, is implicated in tumor cell migration and invasion. 378 LIPC expression is modulated by Enhancer 3 and Enhancer 4, while tissue-specific chromatin loops form progressively during pancreatic cancer distant metastasis, enhancing LIPC expression. 378 Additionally, disease-specific chromatin loop formation has been observed in AML, involving oncogenes such as MYCN, WT1, and RUNX1. 379 For example, a specific interaction between the MYCN promoter and enhancers situated 650 kb downstream is related to AML onset. 379 CONCLUSION AND PERSPECTIVE The 3D genome structure and its functions have long been a focal point. In recent years, significant advances have been made in this area due to the rapid development of chromatin conformation capture techniques and super-resolution fluorescence imaging technologies. In this review, the structural hierarchy of the 3D genome, the effect and mechanisms of cis-regulatory element interactions in the 3D genome for regulating spatiotemporally specific gene expression, the role and mechanisms of dynamic changes in 3D chromatin conformation during embryonic development, and the pathological mechanisms of diseases such as congenital developmental abnormalities and cancer, which are attributed to alterations in 3D genome organization and aberrations in key structural proteins, were systematically discussed. In summary, the 3D genome structure plays crucial roles in cell differentiation and disease development by regulating spatiotemporal gene expression, which may offer some clues for precise diagnosis and treatment of related diseases. Nevertheless, further research is needed to understand the fundamental principles of 3D genome organization and the relationship between 3D genome structure and spatiotemporal gene expression. The recent development of single-cell chromatin conformation capture techniques and genome architecture mapping technologies has enabled a more in-depth exploration of the structural and functional features of the 3D genome. This also includes the mechanisms by which chromatin higher-order structures regulate cell-type-specific gene expression, thereby shedding light on the impact of the genome's spatial organization on cell differentiation and fate determination. Furthermore, SVs and abnormalities in structural proteins can influence the function of cis-regulatory elements, leading to atypical gene expression and, consequently, various diseases. Advancements in chromatin conformation capture techniques and transcriptomics will facilitate a deeper understanding of the pathological mechanisms underlying developmental defects and cancer. This will provide new theoretical insights and research directions for prenatal screening and precision diagnosis and introduce novel therapeutic targets for treating a more comprehensive range of congenital diseases and malignant tumors. C O N F L I C T O F I N T E R E S T S TAT E M E N T The authors declare that they have no conflict of interest. D ATA AVA I L A B I L I T Y S TAT E M E N T Data are available upon request from the authors. E T H I C S S TAT E M E N T Not applicable. O R C I D Jiewen Dai https://orcid.org/0000-0001-7807-3582	2023-07-10T05:04:12.807Z	2023-07-08T00:00:00.000	{ "year": 2023, "sha1": "f183f137046324ad8ecff997086c3c74ebc05ce1", "oa_license": "CCBY", "oa_url": null, "oa_status": null, "pdf_src": "PubMedCentral", "pdf_hash": "f183f137046324ad8ecff997086c3c74ebc05ce1", "s2fieldsofstudy": [ "Biology" ], "extfieldsofstudy": [ "Medicine" ] }
264515514	pes2o/s2orc	v3-fos-license	Solvent based fractional biosynthesis, phytochemical analysis, and biological activity of silver nanoparticles obtained from the extract of Salvia moorcroftiana Multi-drug resistant bacteria sometimes known as “superbugs” developed through overuse and misuse of antibiotics are determined to be sensitive to small concentrations of silver nanoparticles. Various methods and sources are under investigation for the safe and efficient synthesis of silver nanoparticles having effective antibacterial activity even at low concentrations. We used a medicinal plant named Salvia moorcroftiana to extract phytochemicals with antibacterial, antioxidant, and reducing properties. Three types of solvents; from polar to nonpolar, i.e., water, dimethyl sulfoxide (DMSO), and hexane, were used to extract the plant as a whole and as well as in fractions. The biosynthesized silver nanoparticles in all extracts (except hexane-based extract) were spherical, smaller than 20 nm, polydispersed (PDI ranging between 0.2 and 0.5), and stable with repulsive force of action (average zeta value = -18.55±1.17). The tested bacterial strains i.e., Klebsiella pneumoniae, Pseudomonas aeruginosa, Staphylococcus aureus, and Enterococcus faecalis were found to be sensitive to even small concentrations of Ag-NPs, especially P. aeruginosa. The antibacterial effect of these Ag-NPs was associated with their ability to generate reactive oxygen species. DMSO (in fraction) could efficiently extract antibacterial phytochemicals and showed activity against MDR bacteria (inhibition zone = 11–12 mm). Thus, the antibacterial activity of fractionated DMSO extract was comparable to that of Ag-NPs because it contained phytochemicals having solid antibacterial potential. Furthermore, Ag-NPs synthesized from this extract owned superior antibacterial activity. However, whole aqueous extract-based Ag-NPs MIC was least (7–32 μg/mL) as compared to others. Introduction Since the discovery of penicillin, the population across the globe has relied on antibiotics to fight microbial infections.Regrettably, microbes have found a way to resist the action of antibiotics that were once effective due to their misuse, combined use, and over-prescription [1].The development of multi-drug resistant microbes has now become a universal threat with severe results in administering infectious illnesses caused by pathogens [2].The need to overcome this problem has been well-recognized and highly investigated.Researchers are coming up with solutions to combat antibiotic resistance, among which nanotechnology is leading the race.Nanoparticles have emerged as unique particles with size-dependent physicochemical characteristics that are proven to have antibacterial potential [3].Silver nanoparticles (Ag-NPs) have attracted attention as they hold great promise in combating the challenge of antimicrobial resistance due to their extraordinary, broad-spectrum, and strong antimicrobial properties [4].The antibacterial mechanism of Ag-NPs has been extensively studied.In various studies, the mechanism of action of Ag-NPs is revealed to change the structure of bacterial cell wall, increase the permeability of cell membranes, and plasmolysis, inhibit respiration, suppress DNA replication, change the intracellular ATP levels, generate ROS, and disrupt cell integrity.All the ways through which Ag-NPs interact with bacteria are very effective, and the bacteria cannot easily develop resistance to them as compared to antibiotics because, for that, bacteria would have to target multiple mechanisms of action [5].A study reported the potency of Ag+ in Ag-NPs to generate ROS at molecular levels that induce oxidative stress at the cellular level leading to increased levels of calcium in intracellular space, disruption of membranes, phosphatidylserine exposure in the outer membrane, DNA degradation and activation of caspase-like protein [6].A recent study by Buszewski, Rogowska [7] observed a similar inhibitory effect on the growth of S. aureus and P. aeruginosa when treated with bioactive Ag-NPs.Several important steps occur when Ag-NPs penetrate the cell membranes of bacteria.The silver ions alter the DNA's replication process and cause it to cease replicating, leading to cell death [8].Ag-NPs can manage antibiotic resistance because no microbial resistance against Ag+ or Ag-NPs has been observed [9]. For the efficient synthesis of nanoparticles, green methods employing microbial cells, plant extract, and natural polymers have been developed.These methods are cost-effective, environment-friendly, and energy-saving [10][11][12].Among the available biosynthetic methods, the synthesis of nanoparticles using plant extract is the most favorable.Ag-NPs can be synthesized by mixing silver nitrate (metal salt) and plant extract in a proper ratio.Phytochemicals reduce silver ions and stabilize them by capping, resulting in stable Ag-NPs with possibly less toxicity.Various plants have been employed to synthesize Ag-NPs with little or no variation in the process.Conocarpus lancifolius leaf extract [13], Phoenix dactylifera [14], Santalum album leaf extract [15], Mulberry leaves extract, and Amaranthus cruentus [16] are some of the recent examples of plants employed for the synthesis of Ag-NPs with antibacterial applications. Synthesis of Ag-NPs using plant extract is a safe and green method that can be easily scaled up for producing a large amount of nanoparticles [17].Plant metabolites actively participate in the synthesis of nanoparticles by reducing and capping metallic cations.These phytochemicals are known to possess good antimicrobial potential and have been attributed to the increased antimicrobial effect of nanoparticles capped by them in the process of biological synthesis through plant extracts.Alkaloids, acetylenes, coumarins, flavonoids and isoflavonoids, iridoids, lignans, macrolides, phenolics (other than flavonoids and lignans), polypeptides, quinones, steroidal saponins, terpenoids, and xanthones have been identified to possess strong effect against resistant bacteria (Saleem et al 2010).Bearing in mind the antibacterial properties of these compounds, which have been used for centuries, are a source of new therapeutic agents. The extraction of phytochemicals depends on the type of extraction technique and nature of the solvent because the different polarity and chemical characteristics may or may not allow the dissolution of phytochemicals in particular solvents.Thus, various solvents have different affinity for phytochemicals from plants [18].Phenolic acids, flavonoids, tannins, alkaloids, etc., can be extracted in polar or less polar solvents [19].In comparison, components of essential oils and tocopherol can be obtained through extraction in nonpolar solvents [20].Thus, solvent-based plant extracts can help us to reach the specific group of plant compounds that is responsible for the synthesis of nanoparticles with high bactericidal activity, stability, and dispersity. Salvia moorcroftiana, commonly known as "kallijari" in Pakistan is a herbaceous plant known to have medicinal properties to relieve pain, fever, and inflammation [21].The plant contains valuable phytochemical contents like essential oils, flavonoids and polyphenols, tannins, terpenoids, phytosterols, carbohydrates, etc. [22].Polyphenols and flavonoids are natural antioxidants and possess other pharmacological activities such as anticancer, anti-inflammatory, antianxiety, and antimicrobial [23].The targeted extraction of plant metabolites possessing antibacterial properties can be obtained by using suitable solvents (in terms of polarity) for plant extraction [24].The metabolites in S. moorcroftiana possess the potential for Ag-NPs synthesis with varying and enhanced antibacterial activity.To the best of our knowledge, no study is available that attempts to synthesize and study Ag-NPs synthesized through this plant using different solvents.Thus, this research study focuses on the antibacterial activity and bioreduction potential of S. moorcroftiana extracts based on highly polar (Water), less polar (Dimethyl Sulfoxide), and nonpolar (n-hexane) solvents.The purpose of our current study was to determine the effect of the polarity of solvents and the extraction method on the extraction of biologically important phytochemicals of S. moorcroftiana and the synthesis of Ag-NPs (Fig 1).This study further aims to comparatively analyze Ag-NPs synthesized from the extracts of S. moorcroftiana in different solvents, in terms of stability, morphology, functional groups, surface charge, and activity against multi-drug resistant bacteria. Plant material and preparation of extract Fresh Salvia moorcroftiana leaves were collected from the district of Swat, Khyber Pakhtunkhwa, Pakistan.The plant was shade dried and ground into fine powder for extraction.Three types of extracts were prepared in different solvents ranging from polar to nonpolar: aqueous extract (Aq-extract), DMSO-based extract (D-extract), and hexane-based extract (Hextract). Preparation of aqueous extract. Aqueous extract of S. moorcroftiana was prepared by mixing 5 g of plant powder with 100 mL distilled water.The mixture was boiled for 10min and cooled at room temperature.The solution was filtered through Whatman (2.5 μm) filter paper to completely remove the plant remains.The resultant extract was stored at 4˚C for further use. Dimethyl sulfoxide-based extraction. The DMSO-based extract was prepared by mixing 2 g of plant powder with 50 mL DMSO.The mixture was sonicated for 30 min.Whatman (2.5 μ) filter paper was used to filter the mixture to remove plant debris.The dark green viscous plant extract in DMSO was obtained. Hexane-based extraction. The hexane-based extract was prepared by mixing 5 g of plant powder with 100 mL of hexane.The mixture was heated in the water bath at 60˚C for 20 min and then agitated in a shaker for 30 min.The mixture was filtered through Whatman filter paper.The thin hexane-based extract with light green color was obtained. Solvent-solvent fractional extraction of Salvia moorcroftiana. To separate phytochemicals based on their relative solubility liquid-liquid extraction was used.5 g of plant powder was mixed in 100 mL hexane which was then heated at 60˚C for 20 min, agitated for 30 min in a shaker, and then filtered through the Whatman filter to extract the nonpolar components of the plant.The filtrate was stored at 4˚C and the plant debris was used for further extraction.To the remaining plant debris, 100 mL DMSO was added and sonicated for 30 min.The mixture was filtered through the Whatman filter.The filtrate termed DMSO-extract was stored at 4˚C.To the remaining plant debris, 100 mL distilled water was added and boiled for 10 min.Subsequently, the mixture was filtered and Aq-extract was obtained that was stored for further use. Biosynthesis of silver nanoparticles Ag-NPs were synthesized in all fractions of extracts of S. moorcroftiana.For the synthesis, both aqueous extracts i.e. obtained directly from intact plant material (Aq-extract) and through solvent-solvent extraction (Aq-extracts) were mixed with 4 mM solution of Silver nitrate (AgNO 3 ) in clean flasks in a specific ratio (1:2).This means to 100 mL aqueous plant extract 200 mL AgNO 3 solution was added.The solution was incubated in daylight for 3 h.Similarly, DMSO-based extracts obtained directly from intact plant powder (D-extract) and through solvent-solvent extraction (D-extracts) were mixed with AgNO 3 solution in 1:6.This means that 20 mL DMSO-based extract was mixed with 120 mL AgNO 3 solution.The solution was incubated in daylight for 4 h.The hexane-based extract (30 mL) was dried in the oven at 50˚C for about 6 h.After this, an oily semi-dried extract was obtained.Methanol, due to its known ability to extract a broad range of phytochemicals was used to dissolve compounds from dried hexane extract.20 mL methanol was added to the dried hexane extract and the mixture was vortexed for 15 min followed by sonication for 5 min.The resultant methanolic extract was filtered (Whatman filter) to remove clumps.Finally, 30 mL of 4mM AgNO 3 solution (synthesized in methanol) was added to 15 mL methanolic extract and incubated for 24 h. The synthesis of Ag-NPs was observed in each of the extracts by a change in color.Synthesis in aqueous extract changed the color of the solution from light brown to reddish brown; in DMSO extract the color changed from light green to dark brown, and in hexane extract the color changed from lime to light grey.The Ag-NPs synthesized in all the extracts were obtained through centrifugation at 13000 rpm for 10 min.The pellet (nanoparticles) was dried in the oven at 50˚C for 24 h, collected, and stored. Characterization and confirmation of silver nanoparticles Characterization of all types of Ag-NPs was carried out.These nanoparticles included Ag-NPs synthesized in aqueous extract of S. moorcroftiana (Aq-Ag-NPs), in aqueous fraction (Aq-Ag-NPs), in DMSO based extract (D-Ag-NPs), and DMSO based fraction (D-Ag-NPs). UV-Visible spectroscopy. The biosynthesis of Ag-NPs was confirmed by measuring absorbance using UV-Visible spectroscopy (BKS360 BIOBASE Spectrophotometer). 1 mL of each sample of the reaction mixture (after the reaction of AgNO 3 and plant extract) was scanned in a spectrophotometer with the wavelength ranging from 300-600 nm.Each of the respective plant extracts was taken as a control.The solvents used for the synthesis of Ag-NPs were taken as a reference for Ag-NPs synthesized in aqueous extract water was used as a reference. 2.3.2Fourier transform infrared spectroscopy.FTIR characterizes the biomolecules involved in the formation and stabilization of Ag-NPs.To indicate the phytochemicals involved in the capping of Ag-NPs in each of the extracts of S. moorcroftiana including aqueous extracts i.e., Aq-extract, Aq-extract D-extract, and D-extract.All types of Ag-NPs (2 mg) in their dried form were ground with potassium bromide (KBr) pellets.For appropriate results verification, scanning was done at a 1:100 ratio followed by recording a spectrum at a wavelength of 4500 to 500 cm -1 .The different modes of vibrations of functional groups were identified and associated with the belonging biomolecules. 2.3.3Transmission electron microscopy and particle size analysis through Image J. The size and shape of biosynthesized Ag-NPs were calculated and visualized through TEM. 2 μL of Ag-NPs suspension were dropped onto 400 mesh carbon-coated copper grid.The suspension was dried under a lamp in ambient conditions to form a thin layer.Afterward, the samples were subjected to the TEM wherein the interaction of transmitted electrons and the Ag-NPs formed images that were recorded.The average size of nanoparticles was determined by measuring the diameter of more than 50 particles on a 100 nm TEM image by using ImageJ software. 2.3.4Dynamic light scattering of Silver nanoparticles.Dynamic Light Scattering (DLS) is a method used for the determination of the size distribution of the nanoparticles by measuring radiation scattering intensity based on the Brownian motion of particles.DLS is a nondestructive technique used to measure the size distribution and stability of silver nanoparticles in aqueous or physiological solutions.DLS uses light scattering to determine the average hydrodynamic diameter of the sample.A beam of light is put on the dispersion of nanoparticles, which scatter light to the detector.The polydispersity index (PDI) is depicted as the size distribution range of nanoparticles and their stability and uniformity.From this measurement, the mean size of particles inside the sample is obtained along with the correlation between the numbers of particles of a particular size versus the size of the nanoparticles. 2.3.5 Zeta potential determination for stability assessment.Zeta potential analysis of Ag-NPs synthesized in each aqueous and DMSO-based extract was performed to determine their suspension stability.3 mg Ag-NPs/5 mL dH 2 O was prepared and sonicated for 30 min (Power-Sonic 405).The surface charge and hydrodynamic diameter of nanoparticle samples were analyzed using a zeta sizer (Nano-ZS ZEN 3600, Malvern) with a temperature equilibration time of 1 minute at 25˚C. Antibacterial activity 2.4.1 Microbial strains.The antibacterial activity of each type of extract (Aq-extract, Aqextract, D-extract, D-extract, and H-extract) and Ag-NPs (Aq-Ag-NPs, Aq-Ag-NPs, D-Ag-NPs, D-Ag-NPs, and H-Ag-NPs) was performed.The tested microorganisms include Gramnegative Klebsiella pneumoniae and Pseudomonas aeruginosa, and Gram-positive Enterococcus faecalis and Staphylococcus aureus.These are multi-drug resistant and biofilm-forming bacterial strains, thus, they are of high concern [25]. Disk diffusion method. The Kirby-Bauer disk diffusion assay was performed to check the antibacterial activity of nanoparticles and S. moorcroftiana extracts.Each of the bacterial strains was streaked on nutrient agar solidified in sterile Petri dishes.After 24 h of incubation at 37˚C, 3-5 colonies were picked using a sterile metal loop and were inoculated in fresh broth medium contained in a sterile glass tube with screw caps.The broth inoculated with bacteria was vortexed followed by 20 min incubation.The broth suspension was compared with 0.5 MacFarland standard and the turbidity was adjusted accordingly.After adjusting the turbidity of broth suspension according to the MacFarland standard indicates 1×10 8 CFU/mL of bacteria in broth.The bacteria were uniformly spread on nutrient agar in sterile petri-dishes using a sterile cotton swab dipped in broth suspension culture.The empty disks prepared from Whatman filter paper were placed on the plate (containing agar swabbed with bacteria) using autoclaved forceps.8 μL of 2 mg/mL of each Ag-NPs and extract sample was added to each respective disc through sterile micropipette tips.The Petri dishes were sealed with parafilm and incubated for 24 h at 30˚C.The inhibition zones were recorded after 24 h of incubation. Micro-dilution assay for determination of the minimum inhibitory concentration. Micro-dilution assay for each type of nanoparticles was performed to determine the minimum inhibitory concentration.The pure culture of each bacterial strain was obtained by streaking the bacteria which was incubated for 24 h at 37˚C.The colony suspension method was followed to prepare bacterial broth culture.3-5 colonies of bacteria were picked with the help of a sterile metal loop and were inoculated in a fresh broth medium.The bacterial broth culture was compared with 0.5 MacFarland standard by comparing the optical density (OD) at 625 nm.The absorbance was observed in the range of 0.08-0.13which is the desired value and indicates approximately 1×10 8 CFU/ mL. 8 mg/mL stock solution of Aq-Ag-NPs, Aq-Ag-NPs, D-Ag-NPs, D-Ag-NPs, and 4 mg/mL H-Ag-NPs (stock solution) was prepared.The stock solutions were twofold diluted up to the 10 th well.The concentrations of Aq-Ag-NPs, Aq-Ag-NPs, D-Ag-NPs, and D-Ag-NPs started at 4 mg/ml and ended at 7.8 μg/mL.However, H-Ag-NPs the concentrations started from 2 mg/ml and ended at 3.9 μg/mL. The dilutions were added to their respective wells of a 96-well microtitre plate.A separate plate was used for each of the bacteria.50 μL of bacterial broth culture was added from the 1 st to 11 th (growth control) column.50 μL of broth (only) was added to the 11 th column and 100 μL to the 12 th (sterility control).The plate was incubated for 20 h at 37˚C in a shaking incubator.The MIC was observed visually by determining the concentration at which Ag-NPs inhibited the bacterial growth resulting in a clear well. Mechanism of bacterial inhibition by the biosynthesized Ag-NPs. The oxidant sensitive probe 2', 7'-dichlorodihydroluorescein diacetate (H 2 DCFDA) was used to determine the intracellular levels of ROS in cells treated with different concentrations of Ag-NPs (0.5-250 μg/mL).Bacterial cells were grown in an LB medium until OD 600 reaches 0.5.The bacterial cultures were centrifuged to pellet the cells.The cells were washed with 10 mM potassium phosphate buffer (pH 7.0) through vortexing and centrifugation.After washing, the cells were suspended in the same buffer and disrupted by sonication.10 mM H2DCFDA (dissolved in dimethyl sulfoxide) was added at a ratio of 1:2000 (2 microliters in 4 mL buffer), followed by shaking for 30 min at 37˚C.The bacterial cells were again pelleted after incubation through centrifugation.The cells were washed two times with the same buffer to remove the H 2 DCFDA.To cell suspension, different concentrations Ag-NPs were applied.The fluorescence intensity of DCF by fluorescence spectrophotometer at an excitation wavelength of 488 nm and an emission wavelength of 535 nm was measured. Comparative analysis of phytochemical content in extracts and Silver nanoparticles 2.5.1 Determination of total phenolic content.Phenolics are antioxidant molecules with pharmacological activities and reducing capabilities.The total phenolic content in each of the extracts and Ag-NPs synthesized from these extracts were determined quantitatively by using the Folin Ciocalteu method with gallic acid as the standard.Briefly, 20 μL of each extract and Ag-NPs sample (2 mg/mL) were added to a 96-well plate.Then, 90 μL Folin Ciocalteu reagent (1:10 diluted form) was added to the sample.Subsequently, 90 μL of 6% sodium carbonate was added.Gallic acid (4 mg/mL stock in methanol) was used as a standard positive control.The plate was incubated for 30 min and then the optical density of total phenolic content was measured at 630 nm. 2.5.2Determination of total flavonoid content.Similar to phenolics, flavonoids possess great medicinal importance and reducing potential.They also show reduction and stabilization potential in the synthesis of Ag-NPs.The total flavonoid content was measured using the Aluminium chloride (AlCl 3 ) method.Briefly, 20 μL of each of the extract and Ag-NPs (2 mg/mL) sample were added to a 96-well plate.Potassium acetate (10 μL of 98.15 g/L) was added followed by the addition of 10 μL aluminum chloride (10 g/100 mL).Subsequently, 160 μL of distilled water was added.The 20 μL MeOH instead of the sample was used as a negative control.The plate was incubated for 30 min and then the optical density was measured at 405 nm. DPPH (2,2-diphenylpicrylhydrazyl) assay. A DPPH assay was performed to assess the free radical scavenging activity (FRSA) of S. moorcroftiana extracts and Ag-NPs synthesized from it.DPPH is a stable free radical; therefore, it is used to assess whether the extracts and Ag-NPs possess the ability to scavenge this radical.To determine the antioxidant activity 20 μL of sample (extracts and 2 mg/mL Ag-NPs) was added to a 96-well plate.Then, 180 μL of DPPH reagent (3.2 mg/100 mL) was added subsequently.Ascorbic acid (4 mg/mL) was used as positive control which is a naturally free radical scavenging agent.The plate was incubated for 1 hr and FRSA was measured by recording absorbance at 517 nm. 2.5.4Oxidation-reduction potential.The ORP of extracts was determined to know their reduction potential in the synthesis of Ag-NPs.The ORP of Ag-NPs was determined to analyze the potential of nanoparticles as oxidative species that are lethal for bacteria.The ORP was measured through an ORP sensor probe lubricated with potassium chloride.The rod was dipped in each of the extracts and the ORP for each extracted sample was measured three times subsequently their average was considered as a final value.A similar procedure was done to determine the ORP of all Ag-NPs solution (2 mg/ml).The probe was rinsed with distilled water every time it was used for different kinds of extract and Ag-NPs solution. High-Performance Liquid Chromatography (HPLC) for identification of polyphenols in plant extracts HPLC quantification was carried out according to a reported method (Zeb 2015).Briefly, 1-g powdered sample of each of the whole and fractionated extract was added to water and methanol in equal ratios, and the mixture was subject to heating in a water bath at 70˚C for 1 h.The mixture was then filtered through a non-pyogenic 0.4 μm CA syringe filter. To identify and quantify phenolic compounds the Agilent-1260 infinity High-performance liquid chromatography (HPLC) system was used.The HPLC system's essential parts were a quaternary pump, an auto-sampler, a degasser, and a C18 column (Agilent-Zorbax-Eclipse column).The solution (B and C) gradient was such that solvent B was a mixture of acetic acid: methanol: deionized water (20: 100: 180 v/v), and solvent C was a mixture of acetic acid deionized water: methanol (20: 80: 900) v/v.The solvents were provided as a gradient such that they started and gradually decreased the solvent in concentration.Solvent B was given in volume 100, 85, 50, and 30% at 0, 5, 20, and 25 min, finally giving way to 100% solvent C from 30 min onwards till 40 min.The ultraviolet array detector (UVAD) was set at wavelength 250 nm to analyze phenolic compounds, and the chromatograms were recorded.Phenolic compounds were identified by comparing the retention times of obtained HPLC chromatogram with that of the standards. Synthesis of silver nanoparticles in plant extracts obtained in solvents with varying polarity The biosynthesis reaction of Ag-NPs was performed in each of the extracts by allowing it to react with the AgNO 3 solution.After 4 h of reaction, a color change was observed that was different for each of the extracts.The color change is a characteristic determining the reduction of Ag + by plant metabolites [26].This color change is due to the interaction of light and the surface oscillating electrons of silver nanoparticles, the phenomenon is known as surface plasmon resonance [27].Thus, each of the extracts either whole or fraction had the potential of Ag-NPs synthesis except for H-extract.A very slow reaction and thus synthesis of Ag-NPs were observed in hexane because it is nonpolar and is usually used for the extraction of nonpolar compounds.AgNO 3, being polar cannot react with nonpolar compounds.The dried Hextract was dissolved in methanol because of its good extraction capability [28].As a result, Hextract dissolved in methanol when reacted with AgNO 3 solution, and a very low concentration of synthesized Ag-NPs was observed. UV-Vis spectroscopy of silver nanoparticles UV-Vis absorption spectra of the biosynthesized Ag-NPs showed the characteristic surface plasmon resonance of Ag-NPs in the range of 400-500 nm, confirming their synthesis [29].Surface plasmon resonance refers to the phenomenon where electron clouds around Ag-NPs can oscillate and absorb specific wavelengths of electromagnetic radiation.Such absorbance is the result of the size of nanoparticles along with other factors including condition of synthesis, solvent, and surface functionalization [30].Based on the sizes and solvent used, various absorption bands were observed for each kind of nanoparticles (shown in Fig 2). The maximum absorbance (represented as λ (max) ) in the range of 400-500 was recorded to be 498 nm for Aq-Ag-NPs, 432 nm for D-Ag-NPs, and 414 nm for H-Ag-NPs (Fig 2A).[29].Aq-Ag-NPs and D-Ag-NPs showed broader peaks indicating the presence of nanoparticles with varying sizes [31].H-Ag-NPs gave a narrow SPR band with little absorption intensity that shows a small concentration of biosynthesized H-Ag-NPs [31].With the increase in the size of nanoparticles, the absorbance intensity decreases [32] as observed with Aq-Ag-NPs when compared with D-Ag-NPs.Kochkina and Skobeleva [33] synthesized Ag-NPs from the reaction of starches dissolved in DMSO with AgNO 3 and observed the surface plasmon band maximum at 418 nm.The absorption bands of Aq-Ag-NPs (λ (max) = 414 nm) and D-Ag-NPs (λ (max) = 472 nm) within the characteristic range confirms their presence.The UV-Vis spectra of S. moorcroftiana extract in different solvents are taken as control (Fig 2B ) to compare their absorption peaks with that of Ag-NPs.Although the extracts showed absorbance in the UV-Vis range, no characteristic peaks are found to be similar to the peaks of Ag-NPs.The Dshaped absorption curve of D-Ag-NPs in the range of 400-500 is indicative of the presence of Ag-NPs in the sample when compared to D-extract whose absorption showed to be continuously decreasing within this range.According to the absorption spectra of various types of Ag-NPs, most of the Ag-NPs were spherical because their absorption peaks were mostly in the range of 400-470 nm which refers to spherical-shaped nanoparticles, with sizes less than 20 nm [34].These results agree with the results of TEM.Thus UV-Vis spectroscopy results clearly show that the synthesis of Ag-NPs from their respective extracts is achieved. Fourier Transform Infra-Red Spectroscopy (FTIR) Generally, the FTIR spectra of the given plant extracts and Ag-NPs revealed that the compounds present in the extracts have capped the biosynthesized Ag-NPs (as shown in Fig 3) and therefore, gave peaks in the same range of wavenumbers.However, less absorption and more transmittance were observed in the FTIR spectra of Ag-NPs than that of extracts from which they were synthesized. The presence of these functional groups indicates the presence of polyphenols (such as flavonoids and phenolics), terpenoids, carboxylic acid, amines in proteins, aromatic amines, alkenes, alkanes, and organosulfur compounds attached to the surface of Ag-NPs.The presence of these phytochemicals in both the plant extracts and Ag-NPs indicates their role in the synthesis and stabilization of Ag-NPs [35][36][37].It is challenging to point out a single specific class of phytochemicals involved in the synthesis of Ag-NPs, but generally, all these compounds are significant for the synthesis of silver nanoparticles [38].These phytochemicals possess antibacterial properties and can result in the synthesis of Ag-NPs with enhanced antibacterial activity [39].Only D-extract and D-Ag-NPs showed a distinctive peak at 1168 cm -1 and 1178 cm -1 indicating C-O vibration corresponding to esters.In previous studies, it has been shown that ester molecules were obtained in high concentration through extraction in polar solvents as compared to nonpolar [40].Moreover, both aqueous whole and fraction extracts and Ag-NPs based on them revealed the presence of sulfate or sulfoxide (S = O) which is the functional group of organosulfur compounds in plants [41].This compound was not found in DMSO-based extracts and Ag-NPs because they are known to decompose in DMSO [42]. Transmission electron microscopy of silver nanoparticles TEM was used to analyze the morphology of biosynthesized Ag-NPs.The TEM images confirmed the synthesis of Ag-NPs in all the extracts of S. moorcroftiana (Fig 5).The nanoparticles in all the extracts appeared to be polydispersed having both large and small diameters with average diameters of 17.889, 17.11, 18.82, and 17.815 nm for Aq-Ag-NPs, D-Ag-NPs, Aq-Ag-NPs, and D-Ag-NPs, respectively (Table 1).All of the biosynthesized Ag-NPs were almost spherical as determined by TEM micrographs.Ag-NPs with smaller sizes and spherical morphology are more effective against bacteria because they can easily attach and enter the cells via membranes leading to bactericidal effects [4].The close analysis of images showed that clusters of nanoparticles are surrounded by a layer.Such layers are found around nanoparticles mainly synthesized from medicinal plant extracts, where the phytochemicals form a capping layer and help to shape the particles during growth [43]. Zeta potential data for assessing the stability of silver nanoparticles Zeta potential, the electrostatic repulsion or attraction between the particles is determined through Dynamic Light Scattering (DLS).The negative or positive zeta values indicate the repulsive force between the particles showing dispersity and stability.All types of samples showed negative zeta potential values such as -17.6±0.14 mV, -18.5±0.24mV, -16.3±0.26mV, and -21.8±0.16 mV for Aq-Ag-NPs, D-Ag-NPs, Aq-Ag-NPs, and D-Ag-NPs, respectively (Fig 6).The negative zeta potential values indicated that the Ag-NPs possessed a positive charge in the solution that attracted negative charges to surround them [44].The negative zeta values also indicate the repulsive force among the nanoparticles, hence confirming their stability [45].In this context, the DMSO plant extract-based Ag-NPs were found to be more stable than the aqueous one.Furthermore, between the DMSO plant extract-based Ag-NPs, D-Ag-NPs were revealed to be more stable with zeta value = -21.8±0.16mV. Size distribution of silver nanoparticles determined through Dynamic light scattering DLS was used to determine the polydispersity index, referring to the size distribution of nanoparticles in solution.Furthermore, hydrodynamic size or average zeta size was also calculated for four types of Ag-NPs i.e.Aq-Ag-NPs, D-Ag-NPs, Aq-Ag-NPs, and D-Ag-NPs (Fig 7).The values of the mentioned characteristics are shown in Table 1.The results revealed the average zeta size of Aq-Ag-NPs, D-Ag-NPs, Aq-Ag-NPs, and D-Ag-NPs to be 135.7 d. nm, 365 d. nm, 99.36 d. nm, and 377 d. nm, respectively.All the Ag-NPs were found to be less The size determined by DLS was found to be larger than the size determined by TEM because it measures the hydrodynamic size of nanoparticles rather than their physical diameters [46].This means that during the movement of nanoparticles in a liquid medium, the electric dipole layer of the solvent surrounds them which can also influence their motion in the solvent.Thus, hydrodynamic diameter shows the combined diameter of the metal core and the solvent molecules surrounding it [47].In this context, DMSO-based Ag-NPs are probably surrounded by more protective layers than aqueous extract-based Ag-NPs, thus having large hydrodynamic diameter. Comparative analysis of phytochemical content in plant extracts and Silver nanoparticles 3.7.1 Total phenolic content in extracts and silver nanoparticles.The TPC and TFC results showed that high amount of phenolic and flavonoid compounds are present in the extracts which act as reducing and capping agents utilized in the synthesis of stable Ag-NPs [48].The TPC and TFC of Ag-NPs gave evidence of the capping and stabilizing capability of these compounds [49] showing that during synthesis, they attach to the Ag-NPs to stabilize them and provide them enhanced pharmacological properties. TPC of extracts showed diverse results such as 98.92±4.12,255.50±3.13,262.48±0.84,227.76±0.46,and 67.42±1.21μg GAE/mL for Aq-extract, Aq-extract, D-extract, D-extract, and H-extract, respectively (Fig 8).The highest phenolic content was found in D-extract, followed by Aq-extract.The least phenolics were extracted by hexane.The TPC of Ag-NPs was substantially lower than that of extracts ( Fig .This free radical scavenging property is the result of reductants having reducing power that can react with and prevent the formation of free radicals [50].The results of free radical scavenging activity indicate effective antioxidant activities of all types of extracts.This could be linked with the presence of TPC and TFC, because, phenolic compounds possess strong antioxidant potential that are found in considerable amounts in plant extracts [51].Among the extracts, D-extract showed the highest percent of free radical scavenging activity i.e., 80.8±0.72%, which can be attributed to the increased quantity of phenols and flavonoids extracted by DMSO in fraction.DMSO is a preferred solvent for the extraction of phenolic and flavonoid compounds as in both cases DMSO based plant extracts revealed their higher concentration.Demir, Turan [52] used DMSO for extraction of Rhododendron luteum and revealed its high antioxidant activity due to the extraction of higher quantities of polyphenols.This free radical scavenging property is the result of reductants having reducing power that can react with and prevent the formation of free radicals [50].The H-extract also revealed good antioxidant activity having 74.36±1.0percent free radical scavenging activity. Ag-NPs synthesized from these extracts were also found to be good antioxidant agents with the highest antioxidant potential (61±0.50%) of Aq-Ag-NPs followed by D-Ag-NPs with 56 ±1.55% FRSA.This reveals that Ag-NPs synthesized from fractionated plant extracts was higher as compared to the one synthesized from whole extracts.Thus, the strategy of fractionated plant extraction and the synthesis of Ag-NPs from these extracts result in the production of good antioxidant agents. Oxidation Reduction potential of extracts and silver nanoparticles. ORP of extracts was determined to find the reducing agents and their reduction potential for the synthesis of Ag-NPs.The more negative is the value the more reducing agents are present in the extract having the capability to reduce Ag + to Ag-NPs indicating high reduction potential [53].The highest reduction potential was recorded for Aq-extract (-113.667±1.17mV) and the [53].The ORP of Ag-NPs was measured to determine their oxidation potential and find their ability to oxidize bacterial cells.Among Ag-NPs the highest oxidation potential was recorded for Aq-Ag-NPs followed by D-Ag-NPs (Table 2).The high positive ORP value of Aq-Ag-NPs can be linked with that of Aq-extract.The components present in Aq-extract that resulted in increased oxidation potential may have capped the Ag + ions during biosynthesis reaction causing the formation of Aq-Ag-NPs with high oxidizing capacity. Identification of polyphenols through HPLC The results of HPLC revealed the whole and fractionated extracts of S. moorcroftiana to be enriched with phenolic acids and flavonols.Such composition of the extracts make them efficient source for the synthesis of Ag-NPs.From the chromatogram of each of the extract, those with higher peak area (%) are picked for analysis (Table 3).Polyphenols found in Aqextract include malic acid (25.49), caftaric acid (8.387), and myricetin (32.04).D-extract contains gallic acid (14.746), rosmarinic acid (46.227)Plants being the natural reducing agents contain different secondary metabolites in the form of condensed polyphenols which takes part in the reduction of metal salts and stabilization of their nano forms.Identification of these secondary metabolites is vital to understand the possible reaction between the precursor and extract used [54].Various types of polyphenols are identified in each of the extract of S. moorcroftina.With the help of the reaction of these phenols and flavonoids with AgNO 3 , stable Ag-NPs in high concentration were produced.These polyphenols have been reported to possess strong antioxidant and antibacterial activity.Among the extracts, only to the D-extract, all tested strains of bacteria were susceptible.This can also be linked to the results of TPC and TFC that show D-extract possessing higher TFC and TPC values.Despite the presence of polyphenols in Aq and Aq-extract, none of them showed any antibacterial activity.This is similar to the results of a study where the ethanolic extract of Crataegus monogyna showed good activity against Gram positive and negative bacteria while its aqueous extract did not show any.The antibacterial activity was attributed to the presence of high polyphenol content [55].Caffeic acid is a known potent antibacterial agent which in the present study is found in high amount in D-extract [56].In addition, kaempferol and its derivatives have been shown in previous studies to possess good antibacterial potential against Gram positive and negative bacteria [57].Thus, the fractionated DMSO based extract of S. moorcroftiana was able to extract polyphenols that were biologically active against pathogenic bacteria. 3.9 Antibacterial activity of extracts and silver nanoparticles 3.9.1 Disk diffusion assay of extracts and silver nanoparticles.The antibacterial activity of whole and fractionated S. moorcroftiana extracts and Ag-NPs synthesized from these extracts was examined.All types of Ag-NPs showed moderate activity against multi drug resistant human pathogens such as Klebsiella pneumoniae, Pseudomonas aeruginosa, Staphylococcus aureus, and Enterococcus faecalis (Table 4).Among the extracts, D-extract showed good antibacterial potential with inhibition zones having diameter of 11 mm, 9.83 mm, 9.83 mm, and 10.33 mm against K. pneumoniae, P. aeruginosa, S. aureus, and E. faecalis, respectivley.The extraction of phytochemicals depends on method of extraction, solvent, and time duration.Here, solvent is considered as a main factor.On this basis, only DMSO based fraction of S. moorcroftiana rather than its aqueous extracts was able to show antibacterial effect against all the tested strains.This is probably because the plant components active against bacteria may not be effectively extracted in water while the antibacterial aromatic or saturated organic compounds are mostly obtained through organic solvent extraction [58].As compared to D-extract, D-extract had higher antibacterial potential.When the antibacterial activity of D-extract is linked to the results of FTIR they indicate the presence of esters that might be responsible for its antibacterial activity because esters have been reported to possess good antibacterial activity [59].D-extract also possessed concentrated biological important phytochemicals such as polyphenols and flavonoids.Kumar, Nehra [60] showed that E. faecalis is most sensitive to plant compounds such as polyphenols.Therefore, this specific strain (as compared to other strains) in our study showed increased sensitivity to H-extract, D-extract, and D-extract.The largest zone of inhibition was 12 mm that was recorded for Aq-Ag-NPs against K. pneumoniae and D-Ag-NPs against P. aeruginosa.When compared, Aq-Ag-NPs gave larger zone of inhibition than Aq-Ag-NPs but in case of DMSO extract based Ag-NPs, D-Ag-NPs showed higher antibacterial activity than D-Ag-NPs.H-Ag-NPs gave comparatively smaller zones of inhibition i.e. 8 mm, 9.33 mm, 8.67 mm, and 8.60 mm against K. pneumoniae, P. aeruginosa, S. aureus, and E. faecalis, respectivley, revealing it to be weak antibacterial agent.Four types of controls were taken alongside i.e. dH 2 O, DMSO, hexane, and ceftadizime antibiotic (200μg/ml). Each type of Ag-NPs showed activity against all tested MDR strains.H-Ag-NPs were the least effective because H-extract did not contain enough phytochemicals to synthesize and cap Ag-NPs.As in our results we can see that D-Ag-NPs are more stable, have higher antioxidant activity, contain increased amount of phenolic and flavonoid compounds, and also have the high oxidation potential which should how make them most effective against resistant bacteria.However, Aq-Ag-NPs were the most effective with large inhibition zone against each bacterium as compared to all other Ag-NPs.The antibacterial results linked with FTIR analysis revealed the presence of S = O (sulfoxide) on Aq-Ag-NPs that is commonly found in organosulfur compounds having strong antibacterial activity along with other phytochemicals [61].For instance, thiosulfinate (organosulfur compound) such as allicin has been recognized as responsible for the antibacterial property of garlic and onion [62].The contributing role of these compounds with other phytochemicals may be the reason for enhanced antibacterial activity of Aq-Ag-NPs.Though the same group was also found in Aq-Ag-NPs, their antibacterial potential was not that effective because at the end of fractionation (solvent-solvent extraction), water had few remaining phytochemicals to extract and further utilize for synthesis and capping of Ag-NPs.Such organosulfur compounds are immiscible in water and are often more soluble in less polar solvents than water [42,63].Therefore, these compounds were not biologically available in Aq-extract and Aq-extract for interaction with bacteria, hence did not show any antibacterial activity. Minimum Inhibitory concentration (MIC) of Ag-NPs. The minimum inhibitory concentrations of Ag-NPs against all aforementioed bacterial strains were determined through microdilution assay.The bacterial growth inhibition with the application of Ag-NPs was determined through visual observation of turbidity in the wells of microtitre plate.Each kind of Ag-NPs showed different MIC values against each bacterial strain as shown in Table 5.The lowest MIC of various Ag-NPs against the given bacterial strains are: 7.8 μg/mL of H-Ag-NPs against P. aeruginosa, 31.25 μg/mL of Aq-Ag-NPs against P. aeruginosa and Staphylococcus aureus, 31.25 μg/mL of D-Ag-NPs against P. aeruginosa and E. faecalis, 7.8 μg/mL of Aq-Ag-NPs against P. aeruginosa, and 31.25 μg/mL of D-Ag-NPs against P. aeruginosa.Thus, from these results, it is clear that P. aeruginosa is the most sensitive to all kinds of Ag-NPs even at very low concentrations.Similar results were shown by de Lacerda Coriolano, de Souza [64] where P. aeruginosa was sensitive to the concentrations of Ag-NPs between 1 and 200 μg/mL.All types of Ag-NPs were effective in their lower concentrations that ultimately reduce their cytotoxic effects.However, like the results of disk diffusion assay, Aq-Ag-NPs against each bacterium compared to other Ag-NPs revealed to be relatively effective in lower concentrations such as 15.625 μg/mL, 7.8 μg/mL, 31.25 μg/mL, and 15.625 μg/mL against K. pneumoniae, P. aeruginosa, S. aureus, and E. faecalis, respectively. 3.9.3.Demonstration of the mechanism of action of bacterial inhibition by Ag-NPs.One of the main mechanisms through which antimicrobials induce bactericidal effect is the generation of reactive oxygen species.These ROS targets and alter the sites of amino acids and nucleotides, resulting in protein and DNA damage, causing bacterial cell death [65].The results of ROS quantification shows the production of ROS in all three types of bacteria by all types of applied Ag-NPs (Fig 11).The highest level of ROS was generated varyingly in each of the bacterial culture, such as, Aq-Ag-NPs was able to produce increased level of ROS in K. pneumonia and S. aureus, while in P. aeruginosa, D-Ag-NPs produced increased level of ROS.Also, in each bacterial culture, Aq-Ag-NPs consistently generated significant amount of ROS. A similar trend with respect to the relation of concentration of Ag-NPs and fluorescence intensity of ROS was noted in K. pneumoniae and P. aeruginosa.In these bacterial cultures at low concentrations of Ag-NPs (31.2-62.5 μg/mL), the ROS are generated in high numbers, but as soon as the concentration of Ag-NPs increases (reaches 250 μg/mL), the number of ROS falls gradually.This might occur because with the application of higher concentrations of Ag-NPs the bacterial cells die and ultimately they are not able to produce ROS [66].However, in S. aureus, the number of ROS retained to increase with increase in the concentration of Ag-NPs.Thus, all Ag-NPs were shown to be applicable as antibacterial agents because of their reasonably high inhibitive activity even in very less concentrations. Conclusion Green synthesized Ag-NPs because of their ecofriendly synthesis and enhanced antibacterial activity have gained enormous attention of researchers to provide solution for antibiotic resistance.In this study, the extract of S. moorcroftiana (whole and fraction) was used for the synthesis of Ag-NPs in three solvents (water, DMSO, and hexane) with decreasing polarity.The characterization techniques confirmed the synthesis of spherical Ag-NPs in high concentration with average diameter less than 20 nm, having high polydispersity and stability.Phytochemicals like proteins, flavonoids, phenolic compounds and organic acids are identified to be significantly involved in the synthesis of Ag-NPs.These compounds, when extracted in fractions show their targeted role in the synthesis of Ag-NPs and their antibacterial effect, as observed in D*-extract.The plant compounds extracted in DMSO after extraction through hexane showed good reducing, stabilizing, and antibacterial potential.Regardless of identifying the targeted role of phytochemicals, Aq-Ag-NPs were found to be most effective antibacterial agents because of diverse phytochemicals (wholly extracted in water) capped on them. Overall, this study showed that the targeted extraction of phytochemicals from medicinal plant like S. moorcroftiana can help to obtain the antimicrobial plant components that can be applied against bacteria and used for the synthesis of Ag-NPs with enhanced antibacterial activity and reduced toxicity. Fig 8 . Fig 8. Comparative analysis of total phenolic content in different extracts and Ag-NPs based on different solvent-based extracts.Hexane based Ag-NPs synthesis was negligible and could not be analyzed for TPC.https://doi.org/10.1371/journal.pone.0287080.g008	2023-10-28T05:05:42.013Z	2023-10-26T00:00:00.000	{ "year": 2023, "sha1": "70999e0f36240a860b67fdca4d431272a824a39b", "oa_license": "CCBY", "oa_url": null, "oa_status": null, "pdf_src": "PubMedCentral", "pdf_hash": "70999e0f36240a860b67fdca4d431272a824a39b", "s2fieldsofstudy": [ "Biology", "Chemistry", "Environmental Science", "Materials Science", "Medicine" ], "extfieldsofstudy": [ "Medicine" ] }
201130176	pes2o/s2orc	v3-fos-license	Design of high voltage grid-connected switch energy storage circuit control system The paper proposes and designs the control system of the high voltage grid-connected switch energy storage circuit based on ARM, in order to ensure the normal operation of the power system. In view of the problem that the high voltage grid-connected switch controlled by Lab-VIEW software currently takes a long time for spring energy storage and the system response speed is slow, which is not conducive to timely cutting off the fault source. The hollow current transformer and voltage transformer are used to collect the current signal and voltage signal on the main line of the power system respectively. Then the processed signal is input into the internal analog to digital conversion unit of the system microcontroller for analog to digital conversion, and the processed signal is calculated and judged logically by using the system microcontroller and the principle of three-point method measurement. If there is a defect in the energy storage circuit of the grid-connected switch, the corresponding defect treatment method shall be adopted according to the judgment result, and the electrical isolation switch circuit breaker control instruction signal shall be issued at the same time, and then the system drive circuit shall execute the protection action. At the same time, on this basis, the software structure of the system and the specific process of calculating and processing tasks are given. The results of system performance test show that the designed system can not only ensure the high voltage grid-connected switch to meet the mechanical properties and low voltage action characteristics, but also reduce the energy storage time of the spring, shorten the system reaction time, which is conducive to cutting off the fault source in time and avoiding the expansion of the accident range. Introduction High voltage grid-connected switch is an important switch in power system such as power plant and substation. It is mainly used to ensure the safety of high-voltage electrical appliances, work safety during maintenance, and play the role of isolating voltage. Under normal circumstances, high-voltage grid-connected switch can be divided into various types according to the installation location, class difference, mode of action, whether there is a grounding switch, operating mechanism and the number of insulation column included. With the rapid development of intelligent high voltage grid-connected switch, people use computer technology, CAN bus technology, digital signal processing technology and many other advanced technologies to embed high voltage and high voltage grid-connected switch into high voltage electrical products to realize the automatic control of high voltage distribution system [1]. However, in the actual application process, high-voltage grid-connected switch energy storage motor burn-down accidents occur frequently, especially when the high-voltage electrical equipment of unattended substation breaks down, the remote dispatch cannot timely obtain the current ICEMEE2019 IOP Conf. Series: Earth and Environmental Science 295 (2019) 042007 IOP Publishing doi:10.1088/1755-1315/295/4/042007 2 state information of equipment and make correct decisions, which leads to the delay of fault treatment and enlarges the accident scope [2]. In order to solve the above problems and ensure the normal, stable and reliable operation of the power system, power grid and substation, it is necessary to design an effective high-voltage grid-connected switch energy storage circuit control system. High voltage grid-connected switch energy storage circuit control system In this paper, the hollow current transformer and voltage transformer are respectively used to collect the current signal and voltage signal on the main line of high-voltage electrical appliances in the power system, and the collected current signal and voltage signal are preprocessed by setting, voltage dividing and filtering through the analog signal conditioning circuit. The processed signal is input into the internal analog to digital conversion unit of the system microcontroller for analog to digital conversion, and then the current signal and voltage signal are calculated and judged logically by the system microcontroller. If the main circuit is judged to be faulty, corresponding measures shall be taken according to the judgment results, and the electrical disconnecting switch circuit breaker control instruction signal shall be sent out at the same time, and the system driving circuit shall perform the protection action. At the same time, the specific fault types of trunk lines and the occurrence of fault events are stored in the system [3]. In addition, the high voltage grid-connected switch energy storage circuit control system also has an analog trip circuit. When there is a very large short circuit fault in the power system circuit, the analog tripping circuit can directly perform the tripping process and cut off the fault without going through the system microcontroller. The high voltage and high voltage grid-connected switch energy storage circuit control system has two kinds of communication interfaces: RS485 bus and Ethernet RJ45. system analog signal conditioning circuit design The main function of designing the analog signal conditioning circuit of the system is to realize the pretreatment of the current signal and voltage signal on the main line of high-voltage electrical appliances in the power system, such as setting, voltage dividing and filtering, etc., so that the processed signal can meet the processing requirements of the system microcontroller. In the practical application, the measured current and the protection current of the high-voltage grid-connected switch are detected separately in consideration of their normal operation. As shown in figure 2, the system analog signal conditioning circuit diagram is given. In FIG. 2 reference voltage of the high-voltage main line of the power system, which is mainly responsible for raising the reference voltage. The voltage division circuit composed of 5.88 ohms and 1.5 ohms resistance is responsible for voltage dividing treatment between AVIN and AREF; Made up of 750Ω resistance and 0.1μF capacitance of low-pass filter circuit is mainly responsible for dealing with aircore current transformer and voltage transformer output high frequency noise signal; AVOUT is connected with the input channel of ADC in the system microcontroller to convert the acquired signals into analog or digital ones [5]. system clock circuit and temperature detection circuit design In order to accurately record the fault time and the environment temperature on the high-voltage electrical main line of the power system, the real-time clock circuit and the temperature detection circuit are designed by using SD2068A chip and DS18B20 chip, which have the advantages of high data transmission efficiency, digital adjustment, strong anti-interference, low power consumption and so on. defect analysis of grid-connected switch energy storage circuit The reason for the failure of closing of the high voltage grid-connected switch is that the operating time of the system relay is less than the completion time of the energy storage of the high voltage gridconnected switch spring, which results in the failure of the energy storage of the high voltage gridconnected switch spring and the defect of the spring energy storage circuit. grid-connected switch energy storage circuit defect treatment (1) Adjust the setting time of the control system relay, so that the system energy storage motor can set time to complete the relay spring energy storage, and leave enough space. By using a stopwatch to measure the energy storage time required by the high voltage grid-connected switch spring and adjusting the setting time of the control system relay, the setting time delay is greater than the actual energy storage time of the system energy storage motor, and the margin space is fully considered; (2) Because of the energy storage electric pilot run time in the outdoor environment, the transmission mechanism caton that may occur due to the weather and other factors. That makes the grid switch spring energy storage time extension, severe cases may far exceed the system interconnection switch spring energy storage time relay setting time. The transmission mechanism of energy storage spring of grid-connected switch can be lubricated to solve the defect of energy storage not in place. According to the above processing method, the remote and local closing can be performed smoothly, and at the same time, the system microcontroller also stops sending 'abnormal signal of grid-connected switch spring energy storage circuit'. three-point measurement principle of system power parameters The real-time display, three-phase protection, remote monitoring, communication and thermal memory of the control system of the high-voltage grid-connected switch energy storage circuit are realized by software. The reliability and complexity of the software structure determine the final performance of the high voltage grid-connected switch energy storage loop control system. System software part design is the fundamental purpose of when a device or circuit in circuit overload phenomenon, the realization of high voltage grid switch three-phase circuit parameters such as current, voltage of real-time measuring calculation, to cut off the source of trouble in time to prevent the accident range further, at the same time, can also according to the preset interconnection switch trip protection threshold to decide whether to control movement. In order to accurately and quickly judge Taking the three-point method for measuring and calculating the effective value of the current as an example. It is assumed that there is a sinusoidal signal in the operation of the high-voltage electrical appliances. As long as there are three points of amplitude, frequency and initial phase, the characteristic quantity of the sinusoidal signal can be determined. Suppose m I and  represent the amplitude and frequency of the current signal to be measured; t  represents the sampling interval of the current signal to be measured; k represents the sampling sequence of the current signal to be measured; 0  shows the initial phase of the current signal, using a three-point method to measure and calculate the current signal () it ,The specific calculation formula is as follows: The above formula can be summarized as follows: Then, the effective value of the current signal () it to be measured is: ; The above calculation process shows that as long as the set value of current signal sampling interval t  is less than the threshold / 20 T , the higher calculation accuracy can be achieved. If the phenomenon of cos( ) 1 t  appears in the actual application process, it indicates that there is interference in the sampling points. In order to reduce the interference effect, more points should be collected and then digital filtering should be applied. The three-point method is adopted to measure and calculate the effective values of current and voltage, which has a roof with a small amount of calculation, and can greatly improve the response speed of the energy storage circuit of the highvoltage grid-connected switch. high voltage grid-connected switch energy storage circuit control system program structure High-pressure grid switch energy-storage circuit control system to complete the real-time requirements such as sampling calculation, fault diagnosis of high computing tasks, the task even at the same time display, keyboard scanning, and communications, such as real-time demand is not high computing tasks. In order to guarantee system program structure is clear enough, reduce the system server computation, the computing task is divided into four levels, according to the real-time requirements of computing tasks. For example, as slow tasks with low real-time requirements, such as calculation, processing and display, the system program is executed every 300ms and sends data according to user requirements. Keyboard scanning and other quick computing tasks, the system program is executed every 60ms; Sampling, measurement and calculation, fault judgment and other tasks handle task as real-time computing, completed by the system program interrupt; In addition, such as watchdog timer, high voltage detection and other routine computing processing tasks are executed every time the main loop of the system program. System performance test results and analysis The digital high-voltage grid-connected switch characteristic tester produced in the United States, ct-7500, was adopted in the field. The mechanical characteristics of the high-voltage grid-connected switch were tested by the reference [4] system and the arm-based high-voltage grid-connected switch energy storage loop control system proposed and designed in this paper. The comparative test results shown in table 1 and table 2 respectively. According to the comparative test results in table 1 and table 2, it can be seen that the mechanical properties of the high-voltage grid-connected switch are not significantly different between the reference [4] system and the design system. That indicate that the mechanical properties of the grid-connected switch are well preserved in the design system. in table 3 and table 4, it can be found that the grid-connected switch designed in this paper has no influence on the low-voltage action characteristics of the highvoltage grid-connected switch, because the critical action voltage of the high-voltage grid-connected switch is volatile. In order to further test the performance of the designed system, the energy storage time of the highvoltage grid-connected switch of the two systems was compared and tested. The test results are shown in table 5. According to table 5 shows that compared with the present system [4], the designed system can effectively reduce the high voltage grid switch spring energy storage time. This is because the design system controls the setting time of the relay of the system by adjusting the energy storage circuit of the high-voltage grid-connected switch, so that the energy storage motor of the system can complete the energy storage of the relay spring within the setting time of the relay and leave a certain margin. This effectively shortens the energy storage time of the grid-connected switch, and lays a good foundation for the quick and accurate cutting off of faults in the future. After the completion of the high voltage grid-connected switch spring energy storage, if the system detects the overload, short circuit and other fault currents in the equipment or the circuit, the control command signal will be sent out, and the system will drive the circuit to perform the protection action. In order to test the response time of the designed system, a comparative test of fault response time was carried out at different levels of current, in which the rated working current was set as 8A, that is, the input current of the system microcontroller. Experimental results show that when the power system operation in the process of one phase current overload and short circuit faults on the literature [4] the highest system response time of 8ms and 9ms respectively. Design system response time no more than 4ms, highest can satisfy the real-time requirements of different grade computing tasks. The design system takes less time to store energy compared with the spring energy storage system in literature [4]. Once a fault occurs, it can cut off the fault source in the first time, ensuring the normal, safe and stable operation of power system equipment. conclusion In this paper, the control system of energy storage circuit based on ARM high voltage grid-connected switch is proposed and designed, using ARM series microprocessors with high performance, low energy consumption and high cost performance as the control core. In order to solve the problem that when the high-voltage electrical equipment of unattended substation breaks down, the remote dispatch cannot timely obtain the current state information of equipment and make correct decisions. This leads to the delay of fault treatment, the expansion of the scope of accidents, and the failure of the normal, stable and reliable operation of the power system. The system performance comparison test proves the rationality and feasibility of the design system. The system not only reduces the energy storage time of grid-connected switches, but also improves the response speed of the system, which lays a foundation for large-scale application of high-voltage grid-connected switches.	2019-08-22T20:24:40.005Z	2019-07-25T00:00:00.000	{ "year": 2019, "sha1": "200c23cd3830c5bc8d004c648579421126e05a17", "oa_license": null, "oa_url": "https://doi.org/10.1088/1755-1315/295/4/042007", "oa_status": "GOLD", "pdf_src": "IOP", "pdf_hash": "1eb2ee6594095b2acaca044a50f7417cd2f115c9", "s2fieldsofstudy": [ "Engineering" ], "extfieldsofstudy": [ "Computer Science" ] }
237468502	pes2o/s2orc	v3-fos-license	Snakebites in Rural Areas of Brazil by Race: Indigenous the Most Exposed Group Animal stings are environmental hazards that threaten millions annually and cause a significant socioeconomic impact. Snakebite envenoming affects 2.7 million people globally every year, mostly the poorest and rural communities, with approximately 27,000 annual cases in Brazil. This study’s objective is to identify the most exposed racial group for snakebites in rural areas of Brazil and analyze possible differences in the outcome of an accident. A retrospective epidemiological study was conducted using a database of rural snakebite cases from Brazil’s Ministry of Health (2017). Descriptive analysis and a regression model were performed to examine the association of bad outcomes after a snakebite with several covariables. While mixed-race individuals presented the highest number of cases (61.79%), indigenous and white populations were the racial groups with the highest and lowest exposure rates (194.3 and 34.1 per 100,000 population, respectively). The fatality rate was 3.5 times higher in the indigenous population compared to the white population. In the multivariable model, the number of hours between the accident and health care received and the case classification suggested an association with a bad outcome. Snakebite is prominent in Brazil, particularly among indigenous groups. Antivenom is available in the Brazilian Health System; however, efforts need to be made for decentralization. Introduction Animal bites and stings are important environmental hazards that threaten millions of people annually and cause a significant socioeconomic impact at the individual and country level. Snakebite envenoming is one of these examples that affects close to 2.7 million people globally every year and has an estimated mortality close to 100,000 deaths [1,2]. Approximately 95% of snakebite envenoming occurs in low-and middle-income countries, and snakebites disproportionately involve the poorest of the poor, mostly in rural areas [3]. Improperly treated, envenomation poses a large threat to public health in terms of suffering, morbidity, mortality, and long-term disability [2]. In response to the large global burden of disease due to snakebite and the global crisis in antivenom production that has left millions of vulnerable people with no or difficult access to the antivenom, the World Health Organization (WHO)'s Strategic and Technical Advisory Group for Neglected Tropical Diseases recommended the reclassification of snakebite envenomation as a high priority Neglected Tropical Disease (NTD) at the WHO's Because indigenous groups live largely in rural, and oftentimes remote areas and work in high-risk environments such as subsistence farming, hunting, and gathering, they are frequently among the reported victims of snakebite and might be at greater risk of accidents with these animals than other population groups [26,27]. We believe that indigenous groups are more exposed to snakebites than other population groups; however, to date, this burden has not been measured adequately, and very few studies have taken race and ethnicity into account in their analysis of snakebite in Brazil [28,29]. Brazil is a vast country and the differences among its regions are large, not only ecologically but also culturally and ethnically. Prior studies have explored the demographics of snakebite victims in Brazil [9,18,30,31]; however, none has so far focused on the association with race and ethnicity. The Brazilian surveillance system SINAN includes 'race' as one of their variables in the notification form filled out when a person receives health care after a snakebite in the SUS. The variable 'race' collected in SINAN is self-declared, just as in Brazil's census, which is administered every ten years by the Brazilian Institute of Geography and Statistics (acronym in Portuguese, IBGE) [32]. The objective of this study is to identify which racial groups in rural areas of Brazil are more exposed to snakebites and to compare the most exposed group with the less exposed group by region. We also analyze if there are possible differences in the outcome of the snakebite, considering race and other variables related to the accident. Study Area Brazil is the largest country in Latin America with a geographic area of 8,515,767 km 2 and a population of 190,755,799, according to the latest census conducted in 2010 by the Brazilian Institute of Geography and Statistics [32], distributed in five regions, 26 states, and the Federal District. In 2010, the Southeast Region concentrated 42.1% of the total population. The largest population groups in Brazil by race are the white population (47.7%), followed by the mixed-race population (43.1%). Black and Asian Brazilians constitute 7.5% and 1% of the population, respectively. According to the last census, the indigenous population included 896,917 individuals in the entire country, making up only 0.4 percent of Brazil's population [32]. While Brazil's indigenous population is relatively small, it is ethnically diverse and includes 300 indigenous ethnic groups with speakers of over 200 distinct languages [22]. The majority of Brazil's population lives in urban areas, with only 15.6% residing in rural areas. There are, however, important differences among the regions, with close to 27% of the population living in rural areas in the Northeast and North Regions compared to only 7.1% in the Southeast (Table S1). Most indigenous people in Brazil live in over 600 federally recognized reserves, 98% of which are in the Amazon Region, but a substantial number of indigenous individuals (42.3%) live outside of recognized reserves, and most of the indigenous population lives in rural areas (73.8%) [22]. Around 3% of the total population are agriculture workers [32]. Considerable differences in literacy exist, ranging from 95.9% of the population in the South to 83.8% in the Northeast Region, with a country average of 92.0% [33]. Large differences are also apparent with regards to the GDP per capita, ranging from BRL 34,790 in the Southeast to BRL 12,955 in the Northeast Region (Table S1) [34]. The major habitat type in Brazil is tropical and subtropical moist broadleaf forests (TSMBF) [35]. This habitat type includes the Legal Amazon Region and some areas of the Atlantic Forest in the coastal area, where close to 70% of the snakebites occurred [16]. According to the Ministry of Health of Brazil, the venomous genera with public health interest in Brazil are Bothrops, Crotalus, Micrurus, and Lachesis [36,37]. Bothrops is present in most of the country; Crotalus is least present in the Amazon Region; Micrurus is least present in the Northeast Region, and Lachesis is not present in the South Region. According to previous studies, approximately 70% of snakebite envenoming is caused by Bothrops snakes [9,29,30,37]. The Brazilian Unified Health System runs from primary care to high complexity health services countrywide [24]. In case of an accident with a venomous animal, health care is delivered by SUS and the antivenom and other treatments are provided free of charge for the population [37]. Around 1600 hospitals nationwide, located in all 26 states and the Federal District, provide five types of antivenom for different types of venom, including five types of snake antivenoms: antibothropic pentavalent, antibothropic pentavalent and antilachetic, antibothropic pentavalent and anticrotalic, anticrotalic, and antielapidic (Table S1) [36]. However, due to a lack of health professionals and infrastructure in remote areas (e.g., electricity to ensure an adequate cold chain), particularly in the Amazon Region, access to lifesaving antivenom is limited, leading to a delay of patient care or no medical treatment at all [18,37]. Indigenous health is part of the SUS through the Special Secretariat for Indigenous Health (SESAI), with the participation of the state and municipal health authorities [35]. The Special Indigenous Health District (in Portuguese DSEI: Distritos Sanitários Especiais Indígenas) is the decentralized management unit of the Indigenous Health Care Subsystem (SasiSUS). In Brazil, there are 34 DSEI, strategically divided by territorial criteria, based on the geographical location of indigenous communities. The DSEI do not align with the borders of the states but cut across state lines. This structure for indigenous health care includes around 360 basic indigenous health units in the reserves, centers, and the Indigenous Health House at the national level (in Portuguese, CASAI) [38,39]. This structure manages health care of different complexities and includes provisions for transportation (usually by plane) to other places when needed. In the case of an accident with a serpent in the DSEI's jurisdiction, transportation to a hospital with the capacity to administer the antivenom is usually required. Study Design and Source of Information We carried out a retrospective epidemiological study using the SINAN database by individual cases of accidents with venomous animals from the Ministry of Health of Brazil. The database was requested via the Citizen Information System [40]. According to the information access law (in Portuguese "Lei de Acesso a Informacao" number 12,527 from 18 November 2011), government information can be requested (https://www.gov.br/ acessoainformacao/pt-br) (accessed on 19 March 2019); in this case, the protocol number was 25820.001698/2019-76, of 2019. The notification form collects information about the patient's demographic characteristics, epidemiological data about the accident, case classification, the time between the accident and receiving health care, whether the patient received antivenom and the case evolution. All cases are coded and do not reveal any information related to the identification of the person for ethical reasons. The data include information from several types of accidents with venomous animals such as snakes, spiders, and scorpions. For this study, only snake accidents were selected, and only cases reported in the most recent year in the database (2017). From the total number of cases of snakebite accidents in Brazil in 2017, we selected all rural cases, and among them the ones with information about race. Data on the population by race and by state were obtained from the 2010 national census [32]. The Ministry of Health provides guidelines to health care professionals on how to complete the notification forms, and on the data to be gathered. As cited, the variable 'race' collected in SINAN and in the IBGE census is self-declared, and both institutions combine the categories of 'race' and 'ethnicity' into one. Our study population is all snakebite cases in Brazil reported to SINAN in rural areas during the year 2017, which have information about the racial identification of the victim [19]. Part 1: Descriptive Analysis In the descriptive analysis of the epidemiological situation, cases of snakebite and rates per 100,000 population were described by race, by region, and by state in rural areas of Brazil, and the rate ratio by race compared with less exposed racial groups was calculated. To avoid unstable rates with small populations (under 1000 inhabitants in the specific race group), rates were not estimated for small populations and considered 'not applicable' (NA) in the tables by state. Another part of the descriptive analysis presented covariables related to demographics and the accident, including the bite site on the body, the type of snake related to the accident reported, the number of hours between the accident and health care received (the cut-off was pre-defined by the Ministry of Health in the database), case classification, and if antivenom was administered. The definition of case classification according to the Ministry of Health of Brazil [36] is by the genera of the venomous snake and includes recommendations on what type of antivenom and the number of ampoules to be used for treatment (Table S2). This information is available in the notification form that the physician fills out when an accident happens and in the Ministry of Health guidelines [36]. In the last part of the descriptive analysis, possible differences related to the outcome of the accident were presented, including if there were local complications (such as secondary infections, extended necrosis, and amputation), systemic manifestations (such as acute renal failure, sepsis, and shock), and the evolution of the case (cure, death by snakebite, death by other causes or ignored). The total fatality rate by snakebite in rural areas of Brazil was estimated and calculated for the most exposed and less exposed racial groups. Descriptive analysis was conducted using Stata IC/16 and some figures in Microsoft Excel for Microsoft 365 MSO. For better visualization of snakebite case distribution by state, GIS maps were prepared for the total number of cases, and for cases in the most and less exposed groups, using population data from IBGE as background. The maps were created using ArcGIS versus 10.8, utilizing natural breaks for the population and case ranges. Part 2: Statistical Analysis As the number of deaths was much smaller than the number of cases reported, a new variable was created considering the outcome as a combination of three possibilities: the evolution (0 = cure, no death by snakebite; 1 = death by snakebite), and/or systemic complication (0 = no; 1 = yes), and/or local complication (0 = no; 1 = yes). A binomial logistic regression model was used to estimate the association between the outcome of a snakebite accident (0 = good; 1 = bad) and the covariables race, sex, age, bite site on the body, number of hours between the accident and health care received, case classification, and if antivenom was administered. One person could have one or more of the situations considered as a bad outcome (possible complications and death). Only observations that have all the independent variables completed in the database were included. In this part of the statistical analysis, the mixed and black racial groups were included as a combined group under the designation "black" to simplify the results of the logistic regression. Due to the small number of cases in the Asian racial group (127 cases, 0.8%), this group was excluded from the analysis as it was not relevant to the study, given that they were neither the most nor less exposed racial group. We created categories for the variable age, and we combined some of the existing categories of the variables related to bite site on the body and the number of hours between the accident and health care received. A univariate analysis was conducted followed by a multivariable analysis, and estimation of the odds ratio (OR) with a 95% confidence interval (CI). The statistical analysis was performed using SPSS 24. Part 1: Descriptive Analysis In 2017, the total number of snakebites in Brazil reported to SINAN was 28,716, with 16,805 of them (58.52%) recorded in rural areas ( Figure S1). The One Health approach showing the major habitat types, presence of venomous snakes by genera, GDP per capita, and the number of rural cases of snakebites by region is presented in Figure 1. Analyzing this database, 75.34% of the cases were caused by genus Bothrops, and these percentages are similar in all regions. Accidents with Micrurus constituted 61.86% in the Northeast Region, and accidents with Crotalus were more frequent in the Northeast and Southeast Regions (39.01% and 34.46%, respectively). Accident by Lachesis were predominant in the North Region (90.01%) (Table S3). A univariate analysis was conducted followed by a multivariable analysis, and estimation of the odds ratio (OR) with a 95% confidence interval (CI). The statistical analysis was performed using SPSS 24. Part 1: Descriptive Analysis In 2017, the total number of snakebites in Brazil reported to SINAN was 28,716, with 16,805 of them (58.52%) recorded in rural areas ( Figure S1). The One Health approach showing the major habitat types, presence of venomous snakes by genera, GDP per capita, and the number of rural cases of snakebites by region is presented in Figure 1. Analyzing this database, 75.34% of the cases were caused by genus Bothrops, and these percentages are similar in all regions. Accidents with Micrurus constituted 61.86% in the Northeast Region, and accidents with Crotalus were more frequent in the Northeast and Southeast Regions (39.01% and 34.46%, respectively). Accident by Lachesis were predominant in the North Region (90.01%) ( Table S3). The total number of rural observations of snakebites with information on the race of the victim was 15,898, which constitutes our study population. The North Region, which is part of the Amazon Legal Region, has the highest number of snakebite cases in Brazil ( Figure 2). The distribution of snakebite cases by state demonstrates that Para state in the North Region has the highest number of cases (3420 cases), followed by Minas Gerais (1710) in the Southeast Region, and Bahia (1534 cases) in the Northeast Region (Table S4). The total number of rural observations of snakebites with information on the race of the victim was 15,898, which constitutes our study population. The North Region, which is part of the Amazon Legal Region, has the highest number of snakebite cases in Brazil ( Figure 2). The distribution of snakebite cases by state demonstrates that Para state in the North Region has the highest number of cases (3420 cases), followed by Minas Gerais (1710) in the Southeast Region, and Bahia (1534 cases) in the Northeast Region (Table S4). The racial group with the most cases was the mixed racial group with 9823 (61.79%), and the racial group with the lowest number of cases was the Asian group with 127 recorded snakebites (0.80%) ( Figure S1). However, when estimated as rates per 100,000 population, the racial group most exposed to snakebite was the indigenous population, with a rate of 194.3 per 100,000 population (977 cases), and the less exposed racial group was the white population with 34.1 per 100,000 population (3701 cases) ( Table 1). The rate ratio comparing the most exposed and the less exposed racial group was 5.7, suggesting that the probability of an indigenous individual being bitten by a snake is approximately six times higher than that of a white person in Brazil. In the Northeast Region this probability is even higher (8.9 times), whereas it is lower in the Southeast (1.8 times) and South (2.5 times) Regions. The type of snake related to the accident by racial group is presented in Table 2, calling to attention that the genus Lachesis presented higher percentages in the indigenous populations (5.22%) and Crotalus smaller (5.42%), compared with the total population (2.16% and 7.8%, respectively), and with the white populations (0.46% and 9.59%, respectively). The racial group with the most cases was the mixed racial group with 9823 (61.79%), and the racial group with the lowest number of cases was the Asian group with 127 recorded snakebites (0.80%) ( Figure S1). However, when estimated as rates per 100,000 population, the racial group most exposed to snakebite was the indigenous population, with a rate of 194.3 per 100,000 population (977 cases), and the less exposed racial group was the white population with 34.1 per 100,000 population (3701 cases) ( Table 1). The rate ratio comparing the most exposed and the less exposed racial group was 5.7, suggesting that the probability of an indigenous individual being bitten by a snake is approximately six times higher than that of a white person in Brazil. In the Northeast Region this probability is even higher (8.9 times), whereas it is lower in the Southeast (1.8 times) and South (2.5 times) Regions. The type of snake related to the accident by racial group is presented in Table 2, calling to attention that the genus Lachesis presented higher percentages in the indigenous popula-tions (5.22%) and Crotalus smaller (5.42%), compared with the total population (2.16% and 7.8%, respectively), and with the white populations (0.46% and 9.59%, respectively). Table 1. Snakebite rate per 100,000 population; rate ratio of more exposed race group by less exposed race group per state and per region, Brazil, 2017. Legend: NA: Not applicable in the states with less than 1000 population in these race groups. Snakebites affected mostly males (76.4%), and the average age across all populations was 36 years old, with the white population tending to be older than the indigenous population. Half (55.2%) of all bites occurred in lower limbs, particularly in the feet ( Figure 3); for the indigenous population, this percentage was 66.1% compared to 49.4% for the white population. Over 75.2% of the white population received care between 0 and 3 h after being bitten, compared with only 37.4% of indigenous individuals who received care during this timeframe. States Snakebites affected mostly males (76.4%), and the average age across all populations was 36 years old, with the white population tending to be older than the indigenous population. Half (55.2%) of all bites occurred in lower limbs, particularly in the feet ( Figure 3); for the indigenous population, this percentage was 66.1% compared to 49.4% for the white population. Over 75.2% of the white population received care between 0 and 3 h after being bitten, compared with only 37.4% of indigenous individuals who received care during this timeframe. Figure 3. Number of snakebites by site on the body; total (A), the most exposed race group (indigenous) (B), and the less exposed (white) (C), rural population, Brazil, 2017. Related to the case classification, 52.8% of cases were considered mild, 37.6% moderate, and 7.7% severe, without significant differences among the racial groups (Table 3). Among the snakebite cases reported in the SINAN database, 79.7% of the cases received antivenom, with the indigenous population receiving slightly more antivenom than the general population (Table 3). Related to the case classification, 52.8% of cases were considered mild, 37.6% moderate, and 7.7% severe, without significant differences among the racial groups (Table 3). Among the snakebite cases reported in the SINAN database, 79.7% of the cases received antivenom, with the indigenous population receiving slightly more antivenom than the general population (Table 3). Our analysis of clinical outcomes found that local complications were reported in 569 cases (3.9%), with indigenous groups presenting a higher percentage (5.9%) of complications. The most frequent local complication reported was secondary infection (73.5%) ( Table 4). The number of cases with systemic complications was 173 (1.2%), showing almost no differences among the racial groups; the systemic complication most frequently cited was acute renal failure (n = 137, 79.2%) ( Table 4). According to the SINAN database, among the 15,898 cases reported in rural areas in 2017 with information on race, a total of 14,260 cases had information about the case evolution. Of those, a total of 52 snakebite victims died because of their accident occurring in rural areas of Brazil in 2017 (three more cases were reported without race information), indicating a fatality rate of 0.4% in rural areas. The indigenous and white populations recorded the same number of deaths (6 cases in each racial group); however, the fatality rates among the two racial groups were different: an estimated 0.7% for the indigenous and 0.2% for the white group-3.5 times higher in the indigenous group (Table 4). Part 2: Statistical Analysis The full outcome of the binomial logistic regression model with all variables estimates is included in Table S5. In the statistical analysis (N = 13,710), the univariate analysis of the indigenous race group suggested an association with a bad outcome of the snakebite accident (OR = 1.63; CI 95% = 1.19-2.24); also, an association was suggested with several of the covariables (Table 5). However, in the multivariable model, only the number of hours between the accident and health care received and the case classification suggested association (Table 4). Increasing the number of hours between the accident and health care increased the risk of a possible bad outcome: three to six hours (OR = 1.38; CI 95% = 1.12-1.70), six to 24 h (OR = 1.48; CI 95% = 1.18-1.86), and more than 24 h (OR = 2.99; CI 95% = 2.21-4.06). The classification of the accident as moderate (OR = 3.29; CI 95% = 2.65-4.01) or severe (OR = 11.8; CI 95% = 9.29-15.97) was associated with a bad outcome, whereas the variable race lost its significance in the multivariable model. The risk of an accident with a snake resulting in a bad outcome is 4.7% (a total of 644 cases with bad outcome). This percentage increases with the number of hours between the snakebite and the provision of health care, and in severe cases this percentage increased considerably (from 15.2% with health care received until three hours to 35.7% when more than 24 h had passed). The distribution of a bad outcome in the case of snakebite envenoming among the regions suggested that the Central-West, North, and the Northeast Regions presented higher percentages of bad outcomes (5.8%, 5.7%, and 5.7%, respectively) ( Table 6). In two states in the North Region, the percentage of bad outcomes related to the snakebites was higher than 10% of the total accidents reported ( Table 6). Discussion Snakebite envenoming is a major public health problem in Brazil, affecting close to 30,000 people each year. To save lives, a large number of ampoules of antivenom needs to be distributed regularly across the entire country. The Ministry of Health in Brazil has been distributing around 230,000 ampoules of antivenom to approximately 2000 hospitals free of charge annually [16], which is likely one of the reasons for the low mortality rates associated with snakebites in Brazil found in this and prior studies [9,16,[28][29][30]41]. Brazil's case-fatality rate for snakebites of around 0.4%, found in our study and the previous study by Bochner [9], is particularly low if compared with studies in other countries, such as the reported case-fatality rate of 3% in a recent district-level study in Cameroon [42], and an estimated case-fatality rate of 3.2% for in-hospital cases of snakebite envenoming in India [43]. According to Chippaux [44], the model that exists in Brazil regarding antivenom distribution, as well as existing infrastructure that ensures treatment for many people, is a good model for other low-and middle-income countries, including those in Africa and Asia. The most exposed racial group in Brazil are indigenous people, who have a risk of being bitten close to six times higher than the less exposed white population group, and a fatality rate of more than three times that of the white population (0.7% versus 0.2%). The actual risk is probably higher, given that it is likely that a significant percentage of indigenous snakebite victims do not seek health care immediately after the accident. Indigenous peoples hold their own cosmologies, which also guide their interpretations for illness processes and cure options [45]. In the Amazon Region, traditional medical practices such as herbalism and shamanism remain widespread, and the limited access to health care and deeply embedded cultural beliefs mean that rural Amazon communities are frequently using traditional medicine (e.g., plants) to treat ailments and diseases [46], including as the first treatment for snakebite, which may aggravate the conditions at the bite site [18]. A 2017 study by da Silva Souza et al. [29] estimated that the underreporting of deaths due to snakebite envenomation in Amazonia is at nearly 30%, due to lack of follow-up with the patient and underlying health conditions (so that snakebite is the secondary cause of death). The referred study also emphasizes that indigenous populations and remote riverside communities seem to be at higher risk due to agricultural activity and that low accessibility of antivenom due to a greater distance from health care centers contributes to higher case-fatality rates [29]. In Brazil, different medical traditions coexist, including Western biomedicine, popular and traditional medicine involving healers and herbalists, as well as shamans [47]. These different systems of medicine are not mutually exclusive and rural patterns of resort include self-care, visits to traditional healers and the use of indigenous medicines, and utilizing modern medicine as a last resort or in combination with indigenous medicine. The use of traditional herbs and herbal extracts to treat Bothrops envenoming has been observed in riverside communities in the state of Para, which is the state most affected by snakebites in Brazil [47]. Other studies in Latin America demonstrated that as much as 60% of snakebite cases were treated by traditional healers [8]. Indigenous and rural populations are also known to use tourniquets, chemicals such as alcohol (applied or ingested), as well as puncture or suction of the bite site to remove the venom or reduce its effects [26]. Access to health care resources in the Amazon Region is still a challenge in Brazil and its neighboring countries [21,25,26,46,48], and a more inclusive health care system is needed, which focuses on rural, low-income populations, including indigenous people and mixed-race riverside communities. This is particularly urgent, as close to 40% of Brazil's indigenous population lives outside of the designated indigenous reserves and is therefore not covered by the services provided through the Special Indigenous Health Districts. Even though over the past three decades, Brazil's Unified Health System has undertaken efforts to guarantee more equitable access to health services for the indigenous population through the establishment of the Indigenous Health Subsystem, indigenous Brazilians still suffer from the worst health status of any population group [21,22]. Health care should be culturally sensitive, targeted, and tailored to the needs of the indigenous and other low-income populations exposed to snakebites, particularly in remote areas of the North Region. Health education programs should follow the principle of interculturality, incorporating local cultural knowledge to prevent accidents, and to inform the work of primary health agents in innovative activities at the local level, in support of the national program providing the antivenom. Transdisciplinary studies, using the One Health approach, are recommended to better understand snakebite in the context of communities living in close contact with nature, particularly Brazil's indigenous peoples. Understanding indigenous perceptions of the causes and consequences of snakebites and the barriers they face in accessing emergency medical care will help to improve the prevention of accidents and reduce the burden of disease caused by snakebites in this population [16,49]. Although in most states, the indigenous population of Brazil is small compared to other racial and ethnic groups, they constitute a sizable population in the states that belong to the Amazon Region. There are officially 34 Special Indigenous Health Districts, and it is crucial that these units of the Indigenous Health Subsystem always have the antivenom available, that health care personnel are trained to administer it, and have the local conditions in place to respond to health emergencies. However, since rural, multi-ethnic riverside populations in the Amazon are also at higher risk for snakebite [16,26], they too require special attention. Due to the large geographical distances, absence of roads, lack of regular transportation, and remoteness of these communities, provision of adequate primary health care services has been challenging. Over the past decades, the Brazilian health authorities have implemented fluvial mobile units as an alternative health care model to increase access and health care coverage in the Amazon basin [25]. Although access to routine services has improved through the monthly visits of the Fluvial Family Health Teams, due to the intermittent nature of the visits, the fluvial mobile units cannot be relied upon to provide the needed emergency services in the case of severe snakebite accidents. At the same time, the only health professionals permanently stationed in these remote areas are community health workers and sometimes nurses and nurse assistants. These health care workers are often based at smaller health centers that lack access to stable electricity and therefore storage facilities for antivenoms and other drugs, which require refrigeration [25,50]. This means that snakebite victims frequently need to travel for 12 h or more-reportedly up to 30 h-via boat to the nearest town or city to receive the lifesaving antivenom from a larger health care facility, which often leads to serious complications and negative health outcomes [50]. Our results presenting the genera of the snakes related to the accidents demonstrated that among the indigenous population, the percentage of cases caused by Lachesis was more than double compared with the total population; accidents with Lachesis are always classified as moderate to severe [36]. The indigenous reserves are located in areas where Lachesis are more present geographically, mostly in the Amazonian Region, confirming the urgent need to have the antivenom available on the reserves. In our study, the region that presented the highest rate ratio between the two race groups (indigenous and white) was the Northeast Region, which has the lowest socioeconomic indicators and the largest rural population in Brazil. Most studies on indigenous populations in Brazil have been carried out in the Amazon Region [45], and more research should be conducted to understand if the indigenous population living in lowsocioeconomic areas outside of the Amazon Basin is even more vulnerable to snakebites and other health issues. In the states of the Southeast and South of Brazil, which have better socioeconomic indicators, the differences among the race groups are smaller but still present in most of the states. The demographics related to sex and age described in this study are similar to those identified in previous studies [9,14,30] and by the Ministry of Health of Brazil [37]: snakebite cases are most common among males (76.4%) in the productive age group (average age of 36 years). In the study population, the most frequent location of the bites were the lower limbs, particularly the feet, as confirmed in previous studies in Brazil [9,37], whereas studies in other parts of the world also show hands as frequently affected parts of the body [51,52]. This is likely due to the frequent use of sandals in Brazil's tropical climate, as well as to the custom of many indigenous Brazilians to walk barefoot in the Amazon rainforesta behavior which is reflected in the higher percentage of bites in the feet among this population group (66.1% versus 49.4%). This finding suggests that for rural communities in countries with tropical climates, more effective primary prevention strategies are needed, such as wearing practical, and at the same time, culturally appropriate protective gear outdoors. Like previous studies, our study confirmed that the majority of accidents with venomous snakes in Brazil are related to the genus Bothrops [9,36], and that this is similar in the different racial groups. The average percentage of cases that received health care within three hours after the accident was 61.2%, which is a positive finding considering the country's large territory, as well the mobility challenges due to its geography, mainly in the Amazon Region. The difference between the most and less exposed racial groups is great, however, with 75.2% of the white population receiving care between 0 and 3 h after being bitten, compared to only 37.4% of indigenous individuals receiving care during this time. This suggests a delay in medical care due to difficulty in accessing health care facilities with antivenom and other health services for indigenous individuals. According to Chippaux, "mortality results from both the toxicity of the venom, associated with the amount of inoculated venom, and the precocity and effectiveness of the treatments" (p. 9, [30]). In our study, an increase in time to accessing health care was associated with bad outcomes for accidents with venomous snakes, including severity and mortality. This finding, which aligns with previous studies, is particularly relevant in the North Region of Brazil, where higher numbers of cases occur and venomous snakes such as Bothrops atrox and Lachesis are common [16,26,30,53]. As expected, in the final statistical analysis, the variables related to the severity of the case classification presented a higher association with bad outcomes. In our study, around half of the cases that arrived in the health system were classified as mild, a percentage similar to the one reported by previous studies and the Ministry of Health [30,37]. However, even though the Ministry of Health guidelines define the criteria for the case classifications related to severity, it has been suggested that in practice, the severity classification is often empirically established, and we propose that more studies should be conducted at the hospital level [25]. In Brazil's Legal Amazon Region, which includes all seven states of the North Region as well as parts of the Central-West and Northeast Regions, where access to health care is limited and difficult, and the Northeast Region, which has a comparatively low GDP per capita, the percentage of bad outcomes is higher, suggesting that training local health care professionals in the management of envenomation and reviewing the location and number of hospitals and health care centers with the antivenom could substantially improve case outcomes. It is important to remember that the current antivenoms require conservation in adequate facilities, which are not always available in remote settings. In addition, training of health personnel in antivenom administration, side effect management, and case monitoring is required. One of the limitations of this study is the use of secondary data from the Ministry of Health. Data may be underreported, such as the number of cases that received health care following a snakebite. Furthermore, a percentage of cases likely did not seek health care, particularly among the indigenous population living in remote areas. The risk of indigenous Brazilians being bitten by a snake and dying from the accident is likely higher than presented in this study and requires further research to measure with greater accuracy the burden of snakebites in this racial group. Additional research should be conducted to examine the likely sub notifications in the number of deaths. However, the SINAN database was also used as a source of information for previous studies on snakebite envenomation [12,16,30,31] and provides the best secondary data currently available. Since the government requests compulsory notification of all snakebite cases and these data are used by decision-makers at the national level to determine the distribution of antivenom to the states, it is the most reliable and complete data source in the country. Another limitation is the definition of rural areas used by IBGE, which is not very precise but still useful for the purpose of this study, as it provides insight into a particularly vulnerable and often marginalized population, which is considered most at risk for snakebites [2]. The self-declared category of race is another limitation, but since the focus of the analysis was on the comparison between white and indigenous races, the possible bias of classification is lower than if the study had focused on the more ambiguous mixed-race category. The decentralized distribution of the antivenom to more remote places will only be possible if there is a higher number of ampoules available for purchase by the government. Currently, the amount available for government purchase in Brazil is already reduced due to restrictions in the productive capacity of the laboratories, with only four public laboratories capable of producing the antivenom [16], and, as of the publication of this paper, only one public laboratory actually producing the antivenom in Brazil. The ongoing COVID-19 pandemic has led these laboratories to focus their efforts on COVID-19-related research, likely reducing the amount of antivenom produced. This could affect not only the amount of antivenom available for snakebites but also for spider and scorpion envenomation, which affects more than 120,000 Brazilians each year, as well as fatal infectious diseases such as rabies, which also require serum for post-exposure prophylaxis. A survey with the manufacturing institutions was developed to appraise the antivenom production in public laboratories in Latin America during COVID-19 [6]. Among the different aspects analyzed was the demand from the national public health institutions that remained stable during the pandemic; however, five of the twelve laboratories included in the survey did not manufacture during the first half of 2020. "The reasons for halting the production in these cases were the need to carry out improvements in infrastructure for fulfilling the requirements of Good Manufacturing Practices (GMPs) in four cases and the restrictions to do face-to-face work in one institution" (p. 2, [6]). In recent years, Amazon deforestation, as well as more comprehensive reporting of snakebite envenomation has resulted in an increased incidence rate [30]; in previous studies, the association of snakebite and tree loss was demonstrated, supporting this argument and reinforcing the need to provide access to the antivenom in remote areas of the Amazon Region [16]. According to Gutierrez [1] and adopted in the WHO strategy [2], confronting snakebite envenoming at a global level demands the implementation of an integrated intervention strategy, which includes the research community, antivenom manufacturers, regulatory agencies, national and regional health authorities, professional health organizations, international funding agencies, advocacy groups, and civil society institutions. In Brazil, important steps have been taken in this direction since the medical scientist Vital Brazil started the production of antivenom in the early 20th century, including numerous scientific studies on accidents with venomous animals, and the creation of the National Program for Snakebite Control in the 1980s with the countrywide distribution of the antivenom free of charge [9,16]. However, the global manufacturing crisis of antivenom and the reduction of the number of laboratories producing antivenom in Brazil pose a great challenge to the equitable distribution of ampoules across the country, and particularly in remote areas of the Amazon Basin. Conclusions Indigenous people are the most exposed racial group to snakebites in rural areas of Brazil and present higher fatality rates as compared to the white population, which is the racial group less exposed to snakebites. Even though the antivenom is available free of charge in the Brazilian Health system in all states, in 2017 there were still 55 (52 with race information) deaths recorded in rural areas among the roughly 16,000 cases of snakebites officially reported. The fatality rate among the two racial groups compared in this study was around three times higher for the indigenous group (0.7% versus 0.2%) than for the white group. However, in the final multivariate model, the only covariables associated with bad outcomes after a snakebite were the number of hours between the accident and health care received and the severity of the case. When controlling for all selected variables, race was not associated. If the indigenous population would be able to receive the antivenom in a shorter amount of time, the differences in fatality rates among the two race groups and number of bad outcomes would likely be smaller, suggesting that an effort to decentralize the antivenom to indigenous reserves is an important step to save lives due to snakebites in remote areas. Efforts should also be made to make antivenom more accessible, mainly to the low-income and rural population in the Amazon Region, as transportation in this region is often slow due to the geographic conditions, increasing the number of hours required to arrive to a hospital with antivenom. However, despite the country's effort in purchasing and distributing snakebite antivenom for decades, in order to increase this life-saving product decentralization, it needs to be available for purchase. This study clearly shows that snakebite envenoming continues to be an important public health problem in rural areas of Brazil, primarily in indigenous communities, which requires special attention. Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/ijerph18179365/s1, Table S1: Country and region profiles, total population, percentage of population living in rural areas, literacy rate (as % of population 15 years and older), GDP per capita (Reais), number of hospitals with antivenom, Brazil; Table S2: Case classification by genera of the snake according to the Ministry of Health guideline and the number of ampoules recommended; Table S3: Cases of snakebite with reported information on the genus of the snake, in rural population, by state, Brazil, 2017; Figure S1: Flowchart of the study design. Results of the study design with the number of observations of snakebites in each category, Brazil, 2017; Table S4: Populations and cases. Rural population and snakebite cases by race group, Brazil, 2017. Table S5. Estimation of all variables used in the binomial logistic regression model.	2021-09-11T06:17:04.864Z	2021-09-01T00:00:00.000	{ "year": 2021, "sha1": "5105fe475427b07dbc50e855a32a6dc3228ca1dc", "oa_license": "CCBY", "oa_url": "https://www.mdpi.com/1660-4601/18/17/9365/pdf", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "3e4c805b809c24955a3634b49f6ac3fac439fc3c", "s2fieldsofstudy": [ "Medicine", "Environmental Science", "Sociology" ], "extfieldsofstudy": [ "Medicine" ] }
3575923	pes2o/s2orc	v3-fos-license	Is There a Canonical Cortical Circuit for the Cholinergic System? Anatomical Differences Across Common Model Systems Acetylcholine (ACh) is believed to act as a neuromodulator in cortical circuits that support cognition, specifically in processes including learning, memory consolidation, vigilance, arousal and attention. The cholinergic modulation of cortical processes is studied in many model systems including rodents, cats and primates. Further, these studies are performed in cortical areas ranging from the primary visual cortex to the prefrontal cortex and using diverse methodologies. The results of these studies have been combined into singular models of function—a practice based on an implicit assumption that the various model systems are equivalent and interchangeable. However, comparative anatomy both within and across species reveals important differences in the structure of the cholinergic system. Here, we will review anatomical data including innervation patterns, receptor expression, synthesis and release compared across species and cortical area with a focus on rodents and primates. We argue that these data suggest no canonical cortical model system exists for the cholinergic system. Further, we will argue that as a result, care must be taken both in combining data from studies across cortical areas and species, and in choosing the best model systems to improve our understanding and support of human health. INTRODUCTION Cholinergic modulation has been associated with a range of cognitive processes including learning and memory consolidation (Hasselmo and McGaughy, 2004), reward and addiction (Maskos et al., 2005;Kobayashi and Okada, 2007;Chubykin et al., 2013), plasticity (Bear and Singer, 1986;Weinberger, 2003), the sleep/wake cycle and arousal (Jasper and Tessier, 1971), perception (Kosovicheva et al., 2012;Boucart et al., 2015;Gratton et al., 2017), and attention (Everitt and Robbins, 1997;Sarter et al., 2005). In much of the literature exploring a role for cholinergic modulation, there is an implicit assumption that the various model systems used in these studies are equivalent. As such, the results of many experiments-spanning methodologies, species and cortical areas-are combined into unified models of function. This step assumes the existence of a canonical circuit for the cholinergic system in cortex. Even for cases of apparent functional equivalence, as will be discussed, the cholinergic system across model systems fails to exhibit anatomical equivalence. These anatomical differences make for unique cholinergic circuit compositions that may subserve similar processes across systems. With a primary focus on data from the rodent and primate cortex, we will describe anatomical differences in the cholinergic system both across and within species. We focus on these species not because we believe they are the only ones from which we can learn, but because they are the ones for which the published data are most complete. We argue that these existing data reveal that no canonical circuit for acetylcholine (ACh) exists in cortex. Further, they buttress the importance of consideration for species, cortical area and method of origin that inform functional models. STRUCTURE: CHOLINERGIC NUCLEI In all mammalian species studied to date, the cell bodies of cholinergic projection neurons are located in subcortical nuclei of the brainstem and forebrain. Two major clusters of nuclei exist: the brainstem cholinergic system and the basal forebrain. The brainstem system is comprised of the pedunculopontine tegmental nucleus and the laterodorsal pontine tegmentum, which project to the basal ganglia, the thalamus and the basal forebrain. This review emphasizes the effects of ACh in cortex, and therefore focuses on the basal forebrain. In primates, the basal forebrain is made up of the medial septal nucleus that projects to the hippocampus and entorhinal cortex, the diagonal band of Broca that projects to the hippocampus, entorhinal cortex and olfactory cortex, and the nucleus basalis/substantia innominata that provides ACh to the rest of cortex and to the amygdala (Mesulam et al., 1983). Neurons whose cell bodies reside in the basal forebrain are the only source of cortical ACh in adult primates (Mesulam et al., 1983). However, cell bodies immunoreactive for the synthetic enzyme choline acetyltransferase (ChAT) have been reported in the cortex of rats (Houser et al., 1983;Ichikawa and Hirata, 1986), cats (Avendaño et al., 1996), and fetal monkeys (Hendry et al., 1987;discussed below). In primates, the cholinergic projection neurons in the basal forebrain nuclei correspond to partially overlapping cell groups termed Ch1-Ch4. The medial septal nucleus corresponds to Ch1, the vertical limb of the diagonal band of Broca to Ch2, the horizontal limb of the diagonal band of Broca to Ch3, and the nucleus basalis/substantia innominata to Ch4. Similarly, the pedunculopontine tegmental nucleus and the laterodorsal pontine tegmentum of the brainstem projection system correspond to Ch5 and Ch6, respectively (Mesulam, 2004). There has been some debate regarding the extent to which the ''Ch'' nomenclature can extend to species beyond primates (i.e., to rodents and carnivores; Butcher and Semba, 1989), as the differentiation of cholinergic nuclei may be more subtle in other species (Gorry, 1963). Generally, however, the Ch1-Ch4 schema is considered applicable across species (Mesulam et al., 1983). The ''cholinergic'' nuclei of the basal forebrain actually comprise a heterogeneous population of intermingled cholinergic, γ-aminobutyric acid (GABA)-ergic, and glutamatergic neurons (Mesulam et al., 1983;Gritti et al., 1997Gritti et al., , 2006. Importantly, the proportions of the total neuronal population that each of these subpopulations represents are species-specific and differs between cholinergic nuclei within a single species. The methods used to define and quantify the cholinergic population within each nucleus have differed between studies, but the available data suggest a profound species difference between rodents and primates. Mesulam et al. (1983) report the proportion of cholinergic neurons in each of the Ch1-4 nuclei. The within-species variation is striking: 1% cholinergic in Ch3, through ∼10% in Ch1, up to 70% and 90% in Ch2 and Ch4, respectively. The most comparable study in rodents does not report the proportion of cholinergic neurons for each nucleus individually, but the value of 5% given for the entire basal forebrain (Gritti et al., 2006) seems unlikely to arise from a population similar to that found in the primate, given that two large nuclei (Ch2 and Ch4) in the macaque each have a proportion many-fold higher than 5%. The focus of this review is on cholinergic modulation in cortex. As such, the nucleus basalis/substantia innominata complex (NB/SI; the source of most cortical ACh, described above) will be the focus moving forward. The NB/SI in all species studied is heterogeneous. In humans, Ch4 alone contains approximately 220,000 neurons per hemisphere (Arendt et al., 1985). When one considers that Ch4 is one of two large nuclei in the basal forebrain (the other being Ch2) this would lead one to predict well over half a million neurons for the human basal forebrain total. For comparison, in rats, Ch1-4 together (i.e., the entire basal forebrain) contain 355,000 neurons per hemisphere (Gritti et al., 2006). Interestingly, Raghanti et al. (2011) report the number of cholinergic neurons in the human NB/SI to be between 200,000 and 230,000 per hemisphere, strikingly similar to the estimates for the total neuron number in human NB/SI. Accounting for individual variation, this estimate for cholinergic neurons may reflect a finding that in humans-as in macaques, but unlike rodents-90% of NB/SI neurons are cholinergic. Further species differences exist between primates regarding the number of cholinergic neurons in the NB/SI. The chimpanzee NB/SI is estimated to have 150,000-160,000 cholinergic neurons; estimates of that number for macaques range from approximately 90,000-120,000. Further, these estimates differ in tamarins, owl monkeys, capuchins, etc. (estimates are per hemisphere; Raghanti et al., 2011). These differences could be due to brain scaling (i.e., the number of neurons in the NB/SI increases as brain mass increases across animals). This, however, does not seem to be the case as the number of neurons in the NB/SI appears to increase at a slower rate than increases in brain mass across primates species (known as ''hypo-scaling''; Raghanti et al., 2011). Focusing specifically on cortically projecting neurons (and ignoring local circuit interneurons within the basal forebrain itself and neurons that project to other subcortical structures), two studies have combined retrograde tracing from the frontal lobe with visualization of ChAT immunoreactivity. In the rat, Gritti et al. (1997) have found that just under 20% of the cortically projecting basal forebrain neurons express ChAT, while in the macaque, the proportion is strikingly different at 96% (Mesulam et al., 1983). Proportions were similar for projections to the parietal lobe (rats: Gritti et al., 1997) and temporal lobe (primates: Mesulam et al., 1983). If the proportion of cholinergic and non-cholinergic neurons projecting to cortex from the basal forebrain differs between species this profoundly, it seems likely that the cortical modulation arising from activation of the basal forebrain nuclei may also be mechanistically different between species. It is interesting to note here that while there has been an overall hypo-scaling of subcortical areas with respect to cortical expansion over the evolutionary history of mammals (Raghanti et al., 2011;Collins et al., 2013), this increase from 20% to 96% ChAT-immunoreactive neurons suggests a massive functional hyper-scaling of the cholinergic projection to cortex in primates. Comparative studies are under way in our lab to determine what happened in the evolutionary history of the basal forebrain, and when; have rodents gained a large GABAergic population, or have primates lost one? Cortical projections to the NB/SI of the macaque originate in the orbitofrontal cortex, temporal pole, prepyriform cortex, entorhinal cortex, inferotemporal cortex, insula and prefrontal cortex (Mesulam and Mufson, 1984). Subcortical projections to the NB/SI of the macaque originate in the medial hypothalamus, septal nuclei, nucleus accumbens-ventral pallidum and the amygdala (Price and Amaral, 1981;Mesulam and Mufson, 1984). NB/SI afferents in rats and cats are similar to primates (Haring and Wang, 1986;Irle and Markowitsch, 1986;Jones and Cuello, 1989). The NB/SI in rats and monkeys is also the target of extensive input from other neuromodulatory systems including the serotonergic, dopaminergic and noradrenergic systems (reviewed by Mesulam, 2004). Data are needed to describe the status of neuromodulatory inputs to the NB/SI in other species (such as the cat). STRUCTURE: CHOLINERGIC PROJECTIONS Projection neurons of the Ch4 group innervate all of the neocortex, but the densities of these projections differ between and within cortical areas in all species studied to date. This is a notable characteristic of a cortical area because a denser network of cholinergic axons is likely to yield a higher density of cholinergic release sites, leading to more ACh being delivered. Unfortunately, the same type of data are not available for all animals studied, making comparison imperfect and interpretation difficult. There are, however, many examples of both species-and area-specific patterns of cholinergic innervation in cortex (examples briefly summarized in Table 1). In the human and the cat, primary visual area (V1) is less densely innervated by cholinergic axons compared to other primary sensory and motor areas (Mesulam et al., 1992;Avendaño et al., 1996). Cholinergic projections in the macaque also show clear differences between cortical areas, with macaque prefrontal area 4 (primary motor cortex, M1) being more densely innervated than prefrontal areas 9 and 32 (Raghanti et al., 2008). Further, Lysakowski et al. (1989) report temporoparietal association areas (e.g., areas 5 and 7) in the rat are least densely innervated compared to all other cortical regions examined. This study includes rat V1, which-as mentioned above-is the least densely innervated cortical area in both humans and cats, indicating an interesting species difference in relative V1 innervation. From an evolutionary perspective, this is notable as it appears to violate a reasonable assumption that animals phylogenetically closer to each other (e.g., humans and rats) would possess more similar characteristics than animals phylogenetically farther apart (e.g., humans and cats). As demonstrated in Table 1, data are needed to determine where along evolutionary development changes in cholinergic innervation patterns may have occurred. For instance, it could be hypothesized that the innervation pattern in rat cortex is in fact the outlier compared to carnivores and primates. Data from a wider range of species and sampling from a broader range of cortical areas will be needed to resolve questions of the evolutionary origin and significance of observed differences. Looking at clusters of cortical areas, rather than each area individually, distinct cholinergic innervation patterns are again evident between species. For example, unlike in the cat, rat primary sensory areas have innervation patterns that are more similar to each other, but differ from motor and frontal areas (Lysakowski et al., 1989;Avendaño et al., 1996). In the frontal cortex of humans, chimpanzees and macaques, cholinergic axons are most dense in area 4 relative to areas 9 and 32 (Raghanti et al., 2008). This differs from rat cortex, where frontal areas (such as areas 4, 8, 10 and 11) are more uniformly innervated (Lysakowski et al., 1989). Further, while primate species studied all show highest cholinergic density in area 4, area 9 is least densely innervated in humans, area 32 is least densely innervated in macaques, and both areas show comparable density in chimpanzees (Raghanti et al., 2008). V1 and primary auditory cortex A1, the densest cholinergic innervation is found in layer I and at the border between layers IV and V; the least dense innervation is found in layers II/III and VI (Lysakowski et al., 1989). In fact, this study reports as many as 13 different cholinergic laminar innervation patterns across rat cortex. The laminar pattern of cholinergic innervation in cat M1 shows densest innervation in layers I, II and superficial III, while the pattern in V1 is a densely innervated layer I and more uniform innervation in the remaining layers (Avendaño et al., 1996). In macaque A1, cholinergic axons are most dense in layers I, III, and IV (Campbell et al., 1987). However, while layer I is also found to be densely labeled in macaque frontal cortex, layers III and IV are least densely innervated (Lewis, 1991). In all frontal areas in primates, the superficial layers are more densely innervated than are the deeper layers. Interestingly, however, the morphological features of the axons differ between primate species, even where the innervation densities are similar. For example, cholinergic axons forming ''clusters'' are present in both human and chimpanzee frontal cortices, but not in macaque. These clusters have been previously described as possible indicators of cortical plasticity events (Mesulam et al., 1992). Thus, even across species with similar innervation, the exact morphologies of cholinergic axons may differ and contribute to functional circuit differences. FUNCTION: LESIONS OF THE BASAL FOREBRAIN The basal forebrain cholinergic system has long been implicated in modulating cognitive processes and brain states through its widespread cortical projections (described above). Early studies in rats used excitotoxic lesions of basal forebrain nuclei and were interpreted as indicating that the cholinergic system is particularly involved in learning and memory (Dubois et al., 1985;Hepler et al., 1985). However, excitotoxic lesions eliminate all cells types within the basal forebrain, and it is now apparent that cholinergic neurons in the rat represent a minority of basal forebrain neurons. Thus, it was realized that attributions of these effects to the cholinergic system had to be treated with caution. The development and use of the cholinergic-specific toxin 192 IgG-saporin (Wiley et al., 1991) revealed that, in rats, lesions that leave non-cholinergic cells intact result in a somewhat different pattern of cognitive deficits. For instance, performance in learning and memory tasks following cholinergic-specific lesions was either unimpaired or not as severely impaired compared to non-specific lesions (Berger-Sweeney et al., 1994;Wenk et al., 1994). This indicates that non-cholinergic cell types within the basal forebrain may contribute more to some processes, such as learning and memory, than do the cholinergic neurons. While learning and memory appear to be less sensitive to cholinergic depletion, attentional processes are impaired following cholinergic-specific lesions in rats (reviewed by McGaughy et al., 2000). In one study by Waite et al. (1999), the multiple choice reaction time task was used to challenge sustained attention (also known as vigilance). Briefly, rats receive a food reward by selecting one out of a number of available ports, the correct port having been indicated by brief illumination of a light. Some manipulations to the task can increase the ''attentional demand'' such as changing the time between trials (inter-trial interval variability), and increasing the number of response ports (e.g., from five active ports to nine), making the task more difficult. During times of increased demand, task performance is impaired following cholinergic depletion in rats. This indicates that ACh may have a role in vigilance, especially when the task load is greater. Another study by Bucci et al. (1998) reports depletion of cholinergic innervation to posterior parietal cortex in rats (an area thought to be involved in attentive regulation) impairs the ability to increase processing capacity to meet task demands. Here, processing of a visual stimulus (a light) is assessed following manipulation of the predictive relationship between the light and an auditory stimulus (a tone) in relation to a food reward. Rats with depleted cholinergic innervation are impaired at appropriately shifting attention based on the predictive relationships between stimuli compared to control rats. In fact, many studies in rats have demonstrated impairments in vigilance tasks following cholinergic-selective depletion (Turchi and Sarter, 1997;McGaughy and Sarter, 1998;McGaughy et al., 2002). It is important to note here that attention is not a unitary phenomenon and these lesion studies explore a different process (vigilance) than do traditional studies of attention in humans and non-human primates (selective attention). We will discuss evidence for ACh's role in selective attention below. In mammals other than rats and mice, ME20.4 IgG-saporin is used to selectively eliminate cholinergic cells. In the marmoset, both excitotoxic and cholinergic-specific lesions to the nucleus basalis result in learning impairments (Ridley et al., 1986;Fine et al., 1997). In these studies, subjects were trained to perform a simple visual discrimination task in which the location of a food reward is signified by one of two objects; the object marking the reward location must be learned over a series of trials. Marmosets with both lesion types are impaired at this task relative to controls. Given that the primate NB/SI is 90% cholinergic, as discussed above, it is not surprising that the two lesion methodologies do not differ in this species as profoundly as they do in rats. In macaques, a study by Browning et al. (2009) assesses learning and memory following specific cholinergic depletion of the inferotemporal cortex. Here, macaques were trained to identify one of two objects in a complex scene. Identification of the correct object is rewarded with a food pellet (object-in-place scene learning). To assess memory, macaques are trained in a delayed nonmatch-to-sample task. In this task, a sample object is presented briefly then two objects appear, one being the sample object and one being a novel object. Subjects are rewarded with a food pellet for identifying the novel object. Cholinergic depletion does not result in impairments in either task (but see Turchi et al., 2005). Croxson et al. (2011) report no impairment in decision-making or in the object-in-place scene learning task described above, however, spatial working memory is impaired following cholinergic depletion of the prefrontal cortex. The latter was assessed using a spatial delayed response task, which requires subjects to remember the location of a food reward after a delay period. Thus, unlike in rodents, cholinergic depletion in macaques impairs spatial working memory. Macaques who have undergone non-selective excitotoxic lesions show impaired memory in the delayed nonmatch-to-sample task (described above as unimpaired by selective cholinotoxic lesions of cortex), however, the impairment follows combined lesions to the nucleus basalis, medial septal nuclei, and diagonal band of Broca (Aigner et al., 1991). Lesions to the nucleus basalis alone (or any of the cholinergic nuclei in isolation) do not result in any impairment. As demonstrated, cholinergic depletion produces complex results. The resulting impairment (if any) will depend on a number of variables. These include the composition of the basal forebrain in the species under study (e.g., the difference between the proportions of cholinergic neurons in the primate NB/SI vs. the rat), the immunotoxin (e.g., excitotoxic vs. cholinotoxic), the location of the lesion (e.g., the combination of subcortical nuclei and/or the cortical area affected), and the task used to assess the many forms of cognition thought to rely, at least in part, on ACh. Given the considerable variation in study design and resulting impairments following cholinergic depletion, unifying or extending the results of lesion studies to yield descriptions or predictions that hold across species, brain areas, and tasks is challenging, at best. STRUCTURE: CHOLINERGIC INTERNEURONS IN CORTEX A notable distinction between cholinergic systems across species is the presence or absence of cholinergic neurons within cortex. Many studies have reported an absence of cell bodies immunoreactive for ChAT in primate cortex (Hedreen et al., 1983;Mesulam et al., 1983;Raghanti et al., 2008). It is believed that ACh is instead delivered to primate cortex exclusively through projections from the basal forebrain. This, however, is not true for other species studied (cat and rat). Moreover, in animals where cholinergic cortical cell bodies have been reported, there are laminar-and area-specific densities and distributions. For example, in the rat, cholinergic neurons are found in all cortical layers except layer I, and are most dense in layers II and III (Houser et al., 1983;Ichikawa and Hirata, 1986;Mechawar et al., 2000). In the cat, cholinergic cell bodies are similarly most dense in layers II and III, but are also more numerous in motor and somatosensory cortices than in auditory and visual cortices (Avendaño et al., 1996). These neurons are non-pyramidal, and probably co-release GABA and perhaps peptide signaling molecules as well (Eckenstein and Thoenen, 1983;Parnavelas et al., 1986;Bayraktar et al., 1997). STRUCTURE: ACh SYNTHESIS The local regulation of ACh synthesis can influence the capacity for cholinergic signaling. Choline necessary for the synthesis of ACh is taken up by cholinergic neurons via high-affinity choline transporters (ChTs). As such, ChTs are the rate-limiting factor in ACh synthesis (Okuda and Haga, 2003). Regions of tissue with relatively high levels of ChTs likely have the ability to continue to synthesize ACh under high levels of ''demand'' (i.e., axonal activity) and will be able to release ACh in a more sustained fashion than areas with low levels of ChT expression. ChT expression, especially in combination with patterns of cholinergic innervation (described above), affords more finely tuned local regulation of ACh release. Such local differences could lead to marked variation in extracellular ACh levels, both tonically and in response to increased demand. ChT distribution exhibits species and area specificity. While ChT immunoreactivity in fibers and puncta is consistently low across cortical areas in rats (Misawa et al., 2001), levels differ between cortical areas in macaques. For instance, the cingulate cortex exhibits higher levels of ChT immunoreactivity than do the frontal, parietal, temporal and occipital cortices (Kus et al., 2003). This suggests a higher capacity for sustained cholinergic signaling in cingulate cortex than in other cortical areas. Within each cortical area, in both species, there are also laminar differences in ChT expression, with deeper layers showing stronger immunoreactivity compared to superficial layers (Misawa et al., 2001;Kus et al., 2003). This may indicate a need for more sustained signaling in the layers of cortex that interact with subcortical structures than in the layers responsible for the majority of the feedforward and lateral interactions within cortex. STRUCTURE: CHOLINERGIC RECEPTORS The cholinergic system transduces signals through two families of receptors: nicotinic and muscarinic, named for the exogenous ligands to which they are particularly responsive-nicotine and muscarine, respectively (reviewed by Dorostkar and Boehm, 2008). Nicotinic receptors are cation (ionotropic) channels, which mediate a fast depolarization of the receiving cellular membrane. As pentamers (i.e., containing five subunits), they can be characterized by their subunit composition. Heteropentameric nicotinic receptors contain two obligatory α subunits in addition to other subunits, usually β2 in cortex (Gotti et al., 2006;Dorostkar and Boehm, 2008;Albuquerque et al., 2009). The only homopentameric nicotinic receptors described so far in mammals comprise five α7 subunits. Muscarinic receptors are metabotropic. They can act via direct channel coupling or through coupling to various intracellular second messenger systems. Five muscarinic receptor subtypes have been identified (m1-m5) and are characterized according to the class of G protein to which they are coupled (reviewed by Gilsbach and Hein, 2008). The m2 and m4 receptors are preferentially coupled to the Gi/o signaling pathway and correspond to the M2 pharmacological class (i.e., they are insensitive to the antagonist pirenzepine; MacIntosh, 1984). These receptors are typically expressed presynaptically and act as autoreceptors. The m1, m3 and m5 receptors are preferentially coupled to the Gq signaling pathway and correspond to the M1 pharmacological class, although sensitivity to pirenzepine varies within this class. These receptors are more often expressed postsynaptically. While the molecular signaling characteristics of these receptors are almost certainly the same in all species, their expression by neuron type differs. STRUCTURE: CHOLINERGIC RECEPTOR EXPRESSION BY INHIBITORY NEURONS Inhibitory interneurons in cortex (i.e., local circuit neurons that express GABA) exhibit considerable structural and functional diversity. One means of characterizing subpopulations of cortical inhibitory neurons is to classify them by their expression of molecular markers. The most commonly used classification schemes differ between species, as do the interneuron populations themselves. In macaques, the use of immunoreactivity for calcium-binding proteins is the most common scheme, but this may not apply well to other species. For example in rats, the calcium-binding protein parvalbumin (PV) is a marker for the fast-spiking physiological phenotype (Weiser et al., 1995;Sekirnjak et al., 1997), and is expressed by half of V1 inhibitory interneurons (Gonchar and Burkhalter, 1997). However, not all PV-immunoreactive (ir) neurons in the primate are fast-spiking (Härtig et al., 1999;Constantinople et al., 2009). A comparative anatomy study by Disney and Reynolds (2014) reports species differences in cholinergic receptor expression by PV-ir neurons in V1. In humans, macaques, and guinea pigs, the majority of PV-ir neurons (76%-84%) expresses the m1 muscarinic receptor. In ferrets and rats, however, fewer than half of the PV-ir neurons express the m1 receptor. In rats, the low expression is observed in the primary somatosensory cortex (S1) as well as in V1. Further, a study in mice found m1 muscarinic receptor expression by inhibitory neurons, including PV-ir neurons, to be low or undetectable in cortex (Yamasaki et al., 2010). Thus, cholinergic receptor expression by PV-ir neurons differs across model systems, and the neuronal ''class'' defined by PV immunoreactivity may also differ. Other calcium-binding proteins, such as calbindin (CB) and calretinin (CR), are also used to classify inhibitory interneurons, particularly in primates. Disney and Aoki (2008) report 60% of CB-ir and 40% of CR-ir neurons express the m1 receptor in macaque V1. A smaller proportion of these populations expresses the m2 receptor at 23% of CB-ir neurons, 25% of CR-ir neurons, and 31% of PV-ir neurons. In rat V1, one in vitro study reports that ACh depolarizes low-threshold spiking cells via nicotinic receptors (Xiang et al., 1998). While this study does not report any molecular immunoreactivity for these neurons, previous work indicates low-threshold spiking neurons in rodent cortex correspond to neurons immunoreactive for CB (Kawaguchi and Kubota, 1993). In macaques, CB-ir neurons also express nicotinic receptors, as do PV-ir neurons, while expression by CR-ir neurons is rare (Disney et al., 2007). Other molecular markers are often used to describe inhibitory populations in rodents. In a study by Kawaguchi (1997), in vitro recordings in rat frontal cortex show the extent to which GABAergic neurons immunoreactive for PV, somatostatin, vasoactive intestinal polypeptide (VIP) and cholecystokinin respond to cholinergic agonism. In the presence of carbachol (a cholinergic agonist) or muscarine (a muscarinic-specific agonist), VIP-and somatostatin-ir neurons are depolarized and cholecystokinin-ir neurons are either hyperpolarized or depolarized depending on their size. This finding indicates these populations likely express cholinergic receptors. Unlike in rat visual cortex where PV-ir neurons express the m1 receptor (albeit, a small population; discussed above), the PV-ir neurons in this study were unaffected by both carbachol and muscarine. FUNCTION: CHOLINERGIC MODULATION OF INHIBITION In guinea pig, cat and macaque V1, muscarinic agonists can have a suppressive effect on the firing rate of neurons in cortex. At least some of this suppressive effect can be blocked by GABA antagonism (McCormick and Prince, 1986;Müller and Singer, 1989;Disney et al., 2012), indicating that ACh can induce the release of GABA. It is probable that this GABA release is mediated by the PV-ir, m1 receptor-expressing population in these species (discussed above). However, PV-ir neurons in rat visual cortex do not appear to mediate similar effects, as would be expected based on their lack of ACh receptor expression. In vitro studies in rat V1 show that ACh does not depolarize PV-ir neurons (Xiang et al., 1998;Gulledge et al., 2007). In contrast to rat, PV-ir neurons in mouse V1 are both excited and suppressed in response to optogenetic stimulation of basal forebrain cholinergic neurons (Alitto and Dan, 2013)-actions that were attributed to mixed activation of nicotinic and muscarinic receptor subtypes. As such, the capacity for ACh to induce the release of GABA from cortical interneurons may differ between species. This is particularly interesting in the context of the differing extent to which the NB/SI itself releases GABA in these species. Of note here is the apparent conflict between this study and the Yamasaki et al. (2010) finding described above (that m1 receptor expression by PV-ir neurons is low or undetectable). However, the latter study investigated cortex as whole, while the finding in Alitto and Dan (2013) is specific to V1. STRUCTURE: CHOLINERGIC RECEPTOR EXPRESSION BY EXCITATORY NEURONS In macaques, receptor expression by excitatory neurons differs between cortical areas. In V1, the m1 receptor is expressed by over 60% of inhibitory neurons and by less than 10% of excitatory neurons (Disney et al., 2006). In visual area V2, m1 receptor expression by inhibitory neurons is similar to that in V1. However, muscarinic receptor expression by excitatory neurons differs sharply between V1 and V2, with at least double the proportion of excitatory neurons expressing muscarinic receptors in V2 compared to V1 (Disney et al., 2006). In fact, across the visual pathway-from V1, through V2, V3a, V4d, to MT-m1 receptor expression by inhibitory neurons remains roughly constant at 50%-60%, while m1 receptor expression by excitatory neurons differs with an increasing m1 receptor immunoreactive excitatory population moving ''up'' through the visual pathway (Disney et al., 2006unpublished data). Thus, even within a species, cholinergic receptor expression by excitatory neurons differs between cortical areas. The m1 receptor-expressing excitatory population also differs between species. As discussed above, only 10% of excitatory neurons in macaque V1 express the m1 receptor. In contrast, Gulledge et al. (2007) report between 25% and 95% (dependent upon layer) of excitatory neurons across rat cortex (medial prefrontal, somatosensory, and visual cortex) respond to ACh in vitro. Further, qualitative data indicate that excitatory neurons in V1 of ferrets and guinea pigs more frequently express the m1 receptor than do excitatory neurons in V1 of macaques or humans . Regardless of neuron type (excitatory/inhibitory), cholinergic receptor expression in general is known to differ considerably across species. As demonstrated by in vivo recordings, 55% of cells in marmoset V1 respond to ACh (Roberts et al., 2005). This number in cats is between 85% and 90% (Müller and Singer, 1989;Stewart et al., 1999). With such variation not only in the proportion of ACh-responsive cells in an area, but also in the specific proportion of excitatory and inhibitory cells that express cholinergic receptors, effects of ACh observed in various cortical model systems may differ substantially. STRUCTURE: ACh RELEASE The instantaneous extracellular concentration of a given neuromodulator represents a critical feature of cortical circuits. This is because different ambient levels will yield activation of specific receptor subtypes based on receptor-specific affinity for the ligand, and-particularly for ACh-the rate of receptor desensitization. For example, muscarinic receptors m2 and m4 exhibit higher affinity for ACh than do the m1, m3, and m5 muscarinic receptors (Kuczewski et al., 2005). It can be hypothesized, then, that at lower levels of extracellular ACh, the m2 and m4 receptors will be more readily activated than will the m1, m3, or m5 receptors. Because the m2 and m4 receptors usually act as autoreceptors, their activation is likely to result in a different type of circuit regulation (reduced ACh release) than would the activation of the m1, m3 and/or m5 receptors. It is interesting that down-regulation of ligand release will likely be one of the earliest recruited cholinergic mechanisms in any circuit. Similarly, the rate at which receptors desensitize will be related to extracellular ACh. For example, α7-containing nicotinic receptors become desensitized to ACh quickly, resulting in a decreased capacity for continuous cholinergic modulation. Non-α7-containing nicotinic receptors, however, desensitize more slowly, which likely results in prolonged cholinergic modulation. As such, modulatory signaling can differ in type and can range from more rapid to more prolonged, depending on the concentration-dependent recruitment of specific receptor subtypes. Altogether, the extracellular level of ACh will affect dynamics of receptor binding as well as the subtypes to which ACh will bind, and profound differences in these levels have been found across species. In a review by Fitzgerald (2009), studies describing the concentration of ACh (as well as other neuromodulators) in awake macaque and rat prefrontal cortex were compared. These concentrations appear to be much higher in the macaque than in the rat. There are also differences in ACh levels within animals. For example, in vivo microdialysis in anesthetized rats shows baseline ACh levels are highest in the medial prefrontal cortex relative to visual cortex and somatosensory cortex (Fournier et al., 2004; but see Sarter and Bruno, 1997). Considering that a given ACh level will recruit specific receptor subtypes, which are themselves expressed differently by region and species, such variation in extracellular ACh levels may result in strikingly different cholinergic modulation of cortical circuits across and within species. FUNCTION: EFFECTS OF ACh RELEASE IN CORTICAL CIRCUITS It has been proposed that ACh facilitates attentive processes in sensory cortices. A suggested mechanism for this facilitation is the simultaneous enhancement of a sensory input and suppression of intrinsic cortical activation (Hasselmo and Giocomo, 2006). Many studies have described a population of nicotinic receptors in the input layer of a number of thalamic recipient cortical areas across species including rat, cat and primate (Clarke et al., 1984;London et al., 1985;Prusky and Cynader, 1986;Lavine et al., 1997;Disney et al., 2007;Eickhoff et al., 2007). These receptors were found to be located on terminals from thalamic nuclei innervating sensory (Prusky et al., 1987;Lavine et al., 1997;Disney et al., 2007), motor (Lavine et al., 1997), and association (Lavine et al., 1997;Lambe et al., 2003) cortical areas. Nicotinic receptors expressed presynaptically by thalamic terminals are thought to increase the gain of incoming sensory data (Disney et al., 2007;Kawai et al., 2007). The thalamic enhancement/cortical suppression model of attention (Hasselmo and Giocomo, 2006) is based primarily on pharmacological studies in rodent sensory cortex in vitro (e.g., Hasselmo and Bower, 1992;Gil et al., 1997;Kimura et al., 1999;Hsieh et al., 2000). An interesting study from the point of view of species comparisons is that by Gil et al. (1997). In the somatosensory cortex in vitro, they report differences in the receptor population underlying cholinergic regulation of thalamocortical synapses between rats and mice. In both species, thalamocortical synaptic transmission is enhanced by nicotinic receptor agonists, and both thalamocortical and intracortical synapses are suppressed by muscarinic receptor activation. Interestingly, in this study, there was a difference between rats and mice in the nicotinic receptor subtype mediating the thalamic enhancement. Specifically, an α7 nicotinic receptor antagonist blocked the thalamocortical enhancement in rats, but not in mice, suggesting a species difference in receptor subtype expression at the thalamocortical synapse. This is of potential functional importance because the homomeric α7 receptor passes calcium with very high affinity (Vijayaraghavan et al., 1992;Séguéla et al., 1993), allowing nicotinic signaling to be directly coupled to intracellular plasticity mechanisms. Rats are the only species reported thus far to have a predominant α7-mediated nicotinic response reported at the thalamocortical synapse. The suppression of intracortical pathways in these rodent studies has been attributed to a reduction in glutamate release through activation of m2 muscarinic receptors expressed on the axons of excitatory neurons (Hasselmo and McGaughy, 2004). However, another suppressive mechanism has been proposed for primates. In macaque V1, cholinergic suppression of visual responses was shown to be mediated by a strengthening of inhibition (Disney et al., 2012). Here, suppression by ACh is mediated by an increase in GABA release (as opposed to the reduction in glutamate release observed in rats). There is evidence that ACh also increases the strength of inhibition in the cat (Müller and Singer, 1989) and in the guinea pig (McCormick and Prince, 1986). Interestingly, Soma et al. (2012) report a predominant enhancement by ACh (in contrast to the dominant suppression observed by Disney et al., 2012) mediated by muscarinic receptor activation, although both studies report both suppression and enhancement. The difference between these studies is not clear, but may be attributable to the concentration dependence of ACh effects or sampling bias in recording. Concentrationdependent effects will result from differing affinities of muscarinic receptor subtypes (discussed above; Kuczewski et al., 2005;Disney et al., 2012). Both of these studies were conducted under anesthesia and so basal levels of ACh in cortex may differ if the maintained depth of anesthesia differed. Further, differences in ejection barrel geometry and applied iontophoretic currents will yield different levels of delivered drug above that basal level. These factors can combine to activate different populations of receptors for subtly different experimental conditions. Neither study reported the proportion of interneurons in their recorded population, but given the anatomy of macaque V1, a recording bias towards excitatory neurons will also yield a higher proportion of suppression, and a recording bias towards interneurons would yield more apparent enhancement. It is important to note that across the population of neurons in V1, in all species studied, both enhancement and suppression are observed with cholinergic activation in vivo. What may differ between species and with methodology is the proportion of excitatory vs. suppressive effects, and perhaps the mechanism underlying the suppression, when observed (Sillito and Kemp, 1983;Sato et al., 1987;Müller and Singer, 1989;Murphy and Sillito, 1991;Zinke et al., 2006;Disney et al., 2007Disney et al., , 2012Herrero et al., 2008;Soma et al., 2012). In macaques and humans, there have been functional demonstrations of ACh enhancing sensory input relative to intrinsic cortical activity. One study in macaque V1 shows that administration of nicotine (a ligand for nicotinic receptors) improves contrast sensitivity (Disney et al., 2007) for V1 neurons. Here, anesthetized macaques were presented drift-grating stimuli of multiple contrasts with or without iontophoretic application of nicotine in V1. Physiological recordings reveal that in the presence of nicotine, cells in the input layer 4c produce reliable responses to lower-contrast stimuli, indicating that low-contrast detection is improved with nicotine. Similarly, in humans, cholinergic enhancement has been shown to increase signal detection. In a study by Boucart et al. (2015), participants engaged in a two alternative forced choice task in which they are shown two pictures of natural scenes, one of which contains an animal (the target). Participants must indicate which picture contains the target under varying levels of contrast. Before the task, participants were given either a placebo or the drug donepezil. Donepezil limits ACh degradation by inhibiting the ACh metabolizing enzyme acetylcholinesterase, and thereby enhances cholinergic transmission. In the presence of donepezil, signal detection of the target is facilitated. Of course, with systemic drug delivery such as this, there is no way to determine where in the brain the donepezil is acting to produce this behavioral effect. Further data would be needed to assign donepezil's actions to the increased gain at the input to cortex. Beyond the input layer, a decrease in receptive field size and a reduced spread of excitation have been proposed as measurable consequence of suppressing lateral cortical interaction. This is because the size of a receptive field center is thought to be largely determined by inputs arising from the thalamus, while the receptive field surround is provided by lateral and feedback connectivity within cortex (Angelucci and Bressloff, 2006). Silver et al. (2008) report cholinergic enhancement suppresses the spread of excitation in human V1. In this study, subjects passively viewed high-contrast/contrast-reversing checkerboards interspersed with a blank gray screen, while maintaining fixation on a central point. Positive blood-oxygenlevel dependent (BOLD) responses to the checkerboard stimulus relative to baseline (the gray screen) were observed by functional magnetic resonance imaging (fMRI). Prior to fMRI sessions, subjects ingested either a placebo or donepezil (described above). Donepezil administration resulted in a positive BOLD response to stimuli that occupied less cortical surface area compared to placebo. This indicates cholinergic enhancement reduces the spatial spread of excitation. Similar effects have been observed in non-human primates. In a study by Roberts et al. (2005), length tuning of V1 neurons to bar stimuli was studied with and without iontophoretic application of ACh. In the presence of ACh, the neurons' preferred stimulus length shifted toward shorter bars. This phenomenon was modeled as a reduction in the summation area of the neurons' receptive fields, again consistent with a suppressive effect of cortical ACh release. ACh IN ATTENTION The model proposed by Hasselmo and Giocomo (2006) may-at least partially-explain ACh's role in attentive processing. Unfortunately, the current literature does not adequately distinguish between ACh's direct involvement in attention vs. cholinergic modulation of vigilance (which is demonstrated in the lesion studies described above). It may be the case that vigilance-and therefore ACh-is a necessary, but not sufficient component of attention. In this case, the cholinergic modulation of vigilance would act as a condition upon which more selective or focal attention can be achieved. As described earlier, the suggested mechanism for ACh's facilitation of vigilance is the simultaneous enhancement of a sensory input and suppression of intrinsic cortical activation. To date, only one study has used a true selective attention task to investigate the local effects of ACh in the cortex of an awake, behaving primate (Herrero et al., 2008). Here, recordings were made in macaque V1 during performance of a cued contrast change detection task. In the task, macaques must maintain fixation at a central point and detect a luminance change at a cued location (target) while ignoring a luminance change in a non-cued location (distractor). The results show the expected increase in the firing rate of V1 neurons when a target is detected in their receptive fields, that is, when the target is ''attended to.'' Further, they demonstrate that V1 neurons show a greater attentional modulation during application of ACh. These effects were also shown to be the result of muscarinic receptor activation, as the muscarinic antagonist scopolamine reduced the attentional enhancement, while nicotinic antagonists had no effect on attentional modulation (although nicotinic antagonists had a generalized effect on spike rates). This study provides clear evidence that ACh is involved in some form of attentive processing that goes beyond vigilance in showing differences in the attend-to vs. attend-away conditions. The use of inter-hemispheric comparison for attend-to/attend-away, however, leaves open the question of whether ACh is sufficient for selective attentive effects, or whether another molecule supports cholinergic action at a finer scale in cortex. We are not aware of studies of cholinergic action on selective attention in species other than primates (human and nonhuman). Other notable primate (both human and non-human) studies support a role of ACh in attention (e.g., Witte et al., 1997;Furey et al., 2008;Thiel and Fink, 2008); however, these studies cannot address the questions within the scope of this review, as they involve systematic administration of cholinergic drugs. In this case, drugs affect the entire body, leaving little ability to resolve different cholinergic effects across cortical areas, especially at the circuit level. Further, few species comparisons exist, as most of these studies are performed in human subjects. CONCLUSIONS A thorough understanding of the differences that exist in cholinergic signaling across mammalian model systems commonly used in neuroscience is of particular importance for the cholinergic system. This system is traditionally considered broad in its impact throughout cortex because of the relatively small number of cholinergic neurons in the NB/SI, their large branching axonal arbors, and the evidence for volume transmission by ACh. Volume transmission is a diffuse signaling mode wherein a molecule is released from varicosities that are usually not apposed to a specialized receptive surface (demonstrated in both rats and primates; Umbriaco et al., 1994;Mrzljak et al., 1995; but see Turrini et al., 2001). Even synaptically-released ACh can spill over from the synapse and thus participate in volume transmission (Dani and Bertrand, 2007). When it is not confined to a synapse, ACh can diffuse through the extracellular space, activating nearby receptors. This capacity for diffuse signaling often leads to an implicit assumption that the cholinergic system lacks precision. However, variations in the anatomy across the arbor of individual cholinergic axons and differences in the receiving circuits in cortex may introduce regionallyspecific responses to diffuse modulatory signals. Furthermore, differences in expression of acetylcholinesterase and the ChT, which serve to terminate cholinergic signaling, can also induce differences in the temporal precision of a cholinergic signal, whether or not that signal is spatially precise. The product of regional differences in anatomical features that determine the spatiotemporal properties of the cholinergic signal is the existence of unique neuromodulatory compartments across cortex (Coppola et al., 2016). Features such as patterns of axonal innervation to cortex, molecular diffusion, effectiveness of degradation and reuptake pathways, subcellular receptor localization, and patterns of receptor expression across local receiving circuits (among many others) can offer the capacity for local modulation of long-range communication between neurons. A second common assumption is that the various model systems used to study ACh are equivalent and interchangeable, resulting in the possible existence of a canonical circuit for the cholinergic system in cortex. Here, we have discussed important differences in the structure and function of the cholinergic system both across and within species. Even for cases of apparent functional equivalence (such as the common observation of a relative enhancement of sensory input to cortex over intrinsic pathways in the presence of ACh), the cholinergic system across model systems fails to exhibit anatomical equivalence. The anatomical differences across species and areas make for unique cholinergic circuit compositions that may nonetheless yield similar computations across systems (i.e., computational equivalence in the absence of anatomical equivalence). The differences between species should not be viewed as surprising; there is growing evidence from studies in invertebrates that modulatory systems are common sites of evolutionary elaboration (Grashow et al., 2009;Katz and Lillvis, 2014). That this phenomenon extends to vertebrates, specifically mammals, is both likely and-as we have shown here-already supported by the literature. At a conceptual level, anatomical differences may not matter given the evidence for computational or functional equivalence. However, if we are to use the data from studies of cholinergic structure and function to drive improvements in human health and treatment of disease, the precise mechanisms by which these computations are achieved matters, and thus is a critical consideration for future studies. AUTHOR CONTRIBUTIONS AAD conceived of the review topic and scope and assisted JJC in writing the review. FUNDING This work was supported by National Institute of Mental Health grant no. MH093567.	2018-01-31T18:45:55.191Z	2018-01-30T00:00:00.000	{ "year": 2018, "sha1": "99de233b5f7ef8e0c54f35920eeae3f941d78666", "oa_license": "CCBY", "oa_url": "https://www.frontiersin.org/articles/10.3389/fncir.2018.00008/pdf", "oa_status": "GOLD", "pdf_src": "PubMedCentral", "pdf_hash": "99de233b5f7ef8e0c54f35920eeae3f941d78666", "s2fieldsofstudy": [ "Biology" ], "extfieldsofstudy": [ "Biology", "Medicine" ] }
55995908	pes2o/s2orc	v3-fos-license	A magnetic reconnection model for quasi-periodic oscillations in black hole systems The quasi-periodic oscillations (QPOs) in black hole (BH) systems of different scales are interpreted based on the magnetic reconnection of the large-scale magnetic fields generated by the toroidal electric currents flowing in the inner region of accretion disk, where the current density is assumed to be proportional to the mass density of the accreting plasma. The magnetic connection (MC) is taken into account in resolving the dynamic equations of the accretion disk, in which the MC between the inner and outer disk regions, the MC between the plunging region and the disk, and the MC between the BH horizon and the disk are involved. It turns out that the single QPO frequency of several BH systems of different scales can be fitted by invoking the magnetic reconnection due to the MC between the inner and outer regions of the disk, where the BH binaries XTE J1859+226, XTE J1650-500 and GRS 1915+105 and the massive BHs in NGC 5408 X-1 and RE J1034+396 are included. In addition, the X-ray spectra corresponding to the QPOs are fitted for these sources based on the typical disk-corona model. The first convincing QPO of active galactic nuclei (AGNs) was discovered by Gierlinski et al. (2008) in narrow line Seyfert 1 RE J1034+396, which opened a window for the comparative timing studies of stellar-mass and supermassive black holes (BHs). The 1-hour X-ray QPO observed in RE J1034+396 is analogous to the 67 Hz QPO in BHB GRS 1915+105 (Middelton & Done 2010. In earlier years, quasiperiodic signals like QPOs were discovered in some supermassive BHs. For example, the power density spectra of two X-ray flares of Sgr A* observed in 2000 and 2002 show five distinct peaks at periods of 100, 219, 700, 1150219, 700, , 2250219, 700, seconds (Aschenbach et al. 2004) and a quasi-periodic flux modulation with a period of 22.2 minutes was discovered in the X-ray data of the Sgr A* flare in 2004(Blanger et al. 2006. A number of models have been proposed to interpret HFQPOs in BHBs. However, none of them can fully explain the characteristics of QPOs, especially the correlations of spectral and timing properties Maitra & Miller 2010). As is well known, HFQPOs in BHBs are strongly correlated to steep power law (SPL) state. A successful model of QPOs should link the corresponding spectral state, allowing for a highly dynamical interplay between thermal and nonthermal processes with the mechanisms operating over a wide range of luminosity . QPOs observed in massive BHs are probably related to the bright state which is similar to the SPL state of BHBs (Strohmayer & Mushotsky 2003Middelton et al. 2009), and they may have the same origin as the HFQPOs in BHBs (Gierlinski et al. 2008;Bian & Huang 2010). Inverse Compton scattering is generally thought to be the promising radiation mechanism of SPL state (Zdziarski 2000;McClintock & Remillard 2006, hereafter MR06), and disk-corona model is successful in fitting the spectrum of SPL state (Gan et al. 2009, hereafter G09;Huang et al. 2010). Zhao et al. (2009) (hereafter Z09) interpreted HFQPOs in BHBs as the magnetic reconnection of large-scale magnetic fields generated by toroidal electric currents flowing in the disk without considering the spectral state. In this paper, we improve the model of Z09 based on the disk-corona model with the magnetic connection (MC) effects. We fit both the QPO frequencies and the corresponding X-ray spectra of four BH systems of different scales, in which two BHBs (XTE J1859+226, XTE J1650-500), one ULX (NGC 5408 X-1) and one supermassive BH (RE J1034+396) are included. This paper is organized as follows. A description of the model is given in Section 2. The QPO frequencies and X-ray spectra of the four sources are fitted based on the disk-corona model with MC effects in Section 3. And some issues related to our model are discussed in Section 4. Throughout this paper, the geometric units G = c = 1 are used. Origin of magnetic fields in BH systems Large-scale magnetic fields play very important roles in high energy astrophysical phenomena. The relativistic jets from AGNs and BHBs are launched and collimated by invoking the open large-scale magnetic fields related to the BZ and BP processes (Blandford & Znajek 1977;Blandford & Payne 1982), and the broad iron lines observed in BH systems could be interpreted by transferring energy and angular momentum from a spinning BH to its surrounding accretion disk by invoking the magnetic connection via the closed large-scale magnetic fields connecting the horizon with the inner disk (Wilms et al. 2001;Miller et al. 2002;Li 2002a;Wang et al. 2002, hereafter W02). However, a consensus on the origin of large-scale magnetic fields in BH systems has not been reached. Unlike neutron stars, magnetic fields cannot exist on the horizon of an isolated BH, and they should be maintained by the surrounding environment, such as an accretion disk. For BHBs, the magnetic fields probably come from the plasma of the companion. One of the most promising origins of large-scale magnetic fields in BH systems is accretion disk around the BHs, and some authors calculated the magnetic field configurations by assuming the toroidal electric current in the disk (Li 2002b;Z09). However, the origin of the currents remains to be clarified. In this paper, we intend to interpret the origin of the electric current based on the assumption that the accreting plasma deviates somehow from electric neutrality as it flows through the Lagrange point into the Roche lobe of the BHBs, resulting in toroidal electric currents flowing in the accretion disk. In addition, we resolve the dynamical equations of the accretion disk by taking the MC effects into account and figure out the mass density and current density in the disk, and then determine the configuration of the large-scale magnetic fields by considering the interaction between the electric current and the disk with the iterative algorithm. It turns out that we can fit the association of QPOs with spectral states in BH systems of different scales based on our model. The deviation from electric neutrality is described by defining a parameter η as follows, where n e and n p are the number densities of electrons and protons, respectively. The charge density can be expressed as ρ e = ηen p , where e = 4.8 × 10 −10 e.s.u is the electron charge. The large-scale magnetic fields generated by the toroidal electric current may in turn affect the current. For example, the field lines may pipe hot electrons into the corona above the disk therefore change the charge density in the disk. This effect should be strongest in the inner disk because the magnetic field intensity is strongest near the inner edge of the disk and decreases rapidly outwards. For simplicity, we assume this effect decreases with the increasing disk radius as a power law and the surface density of the current in the disk can be expressed as where h, ρ m , and m p are the half height of the disk, the mass density and the proton mass respectively. The power law index n is a free parameter for fitting the QPO frequency, and µ = 0.615 is the weight-average molecular weight of the gas. The toroidal electric current is assumed to distribute from the inner edge of the disk to the outer boundary of the disk-corona system, the radius of which is set at r out = 100M in calculations, since the radiation of the accretion disk mainly comes from the inner region. Following the work of Znajek (1978), Linet (1979), Li (2002b) and Z09, we can calculate the toroidal component of the electric vector potential determined by the current given by Eq. (3) with a given mass density of the disk matter. And the boundaries of inner-outer regions related to the magnetic field configuration can be determined. As shown by Fig.1, there are three types of magnetic field configuration generated-MC of the BH with the disk (MCHD), MC of the plunging region with the disk (MCPD), and MC of the inner and outer disk regions (MCDD). In the MCDD region, the magnetic field lines are frozen in the disk at the inner and outer footpoints. The field lines will twist themselves since the inner and outer footpoints have different angular velocities. In Fig.1, the inner and outer footpoints of the field line in red locate at r in (the inner edge of the disk) and r 0 , respectively. The value of the angular velocity difference between the footpoints of this field line is maximal among all the lines in MCDD region, so this line will twist itself first and trigger the magnetic reconnection, as shown in Fig.2, which releases magnetic energy periodically to generate flares with the frequency of the difference between Keplerian frequencies of the inner and outer footpoints. We interpret this frequency as the QPO frequency, and it reads: where ν 0 ≡ m −1 BH 3.23 × 10 −4 Hz and m BH ≡ M/M ⊙ is the black hole mass in unit of the solar mass. a * ≡ a/M is the dimensionless BH spin. r in is initialized at the innermost stable circular orbit (ISCO). As shown by Eq. (3), parameters n and η determine the current density and therefore magnetic configuration and QPO frequency once the surface density of disk matter is given. The value of η determines the current intensity and therefore magnetic field strength, while that of n determines the concentricity of the currents and magnetic fields. A larger n leads to a smaller MCDD region, a smaller distance btween r in and r 0 , and hence a lower QPO frequency, while a smaller n leads to the opposite results. As an example, a contour of 190 Hz QPO of XTE J1859+226 is plotted in η − n parameter space as shown in Fig. 3. It is found that a maximum η ∼ 10 −12 is required to avoid a too strong magnetic field for a stable solution with the MC effect. The dynamical equations of disk and corona with MC effects The dynamical equations of the accretion flow are modified by considering the MC effects on the transfer of energy and angular momentum (Li 2002a where Q and g are respectively the viscously dissipated energy per unit disk surface and interior viscous torque of the disk, and Q consists of two parts, where Q d is radiated from disk as black body, and Q cor is transported into the corona to heat it by magnetic reconnection of tangled small-scale magnetic fields, and it reads (Liu et al. 2002) where B D , V A and ρ are the intensity of the small-scale magnetic fields, the Alfven speed, and mass density, respectively. The quantity H MC in Eqs. (5) and (6) is the flux of angular momentum and it reads where dT MC and dΨ are the torque exerted on to an infinitesimal annulus of the disk and magnetic flux threading the infinitesimal annulus, respectively. We can calculate a variety of magnetic transfer of energy in BH accretion disk by using an equivalent circuit in an analogous way to Macdonald & Thorne (1982) and W02 as shown in Fig.4. Three types of MC (MCDD, MCPD and MCHD) indicated in Fig.1 are given in Fig.4, in which a series of loops correspond to the adjacent magnetic surfaces arising from the rotation of the closed field lines. The quantities dZ D , dZ P , dZ H and dZ C are the resistances corresponding to the two adjacent magnetic surfaces in the disk, the plunging region, the BH horizon and the corona, respectively. The quantities dε D and dε H are respectively the electromotive forces generated by the rotation of the disk and the BH horizon as given by W02, and dε P is the electromotive force due to the rotation of the plunging region. The quantities Ω i , Ω i+1 and dZ i in Eqs. (6) and (10) are explained for the three different disk regions as follows. (1) MCDD region: Ω i and Ω i+1 are respectively the angular velocities of the inner footpoint i and outer footpoint i + 1 of the closed field line on the disk, and the resistance dZ i is defined by where R cor is the average areal resistivity of the disk and corona, which is related to R H , the surface resistivity of the BH horizon as follows, where η R is a parameter to adjust the value of R cor . For simplicity, we set η R = 0.1 in calculations. (2) MCHD region: Ω i = Ω H and Ω i+1 = Ω D are the angular velocities of the BH horizon and the disk respectively. The resistance where dZ H = 2ρ H dθ/̟. The parameters in Eq. (14) are given by (3) MCPD region: Ω i = Ω P and Ω i+1 = Ω D are respectively the angular velocities of the plunging region and the disk, and Ω P is given as follows (Shapiro & Teukolsky 1983;Wang 2000), The resistance is the same as that in MCDD expressed by Eq. (11). We adopt the same assumption as given by G09 about the corona: the optical depth of the corona τ cor = 1 and the height of the corona l = 10r ms . FITTING THE QPO FREQUENCIES AND X-RAY SPECTRA In our model, the toroidal electric current interacts with the dynamics of the accretion disk. To solve the dynamic Eqs. (5) and (6), we must know the configuration of the large-scale magnetic fields generated by the toroidal electric current, whose distribution is related to the surface density of the disk matter by Eq. (3), and ρ m is in turn figured out by solving the dynamical equations. For simplicity, we assume that initially there is no electric current in the disk, and solve the Eqs. (5) and (6) with H MC = 0 to get the global solution of the disk-corona system. The emerged spectrum of the disk-corona system is then simulated with Monte Carlo method based on the code developed by G09. The free parameters of the disk-corona model, e.g., BH spin a * and mass accretion rateṁ, can be determined by fitting the observed spectrum. So the surface density of the disk matter is obtained. Then we consider the interaction between the electric current and disk-corona with the iterative algorithm which consists of the following steps: (i) Assuming a value of the power law index n in Eq. (3), e.g. n = 5. (ii) Calculating the surface density of the toroidal electric current in the disk and the configuration of the large-scale magnetic fields. (iii) Resolving the disk-corona system by taking the MC effects into account. (iv) Repeating steps (ii) and (iii) until the surface density of disk matter remains stable. (vi) Repeating steps (i)-(v) until the QPO frequency is in accordance with the observations. Steps (ii)-(iii) should be repeated several times before the surface density of disk matter becomes stable. Finally, we simulate the emerged spectrum again until it is unchanged. To avoid the negative radiation flux from the inner disk due to the transfer of energy and angular momentum to the outer disk by the MCDD as argued by Gan et al. (2007), we adjust the radius of the inner boundary of the disk to deviate outwards from ISCO. Although the deviation is less than 10% of the radius of ISCO, it results in ∼30% decrease of QPO frequency. Fitting the HFQPOs in XTE J1859+226 and XTE J1650−500 with SPL state Observations show that HFQPOs in BHBs are associated with the SPL state which is characterized by high luminosity (therefore high accretion rate), strong power-law component and the steep power-law index (Γ > 2.4). We fit the single HFQPOs and the corresponding X-ray spectra of two BHBs XTE J1859+226 and XTE J1650-500. The comparisons between the simulated spectra and the observed ones are shown The fitting parameters are listed in Table 1, in whichṁ and α are respectively the mass accretion rate in terms of the Eddington accretion rate (1.4 × 10 18 m BH g s −1 ) and the viscosity parameter. The BH mass of XTE J1859+226 was estimated in the range 7.6−12M ⊙ (MR06). We fit the spectrum of XTE J1859+226 observed in October 16−18 of 1999 (MR06) with the BH mass of the upper limit 12M ⊙ and of the lower limit 7.6M ⊙ . Since the BH mass of XTE J1650−500 has not been well estimated, we take a larger mass 7.3M ⊙ and a smaller one 4M ⊙ , which are estimated by Orosz et al. (2004) to fit the spectrum. Very large BH spin is needed to fit the spectra for the lower limit to the BH mass, and the lower limit to BH spin corresponding to each BH mass is given in Table 1. The spectrum of XTE J1650−500 is fitted with the lower and upper limits to spin for each BH mass, while the spectrum of XTE J1859+226 can be fitted only for an extreme Kerr BH with the lower limit to the BH mass 7.6M ⊙ . The value of the viscosity parameter α is selected properly in the range 0.1−0.3 to fit the spectrum of each source. The difference of the value arises probably from the difference of the strength of the small-scale magnetic field in the disk if the viscous process is dominated by the tangled small-scale magnetic field. The values of the fitting parameters n and η for QPO frequencies are listed in the last two column in Table 1. We fix η at a certain value for each scale of BH mass. BH systems with smaller mass have stronger magnetic field allowing of larger value of η. A larger value of n is needed with a smaller BH mass, implying that the toroidal electric currents are concentrated very close to ISCO. QPOs in NGC 5408 X-1 and RE J1034+396 One of attractive features of our model is that it is applicable to fit QPOs with corresponding X-ray spectra observed in the BH systems of different scales. Strohmayer et al. (2007) discovered a strong 20 mHz QPO in the ULX NGC 5408 X-1, and the X-ray timing and spectral properties are analogous to those exhibited by Galactic stellar-mass BHs in the 'very high' or SPL state. The fitting parameters for the QPO and X-ray spectrum of NGC 5408 X-1 are listed also in Table 1. The simulated spectra for different parameters and the comparisons to the observed energy spectrum are shown in Fig. 7. However, at present, the BH masses in the ULXs are constrained to a wide range from 10 2 M ⊙ to 10 5 M ⊙ (Miller et al. 2003;Cropper et al. 2004;Roberts et al. 2005;Wu & Gu 2008). We take a very large BH mass 10 5 M ⊙ to fit the data of NGC 5408 X-1 since both the X-ray spectrum and QPO frequency are fitted better for larger BH mass. The QPO frequency and spectrum are both taken from Strohmayer et al. (2007) with the same Hydrogen column density, which are fitted in our model with the lower and upper limits to the BH spin as shown in Table 1. Miniutti et al. (2004). The source distance is set at 2.6 kpc (Homan et al. 2006) and the inclination i = 70 • is assumed. The first convincing QPO of AGNs was reported by Gierlinski et al. (2008) in RE J1034+396. Middelton & Done (2010) suggests that the QPO discovered in RE J1034+396 has an analogy to the 67 Hz QPO seen in the BHB GRS 1915+105 due to their similar 'hot disk dominated' energy spectra. Unlike other HFQPOs, the 67 Hz QPO in GRS 1915+105 is an exceptional case, which appears in thermal-dominant (TD) state (MR06). In this subsection, we fit both the 67 Hz QPO in GRS 1915+105 and the 0.00027Hz QPO in RE J1034+396 as a comparison. The spectrum of GRS 1915+105 showing the 67 Hz QPO is taken from Middleton & Done (2010). And the spectrum of RE J1034+396 showing the 0.00027Hz QPO is taken from Middleton et al. (2009) with the same minimum galactic absorption fixed at 1.31 × 10 20 cm −2 . The 'hot disk dominated' spectra of GRS 1915+105 and RE J1034+396 are both fitted with almost maximum BH spins as shown in Table 1, and their comparisons with the observed spectra are shown in Figs 8 and 9, respectively. The disparities between the simulated spectra and the observed ones in the energy bands 10−30 keV for GRS 1915+105 and 0.5−1 keV for RE J1034+396 indicate probably that a second Comptonization process is needed to generate the TD spectrum as shown in Figs 10 and 11 of Middelton & Done (2010), where a low temperature, optically thick thermal Comptonization is added to fit the spectra. This Comptonization may be generated from the transition layer between the disk and the corona. Inspecting Table 1 and Figs 5−9, we find that the QPOs and the corresponding X-ray spectra of BH systems of different scales can be fitted with the magnetic reconnection of the large-scale magnetic fields based on the disk-corona model. QPOs in massive BHs have similar features with those of BHBs, e.g. very centralized distribution of the electric currents or large value of n, and its inverse proportion to BH mass. They are either related to the SPL state, like XTE J1859+226 and XTE J1650−500, or related to the TD state, like GRS 1915+105. These similar characteristics hint probably that QPOs in BH systems of different scales may have the same origin and are associated with the same spectral state. The simulated spectra of NGC 5408 X-1 with different parameters: (a) m BH = 10 5 , a * = 0.95,ṁ = 0.012 and α = 0.1; (b) m BH = 10 5 , a * = 0.998,ṁ = 0.005 and α = 0.1. The plot style is the same as Figure 5. The observation data are taken from Strohmayer et al. (2007). The source distance is set at 4.8 Mpc (Karachentsev et al. 2002) and the inclination i = 75 • is assumed. The total Hydrogen column density is set as n H = 13 × 10 20 cm −2 (Strohmayer et al. 2007). DISCUSSION In this paper, a toy model for the QPOs in BH systems of different scales is proposed based on the magnetic reconnection of large-scale magnetic fields generated by the toroidal electric currents in the disk. The dynamical equations of accretion disk are resolved based on the interaction between the electric currents with the disk-corona system by using the iterative algorithm. The 190 Hz and 250 Hz single HFQPOs in BHBs XTE J1859+226 and XTE J1650−500 associated with SPL states are well fitted based on the disk-corona model with electric currents flowing in the inner disk. And the similar QPOs observed in ULX NGC 5408 X-1 and Seyfert 1 AGN RE J1034+396 and the corresponding X-ray spectra are also fitted. The spectrum of NGC 5408 X-1 is fitted with strong power-low component and steep power-law index suggesting that the QPO is similar to the HFQPOs in BHBs XTE J1859+226 and XTE J1650-500 and are probably associated with the same spectral state-SPL state. While the QPO in RE J1034+396 is analogous to the 67 Hz QPO in GRS 1915+105 which is associated with TD state. Energy keV E FE keV keV cm 2 s 1 keV 1 Fig. 9 The simulated spectra of RE J1034+396 with different parameters: (a) m BH = 7 × 10 6 , a * = 0.99,ṁ = 0.15 and α = 0.3; (b) m BH = 7 × 10 6 , a * = 0.998,ṁ = 0.14 and α = 0.3; (c) m BH = 2 × 10 6 , a * = 0.998,ṁ = 0.13, and α = 0.3. The plot style is the same as Figure 5. The observation data are taken from Middleton et al. (2009). The source distance is set at 125.9 Mpc (z=0.042, H 0 =100 km s −1 Mpc −1 ) and the inclination i = 40 • is assumed. We only consider the minimum galactic absorption fixed at 1.31 × 10 20 cm −2 (Middleton et al. 2009). Similarities of the QPOs in BH systems of different scales enable us to estimate some physical quantities of the massive BHs which have not been well constrained at present. As suggested by MR06, the frequencies of HFQPOs in BHBs showing the 3:2 frequency pairs are inverse proportional to the BH mass. Abramowicz et al. (2007) used the 1/M scaling to expect QPO frequencies for BHs of different scales and neutron stars. Similarly, we can use this relation to estimate the BH mass in ULX NGC 5408 X-1. Including the three BHBs showing the single HFQPO, we have the relationship, m BH = 1460ν −1 QPO , between the BH mass and the QPO frequency as shown in Fig. 10. The BH mass of NGC 5408 X-1 is then estimated about 7.3 × 10 4 M ⊙ with the 0.02 Hz QPO as shown by the black dot in Fig. 10. The quasi-periodic signals similar to QPOs were also discovered in the X-ray flux of Sgr A. If these signals are indeed QPOs and are triggered by the magnetic reconnection described in this paper, then the lower limit of the BH spin can be constrained because the outer footpoints of the magnetic field lines cannot extend to the infinite distance. We estimate the lower limit of the BH spin of Sgr A as 0.448 by fitting the 22.2 minutes signals discovered in the X-ray flare on 2004 August 31 (Blanger et al. 2006) using the mass 4.4 × 10 6 M ⊙ (Genzel et al. 2010), which is very close to the value a * ≈ 0.44 ± 0.08 estimated by Kato et al. (2010) using the QPO method in the context of disk-seismology. It is noticed that the upper and lower kHz QPOs of accreting X-ray binaries are interpreted as the Alfven wave oscillations with different accreted material mass densities at a preferred radius near the star surface, and this model successfully explains the empirical relation between the upper and lower kHz QPO frequencies and the linear relation between the high and low QPO frequencies of BHs, neutron stars and white dwarfs (Zhang 2004;Zhang et al. 2007). The idea of interpreting QPOs with magnetic reconnection proposed in this paper may also apply to neutron stars and other astrophysical objects. For example, QPOs can be interpreted as the reconnection of magnetic field on the surface of a neutron star, then the model evolves to the beat frequency model (e.g., Miller et al. 1998), providing a physical mechanism for QPO production. Magnetic reconnection is being increasingly recognized as an important process in high energy objects, such as stellar X-ray flares, accretion disk corona, and magnetar flares. In this paper, we apply it to interpret the QPOs in BH systems of different scales based on a possible origin of the large-scale magnetic field in BH accretion disk. The magnetic fields generated in the inner disk by the electric currents are ∼ 10 6 and ∼ 10 3 Gauss for BHBs and AGNs respectively. Although the strength of these fields is much smaller than that of the tangled small-scale magnetic fields in the disk, the magnetic reconnection of the large-scale magnetic fields should have some effects on the heating of the corona, and these should be considered in the future work. Furthermore, the influence of the magnetic reconnection on X-ray spectra is another open question to be solved.	2013-01-02T03:22:22.000Z	2013-01-02T00:00:00.000	{ "year": 2013, "sha1": "2523bfabd80b311e5f0b5c9e09e5ff2513c32b3c", "oa_license": null, "oa_url": "http://arxiv.org/pdf/1301.0162", "oa_status": "GREEN", "pdf_src": "Arxiv", "pdf_hash": "2523bfabd80b311e5f0b5c9e09e5ff2513c32b3c", "s2fieldsofstudy": [ "Physics" ], "extfieldsofstudy": [ "Physics" ] }

Subsets and Splits

No community queries yet

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