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 |
|---|---|---|---|---|---|---|
15368798 | pes2o/s2orc | v3-fos-license | Determinants of anxiety in patients with advanced somatic disease: differences and similarities between patients undergoing renal replacement therapies and patients suffering from cancer
Objective Anxiety is the most frequent emotional reaction to the chronic somatic disease. However, little is known about anxiety and coping strategies in patients with end-stage renal disease (ESRD) undergoing renal replacement therapies (RRTs). The purpose of the study was to assess the intensity and determinants of anxiety in patients treated with different RRTs in comparison with end-stage breast cancer patients and healthy controls. Methods The study involved (1) ESRD patients undergoing different RRTs: 32 renal transplant recipients, 31 maintenance haemodialysis and 21 chronic peritoneal dialysis patients, (2) women with end-stage breast cancer (n = 25) and (3) healthy persons (n = 55). We used State–Trait Anxiety Inventory, Scale of Personal Religiousness, Mental Adjustment to Cancer Scale, Rotterdam Symptom Checklist with reference to medical history. The data thus obtained were analysed using the analysis of variance, the Tukey’s HSD post hoc test and Spearman’s rank correlation coefficient. Results Both ESRD and breast cancer patients revealed higher level of anxiety state and trait than healthy controls; however, there was no statistically significant difference found between both findings. There was a tendency towards higher levels of anxiety state in breast cancer patients when compared to ESRD patients undergoing the RRT treatment and for both groups non-constructive coping strategies correlated with the levels of anxiety state. With ESRD patients undergoing RRTs, the intensity of anxiety state did not depend on the mode of treatment but on the correlation between the levels of anxiety and the general quality of their life, psychological condition and social activity. Conclusion In patients with advanced somatic disease (ESRD and end-stage breast cancer), non-constructive strategies of coping with the disease require further evaluation and possibly psychological support.
Introduction
End-stage renal disease (ESRD) has become a major increasingly prevalent health issue. Patients' permanent loss of renal function requiring a chronic renal replacement therapy (RRT) leads to dramatic life changes. RRTs prolong the lives of ESRD patients but at the same time bring them an enormous burden. A broad range of physical symptoms, activity limitations, altered body image, fatigue and various treatment-related restrictions are stressors precipitating the need for coping efforts and possibly leading to negative outcomes such as depression and anxiety. Therefore, psychological disorders are frequently encountered in such cases [1,2].
Anxiety is an unpleasant emotion affecting patients with advanced somatic disease, and there are various ways of coping with it. Most scientists who deal with the issue of emotional reactions to somatic disease point out that it is anxiety which is the most common emotional response to somatic diseases [3][4][5][6]. Anxiety is a response to either a threatening situation, a risk of losing an important value or a lack of stability, that is, the feeling of insecurity connected with finding oneself in a new unfamiliar situation [5]. We can assess anxiety as a fairly stable anxiety-related trait combined with personality features and a variable anxiety-induced state, which depends on external factors. Getting ill involves the risk of a patient's permanent loss of health, development of disability or even death. At the same time, the disease changes the level of one's self-esteem and one's previous ways of dealing with problems. Thus, the phenomenon of anxiety occurs at the very onset of disease symptoms during diagnostic and treatment stages, which follow, especially if hospitalization is necessary [5].
It has been frequently observed that there is a clear association between the patient's religiousness and spirituality, on the one hand, and clinical outcomes on the other. One's spirituality and religious involvement may play the key role in developing coping mechanisms, which-as has been verified-both improve one's quality of life (QoL) and reduce one's perception of the burden of illness and of its depressive effects [7,8]. Spinale et al. [9] showed that highly developed spirituality resulted in patients' improved chances of survival with regard to the haemodialysis population, but religion as a coping mechanism did not. In end-stage cancer patients, religiousness turned out to be an effective tool for coping with anxiety. Because little is known about anxiety and the coping strategies in patients undergoing renal replacement therapies, it is also not known if religiousness is a determinant of anxiety in ESRD population.
Objectives
The detailed aims of the study were (a) To establish the intensity of anxiety trait and anxiety state in study groups. (b) To find the relationship between physical symptoms, psychological symptoms, social activity and general quality of life and level of anxiety patients undergoing different RRTs (haemodialysis, peritoneal dialysis, renal transplant recipients) and cancer patients. (c) To assess the association between anxiety state and coping strategies in patients. (d) To find out if religiousness has an impact on anxiety levels in the studied groups.
Materials and methods
The study involved 164 subjects (85 women and 79 men), aged between 22 and 80 (average age: 49.17) who comprised three study groups: the first group (I) consisted of patients undergoing different modes of RRT (n = 84) divided into three groups according to method of treatment: 32 renal transplant patients (RTx), 31 maintenance haemodialysis (HD) and 21 chronic peritoneal dialysis (PD) patients; the second group (II) included terminal breast cancer patients (n = 25). These patients found themselves relapsing further into disease and as a result experienced a significant limitation of their professional and social activities; the third group (III) comprised healthy persons (n = 55). The following instruments were used to conduct the study: 1. State-Trait Anxiety Inventory (STAI) in an authorized translation by J. Strelau, M. Tyskarczyk and K. Wrześniewski [10]. It is an adaptation of the American version of the State-Trate Anxiety Inventory developed by Spielberger et al. [11]. This instrument enables one to differentiate and to measure independently two different types of anxiety, namely anxiety understood both as the current state of an individual (''state anxiety'') and as a relatively stable trait of this individual's personality (''trait anxiety''). Both scales contain 20 short statements assessing the patient's subjective feelings. The received results are compared with suitable centile norms. Results under 30 centile show the low level of anxiety, between 30 and 60 centiles reflect the normal level of anxiety and more than 60 centile indicate the high level of anxiety. This inventory is an instrument that lives up to the requirements of modern psychometry. It has high reliability and accuracy indices, respectively. Estimates of reliability (Cronbach's alpha) of the Polish version of the Scale were satisfactory for two scales Anxiety-State (alpha = .90) and Anxiety-Trait (alpha = .88) [10]. 2. Scale of Personal Religiousness (SPR) [12]. This scale was created by Jaworski. It measures religious attitudes and enables one to differentiate between personal religiousness and the apersonal one. Personal religiousness means deeply rooted faith in God that is expressed through religious practices, spiritual development and being a member of a church community following religious rules in daily life. Apersonal religiousness is the opposite of the previous attitude: it reduces faith to tradition, ceremonies and superficial acts of worship. The maximum result-210 pointsindicates the personal character of religiousness. The lowest score-30 points-is the evidence of apersonal religiousness. There is a continuum of varied religiousness concerning assessment of both personal and apersonal attitudes towards religion. This scale was created in Polish. The reliability of the scale (Cronbach's alpha) was proven of retesting method where satisfactory: for particular questions (alpha = .69-.86) [12]. 3. Rotterdam Symptom Checklist (RSCL) [13]. The author of the Polish version of the Scale is Z. Juczyński. It is an adaptation of the American version of the test developed by Watson et al. [15].
The Scale included 29 statements. The results must be estimated according to each of four strategies separately. There are seven statements in each strategy. The range of possible results should be contained within the limit of 7-28 points. The higher the score the larger the intensification becomes of the characteristic modes of behaviour for a specified way of coping with disease. The scores thus obtained should be converted into standardized units-by means of socalled stens (i.e. standardized ten-point scales). Results from 1 to 4 need to be treated as low while those from 7 to 10 should be regarded as high. Results in limits 5 and 6 be classified as mean estimates of reliability (Cronbach's alpha) of the Polish version of the Scale were highly: for helplessness-hopelessness strategy (alpha = .92), fighting spirit attitude (alpha = .90), anxiety preoccupation strategy (alpha = .89), cognitive avoidance strategy (alpha = .87), fatalism (alpha = .64) [16]. 5. Medical records served as the source of data with regard to methods of treatment, comorbidities, social and demographic details and selected laboratory indices (Kt/V for dialysis adequacy in HD and PD patients and eGFR for kidney allograft function), haematocrit levels. 6. Charlson Comorbidity Index (CCI) [17] was used as a single aggregate measure of a patient's comorbid conditions. CCI scores were calculated according to age modification as proposed by Charlson et al. [17].
Statistical methods
Statistical analysis was conducted using the Statistica 9 program. Spearman's rank correlation coefficient was used for analysing the correlation level between variables. The influence of independent variables on the dependent variables was determined by means of variance analysis (ANOVA). Differences among subgroups were assessed using the Tukey's test. A p value \.05 was considered statistically significant. Data are shown as mean ± standard error.
Results
Basic characteristics of studied groups are presented in Table 1. Patients with end-stage breast cancer were women. They displayed no symptoms of comorbidity.
Among patients undergoing RRTs, RTx recipients were younger than both HD and PD patients; standard mean ages of dialysis patients were comparable. Healthy controls were the youngest of all the studied groups. CCI scores were the lowest for RTX patients and the highest for HD patients. These findings bore statistical significance. Statistical analysis confirmed a significant positive relationship between trait anxiety and state anxiety. Analysis of the results showed that patients with advanced somatic illness (ESRD, breast cancer) revealed higher levels of anxiety state and trait in comparison with healthy controls; however, the results did not reach statistical significance. There was a tendency towards higher level of anxiety state in breast cancer patients as compared to ERSD patients undergoing RRT; HD, PD and RTx patients had similar intensity of anxiety state too. Results are presented in Table 2.
Tukey's HSD Post hoc test confirmed no significant relationship (p [ .05) between level of anxiety and gender in ESRD patients and healthy control groups.
The coping strategies as helplessness-hopelessness, anxiety preoccupation correlated with level of anxiety state in ESRD patients undergoing RRT. The coping strategies as fatalism and anxiety preoccupation correlated with levels of anxiety state in women affected by breast cancer.
In contrast to breast cancer patients, in ESRD patients religiousness had no effect on their anxiety state. Spearman's rank correlation coefficient values for the variables state anxiety, religiousness and patient ways of coping are presented in Table 3.
QoL was assessed with the help of RSCL. There were more physical symptoms in breast cancer patients in comparison with dialysis and renal transplant patients. On the other hand, cancer patients evaluated higher their overall QoL when compared with ESRD patients. In RTx patients, general QoL was significantly better than in HD patients. There was a difference of borderline significance in the general QoL evaluation between RTx and PD patients (p = .054). General QoL did not differ between HD and PD patients. Psychological symptoms and activity level did not differ between all groups (cancer and RRT patients). RTx patients suffered from as many physical and psychological symptoms as dialysis patients, yet their self-sufficiency was rated significantly higher than in HD patients. In ESRD group, there was a significant positive correlation between physical symptoms, activity level and overall QoL and CCI scores. RTx recipients with better kidney allograft function assessed by eGFR experienced fewer physical symptoms. There was no such interrelationship with regard to dialysis adequacy assessed by Kt/V in both HD and PD patients (data not shown). The haematocrit (Hct) levels were measured. These results did not differ between studied groups. Table 4 shows the correlation between anxiety state and various domains of QoL in ESRD and cancer patients. There was a correlation between level of anxiety and psychological symptoms and social activity in RRT patients. In RTx patients, level of anxiety was also negatively correlated with general QoL, whereas in HD patients it did not correspond to the level of social activity. There was no correlation between somatic symptoms and level of anxiety in any of ESRD patients, regardless of the mode of RRT. In contrast to RTx, HD and PD patients, in breast cancer patients, the intensity of anxiety corresponded to the results in all domains of RSCL.
There was no significant relationship between level of anxiety and time of treatment in the studied groups.
Discussion
The findings of the research being presented here have confirmed that patients with higher levels of trait anxiety who suffer from advanced somatic disease reveal more intensive state anxiety. It seems to be obvious that people with anxiety as a feature of character will become more anxious when they have to cope with this disease. We observed that patients with advanced somatic illness (ESRD, breast cancer) revealed higher level of anxiety state and trait in comparison with healthy controls; however, the results did not reach statistical significance. These results suggest that anxiety is the frequent emotional reaction to the chronic somatic disease and can be compared with the obtained from previous studies conducted on patients undergoing haemodialysis treatment [18][19][20].
Anxiety trait is understood as a relatively stable trait of an individual's personality. This may suggest that chronic somatic disease can cause a change of personality.
Patients with breast cancer revealed higher levels of anxiety than other studied patients including ESRD patients. In a study by Majkowicz et al. [21], cancer patients expressed significantly higher levels of depression and anxiety while HD patients manifested higher levels of aggression. Moreover, it is typical of end-stage cancer to make existing symptoms become periodically aggravated and cause new ones to appear. New symptoms will tend to arouse anxiety [22]. The fact that cancer patients were all women could be important for the level of anxiety. Some researchers point to gender dimension as one of the factors influencing on anxiety level [23,24]. A study carried out by Spencer et al. [25] showed that female patients, more physically impaired and younger advanced cancer patients are more likely to meet criteria for an anxiety disorder.
As expected, terminal cancer patients rated their somatic condition worst of all study groups. It is surprising that cancer patients scored overall QoL better when compared to dialysis and renal transplant patients. The possible explanation is that renal patients tend to be affected by medical therapy more heavily. In many countries, this is the reason for the withdrawal of aggressive treatment for dialysis patients [26]. Renal transplantation is the ideal choice of treatment for ESRD. Not only does it improve one's chances of survival, but it is believed to be the least stressful of RRTs since it brings patients fewer life style restrictions and improves their physical and psycho-social well-being. Therefore, successful transplantation should mitigate high level of negative emotions observed in dialysis patients. Among studies comparing the condition of mental health between haemoperitoneal dialysis and transplant patients, some have shown an improvement in anxiety after renal transplantation [27,28], while others have revealed no sign of superiority of any of the RRT modes [29].
In our study, HD, PD and RTx patients had similar intensity of anxiety state, while their general evaluation of QoL differed with RTx recipients experiencing best QoL. This was surprising as RTx patients were younger and manifested fewer instances of comorbidity than dialysis patients. We would have expected these patients to express lower levels of negative emotions than HD or PD patients. Our findings are especially surprising as most of the previous studies showed a favourable effect of renal transplantation on the improvement in anxiety state [27,28]. We would well have expected PD patients to have a better QoL than their HD counterparts, even comparable to healthy individuals, for example, because this method of RRT allows increased autonomy [20,21]. On the other hand, in a study by Sayin et al. [29], QoL was similar to that of HD, PD and RTx patients. However, anxiety levels influenced QoL in the case of HD and PD patients alone-RTx recipients remained unaffected. This is consistent with our observation that while all patients undergoing RRTs maintained similar anxiety state levels, RTx patients scored best overall QoL. We can thus hypothesize that RTx patients, who have the highest functional capacity and QoL, become anxious about a possible loss of these values. Although RTx patients suffered from as many physical and psychological symptoms as dialysis patients did, their self-sufficiency, constituting a highly appreciated quality, was enough to boost their QoL.
Intrinsic religiousness was the most effective factor diminishing anxiety in advanced breast cancer patients as opposed to dialysis and transplant ones. Women with intrinsic religiousness show the lower intensity of anxiety; on the other hand, patients with extrinsic religiousness display a higher level of anxiety. Some researchers point to the transcendental dimension as one of the mechanisms for coping with the fear of death [30][31][32][33]. Alvarado et al. [34] showed that people with low levels of the fear of death strongly believe in afterlife. Moreover, a study by Phelps et al. [35] showed that religious coping in advanced cancer patients is associated with receipt of intensive lifeprolonging medical care near death defined as receipt of mechanical ventilation or cardiopulmonary resuscitation in the last week of life.
The analysis of ways of coping with disease revealed that the strategies as helplessness-hopelessness, anxiety preoccupation correlated with level of anxiety state in patients with ESRD undergoing RRT and the strategies as fatalism and anxiety preoccupation correlated with level of anxiety state in women with breast cancer. The coping strategies were assigned to either of two categories: constructive coping strategies (fighting spirit, cognitive avoidance) and non-constructive coping strategies (helplesshopeless, anxious preoccupation, fatalism). We observed that non-constructive ways of coping correlated with higher level of anxiety in studied groups. Our study on 155 women with breast cancer at different disease stages has shown that patients preferring constructive coping strategies revealed lower levels of anxiety when compared to patients manifesting non-constructive forms of coping [36]. Our results concerning patients who receive RRTs are consistent with the report of Takaki et al. [37], who have shown that emotion-oriented coping such as anxiety preoccupation correlated positively with level of anxiety in over 400 HD patients.
In ESRD patients' correlation between level of anxiety and general quality of life, psychological state and social activity depended on mode of treatment. There was no correlation between somatic symptoms and levels of anxiety in the case of dialysis and renal transplant patients. The breast cancer patients' intensity of anxiety correlated with the results in all domains of RSCL. In the phase of recurring cancer, the somatic condition has the most significant influence on the levels of state anxiety. Somatic signs at this stage of the disease worsen patients' health condition and bring about their feeling of threat.
Study limitations include the potential for selfreport response bias and sampling bias because of convenience sampling. We also decided to use tools originally designed for cancer patients that so far have not been verified with ESRD patients. However, this was the only possible way to make comparisons between patients suffering from advanced stages of renal diseases and cancer.
In addition, we cannot exclude the idea that a tendency towards higher levels of anxiety state among cancer-affected patients may be associated with their gender. Despite these limitations, we believe that our study is of significant importance as to our knowledge this is the first study to compare anxiety and coping strategies along with religiousness in patients undergoing three different treatment strategies in ESRD. Moreover, our findings thus obtained are further juxtaposed with those from research done into healthy individuals, on the one hand, and cancer patients on the other.
Conclusions
Patients with a high level of trait anxiety show higher intensity of state anxiety during advanced somatic disease. ESRD patients undergoing RRT present higher level of anxiety than healthy controls and lower than breast cancer patients. The intensity of anxiety is similar irrespectively of the mode of renal replacement therapy (RTx, HD and PD). An elevated level of anxiety trait in patients with advanced disease suggests that the presence of chronic somatic illness modifies this personality feature. The phenomenon of intrinsic religiousness works as a factor most efficiently diminishing the levels of anxiety in advanced breast cancer patients as opposed to dialysis and transplant patients. In patients with advanced somatic disease (ESRD patients and women with breast cancer), non-constructive strategies of coping with the disease require further evaluation and possibly more psychological support. | 2016-05-04T20:20:58.661Z | 2012-11-17T00:00:00.000 | {
"year": 2012,
"sha1": "a488ed3582608ae33bf4795829226dc9c856d858",
"oa_license": "CCBY",
"oa_url": "https://link.springer.com/content/pdf/10.1007/s11255-012-0326-6.pdf",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "a488ed3582608ae33bf4795829226dc9c856d858",
"s2fieldsofstudy": [
"Medicine",
"Psychology"
],
"extfieldsofstudy": [
"Medicine"
]
} |
246072224 | pes2o/s2orc | v3-fos-license | Nanotechnology strategy as antibacterial: A primer for the notice
Background: Multidrug-Resistant (MDR) bacteria cause a serious threat to all fields of bioscience because it may result in treatment failure which might have severe consequences, especially in case of critical or special patients. The development of new antibacterial agents cropped up to be a complex process. structures ranging from approximately 1-100 nm. Nanoparticles (NPs) are of great importance in areas like health care, cosmetics, food and feed, environmental etc. Nanoparticles matrix properties and constituent materials can be used to categorize them into inorganic and organic nanosystems. Techniques use within the synthesis of NPs includes; chemical, physical, chemistry, and biological methods, every methodology has its advantages and disadvantages with common issues of value, quantifiability and uniform particle size. Sir Alexander Fleming, a noble winner aforesaid that “There is perhaps no therapy drug to that in appropriate circumstances the bacteria can't react by in a way deed fastness.” thus, there's high chance that the organism can also become resistance to newly developed medication at later stage. Conclusion: To this end NPs are considered to be good antibacterial agents and may overcome the barrier of MDR owing to their ability to anchor to the bacterial cell wall, multifunctional mechanisms to intervene normal cell functionality and subsequently penetrate it, thereby causing structural changes in the cell membrane permeability leading to cell death.
INTRODUCTION
The development of a new antibacterial agents cropped up to be a complex process in view of the capability to produce effective drugs with no or reduce side effects, in addition to the high production costs and the time required for approval of new drugs that takes regarding 10-15 years (Wang et al., 2017;Patra et al., 2018;Singh et al., 2014).
Nanoparticle has surrounding interfacial layer with particles between 1 and 100 nanometres (nm) in size. The interfacial layer is an integral part of nanoscale matter, fundamentally affecting all of its properties (Singh et al., 2016). Ions, inorganic and organic molecules are major constituents of their interfacial layer typically. Organic molecules coating inorganic nanoparticles are known as stabilizers, capping and surface ligands, or passivating agents (Singh et al., 2016).
Nanomedicine plays an important role in enhancing the effectiveness of existent medicine, by enhancing the chemical science properties and stability of antibiotics, prolongation of antibiotic unharness, providing an opportunity of biofilm acquisition, additionally to the potential of targeted delivery to the location of infection and improved circulation with a ensuant reduction of the connected facet effects compared to the
Prime Scholars Library
Advance Journal of Virology, Epidemic and Pandemic Diseases corresponding free medication (Natan and Banin E, 2017).
NANOPARTICLES
Nanotechnology is a unique and important field of modern research field, it deal with manipulation and synthesis of particle structures starting from about 1-100 nm. Nanoparticles (NPs) have big selection of applications in areas like health care, chemical industries, natural philosophy, nonlinear optical devices, area industries, medical specialty science, drug-gene delivery, energy science, cosmetics, food and feed, environmental health, mechanics, optics, optoelectronics, catalysis, single negatron transistors, light-weight emitters, and photo-electrochemical applications.
Nanoparticles are product of materials of various chemical natures like biomolecules, non-oxide ceramics, polymers, organics, metal oxides, silicates, carbon and metals. Nanoparticles additionally exist in much totally different morphology like spheres, cylinders, platelets and tubes (Kim, 2007). Nanoparticles is generally classified into two namely, organic nanoparticles that embrace carbon nanoparticles (fullerenes) and also the inorganic nanoparticles that embrace magnetic nanoparticles, metallic element nanoparticles (gold and silver) and Semi-conductor nanoparticles (titanium chemical compound and atomic number 30 oxide) (Natan and Banin E, 2017) Nano materials have distinctive chemical and biological properties as compared to their larger counterparts (Shrivastava et al., 2007). The properties of nanomaterial's will greatly influence their interactions with bio molecules and cells, because of their peculiar chemical composition, surface structure, size, shape, charge, solubility and agglomeration. For instance, nanoparticle accustomed turn out exceptional pictures of tumour sites; single-walled carbon nanotubes are used as high-efficiency delivery transporters for biomolecules into cells (Varshney et al., 2009). By additional analysis in technology, it is helpful for each facet of human life such as somatic cell analysis, regenerative drugs and nutraceuticals are among the leading sectors that may be changed by technology innovations. Silver NPs is an area of interest as a result of the distinctive properties (e.g. form and size relying electrical, magnetic properties and optical) which might be incorporated into cosmetics merchandise, composite fibres, antimicrobial applications, biosensor materials, refrigerant superconducting materials, and electronic parts (Thu et al., 2013). It's additionally helpful when incorporated into antimicrobial applications, biosensor materials, refrigerant super-conducting materials, composite fibres, cosmetic merchandise, and electronic parts. Some valuable applications of silver NPs are in medical specialty, medicine, and odontology etc. many physical and chemical ways are used for synthesizing and stabilising silver nanoparticles (Senapati et al., 2005).
CLASSIFICATION OF NANOSYSTEMS
Nanoparticles matrix properties and constituent materials can be used to categorize them into inorganic and organic nanosystems (Baranwal et al., 2018). Inorganic nanosystems are synthesis from inorganic oxides. Their synthesis technique depends on chemical reduction of metallic salts with a reducing agent. Environmental parameters, as an example pH scale and temperature, play a serious operate in determination of the specificities of those materials, aggregation, the in vitro drug release kinetics, consequently affecting their loading capacity and hence their antibacterial affect (Ashik et al., 2018). In addition, organic nanosystems like compound micelles, liposomes, lipid-based nanoparticles and compound nanoparticles have desirable biodegradability and biocompatibility options, making them appropriate candidates for clinical use (Martin-Serrano et al., 2019).
Liposomes nine composition and characteristics of liposomes
Liposomes are the foremost extensively evaluated antimicrobial drug delivery nanosystems. They're characterized by spherical structures of lipid bilayer(s) encompassing associate degree inner liquid house, go in size from 0.02 to 10 nm. The effectuality of antibacterial-loaded liposomes in biofilm demolition depends on the chemical science properties of liposomes that manage their stability and in vivo interactions. Nanoparticle has small unilamellare veiscles of '100 nm which allow antibiotic delivery helping them to display high capability in the eradication of bacterial strains. Liposomes proved to be useful for the management of topical, oculare bacterial infections, pulmonary, and vaginal (Nermin et al., 2020).
Advantages of antibiotics-loaded liposomes as drug delivery agents
Better protection and enhanced antibiotics bio distribution. Selective biofilm targeting affinity. Improved selectivity towards intracellular and extracellular bacterial strains.
Bimetallic NPS
Ag and Au (Np) are also employed to reinforce the results of a drug and cut back the specified dose. They can also be used alone since they possess antimicrobial properties (Singh et al., 2016). AuNPs represent sensible vectors to the delivery of medical specialty compounds. Synergism of Au and Ag alloys is a great solution since they improved stability in advanced biological media provided by gold and combine the antimicrobial result of silver with the convenience of functionalization. Fakhri and colleagues synthetized and functionalized AgAuNPs with an antibacterial drug and over that there exists a synergistic result of the antibiotic with the bimetallic nanoparticle, with greater bactericidal activity of this form in detriment of its free forms (Pedro et al., 2018).
Silver Nanoparticles (AGNPS)
Since the traditional times, silver has been recognized as having antimicrobial effects (Reidy et al., 2013) supported all the proof thus far, AgNPs are as the most promising inorganic NPs which will be used for the treatment of bacterial infections (Natan et al., 2017). These NPs is also synthesized by ancient chemical reduction or via "green" chemistry approaches using plant and/or bacteria extracts (Ribeiro et al., 2018) many mechanisms are planned to grasp AgNPs mediate cellular death, together with cytomembrane disruption (Pedro et al., 2018), oxidization of cellular parts, inactivation of the metabolism chain enzymes, production of ROS, and decomposition of the cellular parts. The adsorption of the NPs leads to alteration of the negative charge of the cell wall and also the depolarization of the cell wall which make them to become more permeable. The entry of AgNPs induces ROS which will inhibit ATP production and DNA replication. However, there's proof that AgNPs will unharness Ag + , acknowledged to exhibit antimicrobial activity, when interacting with thiol-containing proteins, which weaken their functions. The precise technique of the antibacterial mechanism of AgNPs continues to be not utterly understood (Pedro et al., 2018). Previous studies shows that AgNPs exert many antibacterial activity, which can justify why bacteria resistance to silver is rare. Considerations relating to the toxicity and genotoxicity of AgNPs are raised however numerous authors have conducted clinical trials supported AgNPs and no necessary clinical alterations are detected. Curiously, AgNPs are found to exhibit higher antimicrobial activity than antibiotics like vancomycin or gentamicin against P. aeruginosa and MRSA (Ribeiro et al., 2018).
Gold Nanoparticles (AuNPs)
Metallic gold is taken into account as inert and non-toxic, which can vary once it shifts oxidation states (I and II) Gold NPs (AuNPs) is also synthesized by traditional chemical reduction of a gold salt or via "green" chemistry approaches by making use of plant and/or bacteria extracts. Gold nanoparticle alloy of yellow colouration are shown in Figure 1. The foremost used and represented technique is that the chemical synthesis supported the reduction of chloroauric acid by turn (Fernandes et al., 2017). Some studies have addressed the potential of making use of AuNPs as antibacterial agents, however some disceptation still exists in line with Yu H and collaborators, AuNPs AuNPs are usually not bactericidal at low concentrations and sapless antibacterial at high concentrations. This is often presumably because of the result of co-existing chemicals, like gold ions, surface coating agents, and chemicals concerned within the synthesis that weren't utterly removed. The antibacterial mechanism of AuNPs is associated to (i) the collapse within the membrane potential, impeding ATPase activity inflicting a deterioration of the cell metabolism; (ii) impeding of the binding monetary unit of the organelle to transfer RNA; and (iii) Shamaila and co-workers showed that AuNPs may affect the bacterial respiratory chain by attacking nicotinamide (Pedro et al., 2018).
Metal oxides
Metal oxides NPs are among if not the most important, foremost explored and studied family of NPs and are famous to effectively, rising as hopeful candidates to challenge antimicrobial resistance (Kadiyala et al., 2018). Iron compound (Fe 3 O 4 ), flowers of zinc (ZnO), and oxide (CuO) possess antimicrobial properties and may be applied in clinical care. Thanks to the intrinsic photocatalytic activity of the metal oxides they generate ROS and become powerful agents against bacterium these are delineate in additional detail on the subsequent sections (Pedro et al., 2018).
Iron oxide (FE 3 O 4 )
The synthesis of iron compound NPs could also be achieved via totally different routes. The medicine mechanism of those NPs is especially attributed to dissolved metal ions and also the generation of ROS (Wang et al., 2017). Super-paramagnetic iron compound nanoparticle carry out intercellular movement by penetrating the membrane and officious with the negatron. It's been delineating that iron compound NPs will harm macromolecules, proteins, through the formation of ROS and together with DNA.
METHOD
OF SYNTHESIZING OF NANOPARTICLES Many techniques are used for the synthesis of Ag-NPs by exploitation of chemical, physical, chemistry, and biological ways (Figure 2). Every technique has its advantages and disadvantages with common issues of value, measurability, uniform particle size, and also the size distribution (Petra et al., 2018). Historically, metal nanoparticles are created by physical ways like particle sputtering or periodical optical maser ablation and chemical ways like reduction, thermal synthesis, sol-gel technique, hydrothermal and so on. However, recently, the environmentally friendly synthesis ways (by exploitation of natural products) are developed underneath the branch of "green syntheses." relying upon the chosen path of synthesis and totally different experimental conditions, the silver NPs of various morphology, sizes, and shapes are often obtained. Nevertheless, the foremost necessary criteria are that the size distribution that ought to be achieved narrow as possible for the target-specific applications. Three necessary ways (chemical, physical and biological) for the synthesis of nanoparticles are mentioned as follows (Petra et al., 2018).
Chemical methodology of synthesis
Among the present strategies, the chemical strategies are commonest used for the assembly of Ag-NPs. The chemical reduction of metal ions is that the most universal and simple route for the preparation of the metal nanoparticles (Oliveira et al., 2005). The chemical transformation of the silver ions into the silver nanostructures will occur by photochemical methodology, wet chemical synthesis with or while not templates, by using liquid, chemical compound templates, solutionbased methodologies like aspartate reduction and starch-mediated reduction, etc. Generally, the chemical synthesis method of the Ag-NPs in answer typically employs 3 main elements which has metal precursors (for formation of AgNPs: AgNO three AgClO four, AgCl, (PPh three) 3AgNO 3 , CF 3 CooAg), reducing agents and stabilizing/capping (Iravani et al., 2014).
Physical methodology of synthesis
In the physical synthesis method of Ag-NPs, usually, the physical energies (thermal, ac power, and arc discharge) are used to supply Ag-NPs with a slender size particle distribution. This approach can permit us to produce large quantities of Ag-NPs at once (Oliveira et al., 2005). Under the physical strategies, the metallic NPs are often typically contrived by evaporation-condensation method that would be administered during a tube chamber at gas pressure. The large area of tube chamber, consumption of enormous quantity of energy, raising the environmental temperature round the supply material and tons of your time for achieving thermal stability, these are among the few drawbacks of the strategy. Another physical methodology of synthesis of Ag-NPs may be a thermal decomposition methodology used to synthesize the powdered Ag-NPs. This means that the Ag-NPs were ready with a really slender size distribution for synthesizing the metal NPs and by evaporating the source beneath the flow of carrier gas, i.e., air. It had been reported that the mean diameter, the geometric variance, and also the total concentration of spherical NPs while not agglomeration will increase with the temperature of the surface of the heater (Oliveira et al., 2005).
Biological synthesis of nanoparticles
Usually, wet-chemical or physical methodology is employed to arrange the metal nanoparticles. However, the chemicals employed in physical and chemical strategies are typically costly, harmful and combustible however the biogenic strategies are a value effective, energy saver and having environmentally benign protocols technique for inexperienced synthesis of silver nanoparticles from completely different bacteria (yeast, fungi and bacterium, etc.) and plant tissues (leaves, fruit, latex, peel, flower, root, stem, etc.) (Oliveira et al., 2005). Phytochemicals (lipids, proteins, polyphenols, group acids, saponins, aminoacids, polysachccarides amino polyose, enzymes, etc.) gift in plants are used as reducing and capping agent. The utilization of agro waste and bacteria materials not solely reduces the value of synthesis however additionally minimizes the requirement of victimization venturous chemicals and stimulates inexperienced synthesis.
This methodology of synthesis is extremely easy, requiring less time and energy compared to the physical and chemical strategies with predictable mechanisms. The opposite blessings of biological methods is the supply of an enormous array of biological resources, attenuate time demand, high density, stability, and also the ready-to-soluble asprepared nanoparticles in water. Therefore, biogenic synthesis of metal NPs unwraps up huge opportunities for the employment of perishable or waste materials.
Biosynthesis of silver nanoparticles by making use of bacteria
Several bacteria are documented to synthesize silver nanoparticles. Silver nanoparticles at 440 nm peak that corresponded to the surface plasmon resonance of silver nanoparticle were wont to extracellular synthesize silver nanoparticles from nitrate solutions using B. cereus, Bacillus subtilis, Staphylococcus aureus, E. coli, P. syringae and Streptococcus species have been reported to produce silver nanoparticle (Thu et al., 2013). Bacterial antibiotic resistance exhibited by completely different micro-organisms can be earned through numerous intrinsic or nonheritable mechanisms (Reidy et al., 2013). Intrinsic or 'natural' resistance is inherent to a bacterial species and involves inactivation of the antibiotic, cell permeability, the absence of the target or the presence of low-affinity targets and also the presence of effluence mechanisms. On the other hand, non-heritable mechanisms are genes embraced mutations in targeted by the transfer of resistance determinants borne on plasmids, the antibiotics, bacteriophages, transposons, and different mobile genetic materials. In general, this exchange is accomplished through the processes of conjugation (via plasmids) transduction (via bacteriophages), and conjugative transposons), and transformation (via incorporation into the body of body DNA, plasmids, and different DNAs from dying organisms). A really few novel antibiotics are discovered to treat these MDR organisms in past decades. famed life scientist Alexander Fleming aforesaid that "There is maybe no therapy drug to that in appropriate circumstances the bacteria can't react by in how deed fastness." thus there's high chance that the organism can also become resistance to freshly developed medication at later stage, additional these medications are extremely big-ticket. To the present finish nanoparticles are thought of to be sensible medicinal drug agents and should overcome the barrier of MDR because of their multifunctional mechanisms to intervene traditional cell practicality. Nanoparticles are shown to own the flexibility to anchor to the bacteria semipermeable membrane and afterwards penetrate it, thereby inflicting structural changes within the plasma membrane permeableness resulting in necrobiosis. The most economical friendly is Silver NPs (SNPs), it also have antimicrobial property compared with others because of their extraordinarily massive area that provides higher contact with bacteria. SNPs might target at the bacteria membrane, resulting in a dissipation of the nucleon driver that successively cause interference of organic process. Another mechanism concerned in microbicidal activity is that the generation of free radicals by the nanoparticles that have the flexibility to break the plasma membrane and create it porous which might ultimately result in necrobiosis. Metal nanoparticles have the affinity to act with sulphur and phosphorus containing biomaterials gift within the bacteria cell e.g., DNA bases. The metal nanoparticles will act on these soft bases and destroy the DNA which might result in necrobiosis. Nanoparticles are famed to modulate the bacteria signal transduction (Singh et al., 2016). The nanoparticles dephosphorylate the amide substrates on aminoalkanoic acid residues that result in signal transduction inhibition and inhibition of bacteria growth. It's conjointly shown that there can be a unharness of silver ions from conductor nanoparticles and these ions will act with the thiol of the many very important enzymes and inflicting disruption of cellular functions, inactivate them. NPs conjointly exert their medicinal drug activities either by collapsing the membrane potential and inhibiting the nucleotidease activities to decrease the ATP level and also the different is by inhibiting the fractional monetary unit of organelle from binding to tRNA. Many findings on antibacterial result of varied NPs and nanoconjugate systems are in short reviewed in preceding sections (Petra et al., 2018).
ROLE IN OVERCOMING ANTIBIOTIC RESISTANCE
The emergence of resistant's in conjunction with the limitation on the constant production of new effective antibacterial agents has resulted in inefficiency of current antibiotic medical aid with relevant risks on human health. The provision of latest antibacterial agents appeared to be a really advanced method in sight of the potential to provide new effective and safe medicine, additionally to the high production prices and also the time needed for approval of latest medicine that takes concerning 10-15 years (Singh et al., 2016;Natan et al., 2017). In 2016, several antibiotics were clinically tested for the market within the use of America (Petra et al., 2018). Sadly, however, within the last decades, linezolid was the sole approved antibiotic in conjunction with the recently discovered teixobactin (Nermin et al., 2020). Nanomedicine plays an important role in enhancing the effectiveness of existent medical drugs, by enhancing the chemical properties and stability of antibiotics, providing an opportunity of biofilm internalisation, prolongation of antibiotic unharness (Figure 3), additionally to the potential of targeted delivery to the location of infection and improved circulation with a resulting reduction of the connected facet effects compared to the corresponding free medicine (Natan et al., 2017).
Silver nanoparticles
Among metal nanoparticles Ag-NPs are extensively studied and used as effective antimicrobial agents. The bactericidal impact of silver nanoparticles on micro-organisms is incredibly well known; but the bactericidal mechanism isn't fully understood. Studies showed that Ag-NPs attack Gram negative bacteria by penetrating and anchoring the semipermeable membrane, and as a consequence, the leading structural amendment within the membrane morphology (Figure 4). This led to vital increase in membrane porosity and alters transport through the cell membrane leading to death (Singh et al., 2016). It's conjointly been planned that the Ag-NPs powerfully move with thiol teams of important enzymes and phosphoruscontaining
Action of silver nanoparticle on microbes
There are numerous theories on the action of silver nanoparticles on microbes to cause the microbiocidal impact. Silver nanoparticles have the flexibility to anchor on the bacteria semipermeable membrane and after penetrate it, thereby inflicting structural changes on the cell wall like porosity of the cell wall and death of the bacteria cell. There's formation of peats on the cell surface moreover as accumulation of nanoparticles on the cell surface. The formation of free radicals on the bacteria cell surface by the silver nanoparticles perhaps thought-about as another mechanism by that death of the bacteria cell might occur (Kim et al., 2007) (Figure 4).
No releasing nanoparticles
Nitric-oxide-releasing NPs (NO NPs) also possess broad spectrum antibacterial activity which might inhibit the expansion of the many antibiotic resistant and sensitive bacterium like K. pneumoniae, S. pyogenes, E. coli, E. faecalis and P. aeruginosa. NO is a hydrophilic natural gas and lipophilic, and is unstable in an element surroundings. Reactions of NO with element or superoxide impromptu turn out reactive atomic number 7 and element intermediates that are cyanogenic to the cell and act as antimicrobial species (Singh et al., 2016).
Dinitrogen oxide and dinitrosyl-iron
Complexes are generated. NO-associated macromolecule harm has been incontestable with peroxynitrite and gas (nitrogen dioxide). Peroxynitrite mediate macromolecule peroxidation of liposomes that contribute to the antimicrobial activities of NO (Jones et al., 2020). NO interactions with proteins involve reactive thiols, proteome teams, iron-sulphur clusters, phenolic resin or aromatic organic compound residues, tyrosyl radicals, or amines (Nermin et al., 2020). Peroxynitrite and NO•2 conjointly nonspecific ally oxidize proteins at a spread of websites. 73 Studies of NO-related toxicity have incontestable inactivation of enzymes containing Fe-S clusters (e.g., aconitase, NADH dehydrogenase, succinate dehydrogenase) suggesting that NO• (NO radical) may directly unharness iron from metalloenzymes and cause iron depletion. The antimicrobial and healing efficaciousness of sustained unharness gas nanoparticles are investigated against Methicillin-Resistant SA (MRSA) and Acinetobacter baumannii using a murine wound and soft (Singh et al., 2016).
METAL COMPOUND NANOPARTICLES
Metal Oxides Nanoparticles (NPs) like TiO 2 , CuO and ZnO are proverbial to exhibit smart antibacterial properties (Toolabi et al., 2013). One among the common properties they share is their photo catalytic activity because of wide band gap. The photo catalytic activity is attributed to generation of Reactive Oxygen Species (ROS). Copper (CuO) nanoparticles have shown activity against a variety of bacteria pathogens, as well as Methicillin-Resistant Staph Aureus (MRSA) and Escherichia, what is more, the power of CuO nanoparticles to cut back bacteria populations to zero was found to be increased within the presence of silver nanoparticles (Kadiyala et al., 2018).
DISCUSSION AND CONCLUSION
The wide unfold antibiotic resistance has place large pressure on pharmaceutical industries to search for new antimicrobial agents or modification of the existing antimicrobial agents. Nanoparticle has potential to alter the state of this affairs and forestall the unfold of drug resistance. Metallic nanoparticles of biological and chemical origin are shown to be potential agents in antibacterial treatment in previous studies. Nanoparticles not solely incontestable activity against MDR bacterium themselves, however conjointly showed potential for the event of synergistic mixtures that will increase the antibacterial impact of existing antibiotics. They helped to revive the antibacterial activity of previous generation antibiotics against that microbes have developed resistance. The review showcased the impact of varied styles of radiations together with nanoparticles enhancing the antibacterial efficaciousness. More work continues to be needed so as to elucidate the whole mechanism of action of nanoparticles as antibacterial, toxicity of nanoparticles in human and higher delivery of drug within human system by making use of nanodrug carriers. Nanoparticles in addition to either antibiotics and/or irradiation could offer a possible strategic remedy to combat MDR. | 2022-01-21T17:04:06.759Z | 2022-01-19T00:00:00.000 | {
"year": 2022,
"sha1": "6246351450204fcb7190dd2121ea9f502b7df141",
"oa_license": "CCBY",
"oa_url": "https://www.researchsquare.com/article/rs-887162/latest.pdf",
"oa_status": "GREEN",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "14b983d9af41e117d67177848c15b143481d5a6f",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": []
} |
250696634 | pes2o/s2orc | v3-fos-license | Pregnancy outcomes after SARS-CoV-2 infection by trimester: A large, population-based cohort study
Objectives Data regarding women infected with SARS-CoV-2 during early trimesters are scarce. We aimed to assess preterm birth (PTB) and small-for-gestational-age (SGA) rates in a large and unselected cohort by trimester at infection and overall. Design A retrospective cohort study including all women with a positive SARS-CoV-2 RT-PCR test during a non-ectopic singleton pregnancy between February 21st 2020 and July 2nd 2021 (N = 2753). Each infected woman was matched to a non-infected pregnant woman by age, last menstruation date, sector, and socioeconomic status. Methods Logistic regression was conducted to assess the risks of PTB and SGA including an interaction between group and trimester of infection. Multivariable models included underlying diseases, previous abortions and null parity. Subgroup analyses were conducted on symptomatic infected women and matched non-infected women. Results A total of 2753 /2789 (98.7%) eligible women that were infected during pregnancy could be matched, among them, 17.4% and 48.4% were infected during the first and third trimesters, respectively. While first and second trimester infections were not associated with PTB (p>0.8), third trimester infections and in particular after 34 weeks of gestation had a greater risk of PTB with adjusted ORs of 2.76 (95% CI 1.63–4.67) and 7.10 (95% CI 2.44–20.61), respectively. PTB risk was further heightened in symptomatic third trimester infections (OR = 4.28, 95% CI 1.94–9.25). SGA risk was comparable between study groups across all trimesters of infection. Pregnancy loss incidence was similar in both groups (adjusted OR = 1.16; 95% CI 0.90–1.50). Conclusion SARS-CoV-2 infection was associated with increased risk of PTB only among women infected during late pregnancy, particularly among symptomatic women.
Introduction
The current evidence on pregnancy outcomes in COVID-19 patients is limited to studies on women who tested positive for SARS-CoV-2 while hospitalized for delivery or pregnancy termination [1,2], with undetermined time of infection [3] or a small cohort size [2]. While few studies included women infected during different stages of pregnancy [4][5][6], pregnancy outcomes were reported for the cohort as a whole and not by trimester of infection, thus there is a lack of information regarding pregnancy outcomes in women infected during early stages of pregnancy. Timing of viral infection during fetal development may affect birth and other health outcomes [7]. Previous studies on influenza infection during pregnancy found adverse pregnancy outcomes including higher rates of preterm birth (PTB) and caesarian delivery in women infected during late pregnancy [8]. Other Coronaviruses such as severe acute respiratory syndrome coronavirus (SARS-CoV) and Middle East respiratory syndrome coronavirus (MERS-CoV) have been associated with excess risk for preterm delivery and spontaneous abortion [9]. Systematic reviews and meta-analyses of pregnancy outcomes in SARS-CoV-2 positive women reported that PTB and low birth weight (LBW) were more probable in pregnant woman with COVID-19 than pregnant women without COVID-19 [10,11], though less is known regarding the extent to which these associations vary by timing of infection during pregnancy. In this study, we aimed to assess PTB and small-for-gestational-age (SGA) rates in a large and unselected cohort following SARS-CoV-2 infection during pregnancy overall and by gestational age of infection. In addition, we explored pregnancy loss (PL) rates.
Setting
The study was conducted in Maccabi Healthcare Services (MHS), a 2.5 million members statemandated health fund in Israel. According to the Israeli National Health Insurance Act, MHS may not bar membership from any citizen who wishes to join it and therefore every section in the Israeli population is represented in MHS. The MHS computerized database includes extensive demographic data on each patient. SARS-CoV-2 polymerase chain reaction (PCR) tests from nasopharyngeal swabs [12] are highly accessible in Israel and offered to all citizens free of charge. The study utilized MHS pregnancy registry, previously described [13].
the Israel central bureau of statistics data [14]. Retrospective follow-up started at the time of the first positive SARS-CoV-2 test (index date) for the infected and their matched non-infected women.
Data
PTB was defined as less than 37 weeks of gestation. By definition, women infected at gestational age of 37 weeks could not sustain a preterm birth and were therefore excluded from PTB analyses along with their matched non-infected women. SGA was defined as infant body weight at birth less than the gender-specific 10th percentile for the gestational age [15]. LBW was defined as less than 2500 grams at birth. PL included induced abortions, spontaneous abortions up to 20 weeks of gestation and stillbirths from 20 weeks of gestation.
Statistical analyses
Descriptive statistics included chi-square for categorical covariates or Wilcoxon rank sum tests for continuous non-normal covariates. Standardized mean differences (SMD) were calculated as measures of effect size. An SMD value higher than 0.1 was considered a substantial difference. PTB and SGA were both primary outcomes. Logistic regressions were conducted to assess the crude odds ratios (OR) of the primary outcomes, including an interaction between group and trimester at index date. In addition, adjusted OR were assessed by including patient characteristics and underlying diseases. P-values were two-tailed and Bonferroni-corrected Pvalue < .025 was considered statistically significant (corrected for two primary outcomes). First, second and third trimesters were defined as 0-12, 13-26 and 27 or higher gestational weeks, respectively. Subgroup analysis was conducted on symptomatic women and their matched non-infected women. Hospital discharge records were extracted for further exploration of significantly different outcomes between infected and non-infected women.
PL rate was assessed by a nested case-control design. Up to 10 controls were selected for each PL case via incidence density sampling [16], i.e. selected without replacement from all women at risk at the time of PL occurrence. Univariate and multivariate conditional logistic regression (including patient characteristics and underlying diseases) were conducted to assess differences in PL rates between infected and non-infected women. Difference in PL rate was also assessed with a log-rank test and Kaplan Meier cumulative incidence plots. In this observational study, the study sample size was governed by the infection incidence, thus no formal power calculation was carried out. Statistical analysis was performed using SAS statistical software version 9.4.
Ethical statement
This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline [17]. The study was approved by the MHS's institutional review board. Informed consent was waived as identifying details were removed before the statistical analysis.
Results
We identified 43061 non-ectopic singleton-fetus pregnancies that ended during the study period, 2789 (6.5%) of which were positive for SARS-CoV-2 during pregnancy. Non-infected women were matched to 2753 (99%) infected women, of whom 48.4% were infected during their third trimester while only 17.4% were infected during their first trimester (Table 1). Symptomatic disease was reported for 84 women (17.6%) infected during their first trimester, 484 (51.3%) during their second and 622 (46.7%) during their third. Median age of women was 28 years (IQR 25-33), median SES level was 4 (IQR 3-6) and 92.7% of women were Jewish. Infected and non-infected women were well balanced on all matching variables ( Table 1). The median difference of last menstruation date between infected and matched non-infected women was 1 day (IQR 0-1).
Patient characteristics
The prevalence of cancer, diabetes, hypertension and cardiovascular diseases was low and similar in both groups (Table 1). There was a higher rate of null parity among the non-infected women (34.8%) compared to infected (28.0%). History of previous abortions was similar between the groups (SMD<0.1). A higher rate of non-infected women were vaccinated for influenza compared to infected women (36.3% and 32.5%, respectively) while no significant differences were observed between the groups in rates of pertussis vaccination, cytomegalovirus antibodies (CMV IgG)or gestational diabetes during pregnancy (Table 1).
Preterm birth and small-for-gestational-age
Women infected at gestational age of 37 weeks or more (N = 446) and their matched noninfected women were excluded from PTB analyses. Neonatal birthweight was available for 94.7% of women and was lower in infected women (3315 grams; IQR 3025-3610) compared to non-infected (3335 grams; IQR 3060-3636) though with a low SMD of 0.06. Median gestational age at birth was lower in infected (39 weeks; IQR 39-40 weeks) compared to noninfected (40 weeks; 39-40 weeks p<0.001). Rate of PTB was not significantly different in infected and non-infected women in the overall cohort when trimester was not accounted for ( Table 2). A significantly higher rate of PTB was observed in women infected during their third trimester (N = 51, 5.8%) compared to their matched non-infected women (N = 20, 2.3%) with an odds ratio of 2.76 (95% CI 1.63-4.67; Table 3), specifically in women infected after 34 weeks of gestation (OR 7.10, 95% CI 2.44-20.61; Fig 1). Hospital discharge records were available for 72% (N = 51) of women with PTB. The rate of induced labors was higher among infected women (37.8%) compared to non-infected women (14.3%), though the difference was not statistically significant (p = 0. 176). Among infected and non-infected women, 39.1% and 58.3% of spontaneous labors started with premature rupture of membranes, respectively (p = 0.279). Proportions of caesarean sections were similar (p = 0.711) among infected (48.7%) and noninfected women (42.9%). No difference between the groups was observed in SGA rates in the overall cohort (OR 0.98, 95% CI 0.78-1.24) or for women infected in specific trimesters ( Table 3, Fig 2)
Symptomatic COVID-19
Subgroup analyses were conducted on symptomatic women and their matched women. Adjusted OR of PTB in women infected during their third trimester compared to their controls was 4.28 (95% CI 1.94-9.41; Table 3). In contrast, no differences were observed for PTB in other trimesters or for SGA in the whole cohort or in specific trimesters (Table 3).
Pregnancy loss rate
PL incidence was similar in infected (3.9%) and non-infected women (3.8%, Fig 3, log rank p = 0.998). Medically induced abortions accounted for 17.3% of PL in infected women and 8.5% in non-infected women, while 10.6% and 15.1% stillbirth rates were observed among infected and non-infected women, respectively (p = 0.126). A subgroup of the cohort were included in a nested case-control study including all pregnancies that ended with a PL (n = 210) and matched controls (n = 1995). There was no difference in both crude (OR = 1.13; 95% CI 0.88-1.45) and adjusted (OR = 1.16; 95% CI 0.90-1.50) PL rates between infected and non-infected women.
Discussion
Study results indicate a substantially elevated risk of PTB in women infected with SARS-CoV-2 during the third trimester of pregnancy, but not among women infected during earlier Table 2 gestational weeks. These results concur with several prior studies in women with COVID-19 in late pregnancy or time of delivery [1,2]. Other studies did not observe any adverse pregnancy outcomes among these women, possibly due to small sample sizes [2,4]. Similar to previous findings [5,10,[18][19][20][21], the observed increased risk of PTB was particularly high among those infected during the third trimester and presented symptomatic COVID-19. In a metaanalysis of 18 high quality studies COVID-19 during pregnancy was related with a pooled OR of 1.82 to PTB. The lower risk estimates calculated in our analysis can be explained by the higher proportion (more than 50%) of infections during first and second trimester compared to previous studies. For example, in the largest cohort study that was included in the metalanalysis [22] 98% of infections occurred in the third trimester. Further breakdown of PTB indicated that the rate of induced labors was higher in women infected in the third trimester compared to their matched non-infected women, suggesting that clinicians may have been more incline to induce women infected with SARS-CoV-2 nearing the end of their pregnancy. Despite the higher rates of PTB among infected women, no difference in SGA rates was found between infected and their matched non-infected women regardless of trimester of infection or symptomatic status, consistent with current literature [18,23,24]. This result is reassuring that SARS-CoV-2 infection during pregnancy is unlikely to associate with intrauterine growth restriction. During previous outbreaks of other coronaviruses, an increased risk of miscarriage has been reported in pregnant women with SARS and MERS [25]. We did not observe any difference in PL rates among infected and matched non-infected women. Similar results were reported in a case-control study including 225 women for whom the cumulative incidence of infection did not differ between women with spontaneous abortion and women with ongoing pregnancy [26].
Pregnancy outcomes Category Infected n (%) Non-infected n (%) SMD P-Value
Our results indicate that gestational age at infection of SARS-CoV-2 plays a significant role in the pregnancy outcome. Women during their third trimester, specifically after 34 weeks of gestation, should practice social distancing and respiratory protection to reduce risk of adverse pregnancy outcomes [27] with an emphasis on PTB during 35-37 weeks of gestation. While no adverse outcomes were observed in women infected in the earlier trimesters, these parameters should not be regarded solely as reassurance for the impact of COVID-19 as they may not be sufficiently sensitive to reflect other potential differences in the mother or newborn [28]. Prior reports on COVID-19 [28] as well as Zika [29] virus infection indicated no association between birth week and infant weight while maternal and fetal complications were increased for infected patients.
To the best of our knowledge, this is the largest matched population-based cohort study within a nationally representative distribution on pregnancy outcomes in women infected with SARS-CoV-2 during different stages of pregnancy. Our cohort included women from the
PLOS ONE
Pregnancy outcomes post SARS-CoV-2 infection during pregnancy general population as well as women from the Arab community for whom different morbidity outcomes were observed [30] with lower risk for preterm birth [31] and women from diverse socioeconomic statuses [32]. Another strength of this study is that all women are tested for SARS-CoV-2 at hospitalization; therefore, all infected women at time of delivery were identified.
Our study has several limitations. Firstly, women without positive SARS-CoV-2 test results did not necessarily test negative, rather they did not have any positive result, although
PLOS ONE
Pregnancy outcomes post SARS-CoV-2 infection during pregnancy misclassification is minor due to the abundance of SARS-CoV-2 tests in Israel. Secondly, delivery mode was not fully captured in the MHS database and hospital discharge records were not available for the full cohort. Finally, our findings may have limited generalizability to countries populated with several races since there is evidence of disproportionate burden of COVID-19related outcomes among different races [33].
Conclusion
Women infected with SARS-CoV-2 during their third trimester are at higher risk of preterm birth as compared to matched non-infected women. No differences in pregnancy outcomes were observed between women infected during the first two trimesters of pregnancy and noninfected women. Results underline the importance of preventive measures taken against SARS-CoV-2 infection among pregnant women and their families. | 2022-07-21T06:16:22.938Z | 2022-07-20T00:00:00.000 | {
"year": 2022,
"sha1": "0c7a471db33febe7938437538384168607e839dd",
"oa_license": "CCBY",
"oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0270893&type=printable",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "053dffe76e4432af3212918322f6860d295f5de5",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
271166313 | pes2o/s2orc | v3-fos-license | Diagnostic Imaging Performance of Dual-Energy Computed Tomography Compared with Conventional Computed Tomography and Magnetic Resonance Imaging for Uterine Cervical Cancer
Objective This article evaluates the ability of low-energy (40 keV) virtual monoenergetic images (VMIs) in the local diagnosis of cervical cancer compared with that of conventional computed tomography (C-CT) and magnetic resonance imaging (MRI), using clinicopathologic staging as a reference. Methods This prospective study included 33 patients with pathologically confirmed cervical cancer who underwent dual-energy CT and MRI between 2021 and 2022. The contrast-to-noise ratio (CNR) of the tumor-to-myometrium was compared between C-CT and VMI. Additionally, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) for each local diagnostic parameter were compared between C-CT, VMI, and MRI. Interradiologist agreement was also assessed. Results The mean CNR was significantly higher on VMI ( p = 0.002). No significant difference in AUC was found between C-CT and VMI for all local diagnostic parameters, and the specificity of VMI was often significantly less than that of MRI. For parametrial invasion, mean sensitivity, specificity, and AUC for C-CT, VMI, and MRI were 0.81, 0.99, 0.93; 0.64, 0.35, 0.79; and 0.73, 0.67, 0.86, respectively, and MRI had significantly higher specificity and AUC than that of VMI ( p = 0.013 and 0.008, respectively). Interradiologist agreement was higher for VMI than C-CT and for MRI than VMI. Conclusion The CNR of VMI was significantly higher than C-CT and interradiologist agreement was better than with C-CT; however, the overall diagnostic performance of VMI did not significantly differ from C-CT and was inferior to MRI. VMI was characterized by low specificity, which should be understood and used for reading.
Introduction
Cervical cancer is the fourth most common cancer in women, affecting approximately 570,000 individuals and causing 311,000 deaths worldwide in 2018. 1 Patient numbers in developed countries are declining while in developing countries are rapidly increasing. 1 With the revision of the International Federation of Gynecology and Obstetrics (FIGO) 2018, cervical cancer is now staged comprehensively, considering not only physical findings but also imaging. 2,3agnetic resonance imaging (MRI) has been used for local diagnosis due to its superior tissue resolution; however, MRI availability is often limited. 4If computed tomography (CT) can be used for local diagnosis instead of MRI, simultaneous assessment of metastasis becomes possible, leading to shortened treatment time and cost reduction.Low-energy virtual monoenergetic images (VMIs) obtained from dual-energy CT (DECT) improve iodine enhancement in parenchymal tissue and can aid in the detection of various tumors, [5][6][7][8] and VMI of 40 keV is reported to have the highest contrast and detection sensitivity.The objective of this study was to assess the effectiveness of 40 keV VMI in the local diagnosis of cervical cancer compared with conventional CT (C-CT) and MRI.The clinicopathologic staging was used as the reference standard for comparison.
Study Design
This retrospective study was approved by the Institutional Review Board of our institution, and the need for written informed consent was waived (approval number: R03-117).The inclusion criteria were as follows: (1) histopathologically diagnosed uterine cervical cancer by biopsy and (2) examination using CT and MRI including T1-weighed images (T1WIs), T2-weighted images (T2WI), and diffusion-weighted images (DWIs) performed in our hospital between June 2021 and October 2022.
Patients were excluded if they (1) were not examined by contrast-enhanced DECT or (2) examined by contrast-enhanced DECT with reduced iodine dose.Staging of the uterine cancer was determined by pathology if surgery was performed or by clinical examination (i.e., pelvic examination, cystoscopy, and colposcopy) concerning imaging findings if surgery was not performed.
Image Protocol
All CT examinations were performed using an IQon Spectral CT scanner (Philips Medical Systems, Amsterdam, Netherlands).Scanning was performed with the following acquisition parameters: tube voltage, 120 kVp; helical pitch, 0.798; detector collimation, 64 Â 0.625; and rotation time, 0.5 second.Tube current was modulated by automatic exposure control (Dose Right Index, 24-25; Philips Healthcare).A 500 mg/kg dose of contrast medium was injected via a peripheral vein of the upper extremity over 50 seconds using a power injector, followed by a 50-mL saline flush.All scans were captured in a craniocaudal direction from the lung apexes or the hepatic dome to the pubis 80 seconds after intravenous administration of the contrast medium.For image evaluation, images from above the aortic bifurcation to the pubis were selected and reconstructed using C-CT and 40 keV VMI with a slice thickness of 2.0 mm and a reconstruction interval of 2.0 mm.Images of the axial and sagittal planes along the uterine axis were also reconstructed using C-CT and VMI with a slice thickness of 2.0 mm and reconstruction interval of 2.0 mm.Optimal window levels and widths were set at 35 and 350 Hounsfield units (HU) for C-CT, and 140 and 680 HU for VMI.
MRI was performed using 3T or 1.5T equipment (Ingenia, Achieva; Philips Medical Systems, Netherlands).The protocol included T1WI, T2WI, DWI, with a b-value of 0 and 1,000, and apparent water diffusion coefficient maps.Further details of these parameters are listed in ►Table 1.
Quantitative Image Analysis
Quantitative evaluation of the myometrium and cervical cancer on C-CT and VMI was performed using region-ofinterest (ROI) analysis by two board-certified radiologists with 6 years of experience (S.S. and M.Y.).The average of the two measurements was used for the analysis.The mean CT number of the myometrium was measured by placing a circular ROI on the myometrium, carefully avoiding myomas, adenomyomas, and visible vessels.For cervical cancer, the ROI was set to be as large as possible without including the surrounding area.Image noise was quantified as the standard deviation (SD) of the CT number of the anterior abdominal wall fat tissue.The signal-to-noise ratio of the myometrium was calculated by dividing the CT number of the myometrium by the image noise.The contrast-to-noise ratio (CNR) of the tumor-to-myometrium was calculated using the following equation: CNR ¼ | (mean HU of myometrium -HU of cervical cancer)/image noise |. 9 Since the enhancing effect of cervical cancer can be weaker or stronger than that of the myometrium, absolute values were used for CNR.Missing values were set in the following conditions: (1) no tumor could be identified, or (2) the tumor replaced the entire uterus, and the myometrium could not be identified.
Qualitative Image Analysis
This study included four board-certified experienced radiologists (K.M., T.S.,T.I., and T.A., and F.F.) with 29, 19, 12, and 7 years of experience in abdominal radiology; they independently reviewed C-CT, VMI, and MRI in random order.They scored each image by assigning the confidence levels for lateral invasion on a 5-point scale (1, tumor not visualized; 2, no parametrial involvement; 3, parametrial involvement; 4, extension to the pelvic wall or hydronephrosis; and 5, bladder or rectum involvement) and the conspicuity of vaginal wall involvement on a 5-point scale (1, tumor not visualized; 2, absent; 3, equivocal; 4, upper two-third involvement; and 5, lower one-third involvement).Regarding the presence of a tumor, a score of 1 for lateral invasion indicated the absence of a tumor, while any score other than 1 was considered as the presence of a tumor.For vaginal involvement, a confidence score of 4 or 5 was considered definitive.The radiologists were blinded to the pathologic and clinical findings, and there was a minimum of 2 weeks between each image set interpretation.
The criteria for local diagnosis on T2WI were as follows: parametrial invasion is excluded if the outer margin of the cervical stroma of low intensity is preserved; parametrial invasion is diagnosed when the following findings are detected: spiculated tumor-to-parametrial interface, tumor nodule in the parametrium, and/or tumor encasement of parametrial vessels.The presence of pelvic wall extension is indicated by tumor extension into the iliac vasculature or muscles including the internal obturator, piriformis, or levator ani.A dilated ureter obstructed by the tumor is also considered a pelvic wall invasion. 10Bladder/rectal mucosal involvement is diagnosed when the tumor disrupts the low-signal-intensity bladder or rectal wall and extends into the mucosa or lumen. 3Local staging on imaging was performed according to FIGO 2018: stage I as confined to the cervix; stage II as invasion beyond the uterus without extension to the lower one-third of the vagina or pelvic wall; stage III as invasion of the lower one-third of the vagina/hydronephrosis/nonfunctioning kidney/pelvic wall; and stage IVA as extension beyond the true pelvis or involvement of the bladder or rectum mucosa. 10
Statistical Analysis
For quantitative analysis, the mean CNR and SD were calculated, and the Wilcoxon signed-rank test was used to evaluate the difference in CNR between C-CT and VMI.
For qualitative analysis, based on the results of the imaging evaluation, the sensitivity, specificity, and accuracy were calculated for the detection of tumor presence, parametrial invasion, pelvic wall extension, bladder/rectum involvement, upper two-third of the vaginal wall involvement, and lower one-third of the vaginal wall involvement; these were calculated by selecting two options of presence or absence.Receiver operating characteristic (ROC) analysis was also performed to evaluate diagnostic performance.The area under the ROC curve (AUC) interpreted the diagnostic value as follows: 0 to 0.70, poor; 0.70 to 0.90, moderate; and 0.90 to 1.00, high. 11We estimated 95% confidence intervals and performed multiple comparisons using the Friedman test with Bonferroni correction to compare the sensitivity, specificity, accuracy, and AUC.The concordance rate between each reader's C-CT, VMI, and MRI local stage was also calculated.Interobserver agreement was assessed using the κ-statistic.The κ-statistic interpreted the agreement as follows: < 0, none; 0 to 0.20, slight; 0.21 to 0.40, fair; 0.41 to 0.60, moderate; 0.61 to 0.80, substantial; and 0.81 to 1.00, almost perfect. 12All statistical analyses were performed using the Statistical Package for the Social Sciences software (SPSS Statistics 28.0; IBM, New York, New York, United States).Statistical significance was set at p < 0.05.
Results
A total of 33 women (mean age, 56 years; age range, 36-85 years) were evaluated across all the data sets.The flowchart of the study design is shown in ►Fig. 1. ►Table 2 shows the patients' characteristics, pathological type, the clinical FIGO 2018 stage, and the clinicopathological invasion.Among the 33 patients, 22 patients were treated with concurrent chemoradiotherapy (CCRT), and the stage was determined clinically; meanwhile, 11 patients underwent a radical hysterectomy, and the stage was determined pathologically.The smallest lesion was 2 mm in depth, as measured by pathology, and the largest lesion was 13 cm in the greatest dimension, as measured by MRI.
Quantitative Image Analysis
In the ROI analysis, three cases with C-CT and two cases with VMI had missing values because cervical cancer could not be identified.In addition, two cases with C-CT and one case with VMI had missing values because the myometrium could not be separated from the tumor.The mean tumor-to-myometrium CNR was 1.57 AE 1.43 on C-CT and 2.73 AE 2.41 on VMI.The mean tumor-to-myometrium CNR was significantly higher on VMI (p ¼ 0.002).
Qualitative Image Analysis
All readers rated the image quality of C-CT as "acceptable," "good," or "excellent" and that of VMI and MRI as "good" or "excellent." The results in ►Table 3 show variability in sensitivity, specificity, accuracy, and AUC among the different imaging modalities and radiologists.Only cases in which cervical cancer was present were included; therefore, the specificity, accuracy, and AUC were not calculated for the presence of the tumor.In tumor detection, VMI was more sensitive than C-CT in all but one reader and equal to or more sensitive than MRI in all readers.Similarly, in all the other diagnostic parameters, VMI had equal to higher sensitivity than C-CT and MRI, and equal to lower specificity than C-CT and MRI in almost all readers.MRI generally exhibited higher specificity, accuracy, and AUC compared with C-CT and VMI in detecting parametrial invasion and pelvic wall extension.Its sensitivity was higher than that of C-CT, while being comparable to VMI.Additionally, in detecting vaginal wall invasion, even MRI tended to have lower sensitivity and AUC.
►Table 4 shows readers' mean values of sensitivity, specificity, accuracy, and AUC of each local diagnostic parameter and presents the statistical significance of these differences.Among the parameters, VMI had significantly higher sensitivity than that of C-CT and MRI for the lower one-third vaginal involvement, and VMI was significantly less specific than C-CT for all parameters except bladder/rectum involvement and significantly less specific than MRI for the parametrial invasion and upper two-third vaginal invasion.AUC comparisons showed that MRI was significantly higher than that of VMI for detecting parametrial invasion and pelvic wall extension; however, the other parameters were not significantly different.For parametrial invasion, the most important factor in determining treatment, the positive predictive value (PPV) of each reader was 0.68 to 0.91 for C-CT, 0.67 to 0.77 for VMI, and 0.83 to 0.94 for MRI, and the false positive value (FPV) of each reader was 0.08 to 0.69 for C-CT, 0.46 to 0.77 for VMI, and 0.08 to 0.31 for MRI with lowest for MRI, followed by C-CT and VMI.MRI generally exhibited higher sensitivity, specificity, accuracy, and AUC than C-CT and VMI; however, its diagnostic ability for vaginal wall involvement was inferior to that of lateral extension.Staging accuracy by the four readers was 61, 67, 70, and 39% for C-CT; 55, 70, 64, and 48% for VMI; and 64, 85, 79, and 54% for MRI, respectively.
►Table 5 shows the interobserver agreement among radiologists in local diagnostic parameters; the interradiologist κvalues for each modality widely ranged, with generally higher agreement for VMI than for C-CT and MRI than for VMI.
►Figs.2-4 show the C-CT, VMI, and MRI images of IB2, IIA1, and IIIB cervical cancer, respectively.VMI provides better contrast between the uterus and tumor than C-CT, and the tumor is more clearly defined.
Discussion
C-CT, VMI, and MRI were compared to assess the local diagnostic performance of cervical cancer.There was no significant difference in diagnostic performance between C-CT and VMI, with VMI characterized by higher sensitivity, lower specificity, and higher interreader agreement than C-CT.
DECT scanners use more than one peak energy to rapidly acquire images and visualize how tissues and materials preoperative planning.Lv et al reported that the lesion detectability and conspicuity of small hepatocellular carcinoma could be improved by selecting the optimal energy level (40-70 keV) for monochromatic imaging. 6De Cecco et al reported that the diagnostic accuracy in detecting hypervascular liver lesions is improved with 40 keV VMI. 13 Nagayama et al demonstrated that VMI yielded significantly better image quality in multiphasic pancreatic CT than conventional polyenergetic images.In each enhancement phase, 40 keV VMI provided the best quality for the evaluation of pancreatic duct cancer owing to its high pancreastumor contrast and vascular opacification without a relevant increase in image noise. 7For uterus tumors, Rizzo et al evaluated deep (> 50%) myometrial invasion using DECT and transvaginal ultrasound in patients with endometrial cancer, and reported that low keV VMI showed higher specificity and accuracy than that of ultrasound. 14To the best of our knowledge, there have been no DECT studies of uterine cervical cancer, and including other tumors, few studies have evaluated the utility of DECT in assessing the surrounding invasion of tumors.For the staging of cervical cancer, the role of imaging is to evaluate invasion to the parametrium, pelvic wall, and adjacent organs and the assessment of nodal involvement and distant metastasis.The strength of MRI in diagnosing cervical cancer is the high negative predictive value (NPV) of 95% in detecting parametrial invasion, 15,16 and Manganaro et al have documented a 100% NPV of MRI for bladder or rectal invasion, suggesting that MRI can obviate invasive procedures, such as cystoscopy, proctoscopy, and sigmoidoscopy. 3he reported accuracy for the parametrial invasion was 70 to 80% on CT and 87 to 92% on MRI in the 1990s. 17,189][20] The overall staging accuracy of MRI reported between 1988 and 2003 was 76 to 86%, higher than that of CT at 53 to 69%. 15,17,18,21n recent years, there have been few reports on the pathological image correlation beyond stage II in cervical cancer cases, mainly due to the fact that stage IB3 (tumor larger than 4 cm) and higher are typically managed with CCRT.A multicenter study conducted in 2005, focusing on early-stage invasive cervical cancer, found that the accuracy and specificity of both CT and MRI in assessing parametrial involvement and overall staging were quite similar; however, the accuracy of both modalities was relatively low, while the specificity was high. 20A meta-analysis of the literature between 2012 and 2016 has revealed that the sensitivity and specificity of MRI in detecting parametrial invasion, using radical hysterectomy as the reference standard, were 76 and 94%, respectively. 21Similar to our study, Yu et al using clinicopathologic staging as a reference and covering all stages, reported that the accuracy of stage classification of cervical cancer in a single facility from 2010 to 2015 was 80% for MRI and 73% for CT, showing a significant difference. 22As the majority of studies indicate, MRI has consistently shown significantly higher accuracy compared with CT.
Conventionally, low-energy images have high noise and contrast, whereas high-energy images have low noise and contrast.However, in our study, by setting the optimal window width and window level during reconstruction, all readers rated the image quality of VMI as good or excellent, and the noise was less noticeable than that of C-CT, similar to a previous report. 7The diagnostic performance of VMI was not significantly different from that of C-CT in all the diagnostic parameters and was inferior to MRI in the parametrial invasion and the pelvic extension; however, in VMI, the tumor became more distinct due to increased contrast between the tumor and background with a significantly higher CNR than C-CT, leading to improved diagnostic sensitivity, actually, with onethird vaginal involvement, VMI was significantly more sensitive than C-CT.It was also believed that this facilitated easier interpretation, increasing readers' agreement rates.Although On C-CT, only one reader misdiagnosed as no parametrial invasion, and all the readers correctly diagnosed vaginal invasion.On VMI, half of the readers overdiagnosed with bladder and rectal involvement, as well as the lower one-third vaginal involvement.On magnetic resonance imaging (MRI), all readers correctly diagnosed lateral invasion, but only one reader correctly diagnosed vaginal invasion.
not verified in this study, there may have been differences in diagnostic confidence and reading time.On the other hand, the specificity of VMI was significantly lower than that of C-CT and MRI, and VMI had a lower PPV.This is due to the higher FPV of VMI because of the accentuated heterogeneity of the cervical stroma, myometrium, and vaginal wall enhancement.Additionally, the current reading method was developed using C-CT, which was a disadvantage for VMI.Radiologists usually use MRI, not CT, for diagnosing cervical cancer.The FIGO clinical stage often depends on MRI findings, which puts CT at a disadvantage.Previous reports have shown an accuracy of 74 to 93% for vaginal invasion on MRI, 15,23 but in the present study, the diagnostic performance of MRI for vaginal wall invasion was lower than for lateral invasion.This could be improved by adding the sagittal section of DWI, 24 vaginal opacification, 21 or dynamic contracted-enhanced MRI which remains an option in the European Union Special Representatives guidelines.In dynamic contrast-enhanced MRI, cervical cancer typically shows early enhancement with hyperintensity relative to normal cervical stroma in the arterial phase making tumor detection easier. 23This effect is particularly considered useful in detecting small tumors. 25However, the role of imaging in evaluating vaginal involvement may be limited because it is usually clinically assessed.In the case of an exophytic tumor, the vaginal fornix adjacent to the mass appears thin because they are stretched by the tumor, making it difficult to distinguish the thin vaginal fornix from the adjacent tumor; additionally, 39% of the cases in this study had superficial spreading detected only using the Schiller test, which is considered difficult to capture on imaging.The overall staging accuracy in our study was 39 to 70% for C-CT, 48 to 70% for VMI, and 54 to 85% for MRI, which is generally consistent with previous studies, although there are large individual differences.
This study had several limitations.First, the number of patients was small.Second, CCRT is the standard treatment for stage IB3 or higher cervical cancer; therefore, the reference standard of parametrial, pelvic wall, and vaginal invasion was based on clinical examination concerning imaging findings in two-thirds of the cases.Third, because only cervical cancer cases were included in the study, a blinded evaluation of the tumor detection power was not possible.Finally, only 40 keV VMI was used because it was considered to have the highest contrast in previous reports; other lowenergy VMIs were not evaluated.
In conclusion, there was no significant difference in diagnostic performance between C-CT and VMI in the local diagnosis of cervical cancer.However, virtual monoenergetic imaging demonstrated a significantly higher tumor-to-myometrial contrast ratio, higher sensitivity, and greater interreader agreement.Conversely, these images were associated with lower specificity.These characteristics should be understood and utilized in CT reading and experience should be accumulated.
interact with X-ray beams of different energies.A single DECT acquisition can produce several different image data sets, such as VMIs and material-specific iodine images; these data sets provide superior contrast enhancement and reduce artifacts.Low-energy VMIs (40-55 keV) increase iodine density as they approach the K-edge of iodine (33.2 keV).Im-proved iodine prominence in low-energy VMI can be useful for tumor detection and characterization, and the use of VMI at the desired energy level (40-140 keV) improves the lesionto-background contrast and quality of vascular imaging for
Fig. 2 A
Fig. 2 A 73-year-old female patient with cervical cancer (adenocarcinoma), pathologically diagnosed with International Federation of Gynecology and Obstetrics (FIGO) 2018 IB2.(A, B) Conventional computed tomography (C-CT) axial, sagittal images.(C, D) Virtual monoenergetic image (VMI) axial, sagittal images.(E, F) T2-weighted image (T2WI), sagittal images.C-CT shows a tumor with strong enhancement confined to the cervix (A, B; arrows).The contrast is accentuated at VMI (C, D; arrows).On T2WI, the tumor shows high signal intensities (E, F; arrows).On C-CT, one reader could not detect the tumor.On VMI, all readers made the correct diagnosis.On magnetic resonance imaging (MRI), three readers overdiagnosed as extending to the pelvic wall with involvement of the vaginal wall (E; arrowhead, which is considered a combination of vascular and peritoneal thickening).
Fig. 3 A
Fig. 3 A 56-year-old female patient with cervical cancer (squamous cell carcinoma), pathologically diagnosed with International Federation of Gynecology and Obstetrics (FIGO) 2018 IIA1.(A, B) Conventional computed tomography (C-CT) axial, sagittal images.(C, D) Virtual monoenergetic image (VMI) axial, sagittal images.(E, F) T2-weighted image (T2WI) axial, sagittal images.The tumor is strongly enhanced, predominantly at the margins on C-CT (A, B; arrows).On VMI, the tumor enhancement effect is stronger, and its margins are more clearly defined (C, D; arrows).The preserved outer rim of cervical stroma onT2WI indicates no parametrial invasion (E; arrow), and the tumor reaches the upper edge of the anterior vaginal wall (F; arrowhead).On C-CT, one reader failed to detect the tumor.On VMI, three readers overdiagnosed with extension to the pelvic wall, but three readers correctly diagnosed the upper two-third vaginal involvement.On magnetic resonance imaging (MRI), only half of the readers correctly diagnosed lateral and vaginal invasion.
Fig. 4 A
Fig. 4 A 38-year-old female patient with cervical cancer (squamous cell carcinoma), clinically diagnosed with International Federation of Gynecology and Obstetrics (FIGO) 2018 IIIB.(A, B) Conventional computed tomography (C-CT) axial, sagittal images.(C, D) Virtual monoenergetic image (VMI) axial, sagittal images.(E, F) T2-weighted image (T2WI) axial, sagittal images.The tumor shows relative hypoattenuation with strongly enhanced margins (A-D; arrows), predominantly reaching the right pelvic wall (A, C; arrowheads) and invading the upper two-third vaginal wall (B, D; black arrowheads) on CT.T2WI showing spiculated tumor-to-parametrial interface reaching the iliac vasculature (E; arrowhead), and thickening of the anterior vaginal wall (F; black arrowhead).On C-CT, only one reader misdiagnosed as no parametrial invasion, and all the readers correctly diagnosed vaginal invasion.On VMI, half of the readers overdiagnosed with bladder and rectal involvement, as well as the lower one-third vaginal involvement.On magnetic resonance imaging (MRI), all readers correctly diagnosed lateral invasion, but only one reader correctly diagnosed vaginal invasion.
Table 1
Acquisition parameters of MRI
Table 2
Characteristics of patients and lesions Abbreviation: FIGO, International Federation of Gynecology and Obstetrics.
Table 4
Sensitivity, specificity, accuracy, and AUC for local diagnosis parameters in C-CT, VMI, and MRI Abbreviations: AUC, the area under the receiver operating characteristic curve; C-CT, conventional computed tomography; MRI, magnetic resonance imaging; VMI, virtual monoenergetic image.a p < 0.05.
Table 5
Interobserver agreement of local diagnosis parameters between the radiologists Abbreviations: C-CT, conventional computed tomography; MRI, magnetic resonance imaging; VMI, virtual monoenergetic images.Note: Parentheses indicate the mean. | 2024-07-15T15:14:57.691Z | 2024-07-01T00:00:00.000 | {
"year": 2024,
"sha1": "228a6cebed4bc98833e7e1ec3e5162229ee17fb8",
"oa_license": "CCBYNCND",
"oa_url": "http://www.thieme-connect.de/products/ejournals/pdf/10.1055/s-0044-1787780.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "cf6dbf33ef34d4c7d1e45c598abe863567dda683",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
220713969 | pes2o/s2orc | v3-fos-license | Physical fitness assessment in Goalball: A scoping review of the literature
Background Goalball is a Paralympic sport for visually impaired athletes. Although it is widely practiced, a great variety of tests are adopted to evaluate athletes' physical fitness. Therefore, the objective was to identify the physical fitness tests adopted in this sport to find the common aspects between them and, eventually, to propose a standard operating procedure. Methods The Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR) guidelines were adopted. The studies were extracted from PubMed, Web of Science, and Scopus. A selection process by title, abstract, and full-text, according to the inclusion and exclusion criteria, was performed. The results were discussed with narrative synthesis. Results A total of 7 papers and 222 participants were included. A wide variety of tests were adopted and the Brockport Physical Fitness Test (BPFT) was the only battery included to evaluate general athletes' well-being. Conclusions Although few literature exists on Goalball, the BPFT could be the battery for evaluating Goalball athletes though the test battery should be standardized to the characteristics of this sport.
Introduction
Goalball is a Paralympic sport created in 1946 by Hanz Lorenzen and Sepp Reindle to relieve blind war veterans during their rehabilitation. This sport was included in the Paralympics Games program in 1976 becoming popular worldwide over the years. Visual impaired (VI) Goalball athletes present several peculiar features in physical, psychological, and cognitive spheres. The first important effect of practicing this sport is the reduction of body fat mass and body mass index (BMI) (Caliskan et al., 2011;Karakaya et al., 2009). A second, but not least, aspect is the positive effect of this sport on health-related physical fitness characteristics such as cardiorespiratory endurance, muscular strength, muscular endurance, and flexibility (Caspersen et al., 1985;Colak et al., 2004;Karakaya et al., 2009). Positive effects are also on postural control capacity (Colak et al., 2004;Karakaya et al., 2009), and on auditory reaction speed and hearing duration compared to VI sedentary people (Yildirim et al., 2013). All the positive effects of Goalball are important because young VI people show poor neuro-psycho-motor and perceptual developments (Yildirim et al., 2013), lower cardiovascular and muscular endurance, poor flexibility and postural control than people with normal vision (Skaggs and Hopper, 1996).
Goalball is a widely practiced sport by in different levels ranging from school and leisure time (Caliskan, 2010;Caliskan et al., 2011;Colak et al., 2004;Furtado et al., 2016;Karakaya et al., 2009) to the national Paralympics level setting (Akino glu and Kocahan, 2018;Alves et al., 2018;Bednarczuk et al., 2017;Goulart-Siqueira et al., 2019). Regardless of the sporting level, the evaluation of physical performance is fundamental for monitoring both the athletes' health status and the team's preparation for the competition. Lieberman and Mc Hugh (Lieberman and McHugh, 2001), due to the wide variety of tests adopted, have tried to standardize the health-related fitness evaluation in VI children. The authors (Lieberman and McHugh, 2001) proposed to adopt items from the Fitnessgramm health-related fitness test, such as the 1-mile walk/run test for cardiovascular endurance assessment, sit-ups and push-ups for muscular strength and endurance, sit and reach test, back extension test, and shoulder stretch test for flexibility. A similar attempt to create a standard procedure to evaluate youth with mental and physical disabilities, and therefore also for VI people, was performed by Winnick and Short in 2009, who developed the Brockport Physical Fitness test (BPFT) manual (Winnick and Short, 2009). The test battery consists of four to six test items to evaluate aerobic capacity (1-mile walk/run test), body composition (skinfold thickness of triceps plus calf) and musculoskeletal functioning (curl-up, trunk lift, push-up and shoulder stretch tests). Unfortunately, the tests batteries proposed by the literature (Lieberman and McHugh, 2001;Winnick and Short, 2009) are specific for a general population with disabilities and not for Goalball athletes. Considering the importance to have updated sport-specific guidelines, to the best of our knowledge, there are no test batteries for the Goalball athletes' evaluation. For the reason described above, it was important to review the literature to understand if a standard operating procedure (SOP) guidelines clearly described, has been created for the Goalball athletes' evaluation. Hence, through a review of the literature, the objectives of the present manuscript were: (i) to study the tests adopted to evaluate VI Goalball athletes; (ii) to find the common aspects between the testing procedure; (iii) to propose, eventually, a SOP for the evaluation of Goalball athletes and practitioners.
Materials and methods
The present scoping review of literature partially adopted the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR) checklist and explanation (Tricco et al., 2018). The manuscript was not previously registered on PROSPERO or published before even if the protocol was written down before the beginning of the work.
Eligibility criteria
Research articles that met the inclusion and exclusion criteria were selected for review and PICO-S (Population, intervention, comparison, outcomes, and study design) points were followed.
Studies were included if participants were VI Goalball athletes (B1, B2, and B3). All kinds of interventions and comparisons (clinical trials, randomization, observational studies, descriptive studies, and longitudinal) were included if the Goalball was evaluated or adopted to improve participants' physical fitness. The language adopted in the studies was English, regardless of the country of origin.
Studies were excluded if papers were reviews, meta-analyses, abstracts, statements, opinion pieces, citations from scientific conferences, commentaries, editorials, book reviews, books, letters and non-peer reviewed journal articles.
Information sources
Data were sourced from PubMed, Web of Science and Scopus databases, and the screening was carried out between January, 1 st 2000 and March, 4 th 2020. Terms of group 1 ("goalball") and group 2 ("physical fitness", "sports physiology", "performance analysis") were combined with the Boolean indicator "AND" or "OR". "Goalball AND (physical fitness OR sports physiology OR performance analysis)" is an example of a string created and adopted during the systematic literature search.
Data selection and management
Data selection followed a two-step process in which, during the first step, duplicates were deleted through the program EndNote X8 (EndNote version X8; Thompson Reuters, New York, USA), while a manual selection to identify the appropriate manuscript was performed in the second step. The manual selection involved two examiners and it was developed in a three-step process during which the inclusion and exclusion criteria were applied on the title, abstract and full-text. In case of disagreement between the two examiners, a third examiner considered the work and after discussion, it was decided to include or exclude it. This process was presented through a PRISMA flow diagram.
Data collection and synthesis
A Microsoft Excel spreadsheet (Microsoft Corp; Redmond, Washington, USA) was created to extract the following information: first author and year of publication, sample size and gender, participants' age (range, mean and standard deviation), and tests adopted. The health-related components were categorized into the following seven: (i) muscular strength; (ii) aerobic capacity; (iii) postural control; (iv) range of motion; (v) flexibility; (vi) percentage body fat; (vii) battery adopted. The information were extracted from any part of the study.
All data were summarized using descriptive tables and graphs and analysed through a narrative synthesis.
Results
From 71 studies selected after the primary search, a total of 7 papers were included after duplicates removal and eligibility criteria screening (Figure 1).
A total of 222 participants were included, 64 of whom were female and 158 were male. The age ranged between 10 and 42 years. People included were athletes and active or sedentary Goalball practitioners. Visual impairment of the Goalball athletes included ranged from B1 to B3 (Table 1).
Studies characteristics
The seven health-related components considered are analysed one by one below. A summary and details of each health-related component are provided in Table 2. Body fat percentage was evaluated through the skinfold measurements (Akino glu and Kocahan, 2018) in seven (one study), five (one study), or two sites (two studies). Upper limbs strength was measured with the handgrip test (two studies) while the lower limb strength was evaluated with vertical jump tests (three studies) and through isometric devices (two studies). Other tests adopted to evaluate muscular strength were the push-up and the curl-up test (two and one study, respectively). The aerobic capacities were evaluated through field and laboratory tests. Among the laboratory tests, the treadmill incremental test to evaluate the maximal VO 2 consumption was adopted in one paper. Among the field tests, authors adopted the 1-mile walk/run test (two studies) and the Yo-Yo IR1 test (one study). The flexibility was evaluated using the sit and reach test (one study), the back-saver sit and reach test (one study), and the trunk lift test (two studies). The mobility was measured by two research groups with the shoulder stretch test for the shoulders, and, moreover, using the evaluation of the range of motion for shoulders, elbows and wrists (one study). The postural control capacity was valued through laboratory tests (two studies) and field tests such as the Flamingo Balance test (one study).
The only physical fitness test battery adopted was the BPFT (two studies).
Discussion
The results of the present scoping review show that the evaluation of Goalball athletes does not present a standardized test battery, and a wide variety of tests is proposed highlighting the lack of a SOP. Furthermore, a second important finding is the poor literature on this topic, although the health-related physical fitness assessment should be fundamental in individuals with VI.
Considering the lack of a SOP in the literature, an evaluation of the seven health-related tests found in the included studies was performed. The first point to highlight is the suggestion of the BPFT as the basis of the SOP because this test battery was the only one adopted in Goalball athletes (Furtado et al., 2016;Karakaya et al., 2009). A second important consideration is to decide to include only field tests. Indeed, although laboratory tests have more reliable results, field tests are generally easier, faster, and cheaper to administer (Heyward, 1991). These aspects are fundamental especially for fitness components evaluation in population-based studies and in school settings (Artero et al., 2011).
Regarding the tests included in the BPFT, a point by point discussion is provided. Indeed, although skinfold thickness measurement with seven different sites (Goulart-Siqueira et al., 2019) or five (Caliskan et al., 2011) gives more reliable values, this evaluation requires time to be obtained, reducing the test feasibility. For this reason, the two-points skinfold thickness (calf and triceps) adopted by two authors (Furtado et al., 2016;Karakaya et al., 2009) is a faster and feasible solution to obtain information regarding the participants' body fat status. To measure flexibility, the sit and reach test (Colak et al., 2004) is proposed because it presents validity for the hamstring extensibility assessment (Mayorga-Vega et al., 2014). For aerobic and anaerobic capacities, considering the difficulties in the locomotor tasks where changes of direction are required, the 1-mile walk/run test (Furtado et al., 2016;Karakaya et al., 2009) is suggested compared to the Yo-Yo intermittent test where a change of direction is required making this last test not easy to administer to VI people. Muscular strength and endurance were evaluated through different tests (see Table 2). Consequently, a limited number of tests have been decided to include in the SOP based on the feasibility and specificity in the evaluation of Goalball athletes. The tests that could be used for muscular strength evaluation are the curl-up, push-up, trunk lift, or the medicine ball throw, because all of these tests are easy to administer and require no instruments. Although the handgrip test is a valid and reliable test for assessing muscle strength (Artero et al., 2011(Artero et al., , 2012, it is not of interest to the characteristics of Goalball athletes. Consequently, in order to reduce the number of tests proposed, it was not included. In a similar way, although the vertical jump tests are widely adopted to evaluate lower limbs muscle strength (Petrigna et al., 2019), they were not included in the SOP. The evaluation of postural control capacity, especially in the health-related contest, is fundamental because of the loss of the visual field is associated with a greater fear of falling (Ramulu et al., 2012). Consequently, the Flamingo balance test (Colak et al., 2004), the only field test to evaluate postural control adopted by the authors, is suggested to be included in the Goalball athletes' evaluation.
Based on the analysis of the tests adopted to evaluate the healthrelated components in Goalball athletes, a SOP is suggested considering the most adopted and most pertinent test for this population. This include the BPFT as test battery composed by the sit and reach test and shoulder stretch for flexibility; 1-mile walk/run test for aerobic capacity; curl-up and push-up for muscular strength and endurance; and the twopoint skinfold thickness for body fat. Furthermore, as concern the postural control evaluation, the Flamingo balance test could be used.
Strengths, limits and future studies
The strength of the present study was to highlight how poor the literature on Goalball is. Optimistically, based on these results, researchers will develop these information to improve the health of individuals with VI.
The study has some limitations that need mentioning. First, we could not compare the results of the various studies included in the analysis because the researches employed different measurements and data collection protocols. A second limitation was the sample that was too different. Indeed, it was composed of school-children, adolescents, adults, and, as concern the sport level, participants were both active people and national athletes.
Because of the presence of few articles in the literature and based on the benefits of Goalball in VI people, future studies should increasingly investigate the contribution of this sport on physical, psychological, and social benefits in these athletes.
Conclusion
In conclusion, the present manuscript highlights how relatively little research has been systematically conducted to investigate this area and how different are the tests proposed to evaluate Goalball athletes and practitioners. We suggest that the BPFT with standardized tests could be adopted as SOP in Goalball.
Author contribution statement
All authors listed have significantly contributed to the development and the writing of this article.
Funding statement
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. (2018) | 2020-07-23T09:05:24.933Z | 2020-07-01T00:00:00.000 | {
"year": 2020,
"sha1": "99b7a701c6dbaa64770d9618b5c2df6d6663fa87",
"oa_license": "CCBYNCND",
"oa_url": "http://www.cell.com/article/S2405844020312512/pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "87f0f8937a029edb33334edb7fcb495d9561be14",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine",
"Psychology"
]
} |
267039088 | pes2o/s2orc | v3-fos-license | Quantitative Analysis of Inner, Middle, and Outer Retinal Thickness by Optical Coherence Tomography in Children and Adolescents
Purpose: To describe the variance of inner, middle, and outer retinal layer thicknesses (IRT, MRT, and ORT) at the macular area in children and adolescents with normal eyes in different age groups. Methods: This cross-sectional study enrolled subjects aged 5–18 years with normal eyes. The macula was scanned by optical coherence tomography (6 mm × 6 mm AngioScan-Optovue). Four age groups were defined (≤7, 7–10, 11–14, and ≥14 years). The influences of age and gender were analyzed. Results: One hundred and thirty-nine eyes of 69 subjects with a mean age of 10.92 ± 3.51 years were registered. The mean whole macular thickness (MT) was 297.32 ± 11.05 in males and 303.197 ± 13.32 in females (P = 0.01, t-test). The MT in each aging group was 301.47 ± 2.5, 295.53 ± 1.71, 300.81 ± 2.12, and 298.6 ± 1.87, subsequently (P = 0.17, analysis of variance test). Significant differences were found between the sexes at the perifoveal area and mainly in IRT. No correlation between eyes was noted. We observed that the RT fluctuates during growth and that gender has some influence on the evolution of RT. IRT and MRT changed reciprocally in all macular areas, whereas ORT expanded in all age groups of children and adolescents. Conclusions: No subsegmental retinal thickness difference between eyes was observed in pediatric groups in this study, while gender had some influence on perifoveal IRT. Despite the fact that this study is not a longitudinal study, we can get some insight into the developmental changes in retinal thickness and its clinical applications in children and adolescents.
optical coherence tomography (OCT), we now have a better knowledge of normal and pathological retinal development.The technique allows in vivo cross-sectional visualization of biological tissue at micrometer resolution of the retina, retinal nerve fiber layer, and optic nerve head. 6To evaluate pathological changes, it is important to quantify normal retinal thickness. 7In addition, to make an accurate diagnosis of a kid's retinal disease, quantitative measures from the child should be compared to age-matched normal controls.There are few reported normal values from children, and only normative studies of Turkish and American children are accessible in the literature. 8,9The current study uses the spectral domain OCT (SD-OCT) to create a normative pediatric database for inner, middle, and outer retinal layer thicknesses (IRT, MRT, and ORT) at the macular areas in healthy Iranian children and adolescents.
Methods
The study was conducted on healthy children and adolescents who came to Farabi Eye Hospital, Tehran, Iran, for routine eye examinations and students from two schools in Tehran who were invited to participate in the experiment from April 2018 to May 2020.An institutional review board approval was obtained for this cross-sectional study from the Farabi Eye Hospital, Department of Ophthalmology, Tehran University of Medical Sciences (IR. TUMS.VCR.REC.1397.1054).The tenets of Declaration of Helsinki were followed during the study.Participants provided written informed consent.Fifteen eligible children or their parents did not provide informed consent and were thus removed from the study.Patients' data were anonymized before data analysis.
The study included healthy children and adolescents with best-corrected visual acuity of 20/20, a refractive error between −1 and +1 diopters, and intraocular pressure of <21 mmHg.Exclusion criteria were any systemic and ocular disease, any positive past medical, surgical, and drug history.
The patients were divided into four groups based on their age: ≤7, 7-10, 11-14, and 14-18.Imaging and examinations were performed on the same day during working hours (up to 2 PM).On the Optovue RTVue XR Avanti (Optovue Inc., Freemont, California, USA), we obtained OCT angiography images using the split-spectrum amplitude-decorrelation angiography algorithm.A 40 nm wavelength laser provides 70,000 A-scans per second; 304 A-scans form a B-scan, while 304 vertical and horizontal lines were tested in the scanning area to obtain a three-dimensional data cube and eliminate motion artifacts.Volume scans by 3 mm × 3 mm and 6 mm × 6 mm centered onto the fovea were performed in both eyes for each patient using 400 raster lines.
The inbuilt software defined the thickness of the inner retina between the internal limiting membrane (ILM) and the outer boundary of the inner plexiform layer (IPL) (IRT, ILM-IPL), the middle retina from the outer boundary of IPL to the outer boundary of outer plexiform layer (OPL) (MRT, IPL-OPL) and thickness of the outer retina as outer margin of OPL to the outer boundary of retinal pigmentary epithelium (ORT, OPL-BRM).
In this study, the thickness was calculated for different sectors (whole image, central 6 mm diameter at fovea, parafovea, perifovea, and superior and inferior half of the extrafoveal areas) based on the Early Treatment Diabetic Retinopathy Study grid.OCT automatically locates the fovea.A 1.0 mm diameter circle was defined as the foveal region.The parafoveal area was defined as a ring centered on the foveal region with an inner diameter of 1.0 mm and an outer diameter of 3.0 mm.The perifoveal area was defined as a ring centered on the foveal region with an inner diameter of 3.0 mm and an outer diameter of 6.0 mm.The retinal thickness measurements were generated through a 21-line raster scanning protocol that was designed to allow frame averaging to maximize scan quality in this sample of children.
For statistical analysis, the thickness of foveal, parafoveal, perifoveal, and whole 6 mm diameter (WI) and superior (sup) and inferior (inf) half (hemi) data of parafovea and perifovea area from 6 mm × 6 mm images volumetric scans were composed.When two eyes from the same subject were eligible (in terms of the quality of the images), both eyes were included in the analysis.Two ophthalmologists (FG and VH) evaluated the quality and segmentations of OCT images.
Statistical analysis
The sample size was calculated as 97 eyes with a 95% confidence interval, 10% of the vascular density unit change (per percentage-quantitative data), and 50% population proportion.After determining the normality of the distribution with the Kolmogorov-Smirnov test and histogram, all quantitative data were given as mean with standard deviation.The variables with no Gaussian distribution have been reported as median with the range.Statistical software (SPSS software Version 21; SPSS, Inc., Chicago, IL, USA) was used to perform all statistical analyses.Inter-eye correlation was evaluated to see any significant correlation [Supplementary Table 1].For nonparametric and parametric comparisons, the Kruskal-Wallis test and one-way analysis of variance (ANOVA) were used.P values were adjusted for multiple comparisons using the Bonferroni method.Two-way ANOVA on means of retinal thickness was performed to test the main effect and interaction of gender and age.
results
Of the 85 initially examined participants, 16 had to be excluded for further investigation due to poor image quality.A total of 131 eyes of 69 individuals were investigated in this study.The mean age of all subjects was 10.92 ± 3.51 years (range, 5-18 years).70.2% of participants (92 eyes) were male with a mean age of 10.92 ± 3.71 years, and 29.8% (40 eyes) were female with a mean age of 10.9 ± 3.03 years old.Its mean weight was 39.2 ± 15.4 kg and its mean height was 137.8 ± 21.7 cm.In each of the four groups, 26 eyes (19.9%) were in the group under 7 years, 35 eyes (26.7%) were in the group between 7 and 10 years, 43 eyes (32.8%) were in the group from 11 to 14 years, and 27 eyes (20.6%) were in the group over 14 years.
The result of two-way ANOVA showed a marginally insignificant effect of gender on the whole MT [P = 0.04, Table 1].However, The MT was not affected by age [P = 0.1, Table 1].The course of alteration in thickness of inner, mid, and outer retina during childhood for both genders is illustrated in Figure 1a.
The mean central foveal thickness (CFT) was 251.10 μm, with the male having a thicker CFT than the female (252.15 1].
Results of the two-way ANOVA on macular regions showed that the difference in CFT among different age groups was nonsignificant [P = 0.17, Table 1].Furthermore, the CFT was not significantly affected by gender [two-way ANOVA P = 0.31, Table 1].The course of alteration in thickness of inner, mid, and outer foveal thickness during childhood for both genders is reported in Tables 2-4 [Figure 1b].
The two-way ANOVA showed that age [P = 0.14, Table 1] and sex [P = 0.27, Table 1] did not affect PaFT, with no significant interaction between age and sex on PaFT [P = 0.28, Table 1].The course of alteration in thickness of inner, mid, and outer PaFT during childhood for both genders is illustrated in Figure 1c.
In the perifovea, the two-way ANOVA showed a marginally insignificant effect of age on PeFT [P = 0.03, Table 1].However, the effect of gender was significant [P = 0.009, Table 1].However, the PeFT was not affected by the interaction between age and sex [P = 0.09, Table 1].The difference between the PeFT of females and males increased after 7 years old and became significant between 11 and 14 years old (mean PeFT 303.7 ± 12.1 in females and 288.9 ± 12.4 in males) [Figure 1d].However, the difference reduced afterward and became nonsignificant after 14 years old [Figure 1d].The course of alteration in thickness of inner, mid, and outer PeFT during childhood for both genders is illustrated in Figure 1d.
We performed the entire analysis considering one eye per case, and the results were alike.There was no significant correlation between the right and left eyes in our study for any of the study parameters (all P > 0.05) [Supplementary Table 1].The thickness correlation coefficient for both eyes was less than 0.3 in all studied parameters [Supplementary Table 1].
dIscussIon
This is the first study to compare automated layered retina thickness analysis of IRT, MRT, and ORT using Optovue OCT macular segmentation software in normal children and adolescents.The mean central macular thickness (CMT) in children and adolescents aged 5-18 years of age was 251.10 ± 20.54 μm.This value was 211.39 μm for Turkish children aged 6-16 years, 9 and is thinner than results acquired from healthy adults. 10Although CMT increased during childhood, this increase did not reach a significant statistical difference.There was no significant correlation between CMT and age in some studies. 11ncordant with other studies, we found that gender had some influence on total RT at the central 6 mm of the retina, with females having less CMT [Figure 1b and c]. 9,12 Bafiq et al. revealed similar results, with no significant gender differences in CMT. 8,13According to these discrepancies, race or ethnicity appears to have an impact on CMT. 8,14El-Dairi et al. recorded greater values for IRT and CFT in White children compared with Black children. 15Huynh et al. found that white children had more CMT and IRT than East Asians, while Middle Eastern children had greater outer temporal MT than whites. 16cording to our study, in the central 6 mm of the retina, gender has some impact on the course of RT alterations in the age groups [Table 1 and Figure 1a].The difference between girls and boys was more prominent in the IRT of the perifoveal area [Figure 1d].Interestingly, like Grover et al., considering all RT (ILM-BRM) segmentally, we found that the perifoveal area was thicker in females. 17Furthermore, we showed that the RT decreased slightly from <7 years old up to 7 years old.In this period, there was no significant difference between girls and boys.RT then increased from 7 years to 14 years.This increase appeared as a higher slope in the graph of RT changes in the girls and was more prominent in the perifoveal area.This abrupt increase led to significant RT differences between males and females.After 14 years old, RT in boys continued to increase; however, in girls, RT declined suddenly and reached its pregrowth values.
In layer analysis, while in boys, the IRT decreased significantly from <7 years old to 14 years old, the IRT increased in girls in the same period and peaked at 14 years old [Figure 1b].In the same period, the MRT in both sexes underwent thinning.This thinning in girls was significant only in the inferior hemifield.Our data revealed that ORT in boys remained stable during childhood; however, the outer retina in girls grew significantly from 7 to 14 years old.
Puberty starts from 8 to 12 years old in girls and 9-14 years in boys. 18During this period, a variety of hormonal changes leads to significant somatic changes in humans. 19As the observed changes in RT occur during puberty and vary significantly between males and females, we hypothesized that this retinal alteration could be due to sexual hormonal changes.Sex hormones have been shown to have a neuroprotective effect in normal retinas. 20Further animal and human studies are required to evaluate the role of gonadal hormones on the physiology and pathology of the retina.Previous studies demonstrated the gonadal hormone receptors are distributed unevenly throughout the retina. 21This study showed that in the middle retina, the inferior hemifield was more affected during puberty, while in the outer retina, the superior hemifield was affected by puberty.Further study is required to evaluate variation in the presence of gonadal hormone receptors in different layers of the retina.Nevertheless, due to the descriptive nature of this study, the mentioned impact of gender on retinal layers should be carefully considered and have to be confirmed by well-designed animal and prospective human studies.
Histologic and optical coherence studies and tomography studies largely agree on foveal development.Formation of the foveal center and deepening of the foveal pit occur at 31-and 42-weeks' postmenstrual age owing to centrifugal inner retinal cell displacement. 22,23Simultaneously, the inner segment and outer segment bands of the outer retina undergo gradual development or remodeling in the opposite route. 23,24There was no comparable data for comparison in older children.
The assessment of the mid-retina (inner nuclear layer to OPL, [INL-OPL]) in retinal illness is important, as it has been documented in diabetic patients. 25The INL is mainly formed by the nuclei of bipolar and Müller cells and by the association of horizontal and amacrine cells.
At the foveal area, MRT thinning is concordant with IRT thickening.We hypothesize that bipolar, amacrine, horizontal, and Müller cells were modestly decreasing in this area up to the age of 14 years while ganglion cell density was increasing.There is a sudden increase in MRT over the age of 14 years, accompanied by a modest reduction in IRT.Parafoveal and perifoveal alterations were also observed in the fovea.These changes may be caused by changes in mid-retina cells caused by adolescent sex and growth hormones.A new possibility for using growth hormones in retinal regeneration studies has been revealed.Growth hormone acts as a synaptogenic modulator in the chick retina. 26,27 the current investigation, the ORT in the central 1 mm area (165.17 ± 10.63) was much less than the levels informed by Yanni et al.(210.2± 2.4) in children aged 5-16 years using Spectralis SD-OCT (Heidelberg Engineering; Vista, CA). 8hose might be due to different instruments used or ethnic and racial variances.
Studies have revealed that the development of the fovea centralis is a lengthy process that continues for several years after birth. 22istological investigations indicate increasing thickening of the foveal ONL after birth, as cone packing and elongation of IS and OS occur. 24According to Thomas et al., cone-packing density in the fovea only reaches half-adult levels at 45 months. 2 The photoreceptor subcellular structures in the foveal center progress centripetally as the baby grows. 28Agreeing with Thomas, we can assume that the observed cone packing and probably remodeling, and development may continue until adulthood. 2he height of photoreceptors rises steadily from childhood to adulthood, resulting in thin and long cones in the foveal center and long rod outer segments with eccentricity. 28Further OCT investigations on normal foveal development in term newborns and young children revealed that the ONL continues to thicken at least until the age of 12 years, indicating an increasing number of foveal cone photoreceptors. 29e main limitation of our study was the relatively small number of cases and its nonlongitudinal nature.Future studies with longer follow-up periods are needed to evaluate the repeatability and reproducibility of segmented macular layer measurements in children.Our findings were also confined to the macular region of the retina, and while the macula is crucial for central vision, it only accounts for a small fraction of the total retinal area.To better understand the normative morphology of peripheral retinal regions in childhood, and expected changes in morphology over time, wide-field scanning methods are necessary.Furthermore, we did not examine axial length, even though the axial length appears to have minimal influence on macular and retinal nerve fiber layer thickness measurements. 9,30T appears to change regularly, indicating that the area is plastic.As a result of this study, a valid normative database of segmented MTs of healthy Iranian children and adolescents can be generated.The RT of children increased as they grew older, and gender also played a role in layered thicknesses.Up to the age of 14 years, the inner retina of the macular area thickened while the MRT shrank, with reciprocal changes at older ages.
Figure 1 :
Figure 1: Alterations in retinal thickness of normal healthy children in both genders in different age groups.(a) This column shows how whole macular thickness varies from <7 years old to 18 years old in both genders.The first image illustrates changes in whole retinal thickness.The second to fourth images depict changes in inner, middle, and outer retinal thickness; (b) This column shows how foveal thickness varies during childhood.The first image illustrates changes in whole retinal thickness.The second to fourth images depict changes in inner, middle, and outer retinal thickness; (c) This column shows how parafovea thickness varies from less than 7 years old to 18 years old in both genders.The first image illustrates changes in whole retinal thickness during growth.The second to fourth images depict changes in inner, middle, and outer retinal thickness separately for girls and boys.(d) This column shows how perifoveal thickness varies from <7 years old to 18 years old in both genders.The first image illustrates changes in the whole retinal.The second to fourth images depict changes in inner, middle, and outer retinal thickness.The error bars show 95% confidence intervals d c b a vs. 248.61μm, P = 0.47, Mann-Whitney U test).The median CFT in each aging group was 245.8 (236.5-253.6),248.3 (234-257.6),250.7 (238.7-267.2),and 251 (240.1-274.1)μm, subsequently (P = 0.2, Kruskal-Wallis) [Table
Table 1 : Results of two-way analysis of variance analysis for evaluating the effect of age and sex and their interaction on retinal thickness in childhood per micrometer
Yanni et al. and | 2024-01-20T05:10:39.258Z | 2023-12-21T00:00:00.000 | {
"year": 2023,
"sha1": "43b7611fab2e92a201aa7387884239e5d62b97df",
"oa_license": "CCBYNCSA",
"oa_url": null,
"oa_status": null,
"pdf_src": "PubMedCentral",
"pdf_hash": "43b7611fab2e92a201aa7387884239e5d62b97df",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
220839549 | pes2o/s2orc | v3-fos-license | Time-Restricted Feeding Improves Body Weight Gain, Lipid Profiles, and Atherogenic Indices in Cafeteria-Diet-Fed Rats: Role of Browning of Inguinal White Adipose Tissue
Time-restricted feeding (TRF) showed a potent effect in preventing obesity and improving metabolicoutcomes in several animal models of obesity. However, there is, as of yet, scarce evidence concerning its effectiveness against obesogenic challenges that more accurately mimic human Western diets, such as the cafeteria diet. Moreover, the mechanism for its efficacy is poorly understood. White adipose browning has been linked to body weight loss. Herein, we tested whether TRF has the potential to induce browning of inguinal white adipose tissue (iWAT) and to attenuate obesity and associated dyslipidemia in a cafeteria-diet-induced obesity model. Male Wistar rats were fed normal laboratory chow (NC) or cafeteria diet (CAF) for 16 weeks and were subdivided into two groups that were subjected to either ad libitum (ad lib, A) or TRF (R) for 8 h per day. Rats under the TRF regimen had a lower body weight gain and adiposity than the diet-matchedad lib rats, despite equivalent levels of food intake and locomotor activity. In addition, TRF improved the deranged lipid profile (total cholesterol (TC), triglycerides (TG), high-density lipoprotein (HDL-c), low-density lipoprotein (LDL-c)) and atherogenic indices (atherogenic index of plasma (AIP), atherogenic coefficient (AC), coronary risk index (CRI) in CAF-fed rats. Remarkably, TRF resulted in decreased size of adipocytes and induced emergence of multilocular brown-like adipocytes in iWAT of NC- and CAF-fed rats. Protein expression of browning markers, such as uncoupling protein-1 (UCP1) and peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC1α), were also up-regulated in the iWAToftime-restricted NC- or CAF-fed rats. These findings suggest that a TRF regimen is an effective strategy to improve CAF diet-induced obesity, probably via a mechanismthe involving WAT browning process.
Introduction
Obesity is defined as abnormal or excessive fat accumulation that may impair health [1]. The etiology of obesity is quite complicated owing to the involvement of both genetic and environmental factors. Among the environmental factors, diet plays a significant role in the development of obesity [2]. Modern nutritional patterns, termed as "Western diets", including high-fat and -cholesterol, high-protein, high-sugar, and excess salt intake, as well as frequent consumption of processed and fast browning potential in obese diabetic animals [35]. Based on previous findings showing that TRF regimen reduced body weight gain in DOI without altering caloric intake or activity level [8,12,19], we hypothesized that a TRF regimen would prevent obesity via promoting WAT browning. Hence, the present study was aimed to address whethera TRF schedule of 8 h per day during the active phase can attenuate excessive body weight gain inCAF-diet-induced obeserats. There arescarce data available in the literature regarding the effects of TRF on lipid profile and atherogenicity, in animal models of obesity, and therefore, we determined whether TRF can prevent the dyslipidemia produced in response to a CAFdiet. Additionally, the potential role of TRF paradigm as a white fat browning inducer was also explored.
Animals
All research and animal care procedures were authorized and approved by the Institutional Animal Care Committee of the National Administration of Algerian Higher Education and Scientific Research (Ethical approval number: 981-1 law of 22 August 1998).
A total of 24 male albino Wistar rats (Rattusnorvegicus) weighing 120-130 g were obtained from the Pasteur Institute of Algiers, Algeria. Rats were single-housed in polypropylene cages with a stainless steel lid under thermoneutrality conditions (28-30 • C) and a 12 h light: 12 h dark schedule, with lights on at 08:00 a.m. (Zeitgeber time 0, ZT0).Normal rat chow pellets (supplied by ONAB, Algiers) and water were provided ad lib. All animals were acclimatized to their new environment for 15 days before starting the experiment.
Experimental Feeding Schedule and Diets
Upon initiation of the experiment, rats (initial body weight 150 ± 1.5 g) were randomized by weight into four groups (n = 6 per group) to ensure equivalent starting body weight. They were maintained on the normal chow diet (NC) or switched to a pre-selection of highly palatable, energy-dense human foods consisting of cookies, cereals, chocolate, crackers, chips, cheese, processed meat, etc. (cafeteria diet; CAF), and they were subdivided into two groups per diet.One group/diet had ad libitum (ad lib; A) access to food: NC-A (n = 6) and CAF-A (n = 6) rats. The second group/diet had temporal restriction (R) to food: NC-R (n = 6) and CAF-R (n = 6) rats. These feeding regimens continued for 16 wks until the end of the experiment.Rats under the TRF regimen were allowed access to the diet for 8 h per day during the active phase (dark period) from ZT 13 (1 h after lights off) to ZT 21 (3 h before lights on). Food access was monitored by manually switching the rats from cages with food and water to cages with only water, to avoid foraging and coprophagia. Rats fed ad lib were also transferred between feeding cages at the same time to standardize handling stress and minimize experimental variation between groups.The CAF-diet protocol used in this study was adapted from a previous study [28] and consisted of 19 different highly palatable and energy-rich unhealthy human snack foods purchased from a supermarket in Algeria, accompanied by normal rat chow pallets. The list of cafeteria diet food items with nutrient compositions is listed in Supplementary Table S1. The normal chow diet is composed of whole ground corn, soybean meal, and wheat bran and is supplemented with minerals, amino acids, and vitamins and had an energy density of 310 Kcal/100 g (56% carbohydrate, 24% protein, and 16% lipid), whereasthe CAFdiet had an energy density (taken from the daily offering of each component) of 501 Kcal/100g on average (37% carbohydrate, 16% protein, and 47% lipid). Three different items plus normal rat chow pallets were provided daily in excess quantities to each animal. To maintain variety and thus induce sustained hyperphagia, the food items provided were altered daily by replacing the three food items with new fresh ones. The menu was varied daily in a manner that ensuresa similar proportion of fat, protein, sugar, and carbohydrate in each daily set of CAF foods. Normal chow-diet and CAF-diet food items were pre-weighed individually and presented to the rats at ZT 13, and the leftoverswere weighed after 24 and 8 h in ad liband time-restrictedfed rats, respectively. To carefully monitor the CAF food consumption, food spillage was meticulously collected from the cage floor, and the difference between presented and recovered food items was corrected for drying, as previously described by Shafat et al. [31]. Calorie intake was calculated as weekly intake, based on nutritional information provided by the manufacturer (Supplementary Table S1). The body weight was recorded weekly over 16 wks between 09:00 and 10:00 a.m. The experimental protocol is described in Figure 1.
Nutrients 2020, 12, x FOR PEER REVIEW 4 of 17 corrected for drying, as previously described by Shafat et al. [31]. Calorie intake was calculated as weekly intake, based on nutritional information provided by the manufacturer (Supplementary Table S1). The body weight was recorded weekly over 16 wks between 09:00 and 10:00 a.m. The experimental protocol is described in Figure 1. Male Wistar rats were allowed a 15-dayadaptationperiod prior to beginning the experiment and were divided into four groups as follows: NC-A (normal chow,ad libfeeding); NC-R (normal chow time, restricted feeding); CAF-A (cafeteria diet,ad libfeeding); and CAF-R (cafeteria diet, time-restricted feeding).
Time-restricted feeding groups were allowed to feed for 8 h per day during the dark phase between ZT 13 and ZT 21. After sixteen weeks of experiment, animals were sacrificed. CAF: cafeteria diet; NC: normal chow diet; TRF: time-restricted feeding; ZT: Zeitgeber.
Locomotor Activity Assessment
At week fifteen, the locomotor activity was evaluated in the open field by counting the number of squares crossing and occurrences of rearing behavior.Those two variables are indicative of general activity levels [36]. The test was performed in a silent room at night, 2 h after light off (ZT 14), in a four-sided 90 × 90 × 45 cm varnished wooden box, and the floor of the open field was divided into 25 equal-size squares (18 cm). In order to avoid any potential food anticipatory activity, which was reported to occur a few hours prior the arrival of food [37], we first started to assess the locomotor activity in ad lib groups,followedby TRF groups.Data from all groups of rats were collected within approximately 2 h (between ZT 14 and ZT 16). Each rat was placed singly in the center of the open field and given 5 min to explore. A rat was recorded as crossing a square when more than three paws or half of the body crossed the boundary into the nearby squares. Rearing was defined from the moment when both forelimbs raised at least 1 cm above the floor. At the end of each trial, the box was cleaned with water and wiped dry before introducing the next animal [38].
Blood and Tissue Collection
At the end of the experimental period, rats were fasted for ~16h starting from ZT 21, anesthetized by urethane (1 g/kg b.w, i.p.), and scarified. Bloodwas collected via heart puncture, and serum was obtained by centrifugation at 3000 r.p.m for 15 min at 4°C. The recovered serum was keptfrozen at −20°C until analysis. White adipose tissue depots from mesenteric (mWAT), epididymal (eWAT), retroperitoneal (rWAT), and inguinal subcutaneous (iWAT) sites were rapidly removed, rinsed with phosphate buffer (PBS (phosphate-buffered saline), and then weighed, followed by immediate freezing in liquid nitrogen. Serum and iWAT specimens were shipped on Figure 1. Schematic representation of the experimental feeding schedule. Male Wistar rats were allowed a 15-dayadaptationperiod prior to beginning the experiment and were divided into four groups as follows: NC-A (normal chow, ad libfeeding); NC-R (normal chow time, restricted feeding); CAF-A (cafeteria diet, ad libfeeding); and CAF-R (cafeteria diet, time-restricted feeding). Time-restricted feeding groups were allowed to feed for 8 h per day during the dark phase between ZT 13 and ZT 21. After sixteen weeks of experiment, animals were sacrificed. CAF: cafeteria diet; NC: normal chow diet; TRF: time-restricted feeding; ZT: Zeitgeber.
Locomotor Activity Assessment
At week fifteen, the locomotor activity was evaluated in the open field by counting the number of squares crossing and occurrences of rearing behavior.Those two variables are indicative of general activity levels [36]. The test was performed in a silent room at night, 2 h after light off (ZT 14), in a four-sided 90 × 90 × 45 cm varnished wooden box, and the floor of the open field was divided into 25 equal-size squares (18 cm). In order to avoid any potential food anticipatory activity, which was reported to occur a few hours prior the arrival of food [37], we first started to assess the locomotor activity in ad lib groups, followedby TRF groups. Data from all groups of rats were collected within approximately 2 h (between ZT 14 and ZT 16). Each rat was placed singly in the center of the open field and given 5 min to explore. A rat was recorded as crossing a square when more than three paws or half of the body crossed the boundary into the nearby squares. Rearing was defined from the moment when both forelimbs raised at least 1 cm above the floor. At the end of each trial, the box was cleaned with water and wiped dry before introducing the next animal [38].
Blood and Tissue Collection
At the end of the experimental period, rats were fasted for~16 h starting from ZT 21, anesthetized by urethane (1 g/kg b.w, i.p.), and scarified. Bloodwas collected via heart puncture, and serum was obtained by centrifugation at 3000 r.p.m for 15 min at 4 • C. The recovered serum was keptfrozen at −20 • C until analysis. White adipose tissue depots from mesenteric (mWAT), epididymal (eWAT), retroperitoneal (rWAT), and inguinal subcutaneous (iWAT) sites were rapidly removed, rinsed with phosphate buffer (PBS (phosphate-buffered saline), and then weighed, followed by immediate freezing in liquid nitrogen. Serum and iWAT specimens were shipped on dry ice to the laboratory of Prof. Ahmad Agil (UGR, Spain) and then stored at −80 • C for lipid profiles and Western blot analysis.
Measurement of Lipid Profiles and Atherogenic Indices
Serum levels of triglycerides (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-c), and low-density lipoprotein cholesterol (LDL-c) were measured using commercial kits in an automatic analyzer (Roche-Hitani molecular system 902, Branchburg, NJ, USA).
Microscopic Analysis
Following anesthesia, excised inguinal fat was fixed by immersion in 4% formaldehyde in 0.1 M phosphate buffer overnight at 4 • C, washed inPBS, dehydrated in a graded series of ethanol, cleared in xylene, and embedded in paraffin. The paraffin blocks from each group were cut with a microtome into 5 µm thick sections, stained with hematoxylin and eosin (H&E), and inspected under a light microscope (×400, Olympus, Germany) equipped with a digital camera system (Carl Zeiss camera, model Axiocam ERc 5s. Germany). Images of H&E-stained tissue sections were digitized, and adipocyte size and percentage of multilocularity were determined usingAxio Vision software (Carl Zeiss Imaging Solutions). The average adipocyte size was expressed as the average cross-sectional area per cell (µm 2 /cell) of tissue sample, which was calculated based on the values of at least 20 adipocytes in 10 random fields per section. The percentage of total adipocyte population showing multilocularity was determined by quantifying the area of multilocular adipocytes in relation to the entire area of the section. Histological slides were performed and examined by a technician who was blinded to tissue group identity.
Western Blotting Analysis
About 100-200 mg of iWAT was homogenized in lysis buffer (150 mmol NaCl, 5 mmol EDTA, 50 mmol Tris-HCl, and pH 7.4) without Triton X-100 and homogenized with a Teflon pestle. Homogenates were centrifuged (3000× g for 15 min, 4 • C), and the fat cake was removed from the top of the tube. Then, Triton X-100 was added to a final concentration of 1%. After incubating at 4 • C for 30 min, extracts were cleared by centrifugation at 15,000× g for 15 min at 4 • C. One hundred micrograms of total protein was analyzed by SDS-PAGE (sodium dodecyl sulfate polyacrylamide gel electrophoresis). The gels for immunoblot analyses were transferred to a nitrocellulose membrane (Bio-Rad Trans-Blot SD; Bio-Rad Laboratories, Hercules, CA, USA). The membranes were cut at UCP1 and PGC1α molecular weight level (33 kDa), and blots were reacted with a 1:2000 dilution of anti-UCP1 produced in rabbit (Sigma Aldrich, U6382, St. Louis, MO, USA), in blocking solution (PBS, 5% nonfat milk) and anti-PGC1α produced in rabbit. β-actin antibody generated in mouse (Santa Cruz Biotechnology, SC-81178, Santa Cruz, CA, USA) was used as a control. Horseradish peroxidase labeled secondary antibodies were goat anti-mouse IgG and goat anti-rabbit IgG (1:1000, Sigma Aldrich). Proteins were visualized by enhanced chemiluminescence (ECL kit, GE Healthcare Life Sciences, Pittsburgh, PA, USA).
Statistical Analysis
Statistical Package of Social Science (IBM SPSS Software, version 22) was used for statistical analysis. All results are expressed as mean ± standard deviation (S.D.) values. Two-way ANOVA was used to examine two variables (diet and schedule feeding). When a significant F score on two-way ANOVA was recorded, pair-wise comparisons were conducted using aMann-Whitney U-test. Repeated-measures two-way ANOVA was also used to analysis change in body weight over time. Following a significant result, single time point comparisons were made using unpaired Mann-Whitney U-test. Differences between group means were considered statistically significant if p < 0.05.
Effects of TRF on Body Weight, Calorie Intake, Adipose Weight, and Locomotor Activity
To evaluate the effectiveness of a TRF regimen to protect against CAF-diet-induced obesity, we first tested its effect on body weight. As expected, rats fed a CAF-diet on ad libaccess exhibited a significant increase in body weight from the 2nd week of the experiment until the end (16 wks) as compared to rats fed a NCdiet ad lib (p < 0.01; Figure 2a). The final body weight gain in the CAF-A group was significantly higher compared with that in the NC-A group (219.7 ± 18.9 g vs. 126.5 ± 13.0 g; p < 0.01; Figure 2b). The TRF regimen significantly reduced body weight in rats fed with both NC-and CAF-diets, from respectively the 5th and 3rd week until the end of the experiment (p < 0.05; Figure 2a). The final body weight gain in NC-R (75.0 ± 11.0 g) and CAF-R (138.1 ± 12.3 g) groups was significantly lower compared to that in their diet-matched ad lib groups (p < 0.01; Figure 2b). Interestingly, the final body weight gain in the CAF-R group was not significantly different from the NC-A group (p > 0.05; Figure 2b).
Although the final body weight gain was lower in TRF rats, the daily total calorie intake showed no statistically significant difference between time-restricted and ad lib rats that were fed with either NC or CAFdiet (p > 0.05; Figure 2c). Notably, total daily calorie intakes per individual rat were 52.1 ± 8.0, 52.8 ± 10.4, 72.4 ± 12.0, and 75.3 ± 9.6 Kcal in the NC-A, NC-R, CAF-A, and CAF-R groups, respectively.When normalized to body weight, NC-R and CAF-R rats consumed similar calories at the beginning of the experiment (weeks 2-4), as compared to their diet-matched ad lib rats (p > 0.05; Figure 2d). Daily calorie intakes per gram body weight was 0.30 ± 0.07, 0.32 ± 0.05, 0.31 ± 0.10, and 0.34 ± 0.07 Kcal in the NC-A, NC-R, CAF-A, and CAF-R groups, respectively. By the end of the experiment (week 14-16), NC-R and CAF-R rats consumed higher calories than their ad libcounterparts (p < 0.05; Figure 2e). The values were 0.19 ± 0.05, 0.24 ± 0.05, 0.18 ± 0.07, and 0.26 ± 0.05 Kcal/g body weight/day in the NC-A, NC-R, CAF-A, and CAF-R rats, respectively.
To test whether the reduced body weight gain, in the TRF rats, was due to locomotor activity, animals were tested in an open-field arena to evaluate the spontaneous locomotor activity. Notably, the CAF-A group exhibited reduced locomotor activity compared withthe NC-A group, since the CAF-A group showed an overall decrease in the number of crossing squares (53.1 ± 6.9 vr. 62.3 ± 4.9; p < 0.05; Figure 2f) and rears (23.1 ± 7.4 vr. 29.3 ± 4.7; p < 0.05; Figure 2g). Importantly, the TRF schedule had no significant effect on locomotor activity. Indeed, the number of crossing squares and rears in NC-R (62.0 ± 5.9 and 29.4 ± 3.4, respectively, Figure 2f,g) and CAF-R (57.2 ± 6.1 and 26.1 ± 5.6, respectively, Figure 2f,g) groups had no significant differences(p > 0.05)compared totheir ad lib counterparts, although they tended to be lower in the CAF-A group, compared to the CAF-R group.
To identify whether the TRF paradigm had an effect onspecific adipose depots, we weighed adipose tissues from different anatomical locations. As expected from their body weights, the CAF-A group showed significantly higher total adiposity than the NC-A group (p < 0.01; Table 1). The TRF paradigm significantly reduced the weight of the iWAT and all the visceral adipose tissue depots (mesenteric, epididymal, and retroperitoneal) in both NC-and CAF-fed rats (p < 0.05; Table 1). The total adiposity is significantly lower in TRFrats compared to that in ad librats (p < 0.01; Table 1). Two-way ANOVA revealed a significant main effect for diet on number of crossing squares and rears without intersection of diet and schedule feeding (p > 0.05).
Effects of TRF on Lipid Profiles and Atherogenic Indices
To determine whether a TRF regimen could prevent dyslipidemia in the CAF-fed rats, serum lipid profiles and atherogenic indices were evaluated. As shown in Table 2, the CAF-A group showed a significant increase in serum levels of TG, TC, and LDL-c as compared to the NC-A group (p < 0.01). After 16 wks of TRF schedule, serum TG, TC, and LDL-c levels in the CAF-fed rats were significantly decreased by approximately 19.8%, 27.8%, and 43.5%, respectively, as compared to those inthe CAF-A group (p < 0.01; Table 2). Serum HDL-c levels in the CAF-A group weresignificantly lower than thoseinthe NC-A group (p < 0.01; Table 2); however, the TRF paradigm increased serum HDL-c levels by approximately 29.6% ascompared to thoseinthe CAF-A group (p < 0.01; Table 2). Atherogenic and cardioprotective properties of the TRF are also presented in Table 2. The CAF-A group showed significantly higher AIP, AC, and CRI (p < 0.01) than the NC-A group. It was interesting to note that the TRF regimen ameliorated these atherogenic indices in CAF-fed rats.Indeed, the values of AIP, AC, and CRI were significantly reduced in the CAF-R group when compared to those recorded for their ad lib counterparts (p < 0.01; Table 2), and they were similar to those inthe NC-A group.Nevertheless, the TRF had no significant effectoneither lipid profiles (TG, TC, HDL-c, and LDL-c) oratherogenic indices (AIP, AC, and CRI) inNC-fed rats (p > 0.05; Table 2). This is confirmed by two-way ANOVA between groups that showed a significant interaction between schedule feeding and diet on lipid profiles and atherogenic indices (p < 0.05).
Effects of TRF on iWAT Morphology
To determine the effect of TRF on morphological changes and browning induction in iWAT, histological analyses were performed in the inguinal depots. These depots were selected because they are those with a greater browning capacity, according to the literature [42]. Our results showed that adipocyte sizes were significantly higher in the CAF-A group compared to the NC-A group (57 ± 5.6 µm 2 × 100 vs. 39 ± 4.2 µm 2 × 100; p < 0.01; Figure 3a), and that the TRF paradigm decreased adipocytes size inrats fed with both NC-(28 ± 4.0 µm 2 × 100) and CAF-diets (40 ± 4.1 µm 2 × 100) as compared to their diet-matched ad librats (p < 0.01; Figure 3a). Interestingly, adipocyte size in CAF-R group was decreased to nearly that inthe NC-A group.
Effects of TRF on iWAT Morphology
To determine the effect of TRF on morphological changes and browning induction in iWAT, histological analyses were performed in the inguinal depots. These depots were selected because they are those with a greater browning capacity, according to the literature [42]. Our results showed that adipocyte sizes were significantly higher in the CAF-A group compared to the NC-A group(57 ± 5.6 µm 2 × 100 vs.39 ± 4.2 µm 2 × 100; p < 0.01; Figure 3a), and that the TRF paradigm decreased adipocytes size inrats fed with both NC-(28 ± 4.0 µm 2 × 100) and CAF-diets(40 ± 4.1 µm 2 × 100) as compared to their diet-matched ad librats (p< 0.01; Figure 3a). Interestingly, adipocyte size in CAF-R group was decreased to nearly that inthe NC-A group.
As shown in H&E-stained preparation (Figure 3b), the majority of cells in all groups showed an appearance of white unilocular adipocytes with a single lipid droplet. The TRF schedule induced the formation of multilocular brown-like adipocytes dispersed among unilocular adipocytes in both NC-and CAF-fed rats. The percentage of multilocularity represented around 31.1% and 15.2% of total area analyzed in the NC-R and CAF-R group, respectively. As shown in H&E-stained preparation (Figure 3b), the majority of cells in all groups showed an appearance of white unilocular adipocytes with a single lipid droplet. The TRF schedule induced the formation of multilocular brown-like adipocytes dispersed among unilocular adipocytes in both NCand CAF-fed rats. The percentage of multilocularity represented around 31.1% and 15.2% of total area analyzed in the NC-R and CAF-R group, respectively.
Effects of TRF on UCP1 and PGC1α Contents
The molecular signature that identifies brown and brown-like adipocytes is UCP1 [43]. Therefore, we examined UCP1 protein levels in inguinal fat depots from both NC-and CAF-fed rats on TRF using Western blot. As expected from their microscopic aspect, UCP1 protein expression was significantly enhanced in the inguinal fat of NC-R and CAF-R groups by~3 and~2 fold, respectively, as compared to their diet-matched ad lib groups (p < 0.01; Figure 4b). Importantly, a weak signal of UCP1 was observed in the inguinal fat ofad lib-fed rats, whereas TRF rats exhibited a more pronounced UCP1 signal (Figure 4a).
Discussion
Herein, we report, for the first time, that a TRF regimen prevents obesity and associated PCG1α is a master nuclear transcription factor that controls the expression of the thermogenic gene program, including the expression of the UCP1 gene. Its expression level was significantly enhanced in inguinal fat depots of the NC-R and CAF-R groups by~1.5 fold, as compared to their ad lib counterparts (p < 0.01; Figure 4c).
Discussion
Herein, we report, for the first time, that a TRF regimen prevents obesity and associated dyslipidemia in CAF-diet-induced obese rat model. This effect occurred probably via browning of the iWATwithout changing the calorie intake or locomotor activity.
Our results showed that consumption of calorically dense, palatable human foods induced an excessive caloric intake, resultingin rapid and drastic weight gain, abdominal adiposity, and dyslipidemic state in thead lib-fed group, suggesting that the CAF-diet used in the present study successfully induced an animal model of obesity. As shown in Table S1, the CAFdiet contained foods that are high in fat and saturated fatty acids; hence, the weight gain in the CAF-ad lib group could be due to the high rate of acylation of saturated fatty acids into triglycerides that are subsequently stored in the adipose tissue [44]. Besides, post-ingestive effects of high-fat food contribute to weight gain through reduction of satiety signals [45] and attenuation of fatty acid oxidation [46]. Interestingly, TRF significantly reduced the body weight gain in our rats fed on a CAF-diet, suggesting thatTRF is an effective strategy for weight reduction with aCAF-style diet that models the Western pattern diet. Similar findings have also been previously reported in different animal strains after exposure to TRF of other obesogenic diets [8,9,12,[19][20][21]23]. This has also been observed among some overweight or obese humans undergoing a TRF eating pattern [14,15]. It should be noticed that the body weight gain-lowering effect of TRF in the current study took place in the absence of changes in caloric intake and locomotor activity, which is consistent with previous studies [8,9,12,19]. Therefore, the mechanism whereby TRF reduced body weight cannot simply be explained by changes in quantity of diet or physical activity, which represent the locomotor activity of 5 min duration of the dark phase (phase of activity) of the 15th week of the experiment, but rather by other mechanisms, as discussed below. Our present data of calorie intake are not consistent with another previous report in which mice fed an HFD under TRF of 12 h during the dark or light phase were unable to fully compensate for the limited availability of food, and calorie intake was significantly reduced compared with their ad lib counterparts [47]. The reason for this discrepancy is uncertain and requires further investigation.
One potential explanation to account for the attenuated body weight gain without any overall change in calorie intake or physical activity upon TRF is through switching lipid metabolism from storage to oxidative as a strategy to enhance energy expenditure. Although energy metabolism was not measured in the current study, the study by Panda's group showed that 8hTRF during the active phase attenuated the HFD-induced dampening of the daily rhythm of the respiratory exchange ratio and led to an overall increase in oxygen consumption and energy expenditure in mice, indicating lipid utilization for energy metabolism [8]. In another study by the same group, TRF of HFD or normocaloric diet increased PGC1α gene expression in WAT of mice, leading to increased β-oxidation [9], which corroborates our data of PGC1α.
Ourfindingsconcerning body weight gain paralleled those of fat accumulation measurements. Notably, rats fed a CAFdietad lib exhibited drastic increases in fat pad weights (inguinal, mesenteric, epididymal, retroperitoneal) compared to rats fed an NCdietad lib. These observations suggest that the exposure of rats to calorically dense diets facilitated fat accumulation in the abdominal regions due to the high effective energy content of the high-fat foods [48]. Previous studies have observed a marked reduction in adipose tissue mass in rodents under a TRF regimen [8,9,23], which is in agreement with the present finding showing that TRF of NC-or CAF-diet reduced fat accumulation in subcutaneous and visceral fat, resulting in a significant decrease in total adiposity. The reduction of fat accumulation in WAT under TRF may be due to an overall increase in fatty acid oxidation and a decrease in free fatty acid synthesis [9], which paralleledthe current data showing that adipocytes size of iWATinTRF rats were smaller than thatin diet-matched ad lib rats.
The excessive weight gain in the CAF-Agroupwas associated with a dyslipidemic condition, as evidenced by abnormally elevated serum levels of TG, TC, and LDL-c and reduced serum level of HDL-c, which is consistent with the findings of a previous study using a CAFdiet [49]. The dyslipidemia was probably evident due to the overall higher consumption of the CAFdiet, which, while being high in fat, is also high in saturated fatty acids, known to increase the production of TG and LDL-c by the liver [50], and in cholesterol, which can result in a hypercholesterolemia [51] and reduced HDL production or HDL clearance from plasma [52]. Dyslipidemia may also beexplained bya hepatic lipid accumulation, which was associated with disrupted circadian rhythms and food intake patterns in response to high caloric diet [53]. Interestingly, the TRF regimen improved the negative lipid profile in CAF-fed rats by decreasing serum levels of TC, TG, and LDL-c and increasing serum level of HDL-c, suggesting its anti-hyperlipidemic effect.However, TRFunexpectedly had no effect on dyslipidemic parameters when rats were fed with anNCdiet, in spite of a marked effect on body weight. The lack of the TRF effect on dyslipidemic parameters in normocaloric-fed rats may indicate that the influence of TRF may be negligible when the overall content of diet is less caloric, which would require future investigation.To our knowledge, only three studies have been reported so far on the anti-hyperlipidemic potential of TRF in DIO models(Chaix et al. [9], Sherman et al. [23], and Sun et al. [12]). It is important to note that there were considerable variations in experimental design, such as obesogenic diets, animal strain, feeding window, and experimental durations between the present study and the aforementioned studies, which make direct comparison difficult.The mechanism wherebyTRF underlies lipid parameters is unclear. Typically, weight loss is expected to promote lipidimprovements [54]; however, outcomes from thecurrent studydo not consistently support improvements in body weight as, despite reduced body weight gain, no changes in lipid panel were recorded in NC-fed rats. Since the liver is a central organ in lipid metabolism, we suggest that the beneficial effect of a TRF regimen on lipid profiles in CAF-fed ratsmay be due to the enhancement of hepatic lipid metabolism. This is supported by several studies in which TRF has been found to enhance lipid oxidation over lipogenesis [8,12,21,55]. Indeed, TRF has been shown to reduce hepatic expression of PPARγ and SREBF1, the key lipogenic genes, in HFD-fed rats [55]. In addition, TRF regimen reduced lipid droplet associated and lipolysis inhibitor gene Cidec expressions, TG storage-associated protein CD36, and plasma TG marker ApoA 4 in HFD-fed mice [8].Decreased TG in the liver results in reduced serum levels of LDL and very-low-density lipoprotein (VLDL), leading to loss of transported cholesterol and TG within them [54]. The TC-lowering effect of TRF may also be attributed to enhanced cholesterol catabolism to bile acids biosynthesis in theliver, as indicated by increased expression of two rate-limiting enzymes of bile acids biosynthesis, namelyHmgcs 2 and Cyp7α1, in the liver of time-restricted mice [9].
It is commonly thought that unfavorable lipid profiles, synonymous with intake of dietary saturated fat, are the hallmark for the progression of cardiovascular diseases (CVD) and coronary heart diseases [56].Observations from epidemiological studies have confirmed that high TC, LDL-c, and TG and low HDL-c levels are associated with increased risk of CVD [56][57][58][59]. Together, thesedatacorroborate the present findings showing marked increases in AIP, AC, and CRI in CAF-A rats compared to NC-A rats. These indices are strong indicators of the CVD risk in clinical practices by their expression of imbalance between atherogenic and anti-atherogenic lipoproteins [60]. A recent human pilot study reported the effectiveness of 10 h time-restricted eating intervention for 12 wks to lower atherogenic lipids in obese subjects with metabolic syndrome, leading to improved cardiometabolic health [61]. Herein, TRF reduced the calculated values of AIP, AC, and CRIin CAF-fed rats, suggesting its very interesting cardioprotective potential. Since this is the first study to examine the effect of TRF onatherogenic indices in an animal model of obesity, these results should be considered preliminary; therefore, further in-depth investigationsare warranted to confirm the cardioprotective properties of TRF intervention in DIO models.
Since white fat browning is associated with body weight loss and metabolic improvements, we aimed to characterize the morphological and molecular features of iWAT. As hypothesized, we found that, in TRF rats, there was an obvious appearance of multilocular brown adipocytes and increased UCP1 and PGC1α protein expressions in iWAT of NC-and CAF-fed rats, indicating its potential to promote WAT browning.In one interesting study, intermittent fasting, an eating pattern that involves a 24 h fasting period on 2-3 days per week, or on alternate days, was reported to promote inguinal and gonadal fat browning in lean and obese mice [62]. In support of these findings, results from the current study showed that 8hTRF induced iWAT browning in normocaloric and CAF-fed rats. To the best of our knowledge, this report is the first to show the potential of TRF to promote WAT browning.Direct experimental evidence for the role of UCP1 in counteracting obesity has been reported previously in many studies. Of note, an animal model of genetic obesity using adipose tissue-targeted overexpression of UCP1 resulted in a reduced level of obesity [63]. In addition, UCP1 ablation resulted in increased obesity and metabolic deficiency in obesity-resistance mouse strain [64]. In humans, UCP1 gene polymorphism was significantly associated with increased body mass index in obese subjects [65]. Therefore, the present data concerning the iWAT browning and UCP1 expressionmay decipher the possible mechanism behind the body weight gain-lowering effect of a TRF regimen. Although the current data support the possibility that a TRF regimen could act as a WAT browning inducer, further in-depth studies are warranted to understand whether a TRF regimen could promote browningin other WAT depots and to investigate the expression of browning-related genes. Furthermore, the present study is limited by the fact that we did not measure whole-body energy expenditure due to practical reasons; hence, to what extent the effects of a TRF regimen on WAT browning andUCP1 content could be translated into increased whole-body energy expenditure is unclear, and future studies will address this shortcoming.A further limitation is that we did not measure a 24-h profile of locomotor activity. Therefore, the 5-min locomotor activity, which was measured at different times between groups, could induce time effects rather than TRF effects. However, as the locomotor activity for all groups was assessed within a short period of 2 h during the dark phase, it suggests that the locomotor activity was not affected by different times between groups but rather by TRF, although this hypothesis needs to be confirmed.
Conclusions
This present study showed that TRF regimen improved body weight gain and dyslipidemia in CAF-diet model and acted as white fat browning inducer with thermogenic properties, which may explain the anti-obesity effect of TRF. | 2020-07-29T13:06:22.641Z | 2020-07-23T00:00:00.000 | {
"year": 2020,
"sha1": "f0ad90c783ab5a450587594e0efa4f61ad874f99",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2072-6643/12/8/2185/pdf?version=1595497517",
"oa_status": "GOLD",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "b0f487ca18055849dc5a7dabc7d7e4a451034526",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
} |
18669267 | pes2o/s2orc | v3-fos-license | Dynamic Imaging of Individual Remyelination Profiles in Multiple Sclerosis
Background Quantitative in vivo imaging of myelin loss and repair in patients with multiple sclerosis (MS) is essential to understand the pathogenesis of the disease and to evaluate promyelinating therapies. Selectively binding myelin in the central nervous system white matter, Pittsburgh compound B ([11C]PiB) can be used as a positron emission tomography (PET) tracer to explore myelin dynamics in MS. Methods Patients with active relapsing‐remitting MS (n = 20) and healthy controls (n = 8) were included in a longitudinal trial combining PET with [11C]PiB and magnetic resonance imaging. Voxel‐wise maps of [11C]PiB distribution volume ratio, reflecting myelin content, were derived. Three dynamic indices were calculated for each patient: the global index of myelin content change; the index of demyelination; and the index of remyelination. Results At baseline, there was a progressive reduction in [11C]PiB binding from the normal‐appearing white matter to MS lesions, reflecting a decline in myelin content. White matter lesions were characterized by a centripetal decrease in the tracer binding at the voxel level. During follow‐up, high between‐patient variability was found for all indices of myelin content change. Dynamic remyelination was inversely correlated with clinical disability (p = 0.006 and beta‐coefficient = –0.67 with the Expanded Disability Status Scale; p = 0.003 and beta‐coefficient = –0.68 with the MS Severity Scale), whereas no significant clinical correlation was found for the demyelination index. Interpretation [11C]PiB PET allows quantification of myelin dynamics in MS and enables stratification of patients depending on their individual remyelination potential, which significantly correlates with clinical disability. This technique should be considered to assess novel promyelinating drugs. Ann Neurol 2016;79:726–738
demyelinating lesions and subsequent irreversible axonal degeneration. Considerable efforts have been made over past decades to develop immunoactive therapies. These have shown significant effects in reducing the number of clinical relapses; however, they have failed to demonstrate any efficacy in reducing or delaying long-term disability progression. 2 We are therefore assisting to a shift in therapeutic objectives from the development of new immune drugs toward the identification of therapeutic strategies to promote myelin regeneration, an endogenous process that is expected to restore secure and rapid conduction as well as to protect axons from degeneration. 3 In animal models, myelin regeneration is a very effective process that is activated by default in response to any sort of myelin damage, resulting in efficient reconstruction of the area of myelin loss. 4 To date, little is known about the dynamics of remyelination in patients with MS over the course of their disease. Sensitive and specific imaging tools designed to measure myelin in vivo are essential to understand how and why spontaneous remyelination succeeds or fails in MS, as well as to quantify the potential effects of new promyelinating therapies.
Advanced magnetic resonance imaging (MRI) sequences, such as magnetization transfer imaging, diffusion-weighted imaging, and T2 relaxometry, which are able to generate quantitative images exploiting physical properties of the brain parenchyma, have been proposed to gain indirect information about the myelin compartment in the human brain. 5 However, these techniques are not specific for myelin because they are affected to various extents by intra-and extracellular water, axons, edema, and inflammatory infiltration. Positron emission tomography (PET), which allows selective targets to be marked with radiolabeled compounds, is a promising alternative for myelin imaging. Following the pilot demonstration indicating that the stilbene Congo red derivative 1,4-bis(p-aminostyryl)-2-methoxy benzene, could be used as a myelin tracer suitable for PET imaging, 6 a similar affinity for myelin was reported for other stilbene derivatives. [7][8][9][10] These tracers, all previously known as amyloid markers, were hypothesized to bind to proteins characterized by a similar conformation contained in amyloid plaques and myelin. 11,12 On this basis, Pittsburgh compound B (PiB), a thioflavin compound binding to amyloid plaques, was also identified as a promising myelin tracer suitable for human PET studies. 13 In rodent demyelinating lesions, microPET with [ 11 C]PiB showed great sensitivity in capturing remyelination after demyelination. 10 Preliminary data obtained from humans further demonstrated that [ 11 C]PiB PET was sensitive enough to detect myelin loss in MS lesions. 13 A noninvasive parametric voxel-wise quantification procedure based on the extraction of reference regions using a supervised clustering algorithm showing higher reproducibility compared to previously used semiquantitative methods, has recently been shown to allow reliable longitudinal evaluation of [ 11 C]PiB binding in the white matter (WM) of healthy volunteers. 14 Here, we report the results of the first longitudinal study in which PET with [ 11 C]PiB was used to quantify in vivo myelin loss and regeneration in the WM lesions of patients with MS and to explore the clinical relevance of these processes.
Subjects
Twenty patients with relapsing-remitting MS according to the revised McDonald criteria 15 with at least one gadolinium-enhancing (Gd1) lesion (defined as all voxels localized inside a ring-enhancing lesion) with an inplane maximum diameter larger than 6 mm on MRI at study entry (13 women; mean age: 32.3 years; standard deviation [SD]: 5.6) and an age-and sex-matched group of 8 healthy volunteers (5 women; mean age: 31.6 years; SD, 6.3) signed written informed consent to participate in a clinical imaging protocol approved by the local ethics committee (Table 1).
Study Design
At inclusion, all patients were clinically assessed and scored using the Expanded Disability Status Scale 16 (EDSS) and the Multiple Sclerosis Severity Scale 17 (MSSS), which is designed to provide a measure of disease severity by adding the element of disease duration to the EDSS; all patients underwent MRI and PET scan. The 19 patients who completed the study (1 patient withdrew from the study after the first PET scan because of personal reasons) were randomly assigned to two subgroups to repeat the whole protocol after either 1 to 2 months (n 5 9) or 3 to 4 months (n 5 10) after inclusion to explore the best time interval in which to capture and quantify dynamic remyelination and demyelination. All healthy volunteers underwent a second PET scan 1 month after inclusion. No adverse event was observed during the study. resolution of 2.5mm full width at half maximum, with 25-cm axial and 31.2-cm transaxial fields of view. The 90-minute emission scan was initiated with a 1-minute intravenous bolus injection of [ 11 C]PiB (mean 5 358 6 34 MBq). Images were reconstructed using the threedimensional (3D) ordinary Poisson ordered subset expectation maximization algorithm with 10 iterations (considered to represent the appropriate trade-off between image resolution, quality of the data, and reliability of results). 14 An additional smoothing filter implementing the point spread function, which has been shown to be effective in reducing the effect of partial volume in PET data, 18 was applied to the 10-iteration reconstructed image. All the resulting dynamic PET images consisted of 25 time interval (time frames) images: six 1-minute frames for the initial 6 minutes (631), followed by 632-, 433-, 635-, and 3310-minute frames, with a voxel size of 1.22 3 1.22 3 1.22 mm. Interframe subject motion correction was applied by realigning each PET frame to a common reference space by using a previously described procedure. 19 Data were also corrected for carbon-11 decay. To avoid blood sampling in our patient cohort, we chose to quantify our PET scans using a reference region approach. A supervised clustering method already validated for [ 11 C]PiB was used for the extraction of the reference region time activity curve (TAC). 14,20,21 Briefly, this method consists of the multiple regression of all PET image voxel TACs on a predefined set of kinetic classes. After regression, all the voxels associated with the reference class with probability >90% are combined and the average of their TACs defines the reference input function. As described in the validation study performed using healthy controls, three classes were defined: normal gray matter (GM); high specific binding GM; and blood pool. The reference class of choice was normal GM. 14,21 Once the reference TAC was extracted, [ 11 C]PiB binding was calculated by applying a kinetic operator to the measured dynamic PET activity during the last 60 minutes of the PET acquisition. The Logan graphical reference method 22 was then applied at the voxel level on PET scans in native space and returned parametric maps of [ 11 C]PiB binding measured as the distribution volume ratio (DVR; defined as the ratio of the total distribution volume between the target and the reference region; Fig 1).
Native Space Postprocessing. For each patient, T2-w and T1 spin-echo images were registered onto the corresponding 3D-T1 MPRAGE image using a fully affine transformation (12 parameters; performed with FLIRT, FMRIB's Linear Image Registration Tool, which is part of FSL, FMRIB's Software Library; http://fsl.fmrib.ox.ac. uk/fsl/fslwiki/), and the derived transforms were used to align lesion masks with the individual 3D-T1 MPRAGE scans. After performing a "lesion-filling" procedure, 23 Standard Space Postprocessing. For each patient, the 3D-T1 MPRAGE scans acquired at baseline and at follow-up were first registered one onto the other with linear transformation, using FLIRT. The derived transforms were then deconstructed into two half-way transforms, which placed the two original images onto a subject-specific half-way space so that they both suffered the same amount of interpolation-related blurring. The average image of the two half-way 3D-T1 MPRAGE scans was then registered onto a standard brain image (MNI152) using a non-linear transformation (performed with FNIRT, FMRIB's Non-Linear Image Registration Tool, which is also part of FSL, FMRIB's Software Library; http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/). By linking derived transforms, all the previously generated ROIs were taken from the original 3D-T1 MPRAGE scans in native space and placed in standard space. In controls, 3D-T1 MPRAGE scans in native space were directly registered on a standard brain image (MNI152) using a nonlinear transformation performed with FNIRT, and the derived transforms were used to place ROIs in standard space.
PET-MRI COREGISTRATION. For all subjects, each DVR parametric map was linearly registered onto the corresponding 3D-T1 MPRAGE scan with FLIRT; the derived transform was first inverted and then used to align all ROIs with the DVR parametric map in native space at both time points.
Then, the previously generated transforms were linked with one another to align each DVR parametric map in native space to the standard brain image. They were then used to take all ROIs to the DVR parametric maps in standard space.
INDICES OF MYELIN CONTENT CHANGE.
The mean and SD of [ 11 C]PiB binding were computed for each subject and for each ROI from DVR parametric maps in native space at baseline.
Demyelinated voxels inside patients' T2-w lesion masks at baseline were defined on PET scans at study entry once registered to standard space, based on local thresholds in comparison with the PET scans of healthy controls after correction for the effect that the distance from the CSF has on local DVR values to minimize partial volume effects. 14 In particular, one given lesional voxel in patients was defined as demyelinated if its DVR value was 1 SD below the mean DVR value of all the voxels in healthy controls that were localized at the same distance from the CSF as the given voxel (and were therefore potentially affected to the same extent by partial volume). The percentage of demyelinated voxels over the total T2-w lesion load measured at baseline was calculated from each patient's individual DVR map at both time points in standard space.
The difference between the derived percentage at the second time point and the corresponding percentage at baseline was defined as the global index of myelin content change and reflected the subject-specific prevalence of either myelin loss (positive values) or myelin repair (negative values) over the follow-up interval. The index of dynamic demyelination, reflecting ongoing myelin loss, was defined as the proportion of normally myelinated voxels at baseline that were classified as demyelinated at the second time point. The index of dynamic remyelination, which reflected ongoing myelin repair, was defined as the proportion of lesional voxels classified as demyelinated at baseline and reached a myelin level within normal limits at the second time point.
Statistical Analysis
BASELINE GROUP-LEVEL ANALYSIS. To test for differences in mean DVR at baseline between patients' NAWM and GM and healthy volunteers' WM and GM, respectively, two multiple linear regressions were used, with group as the factor of interest and gender and age as covariates. A mixed-effect linear model was used to test for differences in mean DVR between NAWM and PWM, T2-w lesions, Gd1 lesions, and black holes, which included the subject as the random effect and age, gender, disease duration, and T2-w lesion load as covariates. The reduction in mean DVR between two ROIs was calculated and reported as the percent reduction in mean DVR.
To investigate the effect of each lesional voxel's distance in millimeters to the nearest point on the lesional border on its corresponding DVR value, a mixed-effect linear model was used in which the subject was included as the random effect, and age and gender were included as covariates.
CLINICAL CORRELATIONS. The correlation between the PET-derived indices of myelin content change and clinical scores was calculated using two separate multiple linear regressions: the first with absolute values of EDSS and the second with absolute values of MSSS as response variables, and age, gender, and T2-w lesion load as additional covariates. Given that three regressions were run for each of the two clinical variables (one for each of the three indices of myelin content change), a Bonferroniadjusted significance level for these tests has been set at p 5 0.017. To confirm significant results, ordinal regressions were also performed using a proportional-odds cumulative logit model, where EDSS and MSSS values were classified in three groups according to disease severity (mild disability 5 EDSS 1.5 or MSSS 3; moderate disability 5 EDSS 2-3 or MSSS 3-6; severe disability 5 EDSS >3 or MSSS >6) and entered as dependent variables, and age, gender, and T2-w lesion load were included as covariates. An exploratory post-hoc analysis was also performed using significant models, repeating regressions after including treatment status at study entry as an additional covariate.
IDENTIFICATION OF POTENTIAL CONTRIBUTORS TO THE INDICES OF DYNAMIC DEMYELINATION
AND REMYELINATION. Age, gender, disease duration, temporal distance between the two PET scans, treatment at study entry, and volume of T2-w and Gd1 lesion load were considered potential contributors and were included as independent variables in two separate multiple linear regressions, in which the indices of dynamic demyelination and remyelination were entered as dependent variables.
Unless specified otherwise, results are reported as significant at p < 0.05. Regression coefficients have been calculated as the mean change in the dependent variable for one unit of change in the predictor variable while holding other predictors in the model constant, and are reported below as "coefficients." . These box plots show that the lowest myelin content was detected in the "black holes", the hypointense lesions on T1 spin-echo scans that are known to represent the most severely demyelinated lesions in MS brains. A paired t test was used in a within-patient analysis to test for between-region differences in myelin content. (B) [ 11 C]PiB binding values are negatively correlated with the distance from the lesional border. Each point in this scatter plot diagram represents the mean DVR value (y axis) of all the voxels localized at any given distance in millimeters from the lesional border (x axis) in any given patient. Although voxels closer to the lesional border, on average, present higher myelin content values, those located far from the lesional border tend to present lower myelin content values. The correlation between each voxel's distance in millimeters from the lesional border and its corresponding DVR value, which was tested using a mixed-effect linear model in which the subject was included as random effect and age and gender were covariates, was highly significant (p 5 0.00001). DVR 5 distribution volume ratio; MS 5 multiple sclerosis; NAWM 5 normal-appearing white matter; ROI 5 region of interest; WM 5 white matter. SD 5 standard deviation; T2-w 5 T2weighted. Table 2). In T2-w, black holes, and Gd1 lesions, the mean percentage of demyelinated voxels over total lesion load were 53.5% (SD 5 9.1), 68.1% (SD 5 11.7), and 25.9% (SD 5 15.4), respectively.
A negative correlation between the distance of voxels from the corresponding lesional border and their DVR value was found (coefficient 5 -0.002; 95% CI 5 -0.025 to -0.016; p < 0.0001), with lower DVR values for the center of the lesions and higher DVR values for the periphery (Fig 2B).
Between-Patient Variability in Myelin Loss and Regeneration Levels
High between-patient variability was found, with the global index of myelin content change ranging from -11.91 to 113.46 across the cohort (Fig 3A). The index of dynamic demyelination also showed high variability, ranging between 8.4% and 20.5% (representative images in Fig 4; detailed values in Supplementary Table 3). A similar finding was observed for the index of dynamic remyelination, with variability ranging between 7.8% and 22.6% (Fig 4; Supplementary Table 3). These results suggest that patients could be classified into two groups: those showing high myelin regeneration potential and those with prevalent dynamic demyelination during the follow-up period (represented as blue and red bars in Fig 3A).
When the local distribution of dynamically changing voxels was investigated, 85.5% of dynamically demyelinating voxels and 81.7% of dynamically remyelinating voxels were localized at the periphery of T2-w lesions (defined as the 2-mm-thick lesional region inside the border).
The Extent of Myelin Regeneration Inversely Correlates With Clinical Disability
Because no patients were included during a clinical relapse, and because no significant change was detected in either EDSS or MSSS scores during follow-up, clinical scores measured at baseline were used for clinical correlations. At baseline, no correlation was found between the percentage of demyelinated voxels over total T2-w lesion Table 4).
The effect of the global index of myelin content change on EDSS only showed a trend toward statistical significance (p 5 0.045; beta-coefficient 5 0.53; Table 2; significance confirmed with ordinal regression: p 5 0.04). A significant association independent of gender and T2-w lesion load was found between MSSS and the global index of myelin content change, with the latter increasing in more disabled patients (p 5 0.002; beta- Table 3; significance confirmed with ordinal regression: p 5 0.043). A significant association was also detected between MSSS and age, with younger patients showing greater levels of disability (p 5 0.009; beta-coefficient 5 -0.54).
Although no significant association was found between the index of dynamic demyelination and EDSS (p 5 0.72; Fig 5A; Table 2), there was a trend toward a significant correlation with MSSS (p 5 0.08, betacoefficient 5 0.43; Fig 5C; Table 3). The index of dynamic remyelination was a significant explanatory factor for EDSS (p 5 0.006; beta-coefficient 5 -0.67, independent of age and gender; Table 2 Table 3; significance confirmed with ordinal regression: p 5 0.028), together with age (p 5 0.01; beta-coefficient 5 -0.51; Table 3), with patients characterized by higher remyelination potential presenting milder levels of disability. The results remaining significant after correction for multiple comparisons are indicated in Tables 2 and 3 (Bonferroni-adjusted remyelination was found for gender, disease duration, temporal distance between the two PET scans, volume of Gd1 lesions, and disease-modifying treatment (Table 5; Supplementary Material), whereas age was a significant contributor to the index of dynamic demyelination only, with older patients showing more-extensive dynamic demyelination (p 5 0.014; beta-coefficient 5 0.73; Table 5 in the Supplementary Material).
Discussion
In this longitudinal study, for the first time to our knowledge, we used high-resolution [ 11 C]PiB-PET images quantified with a novel non-invasive approach to visualize and measure lesional myelin loss and regeneration developing over a few months in a cohort of patients with active relapsing-remitting MS. Crosssectional analysis at baseline revealed a decrease in regional [ 11 C]PiB uptake that reflected a gradient in myelin concentration from normal-appearing to lesional tissue. Negative and positive changes in [ 11 C]PiB binding were measured in lesions during the follow-up period, which are suggestive of dynamic myelin loss and regeneration. Finally, strong correlations between the remyelination index derived from [ 11 C]PiB PET images and clinical scores were found, supporting the clinical rele-vance of the remyelination process in patients with MS. Interestingly, no significant effect of the temporal distance between the two PET scans (1-2 months or 3-4 months) on the indices of dynamic demyelination and remyelination was found in this study, which allowed us to analyze all the patients together. Mean binding of [ 11 C]PiB was progressively reduced from NAWM to WM lesions, which, in turn, were found to be characterized by a centripetal decrease in [ 11 C]PiB binding at the voxel level. These findings mirror postmortem evidence showing that the lesion edge surrounding areas are characterized by a partial decrease in myelin density with an intermediate level between discrete lesions and NAWM, [28][29][30][31] and that the severity of myelin destruction inside lesions becomes progressively worse from the periphery to the center. 28 As expected, the lowest mean values of [ 11 C]PiB binding were found in the so-called black holes, characterized by extensive demyelination and axonal loss. [32][33][34][35] Interestingly, Gd1 lesions showed intermediate [ 11 C]PiB binding values between NAWM and T2-w lesions, which could reflect the initial stage of myelin destruction in lesions that have recently appeared. 32 Increased permeability of the blood-brain barrier to the tracer in Gd1 lesional tissue could also influence the change in [ 11 C]PiB uptake found in active lesions. Exploring this hypothesis would require full modeling of [ 11 C]PiB kinetics, including the input function measured with arterial catheterization, 36 a procedure that we considered too invasive for patients with MS.
PET With [ 11 C]PiB Identifies Individual Profiles in Myelin Regeneration
In our longitudinal analysis, high between-patient variability was found for all the indices of dynamic myelin content change, particularly the index of dynamic remyelination, reflecting the heterogeneity in remyelination reported in postmortem cases of MS. 37 This result, if confirmed in larger studies with multiple follow-up time points, may support the notion of a patient-specific "remyelination profile," which determines the extent of myelin regeneration of each individual in response to a demyelinating insult. 37 Although age had a significant effect on dynamic demyelination and a borderline impact on dynamic remyelination, no significant effect on either index was demonstrated for disease duration. An age-dependent decrease in remyelination efficiency, possibly resulting from oligodendrocyte progenitor cells migrating and differentiating less efficiently with age in chronic lesions, has been previously highlighted. [38][39][40] However, the relationship between the efficiency of remyelination, age, and disease duration is still controversial, and our results contribute to an ongoing debate regarding the key question of whether the remyelination potential remains constant throughout life or is modified by aging and disease stage. 37,41 The majority of voxels dynamically changing their myelin content in either direction during the follow-up were localized in the peripheral lesional area in accordance to both neuropathological and ultra-high-field MRI studies demonstrating that the lesion edge corresponds to the expanding inflammation where new demyelination takes place and to the privileged lesional area where active remyelination occurs. 37,42 Dynamic Myelin Regeneration Correlates With Disability Scores In our patient cohort, the index of dynamic remyelination was strongly associated with clinical scores, suggesting that an efficient remyelination process taking place in an appropriate time window after a demyelinating insult is one of the discriminating factors in determining a better prognosis in MS, at least during the relapsing phase of the disease. Experimental studies have established that remyelination was effective not only during the shortterm recovery of neuronal function, but also, more importantly, in preventing subsequent axonal degeneration, possibly through mechanisms mediated by axonmyelin interaction. 43 However, the correlation we found between remyelination and clinical scores may also be influenced by the development of a fast and severe process of axonal degeneration that takes place immediately after acute demyelination in the subgroup of patients with a worse prognosis, which precedes a potential remyelinating process and, in turn, results in a reduced percentage of remyelinating voxels. The spatial and temporal link between myelin repair and axonal damage and loss has to be further explored in future longitudinal studies combining imaging measures specific for myelin and axonal damage. Results from such studies may also clarify whether patients with MS can be defined as "good remyelinators" or "bad remyelinators" throughout their entire disease course, or whether the balance between demyelination and remyelination in each individual changes over time, with one process prevailing over the other at different stages of disease or in different brain areas.
In conclusion, this longitudinal pilot study demonstrates that in vivo imaging of myelin loss and regeneration in MS can be successfully achieved with PET. However, this should be considered an exploratory study, because the number of subjects included was limited. Moreover, although the best available methodology has been applied in this study to minimize the partial volume effect, a possible residual effect of this kind cannot be completely excluded and should be taken into account in the interpretation of our results. As a chronic demyelinating disease occurring early in life, MS is among the most appropriate pathologies in which to investigate the remyelination process over time. Despite the reported limitations, results from this study support that such a regenerative process does occur during the relapsing phase of the disease and might promote neuroprotection and improve clinical prognosis. This imaging approach not only provides novel insight for understanding the pathophysiology of MS, but also provides perspective to enable stratification of patients based on their remyelination potential, thereby allowing clinical trials to be shortened and enabling measurement of the effects of novel drugs targeted at promoting myelin regeneration. | 2018-04-03T05:21:58.827Z | 2016-05-01T00:00:00.000 | {
"year": 2016,
"sha1": "f780126d3ead5d62d7ceba9e83e596e82cec56de",
"oa_license": "CCBY",
"oa_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/ana.24620",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "fbe6df67424f6c434c328796b941f458c854dc16",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
11054466 | pes2o/s2orc | v3-fos-license | Generalizing and Hybridizing Count-based and Neural Language Models
Language models (LMs) are statistical models that calculate probabilities over sequences of words or other discrete symbols. Currently two major paradigms for language modeling exist: count-based n-gram models, which have advantages of scalability and test-time speed, and neural LMs, which often achieve superior modeling performance. We demonstrate how both varieties of models can be unified in a single modeling framework that defines a set of probability distributions over the vocabulary of words, and then dynamically calculates mixture weights over these distributions. This formulation allows us to create novel hybrid models that combine the desirable features of count-based and neural LMs, and experiments demonstrate the advantages of these approaches.
Introduction
Language models (LMs) are statistical models that, given a sentence w I 1 := w 1 , . . . , w I , calculate its probability P (w I 1 ). LMs are widely used in applications such as machine translation and speech recognition, and because of their broad applicability they have also been widely studied in the literature. The most traditional and broadly used language modeling paradigm is that of count-based LMs, usually smoothed n-grams (Witten and Bell, 1991;Chen and Goodman, 1996). Recently, there has been a focus on LMs based on neural networks (Nakamura et al., 1990;Bengio et al., 2006;Mikolov et al., 2010), which have shown impressive improvements in performance over count-based LMs. On the other hand, these neural LMs also come at the cost of increased computational complexity at both training and test time, and even the largest reported neural LMs (Chen et al., 2015;Williams et al., 2015) are trained on a fraction of the data of their count-based counterparts (Brants et al., 2007).
In this paper we focus on a class of LMs, which we will call mixture of distributions LMs (MODLMs; §2). Specifically, we define MODLMs as all LMs that take the following form, calculating the probabilities of the next word in a sentence w i given preceding context c according to a mixture of several component probability distributions P k (w i |c): P (w i |c) = K k=1 λ k (c)P k (w i |c). (1) Here, λ k (c) is a function that defines the mixture weights, with the constraint that K k=1 λ k (c) = 1 for all c. This form is not new in itself, and widely used both in the calculation of smoothing coefficients for n-gram LMs (Chen and Goodman, 1996), and interpolation of LMs of various varieties (Jelinek and Mercer, 1980).
The main contribution of this paper is to demonstrate that depending on our definition of c, λ k (c), and P k (w i |c), Eq. 1 can be used to describe not only n-gram models, but also feed-forward (Nakamura et al., 1990;Bengio et al., 2006;Schwenk, 2007) and recurrent (Mikolov et al., 2010;Sundermeyer et al., 2012) neural network LMs ( §3). This observation is useful theoretically, as it provides a single mathematical framework that encompasses several widely used classes of LMs. It is also useful practically, in that this new view of these traditional models allows us to create new models that combine the desirable features of n-gram and neural models, such as: neurally interpolated n-gram LMs ( §4.1), which learn the interpolation weights of n-gram models using neural networks, and neural/n-gram hybrid LMs ( §4.2), which add a count-based n-gram component to neural models, allowing for flexibility to add large-scale external data sources to neural LMs.
We discuss learning methods for these models ( §5) including a novel method of randomly dropping out more easy-to-learn distributions to prevent the parameters from falling into sub-optimal local minima. Experiments on language modeling benchmarks ( §6) find that these models outperform baselines in terms of performance and convergence speed.
Mixture of Distributions LMs
As mentioned above, MODLMs are LMs that take the form of Eq. 1. This can be re-framed as the following matrix-vector multiplication: where p c is a vector with length equal to vocabulary size, in which the jth element p c,j corresponds to P (w i = j|c), λ c is a size K vector that contains the mixture weights for the distributions, and D c is a Jby-K matrix, where element d c,j,k is equivalent to the probability P k (w i = j|c). 2 An example of this formulation is shown in Fig. 1.
Note that all columns in D represent probability distributions, and thus must sum to one over the J words in the vocabulary, and that all λ must sum to 1 over the K distributions. Under this condition, the vector p will represent a well-formed probability distribution as well. This conveniently allows us to 2 We omit the subscript c when appropriate.
In the sequel we show how this formulation can be used to describe several existing LMs ( §3) as well as several novel model structures that are more powerful and general than these existing models ( §4).
n-gram LMs as Mixtures of Distributions
First, we discuss how count-based interpolated ngram LMs fit within the MODLM framework.
Maximum likelihood estimation: n-gram models predict the next word based on the previous N -1 words. In other words, we set c = w i−1 i−N +1 and calculate P (w i |w i−1 i−N +1 ). The maximum-likelihood (ML) estimate for this probability is where c(·) counts frequency in the training corpus.
Interpolation: Because ML estimation assigns zero probability to word sequences where c(w i i−N +1 ) = 0, n-gram models often interpolate the ML distributions for sequences of length 1 to N . The simplest form is static interpolation λ S is a vector where λ S,n represents the weight put on the distribution P M L (w i |w i−1 i−n+1 ). This can be expressed as linear equations (Fig. 2a) by setting the nth column of D to the ML distribution P M L (w i |w i−1 i−n+1 ), and λ(c) equal to λ S .
Probabilities p Static interpolation can be improved by calculating λ(c) dynamically, using heuristics based on the frequency counts of the context (Good, 1953;Katz, 1987;Witten and Bell, 1991). These methods define a context-sensitive fallback probability α(w i−1 i−n+1 ) for order n models, and recursively calculate the probability of the higher order models from the lower order models: (3) To express this as a linear mixture, we convert α(w i−1 i−n+1 ) into the appropriate value for λ n (w i−1 i−N +1 ). Specifically, the probability assigned to each P M L (w i |w i−1 i−n+1 ) is set to the product of the fallbacks α for all higher orders and the probability of not falling back (1 − α) at the current level: Discounting: The widely used technique of discounting (Ney et al., 1994) defines a fixed discount d and subtracts it from the count of each word before calculating probabilities: . Discounted LMs then assign the remaining probability mass after discounting as the fallback probability In this case, P D (·) does not add to one, and thus violates the conditions for MODLMs stated in §2, but it is easy to turn discounted LMs into interpolated LMs by normalizing the discounted distribution: , which allows us to replace β(·) for α(·) and P N D (·) for P M L (·) in Eq. 3, and proceed as normal. Kneser-Ney (KN;Kneser and Ney (1995)) and Modified KN (Chen and Goodman, 1996) smoothing further improve discounted LMs by adjusting the counts of lower-order distributions to more closely match their expectations as fallbacks for higher order distributions. Modified KN is currently the defacto standard in n-gram LMs despite occasional improvements (Teh, 2006;Durrett and Klein, 2011), and we will express it as P KN (·).
Neural LMs as Mixtures of Distributions
In this section we demonstrate how neural network LMs can also be viewed as an instantiation of the MODLM framework.
Feed-forward neural network LMs: Feedforward LMs (Bengio et al., 2006;Schwenk, 2007) are LMs that, like n-grams, calculate the probability of the next word based on the previous words. Given context w i−1 i−N +1 , these words are converted into real-valued word representation vectors r i−1 i−N +1 , which are concatenated into an overall representation vector q = ⊕(r i−1 i−N +1 ), where ⊕(·) is the vector concatenation function. q is then run through a series of affine transforms and nonlinearities defined as function NN(q) to obtain a vector h. For example, for a one-layer neural net- Count-based probabilities and J-by-J identity matrix (b) Neural/n-gram hybrid LMs Figure 3: Two new expansions to n-gram and neural LMs made possible in the MODLM framework work with a tanh non-linearity we can define where W q and b q are weight matrix and bias vector parameters respectively. Finally, the probability vector p is calculated using the softmax function p = softmax(hW s + b s ), similarly parameterized. As these models are directly predicting p with no concept of mixture weights λ, they cannot be interpreted as MODLMs as-is. However, we can perform a trick shown in Fig. 2b, not calculating p directly, but instead calculating mixture weights λ = softmax(hW s + b s ), and defining the MODLM's distribution matrix D as a J-by-J identity matrix. This is equivalent to defining a linear mixture of J Kronecker δ j distributions, the jth of which assigns a probability of 1 to word j and zero to everything else, and estimating the mixture weights with a neural network. While it may not be clear why it is useful to define neural LMs in this somewhat roundabout way, we describe in §4 how this opens up possibilities for novel expansions to standard models.
Recurrent neural network LMs: LMs using recurrent neural networks (RNNs) (Mikolov et al., 2010) consider not the previous few words, but also maintain a hidden state summarizing the sentence up until this point by re-defining the net in Eq. 5 as where q i is the current input vector and h i−1 is the hidden vector at the previous time step. This allows for consideration of long-distance dependencies beyond the scope of standard n-grams, and LMs using RNNs or long short-term memory (LSTM) networks (Sundermeyer et al., 2012) have posted large improvements over standard n-grams and feed-forward models. Like feed-forward LMs, LMs using RNNs can be expressed as MODLMs by predicting λ instead of predicting p directly.
Novel Applications of MODLMs
This section describes how we can use this framework of MODLMs to design new varieties of LMs that combine the advantages of both n-gram and neural network LMs.
Neurally Interpolated n-gram Models
The first novel instantiation of MODLMs that we propose is neurally interpolated n-gram models, shown in Fig. 3a. In these models, we set D to be the same matrix used in n-gram LMs, but calculate λ(c) using a neural network model. As λ(c) is learned from data, this framework has the potential to allow us to learn more intelligent interpolation functions than the heuristics described in §3.1. In addition, because the neural network only has to calculate a softmax over N distributions instead of J vocabulary words, training and test efficiency of these models can be expected to be much greater than that of standard neural network LMs.
Within this framework, there are several design decisions. First, how we decide D: do we use the maximum likelihood estimate P M L or KN estimated distributions P KN ? Second, what do we provide as input to the neural network to calculate the mixture weights? To provide the neural net with the same information used by interpolation heuristics used in traditional LMs, we first calculate three features for each of the N contexts w i−1 i−n+1 : a binary feature indicating whether the context has been observed in the training corpus (c(w i−1 i−n+1 ) > 0), the log frequency of the context counts (log(c(w i−1 i−n+1 )) or zero for unobserved contexts), and the log frequency of the number of unique words following the context (log(u(w i−1 i−n+1 )) or likewise zero). When using discounted distributions, we also use the log of the sum of the discounted counts as a feature. We can also optionally use the word representation vector q used in neural LMs, allowing for richer representation of the input, but this may or may not be necessary in the face of the already informative count-based features.
Neural/n-gram Hybrid Models
Our second novel model enabled by MODLMs is neural/n-gram hybrid models, shown in Fig. 3b. These models are similar to neurally interpolated n-grams, but D is augmented with J additional columns representing the Kronecker δ j distributions used in the standard neural LMs. In this construction, λ is still a stochastic vector, but its contents are both the mixture coefficients for the count-based models and direct predictions of the probabilities of words. Thus, the learned LM can use count-based models when they are deemed accurate, and deviate from them when deemed necessary.
This model is attractive conceptually for several reasons. First, it has access to all information used by both neural and n-gram LMs, and should be able to perform as well or better than both models. Second, the efficiently calculated n-gram counts are likely sufficient to capture many phenomena necessary for language modeling, allowing the neural component to focus on learning only the phenomena that are not well modeled by n-grams, requiring fewer parameters and less training time. Third, it is possible to train n-grams from much larger amounts of data, and use these massive models to bootstrap learning of neural nets on smaller datasets.
Learning Mixtures of Distributions
While the MODLM formulations of standard heuristic n-gram LMs do not require learning, the remaining models are parameterized. This section discusses the details of learning these parameters.
Learning MODLMs
The first step in learning parameters is defining our training objective. Like most previous work on LMs (Bengio et al., 2006), we use a negative log-likelihood loss summed over words w i in every sentence w in corpus W where c represents all words preceding w i in w that are used in the probability calculation. As noted in Eq. 2, P (w i = j|c) can be calculated efficiently from the distribution matrix D c and mixture function output λ c .
Given that we can calculate the log likelihood, the remaining parts of training are similar to training for standard neural network LMs. As usual, we perform forward propagation to calculate the probabilities of all the words in the sentence, back-propagate the gradients through the computation graph, and perform some variant of stochastic gradient descent (SGD) to update the parameters.
Block Dropout for Hybrid Models
While the training method described in the previous section is similar to that of other neural network models, we make one important modification to the training process specifically tailored to the hybrid models of §4.2. This is motivated by our observation (detailed in §6.3) that the hybrid models, despite being strictly more expressive than the corresponding neural network LMs, were falling into poor local minima with higher training error than neural network LMs. This is because at the very beginning of training, the count-based elements of the distribution matrix in Fig. 3b are already good approximations of the target distribution, while the weights of the single-word δ j distributions are not yet able to provide accurate probabilities. Thus, the model learns to set the mixture proportions of the δ elements to near zero and rely mainly on the count-based n-gram distributions.
To encourage the model to use the δ mixture components, we adopt a method called block dropout (Ammar et al., 2016). In contrast to standard dropout (Srivastava et al., 2014), which drops out single nodes or connections, block dropout randomly drops out entire subsets of network nodes. In our case, we want to prevent the network from overusing the count-based n-gram distributions, so for a randomly selected portion of the training examples (here, 50%) we disable all n-gram distributions and force the model to rely on only the δ distributions. To do so, we zero out all elements in λ(c) that correspond to n-gram distributions, and re-normalize over the rest of the elements so they sum to one.
Network and Training Details
Finally, we note design details that were determined based on preliminary experiments.
Network structures: We used both feed-forward networks with tanh non-linearities and LSTM (Hochreiter and Schmidhuber, 1997) networks. Most experiments used single-layer 200-node networks, and 400-node networks were used for experiments with larger training data. Word representations were the same size as the hidden layer. Larger and multi-layer networks did not yield improvements.
Training: We used ADAM (Kingma and Ba, 2015) with a learning rate of 0.001, and minibatch sizes of 512 words. This led to faster convergence than standard SGD, and more stable optimization than other update rules. Models were evaluated every 500k-3M words, and the model with the best development likelihood was used. In addition to the block dropout of §5.2, we used standard dropout with a rate of 0.5 for both feed-forward (Srivastava et al., 2014) and LSTM (Pham et al., 2014) nets in the neural LMs and neural/n-gram hybrids, but not in the neurally interpolated n-grams, where it resulted in slightly worse perplexities.
Features: If parameters are learned on the data used to train count-based models, they will heavily over-fit and learn to trust the count-based distributions too much. To prevent this, we performed 10-fold cross validation, calculating count-based elements of D for each fold with counts trained on the other 9/10. In addition, the count-based contextual features in §4.1 were normalized by subtracting the training set mean, which improved performance.
Experimental Setup
In this section, we perform experiments to evaluate the neurally interpolated n-grams ( §6.2) and neural/n-gram hybrids ( §6.3), the ability of our models to take advantage of information from large data sets ( §6.4), and the relative performance compared to post-facto static interpolation of already-trained models ( §6.5). For the main experiments, we evaluate on two corpora: the Penn Treebank (PTB) data set prepared by Mikolov et al. (2010), 3 and the first 100k sentences in the English side of the ASPEC corpus (Nakazawa et al., 2015) 4 (details in Tab. 1). The PTB corpus uses the standard vocabulary of 10k words, and for the ASPEC corpus we use a vocabulary of the 20k most frequent words. Our implementation is included as supplementary material.
Results for Neurally Interpolated n-grams
First, we investigate the utility of neurally interpolated n-grams. In all cases, we use a history of N = 5 and test several different settings for the models: Estimation type: λ(c) is calculated with heuristics (HEUR) or by the proposed method using feedforward (FF), or LSTM nets.
Distributions: We compare P M L (·) and P KN (·). For heuristics, we use Witten-Bell for ML and the appropriate discounted probabilities for KN.
Input features: As input features for the neural network, we either use only the count-based features (C) or count-based features together with the word representation for the single previous word (CR).
From the results shown in Tab. 2, we can first see that when comparing models using the same set of input distributions, the neurally interpolated model outperforms corresponding heuristic methods. We can also see that LSTMs have a slight advantage over FF nets, and models using word representations have a slight advantage over those that use only the count-based features. Overall, the best model achieves a relative perplexity reduction of 4-5% over KN models. Interestingly, even when using simple ML distributions, the best neurally interpolated n-gram model nearly matches the heuristic KN method, demonstrating that the proposed model can automatically learn interpolation functions that are nearly as effective as carefully designed heuristics. 5
Results for Neural/n-gram Hybrids
In experiments with hybrid models, we test a neural/n-gram hybrid LM using LSTM networks with both Kronecker δ and KN smoothed 5-gram distributions, trained either with or without block dropout. As our main baseline, we compare to LSTMs with only δ distributions, which have reported competitive numbers on the PTB data set (Zaremba et al., 2014). 6 We also report results for heuristically smoothed KN 5-gram models, and the best neurally interpolated n-grams from the previous section for reference.
The results, shown in Tab. 3, demonstrate that similarly to previous research, LSTM LMs (2) achieve a large improvement in perplexity over ngram models, and that the proposed neural/n-gram hybrid method (5) further reduces perplexity by 10-11% relative over this strong baseline.
Comparing models without (4) and with (5) the proposed block dropout, we can see that this method contributes significantly to these gains. To examine this more closely, we show the test perplexity for the 5 Neurally interpolated n-grams are also more efficient than standard neural LMs, as mentioned in §4.1. While a standard LSTM LM calculated 1.4kw/s on the PTB data, the neurally interpolated models using LSTMs and FF nets calculated 11kw/s and 58kw/s respectively, only slightly inferior to 140kw/s of heuristic KN. 6 Note that unlike this work, we opt to condition only on insentence context, not inter-sentential dependencies, as training through gradient calculations over sentences is more straightforward and because examining the effect of cross-boundary information is not central to the proposed method. Thus our baseline numbers are not directly comparable (i.e. have higher perplexity) to previous reported results on this data, but we still feel that the comparison is appropriate. (2), neurally interpolated ngrams (3), and neural/n-gram hybrid models without (4) and with (5) block dropout. three models using δ distributions in Fig. 5, and the amount of the probability mass in λ(c) assigned to the non-δ distributions in the hybrid models. From this, we can see that the model with block dropout quickly converges to a better result than the LSTM LM, but the model without converges to a worse result, assigning too much probability mass to the dense count-based distributions, demonstrating the learning problems mentioned in §5.2.
It is also of interest to examine exactly why the proposed model is doing better than the more standard methods. One reason can be found in the behavior with regards to low-frequency words. In Fig. 4, we show perplexities for words that appear n times or less in the training corpus, for n = 10, n = 100, n = 1000 and n = ∞ (all words). From the results, we can first see that if we compare the baselines, LSTM language models achieve better perplexities overall but n-gram language models tend to perform better on low-frequency words, corroborating the observations of Chen et al. (2015). The neurally interpolated n-gram models consistently outperform standard KN-smoothed n-grams, demonstrating their superiority within this model class. In contrast, the neural/n-gram hybrid models tend to follow a pattern more similar to that of LSTM language models, similarly with consistently higher performance.
Results for Larger Data Sets
To examine the ability of the hybrid models to use counts trained over larger amounts of data, we perform experiments using two larger data sets: WSJ: The PTB uses data from the 1989 Wall Street Journal, so we add the remaining years between 1987 and 1994 (1.81M sents., 38.6M words).
We incorporate this data either by training net parameters over the whole large data, or by separately training count-based n-grams on each of PTB, WSJ, and GW, and learning net parameters on only PTB data. The former has the advantage of training the net on much larger data. The latter has two main advantages: 1) when the smaller data is of a particular domain the mixture weights can be learned to match this in-domain data; 2) distributions can be trained on data such as Google n-grams (LDC2006T13), which contain n-gram counts but not full sentences.
In the results of Fig. 6, we can first see that the neural/n-gram hybrids significantly outperform the traditional neural LMs in the scenario with larger data as well. Comparing the two methods for incorporating larger data, we can see that the results are mixed depending on the type and size of the data being used. For the WSJ data, training on all data slightly outperforms the method of adding distributions, but when the GW data is added this trend reverses. This can be explained by the fact that the GW data differs from the PTB test data, and thus the effect of choosing domain-specific interpolation coefficients was more prominent.
Comparison with Static Interpolation
Finally, because the proposed neural/n-gram hybrid models combine the advantages of neural and ngram models, we compare with the more standard method of training models independently and combining them with static interpolation weights tuned on the validation set using the EM algorithm. Tab. 4 shows perplexities for combinations of a standard neural model (or δ distributions) trained on PTB, and count based distributions trained on PTB, WSJ, and GW are added one-by-one using the standard static and proposed LSTM interpolation methods. From the results, we can see that when only PTB data is used, the methods have similar results, but with the more diverse data sets the proposed method edges out its static counterpart. 7
Related Work
A number of alternative methods focus on interpolating LMs of multiple varieties such as in-domain and out-of-domain LMs (Bulyko et al., 2003;Bacchiani et al., 2006;Gülçehre et al., 2015). Perhaps most relevant is Hsu (2007)'s work on learning to interpolate multiple LMs using log-linear models. This differs from our work in that it learns functions to estimate the fallback probabilities α n (c) in Eq. 3 instead of λ(c), and does not cover interpolation of n-gram components, non-linearities, or the connection with neural network LMs. Also conceptually similar is work on adaptation of n-gram LMs, which start with n-gram probabilities (Della Pietra et al., 1992;Kneser and Steinbiss, 1993;Rosenfeld, 1996;Iyer and Ostendorf, 1999) and adapt them based on the distribution of the current document, albeit in a linear model. There has also been work incorporating binary n-gram features into neural language models, which allows for more direct learning of ngram weights (Mikolov et al., 2011), but does not afford many of the advantages of the proposed model such as the incorporation of count-based probability estimates. Finally, recent works have compared ngram and neural models, finding that neural models often perform better in perplexity, but n-grams have their own advantages such as effectiveness in extrinsic tasks (Baltescu and Blunsom, 2015) and better modeling of rare words (Chen et al., 2015).
Conclusion and Future Work
In this paper, we proposed a framework for language modeling that generalizes both neural network and count-based n-gram LMs. This allowed us to learn more effective interpolation functions for count-based n-grams, and to create neural LMs that incorporate information from count-based models.
As the framework discussed here is general, it is also possible that they could be used in other tasks that perform sequential prediction of words such as neural machine translation or dialog response generation (Sordoni et al., 2015). In addition, given the positive results using block dropout for hybrid models, we plan to develop more effective learning methods for mixtures of sparse and dense distributions. | 2016-06-01T23:26:20.000Z | 2016-06-01T00:00:00.000 | {
"year": 2016,
"sha1": "95cedaeb3178a4671703a05171a144e6b964a819",
"oa_license": "CCBY",
"oa_url": "https://www.aclweb.org/anthology/D16-1124.pdf",
"oa_status": "HYBRID",
"pdf_src": "ACL",
"pdf_hash": "c916f165ae1527c71b9834b90c2f8198d930f3a8",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Computer Science"
]
} |
17065043 | pes2o/s2orc | v3-fos-license | Quantum Cohomology of a Product
The operation of tensor product of Cohomological Field Theories (or algebras over genus zero moduli operad) introduced in an earlier paper by the authors is described in full detail, and the proof of a theorem on additive relations between strata classes is given. This operation is a version of the Kuenneth formula for quantum cohomology. In addition, rank one CohFT's are studied, and a generalization of Zograf's formula for Weil-Petersson volumes is suggested.
associativity equations is given in [D]. For the next definition, remind that M 0n denotes the moduli space of stable curves of genus zero with n labelled pairwise distinct points: see [Ke]. 0.3. Cohomological field theories. A structure of the tree level Cohomological Field Theory on (H, g) is given by a sequence of S n -covariant K-linear maps I n : H ⊗n → H * (M 0n , K), n ≥ 3 (0.2) satisfying the following set of identities (0.3). The values of I n generally are not homogeneous.
Consider an unordered partition σ : {1, . . . , n} = S 1 S 2 , |S i | ≥ 2. It defines an embedding of the boundary divisor ϕ σ : M 0,n 1 +1 × M 0,n 2 +1 → M 0,n . Over the generic point of this divisor the universal curve consists of two components, and the labelled points are distributed between them according to σ. The maps (0.2) must satisfy for all n ≥ 3 the relations: ϕ * σ (I n (γ 1 ⊗ . . . ⊗ γ n )) = ǫ(σ)(I n 1 +1 ⊗ I n 2 +1 )(⊗ j∈S 1 γ j ⊗ ∆ ⊗ (⊗ k∈S 2 γ k )) ( 0.3) where ∆ = Σ∆ a ⊗ ∆ b g ab is the Casimir element, and ǫ(σ) is the sign of the permutation induced on the odd arguments γ 1 , . . . , γ n . There are two other useful reformulations of CohFT. First,dualizing (0.2) we get a series of maps Thus any homology class in M 0,n+1 is interpreted as an n-ary opertaion on H. The relations (0.3) become identities between these operations whose totality means that H is given a structure of an algebra over the cyclic operad {H * (M 0,n+1 , K)} (see [GK]). Second, we can iterate the maps ϕ σ in order to study the restrictions of the classes I n (γ 1 ⊗ . . . ⊗ γ n ) to all boundary strata. These strata are naturally indexed by the (dual) trees of stable curves, which form a category. Both sides of (0.2) extend to the functors on this category, and (0.3) says that {I n } becomes a functor morphism. Below we will extensively use the combinatorial side of this picture explained in [KM].
This transition is markedly more difficult. If I n with the properties (0.5), (0.6) exist at all, they are defined uniquely, because iterating (0.3) one can calculate integrals of I n over all boundary strata of M 0n . Namely, let τ be the dual tree of a stable curve C of genus zero with n marked points. Let us remind that the set of vertices V τ consists of irreducible components of C, edges E τ are (in a bijection with) double points of C, tails T τ (one vertex edges) are marked points. Incidence relations between V τ , E τ , T τ reflect those in C. A flag of τ is a pair (vertex, incident edge or tail); the set of flags is denoted F τ . Let M τ ⊂ M 0n be the submanifold parametrizing curves of type τ and their specialisations. We have M τ ∼ = v∈V τ M |v| . Here |v| = |F τ (v)|, F τ (v) is the set of flags incident to v. The homology classes of all M τ generate H * (M 0n , K), and we have from (0.3): (0.7) We use here the formalism of tensor products indexed by arbitrary finite sets and interpret the argument of the r.h.s. (0.7) as an element of H ⊗F τ and v∈V τ Y |v| as a function on H ⊗F τ . (0.3) is a particular case of (0.7) for an one-edge tree.
In order to use (0.7) for construction of {I n } it remains to check that the r.h.s. of (0.7) satisfies all the linear relations between the classes of M τ in H * (M 0n ). This was made in sec. 8 of [KM] modulo the theorem, describing these relations and proved in § 2 of the present paper. 0.6. The tensor product. Let {H ′ , g ′ , I ′ n } and {H ′′ , g ′′ , I ′′ n } be two CohFT's. Put H = H ′ ⊗ H ′′ and g = g ′ ⊗ g ′′ . We can define a CohFT on (H, g) by where ǫ(γ ′ , γ ′′ ) is the standart sign in superalgebra, and ∧ is the cup product in H * (M 0n , K). One can easily check (0.3). Although (0.8) looks very simple on the level of full CohFT's, it cannot be trivially restricted to calculate the potential. In fact, for potential we must know only the higher dimensional component of I n (see (0.5)), but it involves components of all degrees of I ′ n , I ′′ n (see (0.8)), which are given by (0.7) as functionals on the boundary homology classes.
The only remaining obstruction to calculating Φ of {I n } is thus our incomplete understanding of the cup product in terms of dual boundary classes. Ralph Kaufmann obtained a fairly simple formula for the intersection indices of strata of complementary dimensions. With his permission, we reproduce it in § 2. But at the moment we are unable to invert this Gram matrix (of redundant size). 0.7. Application to quantum cohomology. In the application of this formalism to the quantum cohomology of V we have: H = H * (V, K), g = Poincaré pairing, and M 0n I n (γ 1 ⊗ . . . ⊗ γ n ) is the (appropriately defined) number of "stable maps" (C, x 1 , . . . , x n ; ϕ) where C is a curve of genus 0 with n marked points x 1 , . . . , x n , ϕ : C → V is a morphism, such that any connected component contracted by ϕ (together with its special points) is stable. These Gromov-Witten invariants are of primary interest in enumerative geometry, and we expect that the tensor product of CohFT's described above furnishes an algorithm for calculating them on V × W from those on V and W . Strictly speaking, this must be proved starting with a geometric construction of GW-invariants. In concrete examples it suffices to simply check the coincidence of a finite set of coefficients of the two potentials using The First Reconstruction Theorem 3.1 of [KM]. For example, Φ P 1 ×P 1 3 actually coincides with the tensor square of the simple potential Φ P 1 . Namely, if ∆ 0 ∈ H 0 (P 1 ) and ∆ 1 ∈ H 1 (P 1 ) are respectively the fundamental class and the dual class of a point, we have where N (a, b) is the number of rational curves of bidegree (a, b) on P 1 × P 1 passing through 2a+2b−1 points in general position. As is remarked in [DFI], the structure of N (a, b) which can be derived from the associativity equations looks simpler than that of GW-numbers for P 2 , probably because of this tensor product property.
In the remaining part of this Introduction, we summarize our notation and conventions about the combinatorics of trees and the (co)homology of M 0n . The next section is devoted to the multiplicative properties of strata classes. In § 2 we prove the completeness of the standard linear relations between them. The last § 3 discusses rank one CohFT's. 0.8. Partitions and trees. As in [KM], a tree τ for us is a system of finite sets (V τ , E τ , T τ ), V τ = ∅, with appropriate incidence relations defining F τ (see 0.5 above for notation). A structure of S-tree on τ (where S is a finite set) is given by a bijection T τ → S. Sometimes we identify T τ with S using this bijection. A tree τ is stable if |v| ≥ 3 for all v ∈ V τ . Most our trees are stable.
A stable S-tree τ corresponds to a (family of) stable curve(s) of genus 0 with points labelled by elements of S. One-edge S-trees are in a bijection with unordered partitions of S into two subsets σ : S = S 1 S 2 ; stability means that |S i | ≥ 2; the tails marked by S i belong to the vertex v i , i = 1, 2. We will systematically identify such partitions with the corresponding trees. For such a σ and, say, four elements i, j, k, l ∈ S we use a notation like ijσkl to imply that {i, j} and {k, l} belong to different parts of σ.
If a(σ, τ ) = 3, we sometimes call σ and τ compatible. A family of 2-partitions {σ 1 , . . . , σ m } is called good, if for all i = j, σ i and σ j are compatible. S-trees form objects of several catgeories differing by the size of their morphism sets. The most useful morphisms f : τ → σ contract several edges and tails of τ : f induces a surjection V τ → V σ and injections E σ → E τ , T σ → T τ ; labelling sets for τ and σ may differ. We will mostly consider morphisms of S-trees identical on S (pure contractions of edges, or S-moprhisms). The one-vertex tree is a final object in the category of S-trees and S-morphisms. If a direct product σ ×τ of two S-trees in this category exists, it comes equipped with two contractions σ × τ → σ and σ × τ → τ . A geometrically nice case is when |E σ×τ | = |E σ | + |E τ |. E.g. for one-edge trees this is the case when a(σ, τ ) = 3. For a(σ, τ ) = 4, σ × τ does not exist, and for σ = τ we have σ × σ = σ.
A few more words about the geometry of an individual tree τ . Any flag f = (v, e), v ∈ V τ , e ∈ E τ or e ∈ T τ defines a complete subgraph β(f ) of τ which we will call the branch of f . If e ∈ T τ , β(f ) consist of the vertex v and tail e. Generally, β(f ) includes v, e, and all edges, vertices and tails that can be reached from v by a no-return path starting with f . We denote by T τ (f ) the tails belonging to β(f ), and by S(f ) their labels (if τ is a labelled tree). 0.9. Moduli spaces. For a finite set S, |S| ≥ 3, M 0S parametrizes stable curves of genus zero with a family of pairwise distinct points labelled by S. More generally, for a stable S-tree τ, M τ parametrizes such curves with dual graph (isomorphic to) τ ; M 0S corresponds to the one-vertex S-tree. Any pure contraction τ → σ bijective (but not necessarily identical) on S induces a morphism M τ → M σ . In this way {M σ } form a topological cyclic operad (see [GK]), and {H * (M 0S )} form a linear cyclic operad.
If an S-morphism of S-trees τ → σ exists, it is unique, and M τ → M σ is a closed embedding whose image is called a (closed) stratum of M σ . In particular, In particular, stable one-edge S-trees (and stable 2-partitions) σ bijectively correspond to the boundary divisors. 0.10. Keel's presentation. Fixing S, |S| ≥ 3, we denote by {D σ |σ stable 2-partitions of S} a family of commuting independent variables. Put F S = K [D σ ] (F S = K for |S| = 3). We consider F S as a graded polynomial ring, deg D σ = 1. Define the ideal I S ⊂ F S by means of the following generators: a) For each pairwise distinct foursome i, j, k, l ∈ S: For each pair σ, τ with a(σ, τ ) = 4: Since M 0S is a smooth manifold whose homology and cohomology is generated by algebraic classes, on which homological and rational equivalences coincide, (0.12) describes the homology and the Chow ring as well. In addition, Keel's presentation is very convenient for describing the operadic structure maps. E.g. bijections S ′ → S ′′ (relabelling of points) translate simply by the respective relabelling of D σ 's. For the remaining morphisms, see the next section. § 1. Boundary strata and the multiplicative structure of H * (M 0S ) In particular, D σ and 1 are good.
For each e ∈ E τ , denote by σ(e) the 2-partition of S corresponding to the one edge S-tree obtained by contracting all edges except for e. Then is a good monomial.
Figure 1. Arrows symbolize branches
Let u be the remaining vertex of e, w that of e ′ . Let S ′ be the set of all tails of τ belonging to the branches starting at u but not with a flag belonging to e 0 ; similarly, let S ′′ be the set of all tails of τ belonging to the branches that start at w but not with a flag belonging to e ′ . Finally, let T be the set of all tails on the branches at v 0 , . . . , v r not starting with the flags in e ′ 0 , . . . , e ′ r+1 (we identify tails with their labels). Since τ is stable, all three sets S ′ , S ′′ and T are non-empty. Finally It follows that a(σ(e), σ(e ′ )) = 3 so that m(τ ) is a good monomial. We put m(τ ) = 1, if |E τ | = 0.
Proposition.
For any 1 ≤ r ≤ |S| − 3, the map τ −→ m(τ ) establishes a bijection between the set of good monomials of degree r in F S and stable S-trees τ with |E τ | = r modulo S-isomorphism. There are no good monomials of degree > |S| − 3.
Proof. For r = 0, 1 the assertion is clear. Assume that for some r ≥ 1 the map τ −→ m(τ ) is surjective on good monomials of degree r. We will prove then that it is surjective in the degree r + 1.
Let deg m ′ = r + 1. Choose a divisor D σ of m ′ which is extremal in the following sense: one element, say S 1 , of the partition σ = {S 1 , S 2 } is minimal in the set of all elements of all 2-partitions σ ′ such that D σ ′ divides m ′ . Put m ′ = D σ m. Since m is good of degree r, we have m = m(τ ) for some stable S-tree τ . We will show that m ′ = m(τ ′ ) where τ ′ is obtained from τ by inserting a new edge with tails marked by S 1 at an appropriate vertex v ∈ V τ . (This means that vice versa, there exists a contraction τ ′ → τ of one edge to the vertex having incident flags S 1 and (half of) this edge).
Figure 2. Inserting edge at a vertex
First we must find v in τ . To this end, consider any edge e ∈ E τ and the respective partition σ(e) : {S ′ e , S ′′ e } (obtained by contracting all edges except for e). Since a({S 1 , S 2 }, {S ′ e , S ′′ e }) = 3 and S 1 is minimal, one sees that exactly one of the sets {S ′ e , S ′′ e } strictly contains S 1 . Let it be S ′′ e . Orient e by declaring that the direction from the vertex (corresponding to) S ′ e to S ′′ e is positive. We claim that with this orientation, for any w ∈ V τ there can be at most one edge outgoing from w. In fact, if τ contains a vertex w with two positively oriented flags f 1 and f 2 , then S 1 must be contained in the two subsets of S, S(f 1 ) and S(f 2 ). But their intersection is empty. It follows that there exists exactly one vertex v ∈ V τ having no outgoing edges. Moreover, S 1 is contained in the set of labels of the tails at v by construction. If we now define τ ′ by inserting a new edge e ′ at v so that σ(e ′ ) = σ, we will clearly have m ′ = m(τ ′ ). If r ≤ |S| − 4, the tree τ ′ cannot be unstable because, first, |S 1 | ≥ 2, and second, at least two more flags converge at |v|: otherwise the unique incoming edge would produce the partition {S 1 , S 2 } = σ which would mean that D σ divides already m(τ ).
For r = |S|−3, this argument shows that m ′ cannot exist because all the vertices of τ have valency three.
Assume that this has been checked in degree ≤ r and that deg τ 1 = deg τ 2 = r + 1. Choose an extremal divisor D σ of m(τ 1 ) = m(τ 2 ) as above and contract the respective edges of τ 1 , τ 2 getting the trees τ 1 and τ ′ 2 are isomorphic by the inductive assumption. This isomorphism respects the marked vertices v ′ 1 , v ′ 2 corresponding to the contracted edges because as we have seen they are uniquely defined. Hence it extends to an S-isomorphism τ 1 → τ 2 .
where the direct product is the categorical one in the category of S-trees and Smorphisms. We can identify E σ×τ with E σ E τ , and p 1 : σ × τ → σ (resp. p 2 : ρ × τ → τ ) contracts edges of the second factor (resp. of the first one). b). There exists a divisor D σ ′ of m(τ ), |E σ ′ | = 1, such that a(σ, σ ′ ) = 4. Then Then let e ∈ E τ be the edge corresponding to σ; v 1 , v 2 its vertices, (v i , e) the corresponding flags.
We will write several different expressions for D σ m(τ ) mod I S , corresponding to various possible choices of unordered pairs of distinct flags {ῑ,j} ⊂ Notice that because of stability the set of such choices is non-empty. where tr T,e (τ ) is the tree obtained from τ by "transplanting all branches starting in T to the middle point of the edge e." (An empty sum is zero).
Remark.
We can also describe tr T,e (τ ) as a result of inserting an extra edge instead of the vertex v 1 (resp. v 2 ) and putting the branches T to the common vertex of the new edge and e, similarly to what we have done in the proof of Proposition 1.3. There exists a unique S-morphism tr T,e (τ ) → τ contracting one edge.
Proof.
We choose pairwise distinct labels on the chosen branches i ∈ S(ῑ), j ∈ S(j), k ∈ S(k), l ∈ S(l) and then calculate the element (see (0.10)) ( 1.5) Since ijσkl, for all terms D ρ of the second sum in (1.5) we have a(σ, ρ) = 4 so that D ρ m(τ ) ∈ I S . Among the terms of the first sum, there is one D σ . If ijρkl and ρ = σ, then D ρ cannot divide m(τ ). Otherwise ρ would correspond to an edge e ′ = e, but the 2-partition of such an edge cannot break {i, j, k, l} into {i, j} and {k, l} as a glance to a picture of τ shows. It follows that D ρ m(τ ) = m(ρ × τ ) as in ( 1.2). The projection ρ × τ → τ contracts the extra edge onto a vertex that can be only one of the ends of e, otherwise, as above, the condition ijρkl cannot hold. It should be clear by now that ρ × τ must be one of the trees tr T,e (τ ), and that each tree of this kind can be uniquely represented as ρ × τ for some ρ with ijρkl. But from (1.5) it follows that where τ (resp. ρ) runs over stable 2-partitions of S 1 {.} (resp. S 1 {.}).
1.6. Multiplication formulas II. It may be more convenient to have formulas independent of arbitrary choice of {i, j, k, l}. One way to achieve this is to average (1.4) over all possible choices. We will illustrate this procedure by calculating D 2 σ and some of the "tautological classes." Let σ be a stable 2-partition {T 1 , T 2 } of S.
Proof. We first write |T 1 | 2 |T 2 | 2 identities (1.4) for all possible choices of i, j ∈ T 1 , k, l ∈ T 2 , then sum them up and change the summation order by first choosing subsets T ⊂ T 1 or T 2 , and then i, j or k, l in the complement.
Tautological classes. These classes τ
is the universal curve, x i : M 0n → C is the i-th section, and T * x i is the relative cotangent sheaf to C at x i . In order to calculate τ Notice that deg m(τ ) = |S| − 3 iff |v| = 3 for all v ∈ V τ , and M τ is a point in this case. We put σ 1 , σ 2 = M 0S m(σ 1 )m(σ 2 ) and set to calculate this intersection index for the case when deg m(σ 1 ) + deg m(σ 2 ) = |S| − 3. Generally, we will write m instead of M 0S m. The following notions and results are due to Ralph Kaufmann. We can assume that all pairs of different divisors of m(σ 1 ) and m(σ 2 ) are compatible, otherwise σ 1 , σ 2 = 0. Put τ = σ 1 × σ 2 in the category of Smorphisms. This is a tree with a marked subset of edges E corresponding to D σ 's whose squares divide m(σ 1 )m(σ 2 ). We denote by δ the subgraph of τ consisting of E and its vertices.
Consider an orientation of all edges of δ. Call it good if for all vertices v of τ , the number of ingoing edges equals |v| − 3, where |v| means the valency in τ . Notice that for v / ∈ V δ we interpret this as |v| = 3.
Proposition (R. Kaufmann).
There cannot exist more than one good orientation of δ. If there is none, we have σ 1 , σ 2 = 0. If there is one, we have For a proof, see Appendix.
Remark. Notice that (1.9) depends only on τ whereas E influences only the existence of the good orientation. Curiously, (1.9) coincides with the virtual Euler characteristics of the non-compact moduli space M τ = v∈V τ M 0,|v| . We do not know why this is so. § 2. Boundary strata and the additive structure of H * (M 0S ) 2.1. Basic linear relations. Let |S| ≥ 4. Consider a system (τ, v,ī,j,k,l) where τ is an S-tree, v ∈ V τ is a vertex with |v| ≥ 4 andī,j,k,l ∈ F τ (v) are pairwise distinct flags (taken in this order). Put T = F τ (v) \ {ī,j,k,l}. For any ordered 2-partition of T , α = {T 1 , T 2 }, (one or both T i can be empty) we can define two trees τ ′ (α) and τ ′′ (α). The first one is obtained by inserting a new edge e at v ∈ V with branches {ī,j, T 1 } and {k,l, T 2 } at its edges. The second one corresponds similarly to {k,j, T 1 } and {ī,l, T 2 }. We remind that S(ī) is the set of labels of tails belonging to the branch ofī.
Theorem.
All linear relations modulo I S between good monomials of degree r + 1 are spanned by the relations (2.1) for |E τ | = r.
Proof. For r = 0 this holds by definition of I S . Generally, denote by H * S the linear space, generated by the symbols µ(τ ) for all classes τ of stable S-trees modulo isomorphisms, satisfying the analog of the relations (2.1) Denote by 1 the symbol µ(ρ) where ρ is one-vertex tree.
Main Lemma.
There exists on H * S a structure of H * S -module given by the following multiplication formulas reproducing (1.1), (1.3) and (1.4): , and e corresponds to σ. The notation in (2.5) is the same as in ( 1.4).
(2.5) is well defined.
The r.h.s. of (2.5) formally depends on the choice ofī,j,k,l. We first check that different choices coincide modulo (2.2). It is possible to pass from one choice to another by replacing one flag at a time. So let us consider i ′ =ī,j,k,l and write the difference of the right hand sides of the relations (2.5) written for (τ, v,ī,j,k,l) and (τ, v,ī ′ ,j,k,l). The terms corresponding to those T that do not contain {ī,ī ′ } cancel. This includes all terms with T ⊂ T 2 . The remaining sum can be rewritten as where now T can be empty. We contend that (2.6) is of the type (2.2). More precisely, consider any of the trees tr T ∪{ī ′ } (τ ), tr T ∪{ī} (τ ) and contract the edge whose vertices are incident to the flagsī,j,ī ′ . We will get a tree σ and its vertex v ∈ V τ . The pair (σ, v) up to a canonical isomorphism does not depend on the transplants we started with. In F σ (v) there are flagsī,j,ī ′ and one more flag whose branch contains both k and l and which we denoteh. Then (2.6) is −r(σ, v,ī,j,ī ′ ,h) (see (2.2)). This is illustrated by the Figure 6.
The case a(σ 1 , σ 2 ) = 4 and D σ 1 /m(τ ) is treated in the same way. Hence from this point on we can and will in addition assume that a(σ 1 , σ 2 ) = 3.
Consider first the subcase u 1 = v 1 . Choose someī,j ∈ F τ (u 2 ) andk,l ∈ F τ (u 1 ) in such a way thatl starts a path leading from u 1 to v 1 . Use theseī,j,k,l in a formula of the type (2.5) to calculate D σ 2 µ(τ ) and then D σ 1 (D σ 2 µ(τ )), that will insert an edge instead of the vertex v 1 which survives in all the transplants entering D σ 2 µ(τ ). Then calculate D σ 2 (D σ 1 µ(τ )) by first inserting the edge at v 1 , and then constructing the transplants not movingī,j,k,l. Since by our choice ofl we never transplant the branch containing v 1 , the two calculations will give the same result. Now let v 1 = u 1 . Let {S 1 , S 2 } be the 2-partition of S corresponding to σ 1 . Since σ × τ exists, {S 1 , S 2 } is induced by a partition of F τ (v 1 ) =S 1 S 2 . We denote bȳ S 2 the part to which the flag (v 1 = u 1 , e 2 ) belongs. LetT =S 2 \ {(v 1 = u 1 , e 2 )}. This set is non-empty because otherwise e 2 would correspond to {S 1 , S 2 } and we would have σ 1 = σ 2 . Takeī,j ∈ F 2 (τ ),k ∈S 1 andl ∈ T : see Figure 7.
Figure 7
Now consider D σ 2 (D σ 1 µ(τ )) and D σ 1 (D σ 2 µ(τ )). To calculate the first expression we form a sum of transplants of σ 1 × τ . To calculate the second one, we form transplants of τ , and then insert an edge at v 1 = u 1 .
The transplants corresponding to the branches at u 2 will be the same in both expressions. The transplants corresponding to the subsets T ⊂T \ {l} will also be the same. In addition, the second expression will contain the terms −D σ 1 (µ(tr T,e 2 (τ ))) where T ∩S 1 = ∅. But each such term vanishes. In fact, consider the 2-partition ρ = {R 1 , R 2 } of S corresponding to the edge of tr T,e 2 (τ ) containing the flag (v 1 = u 1 , e 2 ), and let k, l ∈ R 1 . A glance to the third tree of the Figure 7 shows that a(ρ, σ 1 ) = 4, because ift ∈ T ∩S 1 , t ∈ S(t), then ktσ 1 il and klρit. Hence the extra terms are irrelevant.
The case when D σ 2 does not divide m(τ ) is treated in the same way. It remains to consider the last possibility. iv) D σ 1 and D σ 2 divide m(τ ), a(σ 1 , σ 2 ) = 3. Denote by e 1 (resp. e 2 ) the edge corresponding to σ 1 (resp. σ 2 ). Let u 1 , u 2 (resp. v 1 , v 2 ) be the vertices of e 1 (resp. e 2 ) numbered in such a way that there is a path from u 2 to v 1 not passing through e 1 , e 2 (the case u 2 = v 1 is allowed). To calculate the multiplication by D σ 1 chooseī,j ∈ F u 1 (τ ) \ {(u 1 , e 1 )},l on the path from Figure 8).
Figure 8
The critical choice here is that ofl andī ′ . It ensures that calculating D σ 1 (D σ 2 µ(τ )) and D σ 2 (D σ 1 µ(τ )) we will get the same sum of transplanted trees. This ends the proof of (2.7).
Consider the smallest connected subgraph in τ containing the flags i, j, k, l. The Figure 9 gives the following exhaustive list of alternatives. Paths from i to j and from k to l: i) have at least one common edge; ii) have exactly one common vertex; iii) do not intersect.
Consider them in turn. i). Let e be an edge common to the paths ij and kl. Denote by ρ the respective 2-partition. Then ikρjl or ilρkj. Therefore any summand of R ijkl annihilates D ρ so that R ijkl µ(τ ) = 0 in view of (2.4).
ii). Let v be the vertex common to the paths ij and kl. Then exactly the same calculation as in the proof of the Proposition 2.1.1 shows that iii). This is the most complex case. Let us draw a more detailed picture of τ in the neighborhood of the subgraph we are considering ( Figure 10).
Figure 10
Let v 1 be the vertex on the path ij which is connected by a sequence of edges e 1 , . . . , e m (m ≥ 1) with the vertex v m on the path kl so that e a has vertices (v a , v a+1 ) in this order. Let T a be the set of flags at v a which do not coincide with i,j,k,l, and do not belong to e a−1 , e a .
Consider any summand D σ of R ijkl . If jkσil, then D σ µ(τ ) = 0 because each edge e a determines a partition ρ of S such that ijρkl. From now on we assume that ijσkl. Then D σ µ(τ ) can be nonzero if one of the two alternatives holds: a). For some v a , there exists a partition T a = T ′ a T ′′ a ,(with |T ′ a | ≥ 1, |T ′′ a ≥ 1, except for the case a = 1 where T ′ 1 can be empty, and a = m where T ′ m can be empty) such that the following two sets form the 2-partition corresponding to σ. In this case and σ × τ is obtained by inserting a new edge at v a and by distributing T ′ a and T ′′ a at different vertices of this edge. b). For some e a , the two sets form the 2-partition corresponding to σ.
In this case D σ divides m(τ ), and in order to calculate D σ µ(τ ) using a formula of the type (2.5) we must first choose two pairs of flags at two vertices of v a .
Contributions from a) and b) come with opposite signs, and we contend that they completely cancel each other.
To see the pattern of the cancellation look first at the case a) at v 1 . It brings (with positive sign) the contributions corresponding to the following trees. Form all the partitions T 1 = T ′ 1 T ′′ 1 such that T ′′ 1 = ∅, where T 1 = F τ (v 1 ) \ {ī,j, (v 1 , e 1 )}. Transplant all T ′′ 1 -branches to the midpoint of e 1 . Denote the new vertex v ′ 1 . The result is drawn as Figure 11.
Figure 11
Now consider the terms of the type b) for the edge e 1 . If m = 2, we choose for the calculation of D σ 1 µ(τ ) (where σ 1 corresponds to e 1 ) the flagsī,j,k,l. If m > 2, we choose the flagsī,j, (v 2 , e 1 ), t ∈ T 2 . Then we get the sum of two contributions. One will consist of the trees obtained by transplanting branches at v 1 . They come with negative signs and exactly cancel the previously considered terms of the type a). If m = 2, the second group will cancel the terms of the type a) coming from v 2 .
Consider a somewhat more difficult case m > 2. Then this second group of terms comes from the trees indexed by the partitions T 2 = T ′ 2 T ′′ 2 , t ∈ T ′′ 2 , T ′ 2 = ∅. Branches corresponding to T ′ 2 are transplanted to the midpoint v ′ 1 of the edge e 1 . These terms come with negative signs: see Figure 12.
Figure 12
These trees in turn cancel with those coming from the terms of the type a) at the vertex v 2 with positive sign. However, there will be additional terms of the type a) for which t ∈ T ′ 2 . They will cancel with one group of transplants contributing to D σ 2 µ(τ ) where σ 2 corresponds to the edge e 2 of the Figure 10, if for the calculation of D σ 2 µ(τ ) one uses (2.5) with the following choice of flags: (v 2 , e 1 ), t at one end, (v 3 , e 1 ), some t ′ ∈ T 3 at the other end (this last choice must be replaced byk,l, if m = 3).
The same pattern continues until all the terms cancel.
2.2.6. Compatibility with relations (2.2). By this time we have checked that the action of any element of F S /I S on the individual generators µ(τ ) of H * S is well defined modulo the span I * S of relations (2.2). It remains to show that the subspace in ⊕ τ Kµ(τ ) spanned by these relations is stable with respect to this action. But the calculation in the proof of the Proposition 2.2.1 shows that r(τ, v,ī,j,k,l) ≡ m(τ )r ijkl mod I * S , where r ijkl is obtained from R ijkl by replacing m(σ) with µ(σ). To multiply this by any element of H * S we can first multiply it by m(τ ), then represent the result as a linear combination of good monomials, and finally multiply each good monomial by r ijkl . The result will lie in I * S .
This finishes the proof of the Main Lemma and the Theorem 2.2. §3. Cohomological Field Theories of rank 1.
Notation. Rank of a CohFT on (H, g) is the (super)dimension of H.
In this section we consider the case dim H = 1. To slightly simplify notation let us assume that all square roots exist in K. Then H = K∆ 0 , g(∆ 0 , ∆ 0 ) = 1, ∆ = ∆ ⊗2 0 ∈ H ⊗2 . The basic vector ∆ 0 is defined up to a sign. We will consider its choice as a rigidification and without further ado call such rigidified theories simply CohFT's of rank one.
A structure of CohFT on (H, ∆ 0 ) boils down to a sequence of cohomology classes (generally non-homogeneous) where φ σ : M 0,n 1 +1 × M 0,n 2 +1 → M 0n is the embedding of the boundary divisor corresponding to a partition σ (see (0.3)). Put The tensor product formula (0.8) becomes Here are some simple consequences of (3.2) and (3.3).
3.1.1.
The theory c n = c 0 n = [M 0n ] for n ≥ 3 is the identity with respect to the tensor product.
The theory {c
Any invertible theory is a tensor product of one of the type {c n (t)} and one with c (0) n = 1 for all n, and this decomposition is unique. 3.1.4. Assume that c (0) n = 1 and put λ n = log c n ∈ H * (M 0n , K) S n . Then (3.2) becomes ϕ * σ (λ n ) = λ n 1 +1 ⊗ 1 + 1 ⊗ λ n 2 +1 , Vice versa, any sequence of classes λ n ∈ H * (M 0n , K) S n satisfying (3.4) gives rise to a CohFT of rank 1, c n = exp λ n . We can say that {λ n } forms a logarithmic CohFT of rank 1.
3.1.5.
There is a canonical bijection between the set of the isomorphism classes of CohFT's of rank 1 and the set of infinite sequences (C 3 , C 4 , . . . ) ∈ K ∞ given by In fact, this is a particular case of the equivalence stated in 0.4 and 0.5, because any formal series in one variable Φ(x) = C n n! x n satisfies the associativity equations.
where [m(τ )] ∈ H * (M 0n ) is the image of the good monomial m(τ ). We do not know nice formulas for the tensor product in terms of coordinates (C i ). The main goal of this section is to show that there are natural coordinates defined geometrically that are simply additive with respect to the tensor multiplication of invertible theories. This is a reformulation of certain identities from [AC].
Mumford classes.
Consider the universal curve p n : X n → M 0n and its structure sections s i : M 0n → X n , i = 1, . . . , n. Let x i ⊂ X n be the image of s i , ω the relative dualizing sheaf on X n . For a = 1, 2, . . . put ω n (a) := p n * c 1 (ω( ( 3.9) It is proved in [AC] that for any a ≥ 1 (in fact, a = 0 as well) {ω n (a) | n ≥ 3} satisfy (3.4) i.e., form a logarithmic field theory. Hence we can construct an infinitedimensional family of invertible theories of rank one: where P n is a universal polynomial. Hence it remains to check that the coefficient at s n−3 does not vanish. But this follows from the well known fact that ω( Remark. The theories we are considering here are tree level ones in the terminology of [KM]. The general definition of a CohFT given there involves maps I g,n : H ⊗n → H * (M g,n ) for all stable pairs (g, n). The classes ω n (a) given by (3.7) can be automatically defined in this larger generality, and the extension of the property (3.4) is proved in [AC] for all (g, n). Therefore formulas (3.8) in fact define full (any genus) rank one theories.
However we do not know whether Theorem 3.2.3 extends to the general case because it is unclear whether functions C g,n := M g,n c g,n form a coordinate system on the space of full rank 1 theories. In fact, they probably do not, because of the presence of non-trivial cusp classes in H * (M g,n ) having vanishing restrictions to the boundary. ( 3.9) We expect that (3.9) satisfies some interesting differential equations encoding recursive relations between the numbers M 0n a ω n (a) m a , am a = n − 3.
Some partial results are given below. It is even possible that such equations for arbitrary genus are implicit in the relations (conjectured by Witten and proved by Di Francesco, Itzykson, and Zuber) between the numbers denoted in [AC] W (m a ),n a ψ n a i , where W (m a ),n are certain combinatorial classes defined in terms of ribbon graphs. In fact, it is conjectured in [AC] that the dual cohomology classes of W (m a ),n can be expressed as a≥2 ω n (a − 1) m a (2 a (2a − 1)!!) m a m a ! plus terms of lower order and boundary terms (our ω n (a) are denoted k a in [AC]). We hope to return to this problem elsewhere. Here we will treat the case when only one of the coordinates s a is non-zero.
3.4.
Weil-Petersson theory. The noncompact moduli spaces M 0n possess a canonical Weil-Petersson hermitean metric. It is singular on the boundary, but its Kähler form extends to a closed L 2 -current on M 0n thus defining a real cohomology class ω W P n ∈ H 2 (M 0n ) S n (see [W] and [Z]). In [AC] this class is identified as ω W P n = 2π 2 ω n (1). (3.10) The additivity property (3.4) for ω W P n was used by P. Zograf ([Z]) in order to calculate the WP-volumes of M 0n . In our framework, he calculated the coefficients of the potential of the theory {exp(π −2 ω W P n )}. 23 3.5. Weil-Petersson potential. The C n -coordinate of {exp(π −2 ω W P n )} is 1 π 2(n−3) M 0n (ω W P n ) n−3 (n − 3)! = v n (n − 3)! in the notation of [Z]. P. Zograf proved that v 4 = 1, v 5 = 5, v 6 = 61, v 7 = 1379, and generally ( 3.11) The potential (0.6) of the theory is therefore We can rewrite (3.11) as a differential equation for Φ W P (x). Following [M], put Then we have x(x − g)g ′′ = x(g ′ ) 2 + (x − g)g ′ . Define for each n ≥ 3 an S n -invariant function A n on the set of isomorphism classes of n-trees σ with the following property:
A generalization of
σ:|E σ |=a A n (σ)σ, τ = 1 if |E τ | = n − 3 − a and ∃v ∈ V τ , |v| = a + 3, 0 otherwise, (3.13) where σ, τ is defined by (1.9). Presumably, (3.12) can be calculated inductively using the definition (3.7). The solvability of (3.13) will be shown below. To find an explicit solution one has to invert the Poincaré pairing matrix restricted to the S n -invariant part of H * (M 0n ). Hopefully, a version of the Proposition 1.7.1 can be used to do this. | 2014-10-01T00:00:00.000Z | 1995-02-13T00:00:00.000 | {
"year": 1995,
"sha1": "db4213559bc85e380fe897818e0bd72288d90dd2",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/q-alg/9502009",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "c5454b668786da96bc96eab7d1c602e7f6c405ae",
"s2fieldsofstudy": [
"Mathematics"
],
"extfieldsofstudy": [
"Mathematics"
]
} |
24461734 | pes2o/s2orc | v3-fos-license | One-Year Outcomes of an Integrated Multiple Sclerosis Disease Management Program
BACKGROUND: Multiple sclerosis (MS) is associated with high total health care cost, the majority of which is attributable to medications. Patients with MS are less likely to experience relapses, emergency department (ED) visits, and hospitalizations when they are adherent to disease-modifying treatments. Disease management programs are hypothesized to improve medication adherence thereby improving clinical and economic outcomes. OBJECTIVE: To evaluate the clinical and economic effects of a specialty pharmacy and chronic disease management program for patients with MS from a health plan perspective. METHODS: This study was a retrospective analysis using prescription drug claims, medical claims, and electronic medical record information (2013-2015) 1 year before and after enrollment in the disease management program for members with 24 months of continuous health plan coverage. Medication adherence was calculated using proportion of days covered (PDC). Relapse rate was defined as an MS outpatient visit associated with a corticosteroid dispense within 7 days of the visit or an MS hospitalization. Disease progression was assessed using the Modified Expanded Disability Status Scale (mEDSS). Resource use included outpatient visits, ED visits, and hospitalizations. Cost information was collected as health plan-paid amount and was reported in 2013 U.S. dollars. RESULTS: The analysis included 377 patients (mean age 55 years, 76.4% female). After enrollment in the program, 78.7% of the study group had a PDC of ≥ 0.80 compared with 70.0% before enrollment (P < 0.001). There was no difference in MS relapse rate (0.25 after vs. 0.45 before, P = 0.11) or mEDSS score (3.77 after vs. 3.76 before, P = 0.19). Health care resource utilization was minimal and did not change significantly throughout the study period: mean outpatient visits (13.09 after vs. 13.78 before, P = 0.69); mean ED visits (0.18 after vs. 0.16 before, P = 0.60); and mean hospitalizations (0.12 after vs. 0.12 before, P = 1.00). This nonsignificant finding remained when the analysis was limited to MS-related visits only. Average annual health plan spend per patient on MS medications significantly increased ($55,835 after vs. $40,883 before, P < 0.001). CONCLUSIONS: Specialty pharmacy and chronic disease management for patients with MS can increase the proportion of patients adherent to medication. The increase in health plan spend on MS medications is not offset by savings in health care resource utilization.
M ultiple sclerosis (MS) is associated with high total health care cost. In 2014, annual direct costs for the MS population were estimated to be $24,327 higher than the non-MS population, with the majority of costs attributable to medications. 1,2 The cost for MS medications have increased on average 8%-36% annually, a rate significantly higher than prescription drug inflation. 3 Disease-modifying therapies (DMTs) are the current standard of care and help reduce the frequency and intensity of relapses and delay progression to disability. Patients with MS are less likely to experience relapses, emergency department (ED) visits, and hospitalizations when they are adherent to DMTs. [4][5][6][7] Patients are considered adherent, and obtaining benefit from medication therapy, when the medication possession ration (MPR) or proportion of days covered (PDC) is > 0.80. However, only 52%-62% of patients with MS meet this adherence threshold. 8 Disease management programs that promote disease awareness, medication adherence, and side effect management have been adopted to improve care provided to patients with MS. 9 To date, 3 studies have evaluated the effects of such programs. These studies included injectable MS medications and • Patients with multiple sclerosis (MS) are less likely to experience relapses, emergency department visits, and hospitalizations when they are adherent to disease-modifying treatments. • Disease management programs aim to enhance medication adherence, disease awareness, and side effect management; however, literature describing outcomes beyond improved adherence report mixed findings.
What is already known about this subject
• The benefits of a specialty chronic disease management program are diminished for patients already on therapy with high baseline adherence. • Increased adherence drives subsequent increases in health planpaid amount on MS medications. • Contributing to increases in health plan-paid amount to a lesser extent are unit price inflation and progression from less costly medication to more costly medication.
Study Measures
The primary study outcome was the change in proportion of patients who were adherent to MS medication as measured using PDC. 16 The purpose of this study was to evaluate the 1-year outcomes of a specialty pharmacy and chronic disease management program. Findings from this study add to existing literature in 4 important ways. First, the analysis included health record information in addition to administrative claims data. Second, the analysis was conducted after several oral medications became available in the United States. Previous studies on disease management programs for MS included primarily injectable medications (1 study also included fingolimod). [10][11][12][13] Evaluation of oral medications is important as market share shifts from injectables to oral medications. 14 Third, this study applied a recently validated claims-based definition of MS relapse rate. 15 Fourth, the study assessed adherence using PDC, a measure known to more accurately measure adherence than MPR when patients fill medications early. 16
■■ Methods Disease Management Program and Study Design
This study was conducted at Group Health Cooperative (GHC), an integrated health plan and delivery system in the Pacific Northwest. To ensure that its patient population received the most value from MS medication therapy, GHC implemented a specialty pharmacy and chronic disease management program in June 2014. Clinical pharmacists, trained in MS management, provided first fill education and follow-up management within 2 weeks of initiating a medication and every 6 months thereafter. For patients with established therapy before the program, a clinical review was conducted by the pharmacist to ensure that the patients were tolerating their medications and taking them as prescribed. Specialty pharmacy technicians monitored adherence monthly and coordinated dispensing and delivery of medication. Clinical concerns identified by the pharmacy technicians were escalated to a pharmacist for appropriate triage as needed.
To be included in the study, patients were required to be enrolled in the specialty pharmacy and chronic disease management program between June 1 and September 30, 2014, and have continuous health plan enrollment 12 months before and after enrollment. Enrollment in the program was offered to all patients with MS requiring medication therapy. A prepost design was used for all outcomes. GHC prescription drug claims, medical claims, and electronic medical records (EMR) were analyzed retrospectively. The pre-period was defined as the 365 days before the patient's enrollment date, and the postperiod was defined as the 365 days after the patient's enrollment date. This was a quality improvement project exempt from approval by an internal institutional review board. 15,[18][19][20] Outcome measures of resource utilization included outpatient visits, ED visits, and inpatient stays. MS-related resource utilization was identified by the presence of a primary diagnosis code for MS (ICD-9-CM code 340). 19,21,22 Cost of care was collected as the plan-paid amount and reported in 2013 U.S. dollars.
Data Analysis
Baseline characteristics and demographics were described using frequency counts and percentages for categorical variables, and means and standard deviations (SD) for continuous variables. A McNemar's test was conducted to compare the proportion of patients adherent to MS medications 1 year before enrollment in the program, with the proportion of patients adherent 1 year after enrollment. A two-sided paired t-test was conducted to compare all remaining outcome variables 1 year before enrollment in the programs, with the outcome variables 1 year after enrollment. A one-sided Kruskal-Wallis test was conducted to compare the effect of baseline characteristics on the difference between the pre-period and post-period. All tests were at a significance level of P < 0.05 and were carried out using R statistical software (R Foundation for Statistical Computing, Vienna, Austria).
■■ Results Baseline Characteristics
During the study period, 758 patients at GHC had an MS diagnosis and were eligible for enrollment in the specialty chronic disease management program. Of all enrolled patients, 377 had 24 months of continuous health plan enrollment and were included in the analysis. The mean age of the study population was 55 (SD ± 11.8) years, and the majority of patients were female (76.4%) and white (77.2%; Table 1). The baseline mean mEDSS score was 3.8. Most patients had relapsingremitting disease (78.3%) and were established on medication before enrollment (87.5%). At baseline, 15% of the population had a depression diagnosis documented in the EMR.
Outcome Measures
The proportion of patients who were adherent to MS medication increased after enrollment in the disease management program (78.7% 1 year after enrollment vs. 70.0% 1 year before enrollment, P < 0.001). In addition, mean (SD) PDC was higher after enrollment ( Table 2). This nonsignificant finding remained when the analysis was limited to MS-related visits only.
The mean (SD) health plan-paid amount for MS medications was significantly greater 1 year after enrollment ($55,835 [± $24,408] after vs. $40,883 [± $25,900] before, P < 0.001; Figure 1). MS medications accounted for 80.9% of the total cost of care in the 1 year before enrollment in the program and 85.4% in the 1 year after enrollment. The mean (SD) health plan-paid amount for MS-related health care cost did not significantly change 1 year after enrollment compared with 1 year before enrollment ($2,253 [± $6,112] after vs. $2,345 [± $7,052] before, P = 0.730; Figure 1).
■■ Discussion
Adherence is necessary to obtain and optimize the expected beneficial effects from drug therapy. Of the 87.5% of the population established on DMT before enrollment in the program, a higher proportion were adherent to medication therapy (70%) at baseline compared with proportions previously reported in the literature (52%-62%). Furthermore, the mean PDC of 0.85 was higher than the 0.70-0.83 previously reported in the literature. 8,23 A regression analysis of mean change in PDC demonstrated that higher baseline adherence rates were associated with lower change in mean PDC (P < 0.001, analysis not shown). High baseline adherence may be due to the study setting. It is possible patients received education and care coordination in an integrated delivery system that could not be accounted for in the analysis.
Adherence rates > 0.80 are associated with lower rates of severe relapse, hospitalization, ED visits, and lower total costs. [5][6][7] The 70% of the population with a PDC > 0.80 before enrollment in the program may have experienced the benefits of high adherence before the program intervention. This is supported by the low baseline relapse rates, few hospitalizations and ED visits at baseline, and low baseline nonpharmacy-related cost. We hypothesize that the large proportion of patients established on DMT with high adherence before enrollment in the program limited the ability to increase adherence further and subsequently to improve clinical and economic outcomes. That might explain why the 2 percentagepoint increase in mean PDC after enrollment was statistically significant but did not result in improved outcomes at 1 year.
It is possible that the relapse rate remained unchanged because of high baseline adherence, as previously mentioned. In addition, short duration of follow-up and low baseline relapse rates may have limited the ability to observe changes in our population. MS is a chronic and intermittent disease. With nearly 80% of our study population diagnosed with relapsedremitting disease, where relapses occur at a rate of 0.54 relapses per year on average, it may be that these patients would not have had a relapse during the study period, let alone after an intervention aimed at reducing the number of relapses was implemented. 24 Furthermore, the mean relapse rate observed in this analysis (0.25 1 year after enrollment) is similar to rates published in recent clinical trials for oral MS medications (0.16-0.29 in treated arms and 0.36-0.40 in placebo arms). 25-28 A low initial relapse rate suggests a smaller opportunity for improvement through a disease management program.
Inclusion of mEDSS enhanced the ability to assess disease progression beyond what is available in claims data. Absence of significant progression in disability suggests that disease in this study population was well controlled. However, it is possible that patients observed change in disability status not measurable by this questionnaire because of high baseline scores. The mean baseline mEDSS score in the study population approached 4, compared with 2-2.5 in recent clinical trials. [25][26][27][28] At higher initial values, the mEDSS sensitivity to change is lower. 29 This study population represented an older sampling of MS patients, with a mean age of 55 years. It is possible that an older population would have more accumulated disability reflected in a higher baseline mEDSS score. Also, 13% of the study population had nonrelapsing disease. While DMT is frequently continued in patients who progress from relapsing-remitting MS to secondary-progressive MS, DMT does not have an indication for primary-progressive MS. Regression analysis demonstrated that primary-progressive MS was not associated with mean change in any outcome (analysis not shown). However, the small sample size (n = 5) precluded meaningful analysis.
The health plan-paid amount per patient on MS-related medications increased significantly from $40,833 in the year before program enrollment to $55,835 in the year after program enrollment (2013 U.S. dollars). This increase could be attributable to 4 factors: increased adherence, unit price inflation, patients switching from less expensive to more expensive medication, or patients new to medication therapy. We conducted a post hoc analysis of patients who remained on the same medication throughout the entire study period (n = 281). The increase in mean change in health plan-paid amount per patient on MS-related medication diminished from $15,002 in the study population to $10,067 in this subset of patients. We found that 45% of the increase in cost was attributable to inflation, and 55% of the increase in cost was attributable to increased adherence. Specifically, the mean number of 30-day supply medication fills increased from 10 to 11 during the 1 year after enrollment. For the 37 patients that switched medication, 87.5% of them moved from less costly injectable medications (interferon or glatiramer) to more costly oral agents (dimethyl fumarate, fingolimod, or teriflunomide). 3 The increase in plan-paid amount for medications was not offset by savings in other areas of health care resource utilization. While Tan et. al. previously described a decrease in MS-related medical spending during a 12-month follow up that was not replicated in this analysis, they also observed an increase in total cost of care. 12 The decrease in MS hospitalization observed by Tan et al. may have contributed to the observed decrease in MS-related medical spending. Because our analysis did not observe a change in resource utilization, one would not expect to observe a change in spending in relation to those resources. The difference in findings may be partly due to our relatively small sample size, high baseline adherence rate, low MS hospitalization rate, and low baseline nonpharmacy cost throughout the study period.
Limitations
This analysis must be interpreted in the context of several limitations. The pre-post study design does not rule out the possibility that the observed effect is attributable to variables that may change over time. Claims data do not account for disease progression, adverse drug events, or disability status. We attempted to address this by evaluating changes in the mEDSS scores. However, it is possible that our methods were not able to detect the occurance of disease and disability progression. Also, this analysis did not capture change in patient out-ofpocket cost. In response to rising prescription drug prices, many payers have increased cost shares for patients, which may influence their ability to obtain and persist on medication therapy. We estimate that this trend affected roughly 20% of the GHC MS population during the study period.
This study also has several strengths. The inclusion of health record information allowed evaluation of mEDSS scores in addition to outcomes assessed using administrative claims. Also, this study was conducted after several oral medications became available in the United States, allowing for interpretation of program effects on injectable and oral medications. Finally, we applied a recently validated claims-based definition of MS relapse rate, as opposed to relying on patient-reported relapse rate. 15
■■ Conclusions
This retrospective pre-post analysis suggests that a specialty pharmacy and chronic disease management program can improve the proportion of patients adherent to MS medications as measured by PDC > 0.80. Study findings suggest that the benefits of a specialty chronic disease management program are diminished for patients already on therapy with high baseline adherence. Focusing a program on patients new to therapy, or with low baseline adherence, could yield clinically meaningful results. In the short term, an increase in medicationrelated costs can be expected as patients increase their use of prescribed DMTs. Future research should focus on the patient experience, including patient satisfaction, absenteeism, and the effect of side-effects on daily life. | 2018-05-03T01:02:46.580Z | 2018-04-25T00:00:00.000 | {
"year": 2018,
"sha1": "badd62bdf673822fbd51652805820f3ded77b976",
"oa_license": "CCBY",
"oa_url": "https://www.jmcp.org/doi/pdf/10.18553/jmcp.2018.24.5.458",
"oa_status": "BRONZE",
"pdf_src": "PubMedCentral",
"pdf_hash": "dd3a22ea8d7935f5e005d792cf8cfeb0de749dc0",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
129944376 | pes2o/s2orc | v3-fos-license | Estimation of cumulative number of post-treatment Lyme disease cases in the US, 2016 and 2020
Background Lyme disease (LD) is an infectious multi-system illness caused by the bacterial genus Borrelia and spread by bites of infected ticks. Although most patients are successfully treated by timely antibiotic therapy, it is broadly accepted that a sizeable number of patients experience treatment failure and continue to suffer long-term, debilitating symptoms, including pain, fatigue, cognitive dysfunction and other symptoms. This is known as post-treatment LD (PTLD), for which diagnosis is not standardized and treatment remains controversial. The prevalence and societal burden of PTLD is unknown. Methods In an effort to help characterize the LD landscape, we estimated the number of PTLD cases in the US in 2016 and 2020 using Monte-Carlo simulation techniques, publically-available demographic datasets, uncertainty in the inputs and realistic assumptions about incidence and treatment failure rates. Results Depending on the input assumptions, PTLD prevalence estimates for 2016 ranged from 69,011 persons (95% CI 51,796 to 89,312) to 1,523,869 (CI 1,268,634 to 1,809,416). Prevalence in 2020 is predicted to be higher than 2016, and may be as high as 1,944,189 (CI 1,619,988 to 2,304,147) cases. Conclusions The cumulative prevalence of PLTD in the United States is estimated to be high and continues to increase. These findings will be of interest to epidemiologists and health economists studying disease burden in the US and elsewhere, and justify funding to study PTLD diagnosis and treatment.
Background
Lyme disease (LD) is an emerging multi-system infectious illness, caused by bacteria of the Borrelia genus, and is transmitted by bites of black-legged Ixodes ticks [1]. The disease is the most frequent vector-borne illness in the United States, and is increasingly reported in Europe and Asia [2]. Researchers at the United States Centers for Disease Control and Prevention (CDC) have recently calculated annual US incidence at over 329,000 cases; however, true incidence may be higher [3,4]. Early LD is a flu-like illness with or without the pathognomonic erythema migrans rash within thirty days of the bite. Some patients experience early neurological (facial palsy, meningitis) and cardiac (heart block, arrhythmias) manifestations.
Recent studies have shown that despite timely and standard antibiotic treatment of acute LD, treatment failure results in the chronic condition known as Post-Treatment Lyme Disease, or PTLD [5,6]. PTLD is characterized by incapacitating fatigue, pain and neurocognitive dysfunction that persist for more than 6 months [5,7]. Symptoms can be intermittent or constant, and are often subjective and varied in nature. There is no single case definition for PTLD, and its diagnosis is often made based on exclusion of other conditions, such as tick-borne co-infections [8]. However, LD and its sequelae are responsible for significant numbers of school and work absences, and are estimated to cost more than $1 billion per year for healthcare in the US [9,10]. Moreover, the precise societal burden of PTLD specifically, both in real time and in projected amounts, has never been adequately quantified. Without such information, it is impossible to effectively dedicate healthcare resources, education and research efforts. As a critical first step in addressing this absence, we developed a statistical framework to estimate the prevalence of PTLD in the US by using Monte-Carlo simulation techniques and applying it under six settings representing various assumptions about the course of the US Lyme disease epidemic. The settings make use of available published data and demonstrate the wide range of measures provided in the literature. As a means of estimating the cumulative US cases, we chose two specific index dates on which to focus: 2016 and 2020.
Methods
To estimate cumulative cases, we used inputs on Lyme disease incidence and treatment failure rates commonly reported in the peer-reviewed literature. We base our simulations on the technique presented by Crouch et al. [11] and utilize the probability distribution function from classical statistics that most closely represents the type of data used in each step of the simulation, i.e. the binomial distribution for "yes/no" data and the negative binomial distribution for overdispersed count data. We also use a simple deterministic approach as a check for the simulations. The six settings represent three scenarios for LD incidence and two treatment failure rates within each scenario. While any one of the settings may currently be considered more realistic, more research is being conducted that may change our understanding. At that time, an improved framework could be developed that incorporates all uncertainty into one simulation.
Input data sources Duration of epidemic
The first cases of Lyme disease in the United States were diagnosed in the late 1970s in the state of Connecticut [12]. LD has emerged since then, with 96% of diagnosed and reported cases found in 14 states, according to the US CDC.
Yearly incidence of new Lyme infections
Based on direct surveillance reporting to state health departments, there have been approximately 30,000 confirmed cases per year in the United States. However, two recent publications by US CDC researchers suggest that the actual number of new infections is much higher. Hinckley et al. (2014) estimated about 288,000 new infections in 2008 (range 240,000-444,000) based on surveys of seven national commercial labs that performed Lyme disease testing. Due to insensitivity in the diagnostic tests currently used by mainstream medical authorities, incidence estimates based solely on these tests are likely to significantly undercount the numbers of infected. Nelson et al. (2015) used data from a health insurance claims database to estimate there have been approximately 329,000 incident Lyme diagnoses per year during 2005-2010 (range 296,000-376,000). These publications show that reliance on surveillance, reported by local and regional public health authorities results in significant under-reporting.
Age and gender of newly infected individuals
The US CDC has estimated the age and gender distribution of new Lyme disease diagnoses in the United States. The distribution is available online at http://www.cdc. gov/lyme/stats/graphs.html (accessed June 6, 2016). This distribution shows two peaks, one in childhood (ages 5-9) that is higher among boys, and one at ages 45-55 with similar distribution by gender.
Treatment failure
Recent studies have shown that treatment failure rates may range from 10 to 20% [13][14][15]. Given the variability of treatment failure due to regional, geographical differences, socioeconomic factors, co-morbidities, treatment delays, and non-standardized treatment protocols, we chose to encompass both extremes of this range, basing our estimations on either 10 or 20%.
Survival
Since LD is rarely reported as a cause of death [16], we assume death rates for those with PTLD are identical to those of the general US population. Death rates for the general US population are available online at the US CDC website (accessed June 6, 2016) by gender, year (1980-2014) and 10-year age groups (except for age 0-4) [17,18].
Incidence of Lyme disease infections
Mean: The simulations were set up to evaluate three arguably plausible assumptions about disease incidence, since the exact growth rate of the US LD epidemic is not known. All scenarios assumed 0 cases prior to 1981 and allowed linear growth between years for which estimates are available. Scenario A represents LD cases captured for surveillance purposes and assumed linear growth from 0 cases in 1980 to 34,449 cases in 2005, and remained stable with 34,449 annual cases from 2005 onward. These values derive from surveillance data published by the US CDC [19]. Scenario B similarly assumed 0 cases in 1980, linear growth to 329,000 cases in 2005 and then a stable 329,000 cases in years from 2005 onward. Scenario C is the same as scenario B to 2005, but allowed continued linear growth thereafter. Use of linear growth in our predictions is conservative over exponential growth; the latter is a potentially realistic option in an expanding epidemic, due to rapid growth of the vector population and tick habitats [20].
Variability: While the mean (or expected number) of new infections was presented as input, the simulations allowed the actual number to vary stochastically using the negative binomial distribution with the variance set to mu + mu 2 /size, where mu is the mean and size is an overdispersion parameter. This allowed uncertainty to increase with a higher expected number of new infections. The dispersion parameter was estimated by fitting overdispersed generalized linear negative binomial regression models to the number of confirmed cases in the CDC surveillance data from 1997 to 2017, using linear growth over time as the only independent variable [19]. We fit two regressions due to the assumption of Scenario A of linear growth from 1997 to 2005 with constant incidence thereafter and because the CDC employed two different reporting criteria before and after 2008. The dispersion parameter estimated from the 1997-2005 data was 112, and the dispersion parameter estimated from the 2008-2017 data was 127. To be consistent with both models, the dispersion parameter, size, was set to 120 in our simulations.
Age and gender of new infections
Mean: The mean age and gender of new infections was assumed to follow the distribution provided by the CDC and remain stable for the duration of the epidemic.
Variability: Each new infection was assigned to a 5-year age group and gender using a multinomial distribution parametrized using the distribution of the CDC data. Each new infection was randomly assigned an age in years assuming a uniform distribution. After drawing a uniform [0, 1] deviate we assigned an age as follows: if the uniform deviate was [0-.2), we placed the individual in the lowest age in the age group, if it was [0.2, 0.4), the second lowest was used, etc.
Progression to PTLD
Mean: We assumed rates of treatment failure to be 10% or 20%.
Variability: With a wide range of reported rates for transitioning to PTLD from the literature, our study assumed these estimates were based on a relatively small sample size of approximately 500 individuals. To model the uncertainty, we drew a probability from a beta distribution for each year with variability determined by a sample size of 500. The parameters of the beta distribution for a mean probability, p, were alpha = p*500 and beta = (1-p)*500. For a failure rate of 10%, p = 0.10, the mean is 10%, and the 2 standard deviation (SD) range is 7.4 to 12.8%; for a failure rate of 20%, p = 0.20, the mean is 20%, and the 2 SD range is 16.4 to 23.6%.
Survival
Mean: While nine cases of fatal Lyme carditis have been recognized in national surveillance data from 1985 to 2018 in the US [21], a review of death reports and death certificates in the US from 1999 to 2003 cited Lyme disease as a rare cause of death [16,22]. Survival rates for patients with PTLD were, therefore, assumed to be the same as the general US population and survival rates after 2014 were assumed to be as in 2014.
Variability: Survival to the next year was simulated by drawing a binomial random variable for each case of PTLD, parameterized using the CDC survival rates. Our non-parametric survival model differs from Crouch et al., who fit a semi-parametric survival model to survival data specific to their disease population [11].
Simulation algorithm
The simulations estimate N th -year prevalence as described by Crouch et al. [11]. N th -year prevalence is the sum of cases identified in the N-1 years prior to the index date that survived to the index date plus the new cases from the current year. Therefore, 1-year prevalence is simply the number of new PTLD patients each year, and 2-year prevalence is the number of cases from the prior year that survived to the current year plus the current year's cohort, etc. The total number of PTLD cases at an index date of 2016 is the sum of surviving cases since 1981, or the 36-year prevalence, and the total number of cases in 2020 is the 40-year prevalence.
A total of 6 simulations were performed, one for each of the three incidence scenarios at the two PTLD progression rates (3*2 = 6). For each simulation, we performed 5000 runs and provide the median and the 2.5th and 97.5th percentiles as coverage intervals (CI) of the results [23]. All analyses were performed in R version 3.5.1.
For each run, we did the following: Step 1. Drew the number of incident PTLD cases per year by age and gender, specifically a) Drew the number of new infections each year from 1981 to 2020 b) Drew a probability of progressing to PTLD for each year from 1980 to 2020 c) Using the probabilities in (b), randomly assigned each new infection as recovered or PTLD d) Randomly assigned each PTLD case a 5-year age and gender category e) Randomly assigned an age in years to each case Step 2. Allowed the cohorts to age for the duration of the epidemic, specifically a) Determined survivors to the following year using age, gender, and year-specific survival rates b) Increased the age of survivors by one year c) Added the survivors and updated as the N + 1-year prevalence d) Repeated for the duration of the epidemic and appropriately tallied the results Step 3. Steps 1 and 2 were repeated 5000 times and the results were stored for analysis.
Deterministic estimate of prevalence
As a check of the validity of the simulation, the expected prevalence of PTLD cases was also estimated using a deterministic approach. For each year of the epidemic, new infections were expected to follow scenario A, B or C exactly, have an age and gender distribution precisely equal to that presented by the CDC report, exactly 10 or 20% of new cases transitioned to PTLD, and death occurred at 80 years of age. The deterministic values should have approximated the mean from the simulations, but did not have associated variability.
Results
The 2016 and 2020 PTLD prevalence estimates varied for the three incidence scenarios and two treatment failure rates ( Table 1). The 2016 estimates were highly variable and ranged from a low of 69,011 persons (95% CI 51,796 to 89,312) with Scenario A and failure rate of 10%, to 1,523,869 (95% CI 1,268,634 to 1,809,416) with Scenario C and failure rate of 20%. Regardless of incidence scenario or failure rate, all results indicated an increase in the prevalence of PTLD from 2016 to 2020, with 18, 18 and 28% increases over four years under Scenarios A, B and C, respectively. Scenario C, with the highest 2016 prevalence estimates, also showed the largest relative growth over the following four years. Scenario C with a failure rate of 20% predicted a prevalence of about 2 million cases of PTLD in 2020 (1,944,189; CI: 1,619,988 to 2,304,147). Unsurprisingly, doubling the failure rate from 10 to 20% essentially doubled the expected prevalence of PTLD.
The expected number of new PTLD cases per year resulting from new infections that year is the product of the expected number of new Lyme infections and the treatment failure rate; as a result, our findings were sensitive to these input values. However, while the prevalence of PTLD varies by age and gender (Fig. 1), the relative distribution by age and gender was similar across all scenarios (not shown). In 2020 and among females, the highest prevalence PTLD was predicted among those aged 50 to 70 years; and among males, the prevalence was relatively constant among those aged 15 to 65, with peaks of higher prevalence among those aged 15 to 25 and 50 to 65.
Discussion
Although antibiotic therapy cures most LD patients, a significant proportion of patients continue to suffer persisting symptoms that can derail normal life. These include chronic pain, neurological sequelae such as cognitive dysfunction, refractory arthritis and debilitating fatigue. Whether the persistence of symptoms is attributable to LD is the subject of impassioned debate [6,15,24,25]. Due to the subjective nature of many symptoms and the absence of a diagnostic biomarker, the identification of post-treatment LD is often based on clinical diagnosis, and treatment is undefined and challenging. Diagnosing PTLD with objective, quantifiable biomarkers [26][27][28] is complicated and most likely involves some degree of individual variability [29]. Despite the variation in symptoms, it appears that risk factors for PTLD include delayed diagnosis and treatment as well as more symptoms and increased severity of acute LD [30][31][32]. The mechanisms underlying PTLD are not understood. Direct evidence of persistent pathogen replication by culturing B. burgdorferi from patients experiencing long-term illness is difficult. However, in the nonhuman primate model, morphologically intact spirochetes have been observed in the heart, brains, and adjacent to peripheral nerve a year after infection, with Borrelia antigen staining in antibiotic-treated animals [33]. Several months after infection, Borrelia spirochetes have been detected in antibiotic-treated mice via xenodiagnoses [34]. However, studies evaluating long-term, or repeated, antibiotic treatment of PTLD patients have not shown sustained improvement [35][36][37][38], although there is evidence some subgroups may benefit from retreatment [39]. Overall, consistent evidence for continued bacterial replication in PTLD patients is still lacking, and placebo-controlled antibiotic trials are complicated by the heterogeneity and non-quantifiable nature of symptoms in these patients. It is possible that with the development of more advanced diagnostic technology, direct detection of bacterial replication as a measurable endpoint may eventually be possible.
In addition, or alternatively, it is possible that persisting symptoms may be caused by dysregulation of the immune response to Borrelia infection. LD patients with continued symptoms have elevated inflammatory markers or cytokines in blood compared to those who have recovered [26-28, 40, 41], and transcription of genes involved in the response to infectious microbes may differ quantitatively and qualitatively in PTLD vs. cured patients [42].
Confounding the understanding of LD pathogenesis is the absence of a reliable diagnostic test, which causes imprecision in determining LD incidence. A two-stage blood test is recommended by current CDC protocols, and includes an enzyme immunoassay followed by a Western blot to detect antibodies against B. burgdorferi. Its inaccuracy, due to low sensitivity in early infection and inability to capture bacterial strain variation, is overall a dismal 50-60% [43][44][45]. National surveillance of Fig. 1 Prevalence of PTLD in 2020, by age for females (upper plot) and males (lower plot). Data are for Scenario C, with failure rate of 20%. Solid line represents the median and the dashed lines represent 95% coverage intervals LD, based on reporting from local and regional health authorities, is thus likely to be erroneous. Following antibiotic therapy, some patients may remain serologically positive, suggesting that the presence of antibodies is variable and uninformative [43]. Adding to the uncertainty is the lack of suitable biomarkers to indicate treatment endpoint and/or to suggest a transition to PTLD.
Despite the limitations in accurate diagnosis, two studies carried out by the CDC attempted to more accurately determine the incidence of LD, calculating 329,000 cases per year [3,4]. Since the simulation models presented here are based on the expected number of new LD cases each year, this value is critical, and we incorporated the most accurate estimates known to date. We have coupled this information with data on incidence by age and gender, US life expectancy and death rates, and have constructed a simulation modeling framework to estimate the number, along with coverage intervals, of patients who continue to suffer PTLD. Our estimations for numbers of PTLD cases are between 69,011 in 2016, and~2 million by 2020, suggesting a substantial number of patients living with significant health challenges.
One caveat to our models is that although LD is an expanding epidemic [46], its growth rate is not well defined. Thus, which incidence scenario is most realistic is difficult to determine. Scenario A, based on linear growth in the numbers of LD cases until 2005, followed by static low numbers thereafter, seems less plausible than the other two scenarios that include steady, sustained growth. Scenario B proposes growth until 2005, followed by a constant number of LD cases per year. This latter incidence is based on CDC estimates of 329,000 cases per year, rather than surveillance data. Scenario C is based on linear LD incidence growth from 1980 onwards.
We acknowledge and incorporate uncertainty in disease incidence and treatment failure rates by simulating six settings, each providing an estimate of prevalence under the assumption that the setting is correct. As a result, estimates of uncertainty come from the probability distribution function generating the type of data used as input, and do not incorporate uncertainty about the mechanism generating the data (i.e. the setting). The uncertainty in the data-generating mechanism is shown by examining the differences in the results across all six settings.
An alternate approach could have been to incorporate all uncertainty within one scenario, perhaps weighing each scenario based on some knowledge base. The result would be essentially a weighted average of the results presented here with very wide 95% coverage intervals. Alternatively, this question could be cast in a Bayesian framework and could incorporate expert opinion as priors. Unfortunately, as of now, the weights and priors are unknown. An improved diagnostic test, with national surveillance, and research into treatment failure rates will likely provide more precise information to indicate which setting optimally fits the dynamics of the epidemic.
Nevertheless, our findings suggest that there are large numbers of patients living with LD-related chronic illness. Due to the lack of a single case definition or single, shared phenotype for PTLD, private and public health insurance does not include the costs of caring for these individuals, who must personally pay for these expenses. Even during acute infection, LD patients are estimated to cost approximately $3000 per person annually in medical billing [9]. During long-term illness, costs are likely to be much more. A Dutch study estimated more than 1.7 disability-adjusted life years lost per patient, due to persisting Lyme borreliosis, suggesting a significant economic and societal burden [10]. Here, our estimations indicate many chronically ill US patients in present time as well as into the future. Their large numbers, as well as the economic and personal consequences they suffer, warrant further evidence-based, rigorous research to determine the causes of treatment failures, to predict which LD patients are at risk, and to design more effective therapies. Additionally, further studies are urgently needed in the LD field, to improve diagnostic tests, increase medical and public awareness, and to accurately define the numbers of infected and chronically ill.
Conclusions
Using statistical simulation techniques, we have estimated that the cumulative prevalence of PTLD in the US is high and substantially greater than the yearly incidence. We found that prevalence in 2020 is projected to be higher than 2016, and may be as high as 1,944,189 (CI: 1,619,988 to 2,304,147) cases. These findings are relevant to consideration of expected costs for Lyme disease treatment and the care of those with PTLD. | 2019-04-25T07:58:37.393Z | 2019-04-24T00:00:00.000 | {
"year": 2019,
"sha1": "71102f4523a5dd6e361c0c86e3b9edcc95103eed",
"oa_license": "CCBY",
"oa_url": "https://bmcpublichealth.biomedcentral.com/track/pdf/10.1186/s12889-019-6681-9",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "71102f4523a5dd6e361c0c86e3b9edcc95103eed",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
90634382 | pes2o/s2orc | v3-fos-license | Data on spore morphology of Cerosora microphylla ( Anogramma microphylla ) ( Pteridaceae )
We used the method of scanning electron microscopy (SEM) to study the representatives of the genus Anogramma from South-West Asia and the Himalayas. To confirm the presence of Cerosora microphylla (Hook.) R.M. Tryon in China (hence, the absence of Anogramma microphylla (Hook.) Diels), we revealed the features of external morphology of spores that are not inherent in representatives of the genus Anogramma: absence of an equatorial ridge, roundishtriangular shape of the spores in the proximal-polar and distal-polar positions; absence of roller-like bulges extending upward in the form of ridges along the contour of the distal side of the spore; roller-like folds extending upward in the form of ridges both on the proximal and distal sides of the spore. In addition, one feature revealed for C. microphylla – spherical tubercles between the proximal and distal roller-like folds located along the spore equator – indicates its high species-specificity.
Introduction
The genus Anogramma includes about five species widespread in the American tropics from Mexico to Argentina, from the Azores towards the south to the Cape Verde islands, Africa and Madagascar, southwestern and southern Europe, northern India (including Ceylon), southern Taiwan, Australia and New Zealand (Tryon, 1990).
For the flora of China, Zhang Gangmin and Tom A. Ranker report on two types: A. leptophylla (Linn.) Link and A. microphylla (Hook.) Diels . The main species diversity of the representatives of Anogramma occurs in tropical and subtropical America, excluding A. leptophylla, which is found throughout the tropical and subtropical belts in all floristic realms (Gastony and Baroutsis, 1975).
According to the latest data, the genus Anogramma is part of the subfamily Pteridoideae C. Chr. ex Crabbe, the family Pteridaceae E. D. M. Kirchn. (Smith et al., 2006;Christenhusz, 2011). The results on the phylogeny of ferns obtained by independent research groups using molecular genetics show that the representatives of the genus Anogramma can be unambiguously united in the Pteridoid clade, the group of taenitioid (Zhang et al., 2005;Schuettpelz et al., 2007) and hemionitid ferns (Liu et al., 2008).
The representatives of the Pteridoid clade are referred to the ferns of the taenitioids group based on the data on the Taenitidoideae subfamily reported by Roll Trion (Tryon, , 1990. However, in the research devoted to the most complete reconstruction of the molecular phylogenetic relationships of the representatives of the Pteridaceae family of Chinese-Himalayan and Afro-Madagascar origin, Harald Schneider and his colleagues referred the representatives of the genera Anogramma, Cerosora, Cosentinia and Pityrogramma to the Anogramma clade (Schneider et al., 2013).
The representatives of the genus Cerosora exhibit a sufficiently wide disjunct distribution in Borneo, Sumatra and the Himalayas (Tryon, 1990). In 1986, Rolla Trion brought a new combination and singled out the Chinese-Himalayan representative -C.microphylla Schneider et al., 2013).
Despite this, in the Flora of China A. microphylla is distinguished as a representative of the predominantly neotropical genus Anogramma. In this case, the features of the external morphology of A. microphylla, including an obovate form of the final lobes of the frond and the location of the soruses along the central vein of the final frond, are not common to other representatives of the genus Anogramma.
In the relevant research conducted in 2017 that considers the phylogeny of the subfamily Pteridoideae C. Chr. ex Crabbe, the family Pteridaceae, C. microphylla takes an isolated position in the subfamily system (Zhang et al., 2017). At the same time, the group Cerosora, which includes a single representative of the genus, is closely related to the group of the representatives of Pityrogramma and Gastoniella -a new genus, which comprises three species, previously considered in Anogramma (Gastoniella Li Bing Zhang a. Liang Zhang, gen. nov., Zhang et al., 2017).
It should be noted that Rolla Trion transferred the species A. microphylla to Cerosora despite the absence of glandular hairs on the leaf surface common to all representatives of the genus Anogramma (Schneider et al., 2013). This fact testifies to the species-specific features of the external morphology of C. (A.) microphylla.
To confirm the autonomy of the Chinese-Himalayan C. (A.) microphylla, we performed scanning electron microscopy of spores of ferns collected in the Himalayas. In addition, original photographs of the A. leptophylla spore were obtained for specimens from South-West Asia.
Material and methods
Spores for the study were selected from the herbarium material deposited in the Herbarium of the Komarov Botanical Institute of the Russian Academy of Sciences (St. Petersburg). Three specimens collected in South-West and East Asia were investigated. The spores were fixed on a slide with an electrically conductive adhesive tape, sputtered with gold using a sputter coater Quorum Q150R S, and examined under a scanning electron microscope Mini-SEM SNE-4500M (Korea) located in the laboratory of structural and molecular analysis of plants, TSU. The spores surface was scanned in high vacuum at a voltage of 10-30 kV and magnified from 2,500 to 10,000 times. The spore dimensions and the element ornamentation were determined using photographs taken with a scanning microscope.
Results and discussion
The description and original electron micrographs of the representatives of the genus Anogramma are presented below.
A specimen of A. microphylla (Hook.) Diels from Guizhou province was investigated in thesis by Zhang Gangmin "A systematic study on the Chinese cheilanthoid ferns and their related taxa" in 2003 (unpublished). In contrast to our SEM image, the image presented in the thesis exhibits distinctly expressed tubercles on the proximal surface of the spore between the laesura and roller-like folds.
The analysis of the obtained electron micrographs and comparison of the descriptions of external morphology of spores of the representatives of the genus Anogramma from South America, C. sumatrana Holtt. from the island of Sumatra, as well as the distinctive features of the external morphology of representatives of these genera (Tryon and Lugardon, 1991;Giacosa et al., 2004), show that the species A. microphylla can be considered separately from the genus Anogramma. The features typical of the representatives of the genus Anogramma but not of the species A. microphylla are as follows: absence of the equatorial fold, roundish-triangular spores in the proximal-polar and distal-polar positions; absence of roller-like bulges rising in the form of ridges along the contour of the distal side of the spore; highly risen roller-like folds on both the proximal and distal sides of the spore.
The analysis of the clades of the phylogenetic trees obtained for the representatives of the family Pteridaceae (Schneider et al., 2013;Zhang et al., 2017) and the data obtained in the study of morphology of fern spores of this group have shown that the species A. microphylla (Hook.) is Sino-Himalayan C. microphylla (Hook.) R. M. Tryon. In contrast to the external morphology of spores of C. sumatrana (Tryon and Lugardon, 1991), the feature revealed for C. microphylla -spherical tubercles between the proximal and distal roller-like folds located along the spore equator -indicates its high species-specificity. | 2018-12-26T20:35:46.977Z | 2017-05-14T00:00:00.000 | {
"year": 2017,
"sha1": "020595a23a2144efb577a9a3a10966d97629dac0",
"oa_license": "CCBY",
"oa_url": "https://ecology.dp.ua/index.php/ECO/article/download/011721/705",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "020595a23a2144efb577a9a3a10966d97629dac0",
"s2fieldsofstudy": [
"Biology",
"Environmental Science"
],
"extfieldsofstudy": [
"Biology"
]
} |
270554616 | pes2o/s2orc | v3-fos-license | Combined effects of nitrogen fertilizer and biochar on the growth, yield, and quality of pepper
Abstract A pot experiment was conducted to investigate the combined effects of different nitrogen fertilizer levels (5, 25, and 45 kg of pure nitrogen per 667 m²) and biochar concentrations (0, 0.7, 1.4, and 2.1%) on the growth, yield, and fruit quality of pepper. The findings indicated that a combination of 25 kg/667 m2 of nitrogen and either 0.7% or 1.4% biochar significantly enhanced plant growth, yield, and fruit quality. Specifically, the N2 treatment (25 kg of pure nitrogen per 667 m²) increased substrate porosity, alkali-hydrolyzed nitrogen content, and available phosphorus content. It also boosted root activity and superoxide dismutase activity in pepper leaves, resulting in increased yield and better fruit quality. Furthermore, the proper addition of biochar (0.7–1.4% by weight) enhanced the physical and chemical properties of the substrate, including increased chlorophyll content and enzyme activity in plants, thereby leading to improved overall plant growth, yield, and fruit quality.
Introduction
Pepper, an annual or perennial herb of the genus Capsicum belonging to the family Solanaceae, is an herb native to the tropical regions of Central and South America [1].Because of its rich nutrition, distinctive taste, and aroma, pepper plays a crucial role in human food, nutrition, and health [2,3].In China, the annual cultivation area of pepper has remained stable at 2.1 million hm 2 , with a total output of 64 million tons, making it the largest planting area with the highest yield globally [4].Nitrogen is crucial for plant growth and productivity, prompting farmers to utilize nitrogen fertilizer to enhance crop yield.However, plants can only absorb less than 50% of applied nitrogen, resulting in considerable nitrogen loss [5].When combined with nitrogen fertilizer, biochar can mitigate nitrogen loss in the soil and enhance nitrogen retention in the soil [6].
The production and storage of biochar in soils have been considered as potential methods to reduce atmospheric CO 2 concentration [7].Biochar, produced through the oxygen-limited thermal decomposition of organic materials such as straw, wood waste, livestock, and poultry waste, is considered a sustainable biomass [8].Its production process is cost-effective, sustainable, and easily scalable, attracting significant attention [9].Biochar finds applications in enhancing soil physical and chemical properties, water purification, energy and gas storage, and agricultural productivity [10].Additionally, it increases soil nutrient retention and water-holding capacity, facilitating improved plant root absorption while reducing the risk of water infiltration into rivers or underground reservoirs [11].
China, a significant agricultural country globally, produces a substantial amount of straw annually.Previous studies [12,13] showed that the total straw production of major crops in China reached 9.84 × 10 8 tons in 2016.Maize, rice, and wheat accounted for 41.92, 23.23, and 18.36% of total straw production, respectively, making them the primary sources.Straw typically contains 31-41% crude fiber, 3-6% crude protein, and 42% nitrogen-free extract, with digestion energy ranging from 7.79 to 10.46 MJ/kg, making it a valuable biological resource [14].The development, research, and utilization of biochar have become important measures to address waste resource utilization and the problems related to agriculture, rural areas, and farmers, significantly contributing to the efficient, ecological, and sustainable development of agriculture [15].Although the appropriate addition of biochar can enhance the growth and yield of crops such as pepper [16], it is essential to note that biochar generally cannot replace fertilizers completely, as they may not provide all necessary nutrients [17,18].
To date, no attempts have been made to examine the combined effects of nitrogen fertilizers and biochar on pepper production.The following hypotheses were proposed: (i) the combined application of biochar and nitrogen fertilizers increased the yield of pepper by enhancing root activity, superoxide dismutase (SOD) activity, chlorophyll content, and enzyme activity.(ii) Different application conditions of biochar and nitrogen fertilizers had varying effects on the growth and production of pepper.The objectives of this study were twofold: (i) to determine the effects of different amounts of biochar and nitrogen fertilizers on the growth, fruit quality, and yield of potted peppers and (ii) to clarify the optimal application of biochar and nitrogen fertilizers for pepper plants, providing a theoretical basis for their application in pepper cultivation.
Materials
Pepper seeds (Capsicum annuum L. cv.Jinfu 803) were sourced from Tianjin Chaoyan Seed and Seedling Technology Co., Ltd., Tianjin City, China.The seeds underwent evaluation and screening based on their appearance, size, weight, and shape to ensure their qualification for subsequent experiments.The commercial biochar used in the field experiment was derived from corn straw and produced by Liaoning Jinhefu Agricultural Technology Company, Liaoning Province, China.The nitrogen fertilizer tested was urea (N ≥ 46.4%), manufactured by Ordos New Energy Chemical Co., Ltd, Inner Mongolia, China.The physical and chemical properties of the substrate were as follows: alkali-hydrolyzable nitrogen 172.45 mg/kg, available phosphorus content 19.60 mg/kg, and available potassium content 65.18 mg/kg.
Experimental setup
The experiment was conducted at the Facility Agriculture Science and Engineering Base and Horticulture Comprehensive Laboratory of the College of Horticulture, Jilin Agricultural University, China.The pot treatments comprised three nitrogen levels: pure nitrogen 5 kg/667 m 2 (N1), 25 kg/667 m 2 (N2), and 45 kg/667 m 2 (N3) and four biochar rates: 0% (B0), 0.7% (B1), 1.4% (B2), and 2.1% (B3) by weight, resulting in 12 treatments.A randomized block design was used in the experiment, with each treatment replicated three times.The setting of pure nitrogen concentrations allowed for a clearer observation of the impact of biochar on the substrate and pepper quality and yield.
Pepper seeds were sown in a solar greenhouse on March 15, 2017.When the seedlings reached the two-leaf stage, they were transplanted into 10 cm × 10 cm nutrient bowls.The nutrient soil allocation and seedling stage management followed local conventional practices, with a peat-to-vermiculite ratio of 2:1 for the nutrient soil mix to maintain a substrate moisture content of 70-80%.The seedling temperature was managed according to appropriate growth conditions.Pepper seedlings were then transferred to pots filled with a garden soil-turf mixture (4:1, v/v), with each pot containing 12 kg of substrate thoroughly mixed with 1% chicken manure (nitrogen content 6.68 g/kg) and appropriate biochar content according to the treatments.Pepper seedlings were transplanted with a spacing of 40 cm along a row and 60 cm between rows.Nitrogen fertilizer was applied according to the treatments, with recommended doses of P and K (15 and 25 kg/667 m 2 , respectively) applied as diammonium phosphate and potassium sulfate, respectively.
The fruits of the peppers tested matured on July 19 and were harvested successfully.Mature fruits were selected for yield and quality determination.
Substrate measurements
Substrate bulk density was assessed using the cutting-ring method [19].Substrate porosity was calculated as follows [19]: Porosity 1 substrate bulk density /substrate specific gravity 100%.
The content of available phosphorus was determined using the 0.5 M NaHCO 3 method, available nitrogen content was analyzed using the alkali-hydrolyzing nitrogen method, and available potassium content in the substrate was measured using the flame photometry method [19].
Measurements of growth and physiological indexes
The plant height was measured using a ruler, whereas the stem diameter was measured with a Vernier caliper.The entire shoot biomass was harvested by cutting the shoots on the surface of the substrate, and the fresh weight was recorded.The roots were then separated from the soil and weighed.Subsequently, both the shoots and roots were oven-dried at 70°C until a constant dry weight was achieved.The chlorophyll content was determined using the spectrophotometric method, and the root activity was assessed using the TTC method.The SOD activity was determined using the photochemical reduction of nitroblue tetrazolium method, and the peroxidase (POD) activity was measured using the guaiacol method [20].
Fruit yield and quality assessment
At maturity, 10 ripe pepper fruits were randomly selected from each treatment for yield determination and quality analysis.The soluble sugar content (SS) was measured using the anthrone method, the soluble protein content (SP) was determined by the Coomassie Bright Blue G-250 staining method, the vitamin C content (VC) was assessed by molybdenum blue colorimetry, the organic acid content (OA) was determined by alkaline titration, and the free amino acid content (FAA) was measured using ninhydrin coloration method [20].
Statistical analysis
All data were analyzed using two-way analysis of variance (ANOVA), and the means were compared using the least significant difference (LSD) test.The significance between treatments was evaluated at a probability of 0.05.All statistical analyses were performed using DPS software.
Physical and chemical properties of substrate
Substrate bulk density plays a critical role in various biochemical processes such as water movement and salt migration by influencing the water-to-gas ratio, which is an indicator of substrate fertility.As shown in Figure 1, the bulk density of the substrate decreased under all biochar treatments compared with no-biochar treatment.Nitrogen fertilizer did not significantly affect substrate bulk density.The highest substrate bulk density was observed under the combined treatment of 45 kg/667 m 2 nitrogen fertilizer and no biochar (N3B0), whereas the lowest value was recorded Combined effects of nitrogen fertilizer and biochar 3 under the treatment with 5 kg/667 m 2 nitrogen fertilizer and 0.7% biochar amendment (N1B1).
Figure 2 shows the mean values of substrate porosity under three nitrogen fertilizer and four biochar treatments.The substrate porosity values were significantly affected by both nitrogen fertilizer and biochar treatments.The N2 treatment exhibited the highest porosity, surpassing that under the N1 and N3 treatments.Moreover, the B1 and B2 biochar amendments significantly increased porosity compared with no-biochar treatment.The highest porosity value was recorded under the N2B0 treatment, which was significantly higher than under the other treatments.
Alkali-hydrolyzed nitrogen is a crucial indicator reflecting soil nitrogen supply capacity, including both inorganic and organic nitrogen.Its content is closely associated with organic matter content and is unstable in soil.Both nitrogen levels and biochar significantly affected the alkali-hydrolyzed nitrogen content in the substrate (Figure 3).The N2 treatment exhibited comparatively higher alkali-hydrolyzed nitrogen content than N1 and N3 N2 treatments.The highest alkali-hydrolyzed nitrogen content was 187.76 mg/kg under B0 treatment.The highest alkali-hydrolyzed nitrogen content was 198.95 mg/kg under the N2B0 treatment, whereas the lowest content of alkali-hydrolyzed nitrogen was observed under the N2B2 treatment.
The impact of nitrogen fertilizer and biochar on available phosphorus content in the substrate is depicted in Figure 4.The data indicated a significant effect of different nitrogen levels on available phosphorus content.The highest available phosphorus content was 23.83 mg/kg in the N2 treatment.Furthermore, biochar amendment significantly increased available phosphorus content compared with no-biochar control.The highest available phosphorus content was observed under the N2B2 treatment.
Nitrogen fertilizer and biochar treatments exhibited significant effects on available potassium content (Figure 5).The available potassium content in the substrate significantly decreased with the increase in the amount of nitrogen fertilizer applied.The available potassium content under the B1 and B2 treatments was higher than under no-biochar control.The highest available potassium content was observed under the N1B1 treatment, followed by the N1B2 and N2B2 treatments.No significant difference was observed among the three treatments.
Growth parameters
Plant height was significantly affected by both nitrogen fertilizer and biochar treatments (Figure 6).The maximum plant height was observed under the N2 treatment, followed by the N3 treatment.No significant difference was observed between the two fertilizer treatments.The plant height increased significantly under the B1 and B2 treatments compared with the control and B3 treatments.The greatest plant height was 40.13 cm under the N3B1 treatment, which was significantly greater than that under the other treatments.
Both nitrogen fertilizer and biochar treatment had a significant effect on stem diameter (Figure 7).It increased with the increase in nitrogen fertilizer application.Biochar amendment also increased the stem diameter compared with the no-biochar control.Among the 12 treatments, the maximum stem diameter occurred under N3B1 treatment.
Fresh plant weight (FPW) increased significantly in plants treated with biochar (Table 1).Nitrogen fertilizer also had a significant effect on FPW, with the greatest FPW under N2 treatment.The maximum FPW was observed under N2B1 treatment.Both nitrogen fertilizer and biochar had a significant effect on fresh root weight (FRW).The greatest FRW was found under N3B1 treatment compared with other treatments.
Dry plant weight (DPW) and dry root weight (DRW) were significantly affected by both nitrogen fertilizer and biochar treatments (Table 1).DPW was significantly higher under the N1 and N3 treatments compared with the N2 treatment.The B1 and B2 treatments significantly increased DPW compared with the control.Among the combination treatments, N3B2 resulted in the highest DPW, whereas the smallest value was observed under the N2B3 treatment.The largest increases in DRW were observed under the N2B2 and N2B0 treatments.
Nitrogen fertilizer and biochar had a significant impact on chlorophyll content.The chlorophyll content significantly Combined effects of nitrogen fertilizer and biochar 5 increased with higher nitrogen fertilizer levels, and the B1 treatment significantly increased chlorophyll content compared with the control.The N3B1 treatment led to the highest chlorophyll content.
Table 2 illustrates the significant effect of nitrogen fertilizer and biochar on root activity.The root activity increased significantly on increasing nitrogen fertilizer levels from N1 to N3; however, no significant increase was observed under N2 and N3 treatments.Additionally, biochar amendments significantly increased root activity compared with the control, with the highest root activity observed under the N2B2 combination treatment.The influence of nitrogen fertilizer and biochar on POD activity is presented in Table 2.Both nitrogen fertilizer and biochar had significant effects on POD activity, with the highest activity observed under the N3 treatment.Moreover, B1 and B2 treatments increased POD activity compared with the control, with the maximum activity recorded under the N3B2 treatment which was significantly greater than that under other treatments.
Nitrogen fertilizer and biochar both exhibited significant effects on SOD activity.The highest SOD activity was observed under the N2 treatment.A significant increase in the SOD activity was observed under the B3 treatment compared with B0 treatment.The highest SOD activity was recorded under the N2B0 treatment.
Yield
Yield was significantly affected by nitrogen fertilizer and biochar treatments (Figure 8).The yield in the N2 treatment was higher than the N3 and N1 treatments, and biochar amendment increased the yield compared with non-biochar control.The maximum yield was recorded in the N2B2 treatment, followed by the N2B1 and N3B1 treatment.No significant difference was observed among the three treatments.
Fruit quality
As shown in Table 3, nitrogen fertilizer and biochar treatments exhibited significant effects on the SS content.The SS content increased with the increase in the nitrogen fertilizer level.The SS content was higher under B1 and B2 treatments compared with no-biochar control.The SS content under the N2B1 treatment was 2.27%, followed by the N3B1 and N2B1 treatments, which was 2.25% and 2.20%, respectively.No significant difference was observed among the three treatments.
The SP content significantly increased with an increase in nitrogen fertilizer level (Table 3).Biochar amendment also increased SP content compared with the no-biochar treatment.The highest SP values were observed under N3B1, N2B1, and N3B0 treatments.
The composition and content of OAs are crucial factors affecting fruit quality.The measured OA values are presented in Table 3.The OA content under N2 treatment was significantly higher than that under N3 and N1 treatments.It increased significantly with an increase in BA.The highest OA values were recorded under the N2B2 and N2B3 treatments, with no significant difference between the two treatments.The lowest OA content was observed under the N1B2 and N2B0 treatments.
VC is an important antioxidant and a key factor representing fruit nutritional quality [21].The effects of nitrogen fertilizer and biochar on the VC content of pepper fruits were significant (Table 3).The VC content under N2 and N3 treatments was significantly higher than that under N1 treatment.The VC content under the B1 and B3 treatments was significantly higher than that under the B0 and B2 treatments.The highest VC content was observed under the N3B1, N2B1, and N2B3 treatments, with no significant difference among the three treatments.The lowest VC content was found under the N1B1 and N1B0 treatments.
The content of FAAs was significantly affected by nitrogen fertilizer and biochar treatments (Table 3).The FAA content increased with an increase in nitrogen fertilizer application, but no significant difference was observed between N2 and N1 treatments.The FAA content under the B3 treatment was significantly higher than that under the other treatments.The highest FAA content was observed under the N3B3 and N3B1 treatments.
Principal component analysis
Single-index analysis may not fully capture the effects of biochar and nitrogen fertilizer on the quality and production of pepper.
To comprehensively evaluate the impact of different treatments on pepper quality, it is essential to assess various treatments in combination with each quality index.This approach allows for determining the contribution of Combined effects of nitrogen fertilizer and biochar 7 each quality index to overall pepper quality, thereby facilitating a better understanding of which biochar contents and nitrogen fertilizer concentrations are most beneficial for producing high-quality pepper.
The principal component analysis was used to comprehensively analyze six quality-related indicators, such as yield, quality, aboveground dry weight, underground dry weight, aboveground fresh weight, and underground fresh weight, to evaluate the most effective biochar and nitrogen fertilizer treatment.
Initially, the data were dimensionally reduced to obtain standardized data.Subsequently, the standardized data were analyzed to determine the eigenvalues, eigenvectors, and cumulative contribution rate of the correlation matrix (Table 4).The principal components with eigenvalues >1 were selected for further analysis.
The relationship between each quality index and the first four principal components was calculated as follows: First principal component: Second principal component: Third principal component: Fourth principal component: In the formulas, F 1 , F 2 , F 3 , and F 4 correspond to the scores of the first, second, third, and fourth principal components, respectively.X 1 , X 2 , X 3 , X 4 , X 5 , X 6 , X 7 , X 8 , X 9 , and X 10 represent the aboveground fresh weight, underground fresh weight, aboveground dry weight, underground dry weight, yield, SS, SP, OA, VC, and FAA after eliminating the dimensional relationship between variables.
The comprehensive evaluation function utilizes the variance contribution rate corresponding to each principal component as the weight: Comprehensive score 0.4239 0.2542 0.1746 0.1473 .
The standardized values were input into equations (1)-( 4), and then the calculated values were applied into equation (5).Finally, the comprehensive evaluation of pepper quality indexes under biochar and nitrogen fertilizer treatments was conducted (Table 5).
Discussion
The rational application of biochar and fertilizer is vital for efficiently utilizing agricultural straw resources to reduce agricultural pollution, enhance fertilizer utilization rates, boost crop yields, and maintain soil moisture while reducing nutrient loss.The pore structure of biochar can augment soil porosity, decrease soil bulk density, and improve soil texture and tillage performance [22,23].Additionally, biochar incorporation enhances irrigation water efficiency and influences plant growth by improving the matrix medium [24].The findings of the present study revealed that biochar significantly reduced substrate bulk density and increased substrate porosity, aligning with previous studies demonstrating a decrease in soil bulk density with increased peat content [25,26].Besides lower bulk density, adding biochar to the substrate led to even lower densities compared with conventional substrates, potentially due to the use of garden soil and peat as the cultivation substrate.
Biochar serves to restore soil fertility and enhance soil productivity through various mechanisms.First, biochar contains mineral nutrients like phosphorus, potassium, calcium, magnesium, and nitrogen, enriching soil nutrient levels [27].Second, its porous structure and large surface area, along with substantial negative surface, enable it to absorb more water and nutrient ions, facilitating nutrient uptake by plants [28].Third, biochar's ability to restrain and retain soil nutrients reduces leaching and erosion losses, thereby enhancing soil stability and organic carbon content, ultimately boosting total crop yield [29]. Fourth, biochar fosters an optimal environment for soil microorganisms [30], facilitating nutrient cycling in soil ecosystems and contributing to soil quality and health maintenance [31].Although biochar itself contains relatively low mineral element content, its effects on soil nutrients are primarily mediated by its properties and indirect impacts on soil physicochemical properties and microorganisms [32].The present study found that biochar treatments significantly increased available phosphorus content compared with non-biochar treatments, with the highest available potassium content observed in the B2 treatment.Alkali-hydrolyzed nitrogen and available phosphorus content peaked under medium nitrogen application, whereas the available potassium content decreased significantly with the increase in nitrogen application.
The biochar application as a soil amendment holds promise for promoting growth and enhancing crop productivity [33,34].Previous studies demonstrated biochar's growth-promoting effects on various crops like tomatoes [21], peppers, lettuce [35], beans [36], potatoes [37], cowpeas, and radishes [38,39].However, its impact varies depending on biochar type, application rate, and crop species.For instance, biochar with high volatility may hinder nitrogen absorption and crop growth [40,41].However, different biochar extracts can have distinct effects on seed germination.Moreover, the sludge biochar and cow dung biochar extracts showed a more obvious inhibitory effect on the germination of pepper seeds [42].Additionally, the application amount of biochar is also a crucial factor.Biochar application has a specific range of effectiveness [43].Although an appropriate amount of biochar can promote pepper growth and development, higher biochar additions do not significantly promote pepper growth [44].For example, in soils with a lower available nutrients or nitrogen, low amount of biochar is likely to promote crop growth and increase crop yield, whereas a higher amount of biochar can reduce growth and yield [43,45].The impact of biochar varies depending upon the type of crop.Even under identical conditions, the effects of applying the same biochar to different crops can differ in terms of growth and yield [46].
The effect of biochar for soil improvement is closely related to fertilizer management.Kebede et al. found that the combined application of biochar and compost significantly affected pH, OC, TN, P, K, and other soil characteristics and pepper growth [16].However, the impact of biochar on pepper yield is not solely determined by the amount applied [47].Liu et al. conducted field cylindrical tube cultivation experiments on soybean, and the results showed that the combination of biochar and nitrogen fertilizer mainly affected the yield by adjusting the number of grains per plant [48].However, biochar inhibited the accumulation of dry matter weight per plant in the early stage of soybean growth.In the late stage of soybean growth, dry matter accumulation increased continuously.The soybean yield was the highest when 750 kg/hm 2 biochar and 42 kg/ hm 2 nitrogen fertilizer were applied.Li et al. conducted a study which demonstrated that under the application of 22.5 kg/hm 2 nitrogen fertilizer, the addition of 2.4 tons/ hm 2 of biochar significantly enhanced the dry matter accumulation and improved nitrogen utilization efficiency in flue-cured tobacco [49].Reducing nitrogen fertilizer by 40% compared to the normal supply level is more conducive to promoting the growth and development of pepper fruit and improving the accumulation of quality indicators, leading to better absorption of mineral elements by pepper fruit [50].In the present study, the highest yield was observed in the B1 and B2 biochar treatments, indicating that optimum biochar application had a specific range, consistent with previous findings [43,45].The higher yield was associated with the application of high nitrogen fertilizer, with the highest yield observed in the N2B2 and N2B1 combinations.
Numerous studies have investigated the impact of adding biochar on crop quality [19,47,50], as evidenced by various research findings.For instance, biochar could improve the taste of tomatoes by boosting SS values and achieving optimal sugar-to-acid ratios [21].Additionally, in scenarios involving reduced irrigation, biochar influenced the quality of tomatoes [51].Moreover, the SS content in vegetables increased with the increase in nitrogen levels, although excessive nitrogen application led to a decrease in sugar content [52].In our experiment, we observed a gradual increase in SS content with the application of nitrogen fertilizer, consistent with previous research findings.Moreover, an increase in the protein content was observed with the increase in nitrogen application, as nitrogen served as the fundamental building block of amino acids, leading to an increase in protein content within a certain range of nitrogen application [53].
Conclusions
Utilizing combinations of 25 kg/667 m 2 nitrogen fertilizer and 0.7% or 1.4% biochar showed promise in boosting plant growth, yield, and fruit quality in pepper cultivation.Optimized blending of biochar and nitrogen fertilizer fostered a conducive growth environment and enhanced the overall quality of pepper cultivation and output.This optimized approach could provide valuable insights for future fertilizer applications in the pepper industry.
Table 1 :
Effects of nitrogen fertilizer and biochar on the plant growth parameters of pepper DPW, dry plant weight; DRW, dry root weight; FPW, fresh plant weight; FRW, fresh root weight.The data are the means of three replicates.Values within the same columns followed by different letters are significantly different at P < 0.05 as per the LSD test.B, biochar amendments (0, 0.7, 1.4, and 2.1% by weight); N, nitrogen treatments (N1, N2, and N3 treatments of 5, 25, and 45 kg/667 m 2 , respectively).Means followed by the same lowercase letter in each column and the same uppercase letter in each row do not differ at the 0.05 probability level according to Duncan's test.** P < 0.01.
Table 2 :
Effects of nitrogen fertilizer and biochar combination on the physiological index of pepper leaves Note: B, biochar amendments (0, 0.7, 1.4, and 2.1% by weight); N, nitrogen treatments (N1, N2, and N3 treatments of 5, 25, and 45 kg/667 m 2 , respectively).Means followed by the same lowercase letter in each column and the same uppercase letter in each row do not differ at 0.05 probability level by the Duncan's test.** P < 0.01.
Table 3 :
Effects of nitrogen fertilizer and biochar on pepper fruit quality FAA, free amino acid content; N, nitrogen treatments (N1, N2, and N3 treatments of 5, 25, and 45 kg/667 m 2 , respectively); OA, organic acid content; SP, soluble protein content; SS, soluble sugar content; VC, vitamin C content;.The data are represented as means of three replicates.Different letters within the same column indicate significant differences at P < 0.05 as revealed by the LSD test.
Table 4 :
Eigenvalues and variance contribution rates of principal component analysis | 2024-06-18T13:05:50.459Z | 2024-01-01T00:00:00.000 | {
"year": 2024,
"sha1": "02b72603e59c522122ffbee09e7d849675be68ca",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.1515/biol-2022-0882",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "5882387a744230e42b4e42ab9cb1c49eee0722ae",
"s2fieldsofstudy": [
"Agricultural and Food Sciences"
],
"extfieldsofstudy": [
"Medicine"
]
} |
232208449 | pes2o/s2orc | v3-fos-license | Development of a drug screening system using three-dimensional cardiac tissues containing multiple cell types
We hypothesized that an appropriate ratio of cardiomyocytes, fibroblasts, endothelial cells, and extracellular matrix (ECM) factors would be required for the development of three-dimensional cardiac tissues (3D-CTs) as drug screening systems. To verify this hypothesis, ECM-coated human-induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs), ECM-coated cardiac fibroblasts (CFs), and uncoated cardiac endothelial cells (CEs) were mixed in the following ratios: 10:0:0 (10CT), 7:2:1 (7CT), 5:4:1 (5CT), and 2:7:1 (2CT). The expression of cardiac-, fibroblasts-, and endothelial-specific markers was assessed by FACS, qPCR, and immunostaining while that of ECM-, cell adhesion-, and ion channel-related genes was examined by qPCR. Finally, the contractile properties of the tissues were evaluated in the absence or presence of E-4031 and isoproterenol. The expression of ECM- and adhesion-related genes significantly increased, while that of ion channel-related genes significantly decreased with the CF proportion. Notably, 7CT showed the greatest contractility of all 3D-CTs. When exposed to E-4031 (hERG K channel blocker), 7CT and 5CT showed significantly decreased contractility and increased QT prolongation. Moreover, 10CT and 7CT exhibited a stronger response to isoproterenol than did the other 3D-CTs. Finally, 7CT showed the highest drug sensitivity among all 3D-CTs. In conclusion, 3D-CTs with an appropriate amount of fibroblasts/endothelial cells (7CT in this study) are suitable drug screening systems, e.g. for the detection of drug-induced arrhythmia.
Many drugs are withdrawn from the market due to adverse side effects 1 . Cardiotoxicity, including cytotoxicity and proarrhythmic effect, is a major cause of discontinuation of drug development [2][3][4][5] .
The current screening methods are based on animal models, but species-specific differences in drug effectiveness and safety may be observed 6,7 . Moreover, also due to growing political and social pressure, various attempts to develop systems alternative to animal models have recently been made [8][9][10][11] . A particularly promising approach is represented by innovative in vitro drug screening systems based on human cells, including human induced pluripotent stem cells (hiPSCs).
We have previously developed a new drug screening system to test drug-induced cardiotoxicity in vitro, based on hiPSC-derived three-dimensional cardiac tissues (3D-CTs), which showed good response to some cardiovascular agonists 12 . However, the contractile properties of these systems are still poor as compared to those of myocardial tissues in vivo. In addition, we previously observed that although E-4031 administration decreases in vitro contractility, it does not elicit arrhythmia or Torsades de Pointes (TdP), which are normally observed in vivo under the same conditions. In light of the above results, we concluded that 3D-CTs did not completely reproduce the characteristics of in vivo myocardial tissues and that, therefore, appropriate adjustments may be necessary before using 3D-CTs as accurate tools for drug response prediction.
Discussion
In this study, we created three-dimensional cardiac tissues containing various proportions of hiPSC-CMs, www.nature.com/scientificreports/ fibroblasts, and endothelial cells, and investigated their pathophysiology and suitability as a drug screening tool. The expression of cardiac-specific proteins and ion channel-related genes in 3D-CTs increased with the proportion of hiPSC-CMs, whereas the expression of ECM components and cell-adhesion factors depended on the proportion of fibroblasts and the presence of endothelial cells. 7CT exhibited the best contractility in the absence of any cardiovascular agonists. When exposed to low concentrations of proarrhythmic drugs, 7CT and 5CT exhibited a decrease in relaxation velocity and an increase in CRPI and CRDcf compared to 10CT. In particular, 7CT and 5CT showed higher reactivity to these drugs than did the other 3D-CTs. In addition, when exposed to a positive inotropic drug, 10CT and 7CT exhibited the strongest and dose-dependent changes in contraction-relaxation velocity and CRPI. www.nature.com/scientificreports/ Overall, 7CT and 5CT showed better contractility than did 10CT and 2CT, indicating that 3D-CTs with extremely low or high proportions of fibroblasts may not be suitable for drug screening.
Fibroblasts and endothelial cells directly and indirectly regulate cardiac electrophysiology. The interaction between cardiomyocytes and these cells allows for electrical propagation in detached cardiomyocytes, and promotes electrical conduction. The binding of cardiomyocytes to depolarized fibroblasts increases the resting membrane potential, which may result in increased conduction velocity, provided that fibroblasts are present at the appropriate density [17][18][19][20][21][22][23] . In addition, the adhesion of these cells to cardiomyocytes affects the electrical conduction. By producing ECM components and matrix metalloproteinases, fibroblasts and endothelial cells promote cell-to-cell adhesion and fill the gaps between cells [24][25][26][27] . Insufficiently strong cell-to-cell adhesion results in increased cardiomyocyte resting membrane potential and conduction velocity. An excessive proportion of fibroblasts or endothelial cells may result in abnormally strong cell-to-cell adhesion, possibly slowing down, and eventually impairing, electrical conduction, as suggested by previous reports 18,22,24,28 . Therefore, it can be concluded that an optimal proportion of fibroblasts and endothelial cells is crucial to preserve electrical conduction.
Our results showed that 2CT, which is rich in fibroblasts, highly expressed ECM components and exhibited the weakest contractility. 2CT also showed a lower inter-region correlation coefficient than did the other 3D-CTs, indicating the synchronization of cardiomyocyte contraction. 2CT may not be suitable as a model to study contraction, due to the low number of cardiomyocytes and the excessive number of fibroblasts, causing insufficient contractility and impaired conduction, respectively.
7CT, in which the proportions of hiPSC-CMs, CFs and CEs were 70%, 20% and 10%, respectively, showed the best contractile performance. In this 3D-CT, the proper strength of cell-to-cell adhesions and the appropriate proportions of ECM factors, fibroblasts, and endothelial cells ensured an optimal balance between contraction and electrical conduction.
In myocardial tissue in vivo, cardiomyocytes and fibroblasts account for 30% and 70% of the total cell number, respectively, with 70% of the cell volume accounted for by cardiomyocytes and 30% by fibroblasts 15 . In our optimal 3D-CT model, 70% of the cells were cardiomyocytes and 20% were fibroblasts, while the 80% of the cell volume was constituted by cardiomyocytes and 20% by fibroblasts, presenting a model histologically comparable to the native heart 15,18 .
Next, the suitability of the different 3D-CTs as drug screening models was explored. To evaluate proarrhythmic and negative chronotropic effects, a HERG potassium channel blocker, E-4031, acting on ion channels at the cell surface, was employed. Our results showed that 7CT and 5CT, with CF proportions of 20% and 40%, respectively, reproduced proarrhythmic and negative chronotropic effects in an E-4031 dose-dependent manner, whereas 10CT, containing only hiPSC-CMs, showed little response to E-4031. Previous studies reported that TdP do not occur in cardiac tissues that contain only cardiomyocytes, in line with our findings [28][29][30] . In the present study, the exposure of 2CT, containing 70% CFs, to a high concentration of E-4031 resulted in progressive weakening and in the eventual stop of contraction. The block of action potential in 2CT might be due to the lack of depolarization in the presence of high E-4031 concentrations. This behavior was different from that observed in vivo, suggesting that 2CT was not suitable for the evaluation of drugs causing QT prolongation.
To evaluate positive inotropic effects, we used isoproterenol, a non-selective beta adrenoceptor agonist, which increases cardiac contractility via catecholamine receptors. Since catecholamine receptors are present in cardiomyocytes, we reasoned that 3D-CTs with a high content of hiPSC-CMs would show a better contractile performance than those with lower hiPSC-CM proportions. As expected, 10CT and 7CT, containing an hiPSC-CM proportion of 100% and 70%, respectively, as well as a high content of cardiomyocytes, exhibited a concentration-dependent increase in contractility after treatment with isoproterenol. Although 7CT exhibited a sensitive response to typical positive and negative inotropic drugs in this study, it is necessary to test this system using more drugs that act through various mechanisms. In addition, we plan to further investigate drug response by analyzing calcium transient, extracellular potential, and contraction force, in addition to tissue contractile motion, to provide a more complete analysis of cellular physiology. In this study, 5CT showed a drug response that showed proarrhythmic and negative chronotropic effects as sensitively as 7CT, suggesting that it may be used as a cardiac fibrosis model in the future. Standardization of this model for drug screening requires homogeneous cell quantity and quality. In the construction of the described 3D-CTs, the cell proportions can be precisely controlled and, therefore, tissue constructs with approximately the same number of cells can be obtained. Regarding the cellular quality, further optimization is needed to ensure phenotypic homogeneity and controlled maturation of cardiomyocytes after cardiac differentiation.
In conclusion, 3D-CTs with various cellular compositions were tested and 7CT (70% hiPSC-CMs, 20% CFs, 10% CEs) showed optimal contractility and good response to inotropic drugs. In vitro cellular models capable of recapitulating cardiac physiology are strongly desired for drug screening. Thus, 3D-CTs with appropriate cellular combinations are reliable systems for in vitro drug screening.
Cardiac differentiation of hiPSCs. Cardiac differentiation was induced as previously described with modifications 12 . In brief, hiPSCs were dissociated using Accumax (Innovative Cell Technologies, San Diego, Histological stain. The 3D-CTs were fixed with formalin and embedded in paraffin. The paraffin 3D-CTs blocks were then sectioned (0.5 µm thickness) using a microtome (HM430; ThermoFisher Scientific) and the sections were then stained with hematoxylin-eosin and assessed using a microscope (DM4000B; Leica, Germany). Quantitative real-time polymerase chain reaction. Total RNA was extracted from 3D-CTs using a PureLink RNA Mini Kit (ThermoFisher Scientific). Quantitative real-time polymerase chain reaction (qPCR) were performed as previously described 12 . All data were normalized using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as a control, and assessed using the delta CT method. The expression of β-MHC and α-MHC were normalized using cTnT. All qPCR data are shown as relative values compared to 10CT. The primer sequences are listed in Table 1.
Cell motion analyses.
Cell motion analyses were conducted using a Cell Motion Imaging System (SI8000; SONY, Tokyo, Japan). Motion images of the 3D-CTs were recorded at a rate of 150 frames per s, a resolution of 1024 × 1024 pixels, and a depth of 8 bits. Cell motion was recorded for 10 s after cumulative exposure for 10 min to each E-4031 concentration (1, 3, 10, 30, 100, 300, 1000 nM; Calbiochem Merck Millipore, Darmstadt, Germany) or isoproterenol (Calbiochem Merck Millipore). The beating rate, contraction velocity, relaxation velocity, contraction-relaxation peak interval (CRPI), contraction-relaxation duration (CRD), and inter-region correlation coefficient were analyzed using an SI8000C analyzer (SONY). The value of CRD was normalized to the beat rate, using Fredericia correction (CRDcf) 34 . The beating rate, contraction velocity, relaxation velocity, CRPI, and CRDcf are shown as relative values compared to vehicle-treated controls. The effects of treatments on these parameters were examined by comparison with vehicle-treated controls. When an abnormal waveform due to the addition of E-4031 was observed, the subsequent values were excluded from the analysis (Fig. 5B).
Since the inter-region correlation coefficient indicates cooperative cardiomyocyte contraction 35 , only the negative value, but not the abnormal waveform, was excluded.
Statistical analyses. The data were expressed as the mean ± standard deviation. Statistical significance was determined by Student's t-test (2-tailed) adjusted with the Bonferroni correction in the experiments for gene expression, in comparison with 10CT. Dunnett test was used to assess significance in the assessments of contractile properties, performed using R software (freely available from https ://www.r-proje ct.org/). *P < 0.05 was considered statistically significant and **P < 0.01.
Data availability
All data generated or analyzed during this study are included in this published article. | 2021-03-13T06:16:44.398Z | 2021-03-11T00:00:00.000 | {
"year": 2021,
"sha1": "b3bf310de41d9cf00a4c112f8cccc2f5c13f698b",
"oa_license": "CCBY",
"oa_url": "https://www.nature.com/articles/s41598-021-85261-y.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "a249f9b353f82ebeaca6186c6ec7904f164d9e64",
"s2fieldsofstudy": [
"Medicine",
"Engineering"
],
"extfieldsofstudy": [
"Medicine"
]
} |
58598329 | pes2o/s2orc | v3-fos-license | Nontrivial topology of bulk HgSe from the study of cyclotron effective mass, electron mobility and phase shift of Shubnikov-de Haas oscillations
In this paper, the authors report the results of an experimental study of effective mass, electron mobility and phase shift of Shubnikov de Haas oscillations of transverse magnetoresistance in an extended electron concentration region from 8.8*10^15 cm^-3 to 4.3*10^18 cm^-3 in single crystals of mercury selenide. The revealed features confirm the existence of a Weyl semimetal phase in HgSe at low electron density, which has been indicated by previous magnetotransport studies. However, the most significant result is the discovery of an abrupt change of Berry phase of Pi at electron concentration 2*10^18 cm^-3, which we explain in terms of a manifestation of topological Lifshitz transition in HgSe that occurs by tuning Fermi energy via doping.
I. Introduction
The increasing interest in Weyl semimetals (WSMs) in recent years has not only been in terms of fundamental research but also with respect to their potential for future engineering applications [1,2]. A WSM is a three-dimensional (3D) topological phase protected by a crystal symmetry and characterised by linear dispersion around pairs of nodes having opposite chirality.
These so-called Weyl nodes are separated in momentum space; their only connection is through the sample boundary by exotic non-closed surface states known as Fermi arcs [3]. In paper [4] we reported spectacular transport properties in a HgSe sample with low (2.5×10 16 cm -3 at Т= 4.2K) electron concentration, including strong non-saturating transverse magnetoresistance (MR), negative longitudinal MR, which could be related to a chiral anomaly, and unusual for gapless semiconductors maxima of the temperature dependence of electrical resistivity. These features indicate the possibility of the coexistence of a WSM phase with a trivial gapless semiconductor phase in this material with zinc-blend structure. There are other experimental (ARPES investigation [5]) and theoretical [6,7] justifications for proposing mercury selenide, a material with strong spin-orbit coupling, as a possible candidate in non-centrosymmetric WSMs. Time reversal symmetry forces the total amount of a Weyl nodes in non-centrosymmetric system to be a multiple of four, e. g. there must be at least two pairs of Weyl nodes with opposite chirality. On the other hand, cubic symmetry of HgSe includes either mirror planes or 2-fold and 4-fold rotation axes. That is, the high cubic symmetry of HgSe could allow the location of at least four pairs of Weyl nodes with opposite chirality in a single ( , ) plane instead of the two pairs in a single plane in strained HgTe [7] with lower tetragonal 2 -symmetry.
Thus, mercury selenide is of significant interest for topological condensed matter physics as it presumably has four Weyl node pairs of the same type in bulk and surface topological Fermi arcs. This material can be considered as a simpler or "ideal" non-centrosymmetric WSM, the study of which is important for elucidating the exotic transport properties that other noncentrosymmetric WSMs with a more complex Fermi surface topology exhibit [1]. The aim of this paper is to provide arguments for the possible existence of the WSM phase in HgSe based on the concentration dependence of effective mass, Hall mobility and the Berry phase, which is a principal topological parameter of WSM kinetics.
II. Crystal structure and experimental details
Mercury selenide single crystals were grown using the Bridgman technique. Crystal structure of the HgSe ingots was investigated using the high-resolution x-ray diffractometer with As a result of the diffractogram analysis, 4 � 3 space symmetry group of the zinc-blende structure of HgSe was confirmed with unit cell parameter 6.0875 Å and with Wyckoff positions 4a (000) and 4c (¼,¼,¼) related to Hg and Se, respectively.
Samples under investigation were cut from the homogeneous part of the single crystal ingots perpendicular to the direction of growth. Electron concentration in the range of (10 16 ≲ < 10 18 ) cm -3 was obtained by varying the annealing time of the samples in Se vapour from ≈ 24 to ≈ 600 hours at temperatures of (150 -200)°С [8]. Samples with ≈ (1 -5)×10 18 cm -3 were as-grown.
In order to increase above 2×10 18 cm -3 , samples were doped with donor impurities Ga or Fe [10]. All samples had the shape of a rectangular parallelepiped whose dimensions varied between ≈ (3.5×1×0.7) mm 3 and ≈ (8×2×1) mm 3 and which were not specifically oriented. The procedure of etching and electrical contact preparation is described in [4].
Transverse MR was measured using a standard four-probe method at temperatures 2,
A. Electron effective mass and Hall mobility
The effective mass which yields the cyclotron effective mass [4]. In order to determine the electron concentration , we used the formula: where F -FFT frequency of SdH oscillations, is the absolute value of the electron charge, ℏ is the Plank constant (ℏ = ℎ 2 ⁄ ). well as the data for / 0 from [11,12], where 0 is a free electron mass. It can be seen that, within the margin of error, effective mass does not depend on electron concentration in the range of ≲ 1×10 17 cm -3 down to the lowest concentration for our samples ≈ 9×10 15 cm -3 . On the one hand, this important result is not consistent with the generally accepted view of the non-parabolicity of the isotropic band Γ 8 in HgSe [13]. On the other hand, it can be interpreted as a confirmation of hypothesis [4] for the predominance of a topological semimetal phase in HgSe kinetics at low because a rather long annealing time could lead to a depopulation of the conductivity band Γ 8 . It can be assumed that the Weyl fermion dispersion relation ( ) is not an ideal Dirac-like dispersion, which generally deviates from the ideal Dirac cone, but can be described as in [14,15]: where is the energy at the Dirac point , and is the velocity at the Dirac point. As a result of (2), we detect nondependent on electron effective mass at low . Only at ≳ 5×10 17 cm -3 can the dependence ( ) be satisfactorily described using Kane's [16] non-parabolic isotropic dispersion law (solid line in Fig. 3). To fit the dependence ( ) in two-band approximation (see formula (15) in [11]), we used the following parameters: [4].
is the energy gap between light hole band Γ 6 and conduction band Γ 8 for trivial HgSe band structure, and = 7.7×10 -8 eV cm which is the interband momentum matrix element. Thus, we can assume that, at a sufficiently large , the concentration dependence of the effective mass points to the dominance of a trivial gapless semiconductor phase (conduction band Γ 8 ).
Unlike the effective mass, which is a parameter of the band spectrum, the mobility = ⁄ additionally characterises the electron scattering processes through the transport relaxation time . We experimentally determined electron mobility using formula for Hall Fig . 4 shows the concentration dependence of at Т = 4.2K, consisting of the data of [9,11,[17][18][19] combined with our own findings. After analysing this data, we distinguished three regions of ( ) (Fig. 4). Region I corresponds to > ~ 10 18 cm -3 , where is the concentration of the Lifshitz transition (LT), the discovery of which in HgSe is discussed below.
Electron topological LT [20] is the abrupt topological change in the connectivity of the Fermi surface of material, giving rise to anomalies of its physical properties. With regard to the WSMs, bulk LT should be considered as a recombination of two non-trivial Fermi surfaces enclosing one Weyl node into a single trivial Fermi surface that encloses both Weyl nodes of opposite chirality [21][22][23]. This means that chirality is not well-defined, i.e. the total Berry flux through the Fermi surface equals zero. In Region I, the dependence ( ) seem to be reasonably well described by the theory of ionised donor scattering (solid line on Fig. 4 from [18]) and transport properties should be determined by the trivial conduction band Γ 8 as mentioned above. Region II is limited to concentration interval of 10 17 ≲ ≲ 10 18 см -3 below LT. Hence, Weyl semimetal phase already [11], open triangles are from [12], the solid line is the calculated effective mass for non-parabolic isotropic dispersion in two-band approximation.
exists in region II. This region is characterised with well-defined chirality, i.e. the formation of a separate Fermi surface for each Weyl node. It enables the additional mechanism of momentum relaxation -intervalley scattering or backscattering [23,24], thereby leading to a deviation of the mobility from the theoretical value for ionised donor scattering (Fig. 4). However, with reduced the corresponding decrease in is accompanied by a decrease in (Fig. 3), similar to region I. As a result, depends weakly on in Region II. Thereby, trivial gapless phase still plays noticeable role in carrier kinetics in this region.
In that sense, this region could be considered as a "crossover" between the Region I and Region III ( < ≲ 10 17 cm -3 ). In Region III, the transport and other properties of HgSe are fully determined by the topological semimetal phase. For example, in all samples from Region III, we observed negative longitudinal MR induced by chiral anomaly and non-trivial manifestation of topological properties in terms of the Hall effect which we will report in future papers. In Region III, does not depend on (Fig. 3). The decrease in with a decrease in concentration of intrinsic donor-type defects in this region should lead to a decrease in the probability of backscattering and therefore to a strong increase in mobility. As seen in Fig. 4 the mobility reaches 2.2×10 5 cm 2 /V s at ≈ 9×10 15 cm -3 , which is a record value for HgSe and is compatible to that of a TaAs when the Fermi level is located near the Weyl node W1 [23]. For the foregoing reasons, the above-mentioned concentration formally corresponds to a boundary, below which the effective mass should tend to zero -as was observed, for example, in graphene when approaching [9], crosses [17], open squares [11], stars [18], triangles [19]. The solid line is the calculated mobility for single-ionised-donor-scattering mode from [18]. Vertical dashed lines are drawn to separate three regions: Region I -the trivial gapless phase; Region III -the Weyl semimetal phase; Region II -"crossover" (for details see text).
FIG. 4. Electron mobility in HgSe as a function of electron concentration at 4.2K. Filled circlesthis work, open circles
the Dirac point [25]. It is reasonable to assume that reaching in HgSe is unlikely due to the high defect density in this material.
In the conclusion of this subsection, we present the basic electronic parameters obtained from the analysis of the SdH oscillations for HgSe samples from the region III (see Table 1), in which it is most likely to expect the manifestation of the topological properties. Here -frequency of SdH oscillations, -electron concentration, Table 1, the location of Weyl points is ≈ (10-40)meV below the Fermi level within the region III, which is similar to the location of the Weyl nodes in TaAs samples (see Table III in [23]) when the Fermi-level approaches the W1 node.
B. Oscillation phase
A distinguishing feature of the Dirac quasiparticle is the nontrivial Berry phase of π associated with its cyclotron motion [26,27]: function. When the Berry phase is nonzero, the Onsager semiclassical quantization rule [28] becomes [26]: where is an integer number. In case of parabolic dispersion law Φ = 0, that means phase parameter = 1/2. For an ideal Dirac dispersion Φ = and = 0. A straightforward way to detect the Berry phase in systems with a non-trivial topology is to study SdH oscillations since the basis of this phenomenon consists in a quantisation of orbital motion. Experimentally, a non-trivial Berry phase was obtained from the analysis of SdH oscillations in topological insulators [15,[29][30][31], Dirac [32,33] and Weyl [34] semimetals, bulk Rashba semiconductor [35], as well as in graphene [25].
Let us briefly discuss the procedure of extracting the Berry phase from SdH oscillations of transverse MR. We will utilize the following expression for the transverse oscillatory component of MR tensor, adopted from [36] for 3D electrons: where 2 is an additional phase shift with = 0 for 2D and ± 1/8 for 3D Fermi surfaces.
According to [36], the origin of the minus before cosine follows from the antiphase of the lst harmonic of magnetic susceptibility with respect to the MR. Also, the minus before cosine can be deduced directly from the first expansion coefficient in the Fourier series of the transverse oscillatory component ∆ of the conductivity tensor [37]. Additionally, it should be taken into account that in a strong magnetic field ( >>1, where = ⁄ is the cyclotron frequency) ∆~∆ in the lowest order approximation for the small parameter ( с ) −1 [37][38][39]. As it follows from (5), the n-th maximum in occurs at 1 when: and the n-th minimum in ∆ occurs at 1 when: We obtained where integer (half-integer) n marks minima (maxima) of ∆ ( ) as it was performed in [25,32,40]. In accordance with (6) and (7), this intercept yields the ( + ). The values of between and 2×10 18 cm -3 were determined from the analysis of SdH oscillations that were reported in [41].
As seen in Fig. 6, there is a pronounced abrupt change in A similar drop in the phase shift of quantum oscillations near the LT point was recently predicted theoretically in 3D topological semimetals in [39]. It is essential that the flattening of ( ) (Region II in Fig. 4) occurs below approximately the same electron density as was mentioned above. Here it is important to note that, in contradistinction to the TaAs family of WSMs [34], only the W1-type of Weyl cones is supposed in HgSe [4]. Therefore, the phase shift analysis of Furthermore, the type of Weyl nodes is of particular interest. WSMs can be classified into type-I, that exhibit point-like Fermi surface, and type-II with strongly tilted Weyl cones [43].
According to calculations [7], HgTe is in type-II Weyl semimetal phase when the strain is small enough or even zero. However, such calculations were not performed for HgSe. The experimental data presented in current paper also does not provide the opportunity to reach a conclusion on the type of Weyl nodes (type-I or type-II) supposed in unstrained HgSe. In order to elucidate this problem, additional experiments are necessary, for example the measurements of the photocurrent induced by circularly polarized light [44,45].
IV. Conclusion
In conclusion, we have demonstrated that studies of the effective mass, electron mobility and phase shift of SdH oscillations in HgSe complement each other. Effective mass as a band spectrum parameter points to a deviation of the dispersion relation ( ), given by (2), from an ideal linear dispersion at low energy. In contrast to the case of surface states of topological insulators [14], this deviation seems to be unsubstantial and does not shift the Berry phase from π.
The non-trivial Berry phase points to the existence of band touching points in HgSe such as Weyl nodes and related Weyl cones with linear dispersion. The main result is the discovery of the abrupt change ≈ of the Berry phase (Fig. 6), which we associate with the LT in HgSe that occurs by tuning the Fermi energy via doping. It is relevant argument for a topological semimetal phase in HgSe and is consistent with the results of our previous magnetotransport study [4] of HgSe with low electron concentration -non-centrosymmetric WSM candidate. However, it is obvious that direct experiments, such as ARPES, are necessary to clarify the existence of the Weyl semimetal topological phase in HgSe along with the location and type of Weyl nodes. | 2019-01-10T06:34:01.000Z | 2018-07-05T00:00:00.000 | {
"year": 2018,
"sha1": "7497fab0bcaab01bf2560e0fa7c85d078c0dd006",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/1807.01944",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "7497fab0bcaab01bf2560e0fa7c85d078c0dd006",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics",
"Medicine",
"Materials Science"
]
} |
14022188 | pes2o/s2orc | v3-fos-license | The Role of Muscarinic Receptors in the Pathophysiology of Mood Disorders: A Potential Novel Treatment?
The central cholinergic system has been implicated in the pathophysiology of mood disorders. An imbalance in central cholinergic neurotransmitter activity has been proposed to contribute to the manic and depressive episodes typical of these disorders. Neuropharmacological studies into the effects of cholinergic agonists and antagonists on mood state have provided considerable support for this hypothesis. Furthermore, recent clinical studies have shown that the pan-CHRM antagonist, scopolamine, produces rapid-acting antidepressant effects in individuals with either major depressive disorder (MDD) or bipolar disorder (BPD), such as bipolar depression, contrasting the delayed therapeutic response of conventional mood stabilisers and antidepressants. This review presents recent data from neuroimaging, post-mortem and genetic studies supporting the involvement of muscarinic cholinergic receptors (CHRMs), particularly CHRM2, in the pathophysiology of MDD and BPD. Thus, novel drugs that selectively target CHRMs with negligible effects in the peripheral nervous system might produce more rapid and robust clinical improvement in patients with BPD and MDD.
INTRODUCTION
Major depressive disorders (MDD) and bipolar disorders (BPD) are psychiatric illnesses characterized by inappropriate control of mood or emotions. The causes of MDD and BPD are not known, but it is now thought a combination of genetic, environmental and biochemical factors may contribute to the onset of MDD [1,2] and BPD [3,4]. MDD, also called unipolar depression or clinical depression, has a 12 month prevalence rate of over 6.5% and a lifetime prevalence of approximately 16% [5]. The disorder has wide reaching effects on the sufferers' lives, affecting personal and social relationships at work or school. BPD, formerly referred to as manic depression, is the second major mood disorder and is characterised by cyclic episodes of depression and mania. According to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) [6], there are four subtypes; BPD-I, BPD-II, cyclothymia and bipolar disorder not otherwise specified (BPD-NOS). The lifetime prevalence rates amongst the disorder subtypes are 0.6% for BPD-I, 0.4% for BPD-II, and 1.4% for subthreshold BPD, which is characterized by symptoms of hypomania without a depressive episode, with a 2.4% prevalence rate for bipolar spectrum disorder (BPS) including four subtypes of BPD [7]. Current antidepressant drugs, mood stabilisers and antipsychotics can alleviate the symptoms of mood disorders in some people but in many cases treatment is only partially effective with many sufferers being refractory to treatment with current drugs [8,9]. Therefore, there is an urgent need to discover mechanisms that may be involved in the etiology of mood disorders which are suitable for targeting by therapeutic agents. Recent clinical studies, which show that the pan-muscarinic receptor antagonist, scopolamine [10], elicits a rapid anti-depressant response in patients with MDD and BPD [11,12], highlight a growing body of evidence suggesting the cholinergic system plays an important role in the pathophysiology of mood disorders [13][14][15][16] and may present a novel target for drug development. The question arises why the muscarinic receptor antagonist appears to have some antidepressant effects in clinical studies. Thus, in this review, we highlight the potential involvement of muscarinic cholinergic receptors in the pathophysiology of MDD and BPD and review data that suggests they may be a viable therapeutic target.
MUSCARINIC CHOLINERGIC SYSTEM IN THE CENTRAL NERVOUS SYSTEM
Acetylcholine (ACh) plays an important role as a neurotransmitter in the central and peripheral nervous systems. ACh binds to two classes of receptors [17], the muscarinic cholinergic receptors (CHRMs), which are metabotropic receptors, and the nicotinic cholinergic receptors (CHRNs), which are ligand gated ion channels. Both CHRMs and CHRNs have been acknowledged to play crucial roles in mood disorders. The possible role of CHRNs in mood disorders has been reviewed elsewhere [18,19], hence this review will focus on CHRMs.
A. Gibbons
CHRMs are a family of seven-transmembrane domain receptors, consisting of five receptor subtypes (CHRM1-5). As members of the G-protein coupled receptor (GPCR) superfamily, they associate with heterotrimeric G-proteins to translate the extracellular signal into an intracellular signal transduction cascade [20,21] (Fig. 1). In general, the canonical downstream signalling of CHRM1, 3 and 5 receptors, which are coupled to G q/11 class of G-proteins, is initiated by activation of a Gα subunit protein [22,23]. Phospholipase C-β (PLC-β), activated by the Gα subunit once it dissociates from the Gαβγ protein complex, hydrolyses phosphatidylinositol 4,5-bisphosphate (PIP 2 ) into two second messengers, inositol 1,4,5-trisphosphate (IP 3 ) and diacylglycerol (DAG) [22,24]. The IP 3 binds to inositol 1,4,5-trisphosphate receptors (IP 3 R) on the endoplasmic reticulum and induces their opening, causing an influx of Ca 2+ into the cytoplasm [25]. The other second messenger, DAG, activates several isoforms (α, β, γ, δ, ε, η and θ) of protein kinase C (PKC) [26] which have different distributions or kinetics and thus different functions. Activated PKC can participate in a negative feedback of PLC-β signal transduction [27,28] as well as activating other intracellular proteins, including the phosphorylation of protein kinases or transcription factors. By contrast, CHRM2 and CHRM4 receptors are canonically coupled to G proteins of G i/o class [21,22]. The activated Gα subunit, released from Gβγ, inhibits adenylate cyclase (AC), which is responsible for catalysing the conversion of ATP to cAMP [29,30]. Decreased intracellular cAMP levels in the cytoplasm induces the inactivation of protein kinase A (PKA), regulating ion channel activity, either opening Ca 2+ and K + or closing the Na 2+ channels in the plasma membrane. Consequently, inactivated PKA decreases the rate of internalization of Ca 2+ into the cell and increases outward K + currents [31,32].
The CHRMs have distinct functions in the central nervous system (CNS). Pharmacological research with agonists, antagonists and allosteric modulators has significantly contributed to defining the roles of specific muscarinic receptors [14, [33][34][35]. Gene knockout (KO) studies have also greatly advanced our understanding of the functional roles and the signalling mechanisms of CHRMs in the central cholinergic system [36]. From such studies, it has been shown that CHRM1 and CHRM3 act as postsynaptic receptors in the CNS, where they are mainly localized in the cortex and hippocampus, and play an important role in cognitive function [37]. In humans, CHRM2 and CHRM4 are reported to act as presynaptic autoreceptors in the cerebral cortex [38], caudate-putamen (CHRM4) [39] and hippocampus (CHRM4) [40], where they inhibit the release of ACh. By contrast, KO studies suggest murine CHRM2 is the cholinergic autoreceptor in the cerebral cortex and hippocampus, whilst CHRM4 acts as an autoreceptor in the striatum [41]. Finally, the distribution of CHRM5 is restricted to the hippocampus and substantia nigra [42], where they are involved in modulation of dopamine release [42,43].
Shaw reported that among 16 people who had been exposed to organophosphorus insecticides for up to 10 years; 7 had depression and 5 had schizophrenia with impaired memory, paranoia or/and auditory hallucinations [45,46]. It was later reported that the infusion of the cholinesterase inhibitor, physostigmine, exacerbated the depressive symptoms in patients with BPD [47,48]. In response to these data, coupled with observations that cholinergic stimulation can inhibit adrenergic-stimulated, arousal behaviour [49], Janowsky and colleagues hypothesized that an imbalance between central cholinergic and adrenergic neurotransmitter activity, released from cholinergic nerves in the parasympathetic nervous system and adrenergic nerves in the sympathetic nervous system respectively, could induce mania and depression [16]. Specifically, their hypothesis suggested that a hypocholinergichyperadrenegic state could cause mania, whereas a hypercholinergic-hypoadrenegic state could contribute to the symptoms of depression.
Further supporting this hypothesis, activation of the cholinergic system by physostigmine induces hyperactivity of the hypothalamic-pituitary-adrenal axis (HPA or HTPA axis) in premenopausal female, but not male, patients with depression [50]. The HPA axis is regulated by the release of the hypothalamic peptide corticotrophin releasing hormone (CRH), suggesting cholinergic super-sensitivity might be involved in exacerbating the depressive symptoms via the HPA axis activation. Moreover, the cholinergic system in the brain stem influences the activation of rapid eye movement sleep (REM) and the non-REM sleep cycle [51]. It is therefore significant that the sensitivity of REM sleep response to cholinergic agonists such as RS86, arecoline and physostigmine is increased in patients with major depression [51][52][53]. These effects contrast those of scopolamine, a muscarinic antagonist, which decreases REM sleep and increases REM sleep latency [54,55] in patients with depression [56], suggesting the REM variables are regulated by cholinergic tone in depression.
Animal Models of Mood Disorders
The evidence from human pharmacological studies suggesting the cholinergic system is involved in mood disorders is further supported by animal studies that investigated the psychopharmacological effects of antidepressants on the central cholinergic nervous system. The Flinders Sensitive Line (FSL) rat has a behavioural phenotype thought to model some aspects of depression in that it shows a loss of appetite, increases in REM sleep, learning difficulties and immune abnormalities but it does not mimic all biochemical features of depression, such as the noradrenergic, HPA axis or γ-aminobutyric acid (GABA)ergic components [64]. Significantly, this strain of rat was developed by the selective breeding of Sprague-Dawley (SD) rats for high sensitivity to anticholinesterase [57][58][59]. Thus, it is significant that there is no difference in the CNS binding densities of the broadly selective CHRM radioligand [ 3 H]quinuclidinyl benzilate (QNB) in FSL rats compared to their control line [60], despite them being more sensitive to muscarinic agents in terms of their behavioural and physiological effects, including hypothermia, a decrease in locomotor activity, and suppression of operant response for water reward [61], compared to FRL rats [57,62]. Thus, the FSL rats may be of particular interest for investigating the mechanisms of the cholinergic system as an animal model of depression.
Another potential model of depression was developed by selectively breeding CD-1 ® mice for either a high immobility score (>115 seconds) for Helpless (HL) or a low immobility score (<35 seconds) for non-Helpless (NHL) phenotypes using the tail suspension test (TST), an established behavioural model used for identifying the potential antidepressant properties [63,64]. Compared to NHL mice, HL mice show impairments in behavioural tests such as the forced-swim test (FST) and locomotor activity as well as increased REM sleep and a lighter slow wave sleep (SWS: deep or restorative sleep of the non-REM type) [64]. Moreover, acute treatment with tricyclic antidepressants (TCAs), including imipramine and desipramine, which has a direct anticholinergic effect [65], reduced the immobility in the TST in the HL mice [64]. However, of importance to this review, the muscarinic agonist arecoline, reduced REM sleep latency in the HL mice [66], suggesting cholinergic tone plays an important role in the onset of REM sleep.
In conclusion, studies on FSL rat and the HL mouse are useful for behavioural, neurochemical and pharmacological research as models of depressive disorder that have cholinergic super-sensitivity, consistent with the notion that super-sensitive muscarinic cholinergic receptors are an etiologic factor in mood disorders.
Genetic Studies on Human CHRM2 in Mood Disorders
Evidence implicating the CHRMs in the cholinergic dysfunction seen in mood disorders comes from genetic association studies, which point to a role for CHRM2 in BPD and MDD ( Table 1). One study reported variation at the A1890T (rs8191992) single nucleotide polymorphism (SNP) in the 3' untranslated region (3'UTR) of the CHRM2 gene was associated with MDD in women, but not in men, with a significant increase in the frequency of homozygosity at the A1890T SNP reported in women with MDD compared to healthy controls [67]. In addition, modest associations have been found between variation in the CHRM2 gene sequence and populations of European-Americans and African-Americans with affective disorders, where the T-C-A-G-T-T/C-T-C-A-A-A diplotype at rs978437-rs1455858-rs1824024-rs324640-rs324650-rs6962027 was associated with an increased risk of affective disorders [68]. Furthermore, a combination of SNPs (T-T-T haplotype, at rs1824024-rs2061174-rs324650, which has a frequency of over 40% in people with alcohol dependence and major depressive syndrome) at the 5' end of the CHRM2 gene was found to be associated with alcoholism and MDD, based on the collaborative study on the genetics of alcoholism (COGA) [15]. However, Cohen-Woods and colleagues failed to confirm the association with this haplotype at the SNPs of CHRM2 gene and MDD reported by Wang et al. [15] in a cohort of clinical cases with MDD from the Depression Case Control (DeCC) study [16]. Although genetic studies associated with BPD have identified several candidate chromosome regions [69], similar studies failed to find an association between BPD and a total of 93 SNPs in 19 cholinergic genes, including all CHRMs [70].
Neuroimaging and Post-mortem Studies of Human CHRM2 in Mood Disorders
Neuroimaging techniques now provide opportunities to study receptor status in the human CNS and how they may be affected in psychiatric disorders [71]. Such an approach, using positron emission tomography and a selective CHRM2 agonist radioligand, 3-(3-(3-fluoroproply) thio)-1,2,5-thiadiazol-4-yl] -1,2,5,6-tetrahydro-1-methylpyridine labelled with F-18 ([ 18 F]FP-TZTP), reported the distribution volume (DV; DV = K 1 /K 2 , where K1 is the rate of delivery of [ 18 F]FP-TZTP or influx constant from plasma to tissue and K 2 is the rate of clearance) of [ 18 F]FP-TZTP was significantly lower in the anterior cingulate cortex (ACC) from patients with BPD, but not MDD, during depressive episodes. Apparent low levels of binding were also seen in the dorsal and posterior cingulate cortices, the orbital cortex and the visual cortex, but not the amygdala, the hippocampus and the ventral striatum, from patients with BPD but these differences failed to remain significant following statistical correction of the data for multiple comparisons, suggesting changes in CHRM2 may be more widespread in the CNS rather than just being present in the ACC [13]. The smaller distribution volume in the BPD group compared to controls could be due to either a decrease in CHRM2 receptor density, a change in affinity for the ligand or an increase in endogenous ACh levels, as [ 18 F]FP-TZTP binding is sensitive to changes in endogenous ACh concentration due to direct competition between the two ligands [72]. Irrespective of the cause, it is significant that in a follow-up to the study reporting lower density of the PET ligand [ 18 F]FP-TZTP in the ACC from patients with BPD [13], Cannon and colleagues reported that BPD subjects with T-T genotype at the rs324650 SNP showed significantly lower CHRM2 distribution volume compared to both other BPD subjects with A-T or A-A genotype and healthy controls who were also T-homozygotes [73], suggesting a genetic predisposition to lower levels of [ 18 F]FP-TZTP binding. rs324650 lays within the intronic region between exons 5 and 6 of the CHRM2 and variation within the SNP does not alter the protein's sequence. The reasons why variation at this SNP affects CHRM2 levels in BPD is not clear. However, the CHRM2 gene is reported to contain regulatory elements within its intronic regions and this SNP may affect gene regulation or splice variation [74,75]. Alternatively it is possible that the rs324650 was in high linkage disequilibrium with another SNP that affects CHRM2 levels.
In addition, radioligand binding studies using selective muscarinic antagonists have commonly been used to examine the density and regional distribution of muscarinic receptors in human post-mortem brain tissue [14, 33,76] ( Table 2). Contrasting findings from PET studies, a postmortem study has reported no change in the density of the CHRM2/CHRM4 selective radioligand [ 3 H]AF-DX 384 in post-mortem ACC (BA 24a and b, including small amounts of BA 32 and/or 33) from people with BPD [77]. The differences between the post-mortem and the PET study findings may be explained by the different genetic backgrounds of their cohorts. However, the PET and postmortem studies estimated of CHRM2 levels from the radioligand occupation of the receptor's functional vs nonfunctional binding sites, respectively. Therefore, the lower [ 18 F]FP-TZTP binding levels reported in the PET study could also reflect increased occupation of the CHRM2 functional binding sites by higher levels of exogenous ACh [78]. Taken together, these data suggest that regionally specific decreases in the level of cortical CHRM2, but not CHRM1 and CHRM3, could be involved in the pathophysiology of mood disorders.
Physiological Consequences of the Decrease in CHRM2
Our understanding of the consequences of a loss of CHRM2 in the CNS comes predominantly from the Chrm2 KO mice which suggest the receptor plays vital roles in behavioural flexibility [98], working memory [79][80][81] and hippocampal plasticity [80]. Chrm2 KO animals showed an increase in ACh levels in the hippocampus and an impaired performance in the passive avoidance test, which was used to evaluate learning and memory in an aversive stimulus (a mild foot shock)-induced passive-avoidance response, suggesting high levels of ACh by the lack of CHRM2 autoreceptors may contribute to some of cognitive deficits [81]. Chrm2 KO mice showed a significant increase in latencies of the escape response in the Barnes circular maze, during the first 5 days of testing, compared to wildtype (WT) mice. In addition, the correct arm choices for food were decreased in Chrm2 KO mice in T-maze delayed alternation [80], suggesting Chrm2 KO mice show deficits in learning and memory. Cognitive impairment, such as learning and memory, has been widely found in patients with MDD [82] and are a common symptom of BPD [83][84][85].
The PET study [13] suggests lower levels of CHRM2 or a decreased availability of that receptor are present in the ACC, whilst one of 2 post-mortem studies report a lower [ 3 H]AF-DX 384 binding density in the DFPLC from people with mood disorders [14]. These results could indicate [86].
The ACC region is part of the limbic system and is involved in mood changes [87] and in the modulation of emotion [88], whilst the DFPLC is known to play key roles in cognitive processes, including attention, working memory, learning and decision making [89]. These functions could be affected by a loss of CHRM2 in that CNS region. Therefore, lower levels of CHRM2 may contribute to the pathogenesis of the cognitive deficits and mood change associated with MDD and BPD.
NEUROPSYCHOPHARMACOLOGICAL STUDIES IN MOOD DISORDERS
The antidepressant effects of biperiden, which is more selective for CHRM1 as an anticholinergic agent [31,112], showed a significant improvement in depressive patients in an open-label study [54] and in an acute placebo controlled study [4]. However, in a randomized, double-blind, 6-week clinical study, these results did not confirm that biperiden has antidepressant effects compared to the placebo and glycopyrrolate, which is a peripherally effective anticholinergic agent [40].
The evidence supporting the involvement of CHRM2 in mood disorders is pertinent in light of recent reports that scopolamine reduces symptoms of depression in people with either MDD or BPD. Early clinical studies initially failed to show an antidepressant effect of scopolamine. In these studies, elderly patients with depression were treated with different dosages (0.1, 0.25 and 0.5 mg) of scopolamine [76]. 0.5 mg caused significant cognitive impairments in learning, including semantic memory, vigilance, attention, and continuous performance, as well as behavioural changes, including drowsiness, motor restlessness, disjointed speech and anxiety, but the depressive symptoms were not significantly altered an any dose of scopolamine. These data indicated that the symptoms of depression were not strongly influenced by the omnibus activity of CHRMs in human CNS. However, scopolamine (0.4 mg) was observed to reduce REM sleep and prolong REM latency in patients with depression [41], suggesting the REM variables are regulated by cholinergic tone in depression.
More recently, a proof of principle study showed that intravenous infusions of a much lower dose (0.004 mg/kg) of scopolamine significantly reduced scores on the Montgomery-Åsberg Depression Rating Scale (MADRS [75]) and the Hamilton Anxiety Rating Scale (HARS [45,64]) in 8 depressed patients (5 with MDD and 3 with BPD) compared to the placebo group within 3 to 5 days after infusions [36]. Subsequently, 18 depressed patients with MDD (n = 9) and BPD (n = 9) received intravenous scopolamine (0.004 mg/kg) in a double-blind, placebo-controlled trial. Overall, 61.1% of patients achieved remission (11 of 18; HADRS final score ≤ 10), and 55.6% of patients had a full response (10 of 18; HADRS reduction ≥ 50%). However, scopolamine treatment did not result in an improvement in scores on the Young Mania Rating Scale (YMRS [98]) for patients with BPD compared to their baseline score, suggesting scopolamine has an antidepressant effect only [11]. More recently, these authors and colleagues confirmed that scopolamine is effective in the treatment of new patients with MDD (n = 23; 22 were included in analyses) [12], and they found that women (n = 31) are 33% more sensitive than men (n = 19) to the antidepressant effects [99]. Thus, more recent studies of the clinical effects of cholinergic antagonists provide considerable support for the cholinergic imbalance hypothesis of depression and suggest the earlier studies failed because the doses of scopolamine used were not titred to weight.
POTENTIAL PHARMACEUTICAL DEVELOPMENT FOR MOOD DISORDERS
As reviewed above, the CHRM2 as an autoreceptor in the cerebral cortex is involved in the pathophysiology of mood disorders. Thus, if the lower levels of CHRM2 in mood disorders [13,14] are consistent with lower receptor activity, it would be expected there would be an increased release of ACh into the synaptic cleft, resulting in an overall hypercholinergic state. Scopolamine might inhibit the resulting increase in post-synaptic CHRM signalling, which may explain its mechanism of therapeutic action. This notion is supported by animal studies using gallamine, a CHRM2 antagonist, which show an increase in extracellular ACh levels in the nucleus accumbens [86]. However, scopolamine caused an increase in swimming time [86], suggesting an overall effect of that drug at certain doses might be to create a relatively hypercholinergic state due to an increase in the release of ACh. By contrast, the CHRMs agonist arecoline decreased swimming time in the forced swim test and showed a reduction in open-field locomotion activity [86,100], suggesting that cholinergic activation causes depressive-like behaviour. Hence, these data suggest specifically targeting the CHRM2 with an antagonist may be efficacious and produce rapid and robust effects in the treatment of mood disorders.
This review provides evidence in support of the hypothesis that there is an imbalance between the central cholinergic and adrenergic neurotransmitter activity in mood disorders [16]. Data is also presented that argues that drugs targeting specific CHRM2 which regulate the release of ACh could be used for the treatment of mood disorders. Despite their potential therapeutic benefits for mood disorders, the clinical use of agonists or antagonists, which bind to the orthosteric binding site of receptors, could theoretically not be very effective. For example, orthosteric agonists can cause receptor desensitization/internalization and down-regulation [101], resulting in short-lived efficacy. Another disadvantage is that full antagonists, such as a scopolamine, with poor selectivity could influence both presynaptic CHRM2 and other postsynaptic cholinergic receptor subtypes, suggesting that it may not reach maximum therapeutic effect. In order to overcome these disadvantages, an alternative approach is the development of allosteric modulators, which bind to an allosteric site on the receptor [102], either enhance or reduce the effect of the orthosteric ligand and show a great selectivity for the specific CHRM2 subtype [35,103]. The allosteric modulator has shown no influence on binding of orthosteric agonist on the receptor and does not appear to have the problem of receptor desensitization and down-regulation [104] seen with orthosteric agonists. An allosteric modulator targeting presynaptic CHRM2 has different mechanisms of action in comparison with orthosteric agonists [35,104] and produces synergistic effects of the endogenous ACh by enhancing the affinity of ACh for the orthosteric binding site on CHRM2. Thus, it may be possible to develop a positive allosteric modulator selectively targeting CHRM2 to effectively treat patients suffering from mood disorders.
Early studies using cholinesterase inhibitors, such as organophosphorus insecticides and physostigmine, which increase levels of ACh in the synaptic cleft, caused depressive symptoms in healthy people and improved manic symptoms in patients with mania [44,47]. Moreover, the treatment with cholinesterase inhibitors plus a Neuro Psychological Training (TNP) significantly reduced scores on the Geriatric Depression Scale (GDS [105]), compared to their baseline score, resulting in improvement of depressive symptoms and showed significant improvement of cognition in patient with mild cognitive impairment [106]. Results from animal studies of Chrm2 KO showed impaired performance in the passive avoidance test, suggesting CHRM2 is involved with cognitive processes [81]. A CHRM2 antagonist increased extracellular ACh levels in the nucleus accumbens [86], resulting in compensatory higher levels of acetylcholinesterase [107]. An association between lower levels of CHRM2 and cholinesterase functions in mood disorders may contribute to some of cognitive deficits associated with mood disorders. Given that cognitive impairment is widely reported in patients with MDD [82] and a common symptom of BPD [83][84][85], although it is still unclear about these processes between the cholinesterase inhibitors and the selective modulation of CHRM2 because of the possibility of counteraction each other, drug development targeting them would be an important approach to provide enhancement of cognitive function in novel therapeutic mechanism for the treatment of mood disorders.
Finally, some conventional tricyclic antidepressants (TCAs), such as imipramine, which primarily inhibits norepinephrine (noradrenaline) re-uptake [108] but has additional pharmacological activities, have been associated with direct anticholinergic effects [65]. However, the mechanism underpinning this action of TCAs is not fully understood. Although TCAs have adverse side effects due to the antimuscarinic actions [109][110][111], it can be suggested that these drugs could be effective in treating patients with depression through the potent antagonism of muscarinic receptors, as seen with scopolamine [11,99].
CONCLUSION
Data from genetic, neuroimaging and radioligand binding studies support the involvement of CHRM2 in the pathophysiology of MDD and BPD. However, investigating the factors that regulate the expression of CHRM2 remains to be clearly elucidated in the aetiology of these disorders. Thus, further studies are needed to fully understand the regulation of CHRM2 expression in the pathophysiology of mood disorders. First of all, it is important that these data would be replicated with a large number of samples to provide more reliable conclusion. Future studies should investigate an interaction between the SNPs of CHRM2 gene and environmental and biochemical factors or a combination of these factors associated with mood disorders to identify the regulatory mechanism underlying the effects of the allele on CHRM2 expression levels. Lastly, the association with CHRM2 expression and clinical outcomes obtained from biological, psychological and cognitive measures needs to be further confirmed in the future study.
In developing drugs to target the CHRMs to treat mood disorders, it must be recognised that this receptor is abundant in CNS and the peripheral nervous system (PNS), as well as being present in non-neuronal peripheral tissues such as smooth muscle [112] and heart [113]. In order to avoid the adverse effects associated with targeting the peripheral receptors, drugs targeting specific CHRM2 would need to be delivered to the CNS, but not the PNS [114,115]. The main problem in the drug delivery into the CNS is that the bloodbrain barrier (BBB) as a highly selective permeability barrier allows only small-molecule drugs to be carried into the CNS [116]. Thus, to overcome these properties of the BBB in the drug delivery into the CNS, a promising approach for drug delivery into the CNS is the use of invasive techniques, which are to penetrate the BBB by using surgical methods, BBB disruption and direct injection into the CNS, and noninvasive methods, which are to penetrate or to bypass the BBB by using carriers, such as liposomes, nanoparticles, and polymeric micells [114,115,117]. These multiple drug delivery systems give promise for effective drug delivery to the CNS, showing potential for safer and more effective treatment for mood disorders. Hence, drugs selectively targeting the central CHRM2 will have fewer of the common side-effects, such as drowsiness, blurred vision and dry mouth, associated with muscarinic receptor antagonists, including scopolamine, and TCAs. The concept of developing selective ligands targeting the CHRM2 provides a new and feasible approach for the specific CNS drug delivery to more effectively treat mood disorders. | 2018-04-03T01:50:27.246Z | 2015-11-30T00:00:00.000 | {
"year": 2015,
"sha1": "9406298664dab5be7fc17c97e68e6c95e1e26baa",
"oa_license": "CCBYNC",
"oa_url": "https://europepmc.org/articles/pmc4759313?pdf=render",
"oa_status": "GREEN",
"pdf_src": "PubMedCentral",
"pdf_hash": "9406298664dab5be7fc17c97e68e6c95e1e26baa",
"s2fieldsofstudy": [
"Psychology",
"Biology"
],
"extfieldsofstudy": [
"Medicine",
"Psychology"
]
} |
55351862 | pes2o/s2orc | v3-fos-license | Secondary organic aerosol formation from acetylene ( C 2 H 2 ) : seed e ff ect on SOA yields due to organic photochemistry in the aerosol aqueous phase
Secondary organic aerosol formation from acetylene (C2H2): seed effect on SOA yields due to organic photochemistry in the aerosol aqueous phase R. Volkamer, P. J. Ziemann, and M. J. Molina Dept. of Chemistry and Biochemistry and CIRES, University of Colorado at Boulder, CO, USA Dept. of Chemistry and Biochemistry, U.C. San Diego, La Jolla, CA, USA Air Pollution Research Center, U.C. Riverside, 92521, Riverside, CA, USA Received: 1 July 2008 – Accepted: 5 July 2008 – Published: 5 August 2008 Correspondence to: R. Volkamer (rainer.volkamer@colorado.edu) Published by Copernicus Publications on behalf of the European Geosciences Union.
Introduction
Secondary Organic Aerosol (SOA) formation is linked with debates about air quality, visibility, public health, climate, and the oxidative capacity of the atmosphere.Recent experimental findings indicate SOA contributes about 30-60% of the organic component of the aerosols (OA) in urban air, and >70% of OA in rural air and the free troposphere (Zhang et al., 2007).The contribution of SOA to aerosol optical depth (AOD) is estimated to be larger than that of primary organic aerosol (POA) or black carbon (BC) (Tsigaridis et al., 2006).This contribution is likely to increase as (1) current models significantly underestimate SOA production in large and very different compartments of the atmosphere (Heald et al., 2005;de Gouw et al., 2005;Volkamer et al., 2006), (2) current models assume the mass absorption cross-section (MAC) of OA to be zero, while laboratory and field evidence consistently suggests that OA absorbs light (Kirchstetter et al., 2004;Sun et al., 2007;Barnard et al., 2008), and (3) air quality regulations should result into a reduced contribution of sulphate in the future, while SOA precursor emissions are very poorly characterized (Volkamer et al., 2006) and could continue to increase disproportionately compared to sulfate (Tsigaridis and Kanakidou, 2007;Heald et al., 2008).Notably different models agree with respect to the importance of biogenic VOCs for SOA formation on regional and global Published by Copernicus Publications on behalf of the European Geosciences Union.
SOA is defined as the organic aerosol mass formed in the atmosphere by gas-to-particle partitioning of semivolatile products of the atmospheric oxidation of organic vapors.In the widely established view, partitioning theory (Pankow, 1994a) is employed to describe the partitioning of semivolatile organic vapors only between the gas-phase and the organic fraction of aerosol mass (including the aerosol water attracted into that phase from the hygroscopic properties of SOA) (Seinfeld and Pankow, 2003).This view uses as the key input the vapour pressure of gas-phase oxidation products, which decreases with the number of carbon-atoms of the precursor.Early work suggested the existence of a threshold in the number of carbon atoms, n=7, for a precursor to form SOA (toluene forms SOA, benzene does not) (Pandis et al., 1992).However, more recent experimental studies find SOA formation from benzene (n=6) (Martin-Reviejo and Wirtz, 2005) and isoprene (n=5) (Claeys et al., 2004).Some SOA models have also considered partitioning of SOA species to an aerosol water phase (Pun et al., 2003;Griffin et al., 2005).These models only consider the dissolution of the gas-phase reaction products, do not treat aqueous chemistry explicitly, and the mass fraction of SOA that partitions in this way is typically a few percent of the total SOA; the predominant SOA source in current models is from semivolatile vapors.
SOA formation from acetylene (C 2 H 2 , i.e., n=2), the lightest non-methane hydrocarbon (NMHC), questions volatility of the gas-phase oxidation products as the only controlling parameter of SOA formation.C 2 H 2 oxidation products are at least 6 orders of magnitude too volatile to explain partitioning to aerosols based on physical processes alone (i.e., condensation, absorption).These differences are too large to be explained by organic aerosol activity coefficients, which are thought to modify effective vapor pressures by less than one order of magnitude (Bowman and Melton, 2004).However, C 2 H 2 efficiently forms glyoxal (CHOCHO) upon oxidation by OH-radicals (Yeung et al., 2005).In Mexico City, C 2 H 2 oxidation accounts for about 8% to the over-all CHOCHO source (Volkamer et al., 2007).An increasing body of experimental evidence (Karpel vel Leitner and Dore, 1997;Schweitzer et al., 1998;Blando and Turpin, 2000;Sorooshian et al., 2006;Loeffler et al., 2006;Carlton et al., 2007;Altieri et al., 2008a, b) suggests the processing of CHOCHO and CHOCHO precursors in cloud and fog droplets is an important source for SOA (Ervens et al., 2003;Lim et al., 2005;Warneck, 2005;Ervens et al., 2008).CHOCHO uptake to aerosols forms an alternative pathway for SOA formation in the absence of clouds.Compared to clouds, aerosols contain 3-5 orders of magnitude less liquid water, and aerosol water has therefore been easily rejected as "too small" a volume to dissolve a significant amount of gas-phase organics.The experimental evidence for CHO-CHO uptake to particles is more scarce, and far less conclusive (Jang et al., 2002;Kroll et al., 2005;Liggio et al., 2005a;Corrigan et al., 2008).Conflicting evidence exists about the magnitude and mechanism of CHOCHO uptake, as well as the role of acid-catalysis to control CHOCHO reactivity in aerosols.Recent modeling studies deem CHOCHO and methylglyoxal responsible for 5-11 Tg yr −1 SOA formation globally, suggesting this SOA source could be comparable to the sum of SOA formed from monoterpenes, sesquiterpenes, isoprene, and aromatics (Fu et al., 2008;Myriokefalitakis et al., 2008).More than 85% of this SOA is deemed to form in clouds, mostly at altitudes below 1.5 km (Fu et al., 2008), where aerosols are mostly aqueous.
Current state-of-the-art SOA models use empirical fits to simulation chamber data to calculate SOA yields and partitioning coefficients (Odum et al., 1997;Robinson et al., 2007).Conditions in these chambers approximate the polluted atmosphere to a considerable extent, i.e., with respect to VOC/NO x ratios, NO x concentrations, temperatures, the pool of condensable species produced, OH-radical concentrations, light conditions, and oxidation time-scales.However, most SOA observed in chambers is attributed to secondand higher generation oxidation products (Hurley et al., 2001;Kroll and Seinfeld, 2005), while faster SOA formation is observed in the atmosphere (Volkamer et al., 2006;Riddle et al., 2008).Little attention has thus far been paid to the possible effect of the chemical composition of seed aerosols.In particular, most chamber studies of SOA yields to date have either used no seed at all (i.e., homogeneous nucleation of SOA), or added ammonium sulfate (AS) seed to provide a surface area for condensation (Odum et al., 1997;Martin-Reviejo and Wirtz, 2005;Ng et al., 2007;Kroll et al., 2007).Experiments with water containing seed are less common, but have been used to ascertain the effect of ionic strength on the partitioning of semivolatile products (Cocker et al., 2001a, b).More recently it was demonstrated that hydrophobic organic aerosol seed does not participate in the absorption processes that underlie the current concept of vaporpressure driven partitioning of gas-phase oxidation products to the organic aerosol phase (Song et al., 2007), further increasing the discrepancy between SOA measurements and models.There is currently no systematic study of the effect of chemical composition of seed aerosols on SOA yields.
Here we study directly the SOA formation from C 2 H 2 and CHOCHO on different inorganic, organic and mixed inorganic/organic seed aerosols, in a simulation chamber under both photochemical and dark conditions.The relative humidity and chemical seed composition studied resemble an important subset of particles found in the atmosphere.
Experimental
Experiments were conducted in a 7 m 3 indoor smog chamber equipped with 64 blacklights, and made of Teflon, with a surface/volume (S/V) ratio of 3 m −1 .The experimental setup is shown in Fig. 1.The chamber was filled with dry, clean air (Aadco pure air generator, <5 ppbv of hydrocarbons, 0.1% RH), followed by vapour addition of ultrapure water (Milli Q, 18 M ) to the desired relative humidity (RH).For all experiments seed aerosols were added prior to the organic by nebulizing dilute salt or salt/organic solutions, and passing the aerosols via a 210 Po neutralizer connected to the chamber.All chemicals were purchased from Aldrich, unless otherwise noted.The solutions were made up of either pure compounds (0.1 M) or mixtures of equal mass (about 0.1 M for each solute) of ammonium sulfate (AS, purity 99+%), ammonium bisulfate (ABS, Alfa Aesar, 99.9%), Suwannee River fulvic acid (FA, IHSS), Humic acid sodium salt (HA, tech.), and the amino acids (AA) Glycine (>98.5%), and Tyrosine (>98.5%).Succinic acid (SucA) seeds were prepared from vaporizing pure compound (99+%) in a stream of clean air connected to the chamber.C 2 H 2 (Airgass, grade 2.6) was further purified by passing a small flow via an acetone-slush bath held at a temperature T =−85 • C. Concentrations of C 2 H 2 , CH 3 ONO (Taylor et al., 1980), NO (Matheson) were determined from adding known volumes of these gases to the chamber.Typical initial concentrations were [C 2 H 2 ]=20 ppm, [CH 3 ONO]=5 ppm, [NO]=5 ppm.CH 3 ONO/NO/UV-light was used to generate OH-radicals.For CHOCHO, the reversible uptake to chamber walls, mixing fans, or sampling lines, has been identified as an uncertainty if experiments are conducted over extended experiment times (Kroll et al., 2005).This uncertainty was minimized here by homogeneously producing CHOCHO inside the chamber from the photo oxidation of C 2 H 2 with OH-radicals, thus eliminating the need to actively stir or mix the chamber volume, and systematically reducing experiment time.For the dark experiments authentic CHOCHO samples were prepared from flowing He through heated mixtures of glyoxal trimer dihydrate (GTD) with P 2 O 5 .About 1% CHOCHO/He mixtures were stable over several days if stored in the dark.The mixing ratio of CHOCHO in the gas mixture was periodically checked from measuring the distinct absorption features of CHOCHO in the UV and visible spectral range, and comparison to res- olution adjusted spectra of the absorption cross section of CHOCHO (Volkamer et al., 2005b).Known volumes taken from this bulb were added to the chamber.To convince ourselves that the CHOCHO concentration is reasonably constrained in our photochemical experiments, CHOCHO was measured inside the chamber by means of solid-phase microextraction (SPME) with on-fiber derivatization using o-(2, 3, 4, 5, 6-pentafluorobenzyl) hydroxylamine (PFBHA, 98+%) (Baker et al., 2005) in selected experiments.After addition of seed aerosols and chemicals to the chamber the chamber was operated as a static reactor.Temperature, RH, NO, NO y , aerosol size distributions, and in some experiments CO were periodically monitored over the course of an experiment.The Thermo Environmental Instruments NO-NO 2 -NO x analyzer sampling flow was 0.6 l min −1 .Size distributions were measured by a scanning mobility particle sizer (SMPS) system (long DMA and TSI 3010 particle counter).The DMA flow of 3 l min −1 was taken from the chamber; 2.5 l min −1 were filtered and used as sheath flow; 0.5 l min −1 was unfiltered aerosol sampling flow.The RH was measured in the DMA sampling line about mid-way between the chamber and the SMPS.The RH of sheath air was the same as that of the chamber.Diffusion losses in the sampling lines are <1% for diameters >50 nm.The TSI 3010 particle counter flow is 1 l min −1 (0.5 l min −1 aerosol and 0.5 l min −1 filtered make up flow).Number distributions are sampled over a range of 20 to 600 nm mobility diameter about every 120 s.Number and volume concentrations have been compared with a TSI 3936L72 SMPS system, and were found to agree within ±10% (Matsunaga and Ziemann, 2009).Aerosol wall loss was quantified over a period of about an hour prior to each experiment in terms of a volume loss rate τ (τ = ln (V t /V t=o )/ t=1.5 10 −5 s −1 in Fig. 2), and aerosol wall loss was corrected in the further analysis.
Flushing the chamber and filling it with humidified clean air removed NO y to below the detection limit (about 1 ppb; i.e., dilution factor >10 4 ).The dilution ratio was found smaller for CHOCHO, which is consistent with the release of CHOCHO from the walls, as had been observed previously (Kroll et al., 2005).This CHOCHO memory effect was quantified at 3% for the chamber cleaning procedure used between experiments.CHOCHO measurements compared reasonably well (within 30%) with simulations of the C 2 H 2 /CH 3 ONO/NO/light system based on a subset of the Master Chemical Mechanism (Bloss et al., 2005) The experimental conditions were varied to study photochemical and dark reaction systems, and in terms of CHO-CHO concentration, OH radical concentration, RH, seed volume, seed chemical composition, liquid water content (LWC), and the mass of the organic portion of the seed (M om ), and are listed for individual experiments in Table 1.LWC of seed aerosols was calculated using an additive approach that sums the amount of water attracted from inorganic and organic components of the seed.The Aerosol Inorganic Model (Wexler and Clegg, 2002) was initiated at representative temperature and RH conditions to calculate LWC for inorganic seed components, accounting for metastable aerosols.For organic components the hygroscopic properties were taken from the literature: SucA (Ansari and Pandis, 2000;Peng et al., 2001), FA (Peng et al., 2001), AA (Chan et al., 2005); hygroscopic properties of HA were approximated as those of Suwannee River fulvic acid.M om was calculated from subtracting the LWC from the aerosol volume, and converting the volume portions of individual seed components based on their densities d (d AS =1.77 g cm −3 ; d ABS =1.79 g cm −3 ; d SucA =1.56 g cm −3 ; d FA =1.47 g cm −3 ; d HA =1.72 g cm −3 ; d Glycine =1.46 g cm −3 ; d Tyrosine =1.46 g cm −3 ).The density of 40% w/w glyoxal aqueous solution d Gly40% =1.27 g cm −3 corresponds to a dry equivalent density of glyoxal oligomers of 2.13 g cm −3 .In our calculation of SOA mass formed we converted the observed volume using a density d SOA =2 g cm −3 (a conservatively rounded number for the density of glyoxal oligomers in their dissolved equilibrium distribution).The actual density of oligomers is currently not known.It should be noted, however, that the value of density cancels out with relative comparisons between different experiments; any uncertainty in the density of oligomers thus does not affect the conclusions that are reached from relative arguments.Use of a different density value for oligomers will translate into changes of the absolute SOA yields and H eff values.
The SOA yield, Y SOA , was calculated as the mass of SOA formed divided by the mass of C 2 H 2 reacted.The uncertainty in the SOA yields reflects the combined error from: (1) uncertainty in the wall loss corrected volume change, (2) uncertainty in the amount of C 2 H 2 reacted as predicted by our box model, (3) uncertainty due to CHOCHO wall effects (described above); it does not include uncertainty in the density of CHOCHO-SOA.The uncertainty in (1) and (3) is systematically minimized here from using a photochemical CHO-CHO source (which overcomes the need for mixing), and facilitates reduced experiment times, especially in experiments conducted at higher OH concentrations.
Results
Significant SOA mass is formed from C 2 H 2 during all experiments where a seed was present.Oxidation of C 2 H 2 in the The oxidation products from C 2 H 2 are very volatile.The products formed in our reaction system are predicted by the model as summarized in Table 2.For room temperature the vapour pressure for all products is well above 1 kPa (CRC Handbook, 2003), i.e., about 6 orders of magnitude too volatile to explain partitioning to aerosols as the result of physical processes (condensation, sorption into an organic liquid).Also shown are the respective Henry's law coefficients, or lower limits in the case of CHOCHO (Kroll et al., 2005;Sander et al., 2006).As is evident from Table 2, CHO-CHO is the only product that partitions to the aerosol aqueous phase in appreciable amounts.Notably, our conservative estimate of H eff,CHOCHO results in a lower limit for the amount of CHOCHO that is expected to partition to aerosols that accounts for more than 99% of the source for aerosol growth.Higher values for H eff have been reported in the presence of sulphate and from field evidence (>10 9 M atm −1 , Volkamer et al., 2007;Ip et al., 2009), and are in general agreement with the results presented here.The change in the size distribution of aqueous seed aerosols is therefore attributed to CHOCHO uptake to the seed aerosols.A, C, E) and liquid water content, LWC (B, D, F).No obvious correlation is found between Y SOA and M om , but Y SOA correlates well with LWC for all seed-types studied.For pure AS (solid line), AS+FA (dashed line), and pure ABS (dotted line) sufficient experiments were conducted to demonstrate that the increase in Y SOA with LWC is linear.R=0.76), and ABS (Y SOA =0.01+0.008×LWC,R=0.99) seeds.However, no obvious correlation was observed between Y SOA and M om .For example, for AS+FA seed among the highest Y SOA values were observed in experiments with a particularly low M om , while among the lowest Y SOA values were observed at moderately high values of M om (Fig. 3a).For some seed types Y SOA even anti-correlates with M om (e.g., ABS+FA in Fig. 3c).We take this as strong evidence that the aqueous aerosol phase is a chemical reactor for SOA formation from C 2 H 2 .
Dependence of SOA yields on seed chemical composition
No new particle formation was observed from the photochemical oxidation of C 2 H 2 .However, Y SOA from C 2 H 2 was non-zero on all seeds studied here.The actual value of Y SOA was found a strong function of the seed chemical composition, as is evident from the considerable scatter of the plotted data points in Fig. 3: Y SOA varied by a factor >20 (i.e., 1.1%≤Y SOA ≤25.2%) over the range of experimental conditions investigated.For comparable LWC, e.g., LWC=10 µg m −3 , Y SOA varied by more than a factor of 5, and increased in the following sequence: AS+FA+AA (3.2%), AS+FA+AA+SA (3.8%), SucA (4.7%), AS (5%), AS+FA+SA (5.8%), ABS (8.8%), ABS+HA (7.2-25.2%),ABS+FA (9.6-21%), AS+FA (17.2%),FA (17.3%).For this comparison the observed linear relation with LWC was used for AS, ABS, FA and AS+FA; for AS+FA+SA available data points were interpolated; for SucA, AS+FA+AA, AS+FA+AA+SA, ABS+FA, and ABS+HA linear interpolation through zero was applied (to derive a lower limit in the latter two cases); linear extrapolation between available data points was applied for ABS+FA, and this data was scaled for ABS+HA (to derive upper limits).
Interestingly, AS seed -by far the most widely used seed in chamber studies of SOA yields to date -showed among the lowest Y SOA from C 2 H 2 of any seed investigated.For aqueous AS Y SOA compares to that observed for SucA.However, addition of water soluble FA had a pronounced effect on Y SOA from AS.At comparable LWC, Y SOA is found to be 3.75 times larger for mixed AS+FA than for AS seed (Fig. 3b).Y SOA from ABS is slightly higher than for AS (Fig. 3d).However, the addition of FA (and HA) had less of an effect on Y SOA for ABS than for AS.For example, at LWC=5 µg m −3 , the Y SOA for AS+FA is about 9.6%, i.e., is larger than for ABS+FA (Y SOA =7.2%), despite the fact that Y SOA is higher for ABS than AS.This suggests that the relevant Y SOA for mixed ABS+FA (and likely ABS+HA) may be closer to the lower limit number of the range extrapolated above.
The lowest yields are observed for seeds containing amino acids (AA) like tyrosine and glycine.Over the timescales accessible by our setup, the presence of AA reduces Y SOA by a factor of 5.3 for AS+FA seed.It is important to note that this conclusion can only be drawn as the LWC for the AA studied here is well known.AA containing seed showed the largest LWC of all experiments conducted, and therefore yielded the largest Y SOA values observed (Fig. 3f).The apparent effect of AA to reduce Y SOA becomes visible only after LWC has been normalized; only then Y SOA is actually comparable to the other seeds studied.The apparent reduction in Y SOA becomes more pronounced over longer timescales.Furthermore, mild acidification with SA leaves Y SOA essentially unaffected for AS+FA+AA seed, but strongly reduces Y SOA for AS+FA seed.Experiments containing AA and SA yielded Y SOA values on the bottom scale of those observed.Similarly low yields were observed on SucA seed.Additional evidence on the role of acid comes from a comparison of ABS+FA and AS+FA seed.ABS behaves more acidic at lower RH, while AS is neutral at any RH.The LWC dependence of Y SOA is primarily determined from varying RH in our work, and hence ABS+FA seed at LWC=3 µg m −3 is more acidic than at LWC=5 µg m −3 .Interestingly, a factor of 3.5 lower Y SOA is observed for ABS+FA and ABS+HA at LWC=3 µg m −3 , while the difference is smaller (factor 1.5) for the less acidic conditions (Fig. 3d).We do not find evidence that would suggest higher Y SOA under acidic conditions.For comparable LWC and over the timescales accessible in our setup Y SOA values from AA, SA or SucA containing seed were lower than Y SOA from AS seed.Most experiments showed type A growth in the presence of light, followed by type B growth in the dark (see Table 1).This is illustrated in Fig. 4 for experiments that all used the same AS+FA seed solution.In some experiments the seed was acidified mildly by addition of AA (tyrosine and glycine), or SA to the AS+FA seed solution (pH∼4 for the AS+FA+SA seed).Panel b shows only experiments that were conducted at different RH, and used the same AS+FA seed, while in panel c experiments with addition of acids are shown.The data from panels a was normalized in panels b and c according to Eq. (1):
3.3
where V t is the observed volume at time t, V seed is the seed volume before the experiment, [Gly] t is the CHOCHO mixing ratio in units of ppb (normalized to 100 ppb here), and %LWC is the percent contribution of LWC to total seed volume (normalized to 50% here).(V norm,t −1) hence represents the volume change in units of percent, if 100 ppb CHOCHO were present in the gas-phase at 75% RH (for AS+FA seed in Fig. 3b; 50% RH for AS, not shown).For experiments #19, #20, and #27 the conditions varied by about a factor of 2 in seed volume, a factor of 2.5 in gas-phase CHOCHO (70-171 ppb) and span a large RH range (16%≤ RH≤86%), resulting in a factor of more than 6 difference in the observed volume change (see Table 1 and Fig. 4a).Interestingly, Eq. ( 1) fully accounts for these significant differences, and the normalized time profiles (Fig. 4b) are indistinguishable within error bars.Similarly, AS, ABS, FA, SucA, mixed ABS+FA and ABS+HA seed showed type A growth, but the absolute volume change depends on the seed type (see Sect. 3.2 above).Normalization according to Eq. ( 1) accounts for differences in experimental conditions in the case of AS, ABS, FA and AS+FA seed.No dependence was observed on the light level used.Experiments conducted with a factor of 6 reduced light levels gave results that are essentially identical after normalization according to Eq. ( 1).This demonstrates that for these seeds the photochemical volume growth scales directly proportional to the gas-phase CHO-CHO concentration, and the liquid water of the seed, and further suggests that the availability of UV photons does not limit the rate of CHOCHO uptake.After addition of chemicals to the gas-phase, an immediate shift in aerosol volume is observed when C 2 H 2 is photochemically oxidized (twice here), followed by type B growth in the dark.Type B growth as parameterized in Fig. 3 for AS+FA can explain that observed for ABS+HA seed (solid line), see text.The rapid volume growth during periods when the light is turned ON is proportional to the amount of CHOCHO produced, and inversely proportional to LWC of the seed aerosols.
A reference experiment #32 was conducted in the dark chamber, with known amounts of CHOCHO added directly to the chamber.A considerably slower growth rate, and smaller relative volume growth is observed (type B).Notably,experiments #19,#20 (and #12,#17, show a significant volume increase after the lamps were turned OFF.The normalized volume growth rate between dark periods of these experiments agrees quantitatively with that observed in the dark CHOCHO experiment #32, and confirms directly that CHOCHO is a building block for the observed SOA formation.We conclude that type A+B growth is happening for AS+FA seed at high and low RH.Similarly, ABS, FA, mixed ABS+FA and ABS+HA seed showed type A+B uptake behaviour, while AS and SucA seed only showed type A uptake, but did not show continued growth in the dark chamber.Repeated type A volume growth is observed if the lamps were turned ON consecutively, followed by type B uptake during dark periods in between, as illustrated for ABS+HA seed in Fig. 5. Notably, no volume growth is observed if the seed is exposed to light in the absence of reactants in the gas phase (shown for two trials in Fig. 5).The type B growth rate shown in Fig. 5 does not represent a fit to the data, but was calculated using the type B growth rate derived from Fig. 4b for AS+FA seed.Interestingly this rate does explain type B growth rates observed for ABS+HA seed as well, despite different CHO-CHO and seed volume levels, reflecting that the normalization according to Eq. ( 1) applies also to type B growth.While type A growth is significantly lower for ABS+HA compared to AS+FA, type B growth rates seem to be controlled largely by gas-phase CHOCHO and LWC of the seed, and appear to be comparable within the measurement precision for both seed types.
Most notably, seeds containing AA like tyrosine and glycine, and/or SA did not show type A growth, despite the fact that the same photochemical system was used.The presence of these acids reduced the amount, and the rate of volume growth as illustrated in Fig. 4c.The normalized growth rate for these acidified experiments agrees reasonably well with type B growth observed for the dark periods of experiments #20 and #32 (type B growth from panel b is reproduced in panel c).We do not find an enhanced rate of volume growth for seeds that were mildly acidified with SA compared to the non-acidified seed.Seeds containing AA show a slightly faster initial volume growth, but this initial growth seems to level off over extended experiment times, and is below the type B growth rate after about 30 min experiment time.It should be noted however, that the smaller relative volume change associated with type B growth poses an experimental challenge that is not compensated easily by extended measurement times.The uncertainties from the correction of aerosol volume loss to the chamber walls and (reversible) uptake of gas-phase CHOCHO to the chamber walls then become dominant error sources.The reduction in uptake rates and relative volume changes for seeds that contained AA or SA however is beyond experimental doubt, and this observation is taken as evidence that the presence of acids reduces the amount of available water, or scavenges any excited states produced from organic photochemistry in the aerosol aqueous phase.The pronounced acceleration of the rate of volume change.For neutral seeds, the pronounced acceleration of the rate of volume change suggests that additional CHOCHO uptake is triggered as the result of organic photochemistry in the aerosol aqueous phase.The measured Y SOA number establishes as a result of competing uptake and scavenging rates of the intermediates of SOA formation in the aerosol water.
Dependence of effective CHOCHO solubility on relative humidity
Given that photochemical volume growth scales proportional to the gas-phase CHOCHO concentration, and LWC (Fig. 4b and Eq.1), the effective solubility of CHOCHO, H eff , was calculated from type A volume growth, and values are listed in Table 1.In experiments where also type B growth was observed, H eff was calculated based on the largest volume growth observed.In experiments where type B growth did not level off, H eff is listed as a lower limit number in Table 1.We calculate these H eff values primarily as a means of CHOCHO uptake to aerosols; if calculations were based on the overall seed volume the actual H eff value could be underestimated by as much as a factor of 30 (for FA seed, Fig. 6b).This finding presents a deviation from the traditional definition of Molarity as the molar concentration of a solute per volume of solution (water+seed).
The effect of adding FA to AS seed is illustrated in Fig. 7a.The factor of 3.5 higher H eff for AS+FA seed is reflective of the higher Y SOA (Fig. 3b).Interestingly, the two experiments #12 and #20 (AS+FA seed at RH>80%) showed type A+B growth that did not level off over the timescales investigated; the respective H eff values in Table 1 represent lower limit values.For both experiments these lower limit values are higher than the values observed at lower RH (Fig. 7a).Both experiments were conducted above the deliquescence RH (DRH) of AS, and the observed H eff value is found to be approximately the sum of H eff values observed for pure AS and FA seed.For experiments below the DRH of AS, H eff values of mixed AS+FA seed were found essentially identical to those of pure FA seed.The higher H eff value for AS+FA seed at RH>DRH deserves further investigation.
Dependence of effective CHOCHO solubility on OH radical concentrations
It is of interest to explore whether our results depend on the conditions under which C 2 H 2 was oxidized.The OH-radical concentrations predicted by our model over the course of individual experiments varied by about 10% from the average values listed in Table 1; OH was systematically varied by more than a factor of 15 between different experiments.Over the range of OH-radical concentrations probed the H eff values for any given seed type did not depend on the OH concentration, as is shown in Fig. 7b.This is in line with a product distribution that did not vary much between experiments (see Table 2), and the conclusions in Sect.3, that SOA growth is essentially attributed to CHOCHO partitioning to the aerosol phase; in particular, the rate of any gas-phase polymerization reactions would be expected to depend on the OH radical concentration, which is not observed.Notably, studies of the OH radical initiated oxidation mechanism of C 2 H 2 show no change in the CHOCHO formation in the presence and absence of NO x , and suggest that the CHOCHO formation mechanism involves initial OH attack to the triple bond followed by a sequence of rapid O 2 addition/isomerization reactions that recycle the OH-radical without need for RO 2 →RO conversion (Hatakeyama et al., 1986).These simulation chamber experiments are consistent with OH-cycling experiments using laser-flash photolysis (Siese and Zetzsch, 1995), flow tube Chemical Ionization Mass Spectroscopy and computational studies (Yeung et al., 2005) that all show consistently that hydroxyl peroxy radical intermediates involved with the formation of products from C 2 H 2 are very short lived.Contrary to conventional RO 2 radicals, for which their fate could depend on the experimental conditions (reaction with NO vs. HO 2 /RO 2 ), the fate for these RO 2 is exclusively determined by rapid isomerization reactions.Yeung et al. (2005) infer the lifetime with respect to isomerization to be on the order of 700 ns.Assuming a rate constant of 9×10 −12 cm 3 molec −1 s −1 for the RO 2 +NO reaction, and [NO]=5 ppm, the fate of these RO 2 is determined to more than 99.9% by the isomerization reaction.The insensitivity of the OH-radical initiated oxidation mechanism, i.e., to form CHOCHO, is in general agreement with Fig. 7b, and support the robustness of our results also in the absence of NO x .
The lower range of OH radical concentrations in our experiments are comparable to those observed in the atmosphere (Shirley et al., 2006); typical OH concentrations were about 10-50 times higher than the global annual mean OH concentration (Prinn et al., 2001;Heard and Pilling, 2003).The lack of any OH-radical dependence suggests that our results can be extrapolated to atmospheric conditions.
Discussion and atmospheric relevance
There are currently three laboratory studies that investigate the uptake of CHOCHO to aerosols in the dark (Kroll et al., 2005;Liggio et al., 2005a;Corrigan et al., 2008), and one study reporting field evidence in support of the atmospheric relevance of SOA formation from CHOCHO (Volkamer et al., 2007).Our work is unique for three reasons: (1) the widely used concept of SOA yields is linked directly to a particularly volatile CHOCHO precursor NMHC; SOA formation from C 2 H 2 demonstrates directly that the current theory of SOA formation as it is widely used in atmospheric models is incomplete; (2) SOA yields from C 2 H 2 show a strong dependence on the chemical composition of seed aerosols, and raise questions whether a similar dependence of SOA yields exists also for other SOA precursors; and (3) we demonstrate that organic photochemistry in the aerosol aqueous phase is needed to accelerate SOA formation from CHOCHO, and dominates over dark processes.We also observe evidence for continued chemical processing of CHOCHO in the dark, but the amounts of SOA produced, and rate of SOA formation by dark processes are much smaller than by the photochemical reactions.
Comparison with literature values
Ours is the only study of CHOCHO uptake to aerosols conducted in the presence of gas-phase oxidants.The products formed in the presence of UV photons and OH-radicals are likely different from those formed under dark conditions in the absence of oxidants.Kroll et al. (2005) derived H eff based on the total seed volume, and a density of 1 g cm −3 for SOA from CHOCHO; both assumptions lead to an underestimate of the actual H eff value.The re-evaluated H eff value comparable to this work is a factor of 4.6 larger, i.e., H eff,AS,Kroll =1.2×10 8 M atm −1 at 50% RH.We find a 40% higher value of H eff for AS in our photochemical system, but the rate of CHOCHO uptake is enhanced by 2-3 orders of magnitude in our photochemical system.Notably the CHOCHO uptake kinetics for AS followed type A growth over the range of relative humidity conditions studied; no type B growth was observed in the following dark period of experiments.This is in contrast to the continued growth observed over several hours in dark systems (Kroll et al., 2005;Liggio et al., 2005a;Galloway et al., 2008) and UV-light irradiated systems in the absence of OH radical sources (Galloway et al., 2008), and surprising, as type B growth rates as those observed by other authors should have been observable in our system.Notably, the amounts of SOA formed after several hours in their experiments remain 40% below the SOA formed within 90 s in our study.While our results appear to confirm the conclusion for reversible CHOCHO uptake for AS (Kroll et al., 2005) over an extended RH range, the 40% higher uptake in our photochemical system is only marginally within error bars.The difference in amounts and the rate of CHOCHO uptake may reflect either of the following: (1) CHOCHO is irreversibly taken up by photochemical reactions; (2) photochemistry speeds up the reaction rates, but does not alter the reaction products; (3) photochemistry speeds up the reaction rates, and forms essentially different products than dark reactions; or (4) a combination of the above.The question of whether the oxidative CHOCHO uptake to AS is indeed fully reversible, i.e., can indeed be described by Henry's Law, remains to be answered.Kroll et al. (2005) had suggested that the uptake of CHO-CHO to AS is controlled by ionic strength.This appears to be somewhat unlikely based on the anti correlation between ionic strength and the missing CHOCHO sink observed in Mexico City (Volkamer et al., 2007).Based on the results from this work, any effect of ionic strengths is much smaller than the effect of organic photochemistry to enhance CHO-CHO uptake to aerosols.
There is one study available for discussion of SucA.Corrigan et al. ( 2008) did not observe any CHOCHO uptake in the dark under conditions where SucA is believed to be dry and solid (up to 54% RH).Our study was conducted at slightly higher RH (61% RH), and SucA containing about 36% v/v LWC showed type A uptake kinetics in the photochemical system.Consistent with findings by Corrigan et al. (2008) we did not observe type B growth in the dark, when the aqueous seed was exposed to about 90 ppbv CHO-CHO.It appears unlikely that aerosol water associated with dicarboxylic acids catalyzes any irreversible reaction with CHOCHO under dark conditions.Interestingly the uptake kinetics, and amounts of CHOCHO related growth for AS and SucA are comparable.In analogy to AS, we conclude that any irreversible reaction of CHOCHO with SucA in our system must be driven by photochemical reactions.Alternatively, the reaction rate to form oligomers from CHOCHO in the associated water phase could be strongly enhanced by photochemistry, yet follow a reversible mechanism.It remains to be elucidated whether the photochemically enhanced uptake of CHOCHO to AS and SucA follows a reversible or irreversible reaction mechanism.
The reaction of amines to form covalent N-C bonds has been exploited by analytical techniques for the detection of carbonyls including CHOCHO for a long time (Koshy et al., 1975;Yu et al., 1995b).Little is known however about the extent to which gas-phase reactions of amines play a role in the formation of particles, or affect the atmospheric chemistry of CHOCHO.More information is available on amino acids, which have recently been shown to catalyze the formation of oligomers from acetaldehyde via aldol condensation (pathway 1), or a Mannich Pathway (pathway 2) (Noziere and Cordova, 2008).While similar reactions may also trigger aldol condensation of α-dicarbonyls, the rate limiting step for both pathways consists in the formation of enamine, which may proceed also in the case of CHOCHO, as the additional aldehyde group in water undergoes rapid gem-diol chemistry.
For glycine, the most abundant AA in atmospheric particles, the rate of enamine formation is lower by about a factor of four than for arginine (Noziere and Cordova, 2008).Under the molar concentrations of glycine (and tyrosine) used in this study (about 2-4 M), the AA concentration is sufficiently high that any enamine that may form from CHOCHO would likely react via a Mannich pathway.Glycine concentrations used in this study are about two orders of magnitude higher than those found in atmospheric particles (up to 0.02 M) (Zhang and Anastasio, 2003;Matsumoto and Uematsu, 2005;Noziere and Cordova, 2008); however, our concentrations are low compared to those employed in Corrigan et al. (2008), where essentially pure solid AA were used.Under such particular experimental conditions Corrigan et al. (2008) argue that the presence of particle phase nucleophiles, sulfate and acidity all appear to increase CHOCHO uptake.However, growth factors observed on sodium sulfate are larger than growth factors on any AA.Arginine and glutamic acid experiments are the only AA systems that showed enhanced uptake over sodium chloride, which was chosen as a reference seed that only forms glyoxal oligomers in the associated water phase (no reactions with sodium chloride are expected).The observed growth factors on sodium chloride were larger than for glycine and other AA studied, including one of the two glutamic acid experiments.Overall, the enhancements attributed to AA catalysis over salt solutions were, if observable, rather small (Corrigan et al., 2008).In our study addition of glycine and tyrosine to AS+FA seed did reduce the rate of CHOCHO uptake by 2-3 orders of magnitude compared to experiments in the absence of AA in the photochemical system.We observe growth rates in the dark to be comparable in the presence and absence of AA, and do not find evidence that the most abundant AA in atmospheric aerosols, i.e., glycine, enhances the reactivity of CHOCHO in aerosols.
Similarly, acid catalyzed heterogeneous reactions of CHO-CHO are found second to photochemical reactions to enhance the CHOCHO uptake to particles.For the mildly acidified seeds investigated here (see Table 1) the photochemical rate of CHOCHO uptake is reduced by several orders of magnitude compared to identical seeds that were not acidified with SA.Differences in experimental conditions may explain results different from those observed by (Jang et al., 2002); those experiments were conducted with more acidic particles in a photochemical reaction system.An acid effect on CHOCHO uptake was also observed by Liggio et al. (2005a) and Corrigan et al. (2008), while -consistent with our results -no acid effect was observed by Kroll et al. (2005).Notably, in the apparent correlation of the growth rate with pK a (Corrigan et al., 2008) the strongest acid studied (oxalic acid) was an outlier to any apparent acid trend.Oxalic acid and SucA showed negligible CHOCHO uptake in the dark, despite rather high RH.This is consistent with the absence of type B growth on SucA observed here, and our finding that addition of SA had no effect on the rate of CHOCHO uptake in the dark.The apparent correlation of growth rate with pK a could hence equally be interpreted as LWC controlled CHOCHO uptake, and competing acid effects, as concluded in this work.LWC is not well constrained in Corrigan et al. (2008), and such interpretation seems likely also in light of the growth rate data for glycine and glutamic acid, which appears to be internally consistent only if regarded in context to the presumably higher LWC at higher RH.The reduced uptake of CHOCHO to acidified aqueous solutions observed in this work and Corrigan et al. (2008) appears to be analogous to the observation of reduced methylglyoxal uptake on more acidic solutions (55-85% w/w SA) (Zhao et al., 2006), which suggested a competition of acid for liquid water, and recognized the relevance of liquid water rather than acid as catalyst for SOA formation.Our study directly confirms such competition exists also for CHOCHO uptake due to organic photochemistry in aerosol water.At the pH found in atmospheric particles throughout the troposphere, acidity is also an inhibiting factor in SOA formation.In terms of the CHO-CHO reactivity enhancements, we conclude that sulfate is second to organic photochemical reactions; the presence of nucleophiles and SA reduces this reactivity.
The dependence of SOA yields on seed chemical composition demonstrates that the chemical reactions that form SOA depend on the chemical environment provided by the seed.Other than the direct role of providing reactive functional groups, indirect roles of the seed could include catalytic roles of aerosol water, seed chromophores facilitating photolytic processes, and particle pH influencing the reaction rate of CHOCHO processing in aerosol water.
4.2 Do SOA yields depend on the seed chemical composition also for other VOCs ?
Most studies of SOA yields to date used either no seed at all (e.g., Hurley et al., 2001 and references therein), or added AS seed in an attempt to overcome barriers in the Gibbs free energy towards homogeneous nucleation of semivolatile vapors, or ascertain the effect of ionic strength on the partitioning of semivolatile products (Odum et al., 1997;Cocker et al., 2001;Kroll et al., 2007).AS seed has been shown to accelerate the rate of SOA formation, but does not affect the SOA yield compared to experiments conducted in the absence of seed (Kroll et al., 2007).Notably, our observation of rapid photochemical SOA formation on AS could contribute to this acceleration, though AS is among those seeds that forms least SOA in this work.Cocker et al. (2001a, b) did not observe any measurable effect of RH on SOA yields from a-pinene, m-xylene and 1, 3, 5-trimethylbenzene, and attributed this to the existence of two separate SOA and aqueous salt phases in a single particle.This interpretation of phase separation is supported also by thermodynamic model considerations (Chang and Pankow, 2006).Consequently, experiments using seeds composed of inorganic aqueous salts, SOA, or mixtures of both will not be particularly sensitive to any aqueous phase organic photochemistry, as the amount of liquid water associated with SOA is much smaller than that associated with aqueous salts.Notably, the OHradical initiated oxidation of 1, 3, 5-trimethylbenzene does not form CHOCHO as oxidation product, and CHOCHO yields from a-pinene are very small.Also, CHOCHO yields from m-xylene are low compared to other alkylbenzenes (see Table 1 in Volkamer et al., 2007), reducing the sensitivity of all three systems further towards effects as those observed in this study.Regardless, CHOCHO and glyoxylic acid were identified in Cocker et al. (2001b) among the aerosol phase products.There is no currently known mechanism to produce glyoxylic acid via a gas-phase oxidation mechanism, while it is a known product from aqueous phase oxidation of CHO-CHO and other precursors (Ervens et al., 2003).The observation of glyoxylic acid among the constituents in SOA gives direct evidence that the molecular formation routes for SOA formation appear to have been affected by the presence of small amounts of water (either from multiphase chemistry at the interface between aqueous salt phases, or in water associated with SOA); CHOCHO and glyoxylic acid accounted for about 0.1% each of the extracted SOA mass from m-xylene under dry conditions.
To the best of our knowledge ours is the first study that systematically investigated the dependence of Y SOA on seed chemical composition.Unless SOA generated in chambers is equally water soluble as SOA generated in the atmosphere, our findings point to a possible bias in current simulation chamber data: a reduced interface for organic photochemistry in externally mixed SOA and water phases (within the same particle), compared to a greatly enhanced interface in bulk aqueous aerosols containing water soluble organic carbon (WSOC).WSOC is a major component of the seeds studied here, as well as the SOA produced in the atmosphere (Sullivan et al., 2006;Weber et al., 2007).If a similar effect of seed chemical composition was confirmed also for other SOA precursor VOCs, this would significantly increase the predicted amount of SOA formed on local, regional and global scales.4.3 Can CHOCHO uptake be described by Henry's Law ?Whether or not the partitioning of CHOCHO to aerosols can be described by Henry's Law is directly tied to the question whether the uptake mechanism is reversible or irreversible.The relevance of this question is illustrated from discussion in Volkamer et al. (2007): reversible CHOCHO uptake as described by Kroll et al. (2005) can explain several 10 ng m −3 of SOA mass for the typical urban CHOCHO concentrations found in Mexico City (Volkamer et al., 2005a); irreversible CHOCHO uptake (Liggio et al., 2005a) predict several 10 µg m −3 SOA.The missing CHOCHO sink to form SOA is closer to the upper end of this range, and can explain a significant fraction of the missing SOA source (Volkamer et al., 2006(Volkamer et al., , 2007)).Notably, this conclusion was reached without making any assumption about the reaction mechanism, reflecting that the uncertainty related to SOA formation from CHOCHO on aerosols spans several orders of magnitude.
For AS seed, Kroll et al. (2005) reproduced experiments by Liggio et al. (2005a) over extended experiment times, and observed that initial CHOCHO uptake slowed and reached a plateau after several hours.They concluded that CHO-CHO uptake to aqueous AS is reversible.This implies that the chemical bonds formed when CHOCHO reacts with sulfate in the dark (Liggio et al., 2005c), or forms oligomers in aerosol water (Kroll et al., 2005;Hastings et al., 2005;Liggio et al., 2005c) are not stable and appear to be broken already at ambient temperatures.Corrigan et al. (2008) observed continued growth upon exposing CHOCHO to (mostly dry) aspartic acid, sodium sulfate, arginine, glutamic acid aerosols (among others) in a dark chamber, and attribute this growth to reactions of CHOCHO with sulfate, nucleophiles (AA), and oligomer formation in thin water layers associated with mostly dry aerosols.The apparent growth rate appears to decrease with time, and it is not clear at present if it converges towards a plateau.It should be noted that the results by Corrigan et al. (2008) on dry seed (0<LWC<4% w/w showed most growth) and in the dark are somewhat difficult to compare with our results on mostly aqueous aerosols, where AA and/or SA largely reduce the photochemistry enhanced aerosol growth rate (see also Sect. 4.1.).Our findings that photochemical aerosol growth directly scales with (1) the CHOCHO concentration, and (2) the LWC for any given seed type are compatible with Henry's Law behavior, but allow no definite conclusion whether such growth is reversible or irreversible (see Sect. 4.1.).About half our experiments showed continued growth in the dark over accessible time scales; however, about an equal number showed no such growth.While an increasing number of studies now provide evidence that CHOCHO partitions to aerosols in amounts larger than expected, it remains to be demonstrated convincingly whether such CHOCHO uptake proceeds by a reversible or irreversible mechanism.Notably, we choose to present our results in terms of H eff values primarily for means of a quantitative comparison with other studies, and for discussion purposes, and do not mean to imply that photochemical CHOCHO uptake is necessarily reversible.
Atmospheric implications
It is of interest to place the results from the present work into perspective in terms of our understanding of the mechanism of SOA formation in the atmosphere.Traditional SOA models struggle with predicting the fast timing and the large amount of SOA produced in the atmosphere (Heald et al., 2005;de Gouw et al., 2005;Johnson et al., 2006;Volkamer et al., 2006;Riddle et al., 2008).Recent experimental work has further demonstrated that only a portion of the organic aerosol seems to participate in the absorption processes for semivolatile oxidation products (Song et al., 2007).This evidence further increases the already substantial mismatch between SOA predictions and observations.Traditional SOA models contain a high bias with respect to the predicted amount of SOA from semivolatile oxidation products in the atmosphere.Most SOA models represent SOA formation as the result of the gas-phase oxidation of VOCs to form semivolatile products that partition by absorption (dissolution) into an organic particle phase (Odum et al., 1997;Seinfeld and Pankow, 2003;Robinson et al., 2007).These models follow (Pankow, 1994a), according to which only the organic mass fraction of the aerosol (M om ) contributes to absorbing semivolatile gases; the partitioning coefficient for an absorptive process is expressed as a function of physical and thermodynamic properties of the semivolatile compound: where R is the ideal gas constant, T is temperature, M om (see above), MW is the mean molecular weight of the absorbing aerosol phase, ξ i is the activity coefficient of compound i in this phase, and p o i is the vapor pressure of compound i as a pure liquid.M om is defined to include any liquid water that is attracted due to hygroscopic properties of organic material (see Eq. ( 4) in Seinfeld and Pankow, 2003).According to Eq. (2) aerosol components other than M om , such as crustal material, inorganic aerosol, and liquid water attracted by inorganic aerosol do not contribute to SOA formation.In this view, for any given SOA precursor NMHC there is one set of Y SOA and M om data pairs; i.e., Y SOA is treated as a chemically inert quantity that should not depend on the chemical seed composition.The seed role is limited to accelerating SOA formation by providing a surface for condensation.
In this work the Y SOA values for any given seed do not correlate with M om , but instead show an excellent correlation with LWC (see Fig. 3,and Sect. 3.1.).Complementary evidence for the relevance of LWC for SOA formation from recent field observations (Hennigan et al., 2009) is consistent with our results.Moreover, at any given M om the Y SOA is found to vary as a strong function of the chemical seed composition (see Sect. 3.2.),providing direct evidence that SOA formation from C 2 H 2 is controlled by heterogeneous chemistry.Internal mixtures of inorganic and water soluble organic aerosol (fulvic and humic type matter) increase Y SOA over pure inorganic aerosol.This highlights an important function of inorganic aerosol, namely to attract LWC, and thus increase the reactor volume where organic photochemistry causes the Y SOA enhancements over pure salt solutions.Such SOA largely escapes the parameterization of SOA yields derived in simulation chamber experiments of other SOA precursors (see Sect. 4.4.4.).
Vapor pressure as the only parameter to control SOA formation clearly is insufficient to explain SOA formation from C 2 H 2 , and the variability of SOA yields at fixed M om observed in this study.SOA models require a conceptual expansion to also reflect SOA formation via the photochemical and dark solubility -reactivity scale in aqueous aerosols.The implementation of such SOA formation in atmospheric models requires knowledge of partitioning coefficients (e.g., effective Henry's law constants, reactive uptake coefficients), the oxidation mechanism and chemical kinetics of reactions that foster and inhibit SOA formation, ideally under conditions that resemble those found in the atmosphere.
Aerosol vs. cloud water as a chemical reactor
The chemical environment in aerosol water is fundamentally different from that in cloud or fog water, because (1) the liquid water content of aerosols is 3-5 orders of magnitude smaller than in clouds/fog droplets, and (2) the smaller diameter of aerosols results in 2-3 orders of magnitude higher surface to volume ratio for aerosols compared to cloud/fog droplets.The different chemical environment in aqueous aerosols compared to cloud/fog droplets is illustrated in Fig. 8. Based on measurements of the effective Henry's law constants of dilute solutions of CHO-CHO in water H eff =4.2×10 5 M atm −1 (Ip et al., 2009), and seawater H eff =3.6×10 5 M atm −1 (Zhou and Mopper, 1990), CHOCHO should partition in appreciable amounts only to clouds/fog, but not to aqueous aerosols.Even the higher CHOCHO solubility in aqueous AS particles, H eff =2.6×10 7 M atm −1 at 50% RH (Kroll et al., 2005), does not significantly affect the gas-phase CHOCHO budget in the absence of clouds.However, the presence of dissolved organic matter can photochemically enhance the CHOCHO uptake.Using our measured effective solubility of CHOCHO in aqueous fulvic acid particles, H eff =6.1×10 8 M atm −1 , and assuming 10 µg m −3 of aerosol liquid water, about 13% of the CHOCHO will be in the particle phase.For Mexico City, where organic aerosols comprise the largest fraction of fine particulate mass (Salcedo et al., 2006), our laboratory results predict up to 24% of the CHOCHO will partition to aerosols.Clearly, aerosols can affect the gas-phase CHOCHO budget to a measurable extent despite their small LWC.
The aerosol aqueous phase as a chemical reactor is distinctly different from cloud water.Notably, the source of oxidants is different.H 2 O 2 does not partition in appreciable amounts to aerosols; however, OH-radical uptake from the gas-phase provides a significant source of oxidants in particles (Ervens et al., 2003).Also, compared to the microto milli-molar dilute solutions characteristic of cloud or fog water (Carlton et al., 2007), reaction rates of CHOCHO are likely to be accelerated in 0.1-10 M solutions characteristic of aqueous aerosols.The nature of oligomer formation is non-linear (Barsanti and Pankow, 2005), and requires experiments to be conducted under essentially different conditions that resemble aerosols.
Anthropogenic vs. biogenic SOA sources
Global simulations indicate that biogenic NMHC oxidation is the major global source of SOA and this source has been enhanced by humans due to the anthropogenic increases in gas-phase oxidant and primary particle levels (Kanakidou et al., 2005;Fuzzi et al., 2006).This view is supported by 14 C measurements, which attribute a modern carbon age to most OA found in the atmosphere.However, field observations show no correlation between enhancements of OA mass and biogenic precursors (de Gouw et al., 2005;Sullivan et al., 2006;Weber et al., 2007), but instead show an excellent correlation with anthropogenic pollution factors (de Gouw et al., 2005;Volkamer et al., 2006;Sullivan et al., 2006;Weber et al., 2007).Such conflicting evidence from 14 C data and correlative analysis of time series data is difficult to resolve within the purely physical (vapor pressure driven) framework of SOA formation that underlies most SOA models; such models necessarily maintain the 14 C signature of precursors also in the 14 C signature of the semivolatile oxidation products.It should be noted, however, that a primarily modern 14 C age of OA is not necessarily indicative of biogenic NMHC precursors.Rather our interpretation of the information content of 14 C measurements depends on the framework that we choose to think about SOA formation.If, for example, the framework for SOA formation was expanded to include heterogeneous (photo) chemistry, other factors could increase or decrease particulate phase reactivity of SOA building blocks.While the possibility for such bias in the interpretation of 14 C age could in principle be caused by a biogenic or anthropogenic trigger of SOA formation, only an anthropogenic trigger to enhance SOA formation from biogenic NMHC is compatible with the available field evidence (de Gouw et al., 2005;Volkamer et al., 2006;Sullivan et al., 2006;Weber et al., 2007;Riddle et al., 2008).Further, there is no experimental evidence as yet that would demonstrate convincingly that equal amounts of OA are formed from biogenic NHMC precursors in the presence and absence of anthropogenic pollution factors.
The global source of α-dicarbonyls is estimated to be at least 185 Tg yr −1 (Fu et al., 2008).The CHOCHO source is estimated at 45-83 Tg yr −1 , and largely due to secondary sources (17% from C 2 H 2 , 70% from biogenic NMHCs) (Myriokefalitakis et al., 2008;Fu et al., 2008); an additional oceanic source (Wittrock et al., 2006) is currently largely uncertain.Aerosol uptake to form SOA could determine 75-95% of the atmospheric fate of CHOCHO in polluted air, and form about an equal amount of SOA as the sum of semivolatile oxidation products from traditional anthropogenic and biogenic SOA precursor NMHC in Mexico City (Volkamer et al., 2007).CHOCHO, and other α-dicarbonyls, Cl d Kroll et al. (2005) aerosol liquid water.The LWC of clouds (blue line) and fog (orange line) is 3-5 orders of magnitude larger than that of aerosols (pink area).H eff values on the order of 3.5×10 4 M atm −1 (blue dot), 3.5×10 5 M atm −1 (orange dot) and 2×10 9 M atm −1 (black dot) characterize equal parts in the gas-and dissolved phase for clouds, fog and aerosols, respectively.Dotted lines indicate H eff -LWC pairs that favor the gas-phase; dashed lines indicate H eff -LWC pairs where the dissolved phase is the dominant reservoir (the partitioning ratio, defined as dissolved/gas-phase, is indicated next to each line).The vertical lines represent H eff values as measured for CHOCHO in (a) bulk water (Ip et al., 2009), (b) metastable AS aerosols (Kroll et al., 2005), (c) metastable AS+FA aerosols (this work), and (d) aqueous ambient aerosols in Mexico City (Volkamer et al., 2007).Aerosols are distinctly different aqueous/multiphase chemical reactors than clouds/fog.
are deemed responsible for 5-11 Tg yr −1 SOA production on global scales, which is comparable to the sum of SOA formed from monoterpenes, sesquiterpenes, isoprene, and aromatics (Fu et al., 2008;Myriokefalitakis et al., 2008); these estimates underestimate the role of aerosol seeds, and will further increase based on the results of this study.WSOC accounts for a major portion of OA in urban and rural areas (Sullivan et al., 2006;Weber et al., 2007).Our observation of fast SOA formation from WSOC photochemistry of CHO-CHO is consistent with field evidence of fast SOA formation due to first generation oxidation products, and helps explain part of the missing SOA source in Mexico City (Volkamer et al., 2006).However, our results fall short by a factor of 7-14 to explain the CHOCHO sink to form SOA in Mexico City (Volkamer et al., 2007).
Given the large mismatch between predicted and observed SOA on local and global scales, little trust should be placed in the dominance of biogenics on global scales based on these models alone.Clearly, the source of α-dicarbonyls, and the uncertainty associated with their atmospheric fate, www.atmos-chem-phys.net/9/1907/2009/Atmos.Chem.Phys., 9,2009 is large enough to affect the interpretation of 14 C measurements on scales from local to global.Anthropogenic precursors could be responsible for about one third of the global SOA production (Volkamer et al., 2006).Our results support the expanded definition of SOA formation, according to which SOA sources include gas-to-particle partitioning in the organic-and aqueous phase, organic phase processing, aqueous phase processing, and multiphase processing (i.e., surface and bulk) (Volkamer and Kanakidou, 2007).
4.4.4On the issue of double-counting SOA from CHO-CHO Most traditional SOA precursor hydrocarbons are also precursors for CHOCHO, i.e., about 80% of CHOCHO in Mexico City forms from aromatic precursors (Volkamer et al., 2007).The parameterization of SOA yields from traditional SOA precursor hydrocarbons that underlie current SOA models could already include SOA formed from CHO-CHO.If SOA from CHOCHO was treated as an explicit additional pathway in those models, would CHOCHO SOA be double counted?It appears unlikely for the following reasons: (1) the SOA source from CHOCHO closely compares to (and slightly exceeds) that of all other currently studied SOA precursor NMHC (Volkamer et al., 2007;Fu et al., 2008); all SOA would need to be CHOCHO related.This is obviously not the case, as most aerosol products that have been identified to date at the molecular level are unrelated to CHOCHO (Yu et al., 1995a;Forstner et al., 1997;Cocker et al., 2001).(2) Most chamber experiments were conducted under dry experimental conditions, and on AS seed.Such conditions essentially preclude SOA formation via the aerosol aqueous phase.AS showed among the lowest Y SOA values in this study.In the presence of relative humidity there is consistent experimental and theoretical evidence (Cocker et al., 2001a, b;Chang and Pankow, 2006) that strongly suggests that WSOC photochemistry is greatly reduced in current chamber data due to phase separation effects (see also Sect.4.2.).(3) The aerosol mass spectra of SOA generated in the presence of aqueous AS seed in chambers (Bahreini et al., 2005) are different from OOA spectra observed in ambient air (Zhang et al., 2005), and SOA generated from CHOCHO uptake to AS (Liggio et al., 2005a).Notably, CHOCHO uptake to AS accounts only for 2-3% of the missing CHOCHO sink to form SOA in ambient air (Volkamer et al., 2007); WSOC photochemistry is likely to result in different aerosol mass spectra than CHOCHO uptake to AS.Further, the differences in the mass spectra of (Zhang et al., 2005;Bahreini et al., 2005) appear to be largest for masses that have a high contribution from oxygen containing products (e.g., m/z 28, which corresponds to CO). ( 4) The oxygen-to-carbon ratio (O/C) for the di-hydrated CHOCHO monomer is O/C=2, and this ratio approaches O/C=1 for longer chain oligomers; O/C is likely to be high also for the presently unidentified products.As noted in (Volkamer et al., 2007) SOA from CHO-CHO will increase predicted O/C ratios (see also Dzepina et al., 2009).While the high O/C ratios of CHOCHO related SOA compare favorably to the O/C=0.9observed in ambient aerosol, the range of observed values in chamber experiments from toluene, 1, 3, 5-trimethylbenzene, isoprene, α-pinene is much lower, i.e., 0.3<O/C<0.4(Aiken et al., 2008).It is these low O/C ratios that are currently parameterized in atmospheric SOA models.Finally, (5) SOA generated in chambers is more volatile than ambient SOA (Huffman, et al., 2009).Thereby, the more oxidized fraction of the aerosol is least volatile.Typical values for the heat of vaporization H of SOA generated from toluene in chambers range between 11< H<31 kJ mol −1 (Takekawa et al., 2003;Offenberg et al., 2006).The H of SOA generated from single component CHOCHO aerosol is H=13.3 kJ mol −1 (Offenberg et al., 2006).While the volatility of CHOCHO related SOA is difficult to separate from the convoluted volatility of ambient or chamber generated organic aerosol, the CHO-CHO related SOA is likely to lower the effective volatility of ambient aerosol.Conversely, increasing the CHOCHO related portion of SOA in chamber experiments (see Sect. 3.2. and 4.3.)and ambient aerosol (Volkamer et al., 2007) will likely bring the predicted and observed volatility of SOA in closer agreement (see also Dzepina et al., 2009).
Overall, it appears likely that the predominant fraction of CHOCHO related SOA escapes the current parameterizations of SOA yields from chamber experiments (e.g., Odum et al., 1997;Ng et al., 2007) that underlie current estimates of SOA formation from atmospheric models (Kleeman and Cass, 2001;Chung and Seinfeld, 2002;Kanakidou et al., 2005;Fuzzi et al., 2006;Henze et al., 2008).
Conclusions
The formation of Secondary Organic Aerosol (SOA) from C 2 H 2 , the lightest Non Methane HydroCarbon (NMHC), is surprising, because all currently known oxidation products are at least six orders of magnitude too volatile to condense or partition into the aerosol organic phase to any appreciable extent.SOA formation from C 2 H 2 highlights the importance of heterogeneous chemistry in SOA formation.Our results point out the particular role of organic photochemistry in aerosol water as a SOA source from CHOCHO, and possibly other α-dicarbonyls.Our laboratory data is consistent with recent field observations (Hennigan et al., 2009), suggesting that organic photochemistry in LWC is very likely relevant also for other SOA precursor VOCs.This is to our knowledge the first systematic study of the effect of seed chemical composition and photochemistry on SOA yields, and the rate of SOA formation.The seed particles investigated here are representative for a large subset of aqueous atmospheric particles.The SOA yields for mixed inorganic/water soluble organic carbon (WSOC) seed aerosols are about four times larger compared to pure inorganic aerosol, but are comparable for the pure water soluble organic seed aerosols.We conclude that the inorganic aerosol component primarily determines the reactor size, i.e. the liquid water content (LWC), where WSOC photochemical reactions cause the increased SOA yields.Any ionic strength effects of concentrated salt solutions appear to be secondary to providing a sizable photochemical reactor.Further, ours is the only study to date that investigates CHOCHO uptake to aerosols in the presence of OH-radicals and UV light.The rate of SOA formation is about a factor of 500 higher under photochemical conditions compared to dark conditions, i.e., fast enough to compete with rapid gas-phase losses.
The implications for our understanding of atmospheric processes are significant: (1) atmospheric SOA models currently widely represent SOA yields from experiments conducted in the absence of seed, or in the presence of pure ammonium sulfate seed.Here we find the highest SOA yields for internally mixed inorganic/WSOC seed particles due to organic photochemistry in the bulk LWC of the particles.Most SOA precursors for which their chemical identity is known to date are major sources for CHOCHO, and other α-dicarbonyls.Hence, any low bias in the currently available SOA yield data might explain a significant portion of the missing SOA source in current atmospheric models.(2) The effective Henry's Law constants H eff for CHOCHO are sufficiently high to compensate for the 3-5 orders of magnitude smaller amount of liquid water associated with atmospheric particles compared to clouds/fog.We demonstrate that WSOC photochemistry of CHOCHO -and possibly other α-dicarbonyls -is a SOA source already in the absence of clouds/fog.Most notably, the SOA formation associated with these gases does not stop once a cloud droplet has evaporated, as inferred by current SOA models that represent cloud processing, but rather proceeds also in the absence of clouds, possibly even at an enhanced rate as aerosols are more concentrated solutions compared to clouds.(3) The atmospheric relevance of α-dicarbonyls as a source for SOA depends on the competition between the rate of uptake to particles compared to other gas-phase loss rates.Our results demonstrate that organic photochemistry driven by OHradical, and possibly light induced reactions is needed to enhance the rate of CHOCHO uptake to be competitive with rapid gas-phase losses.(4) The rate of CHOCHO uptake to particles in our photochemical system is fostered by most seeds studied but reduced by the presence of atmospherically relevant amino acids like glycine, and sulfuric acid.This finding is difficult to reconcile with the idea of acid catalysis to enhance the rate of CHOCHO uptake, and points to a more complex role of particle pH, i.e., identifies a strong pH dependence of the rate of photochemical reactions in the aqueous aerosol phase.(5) Our results explain up to 430 ng m −3 of the missing SOA source in Mexico City as CHOCHO related (Volkamer et al., 2006), and account for up to 7-14% of the missing CHOCHO sink in Mexico City (Volkamer et al., 2007).While our results reduce the existing imbalance in current gas-phase CHOCHO budgets most of the imbalance in gas-phase CHOCHO cannot be explained as yet based on the currently available laboratory data.
There is an urgent need for further laboratory studies to determine the reaction mechanism, kinetics, and products of this organic photochemistry.Atmospheric SOA models need to be expanded to explicitly represent the aerosol aqueous phase chemistry in form of dark and photochemical reactions in the presence and absence of clouds.
3. 1 Fig. 3 .
Fig. 3. SOA yield (Y SOA ) as a function of organic seed mass, M om (A, C, E) and liquid water content, LWC (B, D, F).No obvious correlation is found between Y SOA and M om , but Y SOA correlates well with LWC for all seed-types studied.For pure AS (solid line), AS+FA (dashed line), and pure ABS (dotted line) sufficient experiments were conducted to demonstrate that the increase in Y SOA with LWC is linear.
photochemical, type A
Fig. 5 .
Fig.5.Time series of the aerosol volume for type A uptake kinetics.Data from experiment #20 (ABS+HA) is shown here.Grey areas indicate times with lamps ON.No effect is observed if the seed is exposed to light in the absence of reagents in the chamber.After addition of chemicals to the gas-phase, an immediate shift in aerosol volume is observed when C 2 H 2 is photochemically oxidized (twice here), followed by type B growth in the dark.Type B growth as parameterized in Fig.3for AS+FA can explain that observed for ABS+HA seed (solid line), see text.The rapid volume growth during periods when the light is turned ON is proportional to the amount of CHOCHO produced, and inversely proportional to LWC of the seed aerosols.
Fig. 6 .
Fig.6.RH dependence of the effective Henry's law constants, H eff , of CHOCHO for different seed aerosols: (A) Ammoniumsulfate mixed with fulvic acid (AS+FA), (B) Fulvic acid (FA), (C) Ammoniumsulfate (AS), (D) Ammoniumbisulfate (ABS).Open symbols represent values of H eff derived under the assumption that CHO-CHO partitions to the overall seed volume (case 1).Solid symbols represent values of H eff derived under the assumption that CHO-CHO partitions only to the portion of the seed volume that corresponds to aerosol liquid water (case 2).The apparent RH dependence of H eff in case 1 is fully accounted in case 2, presenting direct evidence that CHOCHO uptake proceeds through the aerosol aqueous phase.
Fig. 7 .
Fig. 7. Water soluble organic carbon (WSOC) photochemistry enhancement of CHOCHO partitioning to AS+FA and pure FA aerosols compared to pure salt solutions (AS).
4.4.1 Conceptual expansion of SOA models: the photochemical solubility-reactivity scale
Fig. 8 .
Fig.8.Partitioning of water soluble molecules to cloud/fog vs. aerosol liquid water.The LWC of clouds (blue line) and fog (orange line) is 3-5 orders of magnitude larger than that of aerosols (pink area).H eff values on the order of 3.5×10 4 M atm −1 (blue dot), 3.5×10 5 M atm −1 (orange dot) and 2×10 9 M atm −1 (black dot) characterize equal parts in the gas-and dissolved phase for clouds, fog and aerosols, respectively.Dotted lines indicate H eff -LWC pairs that favor the gas-phase; dashed lines indicate H eff -LWC pairs where the dissolved phase is the dominant reservoir (the partitioning ratio, defined as dissolved/gas-phase, is indicated next to each line).The vertical lines represent H eff values as measured for CHOCHO in (a) bulk water(Ip et al., 2009), (b) metastable AS aerosols(Kroll et al., 2005), (c) metastable AS+FA aerosols (this work), and (d) aqueous ambient aerosols in Mexico City(Volkamer et al., 2007).Aerosols are distinctly different aqueous/multiphase chemical reactors than clouds/fog.
Table 2 .
Products formed in the C 2 H 2 /CH 3 ONO/NO x /light system, effective Henry's law coefficients.Values in the right column have been normalized to CHOCHO.
remained constant in all experiments presented in this work.The aerosol volume growth corresponding to this shift in the aerosol size distribution was used to quantify SOA formation.
The timing of SOA formation: photochemical vs. dark systems ). Open symbols represent values of H eff derived under the assumption that CHO-CHO partitions to the overall seed volume (case 1).Solid symbols represent values of H eff derived under the assumption that CHO-CHO partitions only to the portion of the seed volume that corresponds to aerosol liquid water (case 2).The apparent RH dependence of H eff in case 1 is fully accounted in case 2, presenting direct evidence that CHOCHO uptake proceeds through the aerosol aqueous phase.normalization to facilitate a quantitative discussion between different seeds and experimental conditions, and do not mean to imply any conclusion about the reaction mechanism.H eff values for AS, ABS, FA and AS+FA are plotted as a function of RH in Fig.6.H eff values were calculated in two ways:(1) based on the measured seed volume (black open dots) and (2) based on the portion of the seed that was LWC (solid red dots).Calculations of H eff that are based on the measured seed volume give systematically lower values; those H eff values further show an apparent dependence of RH.Notably, this apparent RH dependence is fully accounted after normalization of the LWC portion of the seed.Such calculation gives essentially constant and higher H eff values for all seeds shown, over the entire range of RH probed.This further corroborates that LWC is a crucial quantity for determining the Volkamer et al.: WSOC photochemistry of CHOCHO V seed seed volume before the experiment V norm,t total aerosol volume after normalization, as defined in Eq. (1) WSOC water soluble organic carbon Y SOA SOA yield, Y SOA = M SOA / C 2 H 2 ζ i activity coefficient of compound i t observed aerosol volume at any given time t www.atmos-chem-phys.net/9/1907/2009/Atmos.Chem.Phys., 9, 1907-1928, 2009 R. | 2019-01-19T16:40:41.915Z | 2008-08-05T00:00:00.000 | {
"year": 2008,
"sha1": "b18a78b3f3c635895c2f51904c84a03f6b4a8de8",
"oa_license": "CCBY",
"oa_url": "https://acp.copernicus.org/articles/9/1907/2009/acp-9-1907-2009.pdf",
"oa_status": "GOLD",
"pdf_src": "ScienceParseMerged",
"pdf_hash": "b18a78b3f3c635895c2f51904c84a03f6b4a8de8",
"s2fieldsofstudy": [
"Environmental Science",
"Chemistry"
],
"extfieldsofstudy": [
"Chemistry"
]
} |
236955799 | pes2o/s2orc | v3-fos-license | Cytometric blood phenotyping in free-living white-eyed parakeet (Psittacara leucophthalmus)1
RESUMO.[Fenotipagem citométrica do sangue de maritacas jovens de vida-livre (Psittacara leucophthalmus).] Este estudo teve como objetivo realizar a fenotipagem citométrica com 3,3-di-hexiloxacarbocianina [DiOC6 (3)] de amostras de sangue de maritacas jovens de vida-livre (Psittacara leucophthalmus). As amostras de sangue total, coradas com DiOC6(3) de 19 maritacas de vida livre, foram analisadas por citometria de fluxo e os tipos de células foram distinguidos por sua fluorescência típica no canal laser azul (FL-1) e SSC (dispersão lateral). Foi possível diferenciar eritrócitos (58,3±13,6) de leucócitos (32,4±13,1) e algumas subpopulações de leucócitos: linfócitos/trombócitos (29,7±7,7), monócitos (30,6±8,5) e granulócitos (5,9-26), entretanto, linfócitos e trombócitos não puderam ser diferenciados em duas populações distintas. Nosso estudo determinou que a população predominante P. leucophthalmus foi mononuclear agranulocítica em comparação com a taxa de aquisição da população granulocítica. A metodologia de citometria de fluxo com uso da coloração de DiOC6(3) foi aplicável a amostras sanguíneas das maritacas e pode ser estudado e aplicado para outras espécies de psitacídeos.
INTRODUCTION
Blood evaluations are an important practice in medicine that allows diagnosing diseases, monitoring the patient's health, and interpreting the physiological responses concerning the disease challenge. In avian, the hematological practice is useful for understanding the effects of diseases and stress . However, the automatic count by impedance methods is difficult because nucleated erythrocytes are considered like small lymphocytes or thrombocytes. The availability of antibodies for differentiating blood cells by immunophenotype is a problem when compared to mammalians samples because of antibodies in relation to species-specific and cost antibodies for free-living animals (Uchiyama et al. 2005).
Automated techniques, initially developed for blood evaluation in humans, have also become routine procedures in veterinary medicine. However, automatic blood counting methods, adapted and widely used for diagnosis in mammals have limited use in avian hematology due to the presence of nucleated erythrocytes and thrombocytes (Walberg 2001, Welles et al. 2009).
The flow cytometry offers with speed and accuracy the possibility to analyze and discriminate the cell populations on the basis to size (FSC) and complexity (SSC), and when complemented with a fluorochrome (Morgan et al. 1993) it provides a simple method for differential a cell or your metabolic activities. An example, Inoue et al. (2002), Moritomo et al. (2002) and Uchiyama et al. (2005) described the use of specific fluorescent lipophilic dye, 3,3-dihexyloxacarbocyanine [DiOC 6 (3)] for differentiation through flow analysis with quail, chicken, turkey, and fish blood cells. With this dye is possible to discern the white blood cells from nucleated erythrocytes sin previous cell lysis, and quantify and evaluation cellular behavior like used to CD45 antigen (De Boever et al. 2010). The DiOC 6 (3) can be used a fluorescent dye in the mitochondrial activity, sperm analysis, differentiation and cell growth analysis, in addition to endoplasmic reticulum evaluation (Sabnis et al. 1997).
A detail in the avian clinical practice about blood sampling there is a limited amount of blood by live weight, particularly, when the avian patient is a minimal size or debilitated (Clark et al. 2009). Here, there are principal advantages of cytometry for this species, with few amounts of blood it possible to performer hematology and immunological studies.
The 19 free-living white-eyed parakeets (Psittacara leucophthalmus) used in this study were admitted at the Veterinary Hospital from the FCAV-Unesp of Jaboticabal, São Paulo, Brazil, in the early year on period of hatching of the eggs for this species in this region of the country.
These parakeets frequently nest near or inside human constructions, which provide close contact between the birds and humans. Mostly, these birds have an ingenious behavior and leave the nest before learning to fly then were found on the ground by people, which brings them to the hospital like a rescue service.
All parakeets of this study presented good body condition, normal hydration, normal behaviour and absence of clinical alterations at the time of blood collection. After a hosting period, the birds are released in nature. The gender of the parakeets was not determined, approximately 3-4 months old and weighed between 94 and 172 grams. After physical restraint, a 0.2ml blood aliquot was collected from the right jugular vein with a heparinized syringe. The samples were kept under refrigeration (4-8 o C) until the analysis.
The DiOC 6 (3) (3,3-dihexyloxacarbocyanine, Molecular Probes, Eugene/OR, U.S.A.) staining was conducted following the methodology previously proposed . A stock DiOC 6 (3) solution was previously prepared in ethanol for analysis at 500µg ml-1. Right before staining, the stock solution was diluted ten times in Hank´s Balanced Salt Solution no calcium/magnesium (HBSS) (Thermo Fischer Scientific, Brazil). Fresh blood (10µl) was delivered into a test tube (Falcon 2059, Becton Dickinson, San Jose/CA, U.S.A.) and gently mixed with HBSS 1x (1950µl) and the diluted dye solution (40µl), followed by incubation at room temperature for 10 minutes.
After stained with DiOC 6 (3), the blood cells were analysed using flow cytometry (FACSCanto II, BD BioScience, Philadelphia, 2004) with BD FACSDiva™ software. Forward scatter (FSC), side scatter (SSC), and green fluorescence (FL-1/488nm) of each cell was measured. The workflow for gate strategy can be visualized in Figure 1-6. Ten thousand events in a single cell gate were acquired (FSC-A x FSC-H). After, only positive events [DiOC 6 (3)+ gate] were analyzed in the blue laser channel (FL-1 -488nm) with detection of ≈525nm were labeled FITC-A and FITC-H in the graphs.
The positive events with DiOC 6 (3)+ stain were divided into two gates concerning fluorochrome vs. complexity (FITC-H vs. SSC-H) that were R1 (red blood cells -RBC) and R2 (white blood cells -WBC) like related by Inoue et al. (2002) and Moritomo et al. (2002). Additionally, cell number was expressed as average and percentage related to the population hierarchy.
The white blood cells (WBC) in R2 gate were divided in three gates in relation FSC vs. SSC and the gates were labeled in granulocyte (GRAN), monocyte (MON) and lymphocyte/thrombocyte like related by Inoue et al. (2002) and Moritomo et al. (2002). All values were submitted to Shapiro-Wilk Test (p≤0.05) in the R software (R Core Team 2011) and were described as a descriptive statistic.
RESULTS AND DISCUSSION
To the best of our knowledge, this is first report of dye phenotyping in free-living white-eyed parakeets (Psittacara leucophthalmus). The gates strategic of flow cytometry can visualized in Figure 1-6. In dotplot of FL-1-A x SSC-A, two distinct population were observed: 1) low FL1 (red blood cells -RBC) with 58.3% (44.7-71.9); and, 2) high FL1 (white blood cells -WBC) with 32.4 % (19.3-45.5).
With our results, we found a similar cellularity behaviour as described by Inoue et al. (2002), Moritomo et al. (2002) and Uchiyama et al. (2005). Con basis in the literature, we considered a subdivision in WBC dotpot into three population by the FSC-A x SSC-A. These populations were gated in granulocytes (R2-1), monocytes (R2-2), and lymphocytes/ thrombocytes (R2-3). Following, were created the population hierarchy based in size (FSC) and complexity (SSC) of graph sorting population (Table 1).
This is the first report of the phenotype of blood cells by 3,3-dihexyloxacarbocyanine [DiOC 6 (3)] in flow cytometry of white-eyed parakeets (P. leucophtalmus). We believed this report could to be an inspiration for other studies dedicated in avian hematology, immunopathology and disease effects in avian clinics. Fotopoulos (2012) claims that DiOC 6 (3) can revisited and several new application potentials have emerged in haematological, inflammatory and immune response studies.
The methodology of DiOC 6 (3) stain associated with flow cytometry has managed to get around the hematologic automatization ceilings were the nucleation of erythrocytes and thrombocytes associated the variability of cell morphology among the existing large number of avian species make it difficult the counting and differentiation in species with nucleated blood cells (Walberg 2001). Flow cytometry is widely used on human and mammals samples and provides a rapid and precise for detecting, counting and characterizing cells in mixed populations (Morgan et al. 1993. The DiOC 6 (3) staining was useful for distinguishing the white-eyed parakeets leucocyte from erythrocytes like the reported for common carp (Cyprinus carpio), green iguana (Iguana iguana) and quail (Coturnix coturnix) , Uchiyama et al. 2005. Even though our experiment not used cell sorting for to identify in light microscopy the cellularity, the cytometric behavior was the same found in literature and similar distribution when compared with leukocytograms of dogs and cats based on size and complexity (FSC vs. SSC) (De Boever et al. 2010). Sun et al. (2010) defined three subpopulations peripheral blood of penaeid shrimp (Penaeus vannamei) based on the relative size and complexity -granulocytes, semi-granulocytes and smallest cells that the authors classified as hyalinocytes that can be compared to lymphocytes to the nucleus that fills the entire cytoplasm and without granulations. Morgan et al. (1993) in their results of fish blood marked with acridine orange (AO) gated four population in FSC x SSC plot, erythrocyte, lymphocytes/thrombocytes, monocytes and granulocytes (heterophile/eosinophils). In these stains [AO and DiOC 6 (3)] is problematic cytometric differentiation concerning heterophils from eosinophils; thrombocytes from small lymphocytes and large lymphocytes from monocytes . Currently, there is no standard automated technique for blood count and differentiation in avian species.
An alternative to complement this differentiation is the use of monoclonal antibodies specific for each cell type may allow counting all the leucocyte subtypes in the whole blood (Seliger et al. 2012). There are immunophenotyping that has been reported in commercial poultry (Burgess & Davison 1999, De Boever et al. 2010, Seliger et al. 2012, meantime, the antibody is not commercially available for the different avian species. It is necessary to validate the poultry specific antibodies with other avian species. Bohls et al. (2006) used immunophenotyping of K1 (a thrombocyte and macrophage marker), to distinguish better the thrombocyte from the lymphocyte population, which is positive for K55 (leucocyte marker) in Leghorn chickens. Another example, the authors performed immunophenotyping antiCD-45 in blood samples from commercial laying hens in an attempt to separate lymphocytes from thrombocytes (De Boever et al. 2010) but mostly these markers only available was conjugated with FITC that makes it impossible to use multicolour panel.
The population containing thrombocytes and lymphocytes (R2-3) could not be separated by using DiOC 6 (3) staining. Ours results were similar that discussed in other papers, the DiOC 6 (3) were not effective either in distinguishing these two cell types due to similar cytometry characteristic of them. FACS analysis employing both monoclonal antibodies against thrombocytes and DiOC 18 (3) should be able to count all the cell types in blood (Morgan et al. 1993, Uchiyama et al. 2005).
An earlier study in the literature demonstrated that the contamination of the lymphocyte population with thrombocytes could be as high as 70% in avian blood samples, even in protocols that separate the mononuclear cells by the density gradient (Bohls et al. 2006).
CONCLUSIONS
Our study determined that the predominant population in white-eyed parakeet (Psittacara leucophthalmus) was lymphocytes, thrombocytes and monocytes in the leucocytes gates in comparison the granulocyte population.
We believe that the present study open up an opportunity for the study of leukocytes in parrots and also in other animals with nucleated erythrocytes for which of haematological automation and flow cytometer is limited. With a feasible and easily reproducible methodology, it will be possible to | 2021-08-09T08:09:31.699Z | 2021-01-01T00:00:00.000 | {
"year": 2021,
"sha1": "92e3d22af17b20675284f3df375f074199a6a2b2",
"oa_license": "CCBYNC",
"oa_url": "http://www.scielo.br/j/pvb/a/kymvGkK4yrMnSCjhQJZdWxs/?format=pdf&lang=en",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "92e3d22af17b20675284f3df375f074199a6a2b2",
"s2fieldsofstudy": [],
"extfieldsofstudy": []
} |
234487932 | pes2o/s2orc | v3-fos-license | Liquiritin Attenuates Lipopolysaccharides-Induced Cardiomyocyte Injury via an AMP-Activated Protein Kinase-Dependent Signaling Pathway
Background: Liquiritin (LIQ) is a traditional Chinese medicine that has been reported to regulate inflammation, oxidative stress and cell apoptosis. However, the beneficial effects of LIQ in lipopolysaccharides (LPS)-induced septic cardiomyopathy (SCM) has not been reported. The primary goal of this study was to investigate the effects of LIQ in LPS-induced SCM model. Methods: Mice were pre-treated with LIQ for 7 days before they were injected with LPS (10 mg/kg) for inducing SCM model. Echocardiographic analysis was used to evaluate cardiac function after 12 h of LPS injection. Thereafter, mice were sacrificed to collect hearts for molecular and histopathologic assays by RT-PCR, western-blots, immunohistochemical and terminal deoxynucleotidyl transferase nick-end labeling (TUNEL) staining analysis respectively. AMPKα2 knockout (AMPKα2−/−) mice were used to elucidate the mechanism of LIQ Neonatal rat cardiomyocytes (NRCMs) treated with or without LPS were used to further investigate the roles and mechanisms of LIQ in vitro experiments. Results: LIQ administration attenuated LPS-induced mouse cardiac dysfunction and reduced mortality, based upon the restoration of EF, FS, LVEDs, heart rate, dp/dt max and dp/dt min deteriorated by LPS treatment. LIQ treatment also reduced mRNA expression of TNFα, IL-6 and IL-1β, inhibited inflammatory cell migration, suppressed cardiac oxidative stress and apoptosis, and improved metabolism. Mechanistically, LIQ enhanced the phosphorylation of AMP-activated protein kinase α2 (AMPKα2) and decreased the phosphorylation of mTORC1, IκBα and NFκB/p65. Importantly, the beneficial roles of LIQ were not observed in AMPKα2 knockout model, nor were they observed in vitro model after inhibiting AMPK activity with an AMPK inhibitor. Conclusion: We have demonstrated that LIQ exerts its protective effects in an SCM model induced by LPS administration. LIQ reduced inflammation, oxidative stress, apoptosis and metabolic alterations via regulating AMPKα2 dependent signaling pathway. Thus, LIQ might be a potential treatment or adjuvant for SCM treatment.
INTRODUCTION
Sepsis is a life-threatening inflammatory disease and a major cause of death among patients in intensive care units due to infection-induced multiple organ failure (Rossaint and Zarbock, 2015). Septic Cardiomyopathy (SCM) characterized by cardiac dysfunction (Rello et al., 2017) has high mortality and morbidity in sepsis patients. Even though mechanisms involved in SCM have not been fully elucidated, SCM appears to be the result of inflammatory bursts, apoptosis, dysregulated energy metabolism and oxidative stress (Beesley et al., 2018;Luo et al., 2020;Tan et al., 2021). Strategies to attenuate SCM based upon inflammation, oxidative stress and energy metabolism have not been successful clinically (Rello et al., 2017). Thus, there remains an urgent need to develop new strategies to treat sepsisinduced SCM.
However, it remains to be determined whether LIQ could exert a protective role in LPS-induced SCM mouse model. Thus, this study evaluated the effect of LIQ on the pathogenesis of SCM and clarified the underlying mechanisms.
Ethics Statement
This study was approved by the Animal Care and Use Committee of Renmin Hospital of Wuhan University (approval number: WDRX-2018K010). All studies were carried out in accordance with the Guidelines for the Care and Use of Laboratory Animals published by the United States National Institutes of Health (NIH Publication, revised 2011).
Animals and Lipopolysaccharides-Induced Septic Cardiomyopathy Models
C57BL/6J mice (male, weighing 23.5-25.5 g, and aged 8-10 weeks) were purchased from the Institute of Laboratory Animal Science, Chinese Academy of Medical Sciences (CAMS) & Peking Union Medical College (PUMC) (Beijing, China), and were kept in specific-pathogen-free conditions in the Cardiovascular Research Institute of Wuhan University. Mice had access to food and water freely and were housed with a 12 hour light/dark cycle.
AMPKα2 knockout mice (AMPKα2 −/− ) described previously (Zhang et al., 2018b) were chosen to study the underlying mechanism. Wild-type mice were also used for direct comparison. All mice were pre-treated with LIQ (80 mg/kg) for 7 days and were then injected with 10 mg/kg LPS to induce SCM.
Echocardiography Analysis and Hemodynamics
Twelve hours following LPS administration, echocardiograph and hemodynamic studies were conducted to assess cardiac function as described previously (Zhang et al., 2018a). Mice were anesthetized by inhalation of isoflurane prior to echocardiography [Mylab 30CV (Esaote S.P.A, Genoa, Italy) with a 10 MHz linear array ultrasound transducer]. We selected and measured the short-axis standard view of the left ventricular papillary muscle. Heart rate (HR), left ventricular ejection fraction (LVEF), left ventricular fractional shortening (LVFS), left ventricular end-diastolic dimension (LVEDd), left ventricular end-systolic dimension (LVESd), end-diastolic interventricular septal thickness (IVSd), left ventricular posterior wall thickness (LVPWS) were recorded to evaluate mice cardiac function.
A microtip catheter transducer (SPR-839, Millar Instruments, Houston, TX, United States) was used to measure hemodynamic parameters. Mice were anesthetized with 1.5% isoflurane and the transducer was inserted into the left ventricle through the right carotid artery. Signals were continuously recorded using a Millar Pressure-Volume System (MPVS-400, Millar Instruments).
Neonatal Rat Cardiomyocytes Isolation and Treatment
Neonatal rat cardiomyocytes (NRCM) were prepared as described previously (Zhang et al., 2018a). Briefly, 1-3 days old neonatal SD rats were sacrificed by dislocation, the heart removed and cut into small pieces. Tissue was digested with 0.125% trypsin, centrifuged, the supernatant collected and filtered, and then the filtered cell suspension was seeded into a plate. After 2 hours, the myocyte count in the supernatant was measured using a hemocytometer, and then myocytes were seeded into 6, 24 or 96 well plates, and incubated in a 5% CO 2 incubator at 37°C with Dulbecco's modified Eagle medium (DMEM) (GIBCO, C11995) supplemented with 15% fetal bovine serum (FBS) and 0.1 mM (E)-5-(2-bromovinyl)-2′deoxyuridine (Brdu) to inhibit fibroblast growth. After pretreating with 40 μM LIQ or an equal volume of PBS for 2 hours, 1 μg/ml LPS was added to cells for incubating 12 hours.
To further confirm that LIQ protected against LPS-induced cardiomyocyte injuries via an AMPKα dependent mechanism, Compound C (CC) was used to inhibit AMPKα in NRCM. NRCMs were pre-incubated with CC (10 μM) and LIQ (40 μM) for 2 hours before LPS treatment. Cells were assigned into five groups, CON, LIQ group, LPS group, LPS + LIQ group and LPS + LIQ + CC group. CCK8 was used to examine cell viability according to the manufacturer's protocol.
Western Blotting
Heart tissues and NRCMs were homogenized with RIPA lysis buffer and quantified by BCA assay kit (Thermo, Waltham, MA, United States). 50 μg protein was loaded into 8%, 10% or 12% SDS-PAGE gels according to different molecular weights and transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, Billerica, MA, United States). After blocking with 5% non-fat milk, blots were incubated with the primary antibodies overnight at 4°C. The next day, blots were incubated with peroxidase-labeled secondary antibody (LI-COR Biosciences, 1: 10,000 dilution) at room temperature for 1 hour, and signals were detected with a chemiluminescence ECL kit (Bio-Rad, United States). The results of each blot were analyzed by Image Lab 5.2.1. All proteins were normalized to GAPDH before relative quantitative analysis.
Real-Time Reverse Transcriptase-Polymerase Chain Reaction Analysis
Heart tissue or NRCMs were lyzed in Trizol (Invitrogen, Carlsbad, CA, United States) for RNA extraction. 2 μg of total RNA was reverse transcribed into cDNA by oligo (dT) primers using a cDNA synthesis kit (Roche, Mannheim, Germany). The expression level of mRNA was quantified with SYBR green, and mRNA expression of GAPDH was used as an internal reference before relative quantitative analysis. Primers used in this study were shown in Table 1.
Terminal Deoxynucleotidyl Transferase-Mediated dUTP Nick-End-Labeling Staining
To analyze myocardial apoptosis, Terminal deoxynucleotidyl transferase-mediated dUTP nick-end-labeling (TUNEL) staining was performed with a commercial kit (Millipore, Billerica, MA, United States) according to the manufacturer's instructions, and then the images were assessed by Image Pro Plus 6.0 (Maryland, United States).
Oxidative Stress Examination
Total superoxide dismutase (SOD) Assay Kit with WST-8 (S0101) was purchased from Beyotime Co. (Shanghai, China) and the assay was performed according to the manufacturer's instructions. The activity of NADPH oxidase was detected by commercial kits (Shanghai, China).
Enzyme-Linked Immunosorbent Assay
Mouse IL-6 ELISA kits (Biolegend, 341303), Mouse TNF-α ELISA kits (Biolegend, 430906), and Mouse IL-1β ELISA kits (Biolegend, 432603) were used to examine the level of inflammatory cytokines in mouse heart. The assays were done according to the manufacturer's instructions.
Immunohistochemistry Staining
Hearts were harvested and arrested in 10% KCL, then embedded in paraffin and cut into 10 μm sections. After antigen repairing in a pressure cooker, the heart tissue sections were incubated with 3% H 2 O 2 for 10 minutes, and then TBST was used for tissue blocking for 1 hour prior to incubation with primary antibodies against CD45 (#ab10558) and CD68 (#ab125212). Tissues were incubated overnight at 4°C before application of the secondary antibody, followed by the use of DAB to visualize staining.
Statistical Analysis
All data are presented as Means ± SEM for the number (n) of experiments. A log-rank test was used to compare survival curves. Differences among three or more groups were analyzed by oneway ANOVA with Tukey post hoc test. Data was statistically significant when p < 0.05. All experiments were done randomized and blinded if it is necessary.
Liquiritin Attenuates Lipopolysaccharides-Induced Mouse Cardiac Dysfunction
Mice injected with LPS (10 mg/kg) presented with symptoms of cardiac dysfunction evidenced by increased LVEDs and decreased HR, EF and FS in LPS group compared with the CON or LIQ groups ( Figures 1A-D). LIQ administration (40 mg/kg/d and 80 mg/kg/d) attenuated LPS-induced cardiac dysfunction. However, the lower LIQ dose (20 mg/kg/d) did not significantly protect against LPS-induced cardiac dysfunction compared with the CON or LIQ groups ( Figures 1A-D). Pressure-volume loop analysis was performed to further examine mouse cardiac function. LPS injection significantly reduced dp/dt max and dp/dt min compared to CON or LIQ group (Figures 1E,F). LIQ treatment (40 mg/kg/d and 80 mg/kg/d) markedly restored dp/dt max and dp/dt min compared to LPS group ( Figures 1E,F). The lower LIQ dose (20 mg/kg/d) showed no significant effect for dp/dt max and dp/dt min compared to LPS group Taken together, these data suggested that LIQ administration ameliorated LPS-induced mouse cardiac dysfunction.
Liquiritin Suppressed Lipopolysaccharides-Induced Cardiac Inflammation In Vivo
LPS treatment caused accumulation of CD45-labeled leukocytes and CD68 positive cells in heart tissue (Figures 2A-C), which was effectively attenuated by 40 mg/kg/d or 80 mg/kg/d of LIQ treatment (Figures 2A-C). Moreover, IL-1β, IL-6 and TNFα expression was significantly up-regulated in LPS-treated mice hearts compared to the CON or LIQ group ( Figures 2D-I), and LIQ treatment significantly depressed the expression of IL-1β, IL-6 and TNFα compared to LPS injection group ( Figures 2D-I).
We also measured IκBα/NF-κB signaling pathways that play important roles in regulating inflammatory responses. The data demonstrated that phosphorylated p65/IκBα was significantly up-regulated, and T-IκBα was
Liquiritin Treatment Suppresses Lipopolysaccharides-Induced Oxidative Stress and Apoptosis In Vivo
LPS treatment significantly increased NADPH oxidase activity and decreased SOD activity in LPS-treated mice compared to the CON or LIQ groups (Supplementary Figure S1). LIQ treatment (40 mg/kg/d or 80 mg/kg/d) restored SOD activity to normal and decreased NADPH oxidase activity in LPS + LIQ mice (Supplementary Figure S1). The results show that LIQ treatment ameliorates LPS-induced oxidative stress.
LPS (30 mg/kg) treatment significantly increased mouse mortality rate compared to the CON or LIQ groups ( Figure 3A). LIQ (80 mg/kg) treatment significantly prevented LPS-induced mortality ( Figure 3A). Cardiomyocyte loss appears to play an important role in cardiac dysfunction and increased mortality following LPS administration. As shown in Figures 3B,C, TUNEL staining indicated that LPS (30 mg/kg) treatment significantly induced cell apoptosis compared to CON. BAX, the pro-apoptosis protein, was significantly increased, but BCL2, the anti-apoptosis protein, was significantly decreased by LPS treatment in LPS group compared to CON or LIQ group ( Figures 3D-F). LIQ (40 mg/kg/d or 80 mg/kg/d) treatment significantly reduced TUNEL positive cardiomyocytes, decreased BAX expression and restored Bcl2 expression ( Figures 3B-F). These results indicated that LIQ treatment effectively reduced LPS-induced cardiomyocyte apoptosis.
Liquiritin Treatment Regulates AMPKα2/ mTOR Signaling
We examined several signaling pathways including MAPK, AKT/ GSK3β, JAK/STAT and AMPKα2/mTOR pathway to elucidate Figure S2). LPS treatment significantly decreased AMPKα2 phosphorylation accompanied by decreased ACC phosphorylation but increased mTORC1 phosphorylation compared to the CON or LIQ group (Figures 4A-D). LIQ treatment restored AMPKα2 and ACC phosphorylation and inhibited mTORC1 phosphorylation in LPS + LIQ group compared to the LPS group ( Figures 4A-D). , (E) maximal rate of pressure development (dp/dt max), (F) minimal rate of pressure decay (dp/dt min), (n 10). The data were expressed as the Mean ± SEM. pp < 0.05 compared with the indicated group. The data were compared by one-way ANOVA with Tukey post hoc analysis.
Frontiers in Pharmacology | www.frontiersin.org May 2021 | Volume 12 | Article 648688 5 FIGURE 2 | LIQ treatment suppressed LPS-induced cardiac inflammation in vivo (A) Immunohistochemical staining was performed to detect CD-45 and CD-68 in mouse left ventricle (n 6), (B,C) Counting CD68 and CD45 positive cells (n 6), (D-F) RT-PCR was performed to examine mRNA expression of IL-1β, IL-6 and TNFα (n 6), (G-I) ELISA was performed to detect proteins expression of including IL-1β, IL-6 and TNFα (n 6), (J-L) Representative western blots and relative quantitative results of proteins in NF-κB signaling pathway, including phosphorylated P65 (p-P65), total P65 (T-P65), p-IκBα and T-IκBα (n 6). Mice were treated with a dose of 10 mg/kg LPS, the data are expressed as the Mean ± SEM. pp < 0.05 compared with the indicated group. The data were compared by one-way ANOVA with Tukey post hoc analysis.
Frontiers in Pharmacology | www.frontiersin.org May 2021 | Volume 12 | Article 648688 6 FIGURE 3 | LIQ treatment suppressed LPS-induced oxidative stress and apoptosis in vivo. (A) Mice were treated with LPS (30 mg/kg, intraperitoneal injection) for observing 48 h to draw survival curves (n 20), (B) TUNEL staining was performed to examine cellular apoptosis in mouse hearts (n 6), (C) Calculating the TUNEL positive nucleus (n 6), (D) Representative western blots of BAX, Bcl-2 and GAPDH, (E,F) relative quantitative results of BAX/GAPDH and Bcl-2/GAPDH (n 6). *p < 0.05 compared with the indicated group, the data were expressed as the Mean ± SEM and were compared by one-way ANOVA with Tukey post hoc analysis, A log-rank test was used to compare survival curves.
Liquiritin Treatment Restores Expression of Fatty Acid Metabolism Associated Proteins
AMPKα is a conserved sensor of cellular energy change. Phosphorylated AMPKα could maintain energy balance by decreasing anabolism processes but promoting catabolism to preserve ATP production when senses enhancement of AMP/ ATP or ADP/ATP ratios (Ke et al., 2018). LPS treatment decreased the mRNA expression of PGC1α and PGC1β ( Figures 5A,B), which are involved in regulating mitochondrial metabolism. LIQ treatment restored mRNA expression of PGC1α and PGC1β. (Figures 5A,B). LPS treatment also significantly down-regulated the protein expression of PGC1α and PGC1β compared to the CON or LIQ groups ( Figures 5C-E). LIQ treatment prevented LPSinduced down-regulation of these proteins ( Figures 5C-E).
Liquiritin Treatment Showed No Protective Effects in Lipopolysaccharides Treated AMPKα2 Knockout Mice
AMPKα2 knockout mice (AMPKα2−/−) were used to further determine whether LIQ treatment could prevent LPS-induced cardiac injury. AMPKα2−/− mice were pre-treated with LIQ (80 mg/kg/d) for 7 days and then were injected with LPS (10 mg/kg) for 12 hours. LIQ treatment neither effectively reversed EF, FS, LVEDs nor inhibit LPS-induced myocardial mRNA levels of IL-1β, IL-6, TNFα ( Figures 6E-G), nor inhibit p65 and IKBα phosphorylation and Bax expression, and nor restored Bcl2, PGC1α, and PGC1β expression in AMPKα−/ −mouse heart ( Figures 6H-N). These results suggested that the protective roles of LIQ depend upon activating AMPKα activity.
DISCUSSION
This study demonstrated that LIQ treatment could prevent LPSinduced cardiomyocyte injury in vivo and in vitro by regulating AMPKα2 and IκBα/NF-κB signaling pathway. This study suggested that LIQ may be a useful therapy or adjuvant therapy for SCM. SCM may occur in sepsis patients with a morbidity rate as high as 60% (Rudiger and Singer, 2013). SCM is characterized by left ventricular dilation with low filling pressure and significantly reduced ejection fraction (Beesley et al., 2018). Even though SCM could be reversible within 7-10 days from onset, acute myocardial depression in SCM might cost patient's life.
It has been demonstrated that cytokines, especially IL-1β, IL-6, and TNF-α, are major contributors to the initiation of SCM (Drosatos et al., 2015). In an endotoxin model, cardiac tissue is infiltrated by immune cells, such as lymphocytes, monocytes and macrophages. These immune cells could secret inflammatory cytokines to further exaggerate inflammatory response. LPS is a major component of the outer membrane of Gram-negative bacteria and is used to establish sepsis shock and to mimic SCMassociated heart injury in mice (Chen et al., 2017). Our result showed that LPS treatment significantly induced IL-1β, IL-6 and TNF-α accumulation and immune cell infiltration, which could be effectively inhibited by LIQ treatment. This was consistent with previous studies that showed that LIQ exerts its antiinflammatory effect in several disease models via inhibiting the NF-κB and NLRP3 inflammasome pathways (Min et al., 2015;Yu et al., 2015;Hongyan et al., 2016). We confirmed here that LIQ treatment could inhibit IκB/NF-κB signaling, which was in agreement with previous studies.
Oxidative stress is another very important mechanism involved in SCM pathogenesis, and increased oxidative stress correlates with SCM severity and mortality (Tsolaki et al., 2017). Endotoxin treatment may inhibit antioxidant enzyme activity and promote oxidative stress-associated FIGURE 5 | LIQ treatment improved fatty acid metabolism. RT-PCR was performed to analyze the mRNA expression of fatty acid metabolism associated genes in mouse hearts, including (A) PGC1α, (B) PGC1β, mRNA of GAPDH was used as internal reference (n 6), (C) Representative western blots images of PGC1α, PGC1β and GAPDH, relative quantitative analysis of PGC1α/GAPDH (D), PGC1β/GAPDH (E) (n 6). Data were presented as the Mean ± SEM. *p < 0.05 compared with the indicated group, the data were expressed as the Mean ± SEM and were compared by one-way ANOVA with Tukey post hoc analysis.
Frontiers in Pharmacology | www.frontiersin.org May 2021 | Volume 12 | Article 648688 10 injury and apoptosis. Matsuno showed that the NADPH subunit (NOX1) was up-regulated in SCM, which was accompanied by excessive ROS production and cardiomyocyte apoptosis (Matsuno et al., 2012). NOX1 deficiency significantly reduced LPS-induced cardiomyocyte apoptosis and improved mouse cardiac function (Matsuno FIGURE 7 | LIQ treatment could not protect against LPS-induced cardiomyocyte injuries after inhibiting AMPK activity in vitro (A) CCK-8 kit was used to examine cell viability after treating different concentration of LIQ (0, 1, 10, 20, 40, 80, 160, 320 µM) for 12 h (n 9), (B) CCK-8 was used to examine cell viability in different treatment as indicated in the picture (n 9), NRCMs were pretreated with LIQ for 2 h and then were treated with LPS (1 μg/ml) for another 12 h, RT-PCR was performed to analyze mRNA expression of IL-1β (C), IL-6 (D) and TNF-α (E), all of the mRNA expression were normalized to GAPDH before the relative quantitative analysis (n 6). (F) representative western-blots of phosphorylated AMPKα2 (p-AMPKα2), total-AMPKα2 (T-AMPKα2), p-P65, T-P65, p-IκBα, T-IκBα, BAX, Bcl2, PGC1α, PGC1β and GAPDH (n 6), relative quantitative analysis proteins expression of p-AMPKα2/T-AMPKα2 (G), p-P65/T-P65 (H), p-IκBα/T-IκBα (I), BAX/GAPDH (J), Bcl2/ GAPDH (K), PGC1α/GAPDH (L) and PGC1β/GAPDH (M) (n 6), all proteins were normalized to GAPDH before the quantitative analysis. The cell experiment was repeated 3 times independently. *p < 0.05 compared with the indicated group, the data were expressed as the Mean ± SEM and were compared by one-way ANOVA with Tukey post hoc analysis. , 2012). After the bacterial challenge, Zang also demonstrated sepsis-induced progressive oxidative damage and cytochrome C release in mitochondria accompanying with downregulation of SOD and GPx activity (Zang et al., 2007). Thus, the up-regulation of NADPH activity and downregulation of SOD activity could exaggerate cardiomyocyte apoptosis and cardiac dysfunction. Our study showed that LIQ treatment could decrease NADPH activity but increase SOD activity, which could prevent cardiomyocytes from LPSinduced injuries and apoptosis. SCM is also accompanied by decreased fatty acid oxidation (FAO), which accounts for more than 70% of the energy supply in adult cardiomyocytes (Beesley et al., 2018). A profile of key proteins regulating FAO has been down-regulated in the heart following LPS treatment. LPS or danger-associated molecular patterns (DAMPs) mediated toll-like receptor 4 (TLR4) activation causes PGC-1α and PGC-1β downregulation and mitochondrial energy metabolism disorder and the production of ROS in the heart (Schilling et al., 2011;Martin et al., 2015). Restoration of PGC1α and PGC1β expression could protect against LPS-TLR4 activation and metabolic alterations (Schilling et al., 2011;Martin et al., 2015). PGC1α has also been described as an important regulator of mitochondrial biogenesis (Ventura-Clapier et al., 2008); indicating that finetuning PGC1α expression might be an effective strategy to restore mitochondrial homeostasis, which could protect against pathological cardiac impairment (Zhu et al., 2019). Thus, promoting the expression of these down-regulated proteins might protect against LPS-induced cardiomyocyte energy depletion. Our study demonstrated that LIQ treatment could effectively reverse the down-regulation of proteins involved in cardiac metabolism.
Jung et al. reported that LIQ treatment could protect hepatocytes from serum deprivation-induced oxidative stressassociated injury and mitochondrial dysfunction in mouse liver by activating AMPK (Jung et al., 2017). AMPK exerts a distinct and critical role in regulating malonyl-CoA levels and FAO in the heart. Malonyl-CoA locates at the outer mitochondria membrane where it could inhibit the uptake of fatty acids into the mitochondria. AMPK could phosphorylate acetyl-CoA carboxylase (ACC) and could also phosphorylate and activate malonyl-CoA decarboxylase (MCD) (Hopkins et al., 2003). ACC inhibition and MCD activation could reduce the synthesis of malonyl-CoA resulted in increased fatty acid uptake (Hopkins et al., 2003). In addition, AMPK may also regulate contraction-inducible FAT/CD36 translocation in cardiomyocytes to enhance long-chain fatty acid uptake (Luiken et al., 2003). Activated AMPK may promote mitochondrial biogenesis to adapt to chronic energy deprivation by increasing PGC-1α expression (Zong et al., 2002). Activated AMPK has also been shown to take part in regulating inflammation and oxidative stress in cardiomyocytes. AMPK activation could alleviate LDH activity, ROS content and Nox1 expression resulting in reduced apoptosis and improved cell viability in CoCl2 induced hypoxic injury in cardiomyocytes (Tong et al., 2012). Wang et al. (Wang et al., 2018) showed that AMPK activation prevented cardiomyocyte necroptosis by reducing ROS production, however, these protective effects were disappeared after CC treatment or AMPK dominant-negative . In the LPS-induced SCM model, AMPK activation could inhibit IKKβ/IκBα/NFκB signaling and reduce the mRNA expression of TNF-α, IL-6 and monocyte chemoattractant protein-1 (MCP1) . Jun Ren et al. (Ren et al., 2016) demonstrated that adiponectin deficiency exacerbated LPS-induced cardiac dysfunction, intracellular Ca 2+ defect, inflammation and apoptosis, because adiponectin deficiency depressed activation of CAMKKβ-AMPK-ACC resulted in mTORC1 hyperphosphorylation and AMPK mediated autophagy inhibition. However, administration of the AMPK activator AICAR or mTORC1 inhibitor rapamycin obliterated accentuated cardiac dysfunction, apoptosis and inflammatory response in adiponectin deficiency mouse. This study demonstrated that AMPK activation could attenuate LPSinduced mouse heart injuries at least partly from inducing AMPK mediated autophagy. Although we didn't examine AMPK mediated autophagy induction, it would be a reasonable deduction that LIQ treatment attenuated LPS-induced mouse hear injuries partly depending on AMPK activation mediated autophagy induction. To sum up, these studies indicated that activated AMPK might involve in regulating multiple biological processes, including oxidative stress, cellular apoptosis, inflammation and energy metabolism.-. Our study demonstrated that LIQ could promote AMPKα2 phosphorylation resulted in attenuating LPS-induced myocardial injury and cardiac dysfunction. However, LIQ treatment could not protect against LPS-induced myocardial injury and cardiac dysfunction in AMPKα2 −/− mice. Thus, our study firstly demonstrated that LIQ exerted its protective role at least partly rely on activating AMPKα2 signaling pathways (the mechanic summary is shown in the mechanic diagram).
CONCLUSION
This study exhibited that LIQ treatment attenuated LPS-induced mouse cardiac dysfunction in vivo and in vitro via reducing inflammation, oxidative stress, and apoptosis and improving metabolism. Relying on AMPKα2 knockout mice, This study also demonstrated that LIQ exerted its protective roles depending on activating AMPKα2 phosphorylation. Moreover, LIQ presented to be ineffective in protecting against LPS-induced NRCMs injury in vitro after AMPKα activity inhibition. Although gene knockout mice are important to confirm the potential target of some drugs, it is not entirely clear that LIQ only exerts its protective role by regulating AMPKα2, since the protective effects of LIQ on other pathways may be masked by the deterioration caused by AMPKα2. Moreover, our investigation indicated that some other pathways, including MAPK, AKT/GSK3β, and JAK/STAT, seemed not to involve in LIQ's protective roles in septic cardiomyopathy.
LIQ might be a potential adjuvant drug for treating SCM and inflammation-associated cardiomyocyte injuries. However, additional studies are required to further elucidate mechanisms in improving cardiac remodeling following LIQ treatment. Specially, this study exhibited that LIQ might involve in regulating lipid metabolism. Although we discussed that LIQ-regulated lipid metabolism might be partly associated with activating AMPKα2, it will be more interesting and necessary to provide additional lipidomic data to support and strengthen this finding in future studies.
DATA AVAILABILITY STATEMENT
The raw data supporting the conclusion of this article will be made available by the authors, without undue reservation.
ETHICS STATEMENT
The animal study was reviewed and approved by the Animal Care and Use Committee of Renmin Hospital of Wuhan University.
AUTHOR CONTRIBUTIONS
Q-ZT and Z-YB contributed to the conception and provide the funding. S-QM and Z-YZ performed the experiments and wrote the draft of this manuscript; HF and NZ established the animal model and performed echocardiography analysis; ZL and XA isolated the myocardial cells of neonatal rat; W-JL, WD, and H-HL performed data analysis and revised this manuscript. | 2021-05-14T13:31:12.731Z | 2021-05-14T00:00:00.000 | {
"year": 2021,
"sha1": "1ee8375e7e7d6d3414e27dea4762a9c6d627c5de",
"oa_license": "CCBY",
"oa_url": "https://www.frontiersin.org/articles/10.3389/fphar.2021.648688/pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "1ee8375e7e7d6d3414e27dea4762a9c6d627c5de",
"s2fieldsofstudy": [
"Biology",
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
257857972 | pes2o/s2orc | v3-fos-license | Implantation of engineered adipocytes that outcompete tumors for resources suppresses cancer progression
Tumors acquire an increased ability to obtain and metabolize nutrients. Here, we engineered and implanted adipocytes to outcompete tumors for nutrients and show that they can substantially reduce cancer progression. Growing cells or xenografts from several cancers (breast, colon, pancreas, prostate) alongside engineered human adipocytes or adipose organoids significantly suppresses cancer progression and reduces hypoxia and angiogenesis. Transplanting modulated adipocyte organoids in pancreatic or breast cancer mouse models nearby or distal from the tumor significantly suppresses its growth. To further showcase therapeutic potential, we demonstrate that co-culturing tumor organoids derived from human breast cancers with engineered patient-derived adipocytes significantly reduces cancer growth. Combined, our results introduce a novel cancer therapeutic approach, termed adipose modulation transplantation (AMT), that can be utilized for a broad range of cancers.
Introduction
Tumors are complex tissues composed of cancerous and non-cancerous cells in a hypoxic and nutrient-deprived microenvironment. The tumor microenvironment (TME) contains heterogeneous cell populations, including immune cells, mesenchymal support cells, and matrix components contributing to tumor growth and progression (1). In order to survive this environment, tumors are capable of reprogramming metabolic pathways to better utilize available substrates in the surrounding TME, ultimately becoming dependent on these pathways for continued growth and survival (2). In contrast to normal cells, the main pathway of glucose metabolism in cancer cells is aerobic glycolysis, termed the Warburg effect (3). Glucose uptake and production of lactate is increased in these cells, even in the presence of oxygen and functional mitochondria (3). The increase in glycolytic flux allows glycolytic intermediates to supply subsidiary pathways to fulfill the metabolic demands of proliferating cells. During hypoxia, cancer cells also undergo metabolic reprogramming to increase lipid utilization, as fatty acids produce twice the energy of glucose (4,5).
There are many efforts to target cancer glucose and fatty acid metabolism for therapeutic purposes. For glycolysis, these include drugs that target hexokinase 2 (HK2), which is involved in the initial steps in glycolysis, using ATP from the mitochondria to phosphorylate glucose, or drugs that target glucose transporters (GLUT1 and GLUT4) (2,(6)(7)(8). Several drugs are also used to target lipid metabolism in cancer (9)(10)(11)(12). These include drugs that target lipid uptake (targeting proteins such as LXR, CD36 and FABP4/5), lipogenic enzymes (such as ACC, ACLY, FASN, and SCD1) and proteins involved in intracellular lipid homeostasis (such as CPT1A, PPARG and others) (13,14). In addition, recent work has shown that cold activation of brown adipose tissue (BAT), which dissipates energy by non-shivering thermogenesis, increases adipocyte glucose uptake, lipid metabolism and significantly inhibits tumor growth (15). However, situating cancer patients in cold conditions for extended periods is challenging.
Here, we set out to develop a novel cancer therapeutic approach, termed adipose manipulation transplantation (AMT) that utilizes two unique abilities of white adipose tissue (WAT): 1) They can be readily extracted via liposuction and implanted via reconstructive surgery; 2) they can change into a BAT-like tissue, called browning or beiging (16,17), by upregulating essential transcriptional regulators or enzymes, such as the uncoupling protein 1 (UCP1), PPARG coactivator 1 alpha (PPARGC1A) or PR/SET domain 16 (PRDM16), involved in BAT development and function (18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33). Similar to BAT, beige adipocytes have the capacity to convert energy to heat and contribute to whole-body energy expenditure (34). We show that CRISPR activation (CRISPRa) of either UCP1, PRDM16 or PPARGC1A, induces browning which subsequently increases glucose and fat metabolism in human white adipocytes and adipocyte organoids. Co-culturing of these CRISPRa-modulated adipocytes with various cancer cell lines (breast, colon, pancreatic, or prostate cancer) leads to significantly suppressed cancer cell proliferation as well as decreased glucose uptake, glycolysis and fatty acid oxidation capacity in the cancer cells. Subcutaneously co-transplanting CRISPRa-modulated human adipose organoids and cancer cells (breast, colon, pancreatic, or prostate) in immune-compromised mice leads to significantly reduced tumor size with decreased hypoxia and angiogenesis. Implantation of engineered adipose organoids in pancreatic or breast cancer genetic mouse models significantly reduces cancer progression. Furthermore, in the breast cancer model we show that their implantation both near and distal to the tumor leads to similar results. To further demonstrate the therapeutic potential of this approach, we show that adipocytes isolated from human dissected breast tissues can be similarly manipulated via CRISPRa and inhibit growth of patient-derived breast cancer organoids. Combined, our results introduce a novel cancer therapeutic approach that has the potential to treat numerous cancer types.
CRISPRa browning of human white adipocytes
To induce browning in human adipocytes we utilized CRISPRa to upregulate UCP1, PPARGC1A or PRDM16, all known genes involved in BAT development and function. Using CRISPick (35), we designed five gRNAs targeting each gene's promoter and cloned them into an adeno associated virus (AAV) based expression vector. Differentiated adipocytes derived from human white preadipocytes were co-transfected with the gRNAs along with a Staphylococcus aureus (sa) endonuclease deficient Cas9 (dCas9) fused to the VP64 transcriptional activator. We used an sa dCas9 due to its smaller size and the VP64 transcriptional activator, which carries four copies of VP16, a herpes simplex virus type 1 transcriptional activator (36), as it provides moderate gene upregulation and is small enough to fit into AAV, which has a 4.7 kilobase optimal packaging capacity (37). To generate mature human adipocytes, preadipocytes were subjected to adipocyte differentiation using a cocktail of 3-isobutyl-1-methylxanthine (IBMX), dexamethasone, and insulin before being subjected to CRISPRa. Following four days, we measured gene expression via qRT-PCR, finding several gRNAs to significantly increase the expression levels of UCP1, PPARGC1A, PRDM16 in the transfected cells compared to the control cells (treated with dCas9-VP64 only) (fig.S1A). We selected the top two gRNA for each gene for AAV serotype 9 (AAV9) packaging. We used AAV9, as it was shown to effectively infect various adipose depots (38). We infected human differentiated adipocytes with these viruses along with AAV9 dCas9-VP64, finding for each gene at least one gRNA that significantly increased expression levels compared to dCas9-VP64 only control (fig.S1B). We used the top activating gRNA AAV for all subsequent experiments. (A) qRT-PCR of UCP1, PPARGC1A, and PRDM16 in human white adipocytes transduced with CRISPRa targeting UCP1, PPARGC1A, and PRDM16. Data are represented as mean ± S.D *≤0.05. (B) qRT-PCR of TFAM, DIO2, CPT1b, and NRF1 in CRISPRa-modulated adipocytes. Data are represented as mean ± S.D *≤0.05, ***≤0.001. (C) Oxygen consumption rate (OCR) of CRISPRa-modulated adipocytes measured by the seahorse assay. Uncoupled OCR was measured under oligomycin treatment, while maximal OCR was measured under FCCP. (D) Glucose uptake of CRISPRa-modulated adipocytes with or without insulin. Data are represented as mean ± S.D *≤0.05, **≤0.01, ***≤0.001. (E) OCR of CRISPRa-modulated cells was measured by the seahorse assay in BSA-or BSA-Palmitate-medium. Data are represented as mean ± S.D *≤0.05, ***≤0.001. (F) Exogenous fatty acid oxidation of CRISPRa-modulated adipocytes was calculated by the difference of area under the curve of OCR between BSAand BSA-Palmitate-media upon FCCP treatment. Data are represented as mean ± S.D *≤0.05, **≤0.01. We next examined whether our CRISPRa treatment increases browning in these human white adipocytes. Human adipocytes infected with the top gRNA for UCP1, PRDM16 or PPARGC1A showed significant upregulation of their target genes compared to dCas9-VP64 only infected cells (Fig.1A). In addition, we also observed increased mRNA levels for brown fat marker genes, including TFAM, DIO2, CTP1b, and NRF1 upon upregulation of either of the three genes ( Fig.1B; PRDM16 CRISPRa did not show upregulation of CTP1b, and NRF1). We examined oxygen consumption rate (OCR) in these cells using Seahorse (see Methods), by initially blocking oxygen consumption and then adding 1 μ M oligomycin, followed by the introduction of 1 μ M FCCP to measure the maximal respiratory capacity. We found that CRISPRa-AAV treatment targeting either gene increased overall OCR levels in human white adipocytes with UCP1 gRNA AAV treated cells having the largest increase (Fig.1C). In addition, these cells showed increased uncoupled respiration (under oligomycin, ATPase inhibitor) indicating a brown fat-like phenotype (Fig.1C). CRISPRa treated cells also had elevated maximal respiration following carbonyl cyanide-p-trifluoromethoxy-phenylhydrazone (FCCP) treatment (Fig.1C). Furthermore, these AAV-CRISPRa engineering adipocytes showed increased glucose uptake in both basal and insulin stimulated conditions (Fig.1D). We also tested the fatty acid oxidation (FAO) capacity of these cells by performing a similar Seahorse OCR assay. In both BSA-and BSA conjugated-palmitate (saturated fatty acid complex) media, we found that our CRISPRa-treated adipocytes had an overall OCR increase in BSA conjugated-palmitate (Fig.1E). Under FCCP treatment, in which the increase of OCR in palmitate-containing media compared to BSA-containing media is thought to be due to exogenous fatty acid oxidation, we found that upregulation of either gene increased exogenous FAO capacity in human adipocytes (Fig.1F). Taken together, we show that AAV-based CRISPRa upregulation of UCP1, PPARGC1A or PRDM16 induces browning in human adipocytes, leading them to have increased glucose uptake and fatty acid oxidation.
CRISPRa-modulated adipocytes suppress cancer cell growth in vitro
We next evaluated whether our CRISPRa 'browned' adipocytes could inhibit cancer growth in vitro using a co-culturing system ( Fig.2A). We initially treated differentiated adipocytes with CRISPRa-AAV for either of the three genes (UCP1, PPARGC1A or PRDM16) and replated these adipocytes on the top chamber of a 12-or 24-well transwell plate, which has inserts of 0.4um membrane, so that the adipocytes don't contact the cells on the lower chamber ( Fig.2A). In the lower chamber, we grew five different cancer cell lines: breast cancer cells MCF7 (ER+,PR+,GR+) and MDA-MB-436 (triple negative), colon cancer (SW-1417), pancreatic cancer (Panc 10.05) and prostate cancer (DU-145). As a negative control, we used adipocytes infected with dCas9-VP64 only. After three days, we observed that all five cancer cell lines that were co-cultured with UCP1, PPARGC1A or PRDM16 CRISPRa-treated human adipocytes showed significantly lower cell numbers than cells co-cultured with dCas9-VP64 treated adipocytes (Fig.2B). We found that the number of cancer cells co-cultured with CRISPRa-treated adipocytes were 3-to 5-fold lower than cells cocultured with the control adipocytes (Fig.2C). By qRT-PCR, we observed that all cancer cells co-cultured with CRISPRa-treated adipocytes had significantly decreased levels of MKI67 (other than DU-145 treated with PRDM16 CRISPRa) a proliferation marker, compared to control cells, with UCP1 CRISPRa having the greatest effect (Fig.2D). We also performed a BrdU incorporation assay, finding that after 24 hours the CRISPRa cocultured cancer cells showed significantly reduced proliferation compared to cells cocultured with control adipocytes ( fig.S2A). We next analyzed the glucose and fatty acid metabolism of the cancer cells. To measure glycolysis, we utilized the extracellular acidification rate (ECAR) assay in which the basal glycolytic rate is measured by adding glucose and the maximal glycolytic rate is measured by oligomycin A addition. We found that all five cancer cell lines co-cultured with CRISPRa-AAV treated adipocytes showed lower overall glycolysis than those co-cultured with dCas9-VP64 only ( fig.S2B). These cells also showed a significant reduction in both basal and maximal glycolytic rate ( Fig.2E-F) and lower glucose uptake in both basal and insulin conditions ( Fig.2G-H), other than SW-1417 treated with either PPARGC1A or PRDM16 CRISPRa. Using qRT-PCR, we also found that the expression of key glycolysis genes, such as GCK and GLUT4, a major glucose transporter, were significantly lower in the cancer cells co-cultured with CRISPRa-AAV adipocytes (other than GCK for PRDM16 CRISPRa treated Panc 10.05 and DU-145 cells) compared to the negative control (Fig.2I). We also examined fatty acid oxidation (FAO) using Seahorse. Similar to glycolysis, we found that all five cancer cell lines that were co-cultured with CRISPRa-AAV adipocytes had reduced FAO compared to the negative control (Fig.2J, fig.S2C). Moreover, using qRT-PCR, we found in most CRISPRa conditions that the cancer cells have lower expression of both CPT1b, a key regulator of FAO in the mitochondria and CD36, a fatty acid transporter on the cell membrane, further confirming decreased FAO (Fig.2J). Combined, our data shows that our CRISPRa-modulated adipocytes reduce glycolysis and fatty acid metabolism in five different cancer cell lines and can significantly suppress cancer growth.
CRISPRa-modulated human adipose organoids suppress xenograft growth
To examine whether CRISPRa-modulated adipocytes could inhibit cancer growth in in vivo xenograft models, we co-transplanted four cancer cell lines (MCF7 and MDA-MB-436 (breast), Panc 10.05 (pancreas) and DU-145 (prostate)) with CRISPRa-modulated human adipose organoids. While adipocytes could be used for co-transplantation, adipose organoids offer added advantages, including: 1) providing a 3D-culture that can better recapitulate the heterogeneity of adipose tissue; 2) enhanced response to endogenous stimuli; 3) the ability to form tissue microenvironments that could better integrate with cancer cells following transplantation. We initially established culturing conditions for human adipose organoids. We utilized immortalized human preadipocytes grown using methods that combines features from three different 3D culturing protocols (39,40). In brief, we cultured human preadipocytes in basal DMEM media supplemented with 10% fetal bovine serum in Nunc 96 well plates treated with Nunclon Delta. Organoids formed after 48 hours and were then differentiated to adipose organoids using a differentiation cocktail containing IBMX, dexamethasone, insulin, T3, and rosiglitazone. Adipocytes formed 21-days post differentiation ( fig.S3A). Using qRT-PCR, we analyzed these organoids for various adipogenic markers, including FABP4, PLIN1, and ADIPOQ, finding all to be expressed ( fig.S3B). We further tested our ability to upregulate UCP1, PPARGC1A, and PRDM16 in these organoids using similar methods and AAV gRNAs as in the adipocytes, finding all three to show significant upregulation of the target genes ( fig.S3C). With the establishment of these adipose organoid culturing and CRISPRa conditions, we next set out to test whether they can suppress xenograft cancer growth. To generate xenografts, cancer cells were subcutaneously implanted into immunocompromised SCID mice. Following 6-8 weeks, UCP1-CRISPRa treated human adipose organoids were mixed with Matrigel and co-transplanted adjacent to palpable tumors. For all subsequent assays, we only used UCP1-CRISPRa, as it showed the most optimal results in our cell culture experiments. Due to some reports suggesting that brown fat could be linked to cancer-associated cachexia (loss of skeletal muscle and fat) (41), we measured on a weekly basis the body weight of mice co-transplanted with cancer cells and UCP1-modulated adipose organoids, finding no significant differences between CRISPRa treated and control mice ( fig.S4D). Tumors and human adipose organoids were collected following three weeks (Fig.3A). All tumor types cotransplanted with CRISPRa-modulated human adipose organoids were significantly smaller than dCas9-VP64 only transplanted human adipose organoids (Fig.3B), having over 50% reduction in volume (Fig.3C). Gene analysis showed that tumors cotransplanted with CRISPRa-treated adipose organoids had decreased expression of the proliferation marker gene MIK67 ( fig.S3E), as well as reduced marker gene expression for glycolysis (GLUT4, GCK) and fatty acid oxidation (CD36, CPT1B) ( fig.S3F). Using immunofluorescence, we examined additional tumor marker genes, finding that all tumors co-transplanted with UCP1-CRISPRa modulated human adipose organoids had markedly reduced Ki67+ (Fig.3D). In addition, cancer cells showed decreased levels of hypoxia, identified by having a lower carbonic anhydrase (CA9+) area per image view (Fig.3E). These tumors also showed decreased levels of CD31+ area per image view, indicating decreased microvessel density and suggesting a corresponding lower metastatic potential (Fig.3F). Since brown fat is known to reduce insulin levels, we examined the plasma insulin in these mice. We found that plasma insulin levels of mice that were co-transplanted with UCP1-CRISPRa adipose organoids and cancers were significantly lower than those co-transplanted with the control adipose organoids and were comparable to levels in wild type SCID mice ( fig.S3G). Taken together, these results show that UCP1-CRISPRa modulated human adipose organoids significantly reduce glycolysis and fatty acid oxidation, reduce hypoxia and insulin levels and inhibit tumor growth for four different cancer types in vivo.
AMT suppresses cancer development in pancreatic and breast cancer genetic mouse models
To examine whether our AMT approach can prevent cancer development, we utilized pancreatic and breast cancer genetic mouse models. For pancreatic cancer, we used the KPC mouse model, that upon tamoxifen treatment develops pancreatic ductal adenocarcinoma due to conditional mutations in Kras and Trp53 (42). For breast cancer, we used MMTV-PyMT mice on an FVB background, containing a mouse mammary tumor virus (MMTV) long terminal repeat upstream of the Polyoma Virus middle T antigen (PyVmT), that develop mammary tumors with a mean latency of 53 days (43). We first designed five gRNAs targeting mouse Ucp1 and transfected them into mouse 3T3-L1 adipocytes, finding all five upregulate Ucp1 ( fig.S4A). We next selected the top two gRNA for AAV serotype 9 (AAV9) packaging and infected 3T3-L1 differentiated adipocytes with these viruses along with AAV9 dCas9-VP64, finding that all of them significantly increased Ucp1 gene expression levels compared to dCas9-VP64 only control ( fig.S4B). We used the top activating gRNA AAV9 for all subsequent experiments. As organoids have several of the aforementioned advantages, we generated adipose organoids using mouse preadipocytes, utilizing similar techniques as described for human adipose organoids (see Methods). We then infected them with AAV9 dCas9-VP64 and Ucp1-gRNA and observed both mCherry expression from our gRNA virus and significant Ucp1 upregulation ( fig.S4C).
We first set out to test whether we can suppress pancreatic cancer development using KPC mice. KPC mice were treated with tamoxifen postnatal day 0-4. At four weeks of age, Ucp1-CRISPRa or dCas9-VP64 only (negative control) adipose organoids were orthotopically implanted next to the pancreas (Fig.4A). Following six weeks, pancreases were removed and analyzed. Over these six weeks, we observed no significant difference in body weight between mice implanted with Ucp1-CRISPRa organoids or control organoids (dCas9-VP64 only) ( fig.S4D). We found that KPC mice implanted with Ucp1-CRISPRa upregulated adipose organoids had significantly smaller tumors compared to control dCas9-VP64 implanted mice (Fig.4B) and reduced pancreatic mass (Fig.4C). They also showed reduced expression of the proliferation marker, Mki67, as well as genes involved fatty acid oxidation (Cd36, Cpt1b) (Fig.4D). In addition, we found Ucp1-CRISPRa mice to have lower expression levels of the pancreas' main glucose transporter, Glut2 (Fig.4D). Ucp1-CRISPRa mice also had a lower number of Ki67+ cells compared to control mice as determined by immunofluorescence (Fig.4E, S4E). Similarly, we observed lower CA9+ area per image, and CD31+ area per image view, suggesting reduced hypoxia and angiogenesis (Fig.4E, S4E). In addition, we found Ucp1-CRISPRa adipose organoids to have significantly reduced insulin levels compared to control mice ( fig.S4F).
To examine whether our treatment might provide systematic therapeutic effect in suppressing breast cancer growth, we next used MMTV-PyMT female mice, implanting Ucp1-CRISPRa or dCas9-VP64 adipose organoids near the third nipples (MG) of fourweek-old mice (Fig.4F). As cold treatment leads to widespread BAT activation causing cancer suppression (15), we also examined whether distal implantation of adipose organoids could suppress cancer development by implanting organoids in the back of 4week-old mice (DOR). Following six weeks post implantation, tumors were collected and analyzed. Similarly, we found no difference in body weight of mice implanted with Ucp1-CRIPSRa organoids compared to those implanted with control organoids (fig.S4G). Remarkably, we found both strategies of organoid implantation resulted in significantly reduced tumor size (Fig.4G) and volume (Fig.4H) regardless of site of implantation. Tumors also had decreased expression of the Mki67 proliferation marker and metabolic genes, including Glut4, Gck, Cd36, and Cpt1b (Fig.4I). The tumors of mice implanted with Ucp1-CRISPRa adipose organoids had less Ki67+ cells than control mice (Fig.4J). These tumors also had lower CA9+ area per image, and CD31+ area per image view, indicating reduced hypoxia and angiogenesis (Fig.4I-K). Additionally, mice implanted with Ucp1-CRISPRa adipose organoids had significantly lower plasma insulin levels than control mice ( fig.S4H). Taken together, our results indicate that our CRISPRa-modulated adipose organoids could have systematic therapeutic effects in suppressing cancer growth.
UCP1-CRISPRa upregulation of adipocytes from dissected breast tumors suppresses cancer growth
To further demonstrate the therapeutic potential of AMT, we treated adipocytes obtained from dissected human breast tissues with UCP1-CRIPSPRa AAV9 and tested their ability to suppress tumor progression by co-culturing them with breast cancer organoids generated from dissected breast tumors or metastatic pleural effusions (Fig.5A). Samples were obtained from patients who underwent breast surgery or thoracentesis. For the co-culture experiments, we used five different breast cancer organoids generated from patients with early-stage or metastatic triple-negative breast cancer [estrogen receptor (ER)-negative, progesterone receptor (PR)-negative/human epidermal growth factor receptor 2 (HER2)-negative] or hormone receptor-positive (ERpositive and/or PR-positive)/HER2-negative breast cancer (Table S1). Organoids from breast tumor tissue were generated by digesting cells for one hour in collagenase and organoids from metastatic pleural effusions were generated by isolating tumor spheroids via centrifugation, before in both cases embedding cancer cells in organoid culture using established protocols that enable long-term propagation of tumor organoids (44). In parallel, primary human adipocytes were isolated from human breast tissue using an established protocol (45). For the triple-negative breast cancer organoids derived from primary tumors, we used two different cases, including one where we generated organoids and isolated mammary gland adipose tissue from the same individual (TOR41). Adipocytes were infected with dCas9-VP64 only (negative control) or UCP1-CRISPRa AAV9. After five days, they showed strong mCherry expression, a fluorescent marker that is part of the gRNA AAV9 ( fig.S5A), suggesting that they can be readily infected by our AAVs. In addition, gene analysis showed that both mCherry and UCP1 expression levels were significantly higher in the UCP1-CRISPRa treated adipose organoids compared to the control organoids ( fig.S5B). To test whether UCP1-CRISPRa AAV9 adipocytes can suppress cancer growth, we added them to a hydrogel dome on top of the breast cancer organoids with a fullydefined organoid medium (44) and co-cultured for five days (Fig.5A). In all five cases, we observed significant reduction in cancer organoid size (Fig.5B) and in most organoids also a reduction in organoid number (TOR124 and TOR41 were not significant) upon incubation with adipocytes infected with UCP1-CRISPRa AAV9 compared to the dCas9-VP64 only treated cells (Fig.5C). All five cancer organoid cases co-cultured with UCP1-CRISPRa had significantly lower proliferation marker MKI67 expression compared to negative control (Fig.5D). We also observed significantly decreased levels of GLUT4, GCK, CD36, and CTP1b for the UCP1-CRISPRa cocultured cancer organoids (Fig.5E), suggesting they have lower glycolysis and fatty acid metabolism. Overall, our results demonstrate that human adipocytes from a dissected breast tissue can upregulate UCP1 via CRISPRa AAV9 and are able to reduce glycolysis and fatty acid metabolism and suppress breast cancer organoid growth. In addition, using matched mammary gland adipose tissue and breast cancer samples, we demonstrate the potential clinical utility of an ex vivo autologous transplantation of CRISPRa-modulated adipocytes to treat cancer.
Discussion
Cancer cells are fast-proliferating cells that require large amounts of nutrients, including glucose and fatty acids. They can reprogram metabolic pathways to utilize available substrates in the surrounding environment. Targeting their metabolism can be a potent cancer treatment. Here, we developed CRISPRa-modulated human adipocytes that have increased glucose and fatty acid utilization by upregulating UCP1, PPARGC1A, and PRDM16. These CRISPRa-modulated adipocytes were able to suppress cancer growth in five different cancer cells lines, including two breast lines, colon, pancreas, and prostate. Co-transplanting CRISPRa-treated human adipose organoids and xenografts in mice led to a significant reduction in tumor growth in all five cancer types. These tumors had lower glycolysis and fatty acid metabolism and showed decreased hypoxia and angiogenesis. Implanting CRISPRa-modulated adipose organoids into genetic mouse models of breast and pancreatic cancers also prevented tumor development in mice. Furthermore, we demonstrated that implantation of CRISPRamodulated adipose organoids distant from the cancer tissue leads to similar results. In addition, co-culturing cancer organoids from breast tumors and CRISPRa-modulated mammary adipocytes dissected from patients resulted in lower cancer growth. Combined, our data suggests that AMT has enormous potential to treat a wide variety of cancers.
Adipocytes offer a unique ex vivo therapeutic system with many of the needed procedures already established in the clinic. Liposuction and fat transplantation are commonly used in many surgical procedures, such as aesthetic and reconstructive surgery. Due to successful engraftment, adipose tissue transplantation has progressively evolved, not only in plastic and reconstructive surgery, but also for therapeutic treatments (46). Several reports using rodent models have shown that BAT transplantation has beneficial metabolic outcomes (47)(48)(49)(50). These also include the use of UCP1-CRISPRa modulation in human white preadipocytes to induce browning, followed by their transplantation in mice treated on a high-fat diet, leading to improved body weight, glucose tolerance and insulin sensitivity (34). Our work further showcases how these 'brown' adipocytes and adipose organoids can be utilized for cancer treatment.
There is a growing interest in carrying out human adipose tissue grafting by using adipose stem cells or progenitors due to their resistance to trauma and long-term survival following transplantation. One such example, is the use of CRISPRa to upregulate the relaxin family peptide receptor 1 (RXFP1) in adipose-derived stem cells and their transplantation in a diabetes mellitus induced erectile dysfunction rat model, showing amelioration of the erectile disfunction phenotype (51). Organoids could be advantageous for these cells as they provide a 3D-culture that can better recapitulate the heterogeneity of adipose tissue, respond better to endogenous stimuli and form tissue microenvironments that could more efficiently integrate with cancer cells following transplantation. Several groups have successfully grown adipose organoids from mouse adipose stem cells (33,40,52,53). Here, we were able to culture human adipose organoids from preadipocytes using various conditions from previous studies (33,40,52,53). These human adipose organoids exhibited mature adipocytes markers, including FABP4 and PLIN1. In addition, in our hands they could last up to a year after being subcutaneously implanted in mice, showing the potential for long-term treatment (data not shown). Finally, in our study we show that their CRISPRa modulation could significantly reduce tumor size, glycolysis and fatty acid metabolism and improve hypoxia and angiogenesis in cancer mouse models. Further development of these organoids and their modulation could likely improve these attributes and their therapeutic use for a wide range of diseases.
Our work builds on the recent observation that activating BAT, via cold exposure, increases adipocyte glucose uptake, lipid metabolism and significantly inhibits tumor progression (15). Here, instead of placing tumor models in cold conditions, we took advantage of CRISPRa to increase the expression of BAT key regulators, including UCP1, PPARGC1A, PRDM16 to engineer adipocytes to have increased glucose and fatty acid uptake and metabolism, and used them to deplete resources from cancer cells. Interestingly, and in line with the BAT cold activation results (15), implanting CRISPRa-modulated adipose organoids distal from the tumors also led to suppression of cancer growth, suggesting that complex surgical procedures for tumors with limited access might not be needed for this approach. This phenomenon might be due to additional mechanisms in which CRISPRa-modulated adipose organoids can suppress cancer growth. Previous studies have shown that hyperinsulinemia can lead to cancer growth due to insulin being a powerful mitogen and survival factor (54)(55)(56). For example, administration of Dapagliflozin, a SGTL2 inhibitor that lowers blood glucose, and a controlled-release mitochondrial protonophore (CRMP) suppress cancer growth in mice by reversing hyperinsulinemia (57). Since BAT is widely known to reduce whole-body blood glucose and insulin levels in humans (15,49,58), we hypothesize that the CRISPRa-modulated adipose organoids can reduce cancer progression by lowering plasma insulin levels. Indeed, our data, using both xenograft and genetic mouse models, shows that mice implanted with CRISPRa-modulated adipose organoids have reduced plasma insulin levels compared to dCas9-V64 control mice ( fig.S3G, S4F,H).
Amongst BAT-activating genes, UCP1 showed the most robust effect in terms of cancer suppression. It would be interesting to further develop this AMT approach to upregulate additional genes that could aid in cancer therapy. These could include for example upregulation of GLUT1 and GLUT4, that are the main glucose transporters in adipose cells, with GLUT4 being the most abundant and insulin responsive (59). Glucose metabolism associated genes, such as the transcription factor FOXO1 (60) and the Gcoupled receptors, GPR20 and GPR120, which were implicated in improved glucose uptake and insulin resistance (61,62) , and AIFM2, which promotes glycolysis in BAT (63). Fatty acid oxidation associated genes, including for example the fatty acid transporter CD36, a key transporter for fatty acid oxidation, CPT1b, and the fatty acid breakdown enzyme, ACC1. Additional modifications could also be engineered in these adipocytes/adipose organoids, including for example utilization of their endocrine capabilities to secrete chemotherapeutic drugs or other cancer therapeutic associated compounds.
In this study, we used AAV-based CRISPRa to upregulate genes. However, it is worth noting that both upregulation and delivery could also be carried out using other modalities. For example, gene upregulation could be carried out using zinc fingers, TALENS, generation of specific mutations via regular CRISPR editing or base or prime editing in promoters or enhancers or standard overexpression using a cDNA mammalian expression construct of the gene of interest. Delivery could also be carried out with other viruses, such as lentivirus that is widely used for CAR-T therapy, but has a major caveat of genomic integration, or various non-viral nucleic acid delivery vehicles such as nanoparticles (64) or virus-like particles (VLPs) (65). Various drugs could also be used to upregulate specific genes in adipocytes in a global manner in cancer patients. In addition, downregulation of certain genes in adipocytes or adipose organoids, using CRISPRi, siRNA, CRISPR editing or other techniques could also be utilized for AMT. For example, a recent study used CRISPR to deplete the nuclear receptor interacting protein 1 (NRIP1) to make 'brown' adipocytes which upon implantation in mice decreased their adiposity on a high-fat diet (66).
A link between obesity, excess amount of WAT and cancer development and progression has been established, with nearly 40% of all cancer deaths in the United States being attributed to obesity (67). There are numerous proposed mechanisms in how WAT is linked to cancer development and progression, including chronic inflammation, hyperinsulinemia, steroid hormones, and adipokines (68)(69)(70)(71)(72)(73)(74). In addition, in glucose-rich conditions, cancer cells synthesize de novo FAs from intermediates of the glycolysis-TCA cycle (lipogenesis) (75,76). The synthesized FAs are then utilized to synthesize triglycerides and are stored as lipid droplets in cancer cells. When energy is needed, these lipid droplets undergo lipolysis to release FAs which are subjected for boxidation (76)(77)(78). Since BAT is highly associated with improved glucose tolerance and insulin sensitivity (79), one could envision a personalized treatment development for obese cancer patients, whereby AMT is used not only to target cancer and its unique metabolism in these patients but also treat their metabolic disease. One major hurdle in our approach that needs to be taken into account is cancer-associated cachexia (41). While we did not observe weight loss in our mouse models, a longer treatment time could possibly lead to a reduction in body weight, as was shown for UCP1-CRISPRa mice on a high-fat diet (34). Modification of other genes than UCP1, removal of the implant after a certain time or having a molecular kill switch could all be potential solutions for cachexia if observed.
In summary, our results provide proof-of-principle results for a novel cancer therapeutic approach, termed AMT, that can be further improved and personalized for specific cancers and patients. Similar to chimeric antigen receptor (CAR)-T cell therapy, AMT can be readily used in the clinic as cells can be obtained from cancer patients via liposuction or other procedures, engineered and transplanted back into the same individual for therapeutic benefit. The utilization of adipocytes from dissected breast tissue, as performed in our study, further showcases the clinical utility of such an ex vivo approach. In particular, for breast cancer this could be particularly straightforward as many mastectomies are followed up by reconstructive surgery with autologous tissue (80), which could be manipulated priorto this procedure. Unlike T-cells, adipocytes have a lower immune response (34,81) which could allow more straightforward development of 'off-the-shelf' adipocytes or adipose organoids for cancer and other treatments. Their ease of growth in culture, long lasting and robustness, and lower multiplicity (82) along with existing clinical procedures to remove and transplant them make them an exceptional cell type for cancer and other cellular-based disease therapies. | 2023-04-01T13:13:35.689Z | 2023-03-29T00:00:00.000 | {
"year": 2023,
"sha1": "a3c3a563ca07a3c376a6e06ece4ab23a7e1318ef",
"oa_license": "CCBY",
"oa_url": "https://www.biorxiv.org/content/biorxiv/early/2023/03/29/2023.03.28.534564.full.pdf",
"oa_status": "GREEN",
"pdf_src": "BioRxiv",
"pdf_hash": "a3c3a563ca07a3c376a6e06ece4ab23a7e1318ef",
"s2fieldsofstudy": [
"Biology",
"Medicine"
],
"extfieldsofstudy": [
"Biology"
]
} |
258836676 | pes2o/s2orc | v3-fos-license | Characteristics and impact of interventions to support healthcare providers’ compliance with guideline recommendations for breast cancer: a systematic literature review
Background Breast cancer clinical practice guidelines (CPGs) offer evidence-based recommendations to improve quality of healthcare for patients. Suboptimal compliance with breast cancer guideline recommendations remains frequent, and has been associated with a decreased survival. The aim of this systematic review was to characterize and determine the impact of available interventions to support healthcare providers’ compliance with CPGs recommendations in breast cancer healthcare. Methods We searched for systematic reviews and primary studies in PubMed and Embase (from inception to May 2021). We included experimental and observational studies reporting on the use of interventions to support compliance with breast cancer CPGs. Eligibility assessment, data extraction and critical appraisal was conducted by one reviewer, and cross-checked by a second reviewer. Using the same approach, we synthesized the characteristics and the effects of the interventions by type of intervention (according to the EPOC taxonomy), and applied the GRADE framework to assess the certainty of evidence. Results We identified 35 primary studies reporting on 24 different interventions. Most frequently described interventions consisted in computerized decision support systems (12 studies); educational interventions (seven), audit and feedback (two), and multifaceted interventions (nine). There is low quality evidence that educational interventions targeted to healthcare professionals may improve compliance with recommendations concerning breast cancer screening, diagnosis and treatment. There is moderate quality evidence that reminder systems for healthcare professionals improve compliance with recommendations concerning breast cancer screening. There is low quality evidence that multifaceted interventions may improve compliance with recommendations concerning breast cancer screening. The effectiveness of the remaining types of interventions identified have not been evaluated with appropriate study designs for such purpose. There is very limited data on the costs of implementing these interventions. Conclusions Different types of interventions to support compliance with breast cancer CPGs recommendations are available, and most of them show positive effects. More robust trials are needed to strengthen the available evidence base concerning their efficacy. Gathering data on the costs of implementing the proposed interventions is needed to inform decisions about their widespread implementation. Trial registration CRD42018092884 (PROSPERO) Supplementary Information The online version contains supplementary material available at 10.1186/s13012-023-01267-2.
Introduction
Breast cancer is the most common cancer in women with 2.3 million new cases estimated in 2020, accounting for 11.7% of all cancers [1]. It is the fifth leading cause of cancer mortality worldwide, with 685,000 deaths [1]. Breast cancer diagnosis is more frequent in developed countries [2]. Controlling and preventing breast cancer is an important priority for health policy makers [3].
Treatment procedures have rapidly evolved over recent years. As new and precise diagnosis strategies emerged, early treatment and prognosis of breast cancer patients have shown great progresses [4]. Advances in breast cancer screening and treatment have reduced the mortality of breast cancer across the age spectrum in the past decade [5][6][7]. Although the use of research evidence can improve professional practice and patient-important outcomes, considering also the huge volume of research evidence available, its translation into daily care routines is generally poor [8,9]. It is estimated that it takes an average of 17 years for only 14% of new scientific discoveries to enter day-to-day clinical practice [10].
Clinical Practice Guidelines (CPGs) provide recommendations for delivering high quality healthcare [11,12]. However, the impact of CPGs depends not only on their quality, but also on the way and the extent to which they are used by clinicians in routine clinical practice. Large overviews show that approximately 50% of patients receive from general medical practitioners treatments which differ from recommended best practice [13][14][15][16]. In the area of breast cancer, previous systematic reviews have shown that compliance with breast cancer CPGs [17], as well as for other types of cancer [18][19][20], remains suboptimal. A recent systematic review from our research group [21] found large variations in providers´ compliance with breast cancer CPGs, with adherence rates ranging from 0 to 84.3%. Sustainable use of CPGs is also notably poor: after 1 year of their implementation, adherence decreases in approximately half of the cases [22].
Suboptimal compliance with CPGs recommendations could increase healthcare costs if healthcare resources are overused (e.g., overtreatment, overuse of diagnosis or of screening techniques); but also, if they are underused (i.e., increased costs to cover the additional health care needs that people may face with worsening conditions due to under-used resources). Available evidence suggests that outcomes may improve for patients, healthcare professionals and healthcare organizations if decisionmakers adhere to evidence-based CPGs [23,24]. This is supported by a recent meta-analysis from our group [25], which suggests that compliance with CPGs is probably associated with an increase in both, disease-free survival (hazard ratio (HR) = 0.35 (95% CI from 0.15 to 0.82)) and overall survival (HR = 0.67 (95% CI 0.59 to 0.76). Developing interventions to support clinician uptake of breast cancer CPGs is therefore essential for improving healthcare quality and patient important outcomes. Although several interventions to support compliance with breast cancer CPGs have been proposed, no previous study has systematically examined their characteristics and effects.
The aim of this systematic review is to characterize and evaluate the impact of available interventions to support healthcare providers' compliance with CPGs in breast cancer care.
Design
We conducted a systematic literature review adhering to the PRISMA reporting guidelines [26] (PRISMA 2020 Checklist available at Additional file 1). In this review, we addressed the following two questions: (1) What type of interventions have been used to support healthcare pro-fessionals´ compliance with breast cancer CPGs? and; (2) What type of interventions can effectively support healthcare professionals' compliance with breast cancer CPGs? We registered the protocol in the international prospective register of systematic reviews (PROSPERO registration number CRD42018092884).
Searches
We searched for systematic reviews and original studies in MEDLINE (through PubMed) and Embase (through Ovid) using predefined search strategies from inception to May 2021 designed and implemented by an information specialist (IS) from the Iberoamerican Cochrane Centre (IS). The search strategies (available in Additional file 2) combined MeSH terms and keywords.
Study selection
We applied the following inclusion criteria: • Population: healthcare professionals providing health services related to the prevention or management of breast cancer. All types of healthcare professionals, and from any setting were included. • Intervention: interventions explicitly aimed at supporting or promoting healthcare professionals' compliance with available breast cancer CPGs. Such guidelines may address any specific aspect of breast cancer care, including screening, diagnosis, treatment, surveillance or rehabilitation. • Comparator: any comparator, including also studies not using a comparator group. • Outcome: quality of breast cancer care (based on healthcare professionals' compliance rate with breast cancer CPGs recommendations, but also on their knowledge, attitudes or self-efficacy concerning such recommendations); intervention implementation (fidelity, reach, implementation costs), and; patient health-related outcomes (e.g., survival).
We included experimental (randomized controlled and non-randomized controlled trials), observational (before-after, cohort, case-control, cross-sectional, and case studies), and qualitative or mixed-methods studies. Due to constrained resources, we only included studies published in English. One author (of IRC, DC, APVM) screened the search results based on title and abstract. A second author (ENG, LN, ZSP, EP, DC, APVM, GPM) independently reviewed 20% of all references. Two authors independently assessed eligibility based on the full text of the relevant articles. Disagreements were discussed (involving a third author when needed) until consensus was reached.
Data extraction
One author (ENG, IRC LN, ZSP, EP, DC, APVM, GPM) extracted the following data about the characteristics and results of the included studies using an ad hoc data extraction form which had been piloted in advance: publication year, study design (e.g., randomized controlled trial), study location, setting, number of participants, aim of the study, type of breast cancer guideline (e.g., breast cancer screening), type of intervention (e.g., computerized decision support systems), and outcome(s) assessed (e.g., compliance rate). A second author (ENG, IRC LN, ZSP, EP, DC, APVM, GPM) cross-checked the extracted data for accuracy.
Quality assessment
We used the following tools to determine the risk of bias of the included studies: the Cochrane Collaboration tool for assessing risk of bias in randomized trials (RoB I) [27], the ROBINS I tool for non-randomized controlled before-after studies [28], the Quality Assessment Tool for Before-After (Pre-Post) Studies With No Control Group [29], the Newcastle-Ottawa scale for cohort studies [30], the AXIS tool for cross-sectional studies [31], and the MMAT tool [32] for mixed methods studies. The specific criteria included by each of these tools are available in Additional file 3. One author determined the risk of bias of the included studies, and a second author crosschecked the results for accuracy. Disagreements were solved with support from a senior systematic reviewer.
Data synthesis
We described the characteristics and the effects of the interventions narratively and as tabulated summaries. Findings are synthesized by type of intervention. We applied the Cochrane Effective Practice and Organization Care Review Group (EPOC) [33] taxonomy to classify our findings according to the types of interventions identified. Whereas for the characterization of the interventions we included all the publications identified meeting our eligibility criteria (irrespectively of their design); for the evaluation of the effectiveness of the interventions we focused only on those studies following a suitable design for such purpose [34]: randomized controlled trials (RCTs), controlled before-after studies, non-randomized controlled trials, and interrupted time series. Although we planned to conduct a meta-analysis on the impact of the interventions on compliance rates, this was finally not feasible due to the inconsistent and poor reporting. Instead, we provide a graphical quantitative description of the compliance rates before and after the implementation of the interventions.
Certainty of the evidence
Following the GRADE approach [35], we rated the certainty of evidence as high, moderate, low or very low, taking into consideration risk of bias, imprecision, inconsistency, indirectness, and publication bias. This was done by one researcher. and cross-checked by a second reviewer.
Search results
The eligibility process is summarized in a PRISMA flowchart (Fig. 1). We retrieved a total of 9065 unique citations from database searches, which were reviewed (through screening by title and abstract) along with 416 additional references identified from the thirteen systematic reviews also identified. We selected 145 references for full text revision, from which 35 primary studies (reporting on 24 different interventions) were finally included in our systematic review .
Computerized decision support systems
The use of computerized decision support systems to promote compliance with breast cancer CPGs was described in 12 studies [38-43, 48, 64-68]. Eleven of them reported the same intervention, which consisted of a system developed in France called OncoDoc [38,[40][41][42][43][64][65][66][67][68]. OncoDoc is a computerized clinical decision support system that provides patient-specific recommendations for breast cancer patients according to CancerEst (local) CPGs [71]. A study conducted in Italy reported on the development of a similar system, the OncoCure CDSS [48].
Educational interventions
Seven studies described educational interventions targeting healthcare providers to promote compliance with breast cancer CPGs [44,50,55,[57][58][59]63]. One intervention consisted in the provision of academic detailing on breast cancer screening (based on the American Cancer Society guidelines for the early detection of BC) among primary care physicians in an underserved community in the USA [50]. An intervention in seven hospitals in France consisted in monthly meetings where local opinion leaders presented the relevant sections of the CPGs, which were subsequently sent to all the participating physicians who were expected to use them in their practice [63]. Another intervention consisted in a comprehensive continuing medical education package to address pre-identified barriers to guideline adherence. The intervention followed a multimethod approach to physician education including CME conferences, physician newsletters, CBE skills training, BC CME monograph, "question of the month" among hospital staff meetings, primary care office visits, and patient education materials [57,58]. An educational intervention to improve compliance with radiological staging CPGs in early breast cancer patients [59] consisted of multidisciplinary educational rounds, presenting the Cancer Care Ontario Practice Guidelines [72]. Another intervention, aimed to support compliance with recommendation against serum tumor marker tests and advanced imaging for BC survivors who are asymptomatic for recurrence, consisted in academic detailing for oncologists at regular meetings [55]. Another intervention [44] consisted in an online course to learn to implement and deliver the Strength after Breast Cancer (SABC) guidelines (with recommendations about rehabilitative exercise for breast cancer survivors).
Audit and feedback interventions
We identified two audit and feedback interventions [47,69]. One consisted in sending hospitals a written report with regional benchmark information on nine performance indicators measuring the quality of care based on breast cancer national guidelines [69]. Healthcare professionals attended sessions twice a year, where an anonymous benchmark was presented for each hospital score compared with the regional mean and the norm scores. Another intervention [47] audited patients' medical records according to four agreed indicators. Information from the audit forms was entered into a database, which allowed individualized reports for each participating clinician, providing detailed feedback about their practice, with comparisons across the group and against the agreed criteria. The course also provided all the materials needed to set up the program in clinics.
Online course prepared by the researchers The online course was provided in 2015.
The course was 4-h long. and a brief telephone counseling session for the women.
Researchers
The intervention lasted approximately 9 months. The intervention design was sequential, with progressively more intensive interventions introduced at each stage.
Other types of single component interventions
Five studies described other strategies to promote compliance with breast cancer CPGs [45,49,52,56,61,70]. One intervention consisted on a microcomputer tickler system on the ordering of mammograms [45]. The system displayed the date of the last mammogram ordered in the "comments" section of the encounter form for each visit. An intervention to support compliance with CPGs follow up recommendations in low-income breast cancer survivors [70] consisted in the implementation of a medical home program to support primary care case management. Providers and networks participating in this program received a payment per eligible patient per month for care coordination. Another intervention consisted in the implementation of new clinical pathways supplemented by clinical vignettes [56]. Another intervention consisted in an integrated knowledge translation strategy to be used by guideline developers to improve the uptake of their new CPGs on breast cancer screening [61]. This integrated knowledge translation strategy was based on the Knowledge to Action Framework [73], and involved the identification of barriers to knowledge use. An intervention to support compliance with the Dutch breast cancer guideline [52] consisted of a medical critiquing system (computational method for critiquing clinical actions performed by physicians). The system aimed at providing useful feedback by finding differences between the actual actions and a set of 'ideal' actions as described by a CPG.
Multifaceted interventions
We identified nine multifaceted interventions [36,37,46,49,51,53,54,60,62]. One intervention to increase compliance with mammography screening [37] consisted of (i) audit results and a comparison with the network benchmark; (ii) academic detailing of exemplar principles and information from the medical literature; (iii) services of a practice facilitator for 9 months who helped the practitioners design their interventions and facilitate "Plan, Do, Study, Act" processes; and iv) information technology support. In another intervention [60] to increase screening mammography, primary care providers received (i) a fact sheet providing current information on screening mammography for older women; (ii) telephone follow-up of any questions, and; (iii) copies of a simply written pamphlet on mammography that they could distribute to patients. Another intervention [54] consisted of biannual feedback to primary care providers regarding compliance with cancer screening CPGs and financial bonuses for "good" performers. Feedback reports documented a site's scores on each screening measure and a total score across all measures, as well as plan-wide scores for comparison. Another intervention [51] consisted of an educational intervention accompanied by cue enhancement using mammography chart stickers, and by feedback and token rewards. Another intervention [46] included (i) use of standardized patients to observe and record healthcare professionals' performance followed by direct feedback; (ii) newsletters to inform healthcare providers about screening methods; (iii) posters and cards presenting key points about CBE and the importance of routine screening mammograms, and; (iv) patient education materials. An intervention to improve compliance with new CPGs by the American Society for Radiation Oncology (ASTRO) on the proper use of hypofractionation [49] consisted in implementing five consensus-driven and evidence-based clinical directives to guide adjuvant radiation therapy for breast cancer. Prospective contouring rounds were instituted, wherein the treating physicians presented their directive selection and patient contours for peer-review and consensus opinion. Another intervention combined audit and feedback and education to providers to increase compliance with breast cancer treatment guidelines [62]. Repeated feedback on the performance of the chemotherapy administration, timing and dosing were given to the participants. The feedback consisted of a demonstration of variation in performance between the different hospitals and the region as a whole. The educational component consisted in four consecutive sessions of discussion about relevant literature that became available in that period regarding chemotherapy dose intensity, sequencing of radiotherapy and the importance of adequate axillary lymph node clearance.
An intervention to promote compliance with new National Comprehensive Cancer Network guidelines from routine testing to omission of ordering complete blood cell count and liver function tests in patients with early breast cancer [53] involved (i) provision of educational materials; (ii) audit and feedback; (iii) certification; (iv) patient education; (v) financial incentives and (vi) implementation of alerts in the electronic medical records. Another intervention to promote breast cancer screening CPGs [36] included (i) printed educational materials with the recommendations for breast cancer mammography, (ii) printed educational materials with CPGs recommendations for clinical breast exams and breast self-exams, and (iii) video (12 min) directed at clinicians, exploring strategies for patient discussion around breast cancer screening issues.
Risk of bias
The risk of bias was judged as low in five studies [45,53,54,59,70], moderate in ten [36, 37, 42, 44, 47, 56-58, 62, 63], and high in five [46,49,50,55,65]. In four studies [51,52,60,69] the risk of bias was unclear since there was not enough information available to determine potential biases. We did not assess risk of bias for case studies, due to the lack of appropriate tools available. A detailed description of the risk of bias of the included studies, excluding case studies, is available in Additional file 3.
Impact of the interventions
Six RCTs [37,45,50,51,54,60] and four controlled before-after studies [50,57,58,63] examined the effectiveness of four provider educational interventions, one intervention based on the use of provider reminders, and five multifaceted interventions. In nine of these interventions (90%), the ultimate goal was to improve compliance with breast cancer screening guidelines. Compliance was uniformly measured in terms of mammography rates (e.g., proportion of eligible women undergoing a mammography screening for breast cancer). Except one multifaceted intervention [54], the interventions consistently showed relevant beneficial effects (Fig. 2).
One controlled before-after study showed significant (p < 0.05) improvements in providers' knowledge, attitudes and self-efficacy towards the new CPG screening recommendations [57], whereas another controlled before-after study reported a significant improvement in the number of reported mammography referrals of asymptomatic women aged 50 to 75 years in the intervention group but not in the control group [58]. A controlled before-after study observed an improved compliance to diagnostic and treatment CPG recommendations in the intervention group (from 12% before the intervention to 36% post-intervention; P < 0.001), whereas no significant improvements were observed in the control group [63].
Impact of multifaceted interventions
Five studies examined the impact of multifaceted interventions. A randomized controlled trial observed that, in comparison with usual care, a multifaceted intervention Fig. 2 Compliance rate with guideline recommendations before and after the implementation of the identified interventions (including audit and feedback; provider education; information technology support) increased the proportion of women offered a mammogram (38% vs 53%), and the proportion of women with a recorded mammogram (35% vs 52%) [37]. Another trial observed that a multifaceted intervention (comprising provider education and patient education through pamphlets), did not improve compliance with screening mammography guidelines in the overall sample, but produced significant improvements in specific vulnerable subgroups (elderly, lower educational attainment, black ethnicity and with no private insurance) [60]. A randomized controlled trial observed that a multifaceted intervention (audit and feedback plus financial incentives) doubled screening rates both in the intervention and control groups, with no statistically significant differences observed between groups [54]. A trial examining a multifaceted intervention (provider education, cue enhancement plus feedback, and token rewards) observed that mammography compliance rates significantly improved (p < 0.05) in the intervention (62.8%) in comparison with the control (49.0%) group [51]. A controlled before-after study observed that a multifaceted intervention (including audit and feedback, patient and professional education) improved the demonstration of breast cancer screening, with significantly more women older than 50 receiving mammograms in the intervention than in the comparison group [46].
Certainty of evidence
The results from the assessment of the certainty of evidence concerning the impact of the interventions on compliance with breast cancer CPGs is available in Additional file 5. Based on GRADE criteria, we rated the certainty of evidence as "low" for the four educational interventions targeting healthcare providers. This was due to very serious risk of bias, for which we downgraded the level of evidence two levels. For the only intervention identified consisting in a reminder system for healthcare providers, we rated the certainty of evidence as "moderate" (downgrading one level due to serious indirectness). For the five multifaceted interventions, we rated the evidence as "low", due to serious risk of bias, and serious inconsistency.
Main findings
In this systematic review, we identified 35 studies describing and evaluating the impact of interventions to support clinician compliance with breast cancer CPGs. We described a range of different types of interventions to support adherence of healthcare professionals to breast cancer CPGs. We observed that there is low quality evidence that educational interventions targeted at healthcare professionals may improve compliance with recommendations concerning breast cancer screening, diagnosis and treatment. There is moderate quality of evidence that reminder systems for healthcare professionals improve compliance with recommendations concerning breast cancer screening. There is low quality of evidence that multifaceted interventions may improve compliance with recommendations concerning breast cancer screening. The effectiveness of the remaining types of interventions identified is uncertain, given the study designs available (e.g., cross-sectional, uncontrolled before-after or case studies). There is very limited data on the costs of implementing these interventions.
Strengths and limitations
The main strength of this systematic review is that it addressed a highly relevant question, and provided much needed evidence to help improve providers' compliance with breast cancer guidelines globally. An additional strength is that, contrary to previous systematic reviews, ours was not limited to experimental studies. By including observational, and qualitative and mixed-methods studies, we were able to provide a richer characterization of the available interventions.
This review has several limitations. First, we restricted the bibliographic searches to peer-reviewed publications in English language only. This may have resulted in failing to identify additional relevant data that could have further informed our assessments of the available evidence. However, we think that the impact of this limitation is likely to be small, as suggested by a recent meta-epidemiologic study [74]. Second, the heterogeneity of the reporting of outcome data made meta-analysis not feasible. Third, the heterogeneity in outcomes and the large number of strategies used across studies precluded us to determine the unique influence of each strategy on a given outcome.
Our results in the context of previous research
An important finding of our review is that most of the included studies showed that the interventions were effective in improving compliance to CPGs. This is in line with findings from previous, non-condition-specific reviews, which concluded that guideline dissemination and implementation strategies are likely to be efficient [75,76].
A large proportion of the studies included in our review examined the impact of Computerized Decision Support Systems (CDSS). Previous systematic reviews observed that CDSS significantly improve clinical practice [77,78]. In our review, the evidence about CDSS was only available from observational, uncontrolled studies, and was restricted to two tools in France and Italy in the hospital setting. New studies evaluating other CDSS, and in other settings and countries, are therefore needed.
There is substantial evidence from non-condition specific research that audit and feedback interventions can effectively improve quality of care [79]. A recent systematic review [80] examining the effectiveness of cancer (all types) guideline implementation strategies showed that providing feedback on CPG compliance was associated with positive significant changes in patient outcomes. More research is needed about the impact of audit and feedback interventions on the compliance with breast cancer CPGs.
Educational interventions targeted to providers (both in isolation and in combination with other interventions) have shown to improve outcomes in patients with cancer [80]. Despite the low certainty obtained, the studies in our review consistently showed that educational and multifaceted interventions improve compliance with breast cancer CPGs, supporting also results from previous non-condition specific reviews [16,81], as well as current recommendations from the Institute of Medicine [82].
In line with our finding concerning electronic reminder interventions, a Cochrane systematic review concluded that computer-generated reminders to healthcare professionals probably improves compliance with preventive guidelines [83].
Implications for practice and research
In terms of implications for practice, given that compliance with breast cancer guidelines is associated with better survival outcomes [25], and that there are still a substantial proportion of breast cancer patients not receiving clinical guidelines recommended care [21], it is important that the most effective interventions available are implemented to improve breast cancer guideline uptake by healthcare providers.
In terms of implications for research, as in a previous non-condition-specific review [76], we observed that there is very limited data on the costs of implementing the interventions to support compliance with breast cancer CPGs, as well as a scarcity of studies evaluating the effectiveness of interventions targeting the organization of care (e.g., benchmarking tools). Research in these two areas is urgently needed to allow evidence-based decisions concerning which interventions should be rolled out and implemented widely as part of existing quality improvement programs. Also worth noting is that, up to now, the great majority of the research on this (breast cancer) area has focused on measuring the impact of the interventions on process measures (mostly compliance rates). No study measured the impact on patient outcomes, and only a small minority examined the impact on determinants of compliance behavior (e.g., providers' knowledge, attitudes, or self-efficacy). Future research would benefit from including a broader range of outcomes (including proximal and distal), as this would help to better measure and understand the extent to which the interventions produce the intended benefits.
Future research is also needed to identify the most effective types of interventions in improving CPGs uptake, as well as the "active ingredients" of multifaceted interventions [84]. The characteristics of the CPGs intended users, and the context in which the clinical practice occurs are likely to be as important as guideline attributes for promoting adoption of CPG recommendations. Therefore, future research should focus on gaining a deeper understanding about how, when, for whom, and under which circumstances the interventions identified can effectively support guideline adherence. Using a realist evaluation methodology [85] may prove a valuable strategy in this endeavor. However, as observed in our review, the detailed characteristics of the interventions are very frequently scarcely reported. To allow progress in this area, it is of utmost importance that intervention developers and researchers offer in their published reports a comprehensive characterization of their interventions. The Template for Intervention Description and Replication (TIDieR) guidelines [86] were specifically designed for this purpose.
Conclusion
Promoting the uptake and use of CPGs at the point of care, represents a final translation step, from scientific findings into practice. In this review we identified a wide range of interventions to support adherence of healthcare professionals to breast cancer CPGs. Most of them are based on computerized decision support systems, provision of education, and audit and feedback, which are delivered either in isolation or in combination with other co-interventions. The certainty of evidence is low for educational interventions. The evidence is moderate for automatic reminder systems, and low for multifaceted interventions. For the rest of the interventions identified, the evidence is uncertain. Future research is very much needed to strengthen the available evidence base, concerning not only their impact on compliance, but also on patient important outcomes, and on their cost-effectiveness. | 2023-05-23T14:58:49.941Z | 2023-05-22T00:00:00.000 | {
"year": 2023,
"sha1": "532e7b65d82b6284d45892dcdff32c7317e41f8e",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Springer",
"pdf_hash": "532e7b65d82b6284d45892dcdff32c7317e41f8e",
"s2fieldsofstudy": [
"Medicine",
"Political Science"
],
"extfieldsofstudy": [
"Medicine"
]
} |
215780223 | pes2o/s2orc | v3-fos-license | Portable microfluidic chip for detection of Escherichia coli in produce and blood
Pathogenic agents can lead to severe clinical outcomes such as food poisoning, infection of open wounds, particularly in burn injuries and sepsis. Rapid detection of these pathogens can monitor these infections in a timely manner improving clinical outcomes. Conventional bacterial detection methods, such as agar plate culture or polymerase chain reaction, are time-consuming and dependent on complex and expensive instruments, which are not suitable for point-of-care (POC) settings. Therefore, there is an unmet need to develop a simple, rapid method for detection of pathogens such as Escherichia coli. Here, we present an immunobased microchip technology that can rapidly detect and quantify bacterial presence in various sources including physiologically relevant buffer solution (phosphate buffered saline [PBS]), blood, milk, and spinach. The microchip showed reliable capture of E. coli in PBS with an efficiency of 71.8% ± 5% at concentrations ranging from 50 to 4,000 CFUs/mL via lipopolysaccharide binding protein. The limits of detection of the microchip for PBS, blood, milk, and spinach samples were 50, 50, 50, and 500 CFUs/mL, respectively. The presented technology can be broadly applied to other pathogens at the POC, enabling various applications including surveillance of food supply and monitoring of bacteriology in patients with burn wounds.
Introduction
Sepsis is a significant, life-threatening problem for military personnel and civilians alike. Clinicians typically make an initial diagnosis of infection with an intermittent examination observing changes in temperature, blood pressure, smell, and sight. Depending on the severity of the inflicted wound and infectious agent, this may cause multiple organ dysfunction, hypoperfusion, failure of body systems (including the kidneys, liver, lungs, and central nervous system), and ultimately death. For sepsis diagnosis, daily culture and/or molecular analysis is performed using a small volume of the patient's blood; usually less than 0.1% of entire blood volume (ie, approximately 5 mL). These detection methods are time-consuming and unreliable for identification of pathogens in up to 50% of septic patients. Following pathogen detection, the patient is treated with broad-spectrum antibiotics; this is not an ideal solution, because frequent use of broad-spectrum antibiotics also stimulates development of antibiotic-resistant bacteria. Antibiotic-resistant bacteria are a significant problem for injured military personnel and military treatment facilities. Currently, the sepsis mortality rate is 30%-50% among civilians and is higher in patients with military conditions including burns, trauma, and surgery. 1 Although sepsis biomarkers, inflammatory modulators, and new antibiotics have been identified in past sepsis research, improved identification of various targets has had minimal impact on sepsis related morbidity and mortality. Therefore, there is a significant clinical need for new detection and identification technologies in this area. According to the Centers for Disease Control and Prevention, food-borne diseases cause approximately 76 million illnesses, 325,000 hospitalizations, and 5,000 deaths in US alone each year. 2 The US Department of Agriculture showed that medical costs, productivity losses, and costs of premature deaths caused by food-borne diseases are approximately $6.9 billion per year. 3 In 2011, pathogen-based produce contamination triggered a global concern with the outbreak of the foodborne toxin 'Shiga', which is produced by Escherichia coli (E. coli). 4 On the other hand, sepsis is the tenth leading cause of death in the US, 5 amounting to 24,179 cases in 49 US hospitals over a period of 7 years. 6 As reported, E. coli can contaminate food source 7 and cause sepsis in burn patients. 8,9 Thus, effective E. coli detection would have a positive impact.
Currently, the gold standard detection method for bacteria is agar plate culture. However, this method is limited by the culturing time and volume of sample required to determine the presence of pathogens ( Figure 1). Due to the challenge of obtaining enough sample volume, agar plate cultures give false negative results at rates ranging from 7.2% to 21.2%. 10,11 In addition, the process is complicated by the fact that clinical samples need to go through multiple post-cultural steps for analysis, including Giemsa staining and differentiation on MacConkey plates. 12 The whole process takes 48 to 72 hours. 13 Although polymerase chain reaction (PCR) has high sensitivity and specificity, 14 the need for a thermal cycler makes it unsuitable for point-of-care (POC) testing. 15 Therein lies the niche for which microfluidic technologies are ideal; they have been employed to develop POC testing devices because of low manufacturing cost, reduced consumption of samples and reagents, and shortened assay time. [15][16][17][18][19][20] However, existing microfluidic devices for bacterial detection, either based on PCR 21 or enzyme-linked immunosorbent assay (ELISA), require multiple sample processing steps prior to detection. 22
Device fabrication
The microfluidic device was fabricated as previously reported. 18,24,25 The device was designed with dimensions of 22 mm × 60 mm with three parallel microchannels. To assemble this device, poly(methyl methacrylate) (PMMA) (1.5 mm thick; McMaster Carr, Atlanta, GA) and doublesided adhesive film (DSA) (50 µm thick; iTapestore, Scotch Plains, NJ) were cut using a laser cutter (Versa Laser™, Scottsdale, AZ). The PMMA base and a glass cover slip were then assembled via the DSA. In the assembled E. coli detection device, three microchannels (with dimensions of 50 mm × 4 mm × 50 µm in the DSA layer) were formed with an inlet and outlet (0.565 mm in diameter) at each end of the channels in the DSA layer. Before assembling the chip, glass cover was cleaned with ethanol using sonication. Then, it was washed with distilled water and dried under nitrogen gas. After cleaning steps, the glass cover was plasma treated for 60 seconds. Then, PMMA, DSA, and glass cover were assembled to form the complete microchip ( Figure 2A
Surface chemistry
Two different surface chemistry methods were evaluated to immobilize antibodies for the highest capture efficiency of E. coli on chip ( Figure 2B). Protein G is an immunoglobulinbinding protein and has the ability to immobilize the fragment crystallization (Fc) region of antibodies. NeutrAvidin has strong affinity for biotin-conjugated biomolecules (eg, biotinylated antibodies). In the first method, a biotinylated antibody was immobilized via NeutrAvidin-based surface chemistry. 18 Briefly, a glass slide was first plasma treated to make the surface more hydrophilic. 200 mM of 3-mercaptopropyl-trimethoxysilane (100 µL) dissolved in ethanol was then injected through the channels and incubated for 30 minutes at room temperature. 2 mM of N-(gammamaleimidobutyryloxy) succinimide ester (GMBS; 100 µL) in ethanol, a cross linker of proteins to antibodies, was then incubated in microchannels for 35 minutes at ambient temperature. An ethanol wash of 300 µL was performed to remove the excess of untreated reagents after each incubation step. Before NeutrAvidin binding on the microchannels, 300 µL of PBS was used to wash out ethanol from the chip surface. 30 µL of NeutrAvidin solution (1 mg/mL in PBS) was incubated for 1 hour at room temperature before capture antibody immobilization. Three different experimental designs were performed for the NeutrAvidin studies. In the first design, 30 µL of biotinylated lipopolysaccharide (LPS) binding protein (LBP) Ab solution (5 µg/mL), (goat immunoglobulin G antibody, R&D Systems, Minneapolis, MN) was subsequently immobilized on the microchannel surface. Then, 30 µL of carrier-free recombinant human LBP (R&D Systems) at a concentration of 10 µg/mL was coated on the top of biotinylated anti-LBP antibody (anti-LBP-LBP). In the second design, the same process as performed in the first design was followed by the addition of 30 µL of bovine serum albumin (BSA) to block any remaining reactive spaces on the surface (anti-LBP-LBP-BSA). In the third design, first LBP was directly bound to GMBS to test the effect of protein orientation on E. coli capture, and then the BSA blocking step was performed after LBP binding (LBP-BSA). LBP was incubated in microchannels for 1 hour at room temperature.
In the second surface chemistry method, three nonbiotinylated antibodies (ie, anti-LPS antibody, cluster of differentiation 14 protein (CD14) of human monocyte and antiflagellin antibody) were immobilized onto microchannels via Protein G. 26 In this method, processing steps were same as the first method except for the use of Protein G instead of NeutrAvidin. The Protein G concentration used in the immobilization of alternate antibodies onto the surface was 1 mg/mL. Then 30 µL of 5 µg/mL anti-LPS antibody (Abcam, Cambridge, MA) was immobilized on the surface via Protein G. 30 µL of CD14 (Abcam, Cambridge, MA) was used at 5 µg/mL to capture E. coli by binding to LPS on E. coli surface. Movement of LPS monomers to a binding site on CD14 was catalyzed by LBP. 27,28 Finally, 30 µL of antiflagellin antibody (BioLegend, San Diego, CA) was used at a concentration of 5 µg/mL to capture E. coli via the surface protein flagellin (part of flagellum). 29 These three capture agents were simultaneously incubated in microchannels for 1 hour at room temperature.
Sample preparation
EmGFP-expressing E. coli was spiked into 1 × PBS, blood, milk, and spinach samples with the final concentrations ranging from 50 to 4,000 colony forming units (CFUs) per milliliter for analysis on chip. Discarded de-identified whole blood (purchased from Research Blood Components, LLC, Cambridge, MA) was used in this study. Whole blood was spiked with E. coli and then inverted gently for 1 hour to enable homogenous distribution of E. coli. Spinach and milk samples were obtained from a local grocery store. Spinach leaves were washed and then spiked with E. coli before thorough mixing in a blender with 100 mL of sterile deionized H 2 O. The resultant mixture containing E. coli was submit your manuscript | www.dovepress.com Dovepress Dovepress filtered through a cell strainer (70 µM) to remove the residues of spinach leaves prior to microchip testing. Milk was spiked with E. coli and vortexed for 1 minute before testing.
Device operation and on-chip detection
To optimize the capture efficiency, 75 µL of EmGFPexpressing E. coli sample was pipetted into the functionalized microchannels, and then incubated at ambient temperature for 30 minutes. Following the incubation, microchannels were washed with 300 µL of PBS at a flow rate of 2 µL/minute using a syringe pump (Harvard Apparatus, Holliston, MA) for 150 minutes. For further optimization of processing time 2, 5, and 10 µL/minute wash was performed. After washing, captured E. coli was imaged using an inverted fluorescence microscope (Carl Zeiss, Thornwood, NY) through a GFP fluorescence filter (excitation wavelength 470 nm). For comparison, bright field and fluorescence images of E. coli taken under 10× and 100× magnification were merged ( Figure 2C). The number of E. coli detected under 10× magnification using a GFP filter was counted manually.
Statistical analysis
To evaluate the effect of the capture agent on capture efficiency, we analyzed the experimental results (n = 2-6) using analysis of variance with Tukey's post-hoc test for multiple comparisons with statistical significance threshold set at 0.05 (P , 0.05). To evaluate the effect of flow rate on capture efficiency, we analyzed the experimental results (n = 3-8) using nonparametric Kruskal-Wallis test followed by pair-wise comparisons with nonparametric upper-tailed Mann-Whitney U test, with statistical significance threshold set at 0.05 (P , 0.05). When nonparametric pair-wise tests were performed, Bonferroni correction was used for multiple comparisons. The agreement between chip counts and the E. coli stock concentrations in different media (PBS, blood, milk, spinach) was evaluated by calculating Pearson product-moment correlation coefficient (r) with statistical significance threshold set at 0.05 (P , 0.05). The Bland-Altman comparison analysis was used to evaluate the repeatability of the chip counts using residual analysis in comparison to E. coli stock concentrations. The coefficient of repeatability was calculated as 1.96 times the standard deviations of the differences between chip measurements and E. coli stock concentrations. In the Bland-Altman analysis, a mean difference of zero indicates that the chip measurement is unbiased with respect to the standard. A clinically acceptable range indicates the interval within which the difference would fall approximately 95% of the time. If the mean difference and the limit of agreement are within the clinically acceptable range, then the developed measurement method is deemed comparable to the standard. Statistical analyses were performed with Minitab (Release 14; Minitab Inc, State College, PA).
Results and discussion
We engineered the surface chemistry using immobilized antibodies, where the performance of the chip relies on nanoscale reactions on the microchannel surface. We evaluated antibodies specific to E. coli surface proteins using commonly reported antibody immobilization methods that provide a uniform distribution of antibodies on the capture surface in microfluidic channels, ie, Protein G and NeutrAvidin based methods. To develop a rapid microchip method for E. coli detection, four different capturing agents were immobilized in microchannels via these two surface chemistry methods. As shown in Figure 2B, anti-LPS, antiflagellin, and CD14 were immobilized on the microchannel surface via Protein G and anti-LBP antibody was immobilized on the microchannel surface via NeutrAvidin. To investigate the binding of anti-LBP antibody, and to observe the effect of antibody orientation on E. coli capture. Three different experimental designs were performed. In NeutrAvidin experiments, capture efficiencies of anti-LBP-LBP, anti-LBP-LBP-BSA, and LBP-BSA were observed to be 71.8% ± 5%, 60.7% ± 2%, and 44.5% ± 5%, respectively ( Figure 3A). In comparison, E. coli capture efficiencies of antiflagellin, anti-LPS and CD14 obtained by using Protein G based surface chemistry, were 46.9% ± 3%, 41.5% ± 5%, and 41.0% ± 2%, respectively ( Figure 3A). The capture efficiency via anti-LBP-LBP was observed to be significantly greater (P , 0.05) than the other capture agents that were used with Protein G based surface chemistry ( Figure 3A). In our prior study, Protein G and NeutrAvidin exhibited similar efficiency to immobilize capture agents on microchannel surfaces. 26 Thus, the difference in capture efficiency of E. coli was mainly due to the affinity of capture agents, ie, anti-LPS, antiflagellin, CD14, and LBP. The highest capture efficiency (71.8% ± 5%) was observed in microchannels immobilized with anti-LBP in the presence of LBP. This observation is in accordance with a previous report, in which LBP was shown to bind E. coli with high affinity; anti-LBP antibody helped the protein to take favorable orientation for E. coli capture. 27 BSA blocking was used to prevent nonspecific binding in microchannels in addition to preventing the binding of anti-LBP antibodies and LBP onto the succinimide group submit your manuscript | www.dovepress.com Dovepress Dovepress of GMBS. The use of BSA as a blocking agent did not result in a statistically significant difference in capture efficiency with both anti-LBP-LBP and anti-LBP-LBP-BSA. The LBP-BSA experiment showed that the orientation of the LBP protein is critical to capture of E. coli, which was supported by statistical analysis of experimental results as shown in Figure 3A. As an overall result, Neutr Avidin-mediated anti-LBP antibody-LBP immobilization performed on the microchannel surface to attain high capture efficiency of E. coli.
Statistical analysis on experimental results indicated that flow rate had a significant effect on capture efficiency (nonparametric Kruskal-Wallis test), where 2 µL/min resulted in significantly greater (P , 0.05) capture efficiency compared to 10 µL/min. The lower efficiency observed at higher flow rates may be related to the correspondingly higher shear stress within the microchannels. Additionally, we used food dyes to visualize and qualitatively analyze the wash steps in microchannels. We observed that the selected flow rate (2 µL/min) achieved effective removal of food dye solution from microchannels ( Figure 4).
To determine the microchip's limit of detection for E. coli capture, we used LB agar plate culture as the gold standard for E. coli detection. We correlated agar plate results for a Figure 5). The linearity of the correlation indicates that microchip-based E. coli detection can be used as an alternative to agar plate culture. Figure 5A shows a sensitivity of 50 CFUs/mL that was obtained by the microchip count in PBS. Various sample types including blood, milk, and spinach, were used for rapid on-chip detection of E. coli ( Figure 5C, E, and G). The detection limits of E. coli in blood, milk, and spinach samples were observed to be 50, 50, and 500 CFUs/mL as compared to E. coli dilution count, respectively. According to Bland-Altman analysis, the mean biases were -70 CFUs/ mL, -165 CFUs/mL, -163 CFUs/mL, and -1869 CFUs/ mL for PBS (50 to 500 CFUs/mL), blood (50 to 500 CFUs/ mL), milk (50 to 500 CFUs/mL), and spinach (500 to 4,000 CFUs/mL) samples, respectively ( Figure 5B, D, F, and H). Compared to E. coli spiked in PBS, the capture efficiency and limit of detection slightly decreased in other samples due to the immunological and enzymatic components in blood, milk, and spinach. For instance, there is a significant amount of albumin present in blood; this is commonly used as a blocking agent in immunoassays, and as such may have had a similar effect on our experiment. Milk also tends to have excessive albumin, casein, and other proteins, which may cause further blocking of the antigen-antibody binding. As the data show, these blocking and enzymatic agents can affect the capture efficiency. E. coli detection in spinach samples showed higher limit-of-detection than milk and blood experiments, potentially due to this experiment requiring preprocessing steps to remove the residuals of spinach leaves prior to microchip testing. Preprocessing steps (eg, filtering) caused a loss in E. coli cells before detection. On the other hand, the presented microchip assay achieved a lower detection limit for similar assays than was previously reported by methods such as magnetic separation method, fluorescence staining, and electrical detection (10 4 CFUs/mL and 8 × 10 4 CFUs/mL, respectively). 30,31 Since nucleic acid-based-detection methods require more extensive sample preprocessing, antibody based E. coli detection on a chip presents a simpler alternative method to nucleic acid-based amplification assays on a chip. 32,33 The current culture-based systems are subject to giving false negatives since only a small sample can be cultured compared to the whole sample volume, eg, 0.1% of the total blood volume is cultured in sepsis patients. This leads to 7.2 to 21.2% of such cultures providing false negatives. 10,11 With the microchip system, larger sample sizes can be processed by continuous flow, giving results much faster than a culture system. The sample can also be reprocessed within either the same or new microchannels; this would be expected to increase the number of pathogens captured from a single sample. However, for detection purposes, this would be not necessary granted that accurate detection results are provided by the chip. Capturing larger numbers of pathogens can be beneficial for other applications such as culturing and downstream genomic analysis.
In this study, we used E. coli as a model pathogen because E. coli have been extensively characterized, and as such it provides simple comparison to existing systems. We transformed BL21 Star™ with EmGFPexpressing plasmid and developed a microfluidic device to rapidly capture and quantify E. coli from various biological samples. The GFP-expressing E. coli was used to facilitate the quantification steps. The bacteria can be visualized either under a microscope in a laboratory setting, or for rapid counts in POC settings, with wide field of view lensless imaging systems for rapid counts for POC settings. 20,24,25,34 Also, for characterization of E. coli capture from various biological samples, we employed a GFPexpressing strain for ease of detection under a fluorescence microscope. The extension from a GFP-expressing strain to wild-type strains can be simply achieved using an ELISA-based detection method, as previously reported. 35 The characterized microchip can be coupled with POC detection technologies such as surface plasmon resonance, 36-38 on-chip ELISA, 39 and lensless imaging 20,24 to achieve POC testing without reference to laboratory equipment. Additionally, the presented microfluidic approach can be adapted to detect other microorganisms that cause sepsis such as Gram-negative bacteria (E. coli, Klebsiella, Pseudomonas, Enterobacter, and Hemophilus species) and Gram-positive bacteria (Staphylococcus, Streptococcus, and Enterococcus species). 40 Also, the current material cost, without the labor and other costs of the microchip, includes 10¢ of glass, 1¢ of PMMA, and 70¢ of antibodies. The antibody cost could be lowered with large-scale production and ordering.
In conclusion, we presented a microfluidic immunoassay to capture E. coli from blood, milk and spinach samples. This microchip enables rapid detection of bacteria in blood samples and food supply. The technology can be potentially adapted for on-site real-time food quality monitoring and for diagnosis of sepsis at the POC. | 2016-03-14T22:51:50.573Z | 2012-05-29T00:00:00.000 | {
"year": 2012,
"sha1": "94174afaf72edb627d56faa3b3a6bb51fd62757b",
"oa_license": "CCBYNC",
"oa_url": "https://www.dovepress.com/getfile.php?fileID=12840",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "16fab75593bed6532a5febf81c29e6b122859ff8",
"s2fieldsofstudy": [
"Medicine",
"Engineering"
],
"extfieldsofstudy": [
"Medicine",
"Biology"
]
} |
268371221 | pes2o/s2orc | v3-fos-license | Grape cultivars adapted to hotter, drier growing regions exhibit greater photosynthesis in hot conditions despite less drought-resistant leaves
Abstract Background and Aims Many agricultural areas are expected to face hotter, drier conditions from climate change. Understanding the mechanisms that crops use to mitigate these stresses can guide breeding for more tolerant plant material. We tested relationships between traits, physiological function in hot conditions and historical climate associations to evaluate these mechanisms for winegrapes. We expected a more negative leaf osmotic potential at full hydration (πo), which reduces leaf turgor loss during drought, and either a metabolically cheaper or more osmoprotectant leaf chemical composition, to allow cultivars associated with hot, dry regions to maintain greater gas exchange in hot growing conditions. Methods We measured πo, gas exchange and leaf chemistry for seven commercially important winegrape cultivars that vary widely in historical climate associations. Vines were grown in common-garden field conditions in a hot wine-growing region (Davis, CA, USA) and measured over the hottest period of the growing season (July–September). Key Results The value of πo varied significantly between cultivars, and all cultivars significantly reduced πo (osmotically adjusted) over the study period, although osmotic adjustment did not vary across cultivars. The value of πo was correlated with gas exchange and climate associations, but in the direction opposite to expected. Photosynthesis and πo were higher in the cultivars associated with hotter, less humid regions. Leaf chemical composition varied between cultivars but was not related to climate associations. Conclusions These findings suggest that maintenance of leaf turgor is not a primary limitation on grapevine adaptation to hot or atmospherically dry growing conditions. Thus, selecting for a more negative πo or greater osmotic adjustment is not a promising strategy to develop more climate-resilient grape varieties, contrary to findings for other crops. Future work is needed to identify the mechanisms increasing photosynthesis in the cultivars associated with hot, dry regions.
INTRODUCTION
Climate change is projected to exacerbate heat and drought stress in many agricultural regions worldwide, with detrimental impacts on crop yield and quality (Lobell et al., 2006;DaMatta et al. 2010;Hasegawa et al., 2022).Breeding or genetic engineering of more stress-tolerant cultivars is a promising strategy to mitigate impacts from climate change, but these efforts have been limited by uncertainty around the traits that confer stress tolerance (Vivin et al., 2017;Paleari et al., 2022).Evaluation of trait and climate associations across existing cultivars that are adapted to a diverse range of climatic conditions can identify the traits that have been important for adaptation to hot and dry conditions (Cortés and López-Hernández, 2021).
Two leaf water relationship traits, namely osmotic potential at full hydration (π o ) and osmotic adjustment (Δπ o ), are considered strong predictors of drought performance across cultivars of other crops and wild plant species (Baltzer et al., 2008;Bartlett et al., 2012, 2014, 2016, Blum, 2017) but have not been tested as predictors for stress tolerance in grape cultivars.Both π o and Δπ o impact drought tolerance by affecting leaf vulnerability to damage from dehydration.Adaptations to reduce damage from dehydration are crucial to maintain gas exchange and carbon assimilation in hot and dry conditions.Much of this damage is caused by the cells losing turgor (i.e. the pressure exerted by water pushing out against the cell walls) as they dehydrate.Turgor supports the cell walls and drives cell expansion (Hsiao et al., 1976;Morgan, 1984).Loss of turgor impairs growth and causes the cell walls to collapse and deform, which impedes water and CO 2 transport and causes leaves to wilt (Jones and Turner 1980;Scoffoni et al., 2018).The ability to maintain turgor during dehydration is strongly determined by π o , which is a measure of the potential energy for water influx generated by the cell solutes (Hsiao et al., 1976).Cells with a higher solute concentration exert a stronger driving force for water influx, reducing dehydration and turgor loss.Thus, species or cultivars with higher leaf cell solute concentrations, measured as more negative leaf osmotic potentials at full hydration, typically undergo disruptions in leaf water transport, stomatal closure and wilting under more severe water stress (Baltzer et al., 2008;Bartlett et al., 2016;Scoffoni et al., 2018).Water-stressed plants, including grapevines, can also make leaf osmotic potentials more negative (i.e.osmotically adjust) by accumulating solutes in the leaf cells, which helps to maintain turgor and to reduce leaf vulnerability to wilting, hydraulic dysfunction and stomatal closure (Martorell et al., 2015;Rodriguez-Dominguez et al., 2016;Sorek et al., 2021).Leaf osmotic potentials are typically more negative in plant species adapted to hotter, drier environments, and crop cultivars with greater osmotic adjustment (i.e.larger declines in π o under water stress) typically maintain higher yields under drought (Bartlett et al., 2012, Blum, 2017).
Despite the importance of osmotic potential to drought tolerance in other plants, it is largely unknown how osmotic potential and adjustment vary across grape cultivars or impact grapevine performance in dry conditions.Most studies have focused on one or two cultivars and have shown that grapevines adjust osmotically over the growing season or during drought, and that vines that have undergone adjustment are less vulnerable to (i.e. have more negative leaf water potential thresholds for) leaf hydraulic dysfunction and stomatal closure (Martorell et al., 2015;Rodriguez-Dominguez et al., 2016;Sorek et al., 2021).Also, across three cultivars, a more negative π o measured once in the growing season was associated with less vulnerability to leaf hydraulic dysfunction and stomatal closure (Dayer et al., 2020).However, the largest study comparing π o across cultivars found that osmotic potential was unrelated to stem embolism resistance, raising uncertainty about the importance of this trait to whole-plant drought tolerance (Alsina et al., 2007).Furthermore, other work has found that cultivars typically grown in hotter, drier regions exhibit more water-saving stomatal behaviour, including a lower maximum stomatal conductance (Bartlett and Sinclair, 2021).Modelling work also predicted that osmotic adjustment would increase gas exchange and soil water depletion and cause grapevines to reach critical thresholds for water stress earlier in the growing season (Herrera et al., 2022).These findings suggest that grapevines could use the opposite trait values to wild species (i.e. a less negative osmotic potential and lower osmotic adjustment) to adapt to hotter, drier conditions, if grapevines benefit more from conserving water than maintaining high gas-exchange rates.Evaluation of how these traits contribute to differences in stress tolerance among cultivars would provide insight into whether these traits are worthwhile targets for efforts to improve grapevine cultivars, in addition to the direction in which these traits should be changed.
Previous work has also suggested that the chemical composition of the solutes could impact stress tolerance.Leaf cells can accumulate a wide range of solutes during osmotic adjustment, including inorganic ions, sugars, amino acids and proteins, and solute composition varies widely across species (Zivcak et al., 2016).Synthesizing organic solutes, such as sugars or amino acids, is more resource intensive and energetically expensive than increasing inorganic ion uptake from the soil.Additionally, some organic solutes (e.g.proline) also serve as osmoprotectants, which enhance drought tolerance by stabilizing protein and membrane structures to reduce damage from dehydration (Gagneul et al., 2007;Zivcak et al., 2016).Leaf solute composition has been measured for only a few grape cultivars, and it is unknown whether solute composition contributes to differences among cultivars in drought or heat tolerance (Patakas et al., 2002;Degu et al., 2019).If so, this would indicate that the identification of specific solutes and their role in osmotic adjustment could help to generate new plant material that uses the most effective solutes to achieve optimal values for osmotic potential and osmotic adjustment.
In this study, we tested whether osmotic potential, osmotic adjustment and solute composition vary across Vitis vinifera winegrape cultivars historically adapted to different climatic conditions and are associated with differences among cultivars in vine physiological performance (i.e.gas exchange and water potentials) in hot conditions.Specifically, we tested whether: (1) there are significant differences in osmotic potential, osmotic adjustment and solute composition between cultivars; (2) these differences correspond to differences among cultivars in climate associations (i.e. the typical climatic conditions where each cultivar is grown); and (3) these traits are correlated with vine water potentials and gas exchange.We compared these variables across seven cultivars growing in common-garden conditions in a hot wine region.We hypothesized that cultivars that are typically grown in hotter, drier regions would exhibit greater osmotic adjustment and maintain more negative osmotic potentials.We also hypothesized that these traits would enable these cultivars to undergo greater leaf water stress and maintain greater stomatal conductance and photosynthesis over the hottest, most waterstressed period of the growing season.We also expected solute composition to vary across cultivars and correspond to differences in climate associations, although it was unknown from previous work whether adapting to heat and drought stress would favour ion accumulation, as a metabolically 'cheap' strategy to lower osmotic potentials, or the production of organic osmoprotectants to protect the biochemical machinery from dehydration.We evaluated relationships between these traits, plant physiological performance and historical climate associations in winegrapes, which are an excellent study system for climate adaptation because cultivars have diverse and well-characterized climatic niches (Anderson and Nelgen, 2020).Furthermore, winegrapes are an economically important crop (valued at $70 billion worldwide) under considerable threat from climate change (Jones et al., 2004;Alston and Sambucci, 2019).Addressing these hypotheses should provide crucial insight into the physiological mechanisms adapting winegrapes to stressful growing conditions.
Plant material and growth conditions
We measured leaf water relationships and chemistry on mature vines of seven Vitis vinifera cultivars typically grown in different climatic regions (i.e.Riesling and Pinot Noir from cool regions, Chardonnay, Merlot and Syrah from warm regions, and Zinfandel and Sangiovese from hot regions; n = 3 or 4 vines per cultivar).Leaf osmotic potential at full turgor (π o ) was measured on these same vines on three sampling dates throughout the growing season (Table 1).
The vines are established in an experimental vineyard on the University of California, Davis campus (38.53°N, -121.75°W).Half of the vines of each cultivar were divided between two adjacent blocks.The blocks are established with a north-south row orientation and are all trained using a California vertical shootpositioned trellis system.All vines are grafted onto the same rootstock (420A).Soil types at the site range from a Reiff to a Yolo loam (USGS Web Soil Survey).During the experimental period, all plants received the same irrigation and no precipitation.The vineyard is drip irrigated approximately once per week to replace 80 % of water loss.The replacement amount is based on reference evapotranspiration values generated by the Davis California Irrigation Management Information System (CIMIS) and the seasonal crop coefficient (K c ) values, which are calculated based on equations from the study by Williams et al. (2014).We used the same K c value for all cultivars, because the vines had visually similar canopy sizes, although we did not measure canopy size and thus could have underestimated differences in irrigation demand among cultivars.However, mean predawn water potentials were similar across cultivars (i.e.−0.22 to −0.36 MPa), suggesting that soil water availability was largely consistent across cultivars (Table 2).
We conducted measurements from the onset of berry ripening (veraison) to harvest (from July to September) in 2020, to capture osmotic adjustment during the hottest period of the growing season.The experimental vineyard is located in a hot (Winkler V) growing region.Daily mean and maximum temperatures ranged from 21 to 31 °C and from 26 to 40 °C over the study period, respectively, based on climate data collected by the Davis CIMIS station (https://cimis.water.ca.gov/).The site experienced a severe heat wave in mid-August (14-18 August 2020) that considerably increased atmospheric evaporative demand (Fig. 1).Following standard commercial practices, we increased irrigation by 50 % in the irrigation event before the heatwave.
Climate associations
We defined cultivar climate associations in two ways.First, we represented climate as a set of continuous variables, using the methods from Bartlett and Sinclair (2021).To summarize, we used the 2016 global winegrape dataset from Anderson and Nelgen (2020) to identify the Old World growing regions where each cultivar in our study is well represented.For each cultivar, we defined the well-represented regions as those containing ≥5 % of the total Old World bearing area of the cultivar.We then used this subset to calculate the bearing area fraction in each wellrepresented region, such that the sum of bearing area fractions across well-represented regions equals 100 % for each cultivar.We used the coordinates from Anderson and Nelgen (2020) to extract maximum monthly temperature (T max ) and vapour pressure deficit (VPD max ) for each growing region from the WorldClim dataset, because these variables were the most strongly correlated with gas exchange in a previous meta-analysis (Bartlett and Sinclair, 2021).We then used the bearing area fractions for each region to calculate a weighted average T max and VPD max for each cultivar.We focused on Old World growing regions, where irrigation has historically been banned outright or heavily restricted for winegrapes, to avoid confounding effects of irrigation on the relationships between traits and climate.
Second, to test a common simplified approach, we classified cultivars according to the climate categories from Anderson and Nelgen (2020).This dataset records the global bearing area of each cultivar located in cool, warm or hot growing regions.Mean growing season temperature is <15 °C for cool regions, 17-19 °C for warm regions and >19°C for hot regions.The climate category for each cultivar is defined as the category containing most of its bearing area.Our cultivars were divided among three groups: cool, Riesling and Pinot; warm, Chardonnay, Merlot and Syrah; and hot, Zinfandel and Sangiovese.Similar methods have been used to define regional suitability for cultivars and to predict cultivar responses to future climate conditions (Fraga et al., 2016;Bartlett and Sinclair, 2021;Lamarque et al., 2023).We used both approaches in our study to test whether these methods identify the same relationships between physiology and climate.
Plant water status and gas exchange
We measured leaf water potential (Ψ) and gas exchange at midday (between 1100 and 1300 h) once per week from 16 July to 3 September 2020.We selected healthy, newly expanded mature leaves, 8-12 nodes below the shoot tip, consistently on the east side of the canopy.We measured stomatal conductance and photosynthesis on two leaves per vine with a portable gasexchange system (Li-Cor 6800; Lincoln, NE, USA), using a fan speed of 10 000 rpm, CO 2 concentration of 400 µmol mol −1 and light intensity of 1900 µmol m −2 s −1 .We allowed humidity and air temperature in the sample chamber to match ambient conditions.We selected two adjacent leaves per vine and measured midday water potential with a pressure chamber (PMS Instrument; model 1505D; n = 6-8 leaves per cultivar).Leaves were excised at the base of the petiole, sealed in humidified Whirl-pak bags and either measured immediately or stored in the refrigerator for ≤1 week before measuring.We also measured one leaf per vine for predawn leaf water potential between 0400 and 0600 h at the beginning, middle and end of the experimental period (23 July, 5 August and 3 September 2020).
Osmotic potential at full turgor
We measured leaf osmotic potential at full turgor (π o ) on three sampling dates (15 July, 18 August and 16 September 2020).We excised one shoot per vine, placed the end of the shoot in deionized water and covered the shoots in a dark, humidified plastic bag to rehydrate overnight.We double-bagged two leaves per shoot in humidified Whirl-pak bags at the same time the following morning to standardize the leaf rehydration time.
We then measured leaf osmotic potential following the rapid osmometer method from Bartlett et al. (2012).Briefly, we punctured and froze leaf discs in liquid nitrogen, then sealed the discs in a vapour pressure osmometer (Vapro 5600, Wescor, Logan, UT, USA) to determine the osmotic potential at full turgor.
Sampling for leaf chemistry
To measure leaf solute composition, we collected two leaves per plant from the same shoots used to measure osmotic potential on two of the sampling dates (15 July and 16 September 2020), then flash-froze the leaves in liquid nitrogen.Leaves were cryogenically pulverized to a fine powder using a tissue lyser (Retsch, Newton, PA, USA) with steel jars containing 2-cm-diameter steel balls.Samples were stored at −80 °C until analysis.
Inorganic ions
K, Ca, Mg and Na ion concentrations were measured by the UC Davis Analytical Lab (Davis, CA, USA), following standard analytical methods (Meyer and Keliher, 1992;Sah and Miller, 1992).Briefly, ions were extracted from 0.4 g of dry leaf biomass using nitric acid-hydrogen peroxide microwave digestion and quantified with inductively coupled plasma atomic emission spectrometry (ICP-AES).Each sample was digested with 2 mL 3X deionized water, 2 mL hydrogen peroxide and 1 mL trace metal grade nitric acid, using a microwave digestion system (Mars Xpress, Matthews, NC, USA).Each sample was brought up to a final volume of 15 mL with 3X deionized water (dilution factor ×30), then diluted ×4 again and analysed with a Thermo ICP 6500 (Thermo Scientific, Waltham, MA, USA).Detection limits for this method range from 0.5 to 100 ppm.
Amino acids
Amino acids were extracted from 100 mg of fresh leaf tissue using an EZ:FAAST GC-FID kit (Phenomenex, Torrance, CA, USA) following methods from Wallis et al., (2012).Briefly, 100 mg of fresh leaf tissue was extracted in 500 μL of phosphate-buffered saline (PBS) solution adjusted to a pH of 6.8.Samples were vortexed and shaken overnight at 4 °C.The following day, the samples were centrifuged for 1.5 min at 10 000g.The supernatant was removed, and the pellet was washed with 500 μL of fresh PBS, centrifuged and left overnight at 4 °C once again.The supernatants were then combined to total 1000 μL.One hundred microlitres of the supernatant collected the following day and was used for amino acid quantification, following the user instructions in the EZ: FAAST gas chromatography-flame ionization detector (GC-FID) kit.The column, eluting medium, reagents and standards used to identify amino acids were all supplied by the kit.Samples were prepared and measured the same day with a Shimmadzu GC-2010 system using an FID.
Statistical analyses
We used a type III ANOVA to test the model π o ~ date + variety + date × variety, to determine whether π o varied significantly over the study period (date) and across cultivars (variety) and whether adjustment in π o varied significantly across varieties (date × variety).We repeated this analysis for each of the gas-exchange, water potential and solute concentration variables.We were unable to fit a type III ANOVA for stomatal conductance (g s ) and photosynthesis (A) because of multicollinearity between the main effects and interaction term; therefore, we tested for main effects of date and variety with a type II ANOVA (Supplementary Data Tables S1 and S2), which has more power for models without interaction terms.For consistency, we also used a type II ANOVA to test the main effects for the other dependent variables with insignificant interaction terms, and this did not impact the significance of the main effects for any of these variables.We used Tukey's post-hoc HSD tests to compare differences between varieties.We used the same approach to test differences between climate groups (i.e.π o ~ date + climate group + date × climate group) (Supplementary Data Tables S3 and S4).
We used linear regression to test correlations between π o , gas exchange and the predawn (Ψ PD ) and midday water potentials (Ψ MD ).We tested correlations between values measured in the same week, in order to avoid confounding effects from measuring these variables in highly different environmental conditions.We also tested correlations between osmotic adjustment (Δπ o ) and changes in gas exchange and water potential, and between osmotic adjustment at the water potential at the beginning of each adjustment period, to test whether the more water-stressed cultivars exhibited greater adjustment.Finally, we used linear regression to test correlations between the weighted average climate variables and π o , osmotic adjustment, gas exchange and water potentials.All analyses were conducted with Rstudio (v.4.2.2).
Osmotic potential and osmotic adjustment
All cultivars significantly reduced osmotic potential at full hydration (π o ) over time, and mean osmotic potential was significantly different across cultivars (ANOVA, P < 0.05; Table 1; Fig. 2).However, the interaction between date and variety was not significant, indicating that osmotic adjustment was not different across varieties.Cultivar mean π o values ranged from −1.05 ± 0.06 to −1.48 ± 0.08 (mean + s.e.) at veraison (July) and from −1.68 ± 0.05 to −2.23 ± 0.04 at harvest (September).The mean adjustment in π o across cultivars was larger from July to August (Δπ o = −0.44MPa) than from August to September (Δπ o = −0.22MPa) (Table 1).Notably, the ranking in osmotic potential across cultivars was largely consistent over the season (Fig. 2).Mean π o was consistently the most negative in 1 and 2).However, there was no significant interaction between date and variety or between variety and climate group, indicating that there were no significant differences in osmotic adjustment (Δπ o ) (Tables 1 and 2).
Merlot, followed by Riesling, intermediate in Pinot Noir and Chardonnay, and consistently higher in Sangiovese, Syrah and Zinfandel.
Plant water status and gas exchange
Stomatal conductance (g s ), photosynthesis (A) and midday leaf water potentials (Ψ MD ) were significantly different between sampling dates and cultivars (ANOVA, P < 0.05; Table 2).Cultivar mean g s values from July to September ranged from 0.212 ± 0.010 mmol m −2 s −1 (mean ± s.e.) for coolclimate Riesling to 0.341 ± 0.011 mmol m −2 s −1 for warmclimate Syrah.Mean values of A ranged from 16.26 ± 0.35 µmol m −2 s −1 for cool-climate Pinot Noir to 18.36 ± 0.38 µmol m −2 s −1 for Syrah (Table 2; Fig. 3).Post-hoc tests indicated that g s was higher in Syrah than in the other cultivars, while A was higher in Syrah than in Riesling and Pinot Noir (Tukey's HSD, P < 0.05; Table 2).Midday leaf water potentials ranged from −1.06 ± 0.05 MPa for Pinot Noir to −1.35 ± 0.05 MPa for Chardonnay and were lower for Chardonnay and Riesling than for Zinfandel and Pinot Noir (Tukey's HSD, P < 0.05) (Table 2; Fig. 3).All cultivars experienced the most negative midday leaf water potentials in late August.In response, there was a wide range in midday leaf water potential from −1.44 MPa (Sangiovese) to −1.83 MPa (Merlot) (Fig. 3).In contrast, predawn leaf water potentials were not significantly different between cultivars or sampling dates.
Relationships between osmotic potential, gas exchange and midday water potential
We tested correlations between π o , gas exchange and Ψ MD for each of the three sampling periods when these variables were measured in the same week.The value of π o was significantly correlated with photosynthesis in September (r 2 = 0.51, P < 0.05, n = 8; Table 3; Fig. 4A).Stomatal conductance was ).The values of A and g s varied significantly between date, variety and climate group, but there was no significant interaction between date and variety or between variety and climate group (Table 2; A, B).Midday water potentials also varied significantly between date and variety but not climate group, whereas there was no significant variation in predawn leaf water potential (Table 2; C, D).The value of Ψ MD was significantly correlated with π o early in the season, during the month of July (r 2 = 0.63, P < 0.05, n = 8; Table 3).In contrast, osmotic adjustment was not significantly correlated with changes in gas exchange or midday water potential at the beginning of the adjustment period, but the midday water potential at the end of the adjustment period was significantly correlated with osmotic adjustment (r 2 = 0.055, P < 0.05, n = 6-8; Table 4; Fig. 4B).
Leaf chemical composition
All inorganic ion concentrations, except for Na, changed significantly over time, and mean Ca, Mg and K concentrations were significantly different across cultivars (Table 5; Fig. 5).However, the interaction between date and variety, indicating that cultivars showed different patterns in accumulation, was significant only for Mg (Table 5; Fig. 5).Mean Mg and Ca concentrations increased from July to September, whereas K concentrations decreased.The absolute change in concentration was largest for Ca.
Total amino acid (TAA) content decreased significantly over the season, but mean concentrations were not significantly different across cultivars (Table 6; Fig. 6).
Proline concentrations were also not significantly different across cultivars and did not change significantly over time (Table 6; Fig. 6).
Climate of origin and climate groups
Photosynthesis and π o were significantly correlated with the climate associations of cultivars and were significantly different between categorical climate groups.Photosynthesis was significantly correlated with the weighted maximum growing season temperature (T max , r 2 = 0.85, P ≤ 0.05, n = 8) and vapour pressure deficit (VPD max , r 2 = 0.73, P ≤ 0.05), and π o was significantly correlated with VPD max (r 2 = 0.69, P ≤ 0.05; Fig. 7).Photosynthesis and π o were both higher in the cultivars associated with hot, less humid growing regions.These traits were also significantly higher in the hotclimate cultivars (i.e.Zinfandel and Sangiovese) than the other climate groups (Table 2).Conversely, osmotic adjustment, water potentials and inorganic and organic solute concentrations were not significantly different across the climate groups.
DISCUSSION
We found that mean osmotic potential varied significantly between winegrape cultivars and that all cultivars reduced osmotic potential (i.e.osmotically adjusted) significantly over the ripening period, but that adjustment was largely uniform, preserving cultivar rankings in osmotic potential (Table 1; Fig. 2).Mean osmotic potentials were correlated with cultivar climate associations, but in the direction opposite to expected, with cultivars typically grown in hotter, less humid wine regions exhibiting less negative osmotic potentials (Table 1; Fig. 7).Depending on the sampling date, osmotic potential and osmotic adjustment were either uncorrelated with gas exchange and leaf water stress or correlated in the direction opposite to expected, with a less negative osmotic potential being associated with greater gas exchange (Tables 1 and 4; Figs 4 and 5).Photosynthesis, but not stomatal conductance, was higher in the cultivars typically grown in hotter, less humid regions (Table 2; Fig. 3).Leaf chemical composition varied between cultivars and over the study period, but this variation was not related to climate associations (Table 5; Fig. 5).Altogether, these findings suggest that reducing leaf osmotic potentials has not been a primary mechanism for winegrapes to adapt to hotter, drier regions, contrary to other plant species (Bartlett et al., 2012).Instead, other mechanisms, such as increasing photosynthetic rates in hot conditions, could be more promising targets for developing climate-resilient grape cultivars.More negative osmotic potentials increase leaf drought tolerance by improving maintenance of turgor, which reduces leaf vulnerability to wilting, hydraulic dysfunction and stomatal closure during drought (Patakas et al., 1999;Martorell et al., 2015;Scoffoni et al., 2018;Herrera et al., 2022).Thus, we expected cultivars adapted to hotter, drier regions to exhibit more negative mean osmotic potentials and greater osmotic adjustment.However, we found the opposite patterns.Osmotic potentials were significantly less negative for the hot-climate cultivars than for the other climate groups, and less negative osmotic potentials were significantly associated with a higher maximum growing season vapour pressure deficit (VPD max ) and a higher growing season temperature maximum (T max ) (Fig. 7).These findings could indicate that less drought-resistant leaves are adaptive for winegrapes in hot, dry conditions.We did not find a relationship between π o and g s in the range of Ψ values observed in this study (i.e.mean Ψ MD = −1.1 to −1.4 MPa).However, under more severe water stress, a higher π o could induce earlier stomatal closure or leaf shedding, producing a larger hydraulic safety margin that extends the time to reach critical thresholds for water stress longer into the growing season (Tyree and Ewers, 1991;Hochberg et al., 2016;Charrier et al., 2018;Herrera et al., 2022).Historically, many European wine regions have had strict legal limitations on irrigation, and larger safety margins could have helped hot-climate cultivars to avoid hydraulic damage that leads to long-term disruptions in function, such as stem embolism, during extreme drought or heat events.Hot-climate cultivars could also require a higher π o to produce sufficient safety margins if they are less embolism resistant, which has been found in some studies (Bartlett and Sinclair, 2021) but not others (Lamarque et al., 2023).Alternatively, our findings could indicate that osmotic potential is determined by adaptations beyond drought tolerance.For example, cool-climate cultivars could accumulate more solutes in the leaves during ripening to translocate to the woody tissues before dormancy, to provide greater protection from freezing.Many species use solute accumulation in woody tissues to prevent freezing damage by reducing tissue freezing points and avoiding cellular dehydration (Yuanyuan et al., 2009).Cool-climate cultivars also typically finish ripening and stop translocating sugars and nutrients to the berries earlier in the growing season, which could contribute to greater solute accumulation in the leaves.
All cultivars osmotically adjusted significantly over the ripening period, which is consistent with findings from other field studies for grape (Alsina et al., 2007;Herrera et al., 2022).Most work in other crops has assumed that increasing osmotic adjustment improves drought tolerance (Zivcak et al., 2016;Blum et al., 2017), but we found that osmotic adjustment was not significantly different between climate groups or correlated with climate variables.These findings suggest that osmotic adjustment is not a key trait driving diversification across climates for winegrapes.
Previous work found that cultivars with lower osmotic potentials had more negative water potential thresholds for stomatal closure, and that osmotic adjustment made stomatal and hydraulic conductance less sensitive to leaf water potential over the growing season (Martorell et al., 2015;Sorek et al., 2021;Herrera et al., 2022).Thus, we expected that greater osmotic adjustment and more negative osmotic potentials would allow for greater gas exchange during our study period, in which the vines experienced a record-breaking heatwave at an already hot site.However, π o was mostly uncorrelated with gas exchange or correlated in the direction opposite to expected.Osmotic potential was correlated with gas exchange only in July and September, and a less negative osmotic potential was associated with greater A and unrelated to g s (Table 3).Also, osmotic adjustment was not correlated with changes in gas exchange (Table 4).Altogether, these findings suggest that π o and the capacity for maintenance of leaf turgor is not a main driver of cultivar differences in gas exchange in typical vineyard conditions.Instead, osmotic adjustment and π o might be more closely related to gas exchange during more severe water stress, closer to thresholds for stomatal closure.Alternatively, previous work has suggested that a higher π o allows for a higher maximum g s and A by reducing maximum turgor in the epidermal cells and, thus, turgor limitations on maximum stomatal opening (Henry et al., 2019).This finding is consistent with the positive correlation between π o and A, but the lack of a correlation with g s suggests that a direct effect of π o on stomatal opening is unlikely to drive this relationship.Instead, hot, high VPD conditions might have selected independently for both a higher π o and A.
Photosynthesis was significantly higher in cultivars typically grown in regions with a higher maximum temperature and VPD (Fig. 7).This is the first study to test correlations between typical growing region climate and A, but these findings are largely consistent with previous comparisons of fewer cultivars.Of the six studies where at least two cultivars with published mean growing season temperatures were measured for gas exchange in hot (>30 °C) conditions, both A and g s were significantly higher for the warmer-climate cultivar in three studies (Moutinho-Pereira et al., 2007;Palliotti et al., 2015;Greer, 2018), only A was higher in two (Chaves et al., 1987;Costa et al., 2012), and neither was significantly different in one (Medrano et al., 2003).However, the climate variables were not correlated with stomatal conductance, suggesting that the relationship with A was not driven by stomatal behaviour.Instead, the heat-adapted cultivars could have a more heat-tolerant photosynthetic biochemistry.High temperatures (>35 °C) can limit photosynthesis by reducing maximum rates of carboxylation (V cmax ) and the electron transport chain reactions (J max ) ( Gallo et al., 2021).The V cmax , J max and their temperature dependence vary between cultivars.For example, V cmax and J max were more strongly downregulated as temperatures increased above 35 °C in Grenache than in Syrah (Gallo et al., 2021) and in Chardonnay than in Merlot (Greer et al., 2018).The heatadapted cultivars could have a greater capacity to protect or repair the photosynthetic biochemical machinery from heat stress, allowing these cultivars to maintain a higher J max , V cmax and overall photosynthetic rate at our hot study site.
Leaf chemistry varied between cultivars and changed over the ripening period, but accumulation was significantly different between cultivars only for Mg.Mean Ca, K and Mg concentrations varied significantly between cultivars (Fig. 5).For all cultivars, Ca was the most concentrated mineral at each time point and the most accumulated mineral over time, as observed previously for individual cultivars (e.g.Merlot; Degu et al., 2019).Ca is immobile in the phloem, which limits translocation to the berries and facilitates accumulation in the leaves as berry hydraulics become phloem-dominated at veraison (Hocking et al., 2016).Mg concentrations increased and K concentrations decreased over the season for all cultivars, contrary to previous findings for K accumulating in response to water stress (Patakas et al., 2002;Degu et al, 2019) (Fig. 5).Post-veraison competition between the leaf and berry could have driven the decreases in K, because berry osmotic regulation and demand for K increases near harvest (Monder et al., 2021).K also mediates drought responses by assisting with stomatal regulation (Monder et al., 2021) and, notably, Syrah exhibited the highest K concentrations and gas-exchange rates.Mg and K also compete for plant uptake, and the relatively low soil K/Mg ratio at our site (<0.1) could have contributed to the greater accumulation of Mg.Altogether, our findings show that cultivars growing at the same site and grafted to the same rootstock can vary significantly in nutrient content.The mechanisms driving these differences are poorly understood, and these differences were not explained by climate associations (Supplementary Data Table S2).Finally, TAA and proline content were not significantly different between cultivars or climate groups, contrary to our hypothesis that heat-adapted cultivars would generate osmoprotectant compounds to protect the photochemical machinery from stress.
In sum, contrary to findings for other crops and wild plant species, we did not find that winegrape cultivars have adapted to hotter, drier conditions by increasing osmotic adjustment or reducing osmotic potentials (Bartlett et al., 2012(Bartlett et al., , 2014;;Blum et al., 2017).Instead, osmotic potentials were either unrelated or positively correlated with gas exchange, and heat-adapted cultivars exhibited both higher photosynthetic rates and less negative osmotic potentials (Fig. 4; Tables 1 and 2).These findings suggest that differences among cultivars in gas exchange are driven primarily by besides the capacity for maintenance of turgor, and that osmotic potentials in grape are more closely related to processes other than leaf water relationships.
Increasing photosynthesis in hot conditions emerged as a more promising target for cultivar improvement than reducing osmotic potentials, if breeding programmes build on existing adaptations, but more work is needed to evaluate whether this strategy is beneficial in the new conditions expected from climate change.
CONCLUSION
We tested whether leaf osmotic potential and osmotic adjustment, classical water relationship traits that have been highly predictive of drought tolerance in other crops and naturally occurring plant species, have been important drivers of environmental diversification for winegrapes.We hypothesized that grape cultivars have adapted to hotter, drier growing regions by using greater osmotic adjustment and more negative osmotic potentials to improve maintenance of turgor and reduce vulnerability to wilting, hydraulic dysfunction and stomatal closure.Our seven geographically diverse focal cultivars varied significantly in mean osmotic potentials and osmotically adjusted significantly from the onset of ripening (veraison to harvest), but the cultivars associated with the hottest, driest regions exhibited the least negative osmotic potentials, contrary to our hypotheses.Osmotic potentials were either uncorrelated or positively correlated with gas exchange, indicating that grapevines have not improved gas exchange in hot conditions by increasing the capacity for maintenance of turgor.Instead, grapevine osmotic potentials could be related more closely to nutrient storage or sugar translocation.Photosynthesis, but not stomatal conductance, was significantly higher in the heatadapted cultivars at our hot study site.Future studies should test whether this relationship reflects selection for more heattolerant photochemical machinery in the hot-climate cultivars.Leaf chemistry was not related to climate, indicating that heat-adapted cultivars did not maintain greater photosynthesis through increased production of osmoprotectants.Overall, these findings suggest that maintenance of leaf turgor is not a primary limitation on grapevine adaptation to hot, dry atmospheric growing conditions, and that other traits, including photochemical heat tolerance, would be a more promising focus for efforts at cultivar improvement.
Fig. 1 .
Fig. 1.Maximum daily temperatures at the study site over the summer 2020 study period compiled from the University of California, Davis CIMIS station (station #6; https://cimis.water.ca.gov/).
Fig. 2 .
Fig. 2. Leaf osmotic potential at full hydration (π o ) measurements from July, August and September.Data points represent the mean π o for each cultivar and sampling date (n = 6-8).The value of π o varied significantly between date, variety and climate group (P < 0.05; Tables1 and 2).However, there was no significant interaction between date and variety or between variety and climate group, indicating that there were no significant differences in osmotic adjustment (Δπ o )(Tables1 and 2).
Fig. 3 .
Fig. 3. Photosynthesis (A; A), stomatal conductance (g s ; B), predawn leaf water potential (Ψ PD ; C) and midday leaf water potential (Ψ MD ; D) measurements over the study period.Points are cultivar means (n = 6-8).The values of A and g s varied significantly between date, variety and climate group, but there was no significant interaction between date and variety or between variety and climate group (Table2; A, B).Midday water potentials also varied significantly between date and variety but not climate group, whereas there was no significant variation in predawn leaf water potential (Table2; C, D).
Fig. 4 .
Fig. 4. Correlations between osmotic potential (π o ) and significant gas exchange variable (A) in the month of September (A) and midday leaf water potential (Ψ MD ) in July (B).(A) The value of A was significantly correlated with π o across cultivars, but only in the month of September (r 2 = 0.51, P < 0.05, n = 8).(B) The value of Ψ MD was significantly correlated with π o , but only in the month of July (r 2 = 0.63, P < 0.05).Photosynthesis rates were highest in the cultivars with the least negative π o values, and π o was not significantly correlated with stomatal conductance, contrary to expectation.
Fig. 5 .
Fig. 5. Mean Ca (A), K (B), Mg (C) and Na (D) concentrations, expressed as a percentage of dry leaf sample, at the beginning and end of the sampling period.Bars are means for cultivars and sampling dates (July and September; n = 6-8).Error bars represent the s.e.The Ca, K and Mg varied significantly with date, variety and climate group.Only Mg displayed a significant interaction between date and variety.The Na concentration levels were insignificant across all main effects.
Fig. 6 .
Fig. 6.Mean proline (A) and total amino acids (B) concentrations at the beginning and end of the experimental period (July and September).Error bars are the s.e.Proline did not vary significantly between date, variety or climate group, whereas total amino acids varied significantly only with date.
Fig 7 .
Fig 7. Correlations between π o and gas exchange and cultivar climate associations.Climate associations capture growing season climate conditions in the regions where each cultivar is typically grown.Maximum growing season temperature (T max ) and vapour pressure deficit (VPD max ) were significantly correlated with mean photosynthesis (r 2 = 0.85 and 0.73, respectively, P < 0.05, n = 8) and π o (r 2 = 0.51 and 0.69, P < 0.05) over the study period.Cultivars associated with hotter, drier climates had higher photosynthetic rates in our hot common-garden study, but less negative π o values, contrary to our predictions.
FUNDING
This work was supported by the American Vineyard Foundation (grant number 2284) and the Department of Viticulture and Enology at the University of California, Davis.This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under grant number 2036201.
Table 1 .
Monthly osmotic potential (π o ) measurements shown in megapascals.Values are cultivar means ± s.e.Letters show Tukey's post-hoc HSD test results.
Table 3 .
Linear regressions between osmotic potential and stomatal conductance (g s ), photosynthesis (A) and midday water potentials (Ψ MD ) for each sampling date for osmotic potential.Bold values show significant correlations (P < 0.05).notsignificantly correlated with π o during the study period.
Table 4 .
Linear correlations with gas exchange and Δπ o and with midday leaf water potential (Ψ md ) and Δπ o across all individuals.Bold text signifies significant values.
Table 5 .
Leaf ion concentrations at the beginning and end of the study period.Values are percentages per dry biomass sample ± s.e.Letters show Tukey's post-hoc HSD test comparisons.
Table 6 .
Proline and total amino acid (TAA) concentrations at the beginning and end of the study period. | 2024-03-14T06:18:34.902Z | 2024-03-13T00:00:00.000 | {
"year": 2024,
"sha1": "1362c1b199342e18fb9f5866637e296b4beaa454",
"oa_license": "CCBY",
"oa_url": "https://academic.oup.com/aob/advance-article-pdf/doi/10.1093/aob/mcae032/56956452/mcae032.pdf",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "5c48fd4aa0cafa1e2ab4ece35ac0df3704eee603",
"s2fieldsofstudy": [
"Environmental Science",
"Agricultural and Food Sciences"
],
"extfieldsofstudy": [
"Medicine"
]
} |
255250700 | pes2o/s2orc | v3-fos-license | Comparison of PD-L1 expression in squamous cell cancer of unknown primary and oropharyngeal squamous cell carcinoma
Purpose The tumorigenesis of squamous cell cancer of unknown primary (SCCUP) in the head and neck area has not been decoded so far, while poor survival rates and limited therapeutic options pose a serious challenge. The aim of this project was to investigate immunological characteristics of SCCUPs and compare them to oropharyngeal squamous cell carcinoma (OPSCC). Methods PD-L1 expression (TC) was examined by immunohistochemistry in 50 lymph node metastases of SCCUP and 47 primaries of OPSCC. CD3 + and CD8 + lymphocytic infiltration was measured in 5 high power fields. Expression of p16 and HPV ISH were assessed. Results SCCUP demonstrated a significantly higher expression of PD-L1 than OPSCC. In p16-negative SCCUPs PD-L1 proved to be an independent prognostic factor to prioritize high-risk patients. Conclusions Immunologic differences between SCCUP and OPSCC were detected. A higher PD-L1 expression in SCCUP could potentially facilitate further evaluation of checkpoint inhibitor therapy. Supplementary Information The online version contains supplementary material available at 10.1007/s00405-022-07775-z.
Introduction
The tumorigenesis of squamous cell cancer of unknown primary (SCCUP) in the head and neck area has not been decoded so far, while poor survival rates and limited therapeutic options pose a serious challenge [1,2]. The aim of this project was to investigate immunological characteristics of SCCUPs and compare them to oropharyngeal squamous cell carcinoma (OPSCC); (2) Methods: PD-L1 expression (TC) was examined by immunohistochemistry in 50 lymph node metastases of SCCUP and 47 primaries of OPSCC.
CD3 + and CD8 + lymphocytic infiltration was measured in 5 high power fields. Expression of p16 and HPV ISH were assessed; (3) Results: SCCUP demonstrated a significantly higher expression of PD-L1 than OPSCC. In p16-negative SCCUPs PD-L1 proved to be an independent prognostic factor to prioritize high-risk patients; (4) Conclusions: immunologic differences between SCCUP and OPSCC were detected. A higher PD-L1 expression in SCCUP could potentially facilitate further evaluation of checkpoint inhibitor therapy. The incidence of cancer worldwide is more than 24.5 million cases, resulting in 9.6 million deaths per year. This number increased by more than 30% between 2007 and 2017 [3]. Approximately 2-4% of all cancers are defined as cancer of unknown primary (CUP). Distinct clinical features include early metastatic dissemination in an atypical pattern, poor response to conventional chemotherapy and aggressive progression [3]. In the head and neck region up to 9% of malignancies are considered as CUP, mainly with squamous cell appearance (53-77%). Even though using an extensive diagnostic workup, the corresponding primary tumor of a lymph node metastasis cannot be found. Presumably due to better imaging, the incidence of CUP is decreasing, but the difficulty of diagnosis and treatment still poses a big challenge [4,5]. Thinking of CUP as the metastasis of a primary tumor, identification of the primary tumor is the most important aspect of improving survival and quality of life in patients. The theory of a concealed primary is supported by reports, that a majority of primaries could be found in autopsy studies [6]. The hypothesis of CUP as a metastasis of an (still) occluded primary is also supported by the fact, that around 15% of CUPs resemble a cancer of known origin [4,7].
Similar to HNSCC, a large proportion of CUP cases present themselves for the first time in the form of cervical lymph node metastasis [8]. Due to the lymphatic drainage of the pharynx, the most probable location of a presumed primary of SCCUP in the neck levels II-IV is the pharynx [9]. Most cases initially present with a neck mass or other metastasis-related symptom, such as weight loss, malaise or fatigue [10]. As these symptoms do not point to a specific location, the next steps are thorough clinical and radiological diagnostics and histopathological examination of the lymph node tissue [7]. Epstein-Barr virus-encoded RNA (EBER) and p16, a surrogate marker for HPV, are assessed [11]. p16 is important, as an infection with HPV is common in squamous cell carcinomas of the oropharynx [12]. EBER on the other hand is linked with carcinogenesis of nasopharyngeal carcinoma, another potential site of origin [13]. Since oropharyngeal squamous cell carcinoma (OPSCC) are the most common HNSCC [14], SCCUP were compared to OPSCC in this study. Diagnostic imaging then allows the evaluation of the extent of the disease. Treatment of cervical metastasis is still controversial, since randomized trials are not yet available [15]. Unilateral neck dissection or radiation is often performed as initial treatment. In addition, a tonsillectomy and base of tongue mucosectomy are also viable options [11,12]. In the palliative setting, the response rates to platin-based chemotherapy are limited and alternative therapies are needed to improve the survival of patients [15]. The heterogeneous molecular character of CUPs has defied a "one size fits all" solution in the past [16] and emphasized therapy based on the individual molecular landscape.
In HNSCC recent findings highlighting the importance of the tumor microenvironment and its interactions with the tumor cells paved the way for immunotherapy [17][18][19]. Checkpoint inhibitors targeted against the PD-L1/PD-1 axis demonstrated remarkable success in treatment of various cancer types including HNSCC. PD-L1 is a transmembrane protein on the surface of antigen presenting cells and tumor cells, and on the other hand is expressed on the surface of immune-related lymphocytes. Binding of PD-L1 to PD-1 inhibits proliferation and cytotoxicity of lymphocytes [20][21][22]. To assess the probability of treatment success, the histopathological PD-L1 expression is routinely used to administer checkpoint inhibitor treatment in HNSCC [23]. A meta-analysis of PD-L1 expression and response to checkpoint inhibitor therapy confirmed a better overall survival and tumor response in HNSCC patients with high PD-L1 expression [24]. In SCCUPs checkpoint inhibitors are not routinely administered and the expression of PD-L1 has not been evaluated [25].
Therefore, the aim of this retrospective study was to compare SCCUPs of the head and neck and oropharyngeal carcinomas with regard to the expression of PD-L1 and p16 and the composition of the tumor microenvironment including CD3-and CD8-positive lymphocytes. This could help finding a rationale for targeted therapy of SCCUPs of the head and neck and ultimately improve prognosis of this group of patients.
Patient cohort
In this study a total of 97 patients were included, among them 47 cases with primary oropharyngeal squamous cell carcinomas and 50 cases with lymph node metastases of cancer of unknown primary (CUP). Patients were treated in the Department of Otorhinolaryngology/Head and Neck Surgery, Klinikum rechts der Isar, Technical University of Munich between November 2001 and September 2013. From all patients formalin fixed and paraffin embedded (FFPE) material of resection specimen were obtained from the Institute of Pathology, Technical University Munich, Germany. Clinical and pathological data were gathered in retrospective using patients' files and electronic records. All patients were followed up regularly in the university hospital. The period of time between initial diagnosis and time of death/last follow-up was used for calculating overall survival. The age of patients in this study ranged from 37.12 to 83.11 years (Median age 60.89 ± 11.9 years, Average age 60.81 ± 9.8 years). The average follow-up period was 4.59 ± 2.6 years (Median 4.76 years ± 2.7) for Overall Survival (OS) and 2.85 years ± 2.8 years (Median 2.03 ± 3.7 years) for Progression free Survival (PFS). Ethical approval was obtained from the ethics committee of the Technical University of Munich (reference number 474/18S).
Immunohistochemistry of PD-L1, CD3, CD8
FFPE tissue was cut with the microm HM 355 S (International GmbH, Walldorf, Germany) into 2-3 μm thick sections and deparaffinized at 65 °C. Immunohistochemical staining was conducted with a VENTANA BenchMark GX with PD-L1 primary antibody (US Biological, Salem, MA, USA), CD3 primary antibody (1:400, Cell Marque, Rocklin, CA, USA), CD8 primary antibody (1:25, Thermo Fisher Scientific, IL, USA), p16 primary antibody (US Biological, Salem, MA, USA). Slides were counterstained with hematoxylin. After dehydration by immersion in the ethanol series and xylol (2 min each) the slides were examined under light microscopy by two independent researchers. 3 high power fields of tumor in each section were chosen. Tonsil tissue was used as control. Exemplary images of the PD-L1 staining are depicted in Supplementary Fig. 2. The Tumor Proportion Score (TPS) assessed the membranous expression of PD-L1 in tumor cells. Sections with a percentage of expression ≥ 1% were designated as PD-L1 positive, since this threshold was used in the KEYNOTE-012 study of PD-1 inhibition in HNSCC [21]. Tissue with a strong and diffuse nuclear and cytoplasmic p16 staining of more than 70% of tumor cells were considered p16 positive. The area of tumor covered by CD3-and CD8positive tumor-infiltrating lymphocytes was evaluated in five high-power fields to estimate the impact of the tumor microenvironment on the prognosis of CUP and OPSCC patients. The median number of TILs was used to categorize HNSCC and CUP into high and low infiltration.
To obtain definitive evidence of HPV infection DNA in situ-hybridization (ISH) was applied. The HPV viral types 16,18,31,33,35,45,52,56,58, and 66 were detected with the Inform HPV III Family 16 Probe kit (Ventana Medical Systems, AR, USA) according to the manufacturer's instructions. HPV ISH was interpreted positive when nuclear staining was detected in the infected tumor cells. In addition, the expression of EBV RNA was detected in squamous cell CUPs with the Inform EBER early RNA kit (Ventana Medical Systems, AR, USA) according to the manufacturer's instructions. A threshold of greater than 30% was considered positive.
Statistical Analysis
Kaplan-Meier survival analysis and log-rank testing was used to compare survival rates for different patient groups and clinical characteristics. Associations were tested with Fisher`s exact test and Bonferroni correction. At p < 0.05 the null hypothesis was rejected, and the result was considered statistically significant. Statistical calculation was performed using Prism 8 (GraphPad Software, La Jolla, CA, USA).
Correlation between PD-L1 and pathological data
For the analysis, patients were grouped into PD-L1 positive and negative according to TC score. For the further analysis a threshold of 1% positive stained cells was chosen. The results of staining intensity in percent are depicted in Table 1a. The clinical and pathological characteristics of the cohort are depicted in Table 1b.
Using the TC Score with a cutoff of > 1% positive stained tumor cells, 15 of 47 (31.9%) oropharyngeal squamous cell tumor sections were PD-L1 positive. 34 of 50 (68.0%) CUPs were PD-L1 positive. Exemplary images of negative and positive expression are shown in the Supplementary Fig. 1. Association of CUPs demonstrating a higher expression of PD-L1 than OPSCC was tested with Fisher's exact test and proved to be significant (p = 0.0005). The expression of p16 also differentiated CUPs and OPSCC, with 53.2% of OPSCC being p16-positive compared to 28.2% of CUPs (p = 0.0018). High risk HPV DNA detection with in situ hybridization revealed a high concordance with p16 IHC staining (p = 0.011). An association between HPV-ISH and PD-L1 expression in p16-CUPs was not detected (p > 0.9999). p16-negative OPSCC were more likely to be PD-L1 negative (p = 0.0146). In only 2 out of 50 (4%) squamous cell CUPs EBV RNA could be detected.
Overall survival (OS) of OPSCC was significantly better than OS of CUPs (p = 0.0003) (Supplementary Fig. 1). High expression of PD-L1 in CUP patients was not associated with OS (p = 0.3107) or PFS (p = 0.2249) ( Supplementary Fig. 1). Since the expression of p16 is the most important prognostic factor in HNSCC, the HNSCC and SCCUP cases were stratified into p16-positive and p16-negative SCCUPs. In p16-negative CUPs a high expression of PD-L1 was significantly associated with better OS (p = 0.0080) and PFS (p = 0.0002) (Fig. 1). There was no significant association between p16-positive CUPs and OS or PFS (p = 0.617, p = 0.8572) (Supplementary Fig. 1).
The immune cell infiltration of CUPs and OPSCC was measured by evaluating the percentage of positive stained CD3-and CD8-positive tumor infiltrating lymphocytes (TILs) covering the tumor area in 5 high power fields. A high infiltration of CD3-positive lymphocytes was not associated with a significantly better OS (p = 0.6806) ( Supplementary Fig. 1) in CUP patients. The infiltration of CD8-positive lymphocytes did not differentiate the prognosis of CUP patients significantly (p = 0.7834) (Supplementary Fig. 1). The average tumor area infiltrated by CD3-positive TILs was greater in CUP than in OPSCC (6.2% vs 10.4%). This could also be observed with CD8-positive TILs (3.0% vs 6.3%).
Discussion
In this study, the PD-L1 expression of SCCUPs and OPSCC was compared for the first time. In a second step immune infiltration was assessed to characterize CUPs and get a better picture of underlying immunological mechanisms in this heterogeneous cancer. CUP tissue demonstrated a significantly higher PD-L1 expression (TC) and a stronger infiltration with lymphocytes in this study. This is important, as conventional therapy has high recurrence rates with considerable side effects, whereas the new group of checkpoint inhibitors had remarkable success in treatment of HNSCC and various solid tumors [26][27][28]. The expression of PD-L1 is the key to drug-induced inhibition of PD-L1, effectively hindering the tumor from inhibiting the immune system. A systematic review and meta-analysis of efficacy and safety in metastatic cancer of solid primaries identified PD-L1 inhibitors as a preferable treatment option. The treatment was even more advantageous in male patients, patients under the age of 65, and current or former smokers [29]. Since the peak incidence of CUP is between 60 and 64 years, in contrast to multiple other cancer entities with higher peak incidences between 80 and 84 years [30], and smoking and alcohol consumption further increases the risk of SCCUP [31], checkpoint therapy could improve the outcome in SCCUP patients. However, a clear rationale for a study focusing on checkpoint therapy in SCCUPs is still missing.
In a former study, PD-L1 expression has been evaluated in the large group of general CUPs (including cases outside of head and neck area) and a positive expression has been detected in up to 28% of CUPs using immunohistochemistry [32]. In that analysis, however, only 30 cases (8% of total cases) represented squamous cell carcinoma. In the mentioned study, a cut off for PD-L1 positivity of ≥ 5% and a different antibody was used. This study analyzes to the knowledge of the authors the largest group of head and neck squamous cell CUPs for PD-L1 expression [32]. Since the expression of PD-L1 is predictive for successful checkpoint inhibitor therapy [33], based on our data, substantial proportion of SCCUP patients could potentially benefit from treatment with checkpoint inhibitors. This is even more relevant for the group of p16-negative patients, which is associated with a worse prognosis [34]. In this study PD-L1 expression differentiated p16-negative patients into a high and low risk group. This finding lays the foundation for a prospective study analyzing PD-L1 expression in head and neck squamous cell CUPs and treatment according to PD-L1 status. For the bigger group of general CUPs, a first study investigating the effect of targeted therapy, including the PD-L1 inhibitor Atezolizumab, has already been started. In this Phase II randomized clinical trial (NCT03498521) unfortunately SCCUP are excluded [16,35]. This study demonstrates the high expression of PD-L1 in CUP tissue, giving hope that immune checkpoint inhibition could improve survival in at least a relevant proportion of p16negative squamous cell CUP patients.
To further decipher the tumorigenesis of SCCUP, p16 expression was analyzed. Of all the head and neck malignancies, HPV infection is most common in OPSCC and is, therefore, an important hint toward an oropharyngeal origin of a SCCUP [12,16]. Some reports showed up to 90% of squamous cell CUPs being p16 positive [36,37]. This finding could not be replicated by this study, with only 22% of CUPs, and 53% of OPSCC being p16 positive. A potential explanation for this lies in epidemiological trends regarding the percentage of population infected with HPV. In Germany, the number of OPSCC associated with HPV infection increased from 11.5% between 1988 and 2008 to 55.0% between 2004 and 2009 [38,39]. This finding of high HPV-attributable fractions is important for two main reasons. First, this suggests similarities in the process of carcinogenesis of SCCUPs and OPSCC. In case of SCCUP being a regressed primary with only microscopic remnants, this leads to the oropharynx as the most likely region of origin. Being able to identify the tissue of origin also enables more therapeutic options, including the use of targeted radiation therapy. Radiation of the pharynx and cervical lymph nodes with concurrent chemotherapy led to good regional control in patients with cervical lymph node metastasis from an unknown primary site [40,41]. Similar to previous studies, the results of this study demonstrated a significantly better survival of patients with HPV positive SCCUPs and OPSCC. While HPV infection is associated with better survival in HNSCC, HPV-negative tumors are more difficult to treat and show a high rate of recurrence [42]. HPV-associated oropharyngeal squamous cell carcinoma are more sensitive to chemo-and radiotherapy and demonstrate a higher recurrence-free and overallsurvival [34]. In KEYNOTE-012, immunotherapy with Pembrolizumab led to a response of 32% in HPV-positive, and 18% in HPV-negative recurrent or metastatic HNSCC [21,43]. HPV-negative HNSCC and SCCUPs until this day remain difficult cancer entities to treat, highlighting the need for novel prognostic tools and therapies.
To characterize SCCUPs in more depth and highlight potential similarities and differences between SCCUP and OPSCC, the composition of the tumor microenvironment was analyzed. Investigating the infiltration of CUPs with TILs revealed a strong infiltration of SCCUPs with immune cells compared to OPSCC. An association of immune cell infiltration with better survival could not be detected. This difference in immunological properties may also contribute toward the rationale of a more targeted therapy with checkpoint inhibitors in the future, since the tumor microenvironment, and more specifically, the infiltration with lymphocytes was associated with the response to checkpoint inhibitor therapy in certain types of cancer [44,45]. In a recent publication, the therapeutic effect of checkpoint inhibitor therapy could even be predicted by observing the levels of CD8/CD68/CD163/PD-L1 positive cells in non-small cell lung cancer [46]. A prospective study of SCCUP treatment with checkpoint inhibitors should, therefore, take the composition of the tumor microenvironment into account.
Conclusions
In this study, for the first time, a higher expression of PD-L1 and strong lymphocyte infiltration was observed in SCCUPs compared to OPSCC. These findings support a prospective study evaluating the success of immune therapy in cancer of unknown primary patients of the head and neck.
Ethical statement
The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the ethics committee of the Technical University of Munich (reference number 474/18S).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. | 2022-12-30T16:07:18.894Z | 2022-12-28T00:00:00.000 | {
"year": 2022,
"sha1": "06507a25ec1699cdf00db36ad17510ad3b426fbd",
"oa_license": "CCBY",
"oa_url": "https://link.springer.com/content/pdf/10.1007/s00405-022-07775-z.pdf",
"oa_status": "HYBRID",
"pdf_src": "Springer",
"pdf_hash": "24f84fbb2458e179656c76fe905043360cadc081",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
235377292 | pes2o/s2orc | v3-fos-license | LB4OMP: A Dynamic Load Balancing Library for Multithreaded Applications
Exascale computing systems will exhibit high degrees of hierarchical parallelism, with thousands of computing nodes and hundreds of cores per node. Efficiently exploiting hierarchical parallelism is challenging due to load imbalance that arises at multiple levels. OpenMP is the most widely-used standard for expressing and exploiting the ever-increasing node-level parallelism. The scheduling options in OpenMP are insufficient to address the load imbalance that arises during the execution of multithreaded applications. The limited scheduling options in OpenMP hinder research on novel scheduling techniques which require comparison with others from the literature. This work introduces LB4OMP, an open-source dynamic load balancing library that implements successful scheduling algorithms from the literature. LB4OMP is a research infrastructure designed to spur and support present and future scheduling research, for the benefit of multithreaded applications performance. Through an extensive performance analysis campaign, we assess the effectiveness and demystify the performance of all loop scheduling techniques in the library. We show that, for numerous applications-systems pairs, the scheduling techniques in LB4OMP outperform the scheduling options in OpenMP. Node-level load balancing using LB4OMP leads to reduced cross-node load imbalance and to improved MPI+OpenMP applications performance, which is critical for Exascale computing.
Introduction
On the road to Exascale, we observe that modern and future high performance computing (HPC) systems combine an increasing number of computing nodes and, in particular, cores per node. For example, the top 5 systems on the Top500 list 1 contain thousands of nodes and tens to hundreds of cores per node.
Recent reports 2 indicate that the next update of the Sunway TaihuLight system will include 520 cores per node, double that of its predecessor. Such hardware parallelism increase leads to the challenge of exposing and expressing corresponding degrees of hierarchical parallelism in software to efficiently exploit the hierarchical hardware parallelism.
Load imbalance is a significant performance degradation factor in computationally-intensive applications [1] [2], defined as processors idling while there exist units of computation ready to be executed that no processor has started. This results in uneven execution progress among the parallel processing units, which can emerge from numerous application-, algorithm-, and/or systemic characteristics. Computationally-intensive applications often represent irregular workloads (e.g., due to boundary conditions, convergence, conditions, and branches). Computing systems may consist of heterogeneous processors and may be affected by nonuniform memory access (NUMA) times, operating system noise, and contention due to sharing of resources. Load imbalance can be mitigated by an efficient dynamic scheduling of computation units onto processing units. Finding optimal schedules is NP-hard [3]. Therefore, many scheduling heuristics have been proposed over the years [4] [5].
Scheduling and load balancing that exploit multiple levels of hardware parallelism across and within computing nodes are critical challenges for the upcoming Exascale systems [6] [7]. Dynamic self-scheduling explicitly addresses application-and system-induced performance variations while minimizing load imbalance and scheduling overhead [8] [9] [10].
It has been recently shown that thread-level load imbalance has a significant impact on the performance of hybrid MPI+OpenMP applications [11]. OpenMP is the most widely-used standard for expressing and exploiting node-level parallelism. The OpenMP standard specifies three loop schedule kinds: static, dynamic, and guided. These scheduling options limit the highest achievable performance as they do not cover the broad spectrum of applications and systems characteristics [12] [13] [14]. Furthermore, the absence of a comparative implementation of the multitude of scheduling techniques from the literature hinders research on novel scheduling techniques which typically requires comparison with the scheduling state of the art.
This work builds on recent work on multilevel load balancing [11] by concentrating on thread-level scheduling and deepening the analysis of its performance impact for multithreaded applications executing on hierarchical parallel systems. Specifically, this work provides a broad range of dynamic loop self-scheduling (DLS) techniques, implemented in a unified OpenMP runtime library (RTL), called LB4OMP 3 that can readily be used for MPI+OpenMP applications. Aiming for a wide reach and broad impact, we implemented LB4OMP as an extension of LLVM's OpenMP RTL given its widespread use (e.g., in the US DOE's Exascale Computing Project project [15]), open-source nature, and high compatibility with widely-used compilers (Intel, IBM, PGI, GNU).
The LB4OMP library supports 14 carefully selected dynamic (and adaptive) loop self-scheduling techniques, ready to use in addition to those existing in the standard-compliant OpenMP libraries. Applications using LB4OMP benefit from improved performance due to the portfolio of DLS techniques, of which certain adapt during execution to unpredictable variations in application and systemic characteristics (see Section 3.1). These 14 techniques are selected to cover a broad spectrum of dynamic (and adaptive) scheduling techniques. Specifically, LB4OMP provides: • Five dynamic but non-adaptive self-scheduling techniques: fixed size chunking (FSC) [16], factoring (FAC) [17], the practical variant of factoring (FAC2), tapering (TAP) [18], and the practical variant of weighted factoring (WF2) [19]; • Seven dynamic and adaptive self-scheduling techniques: BOLD [20], adaptive weighted factoring (AWF), its variants (AWF-B,C,D,E) [21], and adaptive factoring (AF) [8]; • Features for performance measurement to measure loop-specific performance metrics for an in-depth analysis of loop scheduling and load balancing.
LB4OMP also provides mFAC and mAF, two improved implementations that reduce the overhead of FAC and AF. The scheduling techniques in LB4OMP differ in the amount of work assigned to a thread at a time, referred to as a chunk of loop iterations. Specifically techniques with: (1) simple chunk calculation include FSC, FAC2, and WF2; (2) profiling-based chunk calculation include FAC, mFAC, and TAP; and (3) adaptive (nonlinear) chunk calculation include BOLD, AWF, its variants AWF-B,C,D,E, and AF and mAF. This work makes the following contributions: 1. A novel systematic and unified implementation of 14 dynamic (and adaptive) scheduling techniques.
2. Advanced features for performance measurement of loop performance and loop-level load imbalance.
3. An in-depth analysis of the performance potential and limitations of the standard and newly implemented scheduling techniques, which were so far only partially known to the non-experts and/or scheduling practitioners.
The novelty of this work lies in providing a standalone and unified implementation of efficient scheduling techniques from literature, which is needed to spur new research in scheduling and load balancing for Exascale systems.
Prior to this work, scheduling research was hindered by the absence of an environment that supports a fair comparison with the existing scheduling algorithms in the literature. Novel scheduling techniques or improved versions of the techniques implemented in this work can now be implemented and compared in this unified testbed. LB4OMP enables and promotes research on automatic selection methods to identify, during execution, the highest performing scheduling technique for a given application-loop-time-step configuration. This work is significant by bridging the gap between the state-of-the-art and the state-of-the-practice of load balancing in multithreaded applications. This will allow the large degrees of heterogeneous node-level parallelism in today's pre-and upcoming Exascale systems to be efficiently exploited for improving applications performance.
This work is organized as follows. Section 2 is a review of the related literature highlighting the differences between prior and the present work. Section 3 describes the LB4OMP design and highlights the required extensions to LLVM's OpenMP RTL. The use of the newly implemented scheduling techniques in OpenMP applications via the LB4OMP library is detailed in Section 3. The experimental design and performance analysis campaign are presented and discussed in Section 4. The work is concluded in Section 5, which also outlines directions for future work.
Related Work
The performance potential of a small number of dynamic and non-adaptive loop scheduling techniques (TSS, FAC2, WF2, and RAND), implemented in the GNU OpenMP RTL, was recently explored [13]. The authors showed cases when applications achieve improved performance beyond the one offered by the scheduling techniques supported in the GNU OpenMP RTL. Another variant of FAC, called BO FSS, was proposed and compared against STATIC, GSS, TSS, FAC2, TAP [18], HSS [22], and BinLPT [23], in another extended implementation in the GNU OpenMP RTL [24]. The scheduling techniques considered in these research efforts does not consider dynamic and adaptive scheduling techniques. In general, other efforts only considered extending the GNU OpenMP RTL, which is not compatible with other compilers, unlike the LLVM OpenMP RTL.
LLVM has gained considerable traction in the software vendor community and improving the open source LLVM compiler and runtime ecosystem is a priority for the US DOE Exascale Computing Project [15]. The LLVM OpenMP RTL was extended only by an implementation of FAC2 [14]. Experiments therein showed improved performance of certain workloads with the newly added DLS technique.
The present work improves over the previous related work by (1) providing a novel systematic and unified implementation of a broader range of dy-namic (and adaptive) scheduling techniques. (2) providing advanced features for performance measurement of loop performance and loop-level load imbalance.
(3) demystifying the performance potential and limitations of the standard and the newly implemented scheduling techniques through an in-depth performance analysis campaign.
Another direction of related work includes efforts that propose generic interfaces to allow users to implement their own loop scheduling techniques in different runtime libraries [25] [26] [27]. These efforts reduce the development challenges associated with the direct modification to the RTL source codes, i.e., developers can implement their scheduling technique via simplified, and ideally, well-documented interfaces. However, these efforts do not exclude the need for extensive scheduling libraries to validate novel scheduling techniques and exploit the increased hardware parallelism of modern HPC systems. Therefore, such efforts [25][26][27] can be seen as potential methods that facilitate the development of another version of the LB4OMP scheduling library in the future.
Considering the vast amount of DLS techniques proposed in the literature, the following non-trivial question arises: What are the criteria to include a particular scheduling technique into a unified scheduling library? A number of research efforts attempted to answer this question [28] [17] [18] [20] [8] [21].
FSC [29], TSS [28], FAC [17], and TAP [18] were introduced by separate research groups. However, the performance of each of these techniques was compared against at least one of three main scheduling techniques STATIC, SS [30], and GSS [31] that nowadays correspond to schedule(static), schedule(dynamic,1), and schedule(guided,1) specified in the OpenMP standard. The chunk calculation in TSS, for instance, is simpler than the non-linear chunk calculation in GSS. Therefore, we state that the simplicity of chunk calculation is an important selection criterion.
FSC, TAP, and FAC are based on probabilistic analyses and use profiling information to calculate the chunk sizes that achieve the most balanced load execution for a given application with a high probability. The profiling information is obtained prior to the applications' execution. We state that the profiling-based chunk calculation is another important selection criterion.
Another set of related research efforts includes BOLD [20], AWF-B,C,D,E [21], and AF [8]. These techniques use profiling information obtained during applications' execution to adapt during execution the calculated chunk sizes to achieve the most balanced load execution for a given application. We state that the adaptivity of chunk calculation is another valuable selection criterion.
Based on the above criteria, FSC [29], FAC [17], TAP [18], WF [19], BOLD [20], AWF-B,C,D,E [21], and AF [8] were selected for implementation into the LB4OMP scheduling library. Other scheduling techniques that meet these criteria can also be considered for inclusion in LB4OMP. The DLS techniques selected in this work can also be applied to schedule OpenMP tasks and taskloops [32] [33]. The use of LB4OMP for scheduling tasks and taskloops is part of separate ongoing work by the authors. 5 3 The LB4OMP Library LB4OMP extends the LLVM OpenMP RTL version 8.0, which is widely used and compatible with various compilers, including Intel, IBM, GNU, and PGI. The choice of extending the LLVM OpenMP RTL meets important research goals and priorities of the HPC community [15]. Figure 1 shows the LB4OMP loop scheduling mechanism which extends the scheduling mechanism in the LLVM OpenMP RTL. The three main functions responsible for the chunk calculation are implemented in the file kmp dispatch.cpp.
Upon initialization, each thread calls the kmp dispatch init algorithm function inside the kmp dispatch.cpp file (init in Figure 1). This function then initializes the needed structures for the selected scheduling technique and calls kmp dispatch next algorithm (next in Figure 1).
The logic of the chunk calculation of all DLS techniques is implemented in the kmp dispatch next algorithm function. The kmp dispatch next algorithm is called each time a thread needs to obtain work.
Since the threads obtain work from a shared queue, kmp dispatch next algorithm relies on different synchronization operations (sync in Figure 1) depending on the scheduling technique in execution. Finally, the threads call kmp dispatch finish (finish in Figure 1) to reset variables or free allocated memory. What is the connection between kmp.h, … kmp_global.cpp and the black arrow coming out of the work sharing loops?
The grey boxes look like unconnected files. Why should they be in this picture? What is their role?
scheduling techniques and features for performance measurement
Chunk of loop iterations
Loop iteration Figure 1: Extension of the OpenMP LLVM RTL scheduling process for worksharing loops with LB4OMP.
Significance of chunk parameter. The OpenMP standard scheduling techniques and the newly implemented scheduling techniques in LB4OMP support the declaration of a chunk parameter which bears different meanings among the scheduling techniques. For schedule(static, chunk) and schedule(dynamic,chunk), the chunk parameter denotes the amount of iterations that the threads should receive for every work request. For the other techniques, the chunk parameter works as a threshold, in the sense that when chunks sizes, calculated by a scheduling technique, fall below this threshold they will be replaced by a chunk sizes equal to the size of the chunk parameter. The chunk parameter was introduced by the OpenMP standard to minimize the scheduling overhead and to improve data locality. Declaring a proper chunk parameter improves performance since threads perform fewer scheduling rounds than without this threshold. This is confirmed by the experiments described in Section 4.
Dynamic Loop Scheduling Techniques
LB4OMP bridges the gap between the literature and the practice in dynamic load balancing of multithreaded applications. It represents an environment for a fair comparison of scheduling techniques and lays the ground for future research in loop scheduling of multithreaded applications. The loop scheduling techniques implemented in LB4OMP are dynamic and a number of them are also adaptive self-scheduling techniques. With self-scheduling techniques, free threads request, calculate, and obtain their own next chunk of units of work (loop iterations) by accessing a central shared work queue containing all iterations of a given loop. The chunk size is calculated according to the loop scheduling techniques in the OpenMP RTL.
Following is a brief description of each scheduling technique in LB4OMP, starting with dynamic and non-adaptive self-scheduling techniques followed by the dynamic and adaptive self-scheduling techniques. More details about the various chunk calculations for these techniques can be found in the literature [34]. Dynamic and non-adaptive self-scheduling. SS (or dynamic,1 in OpenMP) [30] is a dynamic self-scheduling technique wherein the chunk size is always one loop iteration. SS incurs the highest scheduling overhead due to the largest number of chunks (equal to the number of loop iterations). SS can achieve a highly load-balanced execution in highly irregular execution environments.
FSC [16] determines an optimal chunk size that achieves a balanced execution of loop iterations with the smallest overhead. To calculate the optimal chunk size, FSC requires that the variability in iteration execution times and the scheduling overhead of assigning loop iterations are known before applications' execution.
GSS [31] is a trades-off between the load balancing achievable with SS and the low scheduling overhead incurred by STATIC. Unlike FSC, GSS assigns decreasing chunk sizes to balance the loop execution progress among all threads. For every work request, GSS assigns a chunk size equal to the number of remaining loop iterations divided by the total number of threads.
TAP [18] is based on a probabilistic analysis that represents a general case of GSS. It considers the average of loop iteration execution times µ and the standard deviation σ to achieve a higher load balance than GSS.
TSS [28] assigns decreasing chunk sizes similar to GSS. However, TSS uses a linear function to decrement chunk sizes. This linearity results in lower scheduling overhead in each scheduling step compared to GSS.
FAC [17] schedules the loop iterations in batches of equally-sized chunks. FAC evolved from comprehensive probabilistic analyses, and assumes prior knowledge about the average iteration execution times (µ) and their standard deviation (σ). A practical implementation of FAC, denoted FAC2, assigns half of the remaining loop iterations for every batch. The initial chunk size of FAC2 is half of the initial chunk size of GSS. If more time-consuming loop iterations are at the beginning of the loop, FAC2 is expected to better balance their execution than GSS.
WF [19] is similar to FAC, with the difference that each processing unit executes variably-sized chunks of a given batch according to its relative processing weights. The processing weights, W pj , are determined prior to applications' execution and remain constant during execution. WF2 is the practical implementation of WF that is based on FAC2.
mFAC is our improved implementation of FAC. In the original FAC algorithm, the first thread that starts a new batch of iterations locks a mutex and computes the chunk size for the current batch. The subsequent threads simply read and reuse the already computed chunk size until the iterations in the batch have been scheduled. This requires mutex-based synchronization. mFAC avoids such costly synchronization by involving more computation. Specifically, in mFAC, a shared counter is atomically incremented so that threads identify the current batch number. Hence, each thread calculates its own next chunk size depending on the batch counter. Depending on the synchronization and computation overheads in a given computing systems, one may use either FAC or mFAC. Figure 2 depicts an example of the chunk sizes calculated by the dynamic and non-adaptive scheduling techniques and their progression over the work requests for scheduling the iterations of the main loop (L1) of SPHYNX 4 (more details in Section 4.1). STATIC and SS are not shown in Figure 2 as their chunk size progression is constant (straight line), at the size of the chosen chunk parameter. The chunk size progression for the dynamic non-adaptive scheduling techniques follows a decreasing chunk size pattern, wherein the next chunk of iterations is equal to or smaller than the previous. One can note that not only the chunk sizes but also the total the number of chunks allocated varies among the scheduling techniques. A small number of large chunk sizes may not mitigate load imbalance but incur a smaller scheduling overhead due to fewer scheduling operations while a greater number of smaller chunks may improve load balancing at the cost of increased scheduling overhead. Dynamic and adaptive self-scheduling. Adaptive scheduling techniques regularly measure execution performance during the application execution and the scheduling decisions are taken based on this information. The adaptive scheduling techniques incur a higher scheduling overhead compared to non-adaptive techniques but are designed to outperform the non-adaptive ones in highly irregular execution environments.
BOLD [20] is a 'bolder' version of FAC and a further development of TAP. As with TAP, it uses the mean µ and the standard deviation σ of the iteration execution times as well as an estimate of the scheduling overhead. The driving idea behind the BOLD strategy was to increase early chunk sizes such that scheduling overhead is reduced while considering the risk of potentially too large chunks of iterations.
AWF [21] is similar to WF in that each thread executes variably-sized chunks of a given batch according to its relative processing weight. The processing weight is updated during execution based on the performance of each thread. AWF is devised for time-stepping applications and threads processing weights are updated at the end of each time-step. Variants of AWF, namely AWF-B and AWF-C, relaxed this constraint by updating processing weights at the end of every batch and chunk execution, respectively. Additional variants of AWF, namely AWF-E and AWF-D, are similar to AWF-B and AWF-C, respectively. In addition, AWF-E and AWF-D take into account the overhead of scheduling in calculating the relative processing weights.
AF [8] is an adaptive DLS technique derived from FAC. In contrast to FAC, AF learns both µ and σ for each computing resource during application execution to ensure full adaptability to all factors that cause load imbalance. AF adapts the chunk size during application execution based on the continuous updates of the mean loop iteration execution times µ and their standard deviation σ.
mAF is our improved implementation of AF. In the original AF algorithm, the execution time of earlier loop iterations (from the same execution) are collected to calculate the next chunk size. The collected times only consider the execution time of the loop iterations themselves. In LB4OMP, mAF also considers the scheduling overhead. Hence, mAF employs a more precise performance estimation for calculating the chunk size, which is expected to lead to higher load balance and performance. Figure 3 shows the chunk sizes calculated by the dynamic and adaptive scheduling techniques and their progression over the work requests for scheduling the iterations of the main loop (L1) of SPHYNX (more details in Section 4.1). The chunk sizes calculated by the adaptive scheduling techniques do not strictly decrease with each work request, but increase or decrease depending on the requesting thread's performance during execution. This is the quintessence of adaptive self-scheduling and load balancing: threads which require more time to compute receive less work while threads which compute faster receive more work. Dynamic and adaptive self-scheduling and load balancing techniques will be critical for achieving performance on upcoming Exascale systems with heterogeneous architectures in which scheduling needs to dynamically adapt to threads executing on slower or faster processing units. The observation from Figure 2 also holds true for Figure 3 regarding the trade-off between fewer and larger chunks and more and smaller chunks.
Features for Performance Measurement
LB4OMP provides a number of features for performance measurements for target loops associated with the OpenMP schedule clause. These features are crucial for the analysis of loop scheduling and load balancing.
Thread execution time. This feature reports the threads execution times per loop execution instance. This information is important for the estimation of load imbalance in a parallel loop. This feature is enabled by defining the environment variable KMP TIME LOOPS and declaring the path where the measured performance data will be stored.
The thread execution time feature enables the calculation of well-known load imbalance metrics such as coefficient of variation (c.o.v.) [17] and percent imbalance (p.i.) [35]. The c.o.v. and p.i. equations are defined in Table 1, where T loop par denotes the parallel execution time of the loop. In this work, these metrics are calculated based on the data measured with LB4OMP for individual OpenMP loops and presented later in Section 4.
Chunk information. LB4OMP collects and stores the calculated chunk sizes for each thread in each scheduling round. This functionality can be enabled by setting the environment variable KMP PRINT CHUNKS to 1. The collected information is stored at the location defined in KMP TIME LOOPS (see above).
A detailed analysis of the chunk sizes calculated by each scheduling technique for given OpenMP loops is fundamental for understanding their performance. We used this feature for the in-depth performance analysis of the impact of the chunk parameter on all scheduling techniques described in Sections 4.3 and 4.4.
Statistical information about loop iterations execution times. LB4OMP provides a profiling feature that collects the mean of loop iterations execution times (µ) and their standard deviation σ. These measurements are required by the FSC, FAC, TAP, and BOLD scheduling techniques. The profiling feature relieves the (non-expert) user from the burden of collecting such profiling information.
This feature can be enabled by defining schedule(runtime) in the target loop, exporting OMP SCHEDULE=profiling, and setting the environment variable KMP PROFILE DATA to the path where the profiling data will be stored.
Load Balancing Applications with LB4OMP
The use of LB4OMP with an OpenMP application is straightforward and illustrated in Figure 4. First, one must ensure that the target OpenMP loops in the application contain the schedule(runtime) clause. If that is the case, no other changes are required and there is no need to recompile the code. Otherwise, one needs to change (or add, if the loop structure permits) the existing scheduling clause to runtime in all target loops and recompile the application. Next, one needs to add the path to the compiled LB4OMP to the environment variable that the linker uses to load dynamic and shared libraries (e.g., LD LIBRARY PATH on Linux/Unix systems). The workflow in Figure 4 is almost independent of the target system. The only system-related parameter required by LB4OMP is the host CPU clock frequency. This is passed to LB4OMP via the environment variable KMP CPU SPEED as an integer variable in MHz. The adaptive scheduling techniques in LB4OMP use low overhead cycle counters as RDTSCP to measure the execution time of previous chunks of iterations. We use the clock frequency to convert the cycles into time (which are the values expected by the formulas of those techniques). One may argue that the static value defined by the user in KMP CPU SPEED may be inaccurate for a number of modern processors that allow dynamic clock frequency change during execution. The measured performance of the threads is relative to each other, so variations in clock frequency during execution do not affect the chunk calculation. Work is ongoing to automate the process of collecting the clock speed for use in LB4OMP.
Applications with multiple loops may need to employ different scheduling techniques in different parts of the code. LB4OMP uses the schedule(runtime) option available in OpenMP and, therefore, the scheduling technique selected by the user is read from the environment variable OMP SCHEDULE. To select different scheduling techniques for an application in different parts of the code, one can use the function specified in the OpenMP standard omp set schedule(omp sched t kind, int chunk size) 5 . This is used to update the scheduling technique specified with OMP SCHEDULE during execution.
One can also export environment variables directly from the application code to update the configuration of LB4OMP and, if preferred, to update the scheduling technique itself.
Performance Results and Discussion
We use three applications, two microbenchmarks, and three computing node types to evaluate the performance of the existing in LLVM OpenMP RTL and newly implemented DLS techniques in LB4OMP (see Table 1).
The applications 352.nab 6 , SPHYNX, and the microbenchmark DIST 7 were selected since they contain imbalanced and computationally-intensive loops, which are the most promising optimization targets for improved performance with dynamic (and adaptive) self-scheduling techniques [13]. The application GROMACS [36] and the microbenchmark STREAM 8 were selected to provide an overview of the scheduling overhead, ccNUMA effects, and locality loss incurred by the scheduling techniques in LB4OMP. These loops are balanced, of low arithmetic intensity, and mainly perform memory operations which stresses the possibly negative effects of dynamic (and adaptive) self-scheduling.
We will use the following notation to specify details regarding the applications and systems. T denotes number of time-steps, L the IDs of loop with modified schedule clauses, and T ol the time spent by the application outside of loops. The loops for which we modify the schedule clause were parallel and not nested 9 . 10 . The STREAM microbenchmark was executed with its default array size of 80, 000, 000 elements (doubles), memory per array 610.4 MB, which required a total of 1831.1 MB in memory. DIST is synthetic microbenchmark used to show how the scheduling techniques react to different statistical loop workload distributions across iterations. Each loop of DIST follows a different workload distribution as indicated in Table 1.
Design of Factorial Experiments
Each experiment was repeated 5 times (STREAM was repeated 20 times) and the average execution time or memory bandwidth (in MB/s) for STREAM is reported.
The applications 352.nab, SPHYNX, and GROMACS are time-stepping simulations. The computationally-intensive loops with modified schedule clause from each application and microbenchmark are indicated in Table 1. The applications and the LB4OMP library were compiled with the Intel compiler version 19.0.1.144. The characteristics of the computing systems are also indicated in Table 1.
Throughout the performance analysis campaign, all scheduling techniques used the default chunk parameter (1 loop iteration). We used the thread execution time LB4OMP feature (Section 3.2) and measured the loop execution and threads finishing times per loop to derive load imbalance.
Performance Analysis
The goal of this performance analysis campaign is to examine which DLS technique provides the highest performance and lowest load imbalance for each application's loop scheduled with LB4OMP. Figure 4.1 shows the average parallel execution time for each modified loop in 352.nab, SPHYNX, and DIST with the default chunk parameter executing on all node types without hyperthreading. In Figure 4.1, the Best combination of scheduling techniques varies greatly between applications and systems, and outperforms every single technique in most cases. This reinforces the need for additional scheduling options in OpenMP [13] since the Best combination commonly includes the techniques implemented in LB4OMP.
From the results of the dynamic and non-adaptive scheduling techniques in Figure 4 The dynamic and adaptive loop scheduling techniques naturally add overhead. However, they adapt to application and system variations and heterogeneity without requiring profiling information. In Figure 4.1, we observe that, except for BOLD (and for AF, mAF for 352.nab and DIST on miniHPC-KNL), all adaptive scheduling techniques consistently achieved high performance. In numerous cases, the adaptive scheduling techniques are included in the Best combination. For example, AF and mAF, consistently presented high performance for all results with SPHYNX, in which mAF is included in the Best combination for both miniHPC-KNL and Piz Daint-Haswell nodes. The highest achieved performance improvement with a dynamic and adaptive scheduling technique was on SPHYNX with mAF on Piz Daint-Haswell outperforming GSS, the best standard technique in this case, by 10.67%.
The results for SPHYNX on node miniHPC-Broadwell show that AF and mAF reasonably outperformed GSS by approximately 9.28% and 9.59% respectively. SPHYNX executing on miniHPC-Broadwell node also shows that FSC calculated a proper chunk size for both loops, obtaining the highest overall performance, outperforming GSS by 13.32% and mAF by 3.73%. This behavior is consistent among the results on nodes of miniHPC-KNL and Piz Daint-Haswell (Figure 4.1). In the following Section 4.3, we further investigate the performance of the scheduling techniques for the most time-consuming loop of SPHYNX, L1, while varying the chunk parameter. Figure 6 presents the load imbalance metrics, c.o.v. and p.i., calculated for the most time-consuming loop of SPHYNX execution on miniHPC-Broadwell node. These results show that most scheduling techniques achieve nearly perfect load balancing. Although these applications are computationally-intensive, with few memory operations, we can observe that almost perfect load balancing does not directly translate to high performance due to the additional scheduling overhead and loss of data locality. For instance, in Figure 6, AWF-B achieved perfect load balancing while in Figure 4.1 we can observe that the execution time of SPHYNX executing on miniHPC-Broadwell node with AWF-B was 21.40% slower than Best.
Aspects such as scheduling overhead, ccNUMA effects, and data locality cannot directly be observed in Figure 4.1, as the scheduling overhead is absorbed by improvement in the loop execution time due to dynamic and adaptive scheduling. Instead, we use a computationally-inexpensive loop of a widely used molecular dynamics application GROMACS [36], to reveal the overhead of all scheduling techniques. This particular loop in GROMACS has very low arithmetic intensity, regular loop iterations, and initializes three vector data structures which stresses ccNUMA effects and locality issues.
We also use the STREAM microbenchmark, a simple synthetic program that measures sustainable memory bandwidth, to show the memory bandwidth drop caused by ccNUMA effects and the locality issues that arise during dynamic and adaptive self-scheduling. Figure 7 shows the parallel loop execution time for the loop from GROMACS executing on node miniHPC-Broadwell while Figure 8 shows the memory bandwidth (in MB/s) maintained by each kernel from STREAM also executing on node miniHPC-Broadwell. We use miniHPC-Broadwell since it is a two-socket node, making ccNUMA effects and data locality issues more prominent.
Three factors contribute to the scheduling overhead shown in Figure 7: (1) Number of scheduling rounds o sr ; (2) Cost of calculating a chunk size o cs ; and (3) Synchronization cost between threads to obtain or to calculate a new chunk of loop iterations o sync . Note that o cs and o sync are incurred with every scheduling round, therefore, growing with o sr . It is important to note that due to the non-deterministic and stochastic nature of dynamic and adaptive self-scheduling, high o sr potentiates the loss of data locality and the importance of ccNUMA effects, which in this case are compounded with the overhead (compared to STATIC) seen in Figure 7.
STATIC in Figure 7 shows the smallest scheduling overhead, where o sr = 1, o cs is a simple and deterministic division, and o sync = 0, since threads need no synchronization.
SS shows the largest overhead due to o sr ≤ N , o cs is negligible, and o sync can either be negligible if o sr N or significant if o sr = N . In specific situations, some techniques may show higher overhead than SS (e.g. in this case, FAC, mFAC, TAP, AWF-D, and AF). If a scheduling technique with higher o cs than SS calculates the same chunk size as SS it will cause more overhead than SS.
GSS, TSS, and FSC incur less overhead than SS due to assigning larger chunks of loop iterations, which reduces o sr < N , increases data locality, while keeping mFAC has lower o sync than FAC by using atomic operations, leading to lower overall overhead than FAC. FAC2 and WF2 outperform FAC and mFAC in terms of o cs by using a simple chunk calculation function, not requiring profiling information, and using atomic operations for synchronization.
Similar to FAC and mFAC, TAP failed to calculate an appropriate chunk size based on the profiling information due to the very small loop iteration granularity, which resulted in loss of data locality, high o sr , and o cs .
BOLD generates chunk sizes very similar to STATIC but at a very high chunk calculation cost, therefore, incurring high o cs , and low o sr and o sync .
AWF-B and AWF-D do not manage to adapt to the very fine iteration granularity of this GROMACS' loop and assign very small chunks, reducing data locality and increasing o sr , and the cost of adaptation o cs . In contrast, AWF-C and AWF-E only incur from high o cs .
AF and mAF also have very high o cs . However, mAF also considers the time for o cs for the chunk size calculation. Therefore, it increases its chunk size to reduce o sr , offering improvement over AF.
In Figure 7 and Figure 8, one can note that SS causes high scheduling overhead due to high o sr and loss of data locality, which justifies the low memory bandwidth shown in Figure 8 for all STREAM kernels. The low memory bandwidth achieved by FAC and mFAC is justifiable since those techniques not only cause high o cs and o sync overheads but also need to read profiling information collected on a separate execution of the application. For simple kernels, such as those from STREAM and L0 from GROMACS, profiling the execution of each loop iteration may adversely influence execution performance, and may lead FAC and mFAC to calculate very small chunk sizes, increasing o sr and consequently also increasing scheduling overhead, ccNUMA effects, and loss of data locality. This observation is also valid for dynamic and adaptive techniques which measure (during execution) the execution time of previous chunks of iterations to determine the next chunk size. If the loop kernel's arithmetic intensity is low, dynamic and adaptive techniques may measure inaccurate values which may result in small chunk sizes, higher o sr , non-negligible ccNUMA effects and loss of data locality.
Impact of Chunk Parameter Choice
We explore the performance impact of the chunk parameter for all DLS techniques. We experimented with many different values for the chunk parameter per DLS technique, loop, application, and node type/configuration as indicated in Table 1. Figure 9 presents an overview of the results for SPHYNX comparing the Best combination of DLS techniques with the default value of the chunk parameter vs. the most performing combination of DLS techniques with the best value of the chunk parameter. The best value of the chunk parameter is identified by testing the application performance with values of the chunk parameter from N/2P down to 1 (see Table 1).
In Figure 9, the best chunk parameter always improved the performance of the applications. Similar colors indicate that the performance improvement was very low. A carefully selected chunk parameter for SS frequently achieves the highest performance. However, the process of finding such an optimal value requires extensive experimentation (e.g., such as the experiments presented here), and must be performed for each loop and system that the application will execute on. Furthermore, in the case of system variation, the optimal chunk parameter, once found, may no longer provide the highest performance since it would not be adapted during execution. It is impractical to rely exclusively on a manual and extensive experimentation process to find an optimal chunk parameter. This makes dynamically adaptive loop scheduling techniques a highly promising solution, especially on upcoming Exascale systems, which will increasingly be heterogeneous. The existing dynamic and adaptive scheduling techniques offer a first step for performance auto-tuning on a per loop basis against system and application variability. Based on the results in Figure 4.1 and Figure 9, it is interesting to examine the impact of the chunk parameter on the performance of most time-consuming loop from SPHYNX, L1. These results are shown in Figure 4.3, wherein the parallel execution time of L1 (y-axis) is shown for different chunk parameter values (x-axis). The chunk parameter values differ for each system since they are calculated using the available number of threads (see Table 1).
In Figure 4.3 we expect to see SS reaching or outperforming FSC. The dynamic and adaptive loop scheduling techniques are expected to improve the performance by reducing overhead since with a chunk parameter they preserve improved data locality, and are executed fewer times.
The performance of SS indeed reaches and outperforms that of FSC with larger chunk parameters between 12 and 1, 562 for miniHPC-Broadwell, 15 and 122 for miniHPC-KNL, and 10 and 1, 302 for Piz Daint-Haswell. This is due to the improved data locality and scheduling overhead of SS with a larger chunk parameter value. FSC is unaffected since it calculates a chunk size slightly larger than the range of chunk parameter values that achieve highest performance. The performance of FSC is only affected when the chosen chunk parameter value is larger than the chunk size calculated by the technique itself.
All results in Figure 4.3 show that performance degrades with large chunk parameter values, approximately 2, 000 loop iterations. This happens since the L1 loop from SPHYNX is irregular (see Figure 6, L1 of SPHYNX executed with STATIC) and, therefore, certain threads receive more work than others resulting in poor performance due to a load imbalanced execution. We expected that the dynamic and adaptive loop scheduling techniques show improved performance when the chunk parameter is chosen since their overhead would be reduced while preserving data locality. This was not the case. These results are discussed in Section 4.4, where the progression of the chunk sizes of each DLS technique is explored, highlighting why no improvement can be observed for the dynamic and adaptive loop scheduling techniques in this particular case.
Influence of Chunk Size Progression
The chunk size progression for the DLS techniques during the scheduling of the L1 loop from SPHYNX is shown in Figure 11, with chunk ID on the x axis (denoting the number of chunks produced) and their sizes on the y axis. We used LB4OMP with KMP PRINT CHUNKS=1 (chunk information feature, Section 3.2) to collect and report the chunk sizes assigned by each DLS technique in every scheduling round. In general, fewer chunks imply improved data locality and a smaller scheduling overhead due to fewer scheduling operations. Fewer chunks may also result in potentially higher load imbalance since the chunk sizes are larger than with fewer chunks, as observed in Figure 4.3, in Section 4.3. The results for the dynamic and adaptive loop scheduling techniques AWF-B,C,D,E, AF, and mAF in Figure 11, clarify why, in this case, no performance improvements can be ob-served when a chunk parameter is given. Even with a relatively large value of the chunk parameter, such as 3, 125, none of the adaptive loop scheduling techniques reaches the given value early enough to benefit from improved data locality and the lower overhead of executing fewer scheduling rounds. A much larger chunk parameter value would not necessarily improve loop performance due to the potential of load imbalance associated with large chunks.
Apart from AF, mAF, and FSC, all scheduling techniques follow a decreasing chunk size pattern. The dynamic and adaptive loop scheduling techniques AWF-B,C,D,E follow a decreasing chunk size pattern (similar to GSS and FAC2), with the major difference of adapting to system variation by increasing or decreasing their chunk sizes during execution.
Each DLS technique produces a different number of chunks (x-axis) depending on its chunk calculation method (non-adaptive methods) and loop performance during execution (adaptive methods). As described in Section 4.2, the number of chunks is proportional to the number of scheduling rounds, o sr , and contributes to the scheduling overhead associated with a technique.
The first chunks calculated by AF and mAF are small as these DLS techniques perform a warm-up scheduling round where they gather initial information about the loop iterations performance. These first chunks sizes are hard-coded to 10 loop iterations and are unaffected by the declaration of the chunk parameter.
Conclusions and Future Work
We introduced LB4OMP, a novel open-source library for dynamic load balancing of multithreaded applications that use OpenMP, implemented as an extension of LLVM's OpenMP RTL. This work contributes: a systematic and unified implementation of 14 dynamic (and adaptive) loop scheduling techniques; features for advanced performance measurement of loop performance and load imbalance; and an in-depth analysis of the performance potential and limitations of the OpenMP standard and the newly implemented scheduling techniques. Through an extensive performance analysis campaign we showed that for numerous application-systems pairs, the scheduling techniques in LB4OMP outperform those from the OpenMP standard.
With this work we bridge the gap between the state-of-the-art and the state-of-the-practice of load balancing in multithreaded applications. This will allow the efficient exploitation of large degrees of heterogeneous node-level parallelism for improving the performance of applications on upcoming Exascale systems.
LB4OMP represents the first and necessary step for devising automated methods to dynamically select the highest performing loop scheduling techniques during applications execution. Devising such methods is part of ongoing work by the authors.
A possible extension is to expand the selection criteria to include additional DLS techniques in LB4OMP. The study of locality-aware self-scheduling tech-niques is a promising research direction. We plan to patch and up-stream the DLS techniques implemented in LB4OMP to the main LLVM OpenMP RTL, facilitating a broad use and impact for OpenMP applications. Applying LB4OMP to explicit OpenMP task scheduling is also planned as future work. | 2021-06-10T01:15:45.733Z | 2021-06-09T00:00:00.000 | {
"year": 2021,
"sha1": "39b7f262586aaba5fcb00717abe509736662176c",
"oa_license": null,
"oa_url": "https://ieeexplore.ieee.org/ielx7/71/4359390/09524500.pdf",
"oa_status": "BRONZE",
"pdf_src": "Arxiv",
"pdf_hash": "39b7f262586aaba5fcb00717abe509736662176c",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Computer Science"
]
} |
80661882 | pes2o/s2orc | v3-fos-license | Relatives’ experiences of sharing a written life story about a close family member with dementia who has moved to residential care: An interview study
Abstract Aim The aim of this study was to describe relatives’ experiences of sharing a written life story about a close family member with dementia who has moved to residential care. Design An explorative descriptive qualitative design was used. Methods The data were collected using semi‐structured interviews with a purposeful sample of eight relatives and analyzed using a qualitative content analysis. Results Results show that creating and sharing the life story of a close family member could help relatives handle grief and stress. It was perceived as an important, yet difficult, task to ensure that the close family member got good quality care. The creation of a meaningful life story takes time and requires cooperation with family members and other significant people.
| INTRODUC TI ON
Living with dementia means struggling to adjust to changes in life caused by the disease (Holmes, 2012). Age is the main risk factor for dementia, with prevalence increasing exponentially after the age of 60 and to approximately 20% at the age of 80. The number of incidents was estimated to be 35.6 million in 2010 and is expected to nearly double every 20 years, to 65.7 million in 2030 and 115.4 million in 2050 (Berr, Wancata, & Ritchie, 2005;Sosa-Ortiz, Acosta-Castillo, & Prince, 2012). The growing group of people posing thereby a great challenge for the future of healthcare (United Nations, 2015). Dementia is a general term for a syndrome including a range of chronic or progressive organic brain diseases that are characterized by difficulties of short-term memory often followed by non-cognitive symptoms, referred to as behavioural and psychological symptoms of dementia (Song & Oh, 2015; World Health Organization [WHO], 2016). It is known that the use of life story (LS) can enhance person-centred care for people with dementia by enabling the healthcare professionals to know each person's needs, preferences and to see the person behind the illness (McKeown, Clarke, Ingleton, Ryan, & Repper, 2010). Relatives are those significant persons who can tell and create the LS about a close family member during admission to residential care (Hennings, Froggatt, & Payne, 2013). Knowledge of this nature is needed to understand what should be prioritized in the development of dignified and respectful care for people with dementia.
| BACKG ROU N D
People living with dementia have increased vulnerability, which can influence their daily life in a negative way as they may lack understanding from people around them. Living with dementia also makes it difficult for the older person to express and communicate their needs (Miyamoto, Tachimori, & Ito, 2010). Relatives might therefore be of importance telling their LS about a close family member for healthcare professionals during admission to residential care (Karlsson, Sävenstedt, Axelsson, & Zingmark, 2014). A LS can be expressed as a written text with information about the person, photographs and/or personal belongings significant for the person (Eley & Kaiser, 2017). It should include some evaluative points, which communicate moral values of the person and events that have had a special meaning that can be told and retold throughout life (Linde, 1993). The LS may contain events of significant memories from the person's life (McKeown et al., 2010). A LS mirrors in that way identities of people's lives by putting together things, past actions and events, into a meaningful LS and the person's dignity can thereby be preserved (Polkinghorne, 1996). When people are in the midst of their LS, which is true for people living with dementia, both the past and the future become important. What has happened to them in the past cannot be changed, although future events may change the importance of certain events from the past and the thoughts of how the rest of their life will be.
A LS, furthermore, supports the healthcare professionals in communicating with the person living with dementia and makes it possible to get to know the person and interpret and understand the person's needs and desires (Edvardsson, 2010;McKeown et al., 2010). Research (Heggestad & Slettebø, 2015;McKeown et al., 2010) shows that active use of a LS approach is a mechanism to improve person-centred care in residential care. A LS makes it also possible for healthcare professionals to see the person behind the disease, support the person in being seen and heard and help the person feel safe. However, as a relative to write down and share a LS about a close family member can be difficult, as the relative has to decide what to include and exclude, as some things may be too sensitive or hurtful to include. A trustful relationship must therefore be established with the healthcare professionals so they can support the relatives in this process (Ericson, Hellström, Lundh, & Nolan, 2001). Healthcare professionals may also help relatives handle complex emotions that come up when writing the LS (Kellett, Moyle, McAllister, King, & Gallagher, 2010).
According to the literature review, it becomes obvious that relatives are an important source of knowledge when a LS is being created during admission to residential care. Focusing relatives' experiences about sharing a written LS about a close family member during admission to residential care may be used as knowledge for healthcare professionals to improve person-centred care. The aim of this study was to describe relatives' experiences of sharing a written LS about a close family member with dementia who has moved to a residential care.
| Participants and procedure
A purposeful sample of eight relatives participated in the study.
The inclusion criteria were as follows: (a) being a relative to a close family member with dementia in a residential care, (b) having written a LS about the person and shared it with the healthcare professionals in the residential care in the last 6 months and (c) being informed by a registered nurse specialized in dementia (RND) about how to write a LS by using a template. There were five daughters, one son, one husband and one wife. The participants were between the ages of 52-72 years (mean = 61). Four participants had a university education, four a high school education. Three of the participants were retired, and the others were working.
A contact person at a Research Centre forwarded an information letter to RND working in five municipalities in a rural area of southern Sweden, to support the recruitment process. The RND gave eligible relatives an information letter with a consent form. Nine accepted the invitation, but before the interview took place, one of them withdrew.
| Data collection
Between March and April 2016, individual semi-structured interviews were conducted using an interview guide. This design was chosen to obtain data as detailed and complete as possible and to give the participants the opportunity to speak specifically about their experiences (Polit & Beck, 2012)
| Data analysis
To analyze the interviews the researcher used a content analysis (Graneheim & Lundman, 2004 were read and re-read several times to get a sense of the whole text. The text was then divided into meaning units. In content analysis, an important decision is the selection of meaning units to be analyzed-that is, sentences or phrases that are related to the aim. Each meaning unit was then condensed and coded-that is, the text was shortened and labelled with a code that described the content. The codes were then compared and sorted into subcategories based on differences and similarities. The analytical process was characterized by moving back and forward between the whole text and the parts of the texts. All authors discussed the meaning units, the condensation, the codes and finally validated the categories in the original text. According to Graneheim and Lundman (2004), creating categories is the core feature of qualitative content analysis.
| Ethical consideration
All the participants were given verbal and written information about the study and were told that participation was voluntary. They were assured of confidentiality and the right to withdraw from the study without any explanation. The study was conducted in compliance with the established ethical guidelines of the Declaration of Helsinki and received ethical guidance from the Ethical Advisory Board in South-East Sweden (No. 330-2016).
| RE SULTS
The analysis resulted in four categories: The establishment of a LS is a way to process grief and stress; Be the voice of the close family member; Need to understand the meaning of a LS; and Lacks a creative design for the LS. The categories are presented in the text below and illustrated with quotations from the interview texts to verify the categories.
| The establishment of a LS is a way to process grief and stress
The relatives described emotions of grief and stress associated with the progression of the dementia, which made it difficult to write and share the LS of their close family member with the healthcare professionals. They described their own life situation as intolerable since they had been forced to take full responsibility for the close family member's everyday life before the move to the residential care. The relatives experienced stress, fatigue and exhaustion due to the psychological strain that they had lived with for a long time.
They also felt emotionally influenced because of the new and dazing situations they had to handle during their close family member's transition to residential care. To write a LS in this situation was experienced as too much to deal with. The relatives experienced that they first needed to adjust to the new life situation and handle their stress and grief before they could manage to write and share a LS of their close family member. One participant stated: … it was so much with kind of everything then, then when she came there and should move to an accommodation and it was so turbulent before and everything, so you were pretty exhausted as a relative ….
Sharing a written LS was also experienced as a possibility to deal with the grief that came out of the separation caused by the dementia and the close family member's transition to the residential care. The relatives experienced that the writing of a LS was a way to express their own feelings, to look back on and remember their previous life together with the close family member and reach conclusions. To write and share the LS was therefore a way to accept the situation as well as to get an integrated picture of their past life together. One participant stated: maybe this has been some kind of therapy for me, to describe what our life has been like, I have many times felt that, yes, the possibilities for me to go on living in a way has been to think about our earlier life together.
| Be the voice of the close family member
The relatives described that it was important to be the voice of the close family member but that they needed time for the task. It was important to write a suitable LS-one that the close family member could feel proud of. The relatives experienced that writing the LS was difficult because some parts of the close family member's life were unknown to them, but they thought that some of these parts could be of relevance for the healthcare professionals. Therefore, they had to create the LS by putting pieces together in the form of text, photos and letters. In this process, the relatives experienced a need to consult other family members and significant people to get a complete picture of the close family | 279 member. The relatives experienced that it was positive to sit down with siblings or parents to remember and share memories from the past and together create the LS. Some relatives said that they included conversations with other family members who told stories about the person's life history, thereby making the LS richer and mirroring significant people's voices: …a resumé of thoughts, different experiences among the siblings…yes we shared, stories and situations… The relatives felt that if they were given the possibility and had time to investigate their close family member's life history, the LS could be more detailed and informative and thereby be more useful in daily life and contribute more to better quality of care. The relatives wished that the LS would have been written in an earlier stage of the dementia, as they thought it would be more beneficial in the daily care of their close family member. They described also the importance that the close family member then would be able to take part in creating the LS's content which could help them to recognize themselves.
| Need to understand the meaning of a LS
The relatives described that it was important to understand the meaning of a LS but felt that the written information they received from the municipal officials about the importance was insufficient.
Some of them expressed that when information about the LS was submitted together with other information and documents from the municipal office, there was a risk that information was lost in the wealth of information provided. Also, the time at which the information was given was of importance for the possibility of writing a good LS. The relatives experienced that it was difficult to take in so much information in this new and bewildering life situation. They experienced a constant uncertainty about the information they had been given about how to write a LS. They felt that even though the written information had been provided, they did not completely understand it. The relatives described that it would be of importance that the professionals who provided the information ensure that the message had been understood and they requested more opportunities to access information.
Receiving verbal information from the healthcare professionals, especially about the importance of writing a LS and filling in the template at the same time, was positively experienced. It was important to have the possibility to ask questions. One relative described how a registered nurse both gave information and invited the relative to a lecture about the importance of writing a LS. The information and the lecture were perceived to be both clear and inspiring, which motivated the relative to think outside the box and create a LS without filling in the existing template: …she made me go on an information about LS … she described why LS was important and how it was a way to take care of the people in residential care and I probably thought it was good that I, I did it before I received any request from the municipality…then I have done it in my way…I wrote my own without following any template….
The relatives understood that the LS was a working tool for healthcare professionals and that it would help to provide security and good care for their close family member. They expressed a strong desire that healthcare professionals, through the LS, would know the close family member better and understand who the person was before the illness.
It was considered important that healthcare professionals knew what the close family member had liked and not liked earlier in life to be able to deliver good care. The LS was helpful to facilitate communication between them as relatives and healthcare professionals since the close family members could not always understand what was being said or express themselves in words. However, the relatives were unsure whether the healthcare professionals using the LS after it had been written and submitted to them.
| Lacks a creative design for the LS
The relatives expressed a wish that the LS template that they received were designed differently. They felt that the existing template was too long, too stilted, too complicated or that there were too few lines to write on. They experienced a need to simplify and adapt the existing template to each person's needs and preferences.
They suggested that the LS should include images, written stories,
| D ISCUSS I ON
The results from our study show that writing a LS describing a close family member with dementia moving from the home to a residential care was a difficult process filled with stress and grief. The relatives described that it was important to be the voice of the close family member but felt that they need sufficient time for the task as well as more detailed information about how to write and share a LS by answering questions in the existing template. According to Helgesen, Athlin, and Larsson (2015) relatives of people living with dementia are often stressed, as they are responsible for the person's care, which may be experienced as a burden. Being responsible is a physical and psychological stress that causes exhaustion (Graneheim, Johansson, & Lindgren, 2014). Progression of the dementia causes impairments and that leads to the close family members not being able to take care of themselves, their actions or express their will.
Relatives seeing the person being changed may feel grief and become psychologically ill (Helgesen et al., 2015).
Sharing the LS can be a way to process the grief over the changed life situation and accept the separation from the close family member. Research (Kellett et al., 2010) describes relatives' experiences of remembering essential events and moments in terms of a renewed strength and of seeing the earlier life as rich and meaningful. The renewed strength made it easier for the relatives to accept the changed situation, which may also have reduced the stress and grief.
For a moment, it was possible to forget about the dementia, stand by the side and remember the close family member as the person he/ she once was. Relatives feel increased satisfaction by sharing the LS and they feel confirmed and sense that the healthcare professionals value them (Karlsson et al., 2014).
Relatives in our study reported that it was important to be the voice of the close family member. However, the personal commitment required to write down the LS was felt to be too much. The relatives were responsible for practical matters, such as paperwork and official contacts due to the transition to the residential care, which limited their time and energy for LS-writing. Research (Egan et al., 2007) shows that it is difficult for relatives to share a LS about someone else if there is lack of time. Creating a LS means to investigate someone else's life with personal details sometimes unknown for the relative. Therefore, it is an advantage to start writing down a LS early in the stage of dementia, when the close family member can participate (Karlsson et al., 2014). This increases the chance that it is the close family member living with dementia's LS and not an incomplete interpretation made by a relative. In the absence of knowledge about the close family member's life, relatives tend to make assumptions based on the insufficient knowledge they have (Egan et al., 2007). Unfortunately, this often results in a blank and predictable story, both impersonal and inaccurate. A well-written LS on the other hand is one of the most important tools to share and represent the close family member's history and identity.
The result shows further, the importance that the LS describes the family member's personality and earlier life in a proper way, so he/she could receive good care. Person-centred care affirms the identity of the person despite the impairments due to the dementia. The person living with dementia can therefore keep a feeling of identity and the ability to communicate needs and wishes would be maintained (Stein-Parbury et al., 2012). Kellett et al. (2010) stated that the written LS is an aid for professionals in personalizing the nursing care, which can lead to an understanding of the person through the LS and helps maintain their identity. A LS may also support healthcare professionals to understand what brings meaning to the life of the person living with dementia, which also the relatives in this study pointed out. Furthermore, a LS helps healthcare professionals stay focused on the person living with dementia, learn from their life and listen, respond and deal with the symptoms the person expresses (McKeown et al., 2010;Resnick, 2017).
The relatives in our study experienced a need to consult other family members to fill knowledge gaps about the close family member's earlier life. They had positive experiences of sharing memories with the close family member but needed support from other family members to be able to write an extensive LS. Thompson (2011) believes that a person owns his own life history and the written LS should, as far as possible, be designed as the person living with dementia wants and expresses. But severe cognitive impairment can make it necessary for a relative to give a helping hand with the LS.
After that, the LS can develop and follow the person when moving into a residential care. It is, therefore, according to Thompson (2011), important that a LS should be written in an early stage of the dementia, so it can be of use and help for the person living with dementia, the relatives and healthcare professionals. The results in our study show that the LS were asked for too late, when the person living with dementia had already moved into the residential care.
However, when relatives write the LS together with the person living with dementia it may also lead to an improved relationship and reduced negative feelings (Subramaniam, Woods, & Whitaker, 2014).
By creating the LS together, the relatives may bond with the close family member and re-establish contact. It may also contribute to a meaningful communication and quality time between them (Russell & Timmons, 2009;Subramaniam et al., 2014). A LS may be a joy to descendant relatives through positive experiences in making copies | 281 ANDERSSON Et Al. of the LS, through giving copies to other family members for remembrance and through reading the LS about the close family member (Subramaniam et al., 2014). Thompson (2011) implies that a written LS leads to an increased understanding, a deeper sense of belonging.
This gives feelings of safety, reduces anxiety and improves the relationship between the relative, the close family member living with dementia and the healthcare professionals.
Our results show that relatives needed to understand the meaning of a LS and what facts they should have included. However, some of them lack information about when to create a LS and experienced that much information was handed over at the same time, which made it difficult to deal with. The relatives felt unsure and stressed if they received information they did not understand. Kellett et al. (2010) describe how relatives experience confusion, aimlessness and uncertainty when it comes to writing a LS. They are not sure what they are supposed to write, and they do not know which information is important or relevant. Relatives need, therefore, to communicate the LS during the creative process with both the person living with dementia, other family members and healthcare professionals. A well-written LS enables the healthcare professionals to see the person behind the disease, which makes the nursing care more person-centred and this leads to improved relationships (Russell & Timmons, 2009). Relatives in our study described that the template they received was too long, stilted and difficult to fill in and therefore requested a different design of the LS template and wanted it to be written based on the needs of their close family member. Research (Thompson, 2011) shows that a LS can be designed in many ways-for example, as a book, leaflet, collage, memory box or electronically. However, it is important to choose a format that is suitable for all involved. The written LS is often more detailed when described and presented as a book, while a collage is more like a creative project about the person living with dementia and their relatives. Furthermore, memory boxes contain subjects that are of value to the person living with dementia. Clarke, Hanson, and Ross (2003) stated that a combination of written text, personal stories and pictures, often of significant people or events, make the LS alive and can elicit memories. Photos can also facilitate communication and create a stronger bond between the person with dementia and healthcare professionals.
| Strengths and limitations
This study has limitations that need to be discussed. First, the sample size in this study was quite small but according to Sandelowski (1995), the sample size in qualitative research should be large enough to achieve variations in experiences and small enough to permit a deep analysis of the data. The present sample with eight relatives was considered appropriate to maintain depth in the analysis. The participants were also recruited by means of a purposeful sampling, which means that they were selected based on their special knowledge about the studied phenomenon. The included participants had a high educational qualification and that fact may be regarded as a possible selection bias (c.f. Polit & Beck, 2012). However, the level of education was not an inclusion criterion. The participants gave rich descriptions of their experiences and as a result, their stories contained variations of the phenomenon under study. Dependability was ensured as an interview guide was used, and follow-up questions were integrated, which allowed the interview to be more reflective. This data collection method decreased the risk of inconsistency (Polit & Beck, 2012). To achieve trustworthiness of the analysis, the authors discussed the findings until consensus was reached, which strengthens the credibility of the study. The results cannot be generalized as they are limited to the small sample size but can be transferred to similar situations if the results are re-contextualized (Polit & Beck, 2012).
| CON CLUS ION
This study indicates that it is important for relatives to share a close family member's LS with healthcare professionals at the residential care, but at the same time, it is a stressful process that should start as early as possible when a person got a dementia diagnosis. To write and share the LS was experienced as being important in ensuring that the close family member got good quality care and was also a way to handle their own grief and stress over the changed life situation when the close family member moved to the residential care. It was difficult for relatives to share a LS about the close family member, as creating a LS means investigating someone else's life with personal details sometimes unknown to them. However, to create a useful LS means to have enough time and possibilities to include significant people, places, personal things and memories in a creative design to maintain a positive view of the close family member's life. Registered nurses need to involve close relatives to contribute with important information to the LS and thereby preserve personhood and enhance the dignity when a person with dementia is admitted to residential care. A LS is useful and important for the nursing staff to provide person-centred care based on the individual's needs and resources.
ACK N OWLED G EM ENTS
We would like to thank our participants, the contact person at the Research Centre, the registered nurses specialized in dementia who helped with participant recruitment and Stephen Garland for help with language revision.
CO N FLI C T O F I NTE R E S T S
No conflict of interests have been declared by the authors.
AUTH O R CO NTR I B UTI O N S
All authors contributed to design the study, HD and UBJ collected the data. EKA, HD and UBJ analyzed the data. All authors together drafted the manuscript and contributed to editing the final manuscript, revised it critically for scientific content and approved the final version. | 2019-03-18T14:03:29.035Z | 2018-10-04T00:00:00.000 | {
"year": 2018,
"sha1": "98bb88ef885d4856aecfc0e8c7e77342f6fbbd88",
"oa_license": "CCBY",
"oa_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/nop2.208",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "98bb88ef885d4856aecfc0e8c7e77342f6fbbd88",
"s2fieldsofstudy": [
"Psychology",
"Education",
"Medicine"
],
"extfieldsofstudy": [
"Psychology",
"Medicine"
]
} |
119100202 | pes2o/s2orc | v3-fos-license | First radial excitations of heavy quarkonium in a contact interaction
For the flavor-singlet heavy quark systems of quarkonia, we compute the masses of the first radial excitation of mesons in four different channels: pseudo-scalar ($\eta_{c,b}(2S)$), vector ($\psi (2S),\Upsilon(2S)$), scalar ($\chi_{c_0,b_0}(2P)$) and axial vector ($\chi_{c_1,b_{1}}(2P)$), as well as the weak decay constants of the $\eta_{c,b}(2S)$ and $\psi (2S),\Upsilon(2S)$. The framework for this analysis is provided by a symmetry-preserving Schwinger-Dyson equations treatment of a vector$\times$vector contact interaction. The results found for the meson masses are in good agreement experimental data and earlier model calculations based upon Schwinger-Dyson and Bethe-Salpeter equations (BSEs) involving sophisticated interaction kernels.
I. INTRODUCTION
The discovery of the η c (2S) in 2002 [1][2][3] was the cornerstone of a new era in meson spectroscopy. Recent advances in the understanding of quantum chromodynamics (QCD), and the recent flurry of experimental activity have led to the discovery of a bunch of new heavy states, conceiving a golden era for heavy quarkonium physics [4].
For a long time, quarkonia spectra were mainly understood from a phenomenological perspective through the use of potential models to describe the dynamics expected from QCD [5][6][7][8]. The potential-model approach was alluring owing to its simplicity, but became steadily more complicated and inaccurate for the heavier quarkonia states and needed new parameters fitted to data when they confronted relativist effect [9]. However, an innovative unquenched quark model showed that relativistic effects in the coupling for bottomonia was less than 1% [10,11], while for charmonia they were around 2 − 6% [12]. On the other hand, direct calculations performed with Lattice QCD successfully describe much of the quarkonia spectra [13][14][15], though they are objectionably complex and computationally expensive.
Remarkably, in recent years, Schwinger-Dyson equations(SDEs) of quantum chromodynamics(QCD) has contributed to our understanding of these systems; their derivation makes no assumption about the strength of the interaction involved [16]. Thus, since heavy quarkonia systems connect the hard scale of the heavy constituent quarks and the soft scale of the relative momenta between them, they can conveniently to be studied through SDEs [17].
The evolution of the SDEs project to the even more complicated exotic and baryonic states, decay rates and form factors is substantially non-trivial: for instance, in the calculation of elastic form factors (EFFs) [40], and transition form factors (TFF) [41]. It has been demonstrated that brute force numerical evaluation is unable to inspect the large momentum transfer region of form factors and is not considered adequate to make full comparison with already available experimental data. However, a subtle parameterization of the Bethe-Salpeter amplitudes (BSAs) in terms of Nakanishi-like perturbation theory integral representations [42] allows us to reach large space-like momentum transfer region [28,43,44].
Recently, a symmetry-preserving vector-vector contact interaction has appeared as an alternative to full QCDbased explorations [45][46][47][48][49]. Along with the rainbowladder(RL) approximation of the SDEs, which is the leading order in a systematic DSE, one obtains a fully consistent treatment of the simple to implement CI model, that is helpful in providing useful results which can be compared and contrasted with full QCD calculation and experimental data. In this model, confinement is implemented by employing the proper-time regularization scheme. This scheme systematically removes quadratic and logarithmic divergences, ensuring that the axialvector Ward-Takahashi identity (axWTI) is satisfied.
This interaction provides a good description of the masses of meson and baryon ground and excited states for light quarks [45][46][47][48][49]. The results derived from the CI model are quantitatively comparable to those obtained using sophisticated QCD model interactions, [17,30,50,51]. Strikingly fascinating, this simple CI model produces a parity-partner for each ground-state that is always more massive than its first radial excitation, so that, in the nucleon channel, e.g., the first J P = 1/2 − state lies above the second J P = 1/2 + state [49,52].
Our intention is to implement that model [53,54] on the calculation of spectra of radially excited heavyquarkonia states. We compute their mass spectrum by using the dimensionless coupling for the CI model present in Ref. [54]. In addition, we compute their weak decay constants, whose precise knowledge is of huge importance for the hadronic observables measured by LHCb and FAIR-GSI, for example. This paper is organized as follows: in Section II we give the minimum details necessary to the SDE-BSE approach to mesons, employing the contact interaction in the RL approximation, and the consequences that this interaction has for the interaction kernels. In Section III, the model is minimally modified to calculate the first radial excitation. In Section IV, we tabulate our results for the mass spectrum of ground state quarkonia and the decay constants for η c,b (2S), Ψ(2S) and Υ(2S) in the minimally modified CI model. Finally, in Section V, we present our conclusions.
II. CONTACT INTERACTION MODEL
We dedicate this section to recapitulating the CI model and its implementation in a SDE-BSE formalism to study two-particle bound systems. We precise their connection with chiral symmetry breaking and confinement; and we describe the method adopted to obtain the results disclosed in this article. For a more detailed description of the model, see Refs. [49,53].
A. SDE-BSE formalism
The f -flavor dressed-quark propagator S f is obtained by solving the quark SDE [56][57][58][59] where g is the strong coupling constant, D µν is the dressed-gluon propagator, Γ a ν is the dressed-quark-gluon vertex, and m f is the f -flavor current-quark mass.
Since the SDEs form a coupled infinite set of non linear integral equations, a truncation scheme is required in order to characterize a tractable problem. This is achieved once we have specified the gluon propagator and the quark-gluon vertex. For a comprehensive recent review of the SDE-BSE formalism and its applications to hadron physics, see, for example, Ref. [17].
Because high-energy experiments cannot perceive quarks directly, bound states have to be studied in order to test QCD. Specifically, a meson bound-state problem in an explicit J P C channel is determined by its homogeneous BSE [60][61][62], is a momentum-sharing parameter, p (P ) is the relative (total) momentum of the quark-antiquark system; S f is the f -flavor dressed-quark propagator; Γ H (p; P ) is the meson Bethe-Salpeter amplitude (BSA), where H specifies the quantum numbers and flavor content of the meson; r, s, t, and u represent color, Dirac and flavor indices; and K(p, q; P ) is the quark-antiquark scattering kernel. They also specify the kernel in the BSE, Eq. (3), through the axial-vector Ward-Takahashi identity (axWTI) [63] − iP µ Γ 5µ (k; Equation (4), which encodes the phenomenological features of dynamical chiral symmetry breaking (DCSM) in QCD, relates the axial-vector vertex, Γ 5µ (k; P ), to the quark propagator, S(k). This in turn implies a relationship between the kernel in the BSE, Eq. (3), and that in the quark SDE, Eq. (1). This relation must be preserved by any viable truncation scheme of the SDE-BSE coupled system, thus constraining the content of the quarkantiquark scattering kernel K(p, q; P ).
B. Rainbow-ladder truncation and the contact interaction
In a symmetry-preserving vector×vector contact interaction, one considers that the interaction between quarks is not mediated via massless bottom exchange, but instead through the interaction defined by where m g = 800 MeV is a gluon mass scale which is in fact generated dynamically in QCD [64], and α IR is the CI model parameter, which can be interpreted as the interaction strength in the infrared [65,66]. The simultaneous implementation of Eq. (6) for the quark-gluon vertex and Eq. (11) for the scattering kernel is the familiar rainbow-ladder approximation.
Once the kernel has been specifed by Equations (5) and (6) in the quark SDE, Eq. (1), then the the general form of the momentum-independent f -flavored dressed-quark propagator within the context of the rainbow-ladder truncation and a contact interaction is [45-49, 53, 55] Thus, the flavor-dependent fermion constant mass M f is obtained by solving Since the integral in Eq. (8) is divergent, we must adopt a regularization procedure. We employ the proper time regularization scheme [67] to write Eq. (8) as where Γ(a, z 1 , z 2 ) is the generalized incomplete Gamma function: The parameters τ IR and τ UV are infrared and ultraviolet regulators, respectively. A nonzero value for τ IR ≡ 1/Λ IR implements confinement [68]. Since the CI is nonrenormalizable theory, τ UV ≡ 1/Λ UV becomes part of the model and therefore sets the scale for all dimensional quantities. The importance of an ultraviolet cutoff in Nambu-Jona-Lasinio-type models has also been discussed in Refs. [69,70].
In the context of the contact interaction and rainbowladder truncation, Eq. (3) gives where g 2 D µν is given by Eq. (5). Thus, the homogeneous BSE (η = 1) takes the simple form Furthermore, the implementation of axWTI entails [45-49, 53, 55] where (13) states that the axWTI is satisfied if, and only if, the model is regularized so as to ensure there are no quadratic or logarithmic divergences: circumstances under which a shift in integration variables is permitted, an operation required in order to prove Eq. (4) [45][46][47][48][49].
Since the interaction kernel given by Eq. (11) does not depend on the external relative momentum, a symmetrypreserving regularization will give momentum independent solutions. In this case, the general forms of the BSAs for the pseudoscalar, scalar, vector, and axialvector channels, respectively, are given by where M is a mass scale, built from solutions of Eq. (9). Results for physical observables are independent of this choice.
Since the BSE is a homogeneous equation, the BSA has to be normalized by a separate equation. In the rainbow-ladder approximation, this condition is H , Γ H is the normalized BSA, and Γ H is its charge-conjugated version. For the vector and axial-vector channels, there is an additional factor of 1/3 on the right-hand side to account for all three polarizations of a spin-1 meson.
Once the BSA has been normalized, observables can be computed. The pseudoscalar and vector meson decay constants, f 0 − and f 1 − , are defined, respectively, by where m 1 − is the mass of the vector bound state, and the factor of 3 in the denominator on the right-hand side of Eq. (20) comes from summing over the three polarizations of the spin-1 meson.
C. CI running coupling , whereαIR = αIR/m 2 g , for the contact interaction, extracted from a best-fit to data, as explained in Ref. [54]. Fixed parameters are mg = 0.8 GeV and ΛIR = 0.24 GeV.
In a previous paper [54], we explained how the CI can be used to study light and heavy mesons. When studying the heavy sector, a change in the model parameters has to be made: an increase in the ultraviolet regulator, and a reduction in the coupling strength. Subsequently, we figured out that different set parameters are needed in order to study each sector: light, charm and bottom, as displayed in Table I. With these parameters, we define a dimensionless coupling α guided by [69,70] The drop in α, in relation to its value in the light-quarks sector, can be read off from the last column of Table I. Indeed, α is reduced by a factor of 2.1−2.3 on going from the light to the heavy sector, instead of the apparent large factors quoted in Tables III and IV. Moreover, as a reminiscent of the running coupling QCD with the momentum scale at which it is measured, an inverse logarithmic curve can fit reasonably well the functional dependence of α(Λ UV ). The fit reads where a = 0.923 and Λ 0 = 0.357 [54]. With this fit, it is viable to recover the value of the strength coupling α once given a value of Λ UV . We will apply this feature in the future. It is interesting to note that the variation of the coupling α as a function of ΛUV is not far from a logarithmic decrease fitted by the equation.
III. MESON FIRST RADIAL EXCITATION IN A CONTACT INTERACTION
Recently, we have developed a practical CI model suited to calculate observables for light and heavy mesons [54]. Another paper has focused on D mesons using this model [71]. These studies have shown that as the system becomes heavier, a decrease in the coupling and an increase in the ultraviolet cut-off are required. Additionaly, observables for first radial excitations using Table I. this model have been already calculated in the light sector [48,49]. As a continuation of these works, we will adopt their strategy as a starting point from which to calculate first radial excitations in heavy quarkonia systems. Generally, most of the studies that use the SDE-BSE formalism analyze only ground-state systems. When examining radially excited states in fully covariant approaches, a single zero is related to the first radially excited bound-state. This zero is usually seen in the relative-momentum dependence of the leading Tchebychev moment of its dominant Dirac structure [32,[72][73][74]. However, this approximation is not exempt from problems: as the analytic structure of the quark propagator restricts the range of direct calculation [75], some extrapolation may be needed [76]; additionally, some nonphysical solutions may show up, and these cannot be easily related to a physical state [77].
Furthermore, since a single zero cannot be exhibited by a momentum-independent relative-momentum boundstate amplitude, the possibility that the interaction between quarks is momentum-independent vanishes. In other words: if the zero is established at k 2 0 , then a momentum-independent interaction can only produce reliable results for phenomena that probe momentum scales k 2 k 2 0 . In the light sector k 2 0 ∼ 2M 2 ∼ (0.5 GeV) 2 is mostly found [32,72,73]. However, in Refs. [78,79], a way out of this predicament is indicated: in the phenomenological analysis of the contact interaction, a zero by hand is inserted into the analytical expressions of the kernels built up from Eqs. (3). Now we identify the BSE for a radial excitation as the form of Eq. (12) obtained with Eq. (5) and insert a node by hand into the BSE; this reads [47,49] which forces a zero into the kernel at q 2 = 1/d F , where d F is a parameter that will be specified later. Essentially, the mass of the bound-state is increased because the presence of this zero reduces the coupling in the BSE. The mass of a meson radial excitation is elevated by an equivalent mechanism when employing more sophisticated interactions kernels. The presence of this zero needs te replacement C αβ → F αβ of each BSE kernel found in Appendix A of Ref. [53] Because the renormalization condition Eq. (18) involves two amplitudes given by Eq. (23), for an excited state we substitute C αβ → G αβ in each renormalization condition in Appendix B of Ref. [53] where D 01 , E 01 and related terms are explicitly given in Appendix A. Now, it only remains to specify the choice of the new parameter d F introduced in Eq. (23). In Fig. 2, we show the evolution with d f on the excited states masses from the light sector. We immediately notice that there is a switch between the ordering of m π * and m ρ * at 1.5 GeV −2 . Additionally, there is also a maximum value for d f at 8.7 GeV −2 , where the value of the mass for all channels diverges. This maximum value is independent on the current quark mass and the bounds restricted by the infrared and ultraviolet cutoffs, as appears in Figs Table I. Table I. to obtain a value for m π(2S) =1.33GeV. In the next section we test this scheme and calculate the spectrum of first radially excited heavy mesons states.
IV. NUMERICAL RESULTS
Tables II-IV show the masses of ground-states and first radial excitations for different channels that have been studied with the CI. Results of ground-states were obtained and are to be found in the discussion in previous papers [48,49,53,54]; and the light-sector analysis of radial excitations is found in [48,49].
Our results are displayed in Tables III and IV. We notice that, in maintaining the set of parameters used for the ground-states, then radially excited states of pseu-doscalar channels are more massive than those of their respective vector channels, this means m * 0 −+ > m * 1 −− . In contrast, experimental results indicate m * 0 −+ < m * 1 −− . This incorrect ordering also appears in fully-covariant models [32] when analysing the light sector; therefore it is considered a weakness of the rainbow-ladder approximation. Furthermore, as an indication of a naive application to the heavy sector, the first radially excited masses obtained are even lower than their ground-state counterparts in all but the pseudoscalar channel.
In the phenomenological analysis of the contact interaction, in order to obtain an experimental value for the a 1 − ρ splitting, a spin-orbit repulsion has been introduced into the scalar and axial-vector channels to simulate a large dressed-quark anomalous chromomagnetic moment [48,49]. In a recent study, when including opposite-parity diquark correlation in the analysis of the structure of baryons, a second spin-orbit parameter is suggested to match the a 1 −ρ and σ−ρ splitting produced by sophisticated Bethe-Salpeter kernels. The introduction of these parameters simulates the DCSB effects that are crucial for a successful description of the meson spectrum with truncations beyond RL [52]. However, the implementation of that spin-orbit coupling into the BSE kernel has negligible effects in the heavy sector [53].
The analysis of heavy mesons by means of the CI requires a diminution of the coupling accompanied by an increase in the ultraviolet cut-off, as explained in detail in Ref. [53]. Precisely, when the mass of the heavy quarks increases, the further the coupling between them decreases and has to be compensated by an increment in the ultraviolet coupling. On the other hand, in fully covariant models, when the meson mass increases, the effective coupling decreases, and a small correction of the mass may mean a large correction of the binding energy [18].
The sensitivity to the coupling in the heavy sector explains the small values that we obtained in our first attempt to calculate first radially excited heavy quarkonia states with the same parameters as their ground states. Additionally, in pursuing a fix in d F to obtain a correct value for η c (2S) and η c (2S), we get too close at the divergence value of 8.7 GeV −2 , as it can be seen in Figs. 3-4. This means that a small variation on d F is reflected in a big change on the value of the mass and, as a consequence, we have to present a proposal to study first radially excited states with the CI.
With the aid of Eq. (22), we calculate the evolution of quarkonia first radially excited states masses in relation to Λ UV : charmonia spectra are shown in Fig. 5, while those of bottomonia are displayed in Fig. 6. They show a slight increase in the masses, as a result of the evolution of the quark-dressed mass, which increases on reducing the coupling and enlarging Λ UV [53]. Furthermore, we see that, as long as we reduce Λ UV , the mass gap between the pseudoscalar and vector channels decreases slightly; it is therefore expected that this ordering reverses at some point, as can be seen in this figure. A similar behaviour also appears in Ref. [32]: when the infrared length scale of the model therein increases, the ordering is reversed and a correct ordering m * π > m * ρ is presented.
FIG. 5: Contact interaction evolution for cc first radially excited states in terms of ΛUV .
A. Charmonium first radial excitation
In Table III, we present new results for charmonia first radial excitations with new parameters. We propose fixing Λ UV in order to obtain m ηc(2S) = 3.6 GeV, while the remaining channels are predicted values. We found that m χc0(2P ) > m χc1(2P ) , which is in contrast with other model predictions. Even though there are not experimental results for m χc0(2P ) and m χc1(2P ) , we expect a good comparison when data become available. In Fig. 7, we present a pictographical scheme of the spectra and compare this with experimental data when they are available.
Before continuing to the analysis of bottomonia results, it is important to consider the X(3930) [81] and X(3915) [82] states, both with J P C = 0 ++ [83], which suggests that one of them could be the χ c0(2P ) quark model state [84], though there are still some details to MeV. In addition, it is necessary to stress that there is a switched ordering between the scalar and the axialvector channel predicted by means of the CI model and fully covariant models. Moreover, in considering bottomonium experimental results, we consider our results to be better approximation and an acceptable hint for the m χ c0(2P ) − m χ c1(2P ) mass splitting.
B. Bottomonium first radial excitation
Conclusions for bottomonia are a repetition of the results for charmonia. Table IV shows our previous results for ground-state mass spectra in four different chan- Table III. Experimental data are taken from Ref. [80].
nels. These values are in excellent agreement with experimental data [80] and those of other model calculations. However, when calculating the first radial excitation, we again have a drawback: the masses of excited states are lower than their ground-state counterparts. Fig. 8 shows the bottomonium spectrum with the value of Λ UV = 10.17GeV fixed to obtain m Υ(2S) = 9.992GeV; the rest of the predicted results are presented in Table IV. In this case, the experimental mass splitting m χ b0(2P ) − m χ b1(2P ) = 23 MeV is in agreement with our value (9 MeV), whereas other models predict m χ b1(2P ) − m χ b0(2P ) > 150 MeV. Additionally, the CI model predicts the correct mass ordering between these particles, in contrast to other models, which predict a reversed ordering. Table IV. Experimental data are taken from Ref. [80].
Finally, from Fig. 8 and Table IV, we conclude that, when comparing the mass splitting m χ b0(2P ) − m χ b0(1P ) and m χ b0(2P ) − m χ b0(1P ) between the experimental and those of our model, they match perfectly. These results suggest that the CI predicted values in Sec. IV A are close data to the experimental ones. Charmonia BSAs are displayed in Table V. First radial excitation obtained by considering the set of parameters presented in Table III, whereas η c (2S) and Ψ(2S) decay constants are reported in Table VI. Similarly, bottomonia BSAs are given in Table VII, and η b (2S) and Υ(2S) decay constants are shown in Table VIII. It is important to notice that, the canonically normalized amplitude associated with the excited states BSE, are smaller than their corresponding of the ground-states. Otherwise, our results are in reasonably good agreement with experimental results when they are available.
However, the value of the η c (2S) decay constant predicted by our model differs too much from those of other [25] 0.063 0.147 RB2 [25] 0.062 0.162 HGKL1 [36] 0.103 0.096 HGKL2 [36] 0.089 0.121 model predictions; these latter present values from 0.063 to 0.103, while our predicted value is 0.296, which is even greater than the η c (1S) decay constant. We attribute this enormous difference to the fact that the decay constant is sensitive to changes in the model coupling and ultraviolet cut-off Λ UV . It is also important to consider, that this, time we only fitted the parameters to the value of m ηc(2S) . On the other hand, the Ψ(2S) decay constant is in agreement with the experimental value. The bottomonia scene presents a better picture; the predicted values for η b (2S) and Υ(2S) are within 10% concordance with other model predictions, but all of them halve the experimental value. It is expected that future results beyond RL approximation will improve the results of decay constants.
V. CONCLUSIONS
We compute the quark model first radial excitation state spin-0 and spin-1 heavy quarkonia masses, and the decay constants by using a rainbow-ladder approximation of the simultaneous set of SDE and BSE. By means of a CI model of QCD previously developed to the study of ground-state mesons [45-49, 53, 54].
In a fully covariant approach, a radial excitation is a zero in the relative-momentum dependence of the leading Tchebychev momentum of its dominant Dirac structure. This property precludes the possibility that the interaction between quarks is momentum-independent, as an independent relative-momentum bound-state amplitude cannot exhibit a single zero. However, we have inserted a zero by hand into the analytical expressions of the BSE. In this way, the coupling is reduced and the mass of the bound state increases. This idea was first implemented in calculating the spectra of light-mesons first radial excitation. Although, the results provided a good comparison with experimental data, the mass splitting between their ground state and first radially excited state was underestimated by 0.2 GeV [49]. On applying this same idea to heavy quarkonia, we obtained a first radial excited state that was too low for a correct comparison with experimental data, and changing the new parameter does not improve this picture.
In our previous studies of heavy-quarkonia in a contact interaction [53,54], we showed that the extension of the CI model to heavy quarkonia requires a reduction in the model interaction strength, mimicking the high momentum tail of the quark mass function and BSAs [20,86,87] as a consequence of the asymptotic freedom of QCD. This reduction in the interaction strength is appropriately compensated by increasing the ultraviolet cut-off; from these parameters, a dimensionless interaction strength can be defined. With this in mind, we have unified a contact interaction model to study light and heavy mesons; we have also recognized that the coupling can be fitted reasonably well with a inverse logarithmic curve, with an appearance evocative of the QCD running coupling.
In order to obtain a correct mass of the first radially excited state, we suggest that heavy excited mesons need different parameters from those of their groundstate counterparts. This time, instead of use a best fit to obtain the correct values of the pseudoscalar channel, we used the inverse logarithmic curve to obtain a value of the coupling CI model, α IR , in terms of the ultraviolet cut-off Λ UV . This time we find that the masses of the first radially excited-state heavy quarkonium are in good agreement with experimental data [80]. However, we find an incorrect ordering between the pseudoscalar and vector channel, which can be mainly attributed to a defect of the RL approximation.
One important feature of our model is that predicts a mass splitting between m χ b0(2P ) −m χ b0(1P ) and m χ b1(2P ) − m χ b1(1P ) that is consistent with experiment, though each value is underestimated by ≈ 200MeV in relation to experiment. Consequently, if we think the mass splitting values m χ c0(2P ) −m χ c0(1P ) and m χ c1(2P ) −m χ c1(1P ) are correct, and as charmonia ground-state mass values are close to the experimental ones, we can state that the values of m χ c0(2P ) and m χ c1(2P ) might be close to experimental results once these states are detected.
Finally, the decay constants that we have obtained are rather inconsistent in the charmonia sector, our predictions predictions differ too much from those of other model predictions for η c (2S). Nonetheless, we have an excellent agreement with experiment in the value for Ψ(2S). On the other hand, the predicted results for bottomonia are pretty similar for both η b (2S) and Υ(2S), but there is a mismatch of almost 50% in the predicted value for Υ(2S). We expect that future studies beyond RL approximation will improve these values.
This work is part of the series of studies on heavyquarkonia in a contact interaction [53][54][55]71], in which we move towards a comprehensive study of heavy mesons and QCD by using this model. Further steps will involve flavored mesons and baryons, as well as exotic states. Our goal is to provide a unified phenomenological description of light and heavy hadrons within the CI model. | 2018-07-05T08:03:34.000Z | 2018-07-05T00:00:00.000 | {
"year": 2018,
"sha1": "7f0bf63aacd688e3582f1f4e01a5840b9368dc10",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "b237baf69b91f1e3735b369f14ba99f2a523d490",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
} |
37970568 | pes2o/s2orc | v3-fos-license | 2,3-Dimethylanilinium chloride monohydrate
The crystal structure of the title salt, C8H12N+·Cl−·H2O, consists of discrete organic cations, chloride anions and water molecules which are connected by N—H⋯Cl, N—H⋯O and O—H⋯Cl hydrogen bonds. These interactions lead to the formation of layers lying parallel to the ab plane.
The crystal structure of the title salt, C 8 H 12 N + ÁCl À ÁH 2 O, consists of discrete organic cations, chloride anions and water molecules which are connected by N-HÁ Á ÁCl, N-HÁ Á ÁO and O-HÁ Á ÁCl hydrogen bonds. These interactions lead to the formation of layers lying parallel to the ab plane.
Comment
Hydrogen bonding is of interest because of their prevalent occurrence in biological systems. Therefore, it is extremely useful to search simple molecules allowing to understanding the configuration and the function of some complex macromolecules.
Furthermore, the hybrid materials are wealthy in H-bonds and they could be used to this outcome because of their capability emphasis in constructing sophisticated assemblies from isolated molecular or ionic building blocks due to its strength and directionality (Steiner, 2002;Jayaraman et al., 2002).
As shown in Fig. 1, the asymmetric unit of (I) contains a 2,3-dimethylanilinium cation, a chloride anion and a water molecule. Packing diagram of the structure across the a-axis is shown in Fig. 2. It shows that each chloride anion connected to two 2,3-dimethylanilinium cations via N-H···Cl hydrogen bonds and two water molecules. The title complex is a crystalline hydrate containing one water of crystallization, where form layers through N-H···O and O-H···Cl hydrogen bonds (Table 1).
The C-NH 3 , 1.465 (2) Å distance in the organic cations are close to respect to the C-NH 3 , 1.459 (2) Å observed in the crystal structure of 2,3-dimethylaniliniumchloride (Dai & Chen, 2010). Moreover the organic group moiety geometrical features shows the C-C-C and C-C-N angles are in the range usually found for this compound (Abid et al., 2007). Hydrogen bonds, electrostatic and van der Waals interactions participate to the cohesion of the three-dimensional network and add stability to this compound.
Experimental
An initial solution of 2,3-dimetylaniline was made in 10 ml methanol. To a crystallizer vessel initial solution was added in a 1:1 molar ratio of concentrated hydrochloric acid dropwise. For salt formation partnership, the obtained solution was stirrer for 1 h and then gradually evaporated in room temperature. Crystals of the title salt were removed from the crystallizer vessel to yield colorless crystals of the title salt, suitable for X-ray analysis.
Refinement
The H atoms of the protonated nitrogen and water molecule were found in difference Fourier map and refined isotropically.
The water H atoms H1W, H2W were refined with distance restraints of O-H 0.844 (2), 0.860 (2) Å, respectively. The Fig. 1. The asymmetric unit of title compound with displacement ellipsoids drawn at 50% probability level. Refinement. Refinement of F 2 against ALL reflections. The weighted R-factor wR and goodness of fit S are based on F 2 , conventional R-factors R are based on F, with F set to zero for negative F 2 . The threshold expression of F 2 > σ(F 2 ) is used only for calculating Rfactors(gt) etc. and is not relevant to the choice of reflections for refinement. R-factors based on F 2 are statistically about twice as large as those based on F, and R-factors based on ALL data will be even larger. | 2016-05-15T11:25:10.095Z | 2011-10-29T00:00:00.000 | {
"year": 2011,
"sha1": "a004b148997d9776dbe5f32c7b869d652b94895d",
"oa_license": "CCBY",
"oa_url": "http://journals.iucr.org/e/issues/2011/11/00/bt5667/bt5667.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "3e67a622656bd4209b795ffa57749a228a454553",
"s2fieldsofstudy": [
"Chemistry"
],
"extfieldsofstudy": [
"Chemistry",
"Medicine"
]
} |
118382883 | pes2o/s2orc | v3-fos-license | Orientational bond and N\'eel order in the two-dimensional ionic Hubbard model
Unconventional phases often occur where two competing mechanisms compensate. An excellent example is the ionic Hubbard model where the alternating local potential $\delta$, favoring a band insulator (BI), competes with the local repulsion $U$, favoring a Mott insulator (MI). By continuous unitary transformations we derive effective models in which we study the softening of various excitons. The softening signals the instability towards new phases that we describe on the mean-field level. On increasing $U$ from the BI in two dimensions, we find a bond-ordered phase breaking orientational symmetry due to a d-wave component. Then, antiferromagnetic order appears coexisting with the d-wave bond order. Finally, the d-wave order vanishes and a N\'eel-type MI persists.
Unconventional phases often occur where two competing mechanisms compensate. An excellent example is the ionic Hubbard model where the alternating local potential δ, favoring a band insulator (BI), competes with the local repulsion U , favoring a Mott insulator (MI). By continuous unitary transformations we derive effective models in which we study the softening of various excitons. The softening signals the instability towards new phases that we describe on the mean-field level. On increasing U from the BI in two dimensions, we find a bond-ordered phase breaking orientational symmetry due to a d-wave component. Then, antiferromagnetic order appears coexisting with the d-wave bond order. Finally, the d-wave order vanishes and a Néel-type MI persists. Searching for unconventional states of matter and nontrivial elementary excitations (quasiparticles (QPs)) is one of the crucial objectives of the research in strongly correlated lattice models. Outstanding examples range from the quasi long-range ordered Mott insulator (MI) in one dimension (1D) where the neutral spin-1/2 particle "spinon" represents the elementray excitation to quantum spin ice in three dimensions (3D) where magnetic monopoles represent the QPs. In order to find unexpected phases, it is a good idea to focus on parameter regions where two antagonists compensate because then subtle subleading mechanisms can take over.
The present article addresses the IHM in two dimensions (2D) on the square lattice in order to understand which phases possibly arise between the band insulator (BI) and the MI. Its Hamiltonian reads H = δ 2 r,σ (−1) i+j n r,σ + U r (n r,↑ − 1 2 )(n r,↓ − 1 2 ) + t rr ′ ,σ (c † r,σ c r ′ ,σ + h.c.), (1) where r := ix + jŷ spans the square lattice, c † r,σ (c r,σ ) is the fermionic creation (annihilation) operator at site r with spin σ, and n r,σ := c † r,σ c r,σ is the occupation operator. The sum over rr ′ restricts the hopping to nearest-neighbor sites. Initially, the IHM was introduced to describe the neutral-ionic transition in mixed-stack organic compounds 1 and was later used for the description of ferroelectric perovskites 2 .
In 1D, it is well understood that the BI at small Hubbard interaction U is separated from quasi-long-range ordered MI at large U by an intermediate phase with alternating bond order (BO) 3 . The position of the two transition points (U c1 from BI to BO and U c2 from BO to MI) 4,5 and the excitation spectrum 6,7 of the model are determined quantitatively. We highlight that the transition to the BO phase is signaled by the softening of an exciton 3-5 located at momentum π (setting the lattice constant to unity) 6,7 .
In 2D, the limiting BI and MI phases are expected at low and high U , respectively. The MI at large U shows long-range antiferromagnetic order of Néel type (AF) 8 We study the 2D IHM by continuous unitary transformations (CUTs) 12 realized in real space up to higher orders in the hopping t. The flow equations are closed using the directly evaluated enhanced perturbative CUT (deepCUT) 13 . In this way, an effective model is derived in terms of the elementary fermionic QPs. The virtual processes are eliminated as in the derivation of a t-J model from a Hubbard model [14][15][16][17] . The effective model allows us to address the interaction of two of these QPs. A particle and a hole attract each other and may form a bound state, an exciton. The softening of such excitons indicate the instability of the phase towards a condensation of these excitons. In this fashion, we determined the BI-to-BO transition value U c1 of the 1D IHM within 0.3% 6 . We stress that already the second order calculation in the hopping t yields qualitatively the correct result, i.e., the correct symmetries for the condensate phase, in spite of the very small energy scales t 2 /U (relative to δ) driving the transition 6,7 . Thus we adopt the same strategy in 2D where the CUT has been successfully used before 18,19 , in particular for binding phenomena 20 .
We are using the double 1-particle gap ∆ c := 2(E N +1 0 − E N 0 ), the singlet exciton gap ∆ e := E N 1,S=0 −E N 0 , and the spin (triplet) exciton gap ∆ s : stands for the ground state energy and E N 1,S for the first excited state with total spin S for N electrons. At halffilling, the number of electron N equals the number of lattice sites. In the absence of any electron-electron bound state, ∆ c , being twice the minimum of the fermion dispersion, equals the charge gap. We use the term charge gap because it has been used in previous papers on the IHM 4-7 . Fig. 1 summarizes our key results: Upon increasing U in the BI, the exciton gap ∆ e reduces till it vanishes at U c1 where d-wave bond order (dBO) sets in. In this phase, the 90 • rotation symmetry of the square lattice is broken and the d-wave bond order parameter d σ becomes finite. Then, the spin gap ∆ s closes at U c2 beyond which the staggered Néel magnetization m s becomes finite and the d-wave order parameter d σ starts depending on spin (dAF). Eventually, the rotational symmetry is restored at U c3 (d σ = 0) while the Néel order persists (AF).
Technically, we proceed as in 1D 6 by transforming the half-filled IHM into a translationally invariant problem. On the odd sublattice (i + j odd), we apply the electron-hole transformation T (eh) : c † r,σ → h r,σ . For consistency, we then use for particle and hole creation the same operator f ( †) r,σ . The local phase transformation The phase transformation T (l) induces a shift in momentum space by π/2 := (π/2, π/2) for single f -fermions.
The electron-hole transformation T (eh) modifies the spin and the charge on the odd sublattice according to (3) Hence, the AF phase in the original c-fermions appears as translationally invariant ferromagnet in the ffermions. The staggered magnetization reads m s := We apply the deepCUT to (2) eliminating processes which let single f -fermions and pairs of them decay. We disentangle the one-QP and the two-QP subspaces from the remaining Hilbert space by a suitable change of basis such that the 1-QP dispersion and the dispersion of bound states such as excitons become accessible. Technically, we employ the reduced generator n : 2 21 targeting the 0-, 1-, and 2-QP sector 13 . The sign of the terms in the generator, i.e., the sense of rotation, is chosen first according to the change in the number of double occupancy (increase of double occupancy leads to the same sign in the generator as in the Hamiltonian). If this number is not changed, the change in the number of fermions determines the sign. In this way, we map the Hamiltonian (2) to the effective Hamiltonian where H 2 2 stands for interaction terms of two creation and two annihilation operators with coefficients C 2 2 . The effective coefficients C 1 1 and C 2 2 are determined from the flow equation ∂ ℓ H = [η, H] 12,13,21-23 . Higher particle interactions are ignored. The range of processes in (4) is limited by the order in t of the deepCUT. The calculations are restricted to order 6 because for higher orders overlapping continua prevent a well-defined convergence of the flow 21 . Thus the effective Hamiltonian is not fully quantitative, but it is qualitatively correct. For an assessment of the accuracy see Refs. 6 and 7; technical aspects related to simplification rules are given in Appendix A.
The effective Hamiltonian (4) allows us to determine the ground state energy, the 1-QP dispersion, and bound pairs of 2 QPs. Due to the particle-hole transformation T (eh) excitons with odd distance between their con- Recall that the spin configuration shown is a particle-hole singlet in the original c-fermions.
stituents appear in the effective model as pairs of particles. The eigenvalues in the 2-QP subspace are found for fixed total momentum k, total spin S, and total S z component M as in the 1D case 6 . Fig. 2(a) depicts the low-energy spectrum in the BI phase consisting of the fermionic dispersion and a singlet exciton; no triplet exciton is found. Note the rather flat dispersion and lower continuum edge.
Very interestingly, the singlet exciton takes its minimum at k = (0, 0). In 1D, its minimum was at k = π so that its condensation led to dimerized bond order 3,6 . The fact that the singlet exciton softens at a different momentum clearly shows that the condensed phase beyond the BI will no longer be a dimerized BO. Figs. 2(b) and (c) display the charge and the singlet gap vs. U . Generally, the softening of a bosonic excitation implies the divergence of the susceptibility with the same symmetries χ(ω) at ω = 0 because χ(ω) corresponds to the bosonic propagator up to matrix elements. Thus it has a pole at the boson eigen energy moving to ω = 0 upon softening. The direct access to the energies of bound states avoids the necessity to guess the correct symmetry of the diverging susceptibility. In order 4, the singlet gap ∆ e vanishes at U c1 = 1.08δ where the charge gap ∆ c is still finite at 0.003δ. In order 6, we find U c1 = 1.126δ with ∆ c = 0.023δ. Considering all available orders and comparing to the 1D results 6 , we estimate that the transition point U c1 lies between U = 1.08δ and U = 1.126δ. The value of the charge gap at the transition point is also expected to be between the results in order 4 and order 6, similar to 1D findings 7 . Increasing t to 0.1δ the transition point shifts to larger values of U , see Appendix B.
The key observation is that the S = 0 exciton softens before the system would become metallic in analogy to the 1D case, but with other symmetries. The exciton condensation leads to a phase without magnetic order because the condensing exciton has no spin. The condensation does not break translational invariance because the exciton condenses at zero total moment. The analysis of the point group properties of the exciton reveals that its wave function becomes negative upon spatial rotation by 90 • so that the condensed phase should display the same d-wave symmetry.
The mechanism for electron and hole to form a bound state with d-wave symmetry relies on the interaction f † r+x,σ f † r+x+ŷ,σ f r+x,σ f r,σ representing correlated hop-ping, see Fig. 3. Such terms, occurring already in the second order in t with positive sign, have a singlet on a bond inx-direction hop to a bond inŷ-direction 24 . This correlated hopping corresponds to the well-known threesite terms in the t-J model [14][15][16][17] . If the wave function has the opposite sign onŷ-bonds of those onx-bonds this correlated hopping leads to attraction and eventually to a bound state with d-wave symmetry. Similar to the 1D case 6 , we describe this condensed phase by a self-consistent BCS-type MF theory applied to (4), see Fig. 4(a) to the left of U c2 . This approach is exact for U ≤ U c1 where the quantum fluctuations are already considered by the CUT. In particular, the onset of order occurs precisely where the exciton energy vanishes as required by consistency. For U > U c1 the MF approach is a good approximation systematically controlled by the distance U > U c1 . The order of the calculations is 4 in powers of t and t = 0.05δ. Its bond order d σ consists in the difference of the bonds in x and in y direction implying d-wave character. We determine the expectation values f † 0,σ f d,σ and f † 0,σ f † d,σ self-consistently, see Appendix C, comprising also the magnetization m s and the d-wave order parameter d σ .
The singlet and the triplet excitons within the dBO are determined by analysing the 2-QP problem neglecting the off-diagonal terms linking 2-QP states to states of higher particle number, see Appendix C for the details. This leads to the gaps displayed in Fig. 4(b). The exciton gap ∆ e quickly increases till the exciton ceases to exist because it merges with the 2-QP continuum. In parallel, a triplet exciton is formed and softens. The results of order 6, given in Appendix D, are qualitatively similar. Hence the system becomes unstable towards the condensation of a magnetic bosons, i.e., towards magnetic ordering. The softening occurs at momentum π in the original model so that this triplon condensation is the expected Néel ordering.
We study the Néel ordering by allowing for spin dependent MF solutions of (4), i.e., m s becomes finite and d ↑ = d ↓ . In addition, the charge gap becomes also spin dependent ∆ ↑ c = ∆ ↓ c . We find that the d-wave order does not vanish, but that it persists so that antiferromagnetism and d-wave order coexist and a d-wave antiferromagnet (dAF) forms. This solution occurs at and beyond U c2 ≈ 1.083δ as shown in Fig. 4(a) and in the zoom (c). 25 So far, we constructed the phase diagram following the spontaneous symmetry breaking indicated by soft excitons. In the dAF phase, the symmetry is strongly reduced. For large U , however, we expect Néel ordering without d-wave components because the Hubbard model maps to the Heisenberg model since the ionic alternation is only a small perturbation for U ≫ δ. The MF solution retrieves this feature. The BO parameters d σ quickly decrease on increasing U . For U > 1.123δ, d ↑ is smaller than 10 −3 . We conclude that there is a U c3 beyond which the spatial rotation is restored as a symmetry of the IHM: the dAF phase becomes a Néel ordered AF phase. The precise value of U c3 is difficult to pinpoint because d ↑ (U ) is consistent with an exponential decrease ∝ exp(C/(U − U c3 )) with U c3 ≈ 1.24δ. This resembles a Kosterlitz-Thouless type of transition as also proposed in Ref. 10. But the numerics is too involved and the energy scales too small to identify the decrease of d ↑ (U ) unambiguously.
In Fig. 4(c), we also plot the spin-dependent charge gaps in the vicinity of U c2 . Beyond U c2 , ∆ ↓ c increases quickly, but ∆ ↑ c decreases and hits zero at a half-metallic point U HM . Such half-metallic points occurred in previous analyses already 11 . The spin-up charge gap is located at momentum k = π/2 up to U HM , but for U > U HM it moves to incommensurate momenta at the line k x = k y and remains zero. But we cannot exclude that the zero ∆ ↑ c for U > U HM is an artefact of our approach. Any small deviation from the particle-conserving effective Hamiltonian (4) would spoil this feature. Thus we also included ω ↑ π/2 in Fig. 4(c) displaying a more generic behavior.
Summarizing, we studied unconventional phases in the ionic Hubbard model in 2D between the known band insulator at weak interaction and the Néel ordered Mott insulator at strong interaction. We found bond order which spontaneously breaks the orientational symmetry: d-wave bond order (dBO). Even a region of coexistence of bond order and Néel order is identified: d-wave bond order and antiferromagnetism (dAF). Thereby, we provide clarification for the so far ambiguous evidence concerning the influence of strong interactions on band insulators. We stress that no weak-coupling instabilities indicate the existence of these phases. This underlines that these phases are inherent to strong coupling 26 . Against this background, we think that the single-site DMFT 11 does not capture the spatial correlations indispensable for d-wave ordering and the DQMC 9 considers too small clusters at too large temperatures to capture the small energy scales driving the transitions found. The cluster DMFT 10 also advocated bond order, but combined with dimerization and incommensurability. Since the rotational symmetry was not allowed to be broken no dwave ordering could be detected.
We emphasize that the scenario discovered here is different from the bond ordered phases with d-wave symmetry discussed in the literature so far. These previous scenarios imply either a kind of alternation 26,27 or incommensurability 28,29 and they usually occur away from half-filling in contrast to our finding. The dBO phase presently advocated induces no dimerization and breaks only the rotational symmetry of the square lattice.
The conclusions are based on effective models expressed in elementary excitations. This approach was successful in 1D. No ad hoc assumptions about broken symmetries are made, but instabilities are systematically deduced from softening bound states, namely singlet and triplet excitons, the latter being the spin excitations. The quantum numbers and symmetries of the excitonic wave function implies the symmetry of the phase to which the instable phase evolves. In this way, two continuous phase transitions at U c1 and U c2 are identified. Based on symmetry considerations, we expect the first transition to be in the Ising and the second transition to be in the O(3) universality class.
We consider the obtained scenario to be complete because the d-wave order quickly decreases in the dAF phase so that only Néel order remains beyond a third transition interaction U c3 as expected for large interaction U .
The findings indicate for which types of correlations one has to look in order to identify intermediate phases in the wide research field between band insulating and Mott insulating phases, for instance on various lattices.
ACKNOWLEDGMENTS
We are grateful to the Helmholtz Virtual Institute "New states of matter and their excitations" for financial support.
The basic a-posteriori and a-priori simplification rules (SRs) introduced in Ref. 6 can be used for the IHM on the square lattice because they are lattice independent. In addition, we have implemented an extended a-posteriori SR. The aim is to estimate whether a monomial A contributes to the targeted quantities up to a specific order, for example n, or not. We explain the extended a-posteriori SR for the creation operators in A. The annihilation part can be analyzed in the same way.
Similar to the 1D case, one can treat spin-up and spindown operators separately 6 . From the position of creation operators with spin σ, we form a cluster (graph). We consider a vertex for each lattice site and add an edge between two vertices if the corresponding sites are adjacent. The degree of a vertex is defined as the number of edges linked to the vertex. The cluster is divided into a set of linked subclusters. Let us consider one of the subclusters and denote it by C. In the first step, we need to underestimate the number of commutations K[C] necessary to cancel the arbitrary linked cluster C.
The size of C is reduced in an iterative way. We eliminate each vertex with degree 1 and the vertex it is linked to. This costs one commutation for each of these pairs to be omitted. Let us suppose that after p C commutations no vertex with degree 1 is left. This reduces the linked cluster C to isolated vertices and/or to the linked subclusters C ′ which has no vertex with degree 1. Each isolated vertex requires one commutation to be canceled. We denote by s C the number of isolated vertices. One also needs v C ′ 2 commutations to cancel the subcluster C ′ which has v C ′ vertices. ⌈x⌉ is defined as smallest integer not less than x. Hence, we obtain where v r C := C ′ v C ′ . One can use the simpler Eq. (A1b) instead of (A1a) for K [C]. There should be at least two subclusters C ′ each one with an odd number of vertices to change the "equal" sign in (A1b) into "greater than". For the square lattice, this means that the initial linked cluster CF should contain at least 16 (=7+7+2) vertices. One can check that Eq. (A1) simplifies to the 1D counterpart where no subcluster C ′ can exist 6 . As example, let us examine the three clusters in Fig. 5. We find p C = 1, s C = 2, and no C ′ for Fig. 5(a), p C = 2, s C = 2, and no C ′ for Fig. 5(b), and finally p C = 1, s C = 0, and one subcluster C ′ with v C ′ = 4 for Fig. 5(c).
The total number of commutations K c 0 required to cancel all the creation operators of A is given by where the sum runs over all linked spin-up and spin-down clusters of monomial A. Similar to the rules presented in Ref. 13, Eq. (A2) can be extended to K c q where sectors up to q quasiparticles are targeted. The aim is to keep q operators such that the maximal order of A is overestimated 13 . We first keep the operators which can save one commutation. The total number of these operators is given by where we suppose that (A1b) holds for K [C]. The remaining operators (if q > d) can save one commutation per pair. Therefore, we find where d ′ := min(d, q). In the same way, one can analyze the annihilation part of the monomial A to find K a q . The monomial A can be discarded if where O min (A) is the minimal order of A 13 .
Appendix B: Results for Larger Hopping
In order to show that the results obtained in the main text also apply for larger values of the hopping. ear Hamiltonian is diagonalized in momentum space by a Bogoliubov transformation. The relevant expectation values are given by where and ω σ k := c σ The coefficients c σ d and Γ σ d are defined as prefactors of normal-ordered hopping and Bogoliubov operators over the distance d, respectively. Conservation of total charge guarantees that no hopping term over odd and no Bogoliubov term over even distances occurs. One can omit the spin indices from Eqs. (C1) because the SU(2) symmetry is preserved in the d-wave bond order (dBO) phase. The general spin-dependent form helps us to analyze also phases with broken spin symmetry. After some standard calculations, the Hamiltonian (4) be rewritten as where the operator λ ( †) k,σ is defined such that it diagonalizes the non-interacting part of the Hamiltonian. The coefficients of the interaction potential in momentum space C 2 2 are given in terms of the real space interactions C 2 2 and the hopping coefficients c k and the Bogoliubov coefficients Γ k . The "· · · " stand for the quartic off-diagonal interactions H 3 1 ∝ λ † λ † λ † λ and H 4 0 ∝ λ † λ † λ † λ † and their hermitian conjugates.
Taking the ground-state energy of the mean-field approximation as the true ground-state energy neglects the terms H 4 0 + H 0 4 . They are systematically controlled by U − U c1 , i.e., they are indeed small in the parameter regime considered. Considering the 1-particle dispersion ω k in Eq. (C2) and/or the interactions in the 2-particle sector above the mean-field solution, we also neglect the terms H 3 1 + H 1 3 which are again systematically controlled by U −U c1 . The singlet and the triplet 2-particle bound states in Fig. 4(b) of the main text are determined by exact diagonalization in the 2-particle subspace for fixed total momentum k, total spin S, and total magnetic number M . The remaining quantum number is the relative momentum q which is not conserved. There is, however, a subtle point to be noticed. In the 2-particle state |k; q σ1σ2 := λ † k/2+q,σ1 λ † k/2−q,σ2 |0 , states with two electrons (two holes) and one electron and one hole are mixed. In order to stay in the half-filled case, we define |k; q Q=0 σ1σ2 := 1 √ 2 (|k; q σ1σ2 +|k; π−q σ2σ1 ) which has the net total charge Q = 0. The relative momentum q is now restricted, e.g., to 0 < q x < π and π 2 < q y < π 2 . We employ triplet states (S = 1) with specific polarizations rather than specific magnetic numbers. The triplet states are given by |t x = 1 √ 2 (|↑↓ + |↓↑ ), |t y = −i √ 2 (|↑↓ − |↓↑ ), |t z = 1 √ 2 (|↑↑ − |↓↓ ), and the singlet state by |s = 1 √ 2 (|↑↑ + |↓↓ ). Constructing the Hamiltonian matrix in momentum space is time consuming compared to the real space diagonalization. This is the case because for each q i and q f in (C2), one has to perform three summations over weighted real space interaction coefficients to calculate In order to underline that the analysis in the main text is supported by results in higher order of the hop-ping. Figs. 7(a) and (b) represent the same quantities as Fig. 4(a) for U < U c2 and Fig. 4(b), but in order 6. Divergence of the flow equations prevents us to consider interactions beyond U = 1.155δ, so that we can not access the full phase diagram. Figure 7. The same as in Fig. 4(a) for U < Uc2 and Fig. 4(b), but in order 6 of the hopping. | 2016-05-18T10:50:07.000Z | 2015-09-16T00:00:00.000 | {
"year": 2015,
"sha1": "2911ba755362049da5fad04670d9c49b04a4fbac",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/1509.04996",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "2911ba755362049da5fad04670d9c49b04a4fbac",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
} |
104845302 | pes2o/s2orc | v3-fos-license | The course of combustion process under real conditions of work of a traction diesel engine supplied by mixtures of canola oil containing n-hexane
Restrictions for emission of toxic substances of exhaust gases force to conduct research aimed at searching for new types of fuels for powering combustion engines. With reference to diesel engines, the main focus was on acquiring fuels not from petroleum refining, the so-called alternative fuels. While examining these issues, the authors conducted various tests on an engine dynamometer in dynamic and static conditions, in which compression ignition engine was powered by mixtures of canola oil containing n-hexane (in various V/V proportions). Although the results were promising, they did not reflect the work of an engine in real road conditions. This article presents the results of research on traction diesel engine powered by mixtures of canola oil containing n-hexane under real operating conditions, which were reflected in real traffic conditions. Mobile system for engine indicating produced by AVL - Indi Micro 602, which was built in Fiat Qubo 1,3 Multijet, was applied during research. Various thermodynamic parameters of combustion process in various dynamic states, typical for the process of real work of an engine powered by examined mixtures of canola oil containing n-hexane were analysed and compared with parameters obtained while powering research vehicle with diesel fuel.
Introduction
Internal combustion engine is widely used in means of transport. It is projected for the years 2020-2050 that internal combustion engine will be the basic prime mover for cars [1]. Currently, there is a heated debate concerning further possibility of using internal combustion engines powered with diesel fuel. It is mostly connected with restrictions concerning the emission of toxic substances generated by this type of engines [2,3,4,5]. In order to meet the ecological standards, a number of research projects aim at searching for new types of fuel or modifying fuel system for internal combustion engines. It particularly applies to the parameters of compression and fuel injection as these parameters are crucial for the combustion process. With reference to compression ignition engines the research also concerns the use of the so -called alternative fuels which are not obtained from petroleum refining process [Błąd! Nie można odnaleźć źródła odwołania.,Błąd! Nie można odnaleźć źródła odwołania.,Błąd! Nie można odnaleźć źródła odwołania.]. While examining those issues, the authors of this study conducted various tests on an engine dynamometer in dynamic and static conditions, in which diesel engine was powered by mixtures of canola oil with various additives [Błąd! Nie można odnaleźć źródła odwołania.,Błąd! Nie można odnaleźć źródła odwołania.,Błąd! Nie można odnaleźć źródła odwołania.]. One of the choices was the application of a mixture of canola oil containing n-hexane (in various V/V proportions) [Błąd! Nie można odnaleźć źródła odwołania.,Błąd! Nie można odnaleźć źródła odwołania.,Błąd! 2 Nie można odnaleźć źródła odwołania.,Błąd! Nie można odnaleźć źródła odwołania.]. Although the results were promising, they did not reflect the work of an engine in real road conditions. The tests (static and dynamic) were run on an engine dynamometer, where the tested object was AD3.152 engine without forced induction or exhaust gas recirculation (EGR).
The following article presents the test results of a turbocharged engine with the Common Rail direct fuel injection with the EGR system which was fuelled with the mixture of canola oil with nhexane. The engine placed in a vehicle was subjected to indication in different conditions of real traffic. The conditions of the vehicle's movement were selected so that they represented static and dynamic conditions of the engine's work. The aim of the research was not the mutual quantitative reference of the obtained parameters of combustion process in static and dynamic conditions, but the demonstration of the tendencies concerning changes in the course of combustion process parameters for different engine operating conditions resulting from different variants of vehicle movement and being fuelled with tested mixtures. Selected thermodynamic parameters of the combustion process in real road conditions were analysed. The engine was fuelled with the mixture of canola oil with n-hexane andfor comparison -with diesel fuel.
The specificity of the course of the combustion process in the diesel engine fuelled with the mixture of canola oil with n-hexane
Canola oil applied as fuel has certain qualities more useful than diesel oil, the amount of oxygen in the fuel is particularly beneficial. Canola oil contains about 10% of oxygen and it has less carbon and hydrogen than mineral diesel fuel, that is why its calorific value is lower. It also has very good lubricating properties. For diesel engine obtaining required operational parameters such as mean effective pressure, rotation speed, overall efficiency and limiting fumes toxicity and engine noise, depends mainly on proper making of combustible mixture. The mixture needs to be characterised by specific homogeneity and fragmentation of fuel into drops of the smallest possible diameter. In the process of creating the proper combustible mixture, the most important factors are the fuel injection parameters which directly influence the fuel combustion. The combustion process may be divided into four stages such as: the first period of combustion, known as the ignition delay period; the second combustion period which covers the time from creating the first spontaneous combustion moment until the occurrence of the highest combustion pressure; the third period of combustion which commences the moment the pressure in the cylinder reaches its peak value and lasts until the combustion ends; and the fourth period of combustion also called after combustion meaning combustion lasting till expansion stroke. The characteristic feature of combustion in diesel engine is spontaneous fuel ignition preceded by the period of ignition delay, which has direct influence on the further combustion process [Błąd! Nie można odnaleźć źródła odwołania.,Błąd! Nie można odnaleźć źródła odwołania.]. The main focus is on the first stage of ignition delay on which basic parameters which refer to circulation efficiency and ecological parameters -such as the speed of heat development, the speed of pressure and temperature increase -depend mostly. The fuel injection process, which depends on control parameters of the injection system and certain physicochemical qualities of fuels, has a huge influence on ignition delay stage.
Physical properties essential from the point of view of the work of diesel engine are: density, viscosity and surface tension. These parameters greatly influence the diameter of the drops, the reach of the stream of sprayed fuel therefore, they affect the spontaneous ignition delay [21]. The chemical fuel parameters, however, depend on the structural composition of hydrocarbons, among which the most numerous group for diesel oil are paraffin hydrocarbons CnH2n+2 [Błąd! Nie można odnaleźć źródła odwołania.].
While applying plant fuel for powering compression ignition engine it is essential to take into consideration the fact that these engines were constructed for fuels of physicochemical properties characteristic for diesel fuel. It is crucial to aim at changing the physicochemical characteristics of plant fuels so that they are similar to the physiochemical properties of diesel fuels. This effect can obtained 3 by adding n-hexane as a solvent to plant fuels. As a result of data analysis and running preliminary tests, the optimal volume composition of the mixture of canola oil and n-hexane was estimated. Consequently, the following fuels were subjected to further analysis: -diesel fuel -Df -canola oil -Co -canola oil with 10% n-hexane -10%Hex90%Co.
The physicochemical parameters of canola oil with n-hexane were described -table 1 [21]. Mixtures of n-hexane and canola oil start to distil in the temperature of about 60°C and cease to distil in around 100-110 0 C. In the temperature above 250 °C decomposition of canola oil is visible (the amount of the distillate was adjusted to the amount of hexane added to the mixture). N-hexane additive greatly influenced the combustion temperature of the mixture mainly due to its volatility and low ignition temperature ( As mentioned before, in preliminary tests on the engine dynamometer a series of works was done [Błąd! Nie można odnaleźć źródła odwołania.,Błąd! Nie można odnaleźć źródła odwołania.,Błąd! Nie można odnaleźć źródła odwołania.], during which the parameters of the engine work in dynamic conditions and static conditions (imitated with the use of the external exploitative characteristic) were compared. The object of the study was a diesel engine AD3.152 with direct injection characterised with one constant fuel injection per cycle. On the basis of the tests it was stated that: for the mixtures of canola oil with n-hexane there was decrease in the values of mean indicated pressure, while together with the increase of n-hexane in the mixture with Co, those values were reaching those of diesel fuel. With respect to rotational speed up to about 1600 rotation per minute the angle of spontaneous ignition delay was bigger while powering the engine with mixtures of canola oil with n-hexane in comparison with diesel fuel. From rotational speed of about 1600 rotations per minute the tendency was reversed. for the mixtures of canola oil with n-hexane the maximum speed of pressure increase escalated.
With respect to the process of pressing and injection, the acceleration of the angle of the beginning of injection and slight increase of the maximum injection pressure was stated while applying the mixture of canola oil with n-hexane in comparison to applying diesel fuel. These phenomena were caused mainly by residual pressure increase in the high pressure conductors. The tested engine had a traditional rotational fuel injection pump. Due to the development of diesel engines, traditional injection systems (e.g. rotation pump) are replaced with direct fuel injection systems in which the injection strategy (several fuel injections per one work cycle) is immediately corrected by a controller. In the following article the authors intended to estimate the tendency in the changes of the parameters describing the combustion process in dynamic conditions -in comparison to static conditions -during the real vehicle movement in traffic conditions while powering the engine with the fuels tested. The conditions of the 4 engine's work reflected typical conditions of the vehicle functioning. The engine was equipped with Common Rail direct fuel injection. The research showed the influence of the conditions of the engine's work and the type of fuel used on (among others): mean indicated pressure, the maximum Speer of pressure increase, the amount of heat produced and the process of heat production. With respect to the injection process, the injection strategy monitored by the engine controller was described.
The research stand, the researched object
The research was conducted in the Laboratory of the Department of Car Vehicles at the University of Technology in Lublin. The object of tests was Fiat Qubo with a diesel engine fulfilling all emission standards specified by Euro 5 and a five speed gearbox. In the tested vehicle an additional fuel supply system, with an exchangeable fuel tank was installed. The main technical parameters of the tested car engine are shown in table 2. The system for indicating the engine Indi Micro 602 made by AVL with a built-in signal amplifier, cooperating with four analogue input channels and two digital inputs was used in tests. The signals recorded by the AVL Indimicro system are the following: • the course of pressure inside the cylinder -it was recorded by means of a piezoelectric sensor AVL GH13P installed in the glow plug seating of the first cylinder; • the signal of engine crankshaft position; • injection parameters were analysed on the basis of the analogue signal of controlling the electromagnetic injector after converting it into the digital signal. Figure 1 presents the test stand used during the research.
The research methodology
The diesel engine installed in the vehicle was subjected to induction in selected movement conditions reflecting static and dynamic conditions. The tests were run in the vehicle moving in real road traffic. The vehicle was subjected to motion resistance forces mainly rolling resistance force and air resistance force. During the tests, the high -frequency parameters of the engine's work were registered (those parameters were registered for 1000 consecutive cycles of the engine's work) for the following cases of the vehicle movement: • Case 1 -stop -engine neutral gear, conditions of static engine work with rotational speed from 800 to 4000 rpm.
• Case 2 -driving in the fourth gear on even, smooth surface of a straight, asphalt expressway (direct gear)-conditions of static engine work, constant load from rolling resistance forces; • Case 3 -starting the vehicle -on neutral gear, conditions of dynamic engine work;
4.Analysis of the obtained results
In the figures from 2 to 8, the obtained results of tests and calculations are shown. The course of producing heat and the amount of produced heat were analysed for the selected engine work cycles and the course of the rotational speed and mean indicated pressure in consecutive cycles of the engine work. Mean, minimum and maximum values and standard deviations of mean indicated pressure, engine rotational speed, maximum combustion pressure and maximum speed of pressure growth were calculated from all the cycles of engine work. The conducted research refers to static conditions of the engine work (Cases 1 and 2) and dynamic conditions (Cases 3). The analysis of the test results conducted in static conditions of the engine work (case 1 and 2) showed that: In the conditions of driving in the fourth gear (case 2, picture 2B) the mean indicated pressure obtained with powering the engine with plant fuels (for all the registered cycles) was Lower than while powering the engine with diesel oil. Figure 2 shows the values of mean indicated pressure calculated from 1000 cycles of the engine work. During testing in road traffic conditions there was certain inconsistency concerning the regulation of the rotational speed. Despite the fact that all the tests were run on the same part of the road, there is a possibility of slight changes in the vehicle load and its engine with the movement resistance forces. Consequently, in order to analyse the results of the tests shown in picture 2B the cycles of the engine work with the most similar rotational speed were selected and further analysis was conducted on those cases. Those cycles were chosen on the basis of picture 3. Therefore, individual cycles of the engine work such as 214 and 267 with the rotational speed of about 3000 rotations per minute were analysed. For those cycles it was stated that the addition of n-hexane to canola oil had a positive effect on the values of the mean indicated pressure causing its increase: 0.786 MPa for cycle 214 and 0.673 MPa for cycle 267. Respective values for diesel oil were 0.532 MPa and 704 MPa and for canola oil: 0.564 MPa and 495MPa. Plant fuels were characterised by the fact that the beginning of the initial and main injection occurred earlier than for diesel fuel, which is shown in Figure 4. For the fuel with n-hexan 10%Hex90%Co the beginning of the initial injection occurred at 36.5 0 CA before TDC and was about 3.8 0 CA earlier than for diesel fuel. For canola oil the beginning of initial injection occurred 1.4 0 CA earlier than for diesel fuel. A similar relation was observed during the main injection where the beginning of injection for 10%Hex90% Co-occurred 9.9 0 CA before TDC, 0.7 0 CA earlier than for diesel fuel. For canola oil the beginning of the main injection occurred 0.34 0 CA earlier than for diesel fuel. It was estimated that for the selected work cycle the main injection for the fuels with nhexane lasted 2.88 0 CA longer and it ended after the piston reached TDC point. For canola oil the time of the main injection duration was similar to diesel fuel. For all the fuels the duration of the initial injection was similar. For Case 1, with the rotational speed of about 3000 rpm there were three injections (two initial ones and one main one) for one cycle of the engine work. For all the fuels the time of the beginning of the injection occurred about 46 0 CA before TDC, while the main injection was 1.4 0 CA shorter for diesel than for the fuel with n-hexane and 1.6 0 CA shorter than for pure canola oil. In Case 2 ( figure 5) the average values of the maximum speed of pressure increase were higher for powering the engine with diesel fuel when compared with plant fuels. For the engine powered with fuel with n-hexane with the rotational speed of about 3000rpm the maximum speed of the pressure increase was 0.547 MPa/ 0 CA and was 0.053MPa/ 0 CA slower in comparison with diesel fuel. In Case 1 the tendency was reversed -figure 6. 9 of the engine work. C. Minimum, maximum and mean values, average and standard deviation for mean indicated pressure (IMEP1), maximum combustion pressure (PMAX1), maximum speed of pressure growth (RMAX1) and rotational speed of the engine (SPEED)for 1000 cycles of the engine work. A. B.
Figure 7.
A. Case 1 -neutral gear, rotational speed about 3000 rpm, the course of the heat production for the selected 699 cycle of the engine work B. case 2 -driving in real road traffic conditions the rotational speed about. 3000 rpm, the course of the heat production for the selected 214 cycle of the engine work. Additionally, the table shows minimum, maximum and mean values, average and standard deviation for heat production (Q1) for 1000 cycles of the engine work. Black colour -diesel fuel Df, green colour -canola oil Co, purple colour -canola oil with n-hexane 10%Hex90%Co.
The above cycles of the engine work were selected for analysis. The engine was powered with the tested fuels and for those cycles the course of heat production was described Q1. In both cases, the average values of heat production were higher for plant fuels in comparison with diesel fuel. The chart ( Figure 7), presents the courses of heat production for the analysed cycles of the engine work. For the engine in neutral gear (case 1, picture 7A), for diesel fuel the average value of heat production was 3.36 kJ/m 3 deg, for canola oil 3.83 kJ/m 3 deg and for 10%Hex90%Co 3.85 kJ/m 3 deg. For the engine while driving (case 2, picture 7B), for diesel fuel the average value of heat production 8.30 kJ/m 3 deg, for canola oil 9.35 kJ/m 3 deg and for 10%Hex90%Co 13.05 kJ/m 3 deg. In case of maximum values of heat production it was noticed that in Case 1 the relations are similar to those for average values. In Case 2 maximum values of heat production for plant fuels were lower than for diesel fuel ( Figure 7B). For plant fuels (especially for 10%Hex90%Co) high values of the course of heat production were more in the range of the angle of the crankshaft, which can be contributed to a longer injection time. Figure 8. Case 3 -starting the engine in neutral gear. The chart presents the course of the mean indicated pressure (IMEP1), the engine's rotational speed (SPEED). The table shows minimum, maximum and mean values, average and standard deviation for mean indicated pressure (IMEP1), maximum combustion pressure (PMAX1), maximum speed of the pressure growth (RMAX1)and rotational speed (SPEED) for 1000 cycles of the engine work. Black colour -diesel fuel Df, green colour -canola oil Co, purple colour -canola oil with n-hexane 10%Hex90%Co.
Conclusions
Applying pure canola oil for powering compression ignition engines is difficult due to its physiochemical properties, especially its viscosity, particularly in low temperatures. Adding a nonreactive solvent -n-hexane -causes a positive change of the physicochemical qualities of the mixture, especially with respect to density and viscosity. The observation of the courses of parameters of the injection process and combustion parameters shows that: For pure canola oil (Co) and for the oil with n-(10%Hex90%Co) fuel injection is earlier than for diesel fuel (Df) Adding 10% n-hexane to canola oil improved the physicochemical properties of the fuel in comparison with diesel, which may be observed on the example of the course of the parameters of combustion process, particularly in dynamic conditions; the values of the mean indicated pressure (IMEP) were similar to those of diesel fuel but the values of the heat of combustion were bigger (Q1). At the same time, the values of maximum speed of the pressure growth were lower. | 2019-04-10T13:12:18.068Z | 2018-10-11T00:00:00.000 | {
"year": 2018,
"sha1": "9b8d45a6ff1b296462105bcf7a27a15b4e5836bf",
"oa_license": null,
"oa_url": "https://doi.org/10.1088/1757-899x/421/4/042050",
"oa_status": "GOLD",
"pdf_src": "IOP",
"pdf_hash": "5305cbf9cf46e1593f7315cfa64d7fc463418fba",
"s2fieldsofstudy": [
"Engineering"
],
"extfieldsofstudy": [
"Materials Science"
]
} |
231589838 | pes2o/s2orc | v3-fos-license | CDCA4 suppresses epithelial–mesenchymal transtion (EMT) and metastasis in Non-small cell lung cancer through modulating autophagy
Background Cell division cycle associated 4 (CDCA4) has been reported to be engaged into the progression of several cancers. The function of CDCA4 in Non-small cell lung cancer (NSCLC) was unknown. We aimed to explore the critical role of CDCA4 in NSCLC. Methods CDCA4 stably knocking down and overexpression cell lines were established and Western blotting assay was applied to measure relevant protein expression of Epithelial-Mesenchymal Transtion (EMT) and cell autophagy. Staining of acidic vacuoles, transmission electron microscopy and immunofluorescence staining were employed to detect autophagy. The ability of cells to migrate and invade were detected by Transwell migration and invasion assays. The interaction of CDCA4 with CARM1 was identified by immunoprecipitation and Western blotting analysis. Results In the present study, it was found that inhibition of CDCA4 induced EMT, migration and invasion of NSCLC cells while inhibiting autophagy of NSCLC cells. Meanwhile, overexpression of CDCA4 in NSCLC cells showed the opposite function. More importantly, the inhibition of autophagy could promote the EMT, migration and invasion of NSCLC cells, which should be impaired via the activation of autophagy. In addition, CDCA4-inhibited EMT, migration and invasion could be partially aggravated by autophagy activator, rapamycin, and reversed by autophagy inhibitor, 3-MA. Correspondingly, the application of rapamycin or 3-MA to CDCA4 knockdown cells showed the opposite effects. Further investigation suggested that CDCA4 could interact with coactivator associated arginine methyltransferase 1 (CARM1). Autophagy was induced while cell migration and invasion were inhibited in CARM1 knockdown cells. CDCA4 could suppress the protein expression CARM1 and knocking down of CARM1 could alter cell autophagy, migratory and invasive abilities regulated by CDCA4. Conclusion All data indicated that CDCA4 inhibited the EMT, migration and invasion of NSCLC via interacting with CARM1 to modulate autophagy.
Background
Lung cancer is the most commonly diagnosed cancer with the highest mortality among all cancers worldwide [1]. Although diagnostic and therapeutical techniques have been developing in recent years, lung cancer is predicted as a major disease burden continually well through the first half of this century [2]. Non-small cell lung cancer (NSCLC) is the most common form of lung cancer [3]. As cancer recurrence and metastasis are common incidents for patients with late stage of NSCLC, an improved understanding of the molecular mechanisms underlying NSCLC metastasis is of great significance to propose better therapeutic strategies, which are urgently required for prolonging patients overall survival.
Cell division cycle associated (CDCA) gene family serve as important regulators in many biological process, such as cell cycle [4]. Previous studies had indicated that some members of this gene family might have potential diagnostic and porgnostic values for both ovarian cancer [5] and hepatocellular carcinoma [6]. CDCA4, also known as HEPP/SEI-3/TRIP-Br3, is a member of CDCA gene family. Recently, several studies have shown that CDCA4 played various roles in different cancer types. In breast cancer CDCA4 could enhance cancer cell proliferation and reduce apoptosis [7]. Its role as cancer promoter was also reported in hepatocellular carcinoma [8]. However, CDCA4 worked as a tumor suppressor gene in cervical cancer as the knockdown of CDCA4 could promote cell proliferation [9]. The role of CDCA4 in NSCLC has not been well explored. Our study aims at filling in the gap in understanding the contribution of CDCA4 to NSCLC progression, which may provide basic molecular target for developing new therapeutic strategies for NSCLC.
Tumor metastasis is the leading cause of cancer deaths and regarded as an exacerbation of cancer progression. In the early stage of metastasis, cancer cells progress through the epithelial-to-mesenchymal transition (EMT), a process during which epithelial cells obtain semblable properties of mesenchymal cells with altered adhesive properties and motility [10] and thereby achieve increased migratory probability to invade surrounding tissues [11]. Autophagy is a conservative catabolic process wherein dysfunctional or damaged organelles and proteins are metabolically degraded as to maintain macromolecule synthesis [12]. Previous studies have demonstrated that autophagy functioned in cancer progression as a double-edged sword, as it may inhibit tumors in the development of cancers but enables tumor cell survival in stress [13]. Recently several researches indicate that autophagy is intricately involved in tumor metastasis, both with a EMT-promoting or EMT-restraining function [14][15][16]. However, the role and mechanism of CDCA4 in regulating autophagy during NSCLC metastasis are completely unclear.
In this study, we examined the functional relevance of CDCA4-associated autophagy in NSCLC, with a focus on EMT, tumor migration and invasion. Our results revealed that overexpression of CDCA4 could interact with coactivator associated arginine methyltransferase 1 (CARM1) and suppress its expression in NSCLC cells to induce autophagy, and as a result interfere with EMT and metastatic progression. Our data demonstrate for the first time that CDCA4-induced autophagy inhibits tumor cell migration and invasion in NSCLC, suggesting that CDCA4 may be a potential therapeutic target for NSCLC.
Cell culture
Human Non-small cell lung cancer cell lines A549 and H1299 were obtained from Shanghai Cell Bank of the Chinese Academy of Sciences and cultured in RPMI-1640 medium with 10% fetal bovine serum (FBS; WISENT, Inc.). The cells were maintained in an environment of 5% CO 2 at 37℃. Based on this culture environment, the medium of the cells was changed every three days.
Quantitative PCR
Total RNA from cells was extracted using Takara Min-iBEST Universal RNA Extraction Kit (Takara Biotechnology CO., Ltd) according to the manufacturer's protocol. All PCR primers were designed and synthesized by Sangon Biotechnology Co. (Shanghai, China), according to the gene sequences in Genbank. After the concentration and quality of RNA was determined by 260/280 nm absorbance, PrimeScript RT Master mix (Takara Biotechnology CO., Ltd) was applied to reverse transcribe RNA into high-capacity DNA. Quantitative PCR reaction was performed on an ABI 7300 PCR System (Applied Biosystems, CA, USA) using a SYBR Green Master Mix (Thermo Fisher Scientific, USA). Relative expression was normalized to β-Actin as an internal control and calculated using 2 −△△ CT.
Western Boltting analysis
Cells were lysed with ice-cold radio-immunoprecipitation assay (RIPA) buffer (Beyotime Biotechnology, Jiangsu, China) containing proteinase inhibitors. The cell lysate was scraped and centrifuged for 20 min, 12000g at 4 ℃. The supernatants were retained and concentrations were determined using a BCA protein assay kit (Beyotime Biotechnology, Jiangsu, China). Equal amounts of protein were loaded in SDS-PAGE gels and separated by electrophoresis. After electrophoresis, the proteins were transferred on a polyvinylidene fluoride (PVDF) membrane. Then the membranes were blocked at room temperature for 1 h with 5% BSA, and then incubated with anti-CDCA4 antibody (cat. no. 11625-1-AP; washed in TBS-T, membranes were incubated with an HRP-conjugated secondary antibody for 1 h at room temperature. Then the membranes were visualized by chemiluminescent reagents (Millipore, USA).
Trans-well assay
Certain number of 2 × 10 4 cells/well were resuspended in 200 µl of serum-free medium in the upper chamber (8-µm pore size, Coster, Corning, USA) and the lower chamber was filled with 600 µl of medium supplemented with 10% FBS. After incubation for 24 h (migration assay) or 48 h (invasion assay) at 37℃, the cells in the parietal chamber were removed with a wet cotton swab and cells on the submucosal surface were immobilized in 4% paraformaldehyde for 30 min. Then the cells were stained with a crystal violet solution. The number of migrated cells was quantified under a microscope by counting those in four random fields of each membrane. Three independent experiments were performed.
Lentiviral production and stable cell line construction
Lentiviral vectors expressing shRNA and CDCA4 were obtained from Vigene Biosciences China. A549 and H1299 cells were cultured overnight on 6-well plates then transduced by the above lentiviruses. After incubating for 24 h, cells were selected with 2 mg/ml puromycin for A549 cell and 4 mg/ml puromycin for H1299 cell for 3 days. Stable cell lines were harvested.
Staining of acidic vacuoles
Cells were cultured overnight on 6-well flat cover slides, washed with PBS for three times then stained with 1 µM AO (Solarbio, Beijing China) for 15 min at 37 ℃. Then cells were washed with PBS for three times and observed with Axio scope A1.
TEM (Transmission Electron Microscopy)
A549 cells were digested and then fixed with 2.5% glutaraldehyde overnight at 4℃. After washing the cells with PBS (0.1 mol/L) for 4 times (15mins/time), 1% osmium tetroxide was used to fix the cells for 2 h at 4℃ and acetone (50, 70, 90 and 100%) for 15 min each was used to dehydrate the cells. After dehydration, samples were permeated with 100% acetone and epoxy resin (1:1) for 2 h and then 100% acetone and epoxy resin (1:2) for 2 h, following by permeated with pure epoxy resin overnight then embedded in epoxy resin (Epon 812) and polymerized in drying oven (12 h at 37 ℃, 12 h at 45 ℃ and 48 h at 60 ℃). Ultra-thin sections (60 nm) were prepared using an ultramicrotome (Leica Ultracut R) and stained with uranyl acetate and lead nitrate. The images were captured at ×20,000 magnification on a Transmission Electron Microscope (JEM-1010, JEOL) installed in the Analysis and Testing Center in Nanjing Medical University.
Statistical analysis
Statistical analyses were performed using GraphPad Prism version 6 statistical software.All data shown represent the results obtained from triplicate independent experiments with mean ± SD. Student's t-test was used to evaluate comparisons between two groups. P < 0.05 was set to indicate a statistically significant difference.
CDCA4 suppresses EMT, migration and invasion in NSCLC cells
According to the relative expression quantity of CDCA4 on both mRNA and protein level (Fig. 1a, b), H1299 and A549 cell lines were selected to perform subsequent experiments as the abundance of CDCA4 was high in H1299 and medium in A549. We next establish CDCA4 stably knockdown H1299 cell line and CDCA4 stably overexpression A549 cell line. Western blotting and q-PCR confirmed the successful modulation of CDCA4 expression level in these cells (Fig. 1c, d). We then used these cells for in vitro analysis of NSCLC cell migration and invasion. Results showed that CDCA4 knockdown improved migrative and invasive abilities of H1299 cell, while CDCA4 overexpression impaired migration and invasion of A549 cell (Fig. 2a). As the phenomenon of epithelial-mesenchymal transition (EMT) commonly occur during tumor metastasis, we measured the expression pattern of classic EMT markers. Western blot assay results showed that in the shCDCA4 cells epithelial markers such as E-cadherin decreased while mesenchymal markers such as N-cadherin, Vimentin and Snail increased. The results were coincident in CDCA4-overexpressing cell lines, as the expression level of above EMT markers were reversed correspondingly (Fig. 2b).
Collectively, these findings suggested that CDCA4 was a blocker of NSCLC EMT, migration and invasion.
CDCA4 expression levels affect NSCLC autophagy
The association between tumor autophagy and tumor metastasis has been explored in many studies recently. As CDCA4 has previously been shown to serve as an autophagy regulator, we propose a hypothesis that autophagy played a crucial role in the process of CDCA4 regulating EMT, migration and invasion of NSCLC cells. Compared with cells transfected with vector, the formation of autophagosomes was increased in CDCA4 overexpressing A549 cells stained with acridine orange, as more amounts of acidic vacuoles could be counted in sight. In CDCA4 knockdown H1299 cells the formation of autophagosomes showed the opposite change (Fig. 3a). TEM showed the consistent phenomenon of activation of autophagic flux with accumulation of autophagic vacuoles containing intracellular material in CDCA4 overexpressing A549 cells (Fig. 3b). Moreover, an elevated immunofluorescence signal of LC3B and a weakened signal of P62 were observed in CDCA4 overexpressing A549 cell line compared with vector. The changes were correspondingly reversed in CDCA4 inhibition H1299 cells (Fig. 3c). As presented in Fig. 3d, western blotting assay showed that the expression of LC3B decreased and P62 showed the opposite change with the knocking down of CDCA4, along with other autophagy-associated markers such as BECLIN-1, ATG5 and ATG7 decreased. CDCA4 overexpressing cells showed consistent changes in protein levels of above markers. Overall, these results confirmed that CDCA4 positively modulated the autophagy pathway in NSCLC cells.
Autophagy modulates EMT of NSCLC cells
In order to explore the role of autophagy in EMT, A549 and H1299 NSCLC cells were treated with rapamycin or 3-MA, respectively known as autophagy activator and inhibitor, in indicated concentration for 48 h. The expression levels of LC3B increased markedly while P62 protein expression significantly decreased after cells treated with rapamycin (Fig. 4a). With the treatment of 3-MA for 48 h, the expression pattern of both A549 and H1299 reversed (Fig. 4c), indicating that rapamycin could stimulate cell autophagy while 3-MA could suppress cell autophagy in NSCLC cells. As shown in Fig. 4a, c, expression of EMT markers altered in cells treated with rapamycin or 3-MA. In addition, Trans-well assay indicated that activation of autophagy blocked cell migration and invasion while inhibition of autophagy had the opposite effect (Fig. 4b,
Effects of autophagy activation on CDCA4-modulated EMT, migration and invasion in NSCLC cells
To investigate whether autophagy is the pathway through which CDCA4 restained EMT and consequently blocked the migration and invasion of NSCLC cells, we treated CDCA4 overexpressing A549 cells or CDCA4 knockdown H1299 cells with autophagy agnosit respectively. As the primary target of rapamycin (mTOR) is regarted as a key regulator of autophagy, We chose to use rapamycin treatment with the aim of inducing autophagy in treated cells. It was observed that with the treatment of rapamycin, the increased LC3B protein expression while decreased P62 protein expression induced by CDCA4 overexpression were aggravated in A549 cells (Fig. 5a). Moreover, the weakened ability of migration and invasion in CDCA4 overexpressing A549 cells was further diminished by application of rapamycin (Fig. 5b). Western blotting assay showed decreased expression of N-cadherin as well as dramatically downregulated Snail expression, while increased expression of E-cadherin (Fig. 5a). In CDCA4 knocking down cells, the treatment of rapamycin showed a reversed effect on decreased LC3B protein expression and increased P62 protein expression (Fig. 5c). Together with Trans-well assay verified that the shCDCA4-improving abilities of migration and invasion were weakened in H1299 cells (Fig. 5d). Meanwhile, more significant gap in protein expression differences of EMT markers also assisted in verifying that rapamycin could partly reverse CDCA4 knockdown-promoted EMT, migration and invasion in H1299 cells (Fig. 5c). Taken together, our results demonstrated that EMT, migration and invasion inhibited by CDCA4 could be aggraveted by arousing autophagy in NSCLC, while downregulation of CDCA4 promoted.migration could be blocked via autophagy activation.
The inhibition of autophagy reverses CDCA4-suppressed EMT, migration and invasion in NSCLC cells
As we have proved that activation of autophagy had an influence on CDCA4-regulated EMT, migration and invasion in NSCLC, the flipside of autophagy was also investigated using 3-methyladenine (3-MA), which is a class III phosphatidylinositol 3-kinase (Ptdlns3K) inhibitor, to treat both A549 and H1299 cells. In A549 cells, the CDCA4 overexpression-midiated anti-metastasis effects were reversed when cells were treated with the autophagy inhibitor 3-MA (Fig. 6b). In addition, changes in protein expression of assocaited EMT markers were exhibited in Fig. 6a. In H1299 cells co-treated with shCDCA4 and 3-MA, western blotting assay showed further downregulation of LC3B expression and upregulation of P62 expression (Fig. 6c). Enhanced cell migration and invasion were observed in CDCA4 knocking down cells but not control cells (Fig. 6d), together with significantly increased protein expression of markers of mesenchymal cells ( N-cadherin and Snail) and decreased expression of E-cadherin, which marks epithelial cells (Fig. 6c). These data indicated that CDCA4-suppressed EMT, migration and invasion in A549 and H1299 cells could be reversed by the inhibition of autophagy. Collectively, our results suggested that the inhibitory effects of CDCA4 on the migration and invasion abilities of NSCLC cells likely worked through the autophagy-EMT pathway.
CDCA4 interacts with CARM1 and suppresses CARM1 protein expression
We next explored the underlying molecular mechanism of CDCA4 inducing autophagy through performing IP assay with CDCA4 overexpressing A549 cells and subjected the binding proteins to mass spectrometry. After literature search of regulatory roles in autophagy and CO-IP assay to verify the combinational relation with CDCA4, we confirmed that CARM1 as a binding partner of CDCA4 (Fig. 7a) and there was possibility that CDCA4 could regulate autophagy through interacting with CARM1. Western blotting assay showed that CDCA4 could negatively regulate the protein expression of CARM1 (Fig. 7b) and after knocking down CARM1 with siRNA, the decreased expression of P62 and increased expression of LC3B were shown in both A549 and H1299 cell lines (Fig. 7c, d). These data indicated that CARM1 served as a suppressor of autophagy in NSCLC cells, suggesting that CDCA4 may induce autophagy through suppressing the protein expression of CARM1.
With the inhibition of migration and invasion in CARM1
(See figure on next page.) Fig. 4 Activation or inhibition of autophagy alters the EMT, migration and invasion of NSCLC cells. A549 and H1299 cells were treated with dimethyl sulfoxide (control) or rapamycin (100 nM) for 48 h. a Western blotting analysis of P62, LC3B, E-cad, N-cad and Snail protein expression. β-actin was used as a loading control. b Comparison of the migration and invasion of A549 and H1299 cells using transwell compartments. A549 and H1299 cells were treated with phosphate-buffered saline (control) or 3-methyladenine (5 mM) for 48 h. c Western blotting analysis of P62, LC3B, E-cad, N-cad and Snail protein expression. β-actin was used as a loading control. d Comparison of the migration and invasion of A549 and H1299 cells using transwell compartments. Data are means ± S.D. *P < 0.05, **P < 0.01, ***P < 0.005 knocking down cells (Fig. 7e), we proposed a hypothesis that CDCA4 could suppress NSCLC migration and invasion through downregulating CARM1 to promote autophagy.
CARM1 impacts the ability of CDCA4 to control NSCLC migration, invasion and autophagy
To explore if CARM1 was the downstream functional molecule through which CDCA4 could control NSCLC migration, invasion and autophagy, we performed rescue assay via knocking down CARM1 with specific siRNA in both CDCA4 overexpressing cells and CDCA4 knocking down cells to see alterations in migration and autophagy. Results showed that in CDCA4 overexpressing cells, CARM1 knocking down further increased LC3B expression as well as decreased P62 expression (Fig. 8a). Together with autophagy promotion was the more evidently inhibition of migration and invasion (Fig. 8b). Meanwhile, the downregulation of CARM1 could reverse both the inhibition of autophagy or the promotion of NSCLC migration and invasion caused by CDCA4 knocking down (Fig. 8c, d). These data demonstrated that CARM1 could impact the ability of CDCA4 to induce autophagy and suppress migration or invasion in NSCLC, suggesting that CARM1 may serves as the downstream functional molecule of CDCA4. CARM1 has been reported exerting a transcriptional co-activator function on autophagy-related genes under glucose starvation condition. However in this study, we found that with CARM1 knocking down the alterations in mRNA level of a few common autophagy-associated genes showed no consistency in both A549 and H1299 cell lines (Fig. 8e). This suggested that CARM1 may not served as a transcriptional regulator in NSCLC cells under nutritional sufficient conditions. The underlying mechanism of CARM1 regulating autophagy requires further investigation.
Discussion
Cell division is a process during which large numbers of cells are produced for organism living. However, abnormal cell division can contribute to a variety of cancer-promoting errors. The accumulation of these errors could further promote cell division under permissive extracellular environments, and eventually trigger the formation and growth of abnormal cell populations that initiate tumorigenesis [17]. As a result, it is inferred that cell division associated proteins are of great significance in digging in underlying mechanisms of oncogenesis and development of malignant tumors. Cell division cycle associated protein 4 (CDCA4) has been proved as an potential tumor promoter in various cancer types. As an E2F transcription factor family-induced nuclear factor, CDCA4 could function in regulating E2F-dependent transcriptional activation and cell proliferation [9]. Besides, CDCA4 was characterized as a regulator of p53 transcriptional activity but regulates cell growth in a p53-independent way [18]. Although previous studies have demonstrated that CDCA4 could promote cell proliferation in human breast cancer and maligant melanoma, its other potential influence on tumor progression is still unclear. In addition, the potential effects of CDCA4-induced stimulation of transactivation function of p53 are also unexplored. Herein we specifically assessed the role of CDCA4 in NSCLC metastasis and investigated the underlying mechanism. We found that knockdown of CDCA4 enhanced its ability in promoting migration and invasion in an EMT-dependent manner through a mechanism at least partially dependent upon the inhibition of autophagy, with CDCA4 overexpression showing the opposite effect. Beyond that, our results demonstrated that CDCA4 had the ability of inhibiting metastasis in NSCLC cells in a p53-independent manner, as the two NSCLC cell lines we chose to carry out experiments were A549, which contain wild p53 type, and H1299, in which p53 is deficient, and results suggested that CDCA4 had the same function in inhibiting migration and invasion in both cell lines. Although p53 could regulate a few key stages of metastatic progression such as cell migration and invasion [19], distinct p53 expression status in these two cell lines showed little different effects on NSCLC metastasis, indicating that p53 pathway is not the key pathway through which CDCA4 functions as a cancer metastasis regulator. Our further data demonstrated that CDCA4 could inhibit cell migration and invasion via autophagy pathway. To our knowledge, this is the first report specifically report the function of CDCA4 in NSCLC metastasis and its regulatory role in regulating autophagy. In addition, Fig. 5 Effects of CDCA4 on EMT, migration and invasion could be influenced by autophagy activation. CDCA4 overexpressing and control A549 cells were treated with dimethyl sulfoxide (control) or rapamycin (100 nM) for 48 h. a Western blotting analysis was used to assess the expression of the indicated proteins in treated A549 cells. β-actin was used as a loading control. b Cell migration and invasion assays were performed using transwell compartments. CDCA4 knocking down and control H1299 cells were treated with dimethyl sulfoxide (control) or rapamycin (100 nM) for 48 h. c Western blotting analysis was used to assess the expression of the indicated proteins in treated H1299 cells. β-actin was used as a loading control. d Cell migration and invasion assays were performed using transwell compartments. Data are means ± S.D. *P < 0.05, **P < 0.01, ***P < 0.005 (See figure on next page.) this is the first report showing the relevance between CDCA4-inhibited migration and autophagy.
In previous studies, CDCA4 was proved as a tumor promoter in various cancers, representing as an attractive therapeutic target. Its role in promoting cancer cells proliferation in other cancer types have been studied extensively, little is known about the potential role of CDCA4 in cancer progression, especially in the metastasis of NSCLC. The purpose of this article is to explore the effects and underlying mechanisms of CDCA4 on the metastasis of NSCLC. After establishing CDCA4 stably knocking down or overexpression cell lines, we used Tran-swell assays to investigate the role of CDCA4 in migration and invasion. Our results showed that overexpression of CDCA4 could significantly impair NSCLC migration and invasion capability in vitro, whereas its downregulation could dramatically promote NSCLC migration and invasion in vitro. Accordingly these results identify a negatively-regulating role of CDCA4 in NSCLC metastasis. In recent years, increasing incidence of NSCLC implies the significance of identifying more accurate tumor biomarkers and improving current precise treatment strategies. Identification of key molecules of NSCLC metastasis can benefit not only in predicting the metastasis trend of primary tumor, according to which more specific excision extension in lung cancer resection as well as range of radiotherapy can be determined, but also in treating patients with targeted therapies when NSCLC recur with metastasis. Our studies have suggested that CDCA4 is associated with NSCLC migration and invasion in vitro, further studies in vivo should be done to prove CDCA4 as a potential biomarker of NSCLC metastasis.
EMT programmes have been pointed as integral components of the malignant progression of almost all types of carcinoma [20]. During this programme, cancer cells evolve the ability to migrate through the body cavity, with the morphological and functional transformation from epithelial cell to mesenchymal cell [21]. The properties of mesenchymal cell include reduced adhesion and enhanced motility, which allow cancer cells presenting an more aggressive and invasive phenotype [22]. Several molecular markers for EMT have been defined, such as increased expression of N-cadherin and Vimentin, increased production of the transcription factors like Snail and inhibited production of E-cadherin [23]. In this report, we found that overexpression of CDCA4 led to increased expression of E-cadherin and reduced N-cadherin, Vimentin and Snail expression in NSCLC cells, while downregulation of CDCA4 confers the opposite effects. This thus suggested that CDCA4 inhibited the EMT in NSCLC cells, highlight a highly potential link between CDCA4 and the metastasis of NSCLC.
Autophagy is a cellular spontaneous process during which intracellular proteins and organelles are captured and degraded as to sustain metabolism and homeostasis in the starved environment. In cancer, autophagy could be neutral, tumor-suppressive or tumor-promoting in different contexts [24]. Although in most contexts, autophagy facilitates tumorigenesis by supplying materials to cancer cells for their increased metabolic and biosynthetic demands imposed by deregulated proliferation, in some contexts autophagy also suppresses tumorigenesis as autophagy deficiency could cause oxidative stress, activation of the DNA damage response and genome instability, which are known cause of cancer initiation and progression [25]. Autophagy also play a critical role in regulating tumor cell migration and invasion, but the association between autophagy and EMT remains controversial. On the one hand, increasing evidence such as turnover of components of the cell migration machinery and EMT proteins has emerged to demonstrate that autophagy could promote tumor metastasis through inducing EMT [26], on the other hand autophagy modulation was reported with the effect of triggering a molecular switch from a mesenchymal phenotype to an epithelial-like one, consequently impaired migration and invasion in different contexts [27,28]. As it has been reported that CDCA4 served as a regulator of cell apoptosis via autophagy pathway in various cancers [29], we supposed that the regulatory mechanism of CDCA4 on cell migration and invasion partly contained regulation of autophagy. Specifically, our data demonstrated that CDCA4 overexpression upregulated the protein expression of classical autophagy marker such as LC3B along with downregulation of P62. Together with these changes common markers of autophagy pathway such as Beclin-1, ATG5 and ATG7 also altered correspondingly. These (See figure on next page.) Fig. 6 Effects of CDCA4 on EMT, migration and invasion could be affected by autophagy inhibition. CDCA4 overexpressing and control A549 cells were treated with phosphate-buffered saline (control) or 3-methyladenine (5 mM) for 48 h. a Western blotting analysis was used to assess the expression of the indicated proteins in treated A549 cells. β-actin was used as a loading control. b Cell migration and invasion assays were performed using transwell compartments. c,d. CDCA4 knocking down and control H1299 cells were treated with phosphate-buffered saline (control) or 3-methyladenine (5 mM) for 48 h. c Western blotting analysis was used to assess the expression of the indicated proteins in treated H1299 cells. β-actin was used as a loading control. d Cell migration and invasion assays were performed using transwell compartments. Data are means ± S.D. *P < 0.05, **P < 0.01, ***P < 0.005 Fig. 7 CDCA4 interacts with CARM1 and suppresses CARM1 expression. a CO-IP assays were performed in A549 CDCA4 overexpressing cells. b Western blotting. analysis was used to assess the expression of CARM1 and CDCA4 in CDCA4 overexpressing cells or knocking down cells. β-actin was used as a loading control. c Western blotting analysis and qPCR analysis of efficiency of knocking down CDCA4 in both A549 and H1299 cells. d Western blotting analysis of P62 and LC3B protein expression. β-actin was used as a loading control. e Cell migration and invasion assays were performed using transwell compartments. Data are means ± S.D. *P < 0.05, **P < 0.01, ***P < 0.005 (See figure on next page.) Fig. 8 CARM1 knockdown could alter the functions of CDCA4 both in inducing autophagy and suppressing migration. b. CDCA4 overexpressing and control A549 cells were transfected with siCARM1. a Western blotting analysis was used to assess the expression of the indicated proteins in treated A549 cells. β-actin was used as a loading control. b Cell migration and invasion assays were performed using transwell compartments. c.d. CDCA4 knocking down and control H1299 cells were transfected with siCARM1. c Western blotting analysis was used to assess the expression of the indicated proteins in treated A549 cells. β-actin was used as a loading control. d Cell migration and invasion assays were performed using transwell compartments. e qPCR analysis of seven autophagy-related genes. Data are means ± S.D. *P < 0.05, **P < 0.01, ***P < 0.005 results suggested that CDCA4 had the property of inducing autophagy. To further investigate whether autophagy had an effect on NSCLC cell migration and invasion, we respectively dealt cells with rapamycin as autophagy activator or 3-MA as autophagy inhibitor and then detect the migration capacity alteration of both cells. We found that in vitro, induction of autophagy resulting in inhibition of cell migration and invasion in NSCLC, and inhibition of autophagy showed the opposite results. Along with alterations in protein expression, our data demonstrated that autophagy regulation served as an EMT modulator in NSCLC cells. Further investigation was set on whether CDCA4 could inhibit cell migration and invasion through autophagy pathway. We found that inducing autophagy via overexpressing CDCA4 was linked to a partial EMT phenotypic reversion within NSCLC cells, which phenomenon can be strengthened by further activating autophagy, while impairing autophagy in these cells could reverse the observed defects in migration and invasion upon CDCA4 overexpression in vitro. Reducing autophagy via knocking down CDCA4 showed the opposite alteration of EMT markers in NSCLC cells, while 3-MA and CDCA4 knockdown synergistically enhanced the invasive effects of NSCLC cells. In addition, combination therapy significantly promoted EMT compared to treatment alone. As a result, we hypothesized that applying autophagy activators to CDCA4 high expression patients could reverse EMT and significantly prevent NSCLC cell invasion due to its enhanced anti-invasive effect.
CARM1 is known as a crucial component of autophagy in mammals with the transcriptional co-activator function on autophagy-related genes through transcription factor EB (TFEB) [30,31]. Through mediating methylated modification of BAF155 and PKM2, CARM1 could enhance breast cancer progression and metastasis [32,33]. In NSCLC, CARM1 also serves as a tumor promoter [34]. Here, we found that CDCA4 could interact with CARM1 and suppress the protein expression of CARM1. However, with CARM1 knocking down we found that autophagy was promoted in both A549 and H1299 cell lines, which is opposite to previous studies. This is a very interesting phenomenon and we inferred that it is because the condition in previous studies was glucose starvation but in our studies was nutritional sufficient condition. To verify CARM1 as a downstream functional molecule of CDCA4, we performed rescue assay and found CARM1 knockdown could alter effects on cell autophagy, migration and invasion caused by CDCA4. These results demonstrated that CDCA4 could induce autophagy through suppressing expression of CARM1, thus inhibiting NSCLC migration and invasion. Nevertheless, we also found that the knocking down of CARM1 did not alter mRNA expression of several common autophagy-related genes, which means that CARM1 may not serve as the transcriptional co-activator in NSCLC under our nutritional sufficient condition. The specific underlying mechanism of CARM1 inhibiting autophagy requires further investigation.
In conclusion, these results suggest a potential model whereby CDCA4 promotes NSCLC cell autophagy, thus inhibiting the migration and invasion of these cells. For future work, vivo experiments will be conducted to test whether the effect of CDCA4 on migration and invasion and its regulation of autophagy are consistent with the results of in vitro experiments. Overall, the present study not only suggests that CDCA4 could inhibit EMT, migration and invasion of NSCLC cells, but also provides a novel role for autophagy in these processes. In addition, the present study demonstrates the potential of identifying novel molecular targets, providing a basis for future investigations of more beneficial clinical strategies for NSCLC treatment.
Conclusion
The findings of this study demonstrate that CDCA4 serves as a tumor suppressor in NSCLC, with the function of inhibiting migration and invasion of NSCLC through modulating the autophagy pathway. As inactivation of tumor suppresser genes is an important mechanism of tumorigenesis and progression, this study may provide valuable reference for early diagnosis, monitoring progression and evaluating prognosis of NSCLC. | 2021-01-13T14:24:33.989Z | 2021-01-12T00:00:00.000 | {
"year": 2021,
"sha1": "f3958505c19aa19e053feab2b5f2d9c513eb436e",
"oa_license": "CCBY",
"oa_url": "https://cancerci.biomedcentral.com/track/pdf/10.1186/s12935-021-01754-w",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "51ac2c228faf40fbe49e95f672be394051897ada",
"s2fieldsofstudy": [
"Biology",
"Medicine",
"Chemistry"
],
"extfieldsofstudy": [
"Medicine"
]
} |
58663849 | pes2o/s2orc | v3-fos-license | Glycerol confined in zeolitic imidazolate frameworks: The temperature-dependent cooperativity length scale of glassy freezing
In the present work, we employ broadband dielectric spectroscopy to study the molecular dynamics of the prototypical glass former glycerol confined in two microporous zeolitic imidazolate frameworks (ZIF-8 and ZIF-11) with well-defined pore diameters of 1.16 and 1.46 nm, respectively. The spectra reveal information on the modified alpha relaxation of the confined supercooled liquid, whose temperature dependence exhibits clear deviations from the typical super-Arrhenius temperature dependence of the bulk material, depending on temperature and pore size. This allows assigning well-defined cooperativity length scales of molecular motion to certain temperatures above the glass transition. We relate these and previous results on glycerol confined in other host systems to the temperature-dependent length scale deduced from nonlinear dielectric measurements. The combined experimental data can be consistently described by a critical divergence of this correlation length as expected within theoretical approaches assuming that the glass transition is due to an underlying phase transition.
I. INTRODUCTION
Glasses and glass-forming materials are essential components of everyday life.Although the glass transition has been intensely studied for many decades, it is still not fully understood.The mystery of glassy freezing has not been completely unraveled and its investigation remains an important tasks of modern solid state physics. 1,2,3,4In most cases, the temperature evolution of the molecular dynamics of glass-forming liquids does not simply follow a thermally activated (Arrhenius) behavior but rather more strongly slows down when approaching the glass-transition temperature T g .Understanding the molecular origin of this non-canonical behavior represents a great scientific challenge.Often, an increasingly cooperative motion of the glass-forming entities (molecules, ions, polymer segments, etc.) under cooling is presumed to explain the observed super-Arrhenius behavior. 2,5,6,7,8This notion is based on theories assuming that the glass transition is caused by an underlying phase transition into a state with a specific kind of "amorphous order". 5,6,9This would imply the divergence of a characteristic correlation length L corr , characterizing the cooperatively rearranging regions, at a temperature significantly below T g .However, this transition cannot be directly detected because the supercooled liquid falls out of equilibrium when cooling below the glass temperature.Nevertheless, this phase-transition scenario was recently nicely corroborated by nonlinear dielectric measurements from which the relative temperature dependence of L corr was deduced, albeit not providing absolute values. 10,11,12,13Other very useful attempts to learn more about the molecular dynamics and the length scales involved in the glass transition are investigations of supercooled liquids that are confined in spaces of nanometer size. 14,15,16,17,18,19,20,21,22,23,24,25,26In bulk glass formers, the size of the cooperatively rearranging regions is supposed to increase with decreasing temperature, which should lead to an increase of the effective energy barriers and, consequently, to a stronger temperature dependence of relaxation times than in the Arrhenius case.But when the glass former is confined in a pore of defined geometry, below a certain temperature the growing correlation length L corr of the cooperative regions will exceed the pore diameter and, ideally, a crossover from non-Arrhenius to Arrhenius behavior would be expected.
Many different materials have been reported as host materials for measurements in confined geometries (for a thorough overview, see Ref. 23).These can be classified as materials with ordered pores or with disordered pores.Furthermore, the geometry of the pores is also important for the properties of the confined liquid 19 and, thus, it is convenient to differentiate between 3D-(pores), 2D-(layers), and 1D-confinement (channels).Generally, for a meaningful comparison with bulk properties a 3D-confinement seems preferable, thus avoiding dimensionality effects obscuring the cooperativity-related confinement effects.In previous works, confinement measurements on glass formers have been studied for various host materials of different pore sizes. 18,19,20,22,23We have recently reported 25 that metal-organic frameworks (MOFs) are well suited 3D host systems for confinement investigations.
In general, MOFs represent three-dimensional networks composed of metal ions or clusters and organic linkers, thereby forming extended crystalline frameworks with significant porosity and threedimensional pore geometry. 27,28,29Due to their highly ordered structure and design versatility, MOFs with various, well-defined pore sizes are available, in contrast to many other confinement hosts, where the pore dimensions are distributed.The available pore sizes include the region of 1-2 nm, which is not well covered by other materials. 30In Ref. 25 we used three MFU-type MOFs (MFU stands for "Metal-Organic Framework Ulm-University") as host materials with pore sizes between 1.19 and 1.86 nm to study the dynamics of the glass former glycerol in confined geometry. 25In small pores (MFU-4), no bulk dynamics could be observed, indicating that the correlation length in glycerol is larger than the pore diameter of 1.19 nm, even at the highest investigated temperature of 380 K.While for the MOFs with bigger pores the expected confinement effects were partly observed, the interpretation of these results was complicated by the presence of two types of pores in the structure with 1.2 and 1.8 nm (MFU-4l) and by a partial interdigitation of the pore framework (MFU-1).Nevertheless, these earlier investigations have demonstrated that MOFs indeed seem to be ideal host systems for confinement measurements, with weak molecule pore-wall interactions and essentially three-dimensional pores with sizes that are comparable to the cooperativity length scales in typical glass-forming liquids.For the present work, we have selected two special MOFs, namely the microporous zeolitic imidazolate frameworks (ZIFs) ZIF-8 and ZIF-11 with two different, well-defined pore sizes (Fig. 1).ZIFs have zeolite-like topologies. 31,32,33They are composed of tetrahedrally-coordinated transition metal ions [typically Zn(II) or Co(II)] connected by imidazolate linkers.The metal-imidazole-metal link has a large angle of ca.145° which is analogues to the 145° Si-O-Si angle found in zeolites, and thus ZIFs adopt the same framework topologies as zeolites.Additionally, they are known as porous materials with exceptionally high chemical and thermal stability and with strongly hydrophobic internal surfaces, making the formation of surface layers of the guest material very unlikely.ZIF-8 and ZIF-11 crystallize in a cubic crystal system, in the space group I4 3m and Pm3 m, respectively.An important structural feature of these ZIFs is, that they possess rather large pores (ZIF-8: 1.16 nm in diameter, ZIF-11: 1.46 nm), connected by significantly smaller pore apertures (ZIF-8: 0.34 nm, ZIF-11: 0.30 nm).Therefore, these materials exhibit well-defined, independent cavities and we do not assume significant interactions between molecules in different pores.These structural properties (summarized in Table S1 in the supplementary material 34 ) make them attractive host materials for studying dynamics of confined liquids.
In the present work, we investigate the molecular dynamics of supercooled glycerol confined in these hosts via broadband dielectric spectroscopy.We compare our results to previous studies of the cooperativity length scale in this system.Combining the results from confinement measurements of glycerol with the relative temperature dependence of its amorphous-order length scale as recently deduced from nonlinear dielectric spectroscopy 11,12 provides interesting hints at the temperature evolution of cooperativity in this prototypical glass former.
A. Preparation and characterization of glycerol loaded ZIFs
All reagents were used as received from commercial suppliers.ZIF-8 and ZIF-11 were synthesized according to the previously described procedures. 35,36The two ZIFs used in this work consist of the same metal ion, Zn(II), but differ in the organic linkers.ZIF-8 was synthesized using 2methylimidazole (H-MeIM), 35 whereas benzimidazole (H-PhIm) was used in the preparation of ZIF-11. 36The phase purity was confirmed by X-ray powder diffraction (XRPD) measurements (Figs.S1 and S2 in the supplementary material 34 ).The XRPD data were collected in the 2θ range of 4-70° with 0.02° steps using a Seifert XRD 3003 TT diffractometer equipped with a Meteor 1D detector.To load glycerol into the pores, ZIF samples (30 mg) were degassed for 20 h at 120 °C in vacuum and then placed in an open vial into a Schlenk tube containing glycerol (2 mL).The Schlenk tube was heated for 24 h at 85 °C in vacuum (approx. 1 mbar).The amount of glycerol loaded into the pores was determined by thermogravimetric (TG) analysis (Table I and Fig. S3 in the supplementary material 34 ).The TG analysis was performed with a TA Instruments Q500 analyzer in the temperature range of 25-700 °C under nitrogen atmosphere at a heating rate of 5 K min -1 .Models of unit cells of the ZIFs loaded with glycerol (shown in Fig. S4 and S5 in the supplementary material 34 ) were created by the "Sorption Tools" module of Accelrys Materials Studio 2017, employing a Metropolis sampling scheme to find appropriate positions of the glycerol molecules (loading at 298 K to a fixed target pressure of 100 kPa).During the sampling, all framework lattice atoms were fixed at their crystallographic positions.The number of glycerol molecules per unit cell obtained from sorption simulations is given in Table I.
B. Broadband dielectric spectroscopy
Broadband dielectric spectra of the dielectric loss of the samples were collected in the frequency range of 1 mHz -3 GHz combining two experimental techniques. 37For the low-frequency range (1 mHz -1 MHz), a frequency response analyzer (Novocontrol -Analyzer) was used.For the radio frequency and microwave region (1 MHz -3 GHz), two impedance analyzers using a coaxial reflectometric setup were employed.These measurements were performed by an Agilent 4294A impedance analyzer for frequencies below 100 MHz and by an Agilent 4991B for higher ones.In this technique, the sample capacitor is mounted at the end of a specially designed coaxial line. 38Cooling and heating of the samples was achieved by a closed-cycle refrigerator and a nitrogen-gas cryostat.
All measurements were carried out on powder samples filled in parallel-plate capacitors to avoid pressure-induced deterioration of the glycerol-loaded pores that could occur when pressing pellets.
For the capacitors, a plate distance of approximately 100 µm was realized.To minimize voids in the samples, slight pressure was applied to the filled capacitors.After mounting the loaded capacitors into the cooling devices, they were kept under vacuum at room temperature for at least 24 h before starting the dielectric measurements.This procedure ensured that all adsorbed contaminations (mainly water) were removed from the sample surface.
A. Sample characterization
Glycerol was introduced to the pores via vapor diffusion from the gas phase.To confirm that the glycerol loading did not influence the MOF crystal structure, XRPD measurement before and after the glycerol loading were carried out (Figs. S1 and S2 in the supplementary material 34 ).The glycerol content in the loaded samples was determined by TG analysis.The TG curves (Fig. S3 in the supplementary material 34 ) reveal that the glycerol release occurred between 90-210 °C with the mass loss -27.4 % for ZIF-8 and -22.3 % for ZIF-11.As shown in Table I, the experimental loading of glycerol per unit cell of ZIF-8 and 11 closely matches the theoretical values obtained from sorption simulations.
B. Broadband dielectric spectroscopy
As an example for our broadband dielectric-spectroscopy results, Fig. 2 shows the dielectric loss, the imaginary part of the permittivity ′, of glycerol in ZIF-11 as a function of frequency for various temperatures.At first glance, two relaxation processes can be immediately identified, leading to the typical peaks in the dielectric-loss spectra, which shift to lower frequencies upon cooling.(In the following, the faster one will be denoted as process 1 and the slower one as process 2.) A comparison of process 1 with the relaxation in bulk glycerol, 40 i.e., the main reorientational motion of the molecules, which is closely coupled to the glass transition, 41 reveals similar loss-peak frequencies.
Therefore, we ascribe the fastest process detected in the spectra to the relaxation of glass-forming glycerol confined in ZIF-11.At the lowest frequencies and high temperatures, an additional increase in " for decreasing frequency is observed (e.g., below about 1 Hz for the 298 K spectrum).This indicates another contribution (termed process 3), which is only partly resolved due to the low-frequency limit of the measured spectra.Consequently, the spectra of Fig. 2 are fitted by the sum of three relaxation functions.Relaxations 1 and 2 are described by a Havriliak-Negami (HN) function, whereas for process 3 the Cole-Cole (CC) equation was used, the latter being a special case of the HN function with symmetrical broadening. 42Both are empirical functions and often employed to parameterize dielectric relaxation processes.Separate measurements of the empty ZIF-11 at room temperature (not shown) did not exhibit any significant indications of relaxational dynamics.Therefore, we can exclude that residual amounts of solvent or of water within the sample 43 contribute to the detected relaxational response of glycerol enclosed in ZIF-11.
FIG. 2. Frequency dependence of the dielectric loss of glycerol confined in ZIF-11 for different temperatures.
The lines through the points are fits by the sum of two HN functions for the two well-pronounced relaxations and a CC function to formally account for the increase of the loss found at low frequencies and high temperatures.The numbers denote the three contributions to the spectra.
What is the origin of the two low-frequency processes observed in the spectra of Fig. 2? Ionic impurities in the sample that are sufficiently small to pass through the pore apertures, can lead to ionic charge transport, giving rise to dc and ac contributions to the complex conductivity * = ' + i". 44,45As the dielectric loss and the real part of the conductivity are related via " '/, an increase of " towards low frequencies will result as is also observed in bulk glycerol. 46The slowest process (3), which in the spectra of Fig. 2 only appears as an additional low-frequency increase for the highest temperatures (T 253 K), can probably be ascribed to such conductivity contributions.Therefore, it will not be further discussed in the following and the CC function used to fit this spectral feature should be regarded as a purely phenomenological description.
Concerning process 2, two different explanations can be considered: 14,15,16,47,48,49 In confined dipolar systems, in addition to the relaxation arising from the relatively freely reorienting guest molecules, a layer of molecules with slowed down dynamics due to interactions with the pore walls can exist, leading to a separate loss peak at much smaller frequency. 14,48,49However, one should be aware that the inner pore walls of the ZIFs are strongly hydrophobic making strong interactions with the dipolar glycerol molecules unlikely.Alternatively, another process can arise from the heterogeneous nature of the guest-host system: It was shown long ago by Maxwell and Wagner that samples composed of two different dielectric materials can exhibit a nonintrinsic relaxation-like process in the dielectric spectra, nowadays termed Maxwell-Wagner (MW) relaxation. 50,51,52In glycerol in ZIF-11 principle, based on the known dielectric properties of the bulk and host material, theoretical spectra that are modified by the MW effect may be calculated using the following relation: 47,52 (1)
Here
is the permittivity of the glycerol-loaded ZIFs, the permittivity of bulk glycerol, the permittivity of the empty ZIFs, is a factor referring to the geometry of the pores (for spherical cages it is 1/3), and Φ is the volume ratio of the confined material and the host system (Table I).Calculating using Eq. ( 1), leads to theoretical loss spectra that exhibit two relaxation peaks.(Variations of Φ between 38 and 44% only cause a small change of ; therefore this calculation can be taken representatively for both ZIF systems.)One of them arises from the relaxation, which, due to the MW effects, can be shifted towards somewhat higher frequencies compared to the bulk. 47The second one reflects the non-intrinsic, heterogeneity-related MW relaxation mentioned above.The resulting peak positions of the latter do not agree with those of process 2. However, one cannot exclude that the assumptions of the used MW model are oversimplified.Overall, one has to state that the origin of the detected process 2 is not finally clarified yet, but the main interest of the present work lies in the investigation of the -relaxation dynamics of glycerol.Thus, in the following we concentrate on the confinement-induced behavior of process 1.
The dielectric spectra of glycerol confined in ZIF-8 also exhibit two relaxational processes, in addition to the conductivity contribution showing up at low frequencies and high temperatures (see Fig. S6 in the supplementary material 34 ).Processes 1 and 2 both are present, occur in a similar frequency region as for glycerol in ZIF-11, and can be rationalized in a similar way.Especially, relaxation 1 in ZIF-8 again corresponds to the relaxation of glycerol.In contrast to ZIF-11, the symmetrical CC function was sufficient to fit process 2 of glycerol in ZIF-8.FIG. 3. Frequency dependence of the dielectric loss of glycerol in the -peak region (process 1) confined in ZIF-11 (triangles) and ZIF-8 (squares), and of bulk glycerol (circles). 40The temperatures of the confined systems were selected to achieve an approximate match of the peak frequencies of the bulk results.Figure 3 shows typical loss spectra in the -relaxation region (process 1) of glycerol confined in ZIF-11 and ZIF-8, compared to those of the bulk, measured at four different temperatures. 40For the confined samples, the temperatures of the shown spectra were selected to achieve an approximate match of the peak frequencies with those of the bulk spectra.The first insight gained from this plot is the enormous reduction of the relaxation strength Δ induced by the confinement.The performed fits suggest a reduction of of about two orders of magnitude.This phenomenon is a common finding for confined systems. 15,16,19,47Obvious reasons for this behavior are the decreased amount of supercooled glycerol per volume in the confined samples (because part of the volume is occupied by the host material) and the incomplete space filling of the sample powder.However, when considering the volume ratio of the guest-host system of around 38 to 44%, it becomes clear that there has to be another contribution to the found strong reduction of .In fact, as pointed out, e.g., by Richert, 47 the relaxation strengths of heterogeneous mixtures of different dielectrics do not add up in a simple way.The difference in Δ between the two confined systems is rather marginal and probably due to the different combinations of the aforementioned effects in these systems.
Figure 3 also reveals that the relaxations of the confined glycerol are considerably broadened compared to the bulk.At room temperature, the bulk loss-peak features a half width of 1.6 decades, already somewhat smaller than the values of 2.2 in ZIF-11 and 2.1 in ZIF-8.When approaching 200 K, the half width for the bulk only increases to 2.1 decades, whereas in ZIF-11 it even becomes 4.9 and in ZIF-8 3.7 decades.Such a strong broadening is a common finding for confinement measurements and may be explained by a distribution of relaxation times due to interactions of the glycerol molecules with the pore walls and/or a variation in the number of glycerol molecules per pore. 14,15,19,22,25omparing the temperatures in Fig. 3 already provides first hints at the significantly different slowing down of the molecular dynamics in the two confined systems and bulk glycerol: For both confined systems, more or less strong deviations from the bulk temperatures show up implying the modification of molecular dynamics by the confinement.The dynamics of the molecular motions is characterized by the average relaxation time , which is related to the peak frequency via 1/(2 p ). Thus Fig. 3 indicates a confinement-induced shift of the relaxation times of the guest material.This phenomenon certainly is the most interesting outcome of confinement measurements on supercooled liquids.As discussed in the introduction, ideally it should provide insights into the cooperativity length scales in glass forming systems. 15,16,17,18,22,24,25
C. Relaxation times
For a better comparison of the relaxation-time dynamics, (1/T) is plotted in Fig. 4 for the two confined systems.The shown relaxation times represent the inverse circular peak frequency, 1/(2 p ), calculated 53 from the fit parameters used to describe the dielectric spectra with the HN function (e.g., lines in Figs. 2), which is a good approximation of .For the CC function, used for relaxation 2 in ZIF-11, due to its symmetric shape is equal to CC as directly obtained from the fits.
In Fig. 4, the present (T) data are shown together with those for bulk glycerol published earlier (dashed lines). 54he relaxation times of process 2 (squares in Fig. 4) exhibit a linear Arrhenius-like behavior, with small deviations for ZIF-8.They are of comparable order of magnitude for both confined systems.Moreover, for glycerol in MFUs a similar process was found just in this region. 25Thus, it is likely that these relaxations are all of the same origin.As mentioned above, they may be of Maxwell-Wagner origin or arise from layers of glycerol molecules interacting with the pore walls, the latter being unlikely due to the hydrophobic internal surfaces of the ZIFs.In the context of the present work, aiming at the clarification of cooperativity effects of the relaxation, these slow processes are of minor relevance.
The dashed lines in Fig. 4 indicate the temperature-dependence of the -relaxation time of bulk glycerol. 54It deviates from thermally activated Arrhenius behavior, which is a typical feature of glassforming liquids and can be described by the empirical Vogel-Fulcher-Tammann (VFT) law: 55,56,57,58 exp (2) (In fact, for better readability of the figure the dashed lines show the VFT fit curve instead of the actual experimental values. 54) In this equation, corresponds to an inverse attempt frequency, is the Vogel-Fulcher temperature marking a divergence of below the glass transition, and represents the so-called strength parameter, which can be used to quantify the degree of deviation from Arrhenius behavior. 58For T VF = 0, Eq. ( 2) corresponds to Arrhenius behavior.The circles in Fig. 4(a) show the relaxation times for glycerol in ZIF-11.At temperatures above about 213 K, the experimental data can be well fitted by a VFT-law (solid line).However, below this temperature (T) is better described by the Arrhenius function (dash-dotted line).As mentioned in the introduction, such a transition from VFT to Arrhenius behavior at low temperatures in confinement points to growing cooperativity of molecular dynamics under cooling beyond the pore diameter. 15Within this framework, the non-canonical slowing down of molecular motion when approaching the glass temperature is explained by an increasing number of cooperatively moving molecules.This implies a growth of the correlation length scale L corr , characterizing the size of the cooperatively rearranging regions, consistent with a phase-transition-related origin of the glass transition. 5,6,12,9The increasing cooperativity leads to a growing effective energy barrier with decreasing temperature and explains the commonly observed super-Arrhenius behavior of (T).In bulk materials, L corr can get larger without restrictions, but in confinement the pore diameter of the host system limits the maximum correlation length. 15Within this scenario, the observed crossover in (T) of ZIF-11 can be explained by L corr (T) of bulk glycerol exceeding the pore diameter of 1.46 nm below about 213 K.In the confined system, the cooperativity length cannot grow beyond 1.46 nm and remains temperature independent and of the order of the pore size under further cooling.Therefore the energy barrier does not increase anymore and Arrhenius temperature dependence of (T) is observed.It should be noted that an alternative explanation of such a transition from VFT to Arrhenius in confinement was recently suggested, based on the assumption of a separate glass transition of molecules strongly interacting with the pore walls. 59However, such an explanation seems unlikely in the present case, because in ZIFs, in contrast to most other host materials, only weak molecule pore-wall interactions are expected.
At high temperatures, the deduced (T) of glycerol confined in ZIF-11 is by about a factor of 3-4 smaller than for the bulk material [Fig.4(a)].As mentioned above and thoroughly discussed in a review article by Richert 47 , such an acceleration of the relaxation in a confined system may rather trivially arise from the heterogeneity of the guest-host system, not being related to the "real" confinement effects arising from the length scales involved in the glass transition.When using Eq. ( 1) to estimate the shift of the relaxation times expected within this scenario, we find that it indeed is consistent with the experimental observations at high temperatures (1000/T < 4 K -1 ).However, it should be noted that (T) of the confined sample exhibits a somewhat stronger overall temperature dependence than the bulk (at least at T > 213 K), which leads to an approach of the bulk (1/T)-curve in Fig. 4(a) at low temperatures.This cannot be explained by the mentioned heterogeneity effects and demonstrates some confinement-induced variation of the temperature-dependent cooperativity of the bulk-like -relaxation dynamics within the pores.It should be noted that qualitatively similar behavior was previously also reported by Fischer et al. for glycerol in a different MOF (MFU-4l). 25yway, for glycerol confined in ZIF-11 there is a clear transition of (T) from VFT to Arrhenius behavior at low temperatures, just as expected within the framework of a cooperativity-driven glass transition.Notably, until now such a temperature-dependent crossover, triggered by a confinementinduced suppression of a further growth of cooperativity length, was only rarely observed. 15,19,25nstead, very often the whole τ(T) curves in confined systems were found to be completely shifted compared to the bulk, allowing, at best, an estimate of a lower limit of L corr only.
The latter scenario is also found for ZIF-8 [Fig.4(b)], where the temperature-dependent relaxation time (circles) exhibits a clear shift to faster relaxations rates compared to the bulk and Arrhenius behavior in the entire investigated temperature range.At the lowest temperature of 198 K, the deviation from the bulk exceeds three decades.This finding can be well understood by the strong confinement in the small ZIF-8 pores with diameters of 1.16 nm.Obviously, even at the highest measured temperature of 308 K, the correlation length of bulk glycerol still is larger than the pore diameter of ZIF-8.This prevents a temperature-dependent variation of L corr within the small pores of ZIF-8.Consequently, for the confined system there is no indication of the typical super-Arrhenius temperature dependence of (T) of glass-forming systems in the whole temperature range.
This finding is well consistent with those for glycerol in MFU-4 with pore sizes of 1.19 nm, whose 1/ curves are much closer to Arrhenius behavior than bulk glycerol. 25Overall, the present results on glycerol in ZIF-8 provide a lower limit of 1.16 nm for the cooperativity length scale in glycerol at temperatures up to 308 K.
D. Cooperativity length
Figure 5 summarizes the findings concerning the temperature dependence of the cooperativity length scale in glycerol, based on the present work and literature data.To our knowledge, absolute values for L corr , deduced from experimental data, until now were only provided by the dynamic calorimetry investigation of Hempel et al. 60 (stars in Fig. 5) and by the present measurements in ZIF-11 (1.46 nm at 213 K; triangle), the latter revealing a clear crossover from cooperativity-dominated VFT to non-cooperative Arrhenius behavior.Based on multidimensional nuclear-magnetic-resonance experiments, 61 the length scale het of dynamic heterogeneities in glycerol at temperatures around 200 K was found to be about 1 -1.3 nm, 62 of similar order as L corr determined in the present work.
However, as noted, e.g., in Ref. 61, L corr and het do not necessarily have to agree.
In Ref. 11, in agreement with an earlier work using similar methods, 10 information on the temperature dependence of the number of correlated molecules, N corr , was obtained, based on measurements of the third-order harmonic component of the dielectric susceptibility. 63From the definition of N corr , it immediately follows that L corr (N corr ) 1/3 .While no absolute values of N corr (and thus of L corr ) could be obtained by this approach, the value at 213 K from the present work (triangle in Fig. 5) can be used to scale these data as shown by the plusses in Fig. 5.For both ZIF-8 and MFU-4, 25 no super-Arrhenius behavior was found in the whole investigated temperature range.Thus, as discussed above, based on these measurements only lower limits of L corr can be provided.The temperatures of the corresponding data points shown in Fig. 5 represent the maximum T covered by these experiments (only above this temperature, L corr may decrease below this limit).Moreover, from the results on glycerol confined in MFU-4l, reported in Ref. 25, an upper limit of L corr can be derived.This is based on the fact that for this confined system super-Arrhenius behavior of (T) was found in the entire investigated temperature range, i.e., L corr of glycerol obviously did not exceed the pore size of 1.86 nm, even at the lowest covered temperature of 205 K.The corresponding data point (circle in Fig. 5) is shown at the lowest investigated temperature of this study.A similar situation was found for glycerol in nanoporous sol-gel glass with a pore size of 2.5 nm, reported by Arndt et al., 49 which is shown by the "x" in Fig. 5.The shaded areas in Fig. 5 indicate the excluded regions of L corr , as inferred from the shown lower and upper limiting values.Results on glycerol confined in MFU-1, reported in Ref. 25, are not presented in Fig. 5 as their interpretation is hampered by partial interdigitation of the host framework, leading to a distribution of pore sizes. 25G. 5. Temperature dependence of the cooperativity length L corr of glycerol as deduced from the present work and literature data.The closed symbols show the results from confinement measurements performed in our group using ZIF-(present work) and MFU-type MOFs 25 as host materials.For ZIF-8 and MFU-4, only lower limits of L corr can be provided and for MFU-4l an upper limit was reported.The shaded areas indicate the excluded ranges for L corr inferred from these limiting values.Literature data from Arndt et al. 49 (confinement in nanoporous glass; upper limit only) and from Hempel et al. 60 L corr (T-T c ) - , with = 0.28 and T c fixed to the Vogel-Fulcher temperature, i.e., T c = T VF = 132 K. 54 The dashdotted line indicates a possible temperature-independent L corr above T A 288 K, where (T) was reported to cross over from VFT to Arrhenius behavior. 67,71The arrows indicate the Vogel-Fulcher temperature, glass temperature, and VFT-Arrhenius transition temperature (from left to right).
The dashed line in Fig. 5 represents a fit of the temperature-dependent correlation length from Ref. 11 (plusses) with a critical law, L corr (T-T c ) - .Of course, the temperature range covered by this data set is rather restricted and, in principle, should not allow drawing meaningful conclusions on the involved parameters or on the development of (T) at lower and higher temperatures.However, within scenarios assuming a "hidden" phase transition as the true reason of the glass transition, this transition is expected to occur close to the Kauzmann 64 or the Vogel-Fulcher temperature (both are usually of similar magnitude 65 ), where (T) should diverge [cf.Eq. ( 2)].Therefore, for the fit of L corr (T) we could fix T c at T VF = 132 K 54 which strongly enhances the reliability of the obtained fit curve and fit parameters.With a critical exponent = 0.28, resulting from the fit, the experimental data can be reasonably described in this way (dashed line in Fig. 5). 66However, at temperatures above about 300 K, the extrapolated fit curve falls below the lower limit of L corr of about 1.2 nm, derived from our confinement investigations of glycerol in ZIF-8 and MFU-4 (lower shaded region in Fig. 5).To explain this apparent discrepancy, one should be aware that for various glass-forming materials, a transition of (T) from VFT to Arrhenius behavior was proposed to occur above a temperature T A . 67,68,69,70For glycerol, such a crossover was reported to arise at a temperature T A of about 290 K. 67,71 When assuming that the VFT temperature dependence is caused by an increase of the effective energy barrier at low temperatures due to increasing cooperativity, 2,5,6,7 the proposed high-temperature Arrhenius behavior above T A , corresponding to a constant energy barrier, would imply constant (or absent) cooperativity.Indeed, direct interactions with the nearest-neighbor shell should be possible at all temperatures and, in first respect, constitute Arrhenius-type relaxation dynamics.Such a scenario is well consistent with the results of Fig. 5 on L corr (T) when assuming a crossover from the critical law (dashed line) to constant behavior (dash-dotted line) close to T A .When adjusting the level of the latter to obtain a crossover between critical and temperature-independent L corr (T) just at T A = 288 K, 67 we arrive at L corr 1.22 nm for T > T A .This value is well consistent with the lower limits for this length scale provided by our confinement measurements.
At low temperatures, while being consistent with the upper-limit values provided by the MFU-4l 25 and sol-gel-glass confinement investigations, 49 the extrapolated critical law for L corr (T) with T c = T VF does not match the two data points from Ref. 60 (stars in Fig. 5), based on dynamic calorimetry.It is clear that the temperature range of the fitted third-order dielectric data (plusses) is too restricted to allow for meaningful fits without the critical temperature being fixed at T VF .In any case, to account for the data by Hempel et al., 60 T c would have to be much larger than T VF and a divergence of L corr close to T g of about 185 -189 K, 40,60 would have to be assumed.This seems quite unrealistic because T g only marks a purely dynamical effect and the possible phase transition into a state with "amorphous order" that may underlie the glass transition is expected at much lower temperature. 5,6,9
IV. SUMMARY AND CONCLUSIONS
In the present work, we have reported dielectric confinement measurements on glycerol in two different zeolitic imidazolate frameworks with well-defined pore sizes of 1.16 and 1.46 nm.Our findings demonstrate that this class of MOFs can provide well-suited host materials for confinement investigations of glass-forming liquids.In both investigated systems, the relaxation of confined glycerol could be clearly identified.It exhibits the broadening and amplitude reduction as known from other confined systems.Most importantly, depending on temperature and pore size its relaxation time reveals different degrees of deviation from that of the bulk.In contrast to the previous investigation of glycerol confined in MOFs, 25 the ZIFs used here do not suffer from the presence of two types of pores, interdigitation, or partial pore filling.Thus, our present findings allow for significant conclusions concerning the cooperativity length scale L corr of glassy dynamics in glycerol: The measurements of glycerol in ZIF-8 reveal a lower limit L corr (T) > 1.16 nm for T < 308 K while those for the ZIF-11 host material provide an absolute value of 1.46 nm at 213 K.These values are well consistent with earlier estimates of upper or lower limits L corr at different temperatures, based on confinement investigations of glycerol. 25,49All these results can be combined with the information on the temperature dependence of L corr , recently derived from nonlinear dielectric measurements. 11We find that the available data are in good accord with a critical increase of the cooperativity length scale in glassforming glycerol and a divergence close to the Vogel-Fulcher temperature.This supports the notion that there is a growth of amorphous order when approaching the glass transition, which is due to an underlying phase transition arising significantly below T g .However, one should be aware that the overall variation of the deduced length scale between the low-viscosity liquid at high temperatures and the solid glass somewhat below T g is rather limited.Only at temperatures significantly below the glass temperature, L corr varies considerably and finally diverges.However, in measurements with realistic cooling rates, of course this divergence does not come into effect because the system falls out of equilibrium at the glass transition.
Notably, the confinement data seem to indicate that at high temperatures, above about 290 K, L corr becomes temperature-independent and levels off at about 1.22 nm.This would be consistent with the often assumed crossover of the temperature-dependent relaxation time from VFT to Arrhenius behavior and certainly deserves further exploration.When considering the glycerolmolecule size of roughly 0.5 -0.6 nm, 72,73 the deduced limiting high-temperature value of the correlation length of about 1.22 nm seems reasonable, essentially reflecting next-neighbor interactions only.
SUPPLEMENTARY MATERIAL
The supplementary material shows structural properties of the used ZIFs, models of unit cells of the ZIFs loaded with glycerol, TG-analysis results, and dielectric spectra of glycerol in ZIF-8.
FIG. 1 .
FIG. 1. Crystal structures of the ZIFs used in this study; both are drawn to a common scale.Zn ions are represented by polyhedra and the pores by spheres.H atoms are omitted for clarity.
FIG. 4 .
FIG. 4. Arrhenius plot of the average relaxation times of glycerol confined in ZIF-11 (a) and ZIF-8 (b).The framed numbers denote the two detected relaxation processes.For comparison, the dashed line in each frame represents the VFT fit of of bulk glycerol as reported in Ref. 54.The circles indicate the -relaxation times (process 1), the squares those of the slow process 2. The solid line in (a) is a fit of the high-temperature part of the -relaxation-time curve with the VFT function, Eq. (2).The dash-dotted lines are linear fits, indicating Arrhenius behavior.The arrow in (a) indicates the transitions from VFT to Arrhenius behavior.The parameters of the VFT fits are: 3.9 10 s, 16, 132 K (bulk 54 ); 4.0 10 s, 11, 149 K (ZIF-11).The energy barriers for the Arrhenius part of the relaxation are: 1.4 eV (ZIF-11) and 0.69 eV (ZIF-8).
FIG.5.Temperature dependence of the cooperativity length L corr of glycerol as deduced from the present work and literature data.The closed symbols show the results from confinement measurements performed in our group using ZIF-(present work) and MFU-type MOFs 25 as host materials.For ZIF-8 and MFU-4, only lower limits of L corr can be provided and for MFU-4l an upper limit was reported.The shaded areas indicate the excluded ranges for L corr inferred from these limiting values.Literature data from Arndt et al.49 (confinement in nanoporous glass; upper limit only) and from Hempel et al.60 (dynamic calorimetry; absolute values) are also FIG. S2.XRPD patterns of an activated ZIF-11 sample and a glycerol-loaded ZIF-11 sample.
TABLE I .
Overview of simulated and experimentally found amount of glycerol molecules per unit cell of ZIF-8 and ZIF-11.
a From Ref.39. | 2019-01-15T18:54:47.000Z | 2018-11-07T00:00:00.000 | {
"year": 2018,
"sha1": "7e46bd9e1c81a39d9f381ee3e20f92ebffb5ab2e",
"oa_license": null,
"oa_url": "https://opus.bibliothek.uni-augsburg.de/opus4/files/56190/1.5080334.pdf",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "7e46bd9e1c81a39d9f381ee3e20f92ebffb5ab2e",
"s2fieldsofstudy": [
"Materials Science",
"Chemistry"
],
"extfieldsofstudy": [
"Materials Science",
"Medicine",
"Physics"
]
} |
235689914 | pes2o/s2orc | v3-fos-license | Suicidal behaviours and moderator support in online health communities: a scoping review
Objectives Online support can be a crucial source of support for individuals experiencing suicidal behaviours, with forum moderators being pivotal in terms of the role they play in times of personal mental health emergencies. This study identified what is empirically known about the professional practices of health professionals who are online mental health forum moderators and provide support to individuals experiencing suicidal behaviours. Design The Levac, Colquhoun and O’Brien extension of the Arksey and O’Malley scoping review framework was used. Search strategy The Psychology Collection (EBSCO), PsycINFO (EBSCO), Web of Science, Taylor and Francis Online, SAGE Journals and Science Direct databases were searched for articles that featured a result relating to an online forum; included participants who worked as online moderators or facilitators and focused on suicide or self-harm. Results were limited to peer-reviewed articles published in English from 1990 onwards. As a quality assurance measure, grey literature (nonacademic literature) was not included. Reference lists of included articles were hand-searched. Results There were 397 articles initially identified after applying inclusion and exclusion criteria, with five articles included for synthesis. All articles received a moderate quality rating. Only one article featured a moderator who was a qualified health professional; the moderators in the remaining articles were volunteers who undertook preservice training. We found that there is little research that examines the professional working practices of online moderators who support individuals experiencing suicidal behaviours. Conclusions The dearth of research focusing on the professional practices of online forum moderators is cause for concern given that individuals experiencing suicidal behaviours are increasingly turning to online forums when in crisis. Future research should focus on online moderators’ practice through interviewing moderators about their professional practices and by examining online moderator practice as it occurs in situ.
Strengths and limitations of this study ► An established scoping review methodology was used. ► A rigorous search strategy and systematic study selection was carried out by two researchers. ► A quality assessment was undertaken to identify and synthesise knowledge in this area. ► Search criteria were limited to peer-reviewed articles and articles published in English or translated into English. ► Grey literature (nonacademic literature) was excluded.
ABSTRACT
Objectives Online support can be a crucial source of support for individuals experiencing suicidal behaviours, with forum moderators being pivotal in terms of the role they play in times of personal mental health emergencies.
This study identified what is empirically known about the professional practices of health professionals who are online mental health forum moderators and provide support to individuals experiencing suicidal behaviours. Design The Levac, Colquhoun and O'Brien extension of the Arksey and O'Malley scoping review framework was used.
Search strategy The Psychology Collection (EBSCO), PsycINFO (EBSCO), Web of Science, Taylor and Francis Online, SAGE Journals and Science Direct databases were searched for articles that featured a result relating to an online forum; included participants who worked as online moderators or facilitators and focused on suicide or self-harm. Results were limited to peer-reviewed articles published in English from 1990 onwards. As a quality assurance measure, grey literature (nonacademic literature) was not included. Reference lists of included articles were hand-searched. Results There were 397 articles initially identified after applying inclusion and exclusion criteria, with five articles included for synthesis. All articles received a moderate quality rating. Only one article featured a moderator who was a qualified health professional; the moderators in the remaining articles were volunteers who undertook preservice training. We found that there is little research that examines the professional working practices of online moderators who support individuals experiencing suicidal behaviours.
Conclusions The dearth of research focusing on the professional practices of online forum moderators is cause for concern given that individuals experiencing suicidal behaviours are increasingly turning to online forums when in crisis. Future research should focus on online moderators' practice through interviewing moderators about their professional practices and by examining online moderator practice as it occurs in situ.
InTRODuCTIOn
Individuals are increasingly turning to online forums to seek help for mental health concerns including suicidal behaviours. 1 2 Examples of mental health forums include Togetherall (UK), SANE (Australia) and Kooth (USA). The growing popularity of online forums has been attributed to the ease with which information, advice and support can be accessed. [3][4][5] This ease of access, combined with anonymity, can be especially important for individuals experiencing suicidal behaviours; a population that is less likely to seek professional face-to-face support 2 6-8 due to issues such as the stigma that still surrounds suicide. 2 This reluctance is problematic due to the risk of serious harm or death that is associated with suicidal behaviours. 9 10 Recent research framed within the suicide ideation to action framework 11 distinguishes between the separate, yet connected constructs of suicidal ideation, behaviours and suicide. Suicidal ideation refers to the thoughts of killing oneself and can range from fleeting thoughts to detailed plans. 1 Suicidal behaviours refer to the deliberate and intentional acts to kill oneself 12 and suicide is when intentional and self-inflicted death occurs. 2 These constructs are separate in that individuals can experience suicidal thoughts without engaging in suicidal behaviours, or attempting, and dying by suicide. Although previous suicide attempts are the strongest predictor of future suicide Open access attempts 13 and death by suicide, some suicidal individuals can hold a range of intentions at one time. These intentions can potentially shift throughout a suicidal event, with ambivalence experienced as to the outcome of acts where death is a possibility. 14 A further related concept is nonsuicidal self-injury (NSSI), which refers to acts of deliberate self-injury without the intent to die; however, research posits that NSSI can provide emotional relief, which may enable the individual, through increased capability, to inflict more serious self-injury (or attempt suicide) in the future. 15 It is for this reason that NSSI can be viewed as a risk factor for suicidal behaviours.
Some individuals who experience suicidal behaviours engage in online mental health forums for information and support. Online mental health forums are virtual communities where members can interact with one another by asynchronously posting and responding to messages. 16 17 These forums can be open and visible to all internet users or closed and limited to select membership populations with online support forums being accessed by millions of people each day. 18 Research suggests that 20% of online users in the USA and 10% of users in the UK regularly use online mental health forums. 19 While there are no suicide-specific usage data figures that we could locate, in terms of general mental health forum, there has been a noted increase in use, evident in online help seeking for mental health becoming frequent enough for study via large sample national surveys. 20 Support in online forums can be provided by peers or those in formal moderator roles. 21 Online mental health moderators, hereafter referred to as moderators, are concerned with the safety of members and the forum as a whole and will intervene and interact with members as necessary to ensure that the forum remains a safe space. 22 Moderators generally oversee content posted by members to ensure compliance with the forum rules. 23 Specific tasks may include welcoming new members, editing content that contravenes the forum rules, and for some forums, supporting members in crisis. 24 Moderators may be unqualified individuals, often with lived experience of mental illness, who are working in a paid or voluntary capacity, or professionally qualified and employed health workers. 4 There is a body of evidence that reports both benefits and risks associated with talking about suicide on mental health focused online forums. Benefits include reducing stigma and increasing self-disclosure, 2 and possible risks include suicide contagion in others and the promotion of suicide. 25 Despite these identified benefits of online help seeking, there has been no synthesis of research focusing on the professional practices of moderators in supporting individuals. Here, professional practices pertain to the work and conduct of the moderator that ensures the safety of forum members and those at risk of suicidal behaviours. Without such a synthesis, the field risks engaging in research that reproduces rather than advances understandings of online moderator work practices, thereby potentially overlooking opportunities to inform practice recommendations.
A scoping literature review was, therefore, undertaken to identify what research has been conducted on health professionals working as online moderators who engage with, and offer support to, community members experiencing suicidal behaviours. The review sought to map research approaches, limitations and gaps to guide future research.
MeThODS
The scoping review protocol as specified by Perry et al 11 was used in this study. It followed the six-stage scoping review methodology proposed by Arksey and O'Malley, 26 and further developed by Levac et al 27 and the Joanna Briggs Institute (JBI). 28 The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guidelines were used. 29 Stage 1: identifying the question Using the broad population-concept-context mnemonic, 28 initially online moderators with a health qualification were the population. Pilot literature searches yielded only one article that met this definition; hence, the population was broadened to focus on moderators who had received some form of moderator training (ie, external or in house). Given that a scoping review is broad in nature, this change was viewed as consistent with this approach and also with the underlying intent of the study to understand the professional practices of moderators who have training in providing support online. This change is also reflective of the iterative process of scoping reviews. 30 The revised concept focused on moderators who had received some form of moderator training and worked as moderators of online mental health forums where members post about suicidal behaviours, self-harm and NSSI. No geographical limitations were placed on the literature as suicidal behaviours are a global concern. 10 The scoping questions were: ("online community" OR "online health community" OR "online forum") AND moderator OR facilitator AND suicid* OR self harm OR NSSI Published date >1990 Language: English ScienceDirect ("online community" OR "online health community" OR "online forum") AND (moderator OR facilitator) AND ("suicidal ideation" OR suicide OR "self harm" OR NSSI) Published date >1990 Language: English
Research articles only
Medline (Web of Science) ("online community") OR ("online health community") OR ("online forum") AND moderator AND suicid* OR "self harm"OR NSSI From 1990 to 2019 (Basic search) Language: English SAGE Journals "online community" OR "online health community" OR "online forum" AND moderator OR facilitator AND suicid* OR "self harm" OR NSSI From 1990 to 2019 Language: English Taylor and Francis Online "online community" OR "online health community" OR "online forum"~4 AND moderator OR facilitator AND suicid* OR "self harm" OR NSSI
1990-2019 Language: English
Academic Edition and Sociology Ultimate. All of these databases are located within EBSCOhost. ► ScienceDirect. ► Medline. ► SAGE Journals. ► Taylor and Francis Online.
The detailed search strategy and search strings can be located in the published protocol 11 (table 1). The search strategy was conducted independently by two reviewers (AP, DP) and 395 results were returned as well as an additional two articles identified through hand searching of reference lists. Six articles were initially selected for inclusion; however, one article was a systematic review and was excluded at the data analysis stage, resulting in five included studies. 7 16 23 31 32 While the excluded systematic review was not part of this scoping review, the reference list was carefully reviewed to ensure that all relevant studies were included. The reference management programme EndNote (V.9) was used to manage the search results, with duplicate results (n=2) eliminated at the first stage of the review.
Stage 3: study selection
Each remaining article was independently screened for eligibility by two reviewers (AP and DP). The following inclusion criteria were applied to studies identified in step 2. 1. Each study needed to have undergone a peer review process to ensure only high-quality and credible studies were included. As such, grey literature (nonacademic literature) was explicitly excluded. 2. Each study had to be primarily focused on online mental health forums; therefore, comparative studies of online mental health forums and other mediums were excluded. 3. Any study that included moderators of an online health forum as a participant were included. Reviewers independently completed each level of screening as outlined below. The first level of screening was limited to title and abstract review, followed by a fulltext review of the remaining articles. The reference list of each included article was also screened to identify any potential additional articles that may not have been identified in stage 2. A conservative approach was adopted when screening studies to ensure relevant studies were not inadvertently screened out at the title and abstract stage. Thus, for any studies, where it was not clear if the exclusion or inclusion criteria were present, the studies were included for full screening. The reviewers met at the end of each stage of screening to discuss and compare findings. There was one discrepancy when identifying duplicates and two discrepancies at full-text review. In each instance, the discrepancies were resolved through discussion resulting in consensus. The PRISMA flowchart (refer to figure 1) records the number of articles at each stage and provides an overview of the breadth of the literature.
Stage 4: data extraction
The first reviewer (AP) extracted data from the five included studies using an adapted JBI template for evidence details, characteristics and results extraction instrument. 33 Extracted data included bibliographical information (ie, author, year) and study characteristics (ie, aim, methodology, online forum description, key findings, conclusions). To ensure quality and accuracy of the data extraction process, an audit of the extraction process was undertaken by reviewer DP, who directly compared and contrasted the extracted data with the articles of origin. All articles were included in the audit with no errors being identified.
While assessments of research quality are not mandatory for scoping reviews, 28 a quality assessment of included articles was undertaken. AP rated the research quality of studies using an adaption of the JBI critical appraisal tool checklists. 28 As per JBI guidelines, reviewers can adopt a point scoring system to assist in making judgements about the overall quality of studies. To this end, each appraisal item was weighted equally and given 1 point if scored as 'yes' and 0 points if scored as 'no'. Appraisal items not applicable to the study were not counted and deducted from the total tally scores. The total score for each study was calculated and the rating of poor, moderate or high research quality allocated. After consultation with the research team, we established a set of criteria to determine poor, moderate and high research quality. Poorquality research was defined as scoring less than 50%, moderate-quality research scoring between 50% and 80% and high-quality research scoring greater than 80%.
The overall appraisal of quality and supporting comments for each included article were recorded on the data extraction form. These quality assessments were checked for accuracy by DP at the same time as the aforementioned extraction audit was undertaken. No studies were excluded based on quality assessment judgements. As this scoping review sought to map the existing literature Open access rather than critically synthesise and answer a set question, a bias risk assessment was not undertaken. 34 The possibility of bias and how it had been addressed was assessed using criteria such as clarity of the review question and appropriateness of inclusion criteria when making the quality assessments.
Stage 5: Collate, summarise and report the data The reporting of results was guided by PRISMA and extension for Scoping Reviews checklist. A tabular synthesis was used to record the study characteristics, quality assessments, moderator characteristics, study limitations and research gaps. A narrative summary was included to provide an account and interpretation of the findings. 35 Stage 6: consultation While consultation is considered an optional stage 26 Levac et al 27 assert that the consultative process is essential in ensuring methodological rigour, as experts in the field of the review may be able to offer additional perspectives or critique on the findings and suggest additional sources of information for the scoping study. The lead author consulted with 14 moderators, five of whom provided feedback on the findings of the scoping review. These moderators were considered to have expert knowledge of forums and the professional practices of moderators who provide support for individuals experiencing suicidal thoughts and behaviours. The moderators were surprised at the dearth of research that has focused on online mental health forum moderators, and they affirmed the need for more focused research in the future.
Patient and public involvement
No patients or members of the public were involved in this scoping review.
Study characteristics
The aims and characteristics of included studies are outlined in table 2. The study aims of the included articles varied and included reviewing or describing an online forum, 23 31 exploring the experiences or responses of members, 7 32 and comparing trained volunteer moderator responses to lay individuals' responses. 16 Included publications came from a small number of developed countries with no one country dominating results. Publication dates ranged from 2007 31 32 to 2012, 16 reflecting a lack of recent peer-reviewed research. All five articles adopted a qualitative research approach; two studies used content analysis 7 16 and the remaining articles adopted descriptive narratives. 23 31 32 Three articles featured general mental health forums that catered for a range of mental health needs 16 31 32 ; the others focused on suicide prevention, 7 23 with one focusing specifically on distressed adolescents. 23 While all studies where qualitative, there is little methodological similarity between them with content analysis, linguistic analysis, descriptive narrative report, narrative field project report and a descriptive case report all being used. Further the professional practices, or 'the how', were not the key focus of included articles, rather moderator tasks or 'the what', were explored as part of the overview function of the online forum.
Characteristics of moderators
The training and practices of moderators are outlined in table 3. The majority of study participants was moderators without specific professional mental health training or qualifications 7 16 23 31 and were volunteers who had completed preservice training provided by the forum. Preservice training differed in terms of length of time, content and mode of delivery. Training duration ranged from 2 days 23 to 6 months 31 with training content ranging from 1 to 3-day skill-based workshops, 23 60 hours of suicide prevention training, 16 to 16 weekly simulation training sessions. 31 The mode of delivery was not explicitly stated in all included articles, with one article referring to online classroom sessions and simulations 31 and another referring to 12 group sessions. 16 Entry requirements for the moderation role were unclear from the included articles. One study stated that no specific training or qualification was required, however, practical experience of helping others was. 31 The same study outlined a typical selection process where 15 people from an applicant pool of up to 160 people are selected to engage in the 6-month preservice training programme, however, only half the number of selected applicants would successfully complete the training. 31 No explanation was given for the low completion rate other than the challenge of volunteers maintaining discipline and commitment. 31 One article 32 included a qualified health professional (medical doctor) as the sole moderator, with no information provided as to whether the health professional engaged in any training to transfer their clinical in-person skills to the online space. These findings indicate that there does not appear to be standardised training requirements and, therefore, guidance regarding what training and background is needed for individuals to become moderators for forums where suicide content and associated behaviours are discussed.
Across all studies, moderators were required to ensure that the forum was a supportive and safe environment for members; however, what moderators did to enact this differed. A majority of the articles included reference to moderators offering resources and referrals to external sources of support. 7 16 23 31 One article 23 stated that moderators were trained to recognise posts that may be harmful and to respond accordingly, which may consist of alerting a supervisor to potential risk, especially if the moderator is unsure of how to respond. It was not clear how moderators identified what content could be harmful, and what action was required in order to respond accordingly. In other articles, 7 31 32 the moderators themselves responded to distressed members. In one article, 32 any post with an intention or suicide No statement locating the cultural or theoretical orientation of the researchers.
Only one online community was reviewed and therefore may not be representative of all online mental health communities.
As the researcher was a participant-observer, they may be less objective than an independent investigator. Four sessions of personal supervision. Establish and enforce the formal rules for group discussions and foster an environment characterised by a positive attitude towards living. Moderators discourage promotion of suicide and respond to distressed members offering messages emotional support and referrals to external services.
Open access
Webb et al 23 Australia Aged 18 years or older.
Completed at least one Youth Ambassador skills workshops. Compulsory participation on a 2-day scenario-based training session. Moderators to foster a safe and positive environment, and ensure members follow the rules.
Moderators read all posts and respond to any inappropriate content, as well as any referral requests or content that has not received a response from the community in 24 hours (or an adequate response).
Moderators are trained to recognise and respond to any posts which may be harmful that is, include methods or intentions to self-harm or suicide by deleting them and supporting the member.
Moderators must report crisis posts to the supervisor, as well as any posts they are unsure of how to respond to. Barak 31 Israel No required to have a mental health background or training but must have practical experience in some form of helping. Computer skills are essential. Moderators responses are to reflect empathetic understanding and non-judgement. Moderators offer information and external referrals as needed.
Hsiung 32 USA A health professional (medical doctor) The moderator removed any false information, particularly ingenuine expressions of suicidality. Suicide contagion was minimised by electing not to make special announcements.
Open access plan was investigated, and any false reports of suicidality where treated as a breach of forum guidelines and removed. In all articles, the moderators were responsible for discouraging the legitimisation and promotion of suicide; instead fostering a positive attitude towards life. It was unclear from the articles what training equipped moderators to recognise harmful content, respond to distressed members and those at risk of making an attempt on their life and promote a positive attitude towards life nor how they did this. All studies received a moderate quality rating of between 50% and 80% (see table 2). The absence of high-quality studies is a limitation, indicating the need for better-designed studies in the future. Such studies would include more participant information (forum members and moderators) and explore more than one forum and include more information on the training moderators receive to be able to work online with distressed individuals. The absence of recent research is a further limitation given the increasing interest in online interventions and changing patterns of internet use 36 and highlights the need for current research in order to progress the field.
Each study featured one individual online mental health forum. The diversity of mental health forums in terms of concerns catered for, characteristics of forums members and approaches to moderation meant that a singular forum cannot be deemed as representative of all mental health forums. For this reason, the potential for transferability of specific insights into forums is limited given the different methodological approaches and, in particular, the utilisation of descriptive case and narrative reports that are intended to capture a particular experience. However, transferability of broader findings that are not context specific per se is possible. The included studies came from four countries, with a notable lack of representation from areas such as the UK and Asia despite these being locations where online mental help is available. 37 38 This narrow global representation is a limitation as the results from these countries cannot be generalised given that cultural understandings of suicide can influence suicidal behaviours. 39 Moderators featured in all studies; however, as moderators were not the primary research focus, a gap in understanding the professional practices of moderators has resulted. Furthermore, moderators were not asked about their professional practices or experience of working on online forums. This means researchers do not know what it is like for moderators to work with individuals experiencing a suicidal crisis in the online space, in terms of the aspects they enjoy and the challenges they face. The lack of research that focuses on qualified health professionals who work as moderators is a further gap, as only one included study featured a qualified health professional as the moderator.
DISCuSSIOn
The included studies highlighted that the broader work of moderators is to maintain safe and positive online spaces, and that online forums can play a part in improving mental health and are becoming increasing popular. However, despite an increasing number of people turning to online forums to seek help for mental health concerns such as suicidal behaviours, this review has highlighted that little research has explored the professional practices of moderators who are responsible for keeping forums a safe space. Of the research that has been conducted, this review identified that focus has been on volunteer moderators rather than qualified health professionals. It appears that an examination of how those who work online, with individuals experiencing suicidal behaviours provide safety and support, has not yet occurred. The existing research provides little insight as to how and when moderators knew to respond to members who were experiencing a suicidal crisis or how moderators felt about working in such an online space.
This review identified that the professional practices of moderators were often conceptualised as work tasks such as 'read all posts and respond to anything inappropriate'. 23 Viewing the professional practices of moderators as a list of specific task risks adopting a reductionist or simplistic approach to what are often complex and multifaceted decisions and interactions. 23 Equating practices with tasks provides little information about the deeper theoretical and procedural knowledge and skills that are used by a moderator to identify what constitutes inappropriate content and how to respond to the content in ways that mitigate and manage risk, especially if the inappropriate content reflects suicidality. This is an issue as moderators are responsible for working with vulnerable populations online, with little research available that explicates how these professional practices are enacted or what knowledge informs these practices. This makes it difficult to assess the effectiveness of online forums in providing support and help to individuals experiencing suicidal behaviours and ideation. It also makes it difficult to raise awareness of the potential value of this suicide prevention work. This is important because in face-toface therapy, examining how therapy is enacted within the confines of the consultation has provided a better understanding of the process of behavioural, emotional and cognitive change that occurs between client and therapist. 40 This is currently missing in the online space.
This review showed that there is a dearth of research that has specifically focused on qualified mental health professionals who work as moderators. Perhaps this is reflective of the moderator role as one that is more facilitative than therapeutic in nature. This aside, it is unknown how such moderators' transition from working face-toface with individuals displaying suicidal behaviours and ideation to working online; often with additional factors such as anonymity, and the absence of visual cues such as personal presentation and body language. Review findings indicate that there are varying degrees of training Open access and preparedness of online moderators across forums and potentially raises questions as to the efficacy of moderator support for forum members who are experiencing a suicidal crisis. It also raises questions around the safety and well-being of moderators in terms of their level of preparedness to engage with the degree and frequency of risk presented in the online space. While this study did not focus on the risks associated with moderators who provide online mental health support and who were not provided with training (in-house or formal), there is the potential for these untrained moderators to miss key suicidal behaviour indicators when interacting online.
A further gap in the literature is the absence of asking moderators about their experiences as moderators. In this context, experience pertains not only to the tasks moderators undertake but to the professional and procedural knowledge and skills that they call on to complete these tasks and engage with members who are experiencing a suicidal crisis. Critically what is missing is what aspects of working online with those experiencing a suicidal crisis, do moderators find challenging. The field does not know what it is like for moderators to work in the online space, how they perform their work or what skills and knowledge they drawn on to support those members in crisis.
It is vital for clinicians and policymakers to gain a better understanding of the professional practices of moderators to ensure the adequate support and safety of both moderators and forum members. To move the field forward, a starting point may be asking moderators about their moderation experiences to find out what they do (and do not do) in practice in the online space when working with those experiencing a suicidal crisis. Gaining this information can be achieved through interviewing moderators about their online forum moderation work in terms of their training, the aspects of the role they enjoy or find challenging and when and how they know to respond to support members analysing moderator interactions with members who are experiencing suicidal behaviours will also enable researchers to contrast and compare what moderators believe they do in theory to what actually occurs in practice. Gaining insight into the forum moderators is likely to assist in enhancing the recruitment, training and retention of forum moderators to the benefit of forums as organisations, the moderators themselves in terms of job satisfaction and the vulnerable populations that they serve.
The strengths of this study include it being the first study to systematically and rigorously review the work of online forum moderators who work with individuals experiencing suicidal behaviours and ideation. It did this by adopting a scoping review methodology and the completion of quality assessments. A limitation of this study is the exclusion of grey literature and publications that were not subjected to peer review in order to ensure a level of quality assurance that scoping reviews are often criticised as lacking. 41 Furthermore, the inclusion of English-only articles means that potentially relevant sources of information may have been excluded from this review and, therefore, from the analysis, and this may have changed the review outcomes.
COnCluSIOn
This scoping review provides an overview of what is currently known about forum moderators who support members experiencing suicidal behaviours. This review initially sets out to focus on mental health professionals who work as moderators; however, as there was only one available article that focused on this specific population, a wider scope was adopted to review the work of moderators regardless of qualifications. Five articles met the inclusion criteria and were qualitative and descriptive in their approach. No specific information was provided as to how moderators identify those who are experiencing a suicidal crisis or how they know to respond in the way that they do, instead the studies included information on what specific task moderators undertook. We posit that more research is required to understand the professional practices of moderators and that a rigorous and robust research strategy is needed to guide future research. The next steps must include interviewing moderators about their professional moderation experiences and closely examining and analysing moderator interactions with forum members experiencing suicidal behaviours.
Twitter Amanda Perry @mandiperry and Andrea Lamont-Mills @ALamontMills Acknowledgements The authors would like to thank all stakeholders who participated in the review, including Rowena McGregor for guiding with the selection of appropriate databases and the development of the search strategies. Special thanks to the online community moderators who provided consultation on the initial findings. We also acknowledge all stakeholders and we thank Tony Machin, Justin Canty, Raquel Peel, Angela Nicholas, and Brian O'Shea for their insightful peer review comments that improved this article.
Contributors AP led the design and development of this study and wrote the first draft of this paper. AP and DP conducted the searches and applied the selection criteria, with AP completing the data extraction and research quality assessment. DP conducted a quality assessment audit. AL-M, CdP, JdP and DP provided guidance and constructive feedback at all stages of the research. All authors approved the final manuscript and have given consent to the publishing of this article.
Funding The primary author is the recipient of the University of Southern Queensland Research Training Program Stipend Scholarship. The award/grant number is not applicable.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Patient consent for publication Not required.
ethics approval As this study did no involve human participants, human ethics approval was not applicable.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement No data are available. N/A.
Open access
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http:// creativecommons. org/ licenses/ by-nc/ 4. 0/. | 2021-07-01T13:16:42.945Z | 2021-06-01T00:00:00.000 | {
"year": 2021,
"sha1": "641f325917cf8b6c7ac3490b67da5cc7ee83eb10",
"oa_license": "CCBYNC",
"oa_url": "https://bmjopen.bmj.com/content/bmjopen/11/6/e047905.full.pdf",
"oa_status": "GOLD",
"pdf_src": "Highwire",
"pdf_hash": "641f325917cf8b6c7ac3490b67da5cc7ee83eb10",
"s2fieldsofstudy": [
"Psychology"
],
"extfieldsofstudy": [
"Medicine"
]
} |
216203181 | pes2o/s2orc | v3-fos-license | Buffaloes Production and Reproduction Efficiencies as Reviewed for Parity in Nepal
This paper can be downloaded online at http://ijasbt.org & http://nepjol.info/index.php/IJASBT 1 Buffaloes Production and Reproduction Efficiencies as Reviewed for Parity in Nepal Sadikshya Lamsal, Deepak Subedi, Krishna Kaphle Paklihawa Campus, Institute of Agriculture and Animal Science, Tribhuvan University, Siddharthanagar-1, Rupandehi, Nepal Department of Theriogenology, Paklihawa Campus, Institute of Agriculture and Animal Science, Tribhuvan University, Siddharthanagar-1, Rupandehi, Nepal
Introduction
Livestock is the major part of Nepalese farming system and plays significant role in Nepalese economy. It contributes around 11% to the national Gross Domestic Product and more than one-third of the total agriculture gross domestic production. Buffaloes are one of the major livestock raised in Nepal. Buffaloes are raised for milk, meat, draft power and manure. The world population of buffalo is estimated as 199 million (FAO, 2012) with more than 96% of the population located in Asia. Buffaloes are the main source of milk production in Nepal producing about 65.3% of the total annual milk in the country, which is equivalent to 1210 thousand MT of milk (MOLD, 2017) (Fig. 1). According to Ministry of Agriculture Development's Livestock Population Report 2015/16, there are around 5 million buffaloes scattered across the nation.
In Nepal, buffaloes are reared in mixed farming systems and their production systems vary greatly across the agro-eco zones. In Southern Terai, inner Terai and mid-hill, buffaloes are mainly kept under complete stall-feeding during seasons of crop cultivation and are occasionally allowed to graze freely in the crop fields whenever there are no standing crops. In the Himalayan foot hill region, grazing of buffalo in pasture, forests and recently harvested crop fields is more frequently seen, although people keep the milking buffalo within the fence of the homesteads. In the high hills and mountains, they are even reared under migratory herds and they are taken up to the high altitude of alpine pasture, sometimes beyond 3500m. Buffaloes in Nepal are well known for their ability to feed on low quality forage, as they utilize crop residues such as straws of rice, millet, wheat and legumes across all agro-eco zones, but are also fed on green forage. Usually milking buffalo are offered with high quality green fodder and some supplemental grains on regular basis (Rasali, 2000). Buffalo population in Nepal can be broadly classified into three groups based on their breed characteristics-Hill buffalo, Terai buffalo and Indian breeds. Lime, Parkote and Gaddi are the three breeds of Hill buffalo. Lime are found in greater number in the northern areas of high hills and mountains, Parkote are found more towards the southern mid hills. Gaddi buffaloes are found in the Far-western Development Region (Rasali, 1998). Due to lack of systematic study on the population and breed characterization, Terai buffalo are largely considered as the non-descript type. Apart from these indigenous buffalo, about 10 % in the hills and little over 10% in Terai, of the total buffalo population, are said to be of Indian Murrah breed or their crosses which have come into existent in the various pocket areas of the country as a result of crossbreeding programme and occasional imports of buffaloes from India. There has been also introduction of Nili Ravi breed into Nepal (Rasali, 2000). Year wise milk production trend of buffaloes is shown in Fig. 2. There is great variation in production and reproduction ability between these breeds and within these breeds. These variations may arise not only due to genetic factors but also due to different non-genetic factors like management practices, environmental variations, season of calving, nutrition, parity, disease condition etc. Among these factors, parity plays an important role in some of the milk production and reproduction trait.
Production Traits
Different production traits that are considered in this study are Lactation length, Lactation yield, Colostrum period and Days to reach peak milk yield. These traits are considered as economically important and some of them are greatly affected by increasing parity.
Lactation Length (LL)
It is the time period of secretion of milk from mammary glands after parturition until another calving. Lactation length of 305 days is considered as ideal length for buffaloes. Lactation length was high (309 days) in first parity and was low (284 days) in sixth parity (Chaudhry, 1992) compared to LL of 265 days (shortest) and 278 days (longest) in fifth and first parity (Afzal et al., 2007) respectively which was not much different in Nili Ravi buffaloes in Pakistan. LL of Parkote, Lime and Murrah cross was 285 days, 276 days and 269 days respectively in western hill of Nepal which showed shortest LL in Murrah cross and longest LL in Parkote (Shrestha et al., 2005) and average LL of local buffaloes was 326 days and that of Murrah cross was 321 days in eastern hills (Shrestha et al., 1994). LL of 333 days was observed in first and second parity and was increasing in fourth and above i.e. 355 days in Murrah crossbred (Poudel et al., 2017) (Fig. 3). In research performed in western hills, LL of 364 days was observed in first-third parity, 346 days in fourth-sixth parity and 394 days in seventh and above in Lime buffalo (Sharma et al., 2017). LL of 356 days in first-third parity, 364 days in fourth-sixth parity and 394 days in seventh and above was reported in Parkote breed (Sharma et al., 2017). LL of 310, 291 and 268 in first, second and third parity were observed in Venezuela in water buffaloes (Nava-Trujillo, 2018).
Lactation yield
Total milk yield per lactation is one of the most important characteristics of dairy buffaloes. Lactation yield (LY) was high i.e. 2150 kg milk in seventh parity and was low i.e. 1818 kg milk in sixth parity (Chaudhry, 1992) whereas milk yield was least in parity 1 and high in parity 4 and yield of first parity was not much different from fifth, sixth and seventh parity (Afzal et al., 2007) in research conducted on Nili Ravi buffalo in Pakistan. There was least milk reduction in third parity followed by 2,4,1,5 and 6 and this result indicates that third parity is best in dairy buffalo for milk production in Pakistan (Khan, 2008). Total LY as high as 2486 liters in fourth parity and as low as 2061 liters in sixth parity in Haryana in Murrah buffalo was observed (Verma et al., 2017). In Nepal, high yield in second parity (2301 lit) and low yield in fifth-sixth parity (1176 lit) in standard 305 days MY in Murrah and indigenous breed of Terai region was reported (Hayashi, 2006) (Fig. 4). In standard 305 days MY, 948 lit MY in first parity in indigenous buffalo which was lowest and 1067 lit MY in fifth parity which was highest was reported (Shrestha et al., 2005) compared to the report of (Rasali et al., 1996) where, MY was high in seventh and lowest was in ninth parity in western hills of Nepal. In local breed, LY was highest in first parity and lowest in seventh-eighth parity compared to LY of Murrah crossbred which was reported highest in fifth parity and lowest in sixth and above (Shrestha et al., 1994). It was found that, in 305 days MY, 1174 liters was produced in first parity which increased in later parities and was highest in fourth, but it was in decreasing order after fourth in local as well as Murrah crossbred (Rasali et al., 1996). TLY of water buffaloes in Venezuela was found to be 1327,1362 and 1345 liters in first, second and third parity which showed not much variation in total lactation yield (Nava-Trujillo, 2018). Average daily MY of Lime was 1.96 liter in first-third parity, 2.52 liter in fourth-sixth parity which was high and again MY decreased to 1.84 liter in seventh and above compared to Parkote which has high MY in first-third parity and was low in seventh and above in western hills of Nepal (Bhattarai et al., 2018).
Colostrum Period
Colostrum period of 3-5 days is generally reported after calving and then milk composition is normal. Colostrum period of first to fourth parity was reported as 5 days whereas colostrum period above fourth parity was 6 days (Poudel et al., 2017). There was not much variation in colostrum period of Murrah buffaloes in Central Nepal. 3 days of colostrum period in first to sixth parity and 2 days in seventh and above in Lime breed whereas 3 days of colostrum period in Parkote breed in all parity (Sharma et al., 2017). Similar values of colostrum period in both Lime and Parkote was reported and parity had no significant effect on colostrum period. It was suggested that colostrum period of Murrah is up to 5 days irrespective of management practice (Smijisha et al., 2012). There will be higher economical return if there is low number of colostrum days (Poudel et al., 2017).
Days to reach peak milk yield
The time required for maximum milk production after calving is called as days to reach peak milk yield (DPMY). It is considered that shorter the time of DPMY, it is economically more profitable but there must be persistency in milk yield. Days to reach peak milk yield (DPMY) was recorded as 21 days in third-fourth parity but it was 24 days in first, second and above fourth parity in Murrah crossbred in central Nepal (Poudel et al., 2017). It was reported DPMY of 14 days in first to third parity and 13 days in fourth and above in Lime breed but DPMY was high in seventh and above parity in Parkote breed (Sharma et al., 2017). This research suggested there was no difference in the value of DPMY with respect to parity in both Lime and Parkote. The value of DPMY was not different in both Parkote and Lime breed whereas higher value of DPMY (5-6 weeks) was observed in case of Murrah crossbred (Poudel et al., 2011). DPMY of Lime breed was around 36 days. There was not much effect of parity on DPMY and average DPMY was 48 days in Nili Ravi breed in Pakistan (Chaudhry et al., 2000).
Reproduction Traits
Different reproduction traits that are considered in this study are Calving interval, Dry period and Service period. These traits are also considered economically important and greatly affect the lifetime production of animal. Also, these traits are affected by increasing parity.
Calving Interval
In buffaloes, one calf for every 15 months is considered profitable. Available evidence shows that calving interval (CI) of buffaloes is between 12-15 months. CI of 470 days (longest) in first parity and shortest i.e. 451 days in fifth parity (Jamal et al., 2018) and CI of 467 days in first parity and 438 days in third parity (Sanker et al., 2014) was reported in study conducted in India in Murrah buffaloes. The CI for water buffaloes in Venezuela was 497 days in first parity and 418 days in third parity and more (Nava-Trujillo., 2018). CI of 645 days was recorded in the study conducted by Mishra in Gaddi buffaloes in Dadeldhura and Baitadi in 2002. In another study (Rasali, 1996), longest CI of 624 days was reported in tenth parity and more and shortest i.e. 505 days was found in ninth parity and 521 days of CI was reported in fifth parity in indigenous buffaloes in western hills of Nepal. Similarly, longest CI of 646 days was reported in sixth parity and 476 days which was shortest CI was reported in fifth parity in Murrah crossbred and indigenous hill buffaloes in Nepal (Rasali, 1996). Average CI of Murrah cross was shorter (545 days) compared to Lime (600 days) and Parkote (604 days) in research conducted in NARC, Lumle (Shrestha et al., 2005).
Dry Period
Dry period of 175,135 and 105 days in first-third, fourthsixth and above sixth parity was recorded respectively in Lime and Parkote (Sharma et al., 2017). Value of dry period decreased with increase in parity. The dry period is longer (290 days) in 1st parity in Nili Ravi buffalo and there was gradual decrease in value of dry period in later parities (Basir et al., 2015). Similar result i.e. gradual decrease in dry period value with increasing parity was also reported (Kandasamy et al., 1993). Dry period was high (135 days) in first-second parity and low (87 days) in third-fourth parity in Murrah crossbred in central Nepal (Poudel et al., 2017). Dry period of 144 days was reported in different buffaloes which was higher than that of optimum range (Sanker et al., 2014).
Service Period
Optimum service period is considered 60-90 days for buffaloes. Service period of Lime, Parkote and Gaddi was recorded as 190, 175 and 350 days respectively (Neupane et al., 2007). Service period of 198,195 and 212 days in Lime, Murrah cross and Parkote respectively (Shrestha et al., 2005). Service period of 112 days was recorded in first parity and its value was declining and was lowest (57 days) in tenth parity and more in local and Murrah crossbred (Rasali, 1996). Longest service period was 214 days in first parity and shortest was 152 days in sixth parity in Nili Ravi buffaloes in Pakistan (Cady et al., 1983). Longest service period was observed in first parity (157 days) which had significant differences from other parities and shortest service period was recorded in sixth parity (137 days) in Murrah buffalo in India (Jamal et al.,2018).
General Remarks
Generally, past research works conducted in Nepal as well as in India, Pakistan and other countries to evaluate production and reproduction traits of different breed of buffalo clearly shows great range of variation. Research in Nepal showed that the traits such as lactation length and total milk yield were low during early parity and increased mostly during mid parity and then declined gradually in later parities but the traits such as service period, calving interval and dry period had high value in early parities and gradually decreased in later parities. The reasons for variation in production and reproduction traits may be due to differences in climatic condition, management practices, nutrition, proper health care and differences in breed as well. But there was not much variation in colostrum period and days to reach peak milk yield with respect to parity. Also, low performing animals were culled after third and fourth parity and only high performing animals were kept for production and thus this might be the reason for better performance of animals in later parities as shown in some research. Due to culling of animal after fourth and fifth parity, there was low availability of animal for research purpose.
In Nepal, the performance efficiency of Murrah crossbred was found comparatively higher than that of indigenous breed such as Lime, Parkote and Gaddi. There has been some increase in production and reproduction performance as a result of the national programme of crossbreeding the indigenous buffalo with Indian Murrah. Efforts are being made to improve the productivity of indigenous buffaloes. Several researches were conducted in research centers in the past to evaluate and improve the productivity of indigenous buffaloes. Thus, appropriate breeding or mating plans along with proper nutrition supply and good management practice would be a key to improve the reproduction and production efficiency of local buffaloes.
Conclusion
It is clear from the information presented in this review that the some of the performance traits of buffalo breeds of Nepal is highly variable in relation with parity. However, there is good potential to increase production and reproduction efficiency of indigenous breed by breeding with Murrah buffaloes as well as by providing proper nutrition, good management practices and proper health care. Also, it would be more economical to cull the animals after fifth-sixth parity as there was not much production and reproduction efficiency.
Author's Contribution
All authors contributed equally in all stages of research and preparation of manuscript. Similarly, final form of manuscript was approved by all authors. | 2020-04-02T09:04:43.993Z | 2020-03-29T00:00:00.000 | {
"year": 2020,
"sha1": "610a5f27bf1a36c328a6225d464879cb890dc14c",
"oa_license": null,
"oa_url": "https://www.nepjol.info/index.php/IJASBT/article/download/27802/23248",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "75e6a691eaaa840959fffba6bea6ae794b8b8d35",
"s2fieldsofstudy": [
"Agricultural and Food Sciences"
],
"extfieldsofstudy": [
"Biology"
]
} |
228929330 | pes2o/s2orc | v3-fos-license | Regulation’s influence on EU banking efficiency: An evaluation post crisis
Abstract This paper examines the impact of regulatory policies on banking market efficiency using a sample of 678 commercial banks from 21 European Union countries for the post-crisis year 2010, controlling for bank-specific and country-specific variables. Data on regulation, supervision and monitoring variables, and activity restrictions are from the most recent Bank Regulation and Supervision Survey database conducted by the World Bank, published 2012. Besides these we incorporate bank size, equity, market share, government ownership, and growth of Gross Domestic Product per capita, employing an Ordinary Least Squares method. Focus is on two alternative measures of banking market efficiency: net interest margin and overhead costs (operating expenses to assets). Elevated levels of these two ratios should indicate a low level of banking efficiency. The evidence suggests that the link between capital regulation and banking efficiency is not robust enough to control for other regulatory variables. Results confirm that activity restrictions have a negative and significant impact on banking efficiency. Policies encouraging official supervisory power do not enhance efficiency of the banking sector. The only approach positively and statistically significantly associated with efficiency is private monitoring. This leads to the suggestion that government regulation and supervision should be more focused on promoting transparency of information.
Introduction
The global financial crisis has revived the importance of the banking system to the economy. As a result of the crisis, bank regulations have become more stringent to prevent future bank failures and safeguard the economy from negative externalities. In this regard, policymakers concentrate on which regulation is the best to achieve these goals. Thus, different regulatory approaches should be analysed as well as the effects of combining approaches. The most common are capital regulation, activity restrictions, official supervisory power and private monitoring. Following this, the present research studies each of these regulatory measures.
The globalization of banking services, the increasing competition and the technological changes have emphasized the importance of banking efficiency. Since banks act as financial intermediaries between lenders and borrowers, banking efficiency plays a crucial role in the social welfare. It is important that banks work to minimize costs, thus promoting efficient utilization of resources. Therefore, enhancing the efficiency of the banking system is vital to achieve greater social welfare. We rely on two accounting ratios to measure banking efficiency: net interest margin and overhead costs. While imperfect, these measures can signal intermediation inefficiency and excessive market power. Hence, higher net interest margin and higher overhead costs reflect lower levels of banking efficiency.
Nevertheless, banking literature provides conflicting predictions about the impact of bank regulation on banking efficiency. As argued by Barth et al. (2006), the different arguments usually reflect the two divergent views, which are the public interest view and the private interest view. The public interest view supports that government operates to enhance social welfare. The purpose of regulating the banking sector is to promote banking efficiency and prevent bank failure. Conversely, the private interest view states that government uses regulatory policies to satisfy political interests (Shleifer & Vishny, 1998). In line with this argument, regulation may exert adverse effects on banking efficiency.
The most common justification for any regulation stems from market failures due to asymmetries of information, externalities and market power. However, there is still no consensus concerning whether the banks should be regulated, and which regulatory policy generates better results. Some argue that capital regulation decreases the occurrence of bank runs, cushioning the economy from major losses and reduce agency problems between bank owners and depositors, especially through deposit insurance. Others argue that stringent capital requirements encourage risk-taking behaviour (Blum, 1999;Boyd et al., 1998;Furlong & Keeley, 1989). By allowing banks to engage in broader activities, banks may become so important to the economy that they are deemed to be "too big to fail", thus difficult to discipline. The power enjoyed by such complex banks may impede competition in the banking sector. On the other hand, restricting activities impedes the exploitation of economies of scope and scale (Laeven & Levine, 2007). The official supervisory approach stresses that supervisors have the incentives to ameliorate market failures as a consequence of informational asymmetries. Conversely, proponents of private monitoring claim that official supervisors use their power to benefit themselves. Hence, government regulation that promotes private monitoring in terms of more transparency of information may be the most efficacious approach to enhance banking efficiency.
This study assesses empirically the impact of these regulatory measures on banking efficiency. The sample consists of 678 commercial banks from 21 European Union (EU) countries for the year 2010. The information for the regulatory variables is collected from Bank Regulation and Supervision Survey database conducted by the World Bank (Barth et al., 2004(Barth et al., , 2008(Barth et al., , 2013a(Barth et al., , 2006(Barth et al., , 2013bCihak et al., 2012), which provides unique information on how banks are regulated around the world. From the four surveys available, we use the data from the last survey, which is the first examination of this kind conducted after the financial crisis. It covers information on banking policies in 125 countries and was completed in 2012. We also follow the indexes aggregating the responses from the survey of Barth et al. (2006Barth et al. ( ), (2013a).
This study empirically answers the following research question: What is the impact of bank regulatory policies on banking efficiency? In addition to the regulatory variables, bank and countryspecific control variables are included in our analysis. We incorporate bank size, bank equity, bank market share, government-owned banks and growth of Gross Domestic Product (GDP) per capita. To estimate the empirical models, we employ the Ordinary Least Squares (OLS) method.
The study is organised as follows. Section 2 summarises the relevant banking literature and the main findings along with the hypotheses we aim to test. Section 3 presents the data and methodology, which includes data sources, measurement of the variables, descriptive statistics, and empirical model. Section 4 presents and discusses the empirical results. Section 5 concludes. The last section also discusses limitations and recommendations for future research.
Literature review
In this section, we summarize the prior literature related to the effects of Basel II Capital Accord regulations namely, capital requirements (Pillar 1), official supervisory power (Pillar 2) and market discipline (Pillar 3), in addition to activity restrictions on banking efficiency as well as on bank development, soundness, efficiency and risk incentives. Existing theories regarding different regulatory approaches are examined first, followed by the empirical evidence supporting these theories and hypotheses we aim to test.
Capital regulation
There is extensive literature focusing on capital regulations emphasising its importance as a key approach to bank regulation. The capital adequacy requirements stipulate the amount of capital that banks need to hold for a certain level and quality of assets. Consequently, it influences the bankingsector efficiency. In the case of bank owners being required to increase the amount of capital at risk, the gains from increased risk would be counterbalanced by the greater likelihood of capital loss. Hence, capital regulation tends to reduce agency problems between bank owners and depositors (Barth et al., 2006;Kaufman, 1992;Keeley & Furlong, 1990). Moreover, two of the primary functions of capital are risk diversification and lower risk-taking. The former considers capital a buffer which protects the assets from being sold or written off at a very low price. An appropriate level of capital safeguards the debt holders and equity holders. The latter implies that, under more stringent capital regulations, bank managers and owners engage in less risky activities (Chortareas et al., 2012). Based on these arguments, capital regulation is expected to enhance banking efficiency.
However, the theory provides conflicting predictions about the positive effects of capital requirements (Santos, 2001). For instance, Barth et al. (2013b) claimed that the arguments in favour are based on the public interest view and the inherent costs of regulating bank capital as higher barriers and higher taxes charged by governments are not considered. The private interest view opposes capital regulation except when the benefits compensate the costs. Furthermore, some authors find that capital regulation increases risk-taking behaviour (e.g. Besanko & Kanatas, 1996;Koehn & Santomero, 1980). According to Blum (1999), stringent capital requirements increase risktaking behaviour and risk of failure as a result of two effects. First, it lowers profits. If the bank has low profits, the cost of bankruptcy is less costly than the cost of excessive risk-taking. Lastly, it affects the marginal return on risk. In the model presented by the author, the marginal return on risk may increase, and thus the first effect may strengthen, further increasing the risk. Therefore, capital regulation may negatively influence banking efficiency.
Additionally, Pasiouras et al. (2009) highlighted the impact of capital requirements on the quantity and quality of loans, the allocation of asset portfolios and the source of funds. The introduction of regulatory capital requirements on an unregulated banking system may decrease aggregate lending while there is no consensus concerning loan quality (Deli & Hasan, 2017;Kopecky & VanHoose, 2006). Bank managers may also look for other forms of assets rather than loans, affecting banking efficiency as different assets have different returns and require different management. Despite the advantages of portfolio diversification, little is known about the capacity of banks to manage different assets (VanHoose, 2007). Lastly, capital regulation may influence the decision of banks with respect to the trade-off between debt and equity. Barth et al. (2004) analysed the effects of bank regulation and supervision on bank development, stability, and performance for 107 countries based on the first survey from 1999 of the Bank Regulation and Supervision Survey database carried out by World Bank. The results indicated that tightening capital requirements lowers the number of non-performing loans (used as a proxy for bank stability) while capital stringency is not robustly linked with banking-sector development and performance as measured by accounting ratios (net interest margin and overhead costs) when controlling for other regulatory and supervisory practices. Likewise, Beck et al. (2006b) addressed the influence of capital regulation on banking system stability using the same data for regulatory measures and found no statistically significant link between the two variables. However, the link is robust when using the capital regulatory index as a control variable for the relationship between bank concentration and fragility.
Empirical review
More recently, Barth et al. (2013b) investigated the impact of regulation, supervision, and monitoring on banking efficiency using the non-parametric Data Envelopment Analysis (DEA) approach for a sample of 4050 banks over 1999-2007. The regulatory and supervision variables are from the first three World Bank surveys of Barth et al. (2004Barth et al. ( ), (2006Barth et al. ( ), 2008. They found that stringent capital requirements are significantly associated with higher levels of banking efficiency and also the possible existence of a connection between banking efficiency and safety. Similarly, Chortareas et al. (2012) provided evidence of a positive and significant relationship between capital regulation and banking efficiency using the DEA approach 1 as well as using accounting ratios: net interest margin and cost-toincome ratio. The study comprised 22 EU countries for the years 2000-2008 using the information on regulatory variables from the World Bank dataset. These findings are consistent with the view that strengthening capital regulation is associated with higher levels of banking efficiency.
Previous studies (Demirguc-Kunt et al., 2008) on the soundness of the banking system have highlighted the importance of regulation and supervision for sounder banks. For instance, Pasiouras et al. (2006) contributed to this area researching the relationship between the Fitch individual bank ratings and bank regulations (World Bank database) with a sample of 857 banks for the year 2004. Their results revealed that lower ratings are attributed to banks operating in markets with tighter capital regulation. Therefore, stringent capital regulation is negatively associated with bank soundness. This is consistent with the argument that banks in countries that promote tighter capital adequacy requirements take on greater risk but inconsistent with the positive influence that capital regulation exerts on bank efficiency.
Regarding risk-taking, Laeven and Levine (2009) conducted a study of 279 banks in 48 countries analysing the risk-taking behaviour by banks, ownership structures and bank regulations. The empirical evidence of their study has shown that the interaction between capital regulations, assessed by minimum capital requirements and capital stringency index, and risk critically depends on the ownership structure of each bank. Thus, while capital regulation decreases risktaking in widely held banks, it has an adverse effect when the bank has a large owner. This finding confirms that capital stringency reduces the benefits of owning a bank, hence large and powerful owners undertake riskier activities in response to more stringent capital regulations. Moreover, the paper emphasized the importance of including ownership structure in the analysis of the effects of capital regulations on risk-taking behaviour.
In line with previously mentioned studies on banking efficiency and considering the empirical evidence, we aim to test the following hypothesis: Hypothesis 1: Capital regulation does not have a positive or significant impact on banking efficiency.
Activity restrictions
Although activity restrictions are not a pillar of Basel II, they have been stressed as fundamental to the study of the impact of regulation on the proper functioning of the banking sector. The question is whether the existence of financial conglomerates 2 is beneficial to the performance of the banking system. Nevertheless, there is no consensus in theory about the effects of banks engaging in broader activities or restricting their activities.
The information acquired by banks during the lending process may help banks to be more efficient in other activities such as brokering, real estate investment or dealing mutual fund securities. In the same way, by engaging in broader activities, banks acquire information that can be used to improve the process of making loans. Therefore, restricting activities may impede banks from enjoying economies of scale and scope as it limits the exchange of information (Diamond, 1991;Rajan, 1992;Stein, 2002). Moreover, it can lower the franchise value of a bank due to greater difficulty in diversifying revenue streams and thereby narrowing incentives for more efficient conduct. By being responsible for licensing banks and specifying the permitted activities, regulators become more powerful which gives them more opportunities to pursue economic gains (Djankov et al., 2002). Finally, Beck et al. (2006b) presented empirical evidence that tightening activities boosts bank fragility by impeding banks from diversifying portfolio risk. As a consequence of the previous arguments, allowing banks to engage in non-traditional activities is likely to improve banking efficiency.
On the other hand, some (Aron, 1988) argue that expanding the range of financial services may intensify agency problems between owners and shareholders which reduces the market value. Bank managers may engage in broader activities simply because it is more advantageous to them even if it means a lower market value for the financial institution. Therefore, the decision to restrict activities by managers may be due to what benefits them instead of what benefits the firm (Jensen, 1986). Also, moral hazard problems may arise which stimulates risk-taking behaviour (Boyd et al., 1998). Furthermore, it is hard to implement efficient managerial contracts in financial conglomerates as well as handle agency problems (Rotemberg & Saloner, 1994). Another argument is that banks providing a wide range of services may become so important to the economy that they are deemed to be "too big to fail" thus, difficult to discipline. Under these circumstances, activity restrictions may positively affect banking efficiency. According to Laeven and Levine (2007), diversification of activities decreases the market value of the financial conglomerate since the benefits of enjoying economies of scope are not significant enough to offset the negative effects of leading with a variety of activities. Consequently, restricting bank activities may enhance banking efficiency.
Researchers (King and Levine, 1993) focusing on the impact of activity restrictions on banking efficiency and development have concluded mixed results for the association between the two variables. The results are consistent with the view that relaxing restrictions on bank activities improves banking efficiency by increasing opportunities for diversifying sources of income. Barth et al. (2004) found that while regulation of activities negatively affects bank development, there is no robust link between this regulatory measure and net interest margin or overhead costs or non-performing loans. The former finding is notably important since the development of the banking system directly influences economic growth. 3 In addition, their study supported that a banking system with greater freedom on activities is less likely to suffer a major crisis. Kim et al. (Role of financial regulation and innovation in the financial crisis, Kim et al., 2013) carried out a study on the role of regulatory measures in the financial crisis 4 using the surveys on bank regulation for 143 countries. Unlike Barth et al. (2004), they concluded that activity restrictions reduce the likelihood of a banking crisis. Therefore, diversification drives banks to engage in riskier activities, further increasing bank fragility. Chortareas et al. (2012) confirmed that greater restrictions on bank activities exert a positive impact on net interest margin, cost-to-income ratio and inefficiency, which means an overall negative impact on banking efficiency. Barth et al. (2013b) also corroborated that stricter restrictions on bank activities are negatively associated with the efficiency of the banking system, using the DEA approach. Moreover, Demirguc-Kunt et al. (2004) investigated the influence of bank regulations on the cost of financial intermediation using as proxies the net interest margin and the overhead expenditures. The sample comprised 1400 commercial banks in 72 countries, from 1995 through 1999. Similar to the studies mentioned above, the regulatory measures were obtained from the World Bank surveys. The results showed that greater activity restrictions are significantly associated with greater net interest margin, consequently greater operational inefficiency and market power. With respect to the soundness of the banking system, Pasiouras et al. (2006) found that lower ratings are allocated to banks operating in markets with stricter activity regulations. Thus, restricting bank activities negatively influence banking efficiency. Laeven and Levine (2009) proposed in their study that tighter restrictions on bank activities increase risk-taking behaviour when the bank has a sufficiently large owner. If the bank is widely held, the effects of restricting activities on risk are not significant. Since regulation on activities lowers the profits, bank owners may engage in riskier activities to compensate for losses. In this respect, Gonzalez (2005) evaluated the impact of activity restrictions on risk incentives in a study with 251 banks in 36 countries from 1995 to 1999. Nevertheless, the evidence has confirmed that greater regulatory restrictions are related to fewer risk-taking incentives after isolating the relationship between bank charter value 5 and risk-taking incentives and between regulatory restrictions and bank charter value. The findings support that greater freedom on bank activities promotes risk-taking as it provides banks more opportunities to assume greater risk. Hence, the two studies provided conflicting predictions about the effects of activity restrictions on bank risk.
Considering the empirical evidence presented in previous banking literature on the effects of activity restrictions on banking efficiency, we aim to test the following hypothesis: Hypothesis 2: Activity restrictions have a negative and significant impact on banking efficiency.
Official supervisory power
The official supervisory power concerns the degree to which bank supervisors can take actions against bank owners, bank managers and bank auditors in different circumstances. While there is an extensive literature on the importance of empowering official supervision of banks, empirical evidence (Levine, 2005) shows no consensus about the effects on bank performance. Since the costs of supervising banks are excessive, there are few incentives to monitor banks. Proponents of public interest view claim that supervisors have incentives to ameliorate market failures through the monitoring and discipline of banks (Beck et al., 2006a), thus undermining corruption and strengthening the intermediary function of banks. Moreover, the existence of informational asymmetries makes banks vulnerable to contagion risk and possible bank failure. Powerful supervisors can intervene effectively under these circumstances. Besides, with the adoption of deposit insurance, banks may undertake greater risk and depositors may have less incentives to supervise banks. Official supervisors may be capable of preventing such situations (Barth et al., 2004;Beck et al., 2006b). Given the previous arguments, it is expected that official supervisory power enhances banking efficiency.
On the contrary, proponents of private interest view claim that official supervisors use their power for their own interest. As powerful supervisors have access to insider information, they may use it for private and political benefits (Beck et al., 2006b). Political candidates may bribe supervisors to extract votes and donations for their campaign (Djankov et al., 2002;Quintyn & Taylor, 2003). Hence, official supervisory power may be associated with higher levels of corruption, negatively influencing banking efficiency. Additionally, supervisors may pursue private interests instead of public interests to enhance their reputation as being able to monitor banks, especially when there are uncertainties about their capacity to do so (Boot & Thakor, 1993). Thus, powerful supervisors may adversely affect banking operations. Therefore, official supervisory power may positively or negatively affect the corporate governance of banks, the lending process (by encouraging corruption), and the intermediary function of banks.
Nevertheless, the argument that an independent supervisory entity has positive influence on banking efficiency is consensual among the two opposing views. Being independent, the supervisors are less likely to exert power on banks to serve political and private interests. Thus, independent supervisors could monitor the corporate governance of banks more professionally. Moreover, they may better guide banks by giving them advice on how to improve (Barth et al., 2013b).
The empirical evidence indicates that official supervisory power is not significant in explaining bank performance itself. Barth et al. (2004) found that empowering official supervision of banks is not significantly associated with bank development, bank performance or the number of nonperforming loans. The link between supervision and the likelihood of a banking crisis also proved to be non-significant. Similarly, the regressions in the study of Barth et al. (2013b) pointed to an insignificant relationship between official supervisory power and banking efficiency. The study also evaluated the impact of independent supervisory authorities and the interaction between the two variables. Their results suggested that while official supervisory power does not enhance banking efficiency itself, promoting the independence of supervision may strengthen banking efficiency. Furthermore, the interaction term confirmed that empowering official supervision leads to an increase in banking efficiency in countries with greater independence of supervisors. This is consistent with the argument that independent supervisors are less likely to pursue political and private interests rather than social interests, improving the corporate governance of banks. Levine (2005) assessed the impact of supervision, specifically the official supervisory approach and the private monitoring approach, on banking efficiency measured by overhead costs using the data for the supervisory measures from Barth et al. (2004). The connection between official supervision and overhead costs has proven to be statistically insignificant. The research emphasized that the aim of the official supervisory approach may be to lower bank fragility instead of enhancing bank performance as a possible justification. Notwithstanding the main finding, the data indicated that empowering official supervision intensifies corruption in bank lending. Hence, the data is consistent with the private interest view. Different results are suggested by Chortareas et al. (2012). The evidence indicated that official supervisory power has a positive impact on banking efficiency, which confirms the predictions of the public interest view. The results held using inefficiency, net interest margin or cost-to-income ratio as independent variables, with coefficients being negatively significant in all regressions. However, when controlling for private monitoring, the impact turned negative. One explanation may be that strengthening official oversight in less developed countries leads to increased government participation and thus undermines the reliability of the lending process. Beck et al. (2006a) evaluated the association between supervision and corruption in bank lending in a study with 2500 firms from 37 countries using the information on supervision from Barth et al. (2008;. Corruption in bank lending is proxied by the extent to which bank supervisors' corruption is a barrier for firms to get external finance. The results supported that strengthening official supervisory power is positively and significantly related to higher levels of corruption. This suggests that powerful supervisors do not improve social welfare, they improve their own welfare. Therefore, policies that encourage official oversight to enhance banking efficiency may have opposite outcomes, ultimately increasing corruption in the banking system. The analysis of Pasiouras et al. (2006) concluded that more powerful supervision is adversely associated with bank ratings. Thereby, countries that promote official supervision to take specific actions on bank managers and owners obtain lower bank ratings. Under these circumstances, official supervisory power is negatively related to the soundness and efficiency of the banking system and positively related to bank corruption. The study of Demirguc-Kunt et al. (2008) which examined the influence of compliance with the Basel Core Principles on Moody's ratings noted that countries with a higher degree of compliance are better rated, although empowering official supervision per se does not influence bank ratings.
The evidence presented in the literature on the effects of official supervisory approach on banking efficiency is divided into two viewpoints, those that found negative and significant effects and those that found insignificant effects. This leads us to the following hypothesis: Hypothesis 3: Official supervisory power has a negative impact on banking efficiency.
Private monitoring
Private monitoring refers to the extent to which banking regulations require banks to disclose accurate and detailed information to the public and improve contract enforcement mechanisms. It is believed that bank supervision should focus on reducing barriers such as the costs of producing and disclosing information. Thus, private investors could monitor more efficiently, strengthening the corporate governance of banks (Hay & Shleifer, 1998). Proponents of the private monitoring approach often express doubts about the effectiveness of the official supervisory approach. Therefore, the arguments in favour of the former approach tend also to be critiques to the latter approach and vice versa (Levine, 2005).
First, as supervisors are not usually bank shareholders, their incentives may conflict with incentives from private investors regarding bank discipline, consequently agency problems may arise (Barth et al., 2013b). Second, supervisors may be driven by self-interests. Politicians and banks may influence official supervisors to act in their own interests by enriching insiders and channelling credit to some specific companies according to the private interest view. Hence, empowering private monitoring over supervisory power may prevent corruption in bank lending from intensifying, improving banking operations at the same time (Beck et al., 2006a). Third, remuneration of supervisors is low in some countries which incentivises them to move to the banking sector. These circumstances may create different motivations for supervisors regarding bank monitoring (Barth et al., 2004). Lastly, while recognizing the existence of market failures that interferes with the effectiveness of private monitoring, Shleifer and Vishny (1998) emphasized that official oversight may have a more significant negative impact on banks due to the great scale of government failures. Following this logic, the benefits of overcoming market failures may not produce sufficient positive effects to promote official supervisory power over private monitoring.
Nevertheless, there are arguments against empowering private monitoring. Private monitoring may not enhance bank performance in countries with less developed legal systems, poor capital markets and ineffective financial accounting policies as people in these countries do not trust in private monitoring. Therefore, they may not be able to benefit from an efficient monitoring with the private sector, but they may benefit from official oversight. Since the banking system is very complex and opaque, even the private sector in more developed countries may find it hard to monitor banks efficiently (Barth et al., 2004). From this standpoint, strengthening private monitoring may lead to bank customers being misled, and consequently deteriorating banking efficiency. Furthermore, costs related to disclosure of accurate information to the public such as enforcement costs, costs to prepare formal disclosure documents and costs of divulging delicate information to competitors are excessive, thus it negatively impacts banking efficiency (Duarte et al., 2008).
While the arguments mentioned above have supported one supervisory approach over the other between official supervision and private monitoring, the two approaches are not mutually exclusive as stated by Levine (2005). Indeed, countries are able to adopt both approaches, that is, requiring banks to disclose precise information to the public at the same time that official supervisory agencies directly monitor and discipline banks.
Most of the existing studies have concluded that empowering private monitoring in terms of more financial transparency has a good impact on the banking system. For instance, Barth et al. (2004) observed that private monitoring positively influences bank development and negatively influences the amount of nonperforming loans and net interest margin. The study, however, failed to find a strong impact on bank crises when controlling for other variables. Barth et al. (2013b) found that private monitoring exerts positive effects on banking efficiency, specifically they found that strength of external auditor, bank accounting informative and certified audit requirements are related with higher levels of banking efficiency. Similar results were obtained in the study of Levine (2005) which concluded that empowering private monitoring reduces the overhead costs, boosting banking efficiency. One particularity of this study was that the effect lost significance in the regression that included GDP per capita. This means that GDP per capita is highly correlated with the degree of information disclosure. In a broader perspective, the economic development of a country is influenced by the implemented supervisory approach, just as the effectiveness of bank supervision depends on the economic development of the country.
Conversely, Chortareas et al. (2012) provided evidence of a negative relationship between private monitoring and banking efficiency. The results should be interpreted carefully since they used a subsample of countries for regressions that included this variable, excluding economically large countries as UK, Germany, France and Spain. A possible reason for this finding is that producing and disclosing information is costly. Hence, banking efficiency may be adversely affected. Beck et al. (2006a) noticed that empowering private monitoring reduces corruption in bank lending. Their paper also established a positive association between bank lending integrity indicators and the private monitoring approach. Hence, corruption of bank officials is less a barrier for companies to raise capital in countries that promote information provision and countries with developed legal systems and efficient governments. The findings are consistent with the argument that private monitoring reduces bank corruption, enhancing banking efficiency. Pasiouras et al. (2006) analysed the impact of accounting and disclosure requirements and auditing requirements on bank credit ratings and found that both coefficients appear negative in the regressions. This is inconsistent with studies that indicate that empowering private monitoring enhances bank performance as well as with those that indicate that accounting and auditing systems reduce bank risk (Fernandez & Gonzalez, 2005). Nevertheless, the coefficients are only significant at 10% level, which is consistent with the findings of Barth et al. (2004) regarding the impact of private monitoring on the likelihood of a banking crisis. Additionally, the empirical evidence of Demirguc-Kunt et al. (B2008) supported that compliance with dissemination of accurate information is positively associated with bank soundness. Thus, banks in countries that strengthen private monitoring are better rated.
Focusing on the findings of the aforementioned studies about the effects of strengthening private monitoring on banking efficiency, we aim to test the following hypothesis: Hypothesis 4: Private monitoring does not have a positive or significant impact on banking efficiency.
This study investigates the impact of bank regulatory measures namely, capital regulation, activity restrictions, official supervisory power and private monitoring on banking efficiency in the EU. The methodology we follow is closely related to various research on banking efficiency. To measure banking efficiency, we use two accounting ratios: net interest margin and overhead costs. Likewise, Barth et al. (2004), Demirguc-Kunt et al. (2004), and Levine (2005) used accounting ratios as opposed to studies that used efficient frontier analysis (Barth et al., 2013b). Chortareas et al. (2012) measured banking efficiency using accounting ratios as well as the DEA technique. Both methods produced the same results. Therefore, we decide to use accounting ratios. However, it is important to emphasize that bank net interest margin and overhead costs are imperfect measures of banking efficiency. In order to overcome this limitation, we control for bank-specific and country-specific variables.
Data and methodology
In this section, we present information about the data and the methodology employed in our research. We discuss the data collected, sources and sample selection, the definition of the variables and the descriptive statistics. This is followed by the introduction of the econometric models, description of the estimation procedure and the tests performed to obtain accurate results.
Data
This study aims to evaluate the impact of bank regulatory measures on banking efficiency. To assemble the cross-section data, we include commercial banks operating in 21 EU countries, namely: Austria, Belgium, Bulgaria, Croatia, Cyprus, Denmark, Finland, France, Greece, Hungary, Ireland, Italy, Luxembourg, Malta, Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia and Spain.
The dataset is compiled from two primary sources, SNL Financial database provided by S&P Global Market Intelligence and World Bank databases. The bank-specific variables: net interest margin, overhead costs, bank size, bank equity and bank market share are obtained from SNL Financial database. The annual growth rate of GDP per capita is obtained from the World Bank World Development Indicators (World Bank, 2010). The regulatory measures: capital regulation, activity restrictions, official supervisory power and private monitoring, as well as the governmentowned banks, are obtained from the last survey of the Bank Regulation and Supervision Survey carried out by the World Bank (Barth et al., 2004(Barth et al., , 2008(Barth et al., , 2013a(Barth et al., , 2006(Barth et al., , 2013bCihak et al., 2012). As the answers from this survey correspond to the year 2010, we collect data for the same year. Moreover, we follow the indexes constructed and made available by Barth et al. (2006) and (2013a), which aggregates the responses to individual questions into indexes.
Since the availability of data varies by country and by bank, we edit the raw data with the purpose of minimising the errors in our analysis. Therefore, our sample comprises only the observations for which all variables are available. The initial sample included 28 EU countries. 5 The Czech Republic and Sweden did not respond to the survey. Hence, we do not have information on any regulatory variable for these countries. Therefore, they are dropped from our sample, resulting in 1218 banks from 26 countries. After dropping the observations without information for the bank-specific variables, there are 1019 banks from 24 countries, Latvia and Lithuania are excluded. Although the United Kingdom, Germany and Estonia responded to the survey, there are missing values for the regulatory variables. For the United Kingdom, the information is not available for official supervisory power; for Germany, information is not available for activity restrictions; for Estonia, information is not available for private monitoring. Consequently, the final sample used in our study to examine the impact of regulation on banking efficiency consists of 678 banks from 21 EU countries.
Net interest margin and overhead costs
To measure banking efficiency, we use two accounting ratios: net interest margin and overhead costs. We collected data for net interest margin and overhead costs from 2010 since the information for bank regulation is from this year. Net interest margin equals interest income minus interest expense divided by total assets. The net interest margin measures the gap between what banks pay to their lenders and what banks receive from their borrowers, hence it reflects the conventional lending and borrowing operations of the bank. Similar to previous papers, we interpret the net interest margin as a signal of intermediation inefficiency and excessive market power, allowing banks to charge higher interest. Thus, higher values of net interest margin reflect lower levels of banking efficiency.
Overhead costs equal operating expenses divided by total assets. High overhead costs reflect unwarranted managerial perquisites as well as higher market power, which is inconsistent with the efficient intermediation of the banking system. Hence, higher values of overhead costs are associated with lower levels of banking efficiency. It is expected that both dependent variables produce identical results (Barth et al., 2008;Demirguc-Kunt et al., 2004;Levine, 2005). In our econometric model, we use the logarithm values of both net interest margin (LOGNIM) and overhead costs (LOGOC) with the aim of normalising the distribution of the variables and consequently decrease their fluctuation. Graphically, we can perceive that the distribution of the variables before the logarithm values exhibits high skewness and high kurtosis. Using the logarithm, the values of skewness and kurtosis are closer to the values for the standard normal distribution.
Differences across banks in net interest margin and overhead costs may reflect differences in bank activities, business systems, or asset allocations rather than differences in banking efficiency. In order to mitigate possible interpretational issues with the measures of banking efficiency and be able to find the independent relationship between them and bank regulation, we control for the bank and country-specific variables.
Regulatory variables
With the purpose of evaluating the impact of bank regulation on banking efficiency, we include four regulatory measures in our study, namely, capital regulation, activity restrictions, official supervisory power, and private monitoring. We provide the description of these variables and the range of the corresponding indexes constructed by Barth et al. (2006), (2013a)). The regulatory measures data were collected from the last survey. Hence, it refers to the year 2010.
The capital regulatory variable (CAPREG) is an index that incorporates the initial capital stringency and the overall capital stringency. The initial capital stringency concerns whether the sources of regulatory capital can include assets other than cash, government securities or borrowed funds and whether the regulatory/supervisory authorities verify them. The overall capital stringency concerns whether Basel I is the applied capital adequacy regime, whether the capital requirement reflects market or credit risk, whether market value losses are deducted, and the fraction of revaluation gains permitted. The capital regulatory index ranges from 0 to 10 with higher values indicating greater capital stringency.
The activity restrictions variable (ACTRES) is an index that measures the degree to which banks are allowed to engage in non-traditional activities, specifically, securities, insurance and real estate activities. Securities activities include brokering, underwriting, dealing, and all aspects of the mutual fund industry. Insurance activities include underwriting and selling. Real estate activities include management, investment and development. The activity restrictions index ranges from 4 to 16 with higher values indicating greater restrictiveness.
The official supervisory power variable (SPOWER) is an index that concerns whether official supervisors have the power to act on bank owners, auditors, and managers in different circumstances with the intention of preventing and correcting problems. This includes whether the supervisory authority is allowed to share information with external auditors, declare bank insolvency, supersede the rights of shareholders, change bank managers and directors, build a legal case against auditors in case of negligence, require banks to constitute provisions to cover losses, require banks to reduce or suspend dividends, bonuses and other remunerations and whether auditors are required to report unlawful activities. The official supervisory power index ranges from 0 to 14 with higher values indicating a greater power of supervisory authorities.
The private monitoring variable (PMONIT) is an index that measures the degree to which supervisors require banks to disclose accurate and detailed information to officials and the public, promoting the monitoring of banks by the private sector. It involves whether professional external auditors audit banks, whether international and domestic rating entities rate banks, whether there is an explicit deposit insurance scheme and the compensation received by insured depositors the last time a bank failed, whether banks are required to produce consolidated accounts and disclose off-balance sheet items, whether accrued or unpaid interest or principal are included in the income statement on performing and non-performing loans, whether bank directors are legal responsible for the accuracy of the information disclosed and whether subordinated debt is allowed. The private monitoring index ranges from 0 to 12 with higher values indicating greater private monitoring of banks.
Control variables
At the bank-level, we use three key control variables: bank size, bank equity and bank market share. At the country-level, we control for the GDP growth and the government-owned banks. We use 2010 values as the data for the regulatory variables corresponds to this year. Since we find the same results when we omit the control variables, endogeneity problems are not biasing our conclusions regarding the effects of bank regulation on banking efficiency.
Bank size (LOGTA) is measured as the logarithm of total bank assets in US dollars. We control for bank size to capture the effect on bank lending behaviour. Moreover, this variable may have a considerable impact on net interest margin and overhead costs in case of increasing returns to scale in the banking industry. Bank equity (EQTA) is measured as the book value of bank equity divided by total bank assets. We use this variable as previous papers (e.g. Berger, 1995) supported that well-capitalised banks face lower bankruptcy costs, hence lower funding costs. Therefore, higher values of bank equity may imply higher net interest margin and overhead costs. Bank market share (MS) equals the total assets of the bank divided by the total commercial bank assets in the country. Banks with greater market share may experience greater market power, consequently greater net interest margin and overhead costs.
Government-owned banks (GOVERN) equals the fraction of the banking system's assets that is held by banks with 50% or more government-ownership. This variable measures the ownership structure of the banking system. La suggested that government-owned banks tend to allocate resources to benefit politicians and therefore tend to be less efficient. According to this perspective, we expect government-owned banks to exert a positive impact on net interest margin and overhead costs. GDP growth (GDPGR) equals the annual growth rate of GDP per capita. We include this variable to control for business-cycle forces of the macroeconomic environment. Since the demand for financial services depends on the cyclical conditions, GDP growth may have a positive influence on banking efficiency.
Descriptive statistics
This paragraph presents and discusses the key descriptive statistics for the dependent and independent variables used in our regression analysis. Table 1 shows the summary statistics of all the variables used to study the impact of bank regulatory measures on banking efficiency. The average for the LOGNIM is −4.24, with a minimum of −11.10 and a maximum of −1.58. The number of observations of this variable (671) is lower compared to the number of observations of all the other variables (678). The difference refers to the banks from our sample with values of net interest margin equal to 0, since the logarithm of 0 is not computable. The average for the LOGOC is −4.09, with a minimum of 9.95 and a maximum of −0.42. Regarding the regulatory variables, CAPREG has an average of 6.70 with a standard deviation of 1.56, a minimum of 4 and a maximum of 9. As the range of this index is from 0 to 10, we can conclude that the 21 EU countries have a medium-high level of capital stringency. For the ACTRES, the average is 6.21 with a standard deviation of 1.7, a minimum of 4 and a maximum of 11. The banks in our sample are less regulated in terms of engaging in securities, insurance, and real estate activities, compared to the capital as the range of the former is from 4 to 16. The average SPOWER is 11.29 with a standard deviation of 1.9, a minimum of 5 and a maximum of 9. Since this index ranges from 0 to 14, we can conclude that the official supervisory authorities in the considered countries have great power to take actions against bank owners, managers and auditors. Lastly, the average for the PMONIT is 8.32 with a standard deviation of 1.06, a minimum of 6 and a maximum of 11. Given that the range of this index is from 0 to 12 and considering the statistics, we can conclude that the banking regulations in the countries of our sample promote private monitoring of banks. The fact that the countries in our sample promote both official supervisory power and private monitoring is consistent with the theory that the two approaches are not mutually exclusive as discussed in the literature review. Considering the control variables, the average LOGTA is 6.53, with a minimum of 3.42 and a maximum of 9.51. The average for the EQTA is 0.12, with a minimum of −0.01 and a maximum of 1. For the MS, the average is 0.03, the minimum is approximately 0, and the maximum is 0.58, meaning that there is at least one bank in a country that holds more than 50% of the total commercial bank assets in that country. The GOVERN has an average of 0.07, a minimum of 0 and a maximum of 0.51. These statistics reflect that there are one or more countries in our sample without government-owned banks, as well as at least one country where government-owned banks hold more than 50% of total banking assets in that country. The average for the GDPGR is 0.01, with a minimum of −0.06 and a maximum of 0.05, reflecting that there are countries with negative growth rates of GDP per capita and even the highest value reflects positive but low growth rate.
Model description
To assess the effect of bank regulatory measures on banking efficiency while controlling for bankspecific and country-specific variables, we estimate the following two models: Where i refers to bank i, t refers to country t, α is the intercept, B i,t is a vector of bank-specific variables for bank i in country t, C t is a vector of country-specific variables in country t and ε i,t is the residual. The measures of banking efficiency are two accounting ratios: the logarithm of net interest margin and logarithm of overhead costs. In model (1), the dependent variable is LOGNIM i which represents the logarithm of net interest margin for bank i in country t. In model (2), the dependent variable is LOGOC i , which represents the logarithm of overhead costs for bank i in country t. CAPREG t is the capital regulation variable in country t and β 1 is the coefficient which explains the effect of capital regulation on banking efficiency. ACTRES t corresponds to the activity restrictions variable in country t and β 2 is the coefficient describing the impact of activity restrictions on banking efficiency. SPOWER t denotes the official supervisory power variable in country t and β 3 is the effect of official supervisory power on banking efficiency. PMONIT t is the private monitoring variable in country t and β 4 describes the influence of private monitoring on banking efficiency. The vectors B i , and Ci including the bank-specific variables and the country-specific variables, respectively, are defined as follows: The vector B i contains bank size, bank equity and bank market share. Bank size is proxied by the logarithm of total assets for bank i in country t (LOGTA I ). Bank equity is measured as the equity to assets ratio for bank i in country t (EQTA i ). Bank market share is assessed by the ratio between the total assets of the bank and the total commercial bank assets in the country for bank i in country t. The vector C t includes the government-owned banks variable in country t (GOVERN i ) and the annual growth rate of GDP per capita in country t (GDPGR t ).
Previous to the regression of model 1 and 2, it is necessary to check whether the variables are suitable for our cross-section data and whether the assumptions about our data are fulfilled. Therefore, we examine the graphical distribution and the measures of skewness and kurtosis of the variables. The skewness and kurtosis values for standard normal distribution are 0 and 3, respectively. The only variables whose values are significantly different from the standards are net interest margin and overhead costs. As previously mentioned, we use the logarithm values to solve this problem by normalising the distribution of the variables. From Figures 1 and 2, we can perceive the difference before and after the logarithm values.
In order to regress the models, we use the OLS estimation as the method that best fits our data. Since heteroskedasticity is often a problem in cross-section data, we test both models for the presence of heteroskedasticity using the Breusch-Pagan test. The null hypothesis of the test refers to the presence of homoskedasticity. The p-value for both models equals 0.00, hence, we reject the null hypothesis. Thus, we use the heteroskedasticity-robust standard errors in our regressions.
Empirical results
In this section, we present and discuss the main results of the estimated regressions. We compare our findings with the expected findings included in our hypotheses, which consider the evidence presented in previous banking literature. Explanations for possible inconsistencies are also examined. Table 2 presents the regressions results of bank regulation on banking efficiency proxied by the logarithm of net interest margin and overhead costs while controlling for bank-specific and country-specific variables. To recall, we interpret the results as higher values of net interest margin and overhead costs are associated with lower levels of banking efficiency. Our results do not suggest a strong relationship between capital regulation and banking efficiency. While capital regulation enters positively in both regressions, the coefficient is not statistically significant. This is inconsistent with the expected positive and significant effect of capital regulation on banking efficiency described in hypothesis 1. The results are also inconsistent with previous studies (Barth et al., 2013b;Chortareas et al., 2012) that found a significantly positive impact on banking efficiency. On the contrary, Barth et al. (2004) reported an insignificant link between capital regulation and net interest margin or overhead costs, which supports the results in Table 2.
The results also indicate that restricting bank activities is associated with lower levels of banking efficiency, as confirmed by the positive and significant coefficients at the 1% level. The size of the coefficients is substantial. That is, a one unit increase in activity restrictions increases net interest margin by 17.5% (100* (e 0.161 − 1)) and overhead costs by 19.1% (100* (e 0.175 − 1)). Moreover, our findings support hypothesis 2 that activity restrictions have a negative and significant impact on banking efficiency. Similarly, Demirguc-Kunt et al. (2004), Chortareas et al. (2012), and Barth et al. (2013b) provided evidence that greater activity restrictions negatively affect the efficiency of the banking system. Table 2 regressions show that there is a negative relationship between official supervisory power and banking efficiency, which supports hypothesis 3. The link is significant at the 5% level in the second regression when considering overhead costs as the dependent variable. For the first regression, official supervisory power is not significant in explaining net interest margin. This is in accordance with previous banking literature ( Barth et al., 2004Barth et al., , 2013bLevine, 2005) that found no robust impact of official supervisory power on banking efficiency as well as with Chortareas et al. (2012) that found a positive and significant impact on overhead costs when including the private monitoring variable.
As for the private monitoring variable, the coefficient enters significantly and negatively in both regressions. For the net interest margin regression, the coefficient is significant only at the 10% level, while for the overhead costs it is significant at the 1% level. The coefficients are considerably large, hence a one unit increase in private monitoring decreases net interest margin by 12.6% (100* (e 0.119 − 1)) and overhead costs by 18.6% (100* (e 0.171 − 1)). The results confirmed that promoting private monitoring enhances banking efficiency as predicted in our hypothesis 4. This finding is in line with previous studies on banking efficiency (Barth et al., 2004(Barth et al., , 2013bLevine, 2005) that presented empirical evidence in favour of empowering private monitoring. On the other hand, our results are in contrast with the study conducted by Chortareas et al. (2012) that found a positive impact of private monitoring on efficiency levels. Note: This table exhibits the estimates from OLS regressions of model 1 and 2, explaining the relationship between bank regulation and banking efficiency. The dependent variable of model 1 is logarithm of net interest margin (LOGNIM) and model 2 is logarithm of overhead costs (LOGOC). The independent variables of both models are: capital regulation (CAPREG), activity restrictions (ACTRES), official supervisory power (SPOWER), private monitoring (PMONIT), bank size (LOGTA), bank equity (EQTA), bank market share (MS), government-owned banks (GOVERN) and annual growth rate of GDP per capita (GDPGR). The symbols *, **, *** indicate the statistical significance of the estimated coefficients at the 10, 5 and 1% level, respectively. The robust standard errors are in parentheses below the estimated coefficients. The number of observations, R-squared and F statistic of each regression are also presented in this Table. Considering the vector of bank-specific control variables, large banks tend to have lower net interest margin and lower overhead costs. Therefore, bank size is positively and significantly at the 1% level associated with banking efficiency. This may be caused by economies of scale in the banking sector. The finding is in accordance with the evidence presented by Altunbas et al. (2007) and Yildirim and Philippatos (2007) for EU countries. The results for bank equity reveal a positive impact on net interest margin and overhead costs. However, the coefficient is only significant in the second regression at the 10% significance level. This finding confirms that well-capitalised banks have more opportunities for higher net interest margin and overhead costs. Bank market share appears to be an essential driver of banking efficiency as confirmed by the positive and significant coefficients at the 1% level in both regressions. This supports the view that banks with more significant market share experience greater market power and consequently, larger net interest margin and overhead costs.
At the country level, we control for government-owned banks and GDP growth. The government-owned banks positively affect net interest margin and negatively affect overhead costs. Nevertheless, the former relationship is not significant while the latter is significant at the 1% level. The banking literature supports that the presence of the government stimulates corruption in the banking industry, hence promoting banking inefficiency. Our results are inconsistent with this view. Additionally, the annual growth rate of GDP per capita enters negatively in both regressions. The estimated coefficients are significant at the 1% level. This finding indicates that expanding markets, as measured by a higher growth rate, tend to have more efficient banking.
Some of our findings are not consistent with previous literature (Altunbas et al., 2007) on the impact of regulatory measures on banking efficiency, especially regarding capital regulation. Nevertheless, our study differs from the previous ones in a few aspects. First, we are considering the EU countries while most of the studies consider a large cross-country sample. Second, we are using the data from the last survey of the Bank Regulation and Supervision Survey carried out by the World Bank. This is the only study evaluating the impact of bank regulation on banking efficiency using the last survey. For instance, Barth et al. (2013a) and Cihak et al. (2012) presented and discussed the responses to the last survey and compared them to earlier surveys. However, the impact on banking efficiency is not assessed. Lastly, this survey is the first survey conducted after the financial crisis. The global financial crisis has drawn the attention of policymakers to the need to reform banking regulation since it has highlighted problems with the banking sector. Hence, discrepancies with previous studies may exist.
Conclusion
The global financial crisis has emphasised the importance of regulatory policies for the proper functioning of the banking system. In this regard, our study contributes to the existing literature by assessing the impact of the Basel II Capital Accord regulations in addition to activity restrictions on banking efficiency. Specifically, we examine the impact of capital regulation, activity restrictions, official supervisory power and private monitoring. In order to do so, we develop a set of hypotheses in line with our predictions considering the existent empirical evidence. We collected data for a sample of 678 banks from 21 EU countries: Austria, Belgium, Bulgaria, Croatia, Cyprus, Denmark, Finland, France, Greece, Hungary, Ireland, Italy, Luxembourg, Malta, Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia and Spain for the year 2010. We used two alternative accounting ratios to measure banking efficiency, namely net interest margin and overhead costs. We used OLS estimation to regress the empirical models. Also, we employed cross-section data techniques to minimise the errors of our analysis.
The results show that strengthening capital regulation is negatively but not significantly associated with banking efficiency. This finding does not imply that capital regulation is irrelevant for banking efficiency. It merely suggests that the link between capital regulation and banking efficiency is not sufficiently robust when controlling for other regulatory policies, bank and country characteristics. Regarding activity restrictions, we find that tighter activity restrictions are negatively associated with banking efficiency. By promoting the exchange of information, banks that engage in broader activities such as securities, insurance and real estate investment may benefit from economies of scale and scope, hence enhancing the efficiency levels. Similarly, we find that policies that promote official supervision may have adverse effects on banking efficiency. The results indicate that the official supervisory approach is ineffective in reducing overhead costs. However, the estimated coefficient is not significant enough to explain the behaviour of the net interest margin variable. The empirical evidence supports the private monitoring approach. Countries that require banks to disclose accurate and detailed information to the public tend to have a more efficient banking system. Basel II Capital Accord emphasises the importance of capital regulation, official supervisory power and private monitoring. Nevertheless, our findings incline to the negative aspects of capital regulation and official supervisory power.
Moreover, we find that bank-specific and country-specific control variables help explain banking efficiency. Larger banks with lower market share tend to have higher efficiency levels as confirmed by the significant negative coefficient of bank size and the significant positive coefficient of bank market share. Bank equity appears to be an important driver of banking efficiency only in the second model when overhead costs are used as a proxy. Similarly, government-owned banks variable only influences banking efficiency when considering overhead costs. This finding supports that countries that promote government-owned banks are more effective in reducing overhead costs, thus improving banking efficiency. Additionally, the growth rate of GDP per capita is negatively associated with net interest margin and overhead costs. In this perspective, the banking sector in expanding countries is significantly more efficient.
Our evidence overall suggests that the link between capital regulation and banking efficiency as measured by net interest margin and overhead costs is not robust enough to control for other regulatory variables. The results confirm that activity restrictions have a negative and significant impact on banking efficiency. Also, policies that encourage official supervisory power do not enhance the efficiency of the banking sector. The only approach that is positively and statistically significantly associated with efficiency is private monitoring.
This study may be subject to some limitations, therefore our findings should be interpreted carefully. First, we used cross-section data instead of panel data since information on regulatory measures is collected from the last survey conducted by the World Bank. Thus, the data is only available for one point in time (i.e. Bank, 2010). It is argued that banking efficiency is better estimated with panel data (Coelli et al., 2005). Another limitation of this study is the measures of banking efficiency. The literature on banking efficiency highlights the advantages of efficient frontier approaches compared to traditional accounting ratios (Berger & Humphrey, 1997;Fethi & Pasiouras, 2010). Besides, we consider 21 EU countries. Due to data unavailability, we do not consider all EU countries. As explained in the third section, countries economically large as Germany and the United Kingdom are not included, which could influence our results.
Recommendations
Given the negative relationship found between banking efficiency and official supervisory power, the recommendation for policy makers is to re-direct resources away from official supervision and towards private monitoring, a concept that has been promulgated in the Basel II framework. Moreover, supervision should not focus on restricting activities, as this is seen to have a significant and negative impact on banking efficiency. Supervisors should also recognize that, while capital regulation can be important for individual banks, it does very little for banking efficiency, but government-owned banks can contribute to system efficiency.
There are other regulatory policies than those considered in this study, such as financial conglomerate variables, competition, regulatory variables and deposit insurance scheme, which could affect banking efficiency. Larger sample size and additional regulatory variables could be | 2020-10-29T09:06:00.811Z | 2020-01-01T00:00:00.000 | {
"year": 2020,
"sha1": "f39494e60fcd3f6bb78aa970b28afe969cd6707e",
"oa_license": "CCBY",
"oa_url": "https://www.tandfonline.com/doi/pdf/10.1080/23322039.2020.1838735?needAccess=true",
"oa_status": "GOLD",
"pdf_src": "TaylorAndFrancis",
"pdf_hash": "4cfbc5ba2ec8051926d1bbf9f0a638108d66315b",
"s2fieldsofstudy": [
"Economics",
"Business"
],
"extfieldsofstudy": [
"Business"
]
} |
212424671 | pes2o/s2orc | v3-fos-license | Free Movement of Patients in Need for Cardiac Surgery
MIHAELA DACIANA NATEA1, OANA ELENA STOICA2*, FLORIN BUICU3 1University of Medicine, Pharmacy, Science and Technologies of Targu-Mures, Faculty of Sciences and Letters, History and Political Sciences Department, 38 Gh. Marinescu Str., 540139, Targu-Mures, Romania 2University of Medicine, Pharmacy, Science and Technologies of Targu-Mures, Faculty of Dental Medicine, 38 Gh. Marinescu Str., 540139, Targu Mures, Romania 3University of Medicine, Pharmacy, Science and Technologies of Targu-Mures, Faculty of Medicine, 38 Gh. Marinescu Str., 540139, Targu Mures, Romania
The Treaty on the functioning of the European Union (TFEU) establishes in article 6 that protection and improvement of human health is under the competence of member states, except some common public health concerns [1]. This gives states the full authority to act in organizing the health system including the financing and insurance scheme for health service. But accessing health services fall under the provision of free movement of goods and services, meaning that patients should be able to move feely within the European Union(EU).
The foundation for cross-border healthcare is being laid by the EU Directive 2011/24/EU on the application of patients' rights in cross-border healthcare [2] which drafts the legal framework for receiving quality treatment in another EU country, with or without prior authorization and the reimbursement process. In an assessment report on the implementation of the Directive a conclusion was drown over the increase improvement of the legal framework across EU states in managing cross-border health care [3]. However, the Directive does not cover outof-pocket medical care, but sets up the bases for patients' mobility [4].
The European legal norms that lay the foundation for E112 form are: Regulation (EEC) No 1408/71 of the Council of 14 June 1971 on the application of social security schemes to employed persons and their families moving within the Community [5] and Regulation (EEC) No 574/ 72 of the Council fixing the procedure for implementing Regulation (EEC) No 1408/71 on the application of social security schemes to employed persons and their families moving within the Community [6].
According to the National House of Health Insurance (NHHI) there are two conditions in which the Form E112 is released. The first one is in case of transfer of the residence. In this case E 112 may be refused by the health insurance house only if it is established that the displacement of that person would be prejudicial to his / her state of health or medical treatment; The second on is for the condition of referral to medical treatment in casethe treatment in question is among the medical services provided in the basic services package that benefit the insured from the Romanian social health insurance system or this treatment cannot be given within the time period normally required for obtaining the treatment in question in Romania, taking into account the current health status of the patient and the probable evolution of the disease. [7] The E112/S2 form allows access to planned medical treatment in other states, in the same conditions as for residents.
Experimental part
For the experimental part we used official statistics of the European Union regarding access to medical care in Europe andofficial data provided by the NHHI regarding total requests under E112 forms. NHHI also provided, at the request of the University of Medicine, Pharmacy, Sciences and Technology of Târgu Mure', extended data regarding E112/S2 approved requests for the last three years.
Results and discussions
The first aspect that must be taken into consideration is the reason for the need of cross-border health care.According to EUROSTAT there are reported unmet needs regarding medical care due to the cost, distance or waiting list as presented inEUROSTAT data which indicate a decrease both at European and national level with an average of 4% in the EU 28 in 2016 and around 9% in Romania [8] [11]. From the European countries,Estonia, Greece and Poland reported the highest percentage of unmet needs.Romania has a considerable decrease in the reported percentage of total unmet needs in medical but also dental care but is still high comparted to the European average. In the case of dental care this aspect could be linked to the considerable number of dental treatments done in the emergency service as shown in a study conducted over the cases presented in the Dental Office of Mures County Emergency Hospital in Targu-Mures during 2012-2017 [9]. In the case of medical care, the number can indicate also some type of treatments that cannot be provided in the country.
According to NHHI in 2017, 10.055 E forms were received and analyzed. From that number only some received prior authorization [10].
In the case of E112/S2 form, from the data provided by NHHI some interesting facts emerged. In the timeframe from 2016-2017 an average of 650 cases were authorized. We structured the data on Major Diagnostic Category such as presented in Order 587/2013 in order to follow which are the most frequent request for treatmentabroad. As many are complex cases we selected one diagnostic category, the most urgent one, or the one for which the form was released, even if it implied a complex diagnostic. The data resulted are presented in Table 1 and Chart 1.
From the data presented, and comparing 2016 to 2017, we can conclude that the incidence of cases of the circulatory system that require treatment abroad is the highest, followed by diseases and disorders of the musculoskeletal system and connective tissue and diseases and disorders of the nervous system.Most cases involve surgical procedures, 551 cases in 2017, that cannot be done in Romania.
From the 264 cases concerning diseases and disorders of the circulatory system 98 have as main diagnostic or part of the diagnostic septal defect, and all of then required surgical procedures abroad. The age range is presented in the Table 2 which indicate that most of the cases are urgent ones, from newborns, in fact most of neonatal cases with E112/S2 approved treatment concern cases with sever diseases or disorders of the circulatory system, and include a form of septal defect. The fact that so many cases are referred abroad for treatment indicate clearly a lack of national coverage for this type of diseases, due either to lack of infrastructure or medical professionals with competences in this field.
Conclusions
The need to medical healthcare abroad is obvious in the case of septal defect and the European systems offers a chance to life for the patients that can, in this manner, search for medical treatment abroad. The need to invest more in this area at national level is evident, and the investment shouldpursue better equipped hospitals and well-trained professionals. From a legislative viewpoint Romania implemented European laws, but some problems were identified in the payment process, sometimes the state was unable to pay the bills for medical treatment in time. | 2020-01-23T09:07:46.214Z | 2019-12-14T00:00:00.000 | {
"year": 2019,
"sha1": "f64e036fcd4abe49a844cd1074fb3e0cf3970e24",
"oa_license": "CCBY",
"oa_url": "https://revistadechimie.ro/pdf/70%20NATEA%2011%2019.pdf",
"oa_status": "HYBRID",
"pdf_src": "Anansi",
"pdf_hash": "25122f02b8be0abd71ec6d9825cac856bc7270b6",
"s2fieldsofstudy": [
"Medicine",
"Political Science"
],
"extfieldsofstudy": []
} |
267332602 | pes2o/s2orc | v3-fos-license | Unveiling the mechanism of broad‐spectrum blast resistance in rice: The collaborative role of transcription factor OsGRAS30 and histone deacetylase OsHDAC1
Summary Rice blast, caused by Magnaporthe oryzae, significantly impacts grain yield, necessitating the identification of broad‐spectrum resistance genes and their functional mechanisms for disease‐resistant crop breeding. Here, we report that rice with knockdown OsHDAC1 gene expression displays enhanced broad‐spectrum blast resistance without effects on plant height and tiller numbers compared to wild‐type rice, while rice overexpressing OsHDAC1 is more susceptible to M. oryzae. We identify a novel blast resistance transcription factor, OsGRAS30, which genetically acts upstream of OsHDAC1 and interacts with OsHDAC1 to suppress its enzymatic activity. This inhibition increases the histone H3K27ac level, thereby boosting broad‐spectrum blast resistance. Integrating genome‐wide mapping of OsHDAC1 and H3K27ac targets with RNA sequencing analysis unveils how OsHDAC1 mediates the expression of OsSSI2, OsF3H, OsRLR1 and OsRGA5 to regulate blast resistance. Our findings reveal that the OsGRAS30–OsHDAC1 module is critical to rice blast control. Therefore, targeting either OsHDAC1 or OsGRAS30 offers a promising approach for enhancing crop blast resistance.
Introduction
Rice (Oryza sativa L.) is one of the world's most important cereal crops, feeding half of the global population (Zhang et al., 2020).Rice blast is caused by the pathogenic fungus Magnaporthe oryzae and is one of the most serious diseases affecting rice, often causing serious decreases in rice production (Chakraborty et al., 2021).Plants have developed two layers of defence against pathogens, including pathogen-associated molecular pattern-triggered immunity (PTI) and effector-triggered immunity (ETI) (Jones and Dangl, 2006).PTI occurs at the plasma membrane surface and is controlled by pattern recognition receptors (Couto and Zipfel, 2016).Pathogens secrete effector molecules encoded by virulence genes into plant cells to induce pathogen virulence and attenuate plant PTI (Xin et al., 2018).Such pathogen effectors are detected by intracellular plant receptors known as nucleotide-binding leucine-rich repeat resistance (R) proteins, activating ETI (Thordal-Christensen, 2020).To date, over 100 blast R genes in the rice genome and 24 virulence genes in M. oryzae have been identified (Li et al., 2019).Blast R genes interact directly or indirectly with pathogenic effectors following M. oryzae infection, to trigger changes in the expression of defence response-related genes, triggering diverse physiological events including hormone signalling activation, hypersensitivity reactions, biosynthesis of chemical molecules that degrade pathogen cells and production of reactive oxygen species (Liu et al., 2021).However, such resistance is race-specific and is easily lost due to the rapid evolution of pathogen virulence genes (Deng and Naqvi, 2019).In contrast, non-race-specific resistance is broad-spectrum and may fight crop diseases more effectively.To date, a few genes conferring broad-spectrum resistance to pathogens have been identified in plants, including RPW8.1 and RPW8.2 for broad-spectrum resistance to powdery mildew in Arabidopsis (Xiao et al., 2001;Yang et al., 2023;Zhao et al., 2021Zhao et al., , 2023)), STV11 for durable resistance to rice stripe virus (Wang et al., 2014), bsr-d1, OsUBC45 and RBL1 for broad-spectrum blast resistance in rice (Li et al., 2017;Sha et al., 2023;Wang et al., 2023), Lr34 and TapsIPK1 for broad-spectrum resistance to rust fungi in wheat (Fukuoka et al., 2009;Wang et al., 2022) and ZmGLK36 for broad-spectrum resistance to maize rough dwarf disease (Xu et al., 2023).In plants, a growth-defence tradeoff exists due to resource constraints, which requires plants to prioritize either growth or defence according to changes in external and internal factors (Huot et al., 2014).This tradeoff means that most defence genes also affect agricultural traits (He et al., 2022).Thus, broad-spectrum genes with no or little effect on crop growth should be identified and integrated into breeding strategies to optimize the growth-defence balance, maximize crop yields and meet rising global food demands.
Recent reports have shown that epigenetic modificationmediated transcription regulation is involved in many biological processes in plants (Chang et al., 2020;Chen et al., 2020;Liang et al., 2020;Yang et al., 2022).Histone acetylation is one of the most extensively characterized epigenetic modifications and is generally associated with active transcription (Shvedunova and Akhtar, 2022).Histone deacetylases (HDACs) regulate histone acetylation levels and gene transcription by removing acetyl moieties from the acetylated lysines of histones, which are added by histone acetyltransferases (Kumar et al., 2021), crucial players in plant growth and development as well as responses to environmental stimuli (Chen et al., 2020).In addition, HDACs are involved in disease resistance in some plant species.For example, Arabidopsis HDA6 suppresses the expression of defence-related genes (Wang et al., 2017;Wu et al., 2021).HDA19 is required for salicylic acid (SA)-mediated defence responses (Choi et al., 2012), and HDA19 can interact with WRKY38 or WRKY62 to regulate plant basal defence responses (Kim et al., 2008).HD2B is involved in the reprogramming of defence gene expression and innate immunity (Latrasse et al., 2017).In rice, HDA701 negatively regulates broadspectrum resistance to rice pathogens by modulating histone H3K9 acetylation of defence-related genes (Chen et al., 2022a).SRT702 negatively regulates resistance to rice blast, false smut and bacterial blight by modulating histone H4K5 and H4K8 acetylation to affect the expression of disease-resistance genes (Chen et al., 2024).HDT701 negatively regulates plant innate immunity by modulating histone H4 acetylation of defencerelated genes (Ding et al., 2012), and the RNase P protein subunit Rpp30 can target HDT701, conferring resistance to fungal and bacterial pathogens (Li et al., 2021).In wheat, HOS15 can interact with HDA6 to regulate defence responses to powdery mildew (Zhi et al., 2020).However, the molecular mechanisms of HDAC activity in plant immunity remain largely unknown.
Seventeen HDAC genes have been identified in the rice genome (Hou et al., 2021).Here, we demonstrate OsHDAC1 as a negative regulator of immunity against rice blast with no effects on plant height and tiller numbers and identify a novel GRAS transcription factor OsGRAS30, which can interact with OsH-DAC1 and suppress its activity, thereby enhancing rice broadspectrum blast resistance.Further, we have found that OsHDAC1 suppresses the transcription of blast resistance-related genes by binding their promoters and reducing histone H3K27 acetylation.Moreover, we show that chemical HDAC inhibitors confer rice blast resistance, providing a novel strategy for controlling rice blast.Overall, our data reveal the functional characteristics of the OsGRAS30-OsHDAC1 module in mediating broadspectrum blast resistance in rice.
Results
OsHDAC1 negatively regulates broad-spectrum resistance to M. Oryzae in rice Recently, chromatin histone modifications have been reported to play a role in the regulation of plant disease defences (Chen et al., 2022a;Li et al., 2021;Niu et al., 2019).In our previous work, to examine the function of OsHDAC1, we constructed OsHDAC1 RNA interference (RNAi) lines in rice, as homozygous OsHDAC1 mutant seeds could not be obtained, and also generated overexpression (OE) transgenic plants (Chung et al., 2009;Hou et al., 2022).In this study, to explore whether OsHDAC1 is involved in rice blast resistance, we first challenged rice plants with the virulent M. oryzae Guy11.To detect changes in OsHDAC1 protein levels in these rice plants, we produced an OsHDAC1-specific antibody (Figure S1).Analysis of changes in OsHDAC1 expression after M. oryzae infection showed that the transcript (Figure 1a) and protein (Figure 1b) levels of OsHDAC1 were downregulated.To investigate the possible involvement of OsHDAC1 in blast resistance, two OsHDAC1 RNAi lines and two OsHDAC1 OE lines, along with wild-type (WT) Nipponbare rice, were inoculated with M. oryzae Guy11 by spraying with conidia.The two OsHDAC1 RNAi lines (Ri2 and Ri3) developed only small, scattered lesions, while WT Nipponbare showed typical blast lesions (Figure 1c), indicating that silencing of OsHDAC1 enhances resistance to blast.Fungal DNA quantification showed that the two OsHDAC1 RNAi lines harboured only 10% and 30% as much M. oryzae as WT, respectively (Figure 1d).In contrast, the OsHDAC1 OE5 and OE6 lines had larger lesions compared with WT (Figure 1e).The two OE lines had significantly greater fungal DNA content, which was increased by approximately 3-to 4-fold (Figure 1f).We then inoculated the OsHDAC1 RNAi and OE lines and WT by spraying conidia from mixed M. oryzae isolates, including AH4, Sc09-153-07, TM3-2, RB1, RB3, NC1, HLJ1-3, HLJ09-17-1 and HLJ5-3 from different regions of China (Figure S2).The OsHDAC1 RNAi lines showed enhanced resistance to multiple virulent isolates of M. oryzae, whereas the OE lines showed higher susceptibility, compared with WT (Figure 1g).Relative fungal biomass was reduced by approximately 3-fold in the RNAi lines and increased by 1.7-and 1.5-fold in the OE lines (Figure 1h).In addition, field tests demonstrated that the OsHDAC1 OE and RNAi lines had no effects on plant height, tiller numbers, 1000 seed weight and seed setting rate (Figure S3).These results indicate that OsHDAC1 negatively regulates rice broad-spectrum blast resistance.
OsGRAS30 interacts with OsHDAC1
To gain mechanistic insights into OsHDAC1 function, we attempted to identify proteins interacting with OsHDAC1 in a yeast screening experiment.We identified the gene LOC_Os05g49930, named OsGRAS30, which belongs to the GRAS transcription factor family and interacts with OsHDAC1 (Figure 2a).Previous studies indicated that overexpression of the GRAS transcription factor family proteins GIGR2 and OsSLR1 promotes resistance to M. oryzae (De-Vleesschauwer et al., 2016;Tanabe et al., 2016).Therefore, we speculated that OsGRAS30 might be involved in OsHDAC1-mediated M. oryzae resistance.To confirm the OsHDAC1-OsGRAS30 interaction in vivo, OsHDAC1-HA and OsGRAS30-Myc were co-expressed in rice protoplasts and subjected to coimmunoprecipitation (Co-IP) analysis.OsGRAS30-Myc could be co-purified with OsHDAC1-HA using an anti-HA antibody, but no OsGRAS30-Myc was obtained from the control (Figure 2b).We conclude that OsHDAC1 interacts with OsGRAS30 in vivo.Further, in a splitluciferase (Split-LUC) complementation imaging assay, we transiently co-infiltrated Nicotiana benthamiana leaves with OsHDAC1-nLUC and cLUC-OsGRAS30 fusion constructs, and LUC fluorescence was apparent in N. benthamiana leaves, demonstrating OsHDAC1-OsGRAS30 interaction in planta (Figure 2c).To confirm the OsHDAC1-OsGRAS30 interaction in vitro, we expressed and purified recombinant GST-OsHDAC1 and MBP-OsGRAS30-His protein from Escherichia coli and performed an in vitro pull-down assay.MBP-OsGRAS30-His was pulled down by GST-OsHDAC1 but not by GST itself (Figure 2d), suggesting that MBP-OsGRAS30-His interacts with GST-OsHDAC1 in vitro.In addition, laser confocal microscopy showed that OsHDAC1-sGFP and OsGRAS30-mCherry were co-localized in the nucleus of rice protoplasts (Figure 2e).
We attempted to delineate the interaction interface by testing three OsHDAC1 fragments for interaction with three OsGRAS30 fragments in a yeast two-hybrid (Y2H) assay.First, we divided OsGRAS30 into three parts: the N-terminal region (amino acids 1-85), GRAS domain (amino acids 86-448) and C-terminal region (amino acids 449-501).The results showed that the GRAS domain region of OsGRAS30 was important for its interaction with OsHDAC1 (Figure 2f).Then, we divided OsHDAC1 into three parts: the N-terminal region (amino acids 1-22), histone deacetylase domain (amino acids 22-399) and C-terminal region (amino acids 400-518).As shown in Figure 2g, the histone deacetylase domain of OsHDAC1 interacted with OsGRAS30.We further found that the histone deacetylase domain of OsHDAC1 interacted with the GRAS domain region of OsGRAS30 (Figure 2h).These results indicate that the histone deacetylase domain of OsHDAC1 and the GRAS domain region of OsGRAS30 are essential for their interaction.
OsGRAS30 functions as a positive regulator of rice blast resistance
To investigate the role of OsGRAS30 in the response to M. oryzae, we first assessed the change in its expression levels after blast infection.The RNA level began to decrease at 3 h post-infection (hpi) with blast and was further reduced at 12 hpi, reaching 40% of the level in plants with no blast infection (Figure 3a).We then generated loss-of-function and gain-of-function transgenic plants.We used clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9 technology to create OsGRAS30-knockout mutants, with one site in the exon selected for guide RNA (gRNA) design, as shown in Figure 3b.DNA sequencing revealed that among the mutant lines isolated in this study, osgras30 #1 has a single nucleotide deletion, and osgras30 #2 has a two-nucleotide deletion at the gRNA site (Figure 3b).Each mutation causes a frameshift in the OsGRAS30coding sequence.We generated rice transgenic rice plants overexpressing OsGRAS30-MYC using the pCXUN vector containing OsGRAS30 cDNA fused to the MYC tag driven by the Ubiquitin promoter in the WT background (Figure S4a).Of 10 OE lines, 4 exhibited increased OsGRAS30 RNA levels to varying degrees (Figure S4b).We selected lines OE4 and OE8, which displayed the highest OsGRAS30 RNA levels, for further study.
Next, two OsGRAS30 mutant lines (osgras30 #1 and #2), the OsGRAS30 OE lines, and WT were inoculated with blast by spraying Guy11 conidia.The results showed that the osgras30 #1 and osgras30 #2 lines exhibited higher susceptibility to blast treatment compared with WT plants (Figure 3c), and the fungal DNA contents in the OsGRAS30 mutant lines were increased by 2.7-and 3.3-fold compared with WT (Figure 3d).In contrast, we observed enhanced resistance to blast after inoculating the OsGRAS30 OE4 and OE8 lines and WT (Figure 3e).These two OE lines contained lower fungal DNA content than that of WT (Figure 3f).We further inoculated the OsGRAS30 mutant and OE lines along with WT by spraying conidia from mixed M. oryzae isolates including AH4, Sc09-153-07, TM3-2, RB1, RB3, NC1, HLJ1-3, HLJ09-17-1 and HLJ5-3 (Figure S2).The OsGRAS30 OE lines showed enhanced resistance to multiple virulent isolates of M. oryzae, whereas the mutant lines showed higher susceptibility, compared with WT (Figure 3g).Relative fungal biomass increased by 2-and 2.5-fold in the mutant lines and decreased by 3-fold in the OE lines (Figure 3h).Taken together, these data from the mutant and OE lines suggest that the OsGRAS30 gene positively regulates broad-spectrum resistance to rice blast.
OsHDAC1 acts genetically downstream of OsGRAS30
To analyse the genetic interactions between OsGRAS30 and OsHDAC1, we used Agrobacterium tumefaciens strain EHA105 containing OsHDAC1-RNAi constructs (Hou et al., 2022) to transform the background of osgras30 #1 plants and thereby generated homozygous double OsHDAC1 RNAi/osgras30 lines.Through this process, we obtained 10 putative OsHDAC1 RNAi lines of osgras30 #1 rice with reduced OsHDAC1 expression; OsHDAC1 RNA levels were reduced by approximately 90% in the OsHDAC1 RNAi #5 and #9 lines compared with the osgras30 #1 line (Figure 4a).Then, two OsHDAC1 RNAi lines with the osgras30 #1 background, the osgras30 mutants and WT were inoculated with conidia from mixed M. oryzae isolates.The results indicated that the OsHDAC1 RNAi #5 and #9 lines in the osgras30 #1 background showed enhanced resistance to multiple virulent isolates of M. oryzae compared with OsGRAS30 mutants, whereas OsHDAC1 RNAi #5 and #9 lines in the osgras30 #1 background showed similar blast resistance phenotype observed in the WT (Figure 4b).Measurement of M. oryzae DNA content confirmed the phenotype results (Figure 4c).This demonstrates that OsGRAS30 acts, at least in part, genetically upstream of OsHDAC1 to positively regulate blast resistance.
OsGRAS30 represses OsHDAC1 enzyme activity to induce H3K27ac hypoacetylation
To further investigate how OsGRAS30 functions in OsHDAC1mediated immunity in response to M. oryzae.We first detected the abundance of OsHDAC1 proteins in WT, OsGRAS30 mutant and OsGRAS30 OE lines by immunoblotting using an antibody against OsHDAC1.The results indicated that the OsHDAC1 protein level remained unchanged in the OsGRAS30 mutant and OsGRAS30 OE lines (Figure S5a), and the transcript level of OsHDAC1 was also unchanged in OsGRAS30 mutant (Figure S5b, c), suggesting that OsGRAS30 did not affect OsHDAC1 abundance or stability.Thus, we speculated that OsGRAS30 interferes with the deacetylation activity of OsHDAC1.To confirm this hypothesis, we analysed HDAC activity in WT, OsGRAS30 mutants, and OE plants, and the results showed that HDAC activity was significantly elevated in osgras30 #1 and #2 lines and markedly reduced in OsGRAS30 OE4 and OE8 lines compared with WT (Figure 5a).In vitro activity analysis revealed that OsGRAS30 inhibited the deacetylation activity of OsHDAC1 by binding to OsHDAC1 (Figure 5b).
HDACs function to remove the acetyl group from acetylated lysines in histones (Kumar et al., 2021).Thus, we assessed overall histone acetylation in WT, OsHDAC1 RNAi and OsHDAC1 OE lines by immunoblotting, and the results showed that the overall histone 3 acetylation (H3ac) level was dramatically elevated in OsHDAC1 RNAi lines and dramatically reduced in OsHDAC1 OE lines compared with WT (Figure 5c).Previous reports have indicated that H3K27me3 and H3K27ac regulate functional genes in M. oryzae (Lin et al., 2022;Zhang et al., 2021), and H3K9ac plays an essential role in plant defence responses in rice (Chen et al., 2022a).H3K9ac and H3K27ac are the most characterized epigenetic marks invariably associated with transcriptional activation and have been shown to affect many developmental and biological processes in higher plants ( Dasgupta et al., 2022;Zhou et al., 2010).This prompts us to further assess the level of histone H3 acetylation at lysines 9 and 27 by immunoblotting using specific antibodies.The results indicated that the H3K27ac levels were higher in the OsHDAC1 RNAi lines and lower in the OsHDAC1 OE lines compared with WT rice plants (Figure 5c), suggesting that H3K27ac is a critical regulatory site of OsHDAC1.We then challenged rice plants with the virulent M. oryzae Guy11, and observed that H3K27ac levels were upregulated after M. oryzae infection (Figure S6).These results confirmed that H3K27ac was involved in blast defence responses in rice.Further, we detected total acetylation levels of histone 3 and 4 in OsGRAS30 mutant lines, OsGRAS30 OE lines, OsHDAC1 RNAi /osgras30 #1 lines, and WT plants to assess whether OsGRAS30 impacts OsHDAC1-mediated H3 acetylation level.The results showed that overall H3ac levels were decreased in OsGRAS30 mutant lines, but OsHDAC1 RNAi lines in the osgras30 #1 background restored this decrease, while H3ac levels were elevated in the OsGRAS30 OE lines (Figure 5d).We specifically examined H3K9ac and H3K27ac levels in these lines and found that H3K27ac shows a similar trend to H3ac (Figure 5d).These results demonstrate that OsGRAS30 regulates H3K27ac levels by suppressing the deacetylated activity of OsHDAC1.In addition, we applied a series of different concentrations of HDAC inhibitors including trichostatin A (TSA) and sodium butyrate (NaB) to WT plants inoculated with spore suspensions for 5 days and found that the resistance of rice plants to blast infection increased with the HDAC inhibitor concentration (Figure S7).Furthermore, treatment of WT and osgras30 #1 rice plants with HDAC inhibitors greatly improved their symptoms, and osgras30 #1 treated with HDAC inhibitors showed greater blast resistance compared with untreated WT plants (Figure 5e,f).Taken together, these observations demonstrate that OsGRAS30 enhances blast resistance by suppressing OsHDAC1 in rice.
OsHDAC1 and H3K27ac co-occupy targets in rice genome
To elucidate the molecular determinants and targets of OsH-DAC1, we performed chromatin immunoprecipitation sequencing (ChIP-Seq) to detect OsHDAC1 binding sites in the genome of the OsHDAC1 OE5 line using an antibody against the HA tag (Figure S8) because the HA antibody performed better than did our customized OsHDAC1 antibody in ChIP and two biological replicates showed a strong correlation (Figure S9).A total of 21, 471 OsHDAC1-binding sites, corresponding to 15, 420 genes, were identified (Dataset S1).Analysis of the OsHDAC1 distribution in the rice genome showed that among the OsHDAC1 targets, 60.19% of binding peaks occurred in gene regions and OsHDAC1 enrichment was mostly detected near transcription start site (TSS), indicating that OsHDAC1 were highly enriched in the promoter region surrounding the TSS to regulate transcriptional initiation (Figure 6a,b).Examination of the correlation between OsHDAC1 binding and gene expression levels revealed that OsHDAC1 was more enriched in genes with hightranscription levels than in genes with low-transcription levels, and which was located mainly around the TSS of these genes (Figure 6c).We further identified potential DNA-binding motifs in OsHDAC1-binding sites using DREME, and a significantly enriched consensus G-box motif was identified (Figure 6d), which is similar to the report that HAD9 in Arabidopsis could bind to the G-box (Chen et al., 2016).Therefore, OsHDAC1 is associated mainly OsGRAS30-OsHDAC1 regulates blast resistance 1747 with active genes and is preferentially enriched in their promoter regions.
Next, we investigated genome-wide targets of H3K27ac by ChIP-Seq using an antibody against H3K27ac in WT plants with two biological replicates, whose results showed a strong correlation (Figure S10).A total of 22, 092 H3K27ac peaks, corresponding to 19, 256 genes, were identified (Dataset S2).Analysis of the H3K27ac distribution in the rice genome showed that H3K27ac peaks were located mainly in gene regions (83.49%), especially in promoter regions surrounding the TSS, and H3K27ac enrichment was mostly detected around the TSS (Figure 6e,f).In addition, we observed that H3K27ac was enriched in genes with high-transcription levels and occurred mainly around the TSS of these genes, whereas silent genes showed a low level of H3K27ac enrichment (Figure 6g).Interestingly, analysis of the relationship between OsHDAC1 binding and acetylation levels showed that OsHDAC1 tends to be positively correlated with the H3H27ac level at the target sites (Figure 6h), and 3, 856 genes were co-bound by OsHDAC1 and H3K27ac (Figure 6i).In humans, HDAC levels have also been shown to positively correlate with histone acetylation levels, and it is suggested that HDACs function to remove the acetyl group added by HATs in active genes and to reset chromatin modification after gene activation (Wang et al., 2002(Wang et al., , 2009)).Finally, a randomly selected region on chromosome 2 confirmed that OsHDAC1-occupied peaks indeed colocalized with H3K27ac (Figure 6j), which further supports the correlation between OsHDAC1 and H3K27ac.Overall, we discovered that H3K27ac is associated with expressed genes in the rice genome and that OsHDAC1 functions to remove the acetyl moieties from H3K27ac to control the H3K27ac level in co-occupied target genes.
OsHDAC1 modulate blast resistance by directly suppressing blast resistance-related genes
To further investigate the effect of OsHDAC1 on gene expression associated with rice resistance to blast.We first profiled mRNA from 2-week-old rice leaves of WT and OsHDAC1 Ri2 plants via RNA-Seq.There are three biological replicates, whose results showed a strong correlation (Figure S11).In total, 2, 398 upregulated and 2, 953 downregulated genes (expression level fold change >2 and FDR <0.05) were identified in the OsHDAC1 RNAi lines compared with corresponding WT plants (Figures 7a and S12; Dataset S3).Next, we investigated the association between OsHDAC1 and H3K27ac binding and altered gene expression in OsHDAC1 Ri2 plants.The results showed that the expression of 220 genes bound by OsHDAC1 and enriched by H3K27ac was significantly increased in the OsHDAC1 Ri2 lines, while the expression of 223 genes bound by OsHDAC1 and enriched by H3K27ac was significantly decreased in the OsHDAC1 Ri2 lines (Figure 7b).
The differentially expressed genes (DEGs) were functionally characterized based on Gene Ontology (GO) analysis.The main GO terms assigned to downregulated DEGs affected by OsHDAC1 were the biosynthetic process and metabolic process (Figure 7c; Dataset S4).The KEGG analysis revealed that these downregulated DEGs enriched in fatty acid degradation and biosynthesis (Figure S13a; Dataset S5).Prior study revealed that fatty acid was involved in broad-spectrum blast resistance in rice (He et al., 2020), suggesting that OsHDAC1 negatively regulates blast resistance by modulating fatty acid contents.Suppression of the fatty acid desaturase gene OsSSI2 has been reported to enhance resistance to blast by regulating fatty acid content (Jiang et al., 2009), and we verified that OsHDAC1 could increase the expression of OsSSI2 by binding its promoter region and altering the H3K27ac levels (Figure 7e-h).Similarly, OsF3H has been reported to involve in blast resistance by mediating the biosynthesis, signalling and transport of SA (Chen et al., 2022b;Wu et al., 2022).Our result showed that OsF3H is the direct target of OsHDAC1 by RT-qPCR and ChIP-PCR experiments (Figure 7e-h).Conversely, we found that upregulated DEGs involved in biotic stimulus were enriched (Figures 7d and S13b, Dataset S6 and S7), and genes related to blast resistance (OsRLR1 and OsRGA5) were upregulated in OsHDAC1 Ri2 plants and downregulated in OsHDAC1 OE5 plants (Figure 7e).Among, OsRLR1 confers broad-spectrum resistance in rice by increasing hydrogen peroxide (H 2 O 2 ) content (Du et al., 2021); OsRGA5 recognizes the M. oryzae effector AVR1-CO39 to relieve OsRGA4 to promote blast resistance (Césari et al., 2013(Césari et al., , 2014;;Okuyama et al., 2011), supporting the role of these genes in blast resistance.Further, we found that OsHDAC1 was significantly enriched in OsRLR1 and OsRGA5 via ChIP-qPCR, and the H3K27ac levels in these target genes were significantly elevated in the OsHDAC1 Ri2 lines compared with WT plants (Figure 7f-h).Together, these results suggest that OsHDAC1 binds to blast resistance-related genes and alters their H3K27ac levels to regulate their expression, revealing a link between blast resistance and OsHDAC1.
Discussion
Rice feeds half of the world's population, and rice blast is a very serious disease that affects rice production (Chakraborty et al., 2021).Functional characterization of non-race specific resistance-related genes is particularly important for molecular breeding to improve broad-spectrum disease resistance and the development of strategies for effective control of this disease in rice.In this study, we report the functional identification of the OsGRAS30-OsHDAC1 module in rice blast resistance.OsGRAS30 acts genetically upstream of OsHDAC1 and interacts with OsHDAC1, repressing its deacetylation activity; inactivated et al., 2012;Li et al., 2021) in rice; and HDA6 in wheat (Zhi et al., 2020).Although HDA701 and HDT701 have been demonstrated to confer disease resistance in rice, the downstream factors related to their role in plant immunity remain unclear (Chen et al., 2022a;Li et al., 2021).Previous studies showed that homozygous OsHDAC1 mutant seeds could not be obtained and that partial inhibition of OsHDAC1 repressed root growth (Hou et al., 2022).In this work, the protein abundance of OsHDAC1 was downregulated at the early stage of blast infection (Figure 1b) and we further demonstrate that partial inhibition of OsHDAC1 contributes to plant defence against multiple isolates of M. oryzae without affecting plant height and tiller numbers (Figures 1 and S3).Similarly, mutations in a few genes are reported to constitutively promote the nucleotide-binding leucine-rich repeat immune response, as well as to cause plant growth retardation (Palma et al., 2010;Shirano et al., 2002).These findings suggest that OsHDAC1 is important for rice growth but knockdown of OsHDAC1 enhances broad-spectrum resistance to rice blast fungus M. oryzae, making it a prime candidate gene for manipulation in breeding programs.
OsGRAS30 suppresses OsHDAC1 activity, leading to enhanced blast resistance in rice
OsGRAS30 is a member of the GRAS family, and previous studies have revealed that GRAS transcription factors function in plant growth, development, and resistance to various biotic and abiotic stresses (Jaiswal et al., 2022).Our data suggest that OsGRAS30 positively regulates blast resistance in rice, which is supported by our finding that loss of function and overexpression of OsGRAS30 decreased and increased blast resistance, respectively (Figure 3).Previous reports showed that EAR domain-containing transcription factors trigger PRC2-mediated chromatin marking in Arabidopsis (Baile et al., 2021).HDA9 and WRKY53 transcription factors are mutual antagonists (Zheng et al., 2020), and MADSdomain transcription factor AGL15 involves in the recruitment of histone deacetylase complex components (Hill et al., 2008).Thus, the interaction of transcription factor and chromatin remodelling components is valuable for understanding the transcription factor as well as the chromatin remodelling mechanism.Our in vitro and in vivo assays both suggest that interaction between OsGRAS30 and OsHDAC1 (Figure 2), and overexpression of OsGRAS30 led to reduced OsHDAC1 deacetylation capacity and enhanced blast resistance, while mutation of OsGRAS30 increased blast sensitivity (Figures 3 and 5b,c), indicating that OsGRSA30 functions in blast resistance by inhibiting OsHDAC1 activity.This revealed that a GRAS transcription factor is involved in plant resistance responses by modulating the activity of an epigenetic modification enzyme.Additionally, the GRAS domain of OsGRAS30 binds to the histone deacetylase domain of OsHDAC1 (Figure 2f-h), and GRAS family DELLA proteins were found to interact with the DNA-binding domains of EIN3/EIL1, which blocks DNA binding domain of EIN3/EIL1 to inhibit its transcriptional regulation of downstream genes in Arabidopsis (An et al., 2012).This implies that OsGRAS30 could block histone deacetylase domain of OsHDAC1 to inhibit its deacetylation capacity.Meanwhile, we did not detect that MBP-OsGRAS30-His was acetylated in vitro, suggesting OsGRAS30 is not a substrate for OsHDAC1.Further, the blast resistance phenotype of the OsGRAS30 mutant lines was reversed by OsHDAC1 RNAi (Figure 4b,c), validating that OsGRAS30 can directly target OsHDAC1 to decrease OsHDAC1 enzyme activity.HDACs function to remove the acetyl group from acetylated lysines in histones (Kumar et al., 2021), and we revealed that H3K27ac is a favoured substrate of OsHDAC1 (Figure 5c), while OsGRAS30 induces H3K27ac hyperacetylation by regulating the activity of OsHDAC1 (Figure 5d).
It has revealed that acetylation plays an essential role in crop defences during plant-pathogen interactions.For example, infection of maize by the fungus Cochliobolus carbonum increases the acetylation level of histones H3 and H4, which affects the expression of defence-related genes (Walley et al., 2018).In rice, the HD2-type HDAC HDT701 regulates plant innate immunity by modulating H4K5 acetylation in defence-related genes (Ding et al., 2012), and another report revealed that OsRpp30 confers bacterial, fungal and viral resistance by targeting HDT701 to regulate the H4K5 acetylation level (Li et al., 2021).Similarly, the RPD3-type HDAC HDA701 affects disease resistance by regulating H3K9 acetylation (Chen et al., 2022a).Our data reveal that the RPD3-type HDAC OsHDAC1 is associated with the removal of acetyl groups from H3K27, but not H3K9 (Figure 5c), suggesting that OsHDAC1 may use a different mechanism than HDA701 to regulate blast resistance.
Genome-wide mapping of OsHDAC1 targets reveals its regulatory mechanism
HDACs are generally associated with transcriptional repression (Berger, 2007;Robert et al., 2004).However, recent genomewide occupancy studies in plants including Arabidopsis and maize (Chen et al., 2016;Yang et al., 2016), as well as humans (Wang et al., 2009), indicate that HDACs mainly target transcriptionally active genes.In accordance with this conclusion, our data provide evidence that the rice RPD3-type HDAC OsHDAC1 is preferentially enriched in promoter regions and positively related to gene expression (Figure 6a-c).In addition, OsHDAC1 can bind to the G-box (Figure 6d), which is consistent with the result obtained for HAD9 in Arabidopsis (Chen et al., 2016).These findings suggest that the interaction of HDACs with active genes may be conserved across plants and mammalian cells.Interestingly, our data show that the level of OsHDAC1 binding tends to be positively correlated with the enrichment of H3K27ac (Figure 6h, j).Similarly, the level of HDA101 binding also appears to be positively correlated with the amount of H4K5ac in maize (Yang (e) The transcript level of blast resistance-related genes in OsHDAC1 OE and RNAi lines compared with WT.Total RNA was extracted from two-week-old rice leaves and subjected to RT-qPCR.Expression levels of genes in the WT were set to 1.00.Values are mean AE SEM (n = 3).Experiments were repeated three times with similar results.(f) IGV views showing OsHDAC1 and H3K27ac enrichment in blast resistance-related genes.Input serves as the negative control.(g) ChIP-qPCR results of OsHDAC1 binding levels at blast resistance-related genes in OsHDAC1 OE plants.ChIP-qPCR results of IgG binding levels of genes were set to 1.00.Values are mean AE SEM (n = 3).Experiments were repeated three times with similar results.(h) ChIP-qPCR results of H3K27ac levels at blast resistance-related genes in OsHDAC1 Ri2 line compared with WT plants.ChIP-qPCR results of H3K27ac levels of genes in the WT were set to 1.00.Values are mean AE SEM (n = 3).Experiments were repeated three times with similar results.Asterisks mark significant changes compared with WT (e and h) or IgG (g) based on Student's t test: *P < 0.05, **P < 0.01, ***P < 0.001.(i) A proposed working model showing the role of OsGRAS30 in manipulating the OsHDAC1-mediated broad-spectrum blast resistance in rice.WT, wild-type; Ri, RNA interference; OE, overexpression., 2016).The probable explanation is that HDACs function to remove acetyl groups at active genes added by HATs and to reset chromatin modification after gene activation (Wang et al., 2009).25.01% of OsHDAC1-bound genes happen H3K27ac enrichment (Figure 6i), implying that OsHDAC1 bind to these target genes and controls the H3K27ac level.In addition, the more direct targets of OsHDAC1 show no H3K27ac enrichment (Figure 6i).A possible reason for this finding is that other target genes that bind to OsHDAC1 undergo acetylation at sites other than H3K27.Furthermore, our RNA-Seq data indicated that the knockdown of OsHDAC1 did not affect the transcript levels of most genes bound by OsHDAC1 (Figures 6i and 7b), in accordance with HDACs identified in mammalian cells, maize and Arabidopsis (Chen et al., 2016;Wang et al., 2009;Yang et al., 2016).
Plant disease resistance responses involve transcriptional reprogramming and expression changes in many genes (Tsuda and Somssich, 2015).Rice blast resistance is a complex process in which the expression of numerous genes must be fine-tuned (Tsuda and Somssich, 2015).Blast R gene-mediated resistance depends on interactions between R proteins and corresponding pathogenic Avr proteins (Lazar et al., 2022), such as OsRGA5 positively regulates rice blast resistance by affecting the recognition of pathogenic Avr-CO39 protein (Cesari et al., 2013).In addition, OsRGA4 and OsRGA5 are a pair of R genes for blast resistance, and both are required for blast resistance (Césari et al., 2013;Okuyama et al., 2011), which upon recognition of the pathogen effector AVR-Pia by direct binding to OsRGA5, OsRGA4 is relieved to induce cell death (Césari et al., 2014).Upregulation of OsRGA5 due to OsHDAC1 knockdown may possibly contribute to the recognition of the M. oryzae effectors.OsRLR1 improve broad-spectrum blast resistance by regulating H 2 O 2 accumulation (Du et al., 2021).OsRGA5, along with OsRLR1 were identified downstream of OsHDAC1.Our results indicate that OsHDAC1 bind to the promoter region of these genes and alters the H3K27ac levels, inhibiting their expression to suppress immune activation (Figure 7).Thus, an important finding of our study is that positive regulators of blast resistance-related genes are associated with OsHDAC1 binding and are inhibited by OsHDAC1 (Figure 7).Conversely, negative regulators of blast resistance-related genes associated with OsHDAC1 binding are upregulated by OsHDAC1 (Figure 7).These genes including OsF3H and OsSSI2 caused the decreased production of SA and fatty acid, and their inhibition could achieve broad-spectrum resistance to M. oryzae in rice (Chen et al., 2022b;Jiang et al., 2009).Overall, our findings revealed that OsHDAC1 directly affect these targets and thus alters their expression to regulate blast resistance in rice.
A chemical intervention strategy reduces HDAC activity to increase blast resistance in rice
Rice blast, caused by M. oryzae, is an important disease affecting rice production, and thus it threatens global food security (Chakraborty et al., 2021).The symptoms of rice blast can occur throughout the rice growth period (Chakraborty et al., 2021).Genetic breeding of durably resistant plants is considered an effective strategy for controlling plant diseases, although it is difficult and time consuming to accomplish without affecting plant growth due to the limited set of currently available broadspectrum resistance genes without detrimental effects (He et al., 2022).Some genes have been engineered to increase disease resistance directly.The present study showed that knockdown of OsHDAC1 via RNAi conferred rice blast disease resistance, and our preliminary field test results showed no difference in plant height and tiller numbers, between OsHDAC1knockdown and WT rice (Figures 1 and S3), revealing that the development of resistant rice can be achieved by decreasing the expression, rather than inducing loss of function, of a target gene.Another strategy for disease resistance is chemical intervention, i.e., application of small-molecule chemicals or drugs to enhance resistance via interference with the functions of target genes.HDAC inhibitor treatment enhanced rice blast resistance (Figure S7) provides a foundation for further application of such chemicals to disease intervention in vitro.
Plasmid construction and generation of transgenic plants
For the purification of OsHDAC1-His and MBP-OsGRAS30-His recombinant proteins, a fragment containing the OsHDAC1 coding sequence (CDS) was amplified and cloned into the EcoRI site of the pET32a vector, and the OsGRAS30 CDS fragment was amplified and inserted into the EcoRI site of the pMalC2x vector.To generate gene constructs fused to the DNA binding domain or activation fragment for Y2H assays, the OsHDAC1 CDS fragments, as well as the N-terminal region (nucleotides 1-63), histone deacetylase domain region (nucleotides 64-1197), and C-terminal region (nucleotides 1198-1557) of OsHDAC1 were amplified and cloned into the EcoRI site of the pGBKT7 vector; the OsGRAS30 CDS fragments, as well as the N-terminal region (nucleotides 1-255), GRAS domain region (nucleotides 256-1344) and C-terminal region (nucleotides 1345-1503) of OsGRAS30 were amplified and cloned into the EcoRI site of the pGADT7 vector.To generate the OsHDAC1-nLUC and cLUC-OsGRAS30 constructs, the OsHDAC1 CDS fragments were amplified and cloned into the KpnI and SacI sites of the pCambia1300-nLUC vector, while OsGRAS30 CDS fragments were amplified and cloned into the KpnI and SalI sites of the pCambia1300-cLUC vector.To generate the OsGRAS30-mCherry constructs, the amplified OsGRAS30 CDS fragments were fused to a mCherry tag to generate fusion genes and cloned into the BamH1 and SalI sites of the pBI121 vector.To generate OsGRAS30 OE plants, the amplified OsGRAS30 CDS fragment was fused to a MYC tag to generate fusion genes and cloned into the XcmI site of the pCXUN vector.To generate OsGRAS30 mutant plants, guide sequences were designed using CRISPR-P (http://cbi.hzau.edu.cn/crispr) and inserted into the sgRNA expression vector backbone pYLsgRNA-OsU6 using annealed oligonucleotides.Then, the sgRNAs were inserted into the pYLCRISPR/Cas9Pubi-H vector as described previously (Hou et al., 2022).The OsGRAS30-MYC and CRISPR/Cas9-OsGRAS30 constructs were transfected into A. tumefaciens strain EHA105 for transformation of WT (Nipponbare) and OsHDAC1-RNAi constructs (Hou et al., 2022) were transfected into EHA105 for transformation in the background of osgras30 #1.These transformations were performed by Wuhan BioRun Biosciences Co., Ltd.(Wuhan, China).
Plant materials and growth conditions
All rice plants used in this study, including Nipponbare (Oryza sativa L. ssp.japonica), the OsHDAC1 RNAi and OE lines in the background of Nipponbare (Hou et al., 2022), the OsGRAS30 mutants and OE lines, and the OsHDAC1 RNAi lines in the background of osgras30 #1 were grown in hydroponic cultures in Kimura B nutrient salts under a photoperiod of 14 h light at 28 °C and 10 h dark at 26 °C with 70% humidity.For growth in the OsGRAS30-OsHDAC1 regulates blast resistance 1751 field, germinated rice seedlings were transplanted into the Huashan rice experimental paddy fields at the Wuhan University in Wuhan, Hubei province, China in mid-May; grains were harvested in October.Urea was used as the nitrogen source, applied at 2 kg N/100 m 2 .The rice plants were transplanted at a density of three rows of 38 plants with spacing between rice plants of 20 cm.The N. benthamiana plants used in this study were grown in soil under a photoperiod of 14 h light at 22 °C and 10 h dark at 22 °C with 70% humidity.
Inoculation of rice leaves with M. Oryzae strains
The M. oryzae strains Guy11, AH4, Sc09-153-07, TM3-2, RB1, RB3, NC1, HLJ1-3, HLJ09-17-1 and HLJ5-3 were used in this study.M. oryzae cultivation and inoculation assay were performed using methods described previously with modifications (Li et al., 2017).Briefly, M. oryzae strains were gown on complete agar medium in a growth chamber at 25 °C under a 12-h light/12-h dark photoperiod to produce spores, which were collected with sterile water at a final concentration of 5 × 10 5 spores/mL in suspension.For spray inoculation, 2-week-old rice seedlings were sprayed in a dew growth chamber with spore suspensions from various M. oryzae strains.At 5 days after inoculation, the lesion area was measured using Photoshop according to a previously reported method (Li et al., 2017), and relative fungal biomass was determined by qPCR analysis of the M. oryzae POT2 gene, the expression of which was normalized to that of the rice Ubiquitin gene, as described previously (Wang et al., 2018).All inoculation experiments were repeated three times independently.
HDAC inhibitor treatments
TSA (Selleck Chemicals, Cat.# S1045, Houston, TX, USA) was dissolved in dimethyl sulfoxide.NaB (Beyotime, Cat.# ST1636, Shanghai, China) was dissolved in double-distilled water.Twoweek-old rice seedlings were spray-inoculated with M. oryzae spore suspensions and transferred to Kimura B nutrient salts containing the indicated concentration of TSA or NaB for 5 days.Control plants were grown in an identical volume of dimethyl sulfoxide or double-distilled water as a mock treatment.
RNA extraction and RT-qPCR
RNA was extracted from the leaves of 2-week-old seedlings with Trizol reagent (ThermoFisher, Cat.# 15596026, Waltham, MA, USA) and then reverse-transcribed into cDNA using the Super-Script III transcriptase kit (ThermoFisher, Cat.# 18080093).RT-qPCR analysis was conducted using SYBR green master mix (Bio-Rad, Cat.# 1725150, Hercules, CA, USA) on the CFX96 realtime detection system 690 (Bio-Rad, USA).The rice Ubiquitin (LOC_Os03g13170) gene was used as an internal control (Wang et al., 2018).The mRNA levels relative to the Ubiquitin level were calculated according to the 2 ÀΔΔCt method.
Protein extraction and immunoblotting
The leaves of 2-week-old seedlings were ground into powder in liquid nitrogen.Then, histones were extracted using a protein extraction kit (Bestbio, Cat.# BB-31171, Shanghai, China), and total protein was extracted using another protein extraction kit (Bestbio, Cat.# BB-319815).Protein concentrations were measured using a bicinchoninic acid protein assay kit (Bio-sharp, Cat.# BL521A, Hefei, China).Subsequently, proteins were heated for 10 min at 95 °C, separated using a 15% or 12% sodium dodecyl sulfate-polyacrylamide electrophoresis gel, transferred to polyvinylidene fluoride membranes, and detected as described previously (Hou et al., 2017).All western blots were developed using the ECL Plus immunoblotting detection system (GE Healthcare, Little Chalfont, UK).
Co-IP assay
Co-IP assays were performed as described previously (Zhang et al., 2017).The OsHDAC1-HA and OsGRAS30-MYC constructs or HA and OsGRAS30-MYC constructs were transformed into rice protoplasts using a plant protoplast preparation and transformation kit (Real-Times, Cat.# RTU4052, Beijing, China) in accordance with the manufacturer's instructions.Then, total proteins were extracted from rice protoplasts and incubated with rProtein A sepharose (GE Healthcare, USA) and an anti-HA antibody.Proteins bound to sepharose were detected using an anti-MYC antibody.
Split-LUC assay
Split-LUC assays were performed according to a previously described method (Hou et al., 2022).The OsHDAC1-nLUC and ª 2024 The Authors.Plant Biotechnology Journal published by Society for Experimental Biology and The Association of Applied Biologists and John Wiley & Sons Ltd., 22, 1740-1756 cLUC-OsGRAS30 constructs were transformed into A. tumefaciens strain GV3101(pSoup-p19).Equal amounts of bacterial suspension were infiltrated into the leaves of 5-week-old N. benthamiana plants.The inoculated leaves were sprayed with 0.5 mM luciferin (Coolaber, Cat.# CL6928) after 3 days and then kept in the dark for 5 min.Luminescence images were taken using the Tanon 5200 imaging apparatus (Shanghai, China).
In vitro pull-down assay
Pull-down assays were performed as described previously (Hou et al., 2022).Glutathione resin containing GST or GST-OsHDAC1 was incubated with MBP-OsGRAS30-His in phosphate-buffered saline at 4 °C for 2 h.The proteins eluted from the resin were detected by immunoblotting using an anti-His antibody.
Subcellular localization assay
To assess the co-localization of OsHDAC1 with OsGRAS30 in rice protoplasts, the OsHDAC1-sGFP (Hou et al., 2022) and OsGRAS30-mCherry constructs were transformed into rice protoplasts using a plant protoplast preparation and transformation kit in accordance with the manufacturer's instructions.The transformed protoplasts were observed using a confocal laser scanning microscope (Leica SP8, Wetzlar, Germany).Images were analysed using LAS-AF software.The settings used for confocal microscopy were as follows: for sGFP, excitation 488 nm, emission 500-550 nm; for mCherry, excitation 552 nm, emission 575-625 nm; fluorescence intensity, 1%-10%; gain value of the HyD detector, 100-200.
HDAC activity assay
To assess HDAC activity in vivo, leaves of 2-week-old seedlings from WT, OsGRAS30 mutant and OE lines were ground into powder in liquid nitrogen and extracted using a protein extraction kit (Bestbio, Cat.# BB-319815).Protein concentrations in each sample were measured using a bicinchoninic acid protein assay kit (Bio-sharp, Cat.# BL521A), and approximately 4 μg total protein was used to assess in vivo HDAC activity.For the in vitro HDAC activity assay, purified recombinant GST-OsHDAC1 ($2 μg) and MBP-OsGRAS30-His ($2 μg) or MBP-His ($2 μg) proteins were incubated at 37 °C for 1 h, and the mixed proteins were used to assess in vitro HDAC activity.Subsequently, the mixed proteins were subjected to in vivo and in vitro HDAC activity assays using the Epigenase HDAC activity/inhibition direct assay kit (Epigentek, Cat.# P-4034, Farmingdale, NY, USA) in accordance with the manufacturer's instructions.After the reactions were complete, the optical density at 450 nm was measured using a SpectraMax iD5 microplate reader (Molecular Devices, Sunnyvale, CA, USA).The activity of the HDAC enzyme in proportion to optical density (measured as intensity in relative fluorescence units [RFU]) was calculated using the formula: ChIP-Seq or -PCR Two-week-old seedlings were ground into a fine powder and suspended in a nuclear isolation buffer (10 mM HEPES [pH 8.0], 1 M sucrose, 5 mM KCl, 5 mM MgCl 2 , 0.6% Triton X-100, 0.4 mM phenylmethylsulfonyl fluoride, 1× protease inhibitor cocktail, 1% formaldehyde) for 20 min for crosslinking.ChIP assays were performed using a ChIP kit (Bersinbio, Cat.# Bes5002, Guangzhou, China) in accordance with the manufacturer's instructions.Finally, the eluted DNA samples were digested with proteinase K (Invitrogen, Cat.# 25530049, Waltham, MA, USA) and RNase A and purified for subsequent sequencing or qPCR.ChIP-Seq for analysis of OsHDAC1 occupancy was performed using an antibody against the HA tag in OsHDAC1 OE lines, with two replicates.The input was used as a negative control, and IgG was used as an experimental control for the HA antibody.ChIP-Seq for genome-wide H3K27ac analysis was performed using an antibody against H3K27ac in WT plants, with two replicates.Immunoprecipitated and purified DNA was used for library construction and sequencing, which were performed on the Illumina Novaseq 6000 sequencing system by Seqhealth Biotech (Wuhan, China).
ChIP-qPCR was performed as described previously (Ke et al., 2020).In brief, the immunoprecipitated and purified DNA was subjected to qPCR to amplify the target sequences.The Ubiquitin (LOC_Os03g13170) promoter was used as an internal reference control.
Analysis of ChIP-Seq data
A total of 10 Gb high-quality 150-bp paired-end reads were generated from each sample.Raw data obtained from Illumina sequencing were processed and filtered using FastP to generate clean reads.These reads were mapped to the Nipponbare genome (MSU 7.0) using Bowtie2 with the default parameters, and only unique alignments were retained.The SAM-formatted output files generated by Bowtie2 were transformed to BAM format, followed by sorting and indexing using Samtools.Peaks were called using MACS2, and all duplicate reads were retained for OsHDAC1 and H3K27ac ChIP-Seq analysis.The q-value cutoff applied to calculate statistical significance was <0.01.Other parameters were set to their default values.The abundance of OsHDAC1 and H3K27ac ChIP-Seq reads was normalized to the corresponding input level.The genomic regions enriched in OsHDAC1 and H3K27ac were determined by comparison of the corresponding ChIP library with its input DNA library.All statistical analyses and figure construction were conducted using R.
RNA-Seq and data analysis
Total RNA was isolated from 2-week-old seedlings of WT and OsHDAC1 Ri2 plants in three biological replicates using a TRIzol kit according to previously reported methods (Hou et al., 2021).Subsequently, mRNA enrichment, cDNA synthesis, library construction and sequencing were performed using the Illumina Novaseq 6000 system by Seqhealth Biotech (Wuhan, China).
Finally, 6 Gb high-quality 150-bp paired-end reads were generated from each sample.Raw data obtained from Illumina sequencing were processed and filtered using Fastp to generate clean reads.These reads were aligned to the Nipponbare reference genome (MSU 7.0) using the splice junction mapper TopHat2.HTSeq was used to obtain the read counts mapped to each gene.The read counts were normalized as reads per kilobase per million reads to assess relative expression levels.EdgeR was used for differential expression analysis.Genes showing an absolute value of log2 fold change (OsHDAC1 Ri2/WT) > 1 and adjusted P-value <0.05 were considered differentially expressed.All the statistical analyses and figure construction were conducted using R.
Figure 1
Figure 1 OsHDAC1 regulates blast resistance in rice.(a) Transcript variation of OsHDAC1 in response to M. oryzae Guy11 treatment was analysed by RT-qPCR using total RNA extracted from leaves at 3, 6, 12, 24 or 48-h post-infection (hpi).Gene expression levels in the WT without M. oryzae treatment were set to 1.00.Values are means AE SEM (n = 3).Experiments were repeated three times with similar results.(b) Protein level of OsHDAC1was reduced in response to M. oryzae Guy11 treatment, as shown by immunoblotting analysis.Water treatment was used as a control, and the actin protein was applied as an equal loading control.(c) Disease symptoms after spray-inoculation with a spore suspension of M. oryzae Guy11 for 5 days post inoculation (dpi) in OsHDAC1 RNAi lines and WT.Scale bar = 1 cm.Representative data from three independent experiments.(d) Relative fungal biomass in OsHDAC1 RNAi lines compared with WT determined by qPCR for the M. oryzae Pot2 (MoPot2) gene normalized to rice Ubiquitin.Fungal biomass in the WT was set to 1.00.Values are means AE SEM (n = 3).Experiments were repeated three times with similar results.(e) Disease symptoms after sprayinoculation with a spore suspension of M. oryzae Guy11 for 5 dpi in OsHDAC1 OE lines and WT.Scale bar = 1 cm.Representative data from three independent experiments.(f) Relative fungal biomass in OsHDAC1 OE lines compared with WT determined using qPCR for the MoPot2 gene normalized to rice Ubiquitin.Fungal biomass in the WT was set to 1.00.Values are means AE SEM (n = 3).Experiments were repeated three times with similar results.(g) Disease symptoms of the WT, OsHDAC1 OE and RNAi rice leaves inoculated with mixed isolates including AH4, Sc09-153-07, TM3-2, RB1, RB3, NC1, HLJ1-3, HLJ09-17-1 and HLJ5-3.Scale bar = 1 cm.Representative data from three independent experiments.(h) Relative fungal biomass in OsHDAC1 RNAi and OE lines compared with WT determined using qPCR for the MoPot2 gene normalized to rice Ubiquitin.Fungal biomass in the WT was set to 1.00.Values are means AE SEM (n = 3).Experiments were repeated three times with similar results.Asterisks represent significant differences determined by Student's t-test (*P < 0.05; **P < 0.01; ***P < 0.001).WT, wild-type; Ri, RNA interference; OE, overexpression.
Figure 2
Figure 2 OsHDAC1 interacts with OsGRAS30.(a) Yeast two-hybrid (Y2H) assay showing OsHDAC1 and OsGRAS30 interaction in yeast.Empty BD and AD were used as the negative controls.(b) Coimmunoprecipitation assay showed that OsHDAC1 interacted with OsGRAS30 in rice protoplast.+ or -, presence or absence of protein.(c) Split-luciferase complementation assay showing the interaction between OsHDAC1 and OsGRAS30 in N. benthamiana leaves.Empty nLUC and cLUC were used as negative controls.(d) GST pull-down assay showing the interaction between OsHDAC1 and OsGRAS30 in vitro.GST-OsHDAC1 and GST were stained with CBB as loading controls.GST and MBP tags were used as the negative control.+ or -, presence or absence of protein.(e) Confocal microscopy result of OsHDAC1-sGFP co-localizing with OsGRAS30-mCherry in rice protoplasts.Scale bar = 7.5 μm.(f-h) Y2H assays showing the interaction between different regions of OsHDAC1 and different regions of OsGRAS30 in yeast.Schematic of the domain structure of OsHDAC1 (f) and OsGRAS30 (g) are indicated.SD, synthetic dropout medium; BD, binding domain; AD, activation domain; TL, SD medium lacking leucine and tryptophan; TLHA, SD medium lacking leucine, tryptophan, histidine and adenine; X, X-α-gal.HA, HA epitope tag sequence; MYC, MYC epitope tag sequence; GST, GST epitope tag sequence; His, His epitope tag sequence; MBP, MBP epitope tag sequence; CBB, Coomassie brilliant blue; LUC, luciferase gene; N-ter, N-terminal region; C-ter, C-terminal region; HD, histone deacetylase domain; GRAS, GRAS domain.
Figure 3
Figure 3 OsGRAS30 is correlated with blast resistance in rice.(a) Transcript variation of OsGRAS30 in response to M. oryzae.Total RNA was extracted from leaves after 3, 6, 12, 24 or 48-h M. oryzae infection and subjected to RT-qPCR.Expression levels of genes in the WT without M. oryzae treatment were set to 1.00.Values are mean AE SEM (n = 3).Experiments were repeated three times with similar results.(b) Gene structure and mutation sites of OsGRAS30.sgRNA, small guide RNA; PAM, protospacer adjacent motif.(c) Disease symptoms after spray-inoculation with a spore suspension of M. oryzae for 5 dpi in OsGRAS30 mutants and WT.Scale bar = 1 cm.Representative data from three independent experiments.(d) Relative fungal biomass in OsGRAS30 mutants compared with WT determined using qPCR for the MoPot2 normalized to rice Ubiquitin.Fungal biomass in the WT treated with M. oryzae was set to 1.00.Values are means AE SEM (n = 3).Experiments were repeated three times with similar results.(e) Disease symptoms after sprayinoculation with a spore suspension of M. oryzae for 5 dpi in OsGRAS30 OE lines and WT.Scale bar = 1 cm.Representative data from three independent experiments.(f) Relative fungal biomass in OsGRAS30 OE lines compared with WT determined by qPCR for the MoPot2 normalized to rice Ubiquitin.Fungal biomass in the WT treated with M. oryzae was set to 1.00.Values are means AE SEM (n = 3).Experiments were repeated three times with similar results.(g) Disease symptoms of the OsGRAS30 mutant and OE lines in rice leaves inoculated with mixed isolates including AH4, Sc09-153-07, TM3-2, RB1, RB3, NC1, HLJ1-3, HLJ09-17-1 and HLJ5-3 compared with WT.Scale bar = 1 cm.Representative data from three independent experiments.(h) Relative fungal biomass determined in OsGRAS30 mutant and OE lines compared with WT determined using qPCR for the MoPot2 gene normalized to rice Ubiquitin.Values are means AE SEM (n = 3).Experiments were repeated three times with similar results.Asterisks mark significant changes compared with WT without M. oryzae treatment (a) or WT with M. oryzae treatment (d, f and h) based on Student's t-test: *P < 0.05, **P < 0.01, ***P < 0.001.WT, wild-type; OE, overexpression.
Figure 4
Figure 4 OsGRAS30 promotes blast resistance in an OsHDAC1-dependent manner.(a) Expression levels of OsHDAC1 in OsHDAC1 RNAi lines (from #1 to #10) with the osgras30 #1 background.Total RNA was extracted from leaves and subjected to RT-qPCR.Expression levels of OsHDAC1 in the osgras30 #1 plants were set to 1.00.Values are mean AE SEM (n = 3).Experiments were repeated three times with similar results.(b) Disease symptoms after spray-inoculation with a spore suspension of M. oryzae for 5 dpi in OsHDAC1 RNAi lines with the osgras30 #1 background compared with WT and OsGRAS30 mutants.Scale bar = 1 cm.Representative data from three independent experiments.(c) Relative fungal biomass in OsHDAC1 RNAi lines with the osgras30 #1 background compared with WT and OsGRAS30 mutants determined using qPCR for the MoPot2 gene normalized to rice Ubiquitin.Fungal biomass in the WT plants infected with M. oryzae was set to 1.00.Values are means AE SEM (n = 3).Experiments were repeated three times with similar results.Asterisks mark significant changes compared with osgras30 #1(a and c) based on Student's t-test: **P < 0.01, ***P < 0.001.
Figure 5
Figure 5 OsGRAS30 represses OsHDAC1 to induce H3K27ac hypoacetylation.(a) In vivo HDAC activity assays in WT, OsGRAS30 mutant and OE lines.Values are mean AE SEM (n = 3).Experiments were repeated three times with similar results.(b) In vitro analysis of the effect of OsGRAS30 on OsHDAC1 activity.Values are mean AE SEM (n = 3).+ or -, presence or absence of proteins.Experiments were repeated three times with similar results.(c) Overall H3ac, H4ac, H3K9ac and H3K27ac levels in WT, OsHDAC1 RNAi and OE lines using immunoblotting.Histone H3 was applied as an equal loading control.Representative data from three independent experiments.(d) Immunoblotting analysis of the overall H3ac, H4ac, H3K9ac and H3K27ac levels in WT, OsGRAS30 mutant lines, and OsHDAC1 RNAi lines in the background of osgras30 #1 and OsGRAS30 OE lines (line numbers are given above the graph).H3 protein was applied as an equal loading control.Representative data from three independent experiments.(e) Disease symptoms after spray-inoculation with a spore suspension of M. oryzae for 5 dpi in osgras30#1 and WT treated with HDAC inhibitors including TSA and NaB.Untreated was used as a control.Scale bar = 1 cm.Representative data from three independent experiments.(f) Relative fungal biomass was determined using qPCR for the MoPot2 gene normalized to rice Ubiquitin.Values are means AE SEM (n = 3).Experiments were repeated three times with similar results.Asterisks mark significant changes compared with WT (a and b) or WT with M. oryzae treatment (f) based on Student's t test: *P < 0.05, **P < 0.01, ***P < 0.001.n. s., not significant change.WT, wild-type; Ri, RNA interference; OE, overexpression.
ª
2024 The Authors.Plant Biotechnology Journal published by Society for Experimental Biology and The Association of Applied Biologists and John Wiley & Sons Ltd., 22, 1740-1756
Figure 6
Figure 6 OsHDAC1 co-occupies with H3K27ac at genome wide.(a) Metagene plot and heatmap showing OsHDAC1 binding levels on the 2.0-kb flanking genomic region surrounding TSS in rice.(b) Pie chart showing the distribution of OsHDAC1 binding peaks at annotated genic and intergenic regions in the rice genome.(c) Metagene plot showing OsHDAC1 binding profiles on the 1.0-kb flanking genomic region surrounding TSS of genes with high, medium, low and silent expression.(d) DREME identified representative DNA motifs in OsHDAC1 binding sites.(e) Metagene plot and heatmap showing H3K27ac enrichment levels on the 2.0-kb flanking genomic region surrounding TSS in rice.(f) Pie chart shows the distribution of H3K27ac enrichment peaks at annotated genic and intergenic regions in the rice genome (g) Metagene plot showing OsHDAC1 binding profiles on the 1.0-kb flanking genomic region surrounding TSS of genes with high, medium, low and silent expression.(h) Correlation between OsHDAC1 binding and H3K27ac level in OsHDAC1 target genes.(i) Venn diagram showing overlap of OsHDAC1 binding genes and H3K27 acetylated genes.(j) IGV screen shoot showing the co-localization between OsHDAC1 and H3K27ac at a representative region in chromosome 2. TSS, the transcription start site.
Figure 7
Figure 7 OsHDAC1 regulates the expression of blast resistance-related genes.(a) Volcano plot showing differentially expressed genes (DEGs) associated with OsHDAC1 in the OsHDAC1 Ri2 line compared with the WT.(b) Venn diagram showing overlap of DEGs in the OsHDAC1 Ri2 line, OsHDAC1 binding genes and H3K27ac enrichment genes.(c) Go results showing the related biological pathway enriched from downregulated DEGs in OsHDAC1 Ri2 line compared with WT.(d) Go results showing the related biological pathway enriched from upregulated DEGs in OsHDAC1 Ri2 line compared with WT.(e)The transcript level of blast resistance-related genes in OsHDAC1 OE and RNAi lines compared with WT.Total RNA was extracted from two-week-old rice leaves and subjected to RT-qPCR.Expression levels of genes in the WT were set to 1.00.Values are mean AE SEM (n = 3).Experiments were repeated three times with similar results.(f) IGV views showing OsHDAC1 and H3K27ac enrichment in blast resistance-related genes.Input serves as the negative control.(g) ChIP-qPCR results of OsHDAC1 binding levels at blast resistance-related genes in OsHDAC1 OE plants.ChIP-qPCR results of IgG binding levels of genes were set to 1.00.Values are mean AE SEM (n = 3).Experiments were repeated three times with similar results.(h) ChIP-qPCR results of H3K27ac levels at blast resistance-related genes in OsHDAC1 Ri2 line compared with WT plants.ChIP-qPCR results of H3K27ac levels of genes in the WT were set to 1.00.Values are mean AE SEM (n = 3).Experiments were repeated three times with similar results.Asterisks mark significant changes compared with WT (e and h) or IgG (g) based on Student's t test: *P < 0.05, **P < 0.01, ***P < 0.001.(i) A proposed working model showing the role of OsGRAS30 in manipulating the OsHDAC1-mediated broad-spectrum blast resistance in rice.WT, wild-type; Ri, RNA interference; OE, overexpression.
ª
2024 The Authors.Plant Biotechnology Journal published by Society for Experimental Biology and The Association of Applied Biologists and John Wiley & Sons Ltd., 22, 1740-1756 et al.
ª
2024 The Authors.Plant Biotechnology Journal published by Society for Experimental Biology and The Association of Applied Biologists and John Wiley & Sons Ltd., 22, 1740-1756
Figure S4
Figure S4 Diagrammatic representation of vector construction and expression analysis of OsGRAS30 in transgenic plants.Figure S5 Protein abundance and transcript levels of OsHDAC1 in OsGRAS30 overexpression lines, mutants and osgras30 #1/OsH-DAC1 RNAi lines.Figure S6 H3K27ac level was increased in response to M. oryzae Guy11 treatment, as shown by immunoblotting analysis.Figure S7 Rice blast disease symptoms after treatment with HDAC inhibitors.Figure S8 Specificity testing of the anti-HA antibody for chromatin immunoprecipitation sequencing (ChIP-Seq) analysis in OsHDAC1 OE5 plants.Figure S9 The correlation analysis of the two biological replicates of OsHDAC1 in ChIP-seq.Figure S10 The correlation analysis of the two biological replicates of H3K27ac in ChIP-seq.Figure S11 The correlation analysis of the three biological replicates of OsHDAC1 in RNA-seq.Figure S12 Heatmap showing different expression of numerous DEGs in rice seedlings between the OsHDAC1 Ri2 line and WT.Figure S13 Significantly enriched KEGG terms among OsHDAC1regulated DEGs.Table S1 Primers used in this study.Dataset S1 OsHDAC1 binding sites identified by ChIP-Seq.Dataset S2 H3K27 distribution sites identified by ChIP-Seq.Dataset S3 List of DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S4 GO pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S5 KEGG pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S6 GO pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S7 KEGG pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.
Figure S5
Figure S4 Diagrammatic representation of vector construction and expression analysis of OsGRAS30 in transgenic plants.Figure S5 Protein abundance and transcript levels of OsHDAC1 in OsGRAS30 overexpression lines, mutants and osgras30 #1/OsH-DAC1 RNAi lines.Figure S6 H3K27ac level was increased in response to M. oryzae Guy11 treatment, as shown by immunoblotting analysis.Figure S7 Rice blast disease symptoms after treatment with HDAC inhibitors.Figure S8 Specificity testing of the anti-HA antibody for chromatin immunoprecipitation sequencing (ChIP-Seq) analysis in OsHDAC1 OE5 plants.Figure S9 The correlation analysis of the two biological replicates of OsHDAC1 in ChIP-seq.Figure S10 The correlation analysis of the two biological replicates of H3K27ac in ChIP-seq.Figure S11 The correlation analysis of the three biological replicates of OsHDAC1 in RNA-seq.Figure S12 Heatmap showing different expression of numerous DEGs in rice seedlings between the OsHDAC1 Ri2 line and WT.Figure S13 Significantly enriched KEGG terms among OsHDAC1regulated DEGs.Table S1 Primers used in this study.Dataset S1 OsHDAC1 binding sites identified by ChIP-Seq.Dataset S2 H3K27 distribution sites identified by ChIP-Seq.Dataset S3 List of DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S4 GO pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S5 KEGG pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S6 GO pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S7 KEGG pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.
Figure S6
Figure S4 Diagrammatic representation of vector construction and expression analysis of OsGRAS30 in transgenic plants.Figure S5 Protein abundance and transcript levels of OsHDAC1 in OsGRAS30 overexpression lines, mutants and osgras30 #1/OsH-DAC1 RNAi lines.Figure S6 H3K27ac level was increased in response to M. oryzae Guy11 treatment, as shown by immunoblotting analysis.Figure S7 Rice blast disease symptoms after treatment with HDAC inhibitors.Figure S8 Specificity testing of the anti-HA antibody for chromatin immunoprecipitation sequencing (ChIP-Seq) analysis in OsHDAC1 OE5 plants.Figure S9 The correlation analysis of the two biological replicates of OsHDAC1 in ChIP-seq.Figure S10 The correlation analysis of the two biological replicates of H3K27ac in ChIP-seq.Figure S11 The correlation analysis of the three biological replicates of OsHDAC1 in RNA-seq.Figure S12 Heatmap showing different expression of numerous DEGs in rice seedlings between the OsHDAC1 Ri2 line and WT.Figure S13 Significantly enriched KEGG terms among OsHDAC1regulated DEGs.Table S1 Primers used in this study.Dataset S1 OsHDAC1 binding sites identified by ChIP-Seq.Dataset S2 H3K27 distribution sites identified by ChIP-Seq.Dataset S3 List of DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S4 GO pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S5 KEGG pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S6 GO pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S7 KEGG pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.
Figure S7
Figure S4 Diagrammatic representation of vector construction and expression analysis of OsGRAS30 in transgenic plants.Figure S5 Protein abundance and transcript levels of OsHDAC1 in OsGRAS30 overexpression lines, mutants and osgras30 #1/OsH-DAC1 RNAi lines.Figure S6 H3K27ac level was increased in response to M. oryzae Guy11 treatment, as shown by immunoblotting analysis.Figure S7 Rice blast disease symptoms after treatment with HDAC inhibitors.Figure S8 Specificity testing of the anti-HA antibody for chromatin immunoprecipitation sequencing (ChIP-Seq) analysis in OsHDAC1 OE5 plants.Figure S9 The correlation analysis of the two biological replicates of OsHDAC1 in ChIP-seq.Figure S10 The correlation analysis of the two biological replicates of H3K27ac in ChIP-seq.Figure S11 The correlation analysis of the three biological replicates of OsHDAC1 in RNA-seq.Figure S12 Heatmap showing different expression of numerous DEGs in rice seedlings between the OsHDAC1 Ri2 line and WT.Figure S13 Significantly enriched KEGG terms among OsHDAC1regulated DEGs.Table S1 Primers used in this study.Dataset S1 OsHDAC1 binding sites identified by ChIP-Seq.Dataset S2 H3K27 distribution sites identified by ChIP-Seq.Dataset S3 List of DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S4 GO pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S5 KEGG pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S6 GO pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S7 KEGG pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.
Figure S8
Figure S4 Diagrammatic representation of vector construction and expression analysis of OsGRAS30 in transgenic plants.Figure S5 Protein abundance and transcript levels of OsHDAC1 in OsGRAS30 overexpression lines, mutants and osgras30 #1/OsH-DAC1 RNAi lines.Figure S6 H3K27ac level was increased in response to M. oryzae Guy11 treatment, as shown by immunoblotting analysis.Figure S7 Rice blast disease symptoms after treatment with HDAC inhibitors.Figure S8 Specificity testing of the anti-HA antibody for chromatin immunoprecipitation sequencing (ChIP-Seq) analysis in OsHDAC1 OE5 plants.Figure S9 The correlation analysis of the two biological replicates of OsHDAC1 in ChIP-seq.Figure S10 The correlation analysis of the two biological replicates of H3K27ac in ChIP-seq.Figure S11 The correlation analysis of the three biological replicates of OsHDAC1 in RNA-seq.Figure S12 Heatmap showing different expression of numerous DEGs in rice seedlings between the OsHDAC1 Ri2 line and WT.Figure S13 Significantly enriched KEGG terms among OsHDAC1regulated DEGs.Table S1 Primers used in this study.Dataset S1 OsHDAC1 binding sites identified by ChIP-Seq.Dataset S2 H3K27 distribution sites identified by ChIP-Seq.Dataset S3 List of DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S4 GO pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S5 KEGG pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S6 GO pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S7 KEGG pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.
Figure S4 Diagrammatic representation of vector construction and expression analysis of OsGRAS30 in transgenic plants.Figure S5 Protein abundance and transcript levels of OsHDAC1 in OsGRAS30 overexpression lines, mutants and osgras30 #1/OsH-DAC1 RNAi lines.Figure S6 H3K27ac level was increased in response to M. oryzae Guy11 treatment, as shown by immunoblotting analysis.Figure S7 Rice blast disease symptoms after treatment with HDAC inhibitors.Figure S8 Specificity testing of the anti-HA antibody for chromatin immunoprecipitation sequencing (ChIP-Seq) analysis in OsHDAC1 OE5 plants.Figure S9 The correlation analysis of the two biological replicates of OsHDAC1 in ChIP-seq.Figure S10 The correlation analysis of the two biological replicates of H3K27ac in ChIP-seq.Figure S11 The correlation analysis of the three biological replicates of OsHDAC1 in RNA-seq.Figure S12 Heatmap showing different expression of numerous DEGs in rice seedlings between the OsHDAC1 Ri2 line and WT.Figure S13 Significantly enriched KEGG terms among OsHDAC1regulated DEGs.Table S1 Primers used in this study.Dataset S1 OsHDAC1 binding sites identified by ChIP-Seq.Dataset S2 H3K27 distribution sites identified by ChIP-Seq.Dataset S3 List of DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S4 GO pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S5 KEGG pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S6 GO pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S7 KEGG pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.
Figure S4 Diagrammatic representation of vector construction and expression analysis of OsGRAS30 in transgenic plants.Figure S5 Protein abundance and transcript levels of OsHDAC1 in OsGRAS30 overexpression lines, mutants and osgras30 #1/OsH-DAC1 RNAi lines.Figure S6 H3K27ac level was increased in response to M. oryzae Guy11 treatment, as shown by immunoblotting analysis.Figure S7 Rice blast disease symptoms after treatment with HDAC inhibitors.Figure S8 Specificity testing of the anti-HA antibody for chromatin immunoprecipitation sequencing (ChIP-Seq) analysis in OsHDAC1 OE5 plants.Figure S9 The correlation analysis of the two biological replicates of OsHDAC1 in ChIP-seq.Figure S10 The correlation analysis of the two biological replicates of H3K27ac in ChIP-seq.Figure S11 The correlation analysis of the three biological replicates of OsHDAC1 in RNA-seq.Figure S12 Heatmap showing different expression of numerous DEGs in rice seedlings between the OsHDAC1 Ri2 line and WT.Figure S13 Significantly enriched KEGG terms among OsHDAC1regulated DEGs.Table S1 Primers used in this study.Dataset S1 OsHDAC1 binding sites identified by ChIP-Seq.Dataset S2 H3K27 distribution sites identified by ChIP-Seq.Dataset S3 List of DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S4 GO pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S5 KEGG pathways of downregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S6 GO pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants.Dataset S7 KEGG pathways of upregulated DEGS in OsHDAC1 Ri2 plants vs WT plants. | 2024-02-01T06:17:23.775Z | 2024-01-31T00:00:00.000 | {
"year": 2024,
"sha1": "608bd915148f99adad54ff42354426125661ae15",
"oa_license": "CCBY",
"oa_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/pbi.14299",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "52b8ec0ef4738dce55d4eeb4fea7f250e5f7d586",
"s2fieldsofstudy": [
"Biology",
"Environmental Science",
"Agricultural and Food Sciences"
],
"extfieldsofstudy": [
"Medicine"
]
} |
56176327 | pes2o/s2orc | v3-fos-license | Genome-wide association mapping of spot blotch resistance in wheat association mapping initiative (WAMI) panel of spring wheat (Triticum aestivum L.)
Spot blotch (SB) caused by Bipolaris sorokiniana, is one of the most important diseases of wheat in the eastern part of south Asia causing considerable yield loss to the wheat crop. There is an urgent need to identify genetic loci closely associated with resistance to this pathogen for developing resistant cultivars. Hence, genomic regions responsible for SB resistance were searched using a wheat association mapping initiative (WAMI) panel involving 287 spring wheat genotypes of different origin. Genome-wide association mapping (GWAM) was performed using single nucleotide polymorphism (SNP) markers from a custom 90 K wheat SNP array. A mixed linear model (MLM) was used for assessing the association of SNP markers with spot blotch resistance in three consecutive years. Three traits were measured: incubation period, lesion number and area under the disease progress curve (AUDPC). Significant SNP markers were found linked to five, six and four quantitative trait loci (QTLs) for incubation period, lesion number and AUDPC respectively. They were detected on 11 different chromosomes: 1A, 1B, 1D, 4A, 5A, 5B, 6A, 6B, 6D, 7A, 7B with marker R2 range of 0.083 to 0.11. The greatest number of significant SNP-markers was found for lesion number and AUDPC on chromosome 6B and 5B, respectively, representing a better coverage of B-genome by SNPs. On the other hand, the most significant and largest SNP markers for incubation period were detected on 6A and 4A chromosomes indicating that this trait is associated with the A-genome of wheat. Although, QTLs for spot blotch resistance have been reported in wheat on these same chromosomes, the association of incubation period and lesion number with SB resistance has not been reported in previous studies. The panel exhibits considerable variation for SB resistance and also provides a good scope of marker-assisted selection using the identified SNP markers linked to resistant QTLs.
Introduction Wheat (Triticum aestivum L.), one of the most widely adapted cereal crops [1], provides around one-fifth of the total calories to the human population. In addition, it also provides 20% of the protein to more than 4.5 billion people in 94 developing countries [2]. In south Asia, wheat is cultivated in over 40 million hectare area, out of which around 25% falls under Eastern Gangetic Plains (EGP) which contributes only 16% to the total wheat production. The demand of wheat in this densely populated region can only be met if annual production is increased by 1.5-2.0% [3,4]. The EGP of south Asia is characterized as the mega environment 5A due to prevalence of high temperature and humidity during the wheat growing season [5,6].In this region, wheat production is significantly impacted due to spot blotch disease [6,7,8] which may cause yield loss up to 30% [9]. The causal organism of spot blotch (SB) is Bipolaris sorokiniana, which is an anamorph (teleomorph Cochliobolus sativus).
Reduction in disease progress is contributed by few components of resistance i.e., latent period and lesion size [10] and epidemic development can be slow down by using genotypes with longer latent period and fewer lesions [11]. Hence molecular mapping for different components of resistance in plant populations to determine robust markers and their subsequent use is imperative for varietal improvement programs through an effective use of appropriate genetic resources [12,13].Due to various limitations of the conventional bi-parental linkage mapping, there has been a greater preference for linkage disequilibrium (LD) based association mapping (AM) [6]. In AM, a collection of variable lines is scanned to understand marker-trait associations using linkage disequilibrium (LD)which is non-equilibrium association between different alleles at various loci [14] and varies across wheat chromosomes [12,15]. Although wheat has a large homeologous genome that shows weak marker coverage compared to other cereal crops [13], AM can identify superior alleles with detailed marker data in large populations which can be utilized immediately in breeding [16].
Due to significant yield losses caused by spot blotch in EGP, a considerable effort has been invested by wheat breeders to develop spot blotch resistant cultivars in this part of South Asia. To continue this effort further, this study was undertaken to identify novel marker-trait associations and to detect the loci conferring spot blotch resistance through GWAM in a WAMI Panel.
Experimental material
The experimental material comprised of 287 genetically diverse, elite spring wheat lines of the wheat association mapping initiative (WAMI) population assembled by CIMMYT in 2009 was used [17].
Experimental design and trait evaluation
The WAMI population was evaluated for spot blotch in three consecutive crop seasons during 2012-13, 2013-14 and 2014-15 at the Agricultural Research Farm of Banaras Hindu University (BHU), Varanasi, India. The experiment was conducted using Alfa lattice design with sowing done between 1 st to 10 th December in the three years. Two replications were planted and each genotype was sown in two rows of two meters maintaining row-to-row and plant-to-plant distance of 20 cm and 5 cm respectively. The lines were allocated randomly to each replication using the Fisher and Yates Random Table [18]. Fertilizer per hectare was applied as 120 kg N: 60 kg P 2 O 5 : 40 kg K 2 O. The full amount of K 2 O and P 2 O 5 were applied at sowing whereas nitrogen was applied in three split doses, half at sowing, one fourth at first irrigation (21 days after sowing-DAS) and the remaining one fourth at second irrigation (45 DAS). Data were recorded on three traits: incubation period (days), lesion number and spot blotch AUDPC.
Inoculation of pathogen
For creating artificial epiphytotic, a pure culture of Bipolaris sorokiniana (NABM MAT1; NCBIJN128877, BHU, Varanasi, India) known to be highly aggressive was used [19]. The isolate was obtained from the Department of Plant Pathology and Mycology, BHU and multiplied on sorghum grain. An adjusted spore suspension (10 −4 spores per ml) in water was applied in the evening and irrigated immediately after inoculation [20]. Frequent irrigations were provided to ensure environmental conditions conducive to spot blotch development in the field.
Recording for incubation period, lesion number and disease score
Incubation period (days) was recorded from the inoculation to first appearance of visible symptoms on five randomly tagged plants in each plot [10]. For number of lesions, five leaves from each plot were selected randomly. Total numbers of lesions present on flag leaves were counted. The leaf was divided into four parts with a marker pen and the number of lesions on each part was counted. Total number of spots of each part were added and recorded as lesions per leaf [21]. For AUDPC, disease severity (%) was recorded in three different growth stages (GS), GS 63 (beginning of anthesis to half complete), GS 69 (anthesis complete) and GS 77 (late milking). Disease severity was assessed by the formula (D 1 /9× D 2 /9) × 100 using the double digit scale (DD, 00-99) [7]. The first digit (D 1 ) refers to vertical disease progress on the plant whereas the second digit (D 2 ) was the disease severity score in the affected leaves. Thus, disease severity was used to estimate the AUDPC by following formula [22].
Where, Y i = disease severity at time t i , (t i+1 −t i ) = Time (days) between two disease scores, n = number of dates at which spot blotch was recorded.
DNA extraction and SNP Genotyping
The DNA was extracted from fresh leaves of each line following the CTAB procedure [23] and genotyped at CIMMYT, Mexico using the Illumina iSelect beadchip assay [17]for wheat having 26814 SNPs. To avoid monomorphic and low-quality SNPs, markers were removed with a minor allele frequency (MAF) <0.05. Nearly 21132polymorphic SNPs were selected and used for AM.
Phenotypic and STRUCTURE analysis
Analysis of variance (ANOVA) was completed using SAS 9.4 to determine genotypic, year and genotypic × year variances among the traits measured over the years. Population structure (Q) was analyzed for filtered marker panel through a clustering approach based on the admixture model implemented in STRUCTURE [24]. Ten iterations of STRUCTURE runs were employed for each subpopulations K that ranged from 2 to 10 with additional parameters of 10,000 burnin length and 10,000 Markov chain Monte Carlo (MCMC) iterations. Clusters with the greatest likelihood of occurrence were identified through the ΔK method [25]. The web-based tool for population structure analysis "Structure Harvester" [26] was done which maximizes ΔK and thus examines the likelihood distribution of K. Further, fixation index (Fst) for subpopulations was estimated by outputs generated from various STRUCTURE runs. Population Matrix Q was also obtained for further analysis. TASSEL 5.0 was used to calculate population Kinship matrix, which is based on scaled IBS (identity by state) method that uses marker data which has passed quality filtering.
Genome-wide association analysis
Marker-trait associations (MTAs) were detected by Mixed Linear Model (MLM) in TASSEL 5.0 (http://www.maizegenetics.net/) [27]. The general equation used for MLM analysis was as follows: Where, y = phenotype vector, β = vector of marker fixed effects; v = vector of fixed effects; u = vector of random effects (the kinship matrix); e = vector of residuals; X = genotype of a marker; Q = Kinship Matrix (obtained using STRUCTURE software) and Z = Identity matrix which considers the familial relatedness between accessions.
The Q matrix was used as a covariate to improve the precision of results by avoiding false positives. The efficient mixed model analysis [28] was used to save computing time, while other MLM parameters were kept as default. MLM, also called Q + K model, takes into account the population structure as well as familial relatedness. Significant marker-trait associations were identified based on P-value and R 2 used for estimating the magnitude of the QTL effects. Significant marker-trait associations were selected on basis of P-values � 1e-06. In addition, pair-wise LD for pairs of markers was calculated using TASSEL 5.0 considering observed and expected allele frequencies of the markers.
Results
The results of the analysis of variance (ANOVA) are presented in Table 1. The traits-incubation period, lesion number and AUDPC revealed significant (P<0.01) differences among genotype, years and genotype × year ( Table 1).
Distribution of phenotype under study has been generated which shows how performance of phenotype varies over the years (Fig 1A, 1B and 1C). More over to get the better description of phenotypes, correlation among traits under study has also been investigated (Fig 2).
Population structure and marker distribution
Optimal number of cluster i.e. K was determined to be 3 based on delta K approach implemented in STRUCTURE [24]. A bar graph representing population structure is also generated (Fig 3). A total 21132 marker which has passed quality filtering were used for the purpose of mapping. Linkage Disequilibrium (LD) plot based on association among markers has also been generated (Fig 4). It is clear from the figure that most of the markers are tightly linked as large chunks of red blocks observed below the diagonal of figure. It suggests that there has been limited scope of recombination between the markers, which facilitates association mapping of all three traits and minimum number of markers required to effectively cover the entire genome. More detailed distribution of marker over chromosome is presented in Table 2.
Genome-wide Marker-traits association with SNP-markers
In the present study, several markers associated with attributes like incubation period, lesion number and AUDPC were identified. The number of SNPs found significant at P-value threshold 1e-06 in the marker-trait association for the three traits, incubation period, lesion number and AUDPC were 13, 8 and 14, respectively (Table 3 and Fig 5A, 5B and 5C). A total of 35 SNPs markers were detected on eleven chromosomes (1A, 1B, 1D, 4A, 5A, 5B, 6A, 6B, 6D, 7A and 7B) with coefficient of determination (R 2 ) ranging between 0.083 and 0.11 (Table 3). For incubation period, 13 SNP markers were detected on five chromosomes in variable numbers; 5 on 4A, 3 on 6A, 2 on 7A, 2 on 5B and 1 on 6D (Table 3)whereas no MTA was found on rest of the chromosomes (Fig 5A). These markers were significant at P values less than 1e-06 and R 2 for significant markers ranged from 0.084 to 0.098 (Table 3). A prominent QTL was identified for incubation period on 6A Chromosome at 199 cM. Out of five markers detected on 4A chromosome, three markers (BobWhite_c20306_111, BobWhite_c11327_185 and BobWhite_c20322_153) were mapped at 0 cM and the other two (BobWhite_c17524_242 and Tdurum_contig81450_90) at 2 cM away. Two SNPs (Excalibur_rep_c69981_75 and wsnp_Ra_c2270_4383252) were mapped at 0 cM on the chromosome 6A while one SNP marker (BS00004466_51) was detected at 9 cM. Two regions (Ku_c15750_761 and TA002294-0887) were significantly associated with incubation period on 7A chromosome at 0 cM. The other three markers were significantly associated with incubation period on chromosome 5B (71 and 68 cM) and 6D (117 cM).
Discussion
Spot blotch disease of wheat is a major concern in warm humid south Asia and also other countries where similar climatic parameters exist. This is despite the significant progress already made in understanding the biology of Bipolaris sorokiniana and also about the cultural practices that can be used to manage this disease. Although some resistance QTL/genes have been identified and resistant cultivars developed, none of these cultivars are completely resistant since resistance is quantitative in nature and controlled by several genes with relatively minor effect. Conventional breeding approaches have contributed immensely to wheat improvement but the progress in developing wheat cultivars resistant to spot blotch has not been satisfactory [29,30] perhaps because of the complex and polygenic nature of resistance [8,29].
Initial exploratory analysis of phenotypic data which shows differential performance of each traits across the years, encouraged the environments under investigation is suitable for further analysis. Correlations among the traits are not having much significance suggesting rare common QTL associated with traits under study. Significant differences among genotype and genotype × year observed for all the three traits i.e. incubation period, lesion number and AUDPC revealed the differential performance of all the selected genotypes among themselves as well as across the years for three important traits. The panel exhibited considerable phenotypic, possibly due to the diverse genetic background of the germplasm. Consequently, this result acclaim the selected lines under study are important for further GWAM analysis. In this study, we analyzed the association between spot blotch incubation period, lesion number and AUDPC with SNPs from 287 spring wheat lines in the WAMI panel. We identified 13SNP markers significantly associated with loci conferring resistance to spot blotch AUDPC. In addition, 8 SNPs were identified for incubation period, and 14 for lesion number. Applicability of SNPs for genome-wide association study were verified with already detected DArT markers [31,32,33]. GWAM was strengthened by the high-throughput SNP genotyping array and a high-density map to identify putative QTLs associated with spot blotch resistance with better resolution [34].Association mapping is considered superior over bi-parental mapping due to high minor allele frequency, low LD and limited population structure to detect QTLs [32,35].
Using the WAMI panel in this study, we identified SNP markers linked to five and six QTLs for incubation period and spot blotch lesion number on chromosomes 4A, 6A, 7A, 5B, 6D and 1A, 1B, ID, 5A, 6B, 7B, respectively. Among them, SNPs markers closely linked to one QTL each on the chromosome 4A, 6A and 7A revealed that QTL on the A genome are strongly associated with the incubation period. Likewise, for lesion number, SNPs closely linked to one QTL on the chromosome 6Bwere significant. The associations of incubation period and lesion number with spot blotch resistance has not been reported in previous linkage or GWAM studies. In addition to incubation period and lesion number, we identified SNP markers linked to four QTLs for spot blotch AUDPC on the chromosomes 1B, 5B, 6A and 6B. Among them, several SNPs closely linked on chromosome 5B. In an earlier association mapping study [32]having an association panel of 566 spring wheat landraces and using 832 Diversity Array Technology (DArT) markers, several genomic regions associated with spot blotch resistance were also identified on four chromosomes (1A, 3B, 7B, 7D). In a follow-up GWAM study, the same group used 528 diverse spring wheat genotypes that were phenotyped for SB and were genotyped using a 9K wheat SNP chip [35], led to the identification of nine associated SNPs located on five different chromosomes (1B, 5A, 5B, 6B, 7B). We detected MTAs that closely corresponded to fourteen loci on four chromosomes that were previously identified and mapped at low resolution by bi-parental QTL analyses using SSRs [22,29,30]. The similar position of earlier mapped QTLs with the SNPs of the present study validated QTL for spot blotch resistance.
Conclusion
SNP markers and GWAS have been useful for QTL discovery in spring wheat. Considerable phenotypic and molecular variation was observed in the WAMI panel used suggesting the diverse genetic background of the germplasm. This study identified 15genomic regions to be associated with resistance to spot blotch disease of wheat. Of these, four regions were for AUDPC, while five and six were for incubation period and lesion number, respectively. A greater number of SNP markers were significantly associated with spot blotch AUDPC than for incubation period and lesion number. This study is the first time association of between markers and incubation period or lesion number were established and QTLs mapped through GWAM. Our data revealed that most of the SNPs were present on the B-genome of wheat. The identified SNP markers linked to resistant QTLs will be useful in augmenting breeding for spot blotch resistance in wheat.
Supporting information S1 Table. List of wheat association mapping initiative (WAMI) panel of spring wheat along with their phenotypic traits IP (incubation period), LN (lesion number) and AUDPC (Area under disease progress curve) to spot blotch disease. (XLSX) | 2018-12-19T14:03:49.955Z | 2018-12-17T00:00:00.000 | {
"year": 2018,
"sha1": "064e7988b1e9853bfc772a4dca45f42592a3032c",
"oa_license": "CCBY",
"oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0208196&type=printable",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "064e7988b1e9853bfc772a4dca45f42592a3032c",
"s2fieldsofstudy": [
"Agricultural and Food Sciences",
"Biology"
],
"extfieldsofstudy": [
"Medicine",
"Biology"
]
} |
117952614 | pes2o/s2orc | v3-fos-license | On the applicability of an equation of motion method at low-temperatures: comments on cond-mat/0309458 and cond-mat/0308413
The equation of motion method (EOM) is one of the approximations to calculate transport coefficients of interacting electron systems. The method is known to be useful to examine high-temperature properties. However, sometimes a naive application of the EOM fails to capture an important physics at low-energy scale, and it happens in recent preprints cond-mat/0309458 and cond-mat/0308413 which study a series of quantum dots. These preprints concluded that a unitarity-limit transport due to the Kondo resonance, which has been deduced from a Fermi-liquid behavior of the self-energy at T=0, $\omega=0$ [A.O., PRB {\bf 63}, 115305 (2001)], does not occur. We show that the EOM self-energy obtained with a finite cluster has accidentally a singular $1/\omega$ dependence around the Fermi energy, and it misleads one to the result incompatible with a Fermi-liquid ground state.
A Hubbard chain of finite size N connected to two noninteracting leads, which is illustrated in Fig. 1, has been studying as a model for a series of quantum dots [1,2,3,4,5]. Here N corresponds to the number of quantum dots and should be small, while the number of the sites in the noninteracting leads at left and right are infinite, t.e., N L → ∞ and N R → ∞. It has been shown for the chain of an odd N that the transmission probability reaches a unitarity-limit value at T = 0 if the system has the electron-hole symmetry (half-filling) together with the inversion symmetry (v L = v R ) [4]. This conclusion has been deduced from a Fermi-liquid behavior of the self-energy summarized simply in Eq. (6).
Recently, the dc conductance has been calculated with an extended version of the equation of motion method (EOM) [6], which is essentially the Hubbard I approximation [7], at zero temperature [1,2]. The results of these preprints contradict with the unitarity-limit transport mentioned above. The purpose of this report is to show that the discrepancy is caused just by a naive use of the EOM self-energy of obtained with a finite cluster, which accidentally has a singularity at low-energies and is not compatible with the Fermi-liquid ground state.
In Sec. II, we describes briefly the results deduced from a Fermi-liquid property Eq. (6) to make our points clear. Then, in Sec. III, the properties of the approximate selfenergy in the EOM are discussed.
A. Formulation
The dc conductance at T = 0 is determined by the value of a Green's function G + N 1 (ω) at the Fermi energy ω = 0 [4]; Here G + ij (ω) is the retarded Green's in the real space (1 ≤ ij ≤ N ), see Fig. 1, and Γ α (ω) = − v 2 α Im g + α (ω) for α = L, R with g + L (g + R ) being an interface Green's function for the semi-infinite lead at left (right). The Dyson equation can be written in a N × N matrix form; Here G 0 = {G 0+ ij } is the noninteracting Green's function for U = 0 defined with respect to the total system connected via tunneling v L and v R . The inverse matrix of the unperturbed Green's function can be expressed as Here 1 is the N × N unit matrix. The self-energy due to the Coulomb interaction Σ is also defined with respect to the total system of continuous energy spectrum. Note that the Green's function G + 1N (0) appearing in Eq. (1) can be obtained by taking the inverse of the matrix in Eq. (2).
B. Fermi-liquid behavior
At T = 0, ω = 0 not only the Fermi-liquid but also some other "liquid" states of interacting electron systems show a behavior, Im Σ(0) = 0 , and Re Σ(0) is finite.
This property of the self-energy, the electron-hole symmetry, and the inversion symmetry are the three keys to lock the conductance for odd N at the unitarity-limit value. No other assumptions have been made in the proof [4]. The perturbation expansion in U of Yamada-Yosida [8] preserves the properties of Eq. (6) also for N ≥ 2 owing to the presence of the noninteracting leads. The contributions of the hybridization Γ L and Γ R in the unperturbed part G 0 make it possible to take into account the virtual processes in which electrons visit inside the noninteracting leads of continuous spectrum during the multi-scattering by U . For even N , the conductance decreases with increasing N showing a tendency towards the development of a Mott-Hubbard insulator gap ∆ gap [4,5]. Although we are interested in the system of small N , a question arises: what can we see in the limit of N → ∞? Note that the results mentioned above hold just at T = 0. To describe the thermodynamic limit properly, we must consider the physics at finite but sufficiently low temperatures. For odd N , the Kondo resonance of width T K is situated in the Hubbard gap. As N increases, T K decreases and finally T K → 0 at N → ∞ because the developing gap disturbs the electrons to form the Kondo singlet. Thus, the value of the conductance g N for odd N (= 2M + 1) must depend on the order to take the two limit N → ∞ and T → 0: • if we take T → 0 first and then N → ∞, we see the physics at T < T K , i.e., the Kondo behavior, • if we take N → ∞ first and then T → 0, we see the physics at T > T K , i.e., Mott-Hubbard behavior, Note that, for even N (= 2M ), the result does not depend on the order to take these two limits, At the temperatures T K < T < ∆ gap , we see the insulating behavior for large N independent of whether N is even or odd. To observe the Kondo behavior at an accessible temperature T , the resonance width T K should be sufficiently large, so that the number of the quantum dots N should be small.
A. Method
In the EOM calculation of Refs.
[1] and [2] the selfenergy has been calculated using a finite cluster of the size N cl decoupled from the total system. Typically, the cluster introduced for the calculation consists of N interacting sites, several noninteracting sites in the left lead N ′ L , and that in the right lead N ′ R ; Using the ground state of the decoupled cluster obtained with an exact diagonalization, a N cl × N cl cluster Green's function g cl has been calculated [1,2]. The EOM selfenergy Σ cl can be obtained from the Dyson equation for the decoupled cluster, Here g 0 cl is the U = 0 Green's function for the decoupled cluster, which describes a discrete energy spectrum. Then, an approximate Green's function which takes into account the rest of the system can be calculated by substituting Σ cl (ω) for the self-energy Σ(ω) defined with respect to the total system. The applicability of this method is determined by the separation of the eigenvalues of the cluster ∆ǫ. This is because Σ cl has been calculated for the discrete energy spectrum, and the correlation effects on the low-lying states below this energy separation have not been taken into account. Therefore, the application of EOM can be justified at high-energies (or high-temperatures) above the energy separation ∆ǫ. The information about the many-body effects on the low-lying states are still missing even though one gets a continuous spectrum after taking into account the connection to the rest of the system. Therefore, if a fine structure is seen in the obtained continuous spectrum, that structure is expected to be real only when the energy scale is consistent with the resolution ∆ǫ. In principle, the energy resolution can be improved systematically by increasing the size of the cluster successively, and the numerical renormalization group (NRG) can be regarded as one of such approaches [9,10]. Since the cluster is introduced just for the approximation and have no physical meanings, the results of the observable quantities must not have a strong dependence on the value of N ′ L and N ′ R .
B. The 1/ω singularity of the EOM self-energy In Refs. [1] and [2], the calculations have been carried out for the cluster with an odd number of sites N cl . At half-filling, the ground sate of the cluster is doublet with the total spin S tot = 1/2. According to the appendix of Ref. [2], the cluster Green's function g cl is defined using the equal-weight sum of S z tot = 1/2 and −1/2 in Ref. [2], while the work in Ref. [1] considers only the cluster state with S z tot = 1/2. Furthermore, in Fig. 9 of Ref. [2], these two procedures are shown to lead quite different results. This means that huge ambiguities exist in the EOM results at low energies. In these two procedures, however, the equal-weight sum seems to be better because it restores the rotational symmetry, so that we will use this convention in the following discussion of the behavior of the EOM self-energy. As typically seen in Fig. 9 of Ref. [2], the results in Ref. [1] have also been affected by the singularity, although the position of the pole might shift due to the absence of the rotation symmetry of the spin in the convention used in Ref. [1].
The presence of the 1/ω singularity of the EOM selfenergy for the cluster of an odd size can be deduced along the discussion similar to that in the section V of Ref. [2]. Using the SU(2) symmetry of the axial charge holding in the electron-hole symmetric case (see Appendix A), the cluster Green's function can be expressed as, × C j (1/2, S g + 1/2; r ′ |0, S g ; r g ) Here S g , J g , and r g are the quantum numbers of the ground state, the energy of which is denoted by E Jg ,Sg (r g ). Note that J g must be zero for U > 0. Furthermore, for U > 0, the excitation energy must be finite ω ± Sg (r ′ ) > 0 in the case of N ≥ 2 as long as the cluster size N cl is finite. Thus, for N ≥ 2, the full Green's g cl (ω) does not have a pole at ω = 0 independent of whether N cl is even or odd. In contrast, the noninteracting Green's function g 0 cl (ω) does has a pole at ω = 0 for odd N cl , while it does not for even N cl . Therefore the EOM self-energy Σ cl (ω), which can be calculated using Eq. (8), must have a pole at ω = 0 for odd N cl because the structure of the energy spectrum of g cl (ω) and that of g 0 cl (ω) are com- pletely different. On the other hand, for even N cl the the energy spectrum of g cl (ω) and g 0 cl (ω) belong to the same category having no zero mode, so that Σ cl (ω) does not have a singularity at ω = 0 and shows a behavior consistent with that of the Fermi liquid Eq. (6). Therefore, the presence of the zero-energy pole can be summarized as Table I (see also Appendix B for N = 1). Note that in the case of "Yes", it does not necessary mean that all the matrix elements have the pole.
Using the EOM self-energy with the Dyson equation which connects the cluster to the remaining system, an approximate Green's function G + 1N can be calculated. Alternatively, the same G + 1N can also be calculated by substituting the N × N partitioned part of Σ cl , which corresponds to the interacting region, into the Dyson equation of N × N form Eq. (2), and then taking the inverse of the matrix. Therefore, if one uses the cluster of odd N cl for calculating the value of G + 1N (ω) at ω = 0, one obtains the result [2]; G + 1N (0) = 0 and g N = 0 for odd N because of the zero-energy pole of Σ cl (ω). In contrast, if one uses the cluster of even N cl in the same situation, the result must be different. This also suggests that the EOM is not suitable for studying the low-temperature properties.
C. Interpretation of the zero-energy pole of Σ cl The oscillatory behavior caused by the cluster size N cl has been seen also in the NRG calculation [9,10]. For even N cl , the ground-state is singlet and the self-energy can be expanded with respect to ω at the Fermi energy ω = 0, which agrees with the local Fermi-liquid theory of Nozières [11] and the perturbation theory of Yamada and Yosida [8]. For odd N cl , the ground state is doublet, and the self-energy has the zero-energy pole. Nevertheless, the energy spectrum of the low-lying states shows the correct Fermi-liquid behavior even for odd N cl [10]. Furthermore, the residue of the zero-energy pole decreases with increasing N cl and vanishes in the N cl → ∞ limit. Therefore, it is quite dangerous to examine the low-temperature properties only by the Green's function obtained with a finite cluster of an odd number sites N cl . Similar oscillation takes place in an antiferromagnetic Heisenberg chain. The spin susceptibility of the chain of an odd number sites has a Curie term. Since the weight of the Curie term is proportional to 1/N cl and vanishes at N cl → ∞, usually the cluster of an even number sites is used for studying the properties in the thermodynamic limit, although the cluster of an odd number sites can also be used if the contribution of the Curie term is subtracted properly.
IV. CONCLUSION
In conclusion, the EOM self-energy obtained for a finite cluster of an odd number of sites has the zero-energy pole, the residue of which must vanish in the limit of large cluster N cl → ∞. The absence of the unitarity-limit transport reported in Refs. [1] and [2] can be explained to be caused by a naive use of this singular self-energy. | 2019-04-14T02:03:50.532Z | 2003-10-07T00:00:00.000 | {
"year": 2003,
"sha1": "9bae176042738dd76108896277d276cd88516211",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "9bae176042738dd76108896277d276cd88516211",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
} |
54990450 | pes2o/s2orc | v3-fos-license | Most Preferred Market Venues for Locally Grown Fresh Produce among Online Shoppers
This paper examines the most preferred market venues online shoppers use when purchasing locally/regionally grown fresh produce. We surveyed 1,205 consumers who made online purchases at least twice within six months prior to participating in this study. We call these consumers “online shoppers”. The study was conducted in the Southern region of the U.S. We collected data using Qualtrics actively managed market research panels and social media such as Facebook, and Twitter. Results indicate that grocery stores are the most preferred market venue online shoppers use to purchase locally/regionally grown fresh produce with a relative probability of 44 percent. Farmers’ markets are the second most preferred market venues with a likelihood of 33 percent. The third market venue is on-farm/CSA programs with relative probability of seven percent. Online markets are currently the fourth most frequented with a likelihood of five percent. 11 percent have no particular most preferred market venue. This analysis is significant to fresh produce growers and marketers. Results suggest new marketing strategies to conveniently make fresh produce accessible among online shoppers. Furthermore, this study is useful for future studies with interests in explaining the preferred market venues for local/regional fresh produce among online shoppers.
Introduction
Internet is shaping the way consumers shop.Online markets are fast expanding as more consumers are increasingly shopping online and businesses are solidifying their presence online.Judith (2012) posited that with the existence of mobile devices, smartphones, and tablets, online shopping has become more convenient.As Kotler and Armstrong (2012), and Rigby (2012) pointed out, nowadays web presence among consumers and organizations is becoming increasingly common.There are various market venues for locally and regionally grown fresh produce.These include farmers' markets, Community Supported Agriculture (CSA) programs, on-farm programs, grocery stores, and online markets.In its research, Neilson Company (2015) indicated that online retailers of fresh produce have tremendous opportunity to do well in the online market environment.With this fast increasing trend, local food marketers are faced with critical questions: Given that consumers are becoming "online shoppers", what are their most preferred market venues for local/regional fresh produce?Will the offline direct-to-consumer market outlets prevail?Are these online shoppers interested in purchasing fresh produce through online markets?
This paper investigates the most frequent market venues that online shoppers are currently getting local/regional fresh produce from.Specific objectives are to: (1) describe consumer characteristics that explain the preferred market venues for local/regional fresh produce among online shoppers, (2) estimate the relative likelihood for online shoppers to purchase fresh produce at a given market venue, (3) explain marginal effects that the consumer characteristics have on each specific preferred market venue for local/regional fresh produce among online shoppers.Online shopping is included in the market venues to estimate its likelihood relative to other market venues.
To our knowledge, this study is the first to look at fresh produce market preferences among online shoppers.Most of the previous research studies directed their attention towards famers' market consumers.Examples of such studies include Conner et al. (2010), Ruelas et al. (2012), Abello et al. (2012), Gumirakiza et al. (2014), Zepeda et al. (2012), Freedman et al. (2014), Racine et al. (2013), and Kraschnewski et al. (2014).Other studies like Curtis et al. (2015), Meyer (2012), Conner et al. (2010), and Woods and Troppy (2015) focused on CSA programs.This study sets a basis for future analyses that will be interested in tracking changes in the preferred market venues for local/regional fresh produce among online shoppers.This analysis is significant to fresh produce growers for it provides information regarding most proffered market venues for their products among the increasingly common type of consumers.A successful fresh produce grower would be interested in coping with consumer trends.Likewise, agricultural marketers in the local food movement; including fresh produce retailers will find this study significant.Results will help the marketers in adopting new marketing strategies to conveniently make fresh produce accessible to online shoppers.
Literature Review
Eating locally grown fresh food is one of the popular recommendations in healthy eating and supporting community farm businesses.There are various ways to obtain locally grown food, such as farmers' markets, CSA programs, on-farm opportunities, online, and in grocery stores.All of these market venues provide consumers with fresh produce.Existing literature provides characteristics of consumers who attend these market venues.But, very little is known about consumers in the online markets.Mirosa et al. (2012) indicated that middle-aged educated individuals are the main consumers of locally grown food.Studies by Curl et al. (2013), andOnozaka et al. (2010) got the same results and added affluent households and white females as other significant factors.A study conducted by Racine et al. (2012) shows that those who lived in rural areas were more likely to purchase locally grown food.Their study also showed that individuals who ate more than five servings of fresh produce daily were more likely to purchase locally grown foods.Feldmann et al. (2015) found that those who lived in close proximity to local food outlets were more likely to buy locally grown foods.Both studies support that many consumers visit farmers' markets out of convenience.
Farmers' markets are one of the most common ways individuals obtain locally grown food.They bring the farm into the community and provide a way for many people to obtain fresh local food.In an effort to increase diversity of consumers of farmers' markets, many markets now accept SNAP (supplemental nutrition assistance program), and senior vouchers.However, in a study that Wetherill et al. (2015) conducted in Oklahoma revealed that several individuals are unaware.Consequently, low-income individuals are reportedly not part of many farmers' market consumers (Crow et al., 2011;Murphy, 2011;Rice, 2015).Nevertheless, Ruelas et al. (2012) showed a correlation of low-income consumers shopping at farmers' markets in Los Angles.
In recent years many studies identified characteristics of families who attend farmers' markets.A study conducted by Cassia et al. (2012) discovered that the majority of farmers' market consumers are those with only two people in the household.Additional studies found that small families, those composed of two to four persons, are more likely to purchase locally grown foods from farmers' markets (Arrington et al., 2010;Gumirakiza et al., 2014;Pascucci et al., 2011).Their studies suggest that those with smaller families can affordably and conveniently provide local produce for their families at farmers' markets.This seems to contribute to the high percentage of small families that take part in farmers' market outlets.
Community Supported Agriculture (CSA) programs fit well for working families.They allow for consumers to obtain locally grown food without the hassle of growing their own.CSA programs function through annual membership fees that consumers pay to growers in advance.When the produce is ready for consumption, members pick up their share of food at the farm or at any other specified place.CSA consumers have been found to be uniform.Studies have proved that most CSA participants are affluent, middle aged women (Lang, 2010;Pole et al., 2015;Vasquez et al., 2016).Many CSA consumers view eating locally grown foods, as well as healthy foods important.It is noteworthy to mention that Okoye et al. (2014) pointed out that many CSA farm participants do not allow consumers with SNAP benefits to use them.This limits access to fresh produce among such consumers.
Studies about local fresh produce available in grocery stores have varying definitions of the word "local" based on the size of the grocery store.Dunne (2011) indicated that the bigger the store the more broad the definition of local is.Blanch et al. (2011) found that barely 27 percent of grocery store shoppers would also use farm-to-consumer market outlets at least weekly.When it comes to purchasing fresh produce from store or farm-to-consumer venues, their study found no significant differences among gender, race/ethnicity, education, and annual household income.A study by Martinez et al. (2010) concluded that older affluent females were those more likely to purchase locally grown foods grocery stores.There are consumers who enjoy visiting farm and/or engaging in agritourism events.Some famers offer market opportunities through roadside stands and/or U-Pick.Others organize events allowing consumers to tour farms and learn more about locally grown food.According to Govindasamy et al. (2014) those who participate in these outlets tend to be local community members that are well educated.Thorp et al. (2015) reported that those who participate in these outlets are tourists, not necessarily community members.
Data Collection Process
This study uses data collected from a stratified randomly selected sample of 1,205 online shoppers.The target population of interest for this study is consumers who have made at least two online purchases of any kind within six months prior to participating in this study.Hence, they are referred to as being online shoppers.This study targeted online shoppers that reside within the Southern region of the U.S. The states that are in this region are: Arkansas, Delaware, Florida, Georgia, Kentucky, Louisiana, Maryland, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, Virginia, West Virginia, and the District of Columbia (U.S. Census Bureau, 2016).
In this study, we used the Qualtrics software to create the survey.The software has integrated capabilities to make possible the accurate tracking, profiling, monitoring of responses for each respondent.It also allowed us to use the design survey questions using advanced branching logic, randomization, question block presentation, and question timing.The benefit from these features was to avoid possible bias that could arise during survey taking.To ensure that respondents are in fact paying attention to each question included in the survey, we added questions requiring respondents to think and provide a correct answer.Those who gave incorrect answers were automatically excluded from the survey.
The data were collected between March and July, 2016.The survey was emailed to respondents that were randomly selected from online shoppers within the Qualtrics actively managed market research panels and to those using social media such as Facebook, and Twitter.Qualtrics is a professional survey software provider that offers sophisticated and advanced online data collection tools combined with respondent panels.Qualtrics partners with over 20 online panel providers to supply a network of diverse and quality respondents to meet specific sample requirements.To ensure quality and data validity, every IP (Internet Protocol) address/location was checked and a sophisticated digital fingerprinting technology was used.
The survey questions that are relevant to this study include six market alternatives for local/regional fresh produce and consumer characteristics.The former is used as a dependent variable while the later are explanatory variables.The six market alternatives were: farmers' markets, CSA programs, on-farm, online markets, grocery stores, and the "none" option.Note on the "grocery store" alternative: respondents were told to check this option only if they read labels before buying to make sure the produce is fresh and grown locally (within state boundaries) or regionally (within a 400 mile-radius from respondent's location).We found that only 1 percent of the online shoppers use CSA programs as the most frequent when purchasing local/regional fresh produce.For this reason, CSA and on-farm were combined together.Consequently, the dependent variable consists of four market alternatives: farmers' markets, on-farm/CSA programs, online markets, grocery stores, and the "none" (no preferred market venue) option.
Model Specification
In order to estimate the relative probabilities of choosing a particular market alternative to be the most frequent, a multinomial logistic model is appropriate.Chan (2005) indicated that the structure of this model allows predicting the probability that the j th alternative in a set of more than two alternatives is chosen.Chan also posited that this model is appropriate when a systematic utility is modeled in terms of characteristics of respondents.He further urged that this model is applied when examining relative probabilities of alternatives that are provided in an unordered way.This analysis relies on three main assumptions.First, online shoppers are rational.This means that they strive to maximize utility/satisfaction and that their preferences are both complete and transitive (Mas-Colell et al., 1995).Second, the Independence from Irrelevant Alternatives (IIA) holds true.This means that adding a new market alternative to the set of alternatives does not affect the relative odds among the existing choice set (Train, 2009).Finally, we assume that a consumer with a finite set of choice alternatives will select the one that he/she believes gives him/her the maximum amount of utility.A consumer's utility derived from a choice set is specified as a linear function of the consumer's characteristics, along with an error term.The situation is that an individual online shopper i is presented J alternatives and is assumed to choose one he/she believes provides the highest utility.The utility function takes the form, Where, V ij is the deterministic component of the utility and ε ij is the random component.
We assume that the random component term is independently and identically distributed (iid) according to an extreme value F(ε ij ) = exp(-exp(-ε ij )) so that logistic model becomes appropriate (Kennedy, 2008).As in Onozaka and Thilmany-McFadden (2011), we assume a linear-in-parameter utility functional form for the deterministic component of utility.
The choice of one of the J unordered market alternatives is driven by a latent variable or indirect utility.The indirect utility for individual i choosing an alternative j is, Where, X ij is a vector of characteristics of the chooser.The parameter β is to be estimated and differs across alternatives.The μ ij is the disturbance that account for unobserved factors.As researchers, we do not observe individual's utility.Instead, we observe the choices that make implying that they choose an alternative that maximizes their utility.The observed choice y i of an individual i is, (3) The probability (P) that an individual i chooses alternative j is expressed as: (4) The β's are identified by setting the β j* = 0 for one reference alternative; j*.The "None" option is the reference category in this analysis.From Equation (4), we generate a direct interpretation of the parameter estimates as follows: (5) which reduces to, in comparison with the reference outcome j*.According to Schmidheiny (2007), a positive parameter β jk for a continuous variable means that the relative probability of choosing j increases relative to the probability of choosing the reference alternative j*.Dummy variable effects are measured and interpreted as the probability difference between X ij values of zero and one.The marginal effect of an independent variable X k on the choice probability for an alternative j is given by, The null hypothesis is that each independent variable has no impact on the relative probability of choosing farmers' markets, on-farm, online markets, and grocery stores as the most preferred market venue for purchasing local/regional fresh produce.That is ; j = 1, …, J for K explanatories and J market alternatives.The alternative hypothesis is that each independent variable has a significant impact on the relative probability of choosing farmers' markets, on-farm, online markets, and grocery store as the most preferred market venue for purchasing local/regional fresh produce.That is; ; j = 1, …, J for K explanatories and J market alternatives.
Descriptive Statistics of Independent Variables
Thirteen consumer characteristics are included in this analysis.Table 1 shows mean values associated with each of the characteristics.Conveniently, we displayed the statistics per each of the five categories of the dependent variable.This was accomplished by running the following Stata command: tabstat Age Urban Female Married Locavore Education GovAssistance FMSpendperVisit InterestLevelLocalFP MonthlySpendFreshProduce, by(Markets) stat(mean) nototal f(%9.2f)."Locavore" is a dummy variable that represents those shoppers who eat more fruits and vegetables regularly believing that they help address their dietary concerns."GovAssistance" is another dummy variable representing respondents who participate in either food stamps, WIC, and/or senior nutrition assistance programs.The "FMSpendperVisit" represents the amount of money respondents spend or would spend per one visit at a farmers' market.The "InterestLevelLocalFP" represents a 5-likert scale of levels of interests respondents have in locally grown fresh produce.Lastly, the "MonthlySpendFreshProduce" represents the average amount of money respondents spend on fresh produce per month.These independent variables are included in the model because we believe they reflect major consumer characteristics that are relevant and important in explaining the primary preferences.These variables are also a result of a model specification we conducted to ensure we use variables that are adequately able to make the overall model significant.Those that were insignificant across the five outcome responses were removed from the model.Overall, these statistics indicate that an average online shopper is 47 years old, completed a 2-year associate's degree, and is interested in attending farmers' markets.This average shopper would spend $38.39 at a farmers' market per one visit and spends $57.81 on locally/regionally grown fresh produce per month.Among those who participated in this study, 76 percent live in areas they consider to be urban, 62 percent are females, 56 percent are married, 72 percent are locavores, and 15 percent are recipients of some form of food assistance.Statistics about each market venue are available in Table 1.For example, an average online shopper whose most preferred market venue for local/regional fresh produce is farmers' market is 47 years old, completed a 2-year associate's degree, and is interested in attending farmers' markets.He/she would spend $40.28 at a farmers' market per one visit; a little bit more than what an overall average online shopper would spend.This shopper spends an average of $62.09 on locally/regionally grown fresh produce per month.Among these farmers' market funs, 73 percent live in areas they consider to be urban, 59 percent are females, 59 percent are married, 79 percent are locavores, and 10 percent are recipients of some form of food assistance.We found that individuals who are in mid-fifties favor online markets while those in fifties like getting their fresh produce through on-farm/CSA programs.This indicates that those who are younger and more experienced using newer technology are the individuals are more likely to purchase fresh, local produce from the online markets.
Regression Results
The dependent variable for the multinomial logistic model regression consists of five discrete categories.These categories are the four market venues for locally grown fresh produce: (1) Farmers' markets, (2) On-farm and CSA programs, (3) Online markets, (4) grocery stores, and (5) the "none" option.Results in Note.The *, **, and *** denote significance at the 10%, 5%, and 1% level, respectively.The MNL model above compares the four major market outlets where people can purchase fresh, local produce.None is the reference category.Thus, the coefficient estimates in the MNL model (Table 2) are interpreted in comparison to this reference category (no preferred market venue).A positive coefficient estimate shows that an increase in the variable is associated with an increase in the relative probability of the corresponding outcome.Specifically, age has a positive significant effect on the likelihood of favoring the on-farm/CSA markets, but an inverse effect on the likelihood for the online markets.Compared to rural residents, urban respondents are less likely to favor farmers' markets and on-farm/CSA programs, but more likely to favor grocery stores.Likewise, online shopper females are less likely to opt for farmers' markets, but more likely to prefer grocery stores.Married online shoppers show significant preferences for grocery stores.Interestingly, locavores significantly prefer all four venues.Results for other variables are presented in Table 2 above.
The relative probabilities and marginal effects for each of the market venues are shown in Table 3.The marginal effects were computed using the following Stata command: mlogit PrimaryMarketLocalProduce Age Urban Female Married Locavore Education GovAssistance FMSpendperVisit InterestLevelLocalFP MonthlySpendFreshProduce.Then a post-estimate command: margins, dydx(*) for each of the five outcomes was applied to estimate marginal effect of variables in varlist.In other words, we used the following: margins, dydx(*) predict(outcome(1)), margins, dydx(*) predict(outcome(2)), margins, dydx(*) predict(outcome(3)), margins, dydx(*) predict(outcome(4)), margins, dydx(*) predict(outcome(5)).The relative probability that a consumer shops for fresh local produce at a farmers' market is 33 percent.There is almost 11 percent likelihood that those online shoppers who receive any type of food-related government assistance would be more likely to prefer purchasing locally grown fresh produce at farmers' markets.This can be attributed to the fact that there are some incentives to attract those consumers on the food-related government assistance at farmers' markets.We furthermore report a five percent increase in the probability that someone would shop at a farmers' market if they had an interest in local food.There are many programs, such as Kentucky Proud, which try to increase consumer knowledge and concern about locally grown fresh produce.There is a seven percent fall in probability of shopping at farmers' markets for females.Locavore are five percent more like to favor farmers' markets as their primary venue for local fresh produce.On the other hand, this study found that females are less likely to shop for fresh local produce at farmers' markets.Likewise, those in urban areas are seven percent less likely to purchase fresh, local food from farmers' markets.Those who live in rural areas are more likely to purchase local food from farmers' markets.
The relative probability that a consumer shops for fresh local produce on a farm or through a CSA is seven percent.
Ten more years old leads to two percent more likely to participate in on farm/CSA programs to obtain fresh, local produce.There is also an increased likelihood that someone participating in on farm outlets or CSA programs has an interest in local food products by two percent.Those who are interested in local food normally seek out ways to obtain it.Those who live in urban areas are four percent less likely to shop on farms or through CSA programs than their rural counter parts.Additional level of education also decreases the probability of shopping on farms or through CSA program by one percent.
The relative probability that a consumer shops for fresh local produce through an online market is five percent.The interest in local fresh produce has a significant positive influence on the preference to purchase produce from the online markets.One level of interest increases the likelihood of purchasing through the online markets by almost three percent.Interestingly enough, we found that those online shoppers, who receive some sort of food-related government assistance, are two percent more likely to consider online marketplace as their primary market venue for local/regional fresh produce.We further found that $1,000 additional annual income translates into a 20 percent likelihood to prefer to primary purchase fresh produce online.This suggests that as people's income increases, it is a good news to online retailers in the local fresh produce industry.Another variable with some effect is age.An increase in age decreases the likelihood of shopping online for fresh produce by roughly 0.2 percent.Likewise, females are less likely to primarily view the online marketplace as their primary market venue for local/regional fresh produce.In fact, a female is two percent less likely to do so.
The relative probability that a consumer shops for fresh local produce at a grocery store is 44 percent.This result indicates that grocery stores are the most frequent venues online shoppers use to purchase locally/regionally grown fresh produce.In this study married females have an increased likelihood to shop for local fresh produce in grocery stores.Females are eight percent more likely to shop at grocery stores for local fresh produce while married shoppers are four percent more likely to do so.Living in urban area increases the likelihood of shopping for local/regional fresh produce at grocery stores by seven percent.Another increase in probability is with the amount spending on produce each month.One dollar an individual spends on produce leads to one percent more likelihood to shop at grocery stores.Finally there is a negative effect concerning the interest level in fresh produce.With an interest in fresh, local produce there is a four percent decrease in the probability that an individual will shop for fresh, local produce at a grocery store.The relative probability that an online shopper has no preferred market when purchasing fresh produce is 11 percent.
Conclusion
Online shopping has gained great popularity in today's society.It is only logical that people are starting to buy their food online.Having an outlet to buy fresh, local produce without having to leave the house and avoiding the crowds is a huge benefit for many consumers.In this study, we examined the most preferred market venues that online shoppers use when purchasing locally/regionally grown fresh produce.We targeted consumers who made at least any two online purchases within six months prior to participating in this study (herein referred to as "online shoppers").We randomly selected 1,205 online shoppers residing within the Southern region of the U.S participated in this study.We collected data using Qualtrics actively managed market research panels and social media such as Facebook, and Twitter.
Key results indicate that grocery stores are the most preferred market venue online shoppers use to purchase locally/regionally grown fresh produce.The relative probability associated with this situation is 44 percent.Farmers' markets are the second most frequent market venues with a relative likelihood of 33 percent.The third market venue is on-farm (road stands, you-pick-your own, and agritourism) and CSA programs with a relative probability of 7 percent.Online market is currently the fourth most frequent with a relative likelihood of 5 percent.11 percent of the online shoppers have no particular market venue they consider as primary.
Adoption of online shopping for locally/regionally fresh produce is still low.When compared to pace of shopping online for other items, this is concerning.Local food growers and/or marketers should find ways to make shopping online for fresh, local produce a success.We also suggest that further studies aiming at explaining reasons behind this be conducted.Marketing strategies by those who promote farmers' markets should target online shoppers.Grocery stores should do likewise.Growers of local/regional fresh produce are encouraged to revise their marking strategies based on the consumer characteristics that are found to increase the likelihood of preferring their specific market venue.Finally, we encourage future research studies to explain clearly reasons why few consumers are participating in online markets for locally grown fresh produce.
Table 1 .
Consumer characteristics by market venues for local fresh produce
Table 2 .
Coefficient Estimates from the Multinomial Logistic Model
Table 3 .
The Marginal Effects from the Multinomial Logistic Model | 2018-12-05T05:35:02.205Z | 2017-09-13T00:00:00.000 | {
"year": 2017,
"sha1": "eaf45013fbefba9860ad765cdaa565b6022ce352",
"oa_license": "CCBY",
"oa_url": "https://www.ccsenet.org/journal/index.php/jas/article/download/68576/38450",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "eaf45013fbefba9860ad765cdaa565b6022ce352",
"s2fieldsofstudy": [
"Business",
"Economics"
],
"extfieldsofstudy": [
"Business"
]
} |
118863312 | pes2o/s2orc | v3-fos-license | Transition from Solitons to Solitary Waves in the Fermi-Pasta-Ulam Lattice
In this paper, we study the smooth transition from solitons to solitary waves in localization, relation between energy and velocity, propagation and scattering property in the Fermi-Pasta-Ulam lattice analytically and numerically. A soliton is a very stable solitary wave that retains its permanent structure after interacting with other solitary waves. A soliton exists when the energy is small, and it becomes a solitary wave when the energy increases to the threshold. The transition could help to understand the distinctly different heat conduction behaviors of the Fermi-Pasta-Ulam lattice at low and high temperature.
I. INTRODUCTION
The history of solitary waves can be traced back to 1834, when it was first discovered by Russell on the Union Canal in Scotland [1]. Several decades later Korteweg-de Vries equation was provided which explained the phenomenon mathematically [2]. More than a century later, Fermi, Pasta and Ulam (FPU) investigated the fundamental problem of energy equipartition and ergodicity in nonlinear systems and discovered FPU recurrence [3], which has led eventually to the soliton concept [4]. Since then solitary waves have been studied in various different fields of physics like solid-state physics, quantum theory, nonlinear optics, fluid dynamics, biophysics, etc. [5,6].
Distinguishing between solitons and solitary waves is sometimes of importance such as studies on energy transport. Solitons are localized solutions of integrable equations, and they are very special types of solitary waves which do not alter form and only have a phase shift after collision with other solitons. Solitary waves are associated with nonitegrable equations, and they are deformed and energy exchange may occur when solitary waves interact with one another [7]. In the long wavelength approximation, the FPU-α and FPU-β lattices lead to the KdV and modified KdV equations, respectively, in which the exact soliton solutions exist. The stability of soliton in the FPU lattice at low energy is demonstrated [8,9,10,11]. At high energy level, it becomes solitary wave, which makes attribution to heat conduction [12,13] because of the inelastic scattering between each other. Up to now, very little work has been done on the study of changing from solitons to solitary waves in the FPU lattice.
In this work, we focus on the transition from solitons to solitary waves in the FPU lattice. Previously in numerical work solitons are constructed by using analytical soliton solution directly [14] or starting with a special initial condition [4]. Here we obtain solitons in the FPU lattice for the first time with momentum excitation, which is the widely used method to excite solitary waves [12,13,15,16,17]. Soliton transforms to solitary wave when energy increases larger than the threshold. Solitons and solitary waves are different in relation between energy and velocity, propagation and scattering behavior. Solitons and solitary waves maintain shape and energy when travelling; kinetic and potential energy remain unchanged for solitons while they are varying periodically for solitary waves. The scattering of solitons is elastic. When two solitary waves collide with each other, both of them lose energy or they exchange energy, and extra wave packets are excited. The solitary wave properties become similar to those in the pure anharmonic lattice when energy tends to infinity. Finally the different heat conduction behaviors of the FPU lattice at low and high temperature are discussed with the concept of solitons and solitary waves.
II. ANALYSIS IN THE WEAKLY NONLINEAR LIMIT
We consider the Hamiltonian for the FPU lattice where p n denotes the momentum and q n denotes the displacement from equilibrium position for the nth atom. Dimensionless units are used, such that the masses, the linear and nonlinear force constants, and the lattice constant are all chosen to be unity. The parameters α and β are the harmonic and anharmonic force constant, respectively. Here α = 1 and β = 1 for the FPU lattice throughout the paper.
The equations of motion of the FPU system arë In the long wavelength limit, the continuum approximation should be applicable. The displacement q n±1 of the (n ± 1) th lattice is expanded as where q (x, t) = q n (t), and x = n. Substituting Eq. (4) into Eq. (3) and neglecting higher order terms, we obtain the continuous counterparts of Eq. : We shall be interested in right-going waves, and introduce the new slow variables and define where ε is a formal small parameter. Substituting Eq. (6) and (7) into Eq. (5) and keeping terms in the order of ε 4 , we obtain the following evolution equation for the function ψ = ∂ϕ/∂ξ: which is the well known modified Korteweg-de Vries equation. Eq. (8) is derived from the FPU chain [18,19] and it results in the soliton solutions [14,20,21]: where c is constant. Eq. (9) and (10) are valid if the higher derivatives are neglected. It can be achieved if the following is fulfilled.
The solutions for q n and p n are kink and bell shaped soliton respectively and the soliton velocity is achieved: The soliton energy is The first term of Eq. (13) denotes kinetic energy of soliton and the second term is potential energy. In the long wavelength limit, potential energy is very close to kinetic energy for soliton at low energy, so soliton energy can be expressed as follow: Since soliton energy is unchanged as soliton propagates, substituting Eq. (10) at the time t = 0 we obtain Replacing the sum by an integral in Eq. (15) due to continuum approximation and using Eq. (11), we finally arrive to Then we can get the relation between soliton energy and velocity from Eq. (12) and (16) in the case of soliton energy E ≪ 1 So soliton velocity is close to acoustic velocity due to the small energy.
III. HARMONIC LIMIT AND ANHARMONIC LIMIT
We obtain the harmonic lattice in the absence of the quartic term, i.e., α = 1 and β = 0 for Eq. (1) and (2). The system is integrable and can be analytically solved. Here acoustic velocity is 1.
It is the pure anharmonic lattice with α = 0 and β = 1. In this case the excitation, propagation and interaction of solitary waves have been studied in Refs. 13,15, 22, 23. The system has an excellent scaling property, from which we can obtain the relation between energy E and velocity v of a solitary wave: a = 0.68198 from numerical simulations [13].
IV. TRANSITION FROM SOLITONS TO SOLITARY WAVES IN THE FPU LATTICE
Solitary waves can be excited by momentum kicks on the lattice. We apply a kick p 1 on the first lattice at t = 0 for static system under the free boundary condition. In Fig. 1 we plot p i versus i after a short time (t = 500) for three kinds of lattices: the harmonic chain at p 1 = 0.7 (a), the pure anharmonic chain at p 1 = 0.7 (b), the FPU chain at p 1 = 1.5 (c) and the FPU chain at p 1 = 0.7 (d). In the harmonic chain the amplitude of the wave profile decreases while the width increases with time continuously. In the pure anharmonic chain a solitary wave is excited accompanied by the tail, which is made of several small solitary waves shown in the inset of Fig. 1 (b) [13]. In the FPU chain, at first the wave front is connect with the other low amplitude excitations. If the imparted kick is large enough, after a certain short time, the wave front separates form the tail, because it moves faster. It propagates forward and becomes a solitary wave, just like that in Fig. 1 (c). If the kick is small, the excited wave packet looks like that in the harmonic chain in a short time ( Fig. 1 (d)). But actually It will show the essential difference after a long time. Consider the lattice which is so long that the wave packet won't reach the other end of the lattice in the long observing time. As the first wave pulse travels, its width becomes wider and its peak becomes lower, as shown in Fig. 1 (e) at the time of t = 10 4 , still like that in the harmonic chain. The speed of the first pulse is calculated and it is a little larger than the acoustic velocity of 1, and so it is not a linear wave. The first pulse travels a little faster than the second pulse, so as long as time is long enough, the first pulse will separate from the second one. Fig. 1 (f) displays the separation of the first pulse from the others at t = 2 × 10 5 , and from then on its shape and energy do not changed any more, and it becomes a solitary wave. We show this solitary wave in the momentum space and configuration space in 10) and (9) in Fig. 2. and the excellent fit shows that the solitary wave excited by the small kick p = 0.7 is a soliton indeed.
We present the width of solitary waves at different energy in Fig. 3 (a). The energy of solitary wave is determined by the kick strength and can be calculated with Eq. (13), where the sum is computed over the lattices on which the solitary wave is localized. At the energy of E < 0.1 the width of solitary wave is relatively wide and exhibits a power-law decay with energy. At E > 0.1 the decay of width becomes slower and the localization of the solitary wave becomes strong as energy increases. At E > 50 solitary wave is always localized on the 4 lattices, the same as that in the pure anharmonic lattice. Fig. 3 (b) shows the relation between energy and velocity of solitary waves in the FPU lattice. If the energy E < 0.1, the width of solitary wave is large and the continuum approximation is applicable, so the relation is consistent with Eq. (17) [see the inset in Fig. 3 (b)] and the solitary wave is a soliton. According to Eq. (17), soliton velocity is only a little larger than the acoustic velocity because its energy is so small. If E > 0.1, the solitary wave is localized on a few lattices and the long wavelength limit is not satisfied. This leads the deviation from the fitted curve of Eq. (17). So with the increase of energy, the nonlinear part in the interaction potential of Eq. (2) becomes more and more obvious and soliton turns into solitary wave. If E > 50, velocity increases with energy as a power-law function, asymptotic to v = aE 1/4 of Eq. (18) in the pure anharmonic lattice. In this case, the linear part in Eq. (2) is negligible while the nonlinear part is significant, and the property of solitary wave becomes similar to that in the pure anharmonic lattice.
In the following we discuss solitary wave scattering dynamics. Scattering property is an important key to determine whether it is a soliton or a solitary wave. Soliton scattering is elastic while solitary wave scattering is inelastic. Fig. 4 show the process of head-on collision of a pair of solitary wave a and b. The energy of a is one half of that of b, i.e., E a /E b = 1/2. Before collision each solitary wave maintains shape and energy, as shown in Fig. 4 (a), (c), (e) and (g). After collision the two solitary waves pass through each other, denoted as a ′ and b ′ , and the scattering behaviors are significant different in the following four cases. In Fig. 4 (b), the two solitary waves a and b with E a = 0.01 and E b = 0.02 do not change their shapes and energy and it is an elastic scattering. So a and b are solitons. In Fig 4 (d), solitary wave a with E a = 0.05 is unchanged, i.e., a ′ is the same as a, while solitary wave b with E b = 0.1 is scattered to solitary wave b ′ with a tapered tail. The energy of solitary wave b decreases and In Fig 4 (f), both solitary wave a and b with E a = 0.1 and E b = 0.2 are scattered. They are scattered to a ′ and b ′ with tapered tails respectively. The energy of a and b decrease.
is the energy of the tail of a ′ . In Fig 4 (h), the scattering behavior of solitary waves with E a = 0.3 and E b = 0.6 is quite different from that in Fig 4 (f). Energy exchange takes place and the scattering effect is enhanced. The large solitary wave b loses energy and the small one a gains energy, and extra wave packets are excited. So the collision process varies from elastic to inelastic and then the scattering effect becomes stronger with the increase of the energy of solitary wave.
Next we calculate the energy scattering rate of solitary wave to describe the scattering behavior quantitatively. The energy of a solitary wave is the sum of kinetic energy and potential energy. In the FPU lattice the soliton energy is small and its width is wide. Kinetic energy and potential energy don't varies with time while soliton travels, as shown in Fig. 5 (a), where soliton energy is 0.01. So after collision, the scattering result does not change with time delay of the two solitons, labeled as phase δ [see Fig. 5 (b)]. As energy increases, soliton becomes solitary wave. Solitary wave is only localized on a few lattices. So due to the discrete nature of the lattice, during a solitary wave is travelling on the lattice, its total energy is unchanged while its kinetic energy and potential energy vary with time periodically, as presented in Fig. 5 (c), in which solitary wave energy is 0.3. The variation period T is the time that the solitary wave spends on moving the distance of one lattice constant. Here the lattice constant is unity, and then T = 1/v. This leads the scattering results also vary with phase δ periodically [13,23] [see Fig. 5 (d)]. For practical application, statistical properties are more significant than the prompt values. Then we investigate the average energy scattering rates, < E ′ a − E a > /E a , < E ′ b − E b > /E b , and the average rate of the energy "loss", < ∆E > /(E a + E b ). Fig. 6 present the average energy scattering rates and the average rate of the energy "loss" with the increase of energy when E a /E b is fixed as 1/2. When the solitary wave energy E < 0.1, the scattering rate is zero and the scattering is elastic. The scattering process is similar to that in Fig. 4 (a) and (b). The shapes and energy of the two solitary waves are unchanged. So the solitary wave with energy E < 0.1 is a soliton. When 0.1 < E < 0.3, the energy scattering rates of a and b are both less than 0 and the scattering is inelastic. The scattering process can be shown by Fig. 4 (e) and (f) approximately. Both of the two solitary waves are scattered into smaller ones with tapered tails and there is no energy exchange between them. The scattering effect is quite weak as the scattering rate is around 10 −3 . Soliton is transformed into solitary wave with the increase of energy. When E > 0.3, the energy scattering rate is more than 0 for the small solitary wave a and is less than 0 for the large solitary wave b, and the scattering is inelastic. The scattering process is like that in Fig. 4 (g) and (h). There is energy exchange between the two solitary waves, i.e., large solitary wave b loses energy and small one a obtains energy on average, and extra wave packets are excited. The stars in Fig. 6 are the scattering results for the pure anharmonic lattice. Any pair of solitary waves with the same ratio E a /E b has the same scattering rates due to the scaling property [13]. In the FPU lattice the scattering effect increases with energy and the scattering results approach those in the pure anharmonic lattice when E → ∞.
V. CONCLUSION AND DISCUSSION
In summary, we explore the transition from solitons to solitary waves in the FPU lattice. Solitons and solitary waves can be excited by momentum kicks. If energy is smaller than 0.1, the excitation is a soliton, which is consistent with analytical results. Solitons maintain shape and energy, and kinetic and potential energy remain unchanged while travelling. The soliton scattering is elastic; solitons do not alter shape and energy after collision with others. If energy is larger than the threshold of 0.1, soliton transforms to solitary wave. When a solitary wave propagates, kinetic energy and potential energy are varying periodically, while total energy does not change. The solitary wave scattering is inelastic; both of solitary waves lose energy, or large solitary wave loses energy and small one gains on average, and extra wave packets are excited. If energy is large enough, the properties of solitary waves tend to those in the pure anharmonic lattice.
The heat conduction behaviors of the FPU lattice are quite different at low and high temperature. The temperature gradient can't be formed like the harmonic lattice at low temperature, while it can be set up at high temperature [12,24,25]. We discuss it with the concept of solitons and solitary waves. In molecular dynamics simulations, the temperature at position n is defined as T n = < p 2 n > /2 and the role of a thermostat can be interpreted as a series of kicks on the ends of the lattice. When the thermostat temperature is low, solitons are excited. The soliton scattering is elastic and there is no energy exchange, and so there is no temperature gradient formed. In the case of high thermostat temperature, solitary waves are obtained. The solitary wave scattering is inelastic and energy exchange takes place, and so temperature gradient is set up. | 2016-09-30T16:31:09.000Z | 2016-09-30T00:00:00.000 | {
"year": 2016,
"sha1": "a6bd19a0dc22afb04ae77ce3b18a15d4c0ec5fba",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "a6bd19a0dc22afb04ae77ce3b18a15d4c0ec5fba",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
} |
14046530 | pes2o/s2orc | v3-fos-license | An Unbounded Nondeterministic Model for CSP-like Languages
The main contribution of this paper is the introduction of unbounded nondeterminism into CSP-like languages with specification, which increases the expressive power of the specification language. This is achieved by extending the finite message set of such mixed languages to an infinite one. A denotational semantics and a refinement order are defined for such a language and the CSP constructors are proved to be monotone on the specification space and continuous on the process space, which are needed for a simple treatment of iterations on both spaces of the mixed CSP-like language
Introduction
There already exist a number of models for CSP-like languages, e.g., [6,7,10,11,12].To exclude unbounded nondeterminism (or discontinuity of language constructors), the communication alphabets, i.e., the set of channels and messages, are restricted to be finite sets.This restriction becomes a nuisance especially when the language is extended to include specification.For example, with an infinite message set, the Fermat Theorem can be specified as follows (we know now such a specification is infeasible): (which can also be specified in CSP using an infinitary parallel composition constructor [5].) var n; a; b; c : N n; a; b; c : true; a n + b n = c n ^2 n ^0 a; b; c : Research has been done in recent years to handle unbounded nondeterminism in the CSP theory, e.g., [5,11].Unbounded nondeterminism refers to the ability of a process to choose locally one from an infinite set of possible behaviours.The difficulties encountered with unbounded nondeterminism in CSP is the loss of the fixed-point theory for defining recursions.Roscoe first tackled this problem in [11] by introducing a complete but coarser partial order on the failures-divergence model in [3].The resulting model can successfully model a process which will, on its first step, nondeterministically choose any integer, but cannot tell between a process which can communicate any finite number of a's and one which may also choose to communicate an infinite number.To overcome this problem, the failures-divergence model is extended to include infinite traces so that any CSP process is represented by F ; D; I where F is its failures (still with finite traces), D is its set of (finite) divergence traces and I is the set of infinite traces it can communicate.The refined mode can model unbounded nondeterminism properly but with the loss of completeness and monotonicity.To cure this, advanced mathematics was used resulting in more difficult proofs and clumsy mathematical treatment.
In [5] Kumar and Pandya extended CSP to include an infinitary parallel composition operator to increase the expressive power of CSP.The well-formedness of the operator in the failures-divergence model was established under the syntactic restriction that an event can occur only in the alphabet of finitely many processes.Therefore, unbounded nondeterminism is excluded without the sacrifice of the elegance of the CSP theory.
In CSP's failures-divergence model, processes can have infinite alphabets without violating the continuity of CSP constructors if no infinitary nondeterministic choice operator is used and the hiding operator does not hide an infinite 2nd Irish Workshop on Formal Methods, 1998 set of communication events.In this paper we follow the same route to extend the finite message set of a CSP-like language to an infinite one to increase the expressive power of the language extended with specifications.We do not, however, want to introduce unbounded nondeterminism into the CSP-like language itself because it cannot be implemented and also causes discontinuity of the program constructors and thus may need some advanced but, in practice, cumbersome techniques such as transfinite induction in [2,9].To avoid unbounded nondeterminism, some syntactic restriction is placed on the hiding operator which can only be used to hide internal channels of a concurrent process, which is the case for most CSP-like languages.Otherwise, unbounded nondeterminism is unavoidable.
For example, with an infinite message set, the following program is equivalent to choosing any value for x.This, therefore, introduces unbounded nondeterminism into the language, resulting the discontinuity of some of the language constructors.c?x nc: By restricting the hiding operator to internal channels, we prove that unbounded nondeterminism can be avoided in the CSP-like language without specification and iterations (also recursions) can be treated in the usual way.
In this paper we follow Morgan's example [8] and treat specifications like program statements, so that there is a unified framework for modelling specifications and programs [7].Doing so, an infinite alphabet will increase the expressive power of the specification language.However, the extended model including the specification statement contains unbounded nondeterminism, resulting in discontinuity of some of the CSP constructors.We prove that all the CSP-like language constructors are monotonic with respect to the refinement order on the extended model.Therefore, the semantics of iterations can still be defined as a fixed point over all ordinals instead of only finite ones.
This paper is organised as follows.Section 2 provides a brief introduction to the mathematics used in the paper.Section 3 defines the model for the CSP-like language.Section 4 introduces a specification statement and extends the above model to include mixed terms.In Section 5 the syntax and semantics of the CSP-like language with specification are defined.Section 6 proves the monotonicity of all the CSP-like constructors.Section 7 concludes the paper.
Mathematical preliminaries
A relation on a set C is a "partial ordering" of C if it is reflexive, anti-symmetric, and transitive; it is a "total ordering" if in addition each pair of elements in C are comparable.We denote the structure consisting of set C and a partial ordering v on C by C ; v.However, when the partial ordering is understood or when the context makes clear whether we are regarding C as a set or as a set with a partial ordering, we do not distinguish between C and C ; v.
A set B C , with C and hence B partially ordered by v, has an "upper bound" u 2 C if x v u for all x 2 B; u is a "least upper bound"-"lub" for short-if in addition u v v for every upper bound v of B. The lub of B when it exists is denoted by F B. An element ? 2 C is a "bottom" or "least element" of C if ?v x for all x 2 C , and 2 C is a "top" or "greatest element" of C if x v for all x 2 C .An element m 2 C is a "minimal element" of C if x v m x = m for all x 2 C .If sets C and D are partially ordered by v C and v D , respectively, then the Cartesian product C D is partially ordered by v, defined by x; y v w; z x v w and y v z for all x; w 2 C and y; z 2 D. Definition (1) can be generalised to n-fold Cartesian products in the obvious way.The product sets in this paper will always be partially ordered by (1) and we will not state this explicitly in each case.
Given sets C and D with D partially ordered by v D , the set C ! D of functions from C to D is partially ordered by v, defined by f v g f x v D gx for all x 2 C : In this paper, functions will be partially ordered by (2) only and we will not state this explicitly in each case.If, in addition C is partially ordered by v C , then f : C ! D is said to be "monotone" if x v C y f x v D f y for all x; y 2 C : 2nd Irish Workshop on Formal Methods, 1998
An Unbounded Nondeterministic Model for CSP-like Languages
We denote by C ! D the set of monotone functions from C to D.
A "complete lattice" is a partially ordered set in which each subset has a lub.It can be shown that every complete lattice has a bottom and a top.We now give some well-known properties of complete lattices, omitting proofs; the reader looking for more details should refer to literature such as [1].
Lemma 1 Any finite totally ordered set is a complete lattice.
Lemma 2 If C and D are complete lattices, then so is C D. Moreover, for B C D, G B = G fx : 9 y : x; y 2 Bg; G fy : 9 x : x; y 2 Bg : Lemma 2 can be generalised to n-fold Cartesian products in the obvious way.
As is conventional we will use small Greek letter to denote ordinals; ! will denote the first infinite ordinal, i.e., the set of natural numbers.For any complete lattice C and f : C ! C , we define f 0 = the identity function on C ; f +1 = f f (functional composition) for successor ordinals +1; f = G ff : g for limit ordinals :
f c v f c for all ordinals and ;
2. There exists a least ordinal such that In Lemma 4 (2) is called the "closure ordinal" of f in C .
As we can safely replace the closure ordinal in f c with any ordinal , we can conveniently work with one "super-closure" ordinal 1 f for each f ; for 1 f take any ordinal containing all closure ordinals of f .For brevity, we will write simply 1, letting context supply the implicit subscript.
Any c satisfying f c = c is called a "fixed point" of f ; if in addition c v d for every fixed point d of f , then c is a "least fixed point" of f .The least fixed point of f when it exists, is denoted by x:f x .
Lemma 5 Given f : C ! C for C a complete lattice with bottom ?, f has a least fixed point satisfying x:f x = f 1 ?.
Monotonicity also applies to sequences in the obvious way: a (possibly transfinite) sequence x 0 ; x 1 ; ; x ; with elements drawn from a set partially ordered by v is said to be "monotone" if x v x for all and .
In the rest of this paper we will employ only v to denote a partial ordering, letting context resolve any ambiguity that might otherwise arise.
2nd Irish Workshop on Formal Methods, 1998
The model
There already are some models for CSP-like languages, e.g., [6,7,10,11,12].In this paper we take the notations in [7] (the reader may also refer to the textbook by Hoare [4]) and model a process as a set of quadruples s 0 ; tr; ref ; s; where the first and last components s 0 ; s2 State are machine states, which map a variable in Var (may be infinite) to a value in Val (may be infinite), i.e., s 0 ; s : Var 2 Property P4 states that selective input is not allowed.This property is not needed in the CSP theory because CSP deals with individual communication events.But it is essential for CSP-like languages where refusals are sets of channels.Property P5 states that if a process has a proper internal state after the current trace, it can terminate successfully in that state without doing any further communications.Property P6 states that if a process does not diverge after a trace tr, it can have only finitely many internal states.This excludes unbounded nondeterminism in processes and also makes sequential composition continuous on the process space Proc.
2nd Irish Workshop on Formal Methods, 1998
Specification statement
A specification of a communicating process with state describes not only its communication behaviours, but also the relationships between communications, initial states and final states.
In this paper we regard a communicating processes with state as a generalised-state tr ; s transformer as in [12], which takes a process from an initial state tr 0 ; s 0 to a final state tr 0 tr; s after having performed the communications in tr.
The where " p " represents successful termination and A a = A f ag.
The intuitive meaning of the specification statement is as follows.Sp, when started in one of the initial generalisedstates satisfying pretr 0 ; s 0 , must be able to engage in any communications tr satisfying I tr 0 ; tr 0 tr; ref ; if Sp terminates, it does so in one of the final states satisfying posttr 0 ; tr 0 tr; s 0 ; s, with the terminating trace tr satisfying I tr 0 ; tr 0 tr; fg; if Sp cannot start in an initial state s 0 , i.e., 8 tr 0 : :pretr 0 ; s 0 , it diverges immediately.A special symbol " p " is added to the domain of the refusals of specification statements, which indicates termination.If f p g is a possible refusal of a specification statement after tr, it may refuse to terminate successfully.The precise meaning of " p " will be given when the semantics of the specification statement is defined.Specifications can also be defined as a set of computations in Comp.The set of all mixed terms, which are a mixture of specifications and programs, is defined as MixT b = PComp, which also contains all the processes defined in Def. 1.Of course, not every mixed term possesses all the properties P1 to P7 and the empty set is an example.
The refinement ordering v on MixT is defined the same as that on Proc: M 1 v M 2 ; if and only if M 2 M 1 ; so that the least element on MixT is also the same as that on Proc, i.e., ?. Theorem 1 MixT ; v; ?and Proc; v; ?are complete partial orders.
Proof.See Appendix.
5 The CSP-like language and its semantics
To demonstrate the effect of introducing unbounded non-determinism into a CSP-like language, we choose the following simple language.
Definition 2 The syntax of mixed terms
The CSP-like language with specification statements, MProg, is defined as follows: M ::= I ; w : pre; post j skip j stop j div j x := e j c?x j c!e j M 1 ; M 2 j if b then M 1 else M 2 j M 1 k M 2 j while b do M od; where the parallel composition construct M 1 k M 2 is the same as the usual CSP construct except that the linked channels between M 1 and M 2 are hidden.
2nd Irish Workshop on Formal Methods, 1998
An Unbounded Nondeterministic Model for CSP-like Languages
We use to denote the main semantic function: which maps a mixed terms to a set of computations.We first define the semantics of the specification statement.
Equation ( 5) defines the terminating computations of Sp, as well as its divergent computations.The specification statement is a total-correctness formula in the sense that, if Sp has some terminating computation s 0 ; tr; ref ; s, there must exist some initial trace tr 0 such that pretr 0 ; s 0 holds for s 0 and its terminating trace tr and final state s, together with tr 0 , satisfies the communication invariant I tr 0 ; tr 0 tr; fg and the postcondition posttr 0 ; tr 0 tr; s 0 ; s.
The state transition from s 0 to s can only be achieved by changing those alterable variables in w.Whenever there exists some proper final state, termination can happen; we have, therefore, no constraints on ref in (5).If there is no initial trace satisfying pretr 0 ; s 0 for an initial state s 0 , Sp diverges.In this case, (5) = ?.
Equation ( 6) defines the non-terminating computations of Sp.To do so, a new element " p " is introduced into the domain of refusals in the communication invariant I .Without it, unwanted nonterminating computation s 0 ; tr; ref ; would also be included in Sp for every terminating computation s 0 ; tr; ref ; s in Sp .If Sp diverges in some s 0 , then 5 = 6.
The semantics of the rest of the language constructs are defined as follows.For expressions e (including Boolean expressions), the value of e, evaluated in a state s 0 , is denoted by es 0 .If anything goes wrong, such as a variable being undefined, the value of es 0 is error.Whenever this happens, produces a divergent process.The divergent process div is an unspecified process on whose behaviour there is no constraint whatsoever.It is identified with the least element ? on MixT : div b = ?: The definitions for skip and stop are already given in (3) and (4) in Section 3, respectively.The assignment statement, x := e, can be defined as follows: x := e b = fs 0 ; ; ref ; s 0 es 0 =x j s 0 2 Stateg; where es 0 6 = error and s v=x is the same as s except that the value of x is v. f s 0 ; tr; ref ; s j : bs 0 ^s 0 ; tr; ref ; s 2 M 2 g: (11) provided that bs 0 6 = error The parallel composition constructor M 1 k M 2 is the key constructor in any parallel language.It is used to construct concurrent systems from individual ones.We stipulate that M 1 and M 2 do not share any program variable other than read-only ones.The communication alphabet of M 1 k M 2 is the union of the communication alphabets of M 1 and M 2 .The common channels of M 1 and M 2 are linked channels on which M 1 and M 2 can communicate.Linked channels are hidden from outside.Therefore, the parallel constructor is a combination of the CSP parallel constructor k and hiding operator n.We define it as follows: i;j2f1;2g^i6 =j fs 0 ; tr; ref ; s j 9 tr 0 tr: s 0 ; tr 0 M i ; fg; ? 2 M i tr 0 M j 2 tracesM j ; s 0 g; where tr A is a trace obtained from tr by removing all the events that happened on channels not in A and is defined as follows: f s 0 ; tr ^tr 0 ; ref ; s j f tr 00 j tr 00 nc = tr tr 00 2 tracesM ; s 0 g is infiniteg; (16) where trnc = tr M , f cg.
The first clause claims that if M is ready to communicate with its environment along some channels other than c, then so can M nc; the second clause states that if M engages in an infinite unbroken sequence of communications along channel c, then M nc diverges.This also includes the case where M diverges.
The while statement has the usual meaning as that in sequential languages except that it also involves communications and may not terminate.It is defined as a fixed-point of the following equation: X = if b then M ; X else skip: 2nd Irish Workshop on Formal Methods, 1998
An Unbounded Nondeterministic Model for CSP-like Languages
However, as our extended model contains unbounded non-determinism, some of the language constructors are no longer continuous.In [2], Boom showed how Dijkstra's definition of the weakest precondition for the while loop can be adapted to permit unbounded nondeterminism.We use the same idea to adapt the definition for recursion so that unbounded nondeterminism is permitted.The basic idea is to take the infinite join over the set of all ordinals, rather than just over the set of finite ordinals.The existence of the fixed-point then depends on the monotonicity of all the language constructors which have been proved in Theorem 3.
We defined the fixed-point over ordinals by Lemma 5: where where is a limit ordinal: However, the CSP-like language without specification statement is continuous with respect to the refinement order on the process space Proc, as indicated in Theo. 2. Therefore, the fixed-point of the the above recursive definition does exist over the set of all finite ordinals and the traditional loop definition works.
Theorem 2 All the processes without specification are well-defined and continuous.
Proof.Since the only differences between our CSP-like language and others, such as those in [6,12], are the infinite message set and the hiding operator which hides only internal channels, the proofs of the well-definedness and continuity of the language constructors are the same as those in [6] except for the hiding operator.We only give details for the proof of the well-definedness of the hiding operator n.The proof of continuity of hiding follows a similar argument and is omitted here.
We prove that Pnc, where c is a linked channel, satisfies conditions P1 to P7.We only give details for P6 which is affected by an infinite message set.
Assume s 0 ; tr; ref ; s 2 Pnc and s 0 ; tr; fg; ?6 2 Pnc .By definition, there exist some traces tr 0 such that tr = tr 0 nc and s 0 ; tr 0 ; ref f cg; s 2 P .P, being a process, has property P6.Hence, fs 0 j s 0 ; tr 0 ; fg; s 0 2 P g is finite for every tr 0 .As tr is finite, it has the form tr = a 1 a 2 a n ; where a i is a communication event on a channel other that c.Therefore, each trace tr 0 must have the form tr 0 = tr 1 c a 1 tr 2 c a 2 a n tr n+1 c ; where tr i c is a finite trace from f c:v j v 2 Valg .As tr 0 is not a diverging trace of P by definition and c is not an input channel (see P4), there exist only finitely many tr i c 's such that tr = tr 0 nc.Hence, tr 0 :tr=tr 0 nc fs 0 j s 0 ; tr 0 ; fg; s 0 2 P g is finite; so is fs 0 j s 0 ; tr; fg; s 0 2 Pnc g as required. 2
Monotonicity
In this section we show that all the statement constructors, sequential composition, conditional composition, iteration, and parallel composition, are monotone with respect to the refinement order.
Sequential composition.Let P i and Q i be mixed terms satisfying P i v Q i for i = 1 ; 2. We prove that P 1 ; P 2 v Q 1 ; Q 2 , that is, Q 1 ; Q 2 P 1 ; P 2 .
2nd Irish Workshop on Formal Methods, 1998 ! Val; tr is a communication trace in Comm , which are finite sequences of communication events of the form c:v2 Comm b = Chan Val, including the empty one ; the third component ref is a set of channel names from Chan, called refusals.Therefore, the terminating process skip that changes nothing but terminates successfully, can be defined as follows: Notice that ref Chan in (3) can be omitted because it is always true.To model divergences, we introduce a special state ?. Thus, s 0 ; tr; ref ; ?represents a divergent computation after tr.To model non-termination, we introduce another special state .Thus, Processes For a given communication channels Chan, a variable set Var, and a value set Val, the process space Proc is a set of all subsets P of Comp, which satisfy the following conditions: For any s 0 2 State, P1 tracesP; s 0 = ftr j 9 s: s 0 ; tr; ref ; s 2 Pg is nonempty and prefix-closed: tr 6 = ^s 0 ; tr 0^t r; ref ; s 2 P = s 0 ; tr 0 ; fg; 2 P; P4 s 0 ; tr ^ c in :v ; ref ; s 2 P = 8 v 0 9 s 0 : s 0 ; tr ^ c in :v 0 ; ref ; s 0 2 P; P5 s 0 ; tr; ref ; s 2 P ^s 2 State = 8 ref 0 : s 0 ; tr; ref 0 ; s 2 P; P6 s 0 ; tr; ref ; s 2 P ^s 0 ; tr; fg; ?6 2 P = f s 0 j s 0 ; tr; fg; s 0 2 Pg is finite; P7 s 0 ; tr; ref ; ? 2 P = 8 tr 0 ; ref 0 ; s: s 0 ; tr ^tr 0 ; ref 0 ; s 2 P, where v and v 0 range over Val, tr and tr 0 over Comm , ref and ref 0 over PChan, s and s 0 over State ?, and c in is an input channel (not a linked channel).
s 0 ; tr; ref ;represents an unfinished computation where the final state is unobservable.The non-terminating and broken process stop can be defined as follows:(4)The enlarged state set is denoted by State ?.The processes are sets of computations ( Comp b = State Comm Chan State ?), which satisfy the following conditions.Definition 1 syntax of the specification statement Sp is defined as follows: Sp :: I ; w : pre; post ; where w is a list of alterable program variables, I , pre, and post are predicates of the types: The input process c?x inputs a message on channel c and stores it in x.Notice that the second set above is infinite as Val is infinite.The semantics of the output process c!e is similar to that of c?x and left with interested readers.If M 1 and M 2 are two mixed terms with the same alphabets, then M 1 ; M 2 is a mixed term which behaves like M 1 , except that if M 1 terminates successfully, it continues behaving like M 2 .b then M 1 else M 2 is the usual if-statement in the sequential language.If states are hidden, it behaves like the nondeterministic choice in CSP.Its definition is if 2 nflinked channelsg; where the parallel composition k and the hiding operator are defined as follow.
If s 1 and s 2 map x into different non-?values, the parallel combination is broken.The disjointness constraint of k guarantees that this won't happen.The hiding operator nc is defined as follows.Let M be a process with c 2 M .The communication alphabet of M nc is that of M minus c. | 2014-10-01T00:00:00.000Z | 1998-07-02T00:00:00.000 | {
"year": 1998,
"sha1": "5a0b7b20485c385153e729dc0efe1a7aeaec3f0d",
"oa_license": "CCBY",
"oa_url": "https://www.scienceopen.com/document_file/e519e2fe-1265-4dbb-9e56-0687cfcf4a90/ScienceOpen/001_Lai.pdf",
"oa_status": "HYBRID",
"pdf_src": "CiteSeerX",
"pdf_hash": "5a0b7b20485c385153e729dc0efe1a7aeaec3f0d",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Computer Science"
]
} |
169651079 | pes2o/s2orc | v3-fos-license | Inequality in Distribution of Health Care Resources in Iran: Human Resources, Health Centers and Hospital Beds
Background: One of the most important goals of health systems is equality in access to health services and one of the important dimensions of health equality in relation to resources distribution in this sector. Objectives: This study was done to examine the equality in resource distribution between provinces of Iran in 2014. Methods: Inthepresentretrospectivestudy,thepopulationincludedthepublichealthsectorhumanresources,hospitalbeds,and health centers in Iran. The data were collected from the Iran statistics center and ministry of health and Medical education. The inequality was estimated by Lorenz curve, Gini coefficient, Hoover index, and by using Excel and Stata 12 software. Results: The Gini coefficients were 0.13797, 0.19980, and 0.1499 for the manpower, health care centers, and hospital beds, respec-tively. Thus, they were significant. In addition, the Hoover index was 7, 8.60, and 6.70 for the three studied resources, respectively. Conclusions: The distribution of the health sector resources between Iran provinces was relatively equal. Therefore, health policy makersshouldbefocusedontheotheraspectsof accessibilitysuchasdistributionof resourcesaccordingtohealthneeds,resources distribution within provinces, and quality of provided services.
Background
One of the most important implicit objectives in health systems is equity in reach to marginal objectives (1,2). Thus, recently, equity in health resources distribution (such as bed, physician, and equipment,) has increasingly been paid attention (3).
An expand in literature is in the health area regarding concept and measure method of equity due to the fact that the subject is a multi-aspect and expand concept, which includes resource allocation (2,4).
This concept is one of the important criteria in resource distribution and access to health services. In view of experts, equity in health resources distribution is guaranteed for health society (5). In other ways, the study of the equity in facility distribution in health sector is a perquisite for fair utilization of health intervention as one of the ways for assessing the government access level in achievement to social equity (6,7).
On the other hand, the demand growth, health costs, as well as the human and financial resource scarcity have created limitation for responsibility to need society health care in different geographical areas and inequity in health resource distribution (8,9). Based on the WHO definition, health inequity is an avoidable inequality in health status or social determinant health different groups of society (10). Despite economic growth and health gains, there are many differences in the society health status as well as individuals within a country (11). In developing countries, resources are often unbalanced due to the lack of appropriate information and poor health planning (12).
In Iran, health promotion and the fair distribution of health facilities have always been of particular importance. Policy makers have followed the achievement to this goal in economic and social development programs (13). However, limited comprehensive studies have been done on the status of health resources distribution, including the human resources employed in the health system, hospital beds, and health centers in Iran. In another study, in evaluating the geographical distribution of ICU beds for non-cardiac diseases in Iran, showed that the distribution of intensive care unit beds in the provinces of the coun-Copyright © 2018, Shiraz E-Medical Journal. This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/) which permits copy and redistribute the material just in noncommercial usages, provided the original work is properly cited try was fair, however, the number of beds was lower than in other countries (14). In addition, In another study, in regards to examining the geographical distribution of access to health care in rural areas, have shown that there is a significant difference in the distribution of health house based on rural mortality rates (15).
Regarding the mentioned cases, the present study aimed to investigate the status of equity in health resources distribution, including the human resources of the health system, hospital beds, and health centers in Iran. The results of this study can inform policy makers on how these resources are distributed among the provinces. Furthermore, the findings of the present study provide appropriate evidence for future planning and management of health sector resources to improve access to health services in the country.
Data and Variables
The present descriptive-analytical study was done in a cross-sectional way in all provinces of Iran in 2014. The analysis variables consist of population, healthcare centers, health human resources, and hospital beds. The population of provinces was a variable collected from the last census by Iran statistics center. Health care centers include public health care centers in each province. It includes a rural health house, rural health center, delivery center, urban health house (payegah-e-behdashti), urban health center, and pubic educational hospitals. Health human health resources include all human resources in mentioned public healthcare centers. Bed variables are limited to active beds in public educational hospital.
The data related to healthcare centers, health human resources, and hospital beds were collected from the budget and surveillance office of ministry of health.
Analysis
The distribution of health sector resources were analyzed using the Gini coefficient, Hoover index, and Lorenz curve by Excel (2013) and Stata 13 software. The most usual inequality index is Gini coefficient. It is the algebraic method to assess the income distribution, which is statistically the ratio of current income inequality to maximum possible inequality ratio in a complete fair distribution. The Gini coefficient is between 0 (perfect equality) and 1 (perfect inequality). Usually, if the index is 0.25 -0.35, the distribution is relatively fair; if it is 0.35 -0.50, the distribution is relatively unequal, and if it is 0.50 -0.70, the distribution is unequal completely (16). It is possible to find Gini and Hoover indices by Lorenz curve. One dimension of the curve is the population cumulative percent (Xs axis) and another dimension is the cumulative share of the considered variable (in this study, it is manpower, health centers, and hospital beds, Ys axis) ( Figure 1).
In this Figure, the 45 degree line indicates the equality in distribution and the distance between the curve and the line indicates the extent of inequality, namely if the distance is more the inequality is also more among the population. In order to examine the exact distribution of manpower, health centers, and hospital beds, it is necessary to measure the distance between the 45 degree line and the curve to be able to measure the Gini coefficient. The Gini coefficient was measured by the following formula: Where: X = cumulative percentage of the population Y = cumulative percent of the variable On the other hand, Hoover index is the extent of the resources redistribution, which is needed to achieve perfect equality. It is equal to the maximal vertical distance between the Lorenz curve and equity line ( Figure 2) (17).
Results
The study findings indicated that the highest ratio of health centers and hospital beds per 1,000 population belongs to south Khorasan province and the highest manpower number/1,000 population belongs to Ilam province ( Table 1).
The present study findings indicated that Kohgilooyeh and Booyer Ahmad, Ilam, South Khorassan, Chaharmahal and Bakhtiari, Mazandaran, and Markazi provinces are the richest in the field of manpower /population ratio and Alborz, Semnan, Tehran, Ghom, Hamedan, and Ghazvin are the poorest. In addition, Kohgilooyeh and Booyer Ahmad, South Khorassan, Ardabil, Mazandaran, The Lorenz curve, Gini, and Hoover coefficients for health manpower, health centers, and hospital beds are presented in Table 2 and Figures 3 -5. By virtue of the numerical rate of the Gini coefficient indices of the manpower, centers, and hospital beds were distributed in the country relatively equitably. In addition, based on Hoover coefficient, 7% of manpower, 8.6% of health centers and 6.7% of the active hospital beds in the richer provinces should be transferred to the poorer ones in order to have a completely equity distribution of resources.
Discussion
Equilibrant and proper distribution of the resources in the health sector is considered as a factor promoting the health indices and more equity and social justice. In this study, having used Hoover and Gini coefficients as an image of the provincial distribution of manpower, health centers, and hospital beds in the governmental sector in Iran was presented to draw the attention of the decision makers and policy makers in the health sector to a way to distribute such resources. The study findings revealed that the manpower distribution has been relatively equal in the provinces of Iran; the manpower resources have always been an essential element in the health system and played the most important role in providing health services (18); thus, the lack of manpower may create a great obstacle against the goals of organizational development in the third millennium; it has economic, social, and direct effects on the rate of access to health services and the quality of such services (19,20). Besides, it is necessary to have a special attention for proper distribution of special-ist manpower in the hospitals and other health centers in developing countries to achieve the goals of permanent development (21). Farahani et al., (22) indicated that the resources distribution (specially manpower) has been effective on the mortality rate of children aged under 1 year in both short and long term and the effects have been more in the long term. Matsumoto et al., (23), study findings indicated that the distribution of manpower was improper in health care sector in Japan, U.S.A. and England. In addition, Zhang et al., (24), in a study in China, reported high levels of inequality in geographic distributions of health workers. Shinjo and Aramaki's study indicated that there was a strong relationship between the ways of distributing manpower and providing services for the patients in the health sector, thus, proper distribution of manpower may create the field appropriate to provide more favorable services with high quality (25). Besides, having examined justice in the distribution of manpower in the ministry of health in 2003 -2007, Mobaraki et al., (5), reported that manpower resources had been distributed equitably in the ministry of health; however, a part of the employment permission should be redistributed in the state universities in order to have equal distribution. Mirsaeid et al., (26), found that the hospitals of Tehran University of Medical Sciences did not have enough manpower and the distribution of manpower was not according to the directions of ministry of health in most units of the hospitals. The present study indicated that the health centers had relatively equal distribution between Iran's provinces. In this regard, in a research in China, it was presented that there were inequality in distributions of institutions for both the hospitals and the primary care sectors (24). Kiadaliri et al., (27), examined the geographical distribution of access to the health care in rural regions of Iran in 2006 -2009 and indicated that there was a considerable difference between the distribution of health centers and mortality rate in the rural regions indicating the relatively low inequity. Also, Rezaei et al., (28) indicated that the distribution of the pharmacies and health centers had become more and more equity in 2001 -2011 and was better than the distribution of other centers. Taghvaei and Zakeri showed that there were some differences concerning available hospital and clinical services between 14 districts of Esfahan city. In such a way, 20 out of 29 hospitals were located in three regions while some city districts had no hospital (29), which is one of the reasons leading to a difference between the Esfahan study and the findings of the present. Thus, in this study the indices were assessed generally in high-level Iran provinces, while the Esfahan study examined only the spatial distribution of the hospital and clinical services in a metropolis.
In addition, the results of the present study indicated that the distribution of active hospital beds in provinces was just. In this regard, the findings of a case study regarding Iran showed that the Gini and Hoover coefficients and Lorenz curve indicated that the distribution process of the hospital beds was equity (30). Zhang et al., (24), indicated that the geographical distribution of the beds in both the hospitals and the primary care sectors was unequal. In their study, Ameryoun et al., (13), assessed the geographical distribution of the beds in the ICU unit for non-cardiac patients and concluded that such distribution was equity in Iran; however, the beds/population ratio was less than other countries. Another study indicated that although there was some difference regarding the medical specialists/population and active beds/population ratios among the governmental hospitals in Iran provinces. The hospital beds have been distributed equitably in Iran governmental hospitals by virtue of Gini coefficient (31). Finally, the resources/population ratio has been relatively equal between Iran provinces. It should be noted that after years now the resources allocation criteria, such as population and the cities' needs play a more important role than other ones, due to the fact that a proportion of different variables under examination in the immigrant provinces is less than the others; namely Tehran and Alborz provinces were at the lowest rates regarding all the three indices (Manpower, active beds, and health and treatment centers). On the other hand, another factor may play an important role in creating such balances is an irregular increase of immigration to Tehran city and suburbs as a metropolis. Notwithstanding many physical, resources, and manpower of the health sector of Tehran province are in Tehran city, the distribution rate of the governmental services has had more disadvantages than profit for the province.
Like other studies, this study has had some limitations; for instance, the variables under the study were general and only the resources of governmental sector were taken into consideration; however, the current study may present a comprehensive view towards geographical distribution of the resources of the health sector to the policy makers in health system because most of the resources in health sector including manpower, physical resources, and hospital beds were analyzed.
The proper and fair distribution of manpower resources, health centers, and hospital beds have effective roles in distributing other healthcare services. The distribution situation of the health sector resources between Iran provinces was relatively equal. Therefore, health policy makers should be focused on the other aspects of accessibility such as distribution of resources according to health needs, resources distribution within provinces, and quality of provided services. | 2019-05-30T23:44:33.117Z | 2018-05-14T00:00:00.000 | {
"year": 2018,
"sha1": "9166fb9ddfd67198f43e4bb93285ed248060e769",
"oa_license": "CCBYNC",
"oa_url": "https://semj.kowsarpub.com/cdn/dl/09deffd0-633b-11e8-a45e-cb78ec78bbbf",
"oa_status": "GOLD",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "f66a5337b8b146946858be43b19a8afc05caf47c",
"s2fieldsofstudy": [
"Medicine",
"Economics"
],
"extfieldsofstudy": [
"Geography"
]
} |
241114081 | pes2o/s2orc | v3-fos-license | Mechanical and Dynamical Properties of High Performance Concrete Using Arecaceae Fiber
The purpose of this endeavor is to acquire learning for arranging the working with negligible exertion material (arecaceae fiber) using normal fibre.Because trademark fiber can be one possible material which are trashy and locally available.Among all regular fiber coir fiber are picked because of its most important malleability quality. The static direct of arecaceae fiber stronghold concrete essential part is in every practical sense cloud .In this work the effect of fiber content on properties of AFRC is considered. To survey the capability of coir fiber in improving the properties of the strong the first class of plain cement is used as reference. The mechanical properties, for instance, compressive quality , part rigidity, modulus of break and quick chloride entrance test were resolved for all ARFC and Plain Cement models. Likewise the fiber in appropriate sum should be selected.(i.e.,) fiber substance of 2% and 3% by weight of bond and having a length of 6.5cm are used to design AFRC.
I. ARECACEAE FIBRE :
Coconut fiber is isolated from the outer shell of the coconut. Basic Name -Coir Logical Name -Cocos Nucifera Plant Name -Arecaceae (Palm) There are two sorts of coconut fiber dim hued fiber expelled from created coconut and white fiber from energetic coconut.Brown fiber are thick, strong and high bendable quality yet white fiber are smoother and better yet additionally weaker. So in building dark colored fiber are generally utilized.
II SCOPE OF THE STUDY
The key purpose of the endeavor is to upgrade the versatile property of bond by including arecaceae fiber. To convey a viable "Green strong material "using typical fiber (arecacea) will assemble the flexible quality and reduction the existence cost of the structure. This endeavor will make others to extend the energy for look at ordinary fiber and addition the use of ordinarily available materials.
A.CEMENT:
Bonding is the most fundamental fixing in cement. One of the basic criteria for the decision of bond is its ability to make improved microstructure in cement. Properties of cement :
B.FINE AGGREGATE:
The sand got from stream beds is utilized as fine total. The fine total alongside the hydrated concrete glue fill the space between the coarse total. Waterway sand was utilized as the common fine total. Its properties were tried according to May be: 2386.The test aftereffects of sifter investigation is exhibited in beneath tables Cube 3 3 3 3 3 3 Cylinder 3 3 3 3 3 3 Beam 3 3 3 3 3 3 Cylinder for RCPT The above diagrams demonstrates the consequences of compressive quality for the examples on both 7 days and 28 days. From the outcome the 3%CFRC achieve more compressive quality when contrasted with control and 2% CFRC.
C.SPLIT TENSILE STRENGTH
Solid barrels of distance across 100mm Dia x 200mm stature were thrown M40 review of cement .The plain concrete, 2% of CFRC and 3% of CFRC were tried for 7days and 28th days split elasticity. The7 days and 28 days comes about were analyzed in underneath graphs.
Fig (V) Comparison Of Results
The above outlines demonstrate the aftereffects of split rigidity for the examples on both 7 days and 28 days. From the outcome the 3%CFRC achieve more Split rigidity when contrasted with control and 2% CFRC . This test technique covers the assurance of the electrical conductance of cement to give a quick sign of its protection from the entrance of chloride particles.
VI. CONCLUSION
The restoring time of 7 days and 28 days comes about exhibits that the 3%CFRC are higher in quality (compressive ,split malleable and flexural) when appeared differently in relation to 2% CFRC and the plain concrete.
The connection exhibited that the volume of 3% coir fiber invigorated concrete had the perfect plan of mechanical properties Coconut fiber fortified bond has exhibited less number of part progressions . Thusly, it very well may be a conventional choice being a developed area. y chloride passage test comes about that the quality property for the plain concrete and coir fiber bolster cement were less vulnerable against chloride ambush .
Based on the examination of weight it is assumed that coconut fiber can as a piece of the age of helper lightweight concrete. Table VIII RESULTS OF RCPT | 2019-09-17T02:44:26.843Z | 2019-08-30T00:00:00.000 | {
"year": 2019,
"sha1": "07ec2cab39beee7568c043c69133ba84b4043433",
"oa_license": null,
"oa_url": "https://doi.org/10.35940/ijeat.f7986.088619",
"oa_status": "GOLD",
"pdf_src": "MergedPDFExtraction",
"pdf_hash": "f7cd46bc088a9b2fa5ae0426b31e649b3baf191d",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": []
} |
14398822 | pes2o/s2orc | v3-fos-license | Clinical Outcomes of Infrapopliteal Angioplasty in Patients With Critical Limb Ischemia
Background and Objectives With recent advances in equipment and techniques, infrapopliteal angioplasty has shown results that are comparable to those of surgical bypass in patients with critical limb ischemia (CLI). In this study, we evaluated the efficacy and the feasibility of infrapopliteal angioplasty in patients with CLI. Subjects and Methods Between March 2002 and May 2008, infrapopliteal angioplasty was performed on 118 limbs of 101 patients (79 males; mean age 66 years) with CLI (Rutherford category 4, 5 or 6). Freedom from reintervention, limb salvage, and overall survival were analyzed. Results The median follow-up duration was 30 months. Initial technical and clinical success rates were 69.5% and 83.1%, respectively. No major complication requiring surgical intervention developed after angioplasty. Among 82 limbs with initial technical success, the rate of freedom from any reintervention at 2 years was 70.7% and that from limb salvage was 97.6%. Young age and Rutherford category 6 at initial presentation were independent predictors associated with poor 2 year primary patency in these patients with CLI. Overall survival at 1 year was 86.4% and that at 2 years 76.3%. A history of cerebrovascular accident was an independent predictor associated with poor 2 year survival in these patients. Conclusion Infrapopliteal angioplasty as a primary choice of treatment in CLI patients showed favorable clinical outcomes and feasibility.
often aged and not optimal candidates for surgical bypass due to medical co-morbidities with increasing perioperative mortality rates and a poor autogeneous conduit. 4)5) As an alternative method of revascularization, infrapopliteal angioplasty is preferred by patients with visible stumps, good runoff of distal outflow vessels or high surgical risk. However, it had been regarded as an inferior treatment option compared to surgical bypass due to bulky catheters, lack of low caliber wires and general lack of clinical experience which frequently resulted in periprocedural complications. 4) [6][7][8] Recently, newly developed devices and technical advances have widened the therapeutic spectrum of angioplasty to more distal and complex lesions with lower complication rates. 9)10) Also, it rarely compromises the option of future surgical procedures because it preserves the saphenous vein for lower extremity distal bypass surgery while at the same time it reduces the cost and shortens the hospital stay. 10) In this study, we evaluated the feasibility and the efficacy of infrapopliteal angioplasty in terms of primary patency and limb salvage as a primary choice of treatment in patients with CLI. Moreover, we evaluated predictors associated with poor primary patency and poor survival in these patients.
Study subjects
Between April 2002 and May 2008, we retrospectively analyzed 118 limbs of 101 patients with CLI (Rutherford-Becker grades 11) 4, 5, or 6) who had been treated by infrapopliteal angioplasty. Demographic, laboratory, technical, and clinical characteristics were retrospectively collected from their medical records. Blood samples were collected one day before or immediately before the index procedure.
Percutaneous transluminal angioplasty procedure
Vascular access for infrapopliteal angioplasty was gained via ipsilateral or contralateral puncture of the common femoral artery. An antegrade approach was preferred when there were no combined lesions requiring proximal iliac or femoral intervention and in patients without obesity, due to better maneuverability of the catheters and better control of the wiring in case of total occlusion. After placement of the 6 to 7 French sheath under local anesthesia, an intra-arterial bolus of 5000 IU of heparin was given, and additional heparin was administered to maintain an activated clotting time between 250 to 300 seconds, if procedure time was lengthened. Infrapopliteal lesions were passed with a 0.014" to 0.035" guide wire. If the initial transluminal recanalization failed, total occlusions were recanalized through the subintimal dissection plane with reentrance into the true lumen using 0.035" hydrophilic guide wires (Terumo, Tokyo, Japan). After crossing the lesion, a 4 Fr to 5 Fr multipurpose catheter was advanced to the distal patent segment in exchange with a 0.014" wire. Percutaneous transluminal angioplasty (PTA) was then performed with adequate sized balloons (2.25 to 4.0 mm) at 6-10 atmospheres. In case of elastic recoil or a flow limiting dissection after balloon dilatation, stents were implanted for bailout purposes. When proximal lesions existed, including ipsilateral iliac, femoral or popliteal artery lesions, concomitant procedures were performed.
Follow-up
Prior to intervention, patients were premedicated with aspirin (100 mg, daily) and maintained on this dose indefinitely after PTA. Followup included a clinical examination during the hospital stay and 1 month after the intervention to document hemodynamic improvement. Subsequent follow-up was considered when there was a question of worsening clinical status. If necessary, repeat peripheral angiography or a computed tomography angiogram was scheduled on their visit. The causes and date of death were examined by chart re-view, telephone contact with family members, or checking with the national statistical office.
Angiographic findings and clinical outcomes
Technical success was defined as PTA resulting in less than 30% residual stenosis with sufficient antegrade flow; a suboptimal result was defined as sluggish flow and/or residual stenosis 30% to 50% after repeated dilatation. Primary clinical success was defined as an improvement of at least one clinical category in the Rutherford-Becker classification. 11) Primary patency was defined as persistent patency without any re-intervention including angioplasty, surgical procedures performed on or at the margins of the treated lesion, or amputation. Limb salvage was defined as prevention of major amputation. Major amputation was defined as limb loss below or above the knee level, while minor amputation was defined as a transmetatarsal or more distal level amputation of the lower extremity.
Statistical analysis
We analyzed data using the Statistical Package for the Social Sciences (SPSS) software (version 15.0; SPSS Inc., Chicago, IL, USA). Data are presented as the mean±standard deviation for continuous variables and as number (percentage) for categorical variables. Student's t-test was used for comparison of continuous variables, and chisquare analysis was used for comparison of categorical variables. Event-free survival curves were constructed by the Kaplan-Meier method. All significant univariate predictors of poor primary patency and poor survival were included in the multivariate logistic regression analyses. A p less than 0.05 was considered significant.
Laboratory data of the study patients
Laboratory data of the patients are listed in Table 1. The mean blood urea nitrogen level was 25.7±15.6 mg/dL, and the mean creatinine level 2.62±2.90 mg/dL. The mean estimated glomerular filtra-
Angiographic findings of the study patients
Angiographic findings of the study patients are shown in Table 2. A total of 161 lesions were treated in 118 limbs of 101 patients with CLI. Anterior tibial artery was the most commonly treated location (53%), and totally occluded arteries were observed in 60 limbs (51%). Concomitant procedures in the proximal lesion as ipsilateral iliac or superficial femoral artery were performed in 51 limbs (43%). Among them, a superficial femoral artery procedure was most commonly performed (36%). Antegrade access was done in 58 limbs (49%), and subintimal angioplasty was done in 11 limbs (9%). Initial technical success was achieved in 82 limbs (69.5%). On final angiography, a suboptimal result was obtained in 26 limbs (22%), which showed diffuse and heavily calcified lesions at initial presentation. There were 10 technical failures due to 8 unsuccessful placements of guide wires, 1 heavy calcified lesion, and 1 arterial rupture. No major com-plication requiring surgical intervention occurred after PTA. Minor procedural complications such as flow limiting dissection, rupture, and embolization occurred in 25 limbs (21%), which were treated by conservative treatment or stent implantation. In 10 cases of elastic recoil or flow limiting dissection, 11 stents were implanted for bailout purpose. Types of stents included 8 bare-metal coronary stents, 2 drug-eluting coronary stents, and 1 balloon-expandable peripheral stent. Among 11 stents, 3 were implanted in the anterior tibial artery, 5 in the posterior tibial artery, and 3 in the peroneal artery.
Clinical outcomes of the study patients
Clinical outcomes of the patients are shown in Table 2. Primary clinical success was obtained in 98 limbs (83.1%); 18 limbs showed persistent gangrene and 2 other limbs persistent pain. The mean follow-up duration was 34 months (range 1-94 months). During followup, re-PTA was performed in 7 limbs (6%), surgery in 44 limbs (37%), major amputation in 6 limbs (5%), and minor amputation in 41 (35%).
Among 82 limbs with initial technical success, primary patency was 70.7% at 1 year and 2 years, and the limb salvage rate was 97.6% at 1 year and 2 years (Table 2). There were no significant differences in terms of 2 year primary patency (70% vs. 63%, p=0.744) and 2 year limb salvage (100% vs. 95%, p=1.000) regardless of the presence or absence of stent implantation. Moreover, there were no significant differences in terms of 2 year primary patency (82% vs. 62%, p=0.322) and 2 year limb salvage (100% vs. 95%, p=1.000) regardless of the presence or absence of subintimal angioplasty. Kaplan-Meier analysis was used to obtain primary patency (Fig. 1A) and limb salvage rates (Fig. 1B) during 2 year follow-up.
Overall survival was 86.1% at 1 year and 75.2% at 2 years ( Table 2). The most common cause of death was acute myocardial infarction (n=10) and renal insufficiency (n=10). There was no significant difference in terms of 2 year survival (60% vs. 78%, p=0.243) irrespective of the presence or absence of stent implantation. Furthermore, there was no significant difference in terms of 2 year survival (55% vs 79%, p=0.126) irrespective of the presence or absence of subintimal angioplasty. Kaplan-Meier analysis was used to calculate survival during the 2 year follow-up period (Fig. 2).
Multivariate analysis associated with poor 2 year survival
Multivariate analysis showed that a history of cerebrovascular acci-dent (OR 5.396, 95% CI 1.792-16.254, p=0.003) was an independent predictor associated with poor 2 year survival in these patients (Table 4).
Discussion
The Bypass versus Angioplasty in Severe Ischemia of the Leg (BA-SIL) trial, a multicenter, randomized controlled trial that compared 6 month clinical outcomes after angioplasty or surgery among CLI pa- Data are expressed as the mean value±SD or percentage of patients. *Rutherford classification can be found in reference 11. EGFR: estimated glomerular filtration rate, SD: standard deviation tients, did not show a significant difference with respect to amputation-free survival, all-cause mortality and quality of life during shortterm follow-up. 13) However, most data about CLI included the proximal segments as femoral or femopopliteal segments, and few reports have focused on the infrapopliteal segment due to high technical failure rates and low patency rates. Recently, with technological advancements in wires and increased clinical experience, infrapopliteal angioplasty has been more frequently attempted and has shown results comparable to surgical bypass. 4)5)14)15) In meta-analyses, which compared infrapopliteal angioplasty with surgical bypass, the limb salvage rates of the two strategies were similar, 16) and our data showed favorable clinical outcomes after infrapopliteal angioplasty. A relatively low technical success rate (70%) in this study may be attributed to a high proportion of suboptimal results (22%). Most of the suboptimal results involved diffuse and heavily calcified lesions, and had 30% to 50% of residual stenosis despite repeated balloon dilatation. Any supplementary intervention such as stent implantation was not performed in those lesions. Stent implantation in infrapopliteal arteries have been avoided due to the diffuse, complex lesion characteristics in the majority of CLI cases, and the risk of crushed stents or thromboses due to flow stasis. 6)17) Recent advancements in coronary stenting technology now allows for the use of stent implantation in more distal crural arteries. Feiring et al. 14) demonstrated that stent-supported infrapopliteal angioplasty is effective, with few adverse events for patients with CLI or lifestyle-limiting claudication. Also, Siablis et al. 9) demonstrated that sirolimus-eluting stents reduced the restenosis rate and a favorable midterm clinical outcome was obtained with a significant decrease in clinically driven re-interventions compared with bare metal stents in suboptimal infrapopliteal angioplasty. It is plausible that more aggressive use of stent implantations might improve angiographic and clinical results. However, our study showed no significant improvements of clinical outcomes, including 2 year primary patency and 2 year survival, in the stent implantation group. It may be explained by the small number of cases in the stent implantation group. Therefore, long-term follow-up data from multiple centers would be needed, and the potential benefit of stent implantation should be considered over the financial cost.
Although the BASIL trial 13) included patients who were good candidates for either infra-inguinal bypass surgery or balloon angioplasty, many patients with CLI are not optimal candidates for surgical bypass, and most likely undergo amputation for their critical limb ischemia. Considering that the objective of catheter-based angioplasty is also to lower the rate of amputations and to support wound healing by ameliorating distal flow, 14) angioplasty can be considered a first line treatment option in CLI patients who otherwise are expected to undergo major amputation.
In our study, 2 year overall survival was 75.2%, and the main causes of death were cardiovascular events, which correspond with the significant associations between CLI and cardiovascular morbidity or mortality in previous reports. [18][19][20][21] Therefore, in the management of CLI patients with poor life expectancy, the maintenance of symptom relief and limb salvage is more crucial than for their quality of life. 4
)22)23)
Hence, infrapopliteal angioplasty is a less-invasive, safe and reproducible method, and could be a better strategy for CLI patients with comorbidities who are at high surgical risk. Several limitations of our study need to be considered. First, this was a retrospective study from a single institution, which can create referral bias. Also, extrapolation of these results to most patients with CLI has limited validity because of the small population size. Second, we did not perform follow-up angiography or other modalities for long-term patency in asymptomatic patients. Regular follow-up based on ankle-brachial index or duplex scan is needed to obtain better patency results in future prospective studies. Third, the efficacy of proximal PTA in cases with combined proximal and distal angioplasty could not be distinguished from those of distal angioplasty alone. Fourth, we did not analyze the relationship between outcomes and prescriptions such as cilostazol after the procedure because of a deficiency of long-term follow-up data.
In conclusion, infrapopliteal angioplasty showed favorable clinical outcomes in terms of primary patency and limb salvage rate, and demonstrated its feasibility as a primary choice of treatment in patients with CLI. Although it is still a technically challenging procedure, more aggressive treatment with advanced devices and more clinical experience would allow us to achieve better results. | 2014-10-01T00:00:00.000Z | 2012-04-01T00:00:00.000 | {
"year": 2012,
"sha1": "8f731736ce2e5f8db08a0d5d00f884a259235059",
"oa_license": "CCBYNC",
"oa_url": "https://europepmc.org/articles/pmc3341423?pdf=render",
"oa_status": "GREEN",
"pdf_src": "PubMedCentral",
"pdf_hash": "8f731736ce2e5f8db08a0d5d00f884a259235059",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
259980037 | pes2o/s2orc | v3-fos-license | Salivary lactate dehydrogenase and salivary total protein as potential biomarkers for screening periodontal disease
Background Timely diagnosis of periodontal disease is crucial for restoring healthy periodontal tissue and improving patients' prognosis. There is a growing interest in using salivary biomarkers as a noninvasive screening tool for periodontal disease. This study aimed to investigate the diagnostic efficacy of two salivary biomarkers, lactate dehydrogenase (LDH) and total protein, for periodontal disease by assessing their sensitivity in relation to clinical periodontal parameters. Furthermore, the study aimed to explore the impact of systemic disease, age, and sex on the accuracy of these biomarkers in the diagnosis of periodontal health. Materials and methods A total of 145 participants were categorized into three groups based on their basic periodontal examination index, with 20 in the periodontally healthy group, 50 in the gingivitis group, and 75 in the periodontitis group. Salivary LDH was measured using the rate of nicotinamide adenine dinucleotide (NADH) oxidation, to measure the kinetics of LDH activity, while total protein was measured using the Lowry method. Descriptive and analytical statistical analyses were performed to examine the associations between the variables and biomarkers. Results The results of the study demonstrated that salivary LDH was 72% sensitive, while salivary total protein was 78% sensitive in correlation to clinical periodontal parameters. The accuracy of the test was not influenced by sex, but age had a significant effect on both biomarkers, particularly LDH. Systemic disease was another factor that significantly affected the accuracy of the test. Conclusions Although salivary LDH and total protein show promise as biomarkers for screening periodontal disease, their interpretation may be impacted by age and systemic disease. To enhance the accuracy of periodontal disease detection, the study suggests combining these biomarkers with more specific indicators and diagnostic tools.
Introduction
Periodontal disease is one of the most prevalent oral diseases among the middle-aged and elderly populations (Gul et al., 2020).With advances in clinical knowledge and techniques, early detection can prevent the progression of periodontal disease.Therefore, periodontal screening may help prevent periodontal disease and improve oral health-related quality of life (Kim et al., 2013).Periodontal disease involves both bacterial and host inflammatory mediators.Pathogens induce the production of biologically active molecules and stimulate host inflammatory mediators leading to tissue and bone destruction (Isola, 2022).
In recent years, there has been a trend toward screening for systemic diseases using human body fluids, such as blood, cerebrospinal fluid, saliva, urine, sputum, peritoneal fluid, and pleural fluid, which have been approved for definitive diagnosis and disease state maintenance (Giannobile, 2012).Saliva is one of the body's fluids secreted by the major and minor salivary glands.In periodontal diseases, the marginal and sulcular gingivae act as a frontline area where the interaction between bacteria and the host immune response occurs by aggregation, adherence, cell-killing, and inhibition of microbial metabolism (Zin et al., 2021).These interactions may be maintained under oral ecological conditions (Scannapieco, 1994).It is hypothesized that quantitative variations in the salivary protein concentration affect the prevalence of oral diseases (Zin et al., 2021).
Albumin acts as a serum ultrafiltrate in the oral cavity (Shaila et al., 2013).Salivary protein concentrations can serve as markers of plasma protein leakage resulting from inflammation (Shaila et al., 2013).Lactate dehydrogenase (LDH), a stable cytoplasmic enzyme found in all cells, is another salivary biomarker in addition to total protein.LDH is rapidly released into the cell culture supernatant when the plasma membrane is damaged, a key feature of cells undergoing apoptosis, necrosis, or other forms of cellular damage (Miyoshi et al., 2018).LDH activity can be easily quantified using nicotinamide adenine dinucleotide (NADH) produced during the conversion of lactate to pyruvate.The amount of LDH in saliva is directly proportional to the number of dead or damaged cells (Azizi et al., 2011).
This study aimed to investigate the diagnostic efficacy of salivary lactate dehydrogenase (LDH) and total protein, for periodontal disease and the effect of systemic disease, age, and sex on the accuracy of these biomarkers.
Materials and methods
The study enrolled 145 individuals with an age range of 19-73 years and a mean age of 37 years Standard deviation (SD) = 14.109.The sex distribution was 55.2% female and 44.8% male.The participants were classified into three groups based on their periodontal health status: periodontally healthy control group (n = 20), gingivitis group (n = 50), and periodontitis group (n = 75), using the basic periodontal examination index (BPE) as defined by the new classification system (Chapple et al., 2018).
This observational case-control study was conducted among patients attending public health centers between February 2022 and July 2022.Inclusion criteria were as follows: participants who had not used antibiotics or anti-inflammatory drugs in the past month.Exclusion criteria included the following: individuals who had eaten, drank, brushed their teeth, or gargled within 1-2 h before saliva collection, because these factors could impact the results.Participants undergoing periodontal treatment, pregnant or lactating women, and smokers were also excluded from the study.
The samples were stored in a cooler box and assigned code numbers for analysis.Clinical examination was performed using the BPE index (Chapple et al., 2018).Probing depth (PD) and bleeding on probing were measured using World Health Organization (WHO)-standard probes.
The saliva samples were centrifuged for 10 min at 1000 rpm.One milliliter of supernatant saliva was stored at À20 °C for total protein determination, and 1 ml was stored at 4 °C for LDH determination within the same week of collection (de la Pen˜a et al., 2004;Zin et al., 2021).
LDH levels were determined using the method suggested by the French Society of Clinical Biology (SFBC) (Giuliani et al., 2019).This technique measures LDH activity kinetics by assessing the rate of NADH oxidation.
Commercial kits (Spinreact's Quantitative LDH in vitro diagnostic (IVD) kits) were used to conduct the LDH test.The working reagent was prepared and combined with 1 ml of the saliva supernatant.After incubation for 1 min, the mixture was transferred to a 1 cm light path cuvette, and the absorbance was measured at 340 nm.Mathematical calculations followed the manufacturer's guidelines, and the results were reported as U/L.
Lowry method, which relies on the Folin phenol reagent for protein measurement, was employed for total protein determination (Waterborg, 2009).Spinreact's Quantitative determination of total urinary and cerebrospinal (CSF) protein IVD kits were used.The colorimetric method was utilized: 200 ml of saliva sample was first mixed with 1 ml of the working reagent.After a 10-min incubation, the mixture was transferred to a 1 cm light path cuvette, and the absorbance was measured at 598 nm.The protein concentration was calculated using the manufacturer's guidelines, and the results were expressed as mg/mL.Sample size calculations were performed using G*Power 3.1.9.2 software, assuming equal SD in the three groups according to a study by Nomura et al., 2006 considering one-way analysis of variance (ANOVA) with an effect size of 0.25, statistical power of 80%, and a significance level of 95% (a < 0.05).Based on these parameters, a minimum of 20 participants were required for each group.
Data analysis was performed using the SPSS statistical software (version 21.0; SPSS Inc., Chicago, IL, USA).Descriptive statistics, such as mean and SD, were used to describe variables, including sex, age group, systemic health status, and the distribution of LDH and total protein among the healthy, gingivitis, and periodontitis groups.
To assess the differences between the three groups, one-way ANOVA tests were conducted.Independent t-tests were used to examine the correlation between LDH or total protein levels and sex.ANOVA was employed to compare LDH and total protein levels across age groups and health status.Pairwise comparisons between age groups were conducted using Scheffe's test to identify significant differences.
Receiver operating characteristic (ROC) analyses were performed to evaluate the performance of LDH and total protein levels.The sensitivities of LDH and total protein were estimated using the area under the curve (AUC).Statistical significance was defined as p < 0.05.
Results
A total of 145 participants were included in this study.The demographic characteristics of the participants are shown in Table 1.
A comparison of the groups in terms of age showed a statistically significant difference among the periodontally healthy, gingivitis, and periodontitis groups (p < 0.05), with the periodontitis group having a higher mean age than the healthy periodontal and gingivitis groups.However, there was no significant difference in sex distribution among the groups (p > 0.05), indicating a relatively equal representation of males and females across all the three groups.Of the participants, 118 had no medical history of systemic diseases, while 27 had medical conditions and were undergoing treatment.The most common categories of medical conditions were diabetes and hypertension (12), followed by hypertension alone (10), diabetes (4), and thalassemia (1).
The one-way ANOVA analysis of LDH and total protein indicated a significant difference among the three groups for LDH (F (2, 142) = F-value 53.562, p < 0.05) and total pro-tein (F (2, 142) = F-value, 24.548, p < 0.05); Figs.1A and 1B show the mean levels for each group.For both biomarkers, post-hoc pairwise comparisons revealed significant differences between the periodontally healthy and gingivitis groups (p < 0.05), as well as between the periodontally healthy and periodontitis groups (p < 0.05).Additionally, a significant difference was observed between the gingivitis and periodontitis groups (p < 0.05).
There was no evidence to suggest that sex influenced LDH or total protein levels, as listed in Table 2.However, age had a strong effect on both the biomarkers (Table 3).Post-hoc analysis revealed a significant association between age group and salivary biomarkers (p < 0.05).In pairwise comparisons, older individuals had significantly higher LDH levels than young (p < 0.05) and middle-aged adults (p < 0.05).Additionally, total protein levels increased significantly with age, with older participants exhibiting significantly higher levels compared to young adults (p < 0.05).
This study also investigated the effect of systemic disease on both LDH and total protein levels.A significant correlation was found between systemic disease and these salivary biomarkers, as shown in Table 4.
Using the ROC curve, the sensitivity of LDH as a diagnostic test was found to be 72%, whereas the sensitivity of the total protein test was 78%.
Discussion
Saliva, an easily collected and noninvasive sample, contains valuable markers of periodontal disease derived from systemic and local sources (Saliem, 2016).Biomarkers, including inorganic and immunological substances, play crucial roles in oral cavity defense and can be used for screening, diagnosis, and disease monitoring (Mohammed et al., 2022).
In this study, LDH and total protein levels were assessed using straightforward colorimetric biochemical techniques that are widely used in various scientific fields because of their versatility, quantitative capabilities, speed, cost-effectiveness, and ease of use.However, they have limitations, such as low sensitivity, inability to analyze colorless compounds, and potential errors from interfering materials with similar colors (Hosseini et al., 2021).
Numerous studies have extensively explored the presence of LDH in saliva, covering various aspects.Studies investigated its association with oral squamous cell carcinoma (Kallalli et al., 2016;Merza et al., 2010).Another study investigated the relationship between LDH levels and gingivitis in patients with diabetes (Chidzhavadze et al., 2006).Studies have further investigated the impact of smoking and the presence of LDH on periodontal health (Ali et al., 2018;Risteska et al., 2021).Moreover, a previous study evaluated the health of the peri- Fig. 1A The mean levels of lactate dehydrogenase (LDH) in saliva among participants categorized into three groups: periodontally healthy, gingivitis, and periodontitis groups.
Fig. 1B The mean levels of total protein in saliva among participants categorized into three groups: periodontally healthy, gingivitis, and periodontitis groups.
Table 2 The effect of sex on LDH and total protein levels.odontium by measuring total protein levels in saliva without considering the potential factors that could affect protein levels (Zin et al., 2021).However, the current study sets itself apart by examining the combined effect of cofactors on the levels of LDH and total protein in saliva.Furthermore, we investigated how these cofactors influence the diagnostic potential of the biomarkers under investigation.The present study supports the use of salivary LDH as a screening biomarker for periodontal health, which is consistent with previous studies (Ansari Moghadam et al., 2022;Nomura et al., 2016Miyoshi et al., 2018).Based on the ROC curves, we determined that AUC for LDH was 0.72.This finding indicates that LDH holds promise as a potential biomarker for periodontal disease.Our study results confirmed the influence of age as a cofactor affecting salivary LDH levels.This finding is consistent with that of a recent study, which indicated that LDH levels vary with age (Farhana and Lappin, 2022).These observations may be attributed to natural changes in salivary glands and saliva composition that occur with age, potentially affecting the levels of enzymes such as LDH in saliva.
Another significant finding of this study was the impact of systemic diseases on salivary LDH levels.This result can probably be attributed to the increased levels of LDH in response to tissue damage or disease (De La Pen˜a et al., 2007).Systemic conditions can influence LDH levels in saliva, potentially indicating oral cavity damage or systemic disorders affecting salivary gland function (De La Pen˜a et al., 2007).Interestingly, this finding aligns with those of previous studies that proposed different LDH levels in patients with diabetes (Chidzhavadze et al., 2006) and that identified LDH as a diagnostic marker for oral squamous carcinoma (Kallalli et al., 2016).
This study revealed a significant correlation between periodontal health and salivary total protein level, the AUC for total protein was determined to be 0.78, indicating a high level of accuracy.This suggests that total protein has the potential to serve as a biomarker for periodontal disease.The increase in salivary total protein is primarily due to the immune response and inflammatory processes that occur in periodontal tissues (Shaila et al., 2013).Several factors contribute to the higher levels of salivary total protein in individuals with periodontitis, gingival inflammation increases vascular permeability, and protein exudates from tissue breakdown products can contribute to an overall increase in salivary total protein levels (Zin et al., 2021).
These findings are consistent with those of a study conducted by Wakde et al., 2018 who also observed significantly higher salivary total protein levels in the gingivitis and periodontitis groups compared to periodontally healthy individuals.Furthermore, our results are consistent with those obtained by Shaila et al., 2013, Karthiga et al., 2017, and Henskens et al., 1993.Additionally, a study by Zin et al., 2021 found that patients with chronic gingivitis and chronic periodontitis had total protein contents that were 1.6 and 4.2 times higher, respectively, than periodontally healthy controls.However, it is worth noting that the salivary total protein levels reported in previous studies, including ours, vary considerably.This variation may be influenced by factors such as protein detection techniques, reagents used, racial differences, and genetic variations in the salivary glands.
Age is an important factor that influences salivary total protein levels.The current study sheds light on the significant relationship between salivary total protein and age and its implications for assessing periodontal health.Several factors could explain this observation.Age-related changes in salivary glands, saliva flow rate, and oral microbial flora are potential cofactors affecting salivary total protein (Xu et al., 2019).
Additionally, salivary total protein exhibits a significant relationship with systemic diseases such as diabetes, cardiovascular disease, and autoimmune disorders (Hu et al., 2007).Elevated levels of salivary total protein in individuals with systemic diseases suggest potential tissue damage or inflammation, possibly due to plasma protein leakage caused by cellular damage (Lawrence, 2002).
One limitation of this study was the small sample size for each systemic disease, which may not have been sufficient to accurately assess the effect of each systemic disease on biomarkers in relation to periodontal health status.
Conclusions
The salivary biomarkers LDH and total protein show promise as screening biomarkers for periodontal disease.However, it is important to note that these biomarkers should be utilized alongside other salivary biomarkers.This is because their levels can be influenced by various cofactors such as age and systemic disease, which can impact their reliability, as evidenced in this study.Therefore, further research is recommended.
Ethical statement
This research was authorized by the Medical Ethical Committee of the University of Baghdad's College of Dentistry, with Ref. number: 641, date 4/8/2022, project No. (641622).Informed consent was obtained from patients for participating in the study.
Table 1
Demographic variables of the study groups.
Table 3
The effect of age on LDH and total protein levels.
* The mean difference is significant at the 0.05 level.
Table 4
The effect of systemic diseases on LDH and total protein levels.
* The mean difference is significant at the 0.05 level. | 2023-07-20T15:01:53.651Z | 2023-07-01T00:00:00.000 | {
"year": 2023,
"sha1": "3cab90bc29569cf6580ce5af97a7604711bdc8fb",
"oa_license": "CCBYNCND",
"oa_url": "https://doi.org/10.1016/j.sdentj.2023.07.011",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "a2565e1cb9f08362f65e562ed19c60cf2492ccc5",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
13431964 | pes2o/s2orc | v3-fos-license | Proteomic analysis reveals that COP9 signalosome complex subunit 7A (CSN7A) is essential for the phase transition of migratory locust
The migratory locust displays a reversible, density-dependent transition between the two phases of gregaria and solitaria. This phenomenon is a typical kind of behavior plasticity. Here, we report that COP9 signalosome complex subunit 7A (CSN7A) is involved in the regulation of locust phase transition. Firstly, 90 proteins were identified to express differentially between the two phases by quantitative proteomic analysis. Gregaria revealed higher levels in proteins related to structure formation, melanism and energy metabolism, whereas solitaria had more abundant proteins related to digestion, absorption and chemical sensing. Subsequently, ten proteins including CSN7A were found to reveal differential mRNA expression profiles between the two phases. The CSN7A had higher mRNA level in the gregaria as compared with the solitaria, and the mRNA amount in the gregaria decreased remarkably during the 32 h-isolation. However, the mRNA level in the solitaria kept constant during the crowding rearing. Finally and importantly, RNA interference of CSN7A in gregaria resulted in obvious phase transition towards solitaria within 24 h. It suggests that CSN7A plays an essential role in the transition of gregaria towards solitaria in the migratory locust. To our knowledge, it’s the first time to report the role of CSN in behavior plasticity of animals.
Scientific RepoRts | 5:12542 | DOi: 10.1038/srep12542 In recent years, many fruitful studies have been carried out to elucidate the intrinsic molecular mechanisms of phase transition in locusts from various aspects such as genomics, transcriptomics and metabolomics. A large scale of transcriptomic sequencing was carried out in the migratory locust using an expressed sequence tag (EST) technique in 2004 11 , and 532 differentially expressed unigenes were identified between the two phases. The transcriptome dynamics in the same species were further analyzed in 2010 based on a newly emerged next-generation sequencing technology 12 . A lot of genes related to neural pathway, such as dopamine receptor, adipokinetic hormone, neurotransmitter synthetase were found to be up-regulated in the gregarious locusts. Another transcriptomic analysis was performed in the desert locust 13 . The solitary locusts up-regulate genes related to antioxidant systems, detoxification and anabolic renewal, whereas gregarious locusts have a greater abundance of transcripts for genes involved in sensory processing and nervous system development and plasticity. After monitoring and comparing transcript profiles between the two phases at various developmental stages, Chen et al. found that a sharp rise in phase differences appeared during the 4 th instar and the high level difference was maintained in all the following stages. Therefore, the 4 th instar stage seems to be a turning point in the process of forming the phase differences in the migratory locust 12 . Some neuronal signaling and sensory activity related genes, such as dopamine receptor 5 , chemosensory protein (CSP) and takeout 3 were proved to play roles during the phase transition. The successful assembly of the migratory locust genome is a milestone in the study of phase transition of locust 14 . The genome is 6.5 giga base pairs (Gb), the largest animal genome sequenced so far. Significant expansion of gene families associated with energy consumption and detoxification were found in the locust genome 14 . Besides, small RNA 15 and metabolomics 4 analysis also disclosed a lot of regulators contributing for the phase transition.
Proteomic researches have also been carried out, but few significant progresses have been made till now. In 1999, polypeptide maps were generated from hemolymph of the desert locust and twenty differential spots were identified between the two phases. However, detailed information about these peptides was not available 15 . Two proteins, a 6-kDa peptide and a serine protease inhibitor were identified to have different expression patterns between the two phases in the desert locust using a combined approach of high-performance liquid chromatography (HPLC) with matrix-assisted laser desorption/ ionization time-of-flight mass spectrometry (MALDI-TOF MS) 16 . These proteomic studies moved very slowly because of the lack of locust genome information at that time.
The behavior plasticity makes locusts to be a good model in the epigenetic researches 17,18 . Two DNA methyltransferase genes were shown to be phase-specific in certain tissues of the desert locust 19 . Further analysis revealed that the methylome of the gregarious desert locust was characterized by CpG-and exon-specific methylation, and the overall methylation levels were substantially higher than other invertebrates 20 . These findings suggest that DNA methylation may be involved in the regulation of locust phase transition. Besides, a cAMP-dependent protein kinase (PKA) was reported to play a role in the transition from solitary to gregarious behavior in the desert locust 21 . Except for these reports, few studies have been further performed in recent years.
In general, large progresses have been made in exploring the mechanisms of phase transition in the migratory and desert locusts. A lot of differentially expressed genes and pathways have been identified based on DNA sequencing techniques. However, the researches in protein areas, such as protein identification and protein modification, have been largely lagged. One of the most key reasons is the lack of genome sequence information. Fortunately, the genome assembly of the migratory locust was just finished 14 , which provides much convenience for protein identification and will do great help for exploring the complex mechanism of phase transition in another viewpoint.
In the present study, we identified 90 differentially expressed proteins between the two phases in the migratory locust by a quantitative proteomic technique. Among them, CSN7A was found to play an essential role in the transition of gregaria towards solitaria.
Results
Proteins identified in the locust head. A total of 4, 895 peptides were identified by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) from the locust head, and they were finally assembled into 1, 387 proteins. After COG classification, 1, 104 proteins were assigned to 25 COG categories, and the "R" cluster (General function predication) and "O" cluster (Posttranslational modification, protein turnover, chaperones) represent the largest two groups, and their amounts are 18% and 15% of the total identified proteins, respectively (Fig. 1). The "O" cluster proteins are mainly heat shock protein chaperones, ubiquitin-dependent proteins, proteasome-related proteins, peptidase activity-related proteins, glutathione S-transferase, protein disulfide-isomerase, and COP9 signalosome complex subunits (Supplementary Table S1). The top 70 most abundant proteins are listed in Table 1 and Supplementary Table S2 Differentially expressed proteins between the two phases. Among the 1, 387 identified proteins, 90 proteins were shown to have different expression levels between the two phases. Sixty-four were up-regulated in the gregaria (as compared with the solitaria), and twenty-six were down-regulated ( Table 2, Supplementary Table S3). Most of the up-regulated proteins are involved in the processes of structure formation (such as cuticle protein, beta-1 tubulin, profiling, and troponin), energy metabolism (electron transfer flavoprotein subunit alpha [ETFA], dehydrogenase/reductase SDR family member 11-like, 3-ketoacyl-CoA thiolase, V-type proton ATPase subunit B and isocitrate dehydrogenase Based on sequence homology, 1, 104 proteins were classified into 25 COG (Clusters of orthologous groups) categories (A). The "R" and "O" clusters represent the largest two groups. Proteins classified in the "O" cluster (Posttranslational modification, protein turnover, chaperones) were further assigned to different categories according to their molecular functions (B). Detailed information about these "O" cluster proteins can refer to Supplementary Table S1. Continued NAD subunit beta), and environmental stress response (heat shock protein 60 [Hsp60] and heat shock protein 20.6 [Hsp20.6]). Besides, four hexamerin-like proteins are also abundant in the gregaria. The down-regulated proteins are mainly related to the processes of digestion and absorption (carboxypeptidase A-like, serine protease-like protein, and 1,4-alpha-glucan-branching enzyme-like) and chemical sensing (takeout-like). In addition, the differentially expressed proteins are also enriched in the class of "Regulation of gene expression" in both the gregaria and solitaria ( Table 2). For example, wingless protein, 3′ -phosphoadenosine 5′ -phosphosulfate synthase (PAPSS), COP9 signalosome complex subunit 7A (CSN7A), and juvenile hormone binding protein (JHBP) were highly expressed in the gregaria, and splicing factor 3B subunit, ubiquitin-conjugating enzyme E2 variant 2-like isoform 1, proteasome subunit alpha type-4, and arginine/serine-rich-splicing factor RSP31 (RSP31) showed higher levels in the solitaria.
Differential expression at mRNA levels. To validate the differential expression, fourteen representative proteins were selected according to their function categories in Table 2. Their mRNA expression profiles were examined in the whole head of the two-phase locusts. Nine protein genes, including CSN7A, JHBP, PAPSS, choline transporter-like protein 4 (CTL-4), two hexamerin-like protein 2 (Hexa2 and Hexa2*), cytoplasmic actin A3a (actinA3a), ETFA and Arylphorin revealed higher mRNA level in gregaria ( Fig. 2). It was in consistent with the protein profiles in Table 2. The brain tissues, the most important part of head, were also studied. Four genes, such as CSN7A, JHBP, PAPSS, CTL-4 and takeout-like, showed similar expression patterns between the mRNA and protein levels. There were still four genes, including V-ATPase subunit B (V-ATPase), ATPsyn-d, RSP31, and NADPH--cytochrome P450 (P450) revealed constant mRNA levels between the two phases ( Supplementary Fig. S1).
Time-dependent mRNA expression during phase transition.
In order to further narrow target proteins that may play a role in the regulation of locust phase transition, CSN7A was chosen and time-dependent mRNA expression dynamics were examined in brain during the phase transition process. The CSN7A had higher mRNA level in the gregaria (Figs 2, 3), the level decreased significantly at 4, 16 and 32 h-isolation and was as low as that in the solitaria at 32 h (Fig. 3). However, the mRNA level did not change during the crowding of solitary locusts (Fig. 3).
RNA interference (RNAi) and behavioral assay.
To validate the function of CSN7A in locust phase transition, RNAi and behavioral assay were carried out. The mRNA level was suppressed by injection of CSN7A dsRNA in the gregaria (Fig. 4A), and the behavioral state shifted from gregaria (dsGFP population) to solitaria (dsCSN7A population) (Fig. 4B). The phase difference between two populations was highly significant (P Mann-Whitney U test = 1.61 × 10 −12 ). For example, 60% and 0% individuals fall into the P greg interval of 0.8-1.0 in the dsGFP and dsCSN7A population, respectively. In addition, significant difference existed in the three key behavioral parameters (attraction index, total distance moved, and total duration of movement) between the two populations (Fig. 4C). These results revealed that phase transition did happen by RNAi of CSN7A in the gregarious locust.
Discussion
The "O" cluster proteins are extremely abundant in the locust head. This phenomenon was also found in the antennae of Batocera horsfieldi based on cDNA library analysis 22 . However, similar phenomenon did not exist in the whole insect bodies [23][24][25][26] . It seems that "O" cluster proteins are mainly abundant in the head as compared with the other parts of insects. It suggests that the proteins related to post-translational modification, protein turnover and chaperone folding are highly involved in the regulation of head function in insects. Locust phase polyphenism is a typical phenomenon of epigenetics 17,19,26 . The existence of high abundant "O" cluster proteins suggests that post-translational modification may play important roles in the locust phase transition. The two locust phases differ in many aspects, especially in the body color and behavioral activity. The gregaria is darker and more active, while the solitaria is shallower and quieter. The proteomic analysis revealed that proteins related to structure formation, melanism and energy metabolism have significantly higher expression level in the gregaria. This is consistent with the facts that gregarious locusts have stronger muscles, darker color and more frequent activity. As compared with the gregaria, the solitaria owns more abundant proteins related to digestion, absorption and chemical sensing. It's apparently that the former two characteristics provide the solitary locusts with higher abilities in digestion and absorption, and the latter one gives them stronger olfactory sensation. This makes them have an advantage over the gregarious locusts in feeding and mating, and then results in higher reproductive capacity 14 .
In the present study, hexamerins and JHBP are abundant in the head of gregarious locust. Similar results have been revealed by EST library analysis in the same species 11 . Both hexamerin and JHBP have been suggested to play a role as juvenile hormone (JH) transporters, and even as regulators of JH levels and action [27][28][29][30] . This explains the involvement of hexamerins in JH-dependent differentiation of caste phenotype in some social insects, including termite Reticulitermes flavipes [31][32][33][34] , honey bee 35 and wasp Polistes metricus 36 . Besides caste-related polyphenism in social insects, JH was also reported to mediate plasticity of aggregation behavior in adult desert locusts 37 . Surgical removal of the corpora allata to terminate JH secretion increased aggregation index and behavioral activity of adult locust. This effect was caused by repressing the responsiveness of olfactory interneurons in the antennal lobe to aggregation pheromone. Thus, hexamerins and JHBP can be involved in the phase plasticity of locust by mediating JH action.
Heat shock proteins (Hsps) are a kind of stress-induced proteins that can be synthesized rapidly in response to various environmental stress signals. Hsps usually function as molecular chaperones and participate in numerous cellular functions such as folding, assembly, intracellular localization, secretion, regulation and degradation of proteins 38,39 . Gregarious locusts live at high population density. Population density can alter the expression of Hsps. For example, the mRNA levels of five Hsps (Hsp20.5, Hsp20.6, Hsp20.7 and Hsp90) are significantly higher in the gregarious locust head as compared with those in the solitaria. The mRNA levels were up-regulated by crowding of the solitary locusts (for 32 h), and down-regulated by isolation of the gregarious locusts 40 . In the present study, Hsp60 and Hsp20.6 were identified to have higher protein levels in the gregarious locust head. The over-expression of Hsps in gregaria seems to be a direct response to high-population gather of locust. It's hard to distinguish whether Hsps play a role to control the phase transition.
In the desert locusts, two phase populations display different sensitivity to aggregation pheromone 10,41,42 . Chemosensory protein (CSP) and takeout are important proteins for olfactory sensing [43][44][45][46][47] . RNA interference combined with olfactory behavioral experiments confirmed that six CSP genes (CSP-1 to 6) and one takeout gene, LmigTO1, are responsible for the formation of gregarious and solitary behaviors, respectively 3 . In our study, another CSP (CSP-7) and three new takeout proteins (TO 4 to 6) were identified from the head of Locust migratoria (Supplementary Fig. S2), and the CSPs revealed higher protein level in the gregaria, while the TOs showed higher protein levels in the solitaria. These protein expression patterns are consistent with the early report at mRNA levels 3 , and further confirm that both CSP and takeout are involved in the phase plasticity of locust. The mRNA expression profiles were examined by qRT-PCR in both the head and brain tissues. The mRNA levels were quantified by standard curves generated with serial (10×) dilutions of plasmid DNAs. The relative expression level of each target gene was normalized against a house-keeping gene (RP49). Differences between treatments were compared by Student's t-test, and two levels (P < 0.05 or 0.01) were adopted to judge the significance of difference. Abbreviations: "G", gregaria; "S", solitaria. The abbreviation for gene names can refer to Table 3.
Scientific RepoRts | 5:12542 | DOi: 10.1038/srep12542 The CSN, an eight protein complex (CSN1-8) 48 was originally discovered as an essential regulator in light-induced development in Arabidopsis thaliana 49 . In Drosophila melanogaster, it also plays an essential role for development. Disruption of one of the subunits caused lethality at the late larval or pupal stages 50 . This role of CSN is partly due to its regulation on Hedgehog signaling by mediating proteolysis of some transcription factors 51 . In the same species, CSN was also reported to be involved in circadian rhythms by controlling the degradation of two clock proteins 52 . Interestingly, our study showed that CSN7A played a role in the phase transition from gregaria to solitaria in the migratory locusts. RNAi of CSN7A triggered the phase shift from gregaria to solitaria within 24 h (Fig. 4). Isolation (gregaria to solitaria) and crowding (solitaria to gregaia) may have different regulation mechanism. The former takes place within 4 h in the migratory locusts, whereas the latter cannot finish until 32 h 3 . In the present study, the mRNA amount of CSN7A in gregaria decreased during the isolation, however, the mRNA level remained constant during the crowding of solitaria (Fig. 3). It suggests that CSN7A may be only involved in one direction transition from gregaria to solitaria rather than in its reverse process.
It is the first time to disclose the role of CSN in behavior plasticity of animals. CSN has been reported to be involved in neural development, and regulates dendritic morphogenesis in Drosophila brain through Cullin-mediated protein degradation 53 . More and more evidences revealed that CSN plays an important role in protein degradation through Cullin-ubiquitin-proteasome pathway [54][55][56] . Therefore, CSN might be involved in the phase transition of locust by mediating ubiquitin-dependent proteolysis. Further studies need to be carried out to explore the detailed mechanism of CSN in the regulation of phase transition.
In conclusion, a total of 1,387 proteins were identified in the locust head in the present study, and a large proportion of proteins are involved in post-translational modification, especially in protein folding, phosphorylation and ubiquitylation. Ninety proteins were identified to differentially express between two phases in the head of the migratory locust. Gregaria reveals higher expression in proteins related to structure formation, melanism and energy metabolism, whereas solitaria owns more abundant proteins related to digestion, absorption and chemical sensing. This is consistent with their differentiation in morphology and physiology. JHBP, hexamerin, Hsp, CSP and takeout are suggested to play a role in behavior formation according to their differential expression profiles between two phases. The most interestingly, RNAi of CSN7A in gregaria made the behavior shift towards solitaria within 24 h. It is the first time to disclose the role of CSN in behavior plasticity of animals. These results provide important information for further exploration of the complex mechanism of locust phase transition, as well as for the study of behavior plasticity of animals. The mRNA expression dynamics were examined by qRT-PCR in the brain during phase transition. To make the gregarious behavior change towards solitaria, the 4 th instar gregarious nymphs were individually reared at the same condition as solitary ones. After 2, 4, 8, 16 and 32 h of isolation, the brains were dissected and sampled. Similar procedure was used to convert solitary individuals towards gregaria. The sampling and mRNA level detecting methods were as same as the isolation of gregaria. The untreated gregarious and solitary locusts were used as controls. Differences between each treatment and the corresponding control (untreated gregaria or solitaria) were compared by Student's t-test, and two levels of significance (P < 0.05 or 0.01) were adopted to judge the significance of difference.
Methods
Sample preparation and iTRAQ labeling. When the locusts developed into the second day of 4 th instar, the heads of 3 to 5 gregarious or solitary nymphs were collected and thoroughly homogenized in 500 μ L cold PBS buffer including 1 mM PMSF, 2 mM EDTA and 10 mM DTT. The samples were centrifuged for 20 min at 25,000 × g, and the supernatant was collected. A total of 100 μ g of protein per sample was reduced, alkylated, and then digested by adding 2 μ g trypsin (1 μ g/μ L) at 37 °C overnight. The digested samples were lyophilized and re-suspended in 100 μ L of 0.5 M TEAB (triethylammonium bicarbonate). The method of isobaric tags for relative and absolute quantitation (iTRAQ) was adopted for sample labelling according to the protocol of iTRAQ ® Reagents-4plex Applications Kit (AB Sciex Pte. Ltd., Foster City, USA). Each sample was labeled with an isobaric tag. The iTRAQ-labeled peptide mixtures were pre-separated by strong cation exchange (SCX) column. For SCX chromatography, the LC-20AB HPLC Pump system (Shimadzu Corporation, Chiyoda-ku, Tokyo, Japan) was used, the peptide sample was reconstituted with 4 mLbuffer A (25 mM NaH 2 PO 4 in 25% ACN, pH2.7) and then loaded onto a 4.6 × 250 mm Ultremex SCX column containing 5-μ m particles (Phenomenex, Torrance, CA, USA). The peptides was eluted at a flow rate of 1 mL/min with a gradient of buffer A for 10 min, 5-35% buffer B (25 mM NaH 2 PO4, 1M KCl in 25% ACN, pH2.7) for 11 min, 35-80% buffer B for 1 min. The system was then maintained in 80% buffer B for 3 min before equilibrating with buffer A for 10 min prior to the next injection. Elution was monitored by measuring absorbance at 214 nm, and fractions were collected every 1 min. The eluted peptides were pooled as 12 fractions, desalted by Strata X C18 column (Phenomenex, Torrance, CA, USA) and vacuum-dried. Each fraction was resuspended in certain volume of buffer A (2% ACN, 0.1% FA).
LC-MS/MS Analysis.
A total of 5 μ g of the above solution was loaded on a LC-20AD nanoHPLC (Shimadzu Corporation, Chiyoda-ku, Tokyo, Japan) equipped with a 2 cm C18 trap column, and the peptides were then eluted onto a resolving 10 cm analytical C18 column. The MS data acquisition was performed with Triple TOF 5600 System (AB SCIEX, Concord, ON) fitted with a Nanospray III source (AB SCIEX, Concord, ON) and a pulled quartz tip as the emitter (New Objectives, Woburn, MA). Data was acquired using an ion spray voltage of 2.5 kV, curtain gas of 30 PSI, nebulizer gas of 15 PSI, and an interface heater temperature of 150 °C. The MS was operated with a resolving power of greater than Three key behavioral parameters (attraction index, total distance moved, and total duration of movement) were compared between the dsCSN7A and dsGFP populations (C). Isolation of gregaria referred to the method in Fig. 3, and the behavioral assay was then performed in a rectangular arena monitored by EthoVision system. Eleven behavioral parameters (such as attraction index, total distance moved, total duration of movement, etc.) were collected to calculate the possibility of gregaria (P greg ), which were used for criterion of phase type. The behavioral data were analyzed by the Mann-Whitney U test. The phase difference between two populations was highly significant (P Mann-Whitney U test = 1.61 × 10 −12 ). Differences between treatments were compared by Student's t-test, and two levels of significance (P < 0.05 or 0.01) were adopted to judge the significance of difference. Individual numbers of gregaria and solitaria were marked directly on top of the figure.
Scientific RepoRts | 5:12542 | DOi: 10.1038/srep12542 or equal to 30,000 FWHM (full width at half maximum). The MS/MS data collection and processing was done on Analyst ® software (version 1.6, AB SCIEX, Concord, ON) with the method of Information Dependent Acquisition (IDA) according to the manual.
Database searching for protein identification. The resulting MS/MS spectra were searched against the locust protein database generated from the newly assembled genome 7 with MASCOT software (Matrix Science, London, UK; version 2.3.02). The carbamidomethylation of cysteine was considered a fixed modification, and the conversion of N-terminal glutamine to pyroglutamic acid and methionine oxidation were considered variable modifications. The minimal peptide length was seven amino acids, and a single missed cleavage maximum was used. A peptide mass tolerance of 10 ppm was allowed for intact peptide masses and 0.05 Da for fragmented ions. A stringent 0.01 false discovery rate (FDR) threshold was used to filter the candidate peptide and protein. Two thresholds were set up to filter the candidate proteins whose abundances were significantly different from others: < 0.05 for a two-tailed P-value test and > 1.5 (or < 1/1.5) for the fold-change. For gene ontology (GO, http://www.geneontology. org/) mapping, BLAST2GO software (version 2.5.0, http://www.blast2go.org) was employed to deal with the BLASTx results and then to perform the functional annotation by GO vocabularies, enzyme classification codes, KEGG metabolism pathways 57 . The default settings of BLAST2GO were used in every annotation step.
Quantitative real-time PCR (qRT-PCR). Total RNAs were extracted from the whole head and dissected brain tissues, respectively using an RNAeasy mini kit (QIAGEN, Hilden, Germany). Three heads or eight brains were used for each RNA isolation, and five biological repeats were performed during sampling. PCR reactions were performed in a 20 μ L volume and the final concentration of primers was 250 nM. PCR amplification was conducted on a Roche Light Cycler ® 480 system (Roche Applied Science, Penzberg, Germany) using SYBR green master mix (Roche Diagnostics Ltd. Shanghai, China). The PCR was initiated with a 10-min incubation at 95 °C, followed by 45 cycles of 10 s at 95 °C, 20 s at 58 °C and 20 s at 72 °C. Five biological replicates were performed for each sample. The standard curves for target genes and reference genes (ribosomal protein 49, RP49) were generated with serial (10× ) dilutions of plasmid DNAs. Efficiency of qRT-PCR and correlation coefficients were determined for the primers of each gene. The relative expression level of each target gene was normalized against RP49. The specificity of amplification was ensured by both melting curve analysis and sequencing of PCR product. The primers for qRT-PCR were listed in Table 3.
RNAi. Double-strand RNA (dsRNA) of the target gene and a negative control gene (green fluorescent protein, gfp) were prepared using the T7 RiboMAX Express RNAi system (Promega, Madison, USA) according to the manufacturer's instruction. The primers for dsRNA preparation were listed in Table 3. A total of 35 ng dsRNA was injected directly into eight brains of the 4 th instar nymphs using Nanoject II nanoliter injector (Warner Instruments, Hamden, CT, USA). Twenty four hours later, the effects of RNAi on mRNA level were detected by qRT-PCR and behavioral assay. Four biological repeats were performed for qRT-PCR. For behavioral assay, the same injection was carried out in gregaria, and 30 and 36 individuals were used for dsGFP and dsCSN7A, respectively. Phase Transition. To make gregarious behavior change towards solitaria, the 4 th instar gregarious nymphs were individually reared at the same condition as solitary ones. After 2, 4, 8, 16 and 32 h of isolation, the brains were dissected and immediately placed in RNAlater Solution (Ambion, Austin, USA) for qRT-PCR analysis. The gregarious nymphs maintained in normal situation (high population density) were used as controls. To avoid the influences of circadian rhythm and sexual difference, all samples were collected at the same time point of a day with a sex ratio of 1:1. Each treatment included five biological replicates. To make a reverse phase transition (solitaria towards gregaria), ten solitary nymphs were marked and moved into an optic perplex-made box (10 × 10 × 10 cm 3 ), and 20 gregarious individuals were then added to maintain high population density. The sampling, mRNA level detecting and other methods were as same as the isolation of gregaria.
Behavioral assay. The behavioral assay was performed in a rectangular arena (40 × 30 × 10 cm 3 ). The wall of the arena is opaque plastic and the top is clear. One of the separated chambers (7.5 × 30 × 10 cm 3 ) contained 20 4 th instar gregarious locusts as the stimulus group, and the other end of the chamber with the same dimensions was kept empty. Both ends of the chamber were illuminated equally to prevent the formation of mirror images. The floor of the open arena was covered with filter paper during the behavioral assay. The locust nymphs were gently transferred by a tunnel to the arena. Each individual was recorded for 6 min using EthoVision system (Noldus Inc. Wageningen, the Netherlands). Eleven behavioral parameters (such as attraction index, total distance moved, total duration of movement, etc.) were collected to calculate the possibility of gregaria (P greg ), which was used for criterion of phase type. Detailed information can refer to the early reported methods 3,6 . Statistical analysis. Differences between mRNA levels were compared by Student's t-test. The relative mRNA levels were presented as mean ± SEM (standard error of the mean). Behavioral data were analyzed by the Mann-Whitney U test. Two levels of significance (P < 0.05 or 0.01) were adopted to judge the significance of difference. All the statistics was analyzed using SPSS 15.0 (SPSS Inc., Chicago, USA). | 2016-05-12T22:15:10.714Z | 2015-07-27T00:00:00.000 | {
"year": 2015,
"sha1": "a46fec151f6883535acc79e79339002b427d54ca",
"oa_license": "CCBY",
"oa_url": "https://www.nature.com/articles/srep12542.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "d10917eef68cfbe8ee7e4c3a7acd0cf5351cfde7",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Biology",
"Medicine"
]
} |
226058369 | pes2o/s2orc | v3-fos-license | Differences in location of cerebral white matter hyperintensities in children and adults living with a treated HIV infection: A retrospective cohort comparison
Cerebral white matter hyperintensities (WMH) persist in children and adults living with HIV, despite effective combination antiretroviral therapy (cART). As age and principal routes of transmission differ between children (perinatally) and adults (behaviorally), comparing the characteristics and determinants of WMH between these populations may increase our understanding of the pathophysiology of WMH. From separate cohorts of 31 children (NOVICE) and 74 adults (AGEhIV), we cross-sectionally assessed total WMH volume and number of WMH per location (periventricular vs. deep) using fluid-attenuated inversion recovery (FLAIR) MRI images. WMH were either periventricular when within 10mm of the lateral ventricles, or deep otherwise. We assessed patient- or HIV-related determinants of total WMH volume (adjusted for intracranial volume) and location of WMH using logistic regression, while stratifying on children and adults. At enrollment, median age of participants was 13.8 years (IQR 11.4–15.9) for children and 53.4 years (IQR 48.3–60.8) for adults and 27/31 children (87%) and 74/74 adults (100%) had an HIV RNA viral load <200 copies/mL. WMH were present in 16/27 (52%) children and 74/74 adults (100%). The prevalence of deep WMH was not different between groups, (16/16 [100%] in children vs. 71/74 [96%] in adults, p = 0,999), yet periventricular WMH were more prevalent in adults (74/74 [100%]) compared to children (9/16; 56%) (p<0.001). Median WMH volume was higher in adults compared to children (1182 mm3 [425–2617] vs. 109 mm3 [61.7–625], p<0.001). In children, boys were more likely to have deep WMH compared to girls. In adults, older age was associated with higher total WMH volume, and age, hypertension and lower CD4+ T-lymphocyte nadir with a higher number of periventricular WMH. Our findings suggest that the location of WMH differs between children and adults living with HIV, hinting at a different underlying pathogenesis.
Introduction
The widespread use of combination antiretroviral therapy (cART) has resulted in a substantial decline in the incidence of severe HIV-related neurological complications, including HIVencephalopathy in children with perinatally acquired HIV infection (PHIV) and HIV-associated dementia in adults with behaviorally acquired HIV infection [1,2].
Cerebral white matter hyperintensities (WMH) have been reported to be more prevalent in children and adults living with HIV on effective cART compared to their respective healthy controls [3][4][5]. In adults, WMH are associated with poorer cognitive performance, namely attention and learning as compared to adults without HIV [6,7], which may lead to poorer cART adherence and increased need of social services [8,9].
The pathogenesis of HIV-associated WMH is not fully understood. Potential pathophysiological mechanisms include direct vascular pathology, ongoing effects from HIV-associated neuroinflammation, lasting effects of neuronal damage induced during untreated infection, and neurotoxic effects of cART [10][11][12].
Magnetic Resonance Imaging (MRI) scans are sensitive in detecting subtle HIV-related brain changes; fluid-attenuated inversion recovery (FLAIR) is commonly used to detect WMH [13,14]. Studies of brain structure and function, including those addressing WMH, are limited in children on cART [13]. One study from our research group found a higher WMH volume in PHIV children compared to healthy matched controls [3], whereas a recent case-control study performed in Zambia found no difference in WMH prevalence [15]. Another study, without a controlled comparison, reported a 50% prevalence of WMH in a group of children who were investigated because of suspected HIV-related brain disease [16].
In contrast to children, more studies have investigated brain structure, including WMH, in adults. In our own study, including middle-aged adults with suppressed viremia on cART, we observed a higher total WMH volume compared to controls, which was associated with poorer global cognitive function [5]. In another study, we did not find a significant difference between adults living with HIV and controls in the change of WMH volume during two years of follow-up [17]. One study reported HIV-status to be an independent risk factor for cerebral small vessel disease, including WMH [4]. Finally, two smaller studies, however, did not report HIVstatus to be associated with prevalence and volume of WMH [18,19].
Thus far, no studies have attempted to compare prevalence, number, volume and distribution of WMH between children and adults living with HIV. To increase pathophysiological understanding of cerebral WMH in people living with HIV, we explored their possible relationship with age and mode of HIV acquisition, by comparing the presence, size and location of WMH between our ongoing cohort studies of children and adults living with HIV. We hypothesized that WMH would be more prevalent and differently distributed in adults compared to children, and that adults would have a greater degree of periventricular WMH in view of the known relationship with age [20].
Study design and participants
We used anonymized data from participants at cohort entry from two cohort studies: the Neurological, cOgnitive and VIsual performance in perinatally HIV-infected ChildrEn (NOVICE) study and the neuroimaging substudy of the AGE h IV cohort study [3,5]. The NOVICE study investigates neurological, cognitive and visual performance in PHIV-infected children compared to controls matched for age, sex and socio-economic status [21]. The AGE h IV cohort study is an ongoing prospective cohort study evaluating the occurrence of age-related comorbidities in HIV-1-positive and demographically and lifestyle-similar HIV-negative adults 45 years of age or older, the details of which have been previously published (22). A neuroimaging substudy evaluated neurocognitive outcomes in a subset of male participants who had sustained HIV-viremia suppression for at least 12 months [5].
The following exclusion criteria had been used for both cohorts: current or past neurological or psychiatric disorders not associated with HIV, a history of traumatic brain injury resulting a loss of consciousness of more than 30 minutes, intracerebral neoplasms and MRI contraindications including metal implants or claustrophobia; the complete criteria for noninclusion have been previously published in detail [3,5]. The ethics committee of the Amsterdam University Medical Centers approved both study protocols. Both cohorts obtained written informed consent from all participants older than 12 years and from parents of participants younger than 18 years of age. The AGE h IV cohort study is registered at Clinical-Trials.gov (identifier NCT01466582) and the NOVICE cohort study at the trial Dutch Trial registry (identifier NTR4074).
Demographic and HIV-related variables
Data on historical HIV-and-cART related parameters were provided by the Dutch HIV Monitoring Foundation [21,22]. Additionally, adult participants completed an extensive standardized questionnaire concerning a wide range of demographics, medical characteristics and lifestyle-related factors [23].
MRI data acquisition
At enrollment one MRI scan with different modalities was performed in all pediatric and adult participants initially with a 3.0 Tesla Intera system and subsequently with a 3.0 Tesla Ingenia system (Philips Healthcare, Best, the Netherlands) due to a scanner software upgrade. Threedimensional fluid-attenuated inversion recovery (FLAIR) scans were performed to image both periventricular and deep WMH, using the following MRI scanning parameters: echo time = 356 ms in children and 355 ms in adults; repetition time = 4,800 ms; inversion time = 1,650 ms; field of view = 240 × 240 mm 2 in children and 250 × 250 mm 2 in adults; 321 sagittal slices of 0.56 mm thickness; 1.0 × 1.0 mm 2 in-plane resolution in children and 1.1 × 1.1 mm 2 in adults (3,5).
MRI image processing
We anonymized all MRI data prior to analysis and evaluated them for eligibility, while excluding those in which excessive head motion occured. A neuroradiologist reviewed all MRI scans for incidentalomas [3,24]. For the purpose of the current analysis, one investigator (JvG) reviewed all FLAIR image segmentations and manually adjusted these segmentations (by labeling and delabeling WMH [Fig 1]) for false negative or false positive lesions using the segmentation software ITK-SNAP (version 3.2, Philadelphia, PA; Salt Lake City, UT, USA) [25].
We assessed the following outcome variables: total volume of WMH, total number of WMH and number of WMH per location (periventricular vs. deep). We assessed the number of WMH by manually counting all individual lesions within an MRI FLAIR. We obtained the total segmented WMH volume through ITK-SNAP. We assessed the location of WMH by defining lesions to be either periventricular or deep WMH. We defined lesions as periventricular lesions when they were located within 10 mm of the lateral ventricles [26]. All other WMH were defined as deep.
Statistical analysis
We performed statistical analyses using R version 3.5.1 (R Core team, Vienna, Austria) [27]. We considered p < 0.05 as statistically significant.
We compared demographic variables between children and adults using the Mann-Whitney U test for continuous data and Fisher's exact test for categorical data.
We adjusted WMH volume for intracranial volumes (ICV), i.e. WMH to ICV ratio, for direct comparison between groups. The ICV had been previously calculated automatically [3,5]. Due to non-normal distributions, we dichotomized WMH volume and the number of WMH per location based on the median level. We used univariable logistic regression to determine associations of high volume and high number per location of WMH, with the following demographic, HIV-and cART-related characteristics: age, gender, hypertension in adults and measured high blood pressure in children, known duration of HIV infection, CD4 + T-lymphocyte nadir, and duration of antiretroviral treatment. We defined high blood pressure in children based on measured blood pressure according to American Association of Pediatrics guidelines [28]. Hypertension in adults was defined as a diastolic blood pressure � 90 mmHg and/or systolic blood pressure � 140 mmHg, or use of antihypertensive medication. We performed additional analyses in children to asses the associations between location of WMH and adoption status, total Intelligence Quotient (IQ) and severity of HIV history defined by Center of Disease Control and Prevention (CDC) classification.
Given the low numbers of participants included in analyses, the risk of obtaining unrealistically high odds ratios (OR) and inflated type I error, i.e. "sparse data bias" [29], is increased. In order to mitigate this bias, we applied penalized regression through the use of data augmentation techniques [29,30]. Briefly, we constructed non-informative prior distributions of the OR based on an F(2,2) distribution (i.e. OR = 1; 95%CI = 1/39-39), which were appended to the actual study data to estimate a posterior OR. This estimate allows ORs to be anchored on realistic estimates in the presence of sparse-data. As the amount of data increases, the influence of the prior distribution becomes weaker. A posterior OR and its associated 95% confidence interval were calculated using the "R PLRDA" and "plogit" programs in R [29]. Multivariable analysis was precluded by the low numbers of participants.
Participant characteristics
From a total of 35 children and 74 adults included, we excluded four scans from NOVICE participants because of excessive head motion leading to poor quality. Thus, we analyzed data from 31 children and 74 adults. Table 1 shows the characteristics of participants at time of enrollment into their respective cohort. Due to missing data on ICV, we excluded one child and two adults from the analysis of adjusted total WMH volume for which the WMH to ICV ratio was required.
The median age at enrollment was 13.8 years (IQR 11.4-15.9) in children vs. 53.4 years (IQR 48.3-60.8) in adults. The median age at HIV diagnosis was 1.2 years in children (IQR 0.6-4.9) vs. 42.6 years (IQR 37.2-46.2) in adults. The overall duration of antiretroviral treatment was comparable (p = 0.564) with a median of 11.8 years (IQR 7.7-14.5) and 10.9 years (IQR 4.3-14.9) in children and adults, respectively. All 74 adults had an undetectable viral load with < 200 copies/ml. Three (10%) of the children were not on cART at time of enrollment due to personal reasons and four children (13%) had a detectable viral load (range: 5485-188525 copies/ml). In the five children in whom we measured a high blood pressure, the cause of the hypertension was stress related and thus interpreted as incidentally high blood pressure. There was no treatment initiated to lower the blood pressure. Additional participants' characteristics-used as variables to assess for determinants in children-are provided as S1 Table. WMH assessment Adults had a higher prevalence of WMH compared to children. Adults had more WMH and a higher total WMH volume compared to children. Whereas the prevalence of deep WMH was Statistical tests a = Mann-Whitney U test b = Fisher's exact test ¤ = age in years ͒ = two adults excluded for not filling in their questionnaire;~= in children this is measured high blood pressure at enrollment and defined per guidelines of American Association of Pediatrics ω = age-adjusted Z score # = measured in cells/μL comparable between adults and children, the median number of deep WMH was higher, albeit non-significantly, in adults. Adults had both a significantly higher prevalence and median number of periventricular WMH than children (Table 2). In the pediatric participants two children had a history of HIV-encephalopathy and one child had an opportunistic Cytomegalovirus-encephalitis. A sensitivity analysis without these three participants did not alter the results in Tables 2 and 3 (S2 and S3 Tables).
Determinants of WMH volume and location
In children, female gender was associated with a lower prevalence of deep WMH. We found no other significant association in children between patient-or HIV-related characteristics and WMH volume or location (Table 3). In adults, higher age was associated with an ICV-adjusted WMH volume above the median. In univariable analysis, higher age and hypertension were associated with a number of periventricular WMH above the median. A higher CD4+ T-lymphocyte nadir (per 100 cells/μL) was associated with a number of periventricular WMH below the median (Table 4).
Discussion
This study compared characteristics of cerebral WMH between children and adults living with HIV. Compared to children, adults showed a higher prevalence of WMH and had more WMH, regarding volume and number. This study further demonstrated that adults had both periventricular and deep WMH, whereas children had predominantly deep WMH.
The prevalence of WMH in children observed herein is similar to what was reported in another study, despite their inclusion of younger children who were suspected of having HIVassociated neurological complications, and were scanned with a less sensitive MRI [16]. Apart
PLOS ONE
Comparison of cerebral white matter hyperintensities between children and adults with a treated HIV infection from our own earlier report [3], a recent study performed in Zambia found a comparable prevalence of WMH (18% vs. 24%) between PHIV children and HIV-negative controls, respectively [15]. The difference in prevalence might be explained by different study population and lower strength of magnetic field. Besides the prevalence, this study does not provide further specification of WMH, such as volume. Uncontrolled studies of HIV-negative children with other morbidities such as migraine, have reported a lower prevalence of WMH (up to 17%) [31,32]. We found no association between WMH and age at treatment initiation in children, Table 3. Univariable logistic regression analyses on ICV-adjusted WMH volume and location in children. Univariable logistic regression models with applied penalized regression using data augmentation. ICV-adjusted WMH volume in children was dichotomized by median split. High blood pressure measured at enrollment defined per guidelines of American Association of Pediatrics; CD4 + nadir Z score is age-adjusted. Table 4. Univariable logistic regression analyses on ICV-adjusted WMH volume and location in adults. which is of interest as white matter maturation is at a crucial stage in early life [33]. We have incomplete data of possible disruption of treatment in the first two years of life, due to the majority of children arriving in the Netherlands at older age. The median WMH volume observed in our adult cohort is similar to what has been reported in another study investigating a comparable group of adults living with HIV [17].
Number of periventricular WMH in adults
A notable finding from our analysis was the increased prevalence and number of periventricular, but not deep, WMH in adults compared to children. As both groups have a similar duration of known HIV infection, the observed prevalence and number of periventricular WMH in adults is likely at least partially explained by the higher age and higher prevalence of hypertension amongst our adult cohort participants [20]. The number of periventricular WMH in adults should be interpreted with caution, as some of these lesions tend to coalesce and are larger than ten millimeters, i.e. the commonly used border to distinguish between periventricular and deep lesions. In contrast to adults [4,34], we failed to identify studies using visual rating scales (VRS), such as the Fazekas scale, to assess WMH in children. This is the reason for not using this scale in our study.
Children have predominantly deep WMH and due to this apparent difference with adults, we performed additional analyses in children to further explore potential associations not investigated in our previous study [3]. It could be hypothesized that cross-sectional associations might exist between certain risk factors (i.e. CDC classification or adoption status) or even clinical outcomes (total IQ), and prevalence of deep WMH; however, we found no such association.
The association with CD4 + T-lymphocyte nadir and periventricular WMH in adults suggests a possible role of HIV in the pathogenesis of these lesions. Previously published studies comparing adults with age-matched controls have generally found that HIV-positive status is independently associated with a higher WMH volume [4,5,17]. Two smaller studies however did not report similar findings [18,35]. Even though the HIV-associated pathophysiology of WMH is incompletely understood, evidence suggests that WMH in HIV-negative participants are a consequence of small vessel disease [36]. Periventricular WMH are associated with older age and higher systolic blood pressure [37]. The etiology of WMH that are specifically localized deeply is not yet unraveled, though differences in clinical outcomes justify its distinction from periventricular WMH [38].
In our analysis, we found male gender to be a positively associated with the presence of deep WMH in children. The few studies investigating WMH in children have not investigated gender-associated WMH differences [3,16]. This finding might suggest a potential protective role of estrogen on brain development. This hypothesis is supported by another study which found higher rates of progression and volume of WMH in older (postmenopausal) HIV-negative women as compared to men [39]. However, our finding could also be the artifact of inflated type 1 error from sparse data, yet we did apply penalized regression techniques to reduce this bias. Nevertheless, larger studies would be needed to further investigate the potential association between gender and WMH in HIV-infected individuals.
Strengths and limitations
A strength of our analysis is that advanced non-invasive MRI imaging in both groups was performed using the same technique. We re-assessed all the segmentation to further harmonize the methodology. Our finding that the presence and location of WMH differs between children and adults living with HIV provides further insight into the distribution of WMH in these groups, hinting toward a different pathogenesis. Our study also has a number of important limitations. Although the independent association between positive HIV-status and higher WMH volume has been reported previously, the question of whether the difference in WMH distribution observed between children and adults, is associated with HIV remains unanswered, as we did not include HIV-negative participants. However, we aimed to exploratively assess the distribution of WMH in HIV-positive participants and hence we explicitly did not include a group of HIV-negative participants. As common to all cross-sectional studies, we were unable to establish whether the identified factors are causally linked to WMH [40]. Although we have recorded the medical history and for children the CDC stages as much as possible, medical historical information (e.g. prematurity or infections at early age) lacks from adults, and from children who arrived in The Netherlands at older age. Furthermore, the small number of pediatric participants reduces the power to detect risk factors. Lastly, the adult group contains only males and number of children was rather limited, both of which could affect the generalizability of our findings.
Conclusions
In our cross-sectional assessment, we observed that adults have a higher prevalence, volume and number of total WMH compared to children living with HIV. This study suggests that the location of WMH differs between children and adults, suggesting a potentially different underlying pathogenesis. Future investigation including female adults could further explore potential gender differences in WMH prevalence. Nowadays, effectively treated children living with HIV in high resource settings are expected to live far into adulthood [41]. Therefore, it would be of interest to continue follow-up of these individuals during their life course and compare them to matched HIV-negative controls as well as to others who behaviorally acquired HIV. This information could help shed light on the evolution of WMH and its possible clinical effects.
S3 Table. Univariable logistic regression analyses on ICV-adjusted WMH volume and location in children (exclusion of three participants with a history of CNS infection). (DOCX)
We thank S. Cohen, J.S. ter Stege and C. Blokhuis (Emma Children's Hospital) for data acquisition and overviewing the NOVICE I study.
Finally, we thank all participants without whom this study would not have been possible.
Author Contributions
Conceptualization | 2020-10-29T09:02:28.787Z | 2020-10-28T00:00:00.000 | {
"year": 2020,
"sha1": "56e9b786ee4b0b9b32d6cd6274e7e2b4112b63da",
"oa_license": "CCBY",
"oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0241438&type=printable",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "d743aafa2f9dbcb912192e5333dc647c22ce35f3",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
187057396 | pes2o/s2orc | v3-fos-license | Factors Influencing Satisfaction on E-Banking
Information technology has made enormous changes in to the way in which the business enterprises operate today. This is very much true for service industries like banks which experienced tremendous change in their business process. On the other side, due to LPG (Liberalization, Privatization and Globalization) era, customers especially bank customers have become more demanding and choosy. In this competitive world, the success of business will be based on how effectively they utilize the information technology tools for satisfying the needs of the customers at the right time, right place and at the right cost. There are number of delivery channels available in the banking industry to reach the customers. The widely used technologies like ATM, Internet Banking, Mobile Banking, SMS banking etc are satisfying the needs of the customers. But the effectiveness of usage of these tools decides the success of the functioning of banks. This research was an attempt to identify customer satisfaction on the e-banking services offered by banks with special reference to online banking. This was assessed with the help of five dimensions (Security, Performance, Training, Reliability and Ease of use) identified through various review of literature. A questionnaire was constructed and circulated and finally 250 questionnaires with complete responses were included in the study to identify their satisfaction about e-banking services provided by the banks in Tamil Nadu. The data collected were analyzed using the statistical tools like Multiple Regression, Chi Square and Simple Percentage analysis. Based on the results, suggestions were given to bankers for offering the best e-banking services to their customers.
Introduction
In this era of intense competition, only the companies which understand the needs of the customers and satisfy the needs alone can survive. Customers are important assets for any business. Companies should adopt right strategies to satisfy the needs of their customers. This process has become very easy in today"s environment because of the enormous growth of information technology tools that can be used in business for satisfying the needs of customers at the faster rate. Service industries especially the banks are taking lot of initiatives to make the best use of information technology tools and techniques. There has been an increase of competition in the banking industry that made other banks try to develop different approaches to electronic banking. In order to remain competitive in the market and for having a larger share in the market a company needs to have loyal customer base. These important issue prerequisites cannot be achieved except by earning their trust. (Akbari et.al. 2012). So high level of customer satisfaction will be essential for the sustainability of banks and this could be achieved when the customer satisfaction by new upgraded services is measured and also the type of banking service that is able to attract customers more than the other competes is recognized. Factors that influence on adoption of various technologies and information technology are different depending on the desired technology of studied users and the existing situation. Customer satisfaction towards internet banking service has become an issue due to stiff competition among the banks in India. Customer satisfaction is critically important for its impact on customer retention and firm profitability. Internet banking service quality and customer satisfaction are inarguably the two core concepts for the researchers. It has been identified through various researches that quality of internet banking service is a key issue to maintain customer satisfaction. (Sharma et al., 2014). Even though many studies have been conducted to identify the customer satisfaction on e-banking, Tamil Nadu level study to analyze the customer satisfaction on e-banking with the widely adopted dimensions, will help the bankers to formulate strategies for improving their quality of services on ebanking.
Theoretical Background
E-banking or Internet Banking or online banking is a convenient way to do banking from the comfort of any one"s home or office. According to Oxford Dictionary ebanking is a method of banking in which the customer conducts transactions electronically via the Internet. (Source: http:// www.oxforddictionaries.com /definition /English/e-banking). Users need not wait in a long queue in front of counters rather can do all the transactions from their convenient place. According to Daniel (1999), electronic banking is the delivery of banks " information and services by banks to customers via different delivery platforms that can be used with different terminal devices such as personal computer and mobile phone with browser or desktop software, telephone or digital television. Electronic banking, therefore, could be categorized into PC banking, Internet banking, TV-based banking, and Telephone based banking. The SBI Retail Internet Banking offers a plethora of products and services, to cater to all the banking demands of the customers like Transfer funds to own and third party accounts, A suite of completely online deposit products (Fixed, Recurring, Flexi, Tax Saving etc.), Airline, Rail, Bus and hotel ticket booking, Online Shopping and instant recharge features, IMPS Funds Transfer, Western Union Service, Credit beneficiary accounts using RTGS/NEFT feature, Generate account statements, scheduling payments, Configure profile settings, E-Tax for online tax payment, E -Pay for automatic bill payments etc., The Corporate Internet Banking (CINB) facility of SBI enables the corporate customer to carry out banking activities anywhere and anytime aided with the power and convenience of the internet. According to ICICI web site, Internet Banking is a convenient way to do banking from the comfort of the customers" home or office. They can avoid the queue or delays by trying their simple and secure Internet Banking facility for an unmatched online banking experience. To experience the internet banking customers have to just login using their user id and password.
Customer Satisfaction (CS) has become a major area of marketing that has received considerable publications from practitioners and scholars in the last two decades. According to Kotler (2016), satisfaction reflects a person"s judgment of a product"s perceived performance in relationship to expectations. If performance falls short of expectations, the customer is disappointed. If it matches customer expectations, they are satisfied and if it exceeds them the customer is delighted. In short satisfaction is a person"s feeling of pleasure or disappointment resulting from comparing a product"s performance (outcome) in relation to his or her expectation. Customer satisfaction can be defined as the overall customer evaluation of a product or service based on purchase and consumption experience over a time period. The quality of services provided by the bankers in the banking sector plays important role in the satisfaction of customer. The increase in customer satisfaction level leads to decrease in operating and service costs. This provides an opportunity for banks to expand their product portfolio and services for the future development. (Sanyal, 2016). Different researchers have attempted to measure the customer satisfaction on e-banking services using different dimensions like Safety, Reliability, Transactions efficiency, Customer support, Service Security, Ease of use, Performance, Accuracy, Timeliness, Confidence, Content, Format, Responsiveness etc.
Literature Review
The customer user satisfaction concept was applied in the area of information systems in the early 1980s. Different authors have studied the concept of customer satisfaction on e-banking using different dimensions. Bailey and Pearson (1983) have developed a measuring instrument to analyze computer user satisfaction. This instrument was adopted by researchers and practitioners for high validity and reliability. They also have identified 38 factors affecting user satisfaction and developed a questionnaire using 7-point semantic differential scale of adjective pairs. Later Torkzadeh and Doll (1988) have modified the instrument from previous studies and reduced the factors into five factors with 12 items. The final developed factors were, content, accuracy, format, ease of use and timeliness. In 1991 they used the term End User Computing Satisfaction as "the affective attitude towards a specific computer application by someone who interacts with the application directly". This means that the interaction could be towards any computer-based systems that has human-computer interaction.
In 2004 Pikkarainen model is proposed by Pikkarainen and colleagues which examines the factors that influence the Internet use of services or goods. The researchers announced six effective factors on Internet transactions of goods and services (usefulness, ease of use, Enjoyment, Providing information, security and confidentiality, Internet connection quality.) This model was designed and tested for acceptance of electronic banking and the model is derived from the Technology Acceptance Model. (Davis.F.D. 2000) Salehi, Rostami, and Mogadam (2010) have studied computer user satisfaction to end user in AIS environment. They have used a modified version of Doll and Torkzadeh (1988) instrument and tested for its validity and reliability. This paper validates accuracy, ease of use, timeliness, content, format, system speed and system reliability. There were many attempts to simplify the factors and to suit it with studies related to computer user satisfaction. Kumphar (2011) conducted a study on demographic profile of customers and its relationship with perception of service quality, service value and overall satisfaction in e-banking. The study showed that demographic characteristics like age, education, profession and income level are influencing perceived service quality, perceived value and overall satisfaction in e-banking. So the author recommended the banks to adopt the right strategies as per the demographic profile of the customers.
Gupta and Bansal (2012) tried to develop an instrument for measuring Internet banking service quality in India and they also analyzed the impact of Internet banking service quality dimensions on the overall Internet Banking Service Quality (IBSQ) and customer satisfaction. They developed a service quality scale of 22 items. Using exploratory factor analysis they reduced the 22 items into five dimensions: Security/Privacy, Reliability, Efficiency, Responsiveness, and Site Aesthetics. Model was further validated through Confirmatory Factor Analysis. They carried out a survey among the Internet banking users of Private, Public and Foreign banks in Delhi. Multiple regression analysis revealed that security/Privacy dimension carries the maximum impact on the overall IBSQ whereas customer satisfaction is mostly affected by the Efficiency dimension as compared to other dimensions.
Rangsan Nochai and Titida (2013) conducted a survey to examine the impact of seven internet banking service quality dimensions on customer satisfaction using multinomial logistic regression analysis. A survey was done on customers of top three banks in the Bangkok who use Internet Banking. Results showed that Safety reliability, Transactions efficiency, Customer support, Service security, Ease of use and Performance have a significant impact on customer satisfaction.
Sharma & Malviya (2014), conducted a study on Internet banking service quality and its impact on Customer Satisfaction in Indore district. They developed a model based on service quality dimensions like ease of use, comfort, accessibility, confidence and responsiveness with the purpose to investigate the impact of service quality on customer satisfaction. Through multiple regression analysis they concluded that all five dimensions were having impact on customer satisfaction. Ease of use, comfort and accessibility were the most important dimensions which influence the customer satisfaction on e-banking. Based on the above reviews, it was clear that different authors used different dimensions to test the customer satisfaction on e-banking. From the above reviews, the five dimensions namely Security, Performance, Training, Reliability and Ease of use were selected to assess the customer/user satisfaction on e-banking. The following study was conducted to identify the relationship between different dimensions of e-banking and the customer satisfaction on e-banking. The study also tried to identify the relationship between demographic profile of the respondents and customer satisfaction on e-banking. (Kumphar 2011).
Objectives
To identify the relationship between e-banking dimensions and customer satisfaction on e-banking. To find the relationship between demographic profile of the respondents and customer satisfaction on e-banking. The broad research question is What are the dimensions affecting the customer satisfaction on e-banking?
Figure 1 Relationship between Selected E-Banking Dimensions and Customer Satisfaction
Based on the framework, the proposed hypotheses are H1: Security will have an impact on Customer Satisfaction on E-banking H2: Performance will have an impact on Customer Satisfaction on E-Banking H3: Training will have an impact on Customer Satisfaction on E-Banking H4: Reliability will have an impact on Customer satisfaction on e-banking H5: Ease of use will have an impact on Customer Satisfaction on E-Banking H6: Demographic profile of the respondents" will have an impact on customer satisfaction on e-banking.
Methodology
Descriptive research design was used for the study. Both primary and secondary data was collected. Structured questionnaire was used as a source of primary data and apart from this some relevant secondary information were obtained from news papers, magazines, records, websites and books. The population of the study was the bank customers in Tamilnadu who uses e-banking/internet banking services. A pilot study was done by administering questionnaire on 50 respondents selected at random. As per the pilot study, minor changes were made into the questionnaire before carrying out the final data collection. The questionnaire was divided into two parts. First part includes questions related to demographic profile of respondents and second part includes the respondents" opinion on e-banking in the service quality dimensions like Security, Performance, Training, Reliability and Ease of use.. The respondents have given their opinion in a five point Likert Scale (5-Strongly Agree, 4-Agree, 3 -Neutral, 2 -Disagree, 1 -Strongly Disagree). The population for the study was list of bank customers who use e-banking facilities in Tamilnadu. The study was undertaken among both public and private sector bank customers, who use internet banking facility. Data was collected from Individual customers of the public and private sector banks who use internet banking services by logging into the banks online banking website using the Use rid and password provided to them by the bank using a desktop or a laptop. Samples were selected using non probability Snow-Ball sampling. 250 complete questionnaires collected from the respondents were included for analysis. The data collection was done during Nov 2013 -Mar 2014. Cronbach"s Alpha Test of Reliability is applied to test whether the items pertaining to scale, each items were internally consistent and whether they can be used to measure the proposed phenomena e.g. e-Banking dimensions like Security, Performance, Training, Reliability and Ease of use. All the values were found to be more than 0.8. The data analysis like Simple percentage analysis, Multiple Regression and Chisquare analysis were carried out using SPSS. The demographic profile of the respondents was analyzed using simple percentage analysis. The hypotheses related to the impact of e-banking five dimensions selected for the study like Security, Performance, Training, Reliability and Ease of use on Customer Satisfaction was tested using Multiple Regression Analysis. The relationship between demographic profile of the respondents and Customer Satisfaction was tested using Chi-square analysis. The results of the analysis were tabulated and interpreted as shown below.
Simple Percentage Analysis
The respondents were asked about their demographic details like age, gender, education, computer knowledge, occupation, income and period of relationship with the banks. The distribution of respondents on the basis of demographic profile was tabulated below.
Multiple Regression
A regression is a statistical tool used to find out the relationship between two or more variables. In simple regression, there will be only two variables; one variable is caused by the behavior of other variable. The variable, which is causing the behavior of other variable, is called independent variable, and the variable which is caused by other variable, is called dependent variable. When there are two or more independent variables, which affect the behavior of one dependent variable such relationship is called multiple regression or linear regression. Linear regression estimates the coefficients of the linear equation, involving one or more independent variables that best predict the value of dependent variable. H1 -H5: Security, Performance, Training, Reliability and Ease of use will have an impact on customer satisfaction on e-banking. Model: Here customer satisfaction was taken as the dependent variable and the five dimensions of e-banking were the independent variables. The OLS Regression model was used to determine the significance level of the variables for the customer satisfaction on e-banking.
The basic model was as follows.
Customer Satisfaction on e-banking= α + ß1* Security + ß2* Performance+ ß3* Training+ ß4* Reliability + ß5 * Ease of use There α is constant and ßs are coefficients to estimate, and e is the error term. The overall regression model and its ANOVA were summarized as follows The ANOVA Test shows the p value -0.000 which was less than 0.05, hence the result was significant at 5% level of significance. It means there was a significant correlation between dependent variable and independent variables. Therefore, customer satisfaction level depends on different dimensions like Security, Performance, Training, Reliability and Ease of use. The overall predictability of the model was shown in the Model summary above. The adjusted R square value of .503 indicates that the model explains roughly about 50.3% of the variability in the dependent variable customer satisfaction on e-banking by the independent variables Security, Performance, Training, Reliability and Ease of use. The remaining 49.7 percent of the variation in customer satisfaction may be due to reasons not related to e-banking dimensions considered in the model. Table 4 depicts significant F values which implies that the model and data were well fit in explaining customer satisfaction in e-banking. Based on the data found in the above Table, it could be interpreted that the independent variables such as Security, Performance, Training, Reliability and Ease of use have strong impact on customer satisfaction. The strongest relationship was between Ease of use and satisfaction followed by Security and satisfaction and Training and satisfaction. The following Table shows the results of hypothesis testing. From the above findings the following regression model was formed: Customer satisfaction on e-banking = 6.071+0.221* Security+0.106* Performance+0.187* Training +0.156 * Reliability +0.713 * Ease of Use From the above model it could be inferred that Ease of use was the dimension which brings about major change in the customer satisfaction on e-banking followed by Security and Training. The last two dimensions were Reliability and Performance. The study identified that ease of use was very important e-banking dimension that affect the customer satisfaction followed by security. So suggestion was given to the banks that e-banking working should be made easy without compromising on security at any cost since that was the second important dimension that affect the customer satisfaction on e-banking. Suggested changes include explaining and enforcing strong passwords, using the same code for temporary and permanent identification, and using a consistent terminology throughout the user interface. (Morten et al., 2004). Banks should take necessary steps to offer the user friendly ebanking facilities with good security. They should do some research on identifying the best technology for combining Ease of use with Security.
Chi-Square Analysis -Demographic Profile of the Respondents and Level of Satisfaction
In order to assess the relationship between the demographic profile of the respondents like age, gender, education, computer knowledge, occupation, income and period of relationship with banks and the level of satisfaction of customers on ebanking a hypotheses was framed and tested using chi-square analysis.
H6: Demographic profile of the respondents will have an impact on customer satisfaction on e-banking. From the above chi-square analysis it is implied that the demographic profile like Age, Educational qualification, Computer knowledge, Occupation, Income and Period of relationship with bank have a positive relationship with level of satisfaction of bank customers on e-banking but Gender was not having any relationship with level of satisfaction on e-Banking.
Hence it can be concluded that the demographic profile of respondents like age, educational qualification, computer knowledge, occupation, income and period of relationship with bank are having a relationship with customer satisfaction on ebanking where as gender is not having any relationship with the same.
Hence the banks should devise the right strategies to attract the different segment of customers with different age group, educational qualification, computer knowledge, occupation, income and period of relationship with bank. Normally, young customers expect more service from e-banking since they are familiar with all latest services. So, banks should enhance their services to their expectations at the same time expectations of the elder customers also should not be ignored. Highly educated people will have the skill to operate with banking technology than the people with low education. Banks should consider this while designing e-banking services. Naturally people with good computer knowledge will feel it easy to work with technology. The people in different professions and different income groups may have different needs and services from e-banking. Banks should develop their ebanking services according to the expectations of people in different occupation and income range. (Kumphar, 2011). In this study, period of relationship with banks also has an impact on customer satisfaction. The customers with long term relationship will definitely have good customer satisfaction. Banks can utilize the e-banking services to retain the customers for long period. Banks should try to improve the ebanking services to attract the differentiated group of customers on the basis of age, income, education, computer knowledge, occupation, period of relationship with bank. They should design the e-banking services based on the needs of the different segments of people.
Conclusion and Limitations
To compete in this competitive marketplace and to sustain the competitive advantage, each and every organization has to adapt to the latest technologies, should identify the modifications to the existing technology by conducting research. This research seeks to make an original contribution to knowledge by investigating the impact of five e-banking dimensions like Security, Performance, Training, Reliability and Ease of use on Customer Satisfaction on e-banking. To introduce the security in e-banking services, the banks can incorporate new technologies like encrypting and verification. The security issues can be moderated through improved authenticating process, protecting password, information about site security, giving absolute loss guarantee (if something goes wrong during transaction banks should correct the errors without much delay), providing accessible customer service (multiple channels like SMS, E-mail and Toll free numbers should be given for customers to file their complaints) and educating customers on features of e-banking. (Maryam, 2013). Most of the studies showed that if Ease of use and Security is high, customer satisfaction on e-banking will increase. This is proved in this study and this will also increase the loyalty of customers towards bank and it will create positive Word of Mouth (WOM) in the context of e-banking services. The e-banking dimensions and its impact on customer satisfaction identified through this study will allow bank managers and policy makers to direct their efforts and resources in most effective and efficient way through which bank business can increase in long run and new customers can be encouraged to adopt and use e-banking. The study revealed that demographic profile of the respondents like age, education, computer knowledge, occupation, income and period of relationship with the bank are having relationship with e-banking satisfaction. Hence banks should give importance to segmenting their customers and offering products as per the needs of the segments. Adoption of e-banking by most of the customers will reduce the operational cost involved in bank transaction which will in turn help the banks to boost their profits. So this study can be used by bankers in deciding their e-banking practices which will increase the customer satisfaction and loyalty of bank customers. Since the study was conducted in Tamilnadu, the findings of the study could not be generalized. The study did not make any comparison between the adoption and satisfaction of ebanking between the public and private sector banks. The focus of the study was only on internet banking and other e-banking services like ATM, Mobile Banking, SMS banking etc., were not included in the study. Further studies can be conducted focusing on above areas. | 2019-06-13T13:13:57.949Z | 2017-05-01T00:00:00.000 | {
"year": 2017,
"sha1": "eb0dbb33f6f45dc492eee803ca3ea6b3d8470f60",
"oa_license": null,
"oa_url": "https://doi.org/10.26573/2017.11.2.2",
"oa_status": "GOLD",
"pdf_src": "MergedPDFExtraction",
"pdf_hash": "65d73e101543a8076eab5807b2da8f2bc7a6224f",
"s2fieldsofstudy": [
"Business"
],
"extfieldsofstudy": [
"Business"
]
} |
258830979 | pes2o/s2orc | v3-fos-license | Knowledge gap in a cross section of Irish general practitioners prescribing denosumab for osteoporosis
Background Denosumab is commonly used by general practitioners (GPs) in Ireland to treat osteoporosis though drug holidays are not recommended with rebound bone loss and risk of vertebral fractures if stopped. We aimed to investigate GP practice and knowledge regarding denosumab including use and reasons for use, therapy duration, blood monitoring and recommended vitamin D status/calcium intake on treatment, staff administering, methods of recall, delays in receiving injections, management of and awarenes of guidelines if stopped, reasons for stopping and concerns about same. Methods GPs were contacted (n = 846) by email and invited to complete an online anonymous survey comprising 25 questions in January 2022. We collated responses and explored for differences between GP principals/trainers and GP trainees. Results There were 146 responses. Sixty-seven percent were female and 50% were GP principal/trainers. Forty-three percent used denosumab as a first line therapy citing convenience in 32% of cases. Half (50%) envisaged therapy for 3–5 years and 15% lifelong use. A fifth (21%) had no concerns about it being stopped (11% trainors vs 31% trainees, P = 0.002). If stopped, 41% cited opting for a drug holiday with monitoring. Forty percent of GPs gave patients a reminder card for the next injection and 27% had an alert system. Conclusion We identified a knowledge gap in denosumab prescribing among a sample of Irish GPs. Findings suggest a need for education to increase awareness around denosumab use and to consider recall systems in GP practices as suggested elsewhere to ensure persistence with therapy.
Introduction
Denosumab is commonly used by general practitioners (GPs) in Ireland in the treatment of osteoporosis [1].It is a monoclonal antibody administered twice yearly as a subcutaneous injection and is often used in patients with contraindications to bisphosphonates including gastro-oesophageal reflux disease and renal impairment (eGFR < 30 ml/min) [1].However, studies suggest that a significant proportion of Irish GPs use denosumab as a first line therapy which is generally recommended for bisphosphonates [2].Importantly, unlike bisphoshonates where drug holidays are appropriate for some patients, they are not compatible with denosumab with the anti-resorptive effects wearing off after 6 months, resulting in rebound bone loss and risk of vertebral fractures [1].In fact, all treatment gains in bone density can be lost within 12-24 months of stopping.For this reason, if denosumab is initiated in primary care, consultation with secondary care colleagues may be advisable given the need to have a long-term personalised osteoporosis management plan in place to enable denosumab to be stopped in a managed way, as necessary [3].
Guidelines have recently advised treatment with zoledronic acid if denosumab is stopped [3,4].Oral bisphosphonates could also be cautiously considered in some patients with a short duration of therapy (< 2.5 years) providing there is monitoring with bone turnover markers (BTM) [4].However, access to these options is difficult for GPs and even with follow-up therapy, bone loss can occur in a significant proportion of patients [1,5].For this reason, the decision to prescribe denosumab needs to be carefully considered given that there is no data on safety and efficacy beyond 10 years [1].Indeed, in patients at high risk of fracture, remaining on therapy indefinitely may be appropriate [6].
Finally, in routine clinical practice, GPs should not prescribe denosumab if there is hypocalcaemia and/or a vitamin D level below 50 nmol/l and also need to consider checking calcium levels post administration in high risk patients [7].Clinical monitoring of serum calcium levels prior to every dose is recommended but is not mandatory and clinical judgement should be applied on an individual patient basis.
We aimed to ascertain GP knowledge and practice with regard to denosumab including (1) use in the last year, (2) use as a first line therapy and reasons for same, (3) duration of treatment, (4) monitoring of blood biochemistry and knowledge of recommendations for vitamin D status and calcium intake on treatment, (5) clinical staff administering the injection, (6) methods of recall for repeat injections and reliance on pharmacists for reminders, (7) knowledge of acceptable delay in receiving injections and perceived delays during COVID-19, (8) clinical practice if denosumab is stopped (after 2.5 years), as well as awareness of guidelines and if concerns about same and (9) percieved reasons for stopping denosumab.
Methods
GP principals/trainers and GP trainees registered with the Irish College of General Practitioners were contacted by email and invited to complete an online anonymous questionnaire in January 2022 using SurveyMonkey (see Table 1).This comprised 25 questions detailing information on GP demographics/practice characteristics and on knowledge and clinical practice regarding denosumab therapy (see Table 1).Differences in categorical responses to questions by GPs and practice type were explored with 2 test and signficance accepted when P < 0.05.Ethical approval was granted by the Irish College of General Practitioners (ICGP) ethics committee (Ref: ICGP_REC-2021-T26) (Table 2).
Over one in five (21%) responded that they did not check serum calcium prior to therapy and this was higher for GP registrars compared to trainers (26 vs 16%, P = 0.01).Approximately one in six (17%) cited checking serum calcium routinely after denosumab injections and 72% in patients at risk of hypocalcaemia.The minority (21%) did not assess total daily calcium intake and of 86 GPs who commented, 54% cited a recommended daily allowance in osteoporosis (≥ 1000 mg).Most (71%) considered that patients should have a vitamin D level of ≥ 50 nmol/l prior to starting denosumab, with 9% citing 30-50 nmol/l, 15% > 75 nmol/l and 20% none of these.
In most cases (63%), denosumab was administered by practice nurses only.Over one-third (40%) gave a reminder card for the next denosumab injection which was more likely in GP practices of ≥ 4 compared to 1-3 GPs (32 vs 27%, P = 0.01).Overall, 27% of GPs had alerts to remind patients of the need for repeat prescription.Systems included SMS text (13), phone call (5), email (3) and nurse call (3).Nearly half (49%) noted that they relied on pharmacists 'some or most of the time' to give reminders for a repeat prescription.
Over one-third (37%) were not aware that denosumab should not be delayed by > 7 months from the last injection with this being more likely in GP registrars compared to GP trainers (45 vs 29%, P = 0.04).The majority (71%) felt there was a delay in denosumab administration some or most of the time during COVID-19.There was no difference in the response by GP status (registrar vs trainer) as what to do if stopping denosumab after 2.5 years: 19% were unclear and might refer for specialist opinion, 41% cited a drug holiday and monitoring with DXA and about a third (38%) noted that an oral bisphosphonate or SERM should be started.Overall, 21% had no concerns about denosumab being stopped but this varied significantly from 11% in trainers to 31% in registrars, P = 0.002.About one in five (18%) were aware of any recent guidelines on what do if stopping denosumab.
The main reason cited why GPs felt patients stopped denosumab was due to concern about being on treatment too long (43%), noncompliance (32%), adverse effects (29%) and perceived lack of benefit (14%).Individual comments included that 'dementia patients family may forget to bring their relative to the practice' and that 'some patients become housebound and current resources do not support house calls for administration'.One GP commented that they do not use denosumab anymore as 'it is too risky'.
Discussion
To our knowledge, this is the only study in Ireland that has explored both GP knowledge and clinical practice with regard to denosumab in the treatment of osteoporosis.We identified a knowledge gap in a number of areas, especially with regard to stopping treatment and follow-up therapy.
While the majority of GPs had used denosumab in the preceding year, close to half had prescribed it as first a first line treatment.In a recent study of 1146 Irish patients prescribed denosumab by GPs between 2012 and 2017, over half had no prior bone therapy suggesting a rate of first line use not explained by contraindications to other therapies [2].We found that in a third of the cases, 'convenience' was cited as an indication for first line use; however, no other Irish studies have explored the reasons for denosumab prescribing.Other studies elsewhere have shown that patients have a preference for a six monthly injection compared to weekly tablets, with convenience being identified as an important factor [8].This is likely to impact on GPs who take patient preferences into consideration when prescribing osteoporosis medications [9].About half of GPs envisaged therapy duration for 3-5 years with practice nurses administering the injection in most cases.However, it is unclear how this compares to practice elsewhere or whether the availability of nurses to administer injections could influence GPs prescribing of denosumab.A minority did not check serum calcium prior to the injection though while assessing serum calcium is recommended before drug administration it is not mandatory.Indeed, during COVID-19, some guidelines waivered the advice to check serum calcium in all patients if normal in the previous year due to difficulties with accessing bloods and advised clinical judgement on an individudal basis [10].On the other hand, about one in six checked calcium routinely post injection which is not necessary.Indeed, at the end of denosumab therapy (i.e. at about 6 months or more after the last injection), there may also be a mild hypercalcaemia associated with rebound phenomenon [11] that could inadvertently lead to delay in the next injection.However, importantly, the majority of GPs did check serum calcium post injection in patients at risk of hypocalcaemia.Just over half of GPs cited a daily calcium intake of 1000 mg or more for patients with osteoporosis.By comparison, in a survey of GP knowledge of osteoporosis in the Czech Republic in 2017, 41% were reported to correctly state the recommended calcium intake [12].
Most GPs had no alert systems to remind patients of their next dose and about half cited relying on pharmacists.While persistence with denosumab in Irish patients has been found to be 57% at 2 years, it has been reported to be higher in those with a medical card [2].Medical card holders in Ireland are entitled to medications at no cost, with reimbursement of their dispensing pharmacist by the Irish Department of Health.Pharmacy oversight of these prescriptions might contribute to this better persistence though avoidance of an 'out of pocket' expense is also likely to be an important factor.
A third were not aware of the need for timely denosumab administration (no longer than 7 months after the last injection) suggesting a lack of knowledge among some GPs of current guidelines.This was also more likely in GP registrars who might be less aware of recommendations.
Perhaps not surprisingly, the majority of GPs felt there was a delay in denosumab injections during COVID-19 as has been reported elsewhere [13].Interestingly, the paradigm of drug holiday was considered by 41% if stopping denosumab despite the vast majority having concerns if there was therapy cessation.However, a significant proportion were unclear as to what to do if stopping and might refer for specialist opinion.By comparison, in a recently published Australian study, GPs expressed uncertainty about when to stop denosumab, what to do when stopping, the risk of stopping without an alternative being prescribed, or what should be prescribed if a patient had previously had problems with bisphosphonates [9].
Consistent with the above, we found that the majority of GPs had no knowledge of recent guidelines on what do if denosumab is stopped with just over one-third citing the use of an antiresorptive therapy after cessation.A previous study in Ireland found that 6% of patients who stopped denosumab were started on alternative treatments by their GP [2] while in Australia, this was reported to be less than 20% [9].However, both studies reported on GP practices at a time when knowledge of the phenomenon of rebound bone loss on denosumab cessation was only emerging [1].GPs felt that the commonest reason for patients wanting to stop denosumab was concern about being on treatment too long which is a similar to what as been identifed for other osteoporosis drugs [14].
We acknowledge that only 17% of GPs contacted replied to our survey which could bias the findings.However, the response rate to Irish GP surveys has been identified to be similarily low in other studies with the same methodology [15][16][17][18].The quality or representativeness of a survey also does not necessarily correlate with its size, and a lower response rate does not necessarily make a survey less accurate [19].Furthermore, previous research suggests that GPs with less interest in a topic may be less likely to engage in surveys [20,21].Therefore, this survey could potentially underestimate the knowledge gap identified.
In conclusion, we identified a knowledge gap with regard to denosumab prescribing among a sample of Irish GPs, particularly with regard to cessation of therapy and followup treatments.Our findings suggest that there is a need for education to increase awareness around denosumab use.It also highlghts the need for reminder or recall systems in GP practices so as to avoid rebound fractures [9].
Funding Open Access funding provided by the IReL Consortium.
Declarations
Ethics approval Ethical approval was granted by the Irish College of General Practitioners (ref letter: ICGP_REC-2021-T26).Informed consent was obtained from all participants in the study.
Table 1
GP questionnaire
10 Do
you check a patients serum calcium prior to denosumab?Yes/No 11 Do you check a patients serum calcium after denosumab?(routinely/patients at risk of hypocalcaemia) 12 Do you check a patients calcium intake before prescribing denosumab ?Yes/No 13 What is recommended daily intake of calcium for patients with osteoporosis?Open answer 14 What level of vitamin D [25(OH)D] should a patient have before receiving denosumab?(30-50 nmol/l, 50-75 nmmol/l, > 75 nmol/l, none of these) 15 Who routinely administers patients denosumab in your practice?(GP, nurse, both) 16 Do you give patients a reminder card of when their next dose of denosumab is due?Yes/No 17 Do you have an alert system in place to alert patients when their denosumab injection is due?Yes/No 18 If you have an alert system, can you comment on what you use?Open answer 19 How often do you rely on the patients' pharmacist to remind them of when denosumab is due? (never, rarely or not often, some of the time, most of the time) 20 How long can you delay a patient's 6 monthly denosumab dose?(Less than 7 months, more than 7 months) 21 How often was there a delay in patients receiving denosumab during the COVID-19 pandemic?(never, rarely or not often, some of the time, most of the time) 22 If denosumab is stopped after 2.5 years of treatment, what do you routinely do next?23 Have you any concerns about a patient's denosumab treatment being stopped?Yes/No 24 Are you aware of any recent guidelines on what to do after stopping denosumab therapy?Yes/No 25 In your opinion, what are the main reason(s) for patients stopping denosumab therapy?Open answer
Table 2 (
Total number who answered survey was 146 except where indicated * (smaller sample size quoted).%, percentage calculated out of total survey sample except where smaller sample SERM selective oestrogen receptor modulator | 2023-05-23T06:17:19.051Z | 2023-05-22T00:00:00.000 | {
"year": 2023,
"sha1": "a0c87d06ef933aada18c4a7041860685ea82040f",
"oa_license": "CCBY",
"oa_url": "https://link.springer.com/content/pdf/10.1007/s11845-023-03383-w.pdf",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "ab05f269e3308d4e01c98514b7bd7e1899931ebe",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
25682848 | pes2o/s2orc | v3-fos-license | Combined duplication of the colon and vermiform appendix in an adult patient
Combined duplication of the colon and vermiform appendix is one of the rare congenital anomalities of the alimentary tract. Only a few cases have been reported in the adult population. A 28-year-old man presented to the clinic with a mass in the right flank. Imaging showed only a hydronephrotic atrophic kidney. The final diagnosis was only available at exploration. Combined duplication of the tubular colon and vermiform appendix was confirmed histopathologically. The patient was treated with nephrectomy and complete resection of the duplicated colon and vermiform appendix. The patient recovered uneventfully, and has done well for the past year. This is believed to be one of the first reports of combined duplication of the tubular colon and vermiform appendix as a cause of hydronephrotic atrophic kidney in an adult patient.
INTRODUCTION
Gastrointestinal duplication is rare, although it may be obser ved throughout the alimentary tract. Most duplication is located in the ileum and the presenting symptoms are bowel obstruction, intestinal bleeding and/ or even perforation. Apart from ileal duplication, colon duplication is a rare congenital anomaly [1][2][3][4] . We present a case of combined duplication of the colon and appendix vermiformis, which led to hydronephrotic atrophy of the kidney.
CASE REPORT
A 28-yr-old man with fever (38.9℃) and intermittent right flank pain was referred to the clinic. Anorexia, nausea and vomiting were also noted. Physical examination revealed a palpable mass in the right flank. No sign of peritoneal irritation was noted. He had no history of surgery, urinary tract infection and/or calculi. Leukocytosis and urine infection were only the findings upon blood and urine analysis.
Ultrasonography (US) revealed a heterogeneous and hyperechoic pararenal mass and g rade 3 to 4 hydronephrosis. Intravenous pyelography (IVP) was diagnostic for a non-functional right kidney. 99m Tcdiethylenetriaminepentaacetic acid ( 99m Tc-DTPA) and 99m Tc-dimercaptosuccinic acid ( 99m Tc-DMSA) nephrograms confirmed decreased right renal function. CT showed hydronephrosis in the right kidney, with a mass adjacent to the right renal pelvis. A continuum of the mass with the colon was noted. Colonoscopy was not diagnostic for intraluminal pathology.
During exploration, a severely scarred and inflamed mass at the inferior medial pole of right kidney was observed, which caused hydronephrosis and atrophy of the right kidney. Nephrectomy was undertaken. The continuum of the mass was followed during dissection. A precast peritoneal window and connection of the mass to the patent cecum was observed. There were two tubular lumens at the base of the cecum. Resection of the mass was completed with careful dissection to spare the normal adjacent structures. Cecal opening at the connection of the mass was closed with simple saturation, and a formal appendectomy was performed. The patient recovered steadily and was discharged from the hospital after 10 d, and he progressed well over the next year.
Gross pathological examination revealed a combined tubular str ucture with fecal material inside, and an appendix-like extra tubular structure with an obstructing fecalith in its lumen ( Figure 1). This duplicated appendix was not perforated. Histologically, a colonic mucosa layer with accumulation of infiltrating lymphocyte nests, and a duplicated appendix was confirmed. Pathological examination revealed no evidence of malignancy ( Figure 2).
DISCUSSION
Gastrointestinal duplication, which may occur throughout the alimentary tract, is a rare congenital anomaly [1] . More than 80% of the cases present before the age of 2 years, as acute abdomen or bowel obstruction [1,2] . Most duplication is located in the ileum and presenting symptoms are bowel obstruction, intestinal bleeding and/or even perforation [3,4] . Apart from ileal duplication, colonic duplication is a rare abnormality. A thorough review of the international literature since 1950 has revealed 83 cases of colonic duplication to date, and only 4%-18% of all gastrointestinal duplications have been reported in the cecum [5][6][7] . Inclusion of the urogenital tract by an intestinal duplication has been reported in two cases [8,9] . Duplication of the tubular colon and vermiform appendix causing hydronephrotic atrophic kidney in adults has not previously been reported.
Preoperative diagnosis of colon duplication is often difficult. Symptoms usually include abdominal pain, nausea and/or vomiting, palpable mass, weight loss, and bleeding, which often confuses colon duplication with other more common diagnoses. The pathogenesis of these lesions has not yet been well characterized. One of the most widely accepted theories is an abnormality of the embryonic gut that results in the formation of a diverticulum, a cyst, or twinning of a bowel segment [10] . Environmental factors such as trauma or hypoxia during early fetal development have also been suggested to play a role [11] . Although intestinal duplication is considered to be a benign lesion and mostly asymptomatic, they may result in significant morbidity and mortality if left untreated [12] . Indications for surgical intervention often arise in an acute setting, in the form of complications. Specifically, patients with previously undetected duplication may present with bowel obstruction or severe gastrointestinal hemorrhage (i.e., ulcerating gastric mucosa within a duplication cyst) [5] . If encountered incidentally, these lesions should be surgically addressed to avoid any future complication, including the possibility of malignant degeneration within the duplication cyst [5,12] .
Duplication of the vermiform appendix is extremely rare. It is found in only 1/25 000 patients (0.004%) operated on for acute appendicitis [13] . The clinical presentation can vary depending on the location of the appendices in the colon. Cave and Wallbridge have classified the observed variants into several types: type A has incomplete duplication with both appendices having a common base; type B has complete duplication with the first appendix arising from its usual location at the confluence of the teniae coli, and the second appendix is located at various sites along the colon; and type C has complete duplication of the cecum, with each part having its own appendix [14,15] . Our case was diagnosed as type B.
Although preoperative diagnosis has been made with the aid of radiological studies such as a barium enema, the majority of cases have been diagnosed at surgery or upon pathological examination. Barium enema studies are considered essential for the diagnosis of tubular colonic duplication, with opacification of two colons being the diagnostic sign. However, in certain cases, barium enema apparently does not show positive findings. A case has been reported in which a barium follow-through study established a complete diagnosis of duplication of the colon and terminal ileum. The findings were confirmed by laparotomy. Although the diagnosis of appendix duplication is a rarity, surgeons should be aware of the possibility, especially when clinical signs and symptoms point to appendicitis, although at laparotomy the appendix looks normal. Routine exploration for a second appendix is definitely not indicated because of its rarity and increased complication rate.
Although duplication of the alimentary tract is rare, the possibility of congenital lesions of the alimentary tract should not be overlooked, even in adults presenting with vague or no gastrointestinal symptoms. | 2018-04-03T05:08:01.129Z | 2008-01-28T00:00:00.000 | {
"year": 2008,
"sha1": "e15f355cde8954dd8f0e39db93d6588e5d41d6a4",
"oa_license": "CCBYNC",
"oa_url": "https://doi.org/10.3748/wjg.14.641",
"oa_status": "HYBRID",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "9bef5205787c62af0602bf2b655ce032b35b355f",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
119219594 | pes2o/s2orc | v3-fos-license | 600-T Magnetic Fields due to Cold Electron Flow in a simple Cu-Coil irradiated by High Power Laser pulses
A new simple mechanism due to cold electron flow to produce strong magnetic field is proposed. A 600-T strong magnetic field is generated in the free space at the laser intensity of 5.7x10^15 Wcm^-2. Theoretical analysis indicates that the magnetic field strength is proportional to laser intensity. Such a strong magnetic field offers a new experimental test bed to study laser-plasma physics, in particular, fast-ignition laser fusion research and laboratory astrophysics.
Introduction
Laboratory generation of large, strong magnetic fields is of significance to many research areas including plasma and beam physics [1], astrophysics [2], material science [3], and atomic and molecular physics [4]. In material science, strong magnetic fields offer new opportunities to observe the magnetization process of the highly frustrated materials [5]. In astrophysics, soft γ-ray repeaters (SGRs) generate extreme surface magnetic fields of ~10 15 gauss [6][7][8][9]. Jet formation [10], magnetic reconnection [11,12] and collisionless shock generation [1] due to strong magnetic fields are also important research topics attracting much interest. Laboratory generation of large, strong magnetic fields makes it possible to study them in controllable experiments [13][14][15].
At present, traditional strong magnets including resistive magnets, superconducting magnets and hybrid magnets can produce stable magnetic field up to 45 T, repetitive pulses up to 100 T and a single-shot pulse up to 300 T [16]. However, the power consumption of resistive magnets is very high, and the high stress caused by high current and strong magnet is also a great challenge for the development of resistive magnets. Superconducting magnets are limited by the special critical current and critical magnetic field effect of superconducting materials. The applied superconducting magnets with about 50-milimeter aperture can produce about 20 T magnet field, and with the increase of the aperture, the magnet field strength produced by superconducting circle will decrease [17].
With the rapid development of high power laser technology, kilo-ampere [18] or even mega-ampere [19] current is generated by laser-plasma interactions. This offers a new opportunity to produce pulsed strong magnetic fields. Now, several generation mechanisms have been proposed in laser-produced plasma experiments. Spontaneous magnetic fields can be generated by nonparallel temperature and density gradients in ablated plasmas [20]. Surface-magnetic fields can be generated by oblique intense laser irradiation [21,22]. Although the spontaneous magnetic fields and the surface-magnetic fields in plasma are very strong, they are not applicable for applications in free space. In 1986, H. Daido et al. demonstrated a strong magnetic field with a capacitor-coil target irradiated by high power CO 2 laser pulses [23]. The field is generated due to the potential difference across two parallel plates induced by hot electrons. The potential drives a circulation of a strong reverse current through the coil connecting the plates. In 2005, C. Courtois et al. indicated that 100-T magnetic field was accessible using a capacitor-coil target [24]. In 2013, Fujioka et al. demonstrated a magnetic field up to 1 kT produced by nanosecond laser irradiating capacitor-coil targets [19].
In this paper, we propose a different mechanism to produce strong magnetic fields. Laser pulses are focused onto a metal target. Escaping of the hot electrons into vacuum from the target will lead to a large potential at the laser focal spot. Then, the cold electrons around is driven to the focus by the target potential. This current of the cold background electrons can excite a magnetic field pulse in free space. Our experiments demonstrate that a strong magnetic field with a maximum of 600 T is generated at laser intensity of 5.7 × 10 15 ⋅ −2 .
Experiment
The experiments were performed at the SG-Ⅱlaser facility, which can deliver a total energy of 2.0 kJ in a nanosecond Gaussian shape pulse. Fig. 1 shows the experimental setup. The target consisted of two parts, a Cu coil and a Cu planar target. The coil was made of a 200-μm-diameter Cu wire. The diameter of the coil was 1.16-mm-diameter. The coil was connected with the planar target. The planar target was perpendicular to the x axis. The coil plane was parallel to x-y plane. Eight heater laser beams at a wavelength of 351 nm were divided into two bunches and focused onto the two sides of the planar target simultaneously. The focal spot was measured to be about 150 μm FWHM in diameter. During the laser irradiating the planar target, hot electrons were generated and escaped into vacuum from the plasma. A large electrostatic potential grew up near the laser focus and charged the target. The potential dragged the background cold electrons to the focus to neutralize the target. This cold electron current would create a strong magnetic field pulse at the coil. The magnetic field was measured by a B-dot, which consisted of two 1.05-mm-diameter, reversed polarity single induction coils. The B-dot was positioned 25-mm away from coil center and 53-degree to the y axis. This type of magnetic probe can mitigate the electrostatic pickup created by the plasma [25]. A pinhole camera with magnification of 5 was used to monitor the X-ray emission from the target.
FIG. 1. Schematic view of the experimental setup at the SG-Ⅱlaser facility
showing the target, laser beams, B-dot and X-ray pinhole camera. Eight heater laser beams at a wavelength of 351 nm were divided into two bunches and focused onto the two sides of the planar target simultaneously. Hot electrons (green small balls) generated and escaped into vacuum from the target, resulting in a large electrostatic potential. A strong return current (purple arrow) due to cold electrons was generated. Then, a strong magnetic field was excited and detected by the B-dot (white coil). The X-ray pinhole camera was used to monitor the X-ray emission from the target. Fig. 2 shows the induced voltage signals measured by two reversed polarity single-turn induction coils at the laser intensity of 5.7 × 10 15 ⋅ −2 . Since the electrostatic potential is identical on both coils, the final signal can be excited by subtracting out the induced common voltage [25]. The integration of ( ) is the magnetic field strength. The peak magnetic intensity, measured at the local B-dot position, is 0.031 T.
Results and discussions
To distinguish the magnetic field generated by the coil with other possible sources, a straight Cu wire target without the coil was irradiated by the laser pulses. The magnetic field strength measured by the B-dot is only 0.006 T, much lower than 0.031 T. This small magnetic field is probably due to the spontaneous field generated by laser-plasma interactions. Therefore, the main source of the free-space magnetic field measured is from the coil target. The dependence of the magnetic field strength on the laser intensity is also investigated. Table 1 summarizes the details of the magnetic fields at different laser intensities. To see the dependence clearly, the variation of the B max and the B field at the coil center as a function of the laser intensity I are also shown in Fig. 4. The field strength increases with the laser intensity I. The B max is obtained at the strongest laser intensity of 5.7 × 10 15 ⋅ −2 . By fitting the data, we find that the is proportional to 1.4±0.2 .
Conclusions
We have demonstrated a maximum 600-T strong magnetic field with a simple coil target. The field strength increases sharply with the laser intensity. Compared with the capacitor-coil target, the generation mechanism of the coil target is straightforward and it is easy to be fabricated. Such a strong magnetic field can be applied to many research areas, in particular, laboratory astrophysics. | 2019-04-13T11:36:19.039Z | 2015-05-04T00:00:00.000 | {
"year": 2015,
"sha1": "6779881763af062a9059fb7c81b2f475dcf49714",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "6779881763af062a9059fb7c81b2f475dcf49714",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
} |
5065824 | pes2o/s2orc | v3-fos-license | Non-mammalian models in behavioral neuroscience: consequences for biological psychiatry
Current models in biological psychiatry focus on a handful of model species, and the majority of work relies on data generated in rodents. However, in the same sense that a comparative approach to neuroanatomy allows for the identification of patterns of brain organization, the inclusion of other species and an adoption of comparative viewpoints in behavioral neuroscience could also lead to increases in knowledge relevant to biological psychiatry. Specifically, this approach could help to identify conserved features of brain structure and behavior, as well as to understand how variation in gene expression or developmental trajectories relates to variation in brain and behavior pertinent to psychiatric disorders. To achieve this goal, the current focus on mammalian species must be expanded to include other species, including non-mammalian taxa. In this article, we review behavioral neuroscientific experiments in non-mammalian species, including traditional “model organisms” (zebrafish and Drosophila) as well as in other species which can be used as “reference.” The application of these domains in biological psychiatry and their translational relevance is considered.
Introduction
The use of model organisms is paramount in the behavioral neurosciences and its ramifications into biological psychiatry. Among these organisms, rats, and mice are still the most widely used, although other fields of the neurosciences use different species. This almost exclusive focus on rodents is problematic from the epistemic point of view, and behavioral neuroscience could profit from the inclusion of more species in its analysis. In fact, since the comparative aspect is a strong argument in favor of using non-human animals in behavioral research, the expansion of species is important to strengthen that argument. Moreover, insights gained from other species-including model organisms and "reference species" (Striedter et al., 2014)-can help understand what is generalizable and what is speciesspecific. In this Review, we highlight the role of model organisms and reference species in the behavioral neurosciences and discuss some advances, advantages and disadvantages of using a few choice species in behavioral research with consequences for biological psychiatry.
Model Organisms in Behavioral Neuroscience
"the fish is a frog. . . is a chicken. . . is a mouse" (Kimmel, 1989)
Foundational Issues
Before the discussion about model organisms advances, some definitions must be settled. In the life sciences, the term "model organism" refers to a species that is used in an attempt to understand particular biological phenomena (Fields and Johnston, 2005). From an epistemic point of view, a model organism acts as a stand-in for other organisms: [. . . ] model organisms are always taken to represent a larger group of organisms beyond themselves, and hence rely on very particular types of claims about their (potential) representational scope. [. . . ] The actual relationships between the model organism and this larger group often are very ill-articulated in the earliest stages of model organism work, and do not necessarily hinge on particular claims about genetic conservation or precise knowledge of the phylogenetic placement of a particular organisms in relationship to others (Ankeny and Leonelli, 2011, p. 318).
Another important epistemic characteristic of model organisms is that, differently from what Ankeny and Leonelli (Ankeny and Leonelli, 2011;Leonelli and Ankeny, 2013) call "experimental organisms, " they target a wide range of systems and processes occurring in living organisms, including genetics, development, physiology, behavior, evolution, and ecology. In this sense, model organisms are "material analogies" (Hesse, 1963) which, although not faithfully mirroring their target, represent other organisms at the most basic levels (Ankeny and Leonelli, 2011;Leonelli and Ankeny, 2013). The concept of model organism is not specific to any field of the life sciences; the concept of "animal model, " in contrast, is more common in the behavioral sciences. This term refers not to an organism, but to the conjunction of apparatuses and manipulations used to represent a given behavioral (dys)function in a different species than the target (van der Staay, 2006;Nelson, 2012). Animal models are most widely used in the fields of behavior genetics, biological psychiatry, experimental psychopathology, and neuropsychopharmacology, where they are used to generate biological (physiological or genetic) hypotheses regarding psychiatric disorders, to investigate the psychological aspects of the disorder, or to screen for potential psychiatric drugs (McKinney and Bunney, 1969;Willner, 1991;Wright, 2002). LaFollette and Shanks (1995) defined two categories of animal models: causal analog models, with which experimenters test causal mechanisms in a model and then extrapolate, by analogy, to the human condition; and hypothetical analogical models, which have the function of generating novel hypotheses. Shanks (1994, 1995) argued that the assumption of interchangeability between non-human animals and humans is weak at best, and therefore animal models are better used to generate novel hypotheses. However, it has been argued that models with predictive validity and construct validity show biological and translational relevance, and therefore can be used as causal analog models (van der Staay, 2006).
While the majority of animal models use model organisms as subjects-especially rats and mice (Griebel and Holmes, 2013) -, that is not a strict requirement to model a given psychiatric disorder. In fact, the definition of an animal model as an experimental preparation is neutral with regard to the extension of what is being modeled and to which species is targeted (Wright, 2002). Nonetheless, most assumptions of model organism research also inform animal modeling Maximino et al., 2010c;de Mooij-van Malsen et al., 2011;Kas et al., 2011;Stewart and Kalueff, 2014). While the aspects of pharmacological isomorphism, ethological consistency, and symptomatology are central to reasoning with animal models (Willner, 1991;van der Staay, 2006;Kalueff et al., 2007;Belzung and Lemoine, 2011), "[t]he arguments for evolutionary relationships, genetic homologies, and physiological similarities also are part of the epistemic infrastructure that supports the use of animal models" (Nelson, 2012, p. 16). In particular, some authors (Blanchard and Blanchard, 1988;Maximino et al., 2010c;de Mooij-van Malsen et al., 2011;Kas et al., 2011;Stewart and Kalueff, 2014) advocate the use of speciesspecific behavioral and physiological phenotypes as endpoints for assessing the effects of manipulations across multiple species. This, of course, necessitates the model to be embedded in a theoretical framework which will guide the choice of endpoints to be analyzed and validated (McNaughton and Zangrossi, 2008;Maximino et al., 2010c).
Expanding the Breadth of Species
The choice of species is usually guided by practical advantagesthroughput, fertility, developmental speed, availability of genomic and transcriptomic data-and, to a great extent, to the existence of well-established research communities and data availability (Fields and Johnston, 2005;Ankeny and Leonelli, 2011;Leonelli and Ankeny, 2012). In the neurosciences, further criteria are the amenability to genetic manipulations and relative simplicity of the nervous system. These advantages are more extensive in molecular neurosciences, including neurogenetics and developmental neuroscience, and historically gave rise to a handful of model organisms-viz, humans, macaques, rats, mice, zebrafish, Xenopus, Drosophila and C. elegans. Extensive databases of gene expression for flies [http://flybase.org/], frogs [http://www.xenbase.org], humans [http://human.brain-map. org/], mice [http://mouse.brain-map.org/], worms [http://www. wormbase.org/], and zebrafish [http://zfin.org/] are already available, allowing for the comparison of basal expression levels in different brain areas. These data can be used in comparative neuroanatomy to refine homology propositions (Engert, 2014;Mitra, 2014;Striedter et al., 2014), which is essential for circuit approaches in behavioral neuroscience. These data can also be used to mine for the neuroanatomical localization of psychiatric disorder-related genes in different model organisms (de Mooijvan Malsen et al., 2011;Kas et al., 2011). These neuroinformatic approaches exemplify the power of current data tools available for well-established model organisms in the neurosciences.
In addition to studying well-established model species, behavioral neuroscience could benefit from focusing on other, carefully chosen species to amplify the field of discovery. Striedter et al. (2014) used the term "reference species" to define "carefully selected species from phylogenetically widely spaced vertebrate and invertebrate groups" for comparative neuroanatomy. These species would then serve two purposes: as substrates for broad comparisons across all animals to identify nervous system fundamentals and as anchors for more fine-grained analyses within their particular taxon to assess the meaning of variation in whole brains and functional subsystems (Striedter et al., 2014, p. 5).
While the authors' focus was comparative and evolutionary neuroanatomy, their conclusions and recommendations can be extended to behavioral neuroscience insofar as both areas can profit from comparing taxa to infer how variations in one domain (gene expression, connectivity, activation patterns) relates to variation in behavior (Engert, 2014;Mitra, 2014;Striedter et al., 2014). Striedter et al. (2014) suggest that well-established model organisms be included among these species due to the availability of resources for their study, but point that other reference species should also be selected based on a few criteria. Thus, reference species are not "models for some other species, but [. . . ] a basis for comparisons that may reveal both similarities and differences" (Striedter et al., 2014, p. 5) The criteria for choosing a reference species are not established a priori, but might include phylogenetic position (Figure 1) and accumulation of significant data and methodological developments (Hale, 2014;Striedter et al., 2014); the ultimate goal is to allow the emergence of a comprehensive understanding of specific behavioral functions in different species and its specific relationships to brain structure and activity (Striedter et al., 2014;Hale, 2014).
The assumption of conservation is also best studied in a comparative framework. While this assumption is essential for projecting even rodent data toward humans, it has not been tested for most behavioral domains which are relevant for biological psychiatry (Panksepp et al., 2002;Pollen and Hofmann, 2008). It has been argued that, at least in anxiety research, construct validity is dependent on the assumption of evolutionary conservation (Maximino et al., 2010c); thus, to increase construct validity in animal models, testing the predictions of this assumption is highly desirable (McNaughton and Zangrossi, 2008). In this sense, while a common practice in behavioral neuroscience, using data from a single species (no matter how basal) to infer the ancestral state is unfeasible (Garland, 2001). As we will see along this article, while currently the most basal vertebrate used in biological psychiatry is zebrafish, a comparison with closely-related species (e.g., goldfish or guppy) as well as more basal vertebrates (e.g., sharks or lampreys) is necessary to establish the ancestral state of a given neurobehavioral trait in vertebrates. Studying species phylogenetically located near the origin of vertebrates, such as acorn worms and amphioxus, could also be useful FIGURE 1 | Phylogenetic context of some "reference species" that can be used in behavioral neurosciences. The figure underlines the position and phylogenetic distances of a few species in relation to each other, and suggest how this information can be used to inform the selection of organisms for research. For example, while most research in the behavioral neurosciences is performed using rodents, selecting a species from an outgroup-for example, chicks or lizards-could inform researchers on evolutionary conservation of biobehavioral traits in mammals. These informations complement the usual criteria for species choice (ease of reproduction, rapid generation time, etc.) and the availability of behavioral and physiological assays.
"for comparisons that span both vertebrate and invertebrate nervous systems" (Striedter et al., 2014). As a result, behavioral neuroscience could clearly profit from the adoption of other, non-model, "reference" species in the same way that neuroanatomy had.
Behavioral Research in Non-mammalian Species: Relevance to Biological Psychiatry Important as it may be for comparative and evolutionary neuroscience, cross-species research in the behavioral neuroscience is still incipient. Some approaches have been proposed which include comparative research in order to clarify genotype-phenotype relationships, thereby increasing the translational value of animal behavior in relation to human neuropsychiatric disorders LaPorte et al., 2008;de Mooij-van Malsen et al., 2011;Kas et al., 2011). The main rationale is based on the utility of model organisms in each specific domain. For example, rodent stress responses are driven by corticosterone, while zebrafish and humans use cortisol (Steenbergen et al., 2011). In addition to this critical but utilitarian aspect, the expansion toward "reference species"especially non-mammalian species-could allow for uncovering conserved and divergent mechanisms underlying pathogenesis Stewart and Kalueff, 2014). An example of conservation at the genomic and functional levels is the regulation of neurosteroidogenesis in relation to the anxiolytic effect of fluoxetine in both zebrafish (Wong et al., 2013) and mice (Pinna, 2010); an example of divergence is the serotonergic system, which show both duplicated genes and nuclei other than the raphe a multiplication of nuclei in fish in relation to rodents but is equally involved in anxiety and fear (Lillesaar, 2011;Herculano and Maximino, 2014). In both cases, a derivative endophenotype ) results from domain interaction that is conserved or divergent.
In what follows, we will exemplify conserved (or at least convergent) phenotypes that can be found in non-mammalian model organisms and in proposed reference species, as well divergent phenotypes which are found in model and reference species which could be useful for neuroscientific research. While not attempting to be a comprehensive review of every possible use of non-mammalian species in biological psychiatry, some specific examples were chosen based on phylogenetic position, data availability, and ease of use in laboratory settings. In most of the cases, pharmacological correlates are the strongest argument for the use of that species and behavioral model, which underlines the need to reinforce construct validity (the theoretical network behind the model) and face validity (the neurobehavioral isomorphism between the model and the target pathology).
Anxiety-like Behavior in Cyprinids
Human psychiatric disorders associated with anxiety, stress and phobic states result from abnormalities in neurobiological processes, inducing characteristic behavioral responses (Panksepp, 2004(Panksepp, , 2006LeDoux, 2012). Recently, diverse behavioral tests for anxiety, fear, and stress were proposed using teleost fish, among which zebrafish and goldfish stand out (Maximino et al., 2010d;Stewart et al., 2011a;Steenbergen et al., 2011). The potential of these species as model organisms for the analysis of genetic and biological mechanisms of fear and anxiety in vertebrates is beginning to be realized Stewart et al., 2014b). Moreover, other fish species-including goldfish and guppies-are being used in behavioral neuroscience, expanding the comparative breadth of teleost fish in this field (Hall et al., 2014).
Zebrafish is a small cyprinid that has long been used as a model organism in developmental biology and genetics (Signore et al., 2009). Its physiology is comparatively simple, making the species adequate for high-throughput investigation in pharmacology, toxicology, behavioral genetics and pharmacogenomics (Kokel and Peterson, 2008;Gerlai, 2010Gerlai, , 2015Kalueff et al., 2014b). Zebrafish also presents neuroanatomical landmarks and neurotransmitter systems which are very similar to those observed in mammals (Rinkwitz et al., 2011;Kalueff et al., 2014b).
Due to the many inherent advantages of zebrafish as a model organism-including low cost, easy manipulation and upkeep in relation to other vertebrate models, and 70-80% genetic homology with humans (Table 1)-zebrafish are increasingly useful in fields such as behavior genetics (Gerlai, 2003;Norton and Bally-Cuif, 2010). Although the degree of genetic homology with humans is not as high as in rodents, it is favorable in comparison with other genetically tractable organisms such as Drosophila melanogaster and Caernohabditis elegans (Kokel and Peterson, 2008). Despite the rising popularity of zebrafish in biological psychiatry (Griebel and Holmes, 2013;Stewart et al., 2014a), behavioral analyses still lack a more torough study (Gerlai, 2003(Gerlai, , 2010(Gerlai, , 2014(Gerlai, , 2015. Experiments using zebrafish larvae have been widely used, given that complex behavior appears in this species from 4 to 5 days post fertilization (dpf); these behaviors include prey capture, avoidance, phototaxis, and thigmotaxis, which are readily quantifiable in automated setups (Ahmad et al., 2012). Many studies suggest that avoidance and thigmotaxis can be used as measures of anxiety in zebrafish, and stimulus control of avoidance and thigmotaxis in zebrafish larvae is similar to that of anxiety in humans (Richendrfer et al., 2012).
Zebrafish possess all the "classic" neurotransmitters found in vertebrates (Rinkwitz et al., 2011), and its neuroendocrine system allows for different physiological stress responses (Steenbergen et al., 2011;Pavlidis et al., 2015). In adult animals, two important methods-the novel tank test (Cachat et al., 2010) and the light/dark test (Maximino et al., 2010b)-have been proposed to analyze anxiety-like behavior. In the first case, the animal is introduced to a novel environment, typically adjusting its spatial distribution in a "diving" response that tends to habituate with decreasing novelty and is accompanied by freezing and erratic swimming (Egan et al., 2009;Cachat et al., 2010Cachat et al., , 2011Wong et al., 2010). In the light/dark test, the apparatus is composed of a preferred black compartment and a nonpreferred white compartment, and the preference for the black compartment (which does not habituate; Maximino et al., 2010a) is accompanied by risk assessment and, when the animal enters the white compartment, erratic swimming, thigmotaxis and freezing (Maximino et al., 2014d). Both tests show considerable pharmacological isomorphism (Cachat et al., 2011;Stewart et al., 2011b,c;Maximino et al., 2014d) and have been successfully adapted to other species, including goldfish (Maximino et al., 2007(Maximino et al., , 2010bKang et al., 2011c;Nakamachi et al., 2014).
Serotonin (5-HT) mechanisms were strongly implicated in anxiety in humans and non-human animals (Maximino, 2012). Although the serotonergic system is not anatomically or genetically conserved (Lillesaar, 2011;Maximino et al., 2013a;Herculano and Maximino, 2014), there is some evidence for functional conservation. Extracellular 5-HT levels are positively correlated with anxiety-like behavior in the light/dark test and negatively correlated in the novel tank test (Maximino et al., 2013c). Moreover, the anxiolytic-like effect of drugs targeting different systems is associated with their ability to decrease serotonin turnover in the zebrafish brain (Maximino et al., 2014d). Drugs which increase serotonin levels increase anxietylike behavior in the light/dark test (Maximino et al., 2013c;Herculano and Maximino, 2014) and decrease it in the novel tank test (Maximino et al., 2013c;Stewart et al., 2013). In the light/dark test, exposure to an aversive olfactory stimulus (alarm • Small size ideal for microscopy (esp. larvae) • Non-invasive brain observation and manipulation due to transparency (larvae and casper mutants) • Conservation of major nuclei/brain regions: arcuate nucleus, preoptic area, hippocampus, amygdala, raphe, etc.
• Compact neuronal network revealable by two-photon or confocal imaging • Live imaging with genetically encoded calcium indicators • GAL4/UAS enhancer trapping for neuroanatomical determination and pharmacogenetic ablation • Small adult brain size allows reduced number of sections for histological analysis • Larger size suitable for ablation techniques and in vivo electrophysiology • Conservation of major nuclei/brain regions comparable to that of zebrafish • Medium adult brain size still allows for reduced number of sections for histological analysis
BEHAVIOR
• Well-established assays for anxiety/fear/stress, learning, impulse control • Larval assays for high-throughput screening • Well-established assays for learning and aversive control
The cholinergic system is emerging as another important target for the pharmacological modulation of anxiety-like behavior in zebrafish. The acetylcholinesterase inhibitor physostigmine has been shown to decrease bottom-dwelling in the novel tank test, an effect consistent with reduced anxiety (Cho et al., 2012). Nicotine, an antagonist at nicotinic cholinergic receptors, produces ample and consistent anxiolytic-like responses in the novel tank test (Levin et al., 2007), while in the aquatic plus-maze no effect was observed (Sackerman et al., 2010). The effects of nicotine in the novel tank test are mediated by the α 7 and α 4 β 2 nicotinic receptors, as antagonists for these receptors block the anxiolytic-like effects of nicotine (Bencan and Levin, 2008).
Adenosine and its receptors were implicated in the pathogenesis of anxiety-like behavior (Ruby et al., 2011). Caffeine is a non-selective adenosine rececptor antagonist, and has consistently been shown to increase anxiety-like behavior in zebrafish (Egan et al., 2009;Wong et al., 2010;Maximino et al., 2011a,b). These effects of caffeine on the light/dark test are mimicked by drugs which block the adenosine A 1 receptor, but not A 2 receptors, suggesting a participation of the first, but not the latter, in zebrafish anxiety (Maximino et al., 2011a). Interestingly, the A 1 receptor has also been shown to protect against the convulsive actions of pentylenotetrazole (Siebel et al., 2015), while both receptors have been implicated in the amnestic effects of scopolamine in the inhibitory avoidance test in zebrafish (Bortolotto et al., 2014). Treatment with IB-MECA, an agonist at A 3 receptors (which so far have not been described in zebrafish) reduces dark preference in a nitric oxide-and serotonin-dependent manner, while the reduction of bottom-dwelling is dependent on nitric oxide but not serotonin (Maximino et al., 2014b).
These results suggest the potential of adult zebrafish for studying the mechanisms of anxiety-like behavior and discovering novel drug targets. In general, a good balance between demonstration of pharmacological isomorphism and seeking novel targets is seen in the zebrafish anxiety literature.
Two other examples illustrate the potential of this species in describing the substrates of anxiety-like behavior.
In the first study, authors capitalized on the knowledge regarding the role of fibroblast growth factor (FGF) receptors on the development of the zebrafish brain (Thisse and Thisse, 2005). spiegeldanio, a mutant with reduced FGF 1A receptor function, was shown to have increased aggressive and exploratory behavior and decreased neophobia and anxiety (Norton et al., 2011). These animals show reduced dual specificity phosphatase enzyme dusp6 and phosphorylated extracellular signal-regulated kinase in the inferior lobe of the hypothalamus (Norton et al., 2011). While a decreased expression of the isoform A of the serotonin transporter in the raphe is also observed, treatment with fluoxetine does not rescue the behavioral phenotype (Norton et al., 2011); instead, histamine N-methyltransferase is upregulated in the brains of spiegeldanio mutants, which show decreased histamine levels in the preoptic area and raphe nucleus (Norton et al., 2011). Indeed, treatment with the histamine Nmethyltransferase inhibitor tacrine rescues not only the hypohistaminergic profile but also the behavioral syndrome associated with reduced FGF1A receptor signaling (Norton et al., 2011).
In another study, knockdown of otpa, a gene which is duplicated in zebrafish in relation to vertebrates is used to circumvent the lethality of homozygous null mutations in mice (Amir-Zilberstein et al., 2012). OTP is a homeodomain protein that is highly expressed in the neuroendocrine hypothalamus (Blechman et al., 2007), suggesting a role in regulating stress responses. In zebrafish, two isoforms are present; null mutants for otpa show normal basal expression of CRF in the brain, but exposure to a stressor does not increase CRF expression in the mutants (Amir-Zilberstein et al., 2012). These animals also show less bottom-dwelling in the novel tank test (Amir-Zilberstein et al., 2012). A series of experiments demonstrated that Otp associates phosphorylated cAMP response elementbinding protein (pCREB) to recruit the crf and a2bp1 promoters in response to stressors; the latter promoter leads to the expression of a short variant of the PAC1 receptor for the pituitary adenylate cyclase-activating peptide (PACAP) (Amir-Zilberstein et al., 2012). Gene knockdown of the short variant of PAC1 leads to increased behavioral and CRF responses to stressor, and overexpression of the short form in otpa-positive neurons in the hypothalamus increases basal and stimulated CRF expression (Amir-Zilberstein et al., 2012).
PACAP and the PAC1 receptor have been implicated in post-traumatic stress disorder (Ressler et al., 2011). An interesting correlation with those results is the observation of the behavioral and neurochemical effects of PACAP intracerebroventricular (i.c.v.) injections in goldfish, a closelyrelated cyprinid . This peptide suppresses food intake and induces a significant increase in the expression of crf mRNA; both effects are blocked by an CRF 1 receptor antagonist (Matsuda et al., 2006a). Moreover, PACAP also decreases locomotion in goldfish, albeit at a concentration higher than that needed to produce an anorexigenic effect (Matsuda et al., 2006a). While this locomotor effect is not necessarily suggestive of stress, an important research program has emerged on the role of hypothalamic neuropeptides in stress and feeding responses in goldfish (Matsuda, 2009;Matsuda et al., 2011a).
Some hypothalamic regions responsible for stress responses are also involved in food intake, and orexigen and anorexigen peptides produced and secreted from those areas regulate feeding. Some regulatory peptides involved in the organization of energetic homeostasis, such as ghrelin, orexin, galanin, thyrotropin-releasing hormone, PACAP, vasoactive intestinal peptide (VIP), and CRF, modulate different types of behavior after central or peripheral administration in mammals and teleost fish (Matsuda, 2009;Matsuda et al., 2011a). Some of these peptides have been linked to psychiatric disorders in humans (Figure 2). The i.c.v. and intraperitoneal injections of PACAP and VIP inhibits feeding and locomotion in goldfish (Matsuda et al., 2006a); PACAP injections also increase the expression of CRF mRNA in the brain, an effect which is mimicked by excessive feeding . While FIGURE 2 | Neuropeptides involved in anxiety disorder in clinical samples (Steckler, 2008). Peptides marked with asteriks (*) have been investigated in goldfish (Carassius auratus) psychomotor activity, anxiety, or feeding assays. AVP, vasopressin; ANP, atrial natriuretic peptide; CCK, cholecystokinin; CRF, corticotropin-releasing factor; DYN, dynorphin; END, β-endorphin; GAD, generalized anxiety disorder; NPY, neuropeptide Y; PD, panic disorder; PTSD, post-traumatic stress disorder; SST, somatostatin.
CRF has been implicated in anxiety and stress (Takahashi, 2001), there is also some evidence for a role in appetite control. CRF injections in the brain, but not in the periphery, is also anorexinogenic in goldfish (De Pedro et al., 1993). However, CRF injections increase locomotion and produce an anxiogenic-like effect in the novel tank test (Matsuda et al., 2013).
The central effects of CRF on feeding in goldfish are independent on cortisol, as injections of this hormone do not alter food intake (de Pedro et al., 1997). CRF receptors also mediate the anorexinogenic effects of α-melanocyte stimulating hormone (α-MSH), since CRF antagonists block the effects of i.c.v. α-MSH injections on feeding but α-MSH antagonists do not block the effects of i.c.v. CRF (Matsuda et al., 2008). Subsequent studies demonstrated that the anorexigenic actions of both CRF and α-MSH are blocked by a gonadotrophin-releasing hormone (GnRH) type I receptor antagonist, suggesting that these stress peptides induce the release of GnRH2 which ultimately mediates their effects on feeding .
Other neuropeptides involved in feeding have been shown to modulate locomotor activity and/or anxiety in goldfish. I.c.v. injections of ghrelin increase locomotion, while intraperitoneal injections decrease it; the intraperitoneal injection is also unable to alter dark preference (Kang et al., 2011c). Further studies using an acylated peptide demonstrated that this isoform increases food intake and increases locomotor activity when injected i.c.v. or intraperitoneally, while des-acyl ghrelin has no effect on locomotion, suggesting that the acylation is necessary for the psychomotor effects of ghrelin (Matsuda et al., 2006b).
The hypocretin/orexin system has been implicated in zebrafish feeding and arousal (Volkoff and Peter, 2006;Chiu and Prober, 2013). Orexin expression is upregulated in the goldfish hypothalamus after food deprivation and downregulated by intraperitoneal glucose injections (Nakamachi et al., 2006). I.c.v. injection of orexin A, but not orexin B, increases food intake in goldfish, while an anti-orexin antibody decreases food intake (Nakamachi et al., 2006). Orexin A, but not orexin B or an antiorexin antibody, increases locomotion (Nakamachi et al., 2006), dark preference, and bottom-dwelling (Nakamachi et al., 2014) when injected i.c.v.; these effect are blocked by pre-treatment with a OX1R antagonist (Nakamachi et al., 2014).
Cholecystokinin (CCK) is a gut polypeptide involved in stimulating the digestion of fat and protein. Post-translational modifications produce hormones with variable number of amino acids, including CCK4 (which acts primarily in the central nervous system, with little effect on the gastrointestinal tract) and CCK8 (which acts both centrally and peripherally) (Fink et al., 1998). In goldfish, i.c.v. CCK8 administration is anorexigenic and increases pomc mRNA levels, but not crf, in the brain (Kang et al., 2010). Pretreatment with CRF antagonists block this anorexigenic effect of CCK8 (Kang et al., 2010). Intraperitoneal CCK8 also produces an anorexigenic effect (Kang et al., 2010) that is blocked by pretreatment with the NMDA receptor antagonist MK-801 (Kang et al., 2011b).
Neuropeptide Y (NPY) has been implicated in feeding and, recently, has been identified as a potential target for the treatment of anxiety disorders (Garner et al., 2009). I.c.v. injection of NPY reduces locomotor activity and dark preference in goldfish (Matsuda et al., 2011a) and increase food intake (López-Patiño et al., 1999;Miura et al., 2006). The orexigenic effect of NPY is blocked by pretreatment with Y 1 receptor antagonists (López-Patiño et al., 1999;Miura et al., 2006), while the anxiolytic-like effect is not (Matsuda et al., 2011a). However, the injection of an Y 4 receptor antagonist mimicks the effects of NPY (Matsuda et al., 2011a), suggesting that feeding and anxiety are mediated by different mechanisms in the goldfish brain. Interestingly, chronic fluoxetine treatment increases npy expression in the zebrafish brain (Wong et al., 2013).
In addition to feeding-related peptides, stress-related peptides were also tested in goldfish for their effects on stress, locomotion, and feeding. Octadecaneuropeptide is an endozepine which acts as an agonist at the translocator protein 18 kDa (formerly known as peripheral benzodiazepine receptor) and at a metabotropic receptor and an inverse agonist at benzodiazepine binding sites at central GABA A receptors (Gamier et al., 1994). I.c.v. ODN administration increases locomotor activity and dark preference in goldfish, an effect which is blocked by pretreatment with benzodiazepine antagonists, but not by antagonists at the metabotropic receptor, suggesting that its anxiogenic-like and locomotor effects are mediated by the benzodiazepine binding site . ODN and derivative peptides also inhibit food intake in this species, an effect which is blocked by metabotropic receptor antagonists but not flumazenil, suggesting mediation by the metabotropic receptor but not by the benzodiazepine binding site (Matsuda et al., 2007). Interestingly, i.c.v. ODN injection increases proopiomelanocortin (pomc) mRNA levels, but not crf, and its anorexigenic effects are blocked by pretreatment with either a α-MSH antagonist or a CRF antagonist .
Aversive Conditioning in Goldfish
In addition to these experiments analyzing the behavioral effects of neuropeptides in goldfish, other experiments also analyzed learning and memory functions of specific brain regions. These experiments capitalize on the increased brain size of goldfish in relation to zebrafish, making ablation experiments and electrode implantation much easier in the first than in the latter cyprinid ( Table 1). While functional considerations are not normally used to establish homology claims, these experiments were crucial in advancing a second wave of discussion regarding the limbic telecephalon of teleost fish (Butler and Hodos, 2005).
The amygdaloid nuclei possess a variety of neurobiological roles, among which its participation in emotional behavior and learning are intensely studied (LeDoux, 2012). Some amygdaloid nuclei, including the basolateral portions (BLA), have been demonstrated to be central for avoidance learning in rodents (Choi et al., 2010;Lázaro-Muñoz et al., 2011). In this form of instrumental conditioning, the animal associates a response (usually shuttling in a box) with avoidance of an aversive consequence, such as an electric shock, that has been signaled by a sound or light stimulus. Two-factor theory proposes that, in avoidance conditioning, the subject first learns that the signaling stimulus predicts the aversive stimulus (Pavlovian conditioning) and then learns that a particular behavior (avoidance) causes termination of both the signal and the aversive stimulus (operant conditioning) (Maia, 2010). In this sense, the conditioned fear to the signal drives learning of the avoidance response.
The medial pallium (MP) of teleost fish has been proposed to be homologous to the mammalian BLA (Maximino et al., 2013b). Interestingly, lesions in this region, but not in the lateral pallium (LP) impair two-way avoidance learning in goldfish (Portavella et al., 2002(Portavella et al., , 2004aPortavella and Vargas, 2005;Vargas et al., 2012). Interestingly, the same region shows increased cfos expression in zebrafish after exposure to the light/dark test (Lau et al., 2011;von Trotha et al., 2014). LP lesions, on the other hand, impair spatial learning, but not active avoidance conditioning, except when a time interval is added between the signal and the electric shock ("trace conditioning") (Portavella et al., 2002;Portavella and Vargas, 2005).
As is the case with avoidance learning, the medial pallium is also involved in conditioned taste aversion (CTA), which involves the tendency to avoid the ingestion of substances that were previously associated with visceral discomfort. CTA is impaired by BLA lesions in rodents (Reilly and Bornovalova, 2005), and is characterized by long interstimulus intervals, high stimulus specificity, and fast acquisition-in fact, CTA can be established by a single trial (Garcia et al., 1955). The medial pallium, especially at precommissural levels, receives gustatory information from the diencephalic tertiary gustatory nucleus in the Rainbow trout Onchorynchus mykiss (Folgueira et al., 2003), althought there is no evidence that general visceral information reach this region (Yoshimoto and Yamamoto, 2010). Goldfish are able to successfully avoid a gustatory stimulus which was followed by lithium chloride; whole-telencephalon ablation and lesions in the medial pallium, but not in the lateral pallium or cerebellum, impair the acquisition of CTA in this species (Martín et al., 2011). Interestingly, when the gustatory stimulus was paired with an electrical shock, a conditioned bradycardia develops that is not impaired by telencephalic ablation (Martín et al., 2011).
The pharmacological or neurochemical bases of these effects are unknown. However, recent experiments using microinjection of drugs in the telencephalon suggest a role for the glutamatenitric oxide pathway in that process (Xu et al., 2003(Xu et al., , 2009). Goldfish were trained in an active avoidance paradigm similar to that proposed by Portavella and colleagues (Vargas et al., 2012). Escape responses were defined as shuttling made after the onset of both the light signal and the electric shock, while avoidance responses were made after onset of light signal but before the electric shock. The injection of D-AP5, an antagonist at glutamatergic N-methyl-D-aspartate (NMDA) receptors, before training did not alter the performance of escape responses, but significantly impaired the acquisition of avoidance responses (Xu et al., 2003). When this drug was injected after training, no amnesic effect was observed, suggesting a participation of this receptor in the acquisition phase, but not in the consolidation of the aversive memory (Xu et al., 2003). In a second set of experiments, Xu et al. (2009) injected nitric oxide synthase or guanylate cyclase inhibitors, impairing the production of nitric oxide or cyclic guanosine monophospate (cGMP), respectively. Injection of these drugs before training impaired the acquisition of avoidance responses, without effects on escape responses. Interestingly, microinjection of these drugs in the medial pallium imediately after training also impaired avoidance responses, suggesting a participation of the nitric oxide-cGMP system on the consolidation of the active avoidance memory (Xu et al., 2009).
Both microinjection and ablation experiments underline the main advantage of goldfish in relation to zebrafish ( Table 1): while zebrafish certainly has important advantages from the point of view of genetics and molecular toolboxes, the size of goldfish allows more easily for "classical" neuroscience techniques (drug microinjection, structure lesion, in vivo electrophysiological recordings). Indeed, while microinjection techniques are being introduced in zebrafish ( Kizil and Brand, 2011;Barbosa et al., 2012), they are much easier to make in an animal with a bigger brain and body. Thus, experiments with goldfish could complement results found in zebrafish, especially regarding localized interventions.
Aggression in the Siamese Fighting-fish
In aggression research, the Siamese fighting-fish Betta splendens was widely used during the 1970s and 1980s due to the welldescribed characteristics of the appetitive and consummatory aspects of its aggressive behavior (Simpson, 1968). Moreover, as is the case with other ornamental fish species, maintenance, and housing costs are low, and keeping fighting-fish in laboratories is relatively simple Kania et al., 2012). Moreover, a stereotaxic atlas of the fighting-fish telencephalon has been constructed, facilitating the description of neural systems underlying behavior (Marino-Neto and Sabbatini, 1988).
Betta splendens usually exhibit salient aggressive behavior toward conspecifics, including a characteristic appetitive element-the aggressive display -, which is characterized by the extension of operculae, extension of medial, and caudal fins, and an intensification of body color (Simpson, 1968). When the adversary counter-displays, an escalation ensues, resulting in attacks that include offensive biting (Bronstein, 1983(Bronstein, , 1994. Betta males also present aggressive display when exposed to a mirror (Miley and Burack, 1977); mirror presentation can therefore be used in lieu of a conspecific since it can simulate and aggressive encounter without harming the animals, since no difference between the strength of aggressive display toward a mirror or a conspecific (Miley and Burack, 1977).
While studying the aggressive display is useful for analyzing aggression, other techniques can also be used to assess the level of aggressive motivation in fighting-fish. Tapping on the classical studies of rodent motivation using runways, some studies analyzed the aggressive readiness by conditioning fighting-fish to swim through an "aquatic runway" in order to gain access to a conspecific or a mirror stimulus. It has been demonstrated that the level of aggressive motivation (as assessed by the time taken to reach the target area) is associated with combat readiness (number of displays) (Hogan and Bols, 1980). "Social reinforcement"-that is, instrumental behavior controlled by the opportunity for aggressive displays-has been used for a wide variety of applications in fighting-fish; interestingly, Betta splendens display self-control for mirror access, choosing delayed access to a mirror stimulus with longer duration instead of immediate access to a short duration of mirror presentation (Collins, 2008). These observations could complement work on the zebrafish three-choice serial reaction time task (Parker et al., 2012(Parker et al., , 2013a to build a research program on impulsive control disorders. There is some evidence that Betta splendens possess significant cerebral laterality-the division of cognitive functions between both brain sides-in relation to other anabantoid fish, given that they demonstrate an eye use preference in aggressive interactions (Clotfelter and Kuperberg, 2007). Lateralized individuals also present higher group cohesion and better performance in spatial tasks in relation to non-lateralized individuals (Clotfelter and Kuperberg, 2007).
Serotonin (5-HT) has been implicated in the modulation of aggressive behavior in different species (Takahashi et al., 2011), including fish (Herculano and Maximino, 2014). Acute treatment with low doses of fluoxetine decreases the duration of aggressive display (Lynn et al., 2007;Dzieweczynski and Hebert, 2012;Forsatkar et al., 2013), suggesting an inhibitory role for 5-HT in fighting-fish aggression. However, neither the 5-HT synthesis inhibitor para-chlorophenylalanine nor the 5-HT precursor Ltryptophan changed display behavior , suggesting that phasic, but not tonic, 5-HT controls aggressive behavior. Consistently with that hypothesis, intramuscular acute injections of 5-HT and 8-OH-DPAT (an agonist at 5-HT 1A and 5-HT 7 receptors) decrease the duration and readiness of aggressive displays, while the 5-HT 1A receptor antagonist did not produce an effect . The effects of chronic treatment with fluoxetine are mixed, with some authors describing decreases in aggressive displays (Kania et al., 2012) while other authors described no effect .
Honeybees: Aversive Control and Impulsivity
Honeybees comprise the genus Apis, which comprises seven species and 44 subspecies (Martín et al., 2011). While distributed in the whole world, honeybees appear to have originated in South and Southeast Asia and Africa (Engel, 1999). The Western honey bee (Apis mellifera) had its genome fully sequenced in 2006 (http://hymenopteragenome.org/beebase/), and at least since the description of the waggle dance by Karl von Frisch it has been proposed as a model organism for ethology and comparative cognition (Smith et al., 2000;Whitfield et al., 2006). Moreover, as eusocial species, honeybees are increasingly being used to understand how the social environment can shape behavior, including social learning, predator cues, and social decision making (Menzel, 1983). A comparative overview of the advantages and disadvantages of using honeybees in neuroscience can be found in Table 2.
Honeybees, especially Apis mellifera, are increasingly demonstrating their potential as models in behavioral studies, following the inclusion of invertebrates in neurobehavioral research (Leadbeater and Chittka, 2007). In particular, honeybees are capable of complex decision making, presenting cognitive biases (Wilson-Sanders, 2011) and self-control .
Honeybees have long been shown to be sensitive to aversive control (Hunt, 2007;Curran and Chalasani, 2012). Abramson (1986) demonstrated that honeybees quickly acquire aversive control in punishment, escape, and avoidance contingencies when an aversive odor (formic acid) is used. In a seminaturalistic setting, honeybees trained to discriminated between two differently colored targets quickly acquire avoidance responses when response to one of the targets is associated with an electric shock in the proboscis (Abramson, 1986). Honeybees also present a sting extension reflex, which is exhibited when the animal is subjected to noxious stimuli; this reflex can be conditioned so that bees learn to extend their sting in response to odorants previously paired with an electric shock (Vergoz et al., 2007;Tedjakumala and Giurfa, 2013). Interestingly, dopamine and 5-HT (but not octopamine or 20hydroxyecdisone) have been shown to decrease the conditioned sting extenstion response, while 5-HT 2 receptor antagonists increase responsiveness (Vergoz et al., 2007;Tedjakumala et al., 2014). While these results suggest a conserved role for specific 5-HT receptors in simple aversive control, other, more complex phenotypes can also be observed. In an interesting set of experiments, Bateson et al. (2011) first trained honeybees to associate a two-component odor mixture with either a reward (sucrose solution), punishment (quinine solution) or a less valuable reward (diluted sucrose solution); after training, animals were presented with unreinforced (test) trials in which three different odors mixtures, with intermediate concentrations of the mixture of the original compounds, were presented in addition to the two original odors. Animals responded to the original odor associated with the reward by extending the proboscis, while the odor that was associated with punishment or a less rewarding consequence did not elicit proboscis extension. Intermediate concentrations of the two-odor mixtures produced a mixed response, with mixtures with a higher concentration of the punishment-associated odors eliciting less proboscis extension responses. Interestingly, when animals were vigorously shaken for 60 s before testing, responding toward these concentrations was further decreased, suggesting a "cognitive bias"-that is, agitated honeybees classified ambiguous stimuli as predicting punishment . Moreover, shaken bees showed decreased levels of dopamine, octopamine and serotonin in the hemolymph . The authors suggested that this demonstration of a state-dependent modulation of categorization in honeybees has more in common with vertebrate behavior than previously thought.
Given that negative affective states such as anxiety and depression are associated with increased punishment expectancy, greater attention to potential threats, and a tendency to interpret ambiguous stimuli as threats, this "cognitive bias" demonstrates that this behavior can be used to model some aspects of psychiatric disorders. The neural mechanisms underlying negative cognitive biases are so far unexplored, but the observation that both dopamine and serotonin are diminished suggests similarities with the mechanisms underlying aversive control. However, this study also demonstrated a decrease in circulating octopamine levels in agitated bees-an effect which is difficult to reconcile with findings in vertebrates, since octopamine is found only in invertebrates. Moreover, in vertebrates the role of 5-HT on aversive control is highly dependent on receptor subtype and site of action, with 5-HT 2 receptors increasing aversive responsiveness in the amygdala and decreasing it in the periaqueductal gray (Maximino, 2012;Zangrossi and Graeff, 2014).
In addition to aversive control and cognitive bias, important experiments demonstrated that honeybees are able to selfregulate their behavioral choices and make an economic choice for a delayed and bigger reward in opposition to an immediate small reward (Cheng et al., 2002). Experiments showing the capacity for "self-control" are important to understand impulsive choice, sometimes indexed by an alteration in the optimal delay discounting behavior described above (Arce and Santisteban, 2006). Cheng et al. (2002) described an experiment in which honeybees were trained to choose between a delayed sweet reward and an immediate less sweet reward, choosing the first over the latter. Food deprivation increases impulsive choice and brain dopamine levels (Mayack and Naug, 2015). In addition, successive negative contrast has been demonstrated in bumble bees (Bombus impatiens)-that is, animals adjust their choice toward less effort when the reward value is downshifted (Waldron et al., 2005). Although pharmacological, genetic, and biochemical experiments are still much needed to elucidate the isomorphism of these responses to vertebrate systems, they point to an exciting possibility of using bees to study impulse control and its social modulation.
Fruit Fly Exploratory Behavior and Aggression
Drosophila melanogaster are widely used in different fields of the biomedical sciences, especially in genetics (Ankeny and Leonelli, 2011). Drosophila share the broad actions of essential neurochemical substrates (specific receptors, signaling enzymes and proteins, neurotransmitters systems) that are involved in emotional behavior (Schafer, 2002;Iliadi, 2009). Genetic techniques produced Drosophila mutants for genes associated with neurodegenerative disorders, making the species suitable for studying the pathological bases of these diseases (Muqit and Feany, 2002). These advances were made possible by the characteristics of fruit flies which turned them into a central model organisms in genetics ( Table 2): low maintenance cost, short generation time (c. 2 weeks), high fertility, and, of course, the availability of a research community and an extensive toolbox to manipulate gene expression in this species (Muqit and Feany, 2002;Schafer, 2002;van Alphen and van Swinderen, 2013).
As is the case with most model organisms, the bottleneck for its introduction in the behavioral neurosciences was the availability of neurobehavioral assays (Iliadi, 2009;van Alphen and van Swinderen, 2013). Insects such as Drosophila and honeybees exhibit some defensive behaviors which can be interpreted as anxiety-like and/or fear-like. In Drosophila, centrophobism/thigmotaxis has been proposed to represent anxiety-or fear-like behavior (Besson and Martin, 2004;Iliadi, 2009).
Thigmotaxis was first observed in Drosophila as an aftereffect of diethylether anesthesia, although it is present at basal levels in non-treated flies (Götz and Biesinger, 1985). It was observed that, when exploring a novel circular arena, flies avoid the arena center, an effect which is exacerbated after anesthesia (Götz and Biesinger, 1985). It was later observed that the preference for the arena boundaries is not controlled only by center avoidance (centrophobism) or wall preference due to tactile stimulation (thigmotaxis), and that flies prefer sheter-like environments (alcoves or dark corners), but only after the initial boundary exploration waned (Soibam et al., 2012). In both cases, an explicit "wall-following" is controlled by a complex combination of variables, including exploratory behavior ("curiosity"-driven) and avoidance due to the novelty of the environment. This is consistent with Montgomery's (1954) proposal that both approach and avoidance control exploratory behavior, the basis for the use of exploratory behavior in rodent models of anxiety.
This "wall-following" behavior is controlled by the mushroom bodies (MBs), one of the most well-studied central brain-like structures in the Drosophila brain. Hydroxyurea ablation of the MBs diminishes wall-following behavior, an effect that is replicated by genetic disruption of the synaptic transmission in γ lobes, but not in α/β lobes (Besson and Martin, 2004). Mutations that affect the cyclic adenosine monophosphaste (cAMP) pathway also decrease centrophobism, suggesting a participation of the cAMP-PKA pathway in that behavior (Lebreton and Martin, 2009). Finally, it has been shown that neurons expressing a substance P-like peptide in the fan-shaped body of the central complex are involved in wall-following, as genetic ablation of these cells increases this behavior (Kahsai et al., 2010).
These results highlight an important feature of Drosophila research: due to the genetic tractability of the model, manipulations which the use of Gal4/UAS lines to drive gene expression in a cell-specific manner (Aso et al., 2009). These powerful techniques, which are increasingly being used in zebrafish (Scott et al., 2007), allow for the expression of photosensitive proteins such as channelrhodopsin and halorhodopsin, as well as proteins such as tetanus toxin light chain (TeTxLC) or KillerRed in zebrafish or the temperaturesensitive protein Shibire in Drosophila, to inactivate a specific circuit or cell type expressed in a specific region. These powerful tools are increasingly being used in circuit neuroscience in genetically tractable organisms. On the other hand, the "allure" of high-technology research produces a tendency, in such organisms, for fundamental "low-tech" research to be ignored. In the case of Drosophila wall-following behavior, basic pharmacological research using clinically effective drugs has not yet been made, and therefore wallfollowing/thigmotaxis/centrophobism lacks pharmacological isomorphism.
Another important set of behaviors which have been studied in Drosophila and have consequences for biological psychiatry is aggression (Chen et al., 2002;Iliadi, 2009;van Alphen and van Swinderen, 2013). As is the case with most vertebrate species, fly aggressive behavior follows a pattern which includes behaviors without physical contact with the opponent, such as wing threat displays, and actions with direct physical contact, including fencing, holding, boxing, and tussling (Chen et al., 2002;Zwarts et al., 2012). Wing threat displays are threatening postures directed toward other males before fat charges. It is not known whether these display postures, as appetitive elements, represent an emotion such as anger (Iliadi, 2009).
Serotonin has been implicated in the control of aggressive behavior in vertebrates (Miczek et al., 2007;Carrillo et al., 2009). In Drosophila, the role of 5-HT in aggression is unclear, as 5-HT treatment does not alter aggression (Baier et al., 2002), and artificial selection for discordant levels of aggressive behavior does not alter the expression of genes involved in serotonergic signaling (Dierick and Greenspan, 2006). Nonetheless, pharmacological or genetic elevation of the serotonergic tonus increases aggression, and genetic silencing of 5-HTergic circuits blocks the effects of pharmacological induction while sparing aggression (Dierick and Greenspan, 2007). Expression of temperature-sensitive dTrpA1 channels in 5-HTergic neurons allows for the acute activation of these cells, accelerating the escalation of fights (Alekseyenko et al., 2010). In general, then, acute increases in 5-HT levels increase aggression, but it is not known whether this mechanism is present physiologically.
Reptile Neuroethology
While zebrafish and Drosophila are important reference species mainly due to their usefulness as model organisms, reptiles are important due to their position in the vertebrate phylogeny; reptile species are positioned at the anamnioteamniote transition (Figure 1), representing an important evolutionary junction that is mostly underrepresented in behavioral neuroscience.
Among reptile species, anole lizards are among the most widely studied in evolutionary ecology and ethology (Greenberg, 2002;Lovern et al., 2004). The genus Anolis is a diverse and widespread New World taxon that includes the green anole Anolis carolinensis, a small, diurnal, insectivorous lizard from the United States southeast that is convenient to observe and easy to maintain in the laboratory (Greenberg, 2002). Behavioral inventories (ethograms) that emphasize social dynamics have been produced for anole lizards (Greenberg and Noble, 1974;Greenberg, 1977) and other reptiles. Antipredator behavior has also been described in different lizard species, and include tonic immobility (Edson and Gallup, 1972;Hennig, 1977Hennig, , 1979Santos et al., 2010;Maximino et al., 2014a), flight (Hennig, 1979;Maximino et al., 2014a), modifications of exploratory behavior (Greenberg, 2002) and of refuge use (López et al., 2005). The advantages of using small reptiles as laboratory models ( Table 3) include its high availability in the wild (allowing for the establishment of laboratory colonies which are continuously replenished with wild stock to obtain field-relevant laboratory studies), the variety of life history traits between species, their close phylogenetic relationship with birds (forming the most basal extant amniotes), and ease of maintenance and costeffectiveness without sacrificing ecological relevance (Lovern et al., 2004). Moreover, brain atlases are also present for A. carolinensis (Greenberg, 1982), Gallotia galloti (Del Corral et al., 1990) and Gekko gekko (Wang et al., 2008).
Socially submissive green anoles show conspicuous physiological and behavioral alterations at both the immediate (catecholamine surges, increased plasmatic levels of corticosterone and α-melanocyte-stimulating hormone (MSH), increased serotonergic activity in the midbrain, hindbrain, hippocampus and nucleus accumbens, body color alterations) and the long term (reduced androgen tone, elevated corticosterone and α-MSH levels, decreased dopaminergic activity in the hindbrain and midbrain, decreased courtship and perch selection behavior). While the endocrinology of social stress responses is well-established in Anolis and other reptile species, a more careful observation of behavioral alterations is still lacking (Greenberg, 2002;Summers et al., 2003;Øverli et al., 2007).
Lizards are prey to many predators and therefore are subject to intense predation pressure, which is largely responsible for the development of multiple defensive strategies (Greenberg, 2002;López et al., 2005;Thaker et al., 2009). Rodríguez-Robles (1995, 1997) analyzed antipredatory responses in two different anole lizard species (Anolis cristatellus and Anolis cuvieri) during encounters with its natural predator, the snake Alsophis portoricensis, and observed 13 different behavioral responses, of which approximately half are also involved in social interactions in these species. A stereotypical sequence of behavioral acts was observed, depending on the phases of predator-prey interaction. Understanding the underlying mechanisms which organize these behavioral displays is essential to systematization and characterization of defensive responses these animals, as well to facilitate comparative studies relating to the establishment of animal models.
As an example, Machado et al. (2007) analyzed flight responses of Tropidurus montanus, another squamate lizard, to feigned attacks. The experimenters observed that maximum flight distance did not differ between the sexes; the presence of a neighbor did not affect flight for males; and that males with neighbors maximized the time of flight when compared to males without a neighbor. Moreover, body length was not predictive of flight behavior. During capture, T. montanus relied on attempted escapes, cloacal discharges, threat displays, tail breakage and tonic immobility (TI), in no particular order. TI duration is also increased in A. carolinensis by the presence of a simulated predator (Hennig, 1977) and is higher at shorter distances from the simulated predator (Hennig et al., 1976). Hennig (1979) studied the effects of the physical environment, time in captivity, and distance between potential predator and prey on defensive behaviors Anolis carolinensis. In addition to TI duration, flight latency was also recorded in order to determine if environmental factors affect these defensive behaviors similarly. The results revealed that the immediate testing environment is more important when a potential predator was nearby, and housing is more important at greater distances and after the first few days of adaptation to the new conditions. TI seems to be responsive to both changes in testing environment and time in captivity, while flight latency was only sensitive to the type of housing environment during captivity.
Many variables such as those mentioned above can potentially change defense responses, but it remains unclear how individual animals from the same population, sex, age, reproductive status, and other similar conditions of risk and costs, differ in their antipredator behavior. López et al. (2005) suggested that the propensity to take risks in a homogeneous group of adult male Iberian rock lizards (Lacerta monticola) varies in a "shy-bold" continuum and observed than the possible source of variation in antipredator behavior might also be related to small subtle differences in morphology, body condition, and health.
Tonic immobility has been equated with thanatosis, a secondary defense mechanism in which the animal displays "death-feigning" in response to external stimuli. This peculiar defensive behavior is assumed by many different lizard species (Edson and Gallup, 1972;Rodríguez-Robles, 1995, 1997;Machado et al., 2007;Santos et al., 2010). Immobility responses to acute and intense threat have been proposed to be a part of acute stress responses (Bracha, 2004;Moskowitz, 2004) and peritraumatic tonic immobility is predictive of development of post-traumatic stress disorder (Maia et al., 2011;Pires and da Costa Maia, 2013). Thus, tonic immobility has a potential translational value for biological psychiatry. In that sense, Maximino et al. (2014a) demonstrated that the Brazilian wall lizard Tropidurus oreadicus has a well-organized pattern of defensive behaviors that emerge after tonic immobility: immediately following the cessation of TI lizards initiate either freezing or (more frequently) a pattern of flight behavior that, in a circular arena, is characterized as "circling"; after that, animals return to careful exploration of the environment, employing thigmotaxis/centrophobic behavior and tongueflicking (Maximino et al., 2014a). Importantly, these authors observed that this pattern is amenable to pharmacological dissociation, with panicolytic drugs (alprazolam, imipramine) decreasing TI duration and post-TI freezing or circling, but not exploratory (risk assessment-like) behavior, while anxiolytic drugs (diazepam) increased tongue-flicking and decreased thigmotaxis (Maximino et al., 2014a).
A careful consideration of the different antipredator behavioral strategies employed by different lizard species in naturalistic, semi-naturalistic, or experimental settings suggest that, as is the case with mammals, reptiles adjust their behavior in relation to "predatory imminence continua" (Fanselow and Lester, 1988). In safe environments such as a nest or a burrow (low predatory imminence), animals do not exhibit defensive behaviors, resuming their normal activity. In a novel environment, predatory imminence increases because the probability of encountering a predator increases, and the animal engages pre-encounter defensive behaviors-including thigmotaxis, tongue-flicking, air-licking, and posture changesand body color changes associated with crypsis. When a threat is present, predator imminence increase further and the animal engages in post-encounter defensive behaviors (freezing). If post-encounter defensive strategies have failed (i.e., during contact with the predator), the animal switches to circa-strike defensive behaviors (flight/threat, tonic immobility, body thrashing, tongue-bunch, tail autotomy). These sequences have been observed in lizards both in the wild (Machado et al., 2007) and in laboratory environments (Hennig et al., 1976;Rodríguez-Robles, 1995, 1997;Maximino et al., 2014a), and have consequences for the development of novel models of anxiety disorders.
The Chick Separation Stress Model
As is the case with other model organisms in developmental biology, laboratory chicks are used for a variety of reasons: availability and low price of fertilized eggs all the year round; fast development, with a duration similar to that of the mouse (21 days); absence of placenta and therefore of maternal effects after egg laying; and well-described developmental genetics ( Table 3).
The chick separation stress paradigm exploits the strong attachment response of neonate fowl (Gallus gallus) and its typical distress vocalizations (DVoc) when socially isolated to model separation anxiety (Panksepp, 2011). Social separation initially produces an anxiety-like phase with high rates of DVoc, peaking in the first 3 min, followed by a decrease to about 50% of the initial rate within 10-25 min (Feltenstein et al., 2004;Sufka et al., 2006). While isolation for 3 min also produces stress-induced analgesia, hypothermia and ventral recumbency posturing (Sufka and Weed, 1994), the first effect is primarily mediated by novelty and not social isolation (Feltenstein et al., 2002). These phases can be pharmacologically dissociated, with anti-panic compounds (benzodiazepines, imipramine, clonidine, meprobamate, pentobarbital) attenuating DVocs in the first phase and antidepressant compounds (imipramine, maprotiline, fluoxetine) attenuating the second phase (Feltenstein et al., 2004;Sufka et al., 2006;Warnick et al., 2009). Moreover, social separation also increases plasmatic corticosterone levels (Feltenstein et al., 2003;Sufka et al., 2006). Interestingly, plasma corticosterone levels are higher when animals are isolated for 5-15 min, the period in which the first phase peaks, and this response is attenuated by isolation periods longer than 20 min (Sufka et al., 2006). Dopamine metabolism is increased in the telencephalon and diencephalon after 30 min (Hamasu et al., 2012). Interleukin-6, on the other hand, is elevated only by isolation for 120 min (Warnick et al., 2009). Chicks socially isolated for 5 min also show a "cognitive bias" in which they show higher responsiveness to aversive ambiguous cues, while animals isolated for 60 min are more responsive to both aversive and appetitive ambiguous cues (Salmeto et al., 2011); both forms of cognitive bias are attenuated by imipramine treatment, while clonidine did not alter responsiveness to any cue (Hymel and Sufka, 2012).
The amount of effort directed to the pharmacological and construct validation of the chick separation stress model is not a coincidence, as this test has mainly been used as a biobehavioral assay (Willner, 1991) to identify pharmacological compounds. In this sense, at least two target systems have been evaluated successfully. A participation of the opioidergic system was suggested by rodent social behavior assays (Burgdorf et al., 2011); in chicks, a participation of µ-opioid receptors is suggested by the observation that an agonist, DAMGO, attenuates DVocs in a 3 min isolation, while drugs acting at δ, κ or orphanin receptors are not effective (Warnick et al., 2005); the opposite effect is not observed by treatment with the non-selective opioid antagonist naloxone or the µ-opioid receptor antagonist CTOP, suggesting a phasic rather than tonic modulation. A role for NMDA receptors is suggested by a series of experiments in which intracerebroventricular injection of L-cysteine (Yamane et al., 2009a), glutathione (Yamane et al., 2007) and NMDA (Yamane et al., 2009b) blocks DVocs during a 10 min isolation test; in addition to this effect, these drugs also induce hypnotic effects. These results are surprising, since activation of the NMDA receptor induces anxiety-and panic-like behavior in rodents (Bergink et al., 2004) and fish (Herculano et al., 2015), and has been suggested to be mediated by NMDA-evoked GABA release.
In general, the chick separation stress test has proved to be an interesting behavioral assay to screen for potential antipanic and antidepressive drugs, being able to detect hits and dismiss drugs which produce false positives in other tests (Sufka et al., 2006;Warnick et al., 2009). Moreover, the neuroendocrine markers associated with stress follow the timecourse expected for acute stress, although their projection to anxiety and mood disorders is still blurred. While very little is known about the brain structures which mediate behavior in this model, the avian brain is well-described (The Avian Brain Nomenclature Consortium, 2005), and Gallus is a traditional model organism in developmental biology (Fields and Johnston, 2005). Thus, as is the case with zebrafish (Norton, 2013;Stewart et al., 2014b), Gallus gallus is well-positioned to produce models in developmental psychopathology. The case of the chick separation stress test is particularly interesting from this point of view, since about one third of adult patients with separation anxiety disorder developed it during their childhood (Black and Grant, 2014) and infant behavioral inhibition, a temperamental trait that is characterized by high reactivity to environmental and social novelty, is predictive of adult anxiety disorders (Fox et al., 2005).
Conclusions
The exclusive reliance of behavioral neurosciences on a small number of species is counterproductive; as a result, different research groups are starting to focus their efforts on a comparative perspective, using non-mammalian organisms in research. As can be inferred from this Review, the most readily transferable models using non-mammalian species are in the domains of anxiety, impulse control and aggression. While our knowledge of the neurobehavioral systems involved in disorders in these domains is far from complete, mammalian data suggests that a number of neurotransmitter and neuromodulator systems are dysfunctional in these pathologies (Figure 3). In Tables 1-3 and throughout this Review, the degree of conservation in these systems has been discussed. In some cases, a seemingly paradoxical situation is seen, in which the neurotransmitter system is not fully conserved from the molecular point of view, but from a functional perspective the degree of conservation is higher (Figure 3). While this state of affairs can simply represent the current limitations in the tools used to address these questions, it is possible that they represent discontinuities in the evolutionary histories of these traits. Future researchespecially of the comparative kind-will answer these open questions.
Each organism brings advantages and disadvantages, some of which we have discussed in the present Review. As is the case with rodents, no single species can be used to answer all research FIGURE 3 | Apparent paradox in the divergence and conservation of monaminergic systems in mammals, fish, and insects. Receptor and enzyme sequences are not conserved (including gene duplication in the case of teleost fish), and the brain nuclei containing monoaminergic neurons are differently distributed throughout the brain in mammals, fish, and insects, but functions appear to be relatively well-conserved. questions, and there is no such thing as a "perfect laboratory organism." Ideally, all behavioral research using these species should be inserted into a comparative framework, assessing the same variables on different species and/or trying to extrapolate findings in different species; nonetheless, while at the present moment most research using non-mammalian organisms relies on data produced in rodents, the inverse is not necessarily true. The widespread adoption of other species is advantageous from a comparative and epistemic point of view, but still needs to go a long way to impact the field. | 2016-05-12T22:15:10.714Z | 2015-09-08T00:00:00.000 | {
"year": 2015,
"sha1": "23b09cd0928184b0cf77e6ed8bca1f612358775c",
"oa_license": "CCBY",
"oa_url": "https://www.frontiersin.org/articles/10.3389/fnbeh.2015.00233/pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "23b09cd0928184b0cf77e6ed8bca1f612358775c",
"s2fieldsofstudy": [
"Biology",
"Psychology",
"Medicine"
],
"extfieldsofstudy": [
"Psychology",
"Medicine"
]
} |
153804908 | pes2o/s2orc | v3-fos-license | RAPE, MURDER, AND FORMALISM: WHAT HAPPENS IF WE DEFINE MISTAKE OF LAW?
The criminal law maxim “ignorance of the law is no excuse” represents a broad doctrine of strict liability in an area of law that usually insists on a culpable state of mind as a prerequisite for liability. For that reason, many scholars have attacked the harsh mistake-of-law rules as incompatible with basic principles of culpability. Other scholars have come to the defense of the maxim, and courts have adhered to it quite strongly even as the list of exceptions to the maxim has slowly grown. Oddly enough, however, this debate has proceeded without a definition of mistake of law. Distinguishing mistake of law, which generally does not excuse, from mistake of fact, which generally does, has proven difficult for scholars and judges alike, with the result that no serious effort to provide a definition of that distinction has yet been made. This article tries to fill that gap. It first defines the distinction by emphasizing law’s status as a special system of linguistic meaning within the larger world of fact. It then uses that definition to argue that some orthodox categories of criminal law, particularly “mistake of noncriminal law” and “unreasonable mistake of fact,” are empty concepts that serve only to obfuscate the line between mistake of law and mistake of fact, obscure the conflicts between at least two competing traditions of criminal intent, and thus impede a principled approach to the problem of whether and when ignorance of the law should excuse. * Associate Professor, Boston University School of Law. I would like to thank Ken Simons for his generous help with this article. Thanks also to Larry Alexander, Hugh Baxter, Guyora Binder, Stan Fisher, Susan Koniak, David Lyons, David Seipp, Kate Silbaugh, Robert Weisberg, Peter Westen, and, for excellent research assistance, Skye Davis, Gueorgui Balaktchiev, and Steve Ehrenberg. For more than a generation now, first-year law students have annually confronted the shocking facts of Regina v. Morgan,1 in which a husband orchestrated the gang rape of his own wife. According to some of the attackers, the husband invited them home to have sex with his wife, assuring them that any struggle on her part would only constitute her “kinky” way of enjoying the proceedings. In the event, she did indeed struggle prodigiously but was overpowered by her husband and his comrades. Needless to say, the resistance that these men encountered constituted the wife’s violent opposition to the attack, not pleasure or consent. At their trial, the defendants raised the defense of mistake of fact, arguing that, if she did not consent, at least they thought she had and so never intended to rape her. More importantly for the study of criminal law doctrine, they claimed that even a grossly “unreasonable” mistake of fact should exculpate a rape defendant. Their trial judge disagreed, and their jury convicted them. The House of Lords, however, agreed with the defendants on appeal, holding that a rape conviction could not stand where the defendants believed they had the woman’s consent, no matter how unreasonable that belief. This ruling sparked a spirited debate in the press as well as in academic circles,2 even though the Lords denied the defendants the benefit of their victory when they ruled that, notwithstanding the legal error at trial, no miscarriage of justice had occurred. In all of the debate about Morgan, however, no one has observed how the unusual disposition of the case raises difficult questions about the conceptual relationship between the reasonableness requirement, mistake of fact, and, surprisingly, 1 1976 App. Cas. 182, 195 (H.L. 1976). 2 See, e.g., SANFORD H. KADISH & STEPHEN J. SCHULHOFER, CRIMINAL LAW AND ITS PROCESSES: CASES AND MATERIALS 322-323 (6th ed. 1995) (indicating public and academic controversy).
For more than a generation now, first-year law students have annually confronted the shocking facts of Regina v. Morgan, 1 in which a husband orchestrated the gang rape of his own wife. According to some of the attackers, the husband invited them home to have sex with his wife, assuring them that any struggle on her part would only constitute her "kinky" way of enjoying the proceedings. In the event, she did indeed struggle prodigiously but was overpowered by her husband and his comrades.
Needless to say, the resistance that these men encountered constituted the wife's violent opposition to the attack, not pleasure or consent.
At their trial, the defendants raised the defense of mistake of fact, arguing that, if she did not consent, at least they thought she had and so never intended to rape her.
More importantly for the study of criminal law doctrine, they claimed that even a grossly "unreasonable" mistake of fact should exculpate a rape defendant. Their trial judge disagreed, and their jury convicted them. The House of Lords, however, agreed with the defendants on appeal, holding that a rape conviction could not stand where the defendants believed they had the woman's consent, no matter how unreasonable that belief. This ruling sparked a spirited debate in the press as well as in academic circles, 2 even though the Lords denied the defendants the benefit of their victory when they ruled that, notwithstanding the legal error at trial, no miscarriage of justice had occurred.
In all of the debate about Morgan, however, no one has observed how the unusual disposition of the case raises difficult questions about the conceptual relationship between the reasonableness requirement, mistake of fact, and, surprisingly, mistake of law. The Lords rejected a reasonableness limitation on the mistake defense, thus opening the door for defendants generally to argue the most grossly unreasonable mistakes to juries. At the same time, however, the Lords restored something like a reasonableness limitation by the back door when they held that no jury could have acquitted in this case--in other words, that these defendants were not entitled to the chance to argue so gross a mistake to a jury. This disposition is usually understood as resting simply on the Lords' evaluation of the defendants' credibility in making their claim, 3 but underlying that characterization is something more interesting and important than evaluations of credibility. In Morgan, the Lords were actually dealing in mistake of law, not mistake of fact. They allowed the reasonableness question, however, to obscure that point. As is illustrated by discussions of Morgan ever since, the Lords were far from alone in stumbling over the distinction between mistake of fact and mistake of law. 4 Judges and commentators have often done so in many settings. And that is true, at least in part, because that distinction has never been supplied with an adequate definition. The purpose of this article is to rectify that surprising omission in criminal law theory.
The traditional mistake of law doctrine is summed up in the slogan "ignorance of the law is no excuse." This general (although far from universal) refusal to consider mistakes of law as defenses is, of course, a species of strict liability and so highly objectionable to many scholars of criminal law. 5 They want, perhaps justly, to think of criminals as morally bad people, people "'confronted with a choice between doing right 3 See, e.g., GEORGE P. FLETCHER, RETHINKING CRIMINAL LAW 699 (1978). 4 As far as I can tell, the literature on Morgan contains no suggestions that a mistake of law is lurking in the case. 5 See, e.g., Richard G. Singer and doing wrong and choosing freely to do wrong.'" 6 While some scholars think that we may, in fact, be moving closer and closer to restricting the notion of criminality to morally culpable conduct, at least in the formal sphere of Supreme Court opinions, 7 there remains an awful lot of criminal law doctrine that suggests the opposite, that criminality can attach to someone for whom no fault has been found at all. 8 Criminal liability without inquiry into morally crucial aspects of the defendant's state of mind remains common enough, 9 and its chief doctrinal instance remains the maxim that "ignorance of the law is no excuse." Many have thus argued that this basic mistake of law doctrine ought to be abolished or modified as inconsistent with the fundamental, 6 Morissette v. United States, 342 U.S. 246, 250 n.4 (1952) (quoting Pound, Introduction xxxvixxxvii, in FRANCIS BOWES SAYRE, CASES ON CRIMINAL LAW (1927)). See also Francis Bowes Sayre, Mens Rea, 45 HARV. L. REV. 974, 1004 (1932) (supposing that criminal blameworthiness "is necessarily based upon a free mind voluntarily choosing evil rather than good"). 7 REV. 111 (1996) (arguing that there is too much strict liability in criminal law to maintain the notion that morality and fault are fundamental to criminal law). 9 For example, most people would probably consider the defendant's awareness of the age of the victim in a statutory rape case to be crucial to the moral quality of the defendant's act, but the traditional and still very common rule is to impose strict liability on that element of the offense. See State v. Yanez, 716 A.2d 759, 763 (R.I. 1998) (finding that majority of courts maintain strict liability for statutory rape); Garnett v. State, 632 A.2d 797, 803-4, 805 (Md. 1993) (holding that statutory rape is a strict liability crime, and that a majority of states retain statutes imposing strict liability); Commonwealth v. Knap, 592 N.E.2d 747, 748-49 (Mass. 1992) (holding that statutory rape is a strict liability crime); State v. Stiffler, 788 P.2d 220, 221 (Idaho 1990) (same); State v. Pierson, 514 A.2d 724, 727 (Conn. 1986) (same); State v. Elton, 680 P.2d 727 (Utah 1984) (same); Commonwealth v. Miller, 432 N.E.2d 463, 464-5 (Mass. 1982) (finding that most other jurisdictions subscribe to strict liability). See also WAYNE R. LAFAVE & AUSTIN W. SCOTT JR., CRIMINAL LAW § 5.1(c) (2d ed. 1986) (discussing "nearly unanimous view that a reasonable mistake of age is not a defense to a charge of statutory rape."); JOSHUA DRESSLER, UNDERSTANDING CRIMINAL LAW § 11.02(c) (2d ed. 1995) (finding that "in many states" statutory rape is a strict liability crime). But see People v. Hernandez, 393 P.2d 673 (Cal. 1964) (holding that reasonable mistake of age is a defense to a charge of statutory rape).
For other examples of strict liability, see United States v. Park, 421 U.S. 658, 676 (1975) (holding that the Federal Food, Drug, and Cosmetic Act imposes strict liability on "responsible corporate officials who . . . have the power to prevent or correct violations of its provisions"); United States v. Freed, 401 U.S. 601, 607 (1971) (holding that the National Firearms Act imposes of the substantive criminal law than a reliable, formal definition of ML as distinct from MF. Without such a definition, judges are thrown on their own undisciplined discretion with respect to a doctrine that is as central to the character of criminal law as any could be. The principle that ML generally is no excuse while MF generally exculpates not only supports the most venerable maxim of criminal law and innumerable convictions and acquittals. It also explains the bedrock criminal-law principle that no crime exists without a union of proscribed conduct and some kind of intent with respect to that conduct. 13 It purports to show what that principle really means: that the existence of a crime depends on the defendant's intention or awareness with respect to the facts of that conduct but not with respect to the law governing that conduct. If so fundamental a principle has no definition adequate to place meaningful constraints on judicial discretion, then a major prop of judges' claimed formalism and objectivity in applying criminal law is exposed as a mask for some other, hidden mode of decision. Of course, few today would suggest that formal legal definitions can ever wholly eliminate political or moral discretion from adjudication, and there is no shortage of JUDICIAL PROCESS, passim and 197-200 (1975) (identifying formalism with law--as against morality--and identifying both with control of judicial discretion); Richard H. Pildes, Forms of Formalism, 66 U. CHI. L. REV. 607, 608 (observing that among other things "formalism" is a name for the system of rules and rule-following that nearly everyone agrees is characteristic of any legal system); Larry Alexander, "With Me, It's All er Nuthin": Formalism in Law and Morality, 66 U. CHI. L. REV. 530, 530, 543-545 (1999) (arguing that "[l]aw is essentially formalistic" and thus must take the form of rules that settle questions determinately); Cass R. Sunstein, Must Formalism Be Defended Empirically?, 66 U. CHI. L. REV. 636, 638-640 (1999) (observing that formalism is hard to define but noting its emphasis on broadly applicable rules in controlling judicial discretion and thus preserving the autonomy of law from the judge's politics and morality); Frank I. Michelman, A Brief Anatomy of Adjudicative Rule-Formalism, 66 U. CHI. L. REV. 934, 936 (1999) (arguing that formalist adjudication requires construction of some positive legal norm as preexisting and controlling the adjudication). 13 From the age of Coke to the age of the student hornbook, no principle has been more hallowed. See EDWARD COKE, THIRD INSTITUTE 107 (1641) ("Actus non facit reum, nisi mens sit rea"); 4 WILLIAM BLACKSTONE, COMMENTARIES *21 ("[T]o constitute a crime against human laws, there must be, first, a vicious will; and, secondly, an unlawful act consequent upon such vicious will."); LAFAVE critics to argue that such discretion and politics are everywhere in law and adjudication. 14 If I were to conclude that the ML-MF distinction was just another mask in the charade of "law," no one would be too shocked in this day and age. But that conclusion is not mine. 15 Many judges and scholars still seem to think that the ML-MF distinction is reliable, meaningful, and useful for determining liability by law rather than by undisciplined discretion. 16 And I think there is good reason for that position, the position that at least a relatively formal 17 definition is available and implicitly in use.
This article, then, is an effort to derive a formal statement of the ML-MF distinction and then to see what implications such an effort has for criminal law theory and doctrine. & SCOTT, supra note ---, § 3.11 (describing the "concurrence" principle); DRESSLER, supra note ---, § 15.01 (same). 14 One version or another of this claim is routinely associated with the legal realists, critical legal scholars, or post-modernists. See the discussions in, e.g., MARK KELMAN, A GUIDE TO CRITICAL LEGAL STUDIES 12-13 (1987); Peter C. Schanck, Understanding Postmodern Thought and Its Implications for Statutory Interpretation, 65 SO. CAL. L. REV. 2505, 2577-81 (1992); RONALD DWORKIN, LAW'S EMPIRE 9 and n.6 (1986). Referring to the legal realists and to critical legal scholars, Dworkin says that, "Some academic lawyers . . . conclude that there is never really law on any topic or issue, but only rhetoric judges use to dress up decisions actually dictated by ideological or class preference." 15 I join Frederick Schauer in acknowledging the power of the "realist challenge" (and the postmodernist challenge) to the formalist notion that rules can fully control decision-making while insisting on the common sense that rules nevertheless do meaningfully constrain decisionmaking in very many cases. See SCHAUER, PLAYING BY THE RULES 191-196 (1991). 16 Examples of judges fill the reports. See just about any case citation in this article. Examples of scholars include the many who discuss the precise scope that ML defenses do or should have, presumably on the premise that the distinction is intelligible. See, e.g., Sharon L. Davies, The Jurisprudence of Willfulness: An Evolving Theory of Excusable Ignorance, 48 DUKE L.J. 341, 396-413 (1998) (arguing that courts have been too quick to read exceptions to the maxim into federal statutes); Kahan, supra note ---(arguing that ML defenses are made available only to those whose prohibited conduct would otherwise be judged moral); ASHWORTH, supra note ---, at 233-237; Husak & von Hirsch, supra note ---; A. T. H. Smith, Error and Mistake in Anglo-American Criminal Law, 14 ANGLO-AMERICAN L. REV. 3 (1985); Zupancic, supra note ---; GROSS, supra note ---, at 271-275; Cass, supra note ---(questioning the fairness but not the intelligibility of the doctrine). 17 What Sunstein says about formalism in statutory interpretation applies, I think, to discussions of formalism in nearly all legal contexts: "The real question is 'what degree of formalism?' rather than 'formalist or not?'" Sunstein, supra note ---, at 640. See also Zupancic, supra note ---, 438 (noting that "there can be no clear line between formalism and purposive legal reasoning" but that any case may present a greater or lesser "degree of formal determinism").
I. BASICS
Let me start with a couple of easy cases to illustrate the basic functioning of the distinction.
The defendant possesses a cylindrical device that can mildly reduce the "report" of a pistol and is charged under the federal statute that criminalizes possession of an unregistered firearm. The device comes within the federal statutory definition of "firearm," since "firearm" includes "silencer" 18 and since "silencer" includes "any device for silencing, muffling, or diminishing the report of a portable firearm, including any combination of parts, designed and intended for use in assembling or fabricating a firearm silencer . . . ." 19 If the defendant argues that she had no idea that the cylinder was designed to muffle gun noise nor that it could serve that purpose, then she is probably claiming MF.
She is saying that, no matter how well she might have known the details of the federal firearm laws, she could not have known she was doing anything wrong because she simply had no idea that this metal cylinder had anything to do with guns, as opposed to, say, woodworking equipment. Under this particular statute, the Supreme Court has specifically held that a conviction requires that the defendant have known the facts that make the object a firearm, so the defendant's MF would exculpate her. 20 This holding was entirely consistent with conventional criminal-law doctrine, which normally considers a MF with respect to a material element a valid excuse.
There is the additional complication, however, that very often a MF must be reasonable before it can provide a defense. 21 have reasonably failed to perceive the facts that made this particular metal cylinder a silencer within the meaning of the statute. If she had, at some point, detached it from the end of the barrel of a pistol or removed it from packaging marked "FIREARM SILENCER," then presumably her failure to perceive the facts that made the object a silencer would be deemed unreasonable. She would therefore get no defense if a reasonableness requirement were in place. As it happens, the Supreme Court appears to have interpreted federal firearms law not to require reasonableness in this respect; 22 so my defendant would simply have to convince the factfinder that she did not know that the cylinder was designed to muffle a gun, however stupid or unreasonable she was in failing to know that.
Apart from such MF arguments, the defendant could try arguments that would properly be analyzed as ML, although, given the absence of a ML defense under this and most statutes, 23 she would be well advised to try spinning the argument as MF.
Thus, she might try to argue that she made the following "MF": that she did not know that the cylinder was a silencer because she thought the device, although obviously designed to be a silencer, would not actually muffle any gun noise. If she were correct that the statutory definition of a silencer included the requirement that the silencer actually function correctly as a silencer, then she might have a genuine MF argument akin to my example above; she would simply have to convince the factfinder that she truly believed that this device would not actually muffle gun noise. Alas, she is not correct. As at least one court has held, the cylinder would count as a "silencer" as long as it was designed and intended to muffle gun noise, regardless of whether it worked. 24 22 See Staples, 511 U.S. at 619. 23 The question whether the device worked is simply irrelevant under this law, so the question whether the defendant knew whether it worked or not is equally irrelevant. Therefore, as far as the court is concerned, the defendant's claim that she did not know that the device was a silencer was a reflection of her failure to understand the law's definition of firearm or silencer, not a reflection of any misunderstanding as to facts that the law actually makes material. Her claim is a ML.
As a claim of ML, this argument is a loser under this statute, as its like would be under most criminal statutes. 25 There are, however, many exceptions to this rule. The criminal provisions of the tax code may be the best known examples, but there are many others as well. 26 Another illustration of the ML-MF distinction can be drawn from the case of Idaho v. Fox. 27 Fox was charged with possession of a "controlled substance" without a prescription. He had known perfectly well that he possessed quite a lot of ephedrine and nothing in the way of prescriptions. What he did not know, insisted his lawyer, was the "fact" that ephedrine was on the list of controlled substances for which Idaho required a prescription. 28 This mistake was, in a colloquial sense, clearly "factual," but, just as clearly, it counted as a ML for the purposes of criminal law. Fox misunderstood that ephedrine was a "controlled substance" in Idaho. He thus mistook the scope of the term "controlled substance." And the court thus held, as any disinterested criminal 25 See PERKINS & BOYCE, supra note ---, at 1028; DRESSLER, supra note ---, at 147; Model Penal Code §2.02(9) also adopts the traditional rule as the Code's basic presumption. 26 Had Fox claimed instead that, when he mail-ordered his 100,000 tablets, he had said to the order-taker "Excedrin" rather than "ephedrine" and had subsequently assumed that the pills that arrived were all aspirin, he would have had a good claim of MF. Since Idaho did not, in fact, require a prescription to buy aspirin (even in such quantities, let's suppose), such a mistake would not have suggested any misunderstanding of the law but only a misperception of the actual nature of the substance in Fox's possession. As above, such a mistake would be MF and likely would provide a good defense. As it happens, Idaho apparently requires any MF under this statute to be reasonable; 30 so my hypothetical Fox who claims to have ordered Excedrin would have to convince the factfinder, for example, that he had good reason to think that the order-taker understood him properly and no obvious reason to believe that the pills that arrived were something different from what he ordered.
II. PRIOR ATTEMPTS AT DEFINING THE ML-MF DISTINCTION
These examples are meant to lay out the basic structure of the criminal law of mistake and show that it is not usually too hard to make the distinction between ML and MF. But the simplicity of some cases does not mean that there haven't been significant controversies about how to categorize other cases.
In fact, both examples of legal mistake above turn out to fit a general category of mistakes that have sometimes been said to be factual or, at least, best treated as factual.
This category comprises those cases often labelled mistake of "civil law," mistake of "different law," mistake of "legal fact," or the like. Most writers have understood that these are MLs, even if a special kind of ML, but some writers have understood them as MF or as some category independent of ML and MF. 31 The most widely recognized example of such a mistake is the claim-of-right defense to larceny. Under that defense, a mistaken view of civil property law can exculpate if it leads the defendant to believe he has a right to the property. 32 Such a mistake may look like a MF, since the defendant is claiming "I thought the property was mine." However, where that mistake is rooted in a misunderstanding of the law of property--for example, landlord-tenant law regarding ownership of a tenant's improvements--the mistake is manifestly a ML. Still, such a mistake of property law is always a defense to larceny, and criminal law writers have often wanted to place such mistakes of "different law" in a distinct category. 33 does a mistake regarding the definition of "firearm" under federal law. Yet virtually no one would suggest that such mistakes are anything but conventional MLs. 35 More importantly, for this article, there are a number of scenarios that even the most careful scholars have found extremely troublesome to categorize. Such cases include, for example, so-called mixed questions of law and fact, cases involving a defendant's total inattention to some apparently factual question, and cases where the statutory language in question establishes no general category but simply identifies some unique object as having some special legal significance. All of these trouble spots can be found together in Paul Robinson's example of a liquor-seller who relies on the advice of a third party in concluding mistakenly that she is outside the boundary lines of a dry town. Is the location of a store relative to the boundary line a question of law, or fact, or a mixed question of law and fact? Does the defendant's total inattention to the questions made material by the statute--since she chooses simply to rely on the judgment of a third party--make her mistake necessarily a ML or might her mistake still be classifiable as MF? Does the fact that the boundary line is a particular thing rather than a general category resolve the question, since one of the essential characteristics of a law is often said to be its generality? These are tricky questions that have prompted some scholars to dismiss the ML-MF distinction as too tenuous to be of any relevance in determining criminal liability. Robinson thus suggests that, if such a defendant's liability turns on whether her mistake is classified as ML or MF, then the ML-MF distinction is performing no morally worthy service and should be dispensed with. 36 Here, Robinson joins those scholars who, relatively unconcerned with specifying any 35 I had to add "virtually" to this sentence, because I have just recently come across a passing suggestion that a mistake regarding the definition of "firearm" could indeed be conceptualized as a mistake of "legal fact." Singer & Husak, supra note ---, at 898. formal distinction between ML and MF, argue that in any case there is no morally sound reason to treat MFs and MLs differently from each other when determining criminal liability. 37 I will put to one side, for now, the question whether anything ultimately should turn on the ML-MF distinction. Instead, I will take the continuing relevance of the ML-MF distinction as the fact of life that it is and try to respond to the difficult cases head on. My object, then, is to work towards a reliable definition of the distinction between ML and MF, a definition that appears to have eluded the generations of scholars who have discussed criminal law's mistake defenses.
What definitions of the distinction have been attempted? I will not discuss all of the ways in which writers have approached the distinction, but a review of a few common approaches will set the stage. Some quite eminent legal thinkers have tossed off unhelpful formulations of the law-fact distinction. More than a century ago, outside the criminal context, James Bradley Thayer noted the problem that law itself can be understood as a kind of fact, but it was that special kind that included "a rule or standard which it is the duty of a judicial tribunal to apply . . ." 38 The front half of this observation--that law is of course a kind of fact--is important, but the back half is useless. Merely equating a question of law with a question regarding a rule or standard does not help much; yet only modest progress has been made since Thayer in defining the ML-MF distinction.
Forty years ago, Glanville Williams tried to advance the ball in the criminal context with this formulation: "Generally speaking a fact is something perceptible by 36 See, e.g., PAUL ROBINSON, CRIMINAL LAW DEFENSES, 380 and n.26 (1984). 37 See above, n.---and accompanying text. 38 James Bradley Thayer, "Law and Fact" in Jury Trials, 4 HARV. L. REV. 147, 152-153 (1890). the senses, while law is an idea in the minds of men." But Williams immediately recognized that the world of fact must include some things that he did not think of as objects of perception, such as someone else's state of mind, since such a fact had to be inferred from other facts rather than simply sensed. 39 As he recognized this difficulty with his definition of "fact," he might also have recognized the problem with his definition of law: that, at least in some senses of his formulation, a law must be an object of sensory perception and thus, as Thayer saw, a kind of fact. Neither of these points, however, moved him to a more refined statement of the distinction. He just moved on to more manageable problems.
Some years later, an extremely careful illustration of the distinction appeared in an article by Graham Hughes, the main task of which was to analyze the impossibility defense in attempt crimes. In that article, Hughes worked carefully through the various mistakes that might count as ML or MF in a variety of cases. 40 accurate and concise statement of what I might call, in deference to Hughes, the "orthodox position." 42 Like Hughes, Simons's main concern was with problems of the impossibility defense, not with providing a rigorous statement of the ML-MF distinction. His analysis did depend on the ML-MF distinction, however, so he needed a working definition to hold at bay those scholars who have questioned its reliability.
Simons said: "Mistake and ignorance of fact involve perceptions of the world and empirical judgments derived from those perceptions. Mistake and ignorance of law involve assessment of whether, given a certain set of facts, the actor would or would not be violating the law." 43 Simons elaborated by saying that you could tell a ML simply by imagining a hypothetical actor who knew all of the facts. If this person nevertheless failed to see that her conduct violated the law, then her mistake was ML. Similarly, you could identify a MF by imagining a hypothetical actor who knew the law perfectly well in its application to any set of facts. If this person nevertheless failed to understand that her conduct violated the law, then her mistake was MF. 44 These are good descriptions, I think, of how good lawyers and judges normally think about this problem. And no case that I know of attempts a definition more rigorous than this one. Simons completed his summary of the common sense of the problem by noting explicitly that any mistake about how the law "applies" to a given set of facts is also a ML. He thus rejected the notion that there is an "'intermediate' category between law and fact" or a category of "'mixed' questions of fact and law" to muck up the basic distinction. 45 41 Id. at 1019-1020. 42 Id. at 1016. 43 Simons, supra note ---, at 469 (citation omitted). 44 Id. at 469-470. 45 Id. at 470-471.
As a concise description of conventional practice, Simons's formulation is well done, but as a formal statement of the distinction it is inadequate. There cannot be much doubt that Simons's definition of fact, "perceptions of the world and empirical judgments derived from those perceptions," fails to exclude law. As Thayer suggested, the question whether a particular law exists or has a particular meaning certainly seems to be a factual question, a question about something in the world that can be perceived just as much as a person's state of mind, for example, can be perceived. 46 Thus it is perfectly intelligible to ask, "Is it the fact that ephedrine is a 'controlled substance'?," even though every lawyer would agree that that is a question of law rather than the question of fact that it purports to be. The conventional approach has no way of telling us why that question is "legal" rather than "factual." Thus, to follow out Simons's elaborations, you can't coherently posit someone who knows all the facts while remaining mistaken about whether her conduct violates the law, at least not without a more refined distinction between law and fact. That is because the perceptible facts "of the world and empirical judgments derived from those perceptions" include the fact that the conduct violates the law. 47 Thus, one who knows all the facts must also, by definition, know the law until we have a definition of the distinction between law and fact. Moreover, if the meaning of the law lies not just in the bare words of the governing statute itself but in how that statute applies to a given set of facts, as Simons rightly affirms, then the escape from this problem is not as readily available as it might at first 46 47 Alexander, supra note ---, at 57-58. seem. 48 That is, if you propose that a ML just be those factual mistakes that pertain to the meaning of the statute, you still must confront the fact that the meaning of the statute lies in the statute's "application to" these facts, which just begs again the question whether the mistake as to the application of law to fact in any particular instance is a mistake with respect to the law being applied or with respect to the facts it is applied to.
If lawyers generally think they know how to distinguish ML from MF but have never been able to offer an adequate definition of that distinction, 49 one might suppose that there cannot be a reliable statement of the distinction. Many have suggested as much. 50 Long ago, adopting an attitude consistent with the legal realists' general skepticism regarding formal distinctions, Jerome Hall seemed to reject the idea that any reliable distinction between ML and MF could be drawn. He did describe some differences between law and fact, such as that law consists of propositions while fact relates to discrete qualities or events, that laws are "generalizations" and "understood in the process of cognition" while facts are "particulars" and are "directly apprehended" 48 or "sensed," and that knowledge of law (as "cognition of propositions") has a kind of fundamental uncertainty about it that knowledge of fact (as simple "perception") does not. 51 But, after all this, Hall offers no definition and seems to suggest that the difference is really just a matter of degree. 52 The same position seems to emerge from Stephen Weiner's efforts, in the civil trial context, to allocate questions between judge and jury. Law is a "question of identifying broadly formulated principles for judging the parties' conduct," whereas fact is a "question of reconstructing acts or events which have actually taken place, or conditions which have actually existed." 53 This formulation would be easy enough to critique, even apart from its apparent vagueness, since it is not clear why "broadly formulated principles for judging the parties' conduct" are not also "conditions which have actually existed." It is also easy to find laws that seem to involve elements that are not broadly formulated principles; for example, again, laws establishing the boundaries of particular towns. But Weiner seems to recognize that he has not offered an adequate definition, since he goes on to insist that the categories of law and fact must be supplemented with a third category, called "law application," and to argue that the allocation of questions to judge and jury does and should have more to do with questions of institutional competence than any formal distinctions between law and fact. 54 He does not really suggest that these three categories are any more definable than 51 Hall, supra note ---, at 344-345. 52 Id. at 346. 53 Weiner, The Civil Nonjury Trial and the Law-Fact Distinction, 55 CAL. L. REV. 1020, 1020 (1967). 54 See Weiner, supra note ---, at 1022-24; Weiner, The Civil Jury Trial and the Law-Fact Distinction, 54 CAL. L. REV. 1867, 1876-1882 (1966). Endorsing this point, Endicott argues that in England it is the "consensus" that there is no analytical distinction between questions of law and questions of fact and that those labels are used in a purely pragmatic or result-oriented way. Endicott, Questions of Law, 114 LAW QUARTERLY REVIEW 292, 292 (1998). I use Weiner and Endicott here as exemplars of the line of civil law scholarship that grapples with the law-fact distinction not in terms of ML and MF, of course, but with respect to allocations of power between judge and jury, the categories "law" and "fact" are. Rather, he seems to think the tripartite approach simply more useful for revealing the essentially discretionary nature of the allocation of "law" to the judge and "fact" to the jury. Moreover, whatever merit there may be in embracing a third category called "law application" or "mixed question of law and fact," which is such a common part of the vocabulary of the law in civil cases, that intermediate category has never found a place in criminal courts where the distinction between MF and ML, however inadequately defined, has occupied the field.
There have nevertheless been academic advocates of this three-category approach even in the criminal context. George Fletcher, for example, has supposed that MFs are mistakes "based on misperceptions of the world," that MLs are mistakes based on "a false belief about the enactment or abolition of a legal norm," and that there is a "middle category" in which "the mistake goes to the application of an existing legal norm to a particular set of facts." 55 But Fletcher and others have been less interested in refining such categories than in subordinating them to what they take to be more important questions. Fletcher's discussion wastes very little ink on stating the distinction but considers at great length other factors that might render a MF or ML culpable or exculpatory. 56 Mark Kelman implies that the ML-MF distinction is entirely arbitrary and merely covers up other between appellate and trial courts, and between courts and administrative agencies. This literature seems to have long accepted the notion that there is an intermediate category between law and fact, unassimilable to either, for which allocation questions must be matters of discretion. questions that the legal system hesitates to confront. 57 Paul Robinson observes that the ML-MF distinction "has proven very troublesome" and should be dispensed with, since, at least sometimes, there is no morally significant distinction between one kind of mistake and the other. 58 While his main point is the latter one, he also seems to want to point out that the distinction has not been formulated well enough to categorize all cases accurately. 59 Finally, Larry Alexander has frankly rejected the notion that any reliable distinction between fact and law can be drawn. Reprising Thayer's approach, he notes that "all mistakes of law are mistakes regarding facts--those facts that are facts about the existence and meaning of law." 60 For example, he suggests that a misperception of a word of a statute, just like a misperception regarding the loadedness of a gun, is obviously a "factual" mistake in any normal use of the term "factual," even though it would routinely be regarded as ML in criminal law. 61 Having suggested how any ML can be recast as MF, he then suggests how any MF can be recast as ML by rejecting any distinction between "those facts that bear on the existence and meaning of the law and those that bear on its particular application." That line, he argues, "can be drawn at an indefinite number of places--meaning and application cannot be distinguished in a world--then factual mistakes are legal ones . . ." 62 For example, I take it that Alexander would argue that if a defendant thought a gun was unloaded when in actuality it was loaded, that apparently factual mistake could be readily re-categorized as a ML, as a mistake regarding the legal proposition that the gun counted as, say, a "deadly weapon." In sum, "all mistakes of law are mistakes regarding facts," and "factual mistakes are legal ones"; the ML-MF distinction is an illusion, for Alexander, or at least a distinction that must be defined functionally rather than conceptually. 63 Alexander's argument here and Robinson's hypothetical regarding the location of a boundary are two of the most pointed exposures of the absence of any reliable definition of the ML-MF distinction in the legal literature. Almost any lawyer, I think, would reflexively reject Alexander's notion that every factual mistake can be restated as a legal one. A mistake regarding the loaded-ness of a particular gun seems patently a MF, a mere misperception about a particular incident. But I think Alexander has shown that calling something a misperception is not by itself enough to take it out of the realm of ML, since the misperception may be a misperception of the existence of the statute, or of the words in the statute, or of the classes of objects or ideas denoted by those words-misperceptions that at least sometimes are obvious examples of ML. And Robinson's boundary example refutes the notion that the vague general-versus-particular distinction does the trick, 64 at least until more precisely defined, since a legal definition 62 Id. at 52. 63 Id. at 52-53. 64 Schauer argues that it is in the nature of a rule (such as any criminal prohibition) to be "general." See SCHAUER, supra note ---, at 17-18. That is probably right, but that does not mean that every element of the rule must be general. A prohibition on selling liquor in a particular town is general in that it prohibits not just a particular person from selling liquor at a particular time but prohibits a category of persons from selling at any of a large number of times (perhaps all times). It is, therefore, a rule rather than a mere command. But that rule has at least one element--presence in one certain town--that is particular. of a boundary (about which at least some mistakes must be MLs) seems to establish no general principles but only a decision to give some particular object in the world a particular legal status, much like a court's decision that a particular gun has the status of "deadly weapon." The next step I want to take is to improve on the current definitions of the ML-MF distinction to see if challenges like Alexander's and Robinson's can be met.
A. The Definition
A definition of the ML-MF distinction should avoid reliance on the notion of misperception as a distinguishing feature of MF and similarly should avoid reliance on the difference between generality and particularity. Both of these seem attractive, but neither has been shown to work and I don't think that either can be made to work, for the reasons sketched above. Instead, the definition should rely centrally on the idea of law as a system of meaning. The alternative formulation that I want to test, then, is the following: Whenever a court determines that the defendant's experience of the world at the time of the offense conduct (from the defendant's perspective) would entail an inference by the factfinder that a specified statutory term had been satisfied--i.e., the D's experience would instantiate the meaning of the statutory term--then any claim by the defendant that she nevertheless failed to draw that inference (whether at the time of the conduct or later) is a claim of ML. 65 On the other hand, whenever a court determines that the defendant's experience of the world at the time of the offense conduct would 65 It is equally a ML to suppose that a certain set of circumsances is consistent with satisfaction of a statutory term when in law that set of circumstances entails the inference that a statutory term not, if accurate, entail an inference by the factfinder that a specified statutory term had been satisfied--i.e., would not instantiate the meaning of the statutory term--then, of course, the defendant's belief that the statutory term has not been satisfied in the case cannot be a ML; if the defendant has made a mistake somewhere, then it must be in one of the underlying beliefs held by the defendant at the time of the offense and must be a MF.
Less formulaically, a ML is a failure to recognize or understand accurately the meanings of the terms of the charging statute, meanings that, as Alexander says, 66 are constituted at least in part by their "applications" in particular factual scenarios. Thus a ML is a failure to draw proper inferences between particular scenarios (e.g., one in which an object has the capacity to fire bullets at high velocity) and the statutory terms (e.g., "firearm") that govern the case. Every scenario and statutory term mutually entail certain inferences and exclude others, and an inference (or failure of inference) is invalid, for legal purposes, when it is inconsistent with the meaning of a term of the charging statute. To put it another way, the meaning of a statutory term is a statement of the semantic reasons why some particular is or is not deemed to be covered by the statutory term. And the reasons offered then constitute a linguistic equivalent of the statutory term. Thus an object that can fire bullets at high velocity may be deemed a statutory "firearm" because it has that capacity, and so a "firearm" is or means an object that can fire bullets at high velocity, even if it can also be or mean other things. 67 And the circumstance that an object can fire bullets at high velocity entails the inference to has not been satisfied. I will not treat this variation separately because I do not think it raises analytically distinct issues. 66 Alexander, supra note ---, at 52. 67 Under federal law, "firearm" includes such things as hand grenades and mere parts of guns as well. "firearm" at the same time that "firearm" entails an inference to a category of things that includes those objects that can fire bullets at high velocity. Such statements of equivalence or mutual inference--rules of law--may be and are sometimes established in advance of a particular adjudication and then more or less mechanically "applied" to the facts of the case. More universally, however, these definitions simply appear as assertions of the equivalence between a certain set of characteristics abstracted from the facts of the case and a certain set of characteristics previously declared to have been equivalent to the statutory term. In a so-called case of first impression, the equivalence will be established directly between the abstracted characteristics of the case and the statutory term. Thus, for example, in a particular scenario under the firearms statute it may be the case that a silencer is broken. If a previous case has established that a broken silencer still counts as a silencer (i.e., "broken silencer" equals or implies "silencer"), then this court will just "apply" that established equivalence or inference to this case. But the first time the question is faced the judge simply asserts, on the basis of her skill as a reader of statutory English, that the label "broken silencer" is the equivalent of and entails the inference to "silencer"; that is, that "silencer" means, among other things, broken silencer.
B. Some Complications and Clarifications
This illustration immediately suggests a complication: that such a definition of "silencer" or "firearm" or anything else--such a specification of meaning--is always incomplete because whatever characteristics or circumstances or properties (like capacity to fire bullets at high velocity) might entail the inference to "firearm"--i.e., might give a meaning of firearm and stand as a rule of legal inference 68 --will only entail that inference as long as there are no unforeseen circumstances that "disturb" the inference. 69 A future case might, for example, present an object that can fire bullets at high velocity but that does not let the bullets leave its sealed barrel. Is the object still a firearm? The previously established "general" rule that capacity to fire bullets at high velocity means "firearm" may turn out not to work in this case, not to define firearm after all. We can never be completely certain that any a priori statement of a legally entailed inference (i.e., a legal meaning or definition) will never encounter a circumstance that will defeat it. 70 Still, even if the task of distinguishing the meaning of a statutory term from the reality that it governs is complicated by this argument for the fundamental instability of all linguistic meaning, the task can be pursued in a coherent way. We can still conclude that, for a court to find a ML in any particular case, it can and must first identify the collection of beliefs--the experience of reality--actually harbored by the defendant at the time of the conduct and deemed by the court adequate to specify a meaning of the statutory term. The fact that the meaning of a particular word is never completely specified for the future does not prevent its meaning from being specified in any single 68 See Friedrich Waismann, Verifiability, in LOGIC AND LANGUAGE 118 (Antony Flew, ed., 1st ser. 1968, orig. 1951). 69 See id. at 127. 70 See id. at 119-129. See also Lech Morawski, Law, Fact and Legal Language, 18 LAW AND PHIL. 461, 472 (1999) (showing that legal terms cannot be completely defined but have only "partial" definitions that are defeasible by new circumstances). The observation that there might, in any particular case, emerge a circumstance that "disturbs" the normal inference--i.e., the a priori statment of the law's required inferences--may mean that every rule of law is actually particular and never general; only a priori, provisional, working rules of inference are general. But that does not make the inference required in any particular case, the inference entailed by the circumstances of that case and the statutory term, any less a matter of law, since the distinction between ML and MF does not depend on any distinction between generality and particularity. case for that case, by the factfinder's specification of the aspects of the defendant's experience that give the word its pertinent meaning. 71 What about MF, then? For a court to find a MF, it must go through the above process of specifying the definition or meaning of the statutory term in the context of the particular case and on that basis determine that when the term is used correctly it does not entail an inference from the defendant's beliefs or experience at the time of the offense to the conclusion that the term has been satisfied. Any mistake that rests on a body of beliefs that would not require a factfinder to infer the satisfaction of a statutory element is a MF. A MF, in other words, is a mistake untainted by an incorrect understanding of what characteristics or circumstances the court takes to be equivalent in meaning to a statutory term. It is a mistake not about the meaning of a statutory term but about the experience of the defendant that is used to specify the meaning of the statutory term. The ML-MF distinction is the distinction between mistakes of "meaning" under the statute (as elaborated above) and all other mistakes. The ML-MF question is always the question whether the mistaken defendant is using the statutory language correctly by drawing the same inferences under that language that a factfinder would be required to draw from exactly the same experience that the defendant believed herself to have confronted. 71 I should stop to note that there is, of course, a large literature on the philosophy of linguistic meaning, to which I cannot do justice here. In that literature, there is a vibrant debate on what it is for a word or sentence to have a meaning, and my usage of "meaning" in this article is not by any means the only usage out there. For a brief, lucid introduction to the philosophy of language or meaning, see WILLIAM P. ALSTON, PHILOSOPHY OF LANGUAGE (1964). For more recent discussions of ongoing debates about the theory of meaning as they relate to law, see, e.g., DENNIS PATTERSON, LAW AND TRUTH (1996) (evaluating realist and anti-realist theories of meaning); Coleman & Leiter, supra note ---, at 601-607 (same). Still, I think that, for my limited purposes, my usage is defensible in its own terms. I've been saying that, to distinguish ML from MF in a particular case, the judge must isolate the defendant's relevant beliefs about her experience at the time of the offense and determine what inferences are entailed by the combination of the statute and those beliefs. The subtleties of distinguishing ML from MF come from the difficulties in distinguishing between statutory meaning and the lived experience of a defendant that is governed (or not) by that statutory meaning, since statutory meaning cannot be specified in a given case except by asserting the statutory term's equivalence or non-equivalence to that experience or, better, to an accurate linguistic description of that experience. I've tried to indicate how to draw that distinction above, but one might wonder if the distinction doesn't collapse if the defendant's "beliefs" or "experience" at the time of the offense conduct include beliefs about or experience of statutory meaning itself. What if the defendant's experience at the time of the offense conduct includes the belief that the characteristics of the situation do not satisfy the terms of any criminal statute, whether because she is completely ignorant of the relevant statute or because she has misinterpreted that statute?
On first look, that belief might seem classifiable as MF under my definition even though it should clearly be a ML. A factfinder presented with the beliefs held by the defendant at the time of the offense, including the defendant's belief that the statute was not satisfied, might seem to have no conclusion to reach but that the statute was indeed not satisfied. Concluding that the circumstances as the defendant (mistakenly) experienced them could not constitute the statutory term since non-satisfaction of the statute was part of the defendant's "experience" of or "belief" about the statute, the factfinder might seem forced to the conclusion that the mistake was MF.
Why is that conclusion wrong? The answer, of course, is that, in an adjudication, the meaning of a statutory term--the experience that constitutes it in that case--cannot depend on what a defendant thinks it means but only on what the properly instructed factfinder thinks it means. A mistake about the existence or meaning of a term cannot change the fact that the defendant was aware of circumstances that did constitute a meaning of--entail an inference to--the term as understood by the factfinder, the ultimate arbiter of meaning in adjudication. Once the factfinder identifies beliefs or experience that entail the term, the entailing may be defeasible by other beliefs or experience that the factfinder deems relevant but not by any old belief or experience and particularly not by the defendant's failure to see the entailing that the factfinder saw-the "meaning" of the term--in the first place.
For example, if a factfinder concludes that the defendant knew some object she possessed could fire bullets at high velocity and so concluded that such an object would be a firearm, that statement of statutory meaning could be defeated by some further kinds of information but not by other kinds. It could be defeated by the further information that the defendant also believed the barrel to be securely sealed, if the factfinder 72 concluded that such an object would not be a "firearm" after all. But, just as the original entailment (from capacity to fire bullets to "firearm") would not be defeasible by such irrelevant further information as that the defendant believed the sun to have been shining, so it could never be defeated by information that he did not believe that original inference to be entailed in "firearm" or did not know of the existence of the statute. Unless the statute itself declared that the term "firearm" shall 72 By "factfinder," of course, I always mean "properly instructed factfinder." The elaboration of statutory meaning in any particular adjudication is a kind of collaboration between judge and mean whatever any defendant chooses for it to mean, the factfinder would have no reason to defer to the defendant's definition of firearm--the defendant's idea of correct usage of that term--but would persist in concluding that the object the defendant thought she had with the characteristic of capacity to fire bullets would count as--entail the inference to--"firearm." Any mistake as to "firearm" therefore would have to be a mistake as to meaning and so a ML. To give the defendant's belief about the inferences entailed by "firearm"any relevance in determining what inferences really were or were not entailed by that term, and thus whether the defendant's mistake was ML or MF, would be to permit the defendant not simply to convert every mistake into MF (her misreading of "firearms" or ignorance of the statute becomes a MF) but actually to give statutory terms whatever meanings she thought they had at the time of the offense, or even no meaning at all (if she thought no such statute existed).
The meanings of words may perhaps be unstable in significant degree, but their meanings do not have to depend on any old person's beliefs about their meanings. At least when there is an authoritative assigner of meaning, as in a court, the meaning of a word does not in any way depend on a defendant's knowing of the existence or meaning of the word. Through the factfinder, the statute establishes a set of meanings for its terms, a language for the legal proceedings, that is independent of what anyone governed by the statute believes about its terms. Thus, in determining whether some set of the defendant's beliefs would require an inference to x or to not x, the judge cannot consider among those beliefs any belief of the defendant's as to the meaning or existence of x itself without abandoning the very notion of law's establishing objective, authoritative meanings independent of the beliefs of those it is meant to govern. jury in a jury trial but the judge's alone in a bench trial. I use "factfinder" or "properly instructed Of course, there is a sense in which a statute can play the role of a circumstance of the offense and in which the defendant's awareness of statutory meaning is, therefore, material. This happens when a statute makes ML a good defense. A defendant may be able to get a good defense in some cases by saying that she did not know of the statute at the time of the offense or that she knew of it but misinterpreted it at the time of the offense. 73 But neither of those claims bears on the question of whether the defendant has made a ML, only on the question of whether the ML is a good defense in that case. Thus a taxpayer who misunderstands the meaning of "income" has a good defense in a criminal tax prosecution, 74 but that does not change the fact that "income" has a fixed meaning in her circumstances and her misunderstanding of that meaning is a ML, however exculpatory. The determination whether a particular mistake is ML or MF is done in the way argued above, and the role of the statute in that process is not that it is an object of the defendant's belief or mistake at the time of the offense but only that it is the source of the language in which the case must be discussed in the legal proceeding. A ML is a mistake in the use of that language, and it remains a ML whether or not it can support a defense under a particular statute. Thus when I say of certain beliefs or experiences that, under the law, they "entail an inference," I do not mean, of course, that there are bad consequences for getting the inference wrong; there are no bad consequences in those cases where ML is a good defense. I mean only that the statutory term--the law--has a particular meaning that reaches the set of circumstances found to have been in the awareness of the defendant.
C. Some Applications
I want to elaborate on and test this analysis now by briefly reconsidering the challenge I take to have been raised by Alexander. In the case I posed above, the defendant thought a gun was unloaded and, therefore, argued that he did not know it was a "deadly weapon." If a gun must be loaded to be a "deadly weapon" under this statute and if this gun was actually loaded, then it is true that, as a matter of law, this gun was a deadly weapon. One could then coherently say that the defendant made a ML when she formed her belief that this gun was not a "deadly weapon," no matter what the reasons for her belief. One could not refute the argument and insist that this mistake is actually MF just by saying that the finding is particular to one specific object.
If that were so, then a statute that, for example, gave the Rosetta Stone a certain status and protection could never give rise to a ML regarding the definition of the Rosetta Stone, since it is a unique object. Nor could one merely point out that the loaded-ness of the gun is a matter of perception or empirical judgment while its status as a "deadly weapon" is not. Both can easily be characterized as matters of perception or empirical judgment.
But I think the argument can be refuted by way of my formulation above: If the court finds that the defendant took the gun to be unloaded (and no disturbing circumstances are found), there can be no ML with respect to "deadly weapon." That is because the statutory meaning or definition of "deadly weapon" does not entail any inference to "deadly weapon" from the defendant's set of beliefs about the gun, preeminently her belief in its unloadedness. If the factfinder is convinced that the defendant believed that the gun was unloaded in the same sense in which the factfinder understands "unloaded," then the factfinder concludes that there is no mistake of meaning in the defendant's inference that the gun was not a "deadly weapon." The inference to "deadly weapon" can only be entailed by a defendant's belief in the loadedness of this gun (or perhaps some other unconsidered circumstance not present in this case), a belief which the court has no reason to attribute to the defendant. This defendant's mistake is not a mistake of inference, not a mistake of language use or meaning, and so not a ML, because the court has identified no part of the defendant's beliefs at the time of the conduct that would entail an inference to "deadly weapon"; i.e., that would instantiate the meaning of "deadly weapon." A similar analysis can be applied to Robinson's case of the liquor-seller who does not know whether he is within the boundaries of the dry town or an adjacent wet town. In Robinson's version, he relies on the opinion of the civil engineer of the wet town. 75 Setting aside the question of whether the defendant should have a defense derived from his reliance on an appropriate official's advice, 76 I can answer the question whether the defendant's ultimate mistake is one of law or fact by first identifying all of his relevant beliefs at the time he started selling liquor and then asking whether those things entailed an inference that the store was within the boundaries of the dry town. If 75 See ROBINSON, supra note ---, at 380 n.26. 76 One of the exceptions to the rule excluding ML defenses is sometimes that reliance on a mistaken interpretation of the law by certain officials will sustain a defense. See, e.g., MODEL PENAL CODE § 2.04(3)(b) (allowing defense when the actor reasonably relies on an official statement of the law contained in (i) a statute; (ii) a judicial decision, opinion, or judgment; (iii) an administrative order or grant of permission; (iv) or an official interpretation of the officer or body charged with responsibility for interpretation, administration or enforcement of the law); N.J. STAT. ANN. § 2C:2-4(c)(2) (West 1999) (same with addition of allowing defense when the actor diligently pursues all means available to ascertain the legality of his conduct and concludes in good faith that it is not an offense in circumstances in which a law-abiding and prudent person would also so conclude); TEX. PENAL CODE ANN. § 8.03(b) (West 1999) (allowing defense only when the official statement relied upon is in a written order or grant of permission by an administrative agency charged with interpreting the law in question, or a written interpretation of the law in an opinion from a court or public official charged with interpreting the law in question); ALA. CODE § 13A-2-6(b) (1999) (allowing defense only when the official statement is all the defendant really knew was what the engineer told him, then, of course, he was aware of nothing that would entail such an inference. If, however, he was aware of or believed something more, such as that the store was on the north side of Elm Street, then such an inference might well be entailed. If being on the north side of Elm Street necessarily places you within the boundaries of the dry town under the statute, then any failure to draw that inference is a ML. But if nothing that the defendant was aware of would entail that inference under the statute, then any failure to draw the inference that the store was within the dry town was a MF.
In the two sections that follow, I will take up some discrete problems of ML. In so doing, I mean to illustrate the utility of the definition suggested here in clarifying those problems. I also mean to clarify that definition by carrying it through some cases.
IV. MISTAKE OF "DIFFERENT LAW"
If my formulation of the ML-MF distinction is correct, then one logical consequence should be the elimination of the concept of "mistake of non-criminal law" or, as I will usually call it, "mistake of different law," from discussions of criminal doctrine. This sub-category of mistake actually goes under quite a few names in addition to the two just mentioned: mistake of "civil law," mistake of "legal fact," mistake of "legal element," and a few others. 77 There are some distinctions among the contained in a statute or the latest judicial opinion from the highest state or federal court to have ruled upon the matter). 77 There is a long list of scholars who have acknowledged or embraced this category of mistake under these and other names on the theory that there really is a difference in kind between these MLs and others. See, e.g., Singer & Husak, supra note ---, at 898 (mistake of "legal fact"); KELMAN, supra note ---, at 66 (mistake of "legal fact"); WILLIAMS, supra note ---, at 332-345 (mistake of "civil law"); HALL, supra note ---, at 364-72 (mistake of "private law"); PERKINS & BOYCE, supra note ---, at 1031 (mistake of "nonpenal law"), 1043 (mistake of "non-offense law"); FLETCHER, supra note ---, at 739-742 (mistake "extrinsic to the criminal law"); DRESSLER, supra note ---, §13.02 [D][1], 154-155 ("different-law mistake"); Simons, supra note ---, at 457 ("legal mistake concerning an element concepts behind these labels, but for my purposes they are essentially similar in that they carve out of the category of ML an arbitrary set of cases, or so I will argue, that are said to merit treatment as MF. These cases are said to form a conceptually coherent category of mistake that is not quite ML and not usually thought to be MF but an independent category of mistake for which the MF rules of exculpation happen to be more appropriate than are the ML rules. This category is sustained by the abiding desire to excuse some of the apparently less culpable MLs, but I think that a careful understanding of the ML-MF distinction makes clear that the category is impossible to sustain conceptually and, in effect, a device for evading direct confrontation with the hard questions raised by standard ML doctrine.
A. The Principle and Its Advocates
The classic Morissette 78 case provides a nice example of how the "different law" doctrine works. I will alter the case a bit to avoid irrelevant complications. A hunter spends the day on private land which has been posted to keep hunters off but which is commonly hunted anyway. While on the land, the hunter discovers a large pile of apparently unwanted metal at the side of the private road that passes through the middle of the property. The metal has clearly been there a long time. It is beginning to rust and become overgrown with weeds. The hunter, taking the metal to be abandoned, carts it off and sells it for scrap. Much to his surprise, he is arrested and prosecuted for larceny. of the offense" or "legal mistake/E"); Dutile & Moore, supra note ---, at 176-181 (mistake of "mixed fact and law"); Keedy, supra note ---, at 95-96 (mistake in "applying law to facts," mistake of "mixed law and fact," "mistake of fact," mistake involving "an element of fact," mistake of "private right"). 78 Morissette v. United States, 342 U.S. 246 (1952).
For conviction, the larceny statute requires that the hunter have taken the metal with an intent permanently to deprive the owner of her property. Courts have universally interpreted this offense to require that the defendant know that the goods legally belong to someone else. 79 Consequently, the mens rea requirement can be negatived in two different ways. First, to take an example of MF, the hunter could have failed to see the postings or other indications that he was on private property. He therefore thought he was encountering these unmarked hunks of metal on the shoulder of a state highway. As long as the property law of the state would have deemed such metal abandoned when found on a state highway but the property of the landowner when found on a private road, his mistaken belief was MF rather than ML. Nothing in his experience at the time of the offense would have entailed an inference to the statutory term "property of another." And he would have the valid defense that he did not know the metal belonged to anyone else when he took it away.
Alternatively, he might have been perfectly aware that he was encountering the metal on private land but still thought that the metal counted legally as abandoned and available for scavenging, perhaps because the land was commonly used by the public and the scrap had lain there undisturbed for more than a year. If the law preserves the right of the landowner to the scrap to the exclusion even of regular users of her land, and even when the scrap has lain undisturbed for more than a year, then the defendant's belief to the contrary is clearly enough a ML. But as long as the defendant genuinely held that belief, he would have a good defense to larceny in any Anglo-American jurisdiction. 79 See DRESSLER, supra note ---, at 521. Why is such a ML defense available? The reason could be just that larceny had some ad hoc ML exception read into it hundreds of years ago, presumably reflecting some pragmatic policy decision about the best way to protect property rights. But the usual argument is that such property-law mistakes in larceny exemplify the general, exculpatory category of mistake of "different law" or some such, a category distinct from ML. 80 That is, they have been thought an instance of the general principle that, when a criminal statute incorporates "civil law" by reference, a mistake defense is available with respect to the application of that other law to the facts of the case. Thus, since larceny incorporates the independent law of property, any material mistake regarding application of the law of property is a good defense, even though a mistake regarding the law of larceny as such is not a good defense.
This aspect of the law of mistake has recommended itself to many writers as a close cousin to MF doctrine. Thus, for example, the question of who owned the scrap metal in Morissette has seemed renderable as a question of fact, 81 even though Morissette appears to have had no misunderstanding of any fact except the "fact" that objects on private property do not count as legally abandoned even if they look unwanted. A similar approach appears in Liparota v. United States. 82 There, the defendant knew all the facts necessary to render his conduct an unauthorized acquisition of food stamps. 83 He knew that he was buying food stamps, that he was paying substantially less than their 80 See, e.g., WILLIAMS, supra note ---, at 332-345. 81 See, e.g., RICHARD J. BONNIE ET AL., CRIMINAL LAW 134 and n.d (1997), where the authors say that "Morissette's defense was based on a mistake of fact," although in a footnote they agree that "[i]t would be possible to characterize Morissette's mistake as not one of fact but of property law." In the actual Morissette opinion, there is no discussion of MF or ML but only of a requirement of "criminal intent" that the Supreme Court thought the trial court had abandoned. The Court thought so even though the jury instructions ruled out only the ML defense traditional to larceny while leaving a MF defense at least theoretically available. See Morissette, 342 U.S. at 275-276 (quoting and discussing the trial court's instructions). 82 Liparota v. United States, 471 U.S. 419 (1985). face values, that he was not in the grocery business, etc. In his defense, he claimed only that he did not know that those facts added up to lack of authorization under the statute. The Court approved this defense lest it attribute to Congress an intention to criminalize "a broad range of apparently innocent behavior." 84 Yet, faced with the dissent's claim that the Court thereby created a ML defense without any textual warrant, the Court insisted that, "Our holding today no more creates a 'mistake of law' defense than does a statute making knowing receipt of stolen goods unlawful." 85 Among commentators (as opposed to judges), there has been somewhat less confusion about whether this defense is a species of ML, 86 but plenty of uncertainty about how to define and justify this category of defense. Glanville Williams, for example, reviewed at length the many cases in which courts indulged a defendant's ML under the rubric of "claim of right." Larceny was a prime example, 87 but he found many others as well. 88 He also found a number of offenses to which the defense had not been applied even though it seemed to him that in principle it should have been, offenses ranging from bigamy to serious assaults. 89 He concluded that logically the claim of right defense should be available as a "general defence" to any criminal charge that depended on an interpretation of civil law, notwithstanding the ML maxim. 90 Even if it makes sense to require citizens "to make themselves acquainted with the criminal law," 83 The relevant provision is 7 U.S.C. § 2024(b)(1). 84 Liparota, 471 U.S. at 426. 85 Id. at 426 n.9. 86 Still, even sophisticated thinkers sometimes resist calling such mistakes "mistakes of law." Fletcher says that "mistakes about whether particular factual situations fall within a prohibition of the criminal law . . . are not, strictly speaking, mistakes about the interpretation of the norm and therefore do not qualify as mistakes of law." FLETCHER, Basic Concepts, supra note ---, at 167. On the other hand, some have taken judges to task for calling mistakes of different law mistakes of fact for the sake of results. See Smith, supra note ---, at 3, 5-6. 87 See WILLIAMS, supra note ---, at 320-327. 88 Id. at 304-331. 89 Id. at 332-341. he argued, still "it is going beyond reasonable social requirements to expect people to assimilate also, under pain of punishment, the whole of the civil law," such as, for example, the law of marriage and divorce that is central to any bigamy prosecution. 91 Williams recognized the tenuousness of the distinction between mistakes of criminal law and mistakes of civil law, but refused to give it up, offering instead this "convenient test": "whether the rule would appear in a well-drafted criminal code." 92 The mere statement of this "test" was enough to send Williams into retreat, where he ultimately rested on this stout conclusion: "It is submitted that the courts must do the best they can with such problems, asking themselves whether the mistaken rule lies, so to speak, at the centre of the crime charged, or belongs only to its general legal background." 93 Hall adopted a similar orientation. He asserted that "the penal law must be restricted to its distinctive field--to norms that prescribe certain punishment on the commission of certain harms," 94 and mistakes of non-criminal law or "private right" thus should be understood as a separate category from mistakes of criminal law.
Although he had a few pages earlier argued strenuously in favor of the traditional ML maxim, 95 he now said that "in some crimes opinion regarding certain private law qualifies the fact-situations. Such opinion functions as a fact, it engenders certain attitudes, and the resultant state of mind is relevant to penal law." 96 Hall did not explain why this statement was any less true when applied to rules that are widely characterized as "criminal" rather than "private." One might ask why the "fact- 90 Id. at 332. 91 Id. at 334. 92 Id. at 343. 93 Id. at 343-344. 94 HALL, supra note ---, at 365-366. 95 See id. at 351-364. 96 Id. at 367. situation" is not qualified by a mistake as to whether ephedrine is a "controlled substance" or not; or whether a fetus is a "human being"; or whether the conditions under which a woman "consents" to sex vitiate that consent. Why does the absence of a so-called "private right" in such examples have anything to do with whether the "fact- Fletcher's subsequent discussion of the fate of this principle in German law makes clear, I think, that the principle is simply that some MLs seem to make the defendant out to be morally non-culpable while others do not. Thus Fletcher discusses an incest prosecution in which the German Supreme Court declined to apply the principle in the face of an entirely plausible claim of mistake regarding the "non-criminal" law of marital affinity.
The defendant had an affair with the daughter of his former wife; that is, his one-time stepdaughter. He did not know that the German civil code rendered this relationship of "marital affinity" permanent, even after the divorce. But the court denied him the 97 Id. mistake defense. It deemed this mistake an inexcusable mistake regarding the meaning of a criminal prohibition, even though it could as easily have treated it as an instance of the supposed principle that mistakes regarding private law are exculpatory. Fletcher condemns neither the Court's decision nor the supposedly general principle that the Court flouted. He only notes that the apparently "sensible" civil-criminal distinction in this area was broken down in Germany by "the weight of casuistic argument" and by recognition that "the distinction between mistakes of criminal law and mistakes of private law was not as workable and dependable as many had thought." 99 From this recognition, Fletcher, unlike Williams and Hall, draws the conclusion that the real question is not how to justify the category of mistake of private law as analytically separate from ML but why any particular ML (civil or criminal) should be recognized as a defense. 100 The only recent analytical advance in this area, I think, has come from Simons.
Simons recognizes that the distinction between criminal law and civil law is inappropriate here because sometimes the "different law" at stake is, in fact, a different criminal law. 101 He offers as an example the scenario in California v. Bray. 102 There, Bray was charged with being a felon in possession of a concealable firearm. In his defense, he claimed that, while he knew that he had previously been convicted of an offense in another state, he did not know that that offense had made him a "felon." The court 98 FLETCHER, supra note ---, at 740. 99 Id. at 740-741. 100 See id. at 742. 101 called this mistake a MF 103 and held it a good defense. Simons sensibly suggests that this mistake actually fits the category of mistake of "different law" and that the different law here was, at least arguably, a criminal law. Simons also sensibly implies that Bray's mistake of different criminal law suggests his non-culpability at least as strongly as most mistakes of civil law suggest non-culpability. He made a mistake on a difficult question of law and gave ample evidence of his positive efforts to comply with the law. Therefore, the case nicely supports the argument that, if there is to be a defense of mistake of "different" law, it should not turn on whether that different law is civil or criminal.
More than this, though, Simons seems to recognize a basic difficulty with separating out "different law" cases at all. He suggests that the real concern underlying that category is that there is a morally important difference between the meaning of the case's "governing law"--the general statement of the conduct to be prohibited--and the narrower, perhaps technical, meanings of particular "legal elements" within that governing law, whether the latter are defined by reference to some different law or not.
Bray can again provide the example. One way to read the case is that the word "felony" in the possession statute was not defined in any other law but had to be interpreted from scratch for the purposes of the possession offense. Although everyone could state the basic statutory idea that felons must not possess guns, neither Bray nor his prosecutor was able confidently or easily to determine whether Bray was a "felon" simply by referring to some law of the jurisdiction in which he had previously been convicted. 104 There was no "different law" that settled the meaning of this element of 103 See Bray, 52 Cal. App. 3d at 499. Note that in the very same case the opinion referred to the same question as a question of law when it was deciding another point. See id. at 498 n.2. 104 Id. at 498. the possession statute. Thus the "different law" principle might seem not to apply. Yet, to Simons, there seems little reason to distinguish this sort of mistake from mistakes that do involve a "different law." Rather, the latter seem like a mere subset of the more general case of mistakes regarding the "legal elements" of the charging statute. Simons thus embraces the need to refine the distinctions among "legal mistakes" even further than Williams, Hall, and others had. His approach would altogether eliminate the requirement that the mistake relate explicitly to some other law besides the charging statute. And he would relabel mistakes of "different law" as mistakes of "legal element" (or "legal mistake/E"). He distinguishes such mistakes from mistakes regarding the "governing law" of the case ("legal mistake/GL"), which correspond to the classic category of ML. 105 Even after this refinement, however, Simons recognizes the continuing fragility of the distinction between a mistake regarding the reach of the governing law, on the one hand, and mistake regarding the meaning of the particular elements of that law, on the other hand. 106 After all, what gives the governing law its meaning and its reach but the meanings of the elements that constitute it? How does one understand the wrong expressed in the phrase "felon in possession" without an understanding or definition of the "legal element" expressed in the word "felon"? Still, Simons wants to retain his refined distinction (at least as long as the basic ML maxim remains in place), because he believes that it more nearly limits criminal liability to cases of moral culpability than would otherwise be the case. 107 105 See Simons, supra note ---, at 457-459. 106 Id. at 493. 107 Id. at 500. I think Simons has advanced the ball with his dismissal of the criminal-civil distinction, and I think he is right to search for a way of accommodating the dual impulses to exculpate for some MLs and yet to retain a pretty broad presumption against ML defenses. 108 Nevertheless, I think this review of the literature suggests that the effort to refine an analytical category of legal mistake separate from ML has so far proven unproductive. Take the firearms statutes again for a first example. Suppose the defendant says that he misunderstood the licensing requirement of the firearms laws. He knew he was selling firearms and knew that he had to have a license, but he thought that having a driver's license and a social security card would satisfy the requirement. He would be charged under 18 U.S.C. §924(d), which establishes penalties for one who "willfully violates any . . . provision of this chapter." To get a §924 conviction in this case, therefore, we would have to refer to another "provision of this chapter," here 18 U.S.C. §922(a)(1)(A), which renders dealing in firearms an "unlawful act" unless done by a "licensed dealer." But what constitutes a "licensed dealer"? To find out, we would have to refer to 18 U.S.C. §923, which details the law of licensing in this area. The §924 offense element "any . . . provision" thus takes its meaning from §922's "licensed dealer" requirement, which, in turn, takes its meaning from §923, a provision of law separate from the charging statute itself ( §924). In principle, this process of moving from the charging or criminalizing statute through one or more other sources of law to figure out whether the meaning of that criminalizing statute has been satisfied is identical to the process by which the meaning of "property" in a standard larceny statute is determined. A "different law" analysis under the firearms statutes would thus suggest that the question whether the defendant had a license, like the question whether certain objects were "property of another," should be treated as if it were a question of fact.
B. Problems with the Principle and Its Rationales
Once you are committed to the "different law" principle, it may prove difficult to tell what counts as the charging statute or "governing law" in the first place. Is it just §924, the penalties provision, which is the only provision that actually links any statutory conduct to criminal process? Or is it also §922, the "unlawful acts" provision?
Or does it further include §923, the arguably civil provision that lays out for §922 and for the gun industry what it is to be a "licensed dealer"? Moreover, even if we can agree how to locate the "governing law" as distinct from some "different law," it is hard to see what significance that should have. If possession of a driver's license and a social security card were not enough legally to constitute the defendant as a "licensed dealer," then the defendant has made a ML as to the definition of "licensed dealer" and thus as to the meaning of the charging statute, the circumstances that satisfy the "any other provision" element of §924. That the law of licensing in this case may have to be found somewhere other than in the code section that actually criminalizes dealing in guns without a license should not obscure that any mistake as to the legal definition of "licensed dealer" is a mistake with respect to the scope of or the definitions in the charging law. In fact, it is easy to imagine the licensing provision being formally merged into the more overtly criminal provisions of the gun-dealing statute. And it is hard to imagine why the formal location of the definition in or out of those code sections should make any difference to what counts as a ML with respect to this offense.
I suppose one could put some weight on the fact that Title 18 as a whole is called "Crimes and Criminal Procedure," suggesting that all of those provisions in the previous paragraph, unlike the property law that larceny law depends on, obviously count both as criminal and as parts of the charging or governing statute. But, even apart from the flimsiness of any argument that relies entirely on labels without a theory of why or how those labels convey a meaningful substantive distinction, I could easily run through a similar exercise for similar statutory provisions that appear in Title 26, the Internal Revenue Code. 109 Section 5871 of Title 26 provides that, "Any person who violates or fails to comply with any provision of this chapter shall, upon conviction, be fined not more than $10,000, or be imprisoned not more than ten years, or both." That 109 See United States v. Syverson, 90 F.3d 227 (7th Cir. 1996) for a case governed by the firearms regulations in Title 26. See also, for example, Justice Stevens's very plausible assertion that "the current National Firearms Act is primarily a regulatory measure. The statute establishes taxation, registration, reporting, and recordkeeping requirements for businesses and transactions involving statutorily defined firearms, and requires that each firearm be identified by a serial Similarly, the fact that the law governing ownership of apparently abandoned property sometimes appears under the heading "property law" rather than "larceny law" should not obscure the fact that any mistake as to the legal criteria of ownership is a mistake as to the criteria of the "property of another" element of the larceny statute.
Again, it is not hard to imagine the entire law of property as an elaboration of the word "property" in the larceny statute. In fact, it is apparently the case that property law has developed in very significant part not as an independent body of law but as a part of the ongoing definition of the crime of larceny, 113 just as the definitions of such terms as "human being" 114 and "deadly weapon" 115 have been developed in large part not as independent bodies of law but simply as parts of the murder and assault statutes. If there had never been any civil law of property at all, such that the idea of property expressed rights that were to be protected entirely or predominantly by criminal rather than civil process, then presumably defendants' mistakes as to the definition of property would be understood unambiguously as simple mistakes of law for which there is presumptively no excuse. Conceptually, the case would be identical to a case in which someone pointed an unloaded gun at another and was later surprised to learn that "deadly weapon" includes unloaded guns as well as loaded ones.
Complementarily, just about any criminal offense that involves identifiable victims will yield parallel civil suits in which overlapping legal terms will get authoritative interpretation. Thus, just as "property" turns out to be understandable simply as an element of a criminal statute (not just as a term of "civil" law), so elements of criminal statutes can become elements of civil suits. Even terms like a murder statute's "human being" element, which no one ever seems to have thought of as implying a potential "different law" mistake, can nevertheless raise difficult problems of interpretation that are simultaneously encountered in civil cases. For example, the case 113 HALL, supra note ---, at 365. For some possible early examples, see J. G. BELLAMY, THE CRIMINAL TRIAL IN LATER MEDIEVAL ENGLAND 78-79 (1998). Bellamy discusses some refinements of the law of larceny that took out of the reach of the offense takings of beasts or trees from their natural habitat and takings of objects of no monetary value. These refinements, made in larceny cases, appear to have been made without reference to any preexisting definitions of property to these effects in civil cases. 114 See, for example, LAFAVE & SCOTT, supra note ---, at 607-611, which discusses the issue of when life begins and ends for purposes of homicide law. There is, of course, no suggestion in the discussion that a defendant's ML regarding the definition of "human being" might be a mistake of "different law" or "legal element" or the like, and the citations unsurprisingly appear to be mostly criminal law cases. 115 LAFAVE & SCOTT, supra note ---, at 696 and n.38.
of Massachusetts v. Cass 116 relied on the definition of "human being" in prior wrongful death cases to hold that a viable fetus was a human being within the meaning of the state's criminal vehicular homicide statute. 117 Although that court made its ruling prospective because the holding was arguably unforeseeable, 118 it gave no hint that the rule's origins in non-criminal law would give rise to a special ML defense.
Subsequently, in Massachusetts v. Lawrence, 119 the rule from Cass was recognized for murder as well, and, while the defendant unsuccessfully argued that again the rule ought to be made prospective, there was no hint that anyone thought the principle of "mistake of legal element" or "different law" ought to apply just because the dispute turned on the meaning of the legal element "human being" or because the meaning of "human being" in this instance derived from the civil law of wrongful death. In this sense, the boundary between criminal and non-criminal law is quite permeable, 120 as is the distinction between the governing law of a case and the elements of that governing law. Neither of these distinctions, therefore, provides an unproblematic basis for distinguishing classic MLs and classic MFs from some third category of mistake.
In short, once one recognizes that all questions in a criminal case are generated by the meaning of the elements of the criminalizing statute, it becomes hard to see any special significance in an element's depending for that meaning on an existing body of law (even less the happenstance that the law is denominated "civil" rather than "criminal") rather than on common usage or a dictionary or the like. Why should that circumstance create a rule or even a presumption that a mistake be treated as a MF rather than a ML when a mistake as to the meaning of another term in the same statute would immediately be characterized as a ML for which there is no excuse? Similarly, why should the infinitely manipulable distinction between a ML and a mistaken "claim of right" yield a rule or presumption that the former is no defense but that the latter generally is? 121 If legal meaning is a matter of specifying all of the relevant inferences entailed by one or another term in that statute, then mistakes of "different law" are simply mistakes of statutory inference and thus conventional MLs, not some conceptually distinct category at all.
Simons, for one, seems prepared to drop the civil-criminal distinction and the "different law" idea as such, but he retains the idea that there is an underlying principle behind those concepts that should not be lost. Simons's notion is that mistakes regarding the general norms embodied in "governing law" are different from mistakes regarding the "legal elements" within that law. 122 But it is at least questionable whether 121 I take Fletcher, for one, to agree that this distinction is highly manipulable. See FLETCHER, supra note ---, at 741. More recently, see FLETCHER, BASIC CONCEPTS OF CRIMINAL LAW 158 (1998). In this latter work, Fletcher also makes the illuminating observation that in German law the general principles embodied in the statutes are that a defense is available where the defendant mistakes the general norm of wrongdoing but never when the mistake regards the details of statutory meaning. Without being too facile about what concepts correspond to what in American and German law, this seems at least roughly to flip the American rules discussed in the text on their head, and Fletcher himself points out that there is a pretty good argument for saying that the German statute got the culpability analysis backwards. See id. at 158, 169 n.27. 122 Fletcher relies on a similar idea, although Fletcher's nomenclature distinguishes between "norms" and applications of those norms or "whether particular factual situations fall within a prohibition of the criminal law." See FLETCHER, Rethinking, supra note ---, at 686; FLETCHER, Basic Concepts, supra note ---, at 167. But this analysis too seems to me little more than an argument that some MLs are non-culpable and so should be exculpatory. I do not see that he produces any particular progress on the question of how to tell when one has passed from the world of "norms" to the world of legal detail or application, nor on the question of how to tell which MLs should be excused other than by throwing the question on the moral judgment of the judiciary. Cf. Wiley, supra note ---, at 1023 (arguing that judges should interpret all statutes to include whatever mens rea is required to render the defendant morally guilty without regard to the ML-MF distinction); Kahan, supra note ---, at 128-130 (endorsing judicial authority to apportion ML there is a meaningful difference between those two kinds of mistakes. If it is criminal for a felon (but not others) to possess a gun, that norm only has meaning insofar as "felon" has meaning, and "felon" only has the meaning courts give it in the course of applying it (or not) to specific scenarios. As in any case, the meaning of the law lies in the statutory terms as they are elaborated on the facts of cases. It is just as hard for me to see the difference between the meaning of a general norm and the meanings of the elements that constitute the norm as it is for me to see the difference between an element of a criminal statute and a point of "civil" law that gives that element its meaning. In the end, the result of the "different law"/"legal element" approach has been to obscure the conceptually stonger ML-MF distinction in a conceptually confusing, however laudable, attempt to soften the ML rules.
C. Can the "Different Law" Principle Be Salvaged?
That is not to say that there should be no softening of the ML rule. Certainly, different MLs might be evaluated differently in terms of moral culpability, whether they are labeled mistakes of governing law or mistakes of "legal element" or whatever. A general excuse for grossly unfair applications of criminal law is at least theoretically in place in American law 123 and is easy to love. Even apart from cases that refer explicitly to the "different law" principle, courts have found a number of ways--not always defenses according to each statute's relationship to general morality); Regina v. Prince 2 Cr. Cas. Res. 154 (1875) (op. of Bramwell, J.) (similarly relying on the moral judgment of the judiciary to determine the scope of mens rea to read into statutes). 123 See Lambert v. California, 355 U.S. 225, 229 (1957). Lambert established the constitutional principle, almost never applied in actual cases, that ignorance of the existence of a law is a good defense where there is little or nothing to suggest to a defendant that her conduct might be criminal. See also Wiley's argument that recent Supreme Court cases have established a Lambertlike principle at the level of statutory interpretation. He believes that the Court will now read into every statute whatever sort of mental element is necessary to prevent the criminal conviction of the morally blameless. See Wiley, supra note ---, at 1023. satisfying--to read ML defenses into occasional statutes in the name of justice, while generally reaffirming a strong presumption against ML defenses. 124 And it is that impulse to find some apparently coherent way to soften the ML rule without eliminating it altogether that seems to have kept the "different law" principle alive.
One can see this impulse clearly enough in the defenders of a general "different law" principle. Without usually calling for a general departure from the ML maxim, they point to the "vicissitudes" 125 or complexities of mastering the "whole of the civil law" 126 and the consequent unfairness of expecting the citizenry to internalize all of that law or, more generally, the simply unjust results that would happen without some version of this defense. 127 One might then suppose that the "different law" principle can be salvaged by recognizing it as a stand-in for analyses of legal complexity and justice to the individual. When considering whether a statute carries a ML defense, courts often do consider questions of complexity and individual fairness; perhaps the "different law" principle has served as a useful label and analytical clue in such cases.
As a matter of fact, however, the "different law" category has not consistently served as the avenue through which courts have found exceptions to the ML rule; nor has it been a consistent stand-in for the considerations of legal complexity that scholars point to as its rationale. Once one has seen the conceptual difficulties with the defense in the first place, none of this is surprising. Rather, it becomes easy to see the empirical doubtfulness of the claim that the law of mistake has developed by the logic of the "different law" doctrine. And it becomes equally easy to doubt that the criterion of complexity distinguishes civil law from criminal law and, more generally, that "different law" defenses have been or promise to be the key to a just set of ML rules.
First, although degrees of legal complexity do seem to be considered by courts when deciding whether to read ML defenses into statutes, 128 courts have not needed to characterize a case as one of "different law" to do so. For example, in Cheek, the Supreme Court reaffirmed the availability of a ML defense in federal tax prosecutions by pointing to the complexity of the Internal Revenue Code and noting that "in 'our complex tax system, uncertainty often arises even among taxpayers who earnestly wish to follow the law.'" 129 But the Court never considered the notion that the charging statute's dependence on the provisions of the tax code might call into play some "different law" principle.
Similarly, in United States v. Baker, the Ninth Circuit recognized "complex regulatory schemes that have the potential of snaring unwitting violators" as a general category of cases which justify a ML defense. The charging statute in that case, the Contraband Cigarette Trafficking Act, was not, however, sufficiently complex to justify such a defense. Rather, even though the definition of "contraband" in the federal charging statute depended in part on Washington state's tax laws, the court held that the law in the case was "quite simple" and so did not warrant such a defense. 130 The 128 See Davies, supra note ---, at 363-373. 129 Cheek v. United States, 498 U.S. 192, 205 (quoting United States v. Bishop, 412 U.S. 346, 360-361 (1973)) (1991). In addition, see Davies's general discussion of complexity as a flexible criterion for determining when a ML defense should be read into a federal statute. There is no indication in her discussion that any of these cases grappled with the "different law" principle but only with the complexity of the legal issues entailed by the charging statute itself, whether those legal issues included reference to different laws or not. 130 See United States v. Baker, 63 F.3d 1478, 1491-92 (9th Cir. 1982) (recognizing mistake of law defense for complex regulatory schemes, but holding that Contraband Cigarette Trafficking Act, 18 U.S.C. § 2341-46, is "quite simple"). Baker cites United States v. Fierros, 692 F.2d 1291, 1295 (9th Cir. 1982), which had previously recognized in principle a ML defense for complex regulatory schemes but had held that 8 U.S.C. § 1324(a), regarding transportation and harboring court had plenty of opportunity to apply the "different law" principle here. The defendants apparently contended that their failure to know that taxes were owing on their cargo of cigarettes or that they needed special permission to bring unstamped cigarettes into Washington should make out a good defense. Such a claim is, of course, a claim of mistake of different law--a mistake about Washington's tax law that would arguably negate the intent required by the federal charge. But the court, while acknowledging that such a category of ML appeared in some cases, saw no need to apply the category here but simply considered the issue of complexity on its own merits as an independent criterion for recognizing a ML defense, as any court may do in response to any ML claim.
Baker is no anomaly. Courts have often denied ML defenses in cases where the law is quite complex, even when the criminal provision required reference to such presumably "different" laws as administrative regulations. Take, for example, the regulatory cases under the Resource Conservation and Recovery Act (RCRA), 131 which criminalizes mishandling of hazardous waste while also establishing a generally civil statutory scheme. These cases have held that as long as the defendants knew the facts that brought the waste products within the statute they satisfied the criminal provision's mens rea requirement, regardless of whether they knew of the criminal provisions or the statutory definitions or the EPA regulations that defined their substances as "hazardous wastes." 132 of illegal aliens, is "not so complex . . . that the unusual defense of ignorance of the law must be read into it." 131 42 U.S.C. §6928(d). 132 See, e.g., United States v. Baytank, Inc., 934 F.2d 599, 612 (5th Cir. 1991) (holding that it was not necessary for defendants to have known that the waste they were storing had been identified by EPA regulations as hazardous under the RCRA in order to support conviction for improper storage of hazardous waste); United States v. Dee, 912 F.2d 741, 745 (4th Cir. 1990) (holding that Does either the complexity of the civil law or the "different law" principle explain the difference between the treatment of larceny and bigamy? The claim of right defense is universally applicable to larceny. And, among the defenders of the "different law" principle, it seems to be an article of faith that the defense ought to apply to bigamy as well, since it would be a gross injustice to hold individuals to a knowledge of the complex details of the civil law of marriage and divorce and thus to the legality of a remarriage undertaken in good faith. 133 The advocates of the "different law" distinction usually take this as a paradigm case for justifying the distinction, but the courts have usually rejected the defense in bigamy cases. 134 The difference in treatment between larceny and bigamy seems to represent an ad hoc conclusion that the criminal law should not go after people who had a good faith belief in their right to a piece of property, even though it should go after people who had a good faith, although mistaken, belief in their right to remarry. defendants "knowingly" violated the criminal provisions of the RCRA even if they did not know that violation of the RCRA was a crime or that regulations existed identifying chemical wastes as hazardous wastes under the RCRA); United States v. Laughlin, 10 F.3d 961, 965 (1993) (holding that knowledge of regulatory requirements is not necessary to convict defendant under RCRA for knowingly disposing of hazardous wastes without a permit). See also the cases cited by Davies, supra note ---, at 370 n.119.
Courts have been willing to entertain complexity-based arguments for ML defenses in a variety of other statutory and regulatory contexts as well without recourse to the "different law Similarly, as I have suggested that there should be no "different law" defense available as to the definition of "firearm" under the federal firearms laws, so the courts seem to have held, even though that term is defined in an extremely non-intuitive way in a provision separate from the criminalizing provisions. 135 And there is no "different law" defense as to the definition of "controlled substance" under the provisions of the federal drug laws, even though that term too is defined in a set of changeable, complex regulations separate from the sections that declare unlawful acts and establish criminal penalties. 136 One final example: in New York v. Marrero, 137 the defendant sought to defend against a weapons charge on the grounds that he had thought he was a "peace officer" within the meaning of the charging statute, a term that was defined by reference to a separate provision in New York's law of criminal procedure, not in its "penal law." But the court refused to instruct on this ML claim, which no one even thought to label a mistake of "different law" or the like. 138 These cases may or may not be unjust, but they all seem to refute the notion that the "different law" principle explains the law as it stands or the related notion that that 135 Staples held that the defendant need only know the facts that bring the weapon within the statutory definition of "firearm"; conviction does not require that the defendant also have been aware that those facts meant "firearm" under the statute. Only very recently have I even heard of someone hazarding the idea of a mistake of "legal fact" defense in this context. Singer and Husak suggest that just such a defense may possibly be implicit in Staples. See Singer & Husak, supra note ---, at 898 and n.125. I don't see how Staples can be read that way, since the opinion repeatedly says that the question at issue is whether the government had to prove awareness of the "facts" or "features" or "characteristics" that made Staples's gun a statutory "firearm" or, alternatively, something less, such as mere awareness that it was a gun (which may or may not be a "firearm" . 250 (1922)). Section 841(a)(1), for example, makes it unlawful to do various things with respect to a "controlled substance." But that term is defined by 21 U.S.C. §802(6), which refers the reader to 21 U.S.C. §812, which establishes various "schedules" and places certain drugs provisionally on those schedules, and by 21 U.S.C. §811, which authorizes the Attorney General to alter the contents of those schedules through an administrative process. 137 New York v. Marrero, 69 N.Y.2d 382 (1987). principle gets at some special kind of complexity in the law. There is no reason to think that criminal law as such is somehow easier to interpret and know than any other body of law, even apart from the slippery question of how you are supposed to tell what laws are "criminal" and what laws are "civil." In the words of the New York Court of Appeals, "There would be an infinite number of mistake of law defenses which could be devised from a good-faith, perhaps reasonable but mistaken, interpretation of criminal statutes, many of which are concededly complex." 139 As every first-year student of criminal law knows, the criminal law does not need to look to civil law to import a complexity that it would otherwise lack. And the courts have given little evidence that the "different law" principle, informed by the supposed complexity or vicissitudes in the meaning of civil law terms, has really been a conceptually appropriate or necessary tool for identifying the instances of complexity that ought to yield a mistake defense. 140
D. "Different Law," Mistake of Law and the Ambiguity of Criminal Intent
This analysis leaves something of a void. If the "different law" doctrine is conceptually unsound, then why have the larceny cases and some others embraced some strictly limited ML defenses while the bigamy cases and others have rejected ML defenses? I think that is less a question about the philosophy of criminal law than about its history. Here, I will not pretend to have a final answer to that historical question, but I would suggest that the beginning of an answer lies in recognition of the different, 138 See id. at 385. 139 Id. at 391-392. 140 I also take the view that the MPC too has freed itself from reliance on any general "different law" principle. Rather, it has reaffirmed the basic ML maxim in its §2.02(9), while opening the door to ML defenses only on a statute-by-statute basis. See MPC §2.04(1); ROBINSON, supra note ---, §62(d), at 262-263. perhaps contradictory, but coexisting principles of criminal intent that have guided courts' attributions of criminality.
Criminal courts have long considered fundamental to criminal law the apparently interchangeable notions of criminal intent, mens rea, guilty knowledge, malice, and the like. But courts and commentators have also long understood that these terms have many and sometimes conflicting meanings. 141 Each can be used to signify the basic principle that criminal law requires inquiry into a defendant's mental state during commission of the criminal conduct. When combined with the principle that ignorance of the law is no excuse, the basic doctrine seems to be that a defendant must have some level of awareness of the facts (or at least some of the facts) that constitute a crime but need not have any awareness of the law that defines that crime. But the "different law" cases, among others, suggest that the notion of "criminal intent" actually signifies, ambiguously, two conflicting principles of criminal liability. Sometimes, as suggested above and as in most formal elaborations of its meaning, criminal intent just means knowledge of the facts that constitute the crime at the time of one's conduct. 142 But other times a court will pass that principle by without so much as acknowledging its existence and discuss criminal intent as a diffuse quality of mind that might be negatived by ignorance or mistake of law--although, again, frequently courts will not 141 The classic statement is in Francis Bowes Sayre, Mens Rea, supra note ---, at 1026. More modern recognitions of the hazy meanings of mens rea terms can be found in, for example, DRESSLER, supra note ---, at 101-102 and ch. 10, passim; LAFAVE & SCOTT, supra note ---, §3,5, 216. 142 See, e.g., Bryan v. United States, 524 U.S. 184, 193 (1998) (holding that "unless the text of [a] statute dictates a different result, the term 'knowingly' merely requires proof of knowledge of the facts that constitute the offense"); Boyce Motor Lines v. United States, 342 U.S. 337, 345 (1952)(Jackson, J., dissenting)(stating that the knowledge requisite to knowing violation of a statute is factual knowledge as distinguished from knowledge of the law); American Surety Co. of New York v. Sullivan, 7 F.2d 605, 606 (2nd Cir. 1925)(Hand, J.)(stating that the term 'willful' in criminal statutes "means no more than that the person charged with the duty knows what he is doing", not that "he must [also] suppose that he is breaking the law"); PERKINS & BOYCE, supra acknowledge that they are endorsing a ML defense but will simply speak in terms of the immediate statutory language 143 or the unexplained language of "criminal intent." 144 Morissette is an excellent example of the latter approach. Justice Jackson writes in ringing terms of the fundamental role that the notion of criminal intent plays in our system of justice and concludes from that principle that Morissette must have a defense. 145 Never for a moment, however, does he advert to the fact that he is reading a ML defense into the statute. 146 Never does he consider that, contrary to the hoary maxim, he is embracing the defendant who would say, "Yes, I knew I was taking goods off someone else's land, but I did not know that counted as stealing--I did not know that that was against the law." I do not mean to suggest that the holding in Morissette was wrong but only that it helps to reveal the persistent ambiguity in the meaning of criminal intent. Jackson's holding was entirely consistent with the interpretive consensus for larceny; and that consensus may well represent good social policy. Like so many other judges, however, 250-263 (1952). 146 In fact, he seems positively to think he is not when he says that a defendant "must have had knowledge of the facts, though not necessarily the law, that made the taking a conversion," even as the rest of the opinion and the universal history of larceny law on which he so heavily depends reasons, perhaps very compelling reasons, why someone who takes valuable property from another person's land should be allowed to defend by simply saying that he did not know there was a law against it, even though the general principle of the law is that such excuses are no excuse at all.
Another remarkable example comes from criminal conspiracy doctrine. Many courts have held that ignorance of the illegality of an act is a good defense to a charge of conspiracy to do that act, at least where the core crime can be said to be "morally innocent" (although prohibited by statute). This doctrine would make easy sense where the core crime itself provided for a ML defense, but it also holds where ignorance of law is no defense to the core crime. 147 The case that seems to have originated the doctrine relied on nothing more than its own understanding of the notion of "criminal intention," as entailed by the term "conspiracy." It considered the principle of criminal intent to require that a ML defense should be available on the conspiracy charge, even as it reaffirmed that no such awareness of the law, but only the facts constituting the offense, was necessary to constitute the intent required to convict for the core crime. 148 Similarly, in Liparota, the Court created a ML defense to the crime of unauthorized transfer of food stamps on the grounds that "mens rea" was a fundamental principle of criminal law, without acknowledging that an entirely conventional understanding of mens rea is mere awareness of the facts--not the law--that constitute the crime. 149 makes clear that he means to exculpate anyone who fails to understand that the property is someone else's, regardless of the source of that belief. See id. at 271. 147 See LAFAVE & SCOTT, supra note ---, § 6.4(e)(5). 148 A final example comes from Bryan v. United States, 150 a prosecution for dealing in firearms without a license. Bryan argued that the prosecution should have to show that he knew of the statutory provision that he violated. The Court denied his claim but accepted that the statute did contain a requirement that a defendant have knowledge that his conduct was "unlawful." 151 Although the Court acknowledged that criminal "knowledge" normally means knowledge only of the facts that constitute the offense 152 and insisted that it would not carve out an exception to the rule that ignorance of law is no excuse, 153 it nevertheless read the statutory requirement of "willfulness" to require "bad purpose" and thus to require that the defendant have known that his conduct was "unlawful." 154 What might it mean to know that some conduct is "unlawful" without knowing the specific law that makes that conduct unlawful? The Court did not enlarge on this rather central question. 155 Under the circumstances, the Court's insistence that it was not creating an exception to the ML maxim suggests that the Court thought that such an exception arises only when the defendant has to have known the citation to the statute or its precise wording. But that cannot be. So why doesn't the requirement that the defendant have known that his conduct was "unlawful" constitute an exception to the ignorance-of-law maxim? Of course, it really does. But it is an intentionally vague exception, designed--or at least destined--to fudge the meaning of the ML-MF boundary, perpetuate the ambiguities of criminal intent, and thus preserve the judiciary's capacity to avoid the injustices of an excessively formalist criminal law 150 Bryan v. United States, 524 U.S. 184 (1998). 151 Id. at 193. 152 Id. at 192-3. 153 See id. at 196. 154 Id. at 196 (1998). transcendently deserved punishment, not to mention those who seek a synthesis of those two perspectives, 159 will persist indefinitely. As that negotiation persists, the multiple meanings and discourses available to judges may well be the tools necessary to the administration of "criminal justice" within unsettled bounds. Scholars who lack the ambition to put an end to that age-old negotiation might do well simply to delineate as clearly as possible the multiple discourses of criminality available to the judges, whether simply to describe as accurately as possible the ways in which discretionary judicial choices get made 160 or to help clear the path to sensible doctrinal development.
V. UNREASONABLE MISTAKES: LAW OR FACT?
The ambiguities of criminal intent suggested above are not confined, of course, to the doctrinal niche of the "different law" mistake. The perspective provided by this article's understanding of ML suggests at least one other doctrinal area--"unreasonable MF"--in which those ambiguities have been obscured by inadequate attention to the meaning of ML. What is an "unreasonable MF"? For one thing, it is an unquestioned category of analysis in criminal law. At the same time, however, it is a contradiction in terms, or so I think the definition of ML in this article suggests. An implication of the ML-MF distinction as defined here is that the concept of ML extends to any material unreasonableness on the part of the defendant. More precisely, any time a mistake is actually found to have been unreasonable, even where the mistake is conventionally understood as a MF, the mistake is really a ML. This claim runs contrary to well entrenched ways of talking about criminal law, but I think the point seems obvious: whenever the law requires a person to make the inferences that a reasonable person would have made in the situation, then any mistake resulting from a failure to draw those inferences is ML. To allow a criminal intent defense in such a situation without acknowledging that one is indulging a ML defense is, again, to obscure the meaning and amibiguity of criminal intent.
This point illustrates and strengthens the larger point of this article: that whenever a factfinder would infer the satisfaction of a statutory element from the defendant's own experience of the offense conduct and circumstances, then the defendant's failure to draw that same inference is a failure to understand the meaning-in perhaps a special legal sense of "meaning"--of the statutory term and is, therefore, a ML. Perhaps the word "meaning" takes on a special legal sense in this article, since the "meaning" of a statutory term is determined exclusively by, or at least given effect exclusively by, the factfinder. The factfinder determines the function or effect of any particular statutory term in any particular scenario and thereby controls its operational meaning. If a certain body of experience would, for the factfinder, mean satisfaction of a statutory term, then that experience or evidence instantiates the meaning of the statutory term, at least for purposes of determining criminal liability. And if the defendant nevertheless believed in the legality of her conduct--failed to draw the inference that the statutory term is satisfied by that body of experience--then she has made a ML. Only if the defendant's experience would not entail an inference to the statutory element would the defendant's mistake count as MF.
What I want to suggest in this section is that, given this understanding of ML, the notion of an unreasonable MF is a contradiction in terms. A finding of unreasonableness can only happen when the law makes reasonableness a standard of guilt or innocence. If the law has made the question of reasonableness material to liability, then the failure to draw the inferences that a reasonable person would draw from known evidence is a ML.
I do not mean, however, to say that the line between reasonableness and unreasonableness is always the same as the line between ML and MF. Rather, I will suggest that, while any materially "unreasonable" mistake must be a ML, the implication does not run the other way; a ML may be "reasonable" or "unreasonable." Thus a "reasonable" mistake may be ML (as long as it is a mistake about meaning) or it may be MF (if it is not a mistake about meaning), even though a finding of unreasonableness always implies a mistake of meaning or a ML. An unreasonable mistake must be a ML because the finding of unreasonableness implies that, given the things of which the defendant was aware, the defendant should have carried out the chain of inferences entailed from those things to the statutory term; that is, the defendant should have seen that the meaning of the statutory term was instantiated in the things of which she was aware. 161 There is no reason, however, why some MLs, 161 There may possibly be the following objection to my argument here (although in the end I think it fails). Take the case suggested in Larry Alexander, Insufficient Concern: A Unified Conception of Criminal Liability, 88 CALIF. L. REV. 931 (2000), where a distracted parent is momentarily "unaware" of his infant's being in a filling bathtub even though he put her there himself only a few minutes before. The distraction is only the arrival of dinner guests who the career-obsessed parent is intent on impressing, but it is enough to put the child out of the parent's mind and result in the child's drowning. In such a case, one might suppose that a factfinder could find that a particular circumstance (the infant's presence in a running tub) is indispensable to giving the meaning of the statutory term ("unjustifiable risk") in the situation--so it is a "fact" in my understanding of things--and then might find the defendant's unawareness of that fact culpably "unreasonable." The parent's mistake would therefore be an "unreasonable MF," the very thing that I have called a contradiction in terms. But I am not sure that a factfinder could some instances of ignorance or misuse of statutory language, cannot be deemed "reasonable" enough to exculpate, as they sometimes clearly are 162 and as MFs usually are. 163 Reasonable mistake, thus, may be ML or MF, because a finding of reasonableness under the statute implies that, whatever inferences the defendant failed to make, those failures were insufficient to sustain a conviction, whether the inferences were entailed by the statutory term (a ML) or not (a MF).
When is this analysis pertinent? There are many instances in which criminal law makes reasonableness material. For example, in most jurisdictions, a rape defendant who claims he thought the victim had consented to sex has no defense unless the find such a parent unreasonably unaware. After all, the very basis of culpability here must be the parent's awareness at some crucial point of the necessary facts. If this is a clear case of criminally negligent behavior (and I'm not so sure everyone would say it was), that is presumably because it seems clear that such a parent's leaving an infant in a running tub under such circumstances is the very definition of a "substantial and unjustifiable risk" of death. And, if you prefer (arbitrarily, I think) to focus on the moment a few minutes later when the danger is at its height and he has completely forgotten about the child, there is still a real sense in which he "knows" that his child is in the running tub. I see no decisive reason why such "latent" knowledge cannot be counted among the known facts that can entail an inference to a statutory term. (A nice discussion of "latent" knowledge in a slightly different context is in R. A. Duff, Caldwell and Lawrence: The Retreat from Subjectivism, 3 Oxford J. of Legal Stud. 77, 88-93 (1983).) If at any point in the sequence the parent has failed to draw the inference of "unjustifiable risk" from the facts the world has presented to him, then he can be said to have failed to see the definition of that statutory term and made a ML, not a MF. Of course, I do not pretend to have dealt with this objection in all of its potential complexity, and I will not do so in this article, but I hope I have said enough to justify my going forward with the positive argument in the text and leaving for another day all the complexities of the meaning of (un)reasonableness. 162 Although rarely applied, there is a constitutional rule in the United States that requires exculpation for ML where the ML was especially hard to avoid. See Lambert v. California, 355 U.S. 225, 229 (1958). And many American jurisdictions provide for ML defenses where the defendant has made sufficiently reasonable efforts to comply with the law. mistake was reasonable. 164 Similarly, a negligent-homicide defendant must have been reasonable in thinking, for example, that her gun was unloaded; that is, that any risk to life she created was not "substantial and unjustifiable." 165 And, in most jurisdictions, a murder defendant who claims self-defense must have been reasonable in thinking, for example, that her victim had been attempting to kill her. 166
A. Why an "Unreasonable Mistake of Fact" is Really a Mistake of Law
Let me begin with an example from rape law. A particular set of facts might, depending on the state of the law, leave open inferences of either consent or nonconsent to sexual intercourse. For example, to adapt the case of MTS, 167 suppose that a woman has invited a man into bed, both of them undressed. They engage in kissing and other sexual touching, but the woman believes she has indicated that intercourse is out of the question. The man genuinely thinks otherwise and penetrates her; she immediately pushes him off and tells him to leave. Let me suppose that the precise details of this encounter, as known to the man, are consistent with either consent or nonconsent in the eyes of the law (until she pushes him off, anyway). That is, the facts do not mean nonconsent, but neither are they enough to preclude a finding of nonconsent should supplementary facts (beyond those available to the defendant) be introduced. A factfinder might accept that the defendant had believed that the woman had "consented" to intercourse. I am not discussing credibility here but statutory meaning.
Thus, if a completely credible defendant said that he took all the voluntary sexual activity to indicate affirmative "consent" to intercourse, a jury would have little reason to think that he misunderstood the meaning of "consent" and misused the term but only that he was unaware of the limits that the woman had meant to impose, facts that would help to constitute her "consent" or "nonconsent" in the situation. The defendant's mistake, therefore, would be a MF, since it would not be a mistake about the meaning of the relevant statutory term, about the inferences entailed by that term, but only about aspects of the situation independent of the meaning of the term.
None of the reasoning in the previous paragraph depended on the existence of a reasonableness requirement. I did not need a reasonableness requirement in order to conclude that the defendant's mistake was a MF. I only needed a way of knowing the objective limits of the meaning of "consent" in given circumstances and thus a way of determining whether the defendant's mistake with respect to that term constituted a misunderstanding of its meaning (ML) or only an unawareness of certain circumstances that were necessary to instantiate the term's meaning in that scenario (MF). If I then suppose that this jurisdiction, like most, does require reasonable mistake for exculpation, then the very same reasoning as in the previous paragraph suggests that the defendant's MF would count as reasonable. The man's mistaken inference of consent reflects no misuse of the language in those particular circumstances and so must be deemed at least as reasonable as another man's inference of nonconsent would be.
Aware of nothing in the situation that could constitute "nonconsent"--entail an inference to "nonconsent"--within the meaning of the law, his failure to infer the woman's nonconsent must be reasonable, and, for the very same reasons, it must be a MF. Reasonableness requirement or no reasonableness requirement, the judge would take the defendant's claim to be that the circumstances as he understood them did not entail an inference to nonconsent because the woman's invitations to sexual contact should be understood to have expressed an invitation to intercourse. If the court has no grounds on which to find the inference from his experience to "consent" inconsistent with the statutory meaning of "consent," then the mistake cannot be a ML but must be a MF--and, for the very same reasons, must be reasonable.
If the reasoning that is necessary to conclude that a mistake is a MF will always imply that the mistake was "reasonable," then every MF is "reasonable," and the idea of "unreasonable MF" is a contradiction in terms. Let me explicate this conclusion by taking an example of "unreasonable MF" from the facts of the famous Morgan 168 case, still supposing (although contrary to the actual holding in Morgan) that a reasonableness requirement is in effect for mistakes regarding consent. Several defendants had intercourse with the victim at the instigation of the victim's husband, who had assured them that his wife's struggling against them would only reflect her enjoyment of the sex. They went through with the act even though the victim's resistance and calls for help were in fact quite violent. Given such circumstances, the law would certainly require an inference of nonconsent, and I have never encountered any expression of doubt that the defendant's claimed mistake as to consent was anything but grossly unreasonable. 169 A non-nullifying jury with no other facts at its disposal would have had to find the element of nonconsent satisfied, and a defendant's mistaken inference from those very facts to a belief in consent would constitute a ML because based on the legally erroneous premise that a woman who is violently resisting can be taken to be consenting just because her husband has said that she likes it that way. Such a collection of evidence cannot signify consent. Such an experience cannot mean consent. 170 Rather, it must be taken by any court to constitute nonconsent and must require any person in the position of the Morgan defendants or their jury to draw the inference of nonconsent.
The possibility that a jury might also have other facts, unknown to the defendants at the time of their conduct, with which to reinforce the finding of nonconsent would not change the fact that the defendants' actual knowledge itself included enough to constitute "nonconsent," to require a finding of nonconsent by defendants and jury alike. Nor would further facts, also unknown to the Ds, that showed that the woman really was consenting, after all, change the fact that the defendants made a ML by indulging in an illegitimate inference, by misusing the word "consent" when they claimed on these facts that they thought she consented. The defendants would not be convicted of rape, of course, because the element of nonconsent would, in the end, be missing, but they would still have made a ML, just as surely as someone who, mistakenly believing a gun to be loaded, would make a ML in concluding that it would not be "murder" to point the gun at a sleeping stranger and shoot to kill. 171 In effect, the law would have defined nonconsent in terms of characteristics that the defendants knew were present in the scenario they encountered. It would have said that the only attitude of mind that could have been inferred from the victim's conduct and other relevant circumstances was "nonconsent," whereas any "consent" the defendants inferred must have been something other than what the law recognizes as "consent," something for which they simply misappropriated the term. 172 If the defendants' mistake was, therefore, a mistake as to the meaning of "consent," then the mistake was a ML.
That the Morgan defendants' claimed mistake did, in fact, constitute a ML is shown, I think, by the actual, peculiar disposition of the case. While the House of Lords held that the trial judge had misinstructed the jury by attaching a reasonableness 171 In this latter case, the defendant could be prosecuted for attempted murder, since her only defense is ML in thinking that it's okay to shoot an innocent person. Similarly, in the former case, there should be no obstacle to prosecution for attempted rape, since there was no exculpatory MF regarding nonconsent. 172 I express no opinion whether it can be said in some discourse or other that active resistance might consist with "consent" to sex, but I feel safe enough in saying that most courts today would find that persistent active resistance precludes any inference of consent unless there is really sufficiently strong evidence that the "resistance" was just a kind of play-acting, i.e., not resistance at all. In Morgan, any such evidence (the husband's assurances were the sum of it) was wholly requirement to the mistake defense, it nevertheless upheld the convictions on the statutory grounds "that no miscarriage of justice has actually occurred." 173 Since the trial jury had already rejected the defendants' evidence of reasonable mistake, the defendants' lawyers had to argue that, if the jury had been properly instructed, it might have at least found an unreasonable mistake. They argued that the jury could have concluded that the defendants thought the victim was consenting even as she resisted, since the husband had told them that her resistance would only signify desire. 174 All members of the Morgan tribunal considered such an argument too preposterous to require detailed refutation, and they all agreed that no reasonable jury could have accepted any such argument because the argument was, in the end, simply "not conceivable." 175 Why? Because "consent" cannot mean anything a lawyer wants it to mean, not even if the lawyer's client was the very model of sincerity in using the word that way. And the most that the lawyers could make out of the evidence was not enough to come within the meaning of "consent." Violent resistance simply could not consist with consent as the judges understood that term. Thus, although the Lords never put it precisely this way, the defendants' mistake claim was rejected because it was a misuse of the statutory language and thus a ML, not the unreasonable MF that they claimed it was and that the Lords' main holding purported to accommodate.
The argument that the mistake was not ML would have to go as follows, I suppose: once the defendants embraced the belief that the victim expressed a desire for intercourse, that belief obviously and reasonably entitled them to the inference that she had consented. They may have been mistaken that the victim actually desired intercourse, but that just meant that they were mistaken about what she wanted, not that they misunderstood the meaning of consent. Such an argument depends, however, on the capacity of the legal term "consent" to accommodate the idea that violent resistance could somehow indicate desire. If "consent" really could accommodate such an idea, then the defendants' mistake would indeed turn out to be a MF. But to suppose such an expansive definition is not far from supposing a definition of red in the traffic laws that could encompass green as well. Since we know the underlying beliefs on which the belief "she expressed a desire for intercourse" is based--i.e., her struggling and calls for help and the husband's bald assurances--and since those underlying facts as known by the defendants were unambiguous evidence of the reality that the law is trying to get at with the requirement of "nonconsent," they legally entail an inference contrary to the defendants'. It was "not conceivable" to the Lords that a properly instructed jury could have accepted the claim argued by the defendants on appeal because the factual scenario they offered simply could not mean consent as the word is used in rape law. Thus, although the Morgan defendants have usually been assumed to have made a claim of unreasonable MF, 176 I think the proper characterization is that they made a claim of ML, because the gist of their claim was that they did not understand what the law meant by "consent." 177 Part of the point here is to say that any "unreasonable" mistake is necessarily a ML, but it is easy to slip from that claim into the erroneous supposition that it is the reasonableness standard itself, the reasonable-unreasonable line, that determines what mistakes are MF and ML. It is easy to suppose that the reasonableness standard is what justifies a court's going behind the defendant's honest statement of belief that, for example, the victim expressed a desire for intercourse. After all, where a reasonableness standard is in place, courts do generally resort to the language of that standard rather than the language of ML and MF. They simply say, for example, that the Morgan defendants' belief that the victim expressed a desire for intercourse was unreasonable in the circumstances, and so they were not entitled to hold or proffer that belief. But why did the circumstances make a belief in "consent" unreasonable? Because the circumstances that the Morgan defendants encountered simply do not mean consent.
Calling their inference unreasonable is no advance on saying that they misunderstood the meaning of the term "consent" in all the circumstances. And no reasonableness requirement is required to equip a factfinder to reach that conclusion; the language of reasonableness may prove convenient, but it is in no way necessary to determining whether the defendants' mistake was one of meaning or of experience. What is necessary is simply the factfinder's determination of whether the defendant's experience was or was not sufficient to instantiate the term. Thus in Morgan the Lords did ultimately abjure any reasonableness requirement but still found that the defendants' argument was "not conceivable," that it was, in effect, simply a misuse of the statutory term, a ML, apparently because the circumstances of which the defendants claimed to be aware necessarily instantiated "nonconsent" in the eyes of the court. sections of this article, I have found no such theory in the literature and have been unable to In sum, it is true that an actual finding that a mistake was "unreasonable" is sufficient to constitute a finding of ML, since a mistake that is not a ML--one that involves no misuse of statutory language 178 in describing the scenario as the defendant experienced it--must be reasonable. But a finding of unreasonableness is not necessary to a finding of ML. That is obviously true in cases where there is no reasonableness requirement and so no opportunity for a formal finding of unreasonableness. It is equally true when a reasonableness requirement is in place, since even a ML might be deemed reasonable, in the sense of exculpatory, in relatively rare cases. The basic limit on what beliefs a defendant can proffer as "factual" or MF is not the rape statute's reasonableness requirement as such but the bare fact that a statutory requirement like "nonconsent" must have an objectively limited scope of meaning. A defendant cannot claim MF just by distorting the meanings of words; he cannot get a MF defense by asserting a belief that resistance means consent any more than by asserting a belief that red can mean green. In either case, the defendant shows only that he misunderstood the meanings of material terms in a particular situation, not that he mistook the real nature of the situation. To misunderstand the meaning of a material term is to make a ML, and one does not need a reasonableness requirement to authorize the court to question the defendant's claim of "MF" when that claim constitutes a misuse of, misapplication of, or misinference with respect to that term.
B. What Does a Reasonableness Requirement Do?
construct one myself. 178 Note that this phrase, "statutory language," really is a redundancy here, since one of the important points of this article is that all language in a criminal adjudication is statutory languge. It is all either the words of the statute, or elaborations of the words of the statute, or else it is So, if a reasonableness requirement is unnecessary to defeat the "MF" claims made by such as the Morgan Ds, what is the effect of requiring that a mistake be reasonable? It is not, as is often supposed, that it inserts an objective standard into the MF inquiry where there would otherwise be none. 179 Even without a reasonableness standard the defendant is normally held to uses of terms that are within some objectively, legally defined meanings of those terms. Nor is a reasonableness requirement simply a stand-in for the ML-MF distinction, for the reasons offered above.
The reasonable-unreasonable line is independent of the ML-MF line, and the more general effect of the reasonableness standard is simply to limit the scope of MLs that will be deemed exculpatory (under the guise of MF). The reasonableness requirement neither separates one kind of MF from another (as is usually supposed) nor separates ML as such from MF but separates all MFs and some (excusable) MLs from the remaining (non-excusable) MLs. The ML-MF line is determined by the factfinder's essentially semantic analysis of the terms of a statute on the facts of the case, but the reasonable-unreasonable line is determined by analyzing culpability, an analysis that can and sometimes does cut across the ML-MF line.
It may seem that, in using Morgan as an example above, I have been making my argument on the basis of an extreme case, one in which the defendants' claims were just wholly implausible. So let me illustrate the function of the reasonableness requirement irrelevant to any substantive question of liability and thus not really in the substantive adjudication at all. 179 See, e.g., Singer, Resurgence, supra note ---, at 460; Celia Wells, Swatting the Subjectivist Bug, 1982 CRIM. L. REV. 209, 210-211 (1982) (stating that "[w]here mistakes are required to be reasonable . . . an element of . . . objectivism is retained"); ASHWORTH, supra note ---, at 341 (arguing for a negligence standard in rape and acknowledging that that would be a "a departure from the general subjective principle of criminal liability"); David Cowley, The Retreat from Morgan, 1982 CRIM. L. REV. 198, 198, 208 (1982) (contrasting Morgan's raising of "the banner of subjectivism" with later cases that "herald[] an alarming and unwelcome return to an era of objective criminal liability").
here and at the same time reinforce the argument above with an example in which the defendant has a somewhat more intuitively plausible claim of consent than did the Morgan defendants. Suppose that a man with a history of violence towards his girlfriend reacts to her rejection of him on one occasion by threatening violence if she does not agree to intercourse. She calms him down and the immediate crisis passes.
But just a little later in the evening he again presses her to agree to intercourse, this time without any overt threats. If she agrees this time--and let me suppose that there is no other evidence available--the law could well conclude that this body of evidence alone necessarily proves nonconsent. Then, if the defendant thought or thinks that this body of evidence does not constitute nonconsent, he has made a ML. The law has defined consent relatively narrowly, and the defendant has failed to draw the inferences entailed by that definition.
But, in such a case, courts are likely to treat the mistaken belief in "consent" as MF, perhaps an "unreasonable MF" but still a MF. After all, if the mistake in Morgan has generally been labeled a MF, this one would seem no less "factual." And courts conventionally would apply the reasonableness standard to this "MF" to determine whether it was a reasonable MF or an "unreasonable MF." If I am right that it is actually ML, however, then the difference made by a reasonableness requirement is not to divide the set of MFs into culpable and non-culpable mistakes but, in fact, to divide the set of MLs into culpable and non-culpable. Or, to put it another way, the effect of inserting a reasonableness requirement where the standard would otherwise be honest mistake is to reduce the scope of legally mistaken inferences (MLs) that the courts will indulge in the name of MF.
Thus, an honest-mistake standard (no reasonableness requirement) would very likely indulge the boyfriend's usage of "consent," even though it is contrary to the legal meaning of the term, since the standard would presumably exculpate any legally mistaken but "conceivable" usage of the term--any usage more intelligible than, say, that of the Morgan defendants. A reasonable-mistake standard, on the other hand, would be somewhat less likely to accommodate the boyfriend's usage. Whether the factfinder would indulge the defendant's usage as reasonable--exculpatory--would not necessarily depend on whether the defendant's usage was consistent with the law's usage, since misunderstanding legal meaning is not necessarily culpable (although getting legal meaning right is necessarily non-culpable). The factfinder's decision regarding liability might depend, instead, on whether the defendant's usage was sufficiently plausible as colloquial usage or sufficiently close to legal usage to merit indulgence as "reasonable" or exculpatory.
Just to keep the picture complete, let me rehearse quickly what a MF would look like in this scenario. Suppose, in contrast to my hypothetical above, the factfinder has evidence that proves the nonconsent element when the defendant did not have such evidence. For example, here the factfinder might decide that the evidence in the defendant's awareness--threats followed by a period of calm before the woman agreed to intercourse--does not quite constitute nonconsent but that other evidence of the victim's fear does. Then the defendant's failure to draw the inference of nonconsent (from the same body of evidence that left the factfinder similarly short of finding nonconsent) cannot reflect a failure to understand statutory meaning. Rather, it must be a MF, and, since a MF reflects no failures of inference in the eyes of the law, it must be deemed reasonable whenever a reasonableness requirement is in place. 180 The effect of a reasonableness requirement with respect to MF, then, is only to provide an occasion to slap the exculpatory lable of "reasonable" on a mistake that would have been exculpatory anyway (as long as the relevant standard is not strict liability). In the absence of a reasonableness requirement, any MF is exculpatory, and in the presence of a reasonableness requirement any MF is still exculpatory, because every MF must be reasonable. Now let me run Morgan through this structure. My discussion of the Morgan scenario in the previous subsection proceeded from the supposition that a reasonableness requirement was in place, but, in actuality, the holding of Morgan was a rejection of any reasonableness requirement with respect to consent. It has since been characterized, therefore, as a decision that chose a subjective standard of MF over an objective standard of MF. 181 I have made clear, however, that I think it actually was a case of ML rather than MF. The defendants were aware of circumstances that legally constituted nonconsent, 182 and so the question really was whether any honest mistake as to the meaning of consent or only a reasonable mistake could exculpate them. Either way, the claim was ML, since it was a mistake of meaning.
Had the Lords read a reasonableness requirement into the statute, they would have been saying that only a limited range of mistakes would be allowed to exculpate: all MFs and perhaps a limited range of MLs. When, instead, they refused to read a 180 In this structure, the only options for MF standards are strict liability versus complete indulgence for MF. If a MF can exculpate at all, then every MF can exculpate for the relevant element since every MF is fully reasonable. A MF can fail to exculpate only if strict liability applies to the relevant element. 181 See supra note ---(citing Singer, Dressler, Ashworth to this effect). reasonableness requirement into the statute, they seemed to say that any mistake at all regarding consent would exculpate, since they rejected the only handy standard for limiting such a defense. Of course, the Lords would have said that the defense was limited to MFs, but I have already argued that their willingness to contemplate defenses based on unreasonable mistakes meant that they were necessarily opening the door to ML defenses. Without a reasonableness limitation, defendants would seem to have an unbounded ML defense. Nevertheless, in applying their non-standard, the Lords unanimously deemed the defendants' claimed mistake not just non-credible but "not conceivable." They treated the defendants' usage of "consent" as if the defendants were speaking a foreign language. The Lords thereby implicitly recognized that consent, like any term, must have an objective, legal meaning, that any misuse or ignorance of that term is not MF (which would have excused) but ML, and that a sufficiently gross misuse of the term may be culpable even when more plausible MLs are not. A reasonableness requirement would have been a signal to factfinders to be relatively intolerant towards those who misunderstand law--not facts--and would have given them a standard (however vague) to apply in that task, perhaps even limiting the mistake defense strictly to MF. Elimination of the reasonableness requirement, on the other hand, was a signal to be relatively indulgent, excusing misuses of statutory terms just as long as the misuse was, perhaps, "conceivable" but refusing to excuse utter ignorance of the existence or meanings of a term. Only a full ML defense could have saved the Morgan defendants by truly abandoning all objective standards and effectively making the law in any case whatever the defendants had thought it was at the time of their conduct. 183 All of this should suggest that the question of what standard to apply to a claim that amounts to ML raises some knotty questions, but the main point of this exercise is to show that the effect of applying a reasonableness standard is to divide the category of MLs into the exculpatory and the non-exculpatory, not to divide the category of MFs at all.
There are, then, two cases, those in which the mistaken defendant was aware of evidence that constitutes the relevant statutory element (ML) and those where he was not (MF). When a reasonableness requirement applies to the mistake, all cases of the latter sort must be non-culpable while the former category will be divided by that requirement into culpable and non-culpable mistakes. In a case of the first type, any mistake must be a ML. The defendant knows everything except the proper legal label to put on the situation. His ML is not necessarily unreasonable, however, because not necessarily culpable; it may yet be deemed reasonable (or unreasonable) if the statute leaves open a defense for reasonable MLs. If the case is of the second type, where the defendant lacks the necessary evidence from which to infer satisfaction of the element, then the reasonableness requirement is essentially superfluous--all genuine mistakes of that sort are necessarily reasonable and are necessarily MFs. The failure to know the situation reflected no material failures of inference because the defendant simply did not have information that entailed those inferences; that failure must, therefore, be deemed reasonable. A finding that the defendant's mistake was unreasonable would necessarily represent a conclusion that the defendant could at least theoretically have evaluated the situation more accurately--thus that the case was of the first type, in which all of the information necessary to a proper legal evaluation was known, and that the failure to so evaluate was culpable. But, if the defendant's mistake is found reasonable, that does not necessarily mean that the mistake is a MF. Where the defendant is aware of the evidence that constitutes the element, any mistake is ML but may yet be deemed reasonable if such a ML is thought non-culpable under that statute.
The reasonableness requirement for "MF" simply limits, diffusely, the scope of MLs-not MFs at all--that will be permitted to exculpate.
C. Some Homicide Examples and the Unacknowledged Erosion of the Mistake of Law Rule
The mistake structure elaborated in the rape examples of the previous subsections can be fruitfully applied to any area of criminal law, including the law of homicide and self-defense. And sometimes that application reveals surprising, unacknowledged erosions of the ML maxim in modern law.
For example, in MPC jurisdictions, negligent homicide and reckless manslaughter share the common element that the defendant must have taken a substantial and unjustifiable risk of causing the death of another. The only difference between the two offenses is that manslaughter requires conscious awareness of the risk, while negligent homicide does not. 184 If the defendant is acquitted of manslaughter but convicted of negligent homicide, then the implication is that the defendant mistook the risk of death under circumstances where such mistake was unreasonable. Maybe she knew that a gun had been loaded earlier in the day but was told by her ten-year-old son that he had unloaded it. She then assumed that there was no risk of death when, without checking for bullets herself, she pointed the gun at passing traffic and pulled the trigger. She made a ML in failing to infer from known facts that aiming at traffic and pulling the trigger constituted an "unjustifiable risk." Such a mistake would nevertheless be a good defense to manslaughter, although not to negligent homicide. 185 That mistake's capacity to defeat a charge of manslaughter shows that the MPC's definition of manslaughter incorporates a requirement of knowledge of illegality and thus a ML defense. 186 The MPC suggests at least one other ML defense to a homicide charge, as well, when it adopts the doctrine of "imperfect justification." It does not label that doctrine a ML defense. Nor does it acknowledge that it might be departing thereby from the basic ML rule (which the MPC itself reaffirms in its section 2.02(9)). Here is the doctrine: When charged with murder, a defendant can seek acquittal by arguing that she killed to defend herself from a threat of deadly force (among other threats). Traditionally, if the (c), (d) (West 1999) (defining culpable mental states of recklessness and negligence); TEX. PENAL CODE § § 19.04, 19.05 (West 1999) (defining manslaughter and negligent homicide). 185 For an example from a real case, take Arizona v. Olsen, 157 Ariz. 603 (1988). There, the defendant appears to have been fully aware that he was across from a motel swimming pool when he shot at a pair of fleeing felons. He missed the fugitives but hit one of the motel patrons, killing her. His only mistake, apparently, was his failure to infer that the risk he took in firing his gun was "unjustifiable" (that is, a gross deviation from the standard of care) under the circumstances. But that mistake on the legal question of which risks are "unjustifiable" and which are not was enough to prevent a manslaughter conviction (or perhaps even a manslaughter charge) and limit his liability to the level of negligent homicide. 186 One objection to this reasoning might be that failure to perceive a risk can sometimes rest not on failure to infer properly from known facts but from failure to keep in mind--to continue to know--facts that would have made the risk clear. I've discussed this briefly above at note ---defendant's belief in the existence of such a threat was unreasonably mistaken, then the defense disappeared altogether, and she would likely be convicted of murder since she intentionally caused the death of another without lawful justification. 187 The Model Penal Code (MPC) and some state courts, however, have viewed this result as inconsistent with basic principles of culpability. 188 The view of the MPC is that an honest but unreasonable--that is, negligent--belief in the existence of justifying circumstances renders the killing merely negligent homicide rather than murder.
To illustrate the distinction, I want to take the case of Kansas v. Simon. 189 There, an elderly white man claimed to have had a fear of his young Asian neighbor, rooted in a belief that the neighbor must be a dangerous martial artist simply because he was of Asian extraction. As a result, he responded to an essentially non-threatening situation by firing a gun at the neighbor. Simon was both wrong and wildly unreasonable to think that he faced any threat at all. Under the traditional rule, therefore, if the victim had died, he should have been convicted of murder. Nevertheless, if that belief was sincere, as the actual jury in Simon seems to have concluded, 190 then the MPC would consider Simon innocent of murder and guilty of no offense worse than negligent homicide.
The drafters of the MPC consider this a common sense result, 191 but of course it opens the door to a ML defense to the charge of murder since even an unreasonable inference can exculpate. Simon's claim, after all, would have to be that, even though the law requires an inference from the facts he knew to the conclusion that his victim was not going to attack him, he failed to make that inference and that failure provides a good defense to the murder charge. We can be confident that the law entails that inference--defines "imminent use of unlawful force" 192 to exclude the collection of circumstances known to Simon at the time--for at least two reasons. For one thing, we know that the law of negligent homicide would reach him, by way of its requirement that the defendant have been reasonable in determining that he faced an "imminent use of unlawful force." Since he would certainly have been convicted of negligent homicide and thereby found to have made "unreasonable" inferences with respect to that element, his mistake was ML. Imperfect justification doctrine thus provides a ML defense to murder.
Note further, however, that, as argued previously, the reasonableness requirement is unnecessary to any of this anlaysis. The law of self-defense defines "imminent use of unlawful force" through the factfinder's giving meaning to that phrase in cases. Even without the reasonableness requirement, therefore, the law would 191 The Commentaries to the MPC suggest that it makes no sense to convict someone for a crime of intention where the individual "labored under a mistake that, had the facts been as he supposed, would have left him free from guilt." Where the crime requires culpability greater than negligence, "it is neither fair nor logical to convict when there is only negligence as to the circumstances that would establish a justification." Thus the Code considers purposeful killings not to be crimes of purpose or knowledge at all, even in the absence of justification, as long as they are "imperfectly" justified; rather, they are crimes of recklessness or negligence. See MODEL PENAL CODE AND COMMENTARIES § 3.09 comment 2 (1985). Mark Kelman argues that this justification is actually entirely arbitrary and that it makes just as much sense to emphasize the intention to kill (under bad circumstances) as it does to emphasize the mere negligence in the defendant's evaluation of the circumstances. Kelman, supra note ---, at 616-618. 192 K.S.A. § 21-3211. certainly determine that Simon's experience entailed an inference that there was no "imminent use of unlawful force." His inference to the contrary, therefore, was a ML.
D. Further Ambiguities of Criminal Intent
These ML defenses to murder and manslaughter constitute, for better or worse, modern, unacknowledged erosions of the rule against ML defenses. In that sense, they are akin to the phenomenon discussed in Part ---above: the modern elevation of the claim of right defense from ad hoc ML defense to the supposedly general principle that mistakes of "different law" somehow "function as" MF. These homicide doctrines, then, like mistake of "different law" and "unreasonable MF," are not unproblematic manifestations of foundational principles of criminal law. They are significant departures from at least one important, continuing, and perhaps dominant approach to criminal law. That approach emphasizes the distinction between MF and ML as central to maintaining the objectivity of the law even at the expense of what might appear to be individualized justice. These homicide doctrines, on the other hand, show some of the ways in which that approach is in constant but often hidden tension with another equally important approach, in which the distinction between ML and MF seems unimportant or irrelevant to the assertedly more fundamental question of individual moral guilt.
As noted above, the doctrine of imperfect justification is a relative novelty.
Traditionally, the absence of reasonable belief in circumstances that constituted justification under the law eliminated the defense completely. One might suppose that that was because, without much analysis, the courts recognized the defense for the ML defense that it was. I'm inclined to think that that is so, although I am not aware of any court putting it that way. But, in any case, the contrast between the traditional rule and the new doctrine helps to illustrate the tensions in the use of mental-state doctrines.
On the one hand, courts in self-defense cases could always have required the sort of "criminal intent" that in many traditionalist cases incorporated a ML defense. 193 That is, they could have sided with those who emphasize intentional moral wrongdoing as an essential prerequisite to criminal liability and thus required acquittal where the claimed self-defender genuinely thought she was legally justified. Simon, for example, would deserve an acquittal since he really thought his circumstances threatened him with an "imminent use of unlawful force." And a logical extension of the argument would give him an acquittal even if he just thought killing was justifiable to prevent a relatively petty wrong, like an act of vandalism on his car. As the modern popularity of the doctrine of imperfect justification suggests, the use of a version of criminal intent that smuggles a limited ML defense into self-defense law might seem highly appropriate, since, even though the killing is intentional, the frame of the typical defendant's mind is arguably non-criminal or at least less culpable or "corrupt" than that of most other intentional killers. 194 We may not approve of killing to prevent mere vandalism, but one who kills in the genuine belief that he is legally justified can certainly be said to deserve acquittal, since he chose to do only what he understood the law positively to endorse.
On the other hand, however, the traditional, harsh approach to the imperfectjustification scenario is consistent with a Holmesian approach to criminal doctrine that survives in many respects to this day. This Holmesian tradition rejects the MPC's fine distinctions among levels of unlawful homicide on the basis of measures of subjective awareness. Instead it prefers a grading structure that focuses on measures of objective risk inferrable from known facts. Thus Holmes himself, while agreeing that conscious awareness of a risk of death was enough to justify criminal liability should the death eventuate, said that the common law's more basic "test of murder is the degree of danger to life attending the act under the known circumstances of the case." 195 The law does not, according to Holmes, "inquire whether [the accused] did actually foresee the consequences or not" but only whether "a man of reasonable prudence would have foreseen" them. 196 Thus, while manslaughter comes in more than one variety, "[o]ne great difference" between murder and manslaughter "will be found to lie in the degree of danger attaching to the act in the given state of facts," not necessarily in the accused's having actually been aware or unaware of the risk of death. 197 Such an approach to homicide tends to eliminate any ML defense by refusing to indulge any unreasonable inferences in any scenario for any purpose. If the known facts entail an inference of unjustifiable risk, then the result is at least manslaughter. But the result could even be murder if the facts known to the defendant entail an inference of a quite high risk of death, even if no such risk ever crossed the mind of the defendant.
Moreover, as I have suggested, this approach to criminal homicide has hardly disappeared in the generations since Holmes wrote. In Massachusetts, for example, the grading structure for homicide elevates an unlawful killing to murder not just when the defendant intended to kill but whenever a reasonable person in the situation would have been aware of a "plain and strong likelihood of death." 198 Massachusetts affirms that this standard is "purely objective" 199 and that it calls simply for the factfinder to measure the objective quantity of risk created by the known circumstances. 200 Similarly, in the famous English case of D.P.P. v. Smith, 201 the defendant's murder conviction was affirmed on appeal even without a finding of any conscious awareness of risk of death.
Smith knew that a policeman was hanging on the side of his car as he put his foot on the gas and fled along a busy two-way street. The policeman lost his grip and was killed by an oncoming car. Obvious as the danger of death would have been to a reasonable person, Smith claimed that in his panic he just never thought about the danger involved and therefore lacked the "knowledge" supposedly required for a murder conviction. 202 The court held that the test of murder was not whether the accused had actually known that death would result from his actions but whether a reasonable person or "ordinary responsible man would, in all the circumstances of the case, have contemplated [death] as the natural and probable result." 203 And the court then concluded that, even if Smith's testimony were believed, the facts were such as to justify an imputation of knowledge to Smith under this test. 204 If the known facts are such that any reasonable person would have inferred a high probability that death would occur, then in this tradition the necessary "knowledge" is proven. To hold other than these cases have held would be to indulge a ML defense, to say that even though a defendant knew everything necessary to entail an inference of high risk of death we will let him defend by saying he did not understand the meaning of the situation or the meaning of the statutory term in that situation.
An even more common instance of this objective approach to homicide law lies in the doctrine of depraved-indifference murder, which is in effect in virtually every jurisdiction. Convicting someone of murder on the basis of conduct that evinces a depraved indifference to the value of human life could rest on a theory that that phrase expresses a certain kind of conscious awareness of practical certainty of death--e.g., that the accused did not know or care exactly who or how he would kill but was practically certain that his actions would kill someone, somehow. But this doctrine might, alternatively, rest on a theory that, even when an accused lacked the requisite awareness that death was the certain result of his conduct, still his conduct might occur under circumstances that would have indicated the high likelihood of death so clearly to an objective observer that the defendant must be held responsible for murder rather than manslaughter or less. 205 At least in some states, this latter theory holds.
In New York, for example, when death is caused "recklessly" the accused is guilty of manslaughter, 206 but the offense is elevated to murder when the "recklessly" caused death occurs "[u]nder circumstances evincing a depraved indifference to human life." 207 In New York v. Register, 208 the Court of Appeals held that that the statutory language does not modify the mens rea of recklessness. Rather, it simply contemplates that some sets of circumstances and conduct will objectively evidence such a high level of risk as to justify a conviction of murder for one who does that conduct in those circumstances. "[T]he Legislature structured the degree of risk which must be present in nonintentional killings by providing that in a depraved mind murder the actor's conduct must present a grave risk of death whereas in manslaughter it presents the lesser substantial risk of death." 209 The accused need not have known of the level of risk.
Rather, the court held "that depraved mind murder is distinguishable from manslaughter, not by the mental element involved but by the objective circumstances in which the act occurs." 210 "[T]he focus of the offense is not upon the subjective intent of the defendant, as it is with intentional murder . . ., but rather upon an objective assessment of the degree of risk presented by defendant's reckless conduct . . ." 211 Formally, the defendant need only have known of the substantial and unjustifiable risk of death that would convict him of manslaughter, and the difference between manslaughter and murder then became, as Holmes would have it, nothing more or less than the objectively measured "degree of danger attaching to the act in the given state of facts." 212 209 Id. at 276. 210 Id. at 278. 211 Id. at 277. 212 HOLMES, supra note ---, at 59. See also New York v. Word, 689 N.Y.S.2d 36, 37, 260 A. 2d 196 (1999), on reconsideration, 94 N.Y.2d 799 (holding that whether a particular act was committed under "circumstances evincing a depraved indifference to human life" is not a mens rea element which focuses upon the subjective intent of the defendant but an objective assessment of the degree of risk presented by defendant's reckless conduct); Maine v. Dodd, 503 A.2d 1302, 1305 (Me. 1986) (finding that "depraved indifference" murder requires conduct which objectively viewed by a reasonable person manifests a depraved indifference to the value of human life) ( that depravedmind murder involves an "unintended death" caused by an evil act that a reasonable person would know would kill). Some jurisdictions are less clear about the applicable standard but may take the objective approach as well. See South Dakota v. Laible, 594 N.W.2d 328, 333 (S.D. 1999) (holding that "whether conduct is imminently dangerous to others and evincing a depraved mind regardless of human life is to be determined from the conduct itself and the circumstances of its The facts known by the defendant in these categories of cases are understood by the law in both traditions to entail an inference of an objectively unjustifiable risk of death. A defendant's claim that she failed to make that inference would constitute a claim of ML, which is not normally thought available for most offenses and certainly not for murder or manslaughter. Even in justification cases under the MPC, ML is not indulged where the mistake is a mistake as to the formal categories of justification even though it is indulged where the mistake relates to whether that category has been satisfied in the circumstances. That is, even though modern jurisdictions disagree on questions such as whether a threat to rob is a good justification for deadly force in selfdefense, 213 the ML maxim holds defendants strictly to a knowledge of the local law on that point. 214 Meanwhile the imperfect justification doctrine excuses them for failure to commission"). See generally LAFAVE & SCOTT, supra note ---, § 7.4 (discussing a split among jurisdictions as to whether subjective awareness is required); DRESSLER, supra note ---, § 31.05(b). Note further that even the level of conscious awareness supposedly mandated by the requirement of recklessness went out the window in Register itself, since New York follows the MPC in allowing recklessness to be inferred even where it might well be absent because of intoxication. See N.Y. PENAL LAW § 15.05 (McKinney 1999). Once Register's trial judge had concluded that the formal mens rea for depraved-indifference murder was only recklessness, Register's jury was not permitted to consider his intoxication as evidence that he lacked the necessary mens rea for murder, since the relevant New York statute precludes evidence of intoxication to negate recklessness. See Register, 60 N.Y.2d at 275. The Court of Appeals endorsed this result. See id. at 279-281. The result, then, was that Register's jury was permitted to convict him of murder even if it concluded that he had lacked conscious awareness of any risk of death. 213 Compare, e.g., N.Y. PENAL LAW § 35.15(2)(b) (McKinney 1999) (allowing use of deadly force to defend against a robbery), with 18 PA. CONS. STAT. ANN. § 505(b)(2) (West 1999) (deadly force justifiable only in response to threat of death, serious bodily injury, kidnapping or sexual intercourse compelled by force or threat). 214 Section 3.04(b) of the MPC says that "use of deadly force is not justifiable . . . unless the actor believes that such force is necessary to protect himself against death, serious bodily harm, kidnapping or [rape]," and §3.09(1) makes clear that imperfect justification is not available when the mistake relates to "provisions of the Code." The Commentaries spell out the point further: "Mistake as to the scope of the penal law, such as the belief that deadly force is justifiable to prevent trespass, is declared to be immaterial by Section 3.09(1)." MODEL PENAL CODE AND COMMENTARIES § 210.03 comment 6 (1980). This rule is merely an instance of the Code's general embrace of the principle that ML is no excuse unless a local provision of the Code so provides. See MPC §2.02(9). Beyond the MPC itself, cases raising the issue of mistake as to the statutory categories of justification for the use of defensive force are so hard to find as to suggest that it does not even occur to lawyers that such a defense might be available. Take, for example, understand what circumstances make for a threat to rob, rape, kill or whatever the case may be. Thus, if you live in Pennsylvania and you kill someone who attempts to rob you, you are better off saying that you mistakenly thought the robber was going to kill you, even if such a fear would have been totally unreasonable, than saying that you thought it was justifiable to use deadly force to stop a robbery. The latter is a completely unexcusable ML that leaves you open to a murder charge, while the former is an excusable ML that defeats the murder charge 215 even though these two self-defense MLs hardly seem distinguishable in culpability--or perhaps even conceptually. 216 In either case, after all, the defendant kills under belief that society and its law endorse such killings, even though in fact--or, rather, in law--they do not.
MPC defenses like imperfect justification may or may not embody the humane approach to homicide law, but it should not be obscured that the MPC achieves its results by sneaking a ML defense into a doctrinal home where it did not traditionally Pennsylvania v. French, 531 Pa. 42 (1992), in which the question was whether the use of force against a police officer was justifiable whenever the officer used unlawfully "excessive force" or only when the officer used "deadly force." The court held the latter, and there was no hint that the defendant might possibly have a defense based on her entirely plausible misapprehension of what categories of unlawful force would justify the use of force in return. Similarly, in California v. Ceballos, 12 Cal.3d 470, 526 P.2d 241 (Cal. 1974), no mistake defense as to the available categories of justification appears to have been argued or considered even though the defendant might quite reasonably have mistaken those categories and thus acted under the impression that his use of deadly force was entirely justified. Ceballos had constructed a spring gun in his garage, which severely injured a young boy who tried to break in. A California statute provided that deadly force was justifiable to prevent a felony, which the court, looking to common law, read to include only "atrocious crimes," which, according to the court, did not include the general category of burglary. Id. at 245-246. Before this decision, Ceballos could hardly have been blamed for thinking that "burglary" was a category of justification for the use of deadly force, but it does not seem to have occurred to anyone in the case to argue for a ML defense on those grounds. In a more recent case, Connecticut v. Harrison, 32 Conn. App. 687, 631 A.2d 324 (Conn. 1993), the issue was whether robbery was, as a matter of law, a justification for the use of deadly force. Connecticut law provides that deadly force is justifiable only to prevent the use of deadly force or great bodily harm. The defendant claimed that robbery was within the "class of crimes in prevention of which a man may . . . repel force by force," but he made no argument as to mistake regarding that class of crimes. 215 Of course, under imperfect-justification doctrine, you remain subject to lesser homicide charges. belong and into a doctrinal neighborhood where ML is still mostly unwelcome. By now, it should be clear that, for better or worse, criminal courts have done a good deal of this smuggling of ML defenses into cases in the name of criminal intent while in other cases announcing the continuing integrity of the ML maxim. So the ambiguities of criminal intent survive along with the ad hoc judgments they require of judges, all for the ironic preservation of such formalismt limits on criminal-law discretion as the distinction between ML and MF and the maxim that ignorance of the law is no excuse. And there has been misuse, in my judgment, of at least two general kinds. These categories of misuse, moreover, help to illuminate two doctrinal traditions of criminal intent that have equipped judges to seize on ML doctrine when it fit their sense of justice in the case and pass it by when it did not. First, the claim of right defense and other variations of mistake of "different law" are simply exceptions to the usual mistake of law rule, but that fact has often been obscured by the scholarly and sometimes judicial transformation of that defense into a general principle of criminal intent, 216 ROBINSON, supra note ---, at 414-420 and nn.55 and 67, discusses this problem at some length. regardless of the unrepudiated principle that mistake of law is presumptively no excuse.
I have no settled position on when a court should read a mistake of law defense into a statute that does not plainly authorize one, but it does seem to me that the "different law" doctrine is a conceptually unsound device for allowing judges to evade that question. Second, the notion of "unreasonable MF" turns out to be a contradiction in terms, not a natural subdivision of mistake of fact. As with the "different law" concept, the notion of "unreasonable mistake of fact" seems to me a device that allows judges and legislative drafters to create exceptions to the usual mistake of law rule without acknowledging that that is what they are doing. The coexistence of these categories with the ignorance of law maxim illustrates the persistence of two different concepts of criminal intent, one in which criminal responsibility depends only on knowledge of the facts of one's conduct and one in which responsibility depends on one's willingness to defy the law (or, more likely, the public morality that is most persons' best guide to the law 217 ).
I do not mean to suggest that one of these two traditions must triumph over the other. In particular cases, of course, they overlap quite a bit. Moreover, it is just as possible to have a system in which different crimes are governed by different theories of intent as it is to have a system with a single overarching theory. Either way, a crucial concept like the distinction between mistake of law and mistake of fact should be as well defined as possible.
This article is only a first effort to define an important concept. I resist the idea that the distinction between mistake of law and mistake of fact is ultimately undefinable, a mere matter of degree, but I can hardly claim to have settled the matter.
It is certainly possible that there remains a way to preserve the category of mistake of "different law" as a conceptually reliable category that really does inform the interpretation of criminal laws in a more than ad hoc way. Similarly, the notion of "unreasonable MF" may prove to have a validity that I do not now detect. In some senses, it must be true that the relationship between law and fact, between the reality around us and the system of legal meaning, between factfinding and the verdict the factfinder ultimately declares, is too subtle and too indefinite to be captured in a definition or an algorithm. Undoubtedly, some readers will insist that there remains a material distinction, which I have resisted for my formalizing purposes in this article, between the legal "meaning" of a term and the circumstances that legally constitute that meaning from case to case. 218 In this first exploration of the problem, I conclude otherwise, but the complexities of the problem are far from exhausted. I look forward to further investigation. | 2019-05-15T14:30:33.403Z | 2000-09-07T00:00:00.000 | {
"year": 2000,
"sha1": "36463f33aee31e3f6636b1236a5f62f0ca5150f0",
"oa_license": "CCBYNCSA",
"oa_url": "https://scholarship.law.bu.edu/cgi/viewcontent.cgi?article=1731&context=faculty_scholarship",
"oa_status": "GREEN",
"pdf_src": "Anansi",
"pdf_hash": "02cdf31228b1f568a74382ab2407bb85fe775645",
"s2fieldsofstudy": [
"Law"
],
"extfieldsofstudy": [
"Philosophy"
]
} |
230655588 | pes2o/s2orc | v3-fos-license | The Mediterranean Diet Benefit on Cardiovascular Hemodynamics and Erectile Function in Chronic Heart Failure Male Patients by Decoding Central and Peripheral Vessel Rheology
Background: Mediterranean diet was evaluated on erectile performance and cardiovascular hemodynamics, in chronic heart failure patients. Methods: 150 male stable heart failure patients were enrolled in the study (62 ± 10 years, New York Heart Association (NYHA) classes I–II, ejection fraction ≤40%). A detailed echocardiographic evaluation including estimation of the global longitudinal strain of the left ventricle and the systolic tissue doppler velocity of the tricuspid annulus was performed. Erectile dysfunction severity was assessed by the Sexual Health Inventory for Men-5 (SHIM-5) score. Adherence to the Mediterranean diet was evaluated by the MedDietScore. Results: The SHIM-5 score was positively correlated with the MedDietScore (p = 0.006) and augmentation index (p = 0.031) and inversely correlated with age (p = 0.002). MedDietScore was negatively associated with intima-media-thickness (p < 0.001) and serum prolactin levels (p = 0.05). Multi-adjusted analysis revealed that the inverse relation of SHIM-5 and prolactin levels remained significant only among patients with low adherence to the Mediterranean diet (p = 0.012). Conclusion: Consumption of Mediterranean diet benefits cardiovascular hemodynamics, while suppressing serum prolactin levels. Such physiology may enhance erectile ability independently of the of the left ventricle ejection fraction.
Introduction
The prevalence of chronic heart failure (CHF) is estimated around 2% of the adult population, mainly older aged men [1]. CHF causes substantial morbidity, influences quality of life and reduces life expectancy. Moreover, coronary artery disease accounts for two-thirds of CHF cases, in developed countries. Central hemodynamics and particularly wave reflection amplification constitutes a vital parameter of CHF physiology. Increased arterial stiffness influences the cardiovascular mechanisms resulting in adverse cardiovascular outcomes and poor quality of life, including sexual performance. Erectile dysfunction (ED) is strongly correlated with sex-hormones and prolactin levels, and has been used as a proxy of quality of life. ED has been related with hemodynamic status and seems to display peripheral vessel rheology, even in patients with atherosclerosis and CHF, while the role of serum prolactin role is still debated in CHF mechanisms is still under consideration [2][3][4][5][6][7][8][9].
Dietary habits are recognized as important and modifiable risk factors of cardiovascular disease (CVD), not only in primary, but also in secondary prevention [1,10]. Since the 1970s, large-scale community-based interventions revealed that favorable dietary changes, as a part of an integrated and multi-dimensional lifestyle approach, have led to notable reduction in CVD morbidity and mortality [11]. The current trend in nutrition-based epidemiology is to examine the diet-health interaction through a holistic approach, instead of focusing on isolated nutrients and food items. Several dietary patterns have been studied in the past 2-3 decades for their effect on CVD risk. Among them, the Mediterranean diet is one of the most well studied dietary patters regarding its influence on CVD risk. Mediterranean diet is characterized by high consumption of monounsaturated fatty acids, mainly from olives and olive oil, and suggests daily consumption of fruits, vegetables, whole grain cereals and low-fat dairy products, weekly consumption of fish, poultry, tree nuts and legumes, a relatively low consumption of red meat, and moderate daily consumption of alcohol, usually with meals. It is a different dietary pattern as compared to other healthy patterns, like the ketogenic diet which is low in carbohydrates (lower than the Mediterranean diet), but high in fat (higher that Mediterranean diet). According to the 2010 definition of the United Nations Educational, Scientific and Cultural Organization (UNESCO), the Mediterranean diet (from the Greek word diaita, which means way of life), constitutes "a set of skills, knowledge, practices and traditions ranging from the landscape to the table, including the crops, harvesting, fishing, conservation, processing, preparation and, particularly, consumption of food". Several clinical trials and epidemiological studies have already showed that the Mediterranean diet's health benefits stem from its anti-inflammatory and vasoprotective properties [10,12,13]. However, the impact of the Mediterranean diet on CVD risk in patients with CHF, and particularly its influence on cardiovascular hemodynamics, has been poorly studied and understood.
The scope of this work was to examine the role of Mediterranean diet on erectile function and cardiovascular hemodynamics, in patients with chronic heart failure, in order to illustrate central and peripheral vessel rheology and prolactin physiology.
Design and Setting
This is an epidemiological study that was conducted at the First Cardiology Clinic of Hippokration General Hospital of Athens, in Greece, during 2019-2020.
Sample
We enrolled 157 clinically stable CHF, male patients (62 ± 10 years) who consecutively visited the outpatient unit of our hospital. Eligibility criteria included age of at least 18 years, with left ventricular ejection fraction of ≤40% and symptoms according to New York Heart Association (NYHA) class II or higher symptoms. Patients were receiving standard medical therapy on an optimal tolerated dose, including an angiotensin-converting enzyme inhibitor or angiotensin-receptor blocker or a sacubitril-valsartan combination, a mineralocorticoid receptor antagonist plus a beta-blocker and diuretic. If indicated, they were treated with appropriate device therapy (i.e., ICD-Implanted cardiac defibrillator). Newly (i.e., within 1 month) diagnosed heart failure patients were excluded, as well individuals with cancer, active infection or other major systemic disease.
Bioethics
The study conformed to the principles of the Declaration of Helsinki and was approved by the Ethics Committee of Cardiology Clinic of the Hippokration Hospital (Institutional Review Board 0001/10/6/2019). All subjects had signed a written informed consent form.
Dietary Assessment and Evaluation of Adherence to the Mediterranean Diet
Nutritional habits assessment was based on a validated semi-quantitative foodfrequency questionnaire [13]. All participants were asked to report the average weekly intake of several foods and beverages they consume; the MedDietScore was used to evaluate the adherence to the Mediterranean dietary pattern [13]. MedDietScore is an 11-item index with questions regarding the frequency of consumption of the main nutrients that characterize the Mediterranean Diet (MedDiet). Individual ratings (from 0 to 5 or the reverse) were assigned in each of the 11 items according to their position in the Mediterranean diet pyramid. For the consumption of items presumed to be close to this pattern (i.e., consumption on a daily basis or more than 18 servings/month of non-refined cereals, fruits, vegetables, legumes, olive oil, fish, and potatoes) a score of 5 was given; the contrary score 0 was given when someone reported no consumption, score 1 when they reported consumption of 1-4 servings/month, score 2 for 5-8 servings/month, score 3 for 9-12 servings/month, and score 4 for 13-18 servings/month. For food habits presumed to be distinct from this dietary pattern (i.e., rare consumption of healthy foods or frequent intake of meat and meat products, poultry and full fat dairy products), scores were assigned on a reverse scale. Concerning alcohol drinking, score 5 was assigned for consumption of <300 mL of alcohol/day, score 0 for consumption of <700 mL/day and scores 4 to 1 for consumption of 300, 400-500, 600 and 700 or 0 mL/day respectively. Higher values of this diet score indicate greater adherence to the Mediterranean diet (theoretical range 0-55).
Measurements
Baseline evaluation included information about socio-demographic characteristics, history of hypertension, dyslipidemia and diabetes mellitus as well as dietary and other lifestyle habits (i.e., smoking status and physical activity). In particular, smokers were defined as those who were currently smoking at least 1 cigarette/day during the past year. Frequency and duration per time of the engagement in physical activities, e.g., organized sport clubs, free style, was recorded and physical fitness levels were calculated in metabolic equivalents per minute/day (METS). Body mass index (BMI) (kg/m 2 ) was estimated by dividing measured weight (kg) by measured standing height squared (m 2 ). Arterial blood pressure was measured in patients in a sitting position, calm for at least 30 min; those with average systolic/diastolic blood pressure levels ≥140/90 mmHg or who were under antihypertensive medication were classified as hypertensives. Hyperlipidemia was defined as a history of hypolipidemic treatment or and total blood cholesterol level ≥200 mg/dL. The presence of diabetes mellitus was defined as receiving hypoglycemic medication or having fasting glucose at least 126 mg/dL or HbA1c > 6.5%. Serum prolactin and testosterone levels were measured by a radio-immuno-assay method. Patients were instructed to abstain from food and smoking for 12 h and blood samples were taken before 9:00 am.
Cardiac Ultrasound and Ergometric Parameters
Echocardiographic assessment was performed in all patients by using a GE 5500 series equipment with a multifrequency transducer (2.5-4 MHz) and tissue doppler imaging (i.e., TDI) technology. Left ventricular ejection fraction (EF) was calculated by the Simpson's method, while right ventricular systolic function was evaluated upon the systolic component of the Tissue Doppler velocity of the tricuspid annulus (SRV). Left atrial (LA) maximal volume and global longitudinal strain of the left ventricle (GLPS) were measured (biplane modified Simpson method). In cardiorespiratory exercise (ergometric bicycle) VO2max (maximum oxygen consumption), VE/VCO2 (ratio of ventilation to carbon dioxide) and the METS (i.e., metabolic equivalents per min/day) achieved were evaluated. The ultra-sonographic examinations were carried out by a group of 4 clinicians who had at least 5-year experience in ultra-sonography in a clinical setting.
Vascular Measurements
Carotid intima-media thickness (IMT) measurements were performed with the patient lying in the supine position and with the neck rotated to the opposite side of the examination; it was measured in both right and left common carotid artery (1 cm proximally to the carotid bulb), the carotid bulb and the internal carotid artery (B-mode ultrasound imaging, 14.0-MHz multifrequency linear array probe, Vivid 7 Pro; General Electric Healthcare, Mil-waukee, WI, USA). Three measurements of the maximal IMT in the far wall during diastole were averaged, and the summary was calculated for both carotid arteries. The presence of plaque was defined as a clearly identified area of focally increased IMT greater than 1.5 mm or 50% increase as compared to adjacent wall IMT [14,15]. Carotid-femoral pulse wave velocity (PWV) that evaluates central arterial stiffness was calculated by measuring pulse transit time and the distance travelled between 2 recording sites (carotid and femoral). A validated, non-invasive device (Complior ® , Artech Medical, Pantin, France) allowing online pulse wave recording and automatic PWV calculation was used. Wave reflection amplification analysis was performed by a radial artery tonometry device (Sphygmocor System-Atcor Medical, Sydney, Australia) [14,15]. Peripheral pressure waveforms were recorded at the radial artery using a hand-held high-fidelity tonometer and calibrated by using arterial pressures measured at the brachial artery. The Augmentation Index (AIx), a measure of wave reflection amplification, was calculated as the ratio of the augmentation of aortic systolic pressure due to the reflected waves divided by the aortic pulse pressure and normalized for a heart rate of 75 bpm.
Evaluation of Erectile Dysfunction
Diagnosis and evaluation of the severity of erectile dysfunction were based on clinical symptoms, duration of the erectile disorder, and the score of the five-item form of the International Index of Erectile Function, the Sexual Health Inventory for Men (SHIM-5) score (less than 21 indicates erectile dysfunction). The lower the score the more severe the erectile pathology [16,17]. The vascular origin of the erectile disorder was based on the absence of any diagnosed psychological, neurologic or anatomic abnormalities, pelvic surgery or trauma, or history of malignancy and endocrine gland disorders.
Statistical Analysis
The working sample of n = 150 participants was found to be adequate, i.e., achieving 83% statistical power, to evaluate 1 standard deviation changes (two-sided hypotheses) on the tested hemodynamic markers by tertile change of the MedDietScore, at 5% significance level. Continuous variables are presented as mean values ± standard deviation. Categorical variables are presented as absolute values and relative frequencies. One-way analysis of variance (ANOVA) was used to evaluate associations between mean values of continuous variables between 3 or more independent groups. The normal distribution of the continuous variables was assessed using the Shapiro-Wilk test. The Kruskal-Wallis test was used as an alternative non-parametric approach in the case on non-normally distributed variables. Correlations between continuous variables were evaluated by the calculation of Pearson's r-coefficient or the partial correlation coefficient, after adjusting for potential confounding factors. Linear regression models were estimated to assess the association of vascular measurements, SHIM-5 score, GLPS and hormone levels with adherence to the Mediterranean diet. Regression estimates are presented as standardized b-coefficients. Mediation analysis, using Baron and Kenny's approach, was applied to evaluate the potential intervening role of the Mediterranean diet on the relationships between prolactin, testosterone levels and cardiac hemodynamic markers. All reported p values are based on two-sided tests. The SPSS statistical software package version 23.0 (Statistical Package for Social Sciences, SPSS Inc., Chicago, Illinois, USA) was used for all statistical calculations.
Results
The majority of the patients, i.e., 89%, were NYHA class II, and the rest 11% were class III. Mean EF was 30.5% (7.9) (according to the study's protocol all patients had EF <40%), mean MedDietScore was 27.9 (3.4) (suggesting moderate adherence to the Mediterranean dietary pattern, i.e., 27.9 out of 55 theoretical total score = 50.7%), and mean body mass index (BMI) was 30.0 kg/m 2 (8.7), whereas 45.8% were considered as obese (i.e., BMI > 29.9 kg/m 2 ). Demographic and clinical characteristics of the participants by level of adherence to the Mediterranean diet are presented in Table 1. Unadjusted analyses revealed an inverse association between the MedDietScore categories of adherence and prolactin levels (p = 0.05), IMT (p < 0.001), and a positive association with testosterone (p = 0.02) and SHIM-5 score (p = 0.006). No other significant associations between Mediterranean diet and cardiovascular hemodynamics markers were observed in the unadjusted analyses. However, residual confounding may exist, therefore, multi-adjusted models were fitted having MedDietScore as independent factor and various hemodynamics markers as outcomes, adjusting for age, BMI, physical fitness levels of the participants and history of diabetes, hyperlipidemia and hypertension (Table 2). Data analysis revealed a positive association of the MedDietScore with SRV (p = 0.019) and the SHIM-5 index (p = 0.002), as well as an inverse association with AIx (p = 0.014), and IMT (p = 0.004). No significant associations were observed between MedDietScore with cardiac hemodynamic markers, as well as testosterone and prolactin levels (all p-values > 0.10).
Moreover, the impact of physical fitness on the hemodynamic variables examined was also evaluated; it was revealed that physical fitness was inversely associated with AIx (p = 0.010), and IMT (p = 0.01), after adjusting for age, BMI, history of diabetes, hyperlipidemia, hypertension, EF.
To further evaluate the role of sex-related hormones, i.e., testosterone and prolactin, on cardiac hemodynamics of CHF patients, multi adjusted correlation analyses were applied (Table 3). Prolactin was positively correlated with SRV (p < 0.001), with Aix (p = 0.03), and inversely correlated with VO 2 at max (p = 0.03), as well as with SHIM-5 score (p = 0.01). Testosterone was inversely correlated with prolactin (p < 0.001), and positively with SHIM-5 score (p = 0.04). Furthermore, SHIM-5 score was inversely correlated with age (r = −0.677, p = 0.002), and positively correlated to AIx (r = 0.338, p = 0.031). To further evaluate the role of Mediterranean diet adherence on the relationships between sex-related hormones and cardiac hemodynamics, a mediation analysis was applied. In particular, when the level of adherence to the Mediterranean diet was taken into account prolactin levels were not associated anymore with SRV (p = 0.959), or VO 2 max (p = 0.528). Moreover, prolactin levels were associated with reduced SHIM-5 score only among participants with low adherence to the Mediterranean diet (i.e., lowest tertile, b-coefficient: −0.559, p = 0.012), whereas among the other higher adherence groups no significant association between prolactin and erectile function was observed. In addition, adherence to the Mediterranean diet seemed to play a mediating role in the relationship between testosterone levels and erectile function. In particular, among participants who reported low adherence, testosterone levels were not associated with SHIM-5 score; however, among those who reported good or very good adherence (i.e., 2nd and 3rd tertile), testosterone levels were positively associated with SHIM-5 score (p-values <0.01). Table 2. Results from regression models that evaluated the role of Mediterranean diet adherence (through MedDietScore, independent variable) on cardiovascular hemodynamics factors (outcome), in chronic heart failure patients.
Discussion
The present study revealed a multifaceted beneficial impact of the Mediterranean diet in middle-aged men with CHF, by ameliorating central hemodynamics, right ventricular function and deescalating parameters of arterial stiffness, while exerting atheroprotective effect with suppressed circulating serum prolactin levels. Such physiology enhances erectile performance, an important component of quality of life in the male population with heart failure.
Several studies underlined the anti-inflammatory properties of the MedDiet and its components and the positive impact on primary and secondary cardiovascular prevention [13,[18][19][20]. In heart failure patients, higher adherence to the MedDiet has been associated with reduced NT-proBNP (NT-pro Brain Natriuretic Peptide levels), oxidized low-density lipoprotein and Lp(a) levels [21] while it seems to reduce the risk of a heart failure incident, as shown in a population-based cohort of 32,921 subjects [22]. Concerning the relationship between the MedDiet and left ventricular dysfunction, we revealed in a previous study of 372 heart failure patients that higher adherence to this dietary model ameliorates left ventricle diastolic function and, in particular, the ventricular filling physiology (left atrium ejection fraction and E/A ratio) [19]. Another study reported that NYHA class symptoms negatively correlate with the adherence to the MedDiet pointing to a modulating role in the severity of clinical presentation [23].
According to previous studies, in patients with systolic heart failure, the presence of right ventricular dysfunction is frequent and has been related to increased mortality. In our study, right ventricular function, as it was evaluated by tissue Doppler imaging, was positively related to MedDiet linking higher adherence to a healthy diet with amelioration of the right ventricular systolic function. Interestingly, the favorable effect of the MedDiet on right ventricular systolic function was independent of the left ventricle ejection fraction. We suggest that the MedDiet regime with the unique anti-inflammatory properties may induce vasodilation and lessen resistance of the pulmonary vascular bed. Such physiology reduces right ventricular afterload and enhances SRV independently of the left ventricular ejection fraction. It is well known that inflammation correlates with diastolic systolic dysfunction, and this may be an explanation for the deterioration from compensated to de-compensated heart failure [24][25][26].
Moreover, in this study, a greater MedDiet score was associated with a higher augmentation index and improved erectile function; while there was a trend with GLPS. These seemingly contradictory findings may be interpreted by the fact that heart failure physiology of the cardiovascular system is fundamentally distinct and that functions of the LV and the peripheral arterial system are interactive (i.e., ventricular/ vascular coupling). Thus, in order to compensate the reduced left ventricular pumping force and consequently the forward pulse wave, peripheral resistance augments to such point that both actions lessen the wave reflection amplification signal. This may be falsely interpreted as reduced arterial stiffness. High adherence to the MedDiet may reduce inflammation, assist vasodilation and help improving peripheral vascular resistance. Such action reassures erectile potency, a nitric oxide phenomenon based on the capacity of the penile arteries to dilate and augment blood flow to the corpus cavernosum. We clinically identified such physiology by the SHIM-5 score, a simple, cost effective and reliable tool that evaluates erectile dysfunction.
Regarding the reflected wave amplification, a higher AIx may be interpreted as the result of improved vascular performance in the periphery that permits first the reflexion of the forward pulsed volume by the falling myocardium and second to such a degree that it finally enhances the pulse signal. Therefore, in such heart failure conditions a higher AIx may be interpreted as a sign of reduced stiffness mainly in the peripheral vascular bed [27]. The fact that pulse wave velocity, typically an index of central arterial stiffness, did not correlate with the MedDiet score further supports such findings. Moreover, such phenomenon may take place independently of the LVEF. Both actions on peripheral physiology and reflected waves may enhance erectile performance, a vital parameter of quality of life in the male population.
Regarding the role of prolactin in heart and vessels diseases, it is increasingly recognized that this hormone is involved in cardiovascular pathophysiology, exerting pleiotropic unfavourable haemodynamic and pro-atherosclerotic actions, raising overall cardiovascular risk [6,9,28,29]. Hyperprolactinaemia has been related to endothelial dysfunction and arterial stiffness [3], low-grade inflammation [2,5], increased thromboembolic risk [7] and dyslipidemia [2]. For example, plasma prolactin levels are elevated in conditions such as the acute phase of myocardial infarction, ischemic stroke, and transient ischemic attack, hypertension and preeclampsia, while evidence suggests a causative role of prolactin in postpartum cardiomyopathy [5][6][7][8]28]. Furthermore, prolactin may accelerate arteriosclerosis in women in early menopause by increasing central as well as peripheral blood pressure and arterial stiffness [9]. Prolactin levels seem to have a prognostic role of future CVD events in men with erectile dysfunction [2] and was associated with all-cause and CVD mortality in the general population, as well as in patients with chronic kidney disease [8]. However, the particular role of prolactin in heart failure in particular has not been fully understood. While small studies have reported elevated serum levels of prolactin in heart failure patients [5], no relationship was detected between prolactin circulating levels and NT-proBNP levels in heart failure elderly patients [29]. No association was also found between CVD mortality and prolactin concentration during 10 years of follow-up [29]. On the contrary, a more recent study indicated that in patients with heart failure of either ischemic or non-ischemic etiology, serum prolactin levels were closely related to functional status and exercise capacity, while acting as an independent prognosticator of total death or hospitalization for cardiac reasons [7].
This study revealed the association between prolactin and IMT, confirming prolactin's atherogenic properties [30]. However, previous studies yield contradictory results. Subjects with untreated prolactinoma presented increased carotid IMT [31], whereas in premenopausal women high-normal prolactin levels were shown to correlate with lower carotid IMT [32]. In post mortem human coronary arteries obtained after endarterectomy, increased expression of prolactin receptor was found in advanced atherosclerotic arterial walls, whereas no expression was detected in non-significant atherosclerotic lesion, suggesting a modulating role of prolactin in atherosclerosis process [9]. According to our findings, the MedDiet may indeed mitigate the atheromatic effect of prolactin, as high adherence to MedDiet was negatively associated both with IMT and circulating prolactin, suggesting a beneficial atheroprotective effect in patients with systolic heart failure and systolic dysfunction.
Heart failure remains the late serious clinical condition of CVD. As cardiovascular system works in a conduit of the heart and the whole arterial tree, any alterations in cardiac function have impact on arterial properties, and vice versa. The measured central aortic pressure waveform represents the sum of the forward arterial wave, generated by the left ventricle during ejection, and the backward moving wave, caused by the reflection of the forward moving wave within the arterial system. Thus, central aortic pressure waveform expresses the systolic function of the left ventricle, the elastic properties of the entire arterial tree and the interaction between ventricle and aorta, described as aortoventricular coupling [33,34]. In the progress of heart failure as left ventricle dilation occurs and the ventricles lose their compensation, the heart muscle cannot generate the necessary extra force to overcome the late systolic augmented pressure. Thus, augmented pressure, systolic and pulse pressure become diminished in parallel with the systolic ejection duration shortening, as the wave reflection poses a negative influence on flow. In this scope, interventions that can modify contractility of the left or right ventricle, can offer a benefit effect on arterial function, improving ventricular-aortic coupling, which will be expressed by an increase in augmentation index and improved clinical status of the patients.
In conclusion, adherence to the MedDiet regime in systolic heart failure male patients posed a significant improvement in cardiovascular physiology including systolic function, parameters of arterial stiffness, atheromatosis and erectile performance. In this particular finding on a male patient population with raised mortality and morbidity MedDiet dietary pattern can be highly recommended as a lifestyle culinary option to maintain cardiovascular health and the sense of wellbeing. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: Data available on request due to restrictions e.g., privacy or ethical. The data presented in this study are available on request from the corresponding author. | 2020-12-31T09:02:26.639Z | 2020-12-30T00:00:00.000 | {
"year": 2020,
"sha1": "4b4bf7d73ee5c37fe95009fc68e609e9779cd6c6",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2072-6643/13/1/108/pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "ad09e54ba278bc72f0504559dc73fe499fda9ee6",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
2961984 | pes2o/s2orc | v3-fos-license | Factors predictive of treatment failure in staphylococcal prosthetic vascular graft infections: a prospective observational cohort study: impact of rifampin
Background There exists considerable debate concerning management of prosthetic vascular graft infection (PVGI), especially in terms of antimicrobial treatment. This report studies factors associated with treatment failure in a cohort of patients with staphylococcal PVGI, along with the impact of rifampin (RIF). Methods All data on patients with PVGI between 2006 and 2010 were reviewed. Cure was defined as the absence of evidence of infection during the entire post-treatment follow-up for a minimum of one year. Failure was defined as any other outcome. Results 84 patients (72 M/12 F, median age 64.5 ± 11 y) with diabetes mellitus (n = 25), obesity (n = 48), coronary artery disease (n = 48), renal failure (n = 24) or COPD (n = 22) were treated for PVGI (median follow-up was 470 ± 469 d). PVGI was primarily intracavitary (n = 47). Staphylococcus aureus (n = 65; including 17 methicillin-resistant S. aureus) and coagulase-negative Staphylocococcus (n = 22) were identified. Surgical treatment was performed in 71 patients. In univariate analysis, significant risk factors associated with failure were renal failure (p = 0.04), aortic aneurysm (p = 0.03), fever (p = 0.009), aneurysm disruption (p = 0.02), septic shock in the peri-operative period (p = 0.005) and antibiotic treatment containing RIF (p = 0.03). In multivariate analysis, 2 variables were independently associated with failure:septic shock [OR 4.98: CI 95% 1.45-16.99; p=0.01] and antibiotic containing rifampin [OR: 0.32: CI95% 0.10-0.96; p=0.04]. Conclusion Results of the present study suggest that fever, septic shock and non-use of antibiotic treatment containing RIF are associated with poor outcome.
Background
Prosthetic vascular graft infections (PVGI) are uncommon and severe, and are associated with high mortality and morbidity [1][2][3]. Staphylococcus species are most commonly associated with PVGI, regardless of the onset (early or late) and localization of the PVGI. Up until now, there has been an absence of evaluation of medical treatment in terms of patient outcome. Most studies reported in the literature were limited by their small population size and non-uniform microbiological criteria for defining infection. Fitzgerald et al. [1] and other authors [4][5][6] suggested 6-week intravenous treatment following oral treatment, for a total of 6 months (or, for certain selected patients, their entire lifetime). In addition, details on antimicrobial agents used, their route of administration and their duration were not mentioned in most published studies [7][8][9][10][11][12]. Antimicrobial agents with bactericidal properties and activities against biofilm, such as rifampin (RIF), were of interest in PVGI [13,14], as reported in treatment of prosthetic valve infective endocarditis [15][16][17][18] and prosthetic joint infection [19][20][21][22].
The role of RIF in systemic antimicrobial therapy for staphylococcal PVGI has not been previously assessed in clinical studies. The aims of this study were therefore to identify factors associated with treatment failure in a cohort of patients with staphylococcal PVGI, and to assess the impact of RIF in systemic antibiotic therapy administered to these patients.
Study design
This was a prospective study of an observational cohort of 84 from 200 patients admitted at two French reference centers for PVGI and treated for staphylococcal PVGI. Results were reported for 84 selected patients which were followed-up for a minimum of one year after the end of treatment in cases of graft implantation (3 months in cases of removal of graft). We compared characteristics of patients according to outcome.
Ethics committee
This study was approved by the institutional review boards of both Dron and university of Lille Hospitals, including the use of the data's patients for publication. All patients included in this study were informed and gave their consent to the medical and surgical treatment. Their consent was drawn in the medical record.
Study population
All patients treated from 2006 to 2010 for PVGI due to Staphylococcus spp. were identified in our database of 200 patients, by searching for "Staphylococcus aureus" and "coagulase-negative staphylococci (CNS)". Ninetyone patients were selected based on microbial results and the delay of follow-up (> one year of follow-up after the end of treatment). Seven were excluded due to absence of data on medical treatment or insufficient follow-up.
Definition of PVGI
We used the same definitions as in previous studies from our group [3,23] which were derived from those proposed by Fitzgerald et al. [1].
Definite PVGI
A patient was considered as having definite PVGI if at least two of the three following criteria were present: (i) positive bacterial culture of intraoperative specimens or blood samples (for potentially contaminating bacteria such as CNS, Propionibacterium acnes and corynebacteria, at least two intraoperative specimens or blood samples, or at least one intraoperative specimen and one blood culture, were required); (ii) clinical signs of infection [general (fever, chills, septic shock) or in the area of the prosthesis (e.g. inflammatory signs in the area of the vascular graft: local pain, erythema or tumefaction, sinus tract infection communicating with PVGI, enteric aortic fistula, intraoperative gross purulence or failure of graft consolidation]; (iii) biological signs of infection (C-reactive protein > 10 mg/L, white blood count > 10 G/L) or other radiological signs of infection (perigraft air or fluid persisting for more than 8 weeks postoperatively, abscess contiguous to the implant). Each case of definite infection was classified as early-onset infection (i.e. occurring within 4 months after surgery) or as late-onset infection (i.e. occurring more than 4 months after surgery).
Presumed PVGI
PVGIs infections were suspected when bacteremia related to a site other than the surgical site occurred within 4 weeks after graft implantation) [24,25].
"Onset of PVGI" was defined as time from implantation of the prosthesis to clinical onset of infection and was categorized into early (< 4 months after the implantation) or late (≥ 4 months after the implantation) PVGI.
Microbiological documentation
Blood specimens for culture were drawn from all febrile patients. Superficial samples were not used. Antibiotic susceptibility patterns were interpreted in accordance with recommendations from the Comité de l' Antibiogramme de la Société Française de Microbiologie (http://www. sfm-microbiologie.org/UserFiles/files/casfm_2003.pdf).
Risk factors for PVGI
Diabetes mellitus was defined according to the international classification [26]; chronic obstructive pulmonary disease (COPD) as per the American Thoracic Society [27]; obesity as body mass index (BMI) ≥30 kg/m 2 and overweight as 25 ≤ BMI < 30 kg/m 2 ; malnutrition as ≤19 kg/m 2 [28]; chronic renal failure according to the value of creatinine clearance determined by the Cockroft-Gault equation [29]; and immunodepression (e.g. steroid therapy > 7.5 mg per day, cancer, AIDS).
Medical and surgical management Empirical antibiotic treatment
Empirical treatment was started immediately after microbiological intraoperative samples were taken in patients without severe sepsis and consisted of a combination of broad-spectrum beta-lactams (i.e. piperacillin-tazobactam, cefepim-metronidazole, etc.) and anti-methicillin-resistant S. aureus (MRSA) agents (i.e. vancomycin, daptomycin, linezolid) ± aminoglycosides. For patients with severe sepsis or septic shock, this empirical treatment regimen was started before surgery. No antibioprophylaxis was given before surgical revision. Empiric postoperative antibiotic therapy was defined as adequate if it contained ≥ 1 antibiotic agent active against the pathogen(s) identified in intraoperative or blood cultures.
Definite antibiotic treatment
Empirical antibiotic treatment was de-escalated as soon as the susceptibility of microbial agents was available, and included RIF according to the antibiotic susceptibility profile of the staphylococcal strains. The daily dose of RIF was 20 mg/kg divided into two doses, without exceeding 1,800 mg/day, initially by the intravenous route (5-7 days) and then by oral administration.
Total duration was 3 weeks in case of removal with or without replacement of the implants. For allo/homografts and prosthetic grafts, total duration was 6 months (including 6 weeks intravenously), and in some cases, was prolonged in the form of suppressive therapy using, in most cases, oral doxycycline once daily at the discretion of infectious physicians.
Surgical treatment
For us, the optimal surgical option was complete debridement of devitalized and infected tissues around the prosthesis, total graft excision and in situ reconstruction with a new prosthesis, autogenous vein or arterial allograft/homograft. Debridement without graft excision was proposed in patients with very early PVGI or in patients with severe co-morbidities. Finally, when revascularization was not possible, amputation was proposed to the patient.
Outcome
Remission was defined as the absence of local or systemic signs of infection assessed during the most recent contact with the patient, along with absence of the need to re-operate or to administer antibiotic therapy directed to the initial infected site, from the end of treatment to the most recent contact. Failure was defined as any other outcome, including new surgery for infection or death related to PVGI during follow-up. Results were reported for patients within a minimum follow-up of 3 months in case of removal of PVGI with venous graft, or within one year for the other cases. In cases of suppressive therapy, we consider a patient in "remission group" if there is no clinical, radiological or biological specific signs of PVGI after one year of follow-up after its theoretical end of treatment.
Statistical analysis χ 2 -test was used to compare qualitative variables and a 2-sample t test to compare continuous variables. A pvalue < 0.05 was considered a significant difference. To determine independent variables associated with outcome, we performed a complete multivariate analysis including prognostic factors associated with P value < 0.10 in bivariate analysis and, if clinically relevant, some forced variables. Then, adjusted ORs were computed with a logistic regression analysis. To compare the survival distribution of the two samples, we used the log rank test to assess the validity of the model. Statistical analysis was performed using STATA V7.
After a mean post-treatment follow-up > 1 year, remission of infection was observed in 63 patients (75%) (Figure 1). Twenty-one patients failed, most of whom had to be admitted in intensive care unit at initial presentation.
In multivariate analysis, only 2 variables were independently associated with failure: septic shock [OR 4.98: CI 95% 1.45-16.99; P = 0.01] and antibiotic treatment containing RIF [OR 0.32: CI 95% 0.10-0.96; P = 0.04]. The quality of the model has been evaluated through its measures of sensibility and specificity with establishment of ROC curve and measure of the area under the curve who was 0.742.
The Kaplan meier estimates of cumulative failure free period according to the treatment group was detailed on Figure 2. The validity of the model was assessed with the log rank test which was 0.039.
Discussion
We report the outcome of 84 patients treated for staphylococcal PVGI. Remission was observed in 75% of our patients despite adequate empirical antimicrobial treatment most often associated with optimal surgery. The results of the present study suggest that shock in the peri-operative period at initial presentation, renal failure, aortic aneurysm, fever and aneurysm disruption are risk factors in failure. Conversely, methicillin resistance and intracavitary PVGI were not associated with worsened outcome. In our previous study [3], 16/71 patients (22.5%) treated for PVGI and who had minimum follow-up of one year died from persistent, recurrent or new infection. Independent prognostic factors in in-hospital mortality were: age > 70 years and aortic PVGI. Chalmers et al. [30] found that MRSA was associated with poor outcome. In the literature, other risk factors in failure were identified, including conservative treatment [31], partial graft preservation [32], reimplantation of a new expanded tetrapolyfluoroethylen (ePTFE) prosthetic graft, another operation [33], inappropriate antibiotic use, mycotic aneurysm [12], Pseudomonas aeruginosa-related PVGI [34] and thoracic localization [35]. These factors were also associated with a high mortality rate, varying from 6 to 100%.
We also found that use of RIF combinations was associated with improved outcome of staphylococcal PVGI when compared to other antibiotic regimens, consistent with results of previous studies on other device-related infections such as those linked to orthopedic devices [20,36,37] and prosthetic valvular endocarditis [17,38]. Use of RIF combinations in these infections is supported by its intense and rapid bactericidal activity against staphylococci, its ability to penetrate into the biofilm and its persistent activity against microorganisms in stationary growth phase that are present at the surface of infected implants, including vascular grafts [13][14][15][16].
In experimental vascular graft infection, Edmiston et al. [14] investigated the activity of 6 antimicrobial agents (i.e. linezolid, RIF, daptomycin, ceftriaxone, vancomycin and gentamicin) against biofilm-forming and non-biofilmforming strains of staphylococci adhering to graft prosthetic surfaces. They suggested that bactericidal activity is influenced by: (i) the composition and structural characteristics of the biomedical device surface (i.e. Dacron or ePTFE); (ii) the selective activity of the antimicrobial agent; and (iii) the presence or absence of an exopolysaccharide biofilm. Daptomycin, RIF, and linezolid demonstrated greater efficacy and speed at eradicating the microbial adherence of staphylococcal isolates in ePTFE infection than in Dacron infection.
In clinical studies on PVGI, medical treatment was heterogeneous in terms of the antimicrobial agents used, use or not of combined antibiotic and duration and administration of treatment, and was mainly based on the physician's experience [1,4,5,39]. The optimal choice of antimicrobial agents in empirical or definitive therapy and the duration of treatment of PVGI remain unclear. Our results do not argue for including RIF in empirical antibiotic regimens (i.e. before definitive microbiological results are available). Indeed, RIF should only be prescribed after definitive bacteriological documentation, and should systematically be associated with another effective molecule, thereby reducing the risk of emergence of RIF-resistant staphylococcus mutants. In the present study, RIF was initiated in combination with another active antistaphylococcal agent chosen on the basis of definitive microbiological results. In addition, RIF was begun only after all devices aimed at draining the surgical site had been removed, so as to reduce the risk of selecting RIF-resistant mutants.
To our knowledge, the present study is the first to include patients with a homogenous definition of PVGI and treated via homogenous approaches; moreover, we performed long-term follow-up. The present study had several limitations. Potential bias might reside in the variety of surgical approaches, inclusion of aortic and limb PVGI in the same analysis, the delay to surgical procedure and, finally, selection of patients treated at a referral center. Its' the reason why we cannot firmly conclude that rifampin is the only factor which improves the outcome. The homogenous approach in the management of these patients, the management in the peri-operative period in operating theater and in the intensive care unit are the key of the treatment of patients with PVGI.
Conclusion
In conclusion, use of RIF combination as targeted treatment for staphylococcal PVGI based on reliable microbiological documentation seems to be associated with improved outcome of PVGI compared to other antibiotic regimens. The present study suggests a novel example of the advantage of RIF combinations administered under optimal conditions for treatment of staphylococcal implant infections. | 2017-06-21T10:39:25.682Z | 2014-04-28T00:00:00.000 | {
"year": 2014,
"sha1": "e3d257198e0bde2ea127fcfe266f02196cff36cc",
"oa_license": "CCBY",
"oa_url": "https://bmcinfectdis.biomedcentral.com/track/pdf/10.1186/1471-2334-14-228",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "c1f77e8b05ec4ceebdaa1865ca461e924d30d9ee",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
217491646 | pes2o/s2orc | v3-fos-license | THE USE OF TRANSITIONS IN THE STUDENTS ARGUMENTATIVE ESSAY
One of academic writing assignment is argumentative essay. A way to make a coherent paragraph is using transitions which should be clearly and demonstrated in an argumentative essay. Transitions can be words and/or phrases that connect the idea in one sentence with the idea in another sentence. By using transitions, the wriring can be read smoothly. The purposes in conducting the study are to know and describe the use of transitions in student’s writing argumentative essay and the awareness of the students about using transitions. This is a qualitative research-content analysis. The subject of the research is the third grade students of English Department in academic year 2015/2016. There were 23 students: 3 good students, 17 average students, and 3 poor students. The instruments are documents, interview, and student’s journal. After collecting the data, the next step is analyzing them by using coding system, tabulation, and describing them. The results of this research are 1) almost all good students used transitions in essay correctly although there were few incorrect usages, then the average students and poor students had more corrections because they put the transitions in the wrong place, chose inappropriate transitions, used wrong grammatical, and over-use transitions, 2) All students in each category had good knowledge about the transitions, such as its definition, its types, and its function, how importance using transitions are. But sometimes they still felt difficult to use transitions. In conclusion, most of students in every category used transitions correctly even though there were some corrections, then the third grade students of English Department were conscious of transition signals, but they need more practice to apply it in the sentences in order to make their writing had better quality, easier to read and more understandable. Key Words : Argumentative Essay, Transitions
BACKGROUND
Writing is the most difficult skill. It is considered as the most complicated language skill to be learned, compared to other language skills. Brown (2004: 218-246) says that among the language skills, writing is the most complicated and the most difficult skill. It is because in writing, there are several rules that have to be mastered by the students such as spelling, grammar and punctuation, coherence and organization of ideas. It means that the students should master writing although it is difficult because writing can be a way to keep a communication between one and one another and the students have to be able to construct the paragraph coherently.
In University, one of academic writing assignments is argumentative essay.
Here, the researcher chooses an argumentative essay because in argumentative essay, the students are able to write as free as possible but still in the context depending on their mind. Oshima and Hogue (2006: 142) support that statement because the argumentative essay is an essay in which students agree or disagree with an issue, using reasons to support their opinion. This essay also forces the students to think on their own. They have to take a stand on an issue, support their stand with solid reasons, and support their reasons with solid evidence.
Knowing about the organizational structure of academic writing, it can be said that in academic writing, the students should give attention not only in its organizations of idea but also in its coherence. According to Oshima and Hogue (2007: 78), a coherent paragraph flows smoothly from beginning to end. Three ways to give paragraph coherence are using nouns and pronouns consistently throughout a paragraph, using transition signals to show relationships among ideas, and setting ideas into some kind of logical order, such as logical division. In this research, the researcher takes only a way to make a coherent paragraph that is using transitions because it is obviously needed in argumentative essay. Transition signals are words and phrases that connect the idea in one sentence with the idea in another sentence. Transitional signals show movement from either the pro side to the con side or the other way around, like "on one hand", "on the other hand", "however", should be clearly and demonstrated in an argumentative essay.
The students in University should aware of using transitions in their writing but the use of transitions (connectives) almost is ignored in teaching and learning process. This causes the learners confused in using and understanding how and where the transitions can be applied in their writing. McCarthy (1998, see Badiozzaman and Gorjian, 2014: 99) believed that the ability to form a coherent discourse is one of the important aspects of second/foreign language learners' communicative competence. Therefore, if language learners want to create a connected and well-structured written discourse, they should gain the right use of cohesive devices, such as connectives, in order to develop their essays in a cohesive way.
Related to the reasons above, the researcher needs to conduct a research entitled "An Analysis of Using Transitions in Writing Argumentative Essay of The English Department Students at University of Nusantara PGRI Kediri Academic Year 2015/2016". In this study, the researcher decides to focus on the third grade students of English Department at University of Nusantara PGRI Kediri, especially some good, average, and poor students in writing. Their writing score can be looked up from first up to fifth semester whether stable or not to put in good, average, and poor category. Good category, here, means the students who got minimum three scores "A" in writing, then average category means the students who got mostly score "B" in writing, and poor category means the students who got score "C" and/or "D" in writing. Then, the researcher considers that using transitions is the best way to make their writing, argumentative essay, coherently.
The objectives of the research are to describe the use of transitions in student's writing argumentative essay and also to describe the awareness of the English Department students about using transitions.
METHODS
This research used content analysis to analyze and describe the use of transitions in student's argumentative writing. The human in qualitative research is the primary instrument. In this research, the researcher asked the documents of student's writing and interviewed the students. Hence, the existence of the researcher was emphasized because the goal of this observation referred to the document and the interview. The research conducted at the third grade students of English Department in University of Nusantara PGRI Kediri. For taking a sample, the purposive sampling was used. There were 23 students: 3 good students, 17 Journal of English Teaching and Research average students, and 3 poor students. The data of the research were students and student's document. The data collecting procedures are as folows; first, taking the student's document about argumentative essay. Second, interviewing some students in each category using guided interview and stimulated recall methodology (SRM) and asking them to write their journal. The following are the steps in using the stimulated recall method based on Hidayati (2008: 29): First, the researchers distribute copies of the text to the subjects to be checked by them.
Then, she asks them to underline words in their writings wherein they thought they have used transitions.Then, if they do not have any idea or have already forgotten about the transition words, researcher gives the subjects the list of transitions by bringing out the subjects' understanding indirectly. The questions such as 'what, how and why' are used to recall their memory. After they have finished underlining the transition words, researcher compares their findings to the researcher's own findings. At last, researcher interprets the result to find out the meaning and insights from the data.
The coding system was used to know the use of transitions in student's writing. Then the data were analyzed using precentage formulation to count the frequency then the data from interview and students' journal ,to know their awareness in using transitions, were described qualitatively.
In qualitative research, the result needs to be checked its validity and reliability. This research had inter-rater to analyze the data, especially in finding the appropriate use of transitions, in order to measure and compare that the range of score between the researcher and inter-rater is not too far. In conclusion, the results of the research are valid and reliable.
FINDINGS AND DISCUSSION
The result of this research can be seen from the result of data analysis. First, to know the use of transitions in student's argumentative essay, the researcher made tabulation below and devided it into three categories of students : show similarities in the number of 31.91%, and the least transitions used by poor students were the use of transitions in order of importance, to explain or restate an idea, to make a stronger statement, and to add a conclusion in the number of 0%.
Besides, it is also found that the correct and incorrect use of using transitions in writing argumentative essay. In good student's criteria, the use of transitions in order of time is 13 transitions (17.57%). It is found that all students used this type correctly with its function to show the time in their sentences, for example, 'Second is the cheater mostly will not able to catch the main point of the lesson'. Then, in average student's criteria, the use of transitions in order of time is 77 transitions (15.31%). It is found that most students used this type correctly, for example, 'The first reason, cheating is unfair to others'; but some students used it both correctly Journal of English Teaching and Research and incorrectly, for example, 'If a college student cheats on all the work they do for their major, when and if they get a job, that person will be totally unprepared and be a total mess at work, yet be paid for something they don't know how to do'.
Then, in poor students criteria, the use of transitions in order of time is 5 transitions (5.32%). From the data, it is found that only two students used this type and they used it correctly. An example of this type is 'This case happens when their parents to pamper their children until their children dare to lie'.
Second, to know the awareness of the students about using transitions, the researcher not only interviewed good student, average student, and poor student, but also asked them to write a journal. Then, there were found three most important points (the knowledge of transitions, the importance of using transitions, and the effects of using transitions), and they were supported by student's journal. From the result of interviewing, the students had knowledge about the transition well; they also understood that transitions can help their writing more coherent and more qualified, but they still felt difficult and confused to use some transitions, then they also have some corrections in grammatical when they put transitions in a wrong place, so it can make the readers confused to interpret what they wanted to express their ideas.
Those findings support Yolanda, Jufri, and Fitriawati's research (2013: 366-367) that good students could do the test on using transitions well. Those students got the highest score and the most correct use transitions. Then, average students had more errors than good students on writing sentences by using transitions.
Also, the researcher, in this study, found that most poor students used transitions correctly but some of them used it incorrectly because of the placement of transitions in the sentences, the grammatical, and over-use using transitions in the text that could make readers confused to read their essay. This finding did not support Yolanda, Jufri, and Fitriawati's study (2013: 366-367) that poor students had many errors almost in every part. On the other hand, this first finding supports Letsoela (2013: 110-111) Second, it is found that all students in each category had good knowledge about the transitions, such as its definition, its types, and its function; they also understood that transitions can help their writing more coherent, more qualified, and more understandable. From the interpretation above, the third grade students of English Department, good, average, and poor student, were conscious of transition signals. This is supported by Cenoz and Hornberger's theory (2008: 3) who state that language awareness is concerned with making learners conscious of, and able to use language patterns successfully. In the other hand, they still felt difficult and confused to use some transitions, then they also have some corrections in grammatical when they put transitions in a wrong place, so it can make the readers confused to interpret what writers wanted to say. This finding also supports Badiozzaman and Gorjian's study (2014: 98) that EFL learners often have problem in the proper use of connectives to organize their sentences into a coherent unified whole. This issue was the result of lack of understanding the proper use of transitions in writing essays.
CONCLUSION
In conclusion, first, most of students in every category used transitions correctly even though there were some corrections, such as the placement of transitions, omitting some repetition words or clauses that were wasting, and the grammatical. the third grade students of English Department, good, average, and poor student, were conscious of transition signals, but they need more practice to apply it in proper use in the sentences in order to make their writing had better quality, easier to read and more understandable.
So, the suggestion are given to English teachers, students, and other researchers. First, the teacher should be able to give detail explanation about transition words to the students when teaching writing. Then, the students should be able to find which transitions are difficult and asks them to their teacher or in the internet, so they will not feel difficult to apply them in the sentences. It is better | 2019-08-20T03:47:02.733Z | 2016-11-09T00:00:00.000 | {
"year": 2016,
"sha1": "2c13ec4ff03da65d7db88063b24e29130fe1bc99",
"oa_license": "CCBY",
"oa_url": "http://ojs.unpkediri.ac.id/index.php/inggris/article/download/482/373",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "d5c336a5ddd4ce832d69a0ed26afdda06ddd7e68",
"s2fieldsofstudy": [
"Education"
],
"extfieldsofstudy": [
"Psychology"
]
} |
19600019 | pes2o/s2orc | v3-fos-license | Influence of proximal femur fractures in the autonomy and mortality of elderly patients submitted to osteosynthesis with cephalomedullary nail☆
Objective To determine the autonomy and mortality of elderly patients submitted to proximal femoral osteosynthesis with cephalomedullary nail after hip fracture. Methods Retrospective study with 61 patients with proximal femoral fractures submitted to cephalomedullary nail osteosynthesis. The authors analyzed the medical records and collected information from the preoperative period. Patients were questioned regarding pain, postoperative autonomy, and degree of satisfaction. The total number of deaths was verified. The results were then correlated. Results The mean age was 84 years, predominantly female (82%). In the postoperative evaluation, 45% of the patients presented worsened levels of autonomy. The majority of patients presented mild pain (61%) on the VAS scale. The mortality rate was 24.6%, and the mean time of preoperative hospitalization was three days. The factors that presented statistical significance regarding postoperative autonomy were the time elapsed from the trauma until the moment of surgery, ASA score, fracture stability, and previous functional status of the patients. The mortality rate was associated with three main factors: advanced age, ASA score, and preoperative hospitalization time. Conclusion The patient's previous autonomy positively influenced the functional outcome and postoperative recovery. Unstable fractures presented worse results for pain and ambulation in a follow-up of 27 months. Hip fracture is a risk factor associated with mortality and decreased independence in patients over 65 years of age.
Introduction
The incidence of proximal femoral fractures has increased significantly in the last decades and is expected to double within the next 25 years, due to increasing life expectancy. Older age and associated comorbidities are responsible for high morbidity and mortality, as well as higher costs of treatment and hospitalization. 1 Among proximal femoral fractures, the transtrochanteric fracture is an important group, since they are common and usually affect debilitated and elderly patients. 2,3 These fractures are defined as those occurring in the area extending from the extracapsular region of the base of the femoral neck to a proximal region along the minor trochanter. 4 This type of fracture is common in the elderly population due to osteoporosis, and is mainly associated with low-energy trauma, such as fall from their own height. It is estimated that nine out of ten trochanteric fractures occur in individuals older than 65 years. 5,6 Hip fracture is a devastating injury in elderly patients, affecting their physical, mental, functional, and social function. It reflects the aging process of the population, and has permanent consequences on patient survival and independence.
Surgical stabilization is the option of choice to manage trochanteric fractures, regardless of deviation type or pattern. The goal of surgical treatment is to achieve stable reduction and fixation, allowing early active and passive mobilization. Since elderly individuals are not always able to walk without placing some load on the fractured limb, due to the preexisting conditions, fracture stabilization must be sufficient to allow a certain load tolerated by the patient. 7 Studies demonstrate that up to 50% of patients with proximal femoral fractures die within the first six months after trauma, and that many patients submitted to surgical treatment do not regain their baseline function and independence. 8 This study aimed to determine the autonomy and mortality of elderly patients who underwent proximal femoral osteosynthesis with cephalomedullary nail after hip fracture.
Methods
A longitudinal, retrospective, and descriptive study was performed with 61 patients with proximal femoral fracture who underwent osteosynthesis with cephalomedullary nail from March 2012 to March 2014, at the Hospital de Ortopedia e Traumatologia, Rio de Janeiro, Tijutrauma.
Patients' medical records were analyzed and information from the pre-operative clinical records (clinical evolution and surgical risk) was collected.
The following data were collected: age; sex; preoperative time; fracture severity (according to the Tronzo I to V classification); preoperative American Society of Anesthesiologists (ASA) score; and pre-trauma autonomy, divided into five walking patterns: (1) Walks outside home community distances without help; (2) Walks outside home community distances with help; (3) Walks household distances without help; (4) Walks household distances with help; (5) Non-ambulatory or wheelchair user. Patients with incomplete medical charts were excluded from the sample.
After collection of complete preoperative data, patients were interviewed and underwent clinical evaluation to assess pain through a visual analog scale (VAS), postoperative walking status, and degree of satisfaction. At follow-up, the total number of deaths after surgical treatment was used to determine the mortality rate. Patients who did not attend the interview or who were lost follow-up were excluded.
Data collected in the pre-and postoperative periods were then correlated with mortality and with influence on the patients' autonomy.
For statistical analysis of the results, tables with absolute and relative frequency distribution were used. The associations were tested by Pearson's chi-squared test or Fisher's exact test. Normality was assessed by the Shapiro-Wilk test, and continuous variables were assessed by Student's t-test or ANOVA.
Results
The mean age of the patients was 84 years (±7.2), ranging from 67 to 97. The male group presented a slightly higher mean age than the female group (86 [±8.0] years vs. 83 [±7.0], respectively, without statistical significance [p = 0.260]). Laterality of the affected hip was similar (50.8% and 49.2%, respectively).
A significant difference was observed in the prevalence between sex (p < 0.001); as most cases (82.0%) were female, that is, on average, surgical procedures for proximal femoral fracture were performed 4.5 times more often in women than in men (Fig. 1).
Mean hospitalization time was 9 days, ranging from 2 to 27. Mean hospitalization time before surgery was 3 days, ran- ging from 1 to 14. Patients classified as Tronzo type V had the highest mean pre-operative hospitalization: 7 days (p = 0.0047) ( Table 2 and Fig. 2).
In the preoperative assessment, 21 patients were able to walk outside home for community distances without help. After surgery, 15 patients (71.4%) maintained this condition and 33.3% worsened their walking status. Of the patients who were able to walk outside their home for community distances with help, six (50%) maintained their condition and 50% worsened. Of the 10 patients who were able to walk for household distances without help, 44.4% maintained their condition and 55.6% worsened. Of the patients who were able to walk for household distances with help, 60.0% maintained their condition and 40.0% worsened. Therefore, when comparing the pre-and post-operative general ambulation status, 45% of the patients presented a decrease in their level of functional independence ( Table 3). The VAS scale was applied after surgery and most patients (60.9%) presented mild pain while 37.0% presented moderate pain. Only one patient (2.2%) reported severe pain. When asked about the degree of satisfaction, 90% of the patients were satisfied with the result of the surgery; the remaining 10% were dissatisfied with not being able to walk or with living in nursing homes, as they did not regain their previous independence.
Patients with a more severe Tronzo score had a higher mean VAS, but the difference was not statistically significant (p = 0.057).
Comparing pain (VAS) with preoperative hospitalization time and total hospitalization time, a significant correlation was observed in both situations: longer preoperative hospital stay was associated with greater pain reported on the VAS scale (p = 0.016), as was total hospitalization time (p = 0.026; Fig. 3). Fifty-eight patients (95.1%) had at least one preoperative comorbidity. Only three patients (4.9%) had no pre-existing disease.
The mortality rate found in the study was 24.6% (15 deaths). Three factors were significantly associated with higher mortality in these patients: advanced age, ASA score, and preoperative hospitalization time (Table 4). Mean age of the patients who died was 87.6 years, versus 82.5 of those patients who were alive (p = 0.016). Patients classified as ASA III or IV presented a higher mortality rate than those classified as ASA I or II: 35.5% versus 13.3%, respectively. Regarding the preoperative hospitalization time, the mortality rate of patients who were operated on within 48 h of the fracture was 14.3%. In patients who were operated on after over 48 h of the fracture, the mortality rate increased to 38.5%. Factors such as Tronzo classification, gender, number of associated diseases, and total hospitalization time were not associated with mortality.
Discussion
Transtrochanteric femoral fractures are the most frequently operated fracture type, and present the highest postoperative mortality rate; they have become a public healthcare issue, due to the high cost of treatment and the difficulty of recovering postoperative functional independence. 4 The epidemiological profile of the individuals in the present sample was not much different from those observed in Brazilian and international studies. In the present study, a predominance of females was observed, at a 4.5:1 ratio; their mean age was 84 years. Comparing with Brazilian studies, Hungria Neto et al. 2 found a predominance of women in the proportion of 2:1 and a mean age of 78.2 years; Ramalho et al. 9 observed a ratio of 3.3:1 and a mean age of 78.5 years.
In the present study, we observed that the degree of autonomy decreased after surgical treatment. 8 Nonetheless, 55% of the patients recovered their functional level. A similar finding was observed by Herrera et al. 10 in their work with 250 patients treated with cephalomedullary nail, in which approximately 50% recovered their previous walking status after one year.
Patients' previous autonomy positively influenced the functional outcome and the postoperative recovery. We observed that 71.4% of patients who were able to walk outside home for community distances without help maintained their level after surgery. This demonstrates that more physically independent patients present better results. Zuckerman et al. 11,12 observed that the group of patients with the best pre-fracture score recovered almost 100% of their functional capacity.
Regarding pain intensity after surgical treatment of transtrochanteric fractures, most operated patients (60%) had mild pain; similar data were retrieved in the Brazilian literature. 13 The mortality rate in the present study was 24.6%, at a mean follow-up of 27 months. In a study with 1448 patients, Rosso et al. 14 found a mortality rate of 18.8% in one year of follow-up, while Dousa et al., 15 in a study with 4280 patients, observed a 30% death rate.
Cooper et al. 16 stated that the consequences of a hip fracture include premature death (around 20% in one year), loss of function (in 30%), inability to walk independently (40%), and loss of at least one independent daily activity (in 80% of the patients treated).
Some factors were statistically significant for mortality, such as time elapsed from trauma to surgery, ASA score, fracture stability, and age at the time of trauma.
In the present study, patients with preoperative hospitalization time greater than 48 h had a 38.5% mortality rate, while patients who underwent surgical treatment in up to 48 h had a lower rate, 14.3%. Regarding the time of surgical delay, Hamlet et al. 17 demonstrated that patients operated within the first 24 h of admission had lower mortality than those operated after 24 h, regardless of the preoperative ASA score. The higher mortality rate associated with surgical delay demonstrates the importance of the early treatment of these fractures, which are now treated as orthopedic emergencies. Some factors undermine the improvement of this treatment, such as overcrowding of public hospitals, bureaucracy for surgical authorization, availability of the necessary materials, and availability of medical staff.
Another factor that was statistically significant for mortality was the ASA score. Only four deaths (6.5%) were observed in patients classified as ASA I and II, versus 11 (18%) in patients classified as ASA III and IV. Michel et al. 18 demonstrated that patients classified as ASA III or IV had a nearly nine-fold greater risk of death in the first postoperative year than those classified as ASA I or II (p < 0.001).
Conclusion
Patient's previous autonomy positively influenced functional outcome and postoperative recovery.
In a follow-up of 27 months, unstable fractures presented worse results for pain and walking status.
Hip fracture is a risk factor associated with mortality and decreased autonomy in patients aged over 65 years. These fractures are devastating injuries that should be treated as orthopedic emergencies. | 2018-04-03T03:47:18.369Z | 2017-08-26T00:00:00.000 | {
"year": 2017,
"sha1": "a7ea70fb7ad8c284b405be146927fda73717ea70",
"oa_license": "CCBYNCND",
"oa_url": "https://doi.org/10.1016/j.rboe.2017.08.014",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "a7ea70fb7ad8c284b405be146927fda73717ea70",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
232314642 | pes2o/s2orc | v3-fos-license | A Method for Detection of Small Moving Objects in UAV Videos
: Detection of small moving objects is an important research area with applications including monitoring of flying insects, studying their foraging behavior, using insect pollinators to monitor flowering and pollination of crops, surveillance of honeybee colonies, and tracking movement of honeybees. However, due to the lack of distinctive shape and textural details on small objects, direct application of modern object detection methods based on convolutional neural networks (CNNs) shows considerably lower performance. In this paper we propose a method for the detection of small moving objects in videos recorded using unmanned aerial vehicles equipped with standard video cameras. The main steps of the proposed method are video stabilization, background estimation and subtraction, frame segmentation using a CNN, and thresholding the segmented frame. However, for training a CNN it is required that a large labeled dataset is available. Manual labelling of small moving objects in videos is very difficult and time consuming, and such labeled datasets do not exist at the moment. To circumvent this problem, we propose training a CNN using synthetic videos generated by adding small blob-like objects to video sequences with real-world backgrounds. The experimental results on detection of flying honeybees show that by using a combination of classical computer vision techniques and CNNs, as well as synthetic training sets, the proposed approach overcomes the problems associated with direct application of CNNs to the given problem and achieves an average F1-score of 0.86 in tests on real-world videos.
Introduction
Convolutional neural networks (CNNs) have improved state of the art results on tasks of object detection in images and videos [1,2]. However, the majority of these algorithms are oriented towards detection of objects that are large compared to the size of a frame and have distinctive visual features that can be used for learning discriminative object representations. However, when target objects are small, e.g., less than 10 × 10 pixels, the obtained results are considerably worse [3]. The main reason for this discrepancy is the lack of distinctive shape and texture on small objects. It precludes learning useful representations of small objects resulting in worse detection performance. Furthermore, general purpose object detectors are trained to predict bounding boxes of objects in image or video, while for small objects only the object coordinates, (x, y), are required in most applications.
Small object detection is an important task with applications that include: surveillance of insects or small animals, detection of small and/or distant objects in search and track systems, sense and avoid functionality in unmanned aerial vehicles (UAVs), traffic monitoring, and detection of dangerous or unusual behavior in overhead imagery and videos etc. In this paper we propose a method for the detection of small moving objects in videos recorded using UAVs and demonstrate its effectiveness for flying honeybee detection. Possible applications of the proposed method include monitoring of flying insects, studying their foraging behavior, using insect pollinators to monitor flowering and pollination of crops, surveillance of honeybee colonies, and tracking the movement of honeybees for various applications, such as humanitarian demining. [4,5]. While in this paper we primarily present experiments with honeybees, the proposed method is applicable to other types of flying insects and small blob-like moving objects in general.
There are a number of challenges associated with building an effective system for detection of small moving objects. For example, in the case of honeybees, a detection method must be non-invasive, meaning that it must not interfere with honeybees flying during recording. Therefore, UAVs must be flown at higher altitudes. As a result, honeybees in recorded video sequences will be very small, with a blob-like shape, and without a readily noticeable texture, as shown in Figure 1a. In addition, the appearance of flying honeybees may change during the sequence due to flapping of wings, shadows, camera gain variations etc. Furthermore, honeybees are fast targets and individual honeybees may appear in only one or two consecutive frames. Finally, backgrounds in video sequences recorded from UAVs in natural conditions usually contain grass or other vegetation moving due to wind or air flow produced by an UAV. Consequently, the foreground/background contrast ratio in frames will be low and there will be motion that does not originate from flying honeybees. As a result, it is hard, even for human observers, to spot honeybees in recorded videos, and almost impossible in still images. However, as can be seen by comparing Figure 1a,b, it is possible to notice the change of appearance at a particular location due to honeybee motion. Similar problems could be identified in other applications involving detection of small moving objects. Based on this, we decided to work with videos and detection of small moving objects based on a fusion of information about their appearance and motion. Although many algorithms for moving object detection exist [6], due to the presence of motion in the background, their direct application is limited and results in a large number of false positive detections. To filter out these false positive detections, we propose using a CNN trained on groups of consecutive frames, which learns a representation of appearance and motion of small objects and outputs confidence maps of presence of moving objects in the middle frame for each group of frames given as an input to it.
Since it is difficult for human observers to detect small moving objects in recorded videos, manual labelling of a large number of videos, necessary to train a detector based on a CNN, is a difficult and error-prone task. Because of that, there are no available training datasets that can be used for this purpose. To solve this problem, we generated a synthetic training dataset, with video sequences containing real-world backgrounds and artificially generated small moving objects.
We use synthetic video sequences to train a CNN model and perform a series of experiments with both synthetic and real-world video sequences. The goal of the experiments with synthetic sequences is to investigate the impact of the parameters used for generation of training data, as well as other design choices on performance of the detector. Most notably, we vary the contrast ratio between artificial objects and background to find the optimal contrast ratio in the training data that will give good results on a range of contrast ratio values in the test sequences. At the task of detection of flying honeybees in real-world videos our CNN-based method, trained on synthetic videos, achieves the average F1-score of 0.86. To the best of our knowledge, this is the first method to use synthetic data for training a small moving object detector and also the first method for detection of small moving objects applied to the detection of flying honeybees in videos recorded using UAVs.
The main contributions of this paper are: (1) An approach that effectively uses both appearance and motion information to detect small moving objects in videos captured using UAVs. (2) Usage of synthetic data for training a CNN-based detector. (3) Evaluation of the impact of the parameters used for generating synthetic training videos, as well as other design choices on performance of the detector. (4) Investigation whether the detector performance on synthetic data can be used as a proxy for performance on real-world video sequences.
This paper is organized in the following way. In Section 2 we review the related work. Section 3 contains descriptions of the datasets used for training and testing the approach. In Section 4, a detailed description of the method for detection of small moving objects is presented. The experimental results are presented in Section 5. In Section 6 we discuss the obtained experimental results. Finally, Section 7 concludes the paper. The data used in this study is publicly available at https://doi.org/10.5281/zenodo.4400650. The code used for the experiments is publicly available at https://github.com/vladan-stojnic/Detection-of-Small-Flying-Objects-in-UAV-Videos.
Related Work
In general purpose object detection, in the past years the state of the art results have been obtained using approaches based on CNNs. The most well known approaches include region proposal based R-CNN [2,[7][8][9] and YOLO [1,[10][11][12] families of models. However, these object detection methods are oriented towards larger objects and learn models based on their shape and appearance, which are not noticeable in small objects. A modification of region proposal based CNN for small object detection is presented in [13] but it does not use motion information.
In detection of objects in videos, motion is an important cue that can be used for discriminating objects from the background. Moving objects detection, also known as foreground detection or background subtraction is a classical computer vision task with a rich history. For comprehensive reviews of the methods, interested readers are refered to [6,14], and [15]. In addition, in [16] background subtraction in various real-world applications is reviewed. Special attention is devoted to applications in intelligent visual observation of animals and insects, most notably honeybee surveillance.
Taking into account the absence of distinguishing visual features in very small objects, in [3] both appearance and motion information are used to improve state of the art in object detection in wide area motion imagery. Similarly, in [17] a deep learning based approach for joint detection and tracking of small flying objects is proposed. These methods also use CNNs for detection, making them similar to our approach. However, for training they use manually labeled real-world video frames, which are not easily obtained in our case. Noting that visual systems of insects are naturally evolved to perceive small moving targets, some papers investigate biologically inspired approaches to small moving target detection in cluttered environments, [18][19][20][21][22].
The problem of detection of small targets is very important in infrared search and track systems. Consequently, there is a considerable body of work dealing with detection of small targets in infrared videos, such as [23][24][25][26][27][28]. Unfortunately, comparison of these algorithms is difficult because, due to security restrictions, there are no publicly available datasets of infrared videos featuring small targets.
In the past years, monitoring honeybees at the hive entrance [29][30][31][32][33][34][35][36], as well as in a hive [37,38], has received considerable attention. Although closely related to the given problem, these approaches are not applicable in our case, because honeybees in the videos captured at hive entrance or inside a hive are larger and have noticeable color, texture, and shape features.
Generally, the literature on detection of flying honeybees is rather scarce, but some experiments, which are usually goal-oriented and where honeybees are used as detectors of some property, could be found. Honeybees have a very developed olfactory system and are able to recognize and detect scents from a large distance. With their ability of spatial orientation, using the Sun as a light source, honeybees proved to be very good detectors for sources of scents of interest. Because of that, researchers in [39][40][41] used detection of flying honeybees in the context of locating land mines, by performing detection of their movement on recordings obtained using a LiDAR. Application of short-time Fourier transform to pixel intensities in high-frame rate video for honeybee activity sensing is proposed in [42]. The main drawback of these methods is their reliance on special imaging techniques, namely LiDAR and high frame rate video. In contrast, the method proposed in this paper is based on videos captured using imaging in the visible part of the spectrum and common frame rates.
A system for detection and tracking of individual animals in videos recorded without using special imaging techniques, named idtracker.ai, is presented in [43]. However, it uses CNNs trained on videos captured in laboratory conditions, with uniform background and good contrast between targets and background, which is not the case in our usage scenario. Visual tracking of individual small animals in natural environments using a freely moving camera is presented in [44]. More similar to our work, the approach for honeybee tracking proposed in [45] uses RGB videos recorded using a fixed-position camera in less controlled conditions, with tree foliage in the background. For moving object detection, frame differencing and background modelling using a recursively updated model are used. However, only qualitative experimental results are presented. As already mentioned, traditional moving object detection methods, such as frame differencing and background subtraction, result in many false positive detections which need to be filtered out, so an important part of our work is devoted to solving this problem.
Training Data
To circumvent the lack of labeled data suitable for training the detector, we generate synthetic training data with backgrounds from real-world videos captured using UAVmounted camera to which we add artificially generated blob-like objects. For the purpose of the experiments in this paper, the movement of the artificially generated objects was derived, under the supervision of an expert, from manually selected and traced honeybees in UAV videos, with small random variations. The resulting objects, named "artificial honeybees" were created in a way to mimic the appearance and flight patterns of honeybees in search for food near the known location of food sources. It should be noted that the described method for generating synthetic training data can be easily adapted to different type of target objects and different real-world backgrounds.
To make the method invariant to background appearance, we captured 3 videos at locations in Croatia. For capturing the videos used as backgrounds, as well as for testing the proposed method, we flew UAVs at altitudes between 7 and 15 meters. We used two very different UAVs, one quadcopter, DJI Inspire 1, and one large, proprietary-built, hexacopter equipped with high-accuracy positioning system: Real-Time Kinematic (RTK) GPS. RTK system, Figure 2, allows a very precise hovering and provides a better output after automatic video-stabilization process. A limitation of this system is its requirement for a base station in the relative vicinity of the rover station, in our case UAV, but for covering a small area this limitation did not cause us problems. DJI Inspire 1 was equipped with Zenmuse 5R digital camera, allowing recording of 4K uncompressed videos with the possibility of using interchangeable lenses. For this purpose we used a 25 mm (equivalent to 50 mm in 35 mm systems) lens. The hexacopter was equipped with a modified GoPro Hero 6 camera with a 47 mm equivalent lens. These two setups provide similar recording performance with the usage of different aerial vehicles to eliminate any equipment-bias and provide different conditions for reproducibility of the experiment. Different lighting conditions were eliminated using automatic camera settings and did not impose problems in automatic processing of the recorded videos. Because of the nature of experiments, we needed to have rather good atmospheric conditions with dry weather and almost no wind, because honeybees avoid foraging in unfavorable weather conditions. All recorded videos have 4K resolution with frame rate of 25 fps. The choice of the parameters for recording real-world videos, were chosen in such a way as to strike a balance between the ability of human observers to detect flying objects in the recorded sequences and not interfering with flying honeybees. Since honeybees are small targets, it is desirable to fly at lower altitudes to obtain as much information about their appearance as possible. However, flying at low altitudes results in a considerable amount of wind produced by the rotors of UAVs, which could interfere with flying honeybees, as well as create moving artifacts from, for example, grass or other vegetation.
Furthermore, flying at low altitudes results in covering only a small part of the surface area in a frame, thereby reducing the ability of both human observers and the system to detect a flying honeybee at several locations in the frame and use its motion as a cue for detection.
From the available recordings we selected only parts without visible honeybees or other insects in order to obtain representative examples of backgrounds. However, since it is very hard to detect small flying objects in videos, it is possible that some residual flying insects exist in several frames of the selected videos. Nevertheless, we do not expect that a small number of residual insects will negatively impact the performance of the detector, since CNNs can tolerate a certain amount of labeling noise [46].
In order to remove global camera motion, i.e., to stabilize the video sequence, we fit the affine transform between each frame in the sequence and the first frame, and then warp all frames into a common reference frame. For estimation of the affine transform between two frames, we first detect keypoints in both frames and compute their descriptors using ORB detector and descriptor [47]. Then, we find matching pairs of keypoints by comparing the descriptors using Hamming distance. Finally, the matches are used for robust transform estimation using RANSAC algorithm.
We crop the frames of the stabilized videos into blocks of 1024 × 1024 pixels with 200 pixels overlap between successive blocks, and we skip 200 pixels from each side to eliminate border effects caused by stabilization. After this procedure we are left with 96 background video sequences with frames of 1024 × 1024 pixels in size. Each background sequence is 3 s long. Some examples of frames from the background sequences are shown in Figure 3. In the next step we add artificial honeybees to the obtained background sequences. Examining the appearance and behavior of real-world honeybees in videos captured using the same setup as described for the background sequences, we decided to represent artificial honeybees as elliptical blobs modelled using 2D Gaussian kernels with standard deviations randomly chosen from the intervals [2, 4.5] and [1, 3.5] for x and y axes, respectively. The number, sizes, initial locations, initial flight directions, and initial velocities of the artificial honeybees are also chosen randomly by sampling from uniform distributions with minimum and maximum values given in Table 1. The texture of artificial honeybees is modelled using Gaussian noise. We create several datasets with varying means of Gaussian noise (texture means) to assess the impact of this hyperparameter on overall detection accuracy. Specific values are given in Section 5 and discussed in the context of the obtained detection results. We use the same value of 0.07 for standard deviation of Gaussian noise in all datasets. In each frame new velocity v t and flight direction θ t of each honeybee from the previous frame are calculated as where ∆ v and ∆ θ are sampled from normal distributions with zero mean and standard deviations 2 and 30, respectively, and v t−1 and θ t−1 are honeybee velocity and direction in the previous frame. New positions of honeybees are then calculated using projections of their velocities onto x and y axes. If the new position of a honeybee is outside of the visible part of the scene we do not add it to the frame. To simulate honeybees flying out of and returning to the visible part of the scene we keep track of the positions of those honeybees but do not add them to the frame. When adding artificial honeybees to a frame we use their pixel values as alpha channel for blending between the background and black pixels. Therefore, artificial honeybees with lower values of the texture mean will appear lighter, i.e., will have low contrast ratio compared to the background, while artificial honeybees with higher values of texture mean will have high contrast ratio compared to the background. Simultaneously, with generating frames for the training sequence, we generate the ground truth frames that will be used as training targets. Ground truth frames are grayscale frames with black background and artificial honeybees added in the same locations as in the training frames. In this case, we add artificial honeybees by setting the pixel values of the ground truth frame in the neighborhood of the honeybee location to the pixel values of an artificial honeybee. In total, 1000 sequences with frames of 1024 × 1024 pixels containing artificial honeybees are created using the described procedure. Of those, we use 500 sequences for training, 250 sequences for validation, and we retain 250 sequences for testing. We train the network by feeding it with sequences consisting of 5 consecutive frames of 256 × 256 pixels in size cropped randomly from the synthetic training sequences. For each of these sequences the network is trained to predict the ground truth frame corresponding to the middle frame of the sequence. Since the number of honeybees in a single sequence is relatively small, a majority of cropped frames will contain a small number of honeybees or no honeybees at all. Therefore, including the cropped sequences into the training set with uniform probability would result in pronouncedly imbalanced numbers of honeybees present in each sequence. Bearing in mind difficulties of training the network with an imbalanced training set, we decided to include the cropped sequences into the training set with probabilities proportional to the number of honeybees present in the cropped part of the frame. In this way, sequences with a large number of honeybees, although sampled less frequently, will be more often included into the training set. In contrast, more frequently sampled sequences with little or no honeybees will be less frequently included into the training set. By sampling from the cropped sequences in this fashion, we obtain a training set with 53,760, a validation set with 12,800, and a test set with 12,800 samples.
Test Data
Besides testing on synthetic videos, we also evaluated the proposed method on realworld videos captured using the same setup as described previously for capturing the background sequences. We placed six hives near the examined area so the expected number of honeybees was significantly larger than the number of other flying insects of similar dimensions, the grass was cut and there were no flowers attractive to other pollinators. In addition, during the recording we monitored the area and did not notice significant presence of other flying insects and, in the labeling phase, we used knowledge about honeybee flying patterns. Therefore, we can assume that the flying objects in the test sequences are honeybees.
To quantitatively assess the performance of the proposed method on real-world videos, we developed a tool for manual labeling of small moving objects in video sequences. The developed tool enables users to move forward and backward through the video frame by frame, and mark locations of target objects in each frame. Since objects of interest are very small in comparison to the frame size, and it is of interest only to detect whether an object is present at a particular location, its spatial extent is disregarded. Therefore, bounding boxes or outlines of target objects are not recorded and only their locations in each frame are saved and used as ground truth in performance assessment.
For testing, we extracted three sequences with durations of around 3 seconds from the available recordings, performed their stabilization, and cropped a part of the frame of 512 × 512 pixels. The cropped regions were selected on basis of honeybee appearances, i.e., we cropped those regions where it was possible for human observers to spot honeybees. More specifically, during manual labeling of honeybees in these sequences, we noted that it is hard for human observers to equally focus on all parts of a large frame, especially with small target objects. This led us to choose the size of the cropped region in such a way as to strike a balance between the desire to use as large region as possible in order to obtain more information about the behavior of honeybees, and the human ability to analyze large frames. We manually labeled all honeybees in these sequences and used the obtained annotations to evaluate the performance of the trained detectors. To conclude, the labeled test sequences contain frames with one to 4 honeybees, as well as frames without honeybees.
Method for Detection of Small Moving Objects
The main steps of the proposed method are shown in Figure 4 and include video stabilization, background estimation and subtraction, frame segmentation using the CNN, and thresholding the output of the CNN. As we already discussed video stabilization in Section 3, here we present the subsequent processing steps.
Background Estimation and Subtraction
In order to emphasize moving objects in each frame, we first estimate means and standard deviations of the pixel values in a temporal window of previous frames. The pixel-wise mean of the frames in a temporal window can be regarded as background estimation, since small moving objects are filtered out by time averaging the frames in the window. In this step, we essentially fit a Gaussian probability distribution function, characterized by its mean and standard deviation, to the values of each pixel in a window of previous frames. Let I(x, y, t) be the frame at time instant t and N the number of frames in the temporal window. We obtain the pixel-wise mean, i.e., background estimation, as and standard deviation as We subtract the estimated pixel-wise mean from each frame and divide the result with the estimated pixel-wise standard deviation After this step, moving objects in the resulting frames will have larger pixel values than the background, as shown in Figure 4. By subtracting the mean and dividing with standard deviation we obtain a measure of dissimilarity of the current pixel value from the mean of the Gaussian normalized by its width, i.e., standard deviation. It is expected that the values of stationary background pixels will be closer to the mean compared to the values of pixels belonging to moving objects so the differences between moving objects and background estimation will be large. This procedure is usually referred to as background subtraction. Thresholding the obtained differences was previously proposed for foreground detection in video [48]. However, as discussed before, frames can contain moving artifacts, such as grass moving due to wind. An example of the result of background subtraction from a frame is shown in Figure 5a. We can see that a simple thresholding of this frame would result in too many false positive detections, which is the reason why we feed the preprocessed frames into a CNN and train it to segment moving objects. As it can be seen in Figure 5b our CNN segmentation model is able to detect a small moving object and remove the unwanted noise, such as grass moving due to wind.
CNN Topology
The frame segmentation CNN performs segmentation of the middle frame of a sequence of 5 consecutive frames into moving objects and background. The CNN topology used in this paper is given in Figure 6, and the hyperparameters of layers are given in Table 2. Inspired by the U-Net CNN topology with skip connections, proposed for medical image segmentation in [49], we chose fully convolutional CNN, which made possible using input images of different sizes. The used CNN model can be divided into two parts: the encoder responsible for learning a representation of input data, and the decoder used for obtaining the output segmentation map of the desired size based on the representation learned by the encoder. The encoder consists of 3 blocks of convolutional layers with 3 × 3 kernels, batch normalization layers, and ReLU activations. Each block is followed by a max-pooling layer. The last layer of the encoder is a 1 × 1 convolution layer, also followed by batch normalization layer and ReLU activation. The output of the encoder is used as input of the decoder. In the decoder the obtained feature maps are upsampled using nearest neighbor interpolation and then fed to the convolutional layer with 3 × 3 kernel, batch normalization layer and ReLU activation. Identical blocks with convolutional layer, batch normalization layer and ReLU activation are repeated two more times. Given that the target objects are small, it is important to make sure that fine details, present in input frames, are used in segmentation. To achieve this, we use symmetric skip connections between outputs of convolutional layers in the encoder and the decoder with feature maps of the same size, as shown in Figure 6. The final layer of the decoder is a convolutional layer with only one 1 × 1 kernel and sigmoid activation. The obtained segmentation map is two times smaller than the original frame. However, we decided not to add another block with upsampling and convolutional layer, because that would increase the number of trainable parameters and computational complexity. Instead, we just upsample the obtained segmentation map using bilinear interpolation. The experimental results show that this simplification does not negatively impact the results. Finally, to obtain an object detection map, we threshold output of the CNN.
CNN Training
During the training we optimize L2 loss between the model outputs and training targets. Training targets are grayscale frames with synthetic honeybees at same locations as in the training frames but with uniform black background. The CNN is fed with 5 consecutive frames and trained to segment the honeybees in the middle frame.
For optimization we use Adam [50] with hyperparameters given in Table 3 and learning rate reduction with a factor of 5 after each 20 epochs. The training is terminated if validation loss has not improved for 10 consecutive epochs.
Experimental Results
In order to evaluate the proposed method on the task of detection of flying honeybees, we perform two experiments. In the first experiment, we evaluate our trained model on synthetic videos with honeybees whose texture is modelled using Gaussian noise with different means, as described in Section 3. For the second experiment, we used real-world sequences with manually labeled locations of honeybees.
In both experiments we trained one CNN for each dataset with a specific honeybee mean texture value. We used mean texture values from the set {0.25, 0.5, 0.75, 1.0}, as well as combined mean texture values of 0.25 and 0.5, and randomly chosen mean texture values from the interval [0. 25, 0.5]. In this way we obtained 6 different frame segmentation CNN models.
Since we are interested only in detections of honeybees, we threshold the CNN output and compute the centroids of the resulting connected components. These centroids are considered as locations of detected honeybees. To evaluate the performance of the detector, we compare these detections with ground truth honeybee positions. If the distance between the detection and labeled position is less than 10 pixels, it is considered that the honeybee is correctly detected. We chose 10 pixels based on the average size of a honeybee and to introduce a degree of tolerance to imprecise human annotations. The performance of the detector is expressed in terms of recall and precision where TP is the number of true positive detections, FN is the number of false negative detections, and FP is the number of false positive detections, aggregated from all frames in a sequence. We also compute the F1-score as
Testing on Synthetic Videos
In the first experiment, we test the trained CNNs on synthetic test videos with varying mean values of honeybee texture. By varying the texture mean in the test set, we vary the contrast ratio between the moving objects and background thus making the detection easier or harder. The goal was to create a controlled environment which would enable us to examine the influence of the artificial honeybee model in training data to detection accuracy when different honeybee models are used for testing, find the best honeybee model for generating training data, and find the test honeybee model which can serve as a good proxy for tests with real-world honeybee videos. To find out whether the proposed detection algorithm benefits from background subtraction, we train and test CNNs, both without and with this step, and compare the obtained results.
In order to find the best value of detection threshold applied to the output of the frame segmentation CNN, we evaluate the performance of the detector on synthetic video sequences for different values of the threshold. The training and testing sequences both contain artificial honeybees generated with texture mean of 0.25. By varying the threshold value, the values of recall and precision vary, resulting in the Precision-Recall curve given in Figure 7. For threshold values above 0.6, no honeybees are detected so recall is zero and precision is not defined. Based on this curve, for subsequent experiments we select the threshold value of 0.1. The obtained experimental results, when synthetic bees are used for both training and testing, are shown in Tables 4 and 5, for the cases without and with background subtraction, respectively. We can see that, for detectors trained on synthetic videos with a single texture mean, overall F1-scores are higher when background subtraction is used, irrespective of the texture mean used for testing. Moreover, when background subtraction is not used, the detector performance deteriorates in cases when texture means of the training and test honeybees differ significantly. This deterioration is somewhat less pronounced when background subtraction is used. Overall, when a single texture mean value is used for creating training sequences, the best results are obtained when it is set to the smallest value of 0.25. In these sequences the contrast ratio between the moving objects and background is low, which enables the frame segmentation CNN to successfully segment out both low and high contrast moving objects. When the contrast ratio between the moving objects and background in the training sequences is high, the frame segmentation CNN cannot segment low contrast objects, resulting in lower detection rates. When combinations of texture means are used for training, the obtained results improve when no background subtraction is applied but are mostly unchanged in the other case. Furthermore, compared to the case without background subtraction, we can see that the performances are similar or even slightly better when background subtraction is not used. However, the resulting difference in performance is very small and may very well be a consequence of stochasticity in training the CNN. We may conclude that, when testing on synthetic videos, both background subtraction and combinations of texture means are effective in reducing the dependence of the detector performance on honeybee texture mean value used in training sequences.
Testing on Real-World Videos
In the second experiment we tested all the trained models using real-world videos with manually annotated flying honeybees. The obtained results are given in Table 6, for the case without background subtraction, and in Table 7, when background subtraction was used. Similarly to the tests with synthetic video sequences, we can see that, when a single value of texture mean is used for modelling honeybees in the training sequences, the overall results are better when background subtraction is used. Furthermore, we can see that, again, the best results are obtained when the texture mean of 0.25 is used for training. Increasing the texture mean results in decreasing recall and increasing precision of the detector. A possible explanation is that the network trained on high contrast artificial honeybees is not able to detect low contrast honeybees, resulting in more false negative and less false positive detections, i.e., higher precision and lower recall. However, training the frame segmentation CNN using low contrast synthetic honeybees results in higher recall, indicating that more honeybees are detected, without a significant decrease of precision. We conclude that, by using the frame segmentation CNN, we succeeded in reducing the number of false positive detections, which is one of the main drawbacks of the classical moving object detection methods based on background subtraction.
Discussion
The results on real-world sequences, shown in Tables 6 and 7, indicate that it is always beneficial to use background subtraction as a preprocessing step. From the results in Tables 4 and 5, we can see that the tests on synthetic sequences benefit from background subtraction only in cases when a single contrast value is used in training sequences. Nevertheless, given that the differences in performances on synthetic test sequences are small, and probably caused by the stochasticity of training a CNN, we may conclude that background subtraction is a useful preprocessing step resulting in improved detection performance.
The results on both synthetic and real-world sequences show that, with background subtraction, the best results are obtained when low-contrast artificial honeybees with mean texture value of 0.25 are used. Therefore, we may conclude that low-contrast artificial honeybees better model the appearance of real-world honeybees than those with higher contrast values and are, thus, better suited for training the frame segmentation CNN. This conclusion is supported by visual inspection of the enlarged portion of the frame containing a real-world honeybee, shown in the top row of Figure 8a. We can see that the contrast ratio between the real-world honeybee and background is very low. Although honeybees have vivid colors in UAV videos they appear featureless and with low contrast because of the large distance from the camera compared to the size of a honeybee and motion blur originating from their quick movements.
We expected that using combinations of texture means would act as a form of training set augmentation and result in better detection performance. Surprisingly, when training sequences contain honeybees with combinations of texture means, the results on real-world sequences are worse than when a single texture mean is used, while when synthetic test sequences are used, the performance of the detector stays unchanged. Since we trained the segmentation CNN and tuned the hyperparameters of our detector on synthetic sequences, this gap between the performances on synthetic and real-world test sequences indicates that the detector has overfit to the training data, resulting in lower real-world detection performance. A possible explanation is that the used model of a honeybee has shortcomings and does not capture all variations that can arise in the appearance of real-world honeybees. We plan to investigate this finding in more detail in future work.
Concerning the question whether tests on synthetic video sequences can be used as a proxy for performance on real-world sequences, we observed that, when synthetic video sequences are used for testing, tests on sequences with a single contrast value are not a good proxy for performance on real-world sequences. However, the average performance obtained using test sequences with different contrast values are better correlated with results on real-world video sequences.
In Figure 8, enlarged portions of frames with true positive, false positive, and false negative detections are shown. To get a better insight into the visual features of these three outcomes of detection, we show both the raw original frames and the same frames after background subtraction. It can be seen that it is very hard to detect honeybees in raw frames but that background subtraction highlights changes in frames in comparison with the estimated background. These changes are visible as bright spots in the bottom row of Figure 8. Visual features of the background subtracted frames in Figure 8a (true positive) and 8b (false positive) correspond well to the elliptical blob honeybee model, used for artificial honeybees, which explains why they were obtained as positive detections. Based on this example, we can conclude that modelling honeybee appearance alone is not enough to achieve high precision and that motion information should be given more significance. A possible approach to achieve this is to consider a larger temporal context by, for example, using recurrent neural network for frame segmentation. However, visual features of the background subtracted frame in Figure 8c do not fit into the elliptical blob model, which suggests that, for avoiding false negatives, the honeybee appearance model should take into account changes of honeybee appearance during flight.
Due to the lack of texture, it is difficult for both human observers and the proposed detector to distinguish honeybees from other flying insects based solely on the information contained in a video. Nevertheless, we believe that the proposed approach could still be useful in applications involving honeybees because often we can safely assume that the number of other flying insects is small enough to not significantly influence the results. Such a detector can also remove the burden of detection of small flying objects from users, enabling them to focus on their flight patterns and discriminate between e.g., honeybees and other flying insects [51].
Conclusions
In this paper we presented a CNN-based method for detection of small moving objects trained on synthetic video sequences with real-world backgrounds and artificially generated moving targets. The proposed approach uses both the appearance and the motion information to detect small moving objects. We tested the trained detector on detection of flying honeybees in both synthetic and real-world video sequences and obtained promising results. In addition, we examined the influence of the parameters used for generating synthetic training sequences and hyperparameters of the detector on detection performance.
An important feature of our work is that it demonstrates the possibility of training an efficient small moving object detector using synthetic training video sequences. This makes usage of CNNs, in applications such as insect video surveillance, in which manually annotating training data is difficult or expensive, possible. Nevertheless, our experiments showed that testing on synthetic data can provide some insights, but cannot be completely relied on to serve as a proxy for the expected effectiveness on real-world data.
Since we train the frame segmentation CNN and tune the hyperparameters of the detector on synthetic sequences, it is essential that the artificial objects in the training sequences mimic the appearance of real-world objects as closely as possible. Therefore, we chose the parameters for generating training data, namely the sizes and other parameters used for generation of artificial objects based on the analysis of real-world recordings. Consequently, it could be expected that the changes in the real-world data caused by different properties of the target objects or different choices of the lenses, flying altitude, etc. would result in a deterioration of detection performance. Although in this work we chose the parameters for generating training data based on visual inspection of real-world sequences, it would be an interesting avenue for future research to explore the possibility to make the system more robust with respect to these parameters. Nevertheless, it should be noted that the methodology presented in this paper may be used to generate synthetic training and test sequences, which can be used to train and validate a small moving object detector adapted to the requirements of a specific real-world problem.
The proposed method for detection of small moving objects in videos captured using UAVs opens up the possibility of its application to various honeybee surveillance tasks such as pollination monitoring or land mine detection. In the future work we plan to investigate possibilities of these applications in more detail. | 2021-03-23T13:13:42.792Z | 2021-02-11T00:00:00.000 | {
"year": 2021,
"sha1": "90745196b2e7bd17c2877bb8f55374c360f445b6",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2072-4292/13/4/653/pdf",
"oa_status": "GOLD",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "2de3fc0158d832898b789e61f61047766bb0e2a4",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Computer Science"
]
} |
208171177 | pes2o/s2orc | v3-fos-license | Artificial Intelligence Applied to Flavonoid Data in Food Matrices
Increasing interest in constituents and dietary supplements has created the need for more efficient use of this information in nutrition-related fields. The present work aims to obtain optimal models to predict the total antioxidant properties of food matrices, using available information on the amount and class of flavonoids present in vegetables. A new dataset using databases that collect the flavonoid content of selected foods has been created. Structural information was obtained using a structural-topological approach called TOPological Sub-Structural Molecular (TOPSMODE). Different artificial intelligence algorithms were applied, including Machine Learning (ML) methods. The study allowed us to demonstrate the effectiveness of the models using structural-topological characteristics of dietary flavonoids. The proposed models can be considered, without overfitting, effective in predicting new values of Oxygen Radical Absorption capacity (ORAC), except in the Multi-Layer Perceptron (MLP) algorithm. The best optimal model was obtained by the Random Forest (RF) algorithm. The in silico methodology we developed allows us to confirm the effectiveness of the obtained models, by introducing the new structural-topological attributes, as well as selecting those that most influence the class variable.
Introduction
The relationship between dietary intake of bioactive antioxidants and health needs new approaches and studies for a better understanding. Research in this field is limited by the high number of bioactive compounds, which also hinders the development of analytical techniques and the availability of benchmarks [1]. Studying the currently growing and dispersed information on dietary phytochemicals is a huge challenge [2].
Several food databases were prepared based on the emerging Food Composition Database (FCDB) [3,4]. These databases focus on the composition of bioactive substances, including flavonoids and other polyphenols. multivariate perspective or using artificial intelligence (AI) techniques [37,38]. In the biomedical field, several unidirectional supervised networks were used, especially based on the MultiLayer Perceptron (MLP). In chemoinformatic studies, researchers used other methods of Machine Learning ML [39]. In the nutrition sciences, the need to use ML models for personalized nutrition has recently been raised [40]. However, as far as we know, these techniques have never been used for the analysis and study of the FCDB. Therefore, current work is focused on obtaining optimal models based on ML methods that allow for predicting the total antioxidant capacity of foods, based on information from the flavonoid composition database and structural topological descriptors of flavonoids.
Conformation of the Data Related to the Food Composition
Information from the dataset was obtained from different FCDB: (a) database for the flavonoid content of selected foods, version 3.1 and (b) isoflavone database released by the USDA in 2008 [3,5]. Therefore, estimation techniques were used to calculate unavailable values and the decision-making procedure described by Bhagwat et al. (2015) [35]. This information was used to prepare the dataset related to the composition of flavonoids in different foods. The standard reference (SR) was used to identify very unique food intake [7].
Prediction Using ML Algorithms
The prediction followed two phases, with different purposes: (i) selection of the attributes that best relate to the class (set A1). Metaheuristic Particle Swarm Optimization + Rougt Set Theory (PSO + RST) techniques were used [41,42], which included obtaining optimal prediction models among the selected ML algorithms using the hierarchical attributes of set A2 and their validation. To facilitate the experimentation of ML algorithms and the optimization capacity, the R language was used. This language also allowed the creation of each of the models corresponding to the three ML algorithms for predicting the antioxidant capacity. The interpolation package train function (Classification and Regression training) was used to evaluate the ML algorithms using the same metric and validation techniques.
Description of the Class Variable. The selected variable (attribute class) to predict was the ORAC value (ORACexp) was expressed in µmol TE/100g. ORAC was selected because it is considered the preferable methodology to evaluate antioxidant capacity. This is due to its correlation with antioxidant efficacy in vivo [43]. This assay was used to measure the antioxidant activity of foods. The assay measures the degree of inhibition of peroxyl radical induced oxidation by the compounds of interest in a chemical medium. The analytical method developed by Prior et al. (2003) was used as a reference method for selected sources [44].
Training Set and Test Set. As an internal validation methodology, the k-fold cross-validation method of k = 10 iterations was used for all algorithms [3].
Selection of Attributes
Attributes Selection. For the attributes, different weights were assigned considering their influence on the attribute class. The attributes (set A1) were:
Structural-topological characteristics (spectral moments, µk w , where w is bonding weights) The experimental parameters were taken from the available scientific literature. TPexp (GAE mg/100 g) was found for each substrate.
The structural-topological attributes used for the study were the molecular descriptors (µk) of the Topological Sub-structural Molecular Design (TOPSMODE) approach [45]. The spectral moments of each flavonoid were calculated from their Simplified Molecular Input Line Entry Specification (SMILES) using MODESLAB software (version 1.0) and weighted for different binding properties. These bonding weights used in the present work describe the n-octanol/water partition coefficient (H), polar surface (PS), polarizability (Pol), Gasteiger-Marsilli charge (Ch), van der Waals atomic radii (vdW), and molar refraction (RM). An extensive dataset was created with the structural-topological information of flavonoids present in foods.
Attributes Hierarchy. The following relationships were analyzed: (i) the relationship between the attributes of set A1 and the class variable was investigated, and (ii) the influence of new attributes related to the structural-topological information of flavonoids in the class was evaluated. The working hypothesis was based on the existence of a relationship between the chemical structure of each flavonoid and the total antioxidant activity of the studied food matrices.
To select the attributes (A2), a ranking ranked according to their relationship with the class was formed. Different weights were assigned to each attribute using the quality measure of a similarity decision system. Weights were assigned manually and using PSO + RST, implemented in Java.
Obtaining and Validating the Optimal ML Models
To develop the training process, the caret package (classification and regression training) was used through the RStudio version 0.99.441 tool. This allowed the R language to be used in all experiments.
For data preparation, the database contained in a .csv file was imported. The data was divided into a training dataset with 75% of the inputs and the remainder with 25% using the createDataPartition () function (createDataPartition (totalData $ total.orac, p = 0.75, list = FALSE)).
Attribute set A2 was selected for this study. The in silico influence of each attribute was considered in the class variable, which results from phase 1. In this phase, four algorithms were implemented: (a) nearest k-neighbor algorithm (KNN) (where the optimized parameter was the integer, such that k € [1,10]. (b) The Support Vector Machine (SVM) algorithm required the use of the kernlab package and the radial base function of the kernel function, which allows the optimization of sigma parameters according to C (evaluated in an incremental range from smallest to highest). (c) The MLP algorithm was used optimizing the size parameter, which represents the network size given by the number of internal layers it has. The values were assigned over a wide range to evaluate the trend following the best predictions and, thus, select the appropriate number for the parameter. The defined vector (c (1,4,3,5,7,9,10,11,12,15,20,25,50)) was performed using TuneGrid function. (d) In the Random Forest (RF) algorithm, mtry and ntree parameters were defined. The optimal value in this case was 3. For a more comprehensive experiment, it was considered that the use of ntree is generally treated with values of 500 or more, depending on the data and vectors seq (3,4,5,6) and seq (500,600,700) for mtry and ntree, respectively.
The resulting optimal models were validated using test suites. The predict function was used. It was found that the models chosen were not adjusted and the best performance model was established. For this, graphical functions and calculation of the metrics present in the R language were used. Experiment 1: Comparison of the outputs of the KNN, SVM, RF, and MLP algorithms generated in training with those generated in predicting the test suite. The goal is to determine the excess of fit in the models and which of the performances is the best. This was done through the plotObsVsPred function belonging to the interpolation package. A graph with the content of the reticular diagrams of each model was generated in the training and test sets. Model error metrics were calculated in the test phase using mmetrics from the rminer package. The parameters were two numerical vectors that represent the original outputs of each instance and the predicted outputs. Experiment 2: Comparison of predictions for new values of total antioxidant capacity in each model. The objective is to determine the accuracy of the antioxidant capacity predictions corresponding to the new compounds, by comparing them with the original ones, and by characterizing the best predicted occurrences. A dataframe was used, containing the output values of each algorithm and those of the original set, generated by the extractPrediction function of the interleaving package. The graphs were generated with the prediction values and their originals by instances, which were represented in a Cartesian coordinate system.
Results and Discussion
This project focused on the idea that dietary antioxidants are substances that significantly decrease the adverse effects of reactive species, such as reactive oxygen and nitrogen species, among normal physiological functions in humans [46,47]. Due to the complexity of food composition, it is not completely known which diet constituents are responsible for health benefits, but antioxidants appear to play an important role [48,49].
Database Description
The database used to create the templates consisted of 991 entries, six different types of attributes, and the class. Therefore, the resulting matrix has a high dimensionality. The studied feeding matrices were divided into 11 groups according to NDB (Nutrient Database) Alimentary Group Number [3]. Vegetables, spices, and herbal herbs are the two groups with the most flavonoid-containing foods, accounting for 39% and 37%, respectively ( Figure 1). In this dataset, high variability in flavonoid content predominated. This has been similar for all dietary polyphenols [50]. Several factors that affect the content of polyphenols in foods have been described [51,52]. were generated with the prediction values and their originals by instances, which were represented in a Cartesian coordinate system.
Results and Discussion
This project focused on the idea that dietary antioxidants are substances that significantly decrease the adverse effects of reactive species, such as reactive oxygen and nitrogen species, among normal physiological functions in humans [46,47]. Due to the complexity of food composition, it is not completely known which diet constituents are responsible for health benefits, but antioxidants appear to play an important role [48,49].
Database Description
The database used to create the templates consisted of 991 entries, six different types of attributes, and the class. Therefore, the resulting matrix has a high dimensionality. The studied feeding matrices were divided into 11 groups according to NDB (Nutrient Database) Alimentary Group Number [3]. Vegetables, spices, and herbal herbs are the two groups with the most flavonoid-containing foods, accounting for 39% and 37%, respectively ( Figure 1). In this dataset, high variability in flavonoid content predominated. This has been similar for all dietary polyphenols [50]. Several factors that affect the content of polyphenols in foods have been described [51,52]. The monomeric food flavonoids present in the data studied (id_flav attribute) belong to the chemical subclasses: flavonols, flavones, flavanones, and flavan-3-ols (Table 1). Quantifying them as aglycones facilitated the analysis but reduced the variety of compounds that could be analyzed. Flavonoids of the anthocyanin subclass can be found in many foods. Total anthocyanidin content in plant sources and extracts was correlated with the ORAC values. Anthocyanins constitute one of the The monomeric food flavonoids present in the data studied (id_flav attribute) belong to the chemical subclasses: flavonols, flavones, flavanones, and flavan-3-ols (Table 1). Quantifying them as aglycones facilitated the analysis but reduced the variety of compounds that could be analyzed. Flavonoids of the anthocyanin subclass can be found in many foods. Total anthocyanidin content in plant sources and extracts was correlated with the ORAC values. Anthocyanins constitute one of the most studied subclasses in the field [53]. Food intake of anthocyanins is high compared to other flavonoids due to their wide distribution in plant materials [54]. However, they were not included in this study because of their structure, which invalidates the application of the TOPSMODE approach [45]. [3,5]. b Extracted from [58]. c Extracted from [14]. d Extracted from [57]. Trolox equivalent antioxidant capacity flavonoid value (TEACexp). Total polyphenol value (TPexp). Nutrient Database Number (NDB No).
Chemical structures, SMILE codes, and some examples of sources of the studied flavonoids are shown in Table 2. Table 3 shows the order of influence of the attributes on the predictor variable (class). This order is associated with a higher "weight" in qualifying for this data matrix (dataset). Total polyphenols is the most important factor in predicting the total antioxidant capacity of foods. Although no history of this correlation is reported by AI algorithms, there are reports in which linear correlation was observed for more limited datasets. For example, positive correlations between ORAC and total phenolic content have also been previously reported [59].
Hierarchy Analysis of Attributes
In addition, the introduction of structural-topological information as new metadata helped to verify the hypothesis that the chemical structure of the food flavonoids is correlated with the total antioxidant capacity. The influence of these topological weights or structural attributes is limited to this database. However, the high dimensionality of the matrix and the fact that the food is compiled in the FCDB led to the suggestion that the scope of these results is correlated with the knowledge currently available in this field. [3,5]. b Extracted from [58]. c Extracted from [14]. d Extracted from [57]. Trolox equivalent antioxidant capacity flavonoid value (TEACexp). Total polyphenol value (TPexp). Nutrient Database Number (NDB No).
Chemical structures, SMILE codes, and some examples of sources of the studied flavonoids are shown in Table 2. Table 2. Examples of the chemical information of flavonoids, and their presence in food, contained in the studied database.
Chemical structures, SMILE codes, and some examples of sources of the studied flavonoids are shown in Table 2. Table 2. Examples of the chemical information of flavonoids, and their presence in food, contained in the studied database.
11260
Myricetin Potatoes, red, flesh and skin, raw (Solanum tuberosum) 11355 a Nutrient Database Number (NDB No) [3]. Table 3 shows the order of influence of the attributes on the predictor variable (class). This order is associated with a higher "weight" in qualifying for this data matrix (dataset). Total polyphenols is the most important factor in predicting the total antioxidant capacity of foods. Although no history of this correlation is reported by AI algorithms, there are reports in which linear correlation was observed for more limited datasets. For example, positive correlations between ORAC and total phenolic content have also been previously reported [59]. Table 3 shows the order of influence of the attributes on the predictor variable (class). This order is associated with a higher "weight" in qualifying for this data matrix (dataset). Total polyphenols is the most important factor in predicting the total antioxidant capacity of foods. Although no history of this correlation is reported by AI algorithms, there are reports in which linear correlation was observed for more limited datasets. For example, positive correlations between ORAC and total phenolic content have also been previously reported [59]. Table 3 shows the order of influence of the attributes on the predictor variable (class). This order is associated with a higher "weight" in qualifying for this data matrix (dataset). Total polyphenols is the most important factor in predicting the total antioxidant capacity of foods. Although no history of this correlation is reported by AI algorithms, there are reports in which linear correlation was observed for more limited datasets. For example, positive correlations between ORAC and total phenolic content have also been previously reported [59]. Table 3 shows the order of influence of the attributes on the predictor variable (class). This order is associated with a higher "weight" in qualifying for this data matrix (dataset). Total polyphenols is the most important factor in predicting the total antioxidant capacity of foods. Although no history of this correlation is reported by AI algorithms, there are reports in which linear correlation was observed for more limited datasets. For example, positive correlations between ORAC and total phenolic content have also been previously reported [59]. Table 3 shows the order of influence of the attributes on the predictor variable (class). This order is associated with a higher "weight" in qualifying for this data matrix (dataset). Total polyphenols is the most important factor in predicting the total antioxidant capacity of foods. Although no history of this correlation is reported by AI algorithms, there are reports in which linear correlation was observed for more limited datasets. For example, positive correlations between ORAC and total phenolic content have also been previously reported [59]. Table 3 shows the order of influence of the attributes on the predictor variable (class). This order is associated with a higher "weight" in qualifying for this data matrix (dataset). Total polyphenols is the most important factor in predicting the total antioxidant capacity of foods. Although no history of this correlation is reported by AI algorithms, there are reports in which linear correlation was observed for more limited datasets. For example, positive correlations between ORAC and total phenolic content have also been previously reported [59]. Table 3 shows the order of influence of the attributes on the predictor variable (class). This order is associated with a higher "weight" in qualifying for this data matrix (dataset). Total polyphenols is the most important factor in predicting the total antioxidant capacity of foods. Although no history of this correlation is reported by AI algorithms, there are reports in which linear correlation was observed for more limited datasets. For example, positive correlations between ORAC and total phenolic content have also been previously reported [59]. The molecular descriptors that most influence the class are presented in Table 3. All molecular descriptors (Table 3) are referred to as the n-octanol/water partition coefficient. For this reason, in the data series analyzed, this link property is the one with the most influence. The hydrophobicity of flavonoid diphenylpyran scaffolding may also influence antioxidant capacity [60]. The improved ORAC test provided a direct measure of hydrophilic and lipophilic antioxidant breaking ability in the presence of peroxyl radicals [61,62].
Hierarchy Analysis of Attributes
The amount of each flavonoid in the food matrix exert less influence (0.0341), as well as the antioxidant activity of the flavonoid compounds, especially TEACexp (0.0109). This may be related to the fact that antioxidant levels in foods do not necessarily reflect their total antioxidant capacity, which also depend on the synergistic and redox interactions between different molecules present in foods, which are not included in the dataset studied [48].
Training Model
For the KNN algorithm and an optimal value for the k = 1 training model, the metrics produce the best results (small RMSE, Root Mean Squared Error) ( Table 4). These results are superior to the models obtained in previous studies (RMSE = 5,475,398) [63]. This may be due to the features offered in the R language, which beneficially contribute to the model validation process and parameter optimization, as well as avoid excessive adjustments. It was also important to include structural-topological information as a highly influential attribute in the variable class. For experimentation with RF, parameters such as mtry and ntree were defined. The optimal value (for regression problems) is known to be given by the third part of the number of descriptors for mtry (in this case, it would be 3). For the ntree, it is common to be treated with values of 500 or more, depending on the date. The vectors seq (3,4,5,6) and seq (500,600,700) were defined for mtry and ntree, respectively, in order to make the experimentation a little more comprehensive. The optimal model was obtained with the values of mtry = 6 and ntree = 500.
The MLP neural network was used for model adjustment. In this case, the size parameter has been optimized, which represents the network size provided by the number of inner layers. The values were assigned over a wide range to evaluate the trend by following the best predictions and, thus, selecting the appropriate number for the parameter. Therefore, the vector c (1,4,3,5,7,9,10,11,12,15,20,25,50) is defined through tuneGrid. From the resulting models, the best predictor was obtained by applying the size parameter with the value 4, even though its performance was lower than in other experiments.
Regarding the analysis performed with the SVM algorithm, the results of the vector were obtained for the values of sigma c (0.03,0. 30,3.30,36,3,399,30) and distribution C (1,10,16,32,64,128,256,512.1024). The statistics for the Radial Basis Function core function experiment were: Sigma (σ) (399.3), C (10), RMSE (1853.446), Rsquared (0.879), RMSE SD (1370.442), and Rsquared SD (0.166). Subsequent analysis of the intervals around the σ and C values led to the definition of a new lower limit for the vector calculation. The optimal value was found for SVM (Table 4). This value was obtained for the new vector of σ and was c (1,11,121,1331). In this case, the optimal model was reached with σ = 121 and C = 10.
External Validation
Validation of optimal models was performed using the test sets. For this, the prediction function was used as a parameter. Error metrics for the results of each model (Table 5) allowed us to indicate the RF algorithm as the best performance in this validation phase, determined by RMSE and R 2 errors. The performance of the RMSE metrics for each of the algorithms in the parameter optimization process is shown. For the KNN algorithm (Figure 2a), as the parameter k increases number of neighbors (#Neighbors), the greater the error becomes. The results for SVM are shown in Figure 2b, where each row represents a value σ, distributed according to Cost (C) across the X axis. In this case, σ = 121 for C = 10, the optimal parameters are shown. Figure 2c corresponds to the RF algorithm. Each line represents the number of trees generated by the algorithm in each case (ntree). Points are models with the corresponding mtry value. The error tends to decrease as you approach a higher level for MLP. Error behavior is observed by varying the size parameter, which tends to increase abruptly from size = 15. The MLP neural network was used for model adjustment. In this case, the size parameter has been optimized, which represents the network size provided by the number of inner layers. The values were assigned over a wide range to evaluate the trend by following the best predictions and, thus, selecting the appropriate number for the parameter. Therefore, the vector c (1,4,3,5,7,9,10,11,12,15,20,25,50) is defined through tuneGrid. From the resulting models, the best predictor was obtained by applying the size parameter with the value 4, even though its performance was lower than in other experiments.
Regarding (Table 4). This value was obtained for the new vector of σ and was c (1,11,121,1331). In this case, the optimal model was reached with σ = 121 and C = 10.
External Validation
Validation of optimal models was performed using the test sets. For this, the prediction function was used as a parameter. Error metrics for the results of each model (Table 5) allowed us to indicate the RF algorithm as the best performance in this validation phase, determined by RMSE and R 2 errors. The performance of the RMSE metrics for each of the algorithms in the parameter optimization process is shown. For the KNN algorithm (Figure 2a), as the parameter k increases number of neighbors (#Neighbors), the greater the error becomes. The results for SVM are shown in Figure 2b, where each row represents a value σ, distributed according to Cost (C) across the X axis. In this case, σ = 121 for C = 10, the optimal parameters are shown. Figure 2c corresponds to the RF algorithm. Each line represents the number of trees generated by the algorithm in each case (ntree). Points are models with the corresponding mtry value. The error tends to decrease as you approach a higher level for MLP. Error behavior is observed by varying the size parameter, which tends to increase abruptly from size = 15. Tables 4 and 5. In all cases, the superiority of the model corresponds to RF, which is followed by SVM and KNN. In the case of the MLP neural network, a very poor performance at both times was recorded.
Predictions have adequate accuracy and low over-fit rate, except for the MLP model ( Figure 3). A comparison between the training moment and the test moment in each model shows similarity in the distribution of the output values around the reference line.
Effectiveness Performance Comparison
Experiment 1. Model prediction results for metrics in the training and testing phases are shown in Tables 4 and 5. In all cases, the superiority of the model corresponds to RF, which is followed by SVM and KNN. In the case of the MLP neural network, a very poor performance at both times was recorded.
Predictions have adequate accuracy and low over-fit rate, except for the MLP model ( Figure 3). A comparison between the training moment and the test moment in each model shows similarity in the distribution of the output values around the reference line. The optimal models obtained demonstrate the good effectiveness that can be achieved using AI algorithms. Only a small set of foods belonging to a specific food group or type was studied. An important and innovative feature of the present study is the size of the matrix, which represents the very large data set and describes various food groups. Prediction of the antioxidant capacity of foods by the ORAC method has not been documented, which makes it difficult to compare different methodologies. In the field of food, the use of data mining techniques is, therefore, untapped. However, there are recent studies that use traditional regression methods to predict a specific antioxidant property [64][65][66][67][68][69].
The complex role of diet in chronic diseases is difficult to understand, since a typical diet provides large amounts and different types of bioactive components. These bioactive molecules can modify a The optimal models obtained demonstrate the good effectiveness that can be achieved using AI algorithms. Only a small set of foods belonging to a specific food group or type was studied. An important and innovative feature of the present study is the size of the matrix, which represents the very large data set and describes various food groups. Prediction of the antioxidant capacity of foods by the ORAC method has not been documented, which makes it difficult to compare different methodologies. In the field of food, the use of data mining techniques is, therefore, untapped. However, there are recent studies that use traditional regression methods to predict a specific antioxidant property [64][65][66][67][68][69].
The complex role of diet in chronic diseases is difficult to understand, since a typical diet provides large amounts and different types of bioactive components. These bioactive molecules can modify a multitude of processes related to these diseases. Due to the complexity of this relationship, a comprehensive understanding of the role of these bioactive components is required in order to assess the role of food in modulating human health and disease. Food composition data alone does not provide this knowledge. However, processing your data and information obtained may be useful for further studies and to complement in vivo and ex vivo studies. Based on the current study, the total antioxidant capacity of foods can be predicted whenever their TPexp and the structural-topological information of the flavonoids they contain are known. The obtained models were automated in a software (PCAT, version 1.0), whose functionalities allow the validation of each model with a new data set and, therefore, new predictions.
Conclusions
The in silico methodology developed allows us to confirm the effectiveness of the models obtained through the introduction of the new structural-topological attributes, as well as the selection of those that most influence the class variable, determined by the calculation of the PSO + RST algorithm. The RF algorithm shows the best quality parameters, both in the training and validation phases, which are the most successful. It is worth mentioning the use of R as the language and work environment, which allows the optimization of the algorithms' parameters that led to the results. These predictions are limited to the FCDB and its metadata. There are new possibilities for learning ML models from new datasets, which is facilitated by their implementation in an automated predictive system in the development phase. The practical utility of the research is directed toward the generation of predictive theoretical knowledge, which is useful in the development of regional or local FCDB, dietary interventions, new nutritional studies, etc. It is an important antecedent in the "omics" disciplines applied to food and nutrition sciences, which lead to the analysis of a complex data system to obtain information using bioinformatic tools. | 2019-11-20T14:04:46.494Z | 2019-11-01T00:00:00.000 | {
"year": 2019,
"sha1": "16c999459532e805ed5ce9f3e27c2164dd31f33e",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2304-8158/8/11/573/pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "6e53b4168081046a43df8e46cb0bca57328360f6",
"s2fieldsofstudy": [
"Computer Science"
],
"extfieldsofstudy": [
"Computer Science",
"Medicine"
]
} |
14567412 | pes2o/s2orc | v3-fos-license | Poincare Invariance of Hamiltonian Semiclassical Field Theory
Semiclassical Hamiltonian field theory is investigated from the axiomatic point of view. A notion of a semiclassical state is introduced. An"elementary"semiclassical state is specified by a set of classical field configuration and quantum state in this external field."Composed"semiclassical states viewed as formal superpositions of"elementary"states are nontrivial only if the Maslov isotropic condition is satisfied; the inner product of"composed"semiclassical states is degenerate. The mathematical proof of Poincare invariance of semiclassical field theory is obtained for"elementary"and"composed"semiclassical states. The notion of semiclassical field is introduced; its Poincare invariance is also mathematically proved.
Introduction
Different approaches to semiclassical field theory have been developed. Most of them were based on the functional integral technique: physical quantities were expressed via functional integrals which were evaluated with the help of saddle-point or stationary-phase technique. Since energy spectrum and Smatrix elements can be found from the functional integral [1,2], this approach appeared to be useful for the soliton quantization theory [1,2,3,4,5].
Another important partial case of the semiclassical field theory is the theory of quantization in a strong external background classical field [6] or in curved space-time [7]: one decomposes the field as a sum of a classical c-number component and a quantum component. Then the theory is quantized.
On the other hand, the axiomatic field theory [18,19,20] tells us that main objects of QFT are states and observables. The Poincare group is represented in the Hilbert state space, so that evolution, boosts and other Poincare transformations are viewed as unitary operators.
The purpose of this paper is to introduce the semiclassical analogs of such QFT notions as states, fields and Poincare transformations. The analogs of Wightman Poincare invariance and field axioms for the semiclassical field theory are to be formulated and checked.
Unfortunately, "exact" QFT is mathematically constructed for a restricted class of models only (see, for example, [21,22,23,24]). Therefore, formal approximate methods such as perturbation theory seem to be ways to quantize the field theory rather than to construct approximations for the exact solutions of QFT equations. The conception of field quantization within the perturbation framework is popular [25,26]. One can expect that the semiclassical approximation plays an analogous role.
To construct the semiclassical formalism based on the notion of a state, one should use the equationof-motion formulation of QFT rather than the usual S-matrix formulation. It is well-known that additional difficulties such as Stueckelberg divergences [27] and problems associated with the Haag theorem [28,19,20] arise in the equation-of-motion approach. There are some ways to overcome them. The vacuum divergences can be eliminated in the perturbation theory with the help of the Faddeev transformation [29]. Stueckelberg divergences can be treated analogously [30] (exactly solvable models with Stueckelberg divergences have been suggested recently [31,32]). These investigations are important for the semiclassical Hamiltonian field theory [33].
The semiclassical approaches are formally applicable to the quantum field theory models if the Lagrangian depends on the fields ϕ and the small parameter λ as follows (see, for example, [4]): where V is an interaction potential. To illustrate the formal semiclassical ansatz for the state vector, use the functional Schrodinger representation (see, for example, [12,13,16,17]). States at fixed moment of time are represented as functionals Ψ[ϕ(·)] depending on fields ϕ(x), x ∈ R d , the field operatorφ(x) is the operator of multiplication by ϕ(x), while the canonically conjugated momentumπ(x) is represented as a differentiation operator −iδ/δϕ(x). The functional Schrodinger equation reads where The simplest semiclassical state corresponds to the Maslov theory of complex germ in a point [34,35,36]. It depends on the small parameter λ as where S t , Π t (x), Φ t (x), t ∈ R, x ∈ R d are smooth real functions which rapidly damp with all their derivatives as x → ∞, f t [φ(·)] is a t-dependent functional. As λ → 0, the substitution (1.3) satisfies eq.(1.2) in the leading order in λ if the following relations are obeyed. First, for the "action" S t one finds, (1.4) Second, Π t , Φ t obeys the classical Hamiltonian systeṁ Finally, the functional f t satisfies the functional Schrodinger equation with the quadratic Hamiltonian . (1.6) There are more complicated semiclassical states that also approximately satisfy the functional Schrodinger equation (1.2). These ansatzes correspond to the Maslov theory of Lagrangian manifolds with complex germs [34,35,36]. They are discussed in section 5. However, the QFT divergences lead to the following difficulties. It is not evident how one should specify the class of possible functionals f and introduce the inner product on such a space via functional integral. This class was constructed in [33]. In particular, it was found when the Gaussian functional belongs to this class. The condition on the quadratic form R which was obtained in [33] depends on Φ, Π and differs from the analogous condition in the free theory. This is in agreement with the statement of [37,38] that nonequivalent representations of the canonical commutation relations at different moments of time should be considered if QFT in the strong external field is investigated in the leading order in λ. However, this does not lead to non-unitarity of the exact theory: the simple example has been presented in [32]. Another problem is to formulate the semiclassical theory in terms of the axiomatic field theory. Section 2 deals with formulation of axioms of relativistic invariance and field for the semiclassical theory. Section 3 is devoted to construction of Poincare transformations. In section 4 the notion of semiclassical field is investigated. More complicated semiclassical states are constructed in section 5. Section 6 contains concluding remarks.
Thus, one introduces [39,40] the structure of a vector bundle (called as a "semiclassical bundle" in [40]) on the set of semiclassical states of the type (1.3). The base of the bundle being a space of sets (S, Π, Φ) ("extended phase space" [39]) will be denoted as X . The fibers are classes of functionals which depend on Φ and Π. Making use of the result concerning the class of functionals [33], one makes the bundle trivial as follows. Consider the Φ, Π-dependent mapping V which defines a correspondence between functionals f and elements of the Fock space F : as follows. LetR(x, y) be an Φ, Π -dependent symmetric function such that its imaginary part is a kernel of a positively definite operator and the condition of ref. [33] (see eq.(3.73) of subsection 3.6) is satisfied. . ByR we denote the operator with kernelR, whileΓ has a kernel i −1 (R −R * ). The vacuum vector of the Fock space corresponds to the Gaussian functional (1.7). The operator V is uniquely defined from the relations (2.1) Here A ± (x) are creation and annihilation operators in the Fock space.
Definition 2.1. A semiclassical state is a point on the trivial bundle X × F → X . An important postulate of QFT is Poincare invariance. This means that a representation of the Poincare group in the state space should be specified. For each Poincare transformation of the form which is denoted as (a, Λ), the unitary operator U a,Λ should be specified. The group property with (a 1 , Λ 1 )(a 2 , Λ 2 ) = (a 1 + Λ 1 a 2 , Λ 1 Λ 2 ).
Heuristic definition
Consider some special cases of Poincare transformations (a, Λ). The transformation (a, 1) is called translation. If a 0 = 0, this is a spatial translation, while the a = 0-case corresponds to the time translation or evolution. The transformation (0, Λ) ≡ Λ is called as a Lorentz transformation. If Λ = Λ τ , the Lorentz transformation is called as x 1 -boost. If Λ = 1 0 0 L the Lorentz transformation is called as a spatial rotation. Let L n be such a spatial rotation that where e 1 is a spatial vector of the form (1, 0, 0, ...). The transformation will be called as n-boost. It does not depend on choice of L n . Namely, let L (1) n and L (2) n be spatial rotations obeying eq. (3.2). Then L (1) n (L (2) n ) −1 e 1 = e 1 . Therefore, the transformations Λ φ and L (1) n (L (2) n ) −1 commute. Thus, (L (1) Lemma 3.1. [20] Let (a, Λ) be a Poincare transformation. It is uniquely presented as (a, Λ) = (a, 1)Λ φ L, where Λ φ is a boost, L is a spatial rotation. Proof. It follows from (2.2) that (a, Λ) = (a, 1)Λ. Let us show that Λ = Λ φ L and this decomposition is unique. Consider the transformation x ′ = Λx: Let L (1) be such a rotation that L (1) γ = || γ||e 1 . Consider the transformation To construct mappings u a,Λ and operators U a,Λ , one may consider first the partial cases (time evolution, spatial translations, x 1 or n-boost, spatial rotations) and then use the group property.
First of all, let us consider the representation of the Poincare groupŨ a,Λ in the functional Schrodinger representation. Formally, they are related with U a,Λ by the relatioñ To construct operatorsŨ a,Λ and mappings u a,Λ , let us use formal expressions for the Poincare transformations in the "exact" field theory. Namely, the (formally) unitary operator U a,Λ corresponding to the Poincare transformation (a, Λ) = (a 0 , 0)(a, 0) exp(α k l 0k ) exp( The momentum and angular momentum operators entering to formula (3.5) have the well-known form (see, for example, [25]) where formally We are going to apply the operator (3.5) to the semiclassical state (1.3). Note that the operators P µ and M µν (3.6) depend on fieldφ and momentumπ semiclassically, It is convenient to consider the more general problem (cf. [35]).Let us find as λ → 0 the state where K S,Π,Φ has the form (1.3), Note that the state functional (3.7) may be viewed as a solution to the Cauchy problem of the form (3.8) at τ = 1. Let us look for the asymptotic solution to eq.(3.8) in the following form: Substitution of functional (3.9) to eq.(3.8) gives us the following relation: Considering the terms of the orders O(λ −1 ), O(λ −1/2 ) and O(1) in eq.(3.10), we obtaiṅ (3.13) Here A 1 is a c-number quantity which depends on the ordering of the operatorsφ andπ and is relevant to the renormalization problem. We see that for the cases A = −P 0 a 0 , A = P j a j , A = −α k M 0k , A = 1 2 θ sm M sm the mapping u a,Λ takes the initial condition for the system (3.11), (3.12) to the solution of the Cauchy problem for this system at τ = 1. The operatorsŨ a,Λ transforms the initial condition for eq. (3.13) to the solution at τ = 1.
Poincare invariance of the classical theory
The purpose of this subsection is to find explicit forms of mappings u a,Λ . Consider some special cases.
Spatial translations
For this case, a 0 = 0, Λ = 1, so that A = P j a j and System (3.11), (3.12) takes the forṁ
Evolution transformation
Let a 0 = −t, a = 0, Λ = 1. Then A[Π, Φ] is a classical Hamiltonian, so that u −t,0,1 is a mapping taking the initial condition for system (1.4), (1.5) to the solution of the corresponding Cauchy problem.
Lemma 3.2. For transformation (3.23), the group property is satisfied.
We obtain the following corollaries. Let us make more precise the definition of the space X . Definition 3.1. X is a space of sets (S, Π, Φ) of a number S and functions Π, Φ ∈ S(R d ) such that there exists a unique solution of the Cauchy problem (3.22) such that the functions Φ(Λx + a)| x 0 =0 and ∂ µ Φ(Λx + a)| x 0 =0 are of the class S(R d ) for all a.Λ.
We see that the transformation u a,Λ : X → X is defined.
Infinitesimal properties
According to formula (A.2) of Appendix A, one can introduce the operators δ[A] on the space of differentiable functionals F of S, Π, Φ for each element A of the Poincare algebra. This operator plays an important role in analysis of algebraic properties of the representation U a,Λ . Elements of Lie algebra of the Poincare group can be identified with sets (b µ , θ µν ), θ νµ = −θ µν . The curve on the Poincare group with the tangent vector (b, θ) can be chosen to be The operator δ[(b, θ)] is a linear combination of b, θ: Let us find the coefficients from eqs. (3.15), (3.19), (1.5), (3.21). Let F be a differentiable functional of S, Π, Φ.
Semiclassical Poincare transformations in the functional representation
We have formally found the operatorsŨ a,Λ . However, it is not easy to check the group property. Therefore, construct the representation of the Poincare algebra according to Appendix A. Then we will check the algebraic property. The group property will be a corollary of the results of Appendix A. Let us construct the operatorsH((b, θ) : S, Π, Φ) (A.7) for elements of the Poincare algebra: Consider some cases.
The n-boost
Let a = 0, Λ = Λ τ n have the form (3.3), so that θ k0 = −n k = −θ 0k . Then A = n k M k0 , A has the form (3.20), so that eq.(3.13) takes the form: Note that the divergences in these operators are to be eliminated by adding c-number quantities to them. [M λµ ,M ρσ ] = −i(g λρM µσ − g µρM λσ + g µσM λρ − g λσM µρ ). (3.28) should be satisfied (eq.(A.11)). The formal check of these relations is straightforward. However, the functional representation is ill-defined, so that one should use the Fock representations and perform a renormalization. Find a relationship between operators H(A : X) andH(A : X) being generators of representation U a,Λ in Fock and Schrodinger pictures. Let g(τ ) be a curve on the Poincare group with tangent vector A. Eq.(3.4) impliesŨ
Properties of infinitesimal transformations
X . Differentiating this relation by τ at τ = 0, we find Therefore, )V X formally obey commutation relations (3.28). However, the divergences and renormalization problem should be taken into account.
Poincare transformations in the Fock representation
The purpose of this subsection is to construct Poincare transformations U a,Λ (u a,Λ X ← X) in the Fock space. First of all, we calculate the explicit form of generators. Then we will renormalize the obtained expressions and check the conditions of Appendix A. Then we will construct operators U a,Λ and check the group property.
First of all, investigate some properties of the operator V X .
Some properties of the operator V
Remind that the operator V taking the Fock space vector Ψ ∈ F to the functional f [φ(·)] is defined from the relation and from formulas (2.1) which can be rewritten as |c| can be formally found from the normalization condition The argument can be chosen to be arbitrary, for example, The operator V is defined form the relations (3.29) -(3.32) uniquely. Namely, any element of the Fock space can be presented [41] via its components, vacuum state an creation operators as Since the operators A ± (x) satisfy usual canonical commutation relations and A − (x)|0 >= 0, we obtain V A ± (x) = A ± (x)V .
The operator V depend on R. It is useful to find an explicit form of the operator V −1 δV . Proposition 4.4. The following property is satisfied: The notations of the type A +B A − are used for the operators like dxdyA + (x)B(x, y)A − (y), wherẽ B(x, y) is a kernel of the operatorB.
To check formula (3.33), consider the variation of the formula (2.1) if R is varied: Therefore, formula (3.33) is correct up to an additive constant. To find it, note that This relation and formula (2.1) imply It follows from the normalization conditions (3.31) and (3.32) that c = (detΓ) 1/4 . Therefore, δlnc = This property implies that
Explicit forms of Poincare generators
These operators do not contain any divergences.
2. Time evolution. Proposition 3.4 and eqs.(2.1) imply that the operatorH(X) is also quadratic with respect to creation and annihilation operators, Here H ±± (X) and H(X) are the following operators: is a (Π, Φ)-independent self-adjoint operator. The divergent number H is formally equal to
The n-boost
For boost transformation, we obtain that Here is a boost generator in the free theory which is a self-adjoint operator. The divergent term is
Check of algebraic conditions and renormalization
Let us write down the requirements which are sufficient for satisfying the properties H1-H6 of the Appendix A. Since the Poincare generators are quadratic with respect to creation and annihilation operators, we will use the results of Appendix B.
By D ⊂ F we denote the domain {ψ ∈ F |||ψ|| T 1 < ∞}. Lemma 3.5. For self-adjoint operators the following properties are satisfied: Proof. The first part of lemma is justified as follows. One should check that the following norms are finite: wherek j = −i∂/∂x j . This statement is a corollary of the following lemma. Lemma 3.6. The operators are bounded if α ≤ 1. This lemma is a corollary of Lemma C.29 of Appendix C.
To prove the second statement of lemma 3.5, represent it in the following form: It is necessary to investigate the Poincare transformation properties of the operatorsx j andk j . Lemma 3.7. The following relations are satisfied: e iωtxl e −iωt =x l +k lω−1 t, e iωtkl e −iωt =k l ; e ik s a sx l e −ik s a s =x l + a l ; e ik s a sk l e −ik s a s =k l ; e The operatorsX l (τ ) = e iL 1 τxl e −iL 1 τ have the following Weyl symbols: To check the properties, it is sufficient to show that they are satisfied at τ = 0 and show that the derivatives of left-hand and right-hand sides of these relations coincide.
Making use of commutation relations [x s , f (k)] = i ∂f ∂k s (k) and result of lemma 3.6, we find that operators (3.40) are bounded uniformly with respect to t ∈ [0, t 1 ]. Lemma 3.5 is proved.
Let the following conditions on R be imposed. Let h(α) be an arbitrary smooth curve on the Poincare group. P1, The property (3.34) is satisfied. P2. The α-dependent operator functions T B k++ (u h(α) X) and T H ++ (u h(α) X) are continuous in the Hilbert-Schmidt topology || · || 2 .
P3. The α-dependent operator functions B k++ (u h(α) X) and H ++ (u h(α) X) are continuously differentiable with respect to α in the Hilbert-Schmidt topology.
P4. 2. Let us check the commutation relations (A.9), i.e. property H3. Note that the divergences arise in terms B k and H only, so that we suppose them to be arbitrary and then find the conditions that provide Poincare invariance.
are satisfied automatically since is also satisfied.
For the relation
We write eqs.(3.41) -(3.43) as follows: (3.49) 6. The commutation relation 7. The most nontrivial commutation relations are They can be rewritten as follows: To perform a renormalization of the semiclassical theory, one should substitute the divergent terms T rΓ and T rx kΓ by finite terms which will be denoted as T r RΓ and T r R x kΓ , Relations P8. The quantities T r RΓ and T r R x kΓ obey the following properties: . Note also that property P6 can be substituted by the following property. P9. The functions T r R Γ(u h(α) X) and T r R x k Γ(u h(α) X) are continuous. Thus, we have formulated the conditions of invariance of the semiclassical field theory under Poincare algebra.
Construction of Poincare transformations
Let us construct now the operators U a,Λ .
1. First of all, consider the case of spatial rotations, a = 0, Λ = L. Denote U 0,L ≡ V L for this case. Let L(τ ) = exp( τ 2 θ sm l sm ). Then the operator V L(τ ) transforms the initial condition for the equation to the solution. The operator V L(τ ) is uniquely defined from the relations The group property for operators V L is obviously satisfied.
2. Let (a, Λ) is an x 1 -boost: a = 0, Λ = Λ τ has the form (3.1). Then the operator U 0,Λτ ≡ W τ takes the initial condition for the Cauchy problem for the equation to the solution of this Cauchy problem. Here S τ , Φ τ , Π τ are obtained from system (3.21).
3.
Let L be such a rotation that Le 1 = e 1 . Then there exists a matrix smooth function L(t) such that provided that Le 1 = e 1 . 4. Let |n| = 1, L n be such a spatial rotation that L n n = e 1 . Introduce the operator W φn as follows, It corresponds to the Lorentz transformation Show that definition (3.61) is correct. Let L (1) n n = e 1 , L (2) n n = e 1 . One has L (1) L . Note that the following property is satisfied.
Let L be a spatial rotation. Then Namely, let L n be such a rotation that L n n = e 1 , where n = φ/|| φ||. Then
5.
Let Λ be an arbitrary Lorentz transformation. Lemma 3.1 implies that it can be uniquely decomposed as follows, Λ = Λ φ L. Set Let us check the group property. Let be Lorentz transformations. Then where φ 3 and L are defined from the construction of proof of lemma 3.1. One should check that Consider the functions α φ 1 and α φ 2 instead of fixed vectors φ 1 and φ 2 . Then so that we can apply the result of lemma A8: Thus, eq.(3.64) is satisfied. We have checked the group property for the Lorentz transformations. 6. Consider the spatial translations, a 0 = 0, Λ = 1, a = 0. Then the operator U 0,a,1 takes the initial condition for the equation to the solution of this equation at τ = 1. Thus, the operator U 0,a,1 is uniquely defined from the relations For the evolution transformation, a 0 = −t, a = 0, Λ = 1, the operator U −t,0,1 takes the initial condition for the equation to the solution of the Cauchy problem. Here S t , Φ t , Π t obey system (1.4), (1.5). 8. Consider an arbitrary space-time translation (a, 1). Set The group property is a corollary of the relation which is correct because of lemma A.8. 9. Consider an arbitrary Poincare transformation (a, Λ). Set One should check the group property since the group properties for translations and Lorentz transformations have been already checked. We see that it is sufficient to justify the property Since the operator U (0,Λ) was defined as it is sufficient to check eq.(3.72) for the following cases: Lemma A.8 imply all these properties. Therefore, we have proved the following statement.
Choice of the operator R
Let us choose operator R in order to satisfy properties P1-P5, P7. We will use the notions of Appendix C (subsection C.5). First, we construct such an asymptotic expansion of a Weyl symbol R N that for (3.73) Next, we will construct another asymptotic expansion of a Weyl symbol R which obeys the condition ImR > 0 and approximately equals to R N at large |k|, so that eqs.(3.73) are satisfied. This will imply that properties P1-P5, P7 are satisfied. Let us define the expansions R N with the help of the following recursive relations. Set (3.74) Lemma 3.9. The following relation is satisfied: Proof. For n = 0, S 0 = V ′′ (Φ(x)), so that statement of lemma is satisfied. Suppose that statement of lemma is justified for n < N. Check it for n = N. One has The following property is obeyed: One has It follows from the definition of the Weyl symbol that However, the property ∂R n ∂x l = δ l P R n which means that eq.(3.34) is satisfied is checked by induction. Lemma 3.10 is proved. Lemma 3.11. The following properties are satisfied: It follows from the results of Appendix C that X l n is an asymptotic expansion of a Weyl symbol. Let degX l n = α. Suppose that α < n. Then the left-hand side of eqs.(3.75) is of the degree α, the degree of the righthand side of eq.(3.75) is greater than or equal to α − 1. In the leading order in 1/|k| the right-hand side has the form one has (−2iω k X l n ) and its degree should be greater than or equal to α . Therefore, degX l n ≥ α + 1. We obtain a contradiction. Suppose α > n. Then the left-hand side of eq.(3.75) is of the degree n, the right-hand side in the leading order in 1/|k| has the form 2iω k x l S n . so that degS n should obey the inequality degS n ≥ n + 1. We also obtain a contradiction.
Thus, α = n. In the leading order in 1/|k| one has up to terms of the degree n, so that deg(X l n − x l S n ) ≥ n + 1. Lemma 3.11 is proved. We see that for N ≥ max{d/2, d − 1} the properties (3.73) are satisfied. Lemma 3.12. Let R (1) and R (2) be asymptotic expansions of Weyl symbols, degR (1) Then We see that deg(X (1)l − X (2)l ) = N. The second statement is checked analogously. Lemma 3.12 is proved.
Let us construct such an asymptotic expansion R that deg(R − R N ) = N + 1 and ImR > 0. We will look for R as follows (cf. [36]), where A and B are real asymptotic expansions. Then are also asymptotic expansions of Weyl symbols. Choose A and B to be polynomials, Proof. It follows from recursive relations (3.74) that Show that B s is uniquely defined. In the leading order in 1/|k|, one has so that B 1 = C N,1 /2. Suppose that one can choose B 1 , ..., B s−1 in such a way that the degree of the asymptotic expansion of a Weyl symbol Choose B s in such a way that degF N,s ≥ 2s − 1. One has Up to terms of the degree 2s + 1, one has Since is uniquely defined. Lemma 3.13 is proved.
Thus, we have constructed the operator R such that properties (3.73) are satisfied. We obtain the following theorem.
Regularization and renormalization of a trace
The purpose of this subsection is to specify functionals T r R Γ and T r R x k Γ of arguments Φ, Π in order to satisfy properties P8,P9. We want the renormalized trace to satisfy properties like these: for such class of operators that is as wide as possible. Under these conditions, properties P8 and P9 are satisfied. However, one cannot specify such a renormalized trace. Namely, one should have Therefore, Choose l = d. Consider i = j in eq.(3.77) and perform the summation over i. Making use of the relation However, the operator with Weyl symbol is of the trace class. Its trace is nonzero, provided that dxf (x) = 0.
However, we can introduce a notion of a trace for asymptotic expansions of Weyl symbols. The trace will be specified not only by operator but also by its asymptotic expansion which is not unique (see remark after definition C.6).
Let A = (A,Ǎ) be asymptotic expansion of a Weyl symbol. Suppose that the coefficients A l of the formal asymptotic expansionǍ (3.78) For α + l 0 + 1 > d, the last integral in the right-hand side of eq.(3.78) converges. To specify trace, it is sufficient then to specify values of integrals for s ≤ d which are divergent. We will define the quantities (3.79), making use of the following argumentation.
1. We are going to specify to specify trace in such a way that Therefore, I s,n = 0 for odd n, while for even n I s,n is defined from eqs.(3.81), for example, I s,2 ij = 1 s δ ij I s,0 . Therefore, it is sufficient to define integrals Let us use the approach based on the dimensional regularization [43,44]. It is based on considering integrals (3.82) at arbitrary dimensionality of space-time. Expression (3.82) appears to be a meromorphic function of d. Substracting the poles corresponding to sufficiently small positive integer values of d, we obtain a finite expression.
Formally, one has If s−d 2 = −N is a nonpositive integer number, one should modify the definition of I s,0 . Change d → d−2ε. One finds: In the MS renormalization scheme [44], one should omit the term O(ε −1 ). There is also an MS renormalization scheme in which one omits also a fixed term of order O(1). Let us omit the term −ln(πm 2 ) + Γ ′ (1). We obtain the following renormalized value of the integral: provided that N = d−s 2 is a nonnegative integer number. Therefore, we have defined the renormalized trace of an asymptotic expansion of a Weyl symbol by formaula (3.78), provided that the coefficient functions are polynomials in k/ω k .
Let us investigate properties of the renormalized trace. Some properties are direct corollaries of definition (3.78).
Lemma 3.15. The following properties are satisfied: Corollary. The property AP9 is satisfied.
Proof. Making use of eq.(C.8), we find Analogously to Appendix C, one finds that (C j ,Č j ) ≡ C j is an asymptotic expansion of a Weyl symbol. It follows from lemma 3.15 that T r R ∂C j ∂k j = 0. We obtain statement of lemma 3.16.
The proof is analogous. Corollary 1. The following relations are satisfied: Corollary 2. Property P8 is satisfied. Thus, we have constructed functionals T r R x kΓ ≡ T r R x k Γ and T r RΓ ≡ T r R Γ such that properties P8 and P9 are satisfied.
Note that the "finite renormalization" [25] can be also be made. One can add quantities ∆T r R x kΓ and ∆T r RΓ to renormalized traces in such a way that This corresponds to the possibility of adding the finite one-loop counterterm to the Lagrangian.
Semiclassical field
An important feature of QFT is a notion of field. In this section we introduce the notion of a semiclassical field and check its Poincare invariance.
Definition of semiclassical field
First of all, introduce the notion of a semiclassical fieldφ(x, t : X) in the functional Schrodinger representation. At t = 0, this is the operator of multiplication by φ(x). For arbitrary t, one has is the operator transforming the initial condition for the Cauchy problem for eq.(1.6) to the solution to the Cauchy problem. The field operator in the Fock representation is related withφ by the transformation (2.1), Making use of eq.(3.4), one findŝ Let us defineφ mathematically as an operator distribution. Let S(R d ) be a space of complex smooth functions u : R d → C such that ||u|| l,m = max We say that the sequence {u k } ∈ S(R n ), k = 1, ∞ tends to zero if ||u k || l,m → k→∞ 0 for all l, m.
We will write Consider the mapping f → φ t {f }, f ∈ S(R d ) of the form Lemma 4.1. φ t is an operator distribution being continous with respect to t. Proof. One has It follows from lemma B.3 and theorem B.15 that this operator distribution is defined on and continous with respect to t. Lemma 4.1 is proved.
The proof is analogous to lemma 4.1.
Poincare invariance of the semiclassical field 4.2.1 Algebraic properties
To check the property of Poincare invariance, notice that it is sufficient to check it for partial cases: spatial translations, rotations, evolution, boost, since any Poincare transformation can be presented as a composition of these transformations. Let g B (τ ) = (a(τ ), Λ(τ )) be a one-parametric subgroup of Poincare group corresponding to the element B of the Poincare algebra. The Poincare invariance property can be rewritten aŝ Let us check relation (4.2). It is convenient to reduce the group property to an algebraic property. The formal derivative with respect to τ of the right-hand side of eq.(4.2) is If the quantity (4.3) vanishes, the property (4.2) will be satisfied since it is obeyed at τ = 0. Making use of the group property g B (τ + δτ ) = g B (δτ )g B (τ ), we find that vanishing of expression (4.3) is equivalent to the property: Proof. Consider the matrix element where Ψ 1 , Ψ 2 ∈ D. Show it to be differentiable with respect to τ . Let us use an auxiliary lemma. Lemma 4.4. Let Ψ ∈ D. Then the vector φ[v g B (τ ) f : u g B (τ ) X]Ψ is strongly continously differentiable with respect to τ .
Proof. One has: To prove lemma 4.3, notice that This quantity tends as δτ → 0 to the matrix element of the bilinear form (4.3) and vanishes under condition (4.4). Lemma 4.3 is proved.
Check of invariance
One should check property (4.2) for spatial translations and rotations, evolution and boost transformations.
For spatial translations and rotations, property (4.2) reads: φ(x, t : X) = U −1 0,a,Lφ (Lx + a, t : u 0,a,L X)U 0,a,L (4.5) It follows from commutativity of U 0,a,L and U t and eqs.(3.68), (3.58), (3.34) that property (4.5) is satisfied. For evolution operator, property (4.2) is rewritten as: Relation (4.6) is a direct corollary of definition (4.1) and group property for evolution operators. Consider now the n-boost transformation. Check property (4.4). It can be presented as Let us make use of property (A.13). For partial case it can be presented as Making use of relations (4.8), (4.9), we take relation (4.7) to the form is a corollary of relation (3.35). The relation is also checked by direct calculation. Thus, we have obtained the following result. 5 Composed semiclassical states
Semiclassical states in quantum mechanics
The most famous semiclassical approach to quantum mechanics is the WKB-approach. It is the following. One investigates the behavior of solutions of semiclassical equation of the form as ε → 0. Here the Weyl quantization is considered. The initial condition is chosen to be where S 0 is a real function. The WKB-result [34] is that the solution of eq.(5.1) at time moment t has the same form (5.2) up to O(ε), ||ψ t − ϕ t e i ε St || = O(ε), provided that S t (x) is a solution to the Cauchy problem for the Hamilton-Jacobi equation while ϕ t (x) obeys the transport equation and initial condition ϕ 0 . However, we are not obliged to choose the initial condition for eq.(5.1) in a form (5.2). There are other substitutions to eq.(5.1) that conserve their forms under time evolution as ε → 0. For example, the wave function used in the Maslov complex-WKB theory [34,35] also approximately satisfies eq.(5.1) iḟ Therefore, for the initial condition K ε S 0 ,P 0 ,Q 0 f 0 the solution for the Cauchy problem for eq.(5.1) will be asymptotically equal to K ε St,Pt,Qt f t up to terms of the order O( √ ε). The wave function (5.2) rapidly oscillates with respect to all variables. The wave function (5.3) rapidly damps at x − Q t >> O( √ ε). One should come to the conclusion that there exists a wave function asymptotically satisfying eq.(5.1) which oscillates with respect to one group of variables and damps with respect to other variables. The construction of such states is given in the Maslov theory of Lagrangian manifolds with comples germ [34,35]. Let α ∈ R k , (P (α), Q(α)) ∈ R 2d be a k-dimensional surface in the 2d-dimensional phase space, S(α) be a real function, f (α, ξ), ξ ∈ R d is a smooth function. Set ψ(x) to be not exponentially amall if and only if the distance between point x and surface Q(α) is One can note that wave functions (5.2) and (5.3) are partial cases of the wave function (5.5). Namely, for k = 0 the manifold (P (α), Q(α) is a point, so that the functions (5.5) coincide with (5.3). Let k = n. If the surface (P (α), Q(α) is in the general position, for x in some domain one has x = Q(α) for some α. Therefore, We obtain the WKB-wave function. Thus, WKB and wave-packet asymptotic formulas (5.2) and (5.3) are partial cases of the wave function (5.5) appeared in the theory of Lagrangian manifolds with complex germ.
The lack of formula (5.5) is that the dependence on α on x is implicit and too complicated. However, under certain conditions formula (5.5) is invariant if α is shifted by a quantity of the order O( √ ε). In this case, the point α can be chosen in arbitrary way such that the distance of x and Q(α) is of the order O( √ ε).
Namely, The form (5.5) of the semiclassical state appeared in the theory of Lagrangian manifolds with complex germ is not convenient for generalization to systems of infinite number of degrees of freedom. It is much more convenient to consider to consider wave function (5.3) as an "elementary" semiclassical state and wave function (5.5) as a "composed" semiclassical state presented as a superposition of elementary semiclassical states: where g(α, ξ) is a rapidly damping function as ξ → ∞. Superpositions of such type were considered in [45,46,47]; the general case was investigated in [36,48]. The composed semiclassical states for the abstract semiclassical theory were studied in [40].
To show that expression (5.9) is in agreement with formula (5.5), notice that the wave function (5.9) is exponentially small if the distance between x and the surface Q(α) is of order . Consider the substitution α = α + √ εβ. We find If the condition (5.7) is not satisfied, this is an integral of a rapidly oscillating function. It is exponentially small. Under condition (5.7) one can consider a limit ε → 0 and obtain the expression (5.5), provided that c ε = C ε ε k/2 and f (α, ξ) = dβe iβs( ∂Pm ∂αs ξm− ∂Qm (5.10) Integral representation (5.9) simplifies substitution of the wave function to eq.(5.1) and estimation of accuracy. Namely, the integrand entering to eq.(5.9) is an asymptotic solution to eq.(5.1), provided that eqs,(5.4) are satisfied. Using the linearity property, we obtain that the wave function (5.9) approximately satisfies eq.(5.1) [36]. Properties (5.7), (5.10) are shown to be invariant under time evolution [36]. It follows from eq.(5.10) that the function f is invariant under the following change of the function g ("gauge transformation"): Thus, the semiclassical state is specified at fixed S(α), P (α), Q(α) not by the function g but by the class of equivalence of functions g: two functions are equivalent if they are related by the transformation (5.11). This fact can be also illustrated if we evaluate the inner product ||ψ|| 2 as ε → 0: The integral over x is not exponentially small if α − γ = O( √ ε). After substitution γ = α + β √ ε, x − Q(α) = ξ √ ε and considering the limit ε → 0, we find The k-dimensional surface {(S(α), P (α), Q(α))} ("isotropic manifols") in the extended phase space has the following physical meaning. Consider the average value of a semiclassical observable A(x, −iε∂/∂x). As ε → 0, one has We see that only values of the corresponding classical observable on the surface {(Q(α), P (α))} are relevant for calculations fo average values as ε → 0. This means that the Blokhintsev-Wigner density function (Weyl symbol of the density matrix) corresponding to the composed semiclassical state is proportional to the delta function on the manifold {(Q(α), P (α))}. Therefore, elementary semiclassical states describe evolution of a point particle, while composed semiclassical states (including WKB-states) describe evolution of the more complicated objectsisotropic manifolds.
Since the inner product (5.16) may vanish for nonzero g, one should factorize the space of composed semiclassical states. Such functionals g that obey the property )]g(α, ·) = 0 (5.17) should be set to be equal to zero, g ∼ 0.
Constrained Fock space
The purpose of this subsection is to investigate the properties of the inner product for the fock vectors Y 1 , Y 2 . Suppose the functions B 1 , ..., B k to be linearly independent. Since the inner product (5.20) resembles the inner products for constrained systems [49], we will call the space under construction as a constrained Fock space. First of all, investigate the problem of convergence of the integral (5.20). Note that the operator is a well-defined unitary operator [41], provided that B ∈ L 2 (R d ), and obey the relations Lemma 5.1. (cf. [36]). The following estimation is satisfied: Applying this identity m times, we obtain: Making use of the result of lemma B.3, || dxB(x)A ± (x)Y || l ≤ ||B||||Y || l+1/2 , we find: Lemma 5.1 is proved. Corollary 1. Let B 1 , ..., B k be linearly independent functions. Then for some constant C 1 > 0 the following estimation is satisfied: Proof. It is sufficient to notice that for linearly independent B 1 , ..., B k the matrix (B m , B s ) is not degenerate, so that || 1 2 m β m B m || 2 ≥ C −1 1 |β| 2 for some C 1 . Applying the property ||Y || s/2 ≤ ||Y || m/2 for s ≤ m, making use of eq.(5.21), we prove corollary 1.
Let us investigate the property of nonnegative definiteness of the inner product (5.20). Lemma 5.2. Let ||Y || m < ∞ for some m > k and Im(B s , B l ) = 0. Then < Y, Y >≥ 0. Proof. Introduce the following "regularized" inner product It follows form estimation (5.22) and the Lesbegue theorem [52] that It is sufficient then to prove that < Y, Y > ε ≥ 0. One has: here the shift of variable β = β ′ − β ′′ is made. We have also taken into account that commute (i.e. Im(B s , B l ) = 0). Formula (5.23) is taken to the form Lemma 5.2 is proved. The expression (5.20) depends on k functions B 1 , ..., B k . However, one may perform linear substitutions of variables β, so that only the subspace span{B 1 , ..., B k } is essential.
Let L k be a k-dimensional isotropic plane with an invariant under shifts measure dσ. Let B 1 , ..., B k be a basis on L k . One can assign then coordinates β 1 , ..., β n to any element B ∈ L k according to the formula B = s β s B s . The measure dσ is presented as dσ = adβ 1 ...dβ k for some constant a. Consider the inner product Thus, the space F [k/2+1] is divided into equivalence classes. Introduce the following inner product on the factor-space F [k/2+1] / ∼: . This definition is correct because of the following statement.
The proof is standard (cf,, forexample, [52]). One has is the completeness of the factor-space F [k/2+1] / ∼ with respect to the inner product (5.25),
Transformations of constrained Fock vectors
Let us investigate evolution of constrained Fock vectors. Consider the Cauchy problem for eq.(B.3) (Appendix B): Proof. Analogously to proof of lemma B.11, one has It follows from Lemmas B.2, B.3 that Let L k be a k-dimensional isotropic plane with invariant measure dσ. Define its evolution transformation L t k as follows. Let (B 1 , ..., B k ) be a basis on L k . Let B t s be solutions to the Cauchy problems they can be expressed as (5.28) Therefore, L t k is also an isotropic plane. Define the measure dσ t on L t k as follows. For the choice of coordinates β 1 , .., β k on L t k according to the formula B = s β s B t s , set dσ = adβ 1 ...dβ k , where a does not depend on t.
Lemma 5.6. The inner product < ·, · > L t k is invariant under time evolution: Proof. By definition, one has Making use of the relation , we obtain statement of lemma 5.6.
It follows from lemma 5.6 that operator U t takes equivalent states to equivalent. Therefore, it can be reduced to the factorspace F [k/2+1] / ∼. Since it is unitary, it can be extended to F (L k ).
Definition of composed semiclassical state and its Poincare transformation
Let us formulate a definition of a composed semiclassical state.
Introduce the following measure dσ(α) on L k (α): (5.30) where β 1 , ..., β k are coordinates on L k (α) which are determined as B = s β s B s . Definition (5.30) is invariant under change of coordinates. Namely, let (α ′ 1 , ..., α ′ k ) be another set of local coordinates chosen instead of (α 1 , ..., α k ). Then so that property l β ′ l B ′ l = s β s B s implies that coordinate sets β and β ′ should be related as follows: Therefore, for the choice of coordinates α ′ one has The invariance property is checked. Introduce the vector (Hilbert) bundle π Λ k as follows. The base of the bundle is the isotropic manifold Λ k . The fibre that corresponds to the point α ∈ Λ k is H τ = F (L k (α)). Composed semiclassical states are introduced as sections of bundle π Λ k . Definition 5.2. A composed semiclassical state is a set of isotropic manifold Λ k and section Z of the bundle π Λ k , such that the inner product
converges.
Poincare transformation of isotropic manifold Λ k = {X(α)} is determined as . This definition is correct because of the resulta of previous subsubsection, provided that It is sufficient to prove property (5.31) for partial cases: spatial translations and rotations, time evolution and x 1 -boost.
where B k and B k++ have the form (3.38). It is satisfied because of eqs. (3.21) and (5.19).
Thus, the composed semiclassical states and their Poincare transformations are introduced.
Conclusions
In this paper a notion of a semiclassical state is introduced. "Elementary" semiclassical state are specified by a set (X, Ψ) of classical field configuration X (point on the infinite-dimensional manifold X , see section 2 and subsection 3.2) and element Ψ of the space F . Set of all "elementary" semiclassical states may be viewed as a semiclassical bundle. The physical meaning of classical field X is evident. Discuss the role of Ψ. In the soliton quantization language [1,2] Ψ specifies whether the quantum soliton is in the ground or excited state. For the Gaussian approach [14,15,16,17], Ψ specifies the form of the Gaussian functional, while for QFT in the strong external classical field [6,7] Ψ is a state of a quantum field in the classical background.
The "composed" semiclassical states have been also introduced (section 5). They can be viewed as superpositions of "elementary" semiclassical states and are specified by the functions (X(τ ), Ψ(τ )) defined on some domain of R k with values on the semiclassical bundle.
Not arbitrary superposition of elementary semiclassical states is nontrivial. The isotropic condition (5.15) should be satisfied. Moreover, the inner product of the "composed" semiclassical states (eq.(5.18)) is degenerate, so that there is a "gauge freedom" (5.17) in specifying composed semiclassical states.
The composed semiclassical states are used [36] in soliton quantization, since there are translation zero modes and solitons can be shifted. They are useful if there are conserved integrals of motion like charges. The correspondence between composed and elementary semiclassical states in QFT resembles the relationship between WKB and wave packet approximations in quantum mechanics.
An important feature of QFT is the property of Poincare invariance. In this paper an explicit check of this property is presented for semiclassical QFT. The Poincare transformations of elementary and composed semiclassical states have been constructed as follows. First, the simplest Poincare transformations like spatial translations and rotations, evolution and boost are considered. The infinitesimal transformations are investigated, the Lie algebraic commutation relations have been checked and the group properties have been justified.
For the "composed" states, conservation of the degenerate inner product and isotropic condition under Poincare transformation have been checked.
An important feature of QFT is a notion of field. In this paper this notion is introduced for semiclassical QFT. The property of Poincare invariance of semiclassical field is checked. This work was supported by the Russian Foundation for Basic Research, project 99-01-01198.
A Symmetries of the semiclassical theory under Lie groups
Let the semiclassical field theory be symmetric under Lie group transformations. Let G be a Lie group. Let X be a smooth (maybe, infinite-dimensional) manifold, H be a Hilbert space. Suppose that () a smooth mapping u : G × X → X is specified; the smooth mappings u g : X → X associated with the mapping u of the form u g X = u(g, X) satisfy the group property u g 1 u g 2 = u g 1 g 2 ; () for each X ∈ X unitary operators U g [X] ≡ U g (u g X ← X) : F → F are specified; the group property U g 1 (u g 1 g 2 X ← u g 2 X)U g 2 (u g 2 X ← X) = U g 1 g 2 (u g 1 g 2 X ← X) (A.1) are satisfied. Let us investigate the properties of infinitesimal transformations.
By T e G we denote the tangent space to the Lie group G at g = e. Let A ∈ T e G, g(τ ) be a smooth curve on the group G with the tangent vector A at the point g(0) = e. Introduce the differenial operator δ[A] on the space of differentiable functionals F on X : Lemma A.1 1. The quantity (A.2) does not depend on the choice of the curve g(τ ) with the tangent vector A.
Define the operator W g on the space of functionals F as The following property is satisfied: The proof is straightforward. Let g = g(τ ) be a curve with the tangent vector A at g(0) = e. Differentiating expression (A.4) by τ at τ = 0, we obtain: Lemma A.3. The following relation is satisfied: Here [A, B] is the Lie-algrbra commutator for the group G.
Proof. Let g(τ ) be a smooth curve on the Lie group G with tangent vector A at g(0) = e. Make use of the property (A.4): remember that the Lie commutator can be defined as Consider the derivatives of sides of eq.(A.6) at τ = 0. We obtain property (A.5). Lemma A.3 is proved. Consider now the infinitesimal properties of the transformation U. Suppose that on some dense subset D of F the vector functions U g [X]Ψ (Ψ ∈ D) are strongly continously differentiable with respect to g and smooth with respect to X. Define operators where g(τ ) is a curve on the group G with the tangent vector A at g(0) = e. The proof is analogous to lemma A.1. Let h(τ ) be a curve with the tangent vector B at h(0) = e. Eq.(A.1) implies: Differentiating this identity by τ at τ = 0, we obtain: Note that the operator (A.10) is an analog of the covariant differentiation operator in the theory of bundles (see, for example, [42]). Thus, the group property (A.1) is reformulated in terms of Lie algebras. Investigate now the problem of reconstructing the group representation if the algebra representation is known. Our purpose is to prove some lemmas which are useful in constructing the representation of the Poincare group.
Impose the following conditions on the operators H(A : X), A ∈ T e G, X ∈ X . H1. Hermitian operators H(A : X) are defined on a common domain D which is dense in F . H2. For each smooth curve h(α) on G and each Ψ ∈ D the vector function H(A : u h(α) X)Ψ is strongly continously differentiable with respect to α.
Let Z ∈ T e G be a subset of the Lie algebra of the group G. Let B ∈ Z, while g B (t) is a oneparametric subgroup of the Lie group G with the tangent vector B at t = 0, g B (0) = e. Denote by U t B (X) the operator taking the initial condition Ψ 0 ∈ D of the Cauchy problem for the equation (∂Ψ t /∂t is a strong derivative) to the solution Ψ t ∈ D of the Cauchy problem, Ψ t = U t B (X)Ψ 0 . This definition is correct under the following condition.
H4. Let B ∈ Z. If Ψ 0 ∈ D, there exists a solution of the Cauchy problem for eq.(A.12).
Uniqueness of the solution is a corollary of the property ||Ψ t || = ||Ψ 0 || which is checked directly by differentiation. The isometric operator U t B (X) can be expanded then from D to F . It satisfies the property Therefore, it is invertible and unitary. Impose also the following conditions. H5. Let B ∈ Z. For each smooth curve h(α) on G and each Ψ 0 ∈ D the quantity || ∂ ∂α H(A : u h(α) X)Ψ t || is bounded uniformly with respect to α, t ∈ [α 1 , α 2 ] × [t 1 , t 2 ] for any finite α 1 , α 2 , t 1 , t 2 .
H6. For ψ ∈ D, B ∈ Z, A ∈ T e G, the following property is satisfied: Under these conditions, we obtain: Lemma A.7. Let B ∈ Z, A ∈ T e G, The following property is satisfied on the domain D: Proof. Let us check that under these conditions the operator (δ[A]U t B )(X) is correctly defined, i.e. the strong derivative exists for all Ψ ∈ D, where h(α) is a curve on G with tangent vector A. Denote Ψ α,t = U t B (u h(α) X)Ψ. This vector obeys the equation Since Ψ α,0 = Ψ α+δα,0 = Ψ, we have Because of unitarity of the operators U t B , the following estimation takes place: Making use of the Lesbegue theorem [52] and condition H5, we find that ||Ψ α+δα,t − Ψ α,t || → δα=0 0, so that the operator U t B (u h(α) X) is strongly continous with respect to α. Furthermore, The following estimation takes place: Making use of the Lesbegue theorem, conditions H2,H5, we find that the quantity (A.16) tends to zero as δα → 0. Thus, the vector (A.14) is correctly defined. It follows from the expression (A.15) that To construct the representation of the Poincare group, the following statement is used in this paper. Let the property g Bn (t n (α))..
. Let s(α) be a smooth curve on the group G, t(α) is a smooth real function, B ∈ T e G. Then the operator function U t(α) B (u s(α) X) is strongly differentiable with respect to α on D and ∂ ∂α U t(α) where ds dα s −1 is a tangent vector to the curve s(α + τ )s −1 (α) at τ = 0. Proof. Let Ψ ∈ D. One has The first term in the right-hand side of eq.(A.20) tends to by definition of the operator U t B (X). Consider the second term. It can be represented as Making use of eq. (A.15), we take this term to the form Making use of the Lesbegue theorem and property H6, we see that the vector (A.21) is strongly continous as δα → 0, so that it is equal to We obtain formula (A. 19).
B Some properties of quadratic Hamiltonians in the Fock space
The purpose of this appendix is to introduce some notations and check some properties of operators in Fock space. Remind that the Fock space F (L 2 (R d )) is defined as a space of sets Ψ = (Ψ 0 , Ψ 1 (x 1 ), ..., Ψ n (x 1 , ..., x n ), ...) of symmetric with respect to x 1 , ..., x n symmetric functions Ψ n such that ||Ψ|| 2 = ∞ n=0 |Ψ n || 2 < ∞. By A ± (x) we denote, as usual, the creation and annihilation operator distributions: By |0 > we denote, as usual, the vacuum vector of the form (1, 0, 0, ...). Arbitrary vector of the Fock space can be presented via the creation operators and vacuum vector as follows [41] Introduce the operator of number of particlesn as (nψ) n = nψ n . Let T be a nonbounded self-adjoint operator in L 2 (R d ) such that T − 1 ≥ 0. By A + T A − we denote the operator of the form Introduce also the following norms in the Fock space: This relation is a corollary of the formula (ψ n , T 2l 1 ⊗ ... ⊗ T 2ln ψ n ) ≥ (ψ n , ψ n ) for all l 1 , ..., l n ≥ 0. The latter formula is obtained from the relation ||T −l 1 ⊗ ... ⊗ T −ln || ≤ 1. Lemma B.1 is proved.
Proof. One has from lemmas B.2 and B.3 that The fact that d dα H α ψ is continuous is checked analogously. Lemma B.4 is proved.
Consider now the Cauchy problem for the equation on the Fock vector Ψ t ; the strong derivative enters to eq.(B.3). Formally, the solution for the initial condition is looked for in the following form [41,50] while operators A + t (x) are chosen to be Namely, the Gaussian ansatz (B.6) formally satisfies eq.(B.3) if Here M t is the operator with kernel M t (x, y), Here F t , G t are operators with kernels F t (x, y) and G t (x, y).
The proof is presented in [41].
Proof. The commutation relations (B.20) are rewritten as Proof. It follows from lemma B.9 that ). Identity (B.24) is then a corollary of definition of the operator U t .
By D ∈ F we denote set of such Fock vectors Ψ that ||Ψ|| T 1 < ∞. Lemma B.11. Let Ψ 0 ∈ D. Suppose that T F t and H ++ are continuous operator functions in the || · || 2 -norm, are continous operator functions in the || · ||-norm. Then the following statements are satisfied.
To check the second statement, note that lemma B.8 imply that where g t is a bounded operator functions, f t is a Hilbert-Schmidt operator function. The derivatives in (B.27) are understood as Lemma B.12. Let Y t be a strongly continous operator function, while ||Z t+τ − Z t || 2 → τ →0 0, for smooth functions a t k . Then there exist a solution to the Cauchy problem (B.27) such that for smooth functions a t k .
where (f t , g t ) is a solution to the Cauchy problem (B.27) with Y t = e iLt H t e −iLt , Z t = e iLt H ++ t e iL * t , f 0 = 0, g 0 = 1. Check of properties (B.34) is straightforward. Let us prove relations (B.33). One has Since we obtain relations (B.33). Property (B.23) is proved analogously to [50]: one should consider the convergent in ||·||-norm series
C Some properties of the Weyl symbol
The purpose of this appendix is to investigate some properties of Weyl symbols of operators which are useful in justification of properties H1-H6 of Appendix A.
C.1 Definition of Weyl symbol
Firs of all, remind the definition of Weyl symbol of operator (see, for example, [34,53]). Let A(x, k), x, k ∈ R d be a classical observable depending on coordinates x = (x 1 , ..., x d ) and momenta k = (k 1 , ..., k d ). To specify the corresponding quantum observable (to "quantize" the observable A), one should substitute the coordinates x i by operatorsx i of multiplication by x i , while the momenta k j should be substituted by the operatorsk j = −i∂/∂k j . However, it is not easy to determine the operator A(x,k) for arbitrary function A, since the coordinate and momenta operators do not commute. Different ways of ordering operatorsx andk are known. In the Weyl approach, one first considers the partial case A = e iαk+iβx (C.1) and sets = e iαk+iβx (C. 2) The operator (C.2) can be defined as a transformation taking the initial condition f 0 (x) for the Cauchy problem for the equation to the solution f 1 (x) to the Cauchy problem at t = 1. Eq.(C.3) is exactly solvable: For an arbitrary function A, one presents it as a superposition of exponents (C.1), Applying the formula for inverse Fourier transformation and making use of formula (C.4), we find We denote the operatoe of the form (C.5) as = W(A). We will also write A = W(Â) if = W(A). Definition C.1. The operator W(A) is called a Weyl quantization of the function A. The function W(Â) is called as a Weyl symbol of the operatorÂ.
Lemma C.4. The following properties are satisfied: k i ∈ B −1 , ω α k ∈ B −α . Proof. Since |k i /ω k | < 1, we obtain the property k i ∈ B −1 . For the function ω α k , one has where P is a polynomial in k i /ω k . Property (C.7) is checked by induction. Therefore, functions (C.6) are bounded for N = 1. Lemma C.4 is proved. Lemma C.5. 1. Let A ∈ B N . Then Then Proof. Property 1 is a corollary of lemmas C.2 and C.4. Property 1 implies property 2. Lemma is proved.
Lemma C.6. 1. Let B N − lim n→∞ A n = A. Then The proof is analogous to the proof of lemma C.3.
The proof is analogous to lemma C.3.
However, for operator F (k) one has ||F (k)|| = sup k ||F (k)||, since in the momentum representation F (k) is the operator of multiplication be F (k). Therefore, Here N is an arbitrary number such that N > d/2. Thus, The first statement is judtified. Proof of the second statement is analogous. Lemma C. 8 whcih can be obtained from definition (C.5). One has The first statement is justified. Proof of the second statement is analogous. Lemma C.9 is proved.
C.3 Properties of *-product
Remind that the Weyl sumbol of the product of operators can be presented as [34,53] Formula (C.8) can be obtained from definition (C.5). Let us investigate some properties of formula (C.8). Let us find an expansion of formula (C.8) as |k| → ∞. Formally, one has Therefore, This is an asymptotic expansion in 1/|k| as |k| → ∞. Let us estimate an accuracy of the asymptotic series.
C.3.1 The k-independent case Definition C. 3. We say that the function f (x), x ∈ R d is of the class C if f is a smooth function such that for each set (i 1 , ..., i l ) there exists m > 0 such that the function Then the only nontrivial term is r N 1 0 which is taken by integrating by parts to the form r (2) . (C.10) Let us prove some auxiliary statements. Lemma C.10. For some constent A 1 the estimation is satisfied.
so that it is sufficient to check estimation (C.11) for p = αk only. For the function 1/f 2 , one has It has the following minimal value (C.12) The quantity (C.12) is bounded below. Thus, lemma is proved.
and integrating by parts, we obtain that For the function ω N k F , one has Choose L 2 to be such a number that L 2 +N 2 is integer, L 2 > d. The property χ ∈ C implies that there exists such K that ∂ m ∂ξ i 1 ...∂ξ in χ(x + ξ) = ((x + ξ) 2 + 1) K f m.i 1 ...im (x + ξ), m = 0, where f m,i 1 ...im are bounded functions. Choose L 1 to be integer and L 1 > K+d 2 . Integrating expression (C.15) by parts, making use of corollary of lemma C.10 and property C ∈ A N , we obtain that ω N k F is a bounded function. Lemma C.11 is proved.
Lemma C.12. Under conditions of lemma C.11 F ∈ A N . Proof. It is sufficient to consider the functions ω N +I k ∂ I ∂k i 1 ...∂k i I x j 1 ...x j R ∂ ∂x s 1 ... ∂ ∂x s P F (C. 16) which are expressed via linear combinations of integrals of the type (C.13). Lemma C.12 is a corollary of lemma C.11. Lemma C. 13. Let A N − lim n→∞ C n = C, χ ∈ C, ϕ ∈ C[0.1]. Then A N − lim n→∞ F n = F .
The proof is analogous to lemmas C.11 and C.12. We obtain therefore the following theorem. Theorem C.14. 1. Let f ∈ C, B ∈ A N . Then 2. Let f ∈ C, A N − lim n→∞ B n = 0. Then A N +K+1 − lim n→∞ (f * B n − f K * B n ) = 0.
C.3.2 The x-independent case
Let A = A(k), A ∈ B M 1 , B ∈ A M 2 . The only nontrivial term is taken to the form: Lemma C. 15. C = C(k), C ∈ B K 1 , K 1 > 0, D ∈ A K 2 , ϕ ∈ C[0, 1]. Then for 2 )D(x + ξ, k)ξ j 1 ...ξ jm the function ω K 1 +K 2 k F is bounded. Proof. Inserting the identity (C.14) and integrating by parts, we obtain that For the function ω K 1 +K 2 k F , one has ω K 1 +K 2 k F (x, k) = 1 0 dαϕ(α) dβdξ Integrating by parts for sufficiently large L 1 , L 2 , making use of lemmas C.10, we check proposition of lemma C.15. Lemma C. 16. Under conditions of lemma C.15 F ∈ A K 1 +K 2 . Lemma C.17. Let A K 2 − lim n→∞ D n = D, C = C(k), C ∈ B K 1 , K 1 > 0, ϕ ∈ C[0, 1]. Then The proof is analogous to lemmas C.12 and C. 13. We obtain then the following theorem. Remark. If the proposition of theorem C.18 is satisfied for K = K 0 , it is satisfied for all K ≤ K 0 . Therefore, the condition K + M 1 + 1 > 0 can be omitted.
The following lemma is a corollary of theorem C.18. Making use of lemma C.9, we prove lemma C.19.
C.3.3 The A N -case
Let A ∈ A M 1 , B ∈ A M 2 . The r-terms can be investigated as follows.
3. Analogously to previous subsubsections, we prove the following theorem. Theorem C.21. 1. Let A ∈ A M 1 , B ∈ A M 2 . Then
Proof. Consider the Fourier transformation of the function A: A(k) = dγÃ(γ)e iγk .
We have obtained the following important statement. Lemma C. 29. Let Then [f (x), g(k)] is a bounded operator.
C.6 Asymptotic expansions of Weyl symbol
To check the property of Poincare invariance, it is important to investigate the large-k expansion of the Weyl symbols. Introduce the correponding definitions. Definition C.4. 1. We say that a smooth function A(x, n), x, n ∈ R d , |n| < 1, is of the calss L if the functions ∂ I ∂n i 1 ...∂n i I x j 1 ... Making use of definition C.2 and lemma C.25, we obtain the following corollary. Corollary. 1. Let A ∈ L. Then the function ω −α k A(x, k/ω k ) is of the class A α . 2. Let L − lim s→∞ A s = 0. Then A α − lim s→∞ ω −α k A s (x, k/omega k ) = 0. Definition C.5. 1. A formal asymptotic expansion is a setǍ of α ∈ R and functions A 0 , A 1 , ... ∈ L. We say that the formal asymptotic expansionsǍ = (α, A 0 , A 1 , ..) andB = (β, B 0 , B 1 , ..) are equivalent if α − β is an integer number and A l−α+β = B l for all l = −∞, +∞ (we assume A l = 0 and B l = 0 for l < 0. We denote formal asymptotic expansions of Weyl symbols aš A ≡ ∞ n=0 ω −n−α k A n (x, k/ω k ).
One can choose α = 2, A 0 (x, n) = m 2 f (x) anf find A(x, k) = ω −2 k A 0 (x, k/ω k ). On the other hand, one can set α = 0, A 0 (x, n) = f (x)(1 − n i n i ) and obtain A(x, k) = A 0 (x, k/ω k ) since ω 2 k − k i k i = m 2 . We see that a degree is a characteristic feature of an expansion rather than of a symbol.
Let
The integral t 2 t 1 A(t)dt is also defined in a standard way. Theorem C.32 imply the following statement. Theorem C.33. | 2014-10-01T00:00:00.000Z | 2001-03-10T00:00:00.000 | {
"year": 2001,
"sha1": "2e6ff228cf37babc29f918b134d16df8ebbe075c",
"oa_license": null,
"oa_url": null,
"oa_status": null,
"pdf_src": "Arxiv",
"pdf_hash": "2e6ff228cf37babc29f918b134d16df8ebbe075c",
"s2fieldsofstudy": [
"Physics",
"Mathematics"
],
"extfieldsofstudy": [
"Physics",
"Mathematics"
]
} |
220283929 | pes2o/s2orc | v3-fos-license | Polymer Transformers: Interdigitating Reaction Networks of Fueled Monomer Species to Reconfigure Functional Polymer States
Abstract Adaptivity is an essential trait of life. One type of adaptivity is the reconfiguration of a functional system states by correlating sensory inputs. We report polymer transformers, which can adaptively reconfigure their composition from a state of a mixed copolymer to being enriched in either monomer A or B. This is achieved by embedding and hierarchically interconnecting two chemically fueled activation/deactivation enzymatic reaction networks for both monomers via a joint activation pathway (network level) and an AB linker monomer reactive to both A and B (species level). The ratio of enzymes governing the individual deactivation pathways (our external signals) control the enrichment behavior in the dynamic state. The method shows high programmability of the reconfigured state, rejuvenation of transformation cycles, and quick in situ adaptation. As a proof‐of‐concept, we showcase this dynamic reconfiguration for colloidal surface functionalities.
Abstract: Adaptivity is an essential trait of life. One type of adaptivity is the reconfiguration of a functional system states by correlating sensory inputs. We report polymer transformers, which can adaptively reconfigure their composition from a state of a mixed copolymer to being enriched in either monomer A or B. This is achieved by embedding and hierarchically interconnecting two chemically fueled activation/deactivation enzymatic reaction networks for both monomers via a joint activation pathway (network level) and an AB linker monomer reactive to both A and B (species level). The ratio of enzymes governing the individual deactivation pathways (our external signals) control the enrichment behavior in the dynamic state. The method shows high programmability of the reconfigured state, rejuvenation of transformation cycles, and quick in situ adaptation. As a proof-of-concept, we showcase this dynamic reconfiguration for colloidal surface functionalities.
Signal correlation is a key function in living organisms to realize adaptation and evolution. [1] For example, stem cells can differentiate into different lineages by correlating mechanical, topographical, and (bio)chemical sensory inputs using (bio)chemical reaction networks. [2] Induced pluripotent stem cells can even regenerate from a mature state and then transform to different functions. [3] Although these abilities for adaptive reconfiguration of functional states are of formida-ble complexity, an abstraction allows to define one subset of adaptive behavior in synthetic systems. This is the ability to form distinct functional outcomes depending on the strength of different signals in complex sensory landscapes, and to be able to reconfigure the outcome quickly and potentially repeatedly.
For future synthetic materials systems, it is important to develop methods to endow systems with capacities for adaptivity and reconfiguration so that they can be transformed into other desired functions when meeting different environments. Dynamically self-assembling systems with controlled subunit exchange are a favorable starting point as dynamics provide an inherent capacity for fast reorganization. This is seen in the dissipative self-assembly of the microtubules in cells, which are in a dynamic flux-like state so that the cytoskeleton can quickly reconfigure into new spatial patterns. [4] Synthetic out-of-equilibrium systems fueled by chemical energy have been targeted for a range of systems, [5][6][7][8][9][10][11][12] but structural dynamics, which is a key for quick reconfiguration, has thus far only been carefully shown in limited systems. [13][14][15][16] Inspired by the abilities for reconfiguration of living systems, we report an abstracted chemical systems approach, termed polymer transformers, which allows the reconfiguration of functional material states by embedding the underlying species into two interdigitated enzymatic reaction networks (iERNs). The initial material state can transform into other states following certain signal conditions, return to the initial state, and then be reconfigured to a different state. Enzymes are used as signaling units to decisively influence the outcome of iERNs. Small DNA-based monomers, M E , M N , M L , that dynamically polymerize into long polymer strands using ERNs of an ATP-powered T4 DNA ligase-mediated polymerization and concurrent restriction/cutting by restriction endonucleases (REases), constitute the (co)polymers for reconfiguration ( Figure 1). DNA monomers are models for synthetic polymers, as they can advantageously be manipulated by enzymes as signal processors in a highly defined fashion. [17] Fluorophores attached to M E and M N describe a first functional output on a polymer level.
The interdigitation of the iERNs occurs on two levels: Individually, M N and M E can operate in their own ERN1 and ERN2, respectively. When mixed into one system, a connection on the ERN level occurs, because the ligase is jointly used. Upon addition of the heterolinker M L , the interconnection occurs additionally on the building block, and, hence, on the structural level ( Figure 1 a; Supporting Information, Figure S1). Previously, for simple, single-species ERNs of this type, we confirmed that i) the degree of ligation in the dynamic steady state (DySS) plateau depends on the ratio of ligase/REase, ii) structural dynamics (constant bond shuffling and subunit exchange) occurs, and iii) the lifetimes depend on the ATP concentration. [16,18,19] The adaptive transformation will operate as follows (Figure 1 b): First, ligase (black) and ATP polymerize the three monomers into a mixed copolymer, serving as a homogenous initial material state. Afterwards, we add two REases (green, red) as signals with a certain ratio to dynamize the system. This leads to a transformation into different copolymer types that enrich selected monomers and sort out the other, depending on the kinetic balance of iERN as instructed by the REases ratio. The quick reconfiguration is allowed by the dynamics of the system, which is in constant balance between ligation and cutting. As ATP is consumed, all copolymers degrade into the initial monomers owing to cleavage by REases. After heat inactivation of the enzymes, a new round of transformations can start.
We first identified two REases, EcoRI and NheI, that can be operated at similar activities (Supporting Information, Figure S2). Before targeting the transformation, we elucidate the kinetics of the ATP-fueled transient growth of M E and M N as a function of the REase ratio in a joint reaction. M L is absent, causing individual homopolymerizations. Since M E and M N share the ligase, the system is mainly controlled by the activity of the REases. Figure 2 a-d displays time-dependent agarose gel electrophoresis (AGE) of transient polymerizations at different ratios of EcoRI and NheI, which can be quantified as mass-weighted average chain length distributions in base pairs (bp w ; Figure 2 e-h, details of analysis in Ref. [16] and the Supporting Information; please see definitions of enzyme activity U in the Supporting Information). When EcoRI cutting activity is higher than NheI (Figure 2 a), red M E can only grow to oligomers (ca. 90 bp), while green M N reaches a longer polymer state (ca. 260 bp). This is due to more frequent cleavage by EcoRI, causing a lower degree of ligation and hence a shortened bp w . Reduction of [EcoRI] to 8.0 and 5.0 U mL À1 (constant [NheI] = 1.0 U mL À1 ) first doubles the M E chain length to ca. 180 bp and then allows M E to reach similar levels as M N (ca. 350 bp). An inversion, meaning high bp w of M N and low bp w of M E , is obtained by inverting the REase ratio (Figures 2 d,h). Both networks, ERN1 and ERN2, are clearly connected via the activation pathway (shared ligase; Figure 2 a-c,e-g). Upon reduction of [EcoRI] from 10 to 5 U mL À1 , both bp w of M E and M N increase noticeably. This is because the ligase is less required in ligating the M E sequences at lower [EcoRI], and hence more available for ligating longer M N chains. These results confirm that tuning the cutting kinetics with different REase ratios enables two DNA polymers to be simultaneously grown to a tunable bp w in the presence of shared ligase. Yet, first level interdigitation effects by joint use of the activation reaction need to be considered.
We next target distinct reconfiguration by incorporating the heterolinker M L . This further interdigitates the two ERNs Figures 3 a-d). Subsequently, we demonstrate its transformation by adding various ratios of EcoRI and NheI, aiming to enrich or deplete the copolymer structure from one of its species. Figure 3 a shows that a high NheI activity transforms the orange P initial to a polymer containing almost exclusive red fluorescence. This means that the balance in iERN kinetics reconfigures the copolymer to be mainly composed of M E , while M N are mostly sorted out as monomers, because the excessive cutting frequency leads to a low degree of ligation. The overall lower bp w values of the dynamic structures after adding the REases compared to P initial originate from entering the DySS polymerization with a ligase/REase balance. Using a REase balance with a stronger EcoRI activity, the situation can be inverted into enrichment in M N (green polymers; Figure 3 b). Also, a balance with a mixed copolymer structure can be obtained for similar cutting frequency of both REases (Figure 3 c). A symmetric reduction of the REase concentrations, while maintaining their ratio, leads to higher bp w , because lower overall [REase] causes higher degree of ligation (Figures 3 c,g/d,h).
These results confirm that this transformation process is governed by the kinetics of REases. Appropriate tuning allows to reconfigure one state (fully mixed copolymer with green and red fluorescence) into a new state with different constitutions and controllable material compositions (largely one type of sorted homopolymer with red/green fluorescence). We also find that more heterolinkers M L or less total monomer concentration can cause a shorter chain length ( Supporting Information, Figures S5, S6). Although fluorescence is used as a model function, it is obvious that conceptually other physical properties can be changed. [20] Next, we integrated the concept into a series to consecutively transform the same system and sort out different enriched states from an originally mixed P initial (Figures 4 a,b). In Round 1, after obtaining P initial , a REase mixture with more EcoRI was added for dynamization. As expected, red M E is sorted out and a green polymer enriched in M N appears. This degrades over time due to ATP consumption. Before Round 2, heating to 85 8C (1 h) was used to inactivate all enzymes. We began Round 2 with ligase and ATP to reobtain P initial followed by dynamization with REases containing less EcoRI, leading to a reconfigured polymer enriched in red M E . For Round 3, after generating P initial , the system is transformed back to the state of Round 1, although the activities of the enzymes have a slight loss owing to the accumulation of glycerol of repeated enzyme addition (Supporting Information, Figure S8).
We further developed a FastTrans protocol for rapid and in situ transformation, in which each step finishes in only one Figure S7) Figure S9): 0. before ligation; 1-6.
Step 1-6. hour and where the enzyme inactivation is omitted (Figures 4 c,d). By adding the enzymes consecutively, a homogeneous polymer with little sorting is generated at the start, as well as for later steps by ligase (re)supply (Steps 1,3,5), whereas the sorting to enriched green M N polymers (Steps 2,6), or red M E polymers (Step 4) requires the indicated REase ratios to favor restriction appropriately. Although the approach is less effective owing to enzyme accumulation, the in situ transformations with sorting into preferred polymers of one color/ monomer and short oligomers of the other color/monomer and re-homogenization are clearly visible in AGE and length evaluation. This approach would be beneficial when handling temperature-sensitive components, where no heating rejuvenation can be done.
To demonstrate a broader applicability, we showcase the reconfiguration of the surface functionality on colloids as important soft matter model systems [21] (Figure 5; Supporting Information, Figures S10-S12). After growing P initial chains on streptavidinylated beads using biotinylated DNA primers, we incubated these beads with different REase ratios to adaptively dynamize the surface-grown polymers. When exposed to a high EcoRI activity, the initial colloid surface with a mixed red and green fluorescence reconfigures to one with stronger green M N and low red M E (Figures 5 a,b). The color channel-specific intensity distributions across a particle clearly show the pronounced changes and underscore that the surface coating has strongly sorted out M E and is enriched in M N . The inverse result with more M E on the colloidal surface is obtained for higher NheI activity (Figure 5 c,f). When realizing similar activities of EcoRI and NheI, a surface layer containing comparable M E and M N is present (Figure 5 d,g). The reconfiguration on the surface can be obtained because the iERN dynamizes the whole polymer structure and leads to constant bond shuffling and unit exchange.
In summary, we introduced iERNs to demonstrate a simple and efficient concept, polymer transformers, that can adaptively reconfigure their state from mixed copolymers to sorted out polymer structures that are enriched in one monomer. REases which act on the individual species are used as signals for the adaptive response. The heterolinker (and to a lesser extent the joint use of the activating reaction) is the important component to interdigitate both otherwise independent ERN-based steady state structures. Driving the system in an ATP-driven DySS with control over the subunit exchange allows to implement high levels of dynamics and fine-tune the sorting out phenomena. Using this method, we can obtain various polymers with different compositions from the same starting material by simply tuning the REase ratio because of the enzymatically controlled process. The method is robust and allows repeated rejuvenation and several consecutive transformation cycles, even with a FastTrans protocol to achieve the in situ transformation in only one hour. As a first proof-of-concept application, we showcased reconfiguration of colloid surface functions. This method will promote concepts for adaptive materials design and assist the systems chemistry-based development of more intelligent materials design and applications, such as robotics or selfregulation systems. On a conceptual level, we believe that related approaches may be feasible in supramolecular chemistry, or classical polymer science, which however requires to improve regulatory networks in these systems. | 2020-07-02T10:33:31.307Z | 2020-07-01T00:00:00.000 | {
"year": 2020,
"sha1": "ea2765cdabc5300340c7d7f07b35c73773a4a8de",
"oa_license": "CCBYNCND",
"oa_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/anie.202006526",
"oa_status": "HYBRID",
"pdf_src": "PubMedCentral",
"pdf_hash": "6660e134bbdd40d3d2d64ac62c5ce6caed67af81",
"s2fieldsofstudy": [
"Materials Science",
"Chemistry"
],
"extfieldsofstudy": [
"Medicine",
"Materials Science"
]
} |
233412381 | pes2o/s2orc | v3-fos-license | The use of tonal coarticulation in segmentation of artificial language speech: A study with Mandarin listeners
Abstract Tonal carryover assimilation, whereby a tone is assimilated to the preceding one, is conditioned by prosodic boundaries in a way suggesting that its presence may signal continuity or lack of a boundary. Its possibility as a speech segmentation cue was investigated in two artificial language (AL) learning experiments. Mandarin-speaking listeners identified the “words” of a three-tone AL (e.g., [pé.tī.kù]) after listening to six long speech streams in which the words were repeated continuously without pauses. The first experiment revealed that segmentation was disrupted in an “incongruent-cues” condition where tonal carryover assimilation occurred across AL word boundaries and conflicted with statistical regularities in the speech streams. Segmentation was neither facilitated nor inhibited in a “congruent-cues” condition where tonal carryover assimilation occurred only within the AL words in 27% of the repetitions and never across word boundaries. A null effect was again found for the congruent-cues condition of the second experiment, where all AL word repetitions carried tonal carryover assimilation. These findings show that tonal carryover assimilation is exploited to resolve segmentation problems when cues conflict. Its null effect in the congruent-cues conditions might be linked to cue redundancy and suggest that it is weighted low in the segmentation cue hierarchy.
A crucial mental process during spoken language comprehension is the segmentation of continuous speech streams into discrete units such as words. Accumulated evidence has demonstrated that listeners exploit a wide variety of cues to facilitate this process (see, e.g., Cutler, 2012;Davis, Marslen-Wilson, & Gaskell, 2002, for reviews). These range from statistical regularities in the speech input to language-specific phonological patterns and acoustic-phonetic details. This study sets out to expand this line of investigation by experimentally testing whether tonal coarticulation, a low-level acoustic-phonetic phenomenon commonly observed in lexical tone languages, is used in speech segmentation. Before showing how tonal coarticulation could be useful, we present an overview of previous findings as awareness of what cues have been shown to support segmentation is helpful for designing an experiment for our purpose.
Despite its continuous nature, speech contains statistical regularities that provide useful boundary information. These regularities are usually expressed in terms of transitional probability (TP), which captures the likelihood that a pair of elements co-occur. In general, two consecutive syllables with a higher TP tend to be wordinternal and are more likely to be perceived as being so, whereas two syllables with a lower TP tend to occur across words and are more likely to be perceived as straddling a boundary (Mirman, Magnuson, Estes, & Dixon, 2008;Saffran, Newport, & Aslin, 1996;Saffran, Newport, Aslin, Tunick, & Barrueco, 1997). Such a segmentation solution is not exclusive to adult listeners as young children and even infants are also able to compute TPs to extract possible word forms (Aslin, Saffran, & Newport, 1998;Estes, Evans, Alibali, & Saffran, 2007;Hay, Pelucchi, Estes, & Saffran, 2011;Saffran et al., 1997;Thiessen & Saffran, 2007). Tracking statistical regularities in speech is therefore thought to be an ontologically early segmentation strategy, permitting discovery of potentially meaningful units before the emergence of an adultlike lexicon.
With these statistical computations supporting segmentation from an early age, listeners further develop various language-specific segmentation solutions through increasing experience with native-language phonological patterns. For example, Dutch phonotactics prohibit word-internal [mr] sequences (e.g., *[mrɒk] is not a possible Dutch word) and Dutch listeners hearing these sequences would assume a word boundary between the two consonants (McQueen, 1998). Vowel harmony in Finnish dictates that word-internal vowels agree in frontness/backness (Karlsson, 1983), inclining Finnish listeners to segment speech in such a way that two syllables belong to a single word when their vowels agree in this feature but to different words when they do not (Suomi, McQueen, & Cutler, 1997;Vroomen, Tuomainen, & de Gelder, 1998). Phonological patterns that promote segmentation solutions may also arise from the distribution of a phonological entity, such as lexical stress. In English, the majority of the words begin with stressed syllables (Cutler & Carter, 1987); thus, English listeners treat stressed or prominent syllables as word onsets and segment speech accordingly (Cutler, 1990;Cutler & Butterfield, 1992;Cutler & Norris, 1998;Tyler & Cutler, 2009).
In addition to phonological patterns, fine-grained acoustic-phonetic aspects of speech sounds may also help resolve segmentation problems. English listeners use subtle durational differences in [l] to disambiguate two lips and tulips (Gow & Gordon, 1995). Besides, due to the prosodic structuring of speech, segments in the initial position of a larger prosodic constituent are produced with stronger articulatory strengthening than those in the initial position of a smaller prosodic constituent (Cho & Keating, 2001;Fougeron & Keating, 1997;Keating, Cho, Fougeron, & Hsu, 2004). The stronger strengthening effect associated with a larger constituent has been shown to facilitate the search of word onsets (Cho, McQueen, & Cox, 2007). Similarly, the degree of coarticulation between segments carries boundary information. Adjacent segments are more strongly coarticulated within words than between words (Byrd, 1996;Byrd & Saltzman, 1998), or across a smaller prosodic boundary than across a larger one (Cho, 2004;Fougeron & Keating, 1997). Listeners can use such coarticulatory information to segment speech (Fernandes, Kolinsky, & Ventura, 2010;Fernandes, Ventura, & Kolinsky, 2007).
Fine-grained acoustic-phonetic details may modulate segmentation behavior to such an extent that they affect or even determine the use of a segmentation strategy that is motivated by a language-specific phonological pattern. Supporting evidence comes from recent studies with the use of tonal or fundamental frequency (F0) cues by Korean and Taiwanese Southern Min (TSM) listeners. Korean is thought to have a prosodic constituent called accentual phrase (AP), which frequently begins with a low (L) tone and ends in a high (H) tone (Jun, 1998) and Korean listeners tend to perceive H-L tone sequences as cueing an AP boundary (Kim, Broersma, & Cho, 2012;Kim & Cho, 2009). Moreover, Tremblay, Cho, Kim, and Shin (2019) suggest that the way in which the tone sequence is acoustic-phonetically realized affects how well it can be exploited for segmentation purposes. They found that Korean listeners' segmentation improved as the L tone in the tonal sequence became phonetically lower and more closely resembled the canonical realization of the AP-initial L tone in Korean. A related case was reported in Ou and Guo's (2019) study with TSM listeners. TSM is a tone language with an extensive tone sandhi process that restricts its only rising tone to the final position of the tone sandhi domain. It may thus be hypothesized that a cue as simple as a final rise in F0 suffices to signal finality for TSM listeners. Yet, given that the domain-final position is associated with phonetic final lengthening, it may be alternatively hypothesized that final lengthening is needed for a final F0 rise to be a sufficient finality cue. Ou and Guo's findings support the alternative hypothesis: final F0 rise alone did not improve TSM listeners' segmentation; instead, it was the combination of final F0 rise and final lengthening that did. Taken together, these two studies demonstrate that while some segmentation strategies are shaped by the distributions of phonological entities (e.g., the H-L tone sequence of the AP in Korean and the domain-final rising tone in TSM), they do not abstract away from the fine details of how those entities are acoustic-phonetically manifested.
To sum up, the literature has revealed at least three types of speech segmentation cues: (a) statistical regularities, (b) native-language phonological patterns, and (c) fine-grained acoustic-phonetic details. It is also found that acoustic-phonetic details may impact the use of phonological patterns, suggesting that they play a nontrivial role in shaping segmentation behavior. Research on acoustic-phonetic cues could thus contribute insight into how sensitive and resourceful listeners are in solving segmentation problems, a question that may inform theories and models seeking to reveal what cues are useful and how they are integrated (e.g., Mattys, White, & Melhorn, 2005). Yet, empirical work to date on this type of cues is mostly concerned with those at the segmental level (e.g., coarticulation of segments). Less attention has been paid to whether fine-grained tonal or F0 information is exploited. Perhaps the study that bears most on this question is that by Tremblay et al. (2019) discussed above. Nevertheless, while Tremblay et al. show that Korean listeners' use of the H-L tone sequence was affected by the phonetic manifestation of the L tone, the segmentation strategy that the listeners employed is phonologically motivated in nature. Subtle acoustic-phonetic changes to the scaling of the L tone do result in better use of the segmentation strategy if they create a more canonical realization of the tone, but they are not what lead Korean listeners to develop the strategy in the first place. However, there are detailed acoustic-phonetic tonal phenomena in speech that do not seem to depend on a particular phonological entity and can potentially promote segmentation strategies on their own. One of them is tonal coarticulation. In this study, we focused on a specific type of tonal coarticulation and investigated its effect on speech segmentation. Below is an introduction to lexical tones and tonal coarticulation.
Tonal Coarticulation
Lexical tones are pitch patterns over a syllable that serve to differentiate word meanings. For example, in Mandarin, [ma] means "mother" when bearing a high-level tone (Tone 1) but "hemp" when bearing a rising tone (Tone 2). The primary acoustic correlate of lexical tones is F0, and the F0 realizations of one tone in connected speech may vary under the contextual influence of adjacent tones, resulting in the so-called tonal coarticulation. Production experiments have revealed much evidence for tonal coarticulation in Mandarin and other lexical tone languages (see, e.g., Chen, 2012;Xu, 2001, for a review). Recently, Hao, Zhang, Xie, and Zhang (2018) proposed a scheme for annotating tonal coarticulation and applied it to speech samples from a Mandarin corpus. About 51% of bitonal syllable sequences in their data were labeled as being tonally coarticulated, suggesting that tonal coarticulation is prevalent in connected speech, at least in Mandarin. The exact effect of one tone on another is commonly described in terms of (a) whether it is assimilatory or dissimilatory and (b) whether its direction is anticipatory or carryover (e.g., Brunelle, 2009;Chang & Hsieh, 2012;Cheng, 1968;Peng, 1997;Potisuk, Gandour, & Harper, 1997;Shen, 1990;Shih, 1988;Xu, 1994;Zhang & Liu, 2011). A logical corollary of this is that there are four theoretically possible types of tonal coarticulation: carryover assimilation, carryover dissimilation, anticipatory assimilation, and anticipatory dissimilation. As an example, the right panel of Figure 1 is a schematic illustration of carryover assimilation, whereby the F0 contour of a tone is partially assimilated to the preceding tone.
Among the possible types of tonal coarticulation, carryover assimilation is of particular interest to the current investigation for two reasons. First, cross-linguistically, carryover effects are generally assimilatory, as evidenced by the fact that carryover assimilation is attested in a broad range of lexical tone languages, including Mandarin (Shih, 1988;Xu, 1997), Taiwanese (Cheng, 1968;Wang, 2002), Tianjin Chinese (Zhang & Liu, 2011), Thai (Gandour, Potisuk, & Dechongkit, 1994), Cantonese (Li, Lee, & Qian, 2004), Vietnamese (Brunelle, 2009), and so on. In contrast, findings on whether anticipatory effects are assimilatory or dissimilatory are Figure 1. Possible transitions between a high-level tone and a rising tone (adapted from Xu, 1997, p. 63). The left figure represents a situation in which there is no tonal coarticulation. The right one illustrates tonal carryover assimilation, whereby the initial portion of the rising tone's F0 contour changes into a falling F0 transition due to assimilation to the preceding high-level tone.
relatively mixed as these effects are reported to vary across languages or even across different tones of the same language (Zhang & Liu, 2011). Second, and more importantly, it has been shown that as with segmental coarticulation, tonal carryover assimilation is conditioned by prosodic boundary strength. In Mandarin, it tends to be stronger when two adjacent tones span the boundary of a smaller prosodic unit than when they span that of a larger unit (Lai & Kuang, 2016). Such a tonal coarticulatory effect may even be completely eliminated when the neighboring syllables straddle a major prosodic break (Zhang & Kawanami, 1999). These suggest that tonal carryover assimilation may be useful for segmenting continuous speech into discrete units.
The Current Study
The goal of the present work is to experimentally test this possibility. As with many empirical studies, we attempt to draw conclusions about a single cue, which, in our case, is tonal carryover assimilation. Nevertheless, cues can rarely be isolated from each other even in well-controlled laboratory speech materials. For example, while Fernandes et al.'s (2007) listeners exploited segmental coarticulation in segmenting an artificial language (AL), TP information was always present in the stimuli. It is therefore instructive to consider two possible scenarios for segmentation in the presence of multiple cues as the considerations could inform the experimental design and result interpretation. One scenario is that the cues operate in cooperation. A possible outcome is that their effects are additive or even synergistic, enhancing segmentation to a greater extent than a single cue alone does. An example is Fernandes et al., in which segmentation was better when segmental coarticulation and TP information were present and congruent with each other than when TPs were the sole cue. Yet, cooperating cues may be redundant and may not produce extra facilitation. Bagou and Frauenfelder (2018) and Kim et al. (2012) showed that although French and Korean listeners benefited from final lengthening and final F0 rise in isolation, conjoining these two prosodic cues does not improve their performance further. The other scenario is when cues operate in conflict and there are again two possible outcomes. One is that the conflict leads to inhibition. For example, Ordin, Polyanskaya, Laka, and Nespor (2017) found that Italian listeners used vowel lengthening to locate word-medial positions; therefore, when it was the vowels in the word-initial positions that were lengthened, TP-based segmentation was disrupted, reducing performance to a level worse than that of a condition with TPs as the only cue. Alternatively, the cue conflict may neither facilitate nor inhibit segmentation.
With these two scenarios in mind, we investigated the use of tonal carryover assimilation with the AL learning technique. It is an experimental paradigm that has been widely adopted to explore how phonological patterns and acousticphonetic details guide segmentation (e.g., Bagou & Frauenfelder, 2018;Fernandes et al., 2007;Kim et al., 2012;Ordin & Nespor, 2016;Ordin et al., 2017;Toro, Pons, Bion, & Sebastián-Gallés, 2011;Tremblay et al., 2019;Tyler & Cutler, 2009). A typical AL learning experiment has a learning (or exposure) phase followed by a test phase. In the learning phase, participants learn an AL by listening to long speech streams in which tokens of the "words" of the AL, which are meaningless syllable sequences, are concatenated without pauses in between. The basic cue for word segmentation is TP. For example, as mentioned, adjacent syllables with a lower TP are more likely to span a boundary. On top of TPs, additional cues may be introduced to examine how they impact segmentation. After listening to the speech streams, participants complete a two-alternative forced-choice test in which they hear a word of the AL and a sequence that is not part of the AL vocabulary and have to select the former. The proportion of correct selections is thought to reflect how well the AL speech streams were segmented during the learning.
Such an experiment has two important advantages for research concerned with phonological or acoustic-phonetic cues. First, as it has been suggested that segmentation is primarily lexically driven (Mattys & Bortfeld, 2017;Mattys et al., 2005), using nonsense speech prevents listeners from segmenting based on lexical knowledge (e.g., by lexical subtraction: White, Melhorn, & Mattys, 2010) and allows the researcher to obviate confounds such as word frequency. Second, with artificial speech, one can precisely control the acoustic-phonetic content of the additional non-TP cues. Two AL learning experiments were conducted in this study. Although their focus is on tonal carryover assimilation, as noted above, it is useful to also consider its possible effects in the presence of TP information, the basic segmentation cue in an AL learning task. We thus constructed conditions corresponding to the two scenarios discussed. Details about the design and the hypotheses to test are presented below.
Experiment 1
Participants Ninety-six adult native speakers of Mandarin (28 males and 68 females) with no self-reported history of hearing impairments were recruited from a university in Southern Taiwan. Their mean age in years was 20.3 (range: 18-23; standard deviation: 1.4). They had been learning English as a compulsory subject in school, and 9 of them had received musical training. 1 They were randomly assigned to one of the three experimental conditions (see the Design and stimuli section). The single-cue, congruent-cues, and incongruent-cues conditions had 31, 33, and 32 listeners, respectively.
Design and stimuli
In the learning phase, participants were exposed to a nonsense tonal language under three conditions, the design of which was modeled after Fernandes et al.'s (2007) study with segmental coarticulation. One was called the "single-cue" condition, in which participants could only rely on TPs to segment the AL speech streams. In addition to TPs, tonal carryover assimilation was introduced to the speech streams in the other two conditions. In the "congruent-cues" condition, tonal carryover assimilation occasionally occurred between the syllables within an AL word but never across word boundaries. Therefore, the tonal coarticulatory cue agreed with the TP cue. In the "incongruent-cues" condition, however, these cues were pitted against each other by letting tonal carryover assimilation occur across AL word boundaries. The congruent-cues and incongruent-cues conditions corresponded to the situation in which the cues are in harmony and the situation in which they are in conflict, respectively.
As with several studies (e.g., Kim et al., 2012;Ordin et al., 2017;Saffran et al., 1996;Vroomen et al., 1998), we created an AL consisting of six trisyllabic words, which were meaningless sequences of consonant-vowel syllables, as listed in the second column of Table 1. The words were formed by four vowels ([a, i, u, e]), three consonants ([p, t, k]), and three level tones (high-, mid-, and low-level tones). These consonants and vowels are cross-linguistically common and occur in Mandarin at least at the phonetic level. Irrespective of the tones, all the used consonant-vowel syllables except for [ki] are phonotactically possible in Mandarin. One constraint imposed during the construction of the AL lexicon was that adjacent syllables in a word had to differ by one tone level. For example, a high-level tone could only be preceded and followed by a mid-level tone, not by itself or by the low-level tone. Thus, there could be only six tone patterns, as shown in the first column of Table 1. In these patterns, neighboring tones always had different tone heights, creating what is referred to in the literature as "conflicting tonal contexts" (e.g., Peng, 1997;Xu, 1994) and allowing us to implement tonal carryover assimilation for every two adjacent tones in an AL word (and also a partword, described below). The fact that adjacent tones differed by one tone level also enabled us to control for the magnitude of carryover assimilation, so that, for instance, there was no carryover assimilation between the high-level and low-level tones, which would span a wider F0 range than that between the middle-level and low-level tones.
The syllables making up the AL words were individually inserted into a carrier sentence (i.e., /wuo ʂuo ____ ʈ ͡ ʂɤ kɤ ʈ ͡ sɹ̩ / "I said the word ____.") and read by a male native speaker of Mandarin with phonetic training in a monotone into a Zoom H4n Handy Recorder. The recorded items were digitized at a sampling rate of 44.1 kHz and stored as a single WAV file. The syllables were excised from the carrier sentence and then underwent manipulations using Praat (Boersma & Weenink, 2018). Their root-mean-squared amplitudes were equalized and their durations were normalized to 335 ms, which was the mean duration of the original, unmanipulated syllables. Note: The acute (´), macron (¯), and grave (`) marks represent high-level, mid-level, and low-level tones, respectively. The dots indicate syllable boundaries.
Next, their F0 contours were flattened, set to 126 Hz (the average F0 of the syllables prior to the manipulations), and resynthesized using the overlap-add method in Praat. The resulting flat F0 contour served as the mid-level tone. Following Caldwell-Harris, Lancaster, Ladd, Dediu, and Christiansen (2015), we created the high-level and low-level tones by shifting the F0 contour up and down, respectively, by 3.5 semitones. The manipulated syllables were concatenated to form the six AL words. In addition, six "partwords," which are listed in the third column of Table 1, were created using the same set of manipulated syllables. They served as the distractor stimuli in the test phase and were trisyllabic sequences derived by combining the last syllable of an AL word with the first two syllables of another AL word, or by combining the last two syllables of an AL word with the first syllable of another AL word. They were constructed under the same constraint for the AL words and therefore had the same six tone patterns. This prevented participants from being able to easily reject the partwords by identifying novel tone patterns.
The learning-phase stimuli were six speech streams in which tokens of the AL six words were concatenated with no pauses in between. Each stream contained a total of 120 tokens, 20 for each AL word. These tokens appeared in a random order but under the restriction that the same word did not occur twice in a row. As in Tyler and Cutler (2009), the first and last 5 s of each stream were faded in and out to prevent participants from hearing the syllables at the beginning and the end of the stream and using them to discover word boundaries. Each stream was 2 min long and the total duration of the six streams (and hence the learning phase) was about 12 min. The TP for each pair of adjacent syllables AB in the streams was calculated using the formula proposed in Saffran et al. (1996); that is, it is equal to the frequency of AB divided by the frequency of A. For each AL word or partword, an average TP was computed by taking the average of the TP between the first and second syllables and that between the second and third syllables. The average TPs for the AL words ranged between 0.75 and 1.00 (mean: 0.88) and those for the partwords ranged between 0.32 and 0.62 (mean: 0.49). The six speech streams differed from each other in the order in which the AL words appeared but were the same across the three conditions except for the presence or absence of tonal carryover assimilation and the way in which it was introduced. In the single-cue condition, there was no carryover assimilatory effect from one tone on the next tone and listeners could only achieve segmentation by tracking TPs.
In the congruent-cues and incongruent-cues conditions, tonal carryover assimilation was introduced as an additional cue. In each stream, a fixed number of trisyllabic sequences was selected to receive tonal carryover assimilation. In the incongruentcues condition, they corresponded to all instances (i.e., 100%) of the partwords in the speech stream. However, the partwords occurred only incidentally and made up just about 27% of the syllables in the stream. To ensure that the incongruent-cues and congruent-cues conditions differed only in the alignment of the tonal coarticulatory cue with word boundaries but not in the number of the syllables with the cue (as in Fernandes et al., 2007), only 27% of the tokens of the AL words in the congruent-cues condition were selected as the trisyllabic sequences that would receive tonal carryover assimilation. This tonal cue was implemented by performing the following F0 manipulations on the selected trisyllabic sequences. First, the F0 onsets of the second and third syllables were raised or lowered to the F0 offsets of their immediately preceding syllables (i.e., the first and second syllables, respectively). Second, between the shifted F0 onset and the 25% time point into the F0 contour of the second or third syllable, a smooth F0 transition was interpolated quadratically using Praat. The erstwhile flat F0 contour of a level tone then had an F0 rise over the initial quarter of its F0 contour if its preceding tone was lower and an F0 fall if its preceding tone was higher. Note that in actual Mandarin tone production, the assimilatory effect exerted by the preceding tone can be far more extensive: for example, it may still be evident even at the 75% time point of the next tone (e.g., Xu, 1997). Manipulating only the initial 25% of the F0 contour of a syllable allowed us to evaluate the influence of tonal carryover assimilation conservatively. Shown in Figure 2 are samples of speech stream from each condition.
The test phase consisted of a two-alternative forced-choice test. In each trial, two stimuli (a word of the AL and a partword) were presented successively with 500 ms of silence in between. The stimuli did not have the tonal carryover assimilation cue and were the same for all conditions. Therefore, the conditions differed only in the learning phase, specifically, in whether and how carryover assimilation was introduced to the speech streams. The orders in which the AL word and partword were presented in a trial were counterbalanced. There were 36 trials in total, yielded by pairing the six AL words exhaustively with the six partwords. E-prime 2.0 software (Psychology Software Tools, 2012) was used to control stimulus presentation and record responses.
Procedure
Participants were tested individually in front of a desktop computer in a soundattenuated booth. They were told to learn an AL by listening to six prerecorded sound files of that language (i.e., the six learning-phase speech streams). They were not given any cues such as the length or number of the words in the AL. They were instructed to pay as much attention as possible to what they heard and made aware of an upcoming test that would assess their knowledge of the AL. They were allowed to take a short break after finishing listening to each sound file. After the learning phase, they immediately proceeded to the test, in which they heard two stimuli in a row in each trial. They were asked to select the one that they thought was a word of the AL by pressing the button on a response box that corresponded to the order of presentation of the word (i.e., button "1" or "2"). There was a 10-s response timeout after the second stimulus. Participants first completed three practice trials presenting nonsense syllable sequences not used in the AL to familiarize themselves with the procedure. They completed the practice by arbitrarily pressing any button on the response box but were reminded that they had to choose the AL words in the test proper.
Hypotheses and predictions
The single-cue condition served as the baseline for comparison with the other two conditions, based on which hypotheses regarding the effect of tonal carryover assimilation were tested. For the congruent-cues condition, the hypothesis was that tonal carryover assimilation in agreement with TPs contributes to segmentation above and beyond TPs. This predicted that listeners' selection accuracy in the test would be significantly higher in the congruent-cues condition than in the single-cue one (as in Fernandes et al., 2007). As for the incongruent-cues condition, where the partwords received tonal carryover assimilation, the hypothesis of interest was that the conflict between the tonal cue and TPs would impede segmentation. Should this be the case, the listeners in the incongruent-cues condition would respond significantly less accurately compared with the single-cue one (much in the same way as the Italian listeners exposed to initial lengthening in Ordin et al., 2017). Findings lending support for the facilitation in the congruent-cues condition or the inhibition in the incongruent-cues condition may be interpreted as evidence for the use of tonal carryover assimilation.
Results and discussion of Experiment 1
The listeners' responses in the test were analyzed. Timeouts (i.e., no responses within 10 s) accounted for about 0.43% of all observations and were excluded. In the remaining data, a response was coded as correct when the AL word in the trial was selected and as incorrect when the partword was selected. Displayed in Figure 3 are the mean percentage of correct responses of each condition along with those of individual participants. A linear mixed-effects logistic regression model was fitted to the data by using the glmer() function from the lme4 package (Bates, Mächler, Bolker, & Walker, 2015) in R (R Core Team, 2017) to examine the effects of the experimental conditions. The dependent variable was the response to each test trial, which was either correct or incorrect. The fixed effect of central interest was condition, with the single-cue condition being the baseline level. Two additional predictors were also entered as fixed effects to partial out their impact on responses. First, the order in which a given trial appeared in the test (trial) was included to control for possible fatigue or practice effects. Second, as recommended in Ou and Guo (2019), the log-transformed reaction time (LogRT) was included to capture any potential trade-offs between response accuracy and latency. Both trial and LogRT were centered and scaled. All the fixed-effect predictors were entered as main effects only. Following Barr, Levy, Scheepers, and Tily's (2013) recommendation, we used the maximal converging random-effects structure supported by the data. For Experiment 1, the random-effects structure consisted of a by-participant random intercept and a by-item random intercept for partwords. Table 2 shows the results of the mixed-effects model. Trial was significant, with responses in later trials being less accurate than those in earlier trials, possibly due to fatigue. LogRT was significant as well, indicating an inverse correlation between response accuracy and latency (i.e., faster responses were more likely to be correct than slower ones). As suggested in Ou and Guo (2019), this correlation might be merely an artifact of response certainty: as the two stimuli were separated by 500 ms of silence, listeners might be ready to respond if they were sure that the first stimulus was an AL word or a partword. Most importantly, there was a significant main effect of condition, which indicated that response accuracy in the incongruent- cues condition (mean: 57.58%) was generally lower than that in the single-cue one (mean: 63.38%). Therefore, pitting the tonal carryover assimilation cue against TP information (by letting the cue-bearing sequences span word boundaries) hinders segmentation, reducing the listeners' performance to a level below that of a condition where TPs are the sole cue. However, the other main-effect term of condition showed that response accuracy in the congruent-cues condition (mean: 65.39%), where tonal carryover assimilation occurred within word boundaries, was not significantly different from that of the single-cue one. There is thus no evidence that the listeners' segmentation is better or worse when the TP and tonal carryover assimilation cues occur in tandem and in a cooperative manner than when TP information is the only cue.
Experiment 1 examined the use of tonal carryover assimilation by Mandarin listeners in speech segmentation with an AL learning task. One proposed hypothesis predicted that relative to that of the single-cue condition, the response accuracy of the incongruent-cues condition would be significantly lower and this prediction was borne out. The finding is consistent with Fernandes et al.'s (2007) study with segmental coarticulation, in which segmentation in the incongruent-cues condition was worse than in the single-cue one. Analogous results have also been reported by AL learning research demonstrating that compared with a TP-only condition, segmentation is disrupted when an additional prosodic cue appears in a position that is unexpected in view of phonological patterns in the listeners' native language (e.g., Ordin et al., 2017). The Mandarin listeners' disrupted segmentation performance in the incongruent-cues condition of the present study may be interpreted as reflecting their attempts to use the tonal assimilation cue, even though such use is in conflict with TP regularities. Such disruption of segmentation should not be possible if the cue had not been exploited at all.
With regard to the congruent-cues condition, we hypothesized that the congruent tonal carryover assimilation cue would facilitate segmentation above and beyond the effects of TPs. The results did not support the hypothesis as there was no significant difference in response accuracy between the single-cue and congruentcues conditions. This is not consistent with Fernandes et al. (2007), who did find that with segmental coarticulation as the additional non-TP cue, segmentation under the congruent-cues condition was clearly better than under the single-cue one. Rather, it seems compatible with an alternative view: adding congruent tonal carryover assimilation would not facilitate segmentation because word boundaries are redundantly cued by the tonal and statistical information. Higher TPs between adjacent syllables within an AL word already signal that a boundary between them is unlikely and there might be no need for conveying similar information via tonal coarticulation. Such a cue redundancy hypothesis may lead one to expect no significant difference in the listeners' accuracy between the single-cue and congruent-cues conditions (as in the case with the final lengthening and F0 rise in Bagou & Frauenfelder, 2018;Kim et al., 2012). However, the lack of a significant effect of cue congruence could potentially be attributed to a confounding factor: cue reliability. Recall that only 27% of the AL word tokens in the congruent-cues condition carried tonal carryover assimilation. This percentage was rather low considering the percentage of bitonal sequences that showed a tonal coarticulatory effect (i.e., about 51%) as reported by Hao et al. (2018) for Mandarin corpus speech. Therefore, an alternative view was that our listeners were unable to benefit from tonal carryover assimilation under the congruent-cues condition simply because it was not reliably present, not because it was redundant with TP information. Specifically, as cue reliability is associated with cue weight (e.g., Mattys et al., 2005;Seidl, 2007;Tremblay, Spinelli, Coughlin, & Namjoshi, 2018), they might have allocated a relatively low weight to the unreliably present tonal carryover assimilation cue. This is not entirely impossible given that, however prevalent tonal coarticulation is in Mandarin, the listeners were instructed to learn a completely novel AL and they might develop a cue hierarchy for that AL, one in which tonal carryover assimilation was so lowly weighted that it failed to produce any extra gain in segmentation performance.
The main goal of Experiment 2 is then to test this cue reliability hypothesis, that is, to examine whether enhancing the reliability of the tonal carryover assimilation cue in the congruent-cues condition would facilitate segmentation and provide an alternative explanation for the null effect in Experiment 1. Currently, it is unclear as to how many cue-bearing tokens would count as sufficient for obtaining the kind of facilitation effects reported by Fernandes et al. (2007), who did not provide the exact percentage of segmentally coarticulated tokens in their congruent-cues condition. Yet, it would be insightful to test an experimental condition in which the reliability of the tonal carryover assimilation cue is maximized, namely, one in which all the tokens of the AL words receive the cue. Such a condition was included in Experiment 2. In addition, Experiment 2 included the same singlecue and incongruent-cues conditions from Experiment 1. This was done to ensure that participants had been randomly assigned to different conditions while the conditions were being compared and to examine whether the findings of Experiment 1 could be replicated.
Participants
Ninety new adult native listeners of Mandarin (35 males and 55 females) from the same population as those in Experiment 1 were recruited. Their mean age in years was 20.7 (range: 18-25; standard deviation: 2.0) and 15 of them had received musical training. They were randomly and equally assigned to the single-cue, congruentcues 100%, and incongruent-cues conditions. None participated in Experiment 1.
Design and stimuli
The overall design of Experiment 2 was as in Experiment 1. The difference was that for the congruent-cues condition, all instances of the AL words (not just 27%) in the learning-phase speech streams received the tonal carryover assimilation cue, as shown in the lower panel of Figure 4. We refer to this condition as "congruent-cues 100%." The test-phase stimuli were the same as in Experiment 1.
Procedure
The procedure was identical to that of Experiment 1.
Hypothesis and prediction
If it was cue (un)reliability that underpinned the lack of a significant effect of cue congruence in the previous experiment, it was hypothesized that the tonal carryover assimilation cue would be effectively exploited if it was always present in the AL words. This predicted that listeners' response accuracy in the test would be significantly higher in the congruent-cues 100% condition than in the single-cue one.
Results and discussion of Experiment 2
Again, responses in the test phase were analyzed. Timeouts (less than 0.13% of the data) were discarded, and a response was correct when the AL word was selected and incorrect when the partword was selected. Shown in Figure 5 are the percentage of correct selections of each participant and the mean of each condition. A linear mixed-effects logistic regression model with the same fixed-effects was fitted; that is, it contained the main effects of condition (baseline: single-cue), trial, and LogRT. As in Experiment 1, the maximal converging random-effects structure was used. This time it included random intercepts for participants, AL words, and partwords.
The results of the analysis are in Table 3. The effects of trial and LogRT were significant and in the same direction as in Experiment 1. Response accuracy was lower in later trials, possibly because of fatigue; faster responses tended to be correct and this might be an artifact of response certainty. Most crucial were the two condition main effects. First, the response accuracy in the incongruent-cues condition (mean: 55.05%) was significantly lower than that in the single-cue one (mean: 63.97%), as in Experiment 1. Second, and more importantly, the response accuracy in the congruent-cues 100% condition (mean: 64.23%) was neither significantly better nor worse than that in the single-cue one, again suggesting no evidence that the congruence between statistical and tonal coarticulatory cues would enhance segmentation performance.
The primary goal of Experiment 2 was to test the cue reliability hypothesis proposed above. In particular, it investigated whether Mandarin listeners' segmentation would be facilitated under a condition in which all occurrences of the words in the AL speech streams carried the tonal carryover assimilation cue, as compared with the single-cue condition. The results provided no support for the hypothesis. Furthermore, the experiment replicated the inhibitory effect of cue incongruence. In the General Discussion section below, we will summarize the results from the two experiments, consider some possible explanations, and offer a unified account for the findings.
General Discussion
This study investigates the role of tonal coarticulation in the segmentation of continuous speech. Previous work has demonstrated that subtle acoustic-phonetic details are exploited in solving segmentation problems (e.g., Cho et al., 2007;Gow & Gordon, 1995) and can even modulate the use of segmentation strategies motivated by the distribution of a phonological entity (e.g., Ou & Guo, 2019;Tremblay et al., 2019). To extend this line of work, we examined the effect of tonal carryover assimilation, a cross-linguistically attested type of tonal coarticulation, on Mandarin listeners' segmentation with two AL learning experiments. The findings are summarized in Table 4. Experiment 1 revealed that response accuracy in the test was significantly lower in the incongruent-cues condition, where tonal carryover assimilation was pitted against TPs, than in the single-cue condition, where TP information was the only segmentation cue. This reflects the listeners' attempts to exploit the tonal cue despite its conflict with TPs. Yet, in the congruent-cues condition, where the tonal cue and TPs agreed with each other, response accuracy did not differ significantly from that of the single-cue condition. This seems consistent with the view that congruent tonal carryover assimilation is unable to improve segmentation further because it is redundant in the presence of TPs. Experiment 2 examined Mandarin listeners' segmentation in the congruent-cues 100% condition, where tonal carryover assimilation in agreement with TPs was introduced to all instances of the AL words. The results again showed no significant accuracy difference between this condition and the single-cue one, discounting the possibility that the null effect of cue congruence can be attributed to cue reliability. Experiment 2 also replicated the inhibitory effect of cue incongruence. Below we explore a few possible explanations for the findings, discuss their implications for tonal coarticulation as a fine-grained acoustic-phonetic cue in speech segmentation, and point out some further issues. One noteworthy finding from Experiment 1 was the null effect of cue congruence. It was found again in Experiment 2, ruling out the cue reliability hypothesis. Yet, one might wonder whether the null effect is simply a methodological artifact stemming from the stimuli used in the test phase. Recall that the test stimuli did not contain the carryover assimilation cue and were identical across the conditions. It could be argued that the listeners did not show significantly higher accuracy in the two congruent-cues conditions due to mismatch between the learned representations and the test stimuli. That is, the representations of the AL words that they built up during the learning phase contained the tonal coarticulatory cue and somewhat deviated from the AL words actually presented in the test. This possibility gains support from episodic or exemplar-based theories of phonology and speech perception (Bybee, 2000;Goldinger, 1996;Goldinger & Azuma, 2004;Johnson, 1997;Pierrehumbert, 2001). These theories hold that representations of categories are constructed from remembered instances, or exemplars, of those categories. Detailed acoustic-phonetic traces associated with these exemplars may be incidentally encoded in memory and affect speech processing, even though they are not crucial for category distinctions. For example, words and sentences presented before are recognized slower or less accurately when they are presented again in a novel voice than in a familiar voice, suggesting that speaker voice information is retained in memory (Craik & Kirsner, 1974;Geiselman & Bellezza, 1977;Palmeri, Goldinger, & Pisoni, 1993). A similar case could possibly be made for the listeners in our study: they did not benefit from cue congruence because all or at least some exemplars in their AL word representations contained tonal coarticulatory distortions, preventing them from effectively recognizing the stimuli in the test.
Such an exemplar-based explanation, however, is untenable in view of the present results for two reasons. First, if the listeners' responses were driven by exemplars, one would expect their test accuracy to be significantly higher in the single-cue condition than in any other condition as the tokens of the AL words presented in the learning phase of the single-cue condition were acoustically identical to those in the test. This expectation was not borne out. Second, several AL learning studies (e.g., Fernandes et al., 2007;Kim et al., 2012;Ordin & Nespor, 2016;Tremblay et al., 2019) also presented "uncued" stimuli (i.e., stimuli that did not contain the cues of interest) in the test and thus the test AL words were always the same as those in the learning phase of their single-cue or TP-only conditions. None of them reported that listeners performed the best in these conditions. What seems to be a more plausible explanation for the null effect in Experiments 1 and 2 is one of cue redundancy. In the two congruent-cues conditions, TPs between syllables already signal the presence or absence of boundaries and similar information is redundantly provided by tonal carryover assimilation. The consequence is that the congruence between the TP and tonal cues does not yield an extra gain in segmentation performance for Mandarin listeners. Comparable findings have been reported in previous studies with F0 and lengthening cues (e.g., Bagou & Frauenfelder, 2018;Kim et al., 2012). Nevertheless, the redundancy hypothesis needs to be reconciled with the findings of Fernandes et al. (2007), who show that, as mentioned, segmental coarticulation in agreement with TPs leads to extra facilitation when compared with TP information alone. In their study, cue redundancy does not seem to detract from the efficacy of segmental coarticulation as a useful segmentation cue.
Before suggesting how this discrepancy may be accommodated, it should be noted that tonal and segmental coarticulation are possibly rather different in nature, and this needs to be borne in mind in directly comparing the two phenomena. Although there are no prior studies analyzing how they would differentially affect segmentation, production experiments have revealed some differences between the two. For example, while tonal coarticulation is restricted to two contiguous tones (as in our AL), segmental coarticulatory effects can extend up to four segments (Shen, 1990). The scope of segmental coarticulation was not controlled in Fernandes et al. (2007). As a result, direct comparisons between tonal and segmental coarticulation would be the most appropriate only when factors like this are kept equal or systematically manipulated. Yet, an explanation for the discrepant findings on the tonal and segmental coarticulation may offer insight for rethinking previous models of speech segmentation cues.
We assume that redundancy affects tonal coarticulation but not segmental coarticulation in the presence of TPs possibly because tonal information is a less powerful cue than segmental information in speech segmentation, at least for Mandarin listeners. It has been shown that for listeners of Mandarin or other lexical tone languages, tonal cues are relatively disadvantaged compared with segmental cues as the former become available at later stages of auditory processing (Cutler & Chen, 1997;Sereno & Lee, 2015;Taft & Chen, 1992;Ye & Connine, 1999). Support for this comes from, for example, Sereno and Lee (2015), whose auditory priming and lexical decision experiments found that the prime facilitated recognition of the target when the two had overlapping segments but mismatching tone. By contrast, no facilitation was found when the prime and targets had overlapping tone but mismatching segments. Tong, Francis, and Gandour (2008) present converging evidence from Mandarin listeners and offer an explanation from an information-theoretic perspective. They argue that the tone disadvantage may be attributed to the fact that the tonal inventory is smaller than the segmental inventory. As a result, tones exert fewer constraints on lexical access and are less informative. Due to their experience with the weaker role of tonal information in speech processing in general, listeners may have allocated a relatively low weight to tonal segmentation cues such as tonal carryover assimilation. Such cues may be ignored when the boundary information they provide is redundant.
The idea that tonal coarticulation is weighted low or at least lower than segmental coarticulation suggests a few refinements to the current understanding of segmentation cue weight. Based on a series of experiments, Mattys et al. (2005) propose a three-tier segmentation cue hierarchy in which lexical cues are top-ranked, followed by segmental cues such as segmental coarticulation and then by prosodic cues such as word stress. It is not clear how tonal coarticulation can fit into this framework. On the one hand, tonal coarticulation is neither a segmental phenomenon nor a phonological or metrical feature like word stress. On the other hand, as with segmental coarticulation, it is a fine-grained acoustic-phonetic detail, and it involves variations in F0, which is also an acoustic correlate of stress (e.g., Beckman, 1986;Fry, 1958;Gay, 1978;Lieberman, 1960). Tonal coarticulation, which seems to have a lower weight than segmental coarticulation, may be accommodated by expanding the hierarchy with an additional cue category between the segmental and prosody tiers. Alternatively, it can be subsumed under the prosody category, but it has to be noted that the prosodic cue in Mattys et al.'s model currently refers to lexical stress and evidence for its relative importance comes only from experimentation with stress in English. Further investigation can be conducted to pinpoint where tonal coarticulation stands in the cue hierarchy. Yet, as far as the present findings are concerned, some modifications to current models of cue weight or ranking may be necessary, especially if they are to be adapted to account for the segmentation behavior of tone-language listeners.
Although the tonal carryover assimilation cue, unlike the segmental coarticulatory one in Fernandes et al. (2007), did not significantly improve segmentation when congruent with TPs, it did have an appreciable effect in the incongruent-cues conditions of the two experiments. In these conditions, tonal carryover assimilation favored the segmentation of the AL speech streams into what were defined as partwords, which contained a dip in TP (as they spanned an AL word boundary). This resulted in a conflict between prosodic and statistical information and hence hampered the listeners' segmentation. It is argued that the mechanism that drives the listeners to use tonal carryover assimilation might be one akin to the prosody analyzer proposed by Cho et al. (2007). The prosody analyzer computes the prosodic structure of an utterance using available suprasegmental information and generates possible segmentation hypotheses. In the case of the present study, the tonal carryover assimilatory effect of one syllable on the next syllable may be analyzed by a similar mechanism as signaling continuity or the lack of a prosodic boundary between the two syllables, therefore leading the listeners to perceive the two syllables as belonging to a unit. Segmentation is disrupted when the boundaries of prosodically and statistically defined units do not align. Similar cases of disruption have been reported in previous studies examining the effects of prosodic grouping and statistical regularities. For example, Shukla, Nespor, and Mehler (2007) exposed listeners to recurrent nonsense words presented in continuous speech with recurrent F0 frames that spanned several syllables. In a subsequent test, it was found that these nonsense words were recognized more poorly if they previously straddled the boundary of two F0 frames than if they did not.
It is concluded that tonal carryover assimilation can be exploited as a speech segmentation cue at least by Mandarin listeners. The use may be the result of a segmentation mechanism similar to the prosody analyzer, one that interprets the cue as signaling the absence of a prosodic boundary. An issue for further exploration concerns the extent to which such a mechanism is cross-linguistic. Tonal carryover assimilation was chosen for investigation because carryover effects tend to be assimilatory across languages (Zhang & Liu, 2011). Yet, there are exceptions. One of them is Malaysian Hokkien. Chang and Hsieh (2012) elicited productions of disyllabic words by speakers of this language and found that the carryover effect of the first tone on the second one was not assimilatory and even slightly dissimilatory. They attribute this to the final prominence in Malaysian Hokkien tone sandhi system, which requires the tone in the final position to be faithfully realized and militates against coarticulatory distortions by the preceding tone. Analogous findings have been reported by Chen, Wiltshire, and Li (2018) for Nanjin Chinese. Thus, it is possible that tonal carryover assimilation would be not be used by Malaysian Hokkien and Nanjin Chinese listeners, whether it is congruent with TPs or not. Support for such a possibility would suggest that the use of some putatively cross-linguistic acoustic-phonetic cues can be overridden by language-specific phonology. There has been some evidence for this. For example, while final lengthening has been thought to be universal and cross-linguistically useful for segmentation (e.g., Hay & Diehl, 2007;Klatt, 1975;Lindblom, 1978;Oller, 1973;Tyler & Cutler, 2009), recent evidence shows that Italian listeners exploit medial but not final lengthening presumably because stress in Italian falls predominantly on the penultimate syllable of a word (Ordin et al., 2017). Given that the use of a segmentation strategy motivated by phonological patterns is affected by subtle acoustic-phonetic details (e.g., Ou & Guo, 2019;Tremblay et al., 2019), it may be of interest to further examine the influence in the opposite direction, that is, how segmentation solutions motivated by fine-grained acoustic-phonetic information are constrained by language-specific phonology.
Still, it has to be recognized that as far as the current findings are concerned, the contribution of tonal carryover assimilation to segmentation is somewhat limited. It is not observed in the case of cue congruence and, as discussed, this might be linked to cue redundancy and the relatively minor role of tonal information in speech processing. Further issues also arise as to whether the limited contribution of tonal carryover assimilation can also be attributed to methodological factors, such as the fact that only the initial 25% of a tone's F0 contour was manipulated. Such manipulation allowed us to evaluate the impact of tonal carryover assimilation conservatively but might also lead us to underestimate it as the carryover assimilatory effect in naturalistic speech can be more extensive (e.g., Xu, 1997). In addition, despite being commonly adopted, AL learning is not the only experimental technique for studying speech segmentation. The AL learning task assesses segmentation hypotheses about a nonsense language using an offline forced-choice test. It would be insightful to further examine whether tonal carryover assimilation would have a more salient effect in the online segmentation of real meaningful speech, which can be captured by using time-sensitive measures such as eye-tracking fixations (e.g., Tremblay et al., 2018) or lexical decision latencies (e.g., Gow & Gordon, 1995;Mattys et al., 2005).
Conclusion
The present work aims to add to the understanding of the role of fine-grained acoustic-phonetic information in speech segmentation by examining how tonal carryover assimilation is used by Mandarin listeners in segmenting continuous speech streams of an AL. Experiments 1 and 2 found that their segmentation performance was hampered in the incongruent-cues conditions, suggesting that the tonal carryover assimilation cue was exploited despite its conflict with TPs. This might reflect a segmentation mechanism that analyzes the assimilatory effect of one tone on the next tone as cueing the absence of a boundary. However, the contribution of such an effect may be limited as Experiment 1 revealed that tonal carryover assimilation did not facilitate segmentation when agreeing with TPs. This finding cannot be attributed to cue reliability, as Experiment 2 suggests. Nor can it be accounted for by an exemplar-based view of phonological representations. It is assumed that tonal carryover assimilation is redundant in the presence of congruent statistical cues, and the discrepancy with previous studies with segmental coarticulation may be linked to the relatively lower weight of tonal information in speech processing. Further work can be done to investigate whether the current findings would be replicated cross-linguistically and to experiment with different acoustic implementations of tonal carryover assimilation and paradigms other than AL learning.
1. Musical background has been widely reported to confer an advantage in lexical tone identification and pitch discrimination (e.g., Delogu, Lampis, & Belardinelli, 2010;Lee & Hung, 2008;Xie & Myers, 2015). It is thus possible that the musically trained participants might exploit the tonal coarticulatory cue more effectively (or at least differently). To test this, we conducted mixed-effects analyses with musical training and its interactions with condition also included as fixed effects. Results indicated that for both experiments, none of these newly added fixed effects was significant and the patterns of statistical significances of the other fixed effects remained the same. There was no evidence that the participants with musical training behaved differently. | 2021-01-07T09:08:53.600Z | 2021-01-05T00:00:00.000 | {
"year": 2021,
"sha1": "746e884f77eda5a332b52f21a24d17cdd5db7e14",
"oa_license": "CCBY",
"oa_url": "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/58CD55F95F67D9DD11E2E739EA82DD5C/S0142716420000818a.pdf/div-class-title-the-use-of-tonal-coarticulation-in-segmentation-of-artificial-language-speech-a-study-with-mandarin-listeners-div.pdf",
"oa_status": "HYBRID",
"pdf_src": "Cambridge",
"pdf_hash": "c583f53522f510061e7c5177304eea65cfd1486f",
"s2fieldsofstudy": [
"Linguistics"
],
"extfieldsofstudy": []
} |
136646041 | pes2o/s2orc | v3-fos-license | Properties of thermoelectric Ce0.09Fe0.67Co3.33Sb12/FeSb2Te multi-layered structures prepared by laser ablation
Multi-layered Ce0.09Fe0.67Co3.33Sb12/FeSb2.1Te structures composed of thin equidistant layers were prepared by Pulsed Laser Deposition on fused silica quartz glass substrates. The structures were prepared at different substrate temperatures (230 °C or 250 °C) applying the laser beam energy density of 3 Jcm−2. In the contribution we present some thermoelectric properties such as the in-plane electrical conductivity, the Seebeck coefficient and the power factor for the multi-layered structures in the temperature range from 300 K to 500 K. Comparison of multi-layered structure's thermoelectric properties with single thin Ce0.09Fe0.67Co3.33Sb12 and FeSb2.1Te layers is given. A cross sectional picture of the multi-layered structure made by Scanning Electron Microscope is presented for the thicker multi-layered structure.
Introduction
Skutterudites have been of high interest as a promising candidate for thermoelectric applications. The key advantage of skutterudites is their possible high figure of merit ZT (ZT=0.8 at T=750 K for bulk Ce 0.12 Fe 0.71 Co 3.29 Sb 12 [1], ZT=1.1 at T=750 K for bulk Ce 0.28 Fe 1.5 Co 2.5 Sb 12 [2] and Ce 0.9 Fe 3 CoSb 12 [3]. Theoretical maximum ZT=1.4 at 1000 K was predicted for Ce 0.9 Fe 3 CoSb 12 [3]). ZT is the essential material property for thermoelectric energy conversion on the Earth and in the cosmos. The dimensionless figure of merit is expressed by a formula: where S, T and are the Seebeck coefficient, the electrical conductivity, the ambient temperature and the thermal conductivity, respectively. The term S 2 is called power factor and can be used for evaluation of thermoelectric material if thermal conductivity is unknown.
The simplest skutterudites crystallize in a body-centered-cubic structure of the spare group Im3. The low thermal conductivity of skutterudites is obtained by filling the voids in the structure with small diameter, large-mass interstitials such as for example Yb, Ce or Ba. The lattice thermal conductivity can be also reduced by substituting Co with Fe as in our case.
A great improvement of thermoelectric properties was mathematically and also experimentally proved by preparing materials in the form of a low dimensional system [4][5][6][7][8][9][10][11][12][13] such as thin layer, superlattice or multi-layered structure, which can be prepared by simultaneous deposition of very thin layers of two or more different materials. Such improvement in comparison with bulk materials was published for the thin skutterudite layers of CoSb 3 [14,15] or for the skutterudite superlattices [16].
Earlier we published results on thin skutterudite layers prepared by PLD in Ar atmosphere from the Ce 0.1 Fe 0.7 Co 3.3 Sb 12 hot pressed target [29]. The layers were well crystalline when deposited with the laser beam density Ds=3 Jcm -2 at substrate temperature Ts in the range from 200 °C to 250 °C. The best thermoelectric properties were obtained on the layers prepared at Ts=250 °C, although intensities of the skutterudite lines in X-ray diffraction analysis (XRD) diffractogram were more intensive for layers prepared at Ts=200 °C. All layers were of P-type electrical conductivity [29]. The Wavelength Dispersive analysis (WDX) proved Ce 0.09 Fe 0.67 Co 3.33 Sb 12 layers stoichiometry.
Recently, we prepared thin skutterudite layers by PLD in Ar atmosphere from the FeSb 2 Te hot pressed target. The best thermoelectric properties were obtained on the layers prepared at Ts=250 °C and Ts=230 °C with Ds=3 Jcm -2 . Such layers were well crystalline and were also of P-type electrical conductivity. The WDX shoved FeSb 2.1 Te layers stoichiometry. Bulk ternary skutterudite FeSb 2 Te had been examined and published in details before [45,46] and was proved to be a good thermoelectric material.
In this contribution, we examine thin thermoelectric multi-layered Ce 0.09 Fe 0.67 Co 3.33 Sb 12 /FeSb 2.1 Te system composed of thin equidistant layers Ce 0.09 Fe 0.67 Co 3.33 Sb 12 and FeSb 2.1 Te each 5 nm in thickness (5 nm period) prepared by PLD on a fused silica quartz glass substrate. The structures were prepared at Ts=230 °C and Ts=250 °C with Ds=3 Jcm -2 . A comparison of multi-layered structure's thermoelectric properties with properties of single thin Ce 0.09 Fe 0.67 Co 3.33 Sb 12 and FeSb 2.1 Te layers are also given. It is expected that the preparation of Ce 0.09 Fe 0.67 Co 3.33 Sb 12 /FeSb 2.1 Te multi-layered structure can be successful because of the similar lattice constant of both materials [45][46][47].
Experimental details
PLD targets of FeSb 2 Te and Ce 0.1 Fe 0.7 Co 3.3 Sb 12 composition where synthesized from individual elements by high-temperature solid-state reactions. Stoichiometric amounts of Fe (99.9 %), Sb (99.999 %), Te (99.999 %) and Ce (99.9 %), Fe (99.9 %), Co (99.9 %) and Sb (99.999 %) were sealed into evacuated carbon-coated silica glass tubes and heated up to 1050 °C for 48 hrs in a furnace. After quenching into a water bath, the same ampoule was placed into the furnace and annealed at 550 °C for 120 hrs. The resultant material was then ground under acetone, pelletized and heated again at 550 °C for 120 hrs. The completion of the solid-state reaction of obtained powder samples was verified by powder XRD.
The final targets for PLD deposition 20 mm in diameter and 2 mm in height were prepared by the hot pressing method (temperature 500 °C, pressure ~60 MPa for 1 hr). The measured density of pressed targets was found about 96-98 % of theoretical density.
The powder X-ray diffraction patterns of FeSb 2 Te ternary compounds used for the target were collected in the Bragg-Brentano geometry on a Bruker D8 Advance diffractometer. CuKα radiation was used. The crystal structure of FeSb 2 Te was refined by the Rietveld method for X-ray powder diffraction data using the FullProf program [23]. XRD spectrogram of target compound is depicted in figure 5 (top view).
The basic schema of the experimental apparatus for PLD is depicted in figure 1. Conceptually and experimentally, PLD is an extremely simple method, probably the simplest of all thin film growth techniques. A high power pulsed excimer KrF laser (COMPexProTM 205 F) radiation (1) is used as an external energy source to vaporize materials of target (5) and to deposit a thin film. A set of optical components is used to focus the laser beam to the target surface (2,3). After the laser pulse irradiation the temperature rises very rapidly (1011 Ks −1 ) and the evaporation becomes non-equilibristic. In our experiment substrates were cleaned from the mechanical dirt in an ultrasonic cleaner. After that the substrates were subsequently cleaned in acetone, toluene and in ethanol. Cleaning in the vapours of boiling ethanol then completed this process. Fused silica substrates were finally annealed in an oven at a temperature around 250 °C. The layers and multi-layered structures were deposited on fused silica quartz glass substrate 10x10 mm. The deposition took place at Ar atmosphere (13 Pa). The distance of the substrate from the target was set to 40 mm.
Series of multi-layered Ce 0.09 Fe 0.67 Co 3.33 Sb 12 /FeSb 2.1 Te structures composed of thin equidistant Ce 0.09 Fe 0.67 Co 3.33 Sb 12 and FeSb 2.1 Te layers of 5 nm in thickness and total thickness of about 60 nm were prepared by PLD at Ts=230 °C and Ts=250 °C with Ds=3 Jcm -2 . The deposition conditions were chosen based on previous results taken on single Ce 0.09 Fe 0.67 Co 3.33 Sb 12 and FeSb 2.1 Te layers as the conditions giving the best thermoelectric properties.
Transport properties, such as the in-plain electrical resistivity and the Seebeck coefficient, were measured on each multi-layered structure and on single layers in the temperature range from 300 K up to 500 K. The power factor was then calculated. Four square shaped contacts for the measurements were prepared by evaporating Ti. Pressed Pt/PtRh thermocouples with diameter of 0.07 mm were used as leads. A conventional DC van der Pauw's method was used for the electrical conductivity measurement. The experimental error of this method is about 10 % for the conductivity measurement.
The Seebeck coefficient was determined from the variation of the electromotive force for different temperature gradients across the layer. The both sides of the sample were in the thermal contact with an independent wire resistant sub-heater that supplies the heat and induces the sample temperature gradient. The thermocouple junctions were bonded to each corner of the square shaped sample. The experimental error of the Seebeck coefficient measurement is about 20 %.
The thickness of the layers was measured using a mechanical stylus profile-meter (-Step 500, Tencor Instruments, U.S.A). The uncertainty in the thickness estimation is about 10 % in the examined range of thicknesses partly because of the fact that the layer thickness in the centre of the sample is higher than on its edge.
Scanning Electron Microscope (SEM) images were measured using Leo 1550 device (magnification up to 500 k). To see the edge of the layer, the sample was scanned under 45. The WDX composition study done on 439 nm thick single layer proved an average stoichiometry of Ce 0.09 Fe 0.67 Co 3.33 Sb 12 . The maximal deviation was in both cases less than +/-3 % for all elements.
Single thin layers
The XRD measurement done on 439 nm thick single layer proved that the layer was well crystalline with CoSb 3 structure.
The temperature dependencies of the electrical resistivity and the Seebeck coefficient for single thin layers prepared from FeSb 2 Te target at Ts=230 °C and Ts=250 °C with Ds=3 Jcm -2 are depicted in figure 3. Both layers showed semi-conducting P-type behaviour of the electrical resistivity (the electrical resistivity decreases with rising temperature). The room temperature values of electrical resistivity of the layers were higher than the value published for bulk FeSb 2 Te material (1.59x10 -3 Ohmcm at 25 °C) [20] and also higher than value measured on PLD target (2.50x10 -3 Ohmcm at 25 °C), see table 1. The room temperature values of Seebeck coefficient of FeSb 2 Te single layers prepared at Ts=230 °C with Ds=3 Jcm -2 and Ts=250 °C with Ds=3 Jcm -2 are about twice higher than published values for bulk FeSb 2 Te material [45]. Prepared layers shoved higher Seebeck coefficients in the whole studied temperature range from room temperature up to 500 K in comparison with values for bulk FeSb 2 Te material [46]. The layer prepared at Ts=230 °C with Ds=3 Jcm -2 had lower room temperature value of power factor 1557.45x10 -9 WK -2 cm -1 than the value published for bulk material [45,46] and also than layers prepared at Ts=250 °C with Ds=3 Jcm -2 . But the power factor's slope for the layer prepared at Ts=230 °C with Ds=3 Jcm -2 suggests that it may be comparable with layers prepared at Ts=250 °C for higher temperatures (above 500 K).
WDX experiments that were done on the thicker layers proved average FeSb 2.1 Te stoichiometry that corresponds very well to the target composition.
XRD experiments done on the thicker layers with thickness 240 nm and 235 nm prepared at Ts=230 °C and Ts=250 °C with Ds=3 Jcm -2 , respectively, proved that both layers were well crystalline with peaks corresponding to the used PLD target.
The resulting dependency of power factor versus temperature for the best single PLD layers is together with the published bulk Ce 0.12 Fe 0.71 Co 3.29 Sb 12 data [1] depicted in figure 4. The power factors of our best prepared single layers are much smaller (roughly twice smaller at room temperature and roughly 5 times smaller at 500K) in comparison with the bulk Ce 0.12 Fe 0.71 Co 3.29 Sb 12 [1]. Temperature dependency of power factor of bulk FeSb 2 Te has never been published. For the bulk material there was only room temperature value of power factor about 1881.8x10 -9 WK -2 cm -1 published [45,46]. figure 5. Both structures showed semi-conducting P-type behaviour of the electrical resistivity (the decrease of electrical resistivity with the increase of temperature). The multi-layered structures had lower electrical resistivity than the best above mentioned single layers Ce 0.09 Fe 0.67 Co 3.33 Sb 12 and FeSb 2.1 Te in the whole studied temperature range. The multi-layered structures showed also much lower in-plain Seebeck coefficient in the whole measured temperature range than the single layers. Due to the low Seebeck coefficient, the resultant power factor of all prepared multi-layered structures is lower than the power factor of the best prepared single layers.
Multi-layered structures
When we compared the measured values of power factor of the Ce 0.09 Fe 0.67 Co 3.33 Sb 12 /FeSb 2.1 Te multi-layered structures with published bulk Ce 0.12 Fe 0.71 Co 3.29 Sb 12 material [1], we found that multilayered system values are roughly four times lower at room temperature and roughly ten times lower at about 500 K. If the decrease of thermal conductivity on multi-layered structure in comparison to bulk material is taken into account, we may speculate that even better values of ZT for prepared multilayered systems may be achieved in comparison with the bulk Ce 0.12 Fe 0.71 Co 3.29 Sb 12 material. It is assumed that cross-sectional electrical conductivity and Seebeck coefficient are not much influenced by interfaces in the multi-layered structures. Temperature dependency of power factor of bulk FeSb 2 Te has never been published, so the power factor of multiple-structures can not be compared. Only the room temperature value of power factor of about 1881.8x10 -9 WK -2 cm -1 was published [45,46]. The SEM study was done on 286 nm thick multi-layered structure composed of altering equidistant layers of Ce 0.09 Fe 0.67 Co 3.33 Sb 12 (thickness 17.5 nm) and FeSb 2.1 Te (thickness 18.3 nm). The SEM cross sectional picture of multi-layered structure is presented in figure 8, where individual layers were visible.
Conclusions
Multi-layered Ce 0.09 Fe 0.67 Co 3.33 Sb 12 /FeSb 2.1 Te structures composed of thin equidistant Ce 0.09 Fe 0.67 Co 3.33 Sb 12 and FeSb 2.1 Te layers (5 nm in thickness) were successfully prepared by PLD at Ts=250 °C and Ts=230 °C with Ds=3 Jcm -2 . The resultant in-plain power factors of multi-layered structures were lower than the power factors of both used materials in a form of single layer. To make an overall evaluation of thermoelectric properties the Harman measurement of thermoelectric figure merit and a measurement of the thermal conductivity, which is expected to decrease doe to number of interfaces in the structure, is necessary. | 2019-04-28T13:06:06.860Z | 2013-12-02T00:00:00.000 | {
"year": 2014,
"sha1": "d685d7e0c8a618a09d9522d440cf1f4b418ab088",
"oa_license": "CCBY",
"oa_url": "http://iopscience.iop.org/article/10.1088/1742-6596/497/1/012038/pdf",
"oa_status": "GOLD",
"pdf_src": "IOP",
"pdf_hash": "65bedc95f84b08ac823190a8c57d9ee5e44ded6f",
"s2fieldsofstudy": [
"Materials Science"
],
"extfieldsofstudy": [
"Physics",
"Materials Science"
]
} |
67821813 | pes2o/s2orc | v3-fos-license | The Solar Neighborhood XXX: Fomalhaut C
LP 876-10 is a nearby active M4 dwarf in Aquarius at a distance of 7.6 pc. The star is a new addition to the 10-pc census, with a parallax measured via the Research Consortium on Nearby Stars (RECONS) astrometric survey on the Small&Moderate Aperture Research Telescope System's (SMARTS) 0.9-m telescope. We demonstrate that the astrometry, radial velocity, and photometric data for LP 876-10 are consistent with the star being a third, bound, stellar component to the Fomalhaut multiple system, despite the star lying nearly 6 degrees away from Fomalhaut A in the sky. The 3D separation of LP 876-10 from Fomalhaut is only 0.77+-0.01 pc, and 0.987+-0.006 pc from TW PsA (Fomalhaut B), well within the estimated tidal radius of the Fomalhaut system (1.9 pc). LP 876-10 shares the motion of Fomalhaut within ~1 km/s, and we estimate an interloper probability of ~10^{-5}. Neither our echelle spectroscopy nor astrometry are able to confirm the close companion to LP 876-10 reported in the Washington Double Star Catalog (WSI 138). We argue that the Castor Moving Group to which the Fomalhaut system purportedly belongs, is likely to be a dynamical stream, and hence membership to the group does not provide useful age constraints for group members. LP 876-10 (Fomalhaut C) has now risen from obscurity to become a rare example of a field M dwarf with well-constrained age (440+-40 Myr) and metallicity. Besides harboring a debris disk system and candidate planet, Fomalhaut now has two of the widest known stellar companions.
Introduction
Fomalhaut is an important nearby A3 V star, containing a large resolved dusty debris disk (Gillett 1986;Kalas et al. 2005) and a candidate extrasolar planet (Kalas et al. 2008(Kalas et al. , 2013Quillen 2006). Fomalhaut has previously had at least two stars suggested to be companions. See (1898) reported a 14th magnitude stellar companion to Fomalhaut at 30" separation, however this star was later deemed a background star by Burnham (1978) 1 . Luyten (1938) reported discovery of a K-type common proper motion companion to 1 See (1898) reported a single observation of a 14th magnitude companion at θ = 36 • .2, separation 29".98, at epoch 1896.706. Dubbed "λ 1 478" by See, this object appears to have largely disappeared from the literature, and does not appear in the modern Washington Double Star catalog. The only subsequent mentions of this companion that we found are in the Burnham (1906) compendium of double stars (Entry #12071 is listed as "See 478"), and in two popular books (Allen 1963;Burnham 1978). Burnham (1978) stated "it appears to be merely a faint field star, having no real connection with Fomalhaut." Based on the van Leeuwen (2007) Hipparcos astrometry for Fomalhaut A, we estimate that Fomalhaut A has moved 35" since See's observation, 3) we were unable to find the star in the Spitzer imagery and other modern catalogs, we conclude that See's reported companion to Fomalhaut was likely spurious. Fomalhaut: TW PsA (HR 8721). The physicality of the Fomalhaut-TW PsA binary system was investigated by Barrado y Navascues et al. (1997) and Mamajek (2012), and both studies concluded that the pair comprise a physical binary. Mamajek (2012) estimated that Fomalhaut and TW PsA have a true separation of only 0.28 pc and share velocities within 0.1 ± 0.5 km s −1 , consistent with constituting a bound system. Mamajek (2012) estimated the age for the Fomalhaut binary system to be 440 ± 40 Myr based upon multiple age indicators, with the isochronal age of Fomalhaut A and the gyrochronology age of Fomalhaut B providing the most weight.
During the preparation of the Mamajek (2012) article, another neighboring star was identified that appeared to share motion with Fomalhaut and TW PsA: LP 876-10 (NLTT 54872, WT 2282, 2MASS J22480446-2422075, PM I22480-2422). LP 876-10 is a high proper motion star first catalogued as such by Luyten & Hughes (1980), situated 5 • .67 NW (20407. ′′ 6; PA = 337 • .91) of Fomalhaut. At the time of writing Mamajek (2012) there was insufficient evidence to test whether LP 876-10 was truly associated with the Fomalhaut binary, with the main evidence being the coincidental proper motion and photometric distance. In this contribution, we combine newly determined accurate astrometric and radial velocity measurements for LP 876-10 to demonstrate that it too, like TW PsA, appears to be a distant companion of Fomalhaut, and should be considered "Fomalhaut C".
Analysis
The stellar parameters for Fomalhaut, TW PsA, and LP 876-10 are summarized in Table 1. Finder charts for LP 876-10 are provided in Figure 1.
Parallax and Proper Motion
The parallax and proper motion of LP 876-10 have been measured during the long-term astrometry program carried out by RECONS at the SMARTS 0.9-m telescope. Jao et al. (2005) describes the astrometry program, however we briefly summarize the program here.
A filter is selected from the Johnson-Kron-Cousins V R RC I RC filterset that provides a well-exposed reference field that, ideally, encircles the target star. Throughout the course of the observations, the same pointing (to within a few pixels) and filter are used. Centroids for the reference field and parallax star are extracted using SExtractor (Bertin & Arnouts 1996) and corrected for differential color refraction using V R RC I RC photometry of the reference and science target stars (see Section 2.1). Relative parallax and proper motion of the target star are solved for using the Gaussfit program 3 . Correction from relative to absolute parallax is done by estimating the mean distance to the reference field stars, again, using V R RC I RC photometry and the photometric distance relations of Henry et al. (2004).
LP 876-10 was included in the RECONS astrometric survey due to its close predicted photometric distance (7.2 ± 0.8 pc; Reid et al. 2003), which is consistent with preliminary parallax solutions from this program (Bartlett 2007;Bartlett et al. 2007) Blinking images suggests that the neighboring high proper motion star LP 876-11 could be a proper motion companion to LP 876-10; LP 876-11 is located 1'.8 away from LP 876-10 at 42 • east of north. However, we determine a proper motion for LP 876-11 of 321.3 ± 0.7 mas yr −1 at PA 143 • .0 ± 0 • .2 east of north, which is inconsistent with the measured motion for LP 876-10. Using twelve color-magnitude relations from Henry et al.
(2004), we estimate a photometric distance to LP 876-11 of 730 ± 120 pc. We measure a trigonometric parallax of LP 876-11 4 of 1 ± 2 mas, consistent with the photometric distance. We conclude that LP 876-11 is not physically associated with LP 876-10.
Radial Velocity
A spectrum of LP 876-10 was taken with the CRIRES spectrograph on the 8.4-m VLT UT1 (Antu) telescope on UT date 16 June 2009 as part of a near-infrared radial velocity survey of nearby late-type M dwarfs (Bean et al. 2010). The CRIRES spectrum has wavelength coverage 2.292-2.349 µm over the effective 4096 x 512 focal plane detector, a mosaic of four Aladdin III InSb arrays (Kaeufl et al. 2004). The slit width was 0 ′′ .2, yielding a resolving power of R ≃ 100,000 (resolution is 3 km s −1 at 2 pixel sampling). The signal to noise ratio in the continuum of the spectrum was ∼170-220. By fitting a broadened and shifted PHOENIX model spectrum from the GAIA V2.0 library (Hauschildt et al. 1999;Brott & Hauschildt 2005) to the spectrum of LP 876-10, we determine a sizeable projected rotation velocity of vsini = 22 ± 2 km s −1 ; a heliocentric radial velocity of +6.5 ± 0.5 km s −1 was also measured. Slit viewer images of LP 876-10 appear point-like, and there is no sign of duplicity in the CRIRES spectrum. A more detailed spectroscopic analysis of LP 876-10 will be presented in a forthcoming paper (Seifahrt et al., in prep.).
Duplicity
While neither the astrometry nor the spectroscopy data are consistent with LP 876-10 being a binary, it is listed as a double star in the Washington Double Star catalog (WDS; Mason et al. 2001) 5 as WDS 22481-2422 and with discovery identifier "WSI 138" 6 . A single observation is reported for epoch 2010, with a reported companion at separation 0".5 at PA = 144 • , with magnitudes 12.80 and 14.80 (presumably V -band, as the combined magnitude [12.64] is similar to the adopted V magnitude in Table 2). We are unable to confirm the existence of the companion reported in the WDS. In the 118 frames taken during 25 nights, with FWHMs in the range 1".2 to 2".8, LP 876-10 appeared to be a point source -with no evidence of elongation. With only a single observation, the possibility remains that the reported WDS companion may be a chance alignment between this high proper motion star and a background star (B. D. Mason 2013, private communication).
However, we believe that a background star is unlikely to explain this discrepancy. Based Delfosse et al. (2000). Given the projected separation (0".5 = 3.8 AU), these values would predict an orbital period of ∼13.5 yr. Assuming zero eccentricity and face-on projection, one would predict orbital motion of ∼27 deg yr −1 and a photocentric amplitude of ∼110 mas.
The predicted photocentric amplitude would be about half (50 mas over 8 years) of the full amplitude (110 mas over ∼13.5 years) during the observations to date. As seen in Fig. 3, the astrometric solution using only parallax and proper motion is quite good, and any gravitational perturbations on LP 876-10 must be at the <10 mas level over ∼8 yr, which easily rules out the predicted signal for the companion reported in the WDS. Table 3 shows that the difference between the long-term proper motions (e.g. SuperCOSMOS, USNO-B1.0, PPMX, UCAC4) are largely within ∼5-10 mas yr −1 (rms) of the 8-year baseline proper motion calculated in this survey, further suggesting that it would be difficult to hide a ∼50 mas yr −1 perturbation of the photocentric motion. As the purported WDS companion should have a period only somewhat longer than the duration of our RECONS astrometric dataset, and with a predicted photocentric amplitude similar in size to the observed parallax, we conclude that it is unlikely that the companion reported in the WDS catalog is real.
Temperature, Luminosity, and Radius
We estimated T eff for LP 876-10 by fitting the photometry in Table 2 to the BT-Settl grid of synthetic stellar spectra which vary by effective temperature, metallicity, and surface gravity (Allard et al. 2012). Twenty-two colors consisting of combinations of the bands V , R RC , I RC , J, H, K s , W 1, W 2, and W 3 were compared to grid interpolations based on models, and the best fit yielded an interpolated temperature of T eff = 3132 K and solar metallicity. We estimated the uncertainty in T eff due to metallicity and surface gravity by individually varying these parameters by one increment (0.2 dex) and measuring the effect on the resultant T eff . The uncertainty in the T eff breaks down approximately as follows: ±33 K from the dispersion in color-based T eff estimates for the best fit, ±50 K due to metallicity uncertainty, and ±25 K due to uncertainty in log(g). Together this yields an overall T eff uncertainty of ±65 K. The systematic error due to the validity of the BT-Settl models is unknown, however our derived T eff should be comparable to M dwarf T eff values derived using the same models (indeed Rajpurohit et al. 2013, similarly derives T eff ≃ 3100-3200 K for M4 dwarfs like LP 876-10 using BT-Settl models). The best fitting BT-Settl synthetic spectrum had T eff = 3100 K, [Fe/H] = 0.0, log(g) = 5.0. From considerations of the star's color-magnitude diagram position (Sec. 2.6), we predict that LP 876-10 has a slightly subsolar metallicity, and lies near the zero-age main sequence for ∼0.2 M ⊙ stars (log(g) ≃ 5.06; Baraffe et al. 1998). The best fitting BT-Settl synthetic spectrum then adjusted via an iterative process to produce a match to the observed photometry. The process determines a small λ-dependent polynomial correction factor that is applied to the synthetic spectrum to cause small modifications in order to produce the best fit to the photometry (details of the technique are described in Dieterich et al. 2013).
Combined with our estimate of the projected rotation velocity vsini (22 ± 2 km s −1 ), this places an upper limit on the rotation period of LP 876-10 of 0.55 ± 0.05 day (see §2.8).
Color-Magnitude Diagram and Metallicity
Using our new parallax and the photometry in Table 2, we estimate absolute magnitudes of M V = 13.21 ± 0.02 and M Ks = 7.81 ± 0.03. From Table 2, we calculate a (V-K s ) color of 5.40 ± 0.02 mag. Using the (V-K s ) vs. M V relations from Henry et al. (2004) and Johnson & Apps (2009), we predict photometric distances of 7.9 ± 1.5 pc and 7.7 ± 1.4 pc, respectively, in excellent agreement with our trigonometric parallax distance.
The agreement between the trigonometric parallax distance and the available photometric distances (Reid et al. 2003, this section) is also indicative that LP 876-10 is unlikely to have an unresolved companion of similar mass, and it is more likely to be a main sequence, rather than pre-main sequence, star.
We can constrain the metallicity and age using the star's color-magnitude data. In (Santos et al. 2004) and -0.20 (Morell 1994). Hence, both the photometric metallicity estimate for LP 876-10 and the spectroscopic metallicity estimates for TW PsA are self-consistent, and consistent with being very slightly subsolar [Fe/H] ≃ -0.1 dex.
Mass and Age Constraints
Using the Delfosse et al. (2000) M V vs. mass calibration for field M dwarfs (i.e. mixed metallicities and ages), the approximate mass of LP 876-10 is ∼0.20 M ⊙ . Interpolating within the Baraffe et al. (1998) tracks, one finds that solar composition stars with masses of greater than 0.163 M ⊙ are not ever predicted to be as faint as M V = 13.21 mag (see Fig. 4). As the tracks are first and foremost tracing luminosity evolution as a function of mass and age, we also examine the constraints that the luminosity of LP 876-10 can provide. Through fitting BT-Settl models to the photometry, we estimate the luminosity to be log(L/L ⊙ ) = -2.337 ± 0.010 dex. We find that the Baraffe et al. (1998) and Dotter et al. (2008) solar composition tracks give essentially identical predictions that no stars with with masses greater than 0.197 M ⊙ are predicted to have luminosities this low. Using those tracks, we estimate that it takes a 0.2 M ⊙ star approximately ∼300 Myr to reach within ∼0.01 mag of the zero-age main sequence (the actual minimum in luminosity and radius occurs around ∼400-500 Myr). The appearance of LP 876-10 on the zero-age main sequence for [M/H] ≃ -0.1 is commensurate with the adopted age for Fomalhaut A and B (440 Myr; Mamajek 2012).
As seen in Fig. 4, the Baraffe et al. (1998) isochrones do not accurately reproduce the empirical main sequence from Johnson & Apps (2009) in this color regime, so our lower bound on the age of LP 876-10 is only approximate. Naively interpolating the mass and age of Fomalhaut C from the evolutionary tracks and isochrones would yield a mass of ∼0.11 M ⊙ and age of ∼60 Myr. However, as can be seen in Fig. 4, a 125 Myr isochrone (log(age/yr) = 8.1) from the same tracks fails to replicate the intrinsic color-magnitude sequence for the ∼125 Myr-old Pleiades (Barrado y Navascués et al. Baraffe et al. (1998) isochrones for age 125 Myr (log(age/yr) = 8.1) predict absolute magnitude M V = 13.57 for V-K s = 5.4 7 , i.e. 1.33 mag too faint! As summarized by Bell et al. (2012), "for all optical colours, no pre-MS models follows the observed Pleiades sequence for temperatures cooler than 4000 K." Estimating isochronal ages using pre-MS evolutionary tracks is quite problematic, with large systematic differences between tracks (see review by Soderblom 2010). For all of these reasons, we do not adopt the pre-MS mass and isochronal age interpolated from the evolutionary tracks and isochrones in Fig. 4, and instead constrain the age based on its proximity to the main sequence, and infer the mass based on main sequence absolute magnitude vs. mass considerations. Given the empirical and theoretical constraints previously discussed, we adopt a mass of 0.18 ± 0.02 M ⊙ for Fomalhaut C. 7 The Baraffe et al. (1998) tracks use the CIT JHK photometric system. We convert the Baraffe et al. (1998) CIT photometry to 2MASS following (Carpenter 2001).
To search for a rotation period, we selected SuperWASP photometry from a single well-sampled season (2008) with V SuperW ASP magnitudes between 12.46 and 12.70, with magnitude and photometric error of less than 0.2 mag, and with a good TAMFLUX2 flag extraction. SuperWASP photometry is calibrated to the Tycho-2 V T system (Pollacco et al. 2006;Høg et al. 2000). There were 3162 points for subsequent analysis. To remove 1-day aliasing effects, all points during a single observing night were adjusted so that their average These tests indicate that the 0.195, 0.242, and 0.318 day peaks are aliasing effects due to the irregular time sampling of the light curve. We conclude that the P = 0.466 d peak is most likely due to the rotation of the star.
Unfortunately, a rotation period of ∼0.5 day for a ∼0.2 M ⊙ star places negligible constraint on its age. Mid-M stars with rotation periods faster than 1 day are a nearly ubiquitous feature of stellar samples between ages of ∼2 Myr and ∼10 Gyr (see Fig. 12 of Irwin et al. 2011). Figure 11 We conclude that attempts to age-date LP 876-10 via gyrochronology/rotation constraints appear fruitless.
Activity
Not only is LP 876-10 fast rotating, but, unsurprisingly, it appears to be a coronally active star as well. Voges et al. (1999) ranked LP 876-10 9 as the most likely optical counterpart of the ROSAT All-Sky Survey (RASS) Bright Source Catalog (BSC) X-ray source 1RXS J224803.5-242240. The X-ray counterpart is 35" away from LP 876-10.
However, the RASS BSC position error is 15", and LP 876-10 is the brightest optical source within 40", indicating that it is the likely X-ray source (Neuhaeuser et al. 1995). 1RXS J224803.5-242240 appears to be the brightest RASS X-ray source within a degree of LP 876-10. The fact that the position of the brightest RASS X-ray source within a degree of LP 876-10 lies within 40" of the rapidly rotating, nearby M dwarf suggests to us that LP 876-10 is almost certainly the optical counterpart of 1RXS J224803.5-242240. The RASS-BSC catalog (Voges et al. 1999) reports a soft X-ray flux of 0.142 ct s −1 (28% uncertainty) with HR1 hardness ratio of -0.23±0.21, detected over a short exposure time of 176 s. Using the energy conversion factor relation from Fleming et al. (1995), this translates to a coronal X-ray flux in the soft X-ray band (0.2-2.4 keV) of roughly 1.01×10 −12 erg/s/cm −2 . At d = 7.57 pc, this corresponds to an X-ray luminosity of L X ≃ 10 27.84 erg/s. This implies log(L X /L bol ) ≃ -3.41, i.e. a very active star close to X-ray saturation. This corroborates the very high projected rotational velocity measured spectroscopically (vsini = 22 km s −1 ), which should induce strong magnetic activity.
Velocity and Interloper Probability
With our best measurements of the proper motion, radial velocity, and parallax, we calculate the 3D Galactic velocity of LP 876-10 to be (U, V, W) = -5.3 ± 0.2, -7.6 ± 0.3, -11.9 ± 0.4 km s −1 . Comparing these values to those for Fomalhaut and Fomalhaut B (TW PsA), we find that LP 876-10's velocity only differs from that of Fomalhaut by 1.1 ± 0.7 km s −1 , and that of Fomalhaut B by 1.1 ± 0.5 km s −1 . Using the LSR velocity ellipsoid for both dM and dMe dwarfs estimated by Reid et al. (2002, their unweighted solution), and adopting the solar peculiar velocity with respect to the LSR from Schönrich et al.
(2010), we naively only expect roughly 1 in ∼55,000 field M dwarfs to have UVW velocities within 1.1 km s −1 of Fomalhaut, and roughly 1 in ∼12,000 field M dwarfs to have a velocity within 2 km s −1 .
implies that within a sphere of radius 1 pc surrounding Fomalhaut, we would expect to find 0.25 M dwarfs. Hence, we estimate the probability that a random M dwarf could appear within 1 pc of Fomalhaut, and sharing its velocity within less than 2 km s −1 , as approximately 1 in ∼10 4.7 (and sharing its velocity within less than 1.1 km s −1 as roughly 1 in ∼10 5.3 ). For comparison, one would expect to have to encircle a sphere ∼36 pc in radius in the local Galactic disk in order to find another M dwarf whose motion randomly agreed with that of Fomalhaut within less than 2 km s −1 . Our probability estimates do not take into account the similarity in the spectroscopic metallicity of TW PsA and the photometric metallicity of LP 876-10, which provides further agreement. We conclude that LP 876-10 appears to be related to Fomalhaut and TW PsA beyond a reasonable doubt.
The Castor Moving Group
Fomalhaut was listed by Barrado y Navascues (1998) as a potential member of the Castor Moving Group (GMG). The co-motion of LP 876-10 with Fomalhaut may be less significant if Fomalhaut is immersed in a swarm of co-moving stars like the purported CMG. The origin and nature of moving groups like CMG is an active field of study (e.g. Famaey et al. 2005;Murgas et al. 2013). That the CMG represents a kinematic group of stars of common age and birthsite is unlikely.
Calculating revised space motions for the 14 CMG "members" ("Y" or "Y?" members) from Barrado y Navascues (1998), using revised Hipparcos astrometry (van Leeuwen 2007) and the best available radial velocities (Barbier-Brossat & Figon 2000;Gontcharov 2006), we find that the CMG stars have median velocity of (U, V , W ) = -11.1 ± 1.9, -8.6 ± 0.8, -9.7 ± 1.0 km s −1 , with standard deviations of 6.1, 3.6, and 4.2 km s −1 . The scatters are much larger than the typical velocity errors, and larger than the one-dimensional velocity dispersions of nearby clusters and associations (<1.5 km s −1 ; Madsen et al. 2002;Mamajek 2010). The velocity for Fomalhaut differs from the CMG median velocity by 5.6 ± 2.3 km s −1 . The list of "final" members in Barrado y Navascues (1998) comprises ∼27 M ⊙ of stars spread out over a volume of ∼55,000 pc 3 , implying that the density of CMG members in the solar neighborhood is roughly ∼0.004× the local disk density (0.12 M ⊙ pc −3 ; van Leeuwen 2007). The stellar systems in the CMG have negligible interaction with one another, and so their motions are completely dominated by the local Galactic potential.
The velocity differences between Fomalhaut and individual CMG members is illuminating, and we discuss the famous CMG members Vega, LP 944-20, and Castor itself, in more detail. Vega is a proposed fellow CMG star of either similar age (455 ± 13 Myr; Yoon et al. 2010) or somewhat older age (700 +150 −75 Myr; Monnier et al. 2012) than that of Fomalhaut (440 ± 40 Myr; Mamajek 2012). Could Vega and Fomalhaut be related?
Another nearby famous CMG "member" is the nearby candidate brown dwarf LP 944-20 10 10 A new RECONS parallax has been measured which places LP 944-20 at a distance of 6.4 pc (Dieterich et al., submitted), making it most likely a star near the H-burning limit rather than a brown dwarf. The new distance revises LP 944-20's space motion to (U, V , W ) = -14.9, -5.9, -1.5 km s −1 , which differs from that of Fomalhaut by 13.5 km s −1 , and does -25 - (Ribas 2003). Adopting the astrometry from Tinney (1996) and a mean radial velocity of +9.0 ± 0.5 km s −1 (based on measurements from Martín et al. 2006), we calculate a velocity for LP 944-20 of (U, V , W ) = -12.2 ± 0.4, -5.6 ± 0.3, -2.8 ± 0.3 km s −1 . LP 944-20 is currently situated 6.6 pc away from Fomalhaut, and its velocity differs from that of Fomalhaut by 10.9 ± 0.8 km s −1 . Only 10 Myr ago, LP 944-20 and Fomalhaut were separated by ∼100 ± 8 pc, and were only more widely separated in the past. We also investigated whether there was any association between Fomalhaut and Castor itself. For the Castor sextuplet system, we adopt the recent parallax estimate from Torres & Ribas (2002), the long-term system proper motion from PPMX (Röser et al. 2008), and the center-of-mass radial velocity estimate from Heintz (1988, we adopt a RV uncertainty of 1 km s −1 in our calculations). These values are consistent with the Castor system having a velocity of (U, V , W ) = -7.5 ± 0.7, -3.7 ± 0.6, -11.5 ± 0.4 km s −1 .
Fomalhaut is currently ∼21 pc away from the Castor system, differing in velocity by a significant margin (4.9 ± 1.1 km s −1 ), and only 10 Myr ago Fomalhaut and Castor were separated by ∼50 ± 5 pc, and were even more distant in the past (more than 700 pc 100 Myr ago).
Despite these stars (the Fomalhaut system, Vega, LP 944-20, Castor system) being young and having somewhat similar velocities, their velocities are well-constrained enough and different enough that it is clear that they were not in the vicinity of one another even in the recent past, let alone a couple of Galactic orbits ago. We conclude that the Castor Moving Group is comprised of stars from different birthsites rather than a coeval system, and hence "membership" to the CMG does not provide useful age constraints for the Fomalhaut system (or Vega, LP 944-20, Castor, or other CMG members).
not change the qualitative conclusions in the text.
A Bound Companion?
One predicts that stellar companions in multiple systems can exist with separations up to their tidal (Jacobi) radius with respect to the Galactic potential. Jiang & Tremaine (2010) parameterize the tidal radius r t as: where G is the Newtonian gravitational constant, M 1 and M 2 are the masses of the stars, Ω is the Galactic angular circular speed (orbital velocity dividied by Galactocentric radius), and A is the Oort parameter. Adopting modern estimates of the relevant Galactic parameters, and rewriting the expression from Jiang & Tremaine (2010), we estimate the tidal radius to be: Summing the masses of the Fomalhaut system components (2.83 M ⊙ ), one predicts a tidal radius of ∼1.9 pc. The tidal radius for Fomalhaut A alone is ∼1.7 pc. Hence, the separation of ∼0.8 pc between Fomalhaut C and the A is not dynamically implausible for a bound system. Recent systematic surveys for wide-separation pairs using modern astrometric databases have started to yield many previously unrecognized parsec-scale common proper motion pairs (e.g. Shaya & Olling 2011), making Fomalhaut C less unusual than it may have appeared even a decade ago. Stable orbits for time scales longer than a Gyr are also possible for separations larger than the tidal radius if the distant companion orbits retrograde to the Galactic rotation (e.g. Makarov 2012). Indeed, higher precision-radial velocities for the components of the Fomalhaut system and taking into account the sub-kilometer second −1 effects of convective blueshift and gravitational redshift, may lend itself to providing a test as to whether Fomalhaut C is orbiting Fomalhaut AB either retrograde or prograde to the Galactic rotation (V. Makarov, priv. comm.).
Fomalhaut A and B are separated by ∆ AB = 57.4 +3.9 −2.5 kAU, and Fomalhaut C is separated by ∆ AC = 158.2 +2.3 −1.2 kAU from A, and by ∆ BC = 203.4 +1.0 −0.8 kAU from B. We calculate the position of the barycenter (center of mass) for the system using the Galactic (X, Y, Z) positions and masses in Table 1: (X, Y, Z) com = 3.08, 1.13, -6.93 pc. Converting this position to the equatorial ICRS coordinate system yields (α, δ) = 344 • .179, -29 • .792, at distance 7.67 pc. We can make a rough estimate of the orbital period of C around the AB pair. C is currently located ∼0.77 pc from the system's center of mass. If C is currently near apastron (not an unreasonable assumption given that binary stars will spend most of their time near apastron), and if C's periastron must almost certainly be larger than B's current separation from the system barycenter (0.24 pc), then a reasonable first estimate of C's orbit is a ∼ 0.5 pc and e ∼ 0.5. For the total mass of the Fomalhaut system (2.83 M ⊙ ), this translates to an approximate orbital period of ∼20 Myr, or ∼5% the system's age. The predicted orbital velocity of LP 876-10 around the Fomalhaut system barycenter would be ∼0.15 km s −1 . Given the masses and configuration of the AB pair, the escape velocity of C is ∼0.2 km s −1 .
How stable is Fomalhaut C's orbit with respect to A and B? Obviously, the orbit of AB and AB-C are not well constrained. We only have fairly accurate estimates of the relevant mass ratios and current separations, while the semi-major axes and eccentricities are unknown. The mass of C is very small compared to that for the AB pair (µ = M C /(M A + M B ) ≃ 0.07), and its current separation from the center of mass for the system is approximately 159 kAU. Based on simulations of test particles in the vicinity of binary systems of varying semi-major axis, mass ratio, and eccentricity, Holman & Wiegert (1999) provided estimates of the widest stable orbit around a member of a binary system (S-type orbits), and the closest orbit around both members of a binary system (P-type orbits).
If the current A-B separation is equivalent to its semi-major axis (assume e = 0; a = 57.4 kAU), then the minimum stable semi-major axis for C is predicted to be ∼135 kAU. Tokovinin (1998) estimates that the mean eccentricity for wide binary pairs is < e > ≃ 2/3. If Fomalhaut B is currently near apastron with e ∼ 2/3, then a ∼ 34 kAU, and the minimum stable semi-major axis for Fomalhaut C is ∼140 kAU. There are plausible ranges of orbital parameters for Fomalhaut B and C that would be dynamically stable over many orbits.
Could LP 876-10 be genetically related to Fomalhaut AB but we are "catching it in the act" of being an unbound escapee of the Fomalhaut system? We argue that this is very unlikely. LP 876-10 has velocity statistically consistent with that of Fomalhaut A and B (∆S = 1.1 ± 0.7 km s −1 ). If the star actually had a velocity difference of >0.2 km s −1 (i.e. above escape velocity), with respect to the Fomalhaut AB barycenter, then it would not spend much time in the vicinity of Fomalhaut or near its tidal radius. The approximate timescale that LP 876-10 would spend within Fomalhaut's tidal radius is approximately t ≃ r t /∆S, where r t ≃ 1.9 pc is the tidal radius of Fomalhaut AB, and we posit that v must be larger than the escape velocity (0.2 km s −1 = 0.2 pc Myr −1 ). Hence: Table 1. Baraffe et al. (1998) are plotted as dotted lines, and the isochrones for the approximate age of the Pleiades (log(t/yr) = 8.1) and Fomalhaut A & B (log(t/yr) = 8.6) are plotted as long dashed lines. The evolutionary tracks do not accurately predict the solar composition main sequence nor Pleiades sequence for this color-magnitude combination. Note that the Pleiades has a well-determined Lithium depletion boundary and main sequence turn-off age consistent with ∼125 Myr (Ventura et al. 1998;Stauffer et al. 1998;Barrado y Navascués et al. 2004), and the main sequence "turn-on" appears to be consistent with this age as well (Barenfeld et al. 2013). Fomalhaut C appears to lie on the empirical main sequence of Johnson & Apps (2009) | 2013-10-02T16:57:19.000Z | 2013-10-02T00:00:00.000 | {
"year": 2013,
"sha1": "aaf5c6835f5a4fcd3e0b38ece3dbd60570aa6cbe",
"oa_license": null,
"oa_url": "http://joy.chara.gsu.edu/RECONS/published30.pdf",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "aaf5c6835f5a4fcd3e0b38ece3dbd60570aa6cbe",
"s2fieldsofstudy": [
"Physics",
"Geology"
],
"extfieldsofstudy": [
"Physics"
]
} |
238702343 | pes2o/s2orc | v3-fos-license | Ferroelectret energy harvesting with 3D-printed air-spaced cantilever design
Vibrational energy harvesters of air-spaced cantilever design, utilizing ferro-electrets as the electroactive element, are a very recent concept. Such systems, based on the 𝑑 31 piezoelectric effect are further studied with harvesters of improved design, partially implemented by additive manufacturing. The focus of the present work is on the dependence of frequency response, resonance frequency, and generated power on the distance of the ferroelectret from the cantilever beam and on the pre-stressing of the ferroelectret. Experimental data are compared with both analytical and numerical evaluations. It is found that the power generated can be increased by one to two orders of magnitude by proper choice of distance. A suitable pre-stress yields another increase of power by a factor of 2 to 10 and linearizes the response.Thus, normalized output powers more than 1000 μ W referred to an acceleration of 9.81 ms 2 and a seismic mass of 3.5 g, can be achieved, which significantly exceeds previous results of cantilever-based energy harvesters.
INTRODUCTION
Currently most mobile electronic devices use rechargeable batteries. Replacing or recharging those represents a challenging task in a number of applications [1][2][3] such as sensor networks used to monitor widely distributed production and process engineering plants, as well as embedded and implanted sensors for medical applications. In order to supply these electronic devices with energy in places where electric power is not readily avail-This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2021 The Authors. Nano Select published by Wiley-VCH GmbH able, energy harvesting systems have become a steadily growing alternative.
Energy harvesting is defined as the extraction of electrical energy from ambient sources such as heat, vibration or air currents for low power consumption devices [4][5][6][7][8][9] . Although energy harvesters have been investigated and even commercialized, more compact and efficient solutions are required.
In vibrational energy harvesting, piezoelectric transducers utilizing the direct piezoelectric effect have been Nano Select 2022;3:713-722.
wileyonlinelibrary.com/journal/nano 713 F I G U R E 1 A, Cross-section of a parallel-tunnel FEP ferroelectret film. B, Schematic representation of the cantilever structure in idle state. C, 3D representation of the cantilever-based ferroelectret energy harvester. The clamps are used to ensure a reproducible replacement of the cantilever structure without varying the initial pre-stress of the film. D, Cantilever beam loaded with an external load in negative -direction at the free end. This illustrates the cantilever beam under exaggerated bending condition. The black dashed line represents the non-deflected state while the thick black line is the deflected neutral axis of the cantilever beam investigated for the last 20 years, where piezoceramics such as lead zirconate titanate (PZT) have been the dominating materials. [10][11][12][13][14] After the classification of lead as a toxic material, the restrictions to its use have been increased. This, in turn, offered the opportunity for research of new lead-free materials such as ferroelectric polymers represented dominantly by polyvinylidene fluoride (PVDF) and its copolymers. [15][16][17] Other polymer categories that can be considered piezoelectric are the voided charged polymers, also called ferroelectrets or piezoelectrets. [18][19][20][21] In this paper the expression ferroelectrets will be used. The early ferroelectrets were polarized cellular polymers, foremost polypropylene (PP), exhibiting high piezoelectric activity that exceeds PVDF. [22][23][24][25][26] Drawbacks of polarized cellular PP are its missing thermal charge stability and the absence of a significant 31 activity. Due to these limitations, alternative ferroelectret materials, mostly fluorocarbons, have been investigated. [27][28][29][30][31][32][33][34][35][36][37][38][39][40][41] As a rule, these systems are characterized by a significant and thermally more stable 33 piezoelectric response. This makes it possible to implement on their basis rather efficient vibration-based micro-energy harvesters. [42][43][44][45][46][47][48][49] A promising new development are parallel tunnel films. [50][51][52] These consist, for example, of two fluorinated polyethylene propylene (FEP) films partly fused together at elevated temperatures with enclosed airfilled tunnels and surface metallization ( Figure 1A). Beside their superior electret properties as well as their far better thermal stability than PP, these structures exhibit not only similar longitudinal piezoelectric coefficients as piezoceramics but also have a large 31 -activity. This allows their use in energy harvesters based on the transverse piezoelectric effect, including cantilever systems. [50,51,53] A recently presented example is an energy harvester where the seismic mass is placed on the center section of a ferroelectret strip, which in turn is fixed at both ends. [50,51] Consequently, when the support structure is subjected to an acceleration , the ferroelectret film undergoes a dynamic stretching due to the inertia of the seismic mass. However, such harvesters, while generating large normalized power outputs, lack the robustness of ceramic implementations.
A more robust system based on an air-spaced cantilever structure was recently reported. [53] This energy harvester consists of a parallel-tunnel ferroelectret, stretched by a cantilever beam which is subjected to an acceleration . The ferroelectret film was mounted at a fixed distance ℎ from the neutral axis of the cantilever beam. A theory was offered which explained the dependency of the output power and resonance frequency on the seismic mass. It was also shown that the resonance frequency mainly depends on the stiffness of the cantilever beam. However, for accelerations above 0.06 , the generated power increases more slowly than expected theoretically, due to an asymmetrical stretching of the film.
In the present paper, significant improvements of the harvester output power are achieved by varying the distance ℎ ( Figure. 1) and utilizing different values of mechanical pre-stress. The theory of the previous work is shown in detail and used to explain the influence of different distances ℎ. The influence of the pre-stress on the output power and the resonance frequency is also addressed.
THEORY
When the cantilever base is subjected to an acceleration in -direction, the tip mass is deflected by inertia into the negative -direction from its position of equilibrium and forms an angle with the cantilever's plane [(see Figure 1D)]. The deflection of the free end of a cantilever depends on the seismic mass and thus is the result of a point force concentrated at its free end pointing into the negative -direction. The thereby obtained deflection is given by [54] where C is the length of the cantilever, its Young's modulus and the moment of inertia of the beam. It should be mentioned that the force is also pointing in the same direction as the inertia induced deflection . For small deflections, the displacement of the tip of the cantilever in the -direction can be neglected. The transverse contraction of the beam and the ferroelectret film is also neglected. The angle can be expressed as [55] = 3 2 C . ( The change of the length of the ferroelectret film is then geometrically determined by the intercept theorem, still neglecting the displacement of the cantilever tip indirection, as suggested by Beléndez et al. [55] Under these assumptions one obtains: Since the axis is defined as the direction of acceleration and the axis is the perpendicular axis, the equation of motion for the cantilever's tip in -direction can be written as:̈+ and for a sinusoidal steady-state motion this yields to where is the exciting angular frequency of the cantilever base, 0 = √ C the resonance frequency of the undamped harmonic oscillator of the beam which depends on the cantilever stiffness C and the damping constant of the beam. In this approach the influence of the spring constant f and the damping constant of the ferroelectret on the resulting cantilever properties are neglected. Substituting in Equation (3) the film's elongation Δ can then be expressed as follows: Using the direct piezoelectric voltage coefficient 31 = ( OC 1 ) ⋅ F linking the generated open circuit voltage OC to the applied force 1 on the ferroelectret film indirection (which differs from the force in -direction) and utilizing the mechanical force constant f of the ferroelectret film, the voltage OC generated in open circuit can be written as: where F is the ferroelectret film width. The rms-value of the voltage R across a load resistance L connected between the electrodes of the ferroelectret film can be written by means of a voltage divider as: where is the static capacitance of the ferroelectret film. Since this equation is only valid for small displacements, F I G U R E 2 A, Frequency responses of the normalized power of energy harvesters with various distances ℎ between the ferroelectret film and the cantilever beam for a seismic mass s = 3.5 . B, the normalized power at resonance frequency as a function of distance ℎ. The theoretical curve (orange dashed) for n is calculated using Equations (10) and (12) the change of film capacitance is much smaller than its static value. Thus, the capacitance can be assumed to be constant. Increasing the acceleration of the cantilever base results in further stretching of the ferroelectret film, and, thus, results in a higher output power, that is, It follows from Equation (9), that if the generated power is h 0 for the smallest experimentally achievable distance ℎ 0 (in our case ℎ 0 = 1 ), the power h corresponding to a distance of ℎ is increasing quadratically according to The experimental output power out generated by the air-spaced energy harvester for different distances ℎ was obtained experimentally (see Experimental Section) from the relation where is the current through the load resistor and the angular frequency of the acceleration . The normalized value n is then given by where = 9.81 2 is the acceleration of gravity.
Influence of film distance on the output power
All measurements were performed at the same acceleration = 0.1 × , the same pre-stress and seismic mass s = 3.5 , for a proper test of the power generated for different values of ℎ. The measurements indicate that the normalized output power increases from 16 to 438 μW for ℎ = 1 to ℎ = 10 ( Figure 2A). The power decreases for larger distances ℎ. In all cases the resonance frequency shifts toward higher frequencies, indicating an increasing influence of the film stiffness f on the cantilever stiffness C .
Note that in the present analysis C and f are assumed to be independent of each other, which is only valid for small angles and small values of ℎ resulting in small film displacements Δ . For larger angles and larger ℎ values, f starts to influence 0 toward higher frequencies.
According to the above presented theory, the film length variation Δ is assumed to be proportional to ℎ.
Increasing ℎ enhances the stiffness of the film-cantilever system. In fact, when the air-tunnels are subjected to more stress in -direction, the stiffness of the solid FEP layers enclosing the air tunnels becomes more and more dominant compared to the air-filled regions. This means that the ferroelectret stiffness f is no longer negligible compared to the beam stiffness C . This leads to an effectively larger C in the above theory which then results in a shift to higher resonance frequencies ( Figure 2A) and at larger distances ℎ to even smaller normalized power see (Figure 2A).
The first reason for the decrease of the power is related to the fact that the voltage generating tunnels of the ferroelectret become more and more rigid with increasing ℎ, leading to a smaller voltage output (smaller 31 ) and thus to a decrease of the power generation. This effect, for the present devices, is noticeable starting from a film distance of ℎ = 5 ( Figure 2B). The second reason for the flattening of the generated output signal at larger ℎvalues is mainly due to the asymmetry of the ferroelectret stretching, when the beam is bent in positive and negative -direction, as was demonstrated in previous work. [53] The reason for such asymmetry is the different mechanical stress the ferroelectret film is exposed to, when the beam moves down (stretching) and up (loosening).
3.2
Influence of pre-stress on output power and device stability In order to avoid such asymmetry and to achieve a symmetric deformation compared to the initial state of the ferroelectret film during operation, a relocation of the operating point of the device by pre-stressing the ferroelectret film is investigated. Two distances ℎ between the neutral axis of the cantilever beam and the ferroelectret film were considered, namely, 1 and 5 mm, respectively, the latter being close to the film distance with maximum normalized power output ( Figure 2B). The pre-stressing of the film has to be carried out using the setup in Figure 1. The clamp at the end of the cantilever beam (see Figure 1C) is fixed and represents the seismic mass s .
On the side of the cantilever base, a movable clamp is used that can slide in a reproducible way in a guide structure. This structure is controllable in the negativedirection by a micrometer screw. By such sliding the produced strain results in a pre-stress of the ferroelectret film. The term pre-stress is used since the performed straining leads to a slight bending of the beam, which even after mechanical relaxation results in a remaining pre-stress on the ferroelectret. If under such conditions the cantilever is set into vibrations, the individual tunnels will be stretched when the cantilever is bent downward and the stress is partially released when bent upward. The difference to a stress-free bending of a displaced ferroelectret is that under upward bending the film still stays under stress which is not the case without pre-stress. Advantageous for the prestress is that the same deformation occurs in both bending directions leading to an equal energy gain in both directions of acceleration. The disadvantage, however, is that an excessive pre-stress of the sample leads to a decrease in the 31 -coefficient, and, therefore, to a decrease in the generated power (see Equation 9). The optimum pre-stress is difficult to predict and has to be determined experimentally. In order to find the optimum pre-stress for our energy harvester, measurements under different pre-stress conditions were conducted using the same seismic mass of s = 3.5 and a fixed acceleration = 0.1 × .
For ℎ = 5 a 2% strained sample leads to a twice as large peak power of 737 μW compared to the unstrained structure. For ℎ = 1 the output power is relatively even more enhanced up to 253 μW by straining to 3.66%, however, still below the value reached for ℎ = 5 ( Figure 3A and B). Unfortunately, one can uniquely specify the prestrain applied to the samples. The physically important property is, however, the pre-stress which cannot be determined quantitatively. Even so, the parameter stress or prestress will be mostly used in the following discussions.
Considering the nonlinear behavior of the stress-strain curve of the used ferroelectret as described before, [50] a reallocation of the working point also influences the stiffness of the ferroelectret film and consequently the 31 coefficient. Since the dynamic displacement Δ increases with the distance ℎ, the stiffness enhancement of the filmcantilever system for larger ℎ is more noticeable (Figure 3C and D).
Despite the fact that the dynamic deformation of the film while vibrating is much larger for larger ℎ (see Equation 3), it can be noted that the maximum power output of the prestressed film for ℎ = 1 corresponds approximately to the output power of the cantilever structure without prestress using a distance of ℎ = 5 ( Figure 3C and D). The comparability of the two maximum power outputs indicates that the limit is determined neither by the distance ℎ of the ferroelectret film nor by the pre-stress but rather by a common source which has to be looked for in the ferroelectret film itself. Since for both cases (increasing of distance or strain) the power output drops below its maximal value, the reason for the decrease must be assumed to be in the film itself. It is most likely originating from the deformation of the tunnels after reaching saturation, where any further deformation starts to strongly decrease the piezoelectric 31 voltage coefficient. Under positive or negative loads the thickness of the tunnels decreases to a value, where the induced strain mainly acts on the already strongly elongated FEP framework, instead of compressing the tunnels any further. The consequence thereof is a strongly reducing additional strain, leading to a reduction of the power generation.
A further interesting benefit of utilizing a pre-stress is the improved temporal stability of the output power as can be seen from Figure 4A. For this experiment, the power output of the energy harvester was determined for an acceleration of = 0.1 × and a seismic mass of s = 3.5 . Approximately 9 million cycles were applied, which correspond to approximately 70 hours of operation at a frequency of 35Hz, which is a reasonable stress test. Pre-stressing of the film results in a stable output power of 720 μW over the measurement duration, whereas the absence of pre-stress leads to a steady increase of the output power. . The measurements are conducted using the same acceleration = 0.1 × and seismic mass of s = 3.5 . B, Normalized output power at resonance frequency for increasing acceleration Another advantage of the pre-stressed ferroelectrets compared to the stress-free ones is the enhancement of the power stability as a function of acceleration as shown in Figure 4B. When no pre-stress is used, the measurements indicate a decay of the normalized power as soon as an acceleration of = 0.04 × is reached. This limit can F I G U R E 5 A, Stress-strain relation of the ferroelectret film, while holding each strain value for 60 minutes. B, Normalized power at the resonance frequency. The first measurement is conducted directly after the adjustment of the pre-strain and the second measurement 24 hour afterwards, while keeping the film under the same pre-strain. C, Hysteresis of the normalized output power at resonance frequency for ascending and descending pre-strain be enhanced up to = 0.1 × , when a pre-strain of 2% is applied to the ferroelectret film. In addition, the output power is enhanced from 270 to 737 μW.
Viscoelastic property
Considering the viscoelastic and plastic properties of solid FEP-films used to manufacture parallel-tunnel ferroelectrets, one can expect that under pre-stress the ferroelectret film behaves accordingly. In order to examine such effects a fixed pre-strain was applied to the ferroelectret and its impact on three aspects was investigated. The first aspect is the constancy of the pre-stress for a given prestrain. It was measured using a universal testing machine (Inspekt table 5 kN, Hegewald & Peschke). Hereby, the prestrain is increased in steps of 0.33% up to 4% while simultaneously measuring the force (as indication for a varying stress) that results from keeping the applied strain constant ( Figure 5A). For each pre-strain the position is maintained for 60 minutes while measuring the force developing while preserving the reached position ( Figure 5A). To avoid over-shooting, a small velocity of 0.1 − 1 was used to move from one strain to another. The measurement shows different creep behavior of the ferroelectret film depending on the pre-strain. By using 0.33% pre-strain to stretch the ferroelectret film, a negligible decay of the resultant initial stress is reached after 60 minutes. On the other hand, when using a pre-strain of 2%, which corresponds to the optimal strain regarding the output power, a decay of approximately 4% of the initial pre-stress is measured after 60 minutes. A drop of 10% of the initial pre-stress has been recorded for an even higher pre-strain of 3.7%.
The second aspect was to find out whether the mechanical relaxation has an effect on the generated power at the resonance frequency of 35Hz. Therefore the constancy of the power was examined, when the harvester is not used continuously. Two measurements were conducted for each pre-strain. The first measurement is directly carried out after the adjustment of the pre-strain and the second measurement is conducted 24 hours later. The pre-strain is increased by the aforementioned steps in order to compare the mechanical and electrical properties under almost the F I G U R E 6 Schematics of experimental setup used for energy harvesting evaluation same conditions. The resulting output powers show almost the same values (see Figure 5B) with a small increase in power for 2% pre-strain.
The third aspect is the hysteresis of the output power for ascending and descending pre-stress ( Figure 5C). The pre-stress is increased by increasing the strain in steps of 0.33% until a strain of 3.66% is reached and decreased by reducing the pre-strain with the same step size down to 0%. For each pre-strain, the frequency response of the harvester is measured and the achieved amplitude at the resonance frequency is noted. Therefore, each pre-strain is maintained during the measurement time of approximately 30 minutes. For an ascending pre-strain, the highest reached amplitude is 737 μW at 2% pre-strain, whereas, a descending pre-strain shows a different amplitude and different optimal pre-strain, namely 1029 μW and 1.66% respectively. Starting from a pre-strain of 2.66% the hysteresis becomes thinner (if the pre-strain increases) and the achieved amplitudes for both ascending and descending pre-stress become more similar.
The increase of the output power for ascending prestrain can be explained by the fact that the stiffness of the film has decreased due to the excessive stretching. The decrease of the stiffness f can be explained by the viscoelasticity of the FEP films used to manufacture the fer-roelectret. This means a larger deformation of the air-filled tunnels can be reached for the same force, resulting in a larger piezoelectric voltage constant 31 . This hysteresis can also be used to improve the performance of the energy harvester without adjustment of the geometry. Before the energy Harvester starts to operate, the ferroelectret film can be stretched and then relaxed until the new optimal working point is reached. When the energy harvester is set into vibration, a maximum power of 1029 μW can be reached. Note that all hysteresis measurements are reproducible for pre-strains not exceeding 4%. An excessive prestrain results in a plastic deformation of the samples and therefore in a completely different behavior. A detailed study about the influence of viscoelasticity on the output power is beyond the scope of this work and will be published separately.
EXPERIMENTAL SECTION
The smallest experimental distance ℎ 0 was achieved by placing the ferroelectret film at a distance of 1 mm above the neutral axis of the cantilever beam. Afterwards, the distance ℎ is varied up to 20 mm. The seismic mass is designed as a hollow structure to minimize its weight. Furthermore, all parts including screws are made of plastic. Since the characteristics of different ferroelectret films are likely to differ slightly, only a single film was used for all measurements presented in this paper. Thereby, the change in the generated power during the experiments is exclusively caused by only one experimental parameter, namely the different distance ℎ. Therefore, all measurements have been conducted using the same acceleration and seismic mass. This suppresses the influence of unknown factors, such as film slipping from the clamp that may be triggered by large accelerations particularly for larger ℎ. For changing the distance ℎ, the cantilever structure is modified accordingly. In order to avoid different mechanical pre-stress of the film, each time the harvester is assembled for a new distance ℎ, the ferroelectret film length is fixed between the two 3D printed clamps by a spacer block with a fixed length allowing for an undisturbed dismantlement and installation. The ferroelectret film used in this study is made of fluorinated polyethylene propylen (FEP) and has the dimensions of 10 × 40 mm 2 . It exhibits static and dynamic piezoelectric 31 -coefficients of 1.9 −1 and 0.8 −1 , respectively. The energy-harvesting device is an air-spaced cantilever arrangement that was mostly made by additive manufacturing technique known as fused deposition modeling (FDM). The used cantilever material hereby was polylactide acid (PLA) with 15% infill and a gyroid fill pattern. The cantilever beam of dimension 10 × 30 mm 2 is responsible for the stiffness of the mechanical resonant system. The ferroelectret film is incorporated into the structure by clamping it between the cantilever base and the seismic mass ( Figure 1B and C). The metal electrodes are connected to an external electrical network consisting of the load resistor L in addition to a charge amplifier and a voltage meter (see Figure 6). The device was tested by exposing it to vibrations with an acceleration generated by a shaker (Brüel&Kjaer type 4809). The generated power was determined as a function of frequency by terminating the ferroelectret with a load resistor L = 1 0 , where 0 is the mechanical resonance frequency of the device and the static capacitance of the ferroelectret.
CONCLUSION
In conclusion, the dependence of the power output of an energy harvester arrangement on the distance between the ferroelectret and the cantilever beam has been investigated. This harvester can be easily manufactured using additive manufacturing technique (3D printing) and provides a modular assembly which ensures reproducibility. Such a harvester generates a normalized output power of more than 1 mW at a resonance frequency of about 35 Hz for a seismic mass of 3.5 g and an acceleration of 0.1 × . It was experimentally verified that the power generated at the resonance frequency can be stabilized over the measurement duration of 70 hours as well as for different accelerations up to 0.1 × by using mechanical pre-stress. The influence of varying the distance between the neutral axis of the cantilever beam and the ferroelectret film shows an excellent agreement with theory up to a distance of ℎ = 10 . Thus, the presented energy harvester proves to be particularly suitable for low acceleration amplitudes up to 0.1 × . It turns out to be a reproducible, environmentally friendly by using polylactide acid for the cantilever structure and particularly robust system with relatively high sensitivity.
The energy output of a cantilever based energy harvester can be improved by using two ferroelectret films to build a bimorph cantilever structure. The ferroelectret films should be placed at the optimal distance from the neutral axis, which can be investigated experimentally. Such a structure offers twice as much output power than when using a single film. When each film is optimally pre-stressed, the output power can be further enhanced to reach the range of several milliwatts. Open access funding enabled and organized by Projekt DEAL.
C O N F L I C T O F I N T E R E S T
The authors declare no conflict of interest.
D ATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request. | 2021-09-27T18:27:46.857Z | 2021-08-18T00:00:00.000 | {
"year": 2021,
"sha1": "a8c29c57d9ced6c3a535961224a777138bcdd9ff",
"oa_license": "CCBY",
"oa_url": "https://tuprints.ulb.tu-darmstadt.de/21209/1/Nano%20Select%20-%202021%20-%20Ben%20Dali%20-%20Ferroelectret%20energy%20harvesting%20with%203D%25u2010printed%20air%25u2010spaced%20cantilever%20design.pdf",
"oa_status": "GREEN",
"pdf_src": "Anansi",
"pdf_hash": "c6e259e90e0716501a93adcd6df970c496caa6b6",
"s2fieldsofstudy": [
"Engineering",
"Physics",
"Materials Science"
],
"extfieldsofstudy": [
"Materials Science"
]
} |
54199356 | pes2o/s2orc | v3-fos-license | Is there any difference between Vitamin D deficiency among heart failure and non-heart failure elderly with cardiovascular disease ?
Introduction In recent years, the pathophysiologic concept of heart failure (HF) has changed from an isolated hemodynamic view to more complex concept involving neurohormonal overactivation and increased concentration of proinflamatory cytokines, such as TNF and interleukin 6 (1,2). Interestingly, experimental studies have shown that the vitamin D can suppress the release of TNF-α (2). Vitamin D deficiency seems to predispose to hypertension, diabetes and metabolic syndrome, left ventricular hypertrophy, chronic vascular inflammation, pathogenesis of atherogenic dyslipidemia and ischemic heart disease (3-5). Possible causal mechanism that may link vitamin D deficiency to heart failure, may involve the regulatory effect of 1,25 (OH)2Vit D on myocardial gen expressions (6).Vitamin D has also been shown to reduce the activation of the systemic and cardiac rennin angiotensin system, which contributes to hypertension and cardiac hypertrophy (4). Abstract Vitamin D deficiency is common among the elderly with and without heart failure (HF). This study was designed to identify the degree of vitamin D deficiency among HF and nonHF elderly with cardiovascular disease. In this study, fifty hospitalized elderly patients with HF (26 women and 24 men, age range 60-90 years) and fifty elderly without HF were randomly selected. The level of serum 25-hydroxy vitamin D was measured using the Eliza method. There were no significantly difference between two groups regarding to the mean age, living location, life style parameters, mean years of disease involvement and disease severity, cardiovascular risk factors and comorbidities, except for renal problems(p = 0.039), hemoglobin levels(p = 0/044), and creatinine levels (p = 0.048). In both groups, the majority of patients, were moderately vitamin D deficient and there was no significantly difference between two groups. Considering the high prevalence of vitamin D deficiency among HF and non-HF elderly with cardiovascular disease, its monitoring and correcting seems essential.
Introduction
In recent years, the pathophysiologic concept of heart failure (HF) has changed from an isolated hemodynamic view to more complex concept involving neurohormonal overactivation and increased concentration of proinflamatory cytokines, such as TNF and interleukin 6 (1,2). Interestingly, experimental studies have shown that the vitamin D can suppress the release of TNF-α (2). Vitamin D deficiency seems to predispose to hypertension, diabetes and metabolic syndrome, left ventricular hypertrophy, chronic vascular inflammation, pathogenesis of atherogenic dyslipidemia and ischemic heart disease (3)(4)(5). Possible causal mechanism that may link vitamin D deficiency to heart failure, may involve the regulatory effect of 1,25 (OH) 2 Vit D on myocardial gen expressions (6).Vitamin D has also been shown to reduce the activation of the systemic and cardiac rennin angiotensin system, which contributes to hypertension and cardiac hypertrophy (4).
Myocardial calcium homeostasis, which is crucial for the contractility and electrophysiology of the heart, is also partially regulated by 1,25(OH) 2 Vit D. This is mediated by its influence on ion channels and enzymatic reactions (7). The results of several studies show that vitamin D supplementation significantly decreases all-cause mortality (6,8). Exposure of arms and legs for about 5 to 30 minutes ( depending of time of day, season, latitude, air pollution, the degree of cloud cover, age, the extent of clothing covering, and skin pigmentation) between the hours of 10 am and 3pm , twice a week is often adequate for conversion 7dehydrocholestrol (existed in the skin) to pre-vitamin D3 (4). It has been demonstrated that regular exposure to UVB increases circulating 25(OH) Vit D above a level of 10 nmol/l (40 ng/l) and also significantly reduces blood pressure by approximately 6 mm/Hg in hypertensive patients (9). Sunlight exposure and dietary intake of vitamin D may not provide adequate levels of vitamin D. Some studies suggest an intake of at least 800 IU of vitamin D daily among elderly, especially in the winter (10). Consumption of some drugs such as anticonvulsants, steroids, antiacids, H2 blockers and proton pomp inhibitors, prevent vitamin D absorption (11). According to several studies , 40 to %100 of US and European community dwelling elderly involved in vitamin D deficiency (6). Older people are prone to develop vitamin D deficiency, because of various risk factors including: decreased dietary intake, diminished sun light exposure, reduced skin thickness, impaired intestinal absorption, and impaired hydroxylation in the liver and kidneys (5,12). Vitamin D deficiency may play a significant role in the pathogenesis of heart failure (6,7,13), though there is insufficient evidence to develop recommendations for strategies that improve vitamin D deficiency for HF patients. In addition, little is known about the degree of vitamin D deficiency among HF and non-HF elderly. The aim of this study was to identify the degree of vitamin D deficiency among HF and non-HF elderly with cardiovascular disease.
Material and methods
In this cross sectional study, fifty hospitalized elderly patients with HF and the same number elderly patients without HF were randomly selected. This study was conducted between October 2011 and March 2012, at the Sari Heart Center located in Mazandaran province in Iran. All non-HF patients had a diagnosis of ischemic heart disease. Their diagnosis were confirmed by the cardiologist and under angiography. Persons who had history of taking vitamin D during past six months, were excluded. Data gathering tool was a questionnaire that was developed by researchers and its validity and reliability were confirmed by expert panel and test -retest (r = 0.79). In this questionnaire, there were some questions related to demographic characteristics (age, gender, location) and factors affecting vitamin D deficiency such as cardiovascular risk factors (history of diabetes, hyper-lipidemia, hypertension, overweight, anemia, hemoglobin level, ejection fraction < 40%, mean of systolic and diastolic pressures, body mass index, disease severity (using the NYHA class), biochemical characteristics (fasting blood sugar, cholesterol, creatinine, phosphorus, calcium), the diseases and drugs predisposing vitamin D deficiency (history of hepatic, renal, and gastrointestinal diseases, consumption of sun-block, anticonvulsants, anti-acids, H2 blockers, steroids), environmental factors (the number of walks per week and its duration, the number of sun exposure per week and its duration, the number of fish consumption per week and the type of fish, the number of dairy product consumption per week).This study was approved by the Research Ethics Committee of the Mazandaran University of Medical Sciences (Number 90-3). After completing of the informed consent by participants, 5 mL of venous blood was taken. Sample centrifuging and serum extraction were done and were immediately frozen. The level of serum 25(OH) Vit D was measured by the Eliza method (IDS ltd kit, UK). Although a consensus regarding the optimal level of serum 25(OH) Vit D has not yet been established, most experts define vitamin D deficiency as a 25(OH) Vit D level of < 50 nmol/l (19 ng/ml) (14). Based on a randomized clustered sampling from the Tehran population that was designed by Endocrinology and Metabolism Research Center of Tehran University, the cut off point for vitamin D deficiency was considered less than 35 nmol/l (15).
Statistically analysis
Logistic regression was performed for qualitative variables and T-test was used for quantitative variables. Values p < 0.05 were considered significant. The results were analyzed by SPSS software version 16.
Results
Both HF and non-HF groups include 26 women and 24 men. The mean of the age were 73.2 ±7.69 years and 70.44 ± 7.8 years in HF and non-HF patients, respectively (p = 0.078).The majority of the elderly with HF (62%) were rural, while the majority of non-HF elderly (54%) were urban (p = 0.16). The mean years of disease involvement were 5.06± 4.7 years in HF patients (with CI 95% 3.72-6.4) and 5.34 ± 4.96 years (with CI 95% 3.93-6.75) in non-HF patients (p = 0.77). Among HF patients, the number of elderly males with vitamin D deficiency were significantly more than females (83.3 versus 53.8%, p = 0.026). The number of the patients with diabetes, hyperlipidemia, overweight, anemia and ejection fraction (EF) < 40% were more in the elderly with HF compared to those without. There was no significantly difference between two groups, except for EF (p < 0.001) ( Table 1). Regarding to severity of disease, the majority of HF(48%) and non-HF patients (40%) were class IV (p = 0.559).
Diseases and drugs
HF patients had a more renal, hepatic, and gastrointestinal problems and a more consumption of H 2 blockers and anticonvulsants drugs, compared to non-HF patients. The usage of sun-block, steroids, anti-acids, were more in non-HF patients compared to HF patients. There was no significantly difference between two groups, except for renal problems (p = 0.039).
Environmental factors
The number of walks per week, its duration and sun exposure were lower in the elderly with HF compared to those without HF (p = 0.172 and p = 0.223 and p = 0.426 respectively). The amount of milk and fish consumption were more in the elderly with HF compared to those without HF (p = 0.420 and p = 0.546 respectively).
Biochemical characteristics, blood pressure and BMI
The mean of two times fasting blood sugar, cholesterol, creatinine, and phosphor levels were more in the elderly with HF compared to those without HF. The mean of two times systolic and diastolic blood pressure during 3 days, BMI, hemoglobin, and calcium levels were lower in the elderly with HF compared to those without HF. There was significantly difference between two groups, regarding to hemoglobin and creatinine levels (p = 0.044 and p = 0.048 respectively).
Vitamin D levels
The mean vitamin D levels were 33.76 ± 2.82 and 41.76 ± 3.06 nmol/l, in elderly persons with and without HF respectively(p = 0.178). In both groups, the majority of patients (38%, 19 persons), were moderately vitamin D deficient ( Table 2). The odds ratio of vitamin D deficiency was more in elderly with HF (OR=1.35; 95% CI 0.87-2.21; p = 0.151) compared to those without HF.
Vitamin D deficiency group (<35nmol/l)
The odds ratio of the overweight was significantly more among HF patients compared to those without HF (OR = 1.73; 95% CI 1.22-2.68; p = 0.031). In addition, the odds ratio of the gastrointestinal disorders (OR = 1.81; 95%CI 1.23-2.68; P=0.022) and renal disorders (OR = 1.74; 95%CI 1.16-2.61; p = 0.031) were significantly more among HF patients compared to those without HF (Table 3). The number of the elderly with HF who never walks per week, were more than those without HF (41.2% versus 33.3%; p = 0.397).The number of the elderly with HF who used sun exposure daily, were lower than those without HF (55.6% versus 44.1%, p = 0.856). There were no significantly differences between the number and type of fish usage, and the number of dairy consumption, among the elderly with and without HF (P = 0.241 and p = 1 and p = 0.194 respectively).
Discussion
The results of this study didn't show significantly correlation between vitamin D levels among the elderly with and without HF. In both groups, there were no statistical relationship between the age and vitamin D deficiency. This finding is contrast with the results of some studies (7,16) and adjusts with other ones (17,18). With increasing of the age, the probability of vitamin D deficiency is increased, because of decreased skin thickness, decreased dietary intake, diminished physical activity out of home, increased prevalence of the kidney, liver and gastrointestinal problems (5). No significantly relationship between the age and vitamin D deficiency may be explained by the relative youth of participants (mean age 72). Among HF patients, the number of men with vitamin D deficiency were significantly more than women. This finding adjusts with the results of one study (19) and is contrast with other ones (18,20). The reason may be a more disease severity among men compared to women. The odds ratio of the overweight was significantly more among HF patients compared to those without HF. Diminished exposure to UV light, associated with reduced outdoor activities and likely physical inactivity may account, in part, for the lower level of serum vitamin D in overweight participants (21). In addition, the lipid solubility of vitamin D modifies its bioavailability and may contribute to the lower level of serum vitamin D in overweight and obese participants (22,23). Among elderly with HF, there was not significantly relationship between the number of walks per week and vitamin D levels. It may be due to relatively small sample size. The odds ratio of the gastrointestinal and renal disorders were significantly more among the elderly with HF compared to those without HF. Vitamin D deficiency in chronic renal disease, may be due to hyper-phosphatemia. Hyperphosphatemia increases fibroblast growth factor 23, which decreases 25hydroxyvitamin D-1α-hydroxylase activity (24).
Limitations
A number of limitations exist with respect to the present study. Firstly, this is a cross sectional study and the correlations can not imply the causation relationships between parameters. Secondly, there are different normal vitamin D ranges for winter and summer in different countries. Thirdly, the sample size is relatively small. More ever, the data represent only the subjects who agreed to participate in this study.
Conclusion
Considering the high prevalence of vitamin D deficiency among the HF and | 2018-12-02T20:24:57.237Z | 2016-02-10T00:00:00.000 | {
"year": 2016,
"sha1": "0f2301baca184ca8ba485f2e48507655264bdcdf",
"oa_license": "CCBYNC",
"oa_url": "http://pbr.mazums.ac.ir/files/site1/user_files_88c428/danial-A-10-26-29-04b5f70.pdf",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "2dbf401dea0deae3ca456015fd022ed9496bef4e",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
246579891 | pes2o/s2orc | v3-fos-license | Innovation in Making Meatballs, Pempek, and Sate Lilit with Snow Mushrooms (Tremella Fuciformis)
Snow mushrooms are typically served directly, such as sautéed or fried with flour coating. More variations on the snow mushrooms-based food is expected to provide new experiences for people who want to enjoy mushrooms with the home flavor foods, particularly for people who do not eat meat. The purpose of this study is to determine the outcomes of the innovation of meatballs, pempek, and sate lilit with snow mushrooms as the basic ingredients. The snow mushrooms were processed as the basic ingredients for making meatballs, pempek, and satay. The foods were evaluated by 25 trained panelists: Culinary Department lecturers and students, in terms of their aromas, textures, colors, and tastes. Panelists were given a checklist to assess the results of the observations, which were then scored using a Likert scale to determine whether the quality of the meatball, pempek, and sate lilit were very good, good, quite good, not good, or very bad. The experiment was carried out in the kitchen of Politeknik Pariwisata Bali. Based on the results of experiments and organoleptic assessments, it can be seen that the innovation of meatballs, pempek, and sate lilit yields very good results, with an average percentage of 80% for meatballs, 83% for pempek, and
Introduction
In the food processing industry, innovation is required to meet the tastes and desires of consumers. In today's world, there is a growing demand for nutritionally dense foods as well as a desire for good health. This is evidenced by the growing of vegetarian lifestyle, which influences the diets of consumers who used to consume meat but now do not.
Mushrooms are vegetable food ingredients that can be processed into a variety of foods due to their chewy texture when cooked, resembling the texture of meat-based foods. According to Sumarmi (Rahayu et al.,2015), mushrooms are becoming increasingly popular among the community because, in addition to their low cost, they have a high nutritional content. Meanwhile, the snow mushroom (Tremella fuciformis) is a species of heterobasidiomycetes (jelly fungi) with a fruiting body that is shaped like an irregular tassel, white, and very clear like jelly (Wikipedia, 2017).
Mushroom is generally used as an ingredient in sautéed, fried, flour-based foods, or as a filling in soup making. However, mushrooms are not commonly used to replace meat in food preparations that are typically made from meat or fish. Some people, particularly vegetarians, use gluten as a meat substitute in cooking in order to retain the texture, taste, and aroma of meat. According to Listiyowati (2015), the protein, vitamins, and minerals found in food mushrooms have the potential to replace beef, whose demand rises year after year in tandem with the Indonesian people's per capita economy.
Food development in food processing is done not only by developing techniques for cooking food ingredients, but also by changing or adding new ingredients to certain types of food with the goal of producing foods that have a high taste, are appealing, have nutritional value, and are safe to consume.
Meatballs, pempek, and sate lilit are among the most popular and widely consumed Indonesian foods. Meatballs are typically made by grinding the main ingredients that is meat, mixing them with other ingredients, shaping them into a ball shape, and then boiling them. Originally, beef is used as the raw material for meatballs, but as the industry has evolved, chicken and fish are also used as raw materials.
Satay or sate in Indonesia are extremely diverse due to the fact that each region has its own specialty. Bali is one of the areas that has satay, which is popular in Indonesia. There are many different types of satay available in Bali, but sate lilit is the specialty of the the island. Sate lilit is a type of satay that is typically made from minced meat (fish, chicken, pork). Meanwhile, pempek is a traditional South Sumatran dish made of ground fish, tapioca flour, water, salt, and spices.
Based on this description, the processing snow mushrooms as the basic ingredients in the manufacture of meatballs, pempek, and sate lilit is expected to be a viable alternative to meeting the food needs of the community members who do not consume meat as their primary source of protein. The texture, aroma, and taste of the foods made with snow mushrooms as the main ingredient to replace meat are expected to resemble processed meat foods like meatballs, pempek, and sate lilit. and communication between countries and continents, the delicacy and distinct flavor of mushrooms has spread all over the world. Mushrooms have evolved into a popular and prestigious dish. Mushrooms are classified as commodities that are inextricably linked to human needs. Bread, tempeh, tape, taoco, oncom, and various drugs, such as penicillin, are all impossible to make without mushrooms. Several types of mushrooms can be used in place of meat, fish, and other expensive and nutritious foods. Mushrooms are the preferred alternative food (preference) of people at all socioeconomic levels. Timlo, mushroom fried rice, mushroom tauco, sukiyaki, and noodles are all popular mushroom dishes in Indonesia. Mushrooms have twice the protein content of asparagus and potatoes, four times that of carrots and tomatoes, and six times that of oranges. At least 6-7 types of amino acids are found in mushrooms, out of the approximately 15 essential amino acids required by the human body (Djarijah, M., Djarijah, S. 2001).
The snow mushroom (Tremella fuciformis) is a heterobasidiomycetes (jelly fungi) species with fruiting bodies that are irregular in shape, white in color, and very clear.
Snow mushroom (Tremella fuciformis) is a high-protein food ingredient that is also high in vitamins and minerals and has a low carbohydrate, fat, and calorie content. Composition of nutrients 11.4 percent moisture; 3.4 percent ash; 5.7 percent protein; 76.6 percent carbohydrates; 4.9 milligrams per gram of sodium; 3.2 milligrams per kilogram of potassium; and 1.1 milligrams per kilogram of calcium (Khondkar in utami et al, 2015). Meanwhile, Tjokrokusumo (2015) claims that mushrooms have high protein levels ranging from 17.5 to 27 percent, low fat levels ranging from 1.6 to 8%, and high levels of good food fiber ranging from 8 to 11.5 percent, making them suitable for use as healthy food ingredients.
Meatball (Bakso)
Meatball is a traditional Indonesian food made by combining meat, salt, onions, and tapioca flour into a dough that is then formed into ping pong-sized balls before being cooked in boiling water (Soeparno, 2005). In addition, Melia (2010) explained that meatballs are food ingredients made from meat as the main ingredient, both beef, chicken, fish, shrimp and duck meat. Meatball is meat that has been mashed and mixed with other additives and spices so that the meatballs are more delicious.
Meatball derives from Chinese-Indonesian culinary arts. This is demonstrated by the term bakso, which is derived from the Hokkien word Bak-So, which literally means "ground meat." Meatballs are more commonly made from halal meat, such as beef, fish, or chicken, because the majority of Indonesians are Muslims.
According to the National Standardization Council listed in SNI No. 01-3818-1995, Meatball is a food product in the form of a circle or other obtained from a mixture of livestock meat (meat content not less than 50%) and starch (cereals) with or without the addition of other food ingredients, as well as permitted food ingredients. The quality of meatballs is very much determined by the quality of the raw materials, especially the type and quality of meat, the type of flour used and the ratio in the dough (Astawan, 1989).
The quality of the meatball is determined by the amount of flour mixture that is added, the more flour used the lower the quality of the meatball. The chemical composition of meatballs is determined by the chemical composition of the constituent materials (Wibowo, 1999). The raw materials will also affect the quality of the meatballs produced. Meanwhile, other factors affect the quality of the meatballs, including the additional ingredients used and how to cook them (Daniati, 2005).
Pempek
According to Sugito and Hayati (2006), pempek is a product of processed fish meat in the form of a type of protein gel that is homogeneously white, has a chewy and elastic texture. Karneta (2013) states that pempek is made from a mixture of the basic ingredients of mashed fish meat, tapioca flour, water, salt, and spices as a flavor enhancer. This mixture can be served in various forms and then cooked by boiling, steaming, frying, or grilling. The serving of pempek is accompanied by soup or cuko pempek as a complement.
Historically, pempek existed in Palembang, which was around the 16th century when Sultan Mahmud Badaruddin II came to power in the Palembang Darussalam sultanate. The name empek-empek or pempek is believed to have originated from the term apek or pek-pek, which is the name for uncle or old Chinese man (Wikipedia, 2020).
In general, a good pempek has the following characteristics: a. Chewy: the elasticity of the pempek can be influenced by the basic ingredients used in its production, such as whether sago starch is used, as well as the addition of water as an additional thickening agent. b. Pempek color: this is influenced during the pempek-making process, where the pempek stirring should be done evenly until the pempek base dough is smooth and white and not too pale. c. Not easily mushy when boiled: this may be influenced by several factors, namely boiling time and being given a little oil in the cooking water so that the pempek dough is not sticky and damaged. d. Not easily mushy when boiled: this can be influenced by a number of factors, including boiling time and the addition of a small amount of oil to the cooking water to keep the pempek dough from becoming sticky and damaged.
e. The aroma of pempek must be good as the typical of pempek in general.
Sate Lilit
In Bali, sate lilit is typically made for offerings when religious ceremonies or large events are held as a form of respect and appreciation for the Hindu gods. Sate lilit is always produced in large quantities or on a large scale, sometimes involving a hundred men. It turns out that the process of making sate lilit has always been entrusted with men, starting from mixing the dough, slaughtering animals, twisting it to burning it. From the process that men tend to do, sate lilit has a strong philosophical meaning in male life and masculinity.
Previously, sate lilit was only made from pork and fish, but due to a lot of demand and adjusting for consumers who could not eat pork, beef and chicken satay was also made. The satay made from minced pork, fish, chicken, beef, or even turtle, will be mixed with grated coconut, coconut milk, and basa genep or basa gede spices.
According to Darmayasa, et al., (2015) sate lilit is made from mashed pork or tuna fish which is then added with salt, a mixture of spices, and wrapped around a twisted satay stick. The seasonings used to make the satay are shallots, garlic, coriander, turmeric, ginger, galangal, grated coconut, lime leaves, pepper, chilies, and salt.
Sate lilit is a typical Balinese food that is processed using meat ingredients, the meats that are often used are pork, fish, chicken, beef and turtle. The meat mashed and mixed with grated coconut, coconut milk, lime, shallots and pepper. Most of the satay is skewered with sharp skewers, sate lilit however requires a skewer with a wide surface and 10-15 cm long so that the meat will easily stick to it (Milasari, 2017).
Based on this understanding, sate lilit is a satay made from mashed meat or minced meat, (fish, chicken, pork) mixed with spices such as shallots, garlic, coriander, turmeric, ginger, galangal, grated coconut, lime leaves, pepper, chilies, and salt. The mixed meat will be wrapped in bamboo, lemongrass, or sugar cane before being grilled.
According to Kruger (2014), the good quality or quality of typical Balinese satay is as follows: a. Taste Sate lilit has a strong spice flavor accompanied by a distinctive taste of coconut and a sweet taste of brown sugar. a. Aroma There's is a burning aroma of coconut, spices, and the distinct aroma of sate lilit. b. Texture It has a soft texture and can stick or wrap perfectly on the skewers. c. Color It looks burnt on the outside and has the distinctive caramel color of roasted brown sugar.
Materials and Methods
Meatballs, pempek, and satay wrapped in snow mushrooms as a substitute for chicken and fish meat were the subjects of this study. The research was carried out in the Bali Tourism Polytechnic practice's kitchen. The data sources used in this study were primary and secondary data. This study's panelists were 25 trained panelists, including 5 lecturers from the Catering Management Study Program and 20 students in the fifth semester of the 2020/2021 Academic Year.
The data collection technique was carried out by using experimental techniques, documentation, and organoleptic tests. The experiment was carried out with the aim of knowing and controlling the process and stages of making meatballs, pempek, and satay lilit with the basic ingredients of snow mushrooms. The documentation technique is used in the process of making the production stages from the beginning of preparation to producing a product that is in accordance with the standards followed by giving samples to the panelists who are then given an organoleptic assessment so that the taste, aroma, texture and color quality of the meatballs, pempek and satay are obtained. the basic ingredients of snow mushrooms.
According to Sugiyono (2014) experimental research is a research method used to find the effect of certain treatments on others under controllable conditions. The data analysis technique was carried out in a quantitative descriptive manner by providing a score value on the answers given by the panelists using a Likert scale so that the quality of the resulting product could be determined in terms of being very good, good, sufficient, not good or not good (Sugiyono, 2014).
This research was carried out in stages. The first stage was an experiment based on a modified standard recipe, namely the substitution of snow mushrooms for the basic ingredients to produce a product that met the standard, which was then organoleptically assessed by 25 trained panelists. The product processing process is described, starting from preparation to achieving results that are in accordance with good standard of meatball, pempek, and satay lilit and then submitted to the panelists to be assessed for the quality of taste, aroma, texture, and color.
Research Result Experiment
Experiments were carried out by applying standards in the processing of meatballs, pempek, and sate lilit according to the modified recipe by replacing the basic ingredients using snow mushrooms, which can be seen in sequence in tables 1, 2, and 3 as follows: How to make snow mushroom meatballs: a. Prepare the necessary materials and equipment. b. Soak the mushrooms for 3-5 minutes in a bowl, then wash and drain. c. Cut mushrooms into small pieces. Then squeeze the mushrooms to reduce the amount of water on them. d. Put the mushrooms in the blender until they become like grains of sand / fine. e. Put the mushrooms in the bowl and prepare another bowl to be used as a place to stir the meatball dough. f. Weigh all ingredients. Then mix it into the bowl that was prepared earlier. Stir the meatball mixture evenly until it is combined. g. If it has hardened into a meatball dough, boil the meatballs in hot water over high heat. The dough should then be weighed at 15gr and shaped into a round shape. The weighed dough yields 20 pieces of meatball balls. h. When the water is boiling, turn off the heat and add the previously weighed meatball mixture. Allow to stand for 3-5 minutes before returning to a low heat. i. Make some ice water. Allow time for the meatball to expand. If it expands, remove it immediately and drain it. Then pour it into the ice water that has been prepared. Allow to stand for 3-5 minutes before draining again. This is useful to keep the meatballs from sticking together. The results of the snow mushroom meatball experiment are shown in the following image: Data(2020) How to make snow mushroom pempek: a. Prepare the necessary materials and equipment. b. The snow mushrooms that have been cut into pieces are grinded using a blender. c. Mix evenly the mashed sago flour and snow mushrooms. d. Combine with flour mixture, salt, and garlic powder. e. If it has become a pempek dough, prepare hot water to boil the pempek. Weigh into 50-60 grams and formed a lenjer. The dough that has been weighed will produce 5 pieces of pempek lenjer. f. When the water is boiling, boil the formed pempek dough. Then remove it when the pempek is floating, which indicates that the pempek is ripe then chill. g. Prepare the hot oil then fry the pempek until the color turns golden brown. The results of the snow mushroom pempek experiment can be seen in the following picture after boiling and frying: How to make snow mushroom satay: a. Steam the snow mushrooms for about 10 minutes to remove moisture in the mushrooms. b. Cool briefly, then wring it out using gauze. c. Mashed the snow mushrooms until smooth. Set aside. d. Prepare the ingredients for the ground spices then saute briefly until wilted. e. Blend the sauteed spices until smooth. f. Saute ground spices, add bay leaves, lemongrass, salt, and sauteed pepper until fragrant and cooked. Set aside. g. Mix together mashed snow mushrooms, ground spices, chopped lime leaves, grated coconut, coconut milk, brown sugar, fried onions, salt, and pepper in a mixing bowl until well blended. h. Take 30 grams of dough, then wrap it on a twisted skewer. i. Grill for about 10 minutes until cooked. j. Satay lilit is ready to be served.
The experimental results of the snow mushroom sate lilit can be seen in the picture:
Organoleptic Test
The recapitulation of the organoleptic assessment results in table 4 shows the results of the organoleptic assessment to determine the quality of the meatballs, pempek, and satay with snow mushrooms based on the color, taste, texture, and aroma of the 25 panelists. After the final score on each instrument is obtained, the processing of the organoleptic test results data is continued into the process of determining the interpretation results. Before determining the interpretation results, it is necessary to know the highest score (Y) and lowest score (X) for each instrument with the formula (Darmadi, 2011) as follows: The lowest and highest scores for each instrument were calculated using the two formulas. Table 5 shows the calculation of the lowest (X) and highest (Y) scores.: Table 5. The Calculation of The Lowest (X) and Highest (Y) Scores.
From the calculation of the data, the highest score (Y) is 125 points and the lowest score (X) is 25 points for each instrument. The next process in determining the interpretation of the score is determining the interval (distance) and the interpretation of the%. To find out the intervals used in this study, the solution to the formula is used as below: As a result, the interval used to determine the interpretation result is 20%. With these results, it is possible to explain why the score interpretation criteria will be used, which are as follows: (2020) The determination of the results of the interpretation of each variable instrument is carried out using the index formula. The index formula is obtained from the total score divided by the highest score of the assessment multiplied by 100%. Recapitulation of variable index values on the sample instruments of meatballs, pempek, and sate lilit with the basic ingredients of snow mushrooms can be seen in table 7 as follows:
Discussion
The results of the research can be presented based on the organoleptic assessment of the color, taste, texture, and aroma of meatballs, pempek, and satay with snow mushrooms, which were tested on by 25 panelists, as shown below: a. Color In the assessment conducted by the 25 panelists, meatballs and pempek got very good results, but on sate lilit, the results were quite good. The color of the food is heavily influenced by the basic ingredients and the cooking method used; in this case, the method used is grilled in accordance with satay standards in general, but because the basic ingredient used is snow mushrooms, the color is too dark and tends to be less appealing. b. Aroma Meatballs received a lower rating (good) compared to pempek and sate lilit (very good). In the cooking method used, boiled meatballs are used so that the aroma is not overpowering, as opposed to pempek, which is boiled and then fried, as well as grilled satay, which has a stronger aroma. c. Texture Based on the assessment conducted by the panelists, the textures obtained from the meatballs, pempek and sate lilit received very good ratings. The texture of food is heavily influenced by the nature of the main ingredients used; in this case, the texture of the snow mushroom, also known as jelly mushroom, is chewy, making it ideal for the resulting product with a chewy texture. d. Taste Meatballs have a lower (good) rating compared to pempek and sate lilit (very good). In general, meatballs are served with a sauce that gives a savory broth taste so that an assessment with only the meatball results in a lower rating compared to pempek and lilit satay. In the spices satay, which is mixed directly into the satay mixture, the satay is richer in taste even though the basic ingredient, namely snow mushrooms, does not have a striking taste character.
Conclusions
Based on the results of experiments and organoleptic assessments that have been carried out, it can be concluded that the results of the innovation of meatballs, pempek, and sate lilit using snow mushrooms as the basic ingredients in their manufacture get very good ratings. The results of the color assessment on the sate lilit got the lowest result, namely 56% in the Quite Good category so that development and quality improvement are needed, especially in terms of the color of the product produced. The best result of the assessment of the three products produced is in the texture category, which is getting very good results. Assessment of the quality of a food is not only seen in terms of color, aroma, texture, and taste, but also very much determined by the nutritional value of the food so that further research needs to be done by looking at the nutritional content of meatballs, pempek, and sate lilit with snow mushroom as basic ingredients. | 2022-02-06T16:06:10.914Z | 2021-12-11T00:00:00.000 | {
"year": 2021,
"sha1": "31e5f6d2939249cdac6831b2e39052221829d85c",
"oa_license": "CCBYNCSA",
"oa_url": "https://ejournal.upi.edu/index.php/gastur/article/download/35837/pdf",
"oa_status": "HYBRID",
"pdf_src": "Anansi",
"pdf_hash": "6f2f2c14492d8862c9aa7db17398965fe22865a2",
"s2fieldsofstudy": [
"Agricultural and Food Sciences"
],
"extfieldsofstudy": []
} |
197548854 | pes2o/s2orc | v3-fos-license | Identification of secondary fatty alcohols in atmospheric aerosols in temperate forests
Abstract. Fatty alcohols (FAs) are major components of surface
lipids (waxes) and can act as surface-active organic aerosols in the
atmosphere, influencing chemical reactions, particle lifetimes, and the
formation of cloud droplets and ice nuclei. However, studies on the
composition and source of the FAs in atmospheric aerosols are very limited.
In this study, we identified five secondary FAs (SFAs) with C27 and
C29 from aerosol samples collected throughout 1 year at two different deciduous forest sites in Japan. Fatty diols, such as
n-heptacosan-5,10-diol, were identified in atmospheric aerosols for the first
time. Among the identified SFAs, n-nonacosan-10-ol was the most abundant
compound, followed by n-nonacosan-5-10-diol at both of the forest sites.
Concentrations of the SFAs exhibited distinct seasonal variation, with
pronounced peaks during the growing season at each forest site. The SFAs
showed significant correlation with sucrose, which is used as a molecular
tracer of pollen. A significant fraction of the SFAs was attributed to the
submicrometer particles in the growing season. The results indicate that
they originated mostly from plant waxes and could be used as useful tracers
for primary biological aerosol particles.
Introduction
Lipids can be effectively used as molecular markers of terrestrial and marine sources of atmospheric aerosols (e.g., Gagosian et al., 1981). In the terrestrial environment, different organisms contain a variety of long-chain alcohols (fatty alcohols; FAs) because of the diversity of pathways used by the biota for the synthesis of molecules and other compounds. Most fatty alcohols occur in biota as waxes covering parts of plants, which are useful for preventing desiccation, protection from bacterial attack, UV screening, and so on (e.g., Dahl et al., 2005). Because the main function of plant waxes is to reduce water loss through evaporation, waxes originating from this source tend to have longer chain lengths. The most prominent chain lengths are C 26 to C 30 (Tulloch, 1976). Once emitted into the atmosphere, FAs in atmospheric aerosols can act as tracers of primary biological aerosol particles (PBAPs) (Simoneit et al., 2004). Previous studies estimated global emissions of PBAPs mostly based on the abundance of fungal spores (e.g., the sugar mannitol as a biotracer) and bacteria (Carslaw et al., 2010). Among possible compounds of PBAPs, however, our knowledge is very limited on the abundance, emission mechanisms, and atmospheric behavior of FAs originated from plants. Possible emission drivers, such as temperature, light intensity, and local wind speed, likely vary with types and sources of PBAPs.
FAs are known as major components of surface lipids (waxes) with chains varying between C 20 and C 34 carbon atoms (Mudge, 2005). Fatty alcohols are composed of longchain hydrocarbons and hydroxyl (−OH) groups, which can congregate at the aqueous surface and can act as surfaceactive organic aerosols (OAs). These aerosols can cause gases to transfer into the aqueous phase as atmospheric particles and prevent evaporation, which influences chemical reactions, particle lifetimes, fog/cloud droplets, and ice nuclei (IN) (Gill et al., 1983;Cantrell and Robinson, 2006;Knopf et al., 2018). In particular, biogenic particles such as fun-Y. Miyazaki et al.: Identification of secondary fatty alcohols in atmospheric aerosols gal spores and bacteria can serve as a source of IN macromolecules, implying that IN macromolecules are therefore derived from a variety of biological or biogenic particles (Pummer et al., 2015). Long-chain alcohols show appreciable IN activity if they are crystallized into well-defined monolayers. This depends on chain length, the position of the OH group, and any substitutions on the side chains (Popovitz-Biro et al., 1994). Oros and Simoneit (2001a) observed n-nonacosan-10-ol, a secondary FA (SFA), in smoke samples from conifers subjected to controlled burning. In addition, an alkanediol nnonacosan-5-10-diol was identified in ambient aerosol samples collected in the western North Pacific during the ACE-Asia campaign (Simoneit et al., 2004). Some common alkanols were also found within ambient aerosols at urban and remote island sites and are thought to have been emitted to the atmosphere via biomass burning (Oros and Simoneit, 2001a, b). These studies also indicate an input of waxes from forests consisting of both soft and hardwoods. However, studies are very limited on the composition and sources of FAs in atmospheric aerosols obtained in the vicinity of source regions, such as forest environments.
In this study, we used gas chromatography-mass spectrometry (GC-MS) to identify five SFAs found in atmospheric aerosols, both in submicrometer particles and total suspended particulate matter (TSP), sampled throughout 1 year within canopies of two different temperate forests in Japan. We discuss their possible sources based on their seasonal variations in mass concentration found at the two forest sites. All the aerosol samples were collected on quartz fiber filters (25 cm × 20 cm), which were precombusted at 410 • C for 6 h to remove any contaminants. Descriptions of each site are given in the following subsections.
Tomakomai (TMK) experimental forest
The TMK site is situated in a cool temperate zone, with its southern boundary adjacent to Tomakomai city and its industrial port area facing the Pacific Ocean. The mixed cool temperate forest consists of mature and secondary deciduous forest and man-made coniferous forest with various types of forest floor cover (Hiura, 2001(Hiura, , 2005. Tree species include Mongolian oak (Quercus crispula); mono maple (Acer pictum); Korean whitebeam (Aria alnifolia); Japanese lin- den (Tilia japonica); and the planted species Japanese larch (Larix leptolepsis), sakhalin fir (Abies sachalinensis), and sakhalin spruce (Picea glehnii) (Hiura, 2005). The height of the main trees ranges between 15 and 20 m, and the leaf area index (LAI) is within a range of 3.3-4.9 m 2 m −2 (Hiura, 2001). The soil consists of shallow and unweathered volcanogenic regosols. The predominant local wind direction in autumn and winter is from the north, originating in the forested areas. In contrast, air transported from the south (the coastal urban area) is dominant in summer.
At the TMK site, both TSP and submicrometer aerosol samples were collected continuously in parallel using two high-volume air samplers (HVASs; model 120SL, Kimoto Electric, Osaka, Japan) at an altitude of ∼ 18 m above the forest floor of the research site (Müller et al., 2017). A cascade impactor (CI; model TE-234, Tisch Environmental, Cleves, OH, USA) was attached to one of the two HVASs to collect size-segregated particles with five stages at a flow rate of 1130 L min −1 . We used analytical results obtained from the bottom stage of the impactor, which collected particles with aerodynamic diameter smaller than 0.95 µm. Aerosol particles within this size range are referred to as submicrometer particles. The sampling was taken continuously with a duration of approximately 1 week for each sample, from January to December in 2015. Collected filter samples were individually stored in glass jars with a Teflon-lined screw cap at −20 • C to limit the chance for chemical reactions on the filter and the loss of volatile compounds. In total, 23 sample sets were obtained at this site for both TSPs and submicrometer particles. The FHK research site is located at the base of Mt. Fuji in Fujiyoshida city, Yamanashi, Japan ( Fig. 1) (e.g., Mochizuki et al., 2015;Urakawa et al., 2015). The urban area of Fujiyoshida city is located about 8 km northeast of the site. The vegetation type is Larix kaempferi plantation, which was planted uniformly 55 years ago over 150 ha. Tree height is approximately 20-25 m. LAI was estimated to be 2.8 m 2 m −2 based on the leaf mass abundance in 2006 . After tree thinning at the site in 2014, the leaf mass abundance in 2015 and 2016 decreased to 80 % and 96 % of that in 2006, respectively. The site is surrounded by coniferous and broad-leaved mixed forests. A meteorological tower of 32 m height and a steel scaffold of 20 m height for tree surveys are situated at the center of the forest site. The soil type is coarse-grained volcanic ash. The forest floor is predominantly covered with Dryopteris crassirhizoma. The FHK site is located on a slope in the northern foothills of Mt. Fuji, and the dominant wind direction is characterized by two modes throughout a year: from the north (valley wind) in daytime and from the south (mountain wind) in nighttime .
TSP sampling was conducted at a height of 16 m, just beneath the larch canopy, using a HVAS (model 120SL, Kimoto Electric, Osaka, Japan) mounted on the 20 m steel scaffold. The aerosol samples were collected from January 2015 to December 2016 with a sampling duration for each aerosol sample of ∼ 2 weeks. The flow rate of the TSP sampling was approximately ∼ 600 L min −1 . In total, 15 and 17 samples were analyzed, obtained at the FHK site for the year 2015 and 2016, respectively. The collected filter samples were kept frozen at −20 • C until analysis. It is noted that the difference in the sampling duration at the two sampling sites might not significantly affect our conclusions, because we discuss the seasonal changes of the aerosol concentrations whose timescale is much longer than that of the aerosol sampling (1-2 weeks).
Analytical procedure with gas chromatography-mass spectrometry (GC-MS)
For the aerosol samples, a filter cut of 3.8 cm 2 was extracted with dichloromethane/methanol (2 : 1, v/v) and the −OH functional groups in the extracted samples were reacted with a mixture of 50 µL of N, O-bis-(trimethylsilyl) trifluoroacetamide (BSTFA), 1 % trimethylsilylchloride, and 10 µL pyridine to form trimethylsilyl (TMS) ethers. The TMS derivatives were then analyzed for the presence of the compounds listed above using a capillary gas chromatograph (GC7890, Agilent) equipped with a fused silica capillary column (DB-5MS, Agilent Technologies, Santa Clara, CA, USA) and coupled to a mass spectrometer (Miyazaki et al., 2012;Müller et al., 2017). From the TMS derivatives, structures of each compound of FAs were elucidated with low-resolution GC-MS (MSD5975C, Agilent), as well as with a high-resolution gas chromatograph-time-of-flight mass spectrometer (GC-TOF-MS; JMS-T100GCV, JEOL) using electron ionization (EI). Identification of each compound was made based on GC retention time, literature mass spectra, and the interpretation of mass spectrometric fragmentation patterns. In addition to FAs, sucrose and levoglucosan were also measured by the same method as described above.
3 Identification of secondary fatty alcohols in the aerosol samples Figure 2 presents an example of a GC-MS total ion chromatogram (TIC) of the TMS extract of the TSP sample collected at the FHK site. In previous studies, sugar compounds have been identified with the same method of derivatization, such as arabitol, fructose, glucose, and sucrose (e.g., Miyazaki et al. 2012;Müller et al., 2017). In the present study, P1-P5 refer to peaks which were identified to be TMS derivatives of SFAs with carbon numbers of 27 and 29, the peaks of which appear in the latter part of the TIC. Figure 3 shows an EI mass spectra of P1-P5 obtained for the TMS derivatives with their molecular structures. On the basis of the detailed interpretation of the EI mass spectral data together with their comparison to literature data as well as the exact mass measurements by GC-TOF-MS, the derivatives were assigned to (a) n-heptacosan-10-ol-TMS (P1; C 27 ), (b) nheptacosan-5,10-diol-diTMS (P2; C 27 ), (c) n-nonacosan-10ol-TMS (P3; C 29 ), (d) n-nonacosan-10-13-diol-diTMS (P4; C 29 ), and (e) n-nonacosan-5-10-diol-diTMS (P5; C 29 ). The spectra of n-heptacosan-10-ol-TMS is characterized by the presence of ion peaks at m/z 73, 229, and 341 (Fig. 3a). Similarly, the mass spectra of n-nonacosan-10-ol-TMS showed that ion peaks at m/z 73, 229, and 369 are significant (Fig. 3c). The ion at m/z 73, corresponding to [Si(CH 3 ) + 3 ], is characteristic for TMS derivatives containing one or more derivatized OH groups. Indeed, the exact mass measurements by GC-TOF-MS confirmed the estimated composition of each peak. For example, the composition of P3 with M-15 was identified as C 31 H 65 OSi, where its theoretical and experimental m/z are 481.4805 and 481.4826, respectively, with an error of 4 ppm.
A primary plant origin of these SFAs is most likely, because the SFAs are known to be present in plant leaves; nnonacosan-10-ol has been identified as a major compound in epicuticular waxes found in gymnosperm species (Tulloch, 1976;Schulten et al., 1986), such as Sequoiadendron giganteum . Moreover, secondary fatty diols have been reported in other plant species, such as Pisum sativum (Vioque and Kolattukudy, 1997). In the atmosphere, n-nonacosan-10-ol was observed in smoke aerosol samples from conifers subjected to controlled burning (Oros and Simoneit, 2001a). Moreover, some of the common alkanols found in urban and remote island sites in spring are thought to have been emitted to the atmosphere by biomass burning (Oros and Simoneit, 2001a, b;Simoneit et al., 2004), while they also indicate an input of waxes from forests consisting of both soft and hardwoods.
In our study, a primary origin of biomass burning is unlikely because the SFAs showed no significant correlation with tracer compounds of biomass burning, such as levoglucosan with R 2 of < 0.02 (data not shown) at both sites. Rather, the seasonal trends of the SFA concentrations are similar to that of sucrose, which is an important primary saccharide of pollen grains (Pacini, 2000) and has been used as a tracer for pollen (e.g., Miyazaki et al., 2012). Specifically, the concentration of n-nonacosan-10-ol showed significant positive correlations with that of sucrose both at the TMK (R 2 = 0.70) and FHK sites (R 2 = 0.61). The average concentrations of n-nonacosan-10-ol were significantly larger than those of sucrose in the growing season at the two sites. These results suggest that the increase in the mass concentrations of n-nonacosan-10-ol and the other SFAs is due to large emissions of plant waxes or components thereof (i.e., fatty alcohols) from the forest leaves into the atmosphere. Our analysis also suggests that the SFAs identified here can act as tracers for primary biological aerosol particles (PBAPs) which are surface-active.
It is noted that the peak concentration of SFAs at FHK appeared in April (Fig. 4b), approximately 1 month earlier than at TMK (Fig. 4a). The difference in the appearance of the peaks is attributable to the difference in the growing season at the two sites: the growth of forest vegetation generally starts about 1 month earlier at FHK (35 • N) than at TMK (42 • N) (Nakaji et al., 2011;Takahashi et al., 2015), which is also evident from the seasonal changes in the net ecosystem production (Saigusa et al., 2008). The primary emissions of these organic compounds clearly depend on the activity of the forest vegetation and, more specifically, plant leaf phenology at these forest sites. Considering that the aerosol sampling duration of about 1-2 weeks is much longer than the timescale 0.33 ± 0.10 21.5 ± 12.9 3.51 ± 0.67 0.69 ± 0.29 0.14 ± 0.10 23.2 ± 14.4 0.12 ± 0.08 0.16 ± 0.03 * Values are averages (ng m −3 ) with ±1 standard deviation. LOD is the limit of detection. Table 2. The average concentrations of n-nonacosan-10-ol, n-nonacosan-5,10-diol, and sucrose in the submicrometer particles and the fractions of their mass to the TSP mass at the TMK site in 2015.
n-Nonacosan-10-ol a Sub-µm / TSP ratio of n-Nonacosan-5,10-diol a Sub-µm / TSP ratio of Sucrose a Sub-µm / TSP ratio n-nonacosan-10-ol b n-nonacosan-5,10-diol b of sucrose b of transport of particles in the forest regions, the aerosol particles collected in this study were likely influenced by biogenic emissions from the whole area of each forest region. In contrast to the growing season, seasons in which the lowest concentrations were observed are not clear from Fig. 4. Large standard deviations of the average values in winter, summer, and autumn at the two sites (Table 1) mean that the difference in the average concentrations is mostly insignificant for these seasons. This is attributable to various factors (e.g., the difference in the emission strength and photochemical activity in the atmosphere) which need to be further investigated in a future study. Figure 5 presents seasonal averages of the mass concentrations and particle size fractions of n-nonacosan-10-ol observed at the TMK site. The average values and the fractions of their mass to the TSP mass at the TMK site in 2015 are also summarized in Table 2. In spring, the mass of n-nonacosan-10-ol in the submicrometer size ranges accounts for ∼ 68 % of the TSP mass, the fraction of which is much larger than that in summer (∼ 10 %) and autumn (∼ 28 %). Similarly, the average mass fraction of submicrometer n-nonacosan-5,10-diol in TSP is 65 % in spring, whereas the mass fraction is 1 % and 15 % in summer and autumn, respectively. Our result is similar to that of Kavouras and Stephanou (2002), who also measured particle size distributions of n-alkanols (< C 26 ) at Eucalyptus and conifer forests in summer to show clear bimodal size distributions with ∼ 40 % of the mass residing in particles with diameters smaller than 0.96 µm. In general, the majority of the mass of PBAPs resides in supermicrometer particles. Indeed, the (Table 2). In contrast to the typical PBAPs tracers, our result suggests that a significant fraction of the SFAs can be attributed to the submicrometer particles in the growing season.
Conclusions
Secondary fatty alcohols with carbon numbers of 27 and 29 were identified in atmospheric aerosol samples collected throughout the year at two different deciduous forest sites. As far as we know, this study is the first to identify fatty diols, such as n-heptacosan-5,10-diol, in atmospheric aerosol samples and to show the seasonal variations in their concentrations in the forest atmosphere.
Among the identified SFAs, n-nonacosan-10-ol was the most abundant molecular compound, followed by nnonacosan-5,10-diol. The SFAs showed pronounced peaks in the growing season at both forest sites. In spring, the concentrations of n-nonacosan-5,10-diol were much larger than those of sucrose. Moreover, the concentrations of the identified SFAs showed significant positive correlations with those of sucrose at the two sites. These results indicate that the SFAs originated mostly from plant wax. The difference in the peak appearance of the SFA concentrations at the two sites indicates that the primary emissions of these organic compounds clearly reflect the activity of the forest vegetation such as plant leaf phenology at the forest sites.
In the growing season, the fraction of n-nonacosan-10-ol in the submicrometer size range accounted for ∼ 70 % of the TSP mass. Similarly, the average mass of submicrometer nnonacosan-5,10-diol in TSP accounted for 65 % of the mass of TSP in spring. These fractions were much larger than those in the other seasons, suggesting that a significant fraction of the SFAs was attributed to the submicrometer particles in the growing season. This study demonstrates that the SFAs identified, particularly n-nonacosan-10-ol and n-nonacosan-5,10-diol, can be used as effective tracers for PBAPs with surface-active organic matter.
Data availability. Data from measurements are available upon request from the corresponding author. | 2019-04-27T13:13:09.443Z | 2019-02-18T00:00:00.000 | {
"year": 2019,
"sha1": "1404361b58dd8be15f47dcbd28524c60f3703783",
"oa_license": "CCBY",
"oa_url": "https://www.biogeosciences.net/16/2181/2019/bg-16-2181-2019.pdf",
"oa_status": "GOLD",
"pdf_src": "Adhoc",
"pdf_hash": "d31e6cae64762af021b63959e538f07c22bdc900",
"s2fieldsofstudy": [
"Environmental Science"
],
"extfieldsofstudy": [
"Chemistry"
]
} |
257340240 | pes2o/s2orc | v3-fos-license | Intraoperative Hypotension Is Associated with Postoperative Nausea and Vomiting in the PACU: A Retrospective Database Analysis
Multiple risk factors for postoperative nausea and vomiting (PONV)—a very distressing and outcome-related complication—have been identified, including female sex, absence of a history of smoking, history of PONV, and postoperative opioid use. Evidence of association of intraoperative hypotension with PONV is contradictory. A retrospective analysis of the perioperative documentation of 38,577 surgeries was conducted. The associations between different characterizations of intraoperative hypotension and PONV in the postoperative care unit (PACU) were investigated. First, the relationship between different characterizations of intraoperative hypotension with regard to PONV in the PACU was investigated. Secondly, the performance of the optimal characterization was assessed in an independent dataset derived via random split. The vast majority of characterizations showed an association of hypotension with the incidence of PONV in the PACU. In a multivariable regression, time with a MAP under 50 mmHg showed the strongest association with PONV in terms of the cross-validated Brier score. The adjusted odds for PONV in the PACU were estimated to be 1.34 times higher (95% CI: 1.33–1.35) when a MAP was under 50 mmHg for at least 1.8 min than when a MAP remained above 50 mmHg. The finding indicates that intraoperative hypotension may yet be another risk factor for PONV and therefore emphasizes the importance of intraoperative blood pressure management not only in patients at risk for cardiovascular complications but also in young and healthy patients at risk of PONV.
Introduction
Postoperative nausea and vomiting (PONV) are very distressing for patients [1]. PONV has an incidence of 30% [2] in the general population and 80% in high risk populations [3]. Known risk factors include female sex, use of volatile anesthesia, a history of previous PONV or motion sickness, non-smoker, lower age, postoperative opioid use, and particular types of surgery [3,4].
Preventing PONV may be facilitated by proper risk stratification using scoring systems such as the Apfel score, which includes the factors of female sex, non-smoker, history of PONV or motion sickness, and postoperative opioid use [3]. The score indicates whether PONV prophylaxis should be administered and anesthesia with a lower risk of PONV be performed, i.e., total intravenous anesthesia [5][6][7].
of 13
Characterizing intraoperative hypotension is difficult with many different approaches found in the recent literature: [12] Intraoperative hypotension can be defined either by absolute or relative time with a MAP under a certain value or by change from baseline in percent [8,10,11,13]. Another option used in the literature is change from patients' preoperative blood pressure values [12]. Choosing different blood pressure thresholds potentially changes resulting outcomes and the association of one characterization of hypotension with one detrimental outcome, such as MINS, is not necessarily generalizable for other outcomes, such as PONV.
Although often seen in clinical practice, only a few small trials investigated the influence of intraoperative hypotension on PONV: Pusch et al. examined the effect of systolic arterial pressure variation on PONV in 300 patients and showed that a decrease in systolic blood pressure was associated with an increased incidence of PONV [14]. Heidari et al. showed that preoperative volume loading had an effect on both the blood pressure variability and PONV [15]. Recently, an association between hypotension and PONV was reported in thyroidectomies in 247 patients [16]. Hypoperfusion of the brainstem, as well as intestinal hypoperfusion, was hypothesized as possible mechanisms for PONV promotion in hypotensive patients [14,15,17].
We present the first broad-scale, single-center, retrospective cohort study of a large perioperative database to explore the association of hypotension and PONV. In the present study, we investigated the associations of different features of perioperative hypotension with PONV and tested the best characterization subsequently in an independent dataset.
Study Design
We conducted a retrospective cohort study using the Medical University of Vienna's perioperative database: Included were all adults (≥18 years) undergoing non-cardiothoracic surgery between 1 September 2013 to 1 September 2020, with a complete record of intraoperative blood pressure, records of PONV in the PACU, and complete data about risk factors of PONV in their perioperative documentation. Risk factors for PONV are collected routinely in the preoperative assessment. Records of PONV in the PACU are documented in the electronic patient data management system. Only patients receiving general anesthesia were included.
Ethics
The University of Vienna's ethical committee (Borschkegasse 8b/6, 1090 Vienna, President: Prof. Zezula) approved this study and waived the need for informed consent on 21.10.2021 (EKNr: 2062/2020). All research was performed in accordance with relevant guidelines/regulations, especially in accordance with the Declaration of Helsinki.
Data Sources
Data were extracted from the perioperative database and the hospital's research database. The perioperative research database is constantly synchronized with the Philips IntelliSpace Critical Care and Anesthesia (Philips, Amsterdam, The Netherlands) patient data management system, recording all patient data perioperatively including the ICU. The database contains data for both vital parameters and manually entered observations/actions by all health care professionals. Since 2017, vital parameters have been stored with a resolution of 15 sec; prior to this date the resolution was 2 min.
The second database used is the Medical University of Vienna's research database: Most of its data are transferred from the hospital information system, an SAP-based (SAP, Walldorf, Germany) database consisting of all patient data except for perioperative and intensive care data. The database includes patient reports, diagnostic codes (International Classification of Diseases, 10th edition), laboratory results and billing information.
Data Processing
Non-invasive blood pressure data were available every time a measurement was initiated (typically every three to five minutes), and invasive blood pressure values every 15 sec or 2 min (as explained above). PONV was documented at the bedside by the nursing staff in either the post anesthesia care unit (PACU) or the intensive care unit (ICU) using four categorical variables: "None", "Nausea", "Need to vomit", "Vomiting". The first 48 h of ICU and total PACU stay were combined. For analysis purposes, this variable was converted into a dichotomous variable (PONV or No PONV) by aggregating the categories "Nausea", "Need to vomit" and "Vomiting". In case of multiple entries, the most severe entity was used: A documentation of "No PONV" was ignored if the entry also contained a documentation of "PONV" as described above. All PONV entries in the first 48 h after end of surgery were included to eliminate other factors such as intensive care therapy.
Artifact removal of blood pressure data was conducted using the following artifact classifications: [11] (1) Systolic pressure greater than or equal to 300 mmHg, (2) Systolic pressure lower than or equal to 20 mmHg, (3) Diastolic pressure lower than or equal to 5 mmHg, (4) Diastolic pressure greater than or equal to 225 mmHg, (5) Systolic pressure lower than or equal to diastolic pressure + 5 mmHg, (6) Values lying outside three standard deviations of each patient's mean blood pressure. Afterwards, blood pressure values were interpolated linearly, to 15 s intervals where necessary. To assess patients' comorbidities, the weighted Charlson comorbidity index [18] was calculated using R [19] (R Foundation for Statistical Computing, Vienna, Austria, comorbidity package 0.5.3).
The primary outcome was the occurrence of PONV in the PACU as documented by the staff. The exposure to IOH was a priori defined by thresholds following research on other postoperative outcomes [8,10,20]. The following MAP characterizations were calculated: (1) absolute measures of IOH: (a) Lowest MAP for 1, 3, 5 and 10 sustained minutes [11], (b) Lowest MAP for 1, 3, 5, and 10 cumulative minutes per patient [11], (c) Absolute time with a MAP under 50, 55, 60, 65, 70, 75, or 80 mmHg [11], and (2) Relative time with a MAP under 50, 55, 60, 65, 70, 75, or 80 mmHg compared to duration of anesthesia. Particular MAP value for a certain duration of a hypotensive episode (1, 3, 5, or 10 min) was defined as the upper "low MAP" limit of the lowest MAP episode for the respective number of minutes.
Statistical Analysis
The aims of the study were twofold: (1) To determine which, if any, characterization of hypotension has the strongest association with the occurrence of PONV, and (2) To estimate the association between hypotension (in terms of the characterization that could be identified in the first step) and PONV. The data on the surgical procedures were divided equally into a "shaping dataset" (for the first aim) and an "estimation dataset" (for the second aim) by randomly allocating each patient to either the shaping or the estimation dataset and including all their surgical procedures in the respective dataset. For the first aim, 23 multivariable, generalized, estimating equation logistic regression models with patient as clustering variable were fitted on the "shaping dataset" using an exchangeable correlation structure. Each of the models included one of the characterizations of hypotension and the following adjustment variables: Charlson comorbidity score >0 (yes/no), type of surgery carrying high risk for PONV according to current guidelines [4] (yes/no), smoking status (yes/no), sex (female/male), use of volatile anesthetics (yes/no), use of ondansetron as PONV prophylaxis (yes/no), use of dexamethasone as prophylaxis (yes/no), ASA (four categories), duration of surgery, age and the interaction of age and sex as independent variables and PONV in the PACU as dependent binary variable. Postoperative opioid use was not used as confounder due to its nature as post-exposure covariate, as our model only uses demographic and intraoperative variables [21]. The interaction of age and sex was used after initial analysis of the cohort showed a non-equal distribution of age between male and female patients. Characterizations of hypotension, age, and duration of surgery were represented by natural cubic spline bases using three degrees of freedom. Since the characterizations of hypotension "absolute time with a MAP under 50, 55, 60, 65, 70, 75, or 80 mmHg" and "relative time with a MAP under 50, 55, 60, 65, 70, 75, or 80 mmHg" normalized to duration of anesthesia led to many observations equal to zero, an additional, binary variable indicating whether the value of the characterization of hypotension was equal to zero was included. Characterizations of hypotension were winsorized at the 99th percentile. For each of these multivariable models, the performance was assessed by calculating the Brier score, the discrimination slope, the c-statistic, and the calibration slope, using 10-fold cross-validation with 40 repetitions. In the cross-validation, multiple surgical procedures on one patient were allocated to the same fold. The "best-fitting" characterization of hypotension was determined as the one included in the model with the smallest cross-validated Brier score. In addition to the multivariable logistic regression models, the univariate association between hypotension (in terms of the 23 characterizations represented by natural cubic spline bases) and PONV was investigated.
For the second step, the association between hypotension and PONV in the PACU was estimated by fitting a generalized, estimating equation logistic regression model on the "estimation dataset" with the best-fitting characterization of hypotension identified in step 1 and the adjustment variables (parametrized as in step 1) as independent variable and PONV as dependent variable. The independence of the shaping and the estimation dataset ensures that the estimation of the association between the best-fitting characterization and PONV in the PACU is not affected by multiple comparison bias. Finally, the contribution of the best-fitting characterization was compared against the contribution of the adjustment variables included in the multivariable model by calculating the increase in cross-validated Brier score when excluding one of the variables from the model.
Results
Application of the inclusion criteria resulted in 38,577 surgeries on 28,262 patients being included in the analysis before splitting the dataset in two datasets of 19,288 and 19,289 procedures. Overall, 54.5% of surgical procedures were performed on females. The mean age was 53 years. Most procedures were in the field of urology/gynecology/general surgery (42%), followed by orthopedic/trauma surgery (25%) and maxillofacial/ENT surgery (24%). Patients experienced PONV in the PACU in 12.4% (4787) of all included procedures. As expected, Apfel scores differed significantly between the PONV and no PONV group (1.9 vs. 2.4). Primarily Ondansetron and Dexamethasone were used for PONV prophylaxis. Patients at risk of PONV were well detected intraoperatively as those received significantly more Ondansetron (1.8 vs. 1.4 mg, p < 0.001) and Dexamethasone (1.8 vs. 2.5 mg, p < 0.001) than those without PONV in the PACU. In the complete cohort, 2154 (5.6%) patients were admitted to the ICU; the rest was treated in the PACU. Mean length of stay (LOS) in the PACU was 2.6 h (SD: 3.0), while in the ICU mean LOS was 4.5 days (SD: 11.8 days). The occurrence of PONV was documented on event, and the absence was retrospectively documented in general every 2 h. Patients in PACU were in the nursing staffs' sight all the time and were thus continuously monitored for PONV episodes. There was a significant difference in total fluid intake (crystalloid and colloid) with patients suffering from PONV receiving 448 mL more (t-test: p < 0.001). Additional demographics are described in Table 1. In the shaping dataset, all characterizations of hypotension were significantly associated with PONV in the PACU in univariate analysis ( Figure S1). However, in the multivariable logistic regression model, no association could be detected for some characterizations such as minutes with a MAP under 75 mmHg (see Figure 1). When comparing the different characterizations of IOH and their association with PONV, generally a clear rise in risk of PONV could be seen for a MAP < 60 mmHg (Figure 1). Details of all analysis can be found in the Supplementary Materials ( Figure S1).
Systematic comparison of characterizations in terms of the cross-validated Brier score achieved by the multivariable model showed that the characterization "time with a MAP < 50 mmHg" had the strongest association with PONV, followed by "lowest MAP for 3 cumulative minutes" and "lowest MAP for 1 cumulative minute" (Figure 2). associated with PONV in the PACU in univariate analysis ( Figure S1). However, in the multivariable logistic regression model, no association could be detected for some characterizations such as minutes with a MAP under 75 mmHg (see Figure 1). When comparing the different characterizations of IOH and their association with PONV, generally a clear rise in risk of PONV could be seen for a MAP < 60 mmHg (Figure 1). Details of all analysis can be found in the Supplementary Materials ( Figure S1). Systematic comparison of characterizations in terms of the cross-validated Brier score achieved by the multivariable model showed that the characterization "time with a MAP < 50 mmHg" had the strongest association with PONV, followed by "lowest MAP for 3 cumulative minutes" and "lowest MAP for 1 cumulative minute" (Figure 2). Using the best-fitting characterization ("time with a MAP < 50 mmHg"), the adjusted odds for PONV in the PACU in the independent estimating dataset were estimated to be 1.34 times (95%CI: 1.33-1.35) higher where the MAP was under 50 mmHg for at least 1.8 min than if the MAP was consistently above 50 mmHg (Supplement S1). It is notable that due to only 9.8% of patients experiencing a MAP below 50 mmHg, the confidence interval is relatively broad for non-zero values of the characterization.
For "time with a MAP <60 mmgHg", which reflects a more commonly used characterization in literature, in the shaping dataset the adjusted odds for PONV in the PACU were estimated to be 1.17 times higher (95% CI: 1.16-1.17) when a MAP was under 60 mmHg for at least 6 min than when a MAP remained above 60 mmHg. For comparison, in the shaping dataset, the adjusted odds for PONV in the PACU were estimated to be 1.26 times higher (95% CI: 1.26-1.27) when MAP was under 50 mmHg for at least 1.8 min than when a MAP remained above 50 mmHg. A small difference between the two randomly split datasets was found.
Analysis of the influence of the different adjusting variables showed that in the multivariable model the duration of surgery made the most significant contribution in terms of change in cross-validated Brier score (0.005), followed by sex and age (Figure 3). Using the best-fitting characterization ("time with a MAP < 50 mmHg"), the adjusted odds for PONV in the PACU in the independent estimating dataset were estimated to be 1.34 times (95%CI: 1.33-1.35) higher where the MAP was under 50 mmHg for at least 1.8 min than if the MAP was consistently above 50 mmHg (Supplement S1). It is notable that due to only 9.8% of patients experiencing a MAP below 50 mmHg, the confidence interval is relatively broad for non-zero values of the characterization.
For "time with a MAP <60 mmgHg", which reflects a more commonly used characterization in literature, in the shaping dataset the adjusted odds for PONV in the PACU were estimated to be 1.17 times higher (95% CI: 1.16-1.17) when a MAP was under 60 mmHg for at least 6 min than when a MAP remained above 60 mmHg. For comparison, in the shaping dataset, the adjusted odds for PONV in the PACU were estimated to be 1.26 times higher (95% CI: 1.26-1.27) when MAP was under 50 mmHg for at least 1.8 min than when a MAP remained above 50 mmHg. A small difference between the two randomly split datasets was found.
Analysis of the influence of the different adjusting variables showed that in the multivariable model the duration of surgery made the most significant contribution in terms of change in cross-validated Brier score (0.005), followed by sex and age (Figure 3). repetitions) between the model with the respective variable omitted and the full model (described in Figure 3). When evaluating the contribution of sex or age, the interaction of sex and age was also excluded from the model. The larger the difference in Brier score, the higher the contribution of the variable.
Discussion
The association between IOH and PONV has been reported in a small number of earlier trials [14,16,17], but it remains controversial [25]. As PONV is very distressing and has a high incidence in the general surgical population, further reduction and prophylactic strategies are of particular importance. The present analysis showed a distinct association between intraoperative hypotension and PONV in the PACU, with an adjusted odds ratio of 1.34 (95%CI: 1.33-1.35) for a MAP under 50 mmHg for at least 1.8 min versus a MAP consistently above 50 mmHg. As demonstrated in the present study, Figure 3. Brier score of different confounders. Relative contribution of the variables in the multivariable model, quantified by the difference in 10-fold cross-validated Brier score (with 40 repetitions) between the model with the respective variable omitted and the full model (described in Figure 3). When evaluating the contribution of sex or age, the interaction of sex and age was also excluded from the model. The larger the difference in Brier score, the higher the contribution of the variable.
Discussion
The association between IOH and PONV has been reported in a small number of earlier trials [14,16,17], but it remains controversial [25]. As PONV is very distressing and has a high incidence in the general surgical population, further reduction and prophylactic strategies are of particular importance. The present analysis showed a distinct association between intraoperative hypotension and PONV in the PACU, with an adjusted odds ratio of 1.34 (95%CI: 1.33-1.35) for a MAP under 50 mmHg for at least 1.8 min versus a MAP consistently above 50 mmHg. As demonstrated in the present study, IOH appears to be a relevant risk factor in addition to well-known factors including female sex, history of PONV, risk surgery, use of volatile anesthesia, non-smoking, duration of surgery, and age [4]. The choice of "time with a MAP under 50 mmHg" as the best fitting characterization of IOH is based on a comparison of various characterizations in the shaping dataset.
The reasons for the association between IOH and PONV may be hypothesized as multifactorial: To date, reduced gut hypoperfusion [26] has been reported as a probable mechanism, as well as orthostatic dysfunction due to hypotension [17]. As PONV is also caused by decreased gastric emptying, which itself is influenced by intestinal perfusion, this could be a contributing factor, although the short duration of IOH necessary to increase PONV contradicts this theory [27]. A known prevention strategy for PONV is supplementation with intravenous fluids, possibly preventing or correcting IOH to some degree [28]. Whether the use of vasopressors to counteract IOH would lead to reduced incidence of PONV remains unclear and needs to be evaluated subsequently.
Undoubtedly, an important risk factor for PONV is postoperative opioid use. However postoperative opioid use was not included in the model, since it comprised intraoperative variables only; thus, postoperative opioid use was invariably a "post-exposure" variable [21]. In contrast, duration of surgery was included in the model, not at least since longer surgeries often require more extensive postoperative pain therapy, which may be one of the reasons for duration of surgery being a promoting factor for PONV. As only a small proportion of patients (5.6%) were admitted to the ICU, this factor was not used as covariable in the model.
Another possible factor would be the route of application of opioids: In the described cohort, standard procedure in the PACU includes bolus application of opioids. In the ICU, opioids are given either via bolus or infusion. PCA usage is mostly used in the wards. PONV prophylaxis in general plays an important role in routine anesthetic care: Typically, guided by scores such as Apfel or Koivuranta, patients at risk are identified and receive a combination of up to four different prophylaxis agents, including 5HT3 receptor antagonists, dexamethasone, antihistamines, dopamine antagonists, and NK-1 receptor antagonists [3,4,27]. Up to now prevention of intraoperative hypotension as a prophylactic precaution was not part of a routine PONV prevention. Further research must prove whether pre-emptive therapy for IOH can serve as additional effective PONV prophylaxis.
The results of the present study are in accordance with a series of recent publications examining the influence of IOH on a variety of postoperative outcomes. First, the association between IOH and myocardial injury was demonstrated, followed by reports of IOH being associated with AKI [8,10,11,20,29]. Outcomes such as delirium, stroke, and overall organ injury have also been shown to be associated with various characterizations of IOH [13,30,31]. Consequently, examining other postoperative outcomes and their association with IOH is of the utmost interest.
Looking at the recent literature clearly indicates a problem with the heterogeneity of definitions of IOH [13]. A wide variety of characterizations is currently being used, ranging from time with a MAP below a certain value to a relative decrease in blood pressure from a patient-derived baseline value [12]. Since it is unlikely that one characterization of IOH is applicable to all postoperative outcomes, our first step was to search for the optimal characterization of IOH to predict PONV. To achieve this, the patient cohort was randomly split in half, with the first set used only to calculate the optimal characterization. The second dataset was used to confirm the association between PONV and this characterization.
Due to incomplete preoperative vital signs, it was not possible to use characterizations applying a decrease from preoperative baseline. However, the current literature suggests that relative changes of blood pressure may be comparably or even less relevant than absolute MAP limits regarding patient outcome [11,32]. As Figure 1 shows, most characterizations analyzed in the present study showed an increased risk of PONV below 60 mmHg, independently of being based on time below a certain MAP or sustained MAP for a certain time.
When comparing these data to the current literature, a cut-off between 60 mmHg and 65 mmHg is found in most of the studies, when absolute characterizations of IOH were used as in the present study [8,11,13]. For sensitivity analysis purposes, the odds ratio of "time with a MAP <60 mmHg" was calculated in the shaping dataset and showed very similar results.
Nevertheless, looking at associations between perioperative hypotension and various patient outcomes, the application and testing of different forms of these characterizations as a first step in comparison to using a single predefined IOH characterization derived from another outcome may be an obvious approach, since a "one-size-fits-all outcome" characterization of perioperative hypotension is unlikely [12].
Another topic that is also still an open research question is the prevention and treatment of IOH. While a variety of treatments for hypotension is well established, and primarily consists of the use of various vasopressors and fluid administration [33], the actual beneficial effect of treating IOH on the reported outcomes remains unclear. As blood pressure is only one part of a set of complex hemodynamic changes during anesthesia, any "optimal" treatment may well be more complex and "personalized" than a general approach [34].
Since the focus of the present analysis was the association of the "net IOH burden" with PONV in the PACU, the inclusion of the numerous possible treatments of IOH (ranging, e.g., from surgical interventions to blood/fluid administration to medications, which may or may not be initiated by the occurrence an IOH episode) would not have elucidated the primary question of the analysis, which is why the authors decided not to include the current multitude of potential IOH therapies. Further research will be necessary to identify the "personalized" treatment option for IOH episodes in individual patients, not at least in order to reduce risk of PONV. Irrespective of the possible treatments, when IOH has already occurred, prediction and prevention may be the best approach-and while the effect on actual patient outcome remains still to be determined, in a first step, some trials have shown the feasibility of predicting and thus preventing IOH [35][36][37]. This approach was mostly successful, but there were also negative results, where blood pressure did not differ between patients with predictive monitoring vs. routine monitoring [33,38,39]. The reasons for this discrepancy vary, ranging from ignoring alarms to improper therapeutic measures. Larger randomized controlled trials focused on outcome and not only on blood pressure are missing to date. Subsequent prospective trials need to be conducted to show potential, personalized IOH treatment options that improve not only blood pressure but also patient outcomes, and to move on from associations to outcome-improving preventive measures.
One possible limitation of this trial is the low overall incidence of PONV-only 12.4% compared to values of up to 30% [2] as reported in the current literature. This may be due to two main causes. Firstly, compared to other studies that also include the first 24-48 h in wards, in this study only information from the PACU and ICU was available. Deducting from the significantly higher doses of intraoperative ondansetron and dexamethasone in the PONV group, a very good risk stratification took place in the described cohort. When looking at PONV in the PACU as this study did, the PONV incidence is quite comparable to the recent literature reporting PONV incidences of 10-20% in cohorts receiving PONV prophylaxis [40][41][42]. As there is no indication that IOH causes only early or delayed PONV, we assumed that while the shorter observational interval does influence total PONV numbers, it does not influence the association between IOH and PONV. The second possible cause is underreporting by PACU staff. This probably happens independently of the patients' intraoperative blood pressure. Again, such underreporting would only reduce the total number of PONV rather than introduce bias into the estimation of the relationship between IOH and PONV. As we included only patients with a non-empty PONV value in the patient data management system, the risk of patients wrongly being classified as not having PONV is likely to be negligible.
When looking at the usage of PONV prophylaxis in the described cohort, good risk stratification by the treating teams is present: Both intraoperative dexamethasone and ondansetron use was significantly higher in the PONV group.
About 80% of our patients had a non-invasive measurement of their blood pressure at 1 to 5 min intervals. Linear interpolation was used to provide reasonable estimates of intervening values. Obviously, this introduces a small decrease in accuracy compared to invasive measurements available in an interval of up to 15 sec, although it seems unlikely that a further increase in resolution would have changed the characterizations of blood pressure. To ensure this statement, blood pressure characterizations with and without interpolation were compared and resulted in the same measures of central tendency.
The analysis of mean arterial pressure values rather than systolic values was based on two considerations. First, the mean arterial pressure is the value representing organ perfusion best. Second, by technical means, when taking an oscillometric blood pressure, only the MAP is measured directly, while systolic and diastolic pressures are calculated [43].
As this is a retrospective analysis, confounding bias is of concern. We addressed this issue by adjusting for a list of potential confounding variables known from the literature. As with every association study, also in the present analysis an association between different characterizations of IOH and PONV was shown, and it is important to bear in mind that this does not necessarily indicate a causal relationship between these two parameters.
As the study's focus was the development of a prediction model using intraoperative variables for a postoperative prediction of PONV at any time after surgery, "post-exposure" covariates that occurred simultaneously to the outcome of interest were not added to the analysis.
Finally a well-known risk factor for intraoperative hypotension and nausea or vomiting is spinal anesthesia for periparturients [4]. As regional anesthesia was excluded in this trial, there is no risk for introducing bias due to this fact.
When applying the results of the present study in clinical practice, it should be remembered that IOH is associated with PONV, even after short exposure to IOH. Therefore, it could be argued that in patients with high risk of PONV, as well as applying pharmacological PONV prophylaxis, keeping blood pressure always above 60 mmHg may be an important element in the toolbox for PONV prevention.
Conclusions
A distinct association between intraoperative hypotension and PONV in the PACU could be demonstrated. Intraoperative hypotension may thus be yet another risk factor for PONV, which further emphasizes the significance of preventing intraoperative hypotension to improve a multitude of postoperative outcomes including PONV.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/jcm12052009/s1, Figure S1: Univariate and multivariate models for all characterizations of IOH and the risk of PONV in the shaping dataset.
Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee (EC #2062/2020).
Informed Consent Statement: Need for informed consent was waved by the ethics committee.
Data Availability Statement: Data and Code are available from the corresponding author upon reasonable request. | 2023-03-05T16:08:55.357Z | 2023-03-01T00:00:00.000 | {
"year": 2023,
"sha1": "aa269e565eed265471864ab76a10da7904edd15f",
"oa_license": "CCBY",
"oa_url": "https://www.mdpi.com/2077-0383/12/5/2009/pdf?version=1677840060",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "779dc3288d1890a340e47f936592bd3d43fdb79e",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": []
} |
118867091 | pes2o/s2orc | v3-fos-license | Determination of the {\eta} mass with the Crystal Ball at MAMI-B
A new precise determination of the {\eta} meson mass is presented. It is based on a measurement of the threshold for the {\gamma}p ->p{\eta} reaction using the tagger focal-plane microscope detector at the MAMI-B facility in Mainz. The tagger microscope has a higher energy resolution than the standard tagging spectrometer and, hence, allowed an improvement in the accuracy compared to the previous {\eta} mass measurement at MAMI-B. The result m{\eta} = (547.851 +- 0.031 stat. +- 0.062 syst.) MeV agrees very well with the precise values of the NA48, KLOE and CLEO collaborations and deviates by about 5{\sigma} from the smaller, but also very precise value obtained by the GEM collaboration at COSY.
Introduction
The mass of the η meson has been a controversial issue in recent years. Before 2000, three different experiments [1,2,3] yielded comparable masses for the η meson. The Particle Data Group (PDG) then used these results to calculate a weighted mean mass m η = (547.30 ± 0.12) MeV [4]. In 2002, the NA48 collaboration published [5] a very precise result, m η = (547.84 ± 0.05) MeV, which deviated significantly from the world average adopted by the PDG. Including the NA48 measurement in the average, the PDG in 2004 [6] obtained the value m η = (547.75 ± 0.12) MeV, almost 0.5 MeV higher than the value reported previously. This created the motivation to repeat the previous Mainz [3] measurement at MAMI, especially after another precise measurement by the GEM collaboration at the COSY facility [7] gave the result m η = (547.31 ± 0.04) MeV, in agreement with the old measurements of the η mass. Fig. 1 Plan view of the tagging bremsstrahlung facility [16,17] and the Crystal Ball detector (not to scale) at Mainz. The tagger microscope detector [12], giving improved resolution, was installed in the focal plane in front of the main detector at the position indicated. The inset shows the geometry of the microscope scintillators.
Since the previous η mass measurement at Mainz [3], the MAMI electron accelerator [8,9] was significantly improved to provide a more precisely known electron beam energy and much higher beam stability. In addition, on-line monitoring of the electron and photon beam positions in the experimental hall was introduced. Furthermore, the old TAPS setup with its limited angular coverage was replaced with the large acceptance Crystal Ball [10,11] photon spectrometer, which allowed an improvement in the detection efficiency for the two most prominent neutral decays of the η meson, η → 2γ (BR=39.31%) and η → 3π 0 (BR=32.56%). Together with the high energy resolution of the tagged photon beam offered by the recently developed tagger focalplane microscope detector [12], these improvements provided the more accurate determination of the η mass presented in this paper.
This work describes the measurement of the η photoproduction threshold E thr [13] from data measured in 2004-2005 with the Crystal Ball/TAPS detector system and the tagger focal-plane microscope detector. From kinematics of the reaction γp → pη the η meson mass m η was calculated using where m p is the proton mass and c is the speed of light in vacuum.
Experimental setup
In Mainz the real photon beam was produced by bremsstrahlung of the 883 MeV electrons from MAMI-B [8,9] on a 100 µm thick diamond radiator. The absolute electron energy, E 0 , of the incident beam was precisely determined [14,15] in the third race-track microtron of MAMI-B with a total uncertainty of about σ 0 = 140 keV. The photon energies were determined using the Glasgow photon tagging spectrometer (tagger) [16,17], which provided a tagged photon flux of roughly 10 5 s −1 MeV −1 at a beam current of about 35 nA. For the first time the tagger focal-plane microscope detector [12] was used to improve the tagged photon energy resolution. The microscope detector was placed in front of the main focal-plane spectrometer (see fig. 1), so that it covered the region around the η production threshold (E thr ≈ 707 MeV) from E γ = 674 MeV to E γ = 730 MeV at an electron beam energy E 0 = 883 MeV. Made of 96 scintillator strips overlapping to one third, it provided 191 tagging channels with a higher energy resolution of about 0.29 MeV per channel compared to approximately 1.8 MeV available from the main focalplane detector.
into a separate experimental area, where they can induce reactions in the experimental target. Making a coincidence between the reaction products in the experiment and the electron in the tagger, one can determine the energy of the photon E γ as the difference between the main beam energy E 0 and the electron energy The 4.76 cm long liquid hydrogen target was located at the center of the Crystal Ball photon spectrometer. The Crystal Ball [10,11], covering polar angles between θ = 20 • and θ = 160 • , consisted of 672 NaI(Tl) crystals (see fig. 2) and had two openings for the beam in forward and backward directions. Each NaI(Tl) crystal had the form of a truncated 41 cm long pyramid and was equipped with an individual photomultiplier. In order to distinguish between neutral and charged particles detected by the Crystal Ball, the system was equipped with a particle identification detector (PID) [18]. PID was a cylindrical detector, consisting of 24 2 mm thick plastic scintillator strips, arranged parallel to the photon beam axis.
The forward wall detector, TAPS [19], had 510 BaF 2 hexagonally shaped crystals, each equipped with a 5 mm thick plastic scintillator for identifying charged particles. A single BaF 2 crystal was 25 cm long and had an inner diameter of 5.9 cm. The TAPS detector, intended for detecting particles in the forward direction (θ = 4 • − 20 • ), was located at a distance of 173 cm from the Crystal Ball center, making it possible to use the time-of-flight analysis for the particle identification.
The experimental trigger comprised two levels. In the first level, the total energy deposited in the Crystal Ball was checked. If the sum of all photomultiplier analog signals exceeded a threshold corresponding to about 390 MeV, the event was accepted. The secondlevel trigger included a condition on the Crystal Ball sector multiplicity. The 672 crystals of the spectrometer were grouped into 45 sectors of up to 16 crystals each. If at least one of the 16 signals exceeded a threshold of 20 to 40 MeV, depending on the relative energy calibration of the photomultiplier signals, the sector contributed to the multiplicity. All events with multiplicity M ≥ 3 and every third event with M ≥ 2 were recorded for further analysis. The latter condition was especially important for detection of the η → 2γ decay.
Energy calibration of the photon beam
Special care was taken of the energy calibration of the tagger microscope with electrons of different known energies from MAMI. The MAMI accelerator can produce beam energies from 180 MeV upwards in steps of 15 MeV, and it is possible to make accurate measurements of the beam energy. Originally designed to produce electrons of maximum energy 855 MeV, MAMI can also produce a beam of energy 883 MeV by slightly increasing the magnetic field of the bending magnets and slightly raising the energy gain per circulation. Precise determination of the MAMI energy is implemented in an automated manner as a standard MAMI operator menu [14,15]. It is based on a precise measurement of the bending radius of the beam in dipole magnet No. 1 (see fig. 3). After the procedure of optimization, the beam in MAMI is centered along the linac axis with the aid of X-Y position-sensitive HF monitors (XYMOs), as shown in fig. 3. By measuring the angular deviations produced by the steerer magnets and the position of the beam in track 73, with the position monitor [20] placed at a precisely measured distance D m from the linac axis, the bending radius R 73 of the beam in the magnet can be found from geometrical calculations. Since the magnetic field B 0 of the magnets is precisely measured, the energy E 73 of the beam in returning track 73 can be calculated using Table 1 List of contributions to the total uncertainty σ(Eγ) of the photon beam energy calibration. In case the uncertainties are shown with symbol ⊕ (summation in quadrature), the second one originates from the fit to the PTRACE data.
where e is the electron charge and c is the speed of light in vacuum. The final energy E 0 of the electron beam after 90 circulations is interpolated using the calculated data from a simulation using PTRACE, a proven particle tracking program used at MAMI that is based on real measured magnetic field profiles. The calibration of the tagger microscope was performed by varying the magnetic field B scan in the tagging spectrometer around the value B exp used in the experiment. This has been done with three different MAMI energies, E scan , to scan across the tagger microscope by increasing the value of B scan in small steps and plotting the measured hit position of the beam in the microscope versus the equivalent energy Variation of the magnetic field B scan caused slight changes in the shape of the latter, thus, the corrections were applied for this effect using the same method as in [12]. The details of this method are described in [13]. Five calibration scans (three made with energy E scan = 180 MeV, one with E scan = 186 MeV and one with E scan = 210 MeV) gave the 107 data points that are shown in figures 4 and 5. The correlation between the data points of different scans, originating from the uncertainty of the MAMI energy, led to a block-diagonal form of the error matrix used in the fit procedure. The fit was performed by a least squares minimization with the aid of the MINUIT package, supposing a linear dependence of the tagging electron energy, E e − , on the microscope channel. With the known energy E 0 of the MAMI electron beam and the energy E e − of the tagging electron, the energy E γ of the bremsstrahlung photon is determined using The determination of E 0 was made four times during the experiment with the average value E 0 = (883.057 ± 0.134 syst. ± 0.040 non-syst. ) MeV. (5) For tagging electron energies equal to the values of E scan , the photon energy E γ , given by eq. (4), is calculated as the difference between two MAMI energies.
Thus, any systematic uncertainty in determination of E 73 cancels and the non-systematic part contributes twice: where E 73 (exp.) is the energy of the experiment, E 73 (cal.) was measured once for each of the calibration scans, and ∆E 0 and ∆E scan are the differences (calculated by PTRACE) between the measured energy E 73 and the output MAMI energies E 0 and E scan , respectively. The main contributions to the total uncertainty of the MAMI energy, σ 0 = 140 keV, are the uncertainty of the distance D m (see fig. 3), measured by the geodetic method, which is a systematic contribution, and the measured magnetic field uniformity of the MAMI magnets. Since the PTRACE simulation already includes the measured profile of the magnetic field, this contribution is also systematic, and the uncertainty due to the measurement of the absolute magnetic field B 0 , made with an NMR system, can be neglected. The non-systematic contribution, due to the uncertainty in the beam position and in the measurement of the angular deviations produced by the steerer magnets, was estimated to be about 38 keV. In order to obtain an objective estimate of the non-systematic contribution, the RMS deviation from the average of the 106 values of the MAMI energy E 73 measured in the period from 04/2004 to 07/2009, was calculated. The calculation resulted in σ 0 (non-syst.) = 40 keV. All contributions to the uncertainty of the photon beam energy calibration are summarized in table 1. The values of ∆E 0 and ∆E scan were calculated by PTRACE with uncertainties 13 keV and 55 keV [14], respectively. The systematic uncertainty of the calibration fit of 27 keV (σ) was obtained from the fit parameters by applying the law of error propagation to the linear fit function. The initial misalignment of the electron beam at radiator at the beginning of the scans introduced the uncertainty of about 40 keV (σ) [13], and the contribution caused by the drift of the beam position during the experiment was estimated to be 20 keV (σ) [13]. Added in quadrature, all contributions resulted in a total uncertainty of σ(E γ ) = 98 keV in the determination of the photon energy E γ . Applying the law of error propagation to eq. (1), the systematic uncertainty of the η mass due to photon beam energy calibration was estimated to be σ(m η ) = 62 keV.
Analysis and results
The η mesons were identified via their two main decay modes, η → 2γ and η → 3π 0 , with the Crystal Ball/TAPS setup, which measured energies and emission angles of particles. The energy calibration of the Crystal Ball was performed by identifying the reaction γp → pπ 0 . The gains of the photomultipliers were adjusted for all crystals so that the peak position in the two photon invariant mass distribution agreed with the well-known π 0 mass. A cluster in the Crystal Ball was formed by a group of adjacent crystals that had registered parts of the electromagnetic shower initiated by a particle. The weighted mean of the vectors of all contributing crystals, using the square root of the energy in each cluster element as weight, was taken as the vector of the cluster, and the energy sum of these elements gave the total cluster energy. Only crystals with energy deposits greater than 2 MeV could contribute to a cluster, and the total energy threshold for clusters was set to 20 MeV. The charged clusters were identified with PID by checking the agreement of the azimuthal angles of Crystal Ball clusters with the angles of hit PID elements. The uncharged clusters were considered to be photons. In the first step of the analysis, events with two and six coincident photons were selected to pick out candidate η events from the two decay modes mentioned above.
The high intensity of the electron beam caused, for each event, several electron hits to be registered in the tagger focal-plane detector or in the tagger microscope. Only one of these electrons could be correlated to the photon that induced the reaction. This ambiguity was resolved using a coincidence analysis. Since the involved electron had a fixed time difference to the event trigger produced by the Crystal Ball, the tagger hits were cut using a time window around the coincidence peak in the tagger time distribution (prompt time window) shown in fig. 6. The contribution caused by random coincidences was subtracted using the tagger hits within an additional time window in the region of random hits (random time window).
The identification of the η → 2γ decay concentrated on events with a trigger multiplicity M ≥ 2 and two clusters detected as photons, ignoring all other particles. The standard invariant mass analysis using with 4-vectors (E 1,2 , p 1,2 ) of the two identified photons showed a peak at the η mass with a resolution of σ ≈ 20 MeV. Using the tagger hits in the prompt and random time windows, the distribution of the missing mass M miss of the undetected particle (proton) was produced (see fig. 7). Cuts were applied on the invariant mass at M γγ > 480 MeV and on the missing mass at M miss > 880 MeV. The two-photon invariant mass M γγ distribution after the cut on the missing mass and subtraction of the random tagger hits is shown in fig. 7 as the solid histogram. The combinatorial background at smaller invariant masses arises mainly from π 0 production, for example, if two photons from different π 0 mesons are detected within the time resolution or if π 0 π 0 events are produced with two escaping photons. The invariant and missing mass distributions of the η → 2γ decay were simulated using a code based on the GEANT 3.21 simulation library, including all features of the target and detector setup. The simulation of π 0 and π 0 π 0 photoproduction showed that the contribution of the background to the η meson candidates was less than 1.5%. Simulated η events generated an invariant mass in good agreement with the measured data. Though the agreement between simulated and measured background was not perfect, the inconsistency did not extend above M γγ ≈ 480 MeV. After subtraction of the background caused by the random coincidences between the Crys- tal Ball and the tagger microscope, the η yield below production threshold was very close to zero. The identification of the η → 3π 0 → 6γ decay concentrated on events with a trigger multiplicity M ≥ 3 and six clusters detected as photons, ignoring all other particles. Among the 15 possible combinations of six photons to be arranged in three pairs, the combination with the minimum χ 2 -value: was assumed to be correct. Here m γγ [i] are the invariant masses of the photon pairs, σ γγ is the width of the invariant mass distributions and m π 0 is the well-known π 0 mass. Figure 8 shows the distribution of the minimum χ 2 and distribution of the photon pairs invariant mass after a cut at χ 2 < 5 was applied. Simulation of the channel with the GEANT code showed that 12% of the simulated η → 3π 0 events generated a χ 2 beyond this threshold. The invariant mass and missing mass distributions of the six photons, shown in fig. 9, were in quite good agreement with the simulation. Additional cuts on the invariant mass at M 6γ > 460 MeV and on the missing mass M miss < 1020 MeV were applied. The main background contribution was caused by the direct 3π 0 production through the reaction γp → 3π 0 p. Since below the η production threshold (E γ < 707 MeV) no other process can produce six or more photons, the contribution of the resonant 3π 0 production can be estimated by measuring the η cross section below threshold, which ideally must be zero. Such an estimate resulted in approximately 0.12 a. u. (see fig. 10) Supposing that the considered process made the same contribution to the η cross section above the η threshold, the contribution of the 3π 0 events to the η events can be estimated to be about 2% in the energy region 707 < E γ < 730 MeV.
Total cross section
The normalization for the total cross section was obtained from the intensity of the photon flux, the simulated acceptance of the Crystal Ball/TAPS, and branching ratios of the η decays. The photon flux was determined by counting electrons detected in the tagger, and measuring the ratio of the number of tagged photons passing to the experimental area, to the number of the tagger electrons (tagging efficiency). The acceptance of the Crystal Ball was determined by analysis of the events simulated with the GEANT code. For the analysis described in this paper, 10 7 events for each of the two considered η decay modes were generated in the range 707 < E γ < 730 MeV. The acceptance of about 25.2% at the η production threshold, smoothly decreasing to about 24.6% at E γ = 730 MeV, was obtained for the decay η → 2γ. The analysis of the η → 3π 0 decay gave the acceptance of approximately 38.5% at the η threshold, smoothly decreasing to about 36.8% at E γ = 730 MeV.
The resulting total η cross section is shown in fig. 10 in arbitrary units (a. u.). As a first step, the total cross section was obtained using the microscope strips and is shown in fig. 10 top in 0.59 MeV steps for the two considered decays of the η meson. The absolute cross sections obtained using the η → 2γ and η → 3π 0 decays agree well with each other, and the shape of the cross section is in good agreement with the prediction of the ETA-MAID isobar model [21]. Though the background caused by the target windows was measured with an empty target and subtracted, it was not possible to avoid some negligible residual background (≈ 0.12 a. u. in average) below threshold, caused by the background processes considered in section 4. The plot in fig. 10 bottom shows the cross section obtained by summing all η events with the full resolution of the microscope of 0.29 MeV per channel. In order to simplify the procedure of the threshold energy determination, the background below threshold was linearly fitted and extrapolated into the η region. In fig. 10 bottom, this background has been subtracted which decreased the cross section by about 2%. The solid line represents the fit to the data with a function considered in the next section.
Threshold energy and the η mass
Since the determination of the η mass requires a very precise measurement of the production threshold, it was necessary to determine the behavior of the cross section near threshold. Due to strong dominance of the S 11 (1535) resonance in the threshold region [22] it is expected for the total cross section The function based on such dependence, Fig. 11 The total η cross section σ, the square of the cross section σ 2 and the integrated cross section S 2/3 near production threshold. Solid lines represent fits for determination of the threshold energy.
gave good agreement with the shape of the cross section and was fitted to the total η cross section to determine the threshold energy E thr . To improve the statistical uncertainty it was decided to fit the cross section obtained by summing the reconstructed events of both decays. Due to the finite size of the energy bins, the number of counts in the center of the bin is proportional to the average value of the cross section over the width of the bin, so that its content will not be zero even if the bin only slightly overlaps the threshold. Therefore, the cen-ter of the first bin with non-zero counts can be below the threshold, which makes the fitting procedure unstable. One possibility to take this effect into account and make the fitting procedure more reliable is not to use this bin in the fit. Another possibility, which was suggested in [3], is to fit the integral of the cross section. The experimental value of the integrated cross section is given by where σ(i) is the cross section in the energy bin i, and ∆(i) is the width of the bin (the odd and even bins have different width). The sum starts at n 0 , which is the first bin with non-zero counts, and gives the value of S at the upper edge of the bin n, which corresponds to the photon energy Due to s-wave-like energy behavior of the cross section it is expected that and therefore S 2/3 should be a linear function of the photon energy E γ . The threshold energy was obtained by fitting the cross section σ, the square of the cross section σ 2 , and the integral S 2/3 . The σ 2 was fitted with a linear function. The integral S 2/3 showed almost linear behavior close to threshold, but at higher energies a small quadratic term appeared [3]; therefore, the S 2/3 data were fitted with a second order polynomial. The fits are shown in fig. 11. They delivered consistent values for the threshold energy, which were converted to η mass using eq. (1) and are summarized in table 2. The values, derived separately from the η → 2γ and η → 3π 0 events, also agree well within the statistical uncertainties. Analysis of the additional experimental data of the MDM experiment [23] also led to a consistent result for the η mass. Figure 12 shows the η mass obtained by fitting σ, σ 2 , and S 2/3 , plotted versus the upper limit of the fit range, showing good agreement with the η mass found by fitting over the full E γ range up to 730 MeV.
The average of the three values listed in the first row of table 2 resulted in the η mass m η = (547.851 ± 0.031 stat. ± 0.062 syst. ) MeV, (14) where the first uncertainty is due only to statistics, and the second originates from the uncertainty in the photon beam energy calibration. The result of this work supports the three most precise measurements by the NA48 [5], KLOE [24] and CLEO [25] collaborations and disagrees by about 5σ with the smaller value obtained by the GEM [7] collaboration. The η mass determined in Table 2 Comparison of the results for the η mass obtained by fitting the total cross section σ and σ 2 , and the integral S 2/3 up to Eγ = 730 MeV. The most precise result was obtained by fitting all of the data, but results from fitting the η → 2γ and η → 3π 0 data separately are also listed. The uncertainties are statistical only. Fig. 12 The η mass obtained using different methods to fit the threshold energy plotted versus the fit range. The dashed line indicates the result of this work. Fig. 13 Overview of previous η mass measurements in comparison to the world average reported by the Particle Data Group [26,27] in 2006 and 2008 and the result of this analysis [13].
this work is plotted in fig. 13 with the other measurements in the order of the year of publication. The disagreement with the previous measurement [3] at MAMI is most probably due to the lack of online beam monitors that resulted in underestimation of the systematic uncertainty for the photon tagger energy calibration.
Conclusion
This article describes the determination of the η mass by measuring the threshold of the reaction γp → pη at the MAMI accelerator. The high resolution tagger microscope was used for the first time to get the total cross section of the reaction. The η mesons were selected by identifying the decay products of the two most prominent neutral decays, η → 2γ and η → 3π 0 , in the Crystal Ball detector. The three precise η mass measurements by the NA48, KLOE and CLEO collaborations were confirmed, though the result presented in this article disagrees with the GEM collaboration measurement. The uncertainty for the new η mass measurement has been improved in comparison to the previous Mainz experiment by a factor of ∼ 3. The disagreement with the previous MAMI measurement is most probably due to the underestimated systematic uncertainty of the old result. | 2012-07-19T20:46:45.000Z | 2012-07-16T00:00:00.000 | {
"year": 2014,
"sha1": "7e0aa1674d89ea90173df01d62a6d73a35525540",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/1207.3710",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "d80cd3e93b5b0aea084c366e3f1614c5f4e5f034",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
} |
24706191 | pes2o/s2orc | v3-fos-license | Poorly differentiated endocrine carcinoma of the pancreas responded to gemcitabine: Case report.
Poorly differentiated endocrine carcinoma (PDEC) of the pancreas is a rare and aggressive tumor. First-line treatment is commonly a combination of etoposide and cisplatin, but there is no consensus regarding further treatment recommendations. In this report, we describe a case of pancreatic PDEC treated with gemcitabine as third-line chemotherapy. A 62-year-old man with pancreatic PDEC was administered etoposide plus cisplatin as first-line treatment; he then received irinotecan for tumor relapse. However, because irinotecan induced ileus in this patient, we chose gemcitabine as third-line chemotherapy. After two cycles of gemcitabine (1000 mg/m(2) on days 1, 8 and 15 every 4 wk), a partial tumor response was noted by computed tomography (approximately 68% reduction in tumor size). Our patient survived for 15 mo after diagnosis. This is a rare case of unresectable pancreatic PDEC, which showed a partial response to gemcitabine after the failure of two other regimens. Gemcitabine could be an effective treatment option for pancreatic PDEC that is resistant to other treatments.
Abstract
Poorly differentiated endocrine carcinoma (PDEC) of the pancreas is a rare and aggressive tumor. First-line treatment is commonly a combination of etoposide and cisplatin, but there is no consensus regarding further treatment recommendations. In this report, we describe a case of pancreatic PDEC treated with gemcitabine as third-line chemotherapy. A 62-year-old man with pancreatic PDEC was administered etoposide plus cisplatin as first-line treatment; he then received irinotecan for tumor relapse. However, because irinotecan induced ileus in this patient, we chose gemcitabine as thirdline chemotherapy. After two cycles of gemcitabine (1000 mg/m 2 on days 1, 8 and 15 every 4 wk), a partial tumor response was noted by computed tomography (approximately 68% reduction in tumor size). Our patient survived for 15 mo after diagnosis. This is a rare case of unresectable pancreatic PDEC, which showed a partial response to gemcitabine after the failure of two other regimens. Gemcitabine could be an effective treatment option for pancreatic PDEC that is resistant to other treatments.
INTRODUCTION
Pancreatic endocrine tumors (PETs) are rare neoplasms with an annual incidence of less than 1 per 100 000 people [1][2][3][4][5][6] . These tumors account for less than 1%-2% of all pancreatic neoplasms [1,7] . Poorly differentiated endocrine carcinoma (PDEC) of the pancreas is characterized by aggressive tumor biology and poor prognosis. The biological behavior of PDEC is similar to that of small-cell lung cancer (SCLC), and metastatic pancreatic PDECs are often treated with the chemotherapy regimens that are used to treat SCLC. The combination of etoposide and cisplatin has been widely used to treat pancreatic PDEC because no promising chemotherapy regimens have been reported for this disease. Effective second-or later-line chemotherapy is still uncertain. Gemcitabine is an active agent against untreated and recurrent SCLC.
In this report, we describe a case of pancreatic PDEC treated with gemcitabine as third-line chemotherapy.
CASE REPORT
A 62-year-old man with Crohn's disease had previously received treatment at a different hospital. In July 2007, his serum carcinoembryonic antigen (CEA) level was found to be elevated. A contrast-enhanced computed tomography (CT) scan of the patient's abdomen showed a tumor in the head of the pancreas and enlarged paraaortic lymph nodes. In September 2007, he underwent exploratory laparotomy, during which peritoneal dissemination was observed, and hence, a biopsy of the paraaortic lymph nodes was conducted. Based on the histological findings, small cell carcinoma of the pancreas was diagnosed.
Because the tumor was unresectable at the time of diagnosis, the patient was treated with a combination of etoposide and cisplatin as first-line chemotherapy in October 2007. The chemotherapeutic response was deemed to be partial, until multiple bone metastases to the skull, vertebrae, and pelvis were detected using CT after five cycles of chemotherapy. The patient was next administered irinotecan monotherapy as second-line chemotherapy, which started in March 2008. Irinotecan was stopped after one cycle because ileus occurred. He was referred to our hospital for further treatment in July 2008.
The patient had no family history of cancer, and the results of a physical examination were unremarkable. The laboratory findings were hemoglobin 11.5 g/dL (normal 14.0-17.0 g/dL), γ-glutamyl transpeptidase 113 IU/L (normal, 10-47 IU/L), glucose 136 mg/dL (normal, 69-104 mg/dL), CEA 12.8 ng/mL (normal, < 4.0 ng/mL), carbohydrate antigen 19-9 14 U/mL (normal, < 37 U/mL), neuron-specific enolase (NSE) 36.2 ng/mL (normal, < 10.0 ng/mL), and pro-gastrin-releasing peptide (pro-GRP) 338 pg/mL (normal, < 46 pg/mL). A contrast-enhanced CT scan of his abdomen revealed a low-density mass, 7.5 cm in diameter, in the head of the pancreas, as well as enlarged para-aortic lymph nodes at the time of admission. The pancreatic tumor did not show contrast enhancement ( Figure 1A). A CT scan of his chest did not show any primary or metastatic pulmonary tumors. We reviewed an excised biopsy specimen of a para-aortic lymph node obtained at the previous hospital. Histological examination of the specimen showed small to intermediatesized cells with a high nuclear-cytoplasmic ratio and fre-quent mitosis, and partial necrosis. Immunohistochemical staining revealed that these cells were strongly positive for NSE, CD56, and keratin; weakly positive for chromogranin A; and negative for vimentin, leukocyte common antigen, S-100, and CD99 ( Figure 2). On the basis of the pathological findings, the para-aortic lymphadenopathy was determined to be caused by metastasis of PDEC. Therefore, pancreatic PDEC with para-aortic lymph nodes and bone metastases was diagnosed.
We chose gemcitabine as third-line chemotherapy. Starting in July 2008, the patient received 1000 mg/m 2 gemcitabine on days l, 8 and 15 every 4 wk.
After two cycles of gemcitabine, a CT scan of his abdomen showed regression of the pancreatic tumor (from 7.5 cm to 2.4 cm in diameter), and his serum NSE and pro-GRP levels had decreased to within the normal range. The chemotherapeutic response was deemed to be a partial response. After four cycles of gemcitabine, an abdominal CT scan showed a pancreatic mass that was 2.0 cm in diameter ( Figure 1B). In November 2008, after day 15 of the fifth cycle, the patient requested that the therapy be stopped because of general fatigue. He died of multiple organ failure in December 2008.
A B
have reported that PDEC was diagnosed in 17 (9.4%) of 180 patients with non-functioning pancreatic endocrine tumors. PDEC is characterized by aggressive tumor biology and poor prognosis. Bettini et al [8] also have reported that all patients with PDEC died within 3.5 years after diagnosis (median, 11.8 mo), and that only 23.5% of the tumors were resectable at the time of diagnosis. Our patient survived for 15 mo after diagnosis. His survival time was longer than the median survival time that was reported by Bettini et al [8] .
The standard treatment for advanced pancreatic PDEC has not yet been established. The initial approach to treatment of pancreatic PDEC is to attempt curative resection. However, liver and lymph node metastases are present in 32.5% and 59.5% of patients at the time of diagnosis [9] . Therefore, curative surgical resection cannot be achieved in most patients, and effective medical treatment to control metastatic lesions is urgently required. Systemic chemotherapy is proposed for patients with inoperable pancreatic PDEC, and adequate organ function and performance status; however, a standard chemotherapeutic regimen has not been established. In our patient, the tumor was inoperable owing to the presence of peritoneal dissemination and para-aortic lymph node metastases, and hence, systemic chemotherapy was administered to the patient.
The biological behavior of PDEC is similar to that of SCLC, and metastatic pancreatic PDECs are often treated with the same chemotherapy regimens that are used to treat SCLCs. Combination chemotherapy with etoposide and cisplatin, one of the standard regimens for SCLC, is commonly used to treat pancreatic PDEC.
Moertel et al [10] have reported that etoposide plus cisplatin produced good therapeutic results in patients with anaplastic neuroendocrine carcinoma (which has been defined as PDEC according to the recent WHO classification [11] ), with an overall regression rate of 67% and a median regression duration of 8 mo [10] . Other investigators have reported similar results, with a median duration of response of 7-9 mo in patients with poorly differentiated endocrine tumors [12,13] .
Since the report of Moertel et al [10] , the combination of etoposide and cisplatin has been considered to be the reference treatment for PDEC; however, confirmatory studies have not been performed because of the rarity of PDEC. If this first-line chemotherapy fails to treat pancreatic PDEC, there is no consensus regarding further treatment recommendations. Irinotecan plus cisplatin is one of the standard regimens for extensive-stage SCLC [14] . In our case, the patient had been administered irinotecan monotherapy as second-line treatment before he was referred to our hospital. However, this therapy had been discontinued because ileus occurred after one cycle.
Several newer anticancer drugs, including paclitaxel [15] , topotecan [16] and gemcitabine [17] , have shown little activity as single agents against neuroendocrine tumors (NETs). Gemcitabine is a nucleoside analog with structural similarities to cytarabine and is widely used in the treatment of advanced pancreatic adenocarcinoma [18] . In a phase Ⅱ trial of gemcitabine for the treatment of metastatic NETs, Kulke et al [17] have reported that, although the treatment was well tolerated, no radiological responses were observed, 65% of the patients (n = 18) experienced disease stabilization, and that the median overall survival was less than 1 year. However, their study included various histological subtypes of NETs, and only two of the 18 patients had poorly differentiated NETs. Thus, the efficacy of gemcitabine for poorly differentiated NET of the pancreas remains unclear.
Gemcitabine is an active agent against untreated and recurrent SCLC [19][20][21] . The response rate to gemcitabine was reported to be 27% in patients with previously untreated SCLC [19] . In patients with previously refractory or recurrent SCLC treated with at least one chemotherapeutic regimen, gemcitabine resulted in response rates of 6%-17% [20,21] .
We believe that gemcitabine is a reasonable treatment option for pancreatic PDEC, and we chose gemcitabine as third-line chemotherapy. After two cycles of gemcitabine, the pancreatic tumor showed marked regression, which resulted in a partial response. Gemcitabine has shown good efficacy as third-line chemotherapy for refractory pancreatic PDEC. The prognosis of pancreatic PDEC is extremely poor because of its highly aggressive behavior, and hence, effective second-and later-line treatments are important for improving prognosis. In light of this, gem-
A B
citabine could be an effective treatment option for pancreatic PDEC that is resistant to other treatment. | 2018-04-03T05:03:33.346Z | 2010-08-14T00:00:00.000 | {
"year": 2010,
"sha1": "3a3b323cddeac5ad5c0759295ad8580029c6c2cb",
"oa_license": "CCBYNC",
"oa_url": "https://doi.org/10.3748/wjg.v16.i30.3853",
"oa_status": "HYBRID",
"pdf_src": "Adhoc",
"pdf_hash": "385bc167228bdababd171c620778fba974b385fb",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
251742882 | pes2o/s2orc | v3-fos-license | A crossover study evaluating the sex‐dependent and sensitizing effects of sleep deprivation using a nociceptive test battery in healthy subjects
Aim We assessed whether total sleep deprivation (TSD) in combination with pain tests yields a reliable method to assess altered pain thresholds, which subsequently may be used to investigate (novel) analgesics in healthy subjects. Methods This was a two‐part randomized crossover study in 24 healthy men and 24 women. Subjects were randomized 1:1 to first complete a day of nonsleep‐deprived nociceptive threshold testing, followed directly by a TSD night and morning of sleep‐deprived testing, or first complete the TSD night and morning sleep‐deprived testing, returning 7 days later for a day of nonsleep‐deprived testing. A validated pain test battery (heat, pressure, electrical burst and stair, cold pressor pain test and conditioned pain modulation [CPM] paradigm) and sleep questionnaires were performed. Results Subjects were significantly sleepier after TSD as measured using sleepiness questionnaires. Cold pressor pain tolerance (PTT, estimate of difference [ED] −10.8%, 95% CI −17.5 to −3.6%), CPM PTT (ED −0.69 mA, 95% CI −1.36 to −0.03 mA), pressure PTT (ED −11.2%, 95% CI −17.5% to −4.3%) and heat pain detection thresholds (ED ‐0.74 °C, 95% CI −1.34 to −0.14 °C) were significantly decreased after TSD compared to the baseline morning assessment in the combined analysis (men + women). Heat hyperalgesia was primarily driven by an effect of TSD in men, whereas cold and pressure hyperalgesia was primarily driven by the effects of TSD observed in women. Conclusions TSD induced sex‐dependent hyperalgesia on cold, heat and pressure pain, and CPM response. These results suggest that the TSD model may be suitable to evaluate (novel) analgesics in early‐phase drug studies.
| INTRODUCTION
Sleep disturbance is a highly prevalent symptom in chronic pain patients as over 50% also report having impaired sleep. 1 Studies in primary insomnia patients or subjects deprived of sleep reported the development of spontaneous pain and increased sensitivity to (experimentally) evoked pain, 2 therefore proper sleep is necessary to maintain the homeostasis of pain-regulatory processes.
The use of pain models in early-phase pharmacological studies may help to reduce decision-making risks during the translational process from preclinical models to patients with pain, and to determine the biological activity of the studied drug. 3 Most models used to evaluate the effects of (novel) analgesics in healthy subjects are evoked pain tasks eliciting nociceptive pain, eg, pressure or heat application to or on extremities, where there are limited to no disturbances in central pain processing. Such models are less suitable to fully assess the analgesic potential of (novel) neuropathic pain treatments that primarily act in the central nervous system (CNS) and target centrally induced lowering of pain thresholds. 4 Easily adoptable models that assess the effects on central pain processing are mainly based on peripheral input (eg, secondary hyperalgesia induced in the capsaicin or ultraviolet-B [UVB] hyperalgesia models rely on peripheral activation), 5 which does not apply to sensitization caused by sleep deprivation. In addition, models that evaluate central effects often have limitations that prevent repetitive use in studies with a crossover design (which in pharmacological pain studies is preferred to limit the required sample size through increased statistical power compared to a parallel design 6 ). Reasons precluding repetitive use of other central-acting models may be ethical or practical. A high subject burden due to the unpleasantness of the test procedure is one example of the former (eg, intradermal capsaicin injection or high/low-frequency stimulation) 4,7 and longterm adverse effects are another (eg, the prevalence of postinflammatory hyperpigmentation with the freeze injury and ultraviolet-B hyperalgesia models). 8,9 Other models, for example secondary hyperalgesia/allodynia induced by topical capsaicin, appear to be only limitedly sensitive to pharmacological interventions or only limitedly reproducible. 5,10 Sleep deprivation in combination with evoked pain tests may be a suitable alternative to study the effects of centrally acting analgesic drugs. 11 By depriving healthy subjects of sleep, central pain pathways are affected: disturbed sleep significantly decreased pain tolerance to peripheral mechanical, heat and cold stimuli. 2,[12][13][14] Sleep deprivationinduced alterations in the endogenous pain inhibition pathway have also been reported, for example the impaired conditioned pain modulation (CPM) response, a centrally acting mechanism. 2 16 and those reported by Eichhorn et al (heat pain solely decreased in women, cold hyperalgesia was induced sex-independently). 13 Given this discrepancy, and to accurately assess a drug's potential analgesic effect, it is necessary that the sleep-deprivation model is first validated without intervention prior to further use in studies with (investigational) analgesics(s).
Here, we investigated the sex-dependent effects of sleep deprivation on pain responses in healthy subjects using a comprehensive and validated evoked pain test battery. 17
What this study adds
• Sleep deprivation may serve as a surrogate pain model for a central pathological mechanism.
• The results presented validate the use of sleep deprivation in combination with evoked pain tests as a model in healthy subjects for the evaluation of altered central pain processes. Our data suggest that the model may be suited to evaluate (novel) central-acting analgesics in an early-phase drug study.
| Study subjects and design
Twenty-four men and 24 women were enrolled. All subjects provided consent before any study procedures took place. Interested subjects were medically screened and enrolled if they were men (part A) or women (part B), 18-35 years of age (inclusive), and excluded if they had sleep disturbances, irregular sleeping patterns (eg, night shifts) or went through a change in time zone(s) 7 days prior to the first test day. Regular smokers (>10 cigarettes or equivalent per day) were excluded from participation, as were those that smoked 24 hours before each visit, used (illicit) drugs, consumed alcohol within 24 hours prior to each visit, consumed >8 units of (methyl-)xanthineholding products per day or consumed these 4 hours before each visit.
Subjects were acquainted with all tests during a training session (section 2.4) that was part of the screening procedure. Subjects who did not understand the instructions, indicated intolerance during the pain test training or achieved tolerance at >80% of the maximum input intensity for the cold, pressure and/or electrical pain task(s) were excluded. This tolerance threshold of 80% was included to ensure that both an increase and decrease in pain response could be determined during the test visits.
The menstrual cycle may influence (evoked) pain perception. [21][22][23] To minimize a possible influence on test results, women (part B) were required to use a reliable hormonal contraception method at least 30 days before the first study day until the end of the study. In addition, one of the following was required for participation: (i) use of the contraceptive pill continuously (no stop week) throughout the study; (ii) use of the contraceptive pill with a planned stop week, in which the study days were >2 days after restart of contraceptive pill use; (iii) in case of other hormonal contraceptives, study days were >2 days after end of withdrawal bleeding. F I G U R E 1 Study design. Schematic study design of parts A and B. Due to the COVID-19 pandemic, subjects in part B, arm 1 were required to report to clinic at 8.00 instead of 19.00 to allow for a day of quarantine and COVID-19 screening, whereas subjects that participated in part B, arm 2 were required to stay an additional full study day at our unit in quarantine, including COVID-19 screening (see section 2.2). The quarantine day in this arm preceded the regular study visits. Part A was completed before the COVID-19 pandemic and thus was unaffected. T, moments when test rounds were performed
| Study intervention: Sleep deprivation
Subjects were deprived of sleep for at least 24 hours at time of the measurements following TSD. The night before TSD, subjects were instructed to go to bed between 22:00 and 23:00, wake up between 07:00 and 08:00 the next day and report to the clinic around 19.00 to start the TSD night. To aid subjects in staying awake for the duration of the TSD, a personal activities schedule was created at the start of the visit together with the subject (eg, playing [video] games, splashing cold water on the face, light exercise) to provide structure during the night. Subjects were allowed to deviate from this schedule if, for example, an activity was found to be effective to promote wakefulness and prevent sleepiness. During the TSD night (19:00-07:00), study staff closely monitored the subject and reinforced motivation to ensure compliance and that subjects did not took any naps. Caffeine use was not allowed during TSD.
| Study procedures: Evoked pain tasks
Pain detection thresholds (PDT) and pain tolerance thresholds (PTT) were evaluated using a multimodal and fixed-sequence pain test battery at prespecified timepoints (each test round is indicated by 'T' in Figure 1). All subjects thus completed three rounds of pain tests.
Details of the procedures have been described extensively elsewhere and therefore are only briefly repeated here. 24,25 Measurements were performed in the following sequence: heat pain task, pressure pain task, electrical pain taskrepeated stimulus ('burst'), electrical pain tasksingle stimulus ('stair') (pre-cold pressor, #1), cold pressor pain task, electrical pain tasksingle stimulus ('stair') (post-cold pressor, #2), ending with the intra-epidermal stimulation test (results reported elsewhere). 20 Pain intensity for all tests except the heat pain test was captured using an electronic visual analogue scale (eVAS) slider: 0, "no pain" to 100, "worst pain tolerable". PDT was defined as eVAS > 0 and PTT as eVAS = 100.
Heat PDTs were measured with a thermode (contact area 30 Â 30 mm; QSense, Medoc, Israel) placed on the volar forearm that gradually increased from 32 C at 0.5 C/s. The test continued until the subject indicated his/her PDT by pushing a button on a hand-held feedback control or when the safety cut-off of 50 C was reached.
The average of three measurements was used for analysis.
For the pressure pain test, an 11-cm wide tourniquet cuff (VBM Medizintechnik GmbH, Sulz, Germany) was placed over the gastroc- 'stair' test directly stimulates the nerve and bypasses nociceptors, 26 whereas the 'burst' paradigm serves as proxy for temporal summation/ wind-up. 27 With the stair test, single stimuli (10 Hz tetanic pulse,
| Study procedures: Questionnaires
Subject-reported sleepiness was collected using the nine-point 30 Both questionnaires were completed at the start of the well-rested measurements and at the start of the TSD measurements ( Figure 1). Note: Statistical analysis of evoked pain task endpoints for both study parts (A and B) combined. EDs, P values and 95% CIs (in parentheses) are referenced for indicated contrasts. Values are presented either in the unit in which they were measured or in % for tests for which the data were log-transformed. Data in bold and italic denote significant effects (P < .05). EDs < 0 are in favour of first mentioned condition of the contrast (eg, SD in SD vs MORN), >0 in favour of second mentioned condition. Abbreviations: AFT, well-rested afternoon condition; 95% CI, 95% confidence interval; CPM, conditioned pain modulation paradigm; ED, estimate of difference; El stair/burst, electrical stair (single stimulus) and electrical burst (repeated stimulus) pain tests; MORN, well-rested morning condition; n, number of subjects; PDT/PTT, pain detection/tolerance threshold; SD, sleep deprived morning condition.
| Statistical considerations and analysis
The CPM response was calculated as the difference between the electrical stair PDT or PTT measured pre-cold pressor test and the same parameter post-cold pressor test. As the CPM response is generally short-lived, 28 only the results of the electrical stair test within 5 minutes after start of the cold pressor test were used for further analysis to ensure actual CPM effects were evaluated. For CPM PDT and PTT, the data of one subject in part B for the well-rested morning and wellrested afternoon states were excluded. For CPM PDT and PTT in part B, the data for two subjects for the well-rested morning state and for three subjects for the well-rested afternoon state were excluded.
To estimate the differences between groups (sleep deprived/ well-rested morning/well-rested afternoon) and sex, and the interaction sex and group, data were analysed with a mixed-model analysis of variance with a fixed factor group, sex and sex by group, and random factor subject. The Kenward-Roger approximation was used to estimate denominator degrees of freedom and model parameters were estimated using the restricted maximum likelihood method.
Parameters were initially analysed without transformation. Except for those from the heat and CPM tasks, all the pain test parameters suggested log-transformation was required and therefore applied. Logtransformed parameters were back-transformed after analysis to allow for interpretation as percentage change.
T A B L E 3 Results of primary evoked pain task endpoints per study part À7.9%, P = .12 (À16.9-2.1%) Note: Statistical analysis of evoked pain task endpoints per study part (part A, men; part B, women). EDs, P values and 95% CIs (in parentheses) are referenced for indicated contrasts. Values are presented either in the unit in which they were measured or in % for tests for which the data were logtransformed. Data in bold and italic denote significant effects (P < .05). EDs < 0 are in favour of first mentioned condition of the contrast (eg, SD in SD vs MORN), >0 in favour of second mentioned condition. Abbreviations: AFT, well-rested afternoon condition; 95% CI, 95% confidence interval; CPM, conditioned pain modulation paradigm; ED, estimate of difference; El stair/burst, electrical stair (single stimulus) and electrical burst (repeated stimulus) pain tests; MORN, well-rested morning condition; n, number of subjects; PDT/PTT, pain detection/tolerance threshold; SD, sleep deprived morning condition.
| Subject characteristics
See Table 1 for subject characteristics. In part A, 28 men were enrolled of whom 23 completed all study assessments. In treatment arm 1 (Figure 1), one subject had a positive drug test and was replaced; one other withdrew consent and was not replaced. The data obtained from these two subjects from visit 1 could be used for analysis. In treatment arm 2, one subject had a positive drug test and was replaced; his replacement got sick following minor food poisoning and was also replaced. One other withdrew consent following a headache and was not replaced following protocol regulations. The mean age of subjects for which data was used for final analysis was 26 years (SD ± 2.2).
In part B, 24 women were enrolled of whom 23 completed all study assessments. In the second treatment arm, one subject withdrew consent during the TSD night following a headache that did not subside. She was not replaced following protocol regulations. The mean age was 25.9 (SD ± 3.0).
The CPM response PTT was significantly decreased in the combined group after TSD compared to MORN (ED À0.69 mA/s, 95% CI À1.36 to À0.03 mA/s, P < .05). No significant effects were observed in the men-only or women-only groups.
No significant effects were noted for any contrast for the electrical burst or electrical stair pain tests (Tables 2 and 3).
| DISCUSSION AND CONCLUSION
We evaluated the effects of TSD on pain perception in healthy men and women subjects. Cold pressor pain, the CPM response, heat pain detection and pressure pain thresholds were significantly lowered; effects were sex-dependent.
The intervention to deprive subjects of sleep was successful, as observed from the KSS (sleepiness) and LSEQ (early morning behaviour) questionnaire results. While sleep deprivation significantly reduces attention and vigilance, 31 this likely did not bias results as others have discussed previously. 11 In case of reduced attention, a parallel shift in pain detection and tolerance is expected, plausibly on each pain modality in approximately the same manner. Both men and women could still distinguish pain detection from tolerance as exemplified by significantly altered tolerance to pressure and cold pain (PTT), but not detection of pressure and cold pain (PDT; Table 2). Our data therefore indicate that subjects were sleep deprived yet sufficiently focused to properly conduct the tests.
Preclinical studies indicate that sleep deprivation-induced hyperalgesia is partly caused by inhibition of the endogenous opioid protein synthesis and lowering of mu-and delta-opioid receptor affinity. 32,33 Other neurotransmitters, such as serotonin (5-HT), appear to play a role in maintaining hyperalgesia. 34,35 This corroborates clinical evidence showing the importance of opioid and serotonergic supraspinal mechanisms in the descending pain inhibitory pathway, and the impaired CPM response we and others observed following sleep deprivation. 13,36 The increased sensitivity to the pressure pain test we observed, which aims to activate deep tissue mechanoreceptors, 37 builds on this hypothesis, given that the descending pain pathway, when not impaired, particularly inhibits neural activity residing in deep tissue. 38 The exact pathophysiology underlying the observed sleep deprivation-induced heat and cold hyperalgesia is less clear. Heatinduced hyperalgesia is typically restricted to peripheral sensitization mechanisms. In cold-induced hyperalgesia, both peripheral and central processes are involved but these do not relate to those proposed for sleep deprivation. 11 It is interesting to note that the cold pressor task is commonly employed to study the analgesic effects of opioids in healthy subjects and in patients. 39 11 A recent follow-up study reported sex-dependent effects of TSD on the CPM response and heat hyperalgesia in women and sexindependent effects on cold and mechanical hyperalgesia. 13 Sexindependent effects of TSD on the cold pressor task were also observed in another study that employed a similar design. 40 Our results confirm that TSD induces heat, pressure and cold hyperalgesia.
The significantly lowered threshold of heat pain detection in the combined group is largely driven by the effects induced in men, while the significantly lowered threshold of pressure and cold pain tolerance thresholds are largely driven by the effects induced in women (Table 3 and Figure 2). As the pathophysiology of heat and cold hyperalgesia caused by TSD is yet to be elucidated, it is difficult to dissect why effects on the cold pressor test were seemingly more sexdependent in our study. The distinct pressure pain protocols used may have played a role in some of the other discrepancies observed.
Where Schuh-Hofer et al and Eichhorn et al evaluated more superficial and local pain using an algometer and used this for evaluation of the CPM response, 11,13 we used a pressure cuff around the leg that also targeted deep tissue nociceptors, which are primarily affected by the descending pain pathway as stated above, and used the electrical stair task as test stimulus for CPM.
Currently available neuropathic pain treatments are only efficacious to a limited extent, resulting in an unmet clinical need. 41 There is great interest in developing drugs for neuropathic pain in general, and in drugs that can treat central sensitization specifically. Here, we showed that TSD lowered pain thresholds and impaired the CPM response, suggesting that TSD alters descending input from the CNS. 13,15,16 In this study we also observed altered neural activity by increased detection probability of a double-pulse electric intraepidermal stimulus, which further suggests increased facilitation or decreased inhibition (results published elsewhere 20 ). While not strictly related to central sensitization, an imbalance between descending inhibition and ascending facilitation of pain signals is one of the changes found in patients with chronic pain states. 42,43 Sleep disruption also alters other central pain processes in a sex-dependent manner: an impaired CPM response and increase in temporal summation (a phenomenon often reported in chronic pain disorders) was observed in females, 13,44 whereas secondary hyperalgesia (a model for neuropathic pain) was induced in males. 44,45 We discussed in the introduction that, while surrogate models for central sensitization in humans do exist, 4 they mostly require peripheral input and are of limited use in experimental pain studies with a crossover design. TSD in combination with nociceptive testing may therefore offer an alternative method to evaluate central mechanisms that play a role in chronic pain 11 and may be suitable to demonstrate and quantify the effects of analgesics aimed at treating central sensitization and neuropathic pain.
We intend to confirm this assumption in a future study that will include the same pain test battery and TSD model in which we will administer drugs used to treat neuropathic pain to both males and females. If a sex-dependent drug response is observed in this next study, the results may also aid clinicians in making drug prescription decisions for pain patients suffering from sleep disorders. No opioids will be tested in that study as sleep restriction, a similar yet distinct model to TSD, has been shown to attenuate morphine analgesia. 46 To the best of the authors' knowledge, that is the only published study which employed the TSD (or similar) model in the context of analgesic drug testing. 46 It thus remains to be seen which nociceptive test following TSD will be most sensitive to drug effects.
The results presented here should be read with the following considerations. First, only the short-term effects of TSD over one night were assessed, while pain, including that due to sleep disturbance, is mostly a chronic process. The study results can therefore be used for method development, but are only of limited use for understanding the pathophysiology of sleep deprivation-induced/conditioned pain. 34 Additionally, the age of the enrolled subjects was relatively young, limiting our conclusions. A young age range was included as this study was part of a larger study in which we also assessed the effects of TSD on driving performance. 31 Recruiting older subjects for that test was considered unsafe. We did not correct for multiple statistical testing, which may have led to false positives in our results. This was deemed acceptable as the study was exploratory, and is a commonly adopted statistica approach in early-phase exploratory drug studies.
In conclusion, TSD induced sex-dependent hyperalgesia on cold, heat and pressure pain, and impaired the CPM response. This confirmed TSD as a method to alter central pain processes. Our data suggest that the model may be used to evaluate (novel) analgesics in experimental pain studies that (partially) target central processes.
Investigators should be aware of a sex-dependent response when using the model. | 2022-08-24T06:17:56.654Z | 2022-08-23T00:00:00.000 | {
"year": 2022,
"sha1": "54802a83a70a86b3357ecc1d8ee63bd96b6daf3a",
"oa_license": "CCBYNC",
"oa_url": "https://doi.org/10.1111/bcp.15505",
"oa_status": "HYBRID",
"pdf_src": "Wiley",
"pdf_hash": "ea79bd5eb5d62fa02d4dbc2d6f5e0fcbe7bbde99",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
263151858 | pes2o/s2orc | v3-fos-license | Kinetic Ferromagnetism and Topological Magnons of the Hole-Doped Kitaev Spin Liquid
We study the effect of hole doping on the Kitaev spin liquid (KSL) and find that for ferromagnetic (FM) Kitaev exchange $K$ the system is very susceptible to the formation of a FM spin polarization. Through density matrix renormalization group (DMRG) simulations on finite systems, we uncover that the introduction of a single hole with a hopping strength of just $t\sim{}0.28K$ is enough to disrupt fractionalization and polarize the spins in the [001] direction due to an order-by-disorder mechanism. Taking into account a material relevant FM anisotropic spin exchange $\varGamma$ drives the polarization towards the [111] direction via a reorientation transition into a topological FM state with chiral magnon excitations. We develop a parton mean-field theory incorporating fermionic holons and bosonic spinons/magnons, which accounts for the doping induced FM phases and topological magnon excitations. We discuss experimental signatures and implications for Kitaev candidate materials.
INTRODUCTION
Kinetic ferromagnetism, resulting from the subtle interplay between the motion of electrons and their interactions, provides a counter-intuitive example of a strong interaction effect in condensed matter physics [1,2].Nagaoka [3] famously showed how the interference of paths from a single hole doped into a half-filled Hubbard model with infinite repulsion U can lead to a ferromagnetic (FM) state in a system that typically supports antiferromagnetic (AFM) order.Given that the required large interaction limit is an experimental challenge, signatures of Nagaosa's ferromagnetism have only recently been observed experimentally in quantum dots [4] and semiconductor heterostructures [5].In this context, intriguing questions are whether similar kinetic magnetism can appear in other correlated models that are experimentally accessible; and whether the kinetic FM state itself can be nontrivial, e.g., host chiral excitations.
In this work, we account for the effects of hole doping by considering an extension of the Kitaev honeycomb model, referred to as the t-K model, where t denotes the hopping strength of doped holes.In the slow hole limit (t ≪ K), the KSL phase is robust and holes only introduce quasi-static vacancies [49,50], see Fig. 1(a).However, in realistic systems, t is typically much larger than the Kitaev exchange K, and previous works based on parton meanfield theories suggested superconducting ground states in the hole-doped regime [51][52][53][54].Yet, our numerics does not support superconductivity at the considered low doping values, see Supplementary Note 2. Recently, a density matrix renormalization group (DMRG) study suggested a charge density wave ground state emerging on top of a hole-doped AFM KSL, in which the superconducting correlations fall off almost exponentially at long distances [55].
We investigate the ground state of the t-K model by DMRG [56,57] and show that the FM KSL is remarkably fragile already for small and slow hole doping.For hopping strengths t of the order of the KSL's flux gap, ∆E v ∼ 0.1K, the system is already partially FM polarized by the itinerant holes, spontaneously breaking the time-reversal symmetry, as illustrated in Fig. 1(b).To account for our numerical results, we develop a parton theory incorporating fermionic holes and bosonic spinons/magnons.It unveils that the hole kinetic term effectively serves as a FM Heisenberg coupling destabilizing the KSL.In addition, our parton theory shows that the resulting FM order along the [001] direction originates from an order-by-disorder mechanism [58,59].We further uncover that the presence of a finite FM off-diagonal exchange, Γ > 0, shifts the magnetization direction from [001] to [111].Remarkably, due to the change of spin polarization direction, the kinetic FM state becomes non-trivial.It spontaneously forms topological FM order with chiral magnon edge modes, akin to the FM state of the Kitaev model induced by a strong external magnetic field [60,61].Furthermore, we show that hole doping can significantly lower the critical field above which the field-polarized FM state appears in Kitaev candidate materials, such as α-RuCl 3 .
Model Hamiltonian
The celebrated Kitaev honeycomb model is defined by the following Hamiltonian where S γ j (γ = x, y, z) are the S = 1/2 spin operators and S γ j S γ k are Ising couplings according to the γ-type of nearestneighbor (NN) ⟨ jk⟩ bonds.Microscopic derivations [32,33,[62][63][64] have shown that the Kitaev interaction of several material candidates is likely FM.Therefore, we focus on positive K and set K = 1 as the unit of energy.There exist conserved plaquette operators W p on each hexagon p as where the lattice sites j = 1, ..., 6 correspond to the elementary hexagon plaquette p shown in Fig. 1(a).The pristine KSL ground state has W p = 1 for all plaquettes.
To describe the physics of hole doping, we introduce the t-K model [51,65,66] where c † s is the creation operator of an electron with spin index s =↑, ↓ and the projector P removes doubly occupied states.
Spin operators, S γ j = c † j σ γ c j /2, are given by the fermionic vectors c † j = (c † j,↑ , c † j,↓ ) and the standard σ γ Pauli matrices.The hole doping is parameterized by δ such that s ⟨c † j,s c j,s ⟩ = 1−δ.Note that the NN hopping is spin-independent and the effects of spin-orbital coupling are nevertheless retained in H S [65], resulting in a space symmetry group of D 3d .Besides, the t-K model is also symmetric under time reversal transformation and preserves the charge-U(1) conservation.
In order to implement DMRG calculations, the system is placed on a two-dimensional cylindrical geometry, dubbed YCL y , with periodic boundary conditions (PBCs) along the short direction (with L y unit cells), while the longer one (L x ) is open.See details about the honeycomb lattice in Supplemental Note 1.
Phase Diagram
First, we focus on the one-hole doped system.The groundstate phase diagram on YC4 cylinders is analyzed by studying the average magnetic moments = ⟨S i ⟩ and the average Z 2 flux W p , as shown in Fig. 1(c,d).For t < t * , the ground state corresponds to a site-diluted KSL, where the kinetic energy of the holes is insufficient to overcome the Z 2 flux excitation gap ∆E v ∼ 0.1K.Instead, the presence of the slow-moving holes can be thought of as quasi-static vacancies within the KSL state [49,50].Although W p slightly decreases as t increases in this phase, the system does not acquire any magnetization.
We observe that a phase transition occurs at t = t * ≈ 0.28K, as visible by kinks in the curves of both the magnetization and Z 2 flux, see inset plots of Figs.1(c) and (d).As one of our main findings, already for t > t * a FM phase appears with a finite magnetization along the [001] direction.We note that the first-order derivative of the ground-state energy also exhibits a kink around t * ≈ 0.28K which suggests that the phase transition is of second order, see Supplementary Note 2.
This FM phase is an example of kinetic magnetism, which spontaneously breaks the time-reversal symmetry and arises solely from the holes' kinetic energy.Remarkably, the KSL correlations are only partially depleted, as signaled by the gradual decrease of the Z 2 flux W p in Fig. 1(d).This coexistence of magnetization and finite plaquette flux is a general feature of the [001] FM phase until the system is fully polarized by a sufficiently large hopping strength.According to Nagaoka's theorem, for a single-hole doped system [3] in the thermodynamic limit, the saturated FM state is only reached for vanishing spin exchange (or in other words, the on-site Hubbard U → ∞).Therefore, the intriguing coexistence of FM order and KSL correlations could generally persist in systems described by the t-K model.
We further investigate whether FM order persists for multiple holes and study the magnetization as a function of hole density δ in Fig. 2. Unlike Nagaoka's ferromagnetism, which may disappear for a thermodynamic density of holes [67], we find that the FM order observed in our work is robust and can be further enhanced with multi-hole doping.First, a larger δ can lower the critical value of t * .For instance, at t = 0.2K, the system remains in the site-diluted KSL phase without any magnetization for δ ≲ 0.02, while a small but finite magnetic order emerges for δ > 0.02.For the hopping strength t which is already capable of polarizing spins at δ ≈ 0.01, a slight increase in δ can significantly enhance the magnetization until it reaches a saturation value at δ ≈ 0.05.Due to a rapid increase in entanglement entropy, our DMRG simulations face convergence problems for systems with even higher doping levels.Nevertheless, our results suggest that the ferromagnetic order is a generic feature of the t-K model, at least in the low doping limit δ ≤ 0.06.
Off-Diagonal Exchanges
Next, we study the hole-doped extended Kitaev honeycomb model, dubbed the t-K-Γ model, as a more realistic system that incorporates additional off-diagonal symmetric spin exchange parameterized by Γ.The spin part of the model now becomes with α β γ for the Γ terms on γ-type bonds.Although the Γ term is typically believed to be AFM [46,68], here we focus on the effect of a FM Γ > 0 as this has the most interesting consequences.
In the absence of Γ, the spin polarization is always along the [001] direction (or by symmetry equivalently, the [010] and [100] directions), which can be attributed to an order-bydisorder mechanism as we will explain below.In Fig. 3(a), we show that the presence of Γ > 0 can shift the spin polarization direction from [001] to [111].Indeed, already a tiny value of Γ * is sufficient to induce a significant change in the polar angle of magnetization, denoted as ϕ ≡ cos −1 (M z /|M|).For Γ = 0.02, the polar angle is already very close to the value of , which corresponds to a FM order along the [111] direction, see Fig. 3(b).At Γ * ≈ 0.01K, a noticeable kink in the first-order derivative of the ground-state energy in Fig. 3(c) indicates that this change in spin polarization direction is accompanied by a phase transition.We will discuss below that this transition leads to topologically nontrivial magnon excitations.The magnetization displays a dip around Γ * , corroborating the findings from the analysis of the ground-state energy.Note that a moderately large Γ can also enhance the magnitude of the magnetization.
Compared to the FM Γ, the perturbation of an AFM Γ has a much smaller impact on the [001] ordered ground state.For instance, a small AFM Γ ≈ − 0.02 only induces a canting field in the [001] order.However, a FM Γ with a similar magnitude is sufficient to shift the polarization direction from [001] to [111].
Parton mean-field theory
In order to understand the selection of the FM moment direction and the impact of Γ, we develop a parton meanfield theory.To effectively describe the low-energy degrees of freedom in the kinetic FM phase, we consider the following parton representation for electron operators [69] with c † j,s ≡ h j b † j,s .Here, b † j,s 's are Schwinger boson operators representing spinons, and h j 's are fermionic holon operators representing empty sites.The local constraint, h † j h j + s b † j,s b j,s = 1, ensures the original three-dimensional physical local Hilbert space.The parton theory is constructed such that it allows for spontaneous symmetry breaking toward ordered states.
The t-K-Γ model takes the form with F jk ≡ s b † j,s b k,s representing short-range FM spin correlations [70].One can observe that the hopping term in Eq. ( 4) connects the hole's kinetic energy P jk ≡ h † j h k with FM correlations F jk , suggesting that finite doping with dominant t dramatically renormalizes the spin-spin interactions and fosters FM order.
The second term in Eq. ( 4), Ĥγ jk , represents spin interactions as defined in Eq. ( 3), but expressed in terms of bosonic spinons instead of electron operators.We develop a large-N mean field theory within a Schwinger-boson approach [70][71][72] in which the spinon occupancy is given by n b = s b † j,s b j,s = 1 − δ (see Supplementary Note 3).Guided by our DMRG results, our interest lies in the FM phase with t ≫ K.We also explored the possibility of a gapped QSL phase but found no evidence thereof in the limit of n b → 1.Hence, the most natural ansatz is a FM state along the general direction of (sin ϕ cos θ, sin ϕ sin θ, cos ϕ).For conveniently describing magnon excitations, we can replace the Schwinger bosons with Holstein-Primakoff (HP) bosons as [1,73] where a † j and a j are the creation and annihilation operators for the HP bosons (magnons), respectively.Then we perform a standard Hartree-Fock decoupling of Eq. ( 4) to obtain a meanfield theory with separate hole and spin parts, dubbed H h and H s , respectively.Here, H h refers to a spinless free-fermion band with renormalized bandwidth by the factor |⟨ F jk ⟩| and filling of δ.The spin part H s is treated in spin-wave theory (SWT) incorporating up to fourth-order Holstein-Primakoff expansions, and the spin amplitude is renormalized to S = n b /2.The expectation values of FM spin correlations F jk and hole's kinetic energy P jk are determined self-consistently.Remarkably, we can now see that for δ ≪ 1 the kinetic energy effectively acts as a FM Heisenberg interaction with a coupling constant J ≡ − ⟨P jk ⟩t/n b ∼ |t|δ.
The semi-classical ground-state energy of the FM Kitaev honeycomb model (t = Γ = 0) turns out to be independent of ϕ and θ, yielding an emergent O(3) manifold of degenerate ground states.This classical O(3) degeneracy can be lifted by quantum fluctuations in a quantum order-by-disorder mechanism [58,59].Within SWT the quantum correction to the ground-state energy is , where N c is the number of the unit cells and ω k,n are the two magnon bands of the honeycomb lattice.Note that ω k,n depend on ϕ and θ implicitly, and thereby so does E s .Our calculations reveal that a magnetization along the [001] direction (or equivalently, the [100] and [010] directions) is energetically preferred in accordance with our DMRG results (see Supplementary Note 3).
A complication arises from the fact that the semi-classical ground state of the pure Kitaev honeycomb model is a classical spin liquid [74] which is manifest in the linear SWT for arbitrary spin polarization directions as a nearly flat magnon band with almost zero energy (see Supplementary Note 3).Thus, an HP expansion that only includes the quadratic terms is insufficient to correctly capture the quantum fluctuations.We employ a nonlinear SWT by including the quartic terms in the HP expansion and treat these within the Hartree-Fock decomposition [75].We indeed find significant renormalizations of the flat magnon bands, for instance, a closing of the gap between formerly flat magnon bands in the nonlinear SWT for Γ = 0.More technical details can be found in Supplementary Note 3.
Including a FM Γ > 0 breaks the semi-classical O(3) degeneracy of the ground state.Our SWT shows that the semiclassical ground state has [111] magnetic order rather than [001], which is again consistent with our DMRG calculations, see Fig. 3(b).From the magnetic field polarized regime of the Kitaev model, it is known that different polarization directions can have qualitatively different types of boundary magnon excitations as first pointed out in Ref. [60,61].For spins polarized along the [001] direction with Γ = 0, the two magnon bands touch linearly and are almost non-dispersive along one of two primitive vectors.The system thereby cannot support nontrivial boundary excitations.On the other hand, the introduction of a finite Γ > 0 modifies the magnon bands from the change of the spin polarization to the [111] direction.Consequently, the two magnon bands are separated by a gap and acquire non-zero Chern numbers (−1, 1) for the lower and upper magnon bands, respectively [60,76,77].Hence, as our second main result, we find that the kinetic FM state of the doped Kitaev model (with small Γ > 0) spontaneously breaks TRS forming a topological magnon insulator.This explains the phase transition found when turning on the off-diagonal Γ interactions, see Fig. 3.
To see how the change in spin polarization can give rise to chiral boundary states, we calculate the magnon bands within SWT on narrow cylinders.With PBCs along y-direction, k y is still a good quantum number and the 2D systems can be treated as L y independent 1D chains.The magnon spectra for the [001] and [111] ordered states are shown in Figs.4(a) and (d).Indeed, only the [111] ordered state has chiral boundary modes (green) as predicted.
Real-time dynamics
So far our observation that a topological FM state is realized for the [111] direction solely relies on the parton+SWT description.
A numerical confirmation by DMRG is complicated by the fact that the ground state is topologically trivial but only the magnon excitation spectrum carries signatures of the magnon Chern bands.Therefore, we study the distinct real-time dynamics of excitations in the [001] and [111] ordered FM states.To obtain an initial state with a magnon-like excitation at the boundary, we apply a local time-reversal operator at lattice site j, which couples to spin excitations for arbitrary spin polarizations, to a one-hole doped ground state |Ψ⟩ as where K denotes the complex conjugate operator.Note that σ y j also annihilates the components of |Ψ⟩ which contain a local hole at site j.As a result, in addition to flipping the spins at site j, it also weakly excites charge excitations and slightly modifies the global magnetization pattern.Starting with the prepared initial state |Ψ( j)⟩, the system is evolved by performing a standard real-time evolution as |Ψ( j, t)⟩ = e −iHt |Ψ( j)⟩.In practice, we approximately represent the shorttime unitary operator, e −iHdt , as a compact matrix product operator (MPO) [78], where a small time step dt = 0.01K has been used.Then the time evolution can be efficiently simulated by applying such an MPO to |Ψ( j, t)⟩ successively.
We create a one magnon defect at site j in the leftmost column of the YC cylinders and follow the spin current for each column x defined as with ⟨..⟩ the expectation value with respect to |Ψ( j, t)⟩ and (x, y) denoting the lattice coordinates.Note that j s (x, t) is obtained by summing over all the sites belonging to the xth column, namely, summing over allowed k y .Additionally, we also keep track of the charge current where n h j = 1 − P s c † j,s c j,s P is the occupancy of holes.The simulations for the dynamics are performed on YC3 cylinders, and the results are shown in Fig. 4. We observe a distinct difference in the spin currents between the [001] and [111] ordered states, as evidenced by the leftmost boundary columns in Figs.4(b) and 4(e), respectively.The spin current in the [001] ordered state shows very little dynamics, as indicated by its small magnitude of j s ∼ 10 −4 in Fig. 4(b).It can be attributed to the absence of distinct boundary modes and the almost vanishing velocity of magnon modes.In stark contrast, for the [111] direction j s (x, t) exhibits a strong and long-live signal in the leftmost boundary as expected for a localized surface excitation.Note that Eq. ( 5) unavoidably involves excitations related to bulk magnon bands, which explains why the boundary spin current can leak into the bulk.We have checked that the magnetic field-induced topological FM in the [111] direction [60,61] shows a very similar response confirming the presence of chiral magnons.
The charge current for the [001] and [111] ordered states are very similar, see Figs. 4(c) and (f), respectively.This is consistent with our parton mean-field theory.Both the [001] and [111] ordered states give rise to similar short-range FM spin correlations (see Supplementary Note 3), which suggests that the holon Hamiltonians are similar.
Implications for the field-polarized state of α-RuCl 3 Unlike the simplified models introduced in Eq. ( 1) and Eq. ( 3), more realistic models for experimentally relevant materials are much more complicated.Here, we focus on α-RuCl 3 and related materials [31,33,79] which are often described by a number of parameters, including {K, J, Γ, Γ ′ , J 3 }.The corresponding Hamiltonian is expressed as where α β γ and ⟨i j⟩ 3 denotes the third NN bonds.10); see also model parameters in Eq. ( 9).The left inset shows a typical zigzag order at (t, h in ) = (0.3, 0.03) and the right one shows a polarized state at (t, h in ) = (0.5, 0.06).
Positive and negative local magnetizations, ⟨S x j ⟩, are represented by red and blue dots, respectively, with their radii scaled according to the magnitude of ⟨S x j ⟩, with ⟨S x j ⟩ ≈ 0.21.The doping level is δ = 0.0625.
we focus on a specific point in the parameter space, given by which is close to a set of parameters identified in Ref. [79] for α-RuCl 3 , resulting in a zigzag-ordered ground state obtained by DMRG simulations.Moreover, an in-plane magnetic field suppresses the long-ranged zigzag order and ultimately polarizes the state for sufficiently strong fields.Furthermore, it has been proposed that an intermediate KSL state may separate these phases.To account for this effect, we introduce an additional Zeeman term in Eq. ( 8).The total spin Hamiltonian is then given by where h in represents the in-plane magnetic field along the [1 10] direction.
We then investigate how hole doping modifies the phase boundary of the zigzag phase.We add the same kinetic energy term for the holes as in Eq. ( 2), to the experimentally relevant Hamiltonian Eq. ( 10) and compute the ground-state phase diagram with DMRG; see Fig. 5. Remarkably, we find that the phase boundary between the zigzag and polarized phases is highly sensitive to hole doping.For a vanishing hopping strength, t → 0, the critical magnetic field required to polarize the spins is approximately h * in ≈ 0.085.However, already for a small hopping t = 0.8K and low hole doping of δ = 0.0625, the critical magnetic field is reduced to h * in ≈ 0.04.A larger value of t ≈ 1.8K can drive the zigzag order to a FM order (not shown).This can be understood by noting that |t|δ effectively acts as a FM Heisenberg exchange constant J.The polarized phase is distinct from the kinetic FM phase, as the latter develops magnetic order spontaneously.The significant effect induced by the kinetic energy of holes, therefore, is a complicating factor in identifying to correct Hamiltonian parameters in comparison to experimental measurements.
DISCUSSION
In summary, we have studied the hole-doped Kitaev-Γ model using DMRG and effective parton descriptions.It is by now well known that the FM KSL is much more fragile with respect to the application of an external magnetic field compared to the AFM KSL [80].Similarly, our DMRG simulations suggest that the AFM KSL is also much more robust to hole doping.Meanwhile, distinct hole spectral functions have been numerically observed in FM and AFM KSLs, in which a dynamical Nagaoka ferromagnetism emerges in the FM KSL only [81].Concentrating on FM Kitaev exchange, we have shown here that small hole doping is sufficient to destabilize the KSL leading to a kinetic FM state generally coexisting with finite plaquette fluxes.We emphasize that, unlike the field-polarized phase in a magnetic field [82], the doping-induced FM order breaks the timereversal symmetry spontaneously.Due to the particle-hole symmetry of our model, our conclusions also hold for charge doping.
We developed a parton mean-field theory, incorporating fermionic holons and bosonic magnons, which indeed shows that hole condensation gives rise to effective FM Heisenberg exchanges.We found that in the absence of off-diagonal exchange Γ, the spin polarization is along the [001] direction due to an order-by-disorder mechanism, which is also supported by DMRG simulations.A FM Γ term switches the polarization direction to [111], which leads to a distinct topological FM phase with chiral boundary magnon excitations.
Our work raises a whole range of questions for future research.First, one could investigate flux threading and topological entanglement of the FM phase coexisting with finite plaquette fluxes to probe the presence of topological order.Second, it would be interesting to search for other kinetic FM phases of Hubbard models hosting chiral magnon excitations.Third, to further investigate the transition between the KSL and kinetic FM phases, it would be desirable to develop a parton theory incorporating bosonic holons and fermionic spinons.In this context, such a theory could be the starting point for a self-consistent random phase approximation [83] or Gutzwiller-boosted DMRG [84][85][86][87] in order to study the ordering instabilities and spin excitation spectra.Forth, in our real-time evolution results we could observe that the excitations of spin induce a (weak) response of charge, and vice versa.This could possibly be captured by a time-dependent parton mean-field theory in future studies to explore the interplay of charge and spin transport [88].Fifth, a study of doping effects for realistic model Hamiltonians could unveil a new mechanism (see Supplementary Note 4) for stabilizing the zigzag magnetic orders observed experimentally.This could also be important for refining the microscopic models of RuCl 3 and its peculiar magnetic field response, both of which remain poorly understood [79].Sixth, the origin of the observed thermal Hall effect of RuCl 3 [89] is hotly debated [90] and it would be very interesting to explore how the topological magnons of the weakly doped Kitaev model studied here may contribute.
In conclusion, the Kitaev honeycomb model continues to be a fertile soil for novel physics.We expect the interplay of Kitaev spin exchange with kinetic electron motion to hold further surprises in the future.
Supplemental Materials for "Kinetic Ferromagnetism and Topological Magnons of the Hole-Doped Kitaev Spin Liquid"
This Supplemental Material includes more details of the DMRG calculations and parton theory.In Supplementary Note 1, we show the considered honeycomb lattice and the corresponding one-dimensional path for DMRG simulations.In Supplementary Note 2, we show additional information about the DMRG results.In Supplementary Note 3, we show details of the parton mean-field theory.In Supplementary Note 4, we provide additional results for the zig-zag order.In Supplementary Note 5, we provide more information about the spin-1/2 moment around the exactly solvable point of t = 0. To employ the MPS-based methods, one must establish a site ordering for the honeycomb lattice.This can be achieved by assigning an integer from 1 to 2N c to each lattice site, where N c is the number of unit cells.As shown in Fig. 6, we utilize a site-labeling scheme for the honeycomb lattice on a YC4 (L y = 4) cylinder with length L x with primitive vectors ⃗ x and ⃗ y.Note that the x-, y-, and z-type bonds for a Kitaev honeycomb model are marked by the blue, green, and red lines, respectively.
Ground-state calculations
For the DMRG simulations we make explicit use of a U(1) quantum number to control the level of hole doping.The bond dimension of DMRG is kept as large as χ = 4000, resulting in a typical truncation error of ≈ 7×10 −7 .To analyze the convergence of our DMRG simulations, we study the scaling behavior of the ground-state energy and magnetization with the inverse DMRG bond dimension 1/χ.As shown in Fig. 7, the ground-state energy varies weakly with χ for χ > 1000, suggesting a faithful convergence in our DMRG simulations.In the slow-hole regime (Kitaev spin liquid phase), the magnetization fluctuates around the magnitude of truncation error, indicating the absence of magnetic order.In the fast-hole regime (FM-ordered phase), the magnetization extrapolates to a finite value when 1/χ → 0, indicating that the kinetic FM order is stable when increasing the bond dimension.
Fig. 8 displays the ground-state energy E and its first-order derivative ∂ t E as functions of hopping strength t.Note that ∂ t E in Fig. 8(b) exhibits a kink at t * ≈ 0.28K which suggests that the phase transition is of the second order.
A. Superconducting correlation functions
The multiple-hole doping allows us to analyze the equal-time pairing-pairing correlation function.A diagnostic of the superconducting order is the pair-pair correlation function, defined as where ∆ † a (r 0 ) is SC pair-field creation operator with pairing index a defined on the γ-type (γ = x, y, z) NN bond r 0 and r is the distance along the cylinder (the x direction).We consider one even-parity pairing as where i 0 and j 0 denote two lattice sites on the NN bond r 0 .Similarly, we define three odd-parity pairing functions as In practice, we set i 0 and j 0 being the reference bond at x ∼ L x /4 to minimize the boundary effect.We study the abovementioned pair-pair correlation functions on the x-, y-, and z-type bonds for δ ≈ 0.04 and as shown in Fig. 9, we find that all of 11)-( 13)] obtained on a YC4 cylinder with Γ = 0, δ ≈ 0.04, and χ = 4000.For (a-c) we set t = 20K and for (d-e) t = 2K.The even-parity correlation functions, which are not shown here, are significantly smaller than the odd-parity functions (Φ e ∼ 10 −12 ).them exhibit exponential decays, The correlation lengths χ a for different pairing types are summarized in Fig. 9.This observation suggests the absence of superconductivity in the weak doping limit.
Real-time dynamics
The real-time dynamics is performed by the standard MPO-evolution algorithm [78].A small time step dt = 0.01[K] −1 has been used and At each intermediate step of the real-time evolution, one needs to truncate the MPO-evolved MPS.In order to estimate the accuracy of the final MPS, we introduce the accumulated truncation error defined by where ϵ j (D) is the sum of the discarded squared singular values at the j-th bond.Meanwhile, we also track the growth of entanglement entropy (EE) of the evolved MPS.As shown in Fig. 10, the truncation errors are close to zero when t < 3 and are always less than 10 −5 during the whole evolution period of 5[K] −1 .Note that EE displays a jump at t = 0 because we introduce a magnon-like excitation at the boundary.Then, the EE gradually increases until it saturates to a plateau value.
SUPPLEMENTARY NOTE 3: PARTON MEAN-FIELD THEORY
In this section, we provide a detailed description of the parton mean-field theory defined in Eq. ( 4) of the main text.We summarize the main results as follows: We decouple the t-K-Γ model in the FM phase into separate hole and spin parts, H ≈ H h + H s , using a mean-field theory.The hole part H h can be effectively described by H h in Eq. (22).For the spin part, the groundstate energy corrected by the zero-point quantum fluctuation is E s in Eq. (34), which leads to a quantum order-by-disorder phenomenon for Γ = 0.The corresponding magnon excitations can be effectively described by a spin-wave theory with Hamiltonian H s ≈ H (2) s + δH (2) s , as defined in Eqs. ( 31) and (36), respectively.As mentioned in the main text, we represent fermionic electron operators in terms of fermionic holons h j and bosonic spinons b j,s : This representation, which is particularly suitable for studying symmetry-breaking states, enlarges the local Hilbert space.In order to restore the original Hilbert space, one needs to impose the local constraints, h † j h j = δ and s b † j,s b j,s = 1 − δ, for every lattice site j.By means of Eq. ( 15), the kinetic terms can be rewritten as We introduce a vector of spinons, b † j = b † j,↑ , b † j,↓ , and express the bilinear spin-spin interactions in a compact form as Then, we can obtain the spin part Hamiltonian on the NN bonds ⟨ jk⟩ ∈ γ, as Assuming a translational invariant ansatz, the Hamiltonians for the hole and spin parts can be further decoupled into several quadratic terms by introducing mean-field parameters on the NN bonds ⟨ jk⟩ ∈ γ, which should be determined self-consistently from Here P γ represents the holon mobility.And generally, F γ and A γ represent the ferromagnetic and antiferromagnetic short-range correlations, respectively.After implementing the following decoupling scheme as the mean-field Hamiltonian for the hole part reads where µ is the Lagrange multiplier to tune the filling number of holons such that ⟨h † j h j ⟩ = δ.Since P γ ∼ δ, the terms in the last line in Eq. ( 22) can be neglected in the small doping limit.
The four-boson terms Ĥγ jk in the spin part Hamiltonian need to be further decoupled in order to achieve a quadratic form.For instance, we employ the following mean-field treatment for Ĥx jk , which enables us to decouple the Kitaev-type interactions as follows: and the anisotropic spin-spin interaction as Similar decomposition schemes are also applied for Ĥy jk and Ĥz jk .For convenience, we further define on a NN bond ⟨ jk⟩ ∈ γ.Since the system exhibits translational invariance, we can perform a Fourier transform on the bosonic spinons and introduce a spinon vector at momentum k as where the additional index l = 0, 1 denotes the two sublattices of a honeycomb lattice.Using D γ , ψ † k , and a Lagrange multiplier λ to tune the condition such that n b ≡ ⟨b † j,s b j,s ⟩ = 1 − δ, we arrive at the mean-field Hamiltonian for the spinon part where two 4 × 4 matrices T γ k and D γ k correspond to bosonic hopping and pairing terms.The large-N limit of H s is taken with a parameter [72] κ ≡ n b /N being fixed.For a small value of κ, the Schwinger-boson mean-field theory may give rise to a gapful QSL, while a magnetically ordered state with bosonic spinon condensation appears at large κ.In our study, we focus on its realistic limit with κ = n b = 1−δ, i.e., N = 1.Here the explicit forms of the terms in Eq. ( 25) are The quadratic Hamiltonians H h and H s can be diagonalized by eigenvalue decomposition and bosonic Bogolyubov transformation, respectively.Here we denote the dispersion relations for the Schwinger bosons (spinons) as ϵ k,n with n = 1, 2, 3, 4.This enables us to numerically obtain the self-consistent solution for Eqs. ( 19)-( 21) through an iterative scheme.For instance, single occupancy leads to where ρ denote the condensate density, and 4 × 4 matrices V 11 k , V 12 k , V 21 k , and V 22 k form the symplectic matrix M k diagonalizing the Hamiltonian in Eq. ( 25), as . When ρ > 0 is a finite value, the Schwinger bosons condense at specific momenta This leads to a long-range magnetic ordered state.For instance, a ferromagnetic order along the z direction means ⟨b † j,↑ ⟩ = ⟨b j,↑ ⟩ = √ ρ = √ κ and similarly, one along the x direction When ρ = 0, all eigenmodes of Eq. ( 25) are gapped and the corresponding ground state is a disordered quantum spin liquid state.When A γ = 0 and thereby V 12 k = 0, one must have ρ = κ = 1 − δ corresponding to a ferromagnetic ordered ground state.
When carrying out the calculations, different randomly generated mean-field parameters are selected to initialize the iterations.However, after careful investigations, we conclude that, in the parameter regime of FM Kitaev coupling and subdominant Γ, there is no gapped QSL solution.It is well-known from the exact solution that the Kitaev honeycomb model exhibits a gapless Z 2 QSL.Hence, it is not surprising that the Schwinger-boson mean-field theory fails to describe the QSL phase for a FM Kitaev honeycomb model and only predicts a ferromagnetic ordered state, which is equivalent to a spin-wave theory.In the following, we would like to adopt a ferromagnetic order ansatz, which allows us to solve the mean-field theory using spin-wave theory in a more straightforward manner. [001] [100] [010]
B. Spin-wave theory for the spin part Hamiltonian
Based on our DMRG results for t > K and the findings of Schwinger boson mean-field theory, we would like to focus on the FM ordered phases.Instead of the Schwinger boson representation, for simplicity, here we use Holstein-Primakoff (HP) representation.For a semi-classical ferromagnetic order along the (sin ϕ cos θ, sin ϕ sin θ, cos ϕ) direction, the "spin operators" (note that they are not the physical S = 1/2 spins) now can be expressed as Ŝ y j = sin θ sin ϕ(κ − 2a † j a j ) + sin θ cos ϕ κ − a † j a j (a † j + a j ) where a † j and a j are the creation and annihilation operators for the HP bosons (magnons), respectively.The Schwinger bosons are closely related to the HP bosons via their single occupancy constraints.When θ = ϕ = 0, it is easy to verify that by eliminating the b ↑ boson using the constraint, the correspondence is [1] b ↓ ↔ a and b ↑ ↔ κ − a † a.
For the generic case, we can obtain that b ↑ → e −iθ/2 cos(ϕ/2) κ − a † a − e −iθ/2 sin(ϕ/2)a, b ↓ → e iθ/2 sin(ϕ/2) κ − a † a + e iθ/2 cos(ϕ/2)a.Substituting Eq. ( 27) into the spin part Hamiltonian, we can obtain a spin-wave mean-field theory up to the fourth-order: Before discussing each term in the above Hamiltonian, it is worth emphasizing that the self-consistent conditions become Eqs.( 19) and ( 29) which are numerically solved in an iterative manner.When carrying out the calculations, different randomly generated mean-field parameters are selected to initialize the iterations.
Here H (0) s is a constant term, namely, the semi-classical ground-state energy, where ℜ[•] denotes taking the real part, N c is the number of total unit cell, and Note that D ≡ D x = D y = D z is ensured by the C 6 lattice rotational symmetry.The first-order term, H (1) s , is linear in the HP bosons: Note that both H (0) s and H (1) s are independent of Kitaev coupling K, and thus, for Γ = 0, there are no constrains on the values of ϕ and θ.This indicates an emergent O(3) manifold of degenerate ground states for the pure Kitaev honeycomb model in the semiclassical limit.With Γ > 0, a stable solution of ferromagnetic order requires H (1) s to vanish, which, in turn, leads to a magnetic order along the [111] direction with ϕ = arccos(±1/ √ 3) and θ = ±π/4.H (2) s is the quadratic term in the Hamiltonian.In the basis of where T k and D k are 2 × 2 matrices: We emphasize that when Γ > 0, the spin polarization direction must be fixed as ϕ = arccos(±1/ √ 3) and θ = ±π/4.It is worth noting that due to the lattice and time-reversal symmetries, in practice, we have P x = ℜ[P x ] = P y = P z .Therefore, we verify from the form of T k that the hole kinetic terms can play a role similar to the ferromagnetic Heisenberg interactions in the spin-wave theory with J ≡ −⟨P x ⟩t/κ.
C. Order-by-disorder
Note that the constant term in Eq. ( 31) arises from the zero-point energy contribution of the bosons, i.e., appears because we express this quadratic Hamiltonian in a bosonic BdG form.Therefore, the ground state energy of H s up to the second order corrections reads where ω k,n are the magnon dispersions.Note that ω k,n still depends on θ and ϕ even when Γ = 0, indicating that the O(3) degeneracy at the Kitaev point is lifted by quantum fluctuations.We find that a [001] (or equivalently [100] and [010]) magnetization is more energetically favorable than other directions, see Fig. 11, indicating the presence of a quantum orderby-disorder mechanism.D. Flat magnon band in H (2) s When we diagonalize H (2) s , we find that the lower magnon band, ω k,0 , is a nearly flat band with almost zero energy, see Figs. 12, regardless of the spin polarization directions.This flat band arises because the semiclassical ground state of the ferromagnetic Kitaev honeycomb model is not necessarily a ferromagnetic order but instead a classical spin liquid.In order words, one can construct numerous zero-energy excitations on top of a ferromagnetically ordered ground state in the semiclassical Kitaev honeycomb model, as demonstrated in Fig. 13.Though we use the [100] and [xy0] order as examples to illustrate the zero-energy excitations, the same mechanism is applicable for arbitrary [xyz] order.Indeed, one can construct the same type of excitations on top of a [xyz] order by fixing the z components of all spins uniform and unchanged.Therefore, one can always observe flat magnon bands for arbitrary [xyz] ordered ground state in the pure Kitaev honeycomb model.
The flat band we discussed above can gain dispersion in two scenarios: (i) a finite positive Γ which also forces the polarization direction to be [111], and (ii) a large t with a finite doping level δ (i.e., a nonzero P γ ).Note that the lower band of H (2) s exhibits gapless Goldstone-like modes even when Γ > 0, as shown in Fig. 14(a), although the system does not have any continuous symmetries to be spontaneously broken.This phenomenon has also been reported in Refs.[40,60].The zero-energy modes disappear in the presence of finite doping δ, as shown in Fig. 14(b).
Note that in this context, we have neglected the fourth-order contributions from the kinetic terms since P γ ∼ δ is relatively small in the small doping limit.Furthermore, the cubic terms vanish for the [001] order.We also find that the correction of cubic terms are not important for the [111] order, also see Ref. [75].
By introducing several Hartree-Fock parameters: the terms in H (4) s on a α-type bond ⟨(r, 0), (r ′ , 1)⟩ are decomposed (r is the unit-cell index) as Then the quartic interaction terms can be replaced with where δE (4) s and δH (2) s are the Hartree-Fock corrections to the ground-state energy and magnon spectrum, respectively.
Substituting the above decomposition into the quartic terms in H (4) s leads to Here the explicit form of δT k is Similarly, And the fourth-order correction to the ground-state energy is In the end, we would like to comment that the inclusion of δH (2) s leads to several important effects in the spin-wave theory.In Fig 15, we show the magnon bands corrected by the fourth-order term δH (2) s .We find that δH (2) s significantly changes the dispersions of both the lower and upper magnon bands.In particular, the lower band gains much more dispersion, as shown in Fig. 15(a) and (b).Furthermore, the energy gap between the lower and upper magnon bands becomes closed due to the fourth-order corrections when Γ vanishes, see Fig. 15(b).It indicates that the topological properties of the magnon bands are also changed by δH (2) s .Finally, a finite Γ can reopen the energy gap, as shown in Fig. 15(c).
SUPPLEMENTARY NOTE 4: ZIGZAG ORDER
Motivated by the experimental Kitaev materials, we discuss the effect of hole doping on the zigzag ordered state.Previous studies [29,46,48] report that the K-Γ model alone is insufficient to support a zigzag order, even with a dominant antiferromagnetic Γ < 0. Our simulations are consistent with these findings, as we observe something like spiral and/or incommensurate orders rather than zigzag order up to Γ = −5K, see Fig. 16(a).On the other hand, the t-K-Γ model at a doping level δ ≈ 0.01 can stabilize the zigzag ordered phase with a moderate t ≈ 0.2K with Γ = −5K, as shown in Fig. 16(b).This is attributed to that the hole kinetic terms act as a ferromagnetic Heisenberg exchange effectively, J ∼ |t|δ, which can prompt a zigzag order in the K-Γ model [66].
We find that a relatively large value of |Γ| > K is required to achieve a zigzag order in the t-K-Γ model, which might be due to the usage of narrow cylindrical geometries in the DMRG simulations [41].We anticipate that this value will significantly decrease in the two-dimensional limit.Alternatively, we introduce a small third NN antiferromagnetic Heisenberg exchange J 3 < 0 on top of the t-K (Γ = 0) model, which has shown to be sufficient to drive the system to a zigzag ordered phase [32,34,47], even on narrow cylinders.We find that in this scenario, a small t can also enhance the stability of the zigzag phase.For instance, when J 3 = −0.07K, a doping level of δ ≈ 0.01 and a hopping strength of t = 0.2K can approximately double the magnitude of zigzag order compared to the case at t → 0. Overall, our results suggest that a zigzag order can be more easily stabilized in the K-Γ (K-J 3 ) model by supplementing it with a moderate hole kinetic term t as well as small hole doping.Nevertheless, it is worth noting that if the hopping strength t dominates over the antiferromagnetic interactions of Γ < 0 (J 3 < 0), the system still undergoes a ferromagnetic ordering transition.Here we clarify the absence of a free spin moment in the gapless phase even in the limit of a static hole, i.e. hopping t = 0.As discussed in Ref. [49] of the main text, the KHM with a vacancy is still exactly solvable with the help of the Majorana representation.Denoting three Pauli matrices as σ γ (γ = x, y, z), we introduce σ γ j = ic 0 i c γ i with c 0 i and c γ i being the matter and gauge Majorana fermions, respectively.The doped hole introduces a static vacancy to the KHM.Without loss of generality, we assume that this vacancy is located at the site i 0 of A sublattice.Then, there are three unpaired gauge Majoranas (namely, Majorana zero modes), c x j , c y k , and c z l , where j, k, and l are three NN sites of i 0 .And there must be one Majorana zero mode γ 0 as a superposition of matter Majorana fermions.
After a Gutzwiller projection those four Majorana zero modes can lead to two-fold ground-state degeneracy.Moreover, in the gapped A phase [14], γ 0 is spatially localized around the static vacancy, as shown in Fig. 17 A phase, those four modes can effectively form a free and local spin-1/2 moment, leading to a finite magnetization of a single free spin-1/2 in the ground state manifold.However, in the gapless B phase, γ 0 is no longer a local zero mode, see Fig. 17 (b).Therefore, although the system exhibits an odd number of spins and the ground states must be two-fold degenerate, there is not necessarily a free spin-1/2 moment.In stark contrast, this effective spin-1/2 degree of freedom, which exhibits zero-energy excitation, strongly interacts with the rest of the system.In this case, even though the ground state has an odd number of spins, it does not give rise to a magnetization of a free spin-1/2 moment.
FIG. 1 .
FIG. 1. Magnetization in the hole-doped Kitaev spin liquid (KSL).(a) Illustration of the t-K model with three types of bonds, γ = x, y, z.Sites j = 1, ..., 6 label the convention for plaquette operators W p .The system sustains the site-diluted KSL ground state with fractionalized spins (W p ≈ 1) only for a slow hole (red ball).(b) For t ⪆ K doping depletes plaquette operators W p while simultaneously polarizing the spins, which leads to FM order along the [001] direction.(c,d) The ground-state phase diagram for the one-hole doped t-K model (K = 1) on a YC4 cylinder with length L x = 12 (corresponding to a doping level of δ ≈ 0.01).The evolutions of (c) the magnetic moment M z and (d) the average flux W p indicate a phase transition from a sitediluted KSL (blue regime in insets) to a FM state (red regime) around t * ≈ 0.28.
FIG. 2 .
FIG. 2. The magnetization as a function of doping level δ for variable values of hopping strength on YC4 cylinders with length L x = 12 and K = 1.
FIG. 3 .
FIG. 3. Effect of ferromagnetic off-diagonal symmetric exchange Γ.(a) The polar angle ϕ ≡ cos −1 (M z /|M|) representing the magnetization direction as a function of Γ, where ϕ = 0 indicates the [001] direction and the black dashed line the [111] direction.Inset: a finite FM Γ > 0 can induce a rotation of the spin polarization direction from [001] (transparent purple arrows) to [111] (solid red arrows) by an angle ϕ.(b) The amplitude of magnetization as a function of Γ for the one-hole doped model with t = 20K on a YC4 cylinder with length L x = 12, corresponding to doping level δ ≈ 0.01.Inset: The first-order derivatives of ground-state energy versus Γ.
FIG. 4 .
FIG. 4. Distinct boundary magnon excitations for the [001] and [111] polarized FM orders.(a) Magnon bands on a cylinder for a [001] ordered state within nonlinear SWT.(b, c) The real-time dynamics of (b) spin current j s (t, x) and (c) charge current j c (x, t) for a [001] ordered state.These results are obtained by DMRG using one-hole doped YC3 cylinders (δ ≈ 0.013).For (a, b, c), we set the offdiagonal exchange Γ = 0 and hopping strength t = 10K.The selfconsistent solution with doping level δ = 0.01 gives kinetic hole energy ⟨ Pjk ⟩ ≈ −0.93 × 10 −2 and FM spin correlation ⟨ F jk ⟩ ≈ 0.79 on the x− (y-) type bonds as well as ⟨ F jk ⟩ ≈ 0.82 on the z-type bonds.(d) Magnon bands within nonlinear SWT on a cylinder for the [111] ordered state.The chiral edge states are marked in green.(e, f) The real-time dynamics of (e) spin current j s (t, x) and (f) charge current j c (x, t) for a [111] ordered states within DMRG.For (d, e, f), we set Γ = 0.05K and t = 10K.The self-consistent solution with δ = 0.01 gives kinetic hole energy ⟨ Pjk ⟩ ≈ −0.93 × 10 −2 and FM spin correlation ⟨ F jk ⟩ ≈ 0.85 on all three type bonds.
SUPPLEMENTARY NOTE 1 :FIG. 6 .
FIG.6.Schematics of the labeling scheme for a honeycomb lattice as a cylindrical geometry with primitive vectors ⃗ x and ⃗ y.The blue, green, and red lines represent the x-, y-, and z-type bonds for a Kitaev honeycomb model, respectively.The periodic boundary condition has been imposed along the y direction, as indicated by the red dashed lines.
( 28 )
We have b † ↑ b ↑ + b † ↓ b ↓ = κ and hence κ = n b = 1 − δ is the density of spinons.Naturally, the short-range ferromagnetic correlation, up to the second-order of HP bosons, becomes
FIG. 13 .
FIG. 13.(a) In a [100] ordered ground state, where the effect of z-type (red dashed line) and y-type (green dashed line) bonds vanishes effectively, a zero-energy excitation can be constructed by flipping two spins (highlighted by the red shadow) connected by one x-type bonds (blue solid lines) into the [ 100] ( 1 = −1) direction.(b) Arbitrary numbers of zero-energy excitations can be created, leading to a flat magnon band for the [100] polarization.(c) In a [xy0] ordered ground state, the effect of z-type bonds (red dashed line) vanishes effectively and the honeycomb lattice is reduced into several independent chains formed by x-and y-type bonds (blue and green solid lines).The mechanism described in (a) and (b) no longer applies to this case.Instead, a zero-energy excitation can be created by rotating all the spins on a single chain (highlighted by the yellow shading) by an arbitrary angle around the z-axis.(d) All the spins on each decoupled chain can be rotated freely around the z-axis without costing energies, leading to numerous zero-energy excitations and thereby a flat magnon bond for a [xy0] polarization.Note that the mechanism described by (c) and (d) is also applicable to the [100] ordered state.
3 FIG. 16 .
FIG. 16.Zigzag order induced by mobile holes.Real-space magnetization ⟨S z j ⟩ (⟨S x j ⟩ and ⟨S y j ⟩ are similar) profiles for (a, b) t-K-Γ model on YC3 cylinders and (c,d) t-K-J 3 model (Γ = 0) on YC4 cylinders.Here J 3 is the 3rd NN antiferromagnetic Heisenberg exchange, K = 1, and the doping level is δ ≈ 0.015.Positive and negative magnetic moments are shown by red and blue dots, respectively, whose radius scales with the magnitude of magnetization |⟨S z j ⟩|, such that |⟨S z j ⟩| ≈ 0.06 in the two middle columns in (b).
FIG. 17 .
FIG. 17.The single-particle wave function of the Majorana zero mode γ 0 in the matter sector on a L y = 6 and L x = 12 honeycomb lattice.Note that we plot the honeycomb lattice as a 2L y × L x square lattice, where factor 2 accounts for the A and B sublattices.(a) The gapped A phase with K z = 3K x = 3K y .(b) The gapless B phase with K z = K x = K y . | 2023-09-28T06:42:26.313Z | 2023-09-26T00:00:00.000 | {
"year": 2023,
"sha1": "9099308dc12362041e1464dc2d930d44af201bf4",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.1038/s41535-024-00678-8",
"oa_status": "GOLD",
"pdf_src": "ArXiv",
"pdf_hash": "5bbe2b7a818d4155d58158c5d9de85fca667b2fe",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
} |
10378070 | pes2o/s2orc | v3-fos-license | Reshaping of Bulbar Odor Response by Nasal Flow Rate in the Rat
Background The impact of respiratory dynamics on odor response has been poorly studied at the olfactory bulb level. However, it has been shown that sniffing in the behaving rodent is highly dynamic and varies both in frequency and flow rate. Bulbar odor response could vary with these sniffing parameter variations. Consequently, it is necessary to understand how nasal airflow can modify and shape odor response at the olfactory bulb level. Methodology and Principal Findings To assess this question, we used a double cannulation and simulated nasal airflow protocol on anesthetized rats to uncouple nasal airflow from animal respiration. Both mitral/tufted cell extracellular unit activity and local field potentials (LFPs) were recorded. We found that airflow changes in the normal range were sufficient to substantially reorganize the response of the olfactory bulb. In particular, cellular odor-evoked activities, LFP oscillations and spike phase-locking to LFPs were strongly modified by nasal flow rate. Conclusion Our results indicate the importance of reconsidering the notion of odor coding as odor response at the bulbar level is ceaselessly modified by respiratory dynamics.
Introduction
Nasal airflow is the natural vector for odorant molecules so that respiration and odorant sampling are indissociable. Nasal airflow is thus a major parameter to take into consideration when studying olfactory processing in mammals, especially because sniffing parameters, such as frequency and flow rate (ml/min), are highly variable [1][2][3]. The importance of nasal airflow dynamics has been revealed at the olfactory epithelium (OE) level. First, olfactory receptor neurons (ORNs) have been shown to be sensitive to air pressure [4]. Second, low versus high flow rates differentially favor sorption of odorant molecules depending on their solubility [5][6][7], resulting in differential influence on activity patterns across the OE [8,9].
Due to the anatomical organization of the epithelio-bulbar projections [10][11][12], a change in OE activity should be reflected in olfactory bulb (OB) activity. Indeed, optical recordings reveal that glomerular activation varies with sniffing frequency [13][14][15]or flow rate [16], and a sniff frequency-dependent attenuation of glomerular inputs has also been described [15]. At the level of individual mitral/tufted cell (M/T) activity, in the anesthetized rat, temporal patterns of M/T cells reflect the phasic stimulation of ORNs at each inspiration [17]. In the behaving rodent, natural changes in sniffing frequency lead to changes in the variability and spatial organization of M/T cell responses [18]. Importantly, temporal patterning of M/T cells is frequently lost during high frequency sniffing [18][19][20]. All of these studies have focused on the effects of sniffing frequency variation on bulbar activity. Conversely, the effects of nasal flow rate variation on M/T cell and local field potential (LFP) responses have not been studied, to our knowledge. It thus appeared essential to characterize nasal airflow-induced changes in OB response. If sampling behavior can change the characteristics of information carrier (spikes and LFP), then it would be necessary to reconsider the notion of odor coding. To investigate these modifications, we used a double cannulation and simulated nasal airflow protocol in anesthetized rats to uncouple nasal airflow from animal respiration. Both unit M/T cell extracellular activity and LFP were recorded during different conditions of nasal flow rate. We found that airflow changes in the normal range were sufficient to substantially reorganize the response of OB and determined that nasal airflow itself modifies spontaneous OB activity.
Results
Under basal flow rate conditions and for the majority of animals, we observed a temporal pattern of LFP signals in response to ISO similar to those of non-tracheotomized rats ( Fig. 1A1 and A2, middle). The few animals in which we did not observe the alternation between beta and gamma oscillations under ISO stimulation at the basal flow rate were not retained for the analysis. Analyses were performed on 12 rats.
Effect of nasal airflow on bulbar odor responses
Analyses were performed on 69 paired trials (23 trials for each flow rate condition) containing a total of 36 cells. To evaluate if nasal airflow shaped odor response at the olfactory bulb level, we first explored its influence on odor-induced LFP oscillations and then on M/T cell responses to odors. Next, we explored the influence of nasal airflow on the phase locking between M/T cell unit activity and LFP oscillations. For most analyses, since odors induced no statistical difference in the global proportion of respiratory-pattern activities, responses to odors were pooled.
LFP activity. Examples of typical signals in response to ISO are presented in Figures 1A1 and A2. In this example, while the classical alternation between gamma and beta oscillations on the respiratory cycle was observed at the basal flow rate (Fig. 1A1, A2, middle), low and high flow rate conditions induced LFP modifications. On the raw signal, the low flow rate (left) induced more numerous beta oscillatory bursts at the expense of gamma oscillatory bursts. In the example of Figure 1A2, the beta oscillatory burst appeared longer under low flow rate than under basal flow rate conditions. For analysis of the whole data set, we compared data from recordings obtained under low, basal and high flow rate conditions, performed at the same site from the same electrode under the three conditions. We first compared the number of beta or gamma oscillatory bursts per electrode and per recording under each flow rate condition (Fig. 1B). We observed that the mean number of gamma oscillatory bursts significantly increased with increasing flow rate (low versus basal: Wilcoxon = 27.5, p,0.01; basal versus high: Wilcoxon = 15, p,0.001). Conversely, the mean number of beta oscillatory bursts significantly decreased (low versus basal, Wilcoxon = 7, p,0.001; basal versus high, Wilcoxon = 17.5, p,0.001), indicating that nasal airflow can quantitatively change the LFP pattern. We next examined if the intrinsic characteristics of oscillations, such as duration, amplitude and frequency, could also be modified by flow rate. The only significant modifications we observed concerned gamma oscillations; the duration of gamma episodes was significantly decreased when flow rate was decreased relative to basal flow rate (Table 1, Wilcoxon = 3, p,0.05), and their amplitude was significantly decreased (Table 1, Wilcoxon = 3, p,0.05).
In summary, the occurrence probability for odor-evoked LFP oscillations was strongly modified by nasal flow rate, while the intrinsic characteristics of oscillations were less or not. The next Figure 1. Modification of LFP oscillatory patterns. A1) Raw data collected from the same electrode in response to ISO under low (left), basal (middle) and high (right) flow rate conditions. A2) Representation of an average respiratory phase-frequency of LFP oscillatory activity under low, basal and high flow rate conditions in response to ISO. Amplitude is color-coded (the calibration scale below is common to the three representations in arbitrary units). Oscillation frequencies are represented relative to their position in the respiratory cycle, indicated in phase from 0 (beginning of inspiration) to 1 (end of expiration). Phase 0.5 indicates the transition between inspiration and expiration epochs. An average was calculated from signals recorded by all electrodes from 12 trials in each condition. B) Mean (6 SEM) number of oscillatory bursts per electrode under low, basal and high flow rate conditions. For each trial, the same electrode was selected for the three flow rate conditions (n = 23 trials). step was to ask whether the M/T cell unit activity in response to odors was modified by nasal flow rate. M/T cell activities. We recorded 36 mitral cells under each flow rate condition. We first compared M/T cell responsiveness under the three flow rate conditions. A cell was considered responsive when its respiratory pattern and/or frequency changed from a spontaneous to an odor period. A decreasing flow rate resulted in a significant decrease in the percentage of responsive M/ T cells (low: 66.7% vs. basal: 88.9%; Chi 2 (1) = 5.142, p,0.05). No significant difference was observed between basal and high flow rate conditions (88.9% vs. 91.7%, respectively; Chi 2 (1) = 0.158, p.0.05). When comparing the mean instantaneous firing frequency under the three flow rate conditions (Fig. 2A1), we observed that it reached approximately 50-60 Hz under the basal (61.58 Hz) and high (51.26 Hz) flow rate conditions, as reported in non-tracheotomized animal [21,22]. Interestingly, it was significantly decreased under the low flow rate condition (low: 42.89 Hz vs. basal: 61.58 Hz). For a detailed view, see Figure S1 showing the matrix of the mean instantaneous frequency for each cell under each flow rate condition. We next compared the mean instantaneous frequency of M/T cells as a function of the respiratory cycle under each flow rate condition. Respiration-triggered histograms showed that M/T cell activity remained modulated by the respiratory rhythm, regardless of the flow rate (Fig. 2A2). Distribution of M/T cell mean instantaneous frequency relative to the respiratory cycle was compared using the Equal Kappa Test. This test revealed a significant difference of distribution between low and basal conditions (p,0.001) while there was no statistical difference between basal and high flow rate conditions.
We then compared the proportion of respiration-related patterns under the three flow rate conditions. Figure 2B1 shows the respiration-related patterns of a series of 36 M/T units recorded under the three flow rate conditions. Even though no significant difference appeared in the total proportion of each pattern between the three conditions ( Fig. 2B2), 58.3 and 50% of the patterns were modified when airflow was decreased and increased, respectively, relative to the basal flow rate (Fig. 2B1). Thus M/T cell respiratory-pattern in response to the same odor is modified depending on imposed nasal flow rate.
We observed that nasal airflow influenced both OB unit and network activities. Finally, we investigated the temporal relationship between spikes and LFP oscillations.
Relation between LFP and unit activity. To characterize the temporal relationships between oscillatory fields and individual spikes, we plotted the spike phase distribution for each cell for both beta and gamma bands under each flow rate condition. Phase histograms were then computed across cells. The significance of spikes/LFP phase locking was tested on the histograms (Rayleigh test; Fig. 3). Our analysis revealed a significant phase locking between spikes and beta oscillations only for the low flow rate condition. Conversely, a significant phase locking was observed between spikes and gamma oscillations for all three flow rate conditions (Rayleigh test, p,0.05). We then compared the strength of phase locking between the three flow rate conditions using the Equal Kappa test. This test revealed a significant difference of phase locking strength between basal and high flow rate conditions for the gamma oscillation (Fig. 3B). Hence, the higher the flow rate was, the stronger the spike phase locking to gamma phase was. Conversely, the lower the flow rate was, the stronger the spike phase locking to the beta phase was (Fig. 3A).
We showed here that changing flow rate modified odor response at the level of the OB. Since, in our experimental conditions, changing airflow modified both air pressure in the nasal cavity and the odorant stimulation (i.e., the quantity and migration of odor molecules), we then wanted to determine the relative influence of nasal airflow itself and that of the odor. Since odorant stimulation cannot be applied without nasal airflow, we attempted to answer this question by examining the effect of nasal airflow on OB spontaneous activity.
Effect of deodorized nasal airflow on bulbar activity
First, we tested the effect of the presence of a respirationmodulated deodorized airflow (nasal airflow ON, 500 ml/min at the basal flow rate) versus its absence (nasal airflow OFF) in the nasal cavity on OB spontaneous activity (96 trials for each condition, with 45 cells detected). Effect of continuous deodorized nasal airflow on LFP activity was also tested on three additional rats. Second, we tested the effect of different nasal flow rates of deodorized air on M/T cell spontaneous activity (23 trials under low, basal and high flow rate conditions, with 36 cells detected).
LFP activity. Since there was no airflow through the nasal cavity in the OFF condition, we used animal respiration as the time base for signal analyses (Fig. 4A). Even though fast LFP oscillations did not appear without odor, the slow rhythm related to respiration (1-3 Hz) was present. As shown in Figures 4A1 and A2 (bottom), when deodorized airflow passed through the nasal cavity, we observed a slow rhythm. Conversely, under the nasal airflow OFF condition, we never observed the slow rhythm ( Fig. 4A1, A2, top). We then tested the influence of flow rate modulation on spontaneous slow LFP activity. Surprisingly, no significant difference appeared in the slow LFP modulation amplitude when flow rate was decreased (200 ml/min) or increased (800 ml/min, data not shown). In addition, even by increasing or decreasing nasal flow rate, fast LFP oscillations did not appear. To complete the study of the effect of nasal airflow on slow LFP modulation activity, we tested the effect of a continuous nasal airflow at 200, 500 and 800 ml/min. On the 27 trials realized in three rats, we never observed a slow rhythm induced by a continuous nasal airflow whatever the flow rate condition ( Figure S2).
M/T cell activity. Forty-five mitral cells were recorded both in nasal airflow ON and OFF conditions (deodorized air). When airflow was switched from OFF to ON (500 ml/min) in the nasal cavity, two M/T cell populations were revealed. First, we found air-insensitive M/T cells, defined as cells that did not change their temporal firing pattern and/or spike frequency when airflow was ON in the nasal cavity compared to the nasal airflow OFF condition. Out of 45 cells, 33 (73%) were insensitive to airflow. A representative example is shown in Figure 4B1 (a). Conversely, 12 cells (27%) were air-sensitive M/T cells and were characterized by a significant change in their spiking activity when airflow was ON in the nasal cavity [see Fig. 4B1 (b)]. This representative cell did not present activity when nasal airflow was OFF, but it presented respiration-related activity when nasal airflow was ON. To recapitulate pattern changes in air-sensitive mitral cells, respiration-related patterns were classified into three groups ( Fig. 4B2): NULL, NS and SYNCHRO (in which S+, S-and Sc patterns were pooled). As shown in Figure 4B2, when nasal airflow was OFF, 75% of cells presented a NULL activity, and 25% presented an NS activity. When nasal airflow was turned ON, the majority of air-sensitive cells adopted an activity synchronized with respiration (SYNCHRO: 58.33%), whereas NULL activity no longer existed.
Mitral cell spontaneous activity was then studied under the three flow rate conditions. A total of 36 cells were recorded under each flow rate condition (deodorized air). When comparing the mean instantaneous frequency rate between flow rate conditions (Fig. 5A1), we observed, as in response to odors, that spontaneous activity had a lower mean instantaneous frequency rate in the low flow rate condition (basal: 41.15 Hz vs. low: 24.06 Hz; p,0.05). No significant difference was observed between basal and high flow rate conditions. For a detailed view, see Figure S1 showing the matrix of the mean instantaneous firing frequency for each cell under each flow rate condition. We also compared the mean instantaneous frequency as a function of respiratory cycle for each flow rate condition (Fig. 5A2). No obvious modulation of the instantaneous frequency was present, with the exception of a slight one at the high flow rate. Equal Kappa test did not reveal any significant difference of instantaneous frequency distribution relative to the respiratory cycle between the three flow rate conditions.
To explore to what extent M/T cell activity pattern modifications in response to odors were modified by nasal airflow variation, we next studied the effect of nasal flow rate modulation on the spontaneous activity patterns of M/T cells in the absence of any odor (Fig. 5B1). In a global view, proportions of the spontaneous activity respiratory patterns differed from those under odor conditions, as the NULL and NS patterns were the patterns most observed regardless of the flow rate condition. As seen in response to odors, even though no specific activity pattern appeared when flow rate was increased or when it was decreased (Fig. 5B1, B2), 50 and 55.6% of these patterns were modified when airflow was decreased and increased, respectively (Fig. 5B1).
Discussion
The aim of this study was to examine the extent that OB odor response was influenced by nasal flow rate variation. For this purpose, we used a double tracheotomy paradigm coupled with a respiratory cycle simulation in the nasal cavity. We extended and corroborated the earlier conclusion of various authors [4,23,24] showing that nasal airflow itself imposes a respiratory rhythm to OB activity. Importantly, we observed that flow rate modifications induced variations of different odor information carriers: LFP oscillatory activity, M/T single cell activity and spike phase locking to LFP oscillations. Nasal airflow tunes OB activity during the respiratory cycle Adrian [23] first demonstrated a periodic LFP activity in the OB related to the animal's respiration. On unitary level, M/T cells fire in relation with air intake in absence of olfactory stimuli [25,26]. Similarly, olfactory receptor neuron terminals are activated by natural sniffing of deodorized air in the awake rat [24]. Recently, Grosmaitre [4] proposed an explanation about respiration-related rhythmicity by showing that ORNs are sensitive to air pressure. Consistent with these results, we showed that airflow suppression in the nasal cavity led to suppression of respiration-related slow LFP oscillation (Fig. 4A). Application of a continuous nasal airflow did not led to respiration-related slow LFP oscillation whatever the flow rate ( Figure S2). At mitral cell activity level, a continuous nasal airflow did not induce a respiration related pattern in mitral cells; conversely, mitral cells adopted a continuous firing [17]. Effect of central structures as a source of respiratory modulation could not be excluded [27]. Periphery and central structures probably act in concert. However, in our conditions, nasal airflow seems to be the most prominent source of respiratory modulation.
Second, we observed a population of M/T cells that adopted a respiration-related pattern of activity when airflow was applied to the nasal cavity (Fig. 4B). These observations appear to confirm the hypothesis according to which mechanosensitivity of ORNs may be a peripheral drive to synchronize OB rhythmic activity with respiration [4]. This respiratory tuning at the level of the network and M/T cells constitutes a major process for synchronizing OB and higher order structures, such as the piriform cortex, which also presents respiration-related activity [28][29][30]. Bulbar respiratory tuning by nasal airflow could permit an optimal coupling between olfactory structures, which may be important for information transmission. Furthermore, this bulbar ''pre-activation'' by nasal airflow could prepare the OB to process odor stimuli [31]. By imposing a basal activity to the OB, nasal airflow may provide an excitatory source to the OB, which in turn could facilitate its response to odors.
Nasal flow rate strongly impacts odor response in the OB Numerous authors have shown that OE odor responses are modified by nasal flow rate [7][8][9]. Such modifications have also been revealed at the glomerular level. In particular, it has been recently reported that glomerular responses are affected by nasal flow rate [16]. Considering this, our data show for the first time that the response to odors of both OB network and output (M/T cells) are modified by nasal flow rate variations ( Fig. 1 and Fig. 2). We showed that modifications of nasal airflow not only influence bulbar input but also its output activity. At the network level, gamma oscillations appeared more often under the high nasal flow rate, while beta oscillations appeared more often under the low flow rate. These results can be simply interpreted in terms of bulbar activation related to the odorant stimulation level which can be induced by each flow rate. Gamma oscillation prevalence is likely due to the high bulbar activation related to a high flow rate. It has been shown that the gamma oscillation is favored by high odor concentration [32] and high vapor-pressure molecules [33]. Conversely, beta oscillation prevalence is likely due to poor bulbar activation (related to a low flow rate), as it has been shown that the beta oscillation is favored by low odor concentrations [32] and low vapor-pressure molecules [33]. Similarly, at the M/T cell level, changes in odor responsiveness as a function of nasal flow rate may also be attributed to the level of bulbar activation. Finally, we observed that spike phase locking to LFP oscillations was also modified as a function of the flow rate condition (Fig. 3). A high flow rate improved spike phase locking to gamma, whereas a low flow rate improved spike phase locking to beta. Different authors have suggested that spike phase locking to LFP oscillations is a key parameter in olfactory coding, at least in the honeybee (for review, see [34]) and fish [35]. If, as we report, an odorant can evoke different phase locking and LFP oscillatory patterns according to nasal flow rate and the animal's sniffing dynamics, then such hypotheses should be reformulated.
Our finding that OB activity is modified by nasal airflow variation raises a new question: Do the observed modifications simply reflect the differential ORN activation due to air pressure variation, or are they due to variations in odorant stimulation because of flow rate variations? Unfortunately, odors cannot be delivered without airflow into the nasal cavity, making it impossible to directly answer this question. Nevertheless, some clues can be found in our data. First, we observed that M/T cells, in which the response to odors changed with flow rate, were not systematically those cells whose spontaneous activity was modified by flow rate ( Figure S1). Second, slow LFP modulation was not modified by flow rate when no odor was delivered. Third, even with flow rate modification, gamma or beta oscillations were never evoked without odor. It thus seems reasonable to assume that the modifications in odor response we observed were not the only consequence of air pressure variation, but rather such modifications are likely due to both variations in nasal airflow pressure and odorant stimulation. If nasal flow rate variation affects odorant stimulation, it can do it in different ways: first by changing odorant concentration and second by modifying odorant migration through the nasal cavity. Further studies will next be required to specify the respective implication of each of these parameters by using both a systematic panel of odorant concentrations and a panel of odorants with different physico-chemical properties.
Functional implications for odor perception
By demonstrating that nasal flow rate induced modifications in bulbar network activity, M/T single cell activity and spike phase locking to LFP, we have presented novel evidence showing that nasal airflow is a key parameter to consider when studying olfactory coding. This result leads to two questions. First, what allows perception stability? Indeed, it has been recently shown that even though glomerular maps are modified by odorant concentration, animal odor perception remains stable [36]. An interesting explanation can be found in Bathellier et al. [37]. They showed that relevant information for odor coding is contained in mitral cell ensemble activity which is robust to changes in sniffing frequency. We could consider the possibility that such a coding scheme should also be robust to changes in flow rate variations.
Second, what is the role of respiratory dynamics in perception? Different authors have considered respiratory dynamics not only as the odorant molecules vector but also as an integral part of the olfactory percept [38,39]. As demonstrated in humans, there are fast adjustments of sniff volume depending on odorant concentration [40]. These fast modulations of respiratory dynamics suggest that olfactomotor control could be similar to that of vision or audition [40]. Variations in sniff parameters would serve to optimize the transport of odorant molecules along the OE pathway [41,42], similar to how eye movements serve to allow acute visual perception [43].
Preparation and recording
Male Wistar rats (200-450 g) obtained from Janvier (Le Genest-Saint-Isle, France) were anesthetized with urethane (1.5 mg/kg, i.p., with additional supplements as needed) and placed in a stereotaxic apparatus. LFP oscillations were used to monitor anesthesia depth. Animals were placed on a heating pad to maintain constant body temperature.
Ethics Statement. All surgical procedures were conducted in strict accordance with the European Community Council directive of November 24, 1986 (86/609/EEC), those of the French Ethical Committee and French Legislation and received approval from the Lyon 1 University Ethics Committee (Direction of veterinary service # 69387473).
Tracheotomy. Once all pain reflexes were abolished, a tracheotomy was performed by inserting a first cannula into the trachea, which allowed the rat to breathe freely (catheter Biotrol, int. 1.57 mm, ext. 2.08 mm, tracheal cannula in Fig. 6). A second cannula was then inserted rostrally through the larynx to the postnasal cavity to allow air to be pushed and pulled through the nasal cavity (catheter Vygon, Venolux 247, int. 0.8 mm, nasal cannula in Fig. 6). . Paired t-test, *p,0.05. A2) Mean instantaneous frequency (6 SEM) as a function of respiratory cycle under the three flow rate conditions (low flow rate: orange, basal flow rate: black, and high flow rate: purple). B1) Matrix representing respiration-related spontaneous patterns of each cell recorded under the three flow rate conditions. Each line represents a flow rate condition, and each column represents a unit. A color was attributed to each pattern: excitatory synchronized (S+, red), suppressive synchronized (S2, blue), complex synchronized (Sc, green), respiration non-related (NS, gray) and null activity (NULL, light gray). Cells are ordered according to their respiratory-related pattern at basal flow rate. The percent of change in activity pattern represents the percent of units exhibiting a different pattern under the two flow rate conditions B2). Percentages of each spontaneous activity pattern as a function of flow rate condition (n = 36). Low and high nasal flow rate conditions were compared to the basal flow rate. Statistical test: Chi 2 . doi:10.1371/journal.pone.0016445.g005 Electrophysiological recordings. The dorsal region of the OB was exposed. Bulbar activity was recorded as a broadband signal (0.1 Hz to 5 kHz) using 16-channel silicon probes (NeuroNexus Technologies, Ann Arbor, MI) with a homemade 16-channel DC amplifier. Data were digitally sampled at 10 kHz and acquired with a PC using an IOTech acquisition system (Wavebook, IOTech Inc., Cleveland, OH). Probes were placed in the lateral or medial part of the OB at such a depth that the maximum number of channels could be located within, or close to, the mitral cell layer. The mitral cell layer was located by a set of criteria: LFP waveform, magnitude of unit action potentials and the inability to record spikes from the granule cell layer. Recordings were performed in the whole antero-posterior axis of the OB.
Odors. Odors (Sigma Aldrich, Fluka) were delivered in a randomized series through a dilution olfactometer (400 ml/min). The odors were isoamyl acetate (ISO), 2-heptanone (K07), 1decanol (A10) and p-Cymen (CYM). We chose to stimulate animals with several odors in order to increase the probability of cell responses. All odors were delivered in front of the animal's nose at a fraction of 18.10 22 of the saturated vapor pressure. The time delay between each odor presentation was at least 1 min. The recording protocol was as follows: 5 s of spontaneous activity, 5 s of odor-evoked activity and 5 s of post-stimulus activity.
Protocols for simulated airflow variations. Airflow was measured by fast response time airflow sensors (bidirectional micro bridge mass airflow sensor, AWM 2000 series, HoneywellH). This setup is extensively described in [44]. We used two sensors (see Fig. 6), one placed in front of the tracheal cannula to measure the animal's respiration and another placed at the entry of the nostril to measure airflow circulating through the nasal cavity.
To simulate respiratory cycles, we used a homemade apparatus that allows the reproduction of both inhalation and exhalation phases. The animal's respiratory signal (collected at the tracheal cannula) was sent to the respiratory signal simulator, which in turn sent a simulated airflow toward the nasal cavity through the nasal cannula (see Fig. 6). Imposed airflow through the nasal cavity was deodorized by passing through an activated carbon filter device (Carbon-Cap tm 75, Whatman filter, Bioblock). The whole system was based on a proportional-integral-derivative (PID) controller Figure 6. Respiratory cycle simulator. This apparatus consists of an electronic part (PID card, airflow sensor and proportional valve), an acquisition card (NI-6008 analog input/output) and a PC (software under LabViewH). Using a constant aspiration and a compressed air input passing through the system allowed the simulation of both inhalation and exhalation phases of the respiratory cycle. The simulated airflow was directly sent to the nasal cavity via a nasal cannula. The tracheal cannula allowed the animal to freely breathe and was used to acquire the animal's respiratory signal. doi:10.1371/journal.pone.0016445.g006 composed of an electronic card, a miniature proportional valve (VSOH model, Parker Precision Fluidics) and an air flow sensor (micro bridge mass airflow sensor, AWM 40000 series, Honey-wellH). The electronic card was controlled via an analog data acquisition card (National InstrumentsH NI-USB 6000 series). This card was connected to a PC, and software developed under LabViewH was designed to perform several tasks. The user was able to choose the simulated airflow parameters (time course and rate). To maintain conditions as physiological as possible, the simulated airflow was synchronized to the tracheal respiratory signal. In some cases, animal respiration was too irregular to permit the device to keep the synchrony between both signals (i.e., the simulated nasal airflow and animal respiration). We thus discarded trials in which both signals were not synchronized. To estimate the synchronization between simulated nasal airflow and animal respiration, we measured the delay between both signals at the inspiration/expiration transition (I/E) points. We considered both signals to be synchronized when the shift was ,10% of the respiratory cycle. In selected trials, the shift was an average 4.63% (63.12%) of the respiratory cycle.
Since the aim of this work was to study the influence of airflow variation, we chose to impose three different nasal airflow rates: low (200 ml/min), basal (500 ml/min) and high (800 ml/min) without varying any other parameter of the respiratory signal. These values refer to the maximum flow rate for the respiratory cycle. A 500 ml/min flow rate was chosen as basal because it reproduced the bulbar LFP signal in response to ISO that is typically recorded under the anesthetized non-tracheotomized condition [21]. This basal flow rate was adjusted to 500 ml/min 650 ml/min depending on the animal state. The three flow rates employed corresponded to the physiological scale in rat measured in behaving animals [1]. To determine how nasal airflow itself could affect olfactory bulb activity, we tested the effect of nasal airflow presence (nasal airflow ON, 500 ml/min, without any odor, modulated at animal respiratory frequency) and absence (nasal airflow OFF) on olfactory bulb activity. In all of these trials, airflow was discontinuous and exactly reproduced animal's breathing frequency. To ensure us that imposed airflow in the nasal cavity was completely deodorized, we included a blank trial in each recording session. We never observed any bulbar response (neither LFP nor mitral cell activity) to the blank. We also tested, on 3 additional rats, the effect of a continuous nasal airflow at 200, 500 or 800 ml/min.
Data processing
All data processing was performed using Openelectrophy openaccess homemade software [45].
Respiratory signal. An important feature of the olfactory signal is its temporal correlation with breathing. We developed a method to represent data as a function of the respiratory phase [44]. Briefly, the recorded respiratory signal was processed to extract each respiratory epoch. The time component of these periods was then converted into a circular phase component defined between 0 and 1, which represented the beginning of the inspiration and end of the expiration, respectively. As a result, electrophysiological signals were no longer represented as a function of time but as a function of respiratory phase. The main advantage of this method is that the phase representation was common to all trials, as opposed to time representation. Electrophysiological recordings were analyzed relative to the respiratory cycle and in particular to the transition points between inhalation and exhalation (I/E). I/E points were automatically detected as zero-crossings of the respiratory signal, corresponding to the point of null airflow of the rising phase.
LFPs. Wavelet transform LFPs were obtained by band-pass filtering the recorded signal at 5-200 Hz. To preserve both time and frequency information, we used a time-frequency representation based on the continuous wavelet transform method.
Wavelet ridge extraction: We previously developed an algorithm [46] to extract phase information from the identified oscillations in the signal. Briefly, for each frequency band of interest, we computed the mean and standard deviation (SD) of the time-frequency map and defined the threshold as the mean +5 SDs of the time-frequency amplitude contained in the pre-stimulus period (between 0 and 5 s). These thresholds were used to define time and frequency boxes centered on points of maximum amplitude in the signal, bounded by small time and frequency ranges. Finally, we ran a high resolution Morlet's complex wavelet transformation on each box. Ridges were extracted on a Morlet scalogram time frequency map. Each time frequency ridge line represented all parameters of one oscillation (phase, frequency, amplitude, starting and ending times) as a function of time.
Spikes. Spike sorting: Signals from individual electrodes were amplified (gain 10006) and filtered from 300 to 5000 Hz. Multiunit activity consisted of a few neurons on each electrode. We chose to use only the well-discriminated units (with a signal-tonoise ratio $5:1) and to sort cells according to their spike amplitudes. We verified that all sorted cells exhibited a minimal 4 ms refractory period. Consequently, the number of units retained for analysis was restricted to 1-3 units per channel. We preferred to use a very strict procedure, which resulted in a limited number of units but was also very safe. With this conservative procedure, we were very confident in the quality of the sorting; all units were well isolated, and there were no duplicates.
Respiratory patterns: As previously described [47], M/T cell activity is well characterized as a function of its temporal pattern along the respiratory cycle. To evaluate such patterns, the time occurrence of each spike was converted into a respiratory phase (0-1). All data were plotted as histograms (divided into 20 bins) that represented the spike rate along the respiratory cycle. Histograms were classified into different types based on the classification described in [47]. For the present study, we reduced this classification to four types: i) non-synchronized patterns (NS), characterized by a uniform distribution of spiking activity along the respiratory cycle; ii) excitatory-simple-synchronized patterns (S+), presenting a single increase in firing activity along the respiratory cycle; iii) suppressive-simple-synchronized patterns (S2), presenting a single decrease or stop in firing activity along the respiratory cycle; and iv) complex-synchronized patterns (Sc), exhibiting multiple firing frequency changes along the respiratory cycle. A period with no or very few spikes during the considered epoch was classified as NULL.
Spike-LFP phase coupling. Our wavelet ridge-based analysis method of LFPs allowed an accurate estimation of the oscillation phase. An absolute phase was assigned to each action potential that occurred during an oscillatory epoch. The mean distribution of action potentials relative to the phase of the LFP oscillation (beta or gamma) was represented by phase histograms (23 bins), where the peak and trough of the wave were assigned to 0 and pi, respectively. From these histograms, circular mean, deviation and dispersion were calculated.
Statistics. Statistical tests were performed using Excel, Statview software or R combined with Python script. The level of significance was set at p,0.05 for all statistical tests (p,0.05 *, p,0.01 ** and p,0.001 ***).
LFP: LFP spontaneous slow modulation was calculated from the difference between the maximum and minimum points on each averaged LFP signal relative to the respiratory cycle. For fast LFP oscillations, average duration, frequency, amplitude and number of oscillatory bursts (defined as the mean number of detected oscillatory bursts per electrode under each flow rate condition) in the OB were compared between flow rate conditions using the Wilcoxon paired test. For all analyses, oscillation characteristics at basal flow rate were taken as the reference point for comparisons.
Spikes: First, M/T cell instantaneous frequency discharges were compared between the three flow rate conditions using a paired ttest. For each cell, instantaneous frequency discharge at the basal flow rate was taken as the reference point for comparisons. Second, distribution of M/T cell instantaneous frequency along the respiratory cycle was compared between the three flow rate conditions using the Equal Kappa test. Third, the probability of M/T cell respiration-related patterns were compared between flow rate conditions during spontaneous and odor-evoked activities using a Chi 2 test. M/T cell activity patterns were also compared in nasal airflow ON vs. OFF conditions using a Chi 2 test.
Spikes-LFP phase coupling: Rayleigh's uniformity test was used to calculate the probability that the spikes were uniformly distributed throughout the entire duration of an oscillatory cycle (null hypothesis). The circular Equal Kappa test was used to test the difference in spike distribution relative to oscillation cycle between the nasal flow rate conditions. Figure S1 Matrices of spontaneous and odor-evoked activities of OB units. A) Matrix representing respirationrelated spontaneous and odor-evoked patterns of each cell recorded under the three flow rate conditions. Each line represents a flow rate condition, and each column represents a unit. A color was attributed to each pattern: excitatory synchronized (S+, red), suppressive synchronized (S-, blue), complex synchronized (Sc, green), respiration non-related (NS, gray) and null activity (NULL, light gray), ordered by cell. B) Matrix representing spontaneous and odor-evoked instantaneous frequency of each cell recorded under the three flow rate conditions. Each line represents a flow rate condition, and each column represents a unit. Gray scale was used to represent firing rate from ,20 Hz to .80Hz, ordered by cell. (TIF) Figure S2 Continuous nasal airflow does not induce respiratory modulation. Example of LFP signal recorded in different airflow conditions from left to right: nasal airflow OFF, continuous 200 ml/min, continuous 500 ml/min and continuous 800 ml/min. LFP signals are averaged over the respiratory cycle. Gray traces: averaged respiratory cycle measured at the tracheal cannula. | 2014-10-01T00:00:00.000Z | 2011-01-26T00:00:00.000 | {
"year": 2011,
"sha1": "2ffc0e31c6db987b78f891d81dd0b1f306376dbf",
"oa_license": "CCBY",
"oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0016445&type=printable",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "2ffc0e31c6db987b78f891d81dd0b1f306376dbf",
"s2fieldsofstudy": [
"Biology",
"Environmental Science"
],
"extfieldsofstudy": [
"Medicine"
]
} |
12402730 | pes2o/s2orc | v3-fos-license | Development of a Universal Nutritional Screening Platform for Plastic Surgery Patients
Background: Plastic surgeons routinely see patients with complex or chronic wounds of all etiology. In a previous study, we found that up to 1 in 4 of these patients is at risk for malnutrition, which may be influencing their ability to heal. The goal of this study was to develop and validate a universal screening protocol that would be fast and accurate and allow for effective intervention and optimization of nutrition before plastic surgery. Methods: To accomplish these goals, we adopted a 2-part screening algorithm using the Canadian Nutritional Screening Tool (CNST) to triage patients in our outpatient clinics and then further screened those identified as being at risk using the Subjective Global Assessment (SGA) tool and blood work. Results: We screened 111 patients with diagnoses related to breast cancer (n = 10; 9.01%), elective surgery (n = 38; 34.23%), emergency surgery (n = 8; 7.21%), fractures (n = 15; 13.51%), and wounds (n = 40; 36.04%). Of the screened subjects, 15.32% (n = 17) were found to be at nutritional risk using the CNST, and 13 were confirmed to be moderately or severely malnourished using the SGA. Importantly, there were no positive correlations between nutritional status and smoking, diabetes, body mass index, or age, indicating that a universal screening protocol is needed to effectively screen a diverse plastic surgery population for malnutrition. Conclusions: Screening patients with both the CNST and the SGA is an effective way to identify patients before surgery to improve outcomes.
INTRODUCTION
Malnutrition is a well-recognized issue in the acute care setting with a prevalence ranging from 30% to 45%, and elderly individuals are recognized to be most at risk. [1][2][3] The consequence of malnutrition is impaired wound healing, decreased cardiac and respiratory functions, and negative health outcomes that often result in prolonged length of hospital stays. [4][5][6][7][8][9] In fact, it is estimated that malnourished patients account for approximately 27% of surgical patients admitted to hospitals. 10 Identification of nutritional deficiency, especially within the surgical field, is crucial to improve outcomes, 10 as malnourished patients are 60% more likely to be read-mitted to hospital within 15 days and have increased mortality (50% versus 10% for well-nourished patients) in the first 3 years postdischarge. 11 Significantly higher in-hospital costs have also been reported for malnourished patients. 11,12 Because nutritional status is so important to outcomes, a number of different tools have been developed for universal screening purposes (the Mini Nutritional Assessment for the elderly, the Malnutrition Universal Screening Tool (MUST), and the Nutritional Risk Score among others 13 ). However, these tools range in complexity and length, and there is currently no consensus in the Canadian Healthcare system as to the best screening tool for routine use. We recently used the Canadian Nutritional Screening Tool (CNST) to profile the nutritional status of the plastic surgery outpatient population at a busy urban tertiary care centre in Toronto, ON. 14 In that study, we found that almost 1 in 4 of our patients were at nutritional risk, cementing the need for universal nutritional screening in our clinical practice. In that study, we found the CNST to be an easily implemented rapid screen for nutritional risk with the potential for universal implementation. How- Background: Plastic surgeons routinely see patients with complex or chronic wounds of all etiology. In a previous study, we found that up to 1 in 4 of these patients is at risk for malnutrition, which may be influencing their ability to heal. The goal of this study was to develop and validate a universal screening protocol that would be fast and accurate and allow for effective intervention and optimization of nutrition before plastic surgery. Methods: To accomplish these goals, we adopted a 2-part screening algorithm using the Canadian Nutritional Screening Tool (CNST) to triage patients in our outpatient clinics and then further screened those identified as being at risk using the Subjective Global Assessment (SGA) tool and blood work. Results: We screened 111 patients with diagnoses related to breast cancer (n = 10; 9.01%), elective surgery (n = 38; 34.23%), emergency surgery (n = 8; 7.21%), fractures (n = 15; 13.51%), and wounds (n = 40; 36.04%). Of the screened subjects, 15.32% (n = 17) were found to be at nutritional risk using the CNST, and 13 were confirmed to be moderately or severely malnourished using the SGA. Importantly, there were no positive correlations between nutritional status and smoking, diabetes, body mass index, or age, indicating that a universal screening protocol is needed to effectively screen a diverse plastic surgery population for malnutrition. Conclusions: Screening patients with both the CNST and the SGA is an effective way to identify patients before surgery to improve outcomes. ever, we were concerned that this tool was not capturing enough information about a patient's nutritional status if they were found to be at risk. As a next step, we have now created a routine nutritional screening algorithm based on the Integrated Nutrition Pathway for Acute Care and recommendations of the American Society for Parenteral and Enteral Nutrition. 1,15 We now report the screening of an additional 111 patients by triaging first with the CNST and secondarily (in those at risk by CNST screen) with the more comprehensive Subjective Global Assessment (SGA) tool. The results of our SGA-based screening support the findings of the CNST, suggesting that universal screening using this easily implemented 2-part screening system would accurately identify those patients at nutritional in our plastic surgery population and allow us to improve outcomes through optimization of nutritional status.
Setting
This cross-sectional study was approved by the St. Michael's Hospital Research Ethics Board (REB16-266) before initiation of recruitment. Enrolment took place at St Michael's Hospital, Toronto, from October until December 2016. Eligibility criteria included (1) referral to a plastic surgeon for assessment of a wound or other surgical concern, (2) ability and willingness to consent, and (3) age above 18 years. Informed and voluntary consent was obtained for all participants.
Survey Instruments
The CNST is a 2-question validated nutritional screening tool, designed to allow rapid nutritional screening by nonspecialists. 16 The CNST has been shown to have good sensitivity (91.7%) and specificity (74.8%) in Canadian adult inpatients. 16 The SGA, our selected assessment tool, includes a history and physical examination and has a nonnumerical weighing scheme prioritizing certain elements like percentage weight loss and obvious physical signs of malnutrition. 17 The SGA has been known for its relevance for predicting clinical outcome and previous use in surgical populations. 13,18
Data Collection and Reporting
Data collection took place in the St Michael's Hospital outpatient clinic. Two trained research assistants recruited patients to the study, obtained their informed, voluntary, and written consent, and collected the first nutritional screening data using the CNST. For purpose of standardization, 1 research assistant trained with a dietician and used the SGA to screen selected patients identified as being at nutritional risk by the CNST as per our study algorithm (Fig. 1). The algorithm selected for this study follows the recommendations developed from the Integrated Nutrition Pathway for Acute Care in that only patients identified at nutritional risk were further assessed with the SGA. 1 Participants identified as being at nutritional risk were then invited to be evaluated with the SGA 17 and were given blood work requisitions and asked to undergo micronutrient analysis (vitamin D, E, C, A, B 12 , selenium, zinc, rbc, folate, ferritin, iron stores, and prealbumin), and referred to a hospital dietician. The primary objective of this study was to validate the diagnostic accuracy of the CNST in our patient population using the SGA. The secondary objective was to describe the characteristics of the plastic surgery patient population identified as being at nutritional risk. In that regard, categorical and continuous variables were presented as means (%).
Accuracy of the CNST
We adopted a 2-part screening protocol for this study, using the CNST to triage patients and the SGA to validate outcome. A total of 111 patients were surveyed with this protocol over a period of 3 months, with an average age of 49 years and average body mass index (BMI) of 26.75 kg/m 2 ( Table 1). There were a total of 55 men and 56 women. Of the enrolled participants, 71 (63.97%) had a surgical diagnosis and had undergone or were planning to undergo a surgical procedure. The remaining 40 (36.04%) patients were seeing a plastic surgeon because of a subacute or chronic wound diagnosis. This patient population differed from that in our previous study, 14 which was more heavily weighted with diabetic foot wounds. Instead, this study focused on including a more medically diverse patient population.
The CNST identified 17 (15.32%) patients from our sample of 111 at nutritional risk ( Table 2). The SGA was performed on 16 of these patients, with 1 refusal. The SGA confirmed the nutritional diagnosis in 13/16 patients: 10 were found to be moderately malnourished (SGA class B) and 3 were found to be severely malnourished (SGA class C; Table 3). These results suggest that the CNST overpredicts nutritional risk in 18.75% of cases but is generally an accurate predictor of macronutrient deficiency in a diverse plastic surgery patient population.
Malnourished Population
The patients identified as at risk for malnutrition by the CNST ranged in age from 24 to 86 years (mean, 46.2 years) and were equally split between genders (8 males and 9 females). The average BMI of at risk individuals was 21.37 (± 4.76). Of note, 3 individuals had BMIs of > 30 (classified as obese), 2 patients had BMIs of 25-30 (classified as overweight), and the remainder had BMIs in the 18.5-24.99 range (classified as normal; Table 3). This result suggests that preconceived notions as to how a malnourished individual looks should be challenged-these patients were not underweight (and some were in fact obese). Also of note, only 2 of our at-risk patients were diabetics, and there was no correlation between smoking status and malnutrition.
We also examined our at-risk population by subcategorizing patients according to clinical diagnosis. Patients were categorized as breast (including oncologic reconstructions, reductions, gynecomastia), elective (including trigger fingers, skin lesions, scar revisions), emergency (including traumatic lacerations, amputations, foreign bodies), fractures (including craniofacial and hands), and wounds (including pressure sores, subacute to chronic posttraumatic, subacute to chronic postsurgical). The number of patients per subcategory identified by the Table 1
Characteristics Study Population*, n (%) Wound Population †, n (%) Surgical Population ‡, n (%)
Age ( CNST as being at nutritional risk was variable (Table 1), with surprisingly few patients at nutritional risk in the wound cohort (4/40; 10%). Of those wound patients, only 1 was a diabetic. Importantly, examining the breakdown of patients according to clinical diagnosis suggests that 14 patients were seeing the plastic surgeon for wounds or elective surgeries could have benefitted from a nutritional program leading up to their clinic visit, and 3 could have benefitted from optimized nutrition following trauma.
Micronutrient Analysis
The CNST and SGA both screen for macronutrient (protein, carbohydrate, and fats) deficiencies, but wound healing is also affected by deficiencies in micronutrients (vitamins and minerals). Of the patients with confirmed macronutrient deficiencies, we found 1 patient who also had iron deficient anemia (iron low (3 µmol/L), RBC low (3.36 × 10 12 / L), Hg low (103 g/L), vitamin A low (0.8 µmol/L), 1 patient with selenium borderline low (1.3 µmol/L), and low vitamin C (17 µmol/L), and a final patient with low vitamin C (15 µmol/L). We also identified a patient who is likely oversupplementing their diet, as they had elevated levels of folate and vitamins E, A, and B 12 . We unfortunately had a response rate of only 8/16 patients given requisitions for blood work, and therefore the number of patients with micronutrient deficiencies may have been higher. Regardless, these results suggest that patients identified as being at risk for macronutrient deficiencies are also at risk for micronutrient deficiencies and should be screened accordingly.
Age
Nutritional status often declines with age, 19,20 and we therefore also subdivided our patient population by age (< 50 years and ≥ 50 years). We found no age-related increase in nutritional risk: the frequency of malnutrition in patients < 50 years was 16.1% (14/87 patients) and in those ≥ 50 years 12.5% (3/24 patients). Interestingly, the 3 patients ≥ 50 years identified at nutritional risk were all seeing the plastic surgeon as a result of pressure sores on the sacrum, ischium, or coccyx. Although this sample size is too small to evaluate the significance of this finding, it is of note that other studies have shown that nutritional variables can predict the development of pressure sores. 21 Albumin levels were one of the most significant independent predictors of poor wound healing outcomes in chronic wounds. 4,22 However, we did not identify any individuals who had albumin levels outside the normal range (3.0-51.0 g/L).
DISCUSSION
Patients being referred to the plastic and reconstructive surgery outpatient clinics present with diverse diagnoses. It is well established that optimizing perioperative patients' healing involves addressing nutritional deficiencies, but there is little consensus as to what screening tool should be used for routine assessment. Our study was designed to follow up on previous work indicating that we have high rates of malnutrition in our plastic surgery population, with the goal of defining a nutritional screening system for routine use in all our outpatients. To be of use for universal screening, our tool must be (1) rapid to administer, translating into universal implementation and (2) accurate, such that few false-positives/negatives would waste clinical resources. Given the success we had using the CNST in our first study, we triaged patients in the current study using the CNST first and then classified those at nutritional risk more thoroughly using a secondary tool (the SGA) and micronutrient analysis. Such a 2-part screen is supported by the American Society for Parenteral and Enteral Nutrition 15 and offers both good screening accuracy and ease of implementation. Our experience in both our previous study and the current study suggests that the CNST can be completed in less than 1 minute and therefore fulfills our first criteria. The SGA has been previously validated in surgical populations 13,18 and in our hands confirmed the CNST's findings in approximately 80% of cases (satisfying the second criteria). Adding blood work for micronutrient analysis to the workup of patients at risk adds depth to our nutritional assessment and allows for tailored and specific optimization before surgery. Our experience in this study confirms that nutritional screening in the outpatient clinic setting is feasible, innovative, and necessary for holistic patient care.
Beyond validating the nutritional screen itself, we were interested in identifying characteristics of our patient population at nutritional risk. Of the patients participating in this study, 71 (63.96%) were in clinic for a surgical intervention and 40 (36.04%) had a wound diagnosis. The CNST identified 17 patients at nutritional risk, the majority of who were moderately or severely malnourished (10 patients SGA class B and 3 SGA class C). Most of these patients were seeing the plastic surgeon for an elective procedure, which indicates an opportunity for optimization before surgery. Importantly, only 2 patients of our entire patient cohort (1.82%) were found to be underweight (< 18.5) as per the World Health Organization's classification but were not found to be at nutritional risk. This discrepancy between BMI and nutritional status underscores the importance of universal nutritional screening in an outpatient setting, as overt signs of malnutrition are easily missed.
Even more easily missed are micronutrient imbalances that can impact cellular functions including immunity and wound healing. 23 The antioxidant vitamins A, E, K, and C have roles in eliminating prooxidants, scavenging free radicals and maintaining an essential balance between oxidative stress and antioxidants. 23,24 Vitamin C is also a catalyst required in the synthesis of collagen part of new tissue creation and wound healing processes. 23 Micronutrient deficiencies impact the elderly most often and are usually deficiencies in vitamin D, zinc, and vitamin B 12 (cobalamine) 25,26 in that population. Unfortunately, our study was limited by sample size and we did not have a sufficient response rate to compare the micronutrient status of our patients according to age. We did however identify 3 patients with deficiencies in micronutrients important to wound healing, including vitamin C. In a follow-up study, we aim to profile the micronutrient status of more patients undergoing outpatient plastic surgeries in a multicenter setting.
CONCLUSIONS
Screening all patients attending plastic surgery outpatient clinics for nutritional status has the potential to improve patient outcomes significantly. In our case, we identified 15% of our patients who were at risk for poor healing due to nutrient deficiency and could have benefitted from nutritional coaching/intervention before surgery. By screening patients with both the CNST and the SGA, we captured the nutritional status of a large number of patients and carefully profiled those at risk. This screening system will be validated in a follow-up multicenter study in a larger patient population to allow us to draw further conclusions about micronutrient deficiencies in our population. | 2018-04-03T00:34:02.832Z | 2017-07-01T00:00:00.000 | {
"year": 2017,
"sha1": "7eb1115d76e412ef8d45b9cc1248f0e89a7cac07",
"oa_license": "CCBYNCND",
"oa_url": "https://doi.org/10.1097/gox.0000000000001342",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "7eb1115d76e412ef8d45b9cc1248f0e89a7cac07",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
264609432 | pes2o/s2orc | v3-fos-license | Structural insights into TRAP association with ribosome-Sec61 complex and translocon inhibition by a CADA derivative
During cotranslational translocation, the signal peptide of a nascent chain binds Sec61 translocon to initiate protein transport through the endoplasmic reticulum (ER) membrane. Our cryo–electron microscopy structure of ribosome-Sec61 shows binding of an ordered heterotetrameric translocon-associated protein (TRAP) complex, in which TRAP-γ is anchored at two adjacent positions of 28S ribosomal RNA and interacts with ribosomal protein L38 and Sec61α/γ. Four transmembrane helices (TMHs) of TRAP-γ cluster with one C-terminal helix of each α, β, and δ subunits. The seven TMH bundle helps position a crescent-shaped trimeric TRAP-α/β/δ core in the ER lumen, facing the Sec61 channel. Further, our in vitro assay establishes the cyclotriazadisulfonamide derivative CK147 as a translocon inhibitor. A structure of ribosome-Sec61-CK147 reveals CK147 binding the channel and interacting with the plug helix from the lumenal side. The CK147 resistance mutations surround the inhibitor. These structures help in understanding the TRAP functions and provide a new Sec61 site for designing translocon inhibitors.
Cells were then fixed in PBS (Thermo Fisher Scientific) containing 1% paraformaldehyde (VWR Life Science brand) before acquisition on a BD FACS Celesta flow cytometer (BD Biosciences) with BD FACSDiva 8.0.1 software.All data were analyzed in FlowJo X (FlowJo Software), version 10.
Cell-free in vitro translation/translocation assay
For the in vitro translocation of protein, constructs were designed to hold the signal (SP) and first region of the mature protein fused to preprolactin (pPL).In particular, for CD86 the SP and 11 residues of the mature protein were fused to pPL, whereas for PTK7, CD58 and CD40 the SP and 10 residues of the mature protein were fused to pPL, The EasyXpress linear template kit (Qiagen, Hilden, Germany) was used to amplify and linearize the DNA of interest from the plasmid using PCR.PCR products were purified with the Nucleospin gel and PCR clean-up kit (Macherey Nagel, Düren, Germany) and transcribed in vitro using T7 RNA polymerase (RiboMAX system; Promega, Madison, WI, USA).RNA was purified from the sample using the Nucleospin RNA clean-up kit (Macherey Nagel).mRNA transcripts were then translated in rabbit reticulocyte lysate (RRL) (Promega) in the presence of L-35S-methionine (PerkinElmer, Waltham, MA, USA).
Translations were performed for 45 min at 30°C, in the presence or absence of ovine pancreatic rough microsomes and CK147, supplemented with RNasin (Promega).Proteinase K (PK) (Roche, Basel, Switzerland) protection assays were performed on ice for 30 minutes and quenched with PMSF (Thermo Fisher Scientific, Waltham, MA, USA).Samples were washed with low-salt buffer (80 mM KOAc, 2 mM Mg(OAc)2, 50 mM HEPES, pH 7.6) and radiolabeled proteins were isolated by centrifugation (10 minutes at 21,382 g, 4°C).The radio-labelled proteins were then separated with SDS-PAGE on a 4-12% Criterion XT Bis-Tris gels (Bio-Rad, Hercules, CA, USA) in MES buffer (Bio-Rad), detected by phosphor imaging (Cyclone Plus storage phosphor system; Perkin Elmer) and quantified using the ImageQuant software.
Movie legend
Movie S1.A 360º rendering showing the binding of CK147 (orange) to the Sec61 channel near the plug helix (pink).The experimental cryo-EM density accounting for the inhibitor CK147 is shown as the transparent gray volume.
Fig. S1 .
Fig. S1.Structure determination and analysis of ribosome-Sec61-TRAP complex.(A) A typical micrograph with particles circled.(B) A 2D class showing the projection of structural features.(C) The overall resolution of the complex is 2.86 Å at FSC 0.143.(D) Local resolution of different parts of the structure.(E) The local resolution for the TRAP (α, β, and δ) complex located in ER lumen varies between 10 -12 Å.The local resolution distribution for Sec61 (F) and TRAP-γ (G).The experimental density showing the interaction between the transmembrane helices of α and γ chains (H), β and γ chains (I), and δ and γ chains (J).(K) The relative positioning of Sec61 and heterotetrametric TRAP in the ER membrane and ER lumen.
Fig. S2 .
Fig. S2.In vitro translocation of truncated huCD4 in the presence of CK147 and sheep RMs.(A) A schematic representation of the in vitro translation/translocation assay.(B) WT huCD4 contains N-linked glycosylation sites in the extracellular immunoglobulin-like domain D3 and D4.A truncated 25 kDa variant of WT huCD4 was used as substrate for the assay, designated huCD4 D1D2, which does not contain the transmembrane region and glycosylation sites.When huCD4 translocates into the ER lumen, the SP is cleaved.(C) Same as for Fig. 3C, but with the proteinase K (PK) treatment shown.Protein species that are not translocated, are degraded by the proteinase.As a result, only the translocated, SP-cleaved huCD4 protein is detected on the gel (right side of the vertical line).The open arrowhead represents the translocated, and thus SP-cleaved, mature huCD4 protein fraction.The solid arrowhead represents the un-cleaved huCD4 preprotein.
Fig. S3 .
Fig. S3.Comparison of CK147 with CADA on protein down-modulation of substrates in transfected HEK293T cells.Same as for Fig 3B.Four-parameter concentration-response curves were fitted to the data from replicate experiments.Values are mean ± SD; n ≥ 3.
Fig. S4 .
Fig. S4.Comparison of CK147 with CADA on protein down-modulation of substrates in CD4 + MT-4 cells.Same as for Fig 3B, but in MT-4 cells.Cells were treated with CK147 or CADA for 24h, collected and stained with specific antibodies to measure the cell surface expression of the respective receptors with flow cytometry.On average, 20,000 cells were analyzed for receptor expression to calculate the mean fluorescence intensity.Four-parameter concentration-response curves were fitted to the data from replicate experiments.Values are mean ± SD; n ≥ 3.
Fig. S5 .
Fig. S5.In vitro protein translocation in the presence of CK147 and sheep RMs.(A) Same as in figure 3C.Radioblot of in vitro translated and translocated substrates in the absence or presence of 10 µM CK147.The open arrowhead represents the translocated mature protein fraction.The solid arrowhead represents the preprotein.Note that the mature CD86 protein is glycosylated.(B) Quantification of (A).Bars represent the translocation efficiency of the 10 µM CK147-treated samples normalized to the DMSO control.Values are mean ± SD; n ≥ 2.
Fig. S6 .
Fig. S6.Distribution of the point mutations in the sequencing reads of clone 16.The three Sec61α mutations of clone 16 consists of a population of reads with 1 of each mutation or a combination of all 3.Each of the mutations were detected in approximately 50% of the reads (R66G in 49% of all reads, P83H in 54%, and V102I in 46%).All sequenced fragments of Sec61α in our study had a coverage of > 500,000X.
Fig
Fig S7.Concentration-response curves for huCD4 expression in Sec61α single mutants treated with CADA.Same as in Figure 3G, but for a concentration range of CADA.Cells were co-transfected with huCD4 and WT or mutant Sec61α and treated with CADA for 42h.Fourparameter concentration-response curves were fitted to the data from three replicate experiments performed on different days.Values are mean ± SD; n ≥ 3.
Fig. S8 .
Fig. S8.TRAP-a interacts with TMH5 and TMH6 region of Sec61a in the ribosome-Sec61-TRAP structure.A cartoon (A) and molecular surface (B) representations showing the interaction between Sec61a (green) and TRAP-a (pink).
Fig. S9 .
Fig. S9.Wall-eyed stereo showing the fitting of Sec61 primed conformation (grey) and Sec61 CK147-bound conformation (cyan protein and orange CK147) to the experimental density obtained from a sample in which CK147 was added to pre-assembled ribosome-Sec61complex.No density for CK147 and the positions of TMH2, TMH3, and TMH7 confirmed that Sec61 exists in primed state with no bound inhibitor.The map is displayed at 1.9s.The data processed to 3.3 Å resolution density map, however, the local resolution of the map section covering Sec61 is lower.
Fig. S10 .
Fig. S10.Analysis of the ribosome-Sec61-CK147 structure.(A) The experimental density covering the translocon region of the ribosome (gray) -Sec61 (green) -CK147 structure; the weaker density for TRAP-γ (blue) compared to that in the apo structure suggest that the binding of the translocon inhibitor impacts the positioning of the TRAP complex.(B) The local resolution distribution of the ribosome-Sec61-CK147 structure.(C) The local resolution distribution for Sec61 varies between 3 -7Å with the map resolution at the CK147 is ~5Å.(D) A 2D projection using Ligplot shows the interactions of CK147 with the surrounding residues.(E) The number of contacts of CK147 with Sec61α residues calculated at a maximum cutoff of 4.2 Å.The x-axis indicates the residue number and the y-axis the number of contacts.
Fig. S11 .
Fig. S11.Single-particle cryo-EM data processing flowchart for obtaining the density map for the structure of ribosome-sec61-TRAP complex.
Fig. S12 .
Fig. S12.Single-particle cryo-EM data processing flowchart for obtaining the density map for the structure of ribosome-sec61-CK147 complex. | 2023-03-05T05:08:26.171Z | 2023-03-01T00:00:00.000 | {
"year": 2023,
"sha1": "bea9a72e8d093b65a97930021d34253955be3a75",
"oa_license": "CCBYNC",
"oa_url": "https://www.science.org/doi/pdf/10.1126/sciadv.adf0797?download=true",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "bea9a72e8d093b65a97930021d34253955be3a75",
"s2fieldsofstudy": [
"Chemistry",
"Biology"
],
"extfieldsofstudy": [
"Medicine"
]
} |
42728663 | pes2o/s2orc | v3-fos-license | Early diagnosis of colorectal cancer
In 1974, almost 100,000 Americans will develop cancer of the colon and rec tum and approximately 48,000 will die of this disease. Only lung cancer kills more men and only breast cancer kills more women. As in all cancers, early diagnosis is the key to higher cure rates and lower mortality. Unlike many cancers, how ever, the detection and localization of colorectal cancer at an early and even asymptomatic stage is possible with the tools already available: digital rectal ex amination, proctosigmoidoscopy and/or colonoscopy, barium enema with air contrast studies and guaiac testing for occult blood. These modalities com bined with a high index of suspicion can, if applied judiciously, detect cancer at an early stage and achieve cure rates of up to 75-80 percent.
In 1974, almost 100,000 Americans will develop cancer of the colon and rec tum and approximately 48,000 will die of this disease. Only lung cancer kills more men and only breast cancer kills more women.
As in all cancers, early diagnosis is the key to higher cure rates and lower mortality. Unlike many cancers, how ever, the detection and localization of colorectal cancer at an early and even asymptomatic stage is possible with the tools already available: digital rectal ex amination, proctosigmoidoscopy and/or colonoscopy, barium enema with air contrast studies and guaiac testing for occult blood. These modalities com bined with a high index of suspicion can, if applied judiciously, detect cancer at an early stage and achieve cure rates of up to 75-80 percent.
Clinical Manifestations
While detection of asymptomatic co lorectal cancer is the ideal, in reality clinical manifestations most often alert the clinician to the possibility of cancer. (Table.)The signs and symptoms of colorectal cancer differ depending on the location, size and stage of the disease but generally include abdominal pain or discomfort, blood in or on the stools and change in bowel habits.
Abdominal pain is the most common symptom of colon cancer. In right colon cancer, pain is vague, dull and annoying and may be confused with gall badder disease or peptic ulcer.The charac teristic feature of the pain is that it is un characteristic. In left colon cancer, pain is usually secondary to obstruction of the colon and produces what the patient often refers to as â€oe¿ gas― or â€oe¿ cramps.― In patients with rectal cancer, pain is a late manifestation.
Rectal bleeding is another significant finding. In right colon cancer, blood is usually dark or mahogany red and mixed in the stools, while in lesions of the left colon and rectum, it is a brighter red and coats the surface of the stool. The pas sage of bloody mucous usually indicates cancer, polyps or, less commonly, ul cerative colitis.
Other clinical manifestations of right colon cancer include anemiaâ€"secon dary to bleedingâ€"and a palpable mass in the right lower quadrant of the abdo men. In left colon cancer, pain and bleeding are accompanied by a decrease in the caliber of stools, change in bowel habits and increased use of laxatives. A common complication of left colon cancer in the adult is acute large bowel obstruction, usually in the region of the sigmoid or rectosigmoid although other sites may be involved. These patients present with increasing abdominal dis tension, abdominal pain. vomiting and constipation progressing to obstipation. Cancer of the cecum or ileocecal valve must always be considered in older pa tients with symptoms of appendicitis or .@ low small bowel obstruction; emer gency barium enema is indicated in these cases. ( Figure.) Signs and symptoms of rectal cancer, other than bright red bleeding, include a sense of incomplete evacuation and ten esmus. Cancer of the anus also produces bright red bleeding and local discomfort and may be confused with hemorrhoids. Never ascribe rectal bleeding in adults to hemorrhoidsâ€"even in those patients with a past historyâ€"without first search ing for additional lesions by proctosig moidoscopy and/or colonoscopy. Enlarged inguinal nodes, secondary to me tastases, may also be found in patients with anal cancer. t-:t Patients with suspicious findings must be thoroughly examined for signs of me tastatic cancer. These include (I ) supra clavicular adenopathy (Virchow's nodes); (2) an enlarged nodular liver; (3) ascites; (4) Blumer's rectal shelf; (5) bi lateral ovarian tumors (Krukenberg tumors); and (6) umbilical metastases.
Diagnostic Methods
Since some rectal lesions are within reach of the examining finger, digital rectal examination, including inspection of the anal area, isessential for diagnosis. Approximately 12-15percent ofall co lorectal cancers may be detected by this technique. To avoid overlooking palpa ble tumors, the examination should not be confined to the area of the prostate but must include the entire circumference of the anorectal segment. Asking the pa tient to strain mildly will cause some Proctosigmoidoscopy is also a vital diagnostic tool which can detect 60-65 percent of colorectal cancers if the ter minal 25cm. of the large bowel is visual ized. It is recommended annually in asymptomatic patients over 45 years of age and is mandatory in patients with co lorectal symptoms. Bolt notes that in a reviewof 18,335proctosigmoidos copies inasymptomatic patients, un suspected cancers were consistently found in approximately one out of every 435 individuals. He also reports a five year cure rate as high as 88 percent in asymptomatic patients, compared to 50 percent for those with symptoms. 4 In an other study, Gilbertsen points out that the removal of all benign polyps and adenomatouslesions detected by procto sigmoidoscopy sharply reduces thein cidence ofrectal cancer.@ Despite current misconceptions. a proctosigmoidoscopy canbequickly and safely performed with relative ease, evenby physicians withlittle special izedtraining. Since1948more than75,-000 annual proctosigmoidoscopies have beenperformedby a widevariety of physicians at the Cancer Detection StudyoftheUniversity ofMinnesota. Only three perforations occurred; the patients werepromptlytreated andthere were no deaths. Basedon this informa tion. Gilbertsen theorizes that a physi cian could perform five to 10 proctosig moidoscopies a week for 25 to 50 years and anticipate only one perforation or less.6 Thus,usualcaremustbeexercised inperforming anyendoscopic proce dure, but the implication that a procto sigmoidoscopy should only be per formed by a specialist or physician with extensive training and experience, does not appear warranted. (For detailed in formation on endoscopic examination, see â€oe¿ Close-up: Endoscopic Colon and Rectum Examinations,― in this issue of
Caâ€"A Cancer Journal for Clinicians.)
Barium enema examination is another essential technique in patients with co lorectal symptoms. With the exception of emergency conditions such as acutely obstructing lesions which are revealed by barium enema alone, the addition of air contrast provides valuable diagnostic information. Air contrast is mandatory in patients with rectal bleeding. Careful X-ray search of the colon may detect many small or large lesions overlooked by palpation or endoscopy. However, lesions located in certain areas of the colon, particularly the flexures, may not be easily detected. A negative barium enema examination should be repeated when suspicious symptoms persist.
With the development of the fiberop tic colonoscope, the entire colonic lumen from the cecum to the anus can be directly examined in a very high per centage of patients. However, the safety and accuracy of this technique when ap plied on a large scale by physicians of varying skill andexperience muststill be determined. Wolff and Shinya have per formed over 2,000 colonoscopies with insignificant morbidity and without mor tality. They routinely use the 106cm. colonoscope which reaches the trans verse colon; the 186cm. scope, which canbe introduced deepinto thececum, is employed only when specifically indi cated.7 According to Overholt, colonos copy can successfully complement barium enema examination of the colon. Under direct visualization, small lesions previously missed by X-ray may be de tected and suspicious lesions can be studied and biopsied. 8 The use of commercially available guaiac impregnated slides is a simple, inexpensive and esthetically acceptable method of testing the feces for blood.
Patients are asked to prepare at home six stool slides from three consecutive daily bowel movements. (Two slides per day.)To increase theaccuracy anddis criminating ability of the stool analysis, Greegor believes that a special meat free, high-residue diet should be imple mented at least 24 hours before the first stool specimen is collected and then dur ing the next three days.9 The diet should contain (a) no meat, fish or chicken; (b) plenty of vegetables, both raw and cooked, especially lettuce, spinach and corn; (c) plenty of fruit, especially prunes, grapes, plums and apples; (
Figure-Diagnostic Guide for AsymptomaticPatients
Over 40 Yearsof Age orectal cancer.'°As further refinements in techniques occur, even more valuable information should be forthcoming.
On the other hand, early enthusiasm for carcinoembryonic antigen (CEA) de terminations as a screening test for co lorectal cancer has waned following evi dence that the test is positive in many cancerous and noncancerous lesions. Its greatest potential value will probably be to help determine the effectiveness of therapy and indicate prognosis. Further immunological studies of CEA are being conducted.11'12
Differential DiagnosisofCommon Conditions
Many conditions must be considered in the differential diagnosis of colorectal cancer. A high index of suspicion, a good history, thorough physical exami nation and biopsy will usually provide adequate information to make a defini tive diagnosis.
Colorectal Polyps
Patients with polyps typically present with rectal bleeding, prolapse through the anus and, rarely, intussusception. The most common polypoid lesion is the adenomatous polyp, occurring almost eight times more frequently than villous adenoma. Juvenile polyps are found principally in children under 12 years of age, but may occasionally develop in adults. Multiple polypoid lesions are usually noted in patients with familial polyposis or Gardner's syndrome. Pa tients with these hereditary disorders will usually develop colorectal cancer by the age of 40 to 50 years unless total colectomy is performed. Thus, families with a history of these conditions must be very carefully followed.
Sigmoidoscopy and barium enema with air contrast are essential in the dif ferential diagnosis of cancer. The larger the polyp, the greater the risk of cancer: lesions larger than 1.5â€"2.0cm. in size are more likely to be malignant than those less than 1cm. insize.2@t―3
Diverticulitis
Divert iculitis produces symptoms of left-sided appendicitis and the patient usually gives a history of pain and ten derness in the left lower quadrant asso ciated with fever, nausea, anorexia, vomiting and leukocytosis. These symp toms are usually controlled with antibi otics and resting the intestinal tract. When acute symptoms subside, sigmoi doscopy and/or colonoscopy and barium enema are performed to differentiate di verticulitis fromcancer. Indiverticu litis, the mucosa is intact and perfora tion, peritonitis, abscess and fistula for mation occur more commonly than in cancer. Both lesions may produce signs or symptoms of obstruction. There is no causal relationship between diverticu litis and the development of colonic cancer, but these lesions may coexist since both are diseases of older age groups. When the differential diagnosis cannot be made with endoscopy and barium enema, exploratory laparotomy is mandatory.
Chronic Ulcerative Colitis
This condition usually develops in younger patients than does cancer and is characterized by repeated attacks of bloody diarrhea, crampy abdominal pain, the passage of sanguinopurulent uncovers symptoms of appendicitis sev eral days or weeks prior to the develop mentof an abdominal mass. Barium enema showsextrinsic pressure on the terminal ileum and cecum without any evidence of intrinsic involvement, thus differentiating an appendiceal abscess from cancer of the cecum.
Rectal Abscess
The manifestations of a loculated pelvirectal, chronic pararectal or re trorectal abscess, particularly after the acute phase has subsided, may be con fused with cancer. However, a previous history of an acute inflammatory reac tion, a persistent nonmucosal mass, cryptitis and pus following incision point to the diagnosis of rectal abscess.
Thrombosed Hemorrhoids
The patient has a history of hemor rhoids and examination reveals a tender, painful, bluish-colored, firm and slightly mobile mass in the anus. A puffy, shiny red swelling beneath the skin and mucous membranes will also be noted. Thrombectomy usually affords dramatic relief.
Hypertrophied Anal Papillae
Enlargement and ulceration of anal papillae at the pectinate line is not un common. Anoscopic examination shows no induration in the anorectum, thus ruling out cancer.
Condyloma Acuminata
These lesions are probably of viral etiology and tend to occur in groups, generally in the perianal skin and perin eum, but may extend into the anal canal. They are seen more frequently in men and present with itching, bleeding. a material from the rectum, fever, malnu trition and dehydration. The symptoma tology and age group involved are of great clinical value in differentiating chronic ulcerative colitis from cancer. Endoscopy, barium enema and biopsy of suspicious lesions usually provide a definitive diagnosis.
Granulomatous Colitis
Granulomatous colitis generally in volves the right colon while ulcerative colitis affects the left colon and rectum, although the entire colon may be in volved in both diseases. The patient presents with chronic diarrhea asso ciated with abdominal cramps, fever, weight loss and often a palpable abdomi nal mass. Grossly bloody stools which are the hallmark of ulcerative colitis are not seen as frequently in granulomatous colitis but abscesses, fistulae and perianal sinuses are common. In patients with granulomatous colitis, sigmoidos copy reveals an essentially normal or edematous mucosa without bleeding. Barium enema shows inconspicuous shortening and segmental involvement with intervening normal segments of colon. Clinical findings, endoscopic evaluation and barium enema exam ma tion can usually differentiate granuloma tous colitis from cancer. In some cases, however, it is extremely difficult to dis tinguish granulomatous from ulcerative colitis.
Appendiceal Abscess
The patient presents with a palpable. tender or nontender mass in the right lower quadrant associated with vague abdominal symptoms. mild fever and leukocytosis. A careful history usually In Sterile Agreement Granted that speeches and articles should not always deal with local particulars, they should always deal with some particulars, if only by way of illustration. Otherwise, writer and audi ence waste their time in sterile agreement over absolute unknowns. In this world, at any rate, no art or science can long afford to invite assent of this kind to what is unproductive, to what will make no difference in actionâ€"including the action of thinkingâ€"i. Barzun. â€oe¿ So Long as Doctors Have to Think, Bulletin of the New York Academy of Medicine 47 234 1971. slight but offensive discharge and occa sionally ulceration. On examination they appear as glistening, pale-pink, wart-like elevations without induration.
Anal Fistula
Fistulae usually follow drainage of an ischiorectal or perianal abscess and may rarely develop into cancer. The most common finding is a brownish, fecal dis charge soiling the underclothes. Cryp titis isseenon anoscopy.
Anal Fissure
Pain is the chief symptom. Palpation reveals a tender ulcer, secondary to cryptitis and associated with fibrosis. On anoscopy the ulcer is seen at the poste rior anal raphe often guarded by in flamed edematous tissueâ€"sentinel tag. The elevated, rolled, overhanging edges found in malignant ulcers are absent.
Summary
The early diagnosis and localization of colorectal cancer is possible with the tools already available: digital rectal ex amination, proctosigmoidoscopy and/or colonoscopy, barium enema with air contrast and guaiac testing for occult blood. These modalities of treatment combined with a high index of suspicion can detect colorectal cancer at an early stage and achieve cure rates of up to 75-80percent. | 2018-04-03T05:24:25.199Z | 1974-05-01T00:00:00.000 | {
"year": 1974,
"sha1": "9dc233322f831cee8930e6837b85c827d79f648a",
"oa_license": null,
"oa_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.3322/canjclin.24.3.152",
"oa_status": "GOLD",
"pdf_src": "Wiley",
"pdf_hash": "e5afb50888585cbb1ad7ed2e448fcdc92b72a795",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
119210170 | pes2o/s2orc | v3-fos-license | Universal Quantum Computation with Continuous-Variable Abelian Anyons
We describe how continuous-variable abelian anyons, created on the surface of a continuous-variable analogue of Kitaev's lattice model can be utilized for quantum computation. In particular, we derive protocols for the implementation of quantum gates using topological operations. We find that the topological operations alone are insufficient for universal quantum computation which leads us to study additional non-topological operations such as offline squeezing and single-mode measurements. It is shown that these in conjunction with a non-Gaussian element allow for universal quantum computation using continuous-variable abelian anyons.
I. INTRODUCTION
The efficient storage and processing of quantum information is a major open problem for quantum computation (QC). One of the main routes to protect quantum information against random errors introduced via environmental decoherence and noise is the use of quantum error correcting (QEC) codes [1][2][3][4]. Essentially, such schemes attempt to correct errors with clever 'software' design. In contrast, recently, a new approach has been suggested in which information is intrinsically protected by storing it as a non-local topological degree of freedom. This approach is known as Topological Quantum Computation (TQC). In these schemes, ideally, the protection is an intrinsic property of the physical system, being inaccessible to noise. Moreover, any quantum gates that are implemented by non-trivial topological operations are also protected.
A recent development in TQC is the realisation of certain exotic states of matter that occur in two dimensions, the so-called anyon states, which use these topological properties of a system to provide a natural medium for storing and manipulating quantum information [5,6]. Anyons are quasiparticle excitations that exhibit fractional statistics, i.e., when two anyons are exchanged the quantum state acquires a phase shift corresponding to the fractional spin of the anyonic states. These anyonic states first emerged in connection with the fractional quantum Hall effect [7], which occurs in a two-dimensional electron gas at low temperatures. Anyons have been extensively studied for their fundamental interest, but recently it has been shown that anyons are a useful resource for fault-tolerant topological quantum computation [8][9][10]. Anyons fall into two general classes; abelian and nonabelian. Non-abelian anyons have received the most attention as it has been shown that some species, notably the Fibonacci anyons [9], are a resource for universal QC over discrete variables using just their braiding and fusion operations alone. However, it has been shown that the operations available to abelian anyons can also provide a universal gate set [11][12][13], but certain operations such as single-qubit rotations, must be carried out using non-topological methods.
A simple model from which one can produce abelian anyons was proposed by Kitaev for spin 1/2 systems [8,14]. However, this surface code model involves suitable combinations of four-body interactions which are difficult to achieve experimentally. In [15] it was shown that this code can be created efficiently from a two-dimensional cluster state by selectively measuring out single spins. This protocol was extended to the Continuous-Variable (CV) regime in [16] where it was shown that from a CV cluster state, a CV analogue of the Kitaev model can be constructed. It was demonstrated that the continuous excitations above this ground state are CV abelian anyons with non-trivial braiding statistics. Here we study the computational power of these CV anyons. We show how to create gates based on their topological properties and find that we can implement single-mode phase-space displacements as well as twomode controlled phase-space displacements, where both the control and the target can be either in the computational or in the conjugate basis. However, analogous to the qubit case, we find that the topological operations for the CV abelian anyons do not form a sufficient gate set for universal QC and so we include certain non-topological operations to supplement the topological operations in order to complete the gate set.
We begin, in Sec. II, with a short introduction to quantum computation over continuous variables. In Sec. III, we review briefly how to construct the CV Kitaev ground state from a CV cluster state and discuss the fusion and braiding properties of the anyonic excitations. In Sec. IV we examine the range of Clifford group operations that are achievable with the CV anyons by topological and non-topological actions. The Clifford gates alone are not enough for universal CV QC and hence in Sec. V, we study the effect on the anyons of applying a cubic phase gate to the Kitaev ground state. In Secs. VI and VII, we investigate the effect of finite squeezing of the resource state on the excitations and computational model. We conclude in Sec. VIII.
II. CONTINUOUS-VARIABLE QUANTUM COMPUTATION
Quantum logic over continuous variables generalizes the qubit Pauli X and Z operators to the Weyl-Heisenberg (WH) group [17], the group of phase space displacements. This is a Lie group with generatorŝ x = (â +â † )/ √ 2 andp = i(â † −â)/ √ 2 representing, for instance, a single quantized mode (qumode) of the electromagnetic field. These operators satisfy the canonical commutation relation [x,p] = i, equivalent to position and momentum. In what follows, we refer tox andp as position and momentum. The single-mode WH operators are defined as X(s) = e −isp and Z(t) = e itx , s, t ∈ R. The WH operator X(s) is the position-translation operator, which acts on the computational basis of position eigenstates {|x ; x ∈ R} as X(s)|x = |x + s . Z(t) is the momentum-translation operator, which acts on the momentum eigenstates {|p ; p ∈ R} as Z(t)|p = |p + t . These operators are non-commutative and obey the identity (1) In the following we will show how a CV computational model can be constructed using CV abelian anyons.
III. GENERATION OF ANYONIC STATES FROM CV GRAPH STATES
We review the protocol of [16], for the generation of the CV Kitaev ground state. This scheme is based on the first Kitaev lattice model which is a spin Hamiltonian for a two-dimensional square lattice. In Kitaev's scheme, a qubit (for instance, a spin 1/2 particle) is associated with each edge of a square lattice, for which the model Hamiltonian is given by where A s = j∈star(s) X j , and Here ∂(f ) denotes the boundary spins of a plaquette and the operators X and Z are the standard Pauli matrix operators σ x and σ z . These four-body interactions can be produced experimentally by cooling the system to its ground state, but a more practical approach suggested in [15] is to create the ground state dynamically from graph or cluster states. In [16] this model was generalized to CV states which we turn to now.
In the ideal case, CV cluster states are prepared from a collection of N zero-momentum eigenstates [18][19][20][21], which we write as |0 ⊗N p , where the p-subscripted kets satisfyp|s p = s|s p . These states are then entangled via a collection of controlled-Z operations, denoted C Z = exp(igx i ⊗x j ), where g ∈ R is the strength of the interactions (we will assume g = 1 throughout). Labelling the nodes of the graph in some arbitrary order, we can define a symmetric adjacency matrix A = A T whose (j, k)th entry A jk is equal to the weight of the edge linking node j to node k (with no edge corresponding to a weight 0). Note in the ideal case the diagonal entries are all zero since we do not allow for self loops in the graph [22]. The collection of controlled-Z operations used to make the CV cluster state is then a function of A, denoted C Z [A]. The CV cluster state with the graph A is then wherex = (x 1 , ...,x N ) T is a column vector of position operators. Ideal CV cluster states in the unphysical limit of infinite squeezing satisfy a set of nullifier relations, which can be written as wherep = (p 1 , ...,p N ) T is a column vector of momentum operators. This represents N independent equations, one for each component of the vector (p−Ax), which are the nullifiers for |ψ A , because that state is a simultaneous zero-eigenstate of them. The nullifiers are written explicitly asĝ where the modes a ∈ G correspond to the vertices of the graph of N modes and the modes b ∈ N a are the nearest neighbours of mode a. An efficient way of representing CV cluster states is via the stabilizer formalism. A state |φ is stabilized by an operator K if it is an eigenstate of K with unit eigenvalue. If such a set exists for a given state, then we call the state a stabilizer state, and we may use the generators of its stabilizer group to uniquely specify it. The stabilizer operator for a cluster state is To construct the CV Kitaev code state from a CV cluster state, we make single-mode measurements on every second mode in either the position or momentum basis, with a subsequent Fourier transform, F , for all the remaining modes which leaves us with a new graph state |ψ (Fig. 1). This new state is described by the set of correlations of the form {â s = (p s,1 +p s,2 +p s, 3 where s and f label star and plaquettes, respectively, and the indices 1, .., 4 of the position and momentum operators denote those modes located at a common star or at the boundary of a common plaquette. The new stabilizer operators that describe this state, are analogous to the first Kitaev model for a two-dimensional spin lattice. Since these new stabilizer operators commute, the new ground state corresponds to an anyonic ground state with A s (ξ)|ψ = |ψ and B f (ξ)|ψ = |ψ for all stars s and plaquettes f , in the limit of infinite squeezing.
We regard continuous excitations above the preprepared ground state |ψ as CV anyons, produced by applying Z and X operators on the ground state. Specifically, the position-translation operator applied to some mode of the lattice creates a pair of m-type anyons on adjacent plaquettes, i.e., |m((−1) d t) = X(t)|ψ (d ∈ {1, 2}), where d = 1 means the relevant mode lies on the vertical edges, and d = 2 refers to the horizontal edges). An e-type pair of anyons is created on adjacent vertices of the lattice by |e(s) = Z(s)|ψ . We define a computational basis |r v/f on vertices, v, or plaquettes, f , with r ∈ R, composed of a pair of anyons as The fusion rules describe the outcome of combining two anyons. By application of a sequence of X and Z operators we can braid the anyons. For example, consider an initial state |ψ ini = Z i (s)|ψ = |e(s) . If an anyon of type m is at a neighbouring plaquette, it can be moved around e along a path generated by successive application of X(t) on the four modes of the star. The final state is The phase factor is known as the topological phase factor, which reveals the presence of enclosed anyons. Note that the same phase factor would be acquired by the state independent of the path the braiding anyon follows. This topological character reveals the potential robustness of operations with CV anyons and their use as a resource for fault-tolerant quantum computation.
IV. CLIFFORD GATES
We now examine the computational power of the CV abelian anyons. In particular we show how the Clifford group operations are achieved using both topological and non-topological means. The set {Z(s), F, P (η), C Z ; s, η ∈ R} generates the Clifford group [17], where P (η) = exp[i(η/2)x 2 ] is a CV squeezing gate, F = exp[iπ/4(x 2 + p 2 )] is the Fourier Transform operator, and C Z = exp(igx i ⊗x j ) is the controlled-Z gate, as defined before. Transformations within the Clifford group correspond to Gaussian transformations mapping Gaussian states onto Gaussian states. We see below that the topological operations available to CV abelian anyons are not sufficient to generate the entire group and we will require nontopological operations to complete the set.
A. Topological Operations
Our first topological operations are quadrature displacements, Z(s) = e isx . Phase-space displacements are achieved through the creation and fusion of anyons. It is easy to see that creation of an anyon results in a displacement away from the ground state. For non-trivial displacements we fuse anyons of the same type, created by displacements on modes i and j. The anyon on site j can then be moved to site i to implement fusion, To act our quadrature displacement on the computational basis we must ensure that both anyons in the produced pair are fused with its counterpart. Hence the effect of our displacement on the computational basis is The change in the computational basis for the m-type anyons follows similarly. We can extend this to the two mode SUM gate (Fig. 2), which is a controlled displacement C X = e −ixi⊗pj , i.e., |x 1 |y 2 → |x 1 |x + y 2 . We affect the SUM gate by fusing one of the anyons from the first mode with an anyon from the second mode. This results in a displacement of the second mode dependant on the state of the first: where we treat the second anyon of mode two as a spectator anyon that will be annihilated at the end of the computation and the control is left in the original state (up to a sign change).
FIG. 2:
The two-mode SUM gate. An anyon from the top pair is fused to produce a controlled displacement on the lower pair.
We can also perform a controlled-Z gate (Fig. 3), which acts on the computational basis as |x 1 |y 2 → |x 1 e iφ |y 2 . This time we affect this transformation by braiding the anyons. For example to perform a controlled displacement in the conjugate basis on an m-type anyon we braid an e-type anyon around it. From equation (7), we see that the state picks up a phase dependant on the anyonic states. Hence, we have found that the controlled shift operation C X , e ixipj , corresponds to partial fusions of etype anyon pairs, while the controlled shift operation C Z , e ixixj , corresponds to partial braiding between e-and mtype anyon pairs. Similarly, we note that the operations e ipipj and e ipixj are achieved by partial braiding between m and e-type anyon pairs and partial fusion of m-anyon pairs, respectively.
B. Non-Topological Operations
Above we saw how we can implement single mode displacements, displacements in the conjugate basis and two-mode controlled displacement gates. Now we go beyond simple first-moment Gaussian operations and consider the manipulation of second moments. In particular, we seek to complete the set of Clifford gates by the inclusion of a squeezer and a Fourier transform. The squeezing operation compresses the position quadrature by a factor η while stretching the conjugate quadrature by 1/η. We cannot directly squeeze our anyons since the only action we can take on the anyons is fusion and braiding. Instead, we squeeze on some mode i of the ground state, Then creating an anyon on this squeezed ground state mode and commuting through gives us We find that squeezing the ground state is equivalent to squeezing the anyon with an additional phase space displacement. This squeezing operation combined with a measurement in the X basis can be used to implement a Fourier transform F . The action of F is to switch between the position and momentum bases, i.e., F |x x = |x p . This corresponds to a generalization of the Hadamard gate for qubits. To perform F on our anyons, we begin by preparing a zero-momentum squeezed ground state |0 p . We proved above that up to a displacement, any squeezer on the ground state acts on the anyonic excitations in the same way. Hence we can produce zero-momentum anyons. For example, consider an m-type anyon in the computational basis |ψ = X(s) dtf (t)|t q . We fuse this anyon with the momentum squeezed anyon: Then, performing a measurement ofp with outcome m on the mode corresponding to the first anyon collapses this to dtf (t)e i(t+s)m |s + t p = X(m)F |ψ .
We see that the effect of this procedure is to apply a Fourier transform modulo a known quadrature displacement. This completes our set of Clifford gates, and so with appropriate non-topological operations we can apply any Gaussian transformation. As stated above, Gaussian transformations are not sufficient for universal QC and we address the question of universality in the next section.
V. NON-CLIFFORD GATES AND UNIVERSALITY
For our abelian anyon computational scheme, we require either a non-Gaussian gate or a non-Gaussian resource. In a similar fashion to our squeezing operation, we apply a transformation known as a cubic phase [23], V (γ) = e iγx 3 , γ ∈ R, to some mode i of the ground state. Then we commute V (γ) through the X and Z operators to find the effect on the anyons: Now applying X and Z operators to create e-and m-type anyons and commuting the cubic phase gate through we find Hence we find that applying a cubic phase to the ground state is equivalent to applying the cubic phase to the anyons with extra displacements and squeezing operations. This produces non-Gaussian anyons that can then be used to obtain a universal gate set. To conclude our discussion, we attempt to provide a classification of our gate operations. We found that robust topological operations correspond to controlled or uncontrolled WH gates. Non-robust, non-topological Clifford operations correspond to symplectic operations. Non-robust, nontopological, non-Clifford operations correspond to non-WH, non-symplectic operations. In group theoretical terms, all operations that are not elements of the normalizer of the WH group (Clifford group) cannot be topologically realized using abelian anyons. Of those operations that are elements of the normalizer of WH, only the elements of the normal subgroup of the Clifford group can be topologically realized using abelian anyons; those Clifford elements that are not elements of the normal subgroup cannot be realized topologically using abelian anyons [24].
VI. PHYSICAL STATES -FINITE SQUEEZING
So far, we have only considered the generation of anyonic statistics on an infinitely squeezed ground state. However, this state is highly unphysical. Here we extend our model to include finite squeezing of the initial state and show what effects this has on our computational model.
A method to extend the graph representations from ideal (infinitely squeezed) CV cluster states to their finitely squeezed Gaussian approximations was given in [22]. There it was shown that the nullifier formalism for CV cluster states can be extended to general Gaussian pure states using the simple replacement of the CV cluster state graph A with the Gaussian graph Z, so that (p − Zq)|φ Z = 0 with the new non-Hermitian nullifiers defined as Then the adjacency matrix Z for a Gaussian pure state is a complex matrix with imaginary diagonal entries, ie −2r k , corresponding to self-loops on the modes, and the remaining entries either 0 or 1 depending on the particular CV cluster state. Starting from an N -mode square cluster state defined by this nullifier and carrying out the measurement pattern described for the ideal cluster, we generate the finitely squeezed Kitaev lattice with complex nullifiersâ s =â s = (p s,1 +p s,2 + p s,3 +p s, for the remaining modes, where s and f label the stars and plaquettes, respectively, and ∂f denotes the boundary of a face. Comparing with the nullifiers in the infinitely squeezed limit, we observe that finite squeezing introduces extra imaginary terms to the plaquette nullifiers. The stabilizers corresponding to these complex nullifiers are A s (ξ) = e −iξâ s = j∈star(s) X s,j (ξ) and These operators correspond to the Kitaev model with an extra complex term, but note that these reduce exactly to the Kitaev stabilizers as the squeezing parameter r → ∞. These new stabilizers still commute and so the new state |φ corresponds to the anyonic ground state with A s (ξ)|φ = |φ and B f (η)|φ = |φ . In order to see the effects of finite squeezing, we apply the unphysical stabilizers to the physical ground state, where we dropped the subscripts f of the momentum operators for simplicity. We observe that finite squeezing of the ground state violates the unphysical stabilizer conditions by an imaginary phase, ∼ iη 2 , and by imaginary position shifts, ∼ iη. Having derived the form of our finitely squeezed CV lattice, we now turn to anyonic excitations and basic braiding operations.
VII. ANYONIC CREATION AND BRAIDING ON FINITELY SQUEEZED LATTICE
By applying single-mode operations to the ground state, we can examine the excitations above the physical ground state. The effects of finite squeezing on the creation of anyonic excitations are revealed when we calculate the violation of the finitely squeezed nullifiers due to the application of X i (t) and Z i (t) on some mode of the physical ground state defined by Note that now, due to the non-Hermiticity of the physical nullifiers, their violations through anyonic excitations may in general be complex. An excitation of the vertex ground state due to Z j (t) corresponds to butâ s =â s , as the finite squeezing has no effect on the vertex nullifier. Hence, Just as in the infinitely squeezed case, this yields a stabilizer violation of t. The Z j (t) applied to the plaquette stabilizers gives This differs from the infinitely squeezed case (where we had no violation at all), with an imaginary nullifier violation of ite −2rj . This time physical anyons may appear as complex violations of the ground-state stabilizers. Applying X j (t) to the ground state yields, so no stabilizer violation occurs. Finally, the plaquette stabilizer gives which corresponds to a violation of ±t.
In summary, finite squeezing gives us a violation of the plaquette ground state stabilizer due to the action of Z(t), whereas it did not exhibit violations in the infinitely squeezed case. The vertex stabilizers are unaffected and yield violations of the same form as in the infinitely squeezed case. This suggests that any topological operations carried out on e anyons are topologically protected from error. However, the only operation possible on star anyons alone is the SUM gate which is not on its own sufficient for quantum computation. The anyons are now represented in general by complex violations. These may then said to be new types of excitations produced through finite squeezing which we call complex anyons.
The effects on the braiding procedure and hence gate operations are determined by generating vertex and plaquette anyons and guiding them around each other in closed loops.
Commuting through to enact the braid yields Using our definition of the ground state, B f (t)|φ = |φ , we obtain We can express the term proportional to t as an imaginary displacement, As in the infinitely squeezed case, we observe a phase change of e ist , but this time, for finite squeezing, we have an extra imaginary displacement and a term proportional to t 2 (corresponding to an extra imaginary phase, similar to what we had before for the groundstate plaquette stabilizers with finite squeezing). We may call the combination of these terms the topological factor for the braiding of finitely squeezed anyons, and we note that this factor would not be obtained if the initial states were unexcited. We can absorb the complex displacement into the definition of the ground state such that the nulliferb f is no longer zero, but has an imaginary violationb f = it j (−1) j+1 e −2rj . Then b f = s + it j (−1) j+1 e −2rj = s is the nullifier corresponding to a finitely squeezed m-anyon. The topological phase produced when braided with an e-type anyon is then This is one of the central results of this paper, extending the simple (infinite squeezing) factor e ist of [16] to the realistic case of finite squeezing. Similar to the infinitely squeezed case, the state can acquire any phase, but now the phase is modified by extra factors due to finite squeezing. This should not affect topological gate operations since these displacements and dampings are taken into account through the definition of the ground state and they depend on the known squeezing r j available at each qumode. Hence we have shown that topological operations on CV abelian anyons are protected from errors due to finite squeezing of the initial ground state. Note that this result goes beyond that of Ref. [16]. There it was argued that in the case of finite squeezing, excited (anyonic) states can be experimentally distinguished from the ground state and, similarly, the effects of braiding loops can still be detected, provided the corresponding phase-space displacements are sufficiently large. In our treatment, such a requirement is unnecessary. Using the complex-nullifier formalism, we find that any finitesqueezing effects on the 1st-moment-shifts can be absorbed into the definition of the excited states as well as into the topological phases. However, note that these re-defined WH frames would always depend on the explicit values of the corresponding anyonic excitations and anyon-braidings [i.e., the t-dependencies in Eqs. (21)(22)(23)(24)(25)(26)(27)(28)(29)(30)].
As a result, all those gates shown to be implementable in a topological fashion turn out to be robust against finite squeezing errors. Nonetheless, non-topological operations such as the Fourier transform do pick up extra errors due to finite squeezing since they rely on the ability to create zero-momentum eigenstates. In fact, all these non-topological gates include 2nd-or higher-moment manipulations which will be affected by the finite squeezing of the initial graph states. Similarly, of course, the entanglement of the ground and excited states, becoming manifest through nonclassical 2nd-moment correlations, does depend on the finite squeezing of the initial states.
VIII. CONCLUSION
We have shown that abelian anyons generated on an ideal CV Kitaev lattice are a useful resource for continuous-variable quantum computation. We have described the quantum gates that can be achieved by topological operations alone and found this did not form a sufficiently powerful gate set. By including offline squeezing, measurements, and cubic phase gates, we have shown that the computational power can be increased significantly. These additional resources provide a universal gate set for the CV anyons. It has therefore been possible for us to give a classification of the topological operations available to CV abelian anyons. We have shown that braiding and fusion only account for controlled and uncontrolled WH gates, non-topological operations are essential to complete the Clifford group, and a further non-Gaussian element is required to achieve universality. We note that the topological controlled displacements only give entanglement (including 2nd-moment non-classical correlations) if we have access to superpositions of anyonic excitations, e.g. in a bi-anyonic vacuum state, ds dt e −s 2 −t 2 |s |t . Such an extension will give rise to a more general model including non-abelian anyons.
We have also considered the effect of finite squeezing on the CV Kitaev lattice and found that the excitations on this physical ground state correspond to Gaussian anyon states. These physical anyons, mathematically represented by complex stabilizer violations, can perform fault-tolerant topological operations that are protected against errors due to finite squeezing. However, we cannot create a fully protected gate set, since some gates require non-topological operations. Nonetheless, we have been able to identify those operations which are protected, and in a future work, we shall attempt to present a universal, topological gate set using nonabelian anyons.
This research has been supported by the EU STREP project COMPAS FP7-ICT-2007-C-212008 under the FET-Open Programme, by the Scottish Universities Physics Alliance (SUPA) and by the Engineering and Physical Sciences Research Council (EPSRC). DFM acknowledges the financial support from the Deutscher Akademischer Austausch Dienst (DAAD). PvL acknowledges support from the Emmy Noether program of the DFG. | 2011-12-22T17:05:07.000Z | 2011-12-22T00:00:00.000 | {
"year": 2012,
"sha1": "da869a9525255b159a3d1fd8339e165502179d8b",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/1112.5385",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "da869a9525255b159a3d1fd8339e165502179d8b",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
} |
139136708 | pes2o/s2orc | v3-fos-license | Model Verification for Material Parameters of Titanium Alloy Ti-6AL-4V and Steel
. To carry out numerical simulation of TC4 titanium alloy blade impacted by foreign objects effectively, this paper takes the test results of steel ball and sandstone impacting titanium alloy flat blade inlet and blade surface as the benchmark, and uses ANSYS/LS-DYNA software to adopt kinematic hardening plasticity model to simulates the impact results and inverts the contact stiffness factor and the other three parameters were obtained by the way of inversion reasoning.
Introduction
The foreign object damage of the blade is of great significance for ensuring the reliability and flight safety of the aero-engine. The research methods are divided into experimental and simulation calculations. At present, there are many kinds of literature on simulation of blade impact damage [1][2][3][4][5][6][7][8][9][10]. And the material models used are also various. Different material models correspond to different material performance parameters, and kinematic hardening plasticity model proposed by Cowper-Symonds is one of the commonly used models. When this model is selected, a total of 1 calculation parameter and 9 material performance parameters are involved. On the one hand, for the same material, the material performance parameters of different materials are not only different but also far apart; on the other hand, most of the literature does not give all the material performance parameters and calculation parameters required for the full-stage simulation calculation, which reduces the reference value of the existing literature simulation results.
Experimental study 2.1 Scheme 1 analysis
According to the blade sample used in the steel ball impact test of the inlet vane of the flat blade in 1999 by Joseph L. Hamrick et al. [1] , the same flat blade geometry model is established here, as shown in Fig. 1. The steel ball has a diameter of 2 mm, and the flat blade has a thickness of 1.27 mm, a width of 4 mm, and a height of 6 mm.
The finite element discretization of the geometric model of the flat blade and the steel ball is performed In the ANSYS/LS-DYNA by using the SOLIDl64 unit. The grid element size of the flat blade is 0.05 mm, and the size of the steel ball mesh unit is 0.1 mm. Fixing the joints at one end of the flat blade, the steel ball impacts the intake side of the flat blade vertically at a speed of 305 m/s, and draws a dip with a depth D and a width W at the inlet side of the blade, as shown in Fig. 2. The flat blade is Ti-6A1-4V material, and the kinematic hardening plasticity model is adopted; the steel ball adopts the piecewise linear constitutive model. The contact way of Erode nodes to the surface of point-toface is adopted, wherein the blade is a target body and the steel ball is a contact body. During the verification process, the hardening parameters, strain rate parameters C, P, and contact stiffness factors were repeatedly adjusted. Finally, it was found that only when the contact stiffness factor was 0.7 and the other three parameters were listed in Table 1, the calculation results were close to the test results. The specific calculation results are shown in Table 2 and Figure 3. During the verification process, the hardening parameters, strain rate parameters C, P, and contact stiffness factors were repeatedly adjusted. Finally, it was found that only when the contact stiffness factor was 0.7 and the other three parameters were listed in Table 1, the calculation results were close to the test results. The specific calculation results are shown in Table 3.
Discussion
The depth of the impact dent calculated in the first scheme is smaller than the experimental value, and the depth of the impact dent calculated in the second scheme is larger than the experimental value. In order to make the calculated values of the dent depth in the two schemes close to the experimental values, a compromise method was adopted in the selection of the material parameters C, P and the contact stiffness factor of titanium alloy and steel. When the material parameters are as shown in Table 1, the calculated dent depth values for both solutions are close to the test values. When the material parameters are changed again, only the calculated dent depth of one of the solutions is closer to the experimental value, and the calculated value of the other scheme is far from the experimental value.
Conclusion
1. Using ANSYS/LS-DYNA software to adopt kinematic hardening plasticity model to simulates the impact results which helps to provides a viable method for testing residual stress and structural strength of materials.
2. By comparing the simulation value with the actual collision effect value, the material parameters are obtained by inversion reasoning, and the residual stress and the mechanical strength are calculated and predicted by using these parameters. | 2019-04-30T13:08:38.764Z | 2018-01-01T00:00:00.000 | {
"year": 2018,
"sha1": "3dac0e48eb96d324c76c90996ca9bcf54d8c3031",
"oa_license": "CCBY",
"oa_url": "https://www.matec-conferences.org/articles/matecconf/pdf/2018/86/matecconf_icct2018_01005.pdf",
"oa_status": "GOLD",
"pdf_src": "Adhoc",
"pdf_hash": "5df6c25a31256e0fe41e8e5ef4285a20d59e1080",
"s2fieldsofstudy": [
"Engineering",
"Materials Science"
],
"extfieldsofstudy": [
"Materials Science"
]
} |
52312827 | pes2o/s2orc | v3-fos-license | An open-source tool to identify active travel from hip-worn accelerometer, GPS and GIS data
Background Increases in physical activity through active travel have the potential to have large beneficial effects on populations, through both better health outcomes and reduced motorized traffic. However accurately identifying travel mode in large datasets is problematic. Here we provide an open source tool to quantify time spent stationary and in four travel modes(walking, cycling, train, motorised vehicle) from accelerometer measured physical activity data, combined with GPS and GIS data. Methods The Examining Neighbourhood Activities in Built Living Environments in London study evaluates the effect of the built environment on health behaviours, including physical activity. Participants wore accelerometers and GPS receivers on the hip for 7 days. We time-matched accelerometer and GPS, and then extracted data from the commutes of 326 adult participants, using stated commute times and modes, which were manually checked to confirm stated travel mode. This yielded examples of five travel modes: walking, cycling, motorised vehicle, train and stationary. We used this example data to train a gradient boosted tree, a form of supervised machine learning algorithm, on each data point (131,537 points), rather than on journeys. Accuracy during training was assessed using five-fold cross-validation. We also manually identified the travel behaviour of both 21 participants from ENABLE London (402,749 points), and 10 participants from a separate study (STAMP-2, 210,936 points), who were not included in the training data. We compared our predictions against this manual identification to further test accuracy and test generalisability. Results Applying the algorithm, we correctly identified travel mode 97.3% of the time in cross-validation (mean sensitivity 96.3%, mean active travel sensitivity 94.6%). We showed 96.0% agreement between manual identification and prediction of 21 individuals’ travel modes (mean sensitivity 92.3%, mean active travel sensitivity 84.9%) and 96.5% agreement between the STAMP-2 study and predictions (mean sensitivity 85.5%, mean active travel sensitivity 78.9%). Conclusion We present a generalizable tool that identifies time spent stationary and time spent walking with very high precision, time spent in trains or vehicles with good precision, and time spent cycling with moderate precisionIn studies where both accelerometer and GPS data are available this tool complements analyses of physical activity, showing whether differences in PA may be explained by differences in travel mode. All code necessary to replicate, fit and predict to other datasets is provided to facilitate use by other researchers. Electronic supplementary material The online version of this article (10.1186/s12966-018-0724-y) contains supplementary material, which is available to authorized users.
physical activity, where segmentation into journeys is important to assess travel behaviour [18][19][20][21]. A focus on journeys often results in short periods of physically active transit behaviour, such as walking between bus-stops, being identified as part of a non-active travel mode. However, for physical activity researchers, quantifying the volume and intensity of physical activity when actively travelling is an essential component of the overall purpose of a journey. Consequently, it is important to identify all data-points that denote active travel, so that all time in active travel modes can be quantified. As a result, we identify the travel mode of each GPS data point (recorded every 10 s), without prior segmentation into journeys.
In recent years, supervised machine learning has shown the potential to identify active travel from physical activity data. Supervised machine learning algorithms are trained on an example data-set, and are then used for prediction to other data-sets. The most promising algorithms appear to be random forests, an ensemble supervised learning algorithm where predictions are taken from a consensus across a large number of decision trees [18,22,23]. A related algorithm, gradient boosted trees, has recently replaced random forests as a leading algorithm for data science tasks, with many machine learning approaches using the XGBoost implementation of gradient boosting instead of random forests [24,25].
Here we present a method to distinguish five travel modes (walk, cycle, motorised vehicle, train, and stationary) using accelerometer and GPS data and the supervised machine learning tool XGBoost. We use survey data from the Examining Neighbourhood Activities in Built Living Environments in London (ENABLE London) study [26] to create a training data-set of combined GPS and accelerometer data of daily commuters on which to develop and test the algorithm. All code necessary to replicate our findings and apply our predictive model to other data-sets is made available as a package of the open-source statistical software environment R [27,28]. We also provide a full usage example, so that researchers inexperienced in coded input tools such as R can apply the model [28].
Data collection
We used data from the ENABLE London study, which is described in detail elsewhere [26]. In brief, the study is examining the effect of the area of residence, including features of the local built and social environment on health behaviours, particularly physical activity levels. Between January 2013 and December 2015, a total of 1278 adult participants were recruited from neighbourhoods largely in the east of London, UK.Participants were asked to wear an accelerometer (ActiGraph GT3X+; Florida, USA) and a GPS receiver (Qstarz BT-1000XT; Taipei, Taiwan) on an elasticated belt worn around their waist for seven consecutive days, removing devices for sleep, swimming and bathing, with 1089 (85%) participants providing both accelerometer and GPS data. Participants also completed a questionnaire to describe their travel patterns to work/place of study. They reported the specific days on which they would be travelling to work/study during the ActiGraph and GPS wear period, and whether they commuted at the same time on each day. Reported travel modes for these journeys to and from work were: tube (underground) / train (overground) / bus, minibus or coach / taxi / motorcycle, scooter or moped / driving a car or van / passenger in a car or van / bicycle / walk / jog / other. Travel modes were re-categorised into walk, cycle, vehicle (taxi, motorcycle, car/van driver and car/van passenger/bus/minibus/ coach) and train (underground and overground rail) for this analysis (Table 1). Insufficient participants consistently jogged to work for us to be able to separate "jog" as an additional mode. In addition, time of leaving and arriving for each journey to and from work was collected. The study was approved by the City Road and Hampstead Ethical Review Board (REC reference number 12LO1031); all participants gave written informed consent.
Data preparation and cleaning
Raw accelerometer data were extracted as csv files using ActiLife 6 software (ActiGraph, Florida, USA). We chose to not use processed accelerometer count data, because count data is processed differently for each device, and so is not comparable between devices. Furthermore, raw accelerometer data is more detailed than count data, allowing us to create more variables for our algorithm to assess. Acceleration data were then summarised per 10 s epoch as the median absolute deviation from the median, 10th percentile, 90th percentile, skewness and kurtosis of each axis of the accelerometer. We calculated the fast Fourier transform of the accelerometer signal and took the mean strength of all acceleration signals for each accelerometer axis per 10 s epoch. Mean strength was extracted following visual inspection of the full range of transforms, with all showing a similar pattern between travel modes. These derived accelerometer characteristics were merged by timestamp to GPS data using a custom R function [28]. We identified non-wear time using the GGIR 1.5-12 package of R, which identifies periods of 60 min where there is a standard deviation on at least two accelerometer axes of less than 13 mg (1 mg = 0.00981 ms − 2 ) [29,30]. A wide variety of non-wear time algorithms have been used, and a 60 min window is recommended to balance accuracy of non-wear time identification with minimising data loss [31]. The 60-min window we use is analogous to that used on count data, but uses small axis deviations instead of count data, because our analysis focusses on raw data. To provide a measure of satellite signal quality we calculated the sum of the signal to noise ratio (sumSNR) from each satellite the GPS device was connected to at each epoch. SumSNR is a measure of the accuracy of signal coming from each satellite to the GPS deviceif the signal is unobstructed then there should be many satellites connected to the GPS device, and each should show a high signal to noise ratio. If there are obstructions to the GPS signal, such as the participant being indoors, then there will be fewer satellites connected to the GPS device, with lower signal to noise ratios. SumSNR gives a single measure of signal quality, rather than the three measures (horizontal, vertical and position dilution of precision) the GPS device outputs, therefore reducing the number of variables to consider.
During normal travel, variables such as speed or accelerometer signals are not stable across every 10 s epoch, there will be variation, e.g., cars must stop for traffic lights, walkers must pause to cross roads, GPS signal can be poor inside trains, leading to loss of accuracy where single data-points can be located over 50 m from train lines. This variation makes all modes more difficult to identify. We smoothed out some of the inherent natural variation in travel by calculating four-minute We also calculated mean sumSNR and mean distance from train lines. We calculated the distance of each data-point from train lines using a combination of Meridian 2 rail network data for the UK, OS OpenMAP data for central London and the spatstat package of R [32][33][34]. If researchers are not used to using GIS data, it is currently more easy to get hold of and utilise than ever before, and included within our code is a demonstration of the acquisition and use of train line data [28]. We also calculated distance travelled over the previous minute and over the next minute. All the variables were chosen because they are likely to differ between travel modes: vehicles and trains should have higher speed than walking and cycling; walking should show greater accelerometer activity than other modes; cycling may show higher accelerometer activity than vehicles and trains; both vehicles and trains have metal structures around the participant which may obstruct GPS signal resulting in lower sumSNR than other travel modes. Either distance over the next minute or distance over the previous minute should be very low while a subject is stationary.
Training data-set creation
The purpose of the training data is to provide reliable examples of how different travel modes are represented in accelerometer and GPS data, at 10 s epochs. As a result, it is important that a data-point in the training data is a true representation of the assigned travel mode. The context around the training data is less relevant, i.e. the purpose is to ensure that a data point is taken from a point in time when someone is walking; whether they were recently in other types of travel behaviours is not relevant to the assignment to a specific travel mode. As a result, we use a conservative methodology to extract reliable points from commute journeys. Importantly, this does not mean that our method can only predict the mode of commutes, rather we train the model using data from commutes and predict to all data. The time of the journey to and from work was extracted from the ENABLE London questionnaire, using the participants reported home departure time and work arrival time (commute to work) and work departure time and home arrival time (commute from work). We only used participants who commuted to and from work using the same mode of transport and who specified the time they usually commuted for both journeys, in total 326 participants ( Table 1, Additional file 1 for demographics). We extracted all combined GPS and accelerometry data during commutes in R and exported this as a shapefile for ArcGIS 10.4, using the sp and rgdal packages of R [35,36]. Assuming a five-day working week for each participant, the total possible number of commutes included within our training data is therefore 3260 (two commutes per day for 5 days). However, not all data were available due to non-wear, GPS signal loss and participants not necessarily working 5 day weeks: hence, our training data-set was based on 1174 commutes from the 326 participants.
We then manually identified all points during (total 131,537) the commute of each participant that conformed to the mode they stated, using a Geographical Information System (ArcGIS 10.4) (Fig. 1a). Only the points relevant to the stated mode were identified, i.e., if the participant used a train and walked to and from the train station we only marked those points along the train line as the mode "train", and the walked points would not be included in the training data (Fig. 1b). Any sequences of points where we could not clearly ascertain what travel mode was in use, e.g. because there was repeated GPS signal loss, were excluded from the a) b) training data. In addition, any points occurring within the period of the commute that showed no directionality and were clustered around a single location were classified as "stationary". As a result, each participant contributed to the mode which they stated they used to commute and, potentially, to the "stationary" mode, if they displayed any non-travel behaviour. Vehicle and train journeys were confirmed by location of appropriate features (e.g., roads/rail tracks). The identification of the "stationary" subset allowed us to predict travel mode for every data point in the full data-set, without first making assumptions about which points represent any form of travel. It is important to note that stationary does not imply inactive. For our purposes we are seeking to identify when participants travel, therefore walking within a building, which will appear stationary in terms of GPS signal, would be classified as stationary using our methodology. This twofold assessment of the training data, both with the participants stating their mode during this time, and then a researcher manually double checking it in GIS, means that we can be sure that our training data-set contains reliable examples of each travel mode.
Model fitting and prediction
The training data-set was split into two different sections for model-fitting. First, we extracted a subset of the training data to test different moving window sizes: if a participant contributed multiple days to the training data, we took the first day to test moving windows. If a participant only contributed a single day to the training data, we extracted half of that day's data to test moving window sizes. Secondly, we used the remaining points to fit a gradient boosted tree, using the XGBoost package of R [37]. Gradient boosted trees are an example of an ensemble learner [38]. To learn how to assign a mode to each point, a gradient boosted tree fits a large number of decision trees to the data, each of which is shallow. Following creation of a decision tree, data-points are re-weighted (the relative importance of points are changed), to emphasize points which were miss-classified last round, and a new decision tree is fit. Each tree is a weak learner i.e. it performs better than random but with poor accuracy, however a consensus across all trees leads to high predictive accuracy. In prediction, each point is assessed for how likely they are to belong to each travel mode, with the highest probability determining the predicted outcome. Where our parameters differ from default, it is to avoid overfitting to the training data as much as possible. Therefore the learning rate is set lower than default (0.1 instead of 0.3). The learning weight is a measure of how much we dilute the default re-weighting of points. We also used subsampling value of 0.2, so that each tree was only fit to 20% of the available data.
We used five-fold cross-validation to assess model accuracy [39]. Participants within the training data were randomly assigned to one of five subsets. Each subset was iteratively removed from the pool of training data, a model was trained on the remaining 80% of the training data and the excluded subset was used as test data. Therefore, our cross-validation contains five separate fitted models and test data-sets, of which we report the overall accuracy scores. Full cross-validation output is available in Additional file 2, along with all model parameters used.
While our training data contains reliable examples of different travel modes when participants are known to be travelling, it is not a true representation of identifying travel from free-living data. As a further test of the predictive accuracy of our method, we compared our predicted mode with the manually identified travel patterns across all time periods from other participants. We randomly selected 21 participants from the ENABLE London study, who contributed 402,749 data points. These participants were not included within the training data-set. We exported each individual's data as a Shapefile for ArcGIS 10.4 and then worked through each day of data, manually classifying the travel-mode of every data-point.
To test the generalisability of our fitted model to other data-sets we also compared predicted values to manually identified data from 10 participants from a second separate data-set (total 210,936 data points), the Sedentary Time and Metabolic Health in People with type-2 Diabetes study (STAMP-2) [40]. Briefly, STAMP-2 was a cross-sectional observational study of sedentary behaviour in adults with newly diagnosed (diagnosis within previous 5-12 months) type-2 diabetes, conducted in two English National Health Service (NHS) Foundation Trusts in South West England. A total of 139 participants were recruited between January 2014 and June through diabetes education days, general practitioner (GP) referral and self-referral. Eligibility criteria were: aged 30 to 70 years, a clinical diagnosis of T2DM (HbA1c > 48 mmol/mol; > 6.5%) within the previous five to 12 months, no ketosis and a body mass index (BMI) of > 25 kg/m 2 . Exclusion criteria were unstable angina, a myocardial infarction within the previous 3 months and a medical condition that precludes PA (e.g. a foot ulcer). Participants wore an ActiGraph accelerometer for 7 days, and those in one centre also wore a GPS receiver for the same period. STAMP-2 received ethical approval from the South West-Central Bristol NHS Research Ethics Committee (13/SW/0187). All participants provided written informed consent before taking part in the study.
STAMP-2 participants were independent of the ENABLE London study and were recruited from a city with different travel options from central London. Furthermore, the STAMP-2 participants had been recently diagnosed with type-2 diabetes and represented an older (age mean, sd = 58.6, 8.6), less healthy demographic (70.1% obese, BMI mean, sd = 34.4, 7.3), than those in the ENABLE London study. Good predictive performance on this data-set would demonstrate that the algorithm can generalise to other populations and geographical contexts.
Measures of test performance
We report several measures of predictive accuracy in this paper. Firstly, we present an overall accuracy score: the percentage of points we correctly predict overall. This is tempting because it reduces the overall prediction accuracy down to a single number. However, if you have uneven numbers of data-points in each category, which we do, this is not necessarily representative of each mode we report. We therefore also report positive predictive value (the percentage of those points we predict as a mode that were observed as that mode), sensitivity (the percentage of points observed to be a travel mode, which we correctly predict) and F1 score (the harmonic mean of positive predictive value and sensitivity) for each mode separately. To fully understand how our model performs all of these values are useful. For those unfamiliar with such terms we recommend interpreting the F1 score for each mode as a measure of accuracy for that mode. For researchers who would like more information we also present the raw confusion matrix. This table compares counts for observed and predicted mode, and can also be used to calculate accuracy scores.
Training data
We identified 66 participants from the ENABLE study who walked to work, 34 cyclists, 94 vehicle users (grouping together car/van drivers, car/van passengers, taxi users, bus/coach users and motorcyclists) and 132 train users (grouping together underground and over ground train users) ( 165.3 h). These training data were split into two parts, as described above. The subset of the data used to test moving window sizes contained approximately a quarter of the training data (33,529 points). Following testing of 1, 2, 3, 4 and 5-min moving windows, we selected a four-minute window, because a four-minute window resulted in the highest predictive accuracy for active travel modes (Additional file 3). The remaining 98,387 points were used to build the cross-validated model.
Model prediction
In model cross-validation, overall, we correctly predicted 97.3% of points. All five travel modes were predicted with high accuracy (lowest F1 score 93.9; Table 2). The F1 score is the harmonic mean of positive predictive value and sensitivity, therefore a high value represents a high rate of correct identification of both true positives and true negatives.
In comparison with manually identified data (n = 21), overall accuracy was still high at 96.0% of predictions being correct, however, this was substantially driven by the fact that most people spent most of their time stationary (83.7% of the time stationary). Our F1 scores for the other modes were lower than in cross-validation, with the lowest at 75.5 for cycling (Table 3).
When compared with manual identification of travel mode from the STAMP-2 study, our predictions performed well considering the different participants and context (Table 4). We correctly predicted the travel mode of 96.5% of points, but again this is driven by our high accuracy on stationary points, which is the dominant mode (86.8% of time stationary). The poorest performing predictive mode was for cycling, with an F1 score of 69.1% (Table 4).
To understand, how the model miss-classifies in some situations we also present the total time predicted and observed in each travel mode (Table 5). In crossvalidation the times are remarkably close, however in
Discussion
We have developed an accurate predictive algorithm, which identifies five travel modes, including the active travel modes walking and cycling, identifying each travel mode correctly over 90% of the time. Our levels of accuracy in cross-validation out-perform recent similar studies [18,23]. The models developed here are made freely available to apply to similar data in the statistical software environment R [28]. It is unsurprising that our accuracy scores are lower when compared with manual identification of both participants within this study and from elsewhere, than in cross-validation of the training data-set. Identification is most likely to be accurate during defined trips, which is what the training data are comprised of. During full days of data there are likely to be other more ambiguous forms of movement, which may be short in duration, and therefore difficult to identify, or which do not fully represent one of the forms of travel we have included. Our model performs least well for the detection of cycling, and therefore is likely overfitted to our training data. We would therefore recommend manual checking of cycling data, in order to improve accuracy. In this study, cycling represents a relatively small amount of total time, therefore manual checking of this subset represents a much smaller time investment than the full analysis.
Applying the algorithm to the STAMP-2 dataset resulted in similar levels of accuracy of travel mode prediction within a very different group of participants, living in a different environmental setting. This finding suggests that our method may be generalizable to other data-sets and could be used by other researchers without the time-consuming steps of creating new training data. However, until more robust tests have been completed, we would recommend a manual identification checking stage similar to our methods, to verify the generalisability of the method.
We exhibit similar levels of accuracy to the PALMS system, which is a freely available method to process physical activity data [21]. The purpose of the output is somewhat different, though, with PALMS identifying journeys and our method identifying each data point. The preferred method will depend on what the research question is. One advantage our R package does have over the server-based system in PALMS, is that it can be run on a researcher's own machine. Data used for the present (and other similar) analyses is subject to strict data privacy and ethical conditions. Running the analyses on a researchers' own machine, rather than uploading to a server for remote processing, can help avoid problems related to data privacy. Furthermore, our method is open source, meaning all code is freely available online [28]. As a result, other researchers can suggest edits and improvement and contribute to the development of our method, and its utility to the research community. Visual inspection of the data in GIS revealed that much of the disagreement between prediction and manual classification was found at the start and end of journeys. This highlights the challenge of identifying modal shift, i.e. when to switch from one mode to another. This is a limitation of the current method. However, the strength of identifying individual data points, rather than journeys, means that imprecision in identifying modal shift results in small numbers of mis-classified points rather than entire mis-classified journeys. Small numbers of mis-classified points will only have a small effect on total time in each travel mode. Our inaccuracy at the start and end of trips mean that any prediction to a data-set where many short-duration trips are expected in high volume would be expected to yield lower accuracy. Conversely, longer journeys should be able to be identified with greater precision.
It is worth highlighting that some disagreement within our test data-sets may not be true errors. For example, if a participant is stationary, but on a train, it is questionable whether they should be classified as using a train or stationary and it may not matter as long as the rule is applied consistently. However, when comparing the predicted vs manually derived methods this causes disagreement, because during manual classification, stationary points on train lines were termed stationary (for example standing at a train station), yet the algorithm identifies them as "train". A number of the missclassifications between our manual identification and predicted data-sets may therefore be open to interpretation, and may not be true miss-classifications..
The prediction of active travel modes should complement existing analyses of physical activity using accelerometers. Accelerometers have been used with great effect to objectively quantify activity, but they are not without limitations. A well-known problem of traditional PA analyses using accelerometers is that cycling is not recognised as a form of moderate to vigorous activity, because cycling generates relatively low readings on a waist-worn accelerometer compared to other active modes. Identification of cycling from combined accelerometer and GPS data will allow better quantification of cycling as a form of physical activity. Traditional PA analyses will still miss cycling as a form of activity, but our method quantifies it, although with moderate precision. Furthermore, assessment of active travel using our method will help understand how people are active. For example, activity at a single location, e.g. the home/gym, will not be classed as active travel using our method, rather as a stationary travel mode, therefore a participant may show high levels of overall physical activity but low levels of active travel. Incorporation of this extra information will help to understand participant's overall physical activity patterns. For example: if a participant shows increased PA but not active travel it is likely that they have increased their activity at locations such as at home or the gym. If we see no change in PA but an increase in active travel then participants have replaced some of their PA at a location with active travel.
There are several limitations to our study. First, we only identify the active modes walking and cycling, there is no consideration of running, or any other activity. This limitation is based on our study sample, where no commuters consistently used these modes, so we could not include them in our training data. However, walking and cycling represent the most frequently used active transport modes, and other modes were rarely reported by our participants. Any form of running will most likely be identified as walking using our method, due to the high acceleration that running causes on an accelerometer, and so will still be identified as active travel. We therefore feel that this limitation will have little impact on our study, although we would recommend caution in applying our method to a data-set where large quantities of running are expected. Secondly, we have demonstrated our capacity to identify major travel modes, but do not discriminate car travel from public transport. Consequently, if a study is attempting to quantify the use of public transport the current method is inadequate. To address this, we have developed a second model that discriminates bus travel. However, this leads to reduced overall accuracy, because the pattern of speed/activity can be confused between buses and other vehicles (Additional file 4). Further work in this area would need to assess the generalisability of our methods. Thirdly, our method does not identify the purpose of the journey, therefore there is no distinction between leisure and transport-related travel. Another method in addition to our identification step will be necessary to determine the context of the journey. Lastly, though we assess different geographical contexts, both are in the UK. We do not know how well the method would perform elsewhere in the world, where other transport options may be available. Furthermore, we have only tested the current model on adults, and further research could assess how well our model performs on older adults or children, to potentially be of use in a wider group of studies.
Conclusion
In summary we have developed a method to identify travel modes from accelerometer, GPS and GIS data for the ENABLE London study, which successfully predicts over 90% of points tested in a range of contexts. This method can be of use to complement existing analyses of physical activity, and assess active travel alongside physical activity. All code necessary to replicate the analysis, apply the method to other data-sets or predict from our models to other data-sets are provided, to facilitate usage by other researchers. | 2018-09-21T21:15:22.780Z | 2018-09-21T00:00:00.000 | {
"year": 2018,
"sha1": "8318666242a983206a852944aefca4fa11f6c5ad",
"oa_license": "CCBY",
"oa_url": "https://ijbnpa.biomedcentral.com/track/pdf/10.1186/s12966-018-0724-y",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "3d39af3edc2a51d063eb53764106f0c8d1f565b8",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine",
"Computer Science"
]
} |
245080680 | pes2o/s2orc | v3-fos-license | Teleguidance Technology for Endotracheal Intubation: A Scoping Review
Supplemental Digital Content is available in the text.
ndotracheal intubation is a critical life-saving procedure, but many rural and remote settings do not have access to experienced airway providers. Both technical and nontechnical skills are required to reliably and safely secure a patient's airway. A hallmark of critical care training is learning to supervise and coach junior trainees or other providers through a successful intubation without resorting to hands-on intervention. Similarly, clinicians working in tele-critical care practices are likewise accustomed to assisting other providers in performing procedures remotely. Thus, teleguidance technology may be a viable alternative for providing expert airway management consultation and guidance to remote providers who need to perform intubation but are not experts in airway management. While tele-observation, telemedicine, and teleguidance are not new technologies in clinical practice with the first published case from 1974 (1), they have reemerged as invaluable tools during the coronavirus disease 2019 (COVID- 19) pandemic to provide large numbers of patients with expert level care while reducing patient and provider exposure and decreasing the need for personal protective equipment (2). The topic of teleguidance facilitated intubation (TFI) has not been rigorously reviewed in the literature. Here, we perform the first scoping review of the literature involving TFI to elucidate: 1) the feasibility of use and barriers to implementation in clinical practice and 2) whether patient safety, complications, and measures of intubation success (i.e., first-pass success, time to intubation) are improved with its use. A scoping review was chosen in this setting due to the emerging nature of this technology and the limited body of published research.
METHODS
This scoping review was conducted by a research team with expertise in anesthesiology, critical care, medical education, telemedicine, and systematic review methodology. This review adhered to the review methodology outlined by Grant and Booth (3) and the checklist for Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (4). Covidence systematic review software (Melbourne, VIC, Australia) was used for all stages of the review process, including title and abstract screening, full-text review, and data extraction. For the purpose of this scoping review, we considered teleguidance, tele-observation, and telemedicine to be synonymous terms. We defined teleguidance as the use of video and audio telecommunications technology to facilitate procedural guidance of a provider performing an intubation by a supervisor who was not physically present in the location where the procedure was being performed.
Research Questions
TFI was defined as the use of audio-visual telecommunication technology to provide remote guidance for the performance of endotracheal intubation. The aim of this scoping review was to summarize the published literature regarding: 1) the feasibility of TFI and barriers to implementation within clinical practice and 2) examine overall patient safety, complications, and examine whether measures of intubation success (i.e., first-pass success, time to intubation) are improved with TFI.
Search Strategy
A librarian-assisted search was performed on November 10, 2020, using three widely used electronic databases: PubMed, Web of Science, and Excerpta Medica data-BASE. The databases were selected as they are comprehensive covering a wide range of clinical disciplines. The following key words were used in the literature search: teleguidance, intubation, teleprocedures, telemedicine, telehealth, procedures, teleintubation, remote intubation, remote supervision. The selected key words were searched individually in each database. The search was limited to articles published between January 2000 and November 2020 to coincide with the development of the relatively new field of telemedicine. Reference lists of included articles were reviewed, and hand searches of relevant appearing references were performed to identify other literature not captured by the electronic database search.
Inclusion and Exclusion Criteria
Articles were included if they: 1) reported outcomes of a research study, quality improvement effort, or program aimed at implementing, or evaluating TFI; 2) the study was conducted in the operating room, ICU, emergency department, prehospital, or simulation settings; and (3) involved adults. Articles published in languages other than English were excluded. We also excluded reviews, commentaries, editorials, abstracts, and conference proceedings that did not provide original data necessary to address the questions sought in our review.
Data Abstraction
Two reviewers (B.S.L., E.A.B.) independently screened titles and abstracts of identified articles to determine eligibility. The same two reviewers then performed full-text review in duplicate, with conflicts resolved by an independent third reviewer (M.G.C.). Reviewers customized data extraction fields in Covidence to align with the aims of the review; a trial of 10 references were used in a pilot run and both reviewers (B.S.L., E.A.B.) agreed unanimously on inclusion and exclusion.
Two reviewers tested extraction fields for consistency and fidelity to project aims (B.S.L., E.A.B.). The same two reviewers independently extracted data from articles for insertion into required fields. Extracted data were compared between reviewers for consensus prior to finalizing the extraction forms. The following data were extracted from included articles: year and country of publication, project aim, design, setting, key components of the intervention, implementation strategies, challenges/barriers to implementation, and clinical/implementation outcomes. Outcomes extracted included time to intubation (as reported), intubation success, technology feasibility, ease of use, and user satisfaction. Studies were classified based on study design and setting, allowing differentiation between simulation-based studies, clinically based studies, and proof of technology from prospective or retrospective designs. The study question, number of participants and their level of training, video conferencing software, airway equipment used, and outcome and results were also extracted.
RESULTS
The combination of search terms with selection criteria and limits yielded 255 studies. Of these articles, 48 duplicates were removed, leaving 226 studies for title and abstract screening ( Fig. 1).
Of these, 165 were excluded for not meeting initial inclusion criteria (i.e., title or abstract indicated nonadult population, noncritical care setting, or literature review), a subsequent 25 were excluded after screening, resulting in 17 articles that were included in the final review.
Twelve of the studies were conducted in a simulation environment (5, 6, 8, 9, 11-16, 18, 20) and five were performed in clinical environments (7,10,17,19,21). Only three of the 11 proof of technology demonstrations were performed in a clinical environment (17,19,21). Of the prospective simulation studies, randomization of participants was performed in four cases (5,8,9,16) and method of participant allocation was not reported in one (6). The single prospective clinical study randomized participants to TFI versus on-scene supervision (7).
The average number of participants for simulation studies, including proof of technology demonstrations, was 18.6, with the three largest prospective simulation trials having 50 (5), 48 (9), and 46 (6) participants each. The average number of patients for clinical studies was 50. The only prospective clinical study had 25 patients (7). The largest study in the clinical environment had 206 patients (10).
Five studies were performed in a clinical environment, three were proof of technology studies (17,19,21); All reported that the technology was feasible for remote supervision of intubation. The one clinically based randomized trial reported no difference in time to intubation or success rate of intubation of emergency department patients when performed by novice operators (emergency medicine residents) with in-person or remote supervision (7). One observational study showed a 71% first-pass success rate for intubations teleguidance (no comparison was available for the success rate of intubations not performed with teleguidance) (10).
All of the simulation-based prospective trials included a measure of time to intubation as an outcome (5)(6)(7)(8)(9)16). Time to intubation was measured very heterogeneously and was not reported in the clinical studies. In the one prospective clinical trial, time to intubation was defined as beginning from when the laryngoscope was picked up to confirmation via auscultation, and success was determined by end-tidal Co 2 monitoring at the discretion of the supervising attending (7). For the two prospective studies that compared tele-presence to no supervision, one study found that TFI had a faster time to intubation (5, 7) and one reported a higher success rate of intubation and faster time to intubation with TFI (5). All four of the prospective studies that compared in person supervision to telepresence reported no difference in time to intubation (6,8,9,16), and one of these four studies also reported no difference in intubation success rate (8). One simulation study did not provide a statistical analysis (16).
Adverse events were not widely reported in the reviewed studies. The one prospective clinical study reported no difference in frequency of esophageal intubation and no difference in complication rate between teleguidance and on-scene supervision (7). The one observation study, which did not have a control group, and in which only 31% of users used teleguidance prior to their first intubation attempt reported an adverse outcome rate of 24%. Adverse events were associated with increasing number of intubation attempts, the most frequent adverse event was hypoxia (18%), followed by hypotension (6%), cardiac arrest following intubation occurred in 2.4% of patients (10).
Seven of the studies used noncommercially available videoconferencing software (Supplemental Table 1, http://links.lww.com/CCX/A858), while the remaining studies used a variety of commercially available videoconferencing software including Skype, Apple FaceTime, Tango, Google glasses, Adobe Connect, Vidyo link, and Microsoft Net Meeting. Among the commercially available software systems used, all are currently available except for Microsoft Net Meeting (17). One study that compared video quality for TFI between three different brands of videoconferencing software (Apple FaceTime, Skype, and Tango) found that while all three video services provided adequate fidelity to supervise intubations, Apple FaceTime was rated as having the highest video and audio quality (19). The type of videolaryngoscope used in the studies varied with Storz C-Mac the most prevalent (53%), followed by Glidescope (23.5%) (Supplemental Table 1, http://links.lww.com/CCX/A858). Two of the twelve simulation studies used direct laryngoscopy with telehealth glasses that afforded the supervising physician a first-person point of view (5,18).
Only three studies assessed user responses to the technology or reported technical issues that were encountered during use. One study that surveyed users of telehealth glasses for TFI reported that the technology was "bulky and uncomfortable, " "distracted from the learning, " and had other technical issues. Another study evaluated educational value of the TFI (rated 9.5/10) and reported that 11 of 12 users had no issues obtaining the laryngoscope view or hearing the guidance provided (18). A similar study surveyed users regarding quality of imaging and reported that the average rates were 7/9 for video and audio quality and 7.8/9 for overall quality (20). Two studies evaluated surveyed users as to whether they would use the technology again and both had high ratings for continued use of TFI (19,20).
DISCUSSION
Teleguidance for medical procedures is a growing technology and an active field of investigation (22)(23)(24)(25)(26)(27). The majority of published literature has focused on proof of technology demonstrations that allude to the benefits that might be achieved with clinical use. While limited, we believe that the body of literature allows us to begin to answer the study questions regarding feasibility, safety, and efficacy of TFI and suggests avenues of future investigation. Many of the studies reviewed were small simulation-based proof of concept studies, the small size of these studies limits the conclusions that can be drawn from them. The prospective simulation studies using airway trainers and randomization of participants provide reassurance to the robustness of the results and the one prospective clinical trial was of moderate size. Overall, the quality of the studies is mixed but provides an adequate starting point for this review.
Is TFI Feasible Given Existing Technology and What Are the Barriers to Implementation?
The studies selected to answer this question span both the clinical realm and simulation environment. These studies demonstrate that TFI is feasible given the technology available at the time that the studies were performed (5-9, 11-15, 20, 21). While limited fidelity of the video and audio technology was reported in three studies, the majority did not indicate this as being a barrier to successful implementation (12,19,20). Ergonomic challenges with tele-glasses were discussed in one study and did present a distraction and a potential limitation for implementation of this technology for TFI (18). While technological limitations were infrequently reported in the studies, it is important to recognize that the majority of the TFI systems evaluated were customized setups that are not commercially available. The lack of availability of these systems without customization poses a barrier for users looking for "off the shelf " technology ready to implement within their own clinical environments.
A challenge in all of the studies examined was seamless integration of the video laryngoscope imaging with the videoconferencing software. However, even without full integration of these devices, the studies demonstrated that available technology was able to provide a view of a video laryngoscope screen as well as bidirectional communication supporting the feasibility of TFI. It is important to recognize that the majority of studies used technology more than 5 years old and that telemedicine technology has advanced markedly since that time especially during the COVID-19 pandemic. Based on the authors' clinical experience direct supervision of inexperienced operators performing intubations outside of the operating room, we believe that the ideal TFI system should not only integrate of video laryngoscopy with videoconferencing but would also provide visualization of the larger clinical environment. This additional visualization (via camera) should allow for hemodynamic monitoring and visualization of patient positioning and intubator technique. While tele-glasses technology (e.g., Google Glass) is appealing because of the point-of-view images provided from the standpoint of the intubator, they lack crucial environmental information that can be important for optimizing intubating conditions (28).
Does TFI Improve Patient Safety and Intubation Efficacy?
While the available literature is limited, studies indicate that TFI appears to improve patient safety with no additional complications reported beyond those that are reported with in person supervision of intubation (7). In the observational study of TFI performed in the emergency department (10), there was no control group without TFI for comparison, and the rates of hypotension and hypoxia were lower than reported studies in similar environments with in person supervision (29,30). The benefit of TFI in this study was further supported by the findings that adverse events were associated with increasing number of intubation attempts and a decreased rate of complications when a www.ccejournal.org December 2021 • Volume 3 • Number 12 teleconsultation was performed prior to the first intubation attempt (10). Consideration of the comparison groups used in the evaluations studies is important for evaluating the efficacy of TFI. Studies that compare TFI to in-person supervision provide information regarding the comparability of the two methods of oversight, while studies that compare TFI to no supervision evaluate the added value of the technology compared with no oversight at all. Two studies performed in a simulation environment indicated that time to intubation and success rate with TFI were no different than with in-person supervision (6,9), and these findings were confirmed in a study performed in the clinical setting (7). The existing body of literature does not allow determination of superiority when comparing teleguidance to in-person supervision. The heterogeneous nature of the technology used and outcomes selected make comparison across studies difficult. For novice operators performing intubation in simulation environments, TFI was demonstrated in two studies to have a faster time to intubation and higher first-pass success rate when compared with not having any supervision at all (5,8). A reduced rate of complications seen in the observational study when TFI was used prior to first intubation attempt also supports these findings (10). Teleguidance could also be valuable in the setting of failure to intubate; the virtual airway expert could direct a novice through a difficult airway algorithm. This virtual role has been suggested as a component of a pediatricspecific difficult airway response team (28). Teleguidance could also be used to guide performance of a surgical airway. This use of teleguidance for this indication has not been reported, but there are examples of using teleguidance to facilitate other emergent surgical procedures (31) and even damage control laparotomies (27). Extrapolating from the TFI literature, teleguidance could help improve outcomes and decrease complications during a difficult airway situation.
Limitations
Our scoping review has a number of limitations. The number of published studies that have evaluated TFI is limited, and they were performed in a small number of environments by only a few investigators. Study selection was limited to available articles contained within three major medical databases. Studies that may be available in the gray literature including technical reports, proceedings, and theses were not examined. Finally, only articles published in English were included.
There are limitations to the studies themselves, which may impact the findings. The proof of technology studies are small, single center and often employed customized equipment. The efficacy studies were limited by small study size, single-center evaluation, and lack of blinding. Adverse events were only reported in two studies (7,10) and which might represent area source of publication bias that could impact conclusions regarding safety of the technologies. None of the studies examined the role of telemedicine in the setting of the difficult airway or if failure to intubate occurred while using TFI.
Future Directions
Given the advances in telemedicine over the last decade, there is a need to reevaluate TFI that incorporates recent technology. There is a need to fully integrate video laryngoscopy, telecommunication technology with imaging of the clinical environment. Evaluations of this integrated TFI technology first be trialed in a simulation environment with novice intubators and then in randomized controlled studies within clinical environments where experts in emergency airway management are immediately available. Objective measures of intubation success (e.g., first-pass success rate, time to successful intubation) should be collected together with feedback from users regarding challenges and areas for improvement. Practical measures of use including portability, time to setup, troubleshooting, and disinfection should also be evaluated. Robustness of the technology under actual clinical conditions in a variety of diverse environments will be essential to establishment of a successful platform for a wide range of users.
CONCLUSIONS
Based on the available literature included in this scoping review TFI appears to be a feasible, safe, and effective approach when in person supervision is unavailable. TFI appears valuable both for novice operators and may serve as an "expert consult" for trained providers. Widespread implementation of a robust TFI technology has the potential to improve patient safety and outcomes in many existing clinical environments where airway management experts are not immediately available 24/7. Teleguidance may be valuable as an educational tool providing trainees graded autonomy while ensuring an appropriate degree of supervision. While much work remains to further develop, critically evaluate, and integrate TFI into clinical care, TFI holds great promise for improving upon the existing practice for critically ill patients. | 2021-12-12T16:50:17.334Z | 2021-12-01T00:00:00.000 | {
"year": 2021,
"sha1": "8e24c15a7fdadabda9f623d893b224bf818f33ce",
"oa_license": "CCBYNCND",
"oa_url": "https://doi.org/10.1097/cce.0000000000000582",
"oa_status": "GOLD",
"pdf_src": "MergedPDFExtraction",
"pdf_hash": "bebe9289a04236ad7946eefc31228777468fa66b",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
213038550 | pes2o/s2orc | v3-fos-license | Sino-Russian Oil Security: Evolution and Structure
Sino-Russian oil cooperation is of great significance to Sino-Russian oil security. This research evaluates the evolution pattern of China’s external oil supply risk and Russia’s export risk in 2005-2016 and the impact of Sino-Russian oil cooperation on Sino-Russian oil security. It was obtained: The concentration degree of China’s oil import is gradually decreasing, Russia plays an important role in balancing China’s external oil supply and reducing its concentration. When bringing in oil export potential (OEP), China’s external oil supply risk shows the trend of wavy increase, and the continued declining OEP of Russia increases the risk in a certain degree. When bringing in oil price (OP), China’s external oil risk shows an M-style decrease, and the sanction against Russia by western countries led to cliff drop of OP, which led to sharp fall of risk in the short term. The concentration degree of Russia’s oil export shows a wavy increase. The increasingly important role of China in the Russian oil export market (OEM)exacerbates the concentration degree. When bringing in oil import potential (OIP), Russia’s oil export risk shows an M-style increase, and the positive effect of China’s exuberant OIP has largely counterbalanced the negative effect of the continued declining OIP of other major Russian oil export markets. When bringing in OP, Russia’s oil export risk shows an M-style increase, especially under the sharp fall of OP in recent years, the risk increases dramatically.
Introduction
Oil resources is the primary energy as well as strategic resource in the world, which plays an important role in regional energy security, economic security and even national security [1][2][3]. With the rapid economic growth, China has become the world's second largest oil consumer and largest oil importer. However, for a long time, about 60% of China's imported oil is from the Middle East and West Africa [4], China's oil import sources (OIS) are too concentrated. Meanwhile, the geopolitical environment in above areas is complicated, which has become a major potential risk factor that restricts China's external oil supply security. Therefore, Chinese government has been implementing the strategy of diversification on oil import, and oil cooperation with Russia is an important part in this strategy [5]. Russia is an energy superpower and one of most leading energy exporters in the world. However, more than 60% of its exported oil are exporting to Europe, the highly concentrated oil export market (OEM) and complicated Russia-European relationship are direct factors that restrict Russia's oil export security. Therefore, strengthening energy cooperation with the Asia Pacific region, especially with China, has become an important strategy for Russia to improve its energy security [4] The energy strategy of China and Russia are highly compatible, with the promotion of the two government, the traded volume of oil has been increasing rapidly. In 2016, the volume of oil that China imported from Russia exceeded 50 Mts., Russia overtook Saudi Arabia and became China's largest oil supplier, meanwhile, China became At present, existing research mainly study Sino-Russian oil cooperation from the qualitative perspective [4,[6][7][8], few studies analyzed the impact of Sino-Russian oil cooperation on Sino-Russian oil security from the quantitative perspective, resulting in lack of objectivity and scientific city of existing results. Therefore, in order to meet above deficiencies, based on HHI and definition of energy security, we scientifically evaluate the evolution pattern of Sino-Russian oil security, and the impact of Sino-Russian oil cooperation on it, which can provide theoretical guidance and scientific support for promoting Sino-Russian oil cooperation.
Models and methods
Most scholars define energy security from the perspective of energy supply and energy use [9][10][11], little attention is given to the energy export security. However, for energy exporters, such as Russia, energy export security is also a crucial factor to its energy security and economic security. Therefore, we believe energy security should be divided into energy supply security of energy importers and energy export security of energy exporters, thus we established the external oil supply risk index for importers and oil export risk index for exporters.
The external oil supply risk index
HHI is the most widely used index to evaluate energy security, it is based on the principle of diversity [12]. Generally, the more diverse the energy import sources or energy export markets, the smaller the HHI, the higher the energy security. HHIr represents energy security risk index of region r, wrj indicates the share of energy region r import from or export to region j to its total imports or exports (equation (1) and (2)), showing the distribution of OIS or OEM.
(1) (2) For oil importers, the distribution of OIS, their corresponding oil export potential (OEP), and oil price (OP) are critical factors that directly affect their external oil supply security. The more diverse of OIS, the greater of OEP and the lower OP, the smaller the external oil supply risk. Therefore, based on HHI, we bring in the factor of OEP and OP into the external oil supply risk index. OEP is indicated by the share of OIS's proven oil reserve to the world proven oil reserve (SPR), and its reserve and production ratio (RPR). OP is indicated by international annual average oil price. i=1…m, which indicates China's OIS.
The oil export risk index
For oil exporters, distribution of OEM, their corresponding oil import potential(OIP)and OP are main influencing factors of their oil export security.OIP is indicated by the volume of oil consumption (OC)and their volume of oil imports (OI). The bigger OC, OI, OP, the smaller oil export risk. j=1…n, indicates Russia's OEM.
China's external oil supply risk that includes oil import sources and oil export potential (HHIC-OEP).
In 2005-2016, HHIC-OEP shows a wavy upward trend, indicating that the export potential of OIS has an important impact on China's external oil supply security ( figure 3). During this period, several main OIS of China show a steady or wavy declining trend. For instance, Saudi Arab, whose SPR declined from 19.22% to 15.61%, RPR declined from 75.02 to 59.11; Russia, whose SPR declined from 7.59% to 6.05%, RPR declined from 29.81 to 25.54; In addition, SPR and RPR of Oman, Angola and Kuwait all show a wavy declining trend, thus leading to an increasing trend of HHIC-PE. It should be noted that with the increasingly important role of Russia in China's oil import market, Russia's decreasing OEP will directly influence China's external oil supply security. According to the evolution of OP, before 2008, OP increased rapidly from $67 a barrel to $108.42a barrel, however, the outbreak of financial crisis in 2008 led to a severe oversupply of oil in international oil market and OP presented a cliff fall, which dropped from $108.42a barrel to $68.99 a barrel in 2009, leading to dramatic decrease of HHIC-OEP-OP. Along with the recovery of the world economy, the demand for oil gradually resumed, the huge shortage of oil supply led to a rapid rebound of OP, in 2011, OP rose to $118.71 a barrel, which exceeded that in 2008. The rapid rise of OP increased the cost and difficulty for China to import oil, causing HHIC-OEP-OP to rise rapidly. However, due to slowdown of worldwide economy, the shale gas revolution and the increase of oil supply in major oil exporters, especially the sanctions against Russia by the western countries because of the Crimea incident in 2014, OP fell dramatically and successively to $43.73 a barrel. The rapid declination of OP greatly reduced HHIC-OEP-OP in the short term.
Russia's oil export risk
3.2.1. Russia's oil export risk that only include oil export markets (HHIR). HHIR firstly express an upward trend before 2010, then a wavy downward to 2013, and an upward again ( figure 6).
Before 2010, Russia's oil exports showed a trend of concentration (figure7), though number of Russia's main OEM did not change very much, but their total share increased from 55.24% to 59.01%.
Russia's oil export risk index that include soil export market sand oil import potential (HHIR-OIP).
HHIR-OIP shows an M-style trend, which reached its peak in 2009 and 2014 respectively, and then started to decline (figure 8). Before 2008, HHIR-OIP shows a small but steady increase, which on one hand was due to concentrated trend of Russia's OEM during this period, on the other is due to the decreasing trend of OIP of Russia's main OEM. During this time, Russia's biggest OEM Netherlands, whose OC decreased from 50. The abrupt drop in OIP led to a sharp rise in Russia's oil export risk index. After 2009, with the gradual recovery of the world economy, the demand for oil recovered slightly, but under the shale gas revolution, saturation of the world's oil market and oil overcapacity, the declining trend in Russia's main OEM failed to change, HHIR-OIP shows a decline shortly and then tend to increase again. After 2014, due to the sanction against Russia by the western countries, OP decreased dramatically, which stimulated the oil demand to a certain extent. Russia's main OEM, such as Netherlands, Italy, Poland, Republic of Korea, etc., their demand for oil started to rebound, OC and OI all tend to increase, OC and OI of Republic of Korea increased from 108 Mts. and 125 Mts. to 125 Mts. and 144 Mts. The rebound of OIP led to downward trend in HHIR-OIP. | 2019-11-28T12:16:43.350Z | 2019-11-22T00:00:00.000 | {
"year": 2019,
"sha1": "e879923859805b7b5da5e8ab4ef14b3983cb870a",
"oa_license": null,
"oa_url": "https://doi.org/10.1088/1755-1315/381/1/012100",
"oa_status": "GOLD",
"pdf_src": "IOP",
"pdf_hash": "905ce1c8063d56a7d80edf75e2e2bbd0b096f32b",
"s2fieldsofstudy": [
"Economics"
],
"extfieldsofstudy": [
"Physics",
"Business"
]
} |
259980009 | pes2o/s2orc | v3-fos-license | The Relationship Between Harmonious Familiy With Loneliness
: The purpose of this study was to know the relationship between harmonious family with loneliness in 10 th Grade 1 High School Pandeglang. Research conducted using quantitative methods with correlational types. The population in this study were students of 10 th Grade 1 High School Pandeglang with a sample of 218 students studied. The sampling technique using probability sampling include simple random sampling. The instrument used to measure harmonious family and loneliness was a questionnaire created by the researcher. By the result of data analysis using statistical method with Product Moment analysis showed that the value of r-count < r-table was 0,01< 0,05 then the null hypothesis (Ho) was rejected and the alternative hypothesis (Ha) was accepted at a significance level of 5% so that it can be interpreted that there is a relationship between harmonious family with loneliness of the 10 th Grade 1 High School Pandeglang in Academic Year of 2022/2023.
Introduction
Students get their first education from the family. Thus, the family plays a very important role in shaping the personality of students according to the values that can help in their development, so that students can grow to be strong, independent, not easily emotional, adaptable and other positive things. Families, especially parents must pay attention to children's personality problems, children must also get enough attention from their parents. With the attention given and parenting that is suitable for children, it can help children optimize their potential (Saputra, 2021: 2) .
The family also has a very important role for students in understanding their significance. According to Openshaw, the quality of interaction between family members is seen from how they socialize positively between family members, which is one of the characteristics of the functioning of a family (Savitri, 2018: 55) . In this regard, Stewert argues that the family has a more complex function, which is contained in the concept ISSN : Print 2598-3199 -Online 2598-3210 (The Relationship Between Harmonious Familiy With Loneliness) of family characteristics, ability to adapt, communication, self-acceptance , social support, affection, quality time , ability to express emotions, problem solving , morality or religiosity, family climate, hatred and concern between each family member (Savitri, 2018:3) .
Septiana explained that a harmonious family is a family where all family members feel happy, this can be characterized by reduced tension, disappointment and feeling satisfied with all circumstances and existence itself which includes physical, mental, emotional and social aspects (Masi, 2021: 215) . The integrity of parents or mothers and fathers in a family is very important, because the influence, guidance, direction and value system instilled can always be respected, influencing attitudes and behavior patterns in their children. According to Nisfianoor & Yulianti in (Masi, 2021: 215) if in a family there is no father or mother, one of them dies, gets divorced or often goes out of town for months, then the family can be said to be incomplete. a touch of affection, a sense of security, attention and make their family a shelter. If the child does not get all of that, then according to Saqinah the child can experience neurotic or emotional disturbances such as withdrawing from peers, appearing lonely because they do not get enough attention, lack of self-confidence, lack of enthusiasm in life and difficult interact with others (Masi, 2021:215) .
As social beings, students need help and guidance from others, therefore students are expected to have a good relationship with the surrounding environment. The relationship in question is a healthy and quality relationship. The absence of a good relationship that exists can allow loneliness in students. As explained by De Jong Gierveld who said that loneliness or loneliness can occur due to situations of a lack of quality relationships forged with other people. Meanwhile Baron and Byrne argue that loneliness is a feeling of unhappiness arising from cognitive and emotional reactions caused by the desire to establish intimate relationships that cannot be achieved (Susanti, 2019: 7) .
Mund & Neyer explained that loneliness in students can have a negative impact, such as a decrease in wellbeing in the form of sleep problems, depression, appetite disorders, to the point of being pathological (Yuditha et al., 2022: 154) . Thus, Hidayati & Muthia explained that disturbed psychology is the result of a lack of quality and quantity in establishing social relationships (Yuditha et al., 2022: 154) .
Reporting from an IDN Times article on July 28 2022 entitled "survey: Loneliness and self-harm are rife in Indonesia" in commemoration of mental health day which is commemorated in May 2021, a suicide prevention community called into the light works together with Change.org Indonesia to conduct a survey. From the survey results, it can be seen that 98% of the 5,211 participants involving adolescents aged 18-24 years and adults aged 25-34 years experienced loneliness (Putra, 2021) . Based on this phenomenon, the researcher conducted a preliminary study by conducting a simple discussion about loneliness with the resource person , Mrs. Hj. Siti Nuraeni, S.Pd as a counseling teacher at SMA Negeri 1 Pandeglang. From the results of the preliminary study the counseling teacher provided information that there were some students who had the characteristics of loneliness , this could be obtained from the observations made by the counseling teacher while providing guidance and counseling services. Loneliness students show s behavior , one of which is by blaming themselves for bad social relations and various things that are beyond their control. Students who experience these characteristics usually have low self-esteem and self-awareness. Students who experience loneliness usually talk more about themselves and pay less attention to other people's stories. Based on the results of this study, guidance and counseling services are needed to reduce the loneliness experienced by students. The counseling services provided are an attempt to reduce the level of loneliness in students.
Method
The research method used in this study is a quantitative method with a correlational type. According to Arikunto in (Pratomo & Gumantan, 2020: 12) the correlational method has the aim of knowing how close the relationship is in the variables studied. The variables in this study consist of the independent variable or X which is about family harmony and the dependent variable or Y is loneliness.
This research was carried out in stages, starting with the preparatory stage to the end in preparing the report, according to Arikunto (2013: 61) the flow of the research stages is as follows: (a) The preparatory stage, contains the formulation of the problem, determines the research variables, looks for descriptions and theoretical foundations with conduct a literature study, determine and compile the measuring instruments to be used. Determining research locations and conducting research instrument tests. (b) Implementation of research, during the implementation of research carried out the process of collecting data on predetermined samples, explaining the objectives of the research conducted and asking for consent from respondents to fill out a research questionnaire. After obtaining the consent of the respondent, the next step is giving a questionnaire to students with the aim of collecting data. At this stage it is included in CHAPTER I, II and III. (c) The final stage of the research, the final stage being the stage where the researcher interprets, then analyzes and manages statistical data results based on theory which then draws a conclusion. This stage is listed in CHAPTER IV and CHAPTER V.
The characteristics of the population in this study were early adolescents in the age range of 12-15 years with male and female gender, totaling 480 students of class X at SMA Negeri 1 Pandeglang. Sampling was carried out using probability sampling techniques including simple random sampling . This study also used the Slovin formula to determine the required sample size, namely 218 class X students at SMA Negeri 1 Pandeglang.
Data collection in this study used a questionnaire technique as an instrument. According to Tohirin, the technique of collecting data using a questionnaire has a way of providing a set of written statements or questions to be filled in by respondents (Wulandari et al., 2020: 11) . The data obtained from the questionnaire instrument can be processed to obtain interval values that can be interpreted in data analysis. The study uses parametric analysis techniques, namely the data obtained from the instrument produces interval and ratio data that meet several criteria. The data to be processed must be normally distributed, homogeneous and linear. The data that has been obtained is then processed using descriptive and inferential statistics. Inferential statistics consist of analysis prerequisite tests and hypothesis testing. While the prerequisite test consists of a normality test and a linearity test. The instrument lattice for family harmony and loneliness can be seen in the table below: The above instruments were then tested again on 35 class X students at SMA Negeri 1 Pandeglang. Calculation of the validity test using SPSS 20 showed that there were 33 items on the family harmony variable and 28 items on the loneliness variable which were declared valid or usable. The instrument reliability test was carried out using the Cronbach Alpha calculation and obtained a result of 0.986 for the family harmony ISSN : Print 2598-3199 -Online 2598-3210 (The Relationship Between Harmonious Familiy With Loneliness) variable and 0.969 for the loneliness variable with the interpretation of both statement data being declared reliable because it has an alpha value > 0.600.
Picture of Family Harmony
The following is a description of the family harmony of class X students of SMA Negeri 1 Pandeglang obtained from the results of collecting data on 218 students of class X SMA Negeri 1 Pandeglang. The table above shows that family harmony in class X students at SMA Negeri 1 Pandeglang in the high category has 137 students or 63%. Family harmony in the moderate category has 81 students or 37%. Whereas with the low category there are as many as 0 students or 0%.
Description of Loneliness in Students
The following is a description of the loneliness of class X students of SMA Negeri 1 Pandeglang obtained from the results of collecting data on 218 students of class X SMA Negeri 1 Pandeglang. The above data was obtained from data collection on students who were the research sample. Students with a high loneliness category have 67 students or 7%, students with a moderate level of loneliness have a total of 144 students or 66%, while students with a low category have 58 students or 27%.
Normality test
The normality test is carried out to identify normally distributed data by looking at the 2-talled significance value for each variable which must have a value > 0.05 . The normality test is obtained using the Kolmogrov sminrov method , decision making is done by comparing Asymp.sig or pu P-values . If Asymp.sig or P-value is greater than 0.05 then the data is normally distributed. On the other hand , if the Asymp.sig or P-value is less than 0.05 , the data is declared not normally distributed. The results of the normality test in this study can be seen in the table below.
Linearity Test
Linearity test was carried out to see a linear relationship from the data obtained. The relationship between variables is declared linear if it has a significance value of more than 0.05. conversely , if the significant value is less than 0.05, then the relationship between variables is declared non-linear. The results of the linearity test in this study can be seen from the table below.
Data on Family Harmony and Loneliness
The table above shows that the linearity test between family harmony and loneliness has a linear relationship because the sig deviation of linearity has a value of 0.216 which is greater than 0.05.
Hypothesis testing Simple Linear Regression Test
This study uses hypothesis testing with a simple regression test technique. The aim is to find the regression equation on the variable family harmony to loneliness on the basis of decision making as follows: (a) If the sig (2-tailed) < 0.05 then it is rejected, which means that variable x has a relationship with variable y. (b) If the sig (2-tailed) value is > 0.05 then it is accepted, which means that variable x has no relationship with variable y. A simple linear regression test was carried out after the research instrument was declared valid and reliable as well as normal and linaer. The results of the simple regression test in this study are as follows: The results of the simple regression test show the acquisition of a significance value of 0.01 <0.05 which means (H0) is rejected and (Ha) is accepted. This means that there is a relationship between family harmony and loneliness.
Correlation Coefficient Test (R) and Determination Coefficient Test (R ² )
Analysis of the correlation value coefficient was carried out to see how much the independent variable (x) had an effect on the dependent variable (y) which was expressed in percentage form. The small value of ( R² ) means that the capacity of the independent variable to explain the very limited dependent variation. If the value is close to 1, then the independent variable provides a lot of data needed to predict variable variations. The results of the correlation coefficient test (R) and the Determination Coefficient Test (R ² ) can be seen in the The table above interprets the results of the correlation test between family harmony and loneliness , which obtains a correlation value (R) of 0.660 based on the correlation coefficient test. Researchers can predict the level of relationship between family harmony and loneliness. The coefficient of determination (R ² ) in this study obtained a value of 0.435. Based on these results, it can be seen that the family harmony variable correlates with loneliness by 43.5%, the contribution of family harmony with loneliness is at least 56.5%, other factors not examined by researchers that can contribute to loneliness .
Next, to find out the direction of family harmony with loneliness in this study can be seen in the regression equation in the table below: The results of the analysis of the regression equation in this study obtained a constant value of 155.476, while the value of family harmony was -0.634 so that the regression equation obtained Y = 155.476 + (-0.634) X. These results mean that a constant value of 155.476 indicates a constant value on the harmony variable family. The value of the regression coefficient X is -0.634 which states that for every 1% addition to the value of family harmony, loneliness decreases. Thus, the regression coefficient is negative, so that the higher the family harmony, the lower the loneliness experienced by students.
Correlation Hypothesis Test
In this study, researchers used Karl Person Product Moment analysis. The aim is to determine the degree of relationship and the contribution of the family harmony variable to the loneliness variable. The basis for making the decision is if the significance value is < 0.05 then it is correlated. Conversely, if the significance value is > 0.05 then it is not correlated. The results of the product moment correlation analysis in this study can be seen in the table below. 1 218 The table above shows the correlation between family harmony and loneliness with a significance value of 0.01 and a personal correlation of -0.660. That is, a significance value < 0.05 means that there is a correlation between family harmony and loneliness. The personal correlation value shows -0.660 which means that the relationship between family harmony and loneliness is a negative relationship at the level of interpretation of a strong relationship, because the value of -0.660 is included in the coefficient interval of 0.60-0.799.
Discussion
This study aims to determine the relationship between family harmony and loneliness in class X students at SMAN 1 Pandeglang. Before doing statistical analysis using product moment analysis . Where before testing the hypothesis the researcher first carried out a prerequisite test where this prerequisite test consisted of a normality test to find out whether the data obtained was normally distributed and a linearity test to find out whether the independent variable has a linear relationship with the dependent variable. The normality test results obtained sig. = 0.200 whose value is > 0.05. Because the significance value is more than 0.05, it can be concluded that the data is normally distributed. The linearity test results obtained sig. = 0.216 > 0.05, which means that the data has a linear relationship.
Based on research data regarding the relationship between family harmony and loneliness in class X students at SMAN 1 Pandeglang , the results show that there is a relationship between family harmony and loneliness . Based on the results of the correlation calculation test, both have a value of r = (-0.660) with a significance value of 0.0 00 1 < 0.05, which means that the two variables are family harmony and loneliness has a negative relationship where the higher the level of family harmony , the lower the loneliness in class X students at SMAN 1 Pandeglang , and vice versa.
The explanation above is in line with the results of research conducted by (Surya, 2021: 77) with the research title "Relationship Between Parental Involvement and Loneliness in Adolescents Living with Single Parents" with the results of the study that there was a significant negative relationship between parental involvement and loneliness in adolescents who live with a single parent, where the higher the involvement of parents, the lower the level of loneliness in adolescents who live with a single parent. Vice versa, the lower the parental involvement, the higher the level of loneliness in adolescents who live with a single parent.
This statement is also supported by the results of research conducted by (Savitri, 2018: 59) which shows that there is a significant negative relationship between family functioning and loneliness . This is evidenced by statistical calculations with a significance value (p), namely 0.038, the value of R = -0.202. Family functioning contributes 4% to the emergence of loneliness in research subjects (r² = 0.04), which means that 96% of other factors contribute to the emergence of loneliness .
Fraenkel in (Subianto, 2013 : 343) stated that schools are not just a place to carry out learning and teaching activities but schools are also a place that is responsible for molding students and students into individuals who have superior knowledge in all aspects including the development of identity, character and personality possessed by students and his female students. For this reason, in handling it, especially in the school environment as a guidance and counseling teacher, it has a very important role and responsibility in forming the self-development of a student .
The results of the study show that the level of family harmony is in the high category of 63%. Meanwhile loneliness is in the moderate category with a score of 66%. Based on the results of this study, guidance and counseling services are needed to reduce the loneliness experienced by students. The counseling service that is carried out is one of the efforts to reduce the level of loneliness in students, then providing assistance from other parties can also support optimal student development so as to avoid loneliness in the long term, the responsible party is the student's parents/guardians, scope of students and other stakeholders (Aditiono et al., 2022 :306 ) . Researchers designed guidance and counseling programs in the personal-social field as the implications of this research. The program created is also expected to help students reduce the loneliness they experience.
Conclusion
Based on research that has been done regarding the relationship between family harmony and loneliness in class X students at SMA Negeri 1 Pandeglang for the 2022/2023 academic year, it is concluded that the general picture of family harmony of class X students at SMA Negeri 1 Pandeglang is included in the high category . The general description of loneliness in class X students at SMA Negeri 1 Pandeglang is included in the moderate category . There is a relationship between family harmony and loneliness in class X students at SMA Negeri 1 Pandeglang with a correlation value of -0.660 and a significance level of 0.01 <0.05 with a strong correlation coefficient. Based on these data, it shows that there is a negative relationship between family harmony and loneliness . This means that the higher the level of family harmony, the lower the loneliness ISSN : Print 2598-3199 -Online 2598-3210 (The Relationship Between Harmonious Familiy With Loneliness) of students, and vice versa. The R square value is 0.435 which means that the relationship between family harmony and loneliness is 43.5% and the remaining 56.5% comes from other variables not examined by researchers which can contribute to loneliness .
There is a negative relationship between family harmony and loneliness , which has a correlation value of -0.660, which means that the higher the value of family harmony, the lower the loneliness of students. This is also corroborated by a picture of family harmony with a high category of 63% of students, a medium category of 37% of students and a low category of 0% of students. An overview of loneliness with a high category of 7% of students, a moderate category of 66% and a low category of 27%. Based on these findings, personal and social guidance services are needed to help students optimize the level of family harmony and reduce loneliness in class X students at SMA Negeri 1 Pandeglang. Therefore, it is necessary to design a personal and social guidance program which is the implication of this research, which can then be implemented by guidance and counseling teachers at SMA Negeri 1 Pandeglang. | 2023-07-20T15:02:16.776Z | 2022-12-30T00:00:00.000 | {
"year": 2022,
"sha1": "c7c0f08fd3b1b28dc462aa6af7be510a9765c40e",
"oa_license": null,
"oa_url": "https://doi.org/10.23887/bisma.v6i3.55879",
"oa_status": "GOLD",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "0aa8b8455de47d3d782f5dfd4caf689f5407f8eb",
"s2fieldsofstudy": [
"Education"
],
"extfieldsofstudy": []
} |
252838789 | pes2o/s2orc | v3-fos-license | Oscillation of neutrino in a vacuum with mixing flavor
: We developed multiple equations to observe the two and three flavors of neutrino oscillation with the mixing angle based on L/E=0.1 to 0.9 in this study. In different settings, the nature of the neutrino oscillation probability was discovered to be varied in different equations. The observation indicates increasing likelihood in one equation and decreasing probability in the other equations in two flavor oscillation neutrinos. To characterize the probability of neutrino oscillation, we use four distinct angles: 5 0 , 10 0 , 15 0 , and 20 0 . The probability of neutrino oscillation was determined to be highest at an angle of 15 0 degrees. However, with increasing mixing angles, the likelihood of oscillation increases on the basis of created equation (25) and decreases on the basis of equations (26) and (27) in the three-flavor neutrino oscillation. From generated equations (25) and (26) the maximum neutrino oscillation of probability is discovered at an angle of 15 0 , however, from equation (27), the maximum probability is observed at 5 0 . The greatest neutrino oscillation is found to be 0.9999 and the minimum is zero in all of these two and three flavors of oscillation.
Introduction
Neutrino physics is one of the most fascinating and hotly debated issues in today's high-energy physics.The question of whether neutrinos can oscillate or not prompted a slew of tests to see if these oscillations might be observed.The number of electron neutrinos detected from the sun is 2-3 times lower than the number expected by the mainstream solar model (SSM).Because neutrinos have no mass and hence cannot oscillate in the standard model of particle physics, their measurement reveals the existence of novel physics beyond the SM.The neutrino masses and the entries in the neutrino mixing matrix are two of the unknown parameters.Giunti, Kim, and Lee first estimated the oscillation probability for neutrinos in a wave packet model, then demonstrated that the state vectors utilized for quantum mechanical description and developed a theoretical framework in quantum field in 1993. Finally, in 1995, Blasone and Vitiello demonstrated that the quantum field theory (QFT) description of mixed particles causes unexpected complications in the interpretation of neutrinos as particles.All of the different ways are still being debated today, but under the premise of relativistic neutrinos with minuscule mass squared variations, all of them get the same outcome.For the two flavour case the possibilities of oscillation is given by ( → ; ) = 1 − sin 2 (2)sin 2 (1.27 ( → ; ) = 1 − sin 2 (2)sin 2 (1.27 Where ∆m 2 is the difference in squared masses as Δ 2 = 2 − 2 and the factor of 1.27 is due to the conversion between units.The neutrinos' energy is E, and the distance between source and detector is L. NO experiments can only measure relative masses, not absolute masses.The oscillation length is defined as = 2.47 ∆ 2 , which is the distance between complete oscillations.The length of the oscillation imposes a significant constraint on the measurement conditions.Because there are no oscillations for if ∼ and the oscillations for ≪ are averaged out due to natural uncertainties for the neutrino energy, the oscillations can only be detected if ≫ . As a result, the NO condition is ( Kruppke,2007) .In quantum physics, state vectors are used to characterize flavor neutrinos and are commonly used to define mixing in terms of state vectors.
Equivalently, the same can be done for the anti-particle state vectors, which is defined as The formulations of state vectors in (3) and ( 4) are the most frequent, and they are used in most quantum mechanics studies dealing with neutrino oscillations.When comparing (3) and (4), it is clear that the difference in neutron and anti-neutrino therapy is just a complex conjugation of the PMNS-matrix.The degrees of freedom for space-time, as well as the energy-momentum relationship and the degree of freedom that distinguishes the neutrino species, must be factorized.Furthermore, quantum mechanics converts the general wave function of a spin-1/2 particle into a spin-independent wave function and a spin vector.
𝜓(𝒙
This leads to a factorization of the Hilbert space into two parts Because the space-time and energy-momentum degrees of freedom for a flavor neutrino are a priori unknown, this type of factorization is only viable for mass neutrinos.The entire Hilbert space for neutrinos can be expressed as a direct product of the ℋ = 2 space for dynamical degrees of freedom and the ℋ ≔ ℂ ; space for mass degrees of freedom; The state vectors |⟩ described in above equation ( 3) can be used as a basis for ℋ, the elements of ℋ will be denoted by |⟩, where the index indicates the mass dependence of the dynamics.
Neutrinos have no mass in the most basic version of the SM.In fact, because neutrino masses are so much less than the masses of the other fermions, this has been a pretty good representation of Nature for a very long time.Adding neutrino masses via connecting neutrinos to the Higgs field, which would give birth to neutrino masses in the same way as other fermion masses, appears to be the most tempting method at first appearance.Some fermions are referred to as Majorana fermions since they lack their own anti-particle, which is referred to as a Dirac fermion.There have been no observations to yet that have been able to determine whether neutrinos are Dirac or Majorana fermions, which is one of the most intriguing potentials for future neutrino studies.Basic concepts of neutrino oscillations are if neutrinos are massive, then they will oscillate which means their reaction occur with charged lepton of a given generation.The probability for the oscillation of a neutrino of a given generation into a neutrino of another given generation can be calculated to be dependent on the difference between the squares of the neutrino masses.Thus, if neutrinos are massless, or if they all have the same mass, then they will not oscillate.It follows that if neutrino oscillations are observed, then neutrinos must be massive with at least two different masses.In addition to this, there must also be mixing in the lepton sector as the amount of mixing will determine the oscillation amplitudes.
The fundamental principle of neutrino oscillations is that if neutrinos are heavy, they will oscillate, which indicates that their reaction will take place with a charged lepton of a certain generation.The difference in the squares of the neutrino masses can be used to compute the chance of a neutrino of one generation oscillating into a neutrino of another generation.As a result, if neutrinos have no mass or have the same mass, they will not oscillate.If neutrino oscillations are detected, then implies that neutrinos must be hefty, having at least two distinct masses.Furthermore, mixing in the lepton sector is required, as the quantity of mixing determines the oscillation amplitudes.
Neutrino oscillations in vacuum
Let us consider how flavors mixing affects the propagation of neutrinos in vacuum and initial state to be the flavors eigenstate is given as, The flavor eigenstates are denoted by , µ, and mass eigenstates 1, 2, 3. Then after travelling certain distance the state of neutrino is given as, Here, and are the energy and momentum associated with the mass eigenstate j.They can be derived from the dynamics of the elementary process in which the neutrino is produced.For example, in the case of pion decay at rest, π + → µ + + νµ or π − → µ − + µ̅ , the center of mass system can be obtained as, Substituting Eq. ( 10) into Eq.( 11) and taking the square and solving to obtain, On solving and arranging we get, By denoting the energy that would be obtained for massless neutrinos by 2 ) and keeping only terms up to second order in mj , the momentum and energy is obtained as, Now, the neutrino oscillation probability is obtained (Kopp,2006) as,
Two-flavor oscillations in vacuum
Now to study the oscillation of two flavor neutrino we consider and initially which is successful in describing many neutrinos oscillation experiments.In two flavors, there are no Dirac phases, implying that the lepton mixing matrix is real, and the only parameters are the mixing angle () and the mass squared difference(∆ 2 = 2 2 − 1 2 ).
Now the lepton mixing matrix for 2D rotation is given as, On substituting the value of U in general equation of neutrino oscillation probabilities in equation ( 17) and solving we get, the two-flavor neutrino oscillation probabilities as, Here, oscillation phase is and can be written as In addition, unitarily condition basically states that no neutrinos are lost during the neutrino evolution.
Neutrino oscillation in vacuum for 3 flavor
Three-flavor oscillations the lepton mixing matrix using the standard parametrization with three mixing angles and one CP-violating phase.In addition, there will be two independent mass squared differences, ∆ 21 2 and , ∆ 31 2 .UMNO is a 3 × 3 matrix analogous to the CKM matrix for the quarks but due to the Majorana nature of the neutrinos it depends on six independent parameters as, The Pontecorvo-Maki-Nakagawa-Sakata (PMNS) mixing matrix is a popular name for this matrix.It's worth noting that when charged leptons have no other interactions beyond SM interactions, their interaction eigenstates can be identified by comparing them to the appropriate mass eigenstates up to phase redefinition.The approach above produces a unitary UMNO matrix for the light states in the example of three light Dirac neutrinos.Because of the masses, spectrum energy, see-saw process, and violation of unitarity, the UMNO matrix for the three flavors mixing cases is unitary, regardless of whether the neutrinos are Dirac or Majorana particles (Gonzales-Gracia,2018).Now the mixing of three generation neutrino is given as, Since there is no direct evidence of non-zero neutrino mass from the laboratory experiments.So, the upper limits from direct kinematic neutrino mass measurements are given as < 15, < 0.17, < 18.2 (Nunokawa,2000;Upadhyay et al., 2013).Now, the neutrino oscillation probabilities for three flavors using U NO in equation (17) obtained as In
Result and discussion Two flavor oscillation
The numerical analysis is based on Δ = 0.1 and the nature is shows in figure 1.The below figure 1, shows that the probability of neutrino oscillation with the mixing angles.In the case of equation ( 19) when the ∆=0.1, the probability of neutrino oscillation decreases with increasing the mixing angles.The maximum probability of oscillation is approximately 0.99998 at the mixing angle of 5 degree.The neutrino oscillation probability gradually decreases and becomes minimum at the angle of 45 degree after that the probability is certainly increases with increasing the mixing angle and becomes maximum at the angle of 58 degree.In the case of equation ( 20) when the ∆=0.1, the probability of neutrino oscillation is increases with increasing the mixing angles.The maximum probability of oscillation is 2.5× 10 −3 at the mixing angle of 45 degree.The neutrino oscillation probability gradually increases from 0.1 to 2.5× 10 −3 and becomes maximum at the angle of 45 degree after that the probability is certainly decreases with increasing the mixing angle and becomes minimum at the angle of 58 degree.
The numerical analysis of developed equation ( 27) is based on Δ 2 = 10 −5 wth consider 23 and = 0.1 to 0.9 and the nature is shows in figure's 4. The figure 4 shows the same nature of probability of oscillation of neutrino in four different angle 5 0 , 10 0 , 15 0 , and 20 0 in all the condition of L/E=0.1 to 0.9 based on Δ 2 = 10 −5 with mixing angles.The neutrino oscillation probability gradually decreases with increasing the mixing angles in degree.The neutrino oscillation probability becomes maximum at the mixing angle of 45degree.Finding the minimum probability certainly change the probability by increasing its value.The maximum probability of neutrino oscillation is at a mixing angle of 5degrees
Conclusion
We analyze the probability of neutrino oscillation when the four flavor of neutrino mixed.First of all, we observed the nature of two flavor neutrino oscillation based on the equation ( 19) and ( 20) only one condition of∆= 0.1.In this observation we found that the neutrino oscillation probability decreases with increasing the mixing angles in equation ( 19) but in equation ( 20) we found reverse probability, increases with increasing the mixing angle.After that we analyze the three-flavor neutrino oscillation based on the Δ 2 = 10 −5 and L/E=0.1 to 0.9 condition respectively in the four different angles 5 0 , 10 0 , 15 0 , and 20 0 .By observation based on the equation ( 25) and the neutrino oscillation probability from → , the neutrino oscillation probability increases with increasing the mixing angles.In all the condition, the maximum probability is shown at the mixing angle of 45 degree in the angle of 15 degree and minimum probability of oscillation in angle of 5 degree.According to the equation ( 26), the oscillation probability of neutrino decreases with increasing the mixing angles when → .The neutrino oscillation probability is maximum in the initial mixing angles after that the probability becomes zero.The maximum probability is seen in the angle of 15 degree by comparing all the four angles followed by all the conditions.Followed by equation ( 27), we also find the same nature but in different character.The neutrino oscillation decreases with increase in mixing angles.But in this case, the maximum neutrino probability given by an angle of 5 degree instead of the angle of 15 degrees. | 2022-10-12T15:03:35.026Z | 2022-08-04T00:00:00.000 | {
"year": 2023,
"sha1": "0b2ff12af423011e9a454e327757f56bb01db500",
"oa_license": "CCBYSA",
"oa_url": "https://jurnal.uns.ac.id/jphystheor-appl/article/download/64492/40840",
"oa_status": "GOLD",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "468d5abc58d36f77ef93c250594db853e572fa94",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": []
} |
25207182 | pes2o/s2orc | v3-fos-license | “Education level” responsible for inequities in oral practices among 15–34‑year‑old individuals in Jizan, Saudi Arabia
Background: Education plays an important role and is a second major (non‑medical dimensional) factor influencing the health status. Individuals demonstrate oral practices that impact the oral health positively or negatively. This study analyzes how self‑reported oral practices are influenced by different educational levels among young adults in Jizan. Materials and Methods: The survey was conducted in four areas around University of Jizan. The survey subjects, aged between 15 and 34 years, completed the self‑administered, structured questionnaire. Inter‑group comparison was done using Chi‑square test. Level of significance was fixed at P < 0.05. Results: Exactly 1597 subjects completed the survey. Of these, 644 (40.3%) had attained higher education, 884 (55.4%) had lower education, and 69 (4.3%) were illiterate. Statistical comparisons showed significant differences among the three groups with respect to the use of toothbrushes ( χ 2 = 88.67, P < 0.001), use of interdental cleaning aids ( χ 2 = 15.04, P < 0.001), siwak use ( χ 2 = 16.31, P < 0.001), cigarette smoking ( χ 2 = 14.28, P < 0.001), and khat chewing ( χ 2 = 38.1, P < 0.001). Siwak use was more among those with low educational level and illiterates. Further, smoking and khat chewing were significantly more prevalent among illiterates. Conclusion: The subjects with low education and the illiterates exercise harmful oral practices. This study indicates that educational level is responsible for inequities in oral practices.
INTRODUCTION
Education plays an important role and is a second major (non-medical dimensional) factor influencing the health status. [1] World illiteracy map closely coincides with the world poverty map, where malnutrition and other unhealthy conditions prevail. It was documented that those subjects who are better educated, wealthier, and live in better circumstances have better health status, with a greater frequency of dental visits than those who are less educated. [2] From the view point of public health, education not only helps to create and spread awareness, but also supports the public health service providers to plan, formulate, implement, and promote oral health in the community.
Individuals demonstrate oral practices that impact the oral health positively or negatively. The use of toothbrushes, interdental cleaning aids, and siwak use can improve the oral health, [3][4][5] whereas cigarette smoking, pipe smoking, and khat chewing [6,7] can affect the oral health adversely. Research studies indicate that the existence of these behaviors among people varies with respect to age, [3,4,8,9] gender, [4,8,9] educational level, [3] and socio-economic status. [10] Against this background, the present study was designed to assess the existing level of oral practices (use of toothbrushes, interdental cleaning aids, cigarettes, pipe smoking, and chewing siwak or khat) among selected, young adults of Jizan, in particular, to analyze how different educational levels influence the self-reported oral practices.
MATERIALS AND METHODS
The survey was conducted among the subjects from a selected population of Jizan (aged 15-34 years). The total population in the age group 15-34 years in Jizan city was around 24,571, which was obtained from the governorate office of Jizan city. This study was conducted for 3 months (January-April 2013). A pilot study was conducted on 100 individuals in order to ensure the level of validity and reliability of the questionnaire (Cronbach's α = 0.81), and the data of the pilot study were not included in the main survey. The sample size for the study was obtained using the formula: The values obtained are z = a point on normal distribution with 95% confidence, the value of P was considered as the prevalence of toothbrushing from an earlier study on urban Saudi Arabians (p = 73%), q = 100 − p, d = admissible error = 10% of prevalence. Four random zones and three areas per zone were selected. Thus, the total sample was calculated as: 4 × 3 × 148 = 1766.
To collect realistic data, subjects were recruited from four survey areas around Jizan University. The areas were: Hay al matar, Mahata khams, Dara ut tawhid, and Suq ad dakhili. The data were collected using a self-administered, structured questionnaire. The survey was conducted by visiting the houses, colleges, and workplaces in these areas. All those people aged 15-34 years who were willing to participate in the study completed the questionnaire. The illiterates were interviewed by the investigator who also filled their information. Informed consent was obtained from the subjects and permission to conduct the study was obtained from the ethics committee of the institution.
The data were transformed from the pre-coded survey form to the computer. A master file was created for the purpose of data analysis. SPSS version 17.0 (SPSS Pvt Ltd, Chicago, IL, USA) was used for statistical analyses. The Kolmogorov-Smirnov test showed values of educational level and oral practices to be normally distributed. Frequency analysis was carried out and Chi-square test was done for comparison between the groups. All tests were carried out with the significance level set at 0.05.
RESULTS
Out of the 1766 recruited subjects, 179 subjects filled the questionnaire incompletely and were excluded. Therefore, the study was conducted with 1597 subjects. Of the total 1597 subjects, 1162 (72.8%) were males and 435 (27.2%) were females. Also, 644 (40.3%) had attained higher education (post-doctoral degree, post-graduate degree, and degree), 884 (55.4%) had attained lower education (completed elementary school or higher school), and 69 (4.3%) were illiterate.
Overall, it was observed that 1372 (85.9%) practiced toothbrushing, of which 1112 (69.6%) brushed once daily and 260 (30.4%) brushed twice or more. Only 323 (21.5%) were found to use interdental cleaning aids, of which only 50 used them regularly and the remaining used them intermittently. A total of 854 (53.5%) subjects chewed siwak, of which 20 subjects chewed only siwak while the majority (n = 834) practiced toothbrushing in addition to siwak use. Also, 214 (13.4%) were regular smokers, of which 95 smoked less than a packet per day while the remaining were found to smoke two packets and more per day. Thirty-three subjects (2.1%) were ex-smokers and 71 (4.4%) were pipe smokers (hookah); however, only 13 of the pipe smokers were regular pipe users. Khat chewing was prevalent in 293 subjects (18.3%) and all were regular chewers with a varying frequency of 1-6 times per day. Table 1 summarizes the distribution of the 1597 subjects according to their educational level and oral hygiene practices. There was a statistically significant difference observed between the groups with different educational levels and their toothbrushing behavior (χ 2 = 88.67, P < 0.001), use of interdental aids (χ 2 = 15.04, P < 0.001), and siwak use (χ 2 = 16.31, P < 0.001). Table 2 summarizes the distribution of study subjects according to their educational level and oral practices.
Graph 1 indicates the percentage distribution of subjects with different educational levels according to the oral practices followed by them. Percentage of subjects practicing toothbrushing and using interdental aids was greater among those with higher educational level. Siwak use was more among the subjects with low educational level and the illiterates. Harmful practices such as smoking and khat chewing were prevalent among the illiterates.
DISCUSSION
The survey was conducted among the subjects (aged 15-34 years) from a selected population of Jizan, as individuals of this age are at potential risk for acquiring adverse habits under the external influences; proper oral health education may rectify such practices.
The study results showed that educational levels had a significant association with the oral practices being followed by the subjects. Oral practices, such as use of toothbrushes and interdental cleaning aids, were commonly followed by individuals with higher education. This finding matches with other study reports on urban Saudi Arabians [3,11] which indicate that for Saudi Arabians with higher education, toothb rushing behavior started earlier in life. Toothbrushes are the preferred cleansing aids [4] and are most commonly used. In this study, overall, 85.9% of the young adults of Jizan practiced toothbrushing. A study among urban Saudi Arabians [12] showed that 73% used a toothbrush daily, while a siwak was used daily by 65%. It is observed that young and educated population mainly uses modern aids like toothbrushes, [11] which may be attributed to awareness about newer aids and techniques.
In this study, the frequency of toothbrushing was not associated with educational levels, which is similar to that reported earlier. [13] On the contrary, a study among the Danish adult population showed that toothbrushing twice a day was related to high level of education. [9] In Jizan, modern techniques of oral hygiene, such as toothbrushing and use of interdental aids, are prevalent, in addition to the traditional method of chewing siwak. The use of siwak is widespread in Saudi Arabia and many other Muslim countries. [3] It is made from the plant Salvadora persica. In this study, 53.5% of the subjects used siwak, in comparison to 65% of urban Saudi Arabians. [12] Majority of them (n = 834) were toothbrush and siwak users. A mere 20 subjects used only siwaks. A study by Tubaishat et al. [14] reported that some toothbrush users believe that siwak chewing in addition to toothbrushing is the most effective way of reducing food debris. In the present study, statistically significant difference was observed between siwak use and educational level. This finding is in consensus with the study results of al-Otaibi, [3] who indicated that the siwak is preferred by less-educated Saudi Arabian people.
Educational level was found to be significantly associated with cigarette smoking and khat chewing in the present study. Chewing of khat leaves, [7] and tobacco [7,15] and pipe smoking [16] are becoming a major public health problem worldwide. Khat is a well-known natural stimulant obtained from the plant Catha edulis and is widely used in certain Red Sea countries, including Yemen and the province of Jizan in Saudi Arabia. Among the students of Jizan, [17] the percentage of khat chewing was found to be 20.5%. While among the Axum university (North Ethiopia) students, [7] the prevalence of khat chewing and cigarette smoking was 27.9% and 9.3%, respectively, and the commonest reasons for its use were to keep alert while reading (40.6%), for relaxation (65.5%), and to relieve stress (37.7%); also, having peer friends who were khat chewers was strongly and positively associated with khat use. Ageely [6] reported that in comparison to the other determinants like age, gender, and residence, the use of khat was significantly associated with school and college education (P < 0.05).
Smoking has been existing in Saudi Arabia for more than 50 years, with an overall dramatic increase in the import of cigarettes. In the present study, it was found that 13.4% young adults were regular smokers, in comparison to 17.3% of school children from Jizan. [18] The educational level showed an inverse relation with smoking, with high prevalence observed among the illiterates and people with low educational level. This finding is also supported by Jarallah et al. [15] Gaffar et al. [18] revealed that the three most important independent predictors of smoking were academic performance [odds ratio (OR): 5.32], having friends who used khat (OR: 3.23) and/or tobacco (OR: 2.88).
There are many ways to use tobacco and one of them is pipe smoking. The prevalence of pipe smoking in this study was 14.4% and no association was found between pipe smoking and the educational level. However, a report [16] reveals that this habit is linked to the social class (preferred by those from the upper, upper-middle, and middle class).
One drawback of this study was that there were not many subjects (4.3%) who represented the illiterate group. Conversely, this also indicates that the literacy level is good in Jizan. In a true sense, the results represented here seem to be a comparison between the subjects with high level and low level of education. However, this database serves as a good reference to plan a similar study in future.
CONCLUSION
The overall results of this study indicate that educational level is accountable for inequities in oral practices. Earlier, Paulander et al. [13] had shown that educational level influences the oral conditions and also, the model proposed by Gupta [19] indicates that educational status is a marker of health inequities and suggests that good educational status can lead to appropriate health-related behaviors. The advantage for a public health service provider to have an educated patient is that he or she can be informed and motivated to understand the concept in a much superior manner, as compared to the person with low education and the illiterates. It can be hypothesized that when the health education session will be carried out for the study Graph 1: Percentage distribution of subjects of different educational levels according to the oral practices subjects, it would be easier to motivate the people with high level of education when compared to the other groups.
Within the limitations of the study, the results indicate that the subjects with low education and the illiterates exercise harmful oral practices. They need to be educated to improve their hygiene practices and quit smoking and khat chewing. | 2018-04-03T02:06:51.796Z | 2015-03-01T00:00:00.000 | {
"year": 2015,
"sha1": "57dcd20a41b5a840645cb4affaca2b00b2a420bc",
"oa_license": "CCBYNCSA",
"oa_url": "https://europepmc.org/articles/pmc4415330",
"oa_status": "GREEN",
"pdf_src": "ScienceParsePlus",
"pdf_hash": "e5d2e432bf3e1747ce19cafaf4d9cf9bfe65cc56",
"s2fieldsofstudy": [
"Medicine",
"Education"
],
"extfieldsofstudy": [
"Medicine"
]
} |
252736681 | pes2o/s2orc | v3-fos-license | Access to Care Limits Lung Cancer Screening Eligibility in an Urban Safety Net Hospital
Purpose: Lung cancer screening (LCS) results in earlier detection of malignancy and decreases mortality but requires access to care to benefit. We assessed factors associated with timing of lung cancer diagnosis in the absence of systematic LCS in an urban safety net hospital. Patients and Methods: Retrospective chart review was performed of patients with pathologic diagnosis and/or staging of lung cancer at our institution between 2015 and 2018. Patient socio-demographics, disease characteristics, factors associated with access to medical care, and time point and process by which the patient accessed care were collected and analyzed. Results: In total, 223 patients were identified with median age of 63 years and 57.8% male predominance. Racial distribution was 22.9%, 20.2%, 17.1%, and 9.4% for Black, White, Asian, and Hispanic, respectively. Stage at diagnosis was 8.1%, 4.5%, 17.0%, and 60.5% for stages I, II, III, and IV, respectively. Medicaid (59.6%) and Medicare/Medicaid (17.1%) were the most common insurance types, while 16.1% had no insurance. A majority (54.3%) had no established primary care provider (PCP), and only 17.9% had an in-network PCP. Patients without PCPs were more likely to have diagnostic evaluation initiated from Emergency Department or Urgent Care settings (95.0% vs 50.1%, P < .01) and present with later stage disease (92.7% vs 77.8%, P < .01). Of the 83 patients that met age and smoking history LCS criteria, only 33.7% (12.6% of total) also had an in-network PCP. Conclusion: Absence of established PCPs is associated with later stage presentation of lung cancer and may limit system- level benefits of LCS implementation.
Introduction
Lung malignancy is the leading cause of cancer-related death, accounting for approximately 1 in 5 deaths; in the United States alone, there are an estimated 236 000 new lung cancer cases and 132 000 lung cancer-related deaths annually. 1 Stage at diagnosis affects available treatment options and survival, making early detection important in the management of the disease. Patients diagnosed with localized non-small cell lung cancer (Stage I) have a 57.4% 5-year survival which contrasts with 5.2% for those diagnosed with metastatic disease (Stage IV). 1 In 2013, the United States Preventive Services Task Force (USPSTF) recommended annual lung cancer screening (LCS) for asymptomatic persons aged 55 to 80 years with a smoking history of at least 30 pack years who are currently smoking or quit within 15 years. 2 This recommendation was largely based on data from the National Lung Screening Trial (NLST) which showed a 20% relative reduction in lung cancer mortality after 3 years of screening with low-dose computerized tomography (LDCT) compared with chest X-rays in high-risk patients. 3 Subsequently, the Dutch-Belgian Randomized Lung Cancer Screening Trial (NELSON) reported a 24% lower risk of lung cancerrelated deaths in high-risk patients randomized to receive low-dose computerized tomography (LDCT) compared to no screening at 10-year of follow-up. 4 Despite this demonstrated benefit, LCS uptake has been slow and variable across the United States. 5 6 Furthermore, geographic variation persists in regards to LCS. LCS centers are largely clustered in urban areas causing variability in availability of LCS within the same region. 5 In addition to variability in availability of LCS, there are racial differences in both smoking behaviors and age at diagnosis that likely contribute to racial disparities in lung cancer detection using uniform screening eligibility. 7 These differences in behaviors and disease result in varying performance characteristics of the NLST criteria notably in minorities prompting broadening of USPSTF guidelines in 2021 to ages 50 to 80 and a 20 pack year smoking history. 8 LCS is initiated at the primary care provider (PCP) level with patient education and shared decision making about the implications of abnormal findings on LDCT. Implementing a LCS program also requires institutional infrastructure and support, including an interdisciplinary team, standardized reporting systems, protocols for patient follow-up, and strategies to provide effective patient counseling and shared decision making. 9 These systematized processes represent a large commitment of institutional resources, such as equipment, personnel, and education of patients and medical providers which can be an institutionlevel barrier to LCS. In addition, there are numerous factors that may influence a patient's access to medical care. The complex interaction between these factors and other determinants of health is highly individualized making it challenging to understand how different patient populations at institutions are affected. Prior evaluations of barriers to medical care in LCS populations have been from high-risk screening patients already within a medical system. However, these descriptions do not provide insight into the magnitude that this problem represents for those who have not yet established continuity care with a PCP who ultimately develop lung cancer and would benefit the most from access to LCS.
Our institution is a key provider of inpatient and outpatient services within a large urban safety-net health care system serving a diverse and largely underserved patient population. The purpose of this exploratory study is to describe the population of patients diagnosed with lung cancer at our institution who differ considerably from LCS study populations and assess how demographic, socioeconomic, and access to health care might influence the diagnosis of lung cancer in the absence of a systematic LCS program. One specific objective is to assess the percentage of lung cancer patients that would qualify for LCS. Secondly, we aim to quantify the impact that lack of access to care has on LCS by determining the number of lung cancer patients who did not have routine medical prior to diagnosis but would otherwise meet LCS eligibility criteria. There are numerous reports of LCS utilization amongst high-risk populations of patients who already have established medical care within a healthcare system. However, quantitation of the number of patients who do not have access to medical care prior to lung cancer diagnosis is not well described in the medical literature. Quantifying the magnitude of the barrier represented by access to care further provides perspective on the percentage of lung cancer patients within the patient population that might benefit from LCS.
Methods
Retrospective chart review was performed on all patients undergoing diagnostic or staging procedures with a new diagnosis of primary lung cancer at our institution from 2015 to 2018 at Harbor-UCLA Medical Center (Torrance, CA). Patients were identified from a hospital database of cancer patients. All patients with pathologic diagnosis or staging of cancer were included for analysis. Data was retrospectively collected from the electronic medical record, which includes records from all levels of care within the Los Angeles County-Department of Health Services (LAC-DHS) system including primary care providers, subspecialty clinics, urgent care, and the emergency department. Socio-demographic data including insurance status and primary language, smoking status, and characteristics of lung cancer and medical care including diagnostic workup and treatment were recorded in de-identified fashion. Patients who obtained care from a primary continuity care setting within our provider network were classified as having an in-network PCP versus those who received continuity medical care externally who were classified as having an out-ofnetwork PCP. Assessment of an established PCP at least 12 months prior to cancer diagnosis was identified by presence of documentation in the electronic medical record by a medical provider, indication of a PCP on hospital intake forms or in medical documentation including outside medical records, or indication of routine access to prescription medications or other recurrent medical care. Insurance status was differentiated as public insurance from national (Medicare) and state (Medicaid) levels versus private insurance provided by third party sources.
Descriptive analysis was performed with median and interquartile range (IQR). Lung cancer staging was performed using the Eighth Edition of the TNM Classification system. 10 Qualification for LCS was determined using the 2013 and 2021 USPSTF lung cancer screening guidelines. 2,8 Distance to the medical center was calculated using Google Maps. Associations between variables was performed using Fisher's exact test with a threshold of P < .05 used to determine statistical significance. This study was granted an exempt determination by The Lundquist Institute Institutional Review Board.
Results
A total of 223 newly diagnosed or staged lung cancer patients were identified during the target time period. Median age at diagnosis was 63 with an IQR of 55 to 69 and a male predominance (Table 1). Of the patients analyzed, 71.7% ranged in ages from 50 to 80. The largest racial group was Black followed by White. The majority of patients were either current or former smokers with 41.3% having at least 20 pack years of smoking and quit <15 years prior.
The majority of patients were stage IV at time of diagnosis. Distribution of disease stage were 8.1%, 4.5%, 17.0%, and 60.5% for stages I, II, III, and IV, respectively. The most common types of lung cancer were adenocarcinoma (48.4%), squamous cell carcinoma (18.8%), small cell carcinoma (9.9%), and mixed type (7.6%). Genetic driver mutations were identified in 59 (26.5%) patients, though testing was not applicable in 90 (40.3%) based on cancer subtype or stage.
With regards to distance to the medical center, 24.7% lived within a five-mile radius, 40.3% lived between 6 and 10 mi, and 35.0% lived more than 10 miles away. At time of diagnosis, 7 (3.2%) patients did not have stable housing. Any type of medical insurance was identified in 187 (83.9%), with Medicaid alone (n = 133, 59.6%) and Medicare/ Medicaid (n = 38, 17.1%) being the most common insurance types; 2 (0.9%) had private insurance while 36 (16.1%) had no health insurance. Only 102 (45.7%) had an established PCP at least 1 year prior to diagnosis, of which 63 (28.2%) had a PCP within our medical network.
Using 2013 USPSTF screening guidelines which were clinically applicable during the study time frame, 61 (27.3%) patients qualified for LCS based on combination of age and smoking history criteria. Of those patients meeting 2013 USPSTF LCS eligibility criteria, only 26 (42.6% of LCS eligible patients) had an established PCP within our healthcare system. When applying updated 2021 USPSTF screening guidelines to the study population, 83 (37.2%) patients met LCS eligibility criteria ( Figure 1) of which 28 (33.7% of LCS eligible patients) had an established in-network PCP.
Early cancer presentation (stage I/II) occurred in 20 out of 90 (22.2%) with an established PCP and 8 out of 110 (7.3%) without an established PCP (P < .01) ( Table 2). For patients with health insurance coverage, 22 of 164 (13.4%) were diagnosed at early stage compared to 6 of 36 (16.7%) with no identified insurance (P = .61). Twenty-three patients lacked a confirmed stage at diagnosis and were not included in the analysis.
Diagnostic evaluation was initiated due to cancer-related symptoms in 173 (77.6%), CT screening by the patient's PCP in 8 (3.6%), and incidental findings on chest imaging performed for other clinical reasons in 42 (18.8%) ( Table 3). The locations of initial medical evaluation and referral for diagnostic workup that eventually confirmed lung cancer were the emergency department (70.4%), primary care clinic (13.0%), and subspecialty clinic (11.2%). Those without an established PCP were more likely to have diagnostic workup initiated through the emergency department or urgent care than those who had an established PCP (95.0% vs 51.0%, P < .01) ( Table 4).
Regarding components of cancer-related care received within our medical system, 147 (65.9%) received all of their care from clinical presentation to initial diagnostic procedure to treatment in our health care system. Forty (17.9%) patients were diagnosed outside of our medical system but completed staging and subsequently established medical care for cancer treatment at our hospital. Thirty-six (16.2%) patients received subsequent medical care outside our medical network after initial diagnosis (13.5%) or staging (2.7%).
Discussion
Lung and bronchus cancers account for the most cancerrelated annual deaths, with a 5-year relative survival of 21.7%. 11 Lung cancer screening with LDCT has been shown to effectively reduce lung cancer mortality. 4 Despite this, implementation of LCS remains low and variable across different populations. Only 5% of high-risk adults underwent screening based on national data from 2015, which is comparable to the 3.3% and 3.4% estimated national screening rates in 2016 and 2017, respectively, which were drawn from comparison of radiology registry data to populationbased surveys, the U.S. Census, and cancer registry data. 5,[12][13][14][15] There was slight improvement in 2018 to 5.0%, however geographic variation persists. Other multi-state samples from 2017 and 2018, demonstrated approximately 15 to 20% of eligible high-risk patients receiving LCS, however again demonstrating significant variability by region and insurance status. [12][13][14][15] In fact, California had the second lowest LCS rate (of eligible patients) with a 0.21 screening rate ratio compared to the national average. 5 LCS sites remain clustered in urban areas and more than a third of the counties with high lung cancer mortality were beyond a 60 min drive from a LCS center. 16 There are a multitude of different barriers to LCS that have been identified and can be broadly categorized as patient-, provider-, and system-level barriers. 17 Much conversation in the medical literature focuses on provider-and system-level barriers to implementation of LCS. One study found that providers practicing at a community or academic hospital order LDCT scans more frequently than those practicing at a safety net hospital. 18 Provider knowledge of LCS guidelines varies and suggests the need for improved provider education at safety net hospitals. 18 Our institution is a tertiary referral center within LAC-DHS providing subspecialty care for the diagnosis and treatment of lung cancer with a catchment area covering the southern portion of the County of Los Angeles. Our patient population shares similarities in many characteristics of disease to national data. The sociodemographic diversity demonstrated in our patient population highlights the challenges of other medical systems and providers across the nation. The NLST study population was over 90% White but lung cancer behaves differently amongst different demographic groups. Therefore, performance of screening guideline eligibility does not align equally for Blacks and Whites. 7 In the NLST, Black patients had greater reduction in both lung cancer and all-cause mortality compared to Whites, even with low participation. 3 Despite having greater lung cancer incidence, Blacks are less likely to be eligible for screening given their lower average cigarette per day consumption and are nearly half as likely to report undergoing LCS than their White counterparts. 7,19 It is possible that with a lower pack year smoking history there will be an increase in the proportion of eligible Black patients, 7 however even the updated 2021 USPSTF guidelines do not take into account racial, ethnic, or socioeconomic differences in smoking patterns and lung cancer risk and may not be optimal for high risk populations like underrepresented minorities. 19 Nationally, lung cancer is most prevalent amongst patients age 65 to 74. Our patients were diagnosed younger, but still skewed toward stage IV metastatic disease. 1,11 The racial distribution in our population was significantly more diverse, with greater Black, Asian, and Hispanic representation. 11 Though only 9.4% of our patients were Hispanic, we suspect that this is an underrepresentation and limitation of the retrospective nature of the data collection as 32.7% listed their preferred primary language as Spanish. Therefore, it is probable that a significant contingent of the 30.5% of patients listed within the Other racial category are of Hispanic origin. Presentation of lung cancer differs by race with earlier onset (median age 67 vs 70 years) and more advanced stage disease (53% vs 49%) being found in Blacks versus Whites and could contribute to the observed differences. 20,21 Cigarette smoking has been attributed to 81% of deaths in lung, bronchus and tracheal cancers and is mirrored in the high prevalence of tobacco use in our population. 22 Nevertheless, almost 30% of our patients were never smokers which may be associated with the high prevalence (44.4%) of genetic driver mutations identified in those who were clinically appropriate for testing. Prevalence of driver mutations varies widely , however, many that are associated with earlier development of lung cancer are more common in certain demographic groups such as Asians and non-or light-smokers. 23,24 Disparities in lung cancer incidence, diagnosis, treatment and mortality are well documented in the medical literature and are associated with demographic and socio-economic groups. 20,25 Patients with low socioeconomic status are more likely to have medical insurance-related limitations resulting in lack of access to LCS, leaving this vulnerable population at increased risk for lung cancer without an equitable way to access screening opportunities. 26 These disparities can often be associated with a complex interaction of variables, such as communication, education and medical literacy, and factors associated with access to care, such as proximity to the clinic or hospital. In our patients, 45.7% had a primary language other than English, with 32.7% of patients being Spanish speaking. The presence of a language barrier creates extra challenges in health literacy as medical explanations can be lost in translation despite the use of interpreters resulting in miscommunications. Housing and proximity represent additional barriers; 3.1% of our patients were unhoused at time of diagnosis, though this likely is an underrepresentation due to the retrospective nature of the study data collection. While 24.7% of patients live within 5 mi of the hospital and 65.0% of patients live within 10 mi of the hospital, 35.0% live more than 10 miles away. Proximity is a recognized indicator of access to care, however other important considerations such as the mode and access to transportation along with the time requirement for commuting to the medical center could not be accurately assessed from this retrospective chart review and would offer more insight into our patients' barriers to care. 27,28 Other studies note that drive time is significant for some patients secondary to the geographic variability of LCS locations, however the method of transportation is not assessed. 16 Patients also face an opportunity cost when obtaining medical care due to other social and financial considerations such as family or work responsibilities but this information is not represented in the available data. Nevertheless, these data represent social determinants of health that may play a role in accessing medical care for patients. Access to care is a challenging issue and it is likely that complex interactions between various factors confounds the ability to adequately describe it in terms of individual variables with our data set. Medical insurance status is another recognized marker of access to care and has been associated with improved lung cancer survival for patients with private health insurance compared to those with Medicaid or no insurance. 29 In our study, 16.1% of patients had no identifiable insurance, 59.6% had Medicaid alone, 83.0% of patients had only public insurance (Medicare and/or Medicaid), and 0.9% had private insurance. This contrasts with national data from 2018 indicating that private insurance coverage (67.3%) is substantially more prevalent than public coverage (34.4%) in the United States. 30 LCS coverage by Medicaid is determined at the state level causing wider variation in coverage. 31 Until recently, Medicaid was one of the only healthcare payer programs not required to cover LCS. 32 Increased prevalence of severe comorbidities in this older patient population may result in fewer recommendations for LCS by PCPs. 18 We did not find a significant association between lack of medical insurance with diagnosis at an earlier stage of disease though other outcome variables such as mortality were not assessed. Even though there was no demonstrated relationship between insurance status and stage of disease in our cohort, it is probable that the study population size and complex interaction with other socioeconomic and demographic factors may limit our ability to detect this association. Lack of medical insurance or housing are likely to be higher than identified as patients qualify for public insurance once diagnosed with cancer. Therefore, when retrospectively collected, status of these variables may not be reflective of the time of diagnosis unless explicitly indicated in the medical record.
Absence of an established PCP was the most pronounced obstacle to allowing our patients to participate in potential LCS. When considering having any established PCP regardless of in-network versus out-of-network status as a gateway to potential LCS access, 121 (54.3%) did not have an established PCP 1 year prior to diagnosis and therefore would not typically have access to routine health maintenance. There was a significant difference in patients presenting with earlier stage disease who had an established PCP as compared to no established PCP (P < .01). This may be influenced by a variety of factors beyond LCS, such as incidental pulmonary findings from workup of other issues, differences in how patients approach their health and medical care amongst those who have established access to care, or social/cultural perspectives regarding medical care.
Out of our total population, 27.4% qualified for LCS based on age and smoking history criteria using the 2013 USPSTF guidelines but 57.4% of eligible patients did not have a PCP within our medical system. Therefore only 42.6% of eligible patients would have benefited from a screening program at our institution with the remainder of eligible patients unable to access LCS without a PCP. Using the 2021 USPSTF guidelines expands eligibility to 37.2% of all patients qualifying for LCS, but only 33.8% of LCS eligible patients had an established in-network PCP. In addition, it is worthwhile to note that expanded 2021 USPSTF LCS eligibility criteria increases the number of eligible patients for screening by 36.1%, however this only translates to a 7.7% increase when accounting for access to care.
Despite the association between PCP status and earlier stages of disease, 73.5% of patients with an established PCP still presented symptomatic from disease. Lack of an established PCP has been associated with worse lung cancer mortality, and implementation of LCS would promote earlier identification of malignancy as well as initiation of workup from primary care settings. 33 After diagnosis, 83.8% of patients stayed in network for continuity of treatment which demonstrates the reliance of our patient population upon our safety net medical center for their medical treatment.
LCS continues to be underutilized nationally, and studies at other safety net institutions report an overall estimated screening rate of only 16%. 25 In a retrospective description from a single-center, Olazagasti et al reported that 35% of lung cancer patients qualified for LCS however only 4.8% of patients actually underwent screening. A significant association was also found between stage at diagnosis and screening with LDCT. 34 To our knowledge, this is the only other study that describes the percentage of lung cancer patients that would qualify for LCS as all other data describe LCS utilization amongst in-network populations of patients who meet eligibility criteria. In addition, factors associated with access to care such as insurance type and geographic relationship to the medical facility have been associated with the utilization of LCS. 28 However, the interplay of factors associated with access to care is complex and does not always directly equate to lack of medical care. Our data is unique in that it quantitates the percentage of LCS-eligible patients diagnosed with lung cancer who would not be able to undergo screening due to lack of access to medical care.
While access to care remains a significant barrier to LCS, this does not diminish the benefit of screening programs for established patients. However, our study highlights the challenge that access to medical care represents for underserved populations and how it may lead to disparities in lung cancer. In addition, the lack of continuity health care prior to lung cancer diagnosis coupled with a high retention rate for cancer-related care afterward also suggests that system-level benefits from implementation of LCS is likely to be reduced unless improvements can be made in establishing earlier access to care. It is important to note that even with medical insurance and access to care, LCS still is only performed in approximately 10% to 25% of eligible patients. 35 Furthermore adherence to the program is also an issue; at a hospital serving a diverse population, annual follow-up was only 23.7% after 1 year and dropped to 2.8% after 2 years. 36 Based on current USPSTF guidelines there is ongoing concern that multilevel barriers to implementation of LCS and uniform eligibility guidelines will not adequately address the disparities in lung cancer outcomes. Most societal screening criteria utilize age and smoking history similar to USPSTF guidelines. There are other proposed risk models that can be used aside from USPSTF guidelines that include factors beyond just age and smoking history, but improvements are still needed. These risk models are based on little data from minorities and thus still underestimate risk by 5% to 25% in these patients. 37 There are several limitations that influence the interpretation of this data. This is a small patient cohort representing a single-center experience, thus limiting generalizability to other institutions or medical care environments. The smaller sample size also limits the evaluation of broader range factors that might affect access to care. Despite this, our medical center and associated clinics provides medical care to approximately one-third of all patients within our regional public safety net system. Therefore, these data provide a representative cross-section of the patients in our urban medical system. This was a retrospective study and all data were collected from a single evaluation of the medical record. Some variables are not discrete entries within the electronic medical record and therefore relied upon documentation within medical and ancillary provider notes. In addition, data in patient health records at time of chart review may have been subsequently updated and does not always reflect the situation at time of diagnosis, and may result in inaccuracies compared to other methods of data collection. 38,39 Review of records for patients receiving components of care external to our medical system were limited to available data in our health records and documentation of patient self-reporting by medical providers.
Through this study we have identified several issues in a diverse safety net population, however further understanding the barriers that influence our population's access to care is critical. To our knowledge this data is unique in that it quantifies the magnitude of the problem represented by lack of access to medical care and the impact it would have on availability of LCS to an underserved patient population. For the future, prospective assessment of high risk patients who recently established care is needed in order to elucidate their barriers to establishing care and difficulties to undergoing LCS. These factors may include medical literacy, cultural and social perceptions of medical care and lung cancer, financial ability to pay, or financial freedom to take time off work to obtain medical care. These prospective studies will provide valuable insights into underserved patients' barriers to accessing care that are difficult to draw from a retrospective study. Based on this data, we will need to assess the value of implementing a LCS program that has a broader eligibility than the USPSTF 2021 guidelines. Risk models will need to be analyzed and evaluated when used in our underserved and diverse patient population given the performance variability in different patient demographics. 40 Additional evaluations of various risk models need to be performed to avoid an underestimated risk of lung cancer in minorities to ultimately further reduce LCS disparities.
Conclusion
Lack of a PCP in underserved and underrepresented urban patient populations is associated with later stage presentation of lung cancer and may limit system-level benefits of LCS implementation. Further efforts are needed to understand and overcome barriers in access to care to reduce disparities in lung cancer.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article. | 2022-10-07T06:17:42.375Z | 2022-01-01T00:00:00.000 | {
"year": 2022,
"sha1": "569341edde577955ca441fc02a492bdc595f47e6",
"oa_license": "CCBY",
"oa_url": null,
"oa_status": null,
"pdf_src": "PubMedCentral",
"pdf_hash": "100c6abd238a091f5102b0b4f65e07f844c57bb1",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
11154501 | pes2o/s2orc | v3-fos-license | www.ann-geophys.net/25/1037/2007/ © European Geosciences Union 2007
Abstract. We examine the statistical relationships between the interplanetary parameters and spectral power of the geomagnetic fluctuations in the 1–4 mHz frequency band at high latitude stations in two hemispheres. For that, we use a cross-correlative analysis of different combinations of parameters to identify a factor contributing most to the ULF power and to choose (whenever possible) the cross-related keynote controlling parameters. Along with the well-known dependencies of the high latitude pulsation power on the solar wind velocity and variations of the SW dynamical pressure and an additional factor – the flux of solar suprathermal ions with energies about several keV, has been established.
Introduction
The broadband fluctuations in the nominal Pc5-6/Pi3 range (from ∼1 mHz to ∼10 mHz) are the most common manifestation of geomagnetic activity at high geomagnetic latitudes.The fluctuations with millihertz frequencies lie at the boundary between lowest eigenmodes of the Earth's magnetosphere and quasi-static variations.The most intense sources of broadband pulsations are associated with the dayside cusp region and nightside auroral oval.The dedicated studies of the long-period pulsations in the polar caps by Yagova et al. (2002Yagova et al. ( , 2004) ) have shown that these geomagnetic fluctuations, coined Pi3 CAP pulsations, are not reduced to a poleward extension of the auroral or cusp activity.The variations of Pi3 CAP spectral power have a significant component specific to the polar cap only.Mechanisms of these high latitude long-period geomagnetic fluctuations have not been firmly established yet, though some theoretical models Correspondence to: N. Yagova (nyagova@yandex.ru)(Allan and Wright, 2000;Pilipenko et al., 2005) may be relevant to at least part of the observed variations.A thorough examination of ULF signatures in the polar cap regions may reveal specific magnetotail or mantle wave modes.A possible control of Pi3 CAP pulsations by the solar wind/IMF has not been examined so far.
The most important interplanetary parameter for the intensity/occurrence of the geomagnetic fluctuations almost at all geomagnetic latitudes and throughout the ULF frequency range from few mHz (Pc5) (Engebretson et al., 1998;Vennerstrom, 1999) to ∼0.1 Hz (Pc2-3) (Yedidia, 1991) is the solar wind (SW) velocity.This fact is commonly interpreted as an indication on the key role of the KH instability at the interface between the magnetosphere and the solar plasma flow in generation of magnetospheric ULF activity.Besides SW velocity, steep or periodical variations of the SW dynamic pressure can stimulate Pc5 pulsations (Kepko et al., 2002(Kepko et al., , 2003;;Kim et al., 2002) and references therein.Under favorable conditions, the high latitude long-period pulsations can be driven by the IMF quasi-periodic oscillations (Prikryl et al., 1999;Pilipenko et al., 2000;Kepko et al., 2002).Important source of the Pc3-4 (∼10-100 mHz) waves in the Earth's foreshock is the suprathermal protons reflected and accelerated at the bow shock (Blanco-Cano and Schwartz, 1997).Foreshock diffuse ions are associated with the hydromagnetic waves at frequencies 0.01-0.05Hz and reflected ions are related to the high frequency (∼1 Hz) waves (Hoppe et al., 1981).To find sources of the high latitude Pi3 pulsations, a comprehensive statistical study of their correlations with the interplanetary parameters is necessary.In this paper we study the statistical relationships of the spectral power of the geomagnetic field fluctuations in the nominal Pc5-6/Pi3 frequency band with interplanetary parameters including the SW velocity, SW dynamic pressure and its fluctuations, interplanetary magnetic field (IMF), and the suprathermal ion fluxes.
Published by Copernicus GmbH on behalf of the European Geosciences Union.Yagova et al. (2004) showed with the use of multi-regression statistical analysis that Pi3 power at cap latitudes can be presented as a sum of four components: auroral, cusp, pure cap, and local noise.The auroral and pure cap components are bigger than the other two.This determines the choice of stations for the present analysis: 3 stations deep in the polar caps -Antarctic base Vostok (VOS), and Arctic stations Alert (ALE) and Qaanaaq (Thule, THL), and 7 auroral stations at ∼70 • CGM latitude: BAS A81 and Mawson (MAW) in Antarctica, and Nuuk (Godhaben, GHB), Bear Island (BJN), Kotelny (KTN), Yellowknife (YKC), and Fort Churchill (FCC) in the Northern Hemisphere.All the stations are equipped with flux-gate magnetometers with 1 min or better time resolution.The data has been filtered and decimated to common 1-min cadence.The information about the observatory coordinates is given in Table 1 We have taken for analysis 23 months with the best data coverage from February 1997 to December 1998.
The data on the SW plasma, IMF, and keV-ions are taken from WIND satellite, located for the majority of intervals more than 100 R E upstream the magnetopause.In our analysis, the following parameters have been used: SW velocity along Sun-Earth axis V , dynamic pressure P , IMF B Z component, and omni-directional ion fluxes J detected by WIND-3DP experiment in 7 energy channels with energies E from 0.07 KeV to 400 KeV.
Data processing technique
We try to reveal possible sources of the high latitude ULF energy.For that we examine the interplanetary factors which may control the variation of ULF spectral power on the ground and in space.Taking into account that there is no es-sential difference in time behavior of narrow-band and broadband pulsations in the polar caps in Pc5-6/Pi3 frequency range (Yagova et al., 2002) and that noise-type fluctuations dominate at high latitudes we use the notation Pi3 for all the pulsation studied and Pi3 CAP for those measured in the polar caps.The boundaries of the frequency band chosen 1-4 mHz are determined by, on the one hand, an expected lowest frequency of the magnetospheric oscillations (1 mHz, T ∼17 min) and, on the other hand, by the requirement of the spectral estimate stability for 1 min sampling rate.We use 2-h running window to calculate the mean values of interplanetary parameters and the Pi3 spectral power, which corresponds to ∼7 periods at lowest possible frequency under study.A routine data pre-processing procedure includes elimination of instrumental spikes, interpolation to a common sampling rate, and detrending.Intervals with long gaps are excluded from the analysis and short (several points) data gaps are filled with interpolated values.A time window is considered valid for analysis if it includes at least 90 min of continuous record.Thus, the actual length of time interval varies from 90 to 120 min and we do not use overlapping windows to avoid the variation in the degree of overlapping.
Two-hour mean values of the SW velocity V , dynamical pressure, P , and ion flux, J , from a specific energy channel in WIND-3DP experiment have been used to calculate crosscorrelation coefficients between ground and space parameters.As a measure of the ground ULF activity we use the total horizontal spectral power density (i.e., the sum of spectral power densities for two horizontal components) in subbands of ∼0.5 mHz width in the frequency band 1-4 mHz.As a measure of the IMF fluctuations the 2-h spectral power of the IMF B z component variations S Bz was used (the use of other IMF components provides a similar results and will not be shown).Similarly, the SW plasma fluctuations are characterized by the spectral power of the dynamical pressure variations S P .Because the spectral powers and proton flux vary in a wide range, their logarithms are used for the statistical analysis.Though the propagation time between the Wind location and the magnetopause is ∼1 h and its influence on the calculation results is weak, we have time-shifted all WIND parameters to account for the SW propagation to the dayside magnetopause.
We are interested in variations at relatively short time scales, about several days.At this timescale, the shortterm irregular variations dominate while seasonal and solarrelated variations (e.g., the 27-days periodicity responsible for false correlations at longer timescales) are not significant.We calculate correlation coefficients C L for all the sub-sets that contain a given number L of points, and , i.e. the averaged over the period of observations value of C L , is used as measure of linear correlation between two parameters at a given timesclale.The 2-h interval is taken as a point in the subset which contains L points.We are not interested in fluctuations of parameters inside this elementary interval and operate only with 2-h mean values.
Actually, many of the solar wind parameters are interrelated.This may lead to an ambiguity in selection of a governing parameter and, thus, in identification of a physical mechanism of the geomagnetic pulsation excitation.For example, there is a well-known anti-correlation of the SW velocity and density (Burlaga and Ogilvie, 1970;Neugebauer, 1976).To avoid this bias, following Yagova et al. (2004) we apply the statistical procedure of subsequent identification of inter-dependent parameters, which is similar to the multivariate linear regression.The primary governing parameter is chosen to be the one with a maximal absolute value of C.Then, the n-th order parameter has the maximal value of C between the (n-1) residuals.
Commonly, it is unclear whether the correlation between two parameters, X and Y, which both depend on parameter Z, is due to this dependence only, or the X-Y relation has also a Z-independent component.To answer this question, the time variations of each parameter are decomposed with the linear regression procedure into two components: proportional to the factor Z and statistically independent on it (or orthogonal).Then the correlation between these Z-independent (hereafter it is often noted as "pure component") fractions is calculated.This correlation indicates a level of X-Y relation with the influence of factor Z excluded.
In the statistical approach that we apply (see for details Loeve, 1960) the variation of Pi3 spectral power at a given station S STN is considered as an independent variable.In essence, we have to study its expansion over orthogonal basis formed from the variations of dependent parameters.For two variables X, Y centered at their mean values the scalar production XY is introduced and the vectors are orthogonal if XY =0.The correlation coefficient C is the normalized scalar product.A basis can be orthogonalized and a projection of an arbitrary vector, say Y on a given basis (X i , i=1, . . ., n) can be found.This projection corresponds to the component of the variation of the independent variable equal to linear combination of basic variables.If this projection is equal to the vector itself then the behavior of the variable is totally controlled by the governing parameters.
The procedure of expanding of the dependent vector Y over basic vectors (X 1 , . . ., X N ) and the orthogonalization of basis is as follows.We use the notation P for the orthogonal basis.The first vector of the new basis P 1 =X j , where j is determined from the condition of maximum of |C Xj Y |.
Then the components of Y and all the other X vectors, orthogonal to P 1 , are calculated and one with the maximal C is chosen as P 2 .Thus, the vector P i ... P n are orthogonal to all vectors P 1 . . .P i−1 and P i is chosen from the condition of the maximal absolute value of correlation with the Y projection on the subspace X i . . .X n .If the initial set of parameters X or at least its projection on Y is degenerate, i.e. includes linearly dependent variables, they will be excluded during the procedure.
Cross-correlation coefficients between ULF power and interplanetary parameters
The variations of Pi3 spectral power density at three nearly conjugated station pairs VOS-THL, A81-GBH and MAW-BJN are shown in Fig. 1 (from top to bottom).Here the lowest frequency sub-band 1-1.5 mHz has been used.However, because the spectral power decreases with frequency (alike "colored" noise) the results of the subsequent correlation analysis for the power integrated over the whole 1-4 mHz band are almost identical.At cap latitudes (upper panel) the seasonal variations are anti-phase in two hemispheres (with a maximal ULF power in a local summer), while at time scales of several days these variations are in-phase.At auroral (1997)(1998) correlation coefficient C between S (1-1.5 mHz band) and V for 7 high latitude stations in dependence on sample length L (in points).Solid (dashed) lines correspond to the Southern (Northern) Hemisphere, respectively.Warm colors denote polar cap stations, and cold colors denote auroral stations.Among the Northern auroral stations only those with the maximal and minimal correlations are shown.Black dashed-dotted line shows the correlation between V and the power of IMF B z fluctuations S Bz .latitudes (middle and bottom panels) almost no seasonal variations are seen and variations are in-phase and coherent for all temporal scales.It is clear from Fig. 1 that variations of Pi3 power and, possibly, their correlation with governing parameters, depend on timescale.Therefore, further we will present the average correlation coefficient C in dependence on the subset length L.
We start the analysis of correlation with the SW velocity V , because it is known to be the main factor which controls the intensity of almost all types of ULF pulsations at auroral and middle latitudes.Averaged correlation coefficients between the Pi3 power and V in respect to the subset length L for the polar cap and auroral stations in both hemispheres are given in Fig. 2.These dependences are similar for all stations: at short sample lengths, the correlation grows with L, and then it saturates at a plateau at L about several tens.In further analysis, to minimize the influence of regular cyclic variations we use as a reference value L=30.
The correlation with V for the polar cap stations turns out to be higher (∼0.5) than for the auroral stations (∼0.4).The correlation of V with the power of IMF B z fluctuations S Bz (black dashed-dotted line) is lower (∼0.25)than those with the Pi3 power on the ground.More effective ULF response on the ground than in space to the SW driving indicates that the mechanism of the K-H instability at the magnetospheric boundary may be involved.Thus, the SW velocity to a great extent controls the ground Pi3 activity, especially in the polar caps.
Then we check whether the SW dynamic pressure P may control the level of the SW/IMF fluctuations and ground ULF pulsations in the Pi3 frequency range.The correlations of P with the pulsation power S for 6 ground stations and Wind are shown in Fig. 3.These correlations demonstrate the behavior analogous to the previous V −S correlation: saturation at a plateau at about several tens of points.The correlation of P with the Pi3 power at the polar cap stations is again higher (C∼0.3)than that at auroral stations (C∼0.2),though lower than the correlation with V (cf.Fig. 2).Contrary to the correlation with V , the SW pressure correlations with the Pi3 power on the ground and in space are nearly the same (C∼0.3).This fact can be related to the higher level of pressure and IMF fluctuations in the high pressure segments of the SW.To discriminate the influence of the SW dynamic pressure P and the spectral power density of pressure fluctuations, we examine separately the correlations of P and S P with the magnetic fluctuation power at the ground (VOS) and at Wind (Fig. 4a).The relevant correlation coefficients show that the magnitude of P and the spectral power density of pressure fluctuations S P correlate well with each other, and both influence the intensity of magnetic field fluctuations S in a similar way.To reveal the influence on the ULF power of each of parameters free from the influence of the other, The same for the correlation between ULF power at VOS (blue) and WIND (magenta) and solar wind dynamical pressure P (solid) and amplitude of its fluctuation S P (dashed).(b) The same, but for "pure" components (orthogonal to the parameters given in the upper indexes).the correlations between "pure" components (orthogonal to S P and P , correspondingly) are calculated as described in Sect.3. The correlation coefficients (Fig. 4b) for the "pure" components are less than for original time series, but significant.A positive correlation between S P -independent components of the SW dynamical pressure and the power of the magnetic field fluctuations on the ground and in space indicates the existence of an additional mechanism for excitation of the IMF fluctuations in a high pressure SW.
Segments of the SW with high particle density and dynamical pressure are known to be a source of accelerated particles (Gloeckler, 1984;Richardson and Zwickl, 1984).We examine the correlation between the WIND ion flux J and ULF power on the ground and in space.The dependence of the S−J correlation coefficient on ion energy, C(E), is shown in Fig. 5. S−J correlation at (E<20 keV) is surprisingly high, ∼0.4-0.5 both in the polar cap and in the interplanetary space.At these energies the correlation coefficient C(E) slowly decreases with E and it drops significantly for E>40 keV.
The correlation of the SW proton flux with ground magnetic variations at different latitudes is examined in Fig. 6, where the correlations C(L) between the Pi3 power on the ground (6 stations) and the Wind flux J (8 keV<E<30 keV) are given.Qualitatively these dependencies are similar to the previous S-P and S-S P correlations (cf.Figs. 3, 4): the maximal correlation of S with J occurs at cap latitude stations (∼0.4), whereas the correlation at auroral stations is lower (∼0.3).The behavior of J correlation with the IMF Upper panel: the correlations between J and S P (solid), J and P (dash-dot), and P and S P (dashed).Other two panels show the correlations of S VOS (middle) and S Bz (bottom) with J (solid), S P (dash-dot), and P (dash) pure components (orthogonal to the parameters given in the upper indexes).S P and P for the correlations with S VOS (middle panel) are recalculated to the dayside magnetopause.related to V : the SW dynamic pressure, the amplitude of pressure fluctuations, and the flux of protons with keV energies.These parameters are inter-related and influence the ULF power on the ground and in space similarly, but in a different way than V .Correlations between these three parameters and their influence on the Pi3 wave power on the ground and in space are shown in Fig. 7.The highest correlation (∼ 0.8) is found between P and J and the correlation between two other pairs are very close (∼0.6).
To exclude a dependent parameter influencing ULF power only as a superposition of two others, correlations of "pure" components (i.e. each of three independent on two others) with ULF power on the ground and in space have been calculated (two bottom panels in Fig. 7).The correlation vanishes for S−P pair, which means that the SW dynamic pressure influences ULF power not directly, but only through the pressure fluctuations and proton fluxes.At the same time, the correlation of the pure S and J (that is, V and S P -independent) components is non-zero and positive (∼0.2), indicating that the SW proton flux is an additional independent factor which contributes to the excitation of magnetic fluctuations on the ground.
The proton flux and pressure both vary almost synchronously indicating a quick thermalization of 8-30 keV protons.Thus, only those protons accelerated at small distances from the observational monitor contribute to the ULF power in space and on the ground.An important feature of protons with E<30 keV is their weak correlation with higher energy protons generated during the solar proton events (SPE) (not shown).SPE-related peaks can be seen in 8-30 keV channel, but these events are relatively rare and give a low contribution to the total statistics of J .The intervals with low J (8-30 KeV) -P correlation are mainly SPE intervals with the increased proton flux in all energy channels.However, even for these intervals the correlation coefficients between J (8-30 KeV) and ULF power are nearly the same as for the total interval.This gives an additional argument in favor of the above conclusion that the magnitude of the SW dynamic pressure acts on the Pi3 pulsation power only as the superposition of two factors: pressure variations and ion flux.
For the majority of subsets the correlation coefficients of the ULF power on the ground and in space with the proton flux and SW pressure fluctuations are approximately equal and both vary in a similar way with time.However, for ∼25% of samples the correlation of J with the ground ULF power is even higher than its correlation with the ULF power measured on WIND.The two year mean value of the correlation between the pure components of S VOS and J (that is independent on V , S P and S BZ ) is ∼0.13.
The above analysis of possible interplanetary factors, which influence the amplitude of ULF fluctuations at high latitudes, may be summarized as follows: the most important factor is the SW velocity.A higher correlation of V with ULF power at the Earth's surface than in the interplanetary space indicates the occurrence of a magnetospheric generation mechanism (probably, KH instability).Then we have a group of coupled parameters weakly dependent on V and related to the SW dynamical pressure P .For them the correlations with the ULF power in the polar caps and at Wind are very close.The multi-regression analysis shows that the influence of P on ULF magnetic fluctuations is, in fact, a superposition of two factors: the spectral power density of the pressure fluctuations S P and suprathermal ion flux J .A weak positive correlation exists between J and ground ULF power independent on V , S P and S Bz throughout the whole high latitude region, but it is higher in the polar caps.
Cross-covariation of ULF power with the interplanetary parameters
In this section, the cross-covariate functions for all the parameters analyzed in the previous paragraph are calculated in order to take into account a possible delay in variations of related parameters.and physical mechanisms for correlations discovered.The cross-covariate function between averaged values of Pi3 spectral power densities at the cap and auroral stations in both hemispheres and the SW velocity is given in Fig. 8.For all the stations the behavior of their cross-covariate functions is similar, though the peak values are higher (∼0.35) for the polar cap stations.The covariate function reaches a maximum at zero time offset indicating the occurrence of instantaneous relation between parameters.At the same time, this function demonstrates an unexpected puzzling asymmetry, as if Pi3 power depends on future values of the SW velocity, or as if Pi3 power enhancements precede somewhat increases of V .This can be due to the dependence of the ULF power at the Earth's surface on space parameters whose variations precede those of the SW velocity.
The covariate functions of the ULF fluctuations in the polar cap with J , S P , P , and S Bz are shown in Fig. 9.They also have a weak asymmetry, but opposite to that of the S − V relationship.Generally, an asymmetric covariate curve with the maximum at zero offset prompts that the delay time between parameters under examination may vary in a wide range from zero to some maximal value.The above features of the covariate functions shows that the delay times between the P -related SW parameters and ULF power at the ground S, as well as between V and S, may vary from 0 to several ten hours.
Examples of Pi3 driving by solar wind proton fluxes
In the above two sections, we have shown statistically that the Pi3 power depends on the SW velocity, SW dynamic pressure fluctuations, and flux of supra-thermal solar protons, whereas the latter two parameters correlate closely with the SW dynamic pressure.Here we present two events for a disturbed interval and a relatively quiet one when the influence of the proton flux on ground ULF activity is seen evidently with a lower nterference of other factors.These events are not unique, because throughout the entire period of observations, during ∼25% of intervals the correlation of ground Pi3 power with J is higher than with S P and V .
During the first event from 3 August 1998 to 8 August 1998 (days 215-220), shown in Fig. 10, two rises of the SW dynamic pressure (third panel) occur at days 215 and 218.The second increase of P up to ∼10 nPa, caused by arrival of a corotating interacting region (CIR), marks the onset of moderate magnetic storm with peak D st =−138 nT.It is interesting that the response of the ULF power in the polar cap (VOS) (upper panel) to this pressure jump is more evident that the variation of ULF power at WIND (second panel).The ULF power in the polar cap enhances nearly simultaneously with the solar proton flux (upper panel), and slightly earlier before the pressure jumps.At the same time, the correspondence between S VOS and V (lower panel) for this interval is low.An example of a relatively quiet interval is shown in Fig. 11 During six days from 8 June 1998 to 13 June 1998 (days 159-164) moderate variations of all parameters occur.Again, a ULF activity in the polar cap, as characterized by S VOS , is highly correlated (C∼0.7) with J , much higher than with V .Thus, at least for some intervals the correlation between S VOS and V is lower than its correlation with S P and J , that is, the influence of V -independent parameters on the Pi3 power dominates.
Stability of correlations
To understand how the correlations found are revealed during disturbed and quiet space weather conditions, we have analyzed the cross correlations between all the SW parameters separately for the intervals including CME, shocks, and CIRs (∼1/3 of all intervals) and those with no space weather events (hereafter "quiet intervals").The correlation coefficients turned out to be very close for these two sub-sets, so the plots in Figs.2-7 are nearly the same for these sub-arrays.Thus, we may conclude that the acceleration of protons in high pressure SW occurs not only in well-formed shocks or CIRs, but also in the SW with very moderate pressure increase.The features of the covariate functions of the SW dynamic pressure and supra-thermal proton flux in respect to amplitude of the magnetic field fluctuations in the interplanetary space and on the ground are similar for a disturbed and quiet SW.Thus, we may conclude that the relation between SW supra-thermal protons (E<30 keV) and the amplitude of the ULF fluctuations in the interplanetary space and on the ground exists permanently, both for the disturbed and quiet intervals.The other concern is the possible false correlations originating from cyclic variations (e.g., solar-related, seasonal, and diurnal).The most direct and effective method to suppress the influence of these variations might be the high-pass filtration of the time series of 2-h parameters at a frequency about 1-2 day −1 .The problem in our particular case of numerous parameters, measured simultaneously by different sensors in the interplanetary space and on the ground, is the existence of data gaps.Actually, there are not enough subsets with complete data coverage.That is why we use crosscorrelation in time domain.To verify the robustness of the averaged correlations we use the following procedure.The whole two-year interval is divided into two one-year intervals, and the average correlation coefficients for all the subsets are calculated for both one-year intervals and between years for all the parameter pairs.The time-shift in cross-year correlations is varied from 0 to 27 days to estimate the maximal correlation of solar-cycle origin.The same procedure is used for "pure" components as well.The summary of correlations is given in Table 2.The columns 1-1 and 2-2 show the correlations for years 1997 and 1998, respectively.The columns 1-2 and 2-1 show the maximal cross year correlations with index 1(2) corresponding to the year 1997(1998).E.g. in the first line of Table 2 column 1-2 demonstrates the correlation between S VOS (1997) and V (1998).All the cross-year correlations demonstrate cyclic behavior with the 27-day period and we give their maximal absolute values, i.e. the amplitude of these variations, in columns 1-2 and 2-1 of Table 2.For the all the correlations, with the exception of "pure" S−P (orthogonal to S P and J ) correlation, zero-shift one-year correlation is at least 1.5-2 times higher than the maximal values of the cross-year correlations.For the correlations S VOS −P and S Bz −P (components orthogonal to S P and J ) the absolute values of zero-shift one-year and maximal cross-year correlations are approximately equal.
One more result, which is influenced sufficiently by cyclic variations, is the existence of the relation between ULF power on the ground and ion flux independent on V , S P and S Bz (last line in the table).Although the two-year averaged correlation for S VOS −J "pure" correlation is higher than both cross-year maximal correlations, only one-year correlation for 1998 is 2 times higher than the maximal cross-year correlation, while the correlation for 1997 is nearly equal to it.Thus this result is doubtful and requires more precise analysis.
Discussion and conclusion
This study has shown that the high latitude Pi3 (1-4 mHz) activity is controlled by several interplanetary parameters.The relationships between the fluctuations in this frequency band and basic interplanetary parameters: the SW velocity, SW dynamic pressure, and IMF fluctuations, are similar, but not the same as the known relationships for common Pc5 pulsations.The ULF response in this frequency range turns out to be more noticeable at the polar cap latitudes than at auroral latitudes.This distinction may be caused by several reasons: -lower background ULF activity in the polar cap region; -occurrence of additional magnetospheric mechanisms of ULF generation in the cusp/auroral regions, masking the IMF/SW control; -easier KH instability excitation and ULF wave penetration on the flanks of the magnetotail lobes.The primary drivers of high-latitude magnetic Pi3 fluctuations, especially in the polar caps, are turned out to be the SW velocity V and the group of parameters almost independent on V , but correlated with the SW dynamic As all the three parameters, i.e.P , S P and J , correlate with each other and their correlation coefficients with Pi3 power are close to each other, we use the following additional test to choose between these parameters those actually which control the Pi3 spectral power.The flux of SW protons with J (E<30 keV) and ULF spectral power S both on the ground and in space have almost no correlation with higher energy (E>1 MeV) solar protons generated during SPE.As it was mentioned above, J increases with the SW dynamical pressure.However, there are time intervals when the J -P correlation is low.These intervals are characterized by a steep increase in proton flux in all energy channels and, thus, they correspond to SPE.The S−J correlation for these intervals is approximately the same as for quiet intervals, while S−P correlation for SPE intervals vanishes.Thus, the flux of suprathermal protons is directly related to Pi3 power, while its correlation with P is an artifact.The inter-relations between all the parameters studied are summarized in Fig. 12, where the average correlation coefficients are shown by different types of lines.The correlations between all the parameters (except P −V pair) are statistically significant.To make the picture readable, we take only one ground station (VOS).The picture for other polar stations (ALE and THL) is very similar to the one depicted in Fig. 12.The correlations between ULF spectral power in the polar caps and in space are close to each other, except the correlation with V , which is lower in the interplanetary space.For V and S P -independent components a significant correlation remains between Pi3 power, both on the ground and in the SW, and the proton flux.The correlation coefficients S VOS −J and S Bz −J are approximately equal, i.e. suprathermal protons influence the power of fluctuations on the ground and in space in a similar way.However, the correlation between the proton flux and Pi3 power components independent also on the power of S Bz is about 0.13.This may indicate the possibility of generation of ULF noise by suprathermal protons not only in the foreshock, but also inside the bow shock.Special studies are necessary to study the reliability of this relation and establish mechanisms for it.
An important result of our analysis is the puzzling asymmetry in S−V covariations corresponding to preceding of the Pi3 power variations to the variations of the SW velocity.Most probably, this asymmetry is associated with the typical structure of CIRs in which high speed SW streams are preceded by the plasma compression.Compressed SW segments are characterized by the enhanced pressure and IMF fluctuations, generating ULF response in the magnetosphere.The examples of events with high peak-to-peak correspondence between the 1 mHz SW dynamic pressure fluctuations and variations of the Earth's magnetic field were presented by Kepko et al. (2003) proving the possibility for the magnetosphere to be a forced low quality oscillator.
CIRs are also the sources of accelerated ions.Acceleration of SW protons to energies E>40 keV takes place in in-terplanetary shocks while CIRs are known to be a source of suprathermal protons of lower energies even if a shock is not formed (Gloeckler, 1984;Richardson and Zwickl, 1984).A CIR necessarily includes a high pressure segment in the zone where a high speed SW meets a slow flow.As was found by Chotoo et al. (2000), maximal flux position of lower energy (less than several 100 keV) ions is shifted anti-sunward as compared with the Mev ions and corresponds to the zone of dense decelerated SW.The 10-30 keV protons are, probably, stochastically accelerated by the SW fluctuations, therefore, the proton flux correlates with the SW density (Trattner et al., 1991).The evolution SW supra-thermal protons and its relation to the ULF fluctuations in the interplanetary space have not been sufficiently explored.Possibly, an enhanced ULF power can be generated by the differential streaming of keV-ions at super-Alfvenic velocities through the SW.A possibility of the MHD wave generation under differential streaming was discussed by Leubner (2004) and Dubinin et al. (2005).
The main mechanism of the Earth's magnetosphere influence on the proton flux upstream of the Earth's bow shock is the acceleration at the bow shock (McKenna-Lawlor et al., 2003).Additionally, a close relation between particle flux fluctuations (electrons in the magnetosheath) and Pc5 oscillations was observed by Sarafopolous et al. (2001), who proposed a leakage of particles from the magnetosphere as a physical mechanism.These two mechanisms may be responsible for proton events measured at interplanetary monitor magnetically conjugate with the magnetosphere.However, these mechanisms cannot interpret our results, because intervals with possible Wind-magnetosphere conjugacy are very rare.More than 90% of 2 h intervals from our dataset correspond to the WIND location at X>30 R E , and among them there are only very few conjugated intervals with low IMF cone angle θ, such that tan(θ )<h/x, where h is the magnetosheath typical lateral scale and x is the GSE Sun-Earth coordinate of WIND.Taking h=30 R E , we get only ∼160 from ∼4000 subsets and these intervals give almost no contribution to the J −S correlation.Thus, the asymmetry of the S−V covariation cannot be explained by the foreshock extension to the Wind location.Thus, the finding of the significant role of supra-thermal ion fluxes in the stimulation of the Pi3 activity at polar latitudes is a challenge to the ULF community.The identification of relevant mechanisms may require a comparison of the differential fluxes in SW, magnetosheath, and magnetotail.
Fig. 1 .
Fig.1.The total horizontal spectral power density (in log scale) for the period 1997-1998 at 3 nearly conjugated stations VOS-THL (polar cap), A81-GHB and MAW-BJN (auroral) from the Southern (red) and Northern Hemispheres (blue).X-scale is day numbers starting from 1 January 1997.Codes for Northern (Southern) stations are given in upper (lower) part of each plot.
Fig. 2 .
Fig. 2. Averaged over two years(1997)(1998) correlation coefficient C between S (1-1.5 mHz band) and V for 7 high latitude stations in dependence on sample length L (in points).Solid (dashed) lines correspond to the Southern (Northern) Hemisphere, respectively.Warm colors denote polar cap stations, and cold colors denote auroral stations.Among the Northern auroral stations only those with the maximal and minimal correlations are shown.Black dashed-dotted line shows the correlation between V and the power of IMF B z fluctuations S Bz .
Fig. 3 .
Fig.3.The same as Fig.2, but for the correlation between S and the SW dynamic pressure P .
Fig. 4 .
Fig. 4. (a)The same for the correlation between ULF power at VOS (blue) and WIND (magenta) and solar wind dynamical pressure P (solid) and amplitude of its fluctuation S P (dashed).(b) The same, but for "pure" components (orthogonal to the parameters given in the upper indexes).
Fig. 5 .Fig. 6 .
Fig.5.Two years averaged correlations between magnetic fluctuations in the polar cap S VOS (solid) and in the interplanetary space at WIND S Bz (dashed) and interplanetary proton fluxes in respect to energy at L=30.
Fig. 9 .
Fig. 9.The same as Fig. 8, but for the covariate functions S VOS − J (8−30 keV), S P , P , and S Bz .Jumps in S VOS −S P covariate function are due to gaps in S P time series.
Fig. 12 .
Fig. 12. Summary sketch of averaged correlations between all the parameters studied.The power spectral density at VOS and in the SW are taken in the low-frequency band (1-1.5 mHz).
Database of ground and spacecraft observations
It may help to identify the cause/reason, Fig. 8. Two year averaged S−V covariate functions C(τ ) for various stations.Only maximal and minimal covariate functions from auroral stations in the Northern Hemisphere are shown.Positive lag corresponds to S leading. | 2014-10-01T00:00:00.000Z | 2007-05-08T00:00:00.000 | {
"year": 2007,
"sha1": "89379eecb9520c67b2b33f231653c4839d5ee769",
"oa_license": "CCBY",
"oa_url": "https://doi.org/10.5194/angeo-25-1037-2007",
"oa_status": "GOLD",
"pdf_src": "MergedPDFExtraction",
"pdf_hash": "89379eecb9520c67b2b33f231653c4839d5ee769",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics"
]
} |
258175299 | pes2o/s2orc | v3-fos-license | Biochemical evaluation of Ceiba pentandra leaf on the amelioration of testosterone induced-benign prostatic hyperplasia in Wister rats
Benign prostatic hyperplasia (BPH) is a considerable public health problem prevalent among older men. Despite its eminent prevalence, the etiology of BPH is not well understood. Accumulating evidence indicates that modifiable risk factors of cardiovascular disease may also increase the risk of BPH and potentially contribute to BPH development. This study evaluated the ameliorative effect of Ceiba pentandra leaf on some lipids profile indices and bilirubin concentration of rats induced BPH. Thirty six Wistar rats were divided randomly into six groups (n = 6). All the experimental groups except the normal control were induced benign prostatic hyperplasia using testosterone propionate and estradiol valerate at a dose of 400 µg and 80 µg respectively, for 21 days. After induction, rats in the negative control (BPH Control) group received no treatment. Rats in the standard control group received finasterides (1mg/kg body weight), while rats in the low dose extract, medium dose extract and high dose extract groups received the plant extract according to their body weights in kilogram at 500 mg/kg, 1000 mg/kg, and 1500 mg/kg respectively. Normal Control group received only feed without any special treatment. The animals in all the groups were allowed access to water and feed ad libitum for 28 days. The rats were anaesthetized after treatment period. Blood samples were collected and serum harvested for analyses using standard methods. Result showed that Ceiba pentandra leaf had ameliorative effect on the lipids profile analyzed.
Introduction
Benign prostatic hyperplasia (BPH), commonly referred to as benign enlargement of the prostate, is an age-and hormone-related condition that causes varying enlargement of the prostate and histological alterations in the prostate gland.Urinary urgency, slow stream, nocturia, and increased daytime frequency are only a few of the symptoms brought on by prostate enlargement (Djavan, 2003;Lee et al., 2012).The quality of life for BPH sufferers is severely impacted by these symptoms (Sagnier et al., 1995;Lee et al., 2016).The mechanism underlying the pathogenesis of BPH remains largely unidentified, however, a number of overlapping and complementary theories have been proposed.The pathophysiology of BPH involves hormonal changes in an aging man, albeit the exact mechanism is not yet entirely understood (Veeresh et al., 2010).The development and growth of normal prostate mainly depends on androgen stimulation, by dihydrotestosterone (DHT), which is produced from the prostrate by 5α-reductase enzyme (Cho et al., 2013;Carson and Rittmaster, 2003).
There are two primary types of treatment for BPH patients: α1-adrenergic receptor antagonists to reduce smooth muscle tone in the prostate and the bladder neck, and 5α-reductase inhibitors to reduce prostate size (Fine and Ginsberg, 2008).The most often prescribed drugs for treating BPH historically have been tamsulosin and finasteride (Lee, 2003).Only 64% of men receiving both medications, according to McConnell et al. (2003) showed a decreased risk of clinical progression, which is indicated by symptoms getting worse, acute urine retention, and incontinence and urinary tract infection.Additionally, these medications caused unpleasant side effects including low libido, erectile dysfunction, postural hypotension, asthenia, and occasional syncope (Akiyama et al., 1999;Lee eta l., 2014).It is highly desirable to develop a α1-adrenergic antagonist or other medication that can selectively suppress the smooth muscle tone of lower urinary tract without vascular effects and decrease prostate volume without sexual dysfunction for the treatment of urinary outlet obstruction in BPH (Akiyama et al., 2002) Natural compounds produced from plant sources have long been the primary sources of novel medications for the treatment of a variety of ailments (Dharmani et al., 2006).The multi-purpose Ceiba pentandra plant, which is native to northern Nigeria, and other tropical West African locations, has been found to have a variety of pharmacological applications.It is a plant from the Bombacaeae family.The presence of secondary metabolites like polyphenol, flavonoids, alkaloids, and saponins in Ceiba pentandra plant is responsible for its promising medicinal properties.Some of these compounds have been identified to possess antioxidant properties; scavenging free radicals produced by oxidation -reduction reactions (Nandeesh et al., 2008).Previous research on the plant's morphology in different areas has confirmed that the plant is also utilized as a hypoglycemic agent and has been proven to be a successful treatment for rheumatism, headaches, and vertigo (Ngounou et al., 2000).Typically, monitoring the progression of BPH involves evaluating specific biochemical markers (Kellogg, 2007).This study evaluated how extract of Ceiba pentandra leaf affected some lipids profile indices in testosterone propionate and estradiol valerate-induced benign prostatic hyperplasia in male Wistar rats.
Collection and preparation of Ceiba pentandra leaf extract
The plant leaves were purchased from a local market in Okuku Local Government Area, Cross River State.The plant were authenticated by a Botanist, Prof. S. Udoh of the Department of Biology, Cross River University of Technology (CRUTECH), Calabar with a voucher number BOT/PA/2022/017.The sample were deposited in the herbarium of the same department.The fresh leaves of Ceiba pentandra were washed with distilled water and then allowed to get dried in a dust-free environment for ten days.The dried leaves were blended using an electronic blender.
Experimental animals
A total of thirty six (36) male Wistar rats weighing 185 -220 grams were obtained from the Department of Medical Biochemistry, Cross River University of Technology, Okuku Campus, and transferred to the Animal House of the Department of Biochemistry Cross River University of Technology (CRUTECH) Calabar, Nigeria.The animals were acclimatized for 7days on pelletized rat chow and water ad libitum and maintained under standard housing conditions of adequate ventilation and room temperature (25 0 C ± 5 0 C) and relative humidity (46% ±5%) with a natural 12hr light-dark regimen.
Drug/Chemicals
Testosterone (Testosterone Propionate) Testosterone propionate (brand name Testost -manufactured by Laborate Pharmaceuticals India Limited) was purchased from Bez Pharmarcy and stores, No. 27 Etta-Agbor Street, Calabar, Nigeria, after full explanation of the purpose for procurement.
Ketamine Injection (Ketamine hydrochloride)
Ketamine is a medication mainly used for starting and maintaining anaesthesia (brand name Ketalar -manufactured by Sular Pharmaceuticals, India) was purchased from Bez Pharmacy and Stores, No. 27 Etta-Agbor Street, Calabar and was used to anaethesize the rats prior to bilateral orchiectomy.
Extraction of Plant Materials
The powdered material (500 g of C. pentandra) was weighed and soaked in 650 ml distilled water and allowed to macerate at room temperature for 48 hours.The solution was filtered after 48 hours into a beaker using Whatman filter paper No 4. The suspension was filtered through Whatman filter paper No 4 and dried in water bath at 55°C.The dried extract was weighed and stored in a clean reagent bottle, then preserved in a refrigerator at 4 o C until its use.
Aqueous concentration of 500 mg/kg, 1000 mg/kg, and 2000 mg/kg of C. pentandra extract was prepared using distill water and used as administrative doses
Experimental design
The animals weighing 185 and 220g were divided into six (6) groups with six (6) animals each.The six groups were: Group 1 (Normal), Group 2 (benign prostatic hyperplasia control), Group 3 (administered 3 mg/kg b. wt.testosterone + 5 mg/kg finasteride), Group 4 (administered 3 mg/kg b. wt.testosterone + 500 mg/kg of C. pentandra leaves extract), Group 5 (administered 3 mg/kg b. wt.testosterone + 1000 mg/kg of the extract), Group 6 (administered 3 mg/kg b. wt.testosterone + 2000 mg/kg of the extract).Benign prostatic hyperplasia was induced in Wistar rats using testosterone propionate for 28 days.Administration of extract was done for another 28 days, and then followed by accurate measurement of the changes.The animals were anesthetized using chloroform, and the blood samples of each animal were collected through the cardiac puncture method.The blood samples were aliquoted into the appropriate sample bottles (Plain and EDTA bottles), liver, kidney, and testes of Wistar rats were harvested into universal bottles containing formalin and transported to the laboratory.The design of the experimental procedure used in the study is shown in Table 1.
Evaluation of biochemical parameters
The rats received treatment for 28 days, and on the first and 28th days, their weights were recorded.They underwent anesthesia at the conclusion of the experiment.The prostate glands were removed and weighed.Half of the group's prostate were sent to the toxicology lab for biochemical analyses, including lactate dehydrogenase (LDH), superoxide dismutase (SOD), malondialdehyde (MDA), and follicle stimulating hormone (FSH) the remaining halves were put in containers with 10% formalin and stained for histology with hematoxylin and eosin.
PSA content
After anesthetizing the rats, blood samples were collected from their hearts and sent to an associate laboratory to evaluate free prostate specific antigen (PSA) in the blood using the PSA ELISA kit Akanni OO, et al., (2017).
Body weight and prostate weight
Body weight were measured in the beginning and the end of the experiment.Also, prostate heights were measured after ketamine hypochlorite combination was used to anesthetize the animals at a dose of 0.15 ml/kg of body weight administered intraperitoneally and removing prostate gland.
The prostate weight to body ratio can be calculated using the prostate index formula.Any changes in this index, when compared between each group and the normal and illness groups, show the degree to which the disease is being prevented from progressing.Prostate index and inhibitory prostate weight are computed using the two measured weights and these formulas.
. Prostate index = Prostate weight Body weight
Percentage of inhibition = 100
Histopathological examination
To investigate the morphological changes, the tissue was cut into 4-nM sections using a rotatory microtome, stained with eosin and hematoxylin and was observed under a 40x microscope (Kim, Sk, Seok H et al., 2015).
Statistical analysis
Data obtained were analyzed using one-way ANOVA followed by least square difference (LSD) post-hoc comparison test to evaluate significant difference between the mean values of the experimental and control groups.Differences at P < 0.05 were regarded as significant.Graphpad prism version 7 and SPSS software package version 23.0 were used for the statistical analyses.
Results
Effect of aqueous leaf extract of C. pentandra on semen parameters of testosterone propionate-induced benign prostatic hyperplasia in male rats Table 2 displays the findings on the impact of C. pentandra aqueous leaf extract on the semen parameters of testosterone propionate-induced benign prostatic hyperplasia in male rats.In comparison to the negative control, male rats in Groups 3 to 6 had considerably larger amounts of semen (p > 0.05).In comparison to the negative control, groups 4 to 6 exhibit an increase in viability, normal activity, and sperm count that is not statistically significant (p > 0.05).When compared to the negative control, Group 3's viability (82.00%), normal (85.00%),activity (86.00%), and sperm count (710.00 ml/106) all increased significantly (p > 0.05).Group 4 to 6 had a non-significant decrease (p > 0.05) in abnormal, Sluggish, and dead sperm cells analyzed when compared to control while Group 3 had a significant decrease (p > 0.05) in abnormal (14.00 %), Sluggish (6.00 %), and dead (8.00 %) sperm cells compared to the negative control.
Effect of aqueous leaf extract of C. pentandra on oxidative stress and hormone levels of testosterone propionateinduced benign prostatic hyperplasia in male rats
Table 3 displays the findings of the study on the impact of C. pentandra aqueous leaf extract on oxidative stress and hormone levels of testosterone propionate-induced benign prostatic hyperplasia in male rats.Male rats in Groups 3 to 6 had non-significantly higher levels of lactate dehydrogenase (LDH) (u/l) compared to the positive and negative controls (p > 0.05).When compared to the negative and positive controls, Superoxide Dismutase (SOD) levels in Groups 3 to 6 were considerably lower (p > 0.05).When compared to the negative and positive controls, malondialdehyde (MDA) levels in Groups 3 to 6 were considerably higher (p 0.05), while Group 5 exhibited a significant rise when compared to Groups 3, 4, and 6.Findings revealed that male rats' levels of Follicle Stimulating Hormone (FSH) (miu/ml).
A significant increase (p < 0.05) in testosterone level was observed in Group 2. However, Groups 4 to 6 also showed a significant decrease (p < 0.05) in testosterone level compared to Group 2.
Effect of aqueous leaf extract of C. pentandra on hematological indices of testosterone propionateinduced benign prostatic hyperplasia in male rats
The results of the effect of aqueous leaf extract of C. pentandra on hematological indices of testosterone propionateinduced benign prostatic hyperplasia in male rats presented in Table 4 showed that there was no significant difference (p < 0.05) in the PCV and HB levels of the test groups compared to the positive and negative control.The same nonsignificant decrease (p < 0.05) was obtained in PLT, MCH, MCV, MCH, RBC, Leucocyte, Neutrophils, Mesophiles, Basophils, and Eosinophils.Values are presented as mean ± SD (n = 3).Mean values with same superscript letters along the column are not statistically significant at p < 0.05.PCV -Packed Cell Volume, HB -Hemoglobin, WBC -White blood cell-, MCV -mean corpuscular volume, MCH -mean corpuscular hemoglobin, MCHC -mean corpuscular hemoglobin content
Effect of aqueous leaf extract of C. pentandra on cardiac and cancer markers (ng/ml) of testosterone propionate-induced benign prostatic hyperplasia in male rats
The results of the effect of aqueous leaf extract of C. pentandra on cardiac markers (ng/ml) of testosterone propionateinduced benign prostatic hyperplasia in male rats presented in Table 4 showed a non-significant decrease (p < 0.05) of CKMB concentration in Groups 3 to 6. D-DIMER concentration showed a non-significant increase (p < 0.05) in Groups 3 to 6. Myoglobin concentration showed a non-significant decrease (p < 0.05) in Groups 3 and 6, and a non-significant increase (p < 0.05) in Groups 4 and 5. Prostate-Specific Antigen (PSA) (ng/ml) of male rats in Groups 3 to 6 show a significant increase (p > 0.05) when compared to the positive and the negative control.Similarly, there was a nonsignificant decrease (p > 0.05) in the level of Carcinoembryonic antigen of rats in Groups 3 to 6 when compared to the positive and negative controls.
Effect of aqueous leaf extract of C. pentandra on liver markers of testosterone propionate-induced benign prostatic hyperplasia in male rats
Results of the effect of C. pentandra aqueous leaf extract on the liver of testosterone propionate-induced benign prostatic hyperplasia in male rats are shown in Table 5.When compared to the negative control, the ALP concentration in male rat Groups 3 to 6 was significantly lower (p 0.05).When compared to the negative control, the AST concentration increased significantly (p 0.05) in Groups 3 and 4 and non-significantly (p > 0.05) in Groups 5 and 6.When compared to the negative control, ALT concentration results showed a substantial increase (p 0.05) in Groups 3 and 4 and a nonsignificant decrease (p > 0.05) in Groups 5 and 6.Total-Protein concentration showed a non-significant increase (p>0.05) in Groups 3 and 4, a significant increase in Groups 5, and a non-significant decrease in Group 6 compared to the negative control.Albumin concentration in Groups 3 and 4 showed a non-significant decrease (p > 0.05).However, there was a non-significant increase in Groups 5 and 6 when compared to the negative control.Values are presented as mean ± SD of the triplicate determinant (n = 3).Mean values with the same superscript letters along the column are not statistically significant at p < 0.05.AST=Aspartate aminotransferase, ALT=Alanine Aminotransferase, ALP= Alkaline Phosphatase
Effect of aqueous leaf extract of C. pentandra on the kidney of testosterone propionate-induced benign prostatic hyperplasia in male rats
The results of the Effect of aqueous leaf extract of C. pentandra on the kidney of testosterone propionate-induced benign prostatic hyperplasia in male rats presented in Figure 1 shows a non-significant increase (p > 0.05) ) in creatinine and urea concentration in Groups 3 to 6 of male rats compared to the negative control.There was a non-significant increase (p > 0.05) in potassium concentration in Groups 3 to 4 compared to the negative control, and a non-significant decrease (p > 0.05) in potassium concentration in Groups 5 to 6. Results showed a non-significant decrease (p > 0.05) in sodium concentration in Groups 3 to 6 compared to the negative control.
Figure 2
Effect of aqueous leaf extract of C. pentandra on the kidney of testosterone propionate-induced benign prostatic hyperplasia in male rats from Groups 1-6.
Discussion
The results of Amadi et al. [16] who investigated the potential ameliorative effects of C. pentandra on sodium fluorideinduced toxicity on hematological indices and fecundity of adult male Wistar rats and discovered that leaf extracts of C. pentandra have no toxic effect on testes of male rats corroborated the findings of C. pentandra having no toxic effect on testes of male rats.. C. pentandra is antioxidative, anticancer, antimalarial, antihelmintic, piscicidal, antiviral, and antibacterial properties, implying a wide range of possible applications.Asimicin, bullatacin, and bullatalicin are three significant acetogenins found in ripe C. pentandra Hormonal level of testosterone propionate-induced benign prostatic hyperplasia in male rats administered with aqueous leaf extract of C. pentandra showed a significant decrease (p > 0.05) in FSH levels for male rats in Groups 3 to 6 treated with C. pentandra compared to the negative control.This agrees with the study of Amadi et al. [16] who investigated the potential ameliorative effects of aqueous leaf extract of C. pentandra.There was a decrease in testosterone and FSH concentrations in Groups 4 to 6 compared to the negative control.This decrease in FSH may be due to C. pentandra toxic effect on the anterior pituitary gland, causing oxidative damage to the anterior pituitary gland's architecture, responsible for the generation of FSH (glycoprotein hormone) and luteinizing hormone.The decline in testosterone levels in groups administered with aqueous leaf extracts of C. pentandra suggested that its phytochemical composition may reduce the effect of the pituitary gland and testosterone synthesis [19] The effect of aqueous leaf extract of C. pentandra on oxidative stress of testosterone propionate-induced benign prostatic hyperplasia in male rats had a non-significant increase (p > 0.05) in Lactate Dehydrogenase (LDH) activity in Groups 3 to 6 compared to the positive and negative control.Superoxide Dismutase (SOD) concentration in Groups 3 to 6 was significantly decreased (p > 0.05) compared to negative and positive control.Malondialdehyde (MDA) in Groups 3 to 6 was significantly increased (p > 0.05) compared to the negative and positive control, and Group 5 had a significant increase compared to Groups 3, 4, and 6.This is in agreement with the findings of Agu et al. [20] who worked on "protective effect of ethanol extract of soursop (annona muricata linn) leaves on cycad induced oxidative stress in male albino wistar rats", and discovered an increase in LDH and MDA concentration in rats administered with C. pentandra compared to that of the negative control, and that the SOD concentration decreased compared to that of the negative control.The observed increase in LDH concentration of Groups 3 to 6 compared to negative control is owing to tissue damage produced by rats' exposure to testosterone propionate, a well-known toxin as described above.Malondialdehyde (MDA) is a measure of polyunsaturated lipid peroxidation caused by reactive oxygen radicals (ROS).These lipids are essential components of biological membrane structures.Membrane lipids are predominantly attacked by free radicals, resulting in their peroxidation [21].This causes membrane instability, breakdown of membrane proteins and other macromolecules, and cellular damage [22].The increased levels of oxidative stress caused by testosterone propionate were reduced in rats administered with C. pentandra.This reduction was via the reduction in levels of SOD, LDH, and MDA in the rats, supporting the findings of Lolodi and Eriyamremu [23] who found a decrease in SOD activity in rats given Cycads.
There was a decrease (p > 0.05) in PSA level in male rats of Groups 3 to 6 compared to the positive and the negative control.Testosterone propionate-induced benign prostatic hyperplasia is a pointer to the ameliorative impact of testosterone propionate-induced benign prostatic hyperplasia.As a direct result of inhibiting 5-reductase, a decrease in PSA levels is linked to the reduction in prostatic hyperplasia [24] Furthermore, the fact that the PSA level in the negative control group remained high indicates that the observed decline in PSA in treated rats was due to the ameliorative effect of C. pentandra.This corroborates the findings of Ogbu et al. [25] who discovered dutasteride and acetogenin-rich fraction isolated from Annona muricata leaves reduced the levels of PSA in rats with testosterone propionate-induced benign prostatic hyperplasia.
The Effect of aqueous leaf extract of C. pentandra on hematological indices of testosterone propionate-induced benign prostatic hyperplasia in male rats showed normal hematology indices in the test groups when compared to the negative control.Previous studies have shown that some plant extracts including leaf extract of A. muricata have the potential to positively affect the erythropoietic system [16,26,27,28].
The effect of aqueous leaf extract of C. pentandra on cardiac markers of testosterone propionate-induced benign prostatic hyperplasia in male rats revealed there was a non-significant decrease (p < 0.05) of CKMB concentration in Groups 3 to 6 compared to that of the negative control.D-DIMER concentration had a non-significant increase (p < 0.05) in Groups 3 to 6 compared to the negative control.Myoglobin concentration had a non-significant decrease (p < 0.05) in Groups 3 and 6, and a non-significant increase (p < 0.05) in Groups 4 and 5.The result of CKMB, D-DIMER, and Myoglobin from this study indicates the aqueous leaf extract of A. muricata has no deleterious effect on the functional integrity of the cardiac system.This corroborates the findings of Niu et al. [29] who worked on "Hyperbaric oxygen improves survival in heatstroke rats by reducing multiorgan dysfunction and brain oxidative stress" and discovered there was an increase of D-dimer concentration in heatstroke rats.
The liver which is responsible for xenobiotic metabolism and detoxification is also susceptible to hepatotoxic chemicals [30].The effect of aqueous leaf extract of C. pentandra on the liver of testosterone propionate-induced benign prostatic hyperplasia in male rats revealed a significant decrease in ALP concentration in Groups 3 to 6 of male rats compared to the negative control, which agrees with the study of Syahida et al. [28] on "Soursop (Anonna muricata): Blood hematology and serum biochemistry of Sprague-Dawley rats" who discovered ALP concentration in rats treated with C. pentandra was significantly decreased compared to their negative control.The AST concentration was significantly increased (p > 0.05) in Groups 3 and 4 and Groups 5 and 6 showed a non-significant increase compared to the negative control.The result showed that ALT concentration was significantly increased (p > 0.05) in Groups 3 and 4.However, there was a non-significant decrease (p < 0.05) in Groups 5 and 6 compared to the negative control.T-Protein had a non-significant increase (p > 0.05) in Groups 3 and 4, a significant increase in Group 5, and a non-significant decrease in group 6 compared to the negative control.Albumin concentration in Groups 3 and 4 had a non-significant decrease (p < 0.05) and a non-significant increase in Groups 5 and 6 compared to the negative control.This liver-function investigation revealed that aqueous leaf extract of C. pentandra has no acute hepatotoxic effect.The enzymes AST and ALT are crucial in the breakdown of amino acids channeled into the Krebs cycle and electron transport chain [31] Because ALT and ALP are concentrated in the liver, they are considered accurate indicators of liver disease [27,32].Changes in membrane-bound ALP negatively impact membrane permeation, disrupting the transportation of metabolites [31].When the liver is diseased, these enzymes are released into the system in excess of a critical concentration [32].
By measuring the serum levels of urea, creatinine, potassium, and sodium, renal dysfunction was examined.In the blood serum of patients with defective kidneys, these indicators are found above normal levels, according to Edmund and David [32].When urea, creatinine, sodium, and potassium levels in groups 3 through 6 were compared to the positive and negative controls, there was no statistically significant difference (p > 0.05) between them and the kidneys of testosterone propionate-induced benign prostatic hyperplasia in male rats.The renal function investigation on groups treated with aqueous leaf extract of C. pentandra compared to that of the negative control revealed a non-toxic effect on kidney and renal function.Onyegeme-Okerenta et al. [34] in their study on the ameliorating potential of Annona muricata on sodium Fluoride-induced toxicity on lives and kidneys of male Wistar rats concluded that concomitant treatment of NaF leaf extract of Annona muricata resulted in mild/moderate amelioration and generation of damaged hepatic and renal tissues
Conclusion
The findings of this study shows that testosterone propionate-induced benign prostatic hyperplasia in male rats might be ameliorated by aqueous leaf extract of C. pentandra when taken over an extended period of time.On testosterone propionate-induced benign prostatic hyperplasia in male rats, the aqueous leaf extract of C. pentandra showed improved sperm count, volume, normalcy, and viability, while decreasing the amount of abnormal, slow, and dead sperm cells.The sex hormones (Follicle Stimulating Hormone and Testosterone) in male Wistar rats may be brought back to normal by C. pentandra.Moreover, it promotes spermatogenesis, which improves the quality of the semen.The administration of C. pentandra's aqueous leaf extract also demonstrated potential for improving hematological indicators, cardiac, renal, and cancer outcomes.
Figure 1
Figure 1 Prostate Specific Antigen
Table 1
Design of experimental procedureEthical approval for the treatment and handling of experimental animal and human subjects was obtained from the Faculty Animal Research Ethics Committee on Use and Care of Experimental Animals, Faculty of Basic Medical Sciences,
Table 2
Effect of aqueous leaf extract of C. pentandra on semen parameters of testosterone Propionate-induced benign prostatic hyperplasia in male rats Values are presented as mean ± SD of triplicate determinant (n = 3).
Table 3
Effect of aqueous leaf extract of C. pentandra on oxidative stress and hormone levels of testosterone propionateinduced benign prostatic hyperplasia in male rats
Table 4
Effect of aqueous leaf extract of C. pentandra on haematological indices of testosterone propionate-induced benign prostatic hyperplasia in male rats
Table 5
Effect of aqueous leaf extract of A. muricata on liver markers of testosterone propionate-induced benign prostatic hyperplasia in male rats | 2023-04-17T15:10:00.335Z | 2023-04-30T00:00:00.000 | {
"year": 2023,
"sha1": "911868e4b905cfb6f5eba679689e80c31f1d80d0",
"oa_license": "CCBYNCSA",
"oa_url": "https://wjarr.com/sites/default/files/WJARR-2023-0623.pdf",
"oa_status": "HYBRID",
"pdf_src": "Anansi",
"pdf_hash": "613d3d0808a71eadb75245df72e3a2dca8a0dd46",
"s2fieldsofstudy": [
"Agricultural And Food Sciences"
],
"extfieldsofstudy": []
} |
12667501 | pes2o/s2orc | v3-fos-license | Is the Linear Modeling Technique Good Enough for Optimal Form Design? A Comparison of Quantitative Analysis Models
How to design highly reputable and hot-selling products is an essential issue in product design. Whether consumers choose a product depends largely on their perception of the product image. A consumer-oriented design approach presented in this paper helps product designers incorporate consumers' perceptions of product forms in the design process. The consumer-oriented design approach uses quantification theory type I, grey prediction (the linear modeling technique), and neural networks (the nonlinear modeling technique) to determine the optimal form combination of product design for matching a given product image. An experimental study based on the concept of Kansei Engineering is conducted to collect numerical data for examining the relationship between consumers' perception of product image and product form elements of personal digital assistants (PDAs). The result of performance comparison shows that the QTTI model is good enough to help product designers determine the optimal form combination of product design. Although the PDA form design is used as a case study, the approach is applicable to other consumer products with various design elements and product images. The approach provides an effective mechanism for facilitating the consumer-oriented product design process.
Introduction
Products have been considered a symbol of occupation, personality, opinion, and other human attributes. Whether a product is successful largely depends on the final judgment of consumers [1]. Therefore, product designers need to comprehend the consumers' needs in order to design successful products (highly-reputable and hot-selling) in an intensely competitive market [2]. Moreover, a successful product should not only possess good functionalities, interface design, and operating performance, but also need to take the product image design into account to satisfy consumers' psychological requirements [3]. The external appearance of a product can represent a product image that evokes consumers' internal resonance and consuming motivation [4]. The product image engages an influential factor in consumers' preference structure [5]. When choosing a product, consumers tend to rely on their own particular perception of the product, which is regarded as something of a black box [6]. As an ergonomic consumer-oriented methodology, Kansei Engineering is developed as integrative design strategies for affective design to satisfy consumers' psychological requirements [7][8][9]. The word "Kansei" indicates the consumers' psychological requirements or emotional feelings of a product. Kansei Engineering has been used to assist product designers in designing product forms that can best match specific product images [10,11].
In this paper, we present a consumer-oriented design approach addressing for challenging issues in designing consumer products, such as personal digital assistants (PDAs). What are the key form elements for a desirable product image? How to use the adequate product form combination 2 The Scientific World Journal to enhance consumers' preference? Is there an optimal combination of product form that best matches a desirable feeling of the consumers? For example, if product designers want to design a product with "simple-to-look" appearance, are there guidelines of product form design to follow? In addition, nonlinear modeling techniques (such as the artificial intelligent system or the soft computing) are defined as "an emerging approach to reasoning and learning the human mind in an uncertainty and imprecision environment" [12,13]. These techniques are supposed to possess humanlike expertise within a specific domain, adapt themselves and learn to do better in changing environments, and explain how they make decisions [9,12]. Hence, are the nonlinear modeling techniques suitable for exploring the relationship between the consumers' perceptions of product images and product form elements? Or are the linear modeling techniques good enough to do so [14]? What specific technique should be used to help product designers determine the optimal form combination of product design for a particular design concept of product image? To illustrate how the approach can be used to answer these research questions, we conduct an experimental study on PDAs, using two linear modeling techniques and one nonlinear modeling technique. Two linear modeling techniques are the quantification theory type I (QTTI) [15] and the grey prediction (GP) [16], and the nonlinear modeling technique is the neural networks (NNs) [17].
The QTTI is a variant of linear multiple regression analysis and can be used to quantify the relationships between product form elements and product images [5], while the GP model can deal with incomplete information effectively and requires only four data sets or more [16]. As such, the GP can be used to predict how a particular combination of product form elements matches a product image, particularly when the information is available only for a limited number of product form elements [10]. Due to the effective learning ability, NNs have been applied successfully in a wide range of fields, using various learning algorithms [18][19][20]. NNs are well suited to formulate the product design process for matching the product form (the input variables) to the consumers' perceptions (the output variables), which is often a black box and cannot be precisely described [10].
In subsequent sections, we first present the quantitative analysis methods used to analyze the experimental data sets for answering the research questions. Then we conduct an experimental study on PDAs to describe how Kansei Engineering can be used to extract representative samples and product form elements as numerical data sets required for analysis. Finally, we discuss the results of applying these techniques and evaluate their performance in order to determine the better model that can be used to help product designers meet consumers' requirements for a desirable product image.
Methods of Quantitative Analysis
In this section, we present a brief outline of the relevant theories and algorithms, including the QTTI, the GP, and the NNs. We use these techniques to examine the relationship between product form elements and product images.
Quantification Theory Type I.
The QTTI can be regarded as a method of qualitative and categorical multiple regression analysis method [15], which allows inclusion of independent variables that are categorical and qualitative in nature, such as product form elements and quantitative criterion variables within Kansei Engineering. In Kansei Engineering, product form elements are typically classified into two levels that correspond to form design element and its treatments, respectively. The QTTI consists of the followings six steps [15].
Step 1. Define the Kansei relational model associated with the Kansei measurement scores of experimental samples with respect to an image word pair.
In Kansei Engineering, the criterion variables represent the product image, and the explanatory variables represent the product form elements. The categorical multiple regression model can be defined as where y k s : the predicted value of the criterion variable for the sth product sample on the kth image word; i: the index of design element, E: the number of design element; j: the index of category; C i : the number of category of the ith design element; ε: a stochastic variable whose expectation value E(ε) = 0; β i j : the category score of the jth style within the ith design element; x i js : the coefficient of the dummy variable that is the explanatory variable or the dummy variable representing the jth style within the ith design element using the sth experimental sample.
Step 2. Calculate the standardized regression coefficients and the standardized constant in the model. The model of categorical multiple regression analysis can be redefined as where β * i j represents the standardized coefficient of explanatory variables and y k s is the standardized constant in the model.
Step 3. Determine the matrix CCR of correlation coefficient of all variables.
Step 4. Calculate the multiple correlation coefficient R that is regarded as the relational degree of external criterion variable and explanatory variables.
The Scientific World Journal 3 Step 5. Calculate the partial correlation coefficients (PCC) of design elements to clarify the relationships between product form elements and a product image.
Step 6. Determine the statistical range of a categorical variable (product form element) by the difference between the maximum value and minimum value of the category score. The range of the categorical variable indicates its contribution degree to the prediction model with respect to a given product image.
Grey
Prediction. The grey system theory [16] has been developed to examine the relationship among factors in an observable system where the information available is grey, meaning uncertain and incomplete (i.e., only part of the information is known). It has been successfully used in a wide range of fields, including some recent application results [10,[21][22][23] highlighting its effective handling of incomplete known information for exploring unknown information. The system that can be built for answering specific research questions in product design with respect to product form and product image is grey in essence, as there is no way to identify all the product form elements that affect a particular product image perceived by consumers [10].
The GP model uses a grey differential model (GM) to generate data series from the original data series of a dynamic system. The data series generated by the GM are converted back to the original data series by a reverse procedure to predict the performance of the system. Since the generated data series are more coherent than the original, the accuracy of the modeling is enhanced. The GM has three basic operations [16]: (1) accumulated generation, (2) inverse accumulated generation, and (3) grey modeling. The accumulated generation operation (AGO) is used to build differential equations. The GM is usually represented as GM(M, N) for dealing with Mth-order differential equations with N variables. Since any higher-order differential equation can be transferred into a first-order differential equation, we use the first-order differential equation in this paper.
The GM(1, 1), a single variable and first-order grey model, is one of the most frequently used grey prediction models. Its procedure involves the following four steps.
Step 1. Denote the original sequence as where x (0) (i) is the time series data at time i (i = 1, 2, . . . , n).
Step 2. Generate a new sequence x (1) by the AGO based on the original sequence x (0) , where x (1) Step 3. Define the first-order differential equation as Step 4. Use the least square method to solve (4) by where The x (1) (k + 1) is the predicted value of x (1) (k + 1) and x (0) (k +1) is the predicted value of x (0) (k +1) at time k +1. We can also use the inverse accumulated generation operation (IAGO) to obtain x (0) (k + 1) as The GM(1, 1) grey model can be extended to the GM(1, N) model [10,16], first-order with N variables ( The differential equation can be defined as where a, b 1 , b 2 , . . . , b N−1 are unknown parameters and can be calculated by where 4 The Scientific World Journal The prediction of x 1 (1) is defined as The
Neural Networks.
NNs are nonlinear models and are widely used to examine the complex relationship between input variables and output variables [17]. In this paper, we use the multilayered feedforward neural networks trained with the backpropagation learning algorithm, as it is an effective and the popular supervised learning algorithm [10]. A typical three-layer network consists of an input layer, an output layer, and one hidden layer, with n, m, and p neurons, respectively (indexed by i, j, and k, resp.) [24]. The w i j and w jk represent the weights for the connection between neuron i (i = 1, 2, . . . , n) and neuron j ( j = 1, 2, . . . , m), and between neuron j ( j = 1, 2, . . . , m) and neuron k (k = 1, 2, . . . , p), respectively. In training the network, a set of input patterns or signals, (x 1 , x 2 , . . . , x n ) is presented to the network input layer. The network then propagates the inputs from layer to layer until the outputs are generated by the output layer. This involves the generation of the outputs (y j ) of the neurons in the hidden layer as given in (14) and the outputs (y k ) of the neurons in the output layer as given in (15).
where f (·) is the sigmoid activation function as given in (16), and θ j and θ k are threshold values: If the outputs (y k ) generated by (15) are different from the target outputs (y * k ), errors (e 1 , e 2 , . . . , e p ) are calculated by (17) and then propagated backwards from the output layer to the input layer in order to update the weights for reducing the errors.
The weights (w jk ) at the output neurons are updated as w jk + Δw jk , where Δw jk is computed by (known as the delta rule) where α is the learning rate (usually 0 < α ≤ 1) and δ k is the error gradient at neuron k, given as The weights (w i j ) at the hidden neurons are updated as w i j + Δw i j , where Δw i j is calculated by where α is the learning rate (usually 0 < α ≤ 1) and δ j is the error gradient at neuron j, given as The training process is repeated until a specified error criterion is satisfied.
Experimental Procedures of Consumer-Oriented Design
We conduct an experimental study using the concept of Kansei Engineering in order to collect numerical data about the relationship between product form elements and a given product image of PDAs. The experimental study involves three main steps: (a) extracting representative experimental samples, (b) conducting morphological analysis of product form elements, and (c) assessing consumers' perceptions for a given product image.
Extracting Representative Experimental Samples.
In the experimental study, we investigate and categorize various PDAs on the market. We first collect 88 PDAs and then classify them based on their similarity degree. To collect opinions The Scientific World Journal 5 regarding the usage, function, and form of PDAs, a focus group is formed by six subjects with at least two years' experience of using the PDA. The focus group eliminates some highly similar samples through discussions. Then the K-means cluster analysis is used to extract representative samples of PDAs. There are 30 representative PDA samples, including 24 samples as the training set and six samples as the test set for building quantitative models in Section 4.
Conducting Morphological Analysis of Product Form
Elements. The product form is defined as the collection of design features that the consumers will appreciate. The morphological analysis [25], concerning the arrangement of objects and how they conform to create a whole of Gestalt, is used to explore all possible solutions in a complex problem regarding a product form. The morphological analysis is used to extract the product form elements of the 30 representative samples. The six subjects of the focus group are asked to decompose the PDA samples into several dominant form elements and form types according to their knowledge and experience. Table 1 shows the result of the morphological analysis, with six product design elements (i.e., top shape, bottom shape, function-keys arrangement, arrow-key style, color treatment, and outline partition style) and 19 associated product form types being identified. The form type indicates the relationship between 6 The Scientific World Journal the outline elements. For example, the "top shape (X 1 )" form element has three form types, including "line (X 11 )," "chamfer (X 12 )," and "fillet (X 13 )." A number of design alternatives can be generated by various combinations of morphological elements [26].
Assessing Consumers' Perceptions of Product Images.
In Kansei Engineering, image assessment experiments are usually performed to elicit the consumers' psychological feelings or perceptions about a product using the semantic differential method. Pairs of image words are often used to describe the consumers' perceptions of the product in terms of ergonomic and psychological estimation. With the identification of the form elements of the product, the relationship between the image words and the form elements can be established.
In this paper, the image word pair used for representing the product image of PDAs is Simple-Complex (S-C) about the visibility aspect, according to our previous study [27]. In Wang et al. [27], we use these 30 representative PDA samples and product images to examine whether the NN model is an effective technique and what structure is better for the product form design among 4 NN models built with different hidden layer neurons. In this study, we use the same experimental data as a basis for addressing new and significant research issues as stated in Section 1.
To obtain the assessed values for the product image of 30 representative PDA samples, a 10-point scale (1-10) of the semantic differential method is used. 52 subjects (30 males and 22 females with ages ranging from 26 to 45, mean = 35.4, SD = 4.4) are asked to assess the form (look) of PDA samples on a simplicity-complexity scale of 1 to 10, where 10 is most simple and 1 is most complex. The last column of Table 2 shows the assessed S-C value of the 30 PDA samples, including 24 samples in the training set and six samples in the test set (asterisked). For each selected PDA in Table 2, the first column shows the PDA number and Columns 2-7 show the corresponding type number for each of its six product form elements, as given in Table 1. Table 2 provides a numerical data source for building quantitative models, which can be used to develop a design support system for simulating the optimal form design process for PDAs.
Experimental Analysis and Results
In this section, we present the results of applying the QTTI, the GP, and the NN models in order to explore the relationship between product form elements and consumers' perceptions for a given product image, using the assessing results summarized in Tables 1 and 2. 4.1. The QTTI Analysis and Results. We use the QTTI analysis to examine the relationship between the six product form elements and the S-C product image. In this paper, six independent variables (i.e., the six product form elements) and one dependent variable (i.e., the S-C product image) are used. The result of QTTI analysis is given in Table 3. In Table 3, the partial correlation coefficients indicate the relationship between the six product form elements (X 1 , X 2 , X 3 , X 4 , X 5 , and X 6 ) and the S-C product image (Y ). The highest variable of the partial correlation coefficient in the "S-C" image is the "arrow-key style" form element (X 4 = 0.42), meaning that "arrow-key style" primarily affects the "S-C" image of the product, followed by the "color treatment" form element (X 5 = 0.37) and the "top shape" form element (X 1 = 0.26). This implies that the product designers should focus their attention more on these most influential elements, when the objective of designing a new PDA is to achieve a desirable "S-C" image. On the contrary, the product designers can pay less attention to the less influential elements such as "bottom shape" form element (X 2 = 0.14), and the "function-keys arrangement" form element (X 3 = 0.16), as these form elements contribute relatively little to the consumers' perceptions of the "S-C" image on the PDAs.
In the last second row of Table 3, R means the correlation between the observed and predicted values of the dependent variable, and R 2 is the square of this correlation. R 2 ranges from 0 to 1. If there is no linear relation between the dependent variable (Y ) and independent variables (X 1 , X 2 , X 3 , X 4 , X 5 , and X 6 ), R 2 is 0 or very small. Otherwise, if all the values fall on the regression line, R 2 is 1. The category grade (form type grade) shown in Table 3 indicates the preference degree of the consumers' perception on each category of independent variables. If the grade is negative, the consumers' perception leans towards the "complex" image. On the contrary, the positive grade indicates that the consumers' perception favors the "simple" image. For example, the category grades of 3 selected values of "outline partition style (X 6 )" in the "S-C" image are −0.20, −0.13, and 1.32, respectively. The result shows that the consumers' perception prefers the "complex" image if the "outline partition style (X 6 )" is "normal partition (X 61 )" or "fitting outline (X 62 )," and favors the "simple" image while "outline partition style (X 6 )" is "fitting surface (X 63 )." As the result of the QTTI analysis, Model (22) indicates the relationship between product form elements and the S-C product image. We can use this model to input the values of six product form variables, and then output the predicted value of the S-C product image. This model can help the product designers understand consumers' perceptions to find out the optimal combination of product form design in terms of a given product image:
The GP Analysis and Results
. The GP is used as a technique for determining the optimal combination of product form elements for matching a desirable product image. The 24 samples in the training set, given in Table 2, are used as the data set for building the GP model. As a GM(1, 7), the GP model uses the six form elements as the comparison series X i and the average S-C values as the reference series X 0 . To build the GP model, we first obtain a new sequence x (1) for each series using (3)-(5) and the AGO as . . .
With the GP model in (25), product designers can input the value of the corresponding form elements, and then obtain a predicted S-C value.
The NN Analysis and Results.
To examine whether the NN model is an effective technique for determining the optimal combination of product form elements for matching a desirable product image, we develop two neural network models, called NN-FE and NN-FT, respectively. The NN-FE uses all the six form elements (FE) as input variables (input neurons), while the NN-FT has 19 input neurons, which are the whole 19 form types (FT) of the six form elements identified from the experimental study. For the NN-FE model, if a PDA has a particular type of form element, the value of the corresponding input neuron is 1, 2, 3, or 4. On the other hand, for the NN-FT model, if a PDA has a particular type of form element, the value of the corresponding input neuron is 1; otherwise the value is 0. Both NN models use a widely used rule [17], (the number of input neurons + the number of output neurons)/2, for determining the number of neurons in the single hidden layer. Table 4 shows the neurons of these two NN models, including the input layer, the hidden layer, and the output layer. The learning rule used is Delta-Rule and the transfer function is Sigmoid [17] for all layers. All of input and output variables (neurons) are normalized before training. The learning rate is 0.2, and momentum is 0.5, based on our previous study [28].
The experimental samples are separated into two groups: 24 training samples and six test samples. Each model is trained ten epochs at each run. When the cumulative training epochs are over 25,000, the training process is completed. The root of mean square errors (RMSE) of the NN-FE model is 0.057, while the NN-FT model is 0.052. This result seems to suggest that the number of input neurons and hidden neurons have little influence on the training effect of NN models. However, after further examination, we find out that if more neurons are in the input or hidden layer, the faster the convergence speed becomes (as shown in Figure 1). In other words, if the input layer or hidden layer has more neurons, then the network converges faster. This result suggests that if the input variable has multiple categories (i.e., the qualitative or categorical variable, such as product form elements), the total number of categories (not the number of variables) should be used as the layer neurons.
The Scientific World Journal
Performance Evaluation and Discussion
To evaluate the performance of the QTTI, GP, NN-FE, and NN-FT models developed in this paper in terms of their prediction ability in determining the optimal design combination of PDA form elements for matching a given S-C image, the six samples in the test set given in Table 2 are used. Table 5 shows the average S-C values of the six test samples assessed by 52 subjects, which are used as a comparison base for the performance evaluation. With the six test samples as the input, Table 5 shows the corresponding S-C values predicted by using the QTTI (i.e., Model (22)), GP (i.e., Model (25)), NN-FE, and NN-FT models, respectively. To evaluate the performance of a model, the root of mean square errors (RMSE) is commonly used, given as
Performance Comparison. The second row of
where X i is the ith output value predicted by the model, and X 0 is the expected values assessed by 52 subjects in the experiment. If there is no difference or error between the predicted value and the expected value, the RMSE is 0. The last column of Table 5 shows the RMSE of these four models in comparison with the assessed S-C values. Table 5 shows that the lowest RMSE is the QTTI model (0.2343), followed by the NN-FE model (0.2663) and the NN-FT model (0.2875), and the RMSE of the GP model is the highest. The result indicates that the QTTI model has the highest predictive consistency (an accuracy rate of 76.57% = 1 − 0.2343) for predicting the value of the S-C image. This is in line with the result of Multiple Comparisons by one-way Analysis of Variance (one-way ANOVA), as shown in Figure 2. Figure 2 shows the mean of error sum of squares (SSE) for these four models. The lower the SSE, the higher the prediction performance. The result of performance comparison suggests that QTTI is the model to be used for matching a given set of product form elements with a specific product image.
Nevertheless, this result is not consistent with the common notion that nonlinear quantitative models or systems are more suitable to simulate human beings' thinking and generally have a better performance for predicting consumers' psychological requirements or emotional feelings, in comparison with linear quantitative models [6, 10-12, 18, 29]. In addition, the NN model usually has a better performance and an effective technique to formulate the product design process for determining the optimal combination of product form elements with respect to a desirable product image [5].
Further Evaluation.
To further examine the prediction performance of the NN model, we conduct a set of analyses by using different learning rate and momentum factors for getting the better structure of the NN model. Another 3 pairs of learning rate and momentum factors are used for different conditions based on the complication of the research problem. For example, if the research issue is very simple, a large learning rate of 0.9 and momentum of 0.6 are recommended. On more complicated problems or predictive networks where output variables are continuous values rather than categories, use a smaller learning rate and momentum, such as 0.1 and 0.1 respectively. In addition, if the data are complex and very noisy, a learning rate of 0.05 and a momentum of 0.5 are used [30]. To distinguish between the NN-FE and NN-FT models using different input neurons and hidden neurons, both models are associated with the learning rate and momentum mentioned above, such as -P, -S, -C, -N, as shown in Table 6.
As described in Section 4.3, 24 training samples and six test samples are used, and the training process is not stopped until the cumulative training epochs are over 25,000. Figure 3 shows the RMSE of these eight NN models and the convergence diagrams in the training process. As shown in Figure 3, the convergence speed of NN-FT models are faster than NN-FE models. This is in line with the result of Section 4.3 that the more neurons in the input or hidden layer, the faster the convergence speed. In addition, we find the "-S" models (i.e., NN-FE-S model and NN-FT-S model, both using the large learning rate of 0.9 and momentum of 0.6 if the research issue is very simple) have larger movements as compared to other NN models, thus indicating that the essentials of consumers' perceptions are complicated, and often a block box and cannot be precisely described [10]. With the six test samples as input, Table 7 lists the predicted S-C image values and RMSE of these eight NN models for the further test set. Table 7 shows that the lowest RMSE is the NN-FE-N model (0.2203). In addition, the average RMSE value of NN-FE (0.3033) is slightly smaller than the value of NN-FT (0.3168). This is in line with the result of Section 4.3 that the number of layer neurons (the input or hidden neurons) has little influence on the performing effect of NN models. Table 7, except the NN-FE-N model (the RMSE being 0.2203), the other 7 NN models are larger than the QTTI (the RMSE being 0.2343) shown in Row 3 of Table 5. Further analysis shows that the QTTI model is a better approach for matching a given set of product form elements with a specific product image, regardless of what learning rate and momentum factors are chosen for constructing the NN model. This result implies that the linear modeling technique is good enough to help product designers determine the optimal form combination of product design for a particular design concept of product image. Consequently, in some product design settings, applying nonlinear modeling techniques may not necessarily produce a better outcome. In some settings, the QTTI model (the linear modeling technique) can be used to better explore the relationship between the consumers' perceptions and product form elements without compromising the prediction performance. According to the experimental analysis and results mentioned above, model (22) can help product designers understand consumers' perceptions of product form for a given product image. This model can also be used to examine the effect of the corresponding product image for a given combination of product form elements. Consequently, the QTTI model enables us to build a PDA design support database that can be generated by inputting each of all possible combinations (972, 3 × 3 × 4 × 3 × 3 × 3) of product form elements to the QTTI model individually for generating the associated image values. Product designers can specify a desirable image value for a new PDA form design, and the database can then work out the optimal combination of form elements. Table 8 shows the design support information for product designers to find out the optimal combination of product form elements in terms of a given product image. In addition, the design support database can be incorporated into a computer-aided design (CAD) system to facilitate the product form in the new PDA development process. To illustrate, we focus the attention more on the most influential elements, such as the "arrow-key style" form element (X 4 ) and the "color treatment" form element (X 5 ), for the desirable "simple" image of PDA. Figure 4 shows two new PDA form designs with the optimal combination of form elements for the desirable "simple" image.
Limitations and Further Suggestions.
In this paper, we use two linear modeling techniques (i.e., quantification theory type I and grey prediction) and one nonlinear modeling technique (i.e., neural networks) to determine the optimal form combination of product design for matching a given product image. In the further studies, other quantitative analysis models should be adopted to test the prediction performance, for example, fuzzy system, genetic algorithm, rough set, multiple regression analysis, and so on. Although PDAs are chosen as the experimental product, the consumeroriented design approach presented can be applied to other consumer products (e.g., smart phones, Tablet PC, etc.) with various design elements (e.g., color, texture, brand, etc.) and product images (e.g., classic or modern, artificial or artistic, etc.).
Conclusion
In this paper, we have conducted an experimental study on PDAs to demonstrate how a consumer-oriented design approach can be used to help determine the optimal form combination for matching a given product image. The consumer-oriented design based on the process of Kansei Engineering has used the QTTI model, the grey model, and the neural network model to predict the desirable simplecomplex image of consumers' perception. The result of the experimental study has shown that the QTTI model has the highest predictive consistency, thus suggesting that the QTTI model is a better methodological alternative for modeling the consumers' perception of a product characterized by a given set of product form elements. Noteworthily, this result has shown that the QTTI model (the linear modeling technique) is good enough to help product designers determine the optimal form combination of product design for a particular design concept of product image. Consequently, in some product design settings, we can use the linear modeling technique to explore the relationship between the consumers' perceptions and product form elements without compromising the prediction performance. Furthermore, the consumeroriented design approach has been built a PDA design support database, in conjunction with the computer-aided design (CAD) system, to help product designers facilitate the product form in the new PDA development process. | 2016-05-12T22:15:10.714Z | 2012-11-11T00:00:00.000 | {
"year": 2012,
"sha1": "1e3fa191e340b54ebf0cb569034dd01dfec25a90",
"oa_license": "CCBY",
"oa_url": "http://downloads.hindawi.com/journals/tswj/2012/689842.pdf",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "38d79936675ade991587394aa25f830b00486967",
"s2fieldsofstudy": [
"Business",
"Engineering"
],
"extfieldsofstudy": [
"Computer Science",
"Medicine"
]
} |
36892002 | pes2o/s2orc | v3-fos-license | Modulational instability of partially coherent signals in electrical transmission lines
We present an investigation of the modulational instability of partially coherent signals in electrical transmission lines. Starting from the modified Ginzburg-Landau equations and the Wigner-Moyal representation, we derive a nonlinear dispersion relation for the modulational instability. It is found that the effect of signal broadbandness reduces the growth rate of the modulational instability.
In this Brief Report, we consider the modulational instabilities of partially coherent electrical pulses that are governed by the MCGL equation [3] i∂ where P , Q j (j = 1, 2, 3), and γ are real transmission line parameters. We note that Eq. (1) has the space independent harmonic solution u where the asterisk denotes the complex conjugate. Separating the perturbation into its real and imaginary parts, according to u 1 = X + iY , and letting X, Y ∝ exp(iKx − iΩt), we obtain from (2) the nonlinear dispersion relation where K and Ω are the wavenumber and the frequency of low-frequency perturbations modulating the carrier signal. For Ω = iΓ − (Q 2 + 2Q 3 )u 2 0 K, we obtain the modulational instability growth rate from (3) when P Q 1 > 0. We see that the effect of the derivative nonlinearity Q 3 is to decrease the instability region, while the higher order nonlinearity coefficient Q 1 tends to increase the instability region. In Fig. 1, the typical structure of the modulational instability growth rate is depicted.
In order to analyze the effects due to partial coherence on the pulse propagation in nonlinear electrical transmission lines, we next introduce the Wigner function, defined as the Fourier transform of the two-point correlation function [5] ρ(t, x, k) = 1 2π dξ e ikξ u * (t, x + ξ/2)u(t, x − ξ/2) , where the angular bracket denotes the ensemble average [7]. The Wigner function defines a generalized phase space distribution function for quasi-particles, which satisfies the relation where I is the pulse intensity. Applying the time derivative on the definition (5) and using the MCGL equation (1), we obtain [5,6] where the sin and cos operators are defined in terms of their respective Taylor expansion. We note that the γ-term drops out, since it contains only the phase information for u. Equation We now analyze Eq. (7) for small perturbations, i.e. we let ρ(t, x, k) = ρ 0 (k)+ρ 1 exp(iKx− iΩt) and I(t, x) = I 0 + I 1 exp(iKx − iΩt), where |ρ 1 | ≪ ρ 0 and |I 1 | ≪ I 0 . Linearizing Eq.
If the background wave function u 0 has a partially coherent phase, the corresponding quasi-particle distribution is given by the Lorentzian [8] ρ 0 (k) = I 0 π where ∆ is the width of the distribution, giving the degree of decoherence of the pulse intensity. Using the distribution (9) in the general dispersion relation (8), we obtain We will assume that P Q 1 > 0 in order to make a comparison to the coherent case. With the normalization Ω = Q 1 I 2 0 Ω, K = (Q 1 /P ) 1/2 I 0 K, ∆ = (Q 1 /P ) 1/2 I 0 ∆, Q 2 = (P Q 1 ) 1/2 Q 2 , and Q 3 = Q 2 Q 3 , we obtain the dimensionless dispersion relation where we have dropped the tilde on all variables, and ǫ = sgn(P ). In Fig. 1 we have plotted the normalized growth rate Γ as a function of the normalized wavenumber K. We have assumed that P < 0, i.e. ǫ = −1. When Q 2 = Q 3 /4 = 1/2 and ∆ = 0, we obtain the full curve in Fig. 1, while ∆ = 0.1 gives the dashed curve. For ∆ = 0.1, but Q 2 = Q 3 /4 = −1/2, we obtain the dotted curve in Fig. 1, and Q 3 = Q 2 /4 = −1/2 gives the dashed-dotted curve.
When ǫ = 1, a reduced distribution width ∆ tend to increase the growth rate, which is unphysical.
To summarize, we have examined the modulational instability of partially coherent pulses in nonlinear electrical transmission lines. For this purpose, we have derived a nonlinear dispersion relation from the MCGL equation by using the Wigner-Moyal representation.
The nonlinear dispersion relation is analyzed for a Lorentzian equilibrium pulse distribution function. It is found that the growth rate of the modulational instability is reduced due to | 2017-09-23T22:48:33.492Z | 2006-04-05T00:00:00.000 | {
"year": 2006,
"sha1": "1076aa9b9a2e40e7a1d018a2e318d3686fdca282",
"oa_license": null,
"oa_url": "http://arxiv.org/pdf/physics/0604041",
"oa_status": "GREEN",
"pdf_src": "Arxiv",
"pdf_hash": "1076aa9b9a2e40e7a1d018a2e318d3686fdca282",
"s2fieldsofstudy": [
"Physics"
],
"extfieldsofstudy": [
"Physics",
"Medicine"
]
} |
233249023 | pes2o/s2orc | v3-fos-license | Acquired Hemophilia B Through Liver Transplantation: A Case Report and Literature Review
Hemophilia B, also called Christmas disease, is an X-linked inherited bleeding disorder that predominantly affects males, and it is caused by deficiency of coagulation factor IX. Factor IX is a coagulation factor produced in the liver. Liver transplantation from hemophilia B donors is neither widely documented nor performed due to presumed risk of developing increased bleeding tendency in the recipient due to potential acquired factor IX deficiency. In this review, we present a case of liver and kidney transplantation from a donor with mild hemophilia B to a recipient with no history of hemophilia B and literature review. This is a relatively rare situation with only two other prior case reports of acquired Hemophilia B through liver transplantation noted in 2015. The 2 cases presented in the literature lead to acquired hemophilia B in liver transplant recipients since donors were not screened due to mild disease and low suspicion level. In our case, the donor was known to have mild Hemophilia B and the recipient had a rapid decline – thus it is imperative to weigh risks and benefits of transplantation from this donor population. While those individuals with mild hemophilia B disease (and low bleeding burden) should be considered as an alternative donor for liver transplantation, appropriate counseling should be done with the recipient about possible bleeding risks and OLT should only occur after patient consent. A multidisciplinary approach should involve hematology in case factor replacement therapy is needed post transplantation and for long term follow up.
Introduction
Hemophilia B, also called Christmas disease, is an X-linked inherited bleeding disorder that predominantly affects males. Patients have factor IX deficiency due to a variation in the factor IX gene [1][2][3]. The prevalence of hemophilia B is estimated to be 1 in 25,000 to 30,000 males. Approximately half have severe disease (factor IX activity <1 percent of normal). Factor IX contributes to secondary hemostasis and formation of fibrin clot via intrinsic pathway leading to factor x activation. Clinical manifestations of hemophilia relate to bleeding from impaired hemostasis, sequelae from bleeding, or complications of coagulation factor infusion. Diagnostic criteria of Hemophilia B includes factor IX levels <40% of normal and presence of a mutated F9 gene. Patients will have a characteristically prolonged partial thromboplastin time (PTT) that corrects on mixing studies.
Factor IX is a coagulation factor produced in the liver [4][5][6][7][8]. Liver transplantation from hemophilia B donors is neither widely documented nor performed due to presumed risk of developing increased bleeding tendency in the recipient due to potential acquired factor IX deficiency. In this review, we present a case of liver and kidney transplantation from a donor with mild hemophilia B to a recipient with no history of hemophilia B and literature review.
Case
The liver recipient was a 72-year-old male with a past medical history of decompensated alcoholic cirrhosis A Case Report and Literature Review complicated by hepatocellular carcinoma, bladder carcinoma, prostate carcinoma, and immune thrombocytopenic purpura. His Model for End-Stage Liver Disease (MELD) score was 24 upon time of evaluation for transplant, and he was in acute renal failure, related to hepato-renal syndrome leading to hemodialysis. The patient was listed for liver-kidney transplantation with discussions to expand patient's pool of donors (US Centers for Disease Control and Prevention high-risk donors and hepatitis B core antibody-positive donors) due to rapidly progressive disease. Prior to transplantation, he did not have any history of recurring bleeding events (no Factor IX activity level prior to pre-transplant evaluation was available, however, PTT was normal at 34 seconds 5 years before presentation). His pre transplant labs are listed in table 1. Notably, factor IX activity was 38%, likely due to his history of end stage liver disease. The liver and kidney donor was a 38-year-old male patient who had congenital mild Factor IX deficiency. He did not have significant history of bleeding. He was not on any type of prophylactic factor replacement. He did not develop any factor IX inhibitor based on his records. No genotype testing was found in our records. He was admitted to the hospital for head trauma with massive intracranial bleed and became brain dead. The donor's factor IX level was 8% at the day of transplant.
We were consulted to discuss risk of bleeding and potential development of acquired Hemophilia B in the donor. Risks of transplantation with a donor with hemophilia B were thoroughly reviewed with the patient and it was understood that he may theoretically develop acquired hemophilia B requiring factor replacements therapy. As the patient's renal failure and liver failure were progressing, and weighing risks and benefits, he decided to accept the hemophilia B donor organs.
Orthotropic liver and kidney transplantation was performed on 3/6/2018. The hematology team closely followed him post-transplant. Following discussion, the decision was made to treat with standard half-life recombinant FIX product with the goal to maintain FIX troughs of 80-100% for the first 5 to 7 days, then troughs greater than 50% afterward for 10-14 days, with peak levels not to exceed 200%. His last factor IX replacement therapy was on 4/30/2018. He was negative for factor IX inhibitor when tested post-op (AB titer <0.6 on 5/2/2018). His factor IX levels and PTT post transplantation were as shown in Figure 1. His other factor activity levels were checked post-transplant, and also were normal, as seen in table 2. His postoperative course was complicated by many events. Immediately postoperatively, kidney graft function was delayed, requiring CVVHD. He had intermittent pneumatic compression for thromboprophylaxis for the first 2 days, then switched to heparin 5000 Units SQ Q 12 hours. On POD 6, he was found to have a femoral DVT and was placed on a heparin drip while receiving factor IX replacement. Our new target for factor IX trough level was to increase it to 70% to minimize bleeding events while on full dose anticoagulation, with peak not to exceed 200%. At the time of diagnosis of his clot, his factor IX level was 52%. Hemoglobin dropped on POD 11. Investigation revealed a peri-nephric hematoma, 7 cm in size. His anti-Xa level at the time was 0.23, and Factor IX level was 57%. Anticoagulation was stopped and an IVC filter was placed on POD 12. Evacuation of the hematoma was performed on POD 13. Embolization of the left L2, L3, L4 lumbar and left iliolumbar arteries was performed by interventional radiology on POD 16 and bleeding was controlled. Unfortunately, the patient developed Pseudomona pneumonia and developed hypoxic respiratory failure. His condition continued to deteriorate and his family made the decision to withdraw care on POD 55, and replacement factor stopped. His factor IX levels eventually dropped to around 20-30% in the following days before he passed away on POD 62.
Discussion
This is a relatively rare situation with only two other prior case reports of acquired Hemophilia B through liver transplantation noted in 2015 [9,10]. The first report in literature was in 2015 by Bergstrom and colleagues who described an OLT in a 19 months old boy due to biliary atresia. He developed mild bleeding post OLT from vein puncture sites as well as bruising that led to work up showing mild hemophilia B with factor IX ranging from 5%-10. His factor IX activity level was 8% with no inhibitor at time of evaluation. F9 gene sequencing on patient's leukocytes did not reveal any abnormality. However, gene sequencing of transplanted liver tissue revealed a F9 mutation. The donor was mildly symptomatic and not known to have a bleeding disorder as per his family members [9]. This was followed by another report in 2015 by Brunetta et al., who detailed a case of a new onset of FIX deficiency, in association with mild coagulation abnormalities following liver transplantation, attributable to an unrecognized hemophilia B in the donor. Active bleeding was noticed after a dental extraction performed 7 months after transplant. PTT was prolonged and factor IX activity was 11%. In speaking with the donor's family, two thigh hematomas were noted once after a sports-related injury, but patient was never officially diagnosed with a bleeding disorder [10].
The hepatocyte is the primary cell type responsible for synthesizing factor IX, as shown in murine models [11]. Their studies showed that liver sinusoidal endothelial cells and extrahepatic cells (including the lung, spleen, kidney, brain, intestine, and tongue) did not express factor IX mRNA. In comparison, factor VIII is produced extrahepatically, and transplant of a liver from a donor carrying the diagnosis of hemophilia A did not lead to development of the bleeding disorder in the recipient [12][13][14][15].
There have been few reports of hemophilia B donors in liver transplantation. Patients with mild hemophilia frequently have a benign course and bleeding occurs only in response to surgery, tooth extractions, or major injuries. The diagnosis can be missed frequently if suspicion level is low. The 2 cases presented in the literature lead to acquired hemophilia B in recipients since donors were not screened due to mild disease and low suspicion level.
In our case, the donor was known to have mild Hemophilia B and the recipient had a rapid decline -thus it is imperative to weigh risks and benefits of transplantation from this donor population.
Conclusion
While those individuals with mild hemophilia B disease (and low bleeding burden) should be considered as an alternative donor for liver transplantation, appropriate counseling should be done with the recipient about possible bleeding risks and OLT should only occur after patient consent. A multidisciplinary approach should be done and it should involve hematology in case factor replacement therapy is needed post transplantation and for long term follow up.
Liver transplantation from hemophilia B donors is not
commonly performed due to the perceived risk of developing hemophilia B in the recipient. 2. We present a case of liver and kidney transplantation from a donor with mild hemophilia B to a recipient with no history of hemophilia B and literature review with development of mild hemophilia B in the recipient. 3. The hepatocyte is the primary cell type responsible for synthesizing factor IX, as shown in murine models. Their studies showed that liver sinusoidal endothelial cells and extrahepatic cells (including the lung, spleen, kidney, brain, intestine, and tongue) did not express factor IX mRNA. 4. While patients with mild hemophilia B disease should be considered as an alternative donor for liver transplantation, appropriate counseling should be done | 2021-04-16T09:07:25.117Z | 2019-06-26T00:00:00.000 | {
"year": 2019,
"sha1": "7b3d0eef8a7da6da47caae46618989a08ebe0da7",
"oa_license": "CCBY",
"oa_url": "http://article.sciencepublishinggroup.com/pdf/10.11648.j.ajim.20200806.21.pdf",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "7b3d0eef8a7da6da47caae46618989a08ebe0da7",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
265624264 | pes2o/s2orc | v3-fos-license | The impacts of nurses’ attitudes towards dementia, critical reflection competency, and nursing work environment on person-centered nursing in acute care hospitals: A descriptive study
Purpose: The purpose of the study was to examine the attitudes of acute care nurses towards dementia, their critical reflection competency, and the level of nursing work environment and to identify the factors relating to person-centered nursing. Methods: The study was conducted for 149 nurses with at least 1 year of experience working in acute care hospitals and providing nursing care to hospitalized elderly patients. The measurements used were the Dementia Attitude Scale Korean version, the Critical Reflection Competency Scale, the Nursing Work Environment Scale, and the Person-Centered Nursing Assessment Tool. The collected data were analyzed with frequency analysis, descriptive statistics, independent t-test, ANOVA test, Pearson’s correlation, and multiple regression using IBM SPSS 29.0 software. Results: The factors influencing person-centered nursing performance were critical reflection competency (β=.49, p <.001), attitudes towards dementia (β=.29, p <.001), and clinical career (β=-.26, p =.024), having the explanation power of 59.6%. Conclusion: The results of this study indicate that critical reflection competency, attitudes towards dementia, and clinical career have an impact on person-centered nursing performance. Specifically, higher levels of critical reflection competency and positive attitudes towards dementia were predictors with higher levels of person-centered nursing performance. Therefore, to enhance person-centered nursing performance among acute care nurses, it is important to prioritize critical reflection training programs for experienced nurses to enhance their critical reflection competency.
Significance of the Study
In Korea, the proportion of adults aged 65 years and older of the total population is 17.5% in 2022 [1].Furthermore, the number of older adults with dementia and the overall older adult population are increasing rapidly.The number of older adults with dementia living in communities receiving hospital care for acute illnesses is increasing, along with the number of nurses caring for older adults with dementia in acute care hospitals [2].Notably, the ratio of hospitalization days to outpatient visits for older adults with dementia in tertiary and general hospitals has increased annually from 27% in 2008 to 55% in 2017 [3].
The unfamiliar hospital environment during acute care stays can lead to behavioral issues in older adults with dementia.A retrospective cohort study found that 45.8% of patients with dementia in 10,014 hospitals developed delirium during hospitalization.Furthermore, older adults with dementia experience higher mortality, higher incidence of falls, decreased function, and an increased risk of malnutrition, pain, fear, and overstimulation compared to the general population [4].Person-centered nursing for older adults with dementia focuses on providing individualized services, understanding psychological needs, creating a physical environment that preserves personal "respect" and "dignity," and establishing an autonomous decision-making structure for older adults with dementia [5].This approach can reduce the behavioral and psychological symptoms of dementia and help maintain and improve cognitive function, making it a critical factor in caring for older adults with dementia [6].However, acute care hospitals primarily concentrate on treating the disease [2].Accessing person-centered nursing with standardized protocols within the current healthcare system for older adults with severe, complex, chronic diseases and a wide range in physical, cognitive, and functional abilities among individuals is either impossible or severely limiting [7].As the demand for specialized nursing care in the aging era increases, the provision of quality care through a comprehensive understanding of the person, in addition to the patient's disease, becomes fundamental to the modern nursing environment [8].Further, there is a growing demand for nurses to provide care to older adult patients with dementia as a medical condition.Therefore, it is essential to analyze the factors that impact nurses' ability to deliver person-centered nursing care.This analysis encompasses examining attitudes toward dementia as an individual factor and assessing critical reflection competency and the nursing work environment as organizational environmental factors.
At the individual level, positive attitudes toward dementia can aid in focusing on care for individuals with dementia and delivering person-centered nursing [2].Moreover, attitudes toward dementia were positively correlated with person-centered nursing [9] -suggesting that positive attitudes toward dementia are essential for improving quality of care.
Critical thinking disposition, the ability to gather information from a situation and draw necessary conclusions through reasoning, has been positively correlated with person-centered nursing [10].Nurses who effectively apply critical reflection develop strategies that promote an individualized, holistic approach to patient care and constantly strive to improve their professional competence [11].Individuals with high critical reflection competencies actively learn to practice professional nursing.They are rewarded with positive feedback to adapt their practice to the needs of individual patients, which can be an essential factor in providing person-centered nursing care.
At the organizational level, the nursing work environment, which includes sufficient time, organizational support, and a collaborative relationship between the client and the healthcare team, is critical for providing appropriate care [8].Positive perceptions of the nursing work environment are associated with higher levels of job satisfaction and lower rates of burnout, leading to higher-quality nursing care [12].In acute care hospitals, where patients with various conditions, frequent emergencies, and high severity are cared for, the work environment, including adequate staffing, resources, and support for individual nurses, may also significantly affect the implementation of person-centered nursing.
Studies have analyzed the factors affecting the performance of person-centered nursing.Nonetheless, these studies are insufficient, and previous studies have reported conflicting results regarding the effects of the nursing work environment on person-centered nursing [13,14].Additionally, since acute care is provided in general hospitals focusing on "disease," there is a lack of research on nurses caring for older people with dementia as an underlying disease.
Therefore, person-centered nursing is an important concept that should be considered in acute care hospitals where the number of older adults with dementia increases with age.Furthermore, the present study analyzes the obstacles to person-centered nursing and proposes improvements.
Research Objectives
This study aimed to identify nurses' attitudes toward dementia, critical reflection competency, and nursing work environment.Furthermore, it sought to identify factors affecting person-centered nursing performance.The specific objectives were as follows: 1) Identify the general characteristics of participants.
2) Determine participants' attitudes toward dementia, critical reflection competency, the nursing work environment, and person-centered nursing performance.3) Identify differences in person-centered nursing performance based on participants' characteristics.
Study Design
This descriptive study used a survey to assess the attitudes of acute care hospital nurses toward dementia, their critical reflection competency, and their perceptions of the nursing work environment.Additionally, it aimed to identify factors that impact person-centered nursing performance.
Study Participants
The participants were nurses in acute care hospitals with at least 1 year of experience providing nursing care to inpatients aged 65 and older.The exclusion criterion was nurses working in departments that did not provide direct care to older patients, such as pediatrics or operating rooms.
The number of participants required for the study was calculated using the G power 3.1 and 9.7 programs.Moreover, the multiple regression analysis yielded a moderate effect size of .169based on the coefficient of determination R² = .14based on previous studies [15].Given a significance level of .05, a power of .80,and 18 independent variables (comprising 15 general characteristics and three main independent variables), a minimum sample size of 135 was required.Therefore, this study recruited 149 participants, considering a 10% dropout rate based on previous studies [14].The data from all 149 participants were used in the final analysis.
1) General Characteristics of Participants
The general characteristics included sex, age, marital status, highest level of education, current department, total clinical experience, type of care provided, average number of patients per shift, dementia-related education, presence of a family member with dementia, and living with a family member with dementia.
2) Attitudes Toward Dementia
This study used the Dementia Attitudes Scale Korean version (DAS-K) developed by O'Connor and McFadden [16] and later translated by Chang et al. [17].The DAS-K comprises 20 questions, divided into a 10-item dementia knowledge scale to assess cognitive attitudes toward dementia and a 10-item social comfort scale to assess emotional and behavioral responses.These questions are further grouped into four subscales: understanding the person with dementia, social discomfort, social comfort, and willingness to provide person-centered nursing.Each item is scored on a seven-point Likert scale, with 1 being "not at all" and 7 being "very much," with six reverse-scored items.The total score ranges from 20 to 140, with higher scores indicating more favorable attitudes toward dementia and older adults with dementia.The Cronbach's α was .83 at the time of tool development [16], .85 for 20 items with .78~.84 for four sub-factors in Chang et al. 's study [17], and .90 in this study.
3) Critical Reflection Competency
Critical reflection competency was measured using the Korean version of the Critical Reflection Competence Measurement Tool developed and translated by Shin et al. [11].This instrument comprises 19 items, each scored on a five-point Likert scale ranging from 1 (not at all) to 5 (strongly agree).The total score ranges from 19 to 95, with higher scores indicating higher critical reflection competence in the clinical nursing environment-Cronbach's α was .85 at the development [11] and .91 for this study.
4) Nursing Work Environment
The Korean version of the Practice Environment Scale of the Nursing Work Index was developed by Lake [18] and validated by Cho et al. [19].The instrument comprises 29 questions, including four questions on "sufficient workforce and material support," nine questions on "foundation for quality nursing care," nine questions on "nurses' participation in hospital management," three questions on "nurses' cooperation with doctors," and four questions on "nursing managers' competence, leadership, and support for nurses."Each item was scored on a four-point Likert scale, with 1 being "not at all true" and 4 being "very true."The score is the average of all items and ranges from 1 to 4. The higher the score, the better the nurses' perceived work environments.Cronbach's α was .82when Lake [18] developed the instrument and .93 with .80~.84 for each subscale in Cho et al.'s [19] study.In this study, Cronbach's α was .94.
5) Person-Centered Nursing Performance
Person-centered nursing performance was measured using the Lee's [20] Person-centered Nursing Assessment Tool.The tool comprises 25 questions in five subdomains: four questions on empowerment, four questions on wholeness, five questions on individualization, five questions on respect, and seven questions on relationships.The tool is a five-point Likert scale, with 1 being "not at all" and 5 being "very much."The score is the average of the total scores and ranges from 1 to 5. A higher score indicated a higher level of person-centered nursing performance.The reliability of this tool was Cronbach's α at .94 with .80~.89 for each factor in Lee's [20] study.In this study, Cronbach's α was .94.
Data Collection
Data collection for this study was conducted anonymously through an online survey from March 11, 2023, to April 5, 2023, after receiving approval from the Institutional Review Board (IRB) of Samsung Medical Center.After explaining the study's purpose and data collection methods thoroughly to the nursing department of the affiliated acute care hospital, the researcher received approval for data collection.Subsequently, a notice was distributed to all hospital departments, and those interested in participating accessed the study using QR codes.Also, Participants were recruited through announcements on the "My Duty" bulletin board, an online application used to manage nurses' work, and Naver cafés dedicated to preparing for graduate school admission.A link to the Google Online Questionnaire was shared, and interested individuals were invited to complete a survey.Participation was contingent upon their voluntary agreement after reading the study description, which included information about the study's purpose, procedures, potential risks and benefits, compensation for losses resulting from participation in the study, voluntary participation, withdrawal of consent, and protection of personal information.
Data Analysis
The data collected in this study were analyzed using SPSS version 29.0 (IBM Corp.).The specific methods of analysis are as follows: 1) The participants' general characteristics were analyzed using frequencies, percentages, means, and standard deviations.2) Participants' attitudes toward dementia, critical reflection competency, the nursing work environment, and person-centered nursing performance were analyzed using means and standard deviations.3) Differences in person-centered nursing performance according to general characteristics were analyzed using an independent t-test or ANOVA.Post-hoc tests were performed using Scheffé's tests.4) Correlations between attitudes toward dementia, critical reflection competency, and person-centered nursing performance according to the nursing work environment were analyzed using Pearson's correlation.5) Factors affecting nurses' person-centered nursing performance in acute care hospitals were analyzed using multiple regression analysis.
Ethical Considerations
Before conducting this study, approval was obtained from the IRB of Samsung Medical Center in Seoul, Korea (IRB No. 2023-02-060, 2023-02-060-001).Permission was obtained from the hospital's nursing department before data collection.Participants were provided with sufficient information regarding the purpose and content of the study.Moreover, the questionnaire was administered if the participants voluntarily agreed to participate in the study.Nurses who participated in the study were given a coupon worth 5,000 won as an appreciation token.The survey content was anonymized and used only for research purposes.Personal contacts collected for rewards were permanently deleted immediately after bonuses were provided.All personal information is kept secure, and research-related materials are stored in a separate place with a lock for a mandatory storage period of 3 years.They will be permanently deleted immediately after the storage period.
Participants' Attitudes Toward Dementia, Critical Reflection Competency, Nursing Work Environment, and Person-Centered Nursing Performance
Participants' attitudes toward dementia, critical reflection competency, nursing work environment, and person-centered nursing performance are shown in
Differences in Person-Centered Nursing Performance by General Characteristics
Differences in person-centered nursing performance according to general characteristics are shown in Table 3.There were significant differences in person-centered nursing performance according to clinical experience (F = 3.61, p = .015),continuing education (t = 4.40, p < .001),and dementia specialty training (t = 5.91, p < .001).Scheffe's post-hoc test for clinical experience showed that those with less than 1~3 years of experience had a higher level of person-centered nursing performance than those with more than 10 years of experience.Additionally, those who completed dementia-related continuing education and specialized dementia education courses had a higher level of person-centered nursing performance than those who had not.
Correlations Between Attitudes Toward Dementia, Critical Reflection Competency and Nursing Work Environment, and Person-Centered Nursing Performance
Correlations between attitudes toward dementia, critical reflection competency, nursing work environment, and person-centered nursing are shown in Table 4. Person-centered nursing performance was significantly and positively correlated with attitudes toward dementia (r = .669,p < .001),critical reflection competency (r = .704,p < .001),and nursing work environment (r = .443,p < .001).Specifically, the more positive the attitude toward dementia, the higher the critical reflection competency.And the better the nurses' perceived nursing work environment, the higher their level of person-centered nursing.
Factors Affecting the Person-Centered Nursing Performance
Factors affecting person-centered nursing performance are shown in Table 5.We analyzed the factors affecting per- values were 1.44~4.82,which were not greater than 10, indicating no risk of multicollinearity.Regression analysis showed that among the variables, critical reflection competency (β = .49,p < .001),attitude toward dementia (β = .29,p < .001),and clinical experience (β = -.26,p = .024)were influential factors.The explanatory power was 59.6% (F = 32.18,p < .001).Specifically, the higher the clinical experience, the lower the person-centered nursing performance.Additionally, the higher the critical reflection competency or the more positive the attitude toward dementia, the higher the person-centered nursing performance.
DISCUSSION
By analyzing the factors influencing person-centered nursing in acute care hospitals, this study provides a rationale for developing strategies to improve person-centered nursing.The score of person-centered nursing in this study was 3.94 ± 0.48 out of 5.This is relatively low among the empowerment and wholeness subdomains.This study focused on shift nurses providing direct care, consistent with a study that found that shift nurses providing direct care had lower levels of perceived structural empowerment than permanent or outpatient nurses [21].Previous studies have shown that an essential factor in providing person-centered nursing is environmental support that can provide immediate help from professional personnel at the right time [22].Thus, general hospitals still experience a vicious cycle of staff shortages and work environments [23].
The factors affecting person-centered nursing performance in this study were critical reflection competency, attitude toward dementia, and clinical career.Among these factors, critical reflection competency was the most crucial in improving person-centered nursing performance, confirming the importance of strengthening critical reflection competency for nurses.It is necessary to provide nurses with opportunities to reflect on their nursing practice and discover the meaning of nursing through a critical reflection training program that comprises the process of understanding the concept of critical reflection, strengthening its strategies, and applying them to enhance critical reflection competency [24].Moreover, the exchange of opinions and reflections based on clinical cases should be encouraged in clinical settings.Additionally, as there is a lack of research on the relationship between critical reflection competency and person-centered nursing performance in clinical nurses, studies that apply critical reflection training programs in practice are needed.The second factor affecting person-centered nursing performance was nurses' attitudes toward dementia.This is similar to previous studies [9] that found that nurses' attitudes toward dementia were significantly associated with person-centered nursing.Nurses' attitudes toward dementia are considerably better when they receive dementia-related education [2].Notably, nurses in nursing homes with a high proportion of dementia patients have positive attitudes toward older adults with dementia [25].Thus, dementia-related education is important for fostering positive attitudes toward dementia among nurses in acute care hospitals.Among the general characteristics in this study, clinical career negatively affected person-centered nursing performance, similar to the results of a previous study [26] that showed that the lower the total clinical career, the higher the degree of person-centered nursing.Nurses with more experience experienced higher job stress [27].A previous study of nursing home nurses [28] showed that job stress was negatively related to person-centered care.Therefore, applying the Stress Management and Resiliency Training (SMART) program to experienced nurses is expected to reduce stress, improve resilience [29], and increase person-centered nursing performance.
Conversely, in this study, the nursing work environment did not significantly influence person-centered nursing, which is consistent with the findings of Kang and Seo [13].Although there was a positive correlation between the two variables, nurse burnout and psychological stress caused by a negative work environment were expected to be more critical variables in person-centered nursing than the nursing work environment itself.Previous studies have shown that nurses' burnout increases when their work environment is poor [8] and that burnout significantly affects person-centered nursing [14].Therefore, future studies are needed to compare the perceived nursing work environment according to hospital size, department, and average number of patients per capita.This includes considering nurses' burnout and psychological stress and verifying their relationship with person-centered nursing performance.
As the population ages, older adults with dementia are often admitted to acute care hospitals for treatment.Hence, acute care nurses should provide person-centered nursing to older adults with dementia.Acute care hospital nurses need to perform person-centered nursing without experiencing psychological burdens when caring for older patients with dementia.Thus, acute care hospitals must offer critical reflection training programs and dementia-related educational opportunities to experienced nurses with relatively low levels of person-centered nursing performance.This should be extended to nurses throughout the hospital to improve individual competencies and lay the foundation for providing person-centered nursing for older adults with dementia in acute care hospitals.
CONCLUSION
This study aimed to investigate the impact of attitudes toward dementia, critical reflection competency, and the nursing work environment on person-centered nursing performance among nurses in acute care hospitals.The goal was to improve the quality of care for older adults, given the increasing hospitalization of older adults with dementia due to an aging society.The findings identified critical reflection competency, attitudes toward dementia, and clinical experience as factors affecting person-centered nursing.Therefore, to improve person-centered nursing performance, it is necessary to implement critical reflection training programs to strengthen nurses' critical reflection competency and expand opportunities for nurses to regularly complete dementia-related educational programs to understand and perceive older adults with dementia positively.Furthermore, clinical nursing guidelines for caring for older adults with dementia are required to reduce nurses' psychological burden and lay the foundation for improving the performance of person-centered nursing.
The significance of this study is that it provides a basis for developing strategies to improve person-centered nursing among nurses in acute care hospitals.This study collected comprehensive data from nurses with experience in caring for older adult patients to analyze the general factors affecting the performance of person-centered nursing for older adults with dementia.However, the person-centered nursing tool used in this study was not developed specifically to care for older adults.Thus, the questions themselves had limitations in reflecting the uniqueness of older patients.Therefore, based on the main findings and limitations, we make the following suggestions: First, it was necessary to verify the effectiveness of the proposed critical reflection training.In particular, this should be applied to nurses' careers and analysis of their person-centered nursing performance before and after training.Second, it is necessary to develop a person-centered nursing scale that considers older adults' characteristics.Third, we recommend a follow-up study involving nurses with expertise in caring for older adults with dementia.and SS; Drafting and critical revision of the manuscript -MK and SS
Table 2 .
The score of attitude toward dementia was a mean of 95.99 ± 17.32 out of 140 points.The scores of subdomains were 33.05 ± 5.55 for an understanding of individuals with dementia, 26.95 ± 6.64 for social discomfort, 14.74 ± 5.40 for social comfort, and 21.24 ± 4.27 for willingness to provide person-centered nursing.Critical reflection competency averaged 77.95 ± 7.83 out of 95.The nursing work environment was 2.86 ± 0.44 out of
Table 2 .
The Degree of Attitudes Toward Dementia, Critical Reflection Competency, Nursing Work Environment, Person-Centered Nursing (N=149) https://doi.org/10.17079/jkgn.2023.00101ized residual was -2.76, and the maximum value was 2.54.The standardized deviation was 0.98, and normality was confirmed using P-P diagrams and histograms.The tolerance limits were 0.21~0.70,more than 0.1, and the Variance Inflation Factor
Table 3 .
Differences in the Performance of Person-Centered Nursing According to Characteristics of Participants (N=149) | 2023-12-05T16:32:37.160Z | 2023-11-30T00:00:00.000 | {
"year": 2023,
"sha1": "08cb278b924210bd4ddc6b008eef606221fa2e22",
"oa_license": "CCBYNC",
"oa_url": "https://www.jkgn.org/upload/pdf/jkgn-2023-00101.pdf",
"oa_status": "GOLD",
"pdf_src": "Anansi",
"pdf_hash": "52c6070e824770494effc38c9943f9b82e6e1b1b",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": []
} |
256937603 | pes2o/s2orc | v3-fos-license | Stationary Trend in Elevated Serum Alpha-Fetoprotein Level in Hepatocellular Carcinoma Patients
Simple Summary In this study, we demonstrated that overall 51.2% of patients with hepatocellular carcinoma (HCC) had elevated alpha-fetoprotein (AFP) levels. The proportion of patients with elevated AFP levels was stationary in the period from 2011 to 2020. The proportion of patients with Barcelona Clinic Liver Cancer classification (BCLC) stages 0–A HCC decreased from 2011 to 2020, whereas the proportion of patients with non-HBV- and non-HCV (NBNC)-HCC increased in the same period. Furthermore, the proportion of patients with early-stage HCC (i.e., BCLC stages 0–A) was lower for NBNC-HCC than for HBV- or HCV-related HCC. Advanced tumor stage, severe underlying liver disease, viral etiology, and female gender are associated with elevated AFP levels in HCC patients. Abstract A recent study from the US showed a decreasing trend in the elevated serum alpha-fetoprotein (AFP) level (i.e., ≥20 ng/mL) in hepatocellular carcinoma (HCC) patients at the time of diagnosis. Furthermore, advanced tumor stage and severe underlying liver disease were associated with elevated AFP levels. We aimed to evaluate this issue in an area endemic for hepatitis B virus (HBV). Between 2011 and 2020, 4031 patients were newly diagnosed with HCC at our institution. After excluding 54 patients with unknown AFP data, the remaining 3977 patients were enrolled in this study. Elevated AFP level was defined as ≥20 ng/mL. Overall, 51.2% of HCC patients had elevated AFP levels; this proportion remained stationary between 2011 and 2020 (51.8% vs. 51.1%). Multivariate analysis showed that female gender (odds ratio (OR) = 1.462; p < 0.001), tumor size per 10 mm increase (OR = 1.155; p < 0.001), multiple tumors (OR = 1.406; p < 0.001), Barcelona Clinic Liver Cancer stages B–D (OR = 1.247; p = 0.019), cirrhosis (OR = 1.288; p = 0.02), total bilirubin > 1.4 mg/dL (OR = 1.218; p = 0.030), and HBV- or hepatitis C virus (HCV)-positive status (OR = 1.720; p < 0.001) were associated with elevated AFP levels. In conclusion, a stationary trend in elevated serum AFP level in HCC patients has been noted in the past 10 years. Advanced tumor stage, severe underlying liver disease, viral etiology, and female gender are associated with elevated AFP levels in HCC patients.
Introduction
Hepatocellular carcinoma (HCC) is one of the leading causes of cancer-related death worldwide [1]. A meta-analysis showed that HCC surveillance is associated with significant improvements in early-stage tumor detection, the receipt of curative therapy, and survival of cirrhotic patients [2]. The American Association for the Study of Liver Diseases (AASLD) guideline recommends HCC surveillance for high-risk populations. The modality recommended for surveillance is ultrasound with or without an alpha-fetoprotein (AFP) serum assay [3]. Ultrasound with an AFP serum assay is recommended for surveillance because a meta-analysis demonstrated that ultrasound alone had low sensitivity in detecting earlystage tumor in cirrhotic patients. The combination of AFP serum assay and ultrasound significantly increases the sensitivity of tumor detection [4]. Currently, the AASLD guideline recommends diagnostic multiphasic magnetic resonance imaging (MRI)/computed tomography (CT) for further evaluation when the AFP level is ≥20 ng/mL on surveillance [3].
Multiple factors, including advanced tumor stage and viral etiology of chronic liver disease, are associated with elevated AFP levels in HCC patients [5]. A recent study from the US found a downtrend in the percentage of HCC cases with elevated AFP levels at the time of diagnosis from 2010 to 2017 in a large cohort from the National Cancer Database. Elevated AFP was defined as ≥20 ng/mL. Furthermore, advanced tumor stage and severe underlying liver disease were associated with elevated AFP levels. The authors suggested that these changes in AFP values at HCC diagnosis were possibly related to the increasing trend in early-stage tumor detection and the shift from viral (i.e., hepatitis B virus (HBV) or hepatitis C virus (HCV)) to nonviral etiology. However, data on the etiology of liver disease were unavailable in the database analyzed in the study [6].
Approximately 90% of HCC cases are associated with a known underlying etiology [7]. In East Asia, the major risk factor is HBV, whereas, in the Western world, it is HCV [7]. The risk of HCC attributed to HCV infection has largely decreased owing to the eradication of the virus with direct-acting antiviral (DAA) agents [8]. Nonalcoholic fatty liver disease (NAFLD), which is usually associated with obesity, metabolic syndrome, or diabetes mellitus, is becoming the fastest growing etiology of HCC, not only in Western countries [9], but also in Asia [10].
Due to the different etiologies of HCC in the East and the West and viral etiology being associated with elevated AFP levels in HCC patients [5], we aimed to evaluate whether there is a downtrend in the percentages of HCC cases with elevated AFP levels at the time of diagnosis and the factors associated with elevated AFP levels in HCC patients in a country from East Asia, where the leading etiology of HCC is HBV.
Materials and Methods
The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board of Kaohsiung Chang Gung Memorial Hospital (reference number: 202201189B0; date of approval: 8 August 2022).
The Institutional Review Board of Kaohsiung Chang Gung Memorial Hospital waived the need for informed consent due to the retrospective and observational nature of the study design. Data were extracted from Kaohsiung Chang Gung Memorial Hospital's HCC registry database, which holds prospectively collected and annually updated data.
From 2011 to 2020, 4031 patients were newly diagnosed with HCC at the institution. After excluding 54 patients with unknown AFP data, the remaining 3977 patients were enrolled in this study.
Variables of Interest
Patient demographics, tumor size and number, clinical tumor-node-metastasis (TNM) stage (seventh edition of the American Joint Committee on Cancer (AJCC)) [11], Barcelona Clinic Liver Cancer classification (BCLC) stage [12], AFP level, cirrhosis, Child-Pugh class [13], creatinine, bilirubin, international normalized ratio (INR), hepatitis B surface antigen (HBsAg), anti-HCV antibody, alcohol use disorder (AUD), and HCC diagnostic method (i.e., clinical vs. pathological diagnosis) were prospectively collected from the HCC registry data. Infection with HBV was defined as being HBsAg-positive. Infection with HCV was defined as being anti-HCV-antibody-positive, irrespective of viremia. An individual with AUD was defined as a habitual drinker. Demographic information in-cluded age, gender, height, and weight. Tumor size was determined according to the results of pathological examination of patients who underwent surgery, whereas it was determined according to the findings of imaging in patients who underwent nonsurgical treatments. Tumor number (solitary vs. multiple) was determined from the findings of imaging. The presence of cirrhosis was indicated by an Ishak score [14] of 5 or 6 in patients who underwent surgery, whereas it was determined according to the findings of imaging in patients who underwent nonsurgical treatments. Cirrhosis was indicated in imaging by small liver size, nodular liver surface, presence of regeneration nodules, left and right lobe liver volume redistribution, etc. [15]. The BCLC stages according to the original version and BCLC stage A were defined within Milan criteria [16].
Statistical Analysis
Variables are presented as number and percentage or median and interquartile range. The Chi-square test was used to compare categorical variables. Mann-Whitney U test was used to compare continuous variables. Whether there was an increasing or decreasing trend of BCLC stages 0-A or non-HBV-and non-HCV (NBNC)-HCC according to the year of HCC diagnosis was examined for a linear trend using the Chi-square test. Univariate analyses were conducted to explore the association between elevated AFP levels and clinical variables. Variables with p-values ≤ 0.1 in univariate analyses were included in a multivariate logistic regression analysis. To avoid collinearity, we examined the correlation between two independent variables using Spearman's correlation test. If two independent variables had a correlation coefficient above 0.5, then they were determined to be highly correlated with each other and thus, collinear. In this case, we only chose one of the variables for multivariate analysis. In this analysis, we used the following cutoff values for continuous variables: for bilirubin, the upper limit of the normal range (i.e., 1.4 mg/dL); for creatinine, the upper limit of the normal range (i.e., 1.2 mg/dL); and for INR, the upper limit of the normal range (i.e., 1.2). Relative risks are presented as odds ratio (OR) with a 95% confidence interval (CI). To compare with a recent US study [6], we used the same method adopted in that study [6] to interpret the temporal trend of elevated AFP levels. We estimated the percentage of elevated AFP using marginal effects (i.e., the average predicted probability) from a logistic regression model [17]. All statistical analyses were performed using SPSS version 22.0 and R statistical software version 4.0.5. Two-tailed significance values were applied, and the level of statistical significance was defined as p < 0.05.
Trends in Early-Stage Tumor Prevalence and Non-Viral Etiology at the Time of HCC Diagnosis
The proportion of patients with early-stage tumor (i.e., BCLC stages 0-A) decreased in the period from 2011 to 2020 (54.2% vs. 42.5%, p < 0.001) ( Figure 2). The proportion of patients with non-HBV and non-HCV (NBNC)-HCC increased between the years 2011 and 2020 (from 20.4% to 28.7%, p < 0.001) ( Figure 3).
Trends in Early-Stage Tumor Prevalence and Non-Viral Etiology at the Time of HCC Diagnosis
The proportion of patients with early-stage tumor (i.e., BCLC stages 0-A) decreased in the period from 2011 to 2020 (54.2% vs. 42.5%, p < 0.001) ( Figure 2). The proportion of patients with non-HBV and non-HCV (NBNC)-HCC increased between the years 2011 and 2020 (from 20.4% to 28.7%, p < 0.001) (Figure 3).
Trends in Early-Stage Tumor Prevalence and Non-Viral Etiology at the Time of HCC Diagnosis
The proportion of patients with early-stage tumor (i.e., BCLC stages 0-A) decreased in the period from 2011 to 2020 (54.2% vs. 42.5%, p < 0.001) ( Figure 2). The proportion of patients with non-HBV and non-HCV (NBNC)-HCC increased between the years 2011 and 2020 (from 20.4% to 28.7%, p < 0.001) (Figure 3).
Patients' Characteristics Categorized by AFP Level
Compared to patients with normal AFP levels, a smaller proportion of patients with elevated AFP were male (p = 0.02), had a pathological diagnosis of HCC (p < 0.001), were in AJCC stage 1 or 2 HCC (p < 0.001), had a solitary tumor (p < 0.001), were in BCLC stage 0 or A HCC (p < 0.001), were in Child-Pugh class A (p < 0.001), and had a low body mass index (BMI) (p < 0.001). Furthermore, patients with elevated AFP levels had larger tumors (p < 0.001), a higher total bilirubin level (p < 0.001), and a higher INR (p < 0.001) and a higher proportion of them were cirrhotic (p = 0.001) and HBsAg-positive (p = 0.002). However, there were no significant differences in age, creatinine level, proportion with AUD, and proportion with anti-HCV-antibody-positive status between the two groups ( Table 1).
Patients' Characteristics Categorized by AFP Level
Compared to patients with normal AFP levels, a smaller proportion of patients with elevated AFP were male (p = 0.02), had a pathological diagnosis of HCC (p < 0.001), were in AJCC stage 1 or 2 HCC (p < 0.001), had a solitary tumor (p < 0.001), were in BCLC stage 0 or A HCC (p < 0.001), were in Child-Pugh class A (p < 0.001), and had a low body mass index (BMI) (p < 0.001). Furthermore, patients with elevated AFP levels had larger tumors (p < 0.001), a higher total bilirubin level (p < 0.001), and a higher INR (p < 0.001) and a higher proportion of them were cirrhotic (p = 0.001) and HBsAg-positive (p = 0.002). However, there were no significant differences in age, creatinine level, proportion with AUD, and proportion with anti-HCV-antibody-positive status between the two groups (Table 1).
Discussion
In this study, we demonstrated that overall 51.2% of patients with HCC had elevated AFP levels. The proportion of patients with an elevated AFP level was stationary in the period from 2011 to 2020. The proportion of patients with BCLC stages 0-A HCC decreased from 2011 to 2020, whereas the proportion of patients with NBNC-HCC increased in the same period. Furthermore, the proportion of patients with early-stage HCC (i.e., BCLC stages 0-A) was lower for NBNC-HCC than for HBV-or HCV-related HCC. Our previous study reported that the HCC cases in our institution accounted for 9.8% of the total cases at the national level [18].
Independent factors associated with elevated AFP levels included female gender, increased tumor size, multiple tumors, BCLC stages B-D, cirrhosis, total bilirubin > 1.4 mg/dL, and viral etiology. Advanced tumor stage and viral etiology were associated with elevated AFP levels. Between 2011 and 2020, the proportion of patients with NBNC-HCC increased (which would have led to a decreasing trend in AFP elevation), which was counterbalanced by the decreased proportion of patients with BCLC stages 0-A HCC (which would have led to an increasing trend in AFP elevation). Ultimately, the proportion of patients with elevated AFP levels was stationary between the years 2011 and 2020 in this study. Furthermore, the decreasing proportion of patients with BCLC stages 0-A HCC during this period may be due to the concurrent increase in the proportion of patients with NBNC-HCC, because the BCLC stages 0-A were less frequently found in NBNC-HCC patients compared to HBV-or HCV-related HCC patients, a result that agrees with the findings of a previous study [19]. In that study, patients with NBNC-HCC presented with larger tumors and at later stages of disease compared to patients with virus-related HCC [19]. This result may be due to the low rate of HCC surveillance in NBNC-HCC patients; furthermore, a significant proportion of NBNC-HCC patients could be NAFLD-related cases [20].
A recent study from the US reported an overall 62.6% of HCC patients with elevated AFP levels (i.e., ≥20 ng/mL) at the time of diagnosis. Between 2010 and 2017, there was a decline in the percentage of HCC patients with elevated AFP levels (68.2% vs. 57.5%). Furthermore, the decline was most evident among patients with early-stage tumors (i.e., seventh edition AJCC stage 1), from 55.7% in 2010 to 40.7% in 2017. However, the authors did not investigate the potential cause of these results. They assumed that these results were likely due to the increasing trend in early-stage tumor detection and the shift from viral to nonviral etiology [6]. In contrast, in this study, an overall 51.2% of patients with HCC had elevated AFP levels. The proportion of patients with elevated AFP levels was stationary in the period from 2011 to 2020 for all patients and across different AJCC stages in the present study. However, it is unclear why there is a discrepancy between this study and the US study [6].
Previous studies have shown that female gender, viral etiology, severe underlying liver disease, and advanced tumor stage are independently associated with elevated levels of AFP [5,6], findings that are compatible with those of the present study. Interestingly, female gender is associated with elevated AFP levels in the present study and previous large-scale studies [5,6]. However, the underlying mechanism is still unknown. It is well known that severe liver disease and advanced tumor stage are associated with elevated AFP levels [5,6,21]. The link between viral etiology and elevated AFP may be due to the former's association with cirrhosis. A previous study demonstrated that NAFLD is the leading cause of non-cirrhotic HCC [22].
In Taiwan, HBV is the leading etiology of HCC [18]. In 1984, Taiwan established a universal HBV vaccination program for newborns [23,24]. HBV vaccination has reduced the incidence of HCC in children and adolescents [23,24]. Antiviral therapy for HBV and HCV has been widely used in Taiwan since 2003. Antiviral therapies reduce the incidence of HBV-and HCV-related HCC [25,26]. Consequently, the incidence of HBV-and HCVrelated HCC was expected to decrease. Indeed, of the 3843 HCC patients from five medical centers in Taiwan enrolled by Chang et al. in their study during 2005-2011, only 10.7% had NBNC-HCC [27], in contrast to 20.4% in 2011 and the gradual rise to 28.7% in 2020 in the present study. Therefore, we infer that the universal HBV vaccination program for newborns and the widely employed antiviral therapy for HBV and HCV have resulted in the decreasing incidence of HBV-and HCV-related HCC in Taiwan.
The etiologies of NBNC-HCC may be AUD and NAFLD; other etiologies, such as autoimmune hepatitis, primary biliary cirrhosis, and primary sclerosing cholangitis, related to chronic liver disease are rare [22]. Patients with NBNC-HCC have an increased risk of metabolic comorbidities [28], which implies that a significant proportion of NBNC-HCC cases could be NAFLD-related.
The incidence of HCC attributed to a nonviral etiology (mainly NAFLD) is rising [29,30]; in addition, it is associated with a decline in the percentage of patients with elevated AFP levels. Other tumor markers more specific to NBNC-HCC are needed for HCC surveillance in this population. Another available tumor marker for HCC surveillance is des-γ-carboxyprothrombin (DCP), which is recommended in clinical practice guidelines by the Japan Society of Hepatology [31]. Previous studies have demonstrated that elevated DCP might be a diagnostic marker for NBNC-HCC [32,33].
Due to the obesity pandemic, the challenge is how to screen for HCC in patients with NAFLD. The American Gastroenterological Association (AGA) guidelines recommend that screening for HCC should be considered for cirrhotic patients due to NAFLD. When the quality of ultrasound is suboptimal for HCC screening (e.g., due to obesity), future surveillance should be performed using CT or MRI, with or without determining the AFP level, every 6 months [34]. However, the cost-effectiveness of HCC surveillance, if CT or MRI replaces ultrasound, remains unknown [34].
Modern molecular biology-based technologies (e.g., liquid biopsy) hold considerable promise for early diagnosis of HCC. To date, there are still no US Food and Drug Administration (FDA)-approved liquid biopsy assays for HCC, mainly due to the lack of survival benefit of such assays [35].
The strength of the present study is the use of a large cohort of patients with HCC with prospectively collected data and limited missing data. However, the study has several limitations. First, we did not use the Alcohol Use Disorders Inventory Test (AUDIT) [36] or AUDIT-C [37,38] (which is recommended by the European Association for the Study of the Liver (EASL) guidelines [39]) to screen HCC patients for AUD. In contrast, we reviewed medical records and used habitual drinking to define AUD, which may underestimate the prevalence of AUD in the present study. Second, the study lacked data on antiviral therapy for HBV and HCV, which can affect AFP levels. Previous studies have shown that the cutoff values of AFP for HCC surveillance are lower in HBV-related cirrhosis patients receiving nucleos(t)ide analogue therapy [40] and HCV-related cirrhosis patients treated with DAAs [41]. Third, the study lacked data on HCV ribonucleic acid (RNA). A proportion of patients with anti-HCV-antibody-positive status could have experienced a past episode of resolved HCV infection. Fourth, etiologies other than HBV, HCV, and AUD were unavailable. Furthermore, we also did not have data on hepatic steatosis and metabolic comorbidities [42]. Therefore, we could not define NAFLD in the present study. Finally, this is a retrospective study.
Conclusions
In the past 10 years, a stationary trend in elevated serum AFP levels in HCC patients was noted in this cohort from an HBV-endemic area. This result may be due to the proportion of patients with early-stage HCC decreasing and the proportion of patients with NBNC-HCC increasing during this period. Furthermore, advanced tumor stage, severe underlying liver disease, female gender, and viral etiology were associated with elevated AFP in HCC patients. | 2023-02-17T16:09:21.051Z | 2023-02-01T00:00:00.000 | {
"year": 2023,
"sha1": "71e2a85c03b9c3e99f7b3e105611e574806bca74",
"oa_license": "CCBY",
"oa_url": null,
"oa_status": null,
"pdf_src": "PubMedCentral",
"pdf_hash": "26d38d583e623cc253ea8c0b33eaafa06d6ab92e",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
245432755 | pes2o/s2orc | v3-fos-license | The Effect of Postmenopausal Osteoporosis on Middle Ear Resonance Frequency
Background: The effect of postmenopausal osteoporosis on the middle ear mechano-acoustic system is unknown. The aim of this study is to investigate whether or not middle ear resonance frequency is affected in females with postmenopausal osteoporosis. Methods: The study included postmenopausal women aged 45-60 years, separated into 2 groups as females with postmenopausal osteoporosis and healthy postmenopausal females (control group). A detailed anamnesis was taken from all subjects and then the ear, nose, and throat examinations were done followed by pure tone audiometry, tympanometry, and multifrequency tympanometry tests. The groups were compared in respect of pure tone average, bone conduction threshold, RF, static admittance, and tympanometric peak pressure values. Results: The mean age of the patients was 59.2 ± 4.53 years (range, 48-65 years) in the postmenopausal osteoporosis group and 57.11 ± 4.27 years (range, 48-65 years) in the control group (P > .05). The mean resonance frequency values for the postmenopausal osteoporosis and control group were 954.41 ± 127.47 and 935.29 ± 126.39 Hz (P > .05 ). The mean static admittance values for the postmenopausal osteoporosis and control group were 0.82 ± 0.33 and 0.85 ± 0.3 mmho, and mean tympanometric peak pressure values were −7.35 ± 18.52 and −6.94 ± 19.52 daPa (P > .05 for both static admittance and tympanometric peak pressure). The mean pure tone averagevalues for the postmenopausal osteoporosis and control group were 20.96 ± 6.82 and 15.60 ± 7.81 dB, and mean bone conduction threshold values were 17.57 ± 6.03 and 12.10 ± 6.52 dB (P < .05 for both pure tone average and bone conduction threshold). Conclusions: The results showed that the middle ear resonance frequency values were not affected in postmenopausal osteoporosis patients, but there was seen to be greater sensorineural hearing loss in females with postmenopausal osteoporosis compared to healthy postmenopausal females.
INTRODUCTION
Ostroporosis is a bone disease defined as increased bone fragility and increased risk of fracture associated with low bone mass and impaired bone microstructure. 1 Females constitute 80% of osteoporosis cases, and the majority of these are in the postmenopausal period. This disease, which develops in females due to estrogen deficiency in the postmenopausal period, is known as postmenopausal osteoporosis (POP). As POP is a systemic disease, all the bones in the body are affected. 2 Previous studies have shown that the temporal bone is affected by POP and this is related to sensorineural hearing loss. 3,4 However, the effect on the middle ear ossicles is unknown. The potential effect of POP on ossicles can cause a deterioration of stiffness and mass in the middle ear mechano-acoustic system, and this potential impairment can be revealed with immittancemetric measurements.
Classic immittancemetry is an economical and non-invasive test allowing the evaluation of acoustic admittance in the middle ear. A change in air pressure in the external ear canal provides a dynamic measurement of acoustic admittance. Classic immittancemetry in current use is usually applied with a 226 Hz probe tone. Multifrequency tympanometry is a method that provides the analysis of tympanograms obtained with probe tones varying between 226 Hz and 2000 Hz. High-frequency probe tones are extremely valuable in the determination of pathologies that change the stiffness and mass effect of the middle ear system. 5 One of the important parameters of multifrequency tympanometry is resonance frequency (RF). RF is the frequency with the lowest resistance and the highest vibration of the middle ear system and is affected by changes in mass and hardness occurring in the middle ear mechanoacoustic system. 6 The aim of this study was to investigate whether or not the RF, TPP, and SA of the middle ear were affected through the application of multifrequency tympanometry tests in POP patients without airbone gap.
MATERIALS AND METHOD
This research was approved by the Baskent University Medical and Health Sciences Research Committee and was supported by the Baskent University Research Fund. Informed consent was obtained from all the study participants. This prospective, controlled clinical study was conducted in the Audiology Unit of the Ear, Nose, and Throat Diseases Department. The participants were separated into 2 groups, as those aged 45-60 years diagnosed with POP, and agematched, healthy, postmenopausal volunteers.
The diagnosis of osteoporosis in the postmenopausal females was made from the determination of the T-score value in the bone mineral density measurements taken with the dual x-ray absorptiometry (DXA) technique. Based on the World Health Organization (WHO) criteria, 7 a T-score of <−2.5 was evaluated as postmenopausal osteoporosis. The control group was formed of postmenopausal women with a T-score of >−1.
The patients with postmenopausal osteoporosis included in the study were those with no external ear canal or tympanic membrane pathology determined in the otoscopic examination, no history of ear disease or ear surgery, had not previously received any treatment which could affect bone metabolism, had middle ear pressure in the limits of ±50 daPa in the electroacoustic immittancemetry evaluation, had a type A tympanogram, and had no conductive or mixed type hearing loss in the audiometric examination. Audiograms with air-bone gap >10 dB at least 1 frequency were evaluated as a conductive hearing loss if the bone thresholds were <20 dB at all frequencies, and mixed hearing loss if the bone thresholds were ≥20 dB at least one frequency, and these patients were excluded from the study. Patients with any systemic chronic disease were excluded from the study. First, an ear, nose, and throat (ENT) examination was applied by an ENT specialist, and a detailed history was taken. Patients deemed suitable for inclusion in the study as a result of this examination were then applied with pure tone audiometry, tympanometry, and multifrequency tympanometry tests.
Pure Tone Audiometry
The pure tone audiometry was applied in quiet rooms according to the standards of the Industrial Acoustic Company (IAC) Inc. using a Clinical Audiometer AC40 device (Interacoustics, Assens, Denmark). The determination of the type and degree of hearing loss was based on pure tone average (PTA). By determining the air and bone conduction thresholds (BCT) at 500, 1000, and 2000 Hz, the PTA values for both ears were obtained separately. The Northern and Downs classification 8 was used for the classification of hearing loss. According to this classification, a pure tone average less than 16 dB is accepted as normal hearing, 16-25 dB slight hearing loss, 26-30 dB mild hearing loss, 31-50 dB moderate hearing loss, 51-70 db severe hearing loss, 70 dB and above profound hearing loss.
Immittancemetric Measurements
The immittancemetric measurements of the study participants were taken using a GSI Tympstar Version 2 (Grason Stadler Inc., MN, USA) electroacoustic immittancemeter. Using a 226 Hz probe tone, the tympanogram and static admittance values were recorded. The tympanogram entries are recorded at air pressure varying between +200 and 400 daPa at the rate of 200 daPa/s. In patients determined with type A tympanogram, multifrequency tympanometry measurements were taken.
The device takes the multifrequency tympanometry measurement in two stages. In the first step, standard tympanometry data are investigated such as static admittance (SA), tympanometric peak pressure (TPP), and gradient value by giving a fixed frequency probe tone, changing the pressure between 200 and −400 daPa, and the tympanogram is obtained. In the second step, by keeping the pressure constant, the middle ear RF value is determined by stimulating each ear in the frequency range of 250-2000 Hz consecutively at 50 Hz intervals. At the end of the test, the outputs were recorded together with other immittancemetric values.
Statistical Analysis
In the biostatistical analysis, it was calculated to be necessary to have at least 32 subjects in the study group and 32 in the control group for the power of the study to be ≥0.90. Data obtained in the study were analyzed statistically using SPSS v. 22.0 (IBM SPSS Corp.; Armonk, NY, USA). All data were recorded, and the arithmetic mean, standard deviation, minimum and maximum values were calculated. Normality of the variables and homogeneity was evaluated with the Levene test. In the comparison of mean values between groups, the Student's t-test was applied. A value of P < .05 was accepted as statistically significant.
RESULTS
Evaluation was made for 146 ears of 73 postmenopausal women, as 74 ears of 37 women in the POP group and 72 ears of 36 control group women. Two women (4 ears) from the POP group and 2 women (4 ears) from the control group were excluded because of mixed hearing loss. One woman (2 ears) from the POP group was also excluded because of conductive hearing loss. As a result, the data of 136 ears of 68 women (34 women in control group and 34 women in study group) were used for analysis.
The mean age of the participants was 59.2 ± 4.53 years (range, 48-65 years) in the POP group and 57.11 ± 4.27 years (range, 48-65 years) in the control group. No statistically significant difference was determined between the groups in respect of age (P = .53).
The mean PTA value was 20.96 ± 6.82 dB in the POP group (68 ears) and 15.60 ± 7.81 dB in the control group (68 ears) (P < .05). The mean BCT value was 17.57 ± 6.03 dB in the POP group (68 ears) and 12.10 ± 6.52 dB in the control group (68 ears) (P < .05). In both the PTA and BCT values, a statistically significant difference was determined in the comparisons of the POP group and the control group (Table 1).
Normal hearing was found in 43 ears (63%), slight hearing loss in 17 ears (25%), mild hearing loss in 3 ears (4.5%), and moderate hearing loss in 5 (7.5%) ears of the control group. In the POP group, normal hearing was found in 20 ears (29.5%), slight hearing loss in 33 ears (48.5%), mild hearing loss in 7 ears (10.5%), and moderate hearing loss in 8 ears (11.5%) ( Figure 1). None of the individuals included in the study had severe or profound hearing loss.
The mean RF value was 954.41 ± 127.47 Hz in the POP group (68 ears) and 935.29 ± 126.39 Hz in the control group (68 ears) (P > .05). In the RF values, no statistically significant difference was determined in the comparisons of the POP group and the control group ( Table 1).
The mean SA value was 0.82 ± 0.33 mmho in the POP group (68 ears) and 0.85 ± 0.30 mmho in the control group (68 ears) (P > .05). The mean TPP value was −7.35 ± 18.52 daPa in the POP group (68 ears) and −6.94 ± 19.52 daPa in the control group (68 ears) (P > .05). In the SA and TPP values, no statistically significant difference was determined in the comparisons of the POP group and the control group (Table 1).
In the PTA, BCT, RF, SA, and TPP values, no statistically significant difference was determined in the comparisons of the right and left ear of the POP group. All mean audiological values of right and left ears in the POP group are shown in Table 2. In the PTA, BCT, RF, SA, and TPP values, no statistically significant difference was determined in the comparisons of the right and left ear of the control group. Mean audiological values of right and left ears in the control group are shown in Table 3.
DISCUSSION
Postmenopausal osteoporosis is a systemic disease, characterized by reduced bone mass and increased bone turnover, which affect all bones in the body. As all the bones of the skull are affected by osteoporosis, the temporal bone and otic capsule are also affected. 9 However, the effect of osteoporosis on the middle ear ossicles and sound transmission is not clear. The internal structure of middle ear ossicles is formed of compact bone, bone lacunae, and cartilaginous nodules. In a histomorphological study of 52 ear ossicles, Sarrat et al 10 showed that there was a great difference between specimens in respect of the degree of these cavitations and these increased with increasing age. In another study, high-resolution CT showed that the middle ear ossicles of patients with osteoporosis were affected. 11 Although Kumar et al reported that the osteoporotic process affected middle ear resonance frequency, no significant air-bone gap was determined in the audiometry test in the same study. 12 In the current study, no significant difference was determined between POP patients without air-bone gap and the healthy postmenopausal women in respect of middle ear RF, SA, and TPP values. This finding demonstrates that the middle ear ossicles were not affected by POP to an extent to change the RF.
Significantly greater sensorineural hearing loss was determined in the POP group than in the control group in this study (P < .05). Several previous studies have shown a relationship between osteoporosis in the postmenopausal period and sensorineural hearing loss. 3,4,[13][14][15] Demineralization of the otic capsule and cochlea has been shown to be responsible for the sensorineural hearing loss in these patients, and as demineralization increases, so hearing loss also increases. 16,17 Pathologies affecting the middle ear ossicle chain cause an air-bone gap in pure tone audiometry. 18 There is also a possibility that the potential loss of mass in the ossicles associated with osteoporosis is a reason for the air-bone gap in audiometry. As it is difficult to differentiate other ossicle chain problems that affect the air-bone gap, subjects with conductive or mixed type hearing loss were excluded from this study to avoid bias. In the current study, patients with air-bone gap >10 dB at least in 1 frequency were excluded from the study. Although this exclusion criteria were relatively strict, only 3 of 37 POP patients were excluded for this reason. It may have been better to discount other ossicle chain pathologies by taking high-resolution CT of patients with conductive or mixed type hearing loss and then they could have been included in the study.
CONCLUSION
The results of this study showed that there was significantly more sensorineural hearing loss in the POP group than in the control group. The data also showed that the middle ear RF, SA, and TPP of POP patients without air-bone gap were not affected. This demonstrates that the postmenopausal osteoporotic process does not affect the middle ear ossicles to an extent that will change the multifrequency tympanometry in POP patients with sensorineural hearing loss. Nevertheless, there is a need for further controlled histomorphological studies to more accurately determine to what extent the middle ear ossicles are affected in POP patients. | 2021-12-24T16:05:22.191Z | 2021-11-01T00:00:00.000 | {
"year": 2021,
"sha1": "0381a389f24471651b57d0e87874c28bf3230c25",
"oa_license": "CCBYNC",
"oa_url": null,
"oa_status": null,
"pdf_src": "PubMedCentral",
"pdf_hash": "68f69597c6076c1aa3ab0158809138507764d174",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
256192391 | pes2o/s2orc | v3-fos-license | Phytoplankton and benthic infauna responses to aeration, an experimental ecological remediation, in a polluted subtropical estuary with organic-rich sediments
Fine-grained organic-rich sediments (FGORS) are accumulating in estuaries worldwide, with multi-faceted negative ecosystem impacts. A pilot experiment was carried out in a residential canal of the Indian River Lagoon estuary (IRL, Florida, USA) using an aeration treatment intended to mitigate the harmful ecological effects of organic-rich sediment pollution. Planktonic and benthic communities were monitored, and environmental data collected throughout the aeration process. Results were compared against control conditions to evaluate the efficacy of aeration in the mitigation of FGORS. During the aeration process, hurricane Irma impacted the study area, bringing heavy rainfall and spawning a brown tide event (Aureoumbra lagunensis). The overall thickness and volume of FGORS, and the organic content of surface sediments did not change during the aeration treatment. Dissolved oxygen was higher and ammonium concentrations were lower in aeration canal bottom water compared to the control canal. During treatment, aeration did facilitate benthic animal life when temperatures dropped below 25°C, likely due to water column mixing and the increased capacity of water to hold dissolved gasses. In general, aeration did not significantly change the planktonic community composition relative to the control canal, but, during the post-bloom period, aeration helped to weaken the brown tide and phytoplankton densities were 35–50% lower for A. lagunensis in aeration canal surface water compared to the control canal. Aeration has important management applications and may be useful for mitigating algal blooms in flow-restricted areas and promoting benthic communities in cooler environments.
This study of aeration in an estuarine residential canal was conducted to provide data in a replicated and controlled in-situ environment. Biological and environmental responses were compared between the aerated canal and an adjacent non-aerated control canal of similar structure and circulation. In this in situ estuarine study, we hypothesized that 1) aeration would reduce hypoxia and eutrophic nutrient levels; 2) aeration would change phytoplankton and benthic infauna abundances, diversity and community compositions.
Study sites and aeration treatment
This study was carried out in the mid-section of the northern IRL system located on the east coast of central Florida, USA. No permit was required for the study since the experiment was conducted in the public waterways. It is in a humid subtropical climate zone with a rainy season from May through September, and a dry season from October to late April. The average annual temperature and precipitation are 22.2˚C and 136.4 cm, respectively. The experiment was conducted in residential canals connected to the larger estuary (IRL) through the Grand Canal with no direct freshwater discharge into the experimental area (Fig 1A, 1C and 1D). Residential canals are long rectangular canals lined with seawalls. They constitute the back yards of single-family houses abutting the canal for convenient estuary access. One canal was aerated and another, 800-m away, served as control. The control canal was selected for its geomorphological and hydrological similarities to the aerated canal (Fig 1A and 1B). One sampling station near the mouth of each canal, but in the Grand Canal, was selected as an intermediate site for comparison (Fig 1B and 1C). The aeration system for this 6070.3 m 2 canal consisted of five, 30.48 cm 2 micro-porous diffusers placed on the benthos, amongst FGORS, with 50-m between diffusers (Fig 1C and 1D). Three additional diffusers were placed next to the fifth diffuser at the mouth of the canal to establish a fine bubble retention curtain to limit the loss or gain of FGORS (Fig 1C and 1D). Diffuser placement and pump size were selected based on standard industry practices. Two 1.5 hp air pumps yielded 2.47×10 −4 hp per m 2 . Diffuser spacing was
Phytoplankton collection and laboratory processing
The canal water column averaged 2.5 m in depth. Phytoplankton samples were collected using a peristaltic pump with acid-washed Tygon tubing at specified depths (0.5m below the surface and 0.5m above the bottom) at the 6 stations ( Fig 1B and 1C). Three replicates were taken at each sampling depth at each station. Each sample was split into two subsamples. One subsample (30 ml) was filtered (53-μm), kept on ice, and analyzed using flow cytometry (BD Accuri C6) within 12 hrs of collection. The other subsample (250 ml) was preserved in 4% formalin pending analysis and identification. Preserved samples were concentrated using the sedimentation technique described in Standard Operating Procedure for Phytoplankton Analysis [27]. The sedimentation process was conducted in 24-cm graduated cylinders and took 6 days. After settling, overlying water was drawn off via syringe. Concentrated samples were counted using a graduated Sedgewick Rafter counting chamber until � 300 cells were counted. Phytoplankton were identified to the lowest possible taxonomic level. Flow cytometry was used to differentiate nano-phytoplankton according to their pigmentation and cell sizes [28]. Counting via compound light microscopy yielded the identities and densities of the larger microphytoplankton.
Benthic infauna collection and laboratory processing
Benthic infauna sampling approach followed Poirrier et al. (2008) [29]. Three benthic infauna samples were collected randomly at each sampling location (Fig 1B and 1C) using a Wildco Petite Ponar Grab (sampling area 225 cm 2 ). Sediment grab volumes, measured via graduated cylinder, enabled determination of grab penetration depth. Samples were sieved through 500μm mesh [30] and retained organisms were frozen for lab analysis. The infauna samples were examined within one month after collection and thawed organisms retained their anatomical structure enabling confident laboratory identification. Aliquots of each sample (1/4 or 1/8) were sorted to ensure at least 100 organisms were counted in each sample. Organisms were identified to the lowest possible taxonomic level and counted via stereomicroscopy (8×-35× magnification).
Environmental data collection methods
All environmental data were collected simultaneously with phytoplankton samples. Secchi Depths were measured using a standard black-white estuary-style Secchi disk. The disk was lowered into the water column with a graduated rope under sunlight. The Secchi Depth is the depth where underwater disk pattern differentiation is no longer possible. Temperature, salinity, pH and DO were measured using Yellow Springs Instruments ProDSS. The Sonde was factory calibrated and then calibrated following manufacturer's guidelines immediately prior to each sampling event. Water depths and FGORS thicknesses were determined at mapped locations (~50 locations in each canal, S1 Fig) using a capped 4-cm diameter polyvinyl chloride (PVC) probe. The probe was first lowered through the water column to the benthic surface and water depth was recorded. The probe was then pushed into the sediment until encountering a firm bottom, and this was recorded as the total (muck plus water column) depth. The thickness of probe-penetrable sediment (muck) was calculated by subtracting the water depth from the total depth. Water depths were verified by sounding with a 20-cm disc that settled onto the surface of soft sediments. Sediment thicknesses were validated for selected sites with sediment cores. The surface areas and volumes of muck were extrapolated by contouring probe data via ArcGIS (S1 Fig). Samples for sediment quality analysis were collected during the FGORS thickness survey using a 0.15 m 2 Ekman grab. Subsamples of the surface sediments (top 3 cm) were placed into polystyrene vials and sealed with parafilm. Samples were then weighed, frozen, freeze-dried (Labconco Freezone 6) and re-weighed to determine water content. Dried sediments were homogenized (Spex 8000 mixer mill) and organic matter content was determined via the Loss on Ignition (LOI) method (loss of mass after combustion at 450˚C for 4 hours) [31].
Water samples for nutrient analyses were collected alongside phytoplankton samples using a peristaltic pump and acid-washed (HCl) Tygon tubing at multiple depths at each sampling station. Water samples were pumped into acid-washed LDPE bottles and stored on ice for transport. In the laboratory, samples were immediately vacuum-filtered through 47-mm diameter 0.4-μM pore-sized polypropylene filters. Analyses of ammonium (NH 4 +), nitrate plus nitrite (NO 3 - , total dissolved phosphorus (TDP) and silica were determined using a segmented continuous flow nutrient autoanalyzer (SEAL AA3) following the manufacturer's methods. Standard reference materials were run with each batch of samples and standards were within 10% of the certified values for all analyses.
Statistical analyses
One-way analysis of similarities (ANOSIM) based on Bray-Curtis similarities was used to detect differences in phytoplankton assemblages between and among different sites, and before and after aeration (α = 0.05, R version 4.1.3). Similarity percentage analysis (SIMPER) was used to determine which species contributed to observed differences (R version 4.1.3). Phytoplankton density data were square-root transformed for ANOSIM and SIMPER tests. Environmental parameters, benthic infauna densities, Shannon's Diversity Indexes and A. lagunensis cell densities were compared between aeration and control canals using 2-way ANOVA and Fisher's LSD post-hoc pairwise comparisons (α = 0.05, GraphPad Prism v9). Spearman's correlation analyses between benthic infauna diversity indexes and abundances and environmental parameters were performed for all sampling stations (α = 0.05, R version 4.1.3).
Phytoplankton community-changes during aeration
Fifty-two micro-phytoplankton taxa, mostly dinoflagellates, were found in canals (S1 Table). Twenty-one nano-phytoplankton taxa were found in canals, including 11 groups of cyanobacteria (S1 Table). Two algal blooms were observed during the study. Scrippsiella sp., a dinoflagellate, formed a small bloom in surface water in September 2017 with an average density > 1×10 3 cells ml -1 in all canals (Fig 2A). A. lagunensis, notoriously referred to as "brown tide", bloomed from February 2018 to June 2018, with the highest density being 3.5 × 10 6 cells ml -1 (Fig 2C and 2D).
Nano-phytoplankton dominated all canals year-round (Fig 2C and 2D). Densities of nano-phytoplankton were 2 orders of magnitude higher than that of micro-phytoplankton during non-bloom periods, and 4-5 orders higher during the bloom (Fig 2). The main differences between the aeration and control canals were 40% and 47% fewer total cell counts for surface nano-phytoplankton in the aeration canal in April and June 2018, respectively ( Fig 2C). This is mainly attributed to the cell density differences of A. lagunensis (Fig 3A). Surface water cell counts for A. lagunensis were significantly lower in the aerated canal relative to the control in April 2018 (p < 0.05) and June 2018 (p < 0.001) (Fig 3A). Meanwhile, average ammonium (Fig 3B, 3E and 3H) and surface water DO concentrations ( Fig 3C) were lower in the aeration canal compared to the control from February to June 2018.
Comparing across months, temporally distinct planktonic communities developed throughout the experimental period (Fig 4). Both of the surface and bottom planktonic communities were more divergent in summer and fall, but converged in winter and spring (Fig 4). Dissimilarities among different months mainly came from nano-sized cyanobacteria and the A. lagunensis bloom (ANOSIM and SIMPER tests, p < 0.001, R > 0.95). After aeration (September 2017-June 2018), bottom water planktonic community assemblages from the aeration canal were more consistent with those of the Grand Canal compare to the control canal ( Fig 5C and 5D). However, in general, aeration treatment differences were overshadowed by seasonal variations (Figs 4 and 5).
Benthic infauna community-changes during aeration
Benthic infauna were present only in late fall to early spring (Table 1). In February 2018, benthic infauna were significantly more abundant in the aeration canal relative to the control (p < 0.05, Fig 6D). Meanwhile, bottom water temperature was below 25˚C (Fig 6B) and DO concentration was significantly higher in the aeration canal compared to the control in February 2018 (p < 0.0001) and April 2018 (p < 0.001) (Fig 6C, Table 1). Correlation analyses between benthic infauna communities and environmental parameters revealed that both abundance and diversity of benthic infauna positively correlated with DO (p < 0.05, rho > 0.5) and negatively correlated with temperature (p < 0.05, rho < -0.5) (Fig 6A).
Environmental conditions-changes during aeration
Concentrations of ammonium and nitrate + nitrite peaked in all canals at >200 μg N L -1 during November 2017 following Hurricane Irma (Fig 7A). Coincident with the A. lagunensis bloom in February 2018, near-complete depletion of ammonium and nitrate + nitrite was observed in all canals. Depletion of Dissolved Inorganic Nitrogen (DIN) coincided with a sharp increase in Dissolved Organic Nitrogen (DON) (Fig 7A). Concentrations of ortho-phosphate showed peak concentrations of >40 μg P L -1 in all canals before aeration, and remained below 40 μg P L -1 during aeration (Fig 7B). Dissolved Organic Phosphorus (DOP) showed the same trend observed for DON, which peaked > 100 μg P L -1 in February 2018 during the A. lagunensis bloom and was around 20 μg P L -1 before and after the bloom period. Silica concentrations remained high during the whole sampling period (6 to 16 mg Si L -1 ) (Fig 7F).
Salinity and temperature followed similar vertical and temporal trends in all canals. Salinity stratifications were observed in summer, with maximum salinity of 25 in the bottom water column in August 2017 ( Fig 7C); however, during cooler months when water temperatures were below 25˚C, salinity was well-mixed. Vertically uniform salinities continued in all canals with values increasing to 19 by April 2019, after which the seasonal pattern began to repeat (Fig 7C). Temperature was uniform with depth all year in all canals with maxima of 29-32˚C during August 2017 and June 2018 and minima of 21-22˚C in November 2017 (Fig 7C). Prior to the onset of aeration (August 2017), bottom water was anoxic in all canals. However, by November 2017, aerated canal water had DO concentrations of~60% saturation throughout the water column (Fig 7D). In contrast, DO was variable in the control canal, with several instances of hypoxia (Fig 7D). Bottom water DO was significantly higher in the aeration canal compared to the control in February 2018 (p < 0.0001) and April 2018 (p < 0.001) (Fig 7C, Table 2), while surface water DO was significantly higher in the aeration canal in September 2017 (p < 0.05), February 2018 (p < 0.01) and June 2018 (p < 0.001) ( Table 2). Secchi Depths were deeper in the non-bloom periods, but only 30 cm during the A. lagunensis bloom in February 2018 (Fig 7E). The first complete FGORS resurvey was carried out before aeration (July 2017) and revealed the aeration canal to have~93% benthic FGORS coverage, with an estimated volume of 4,640 m 3 and a mean thickness of 0.82 m (Table 3). A survey during aeration and eleven days post-hurricane Irma (September 2017) revealed an increase of 168 m 3 of FGORS to the aeration canal following the storm (Table 3). Aeration and the aeration curtain were shut down when the hurricane passed through. In the aeration canal, the post-aeration FGORS survey indicated a mean FGORS thickness increase of 12 cm during the one-year study period, greater than our measurement precision of ± 10 cm, indicative of an actual volume change, whereas the post-aeration FGORS thickness estimate was not significantly different from pre-aeration survey in the control canal (Table 3). FGORS organic content (loss on ignition, LOI), however, was unaffected by time or treatment (Table 3, Fisher's LSD test, p > 0.05 in all months).
Discussion
The canal system is a nutrient-and sediment-polluted shallow brackish waterway with limited water exchange though small connections to the greater estuary. With high concentrations of dissolved nitrogen and phosphorus, high phytoplankton biomass is common. Comparing across months, temporally distinct planktonic communities developed through seasons, but communities evolved similarly in the two canals and changes seemed unrelated to treatment (Fig 4). These data showed that regional and temporal trends overshadowed would-be aeration effects. Our results were inconsistent with the findings of Huisman et al. (2004) which provided strong lab and field evidence that vertical mixing favors the growth of specific phytoplankton groups, such as diatoms [32]. One possible explanation is that the aeration system lacked the strength for sufficient mixing, especially in the summer when the water column was highly stratified [14]. During the brown tide, surface water cell counts of A. lagunensis were 35% and 50% lower in the aeration canal in April and June 2018, respectively (Fig 3A). This difference could be due to aeration canal oxidation of ammonium to chemical species not readily used by A. lagunensis [33,34]. A. lagunensis prefer to assimilate reduced nitrogen, such as ammonium [33]. Indeed, the average concentrations of ammonium in the aeration canal were lower compared to the control canal during the bloom fading period (Fig 3H) when differences of A. lagunensis cell densities were observed (Fig 3G). Meanwhile, surface DO was significantly lower in the aeration canal, possibly being consumed in the oxidation of reduced nitrogen (Fig 3C). This suggests that aeration decreased concentrations of nitrogen available to A. lagunensis, restraining the bloom in the aerated canal. Thus, short-term aeration benefits included significantly fewer A. lagunensis during the bloom fading period (April to June 2018) (Fig 3A). Although subtle relative to seasonal shifts, the potential to impact algal bloom densities has important management implications; aeration may be useful for managing and mitigating algal blooms. During the warm period, the deepest water (>1.5 m) in all canals was anoxic and no benthic infauna were observed (Fig 6D and Table 1). Aeration failed to maintain DO in summer bottom water when oxygen capacity is lowest and benthic respiration is highest. During the colder months (November 2017 to February 2018), however, benthic infauna were more abundant and diverse in the aeration canal (Fig 6D and 6E). FGORS thickness and surface sediment LOI was similar throughout this study, but benthic infauna communities were correlated with bottom water DO (Fig 6A). These data suggest that higher DO concentrations in the aeration canal in winter months promoted benthic infauna. Aeration apparently rendered the environment more habitable to benthic fauna without removing or changing the composition of benthic sediments, although other studies have shown that sediment organic content is a key factor for benthic infauna [35,36]. Our finding is consistent, however, with another study showing that DO, rather than chemical pollutants, controlled benthic infauna abundance [37]. Aeration in this study facilitated colonization by benthic infauna during the colder months (< 25˚C), but was insufficient to overcome bottom hypoxia during summer months (> 25˚C).
Benthic infauna in canals were species adapted to polluted or harsh environments. For instance, Capitella capitata tolerates sewage-impacted sites [38]; polychaete annelids such as Ctenodrilus serratus and Capitella capitata have high tolerance of heavy metal pollutants [39]; Mulinia lateralis is an opportunistic bivalve known to invade rapidly after anoxic events in polluted estuaries [40]. Leptochelia dubia and Capitella capitata were found in an aliphatic hydrocarbon polluted area [41]. Though aeration produced a habitable winter benthos, colonization was mainly by pollution-tolerant species. Higher DO promotes an overall healthier benthic system. Nevertheless, these data suggest that aeration could be employed to promote benthic colonization in areas usually characterized by hypoxic or anoxic sediments. Ecological succession might then lead to the establishment of more diverse estuarine communities over time.
Collectively, these data suggest that when aeration was able to overcome intense stratification, and it stabilized dissolved oxygen concentrations, decreased concentrations of dissolved ammonium in the water column, decreased the intensity of algal blooms and promoted benthic colonization. Based on these modest outcomes, it is possible that more efficient delivery of oxygen using other systems, such as nanobubbles, could overcome the challenges encountered during warmer months in the present study. Overall, aeration has some important management applications and may be useful for mitigating active algal blooms and promoting recovery of benthic communities. Once established, the presence of benthic communities could help to maintain dissolved oxygen in sediments via bio-irrigation and bioturbation. These processes could create a positive feedback loop, helping to sustain the improved conditions and facilitating the return of ecosystem services such as nitrogen removal via denitrification when appropriate conditions exist [42].
Conclusion
The canal system is a highly polluted shallow brackish waterway. One year of aeration did not significantly change nutrient conditions in the form of sediment organic content, nor the inherent potential of sediments to flux nutrients into the water column and precipitate blooms. Dissolved oxygen levels and the biological communities, however, were significantly altered by aeration under some conditions. In general, aeration did not significantly change the planktonic community composition relative to the control canal. During the post-bloom period, however, aeration appeared to help weaken the brown tide, which may have been due to increased bottom DO and decreased bottom ammonium concentrations in the aeration canal. During treatment, aeration facilitated benthic animal life when temperatures dropped below 25˚C, likely due to water column mixing and the increased capacity of water to hold dissolved oxygen. Aeration has important management applications and may be useful for mitigating algal blooms in flow-restricted areas and promoting benthic communities. | 2023-01-25T06:18:13.036Z | 2023-01-24T00:00:00.000 | {
"year": 2023,
"sha1": "a3eb263ce80e8fa64e76b668b553e04279b2c3af",
"oa_license": "CCBY",
"oa_url": null,
"oa_status": null,
"pdf_src": "PubMedCentral",
"pdf_hash": "535587e435da3f2b1e0b456e17d063d5dd21c01f",
"s2fieldsofstudy": [
"Environmental Science"
],
"extfieldsofstudy": [
"Medicine"
]
} |
53028312 | pes2o/s2orc | v3-fos-license | Effect of adherence to growth hormone treatment on 0–2 year catch-up growth in children with growth hormone deficiency
Background Quantifying the association between adherence and the growth response to growth hormone (GH) treatment is hampered by suboptimal methods of measuring adherence, confounders associated with the growth response, and restriction of the outcome parameters to yearly growth velocities. Aim To investigate the effect of adherence on the two-year growth response to GH treatment in prepubertal children with idiopathic isolated growth hormone deficiency (GHD) participating in the easypod connect observational study (ECOS), a 5-year, Phase IV open-label study to continuously assess real-world adherence via the easypod electronic drug-delivery device. Patients and methods Outcome measures were change in height standard deviation score (ΔHSDS), index of responsiveness (IoR), and parameters of two catch-up growth (CUG) curve functions (monomolecular growth curve and second degree polynomial) with adj-HSDS (HSDS minus Target height (TH) SDS) as dependent variable. Inclusion criteria were GHD, naïve to GH treatment, known TH, age <10y in girls and <12y in boys, ≥3 measurements, HSDS <-2 at start, complete data on growth and adherence in the first and second year. Linear regression analyses were performed to test the association between adherence (continuous and high vs. low) and the outcome measures, also adjusted for potential clinical confounders (age at start, adj-HSDS at start, birth weight SDS, gestational age (<37 weeks vs ≥37 weeks), GH dose, GH max (n = 58)). The formula of IoR already adjusts for confounders. Results In total, 95 patients complied with the inclusion criteria. The strongest associations were found between high adherence in the second year (≥91% as cut-off value) and IoR 2y (+0.62), and average adherence and high adherence (≥78%) in the first two years and ΔHSDS 0-2y (+0.11 SD per 1 injection/week, and +0.34 SD for high vs. low adherence). Conclusion Suboptimal adherence negatively affected the growth response in the first two years of GH treatment.
Aim
To investigate the effect of adherence on the two-year growth response to GH treatment in prepubertal children with idiopathic isolated growth hormone deficiency (GHD) participating in the easypod connect observational study (ECOS), a 5-year, Phase IV open-label study to continuously assess real-world adherence via the easypod electronic drug-delivery device.
Patients and methods
Outcome measures were change in height standard deviation score (ΔHSDS), index of responsiveness (IoR), and parameters of two catch-up growth (CUG) curve functions (monomolecular growth curve and second degree polynomial) with adj-HSDS (HSDS minus Target height (TH) SDS) as dependent variable. Inclusion criteria were GHD, naïve to GH treatment, known TH, age <10y in girls and <12y in boys, �3 measurements, HSDS <-2 at start, complete data on growth and adherence in the first and second year. Linear regression analyses were performed to test the association between adherence (continuous and high vs. low) and the outcome measures, also adjusted for potential clinical confounders (age at start, adj-HSDS at start, birth weight SDS, gestational age (<37 weeks vs �37 weeks), GH dose, GH max (n = 58)). The formula of IoR already adjusts for confounders.
Results
In total, 95 patients complied with the inclusion criteria. The strongest associations were found between high adherence in the second year (�91% as cut-off value) and IoR 2y PLOS
Introduction
Growth hormone (GH) treatment for children with GH deficiency (GHD) is efficacious in generating catch-up growth (CUG) in the first years of treatment, usually followed by a maintenance phase and finally leading to a height close to the target height (TH) [1,2]. However, the growth response has been shown to be quite variable. Several groups have investigated the clinical features associated with the growth response [3,4,5,6], and in general approximately 50% of the variance could be explained by features like the severity of GHD (as assessed by the GH peak in provocation tests), age, bone age delay, birth weight, etc [7]. None of these prediction models could include adherence into their models, because there were no tools to assess adherence accurately. However, several studies have shown that adherence has an important effect on growth on GH treatment [8,9,10,11,12], although they could only be conducted over short time periods, and using a methodology based on self-reporting or number of issued and renewed prescriptions.
There are several challenges in studying the association between adherence to growth hormone (GH) treatment versus the growth response to GH. The first challenge is to make a decision on which growth parameter is the most suitable outcome parameter. So far, most studies investigating predictive factors have used yearly growth velocity data, particularly first year's height velocity [3,6]. Others have used the change in height Standard Deviation Score (HSDS) (delta HSD, ΔHSDS) over 1 or 2 years [4,5]. All these parameters depend on the age and height SDS at start of treatment, and do not fully represent the actual shape of the catch-up growth (CUG) in the first years of GH treatment in children with GHD. The second challenge is that, besides adherence, there are many other clinical parameters that influence the growth response to GH, as illustrated by the various prediction models [7]. So, in order to obtain an unbiased effect of adherence, it is necessary to adjust for important clinical predictors that are confounders in an analysis of the effect of adherence on the growth response to GH. The third challenge is to find a reliable method for measuring adherence.
In this study we accounted for these challenges. First, we used a novel way of expressing CUG on GH treatment, using mathematical models of HSDS (adjusted for TH]) over two years of GH treatment. In an earlier study, we showed that mathematical modelling of CUG with the monomolecular growth curve is suitable for celiac disease [13]. A second degree polynomial was also used, because for most children the shape of the HSDS in the first two years is a catch-up in growth followed by stabilization. Second, we entered important potential confounders into the models. Third, we used a novel electronic tool to assess adherence accurately (easypod) [14,15,16]. For this purpose we used data from children naïve to GH treatment participating in the ECOS study, using automated continuous assessment of adherence through easypod. The easypod Connect observational study (ECOS) is a 5-year, Phase IV open label study that started in 2010 in 24 countries to assess 'real-world' adherence via the easypod electronic drug delivery device. [17]
Data
Adherence data were derived from the easypod device combined with physician data entry of outcome measures. Data were collected retrospectively and prospectively. Collected data were analyzed in a multinational pooled analysis. Inclusion criteria for the ECOS study were: GH administered via the easypod electromechanical device; under 18 years of age, or over 18 without fusion of growth plates; parent's or guardian's written informed consent, given before entering data into the registry/observational study (if the child was old enough to read and write, a separate assent form was given as defined in the appropriate jurisdiction of each country). Exclusion criteria for the ECOS study were: Subjects taking GH in whom growth plates have fused (i.e. taking GH for its metabolic effects); contra-indications to Saizen GH; Use of an investigational drug or participation in another interventional clinical study. The study was conducted in accordance with the principles of the Declaration of Helsinki, Good Clinical Practice (ICH-GCP E6) guidelines and applicable national legal and regulatory requirements.
To investigate the effect of adherence on the two-year growth response to GH treatment in prepubertal children with idiopathic isolated growth hormone deficiency (GHD) in the ECOS study, the following selection criteria were applied: • Naïve to GH treatment • Patients diagnosed with Idiopathic Isolated GHD by their physicians • Known parental heights, so that TH could be calculated • Age at end of 2 year study period <10 years in girls and <12 years in boys (i.e. before onset of puberty) • At least three measurements of height available from start treatment • HSDS <-2.0 at start treatment • At least two data measurements on growth and adherence in the first and second year after start of treatment The age limit was used for two reasons. First, we wished to limit the analysis to prepubertal children, because insufficient data were available in the database about pubertal status during GH treatment. Second, beyond these age limits the SD of the population's growth reference charts shows a non-linear pattern of an increase followed by a decrease, due to the varying ages at pubertal onset in the general population. Because HSDS is the main outcome parameter in our analysis, a higher cut-off for age would generate noise and bias.
Statistical analyses
Adherence was calculated as the number of injections received divided by the number of planned injections during the considered period, expressed as a percentage. All adherence rate analyses were based on periods of complete weeks. Data are presented over the first two years.
Outcome measures for the growth response to GH treatment in the first two years were: • Change in HSDS in the first (ΔHSDS 0-1 y), second (ΔHSDS 1-2 y) and first two years (ΔHSDS 0-2 y) at start GH (mg/kg/wk)))-0.6 � (HSDS 1y-TH SDS)+0.28 � weight SDS 1y))/1.72, and IoR second year = (height velocity second year-(5.69-0.09 � Age at start GH+0.63 � (log(3 � GH dose at start GH (mg/kg/wk)))+0.24 � weight SDS 2y+0.31 � height velocity first year))/1.19 • Parameters of the monomolecular growth curve for catch-up growth [13]: Monomolecular growth curve: adj-HSDS = A � (1-B � exp(-k � x))-5, with x age after start GH • Parameters of a second degree polynomial: second degree polynomial: adj-HSDS = c + D � x + E � x 2 , with x age after start GH Mixed-effect models were used to fit the monomolecular growth curve and the second degree polynomial on adj-HSDS (HSDS minus TH-SDS). Each curve summarizes the first phase of CUG in three parameters. For the monomolecular growth curve, these parameters are A-5 = attained adj-HSDS after two years, A � (1-B)-5 = adj-HSDS at start, k = growth rate. For the second degree polynomial, these are c = intercept (adj-HSDS at start), D = slope (CUG), E = deceleration (deceleration part after CUG). A higher value of D implies a steeper curve between 0-2 years and a higher value of E implies a stronger CUG after approximately 1 year. If E >0, then there is a stronger CUG in the second year, and if E<0 then CUG is stronger in the first year.
Linear regression analyses were performed to test the association between adherence in the first, second, and the first two years and the outcome measures. Adherence is highly skewed with peak values near or at 100%. We therefore analyzed this independent variable both as a continuous variable and categorized into a high and low level. Recursive partitioning [18] in software R (package rpart) was used to find the cut-off point (the first split) for high and low adherence that maximizes the correlation between adherence and height gain.
The parameter B from the monomolecular growth curve and c from the second degree polynomial were not used in the analyses as outcome measures, because these are related to the adj-HSDS at start and only of interest as a potential confounder. We adjusted the associations for potential clinical confounders that were significantly related to adherence in the first, second or first two years. The potential clinical confounders were age at start, adj-HSDS at start (observed data, not derived from the monomolecular growth curve model), birth weight SDS, gestational age (<37 weeks vs. �37 weeks), GH dose, and GH max. The formula of IoR already adjusts for confounders. We, therefore, did not adjust the IoR for the potential clinical confounders. HSDS at exactly 1.0 and 2.0y were obtained by linear interpolation using the nearest measurements around these ages. WHO growth references were used to calculate the SDS values. These references were obtained from the WHO Multicentre Growth Reference Study (MGRS) (ages 0-5 years) and the WHO 2007 reference (5-19 years for height) [19]. TH-SDS was calculated as 0.72 x (father's HSDS + mother's HSDS)/2 [20]. This definition corrects for assortative mating and parent-offspring correlations. We also used the first and second year Index of Responsiveness (IoR) for the KIGS prediction model as outcome measures [3], but using the WHO references for weight SDS at birth. Fig 1 shows the sample size of this study after applying the inclusion criteria. In total 490 naïve patients with GHD were available within the ECOS study. The sample size of patients with an Organic GH deficiency (OGHD) due to congenital/anatomical or tumor origin was low and therefore excluded from the analysis in order to analyse a homogeneous group. In total, 95 patients with Idiopathic Isolated GH deficiency (IIGHD) were available for analysis. Table 1 shows the descriptive statistics of the background characteristics and the clinical parameters of these 95 patients. Although one patient exceeded the (arbitrary) cut-off of 10 ug/L at the GH peak after stimulation (GH peak-1 = 9.97 and GH peak -2 = 12.5 ug/L), we included this patient in our study because the treating physician interpreted this patient as GHD. Fig 2 shows the median and P25-75 adherence % in the first two years of GH treatment. In the first year, median adherence was high with relatively little variation. In the second year, median adherence decreased, while the variation increased. Recursive partitioning for adherence resulted in a split at �98% in the first year, �91%, in the second year, and �78% in the IoR second year -0. first two years. With these splits, 32 children (34%) had a high adherence (�98% as cut-off value) and 63 children (66%) a low adherence (<98% as cut-off value) in the first year. For the second year, these figures were 50 (53%) for a high adherence (�91% as cut-off value) and 45 (47%) for a low adherence (<91% as cut-off value). For the first two years, these figures were 68 (72%) for a high adherence (�78% as cut-off value) and 27 (28%) children (<78% as cut-off value). Table 2 shows the results of the effect of adherence on ΔHSDS, the parameters of the growth curves and IoR. Ten out of eighteen associations from the final model were statistically significant. The strongest associations were found between high versus low adherence in the second year and IoR 2y, between adherence (continuous, in %) and ΔHSDS 0-2y, and between high versus low adherence in the first two years and ΔHSDS 0-2y. Mean SD) IoR at 2y was 0.19 (0.99) in the high adherence group and -0.44 (1.04) in the low adherence group. Without adjustment, mean (SD) ΔHSDS 0-2y was +1.16 (0.52) in the high adherence group and +0.88 (0.41) in the low adherence group. After adjustment for the clinical confounders, the difference in mean height between the high and low adherence group increased from +0.28 SD to +0.34 SD. The size of the effect of adherence (continuous) in the first two years on HSDS 0-2y can be calculated by multiplying the model estimate by the percentage gain of adherence. For example, a 14% lower adherence (missed 1 injection/week) in the first two years is associated with 14 x 0.00792 = 0.11 SD less height gain, and a 28% lower adherence (missed 2 injections/ week) with 0.22 SDS less height gain.
Discussion
Our study is the first to show the effect of suboptimal adherence in a large group of patients with GHD using a reliable method to automatically assess adherence through an electronic device over the first two years of GH treatment (Easypod). The effect size of suboptimal adherence is dependent on the percentage of missed injections; if a cut-off of 78% is used, the loss of height gain was 0.34 SD, approximately 2-2.5 cm after two years of GH treatment.
It is noteworthy that average adherence in the first year was not associated with first year growth response, in contrast to the second year and both years combined. We believe that there are three explanations for this observation. The first is that average adherence in the first year is very high with little variation, which limits the power to detect statistically significant associations. Second, adherence has a skewed distribution, which may reduce the associations. Third, the effect of GH on growth is most prominent in the first year of treatment, with a relatively strong effect on growth velocity with relatively little effect of GH dose [3].
Our data show that in order to obtain a good insight into the effect of suboptimal adherence, two conditions have to be met. First, an outcome parameter has to be used that represents the natural shape of CUG. In the first years of treatment, CUG has the shape of a sharp increase of HSDS in the first year, followed by gradually decreasing height velocities until HSDS adjusted for TH-SDS is close to 0. HSDS remains stable until the onset of puberty, when the pubertal growth spurt begins. This implies that the choice for the outcome parameter is dependent on the number of years after start of GH treatment. In the first years, growth should be compared with a model for CUG, thereafter to a stable HSDS, and from pubertal onset the effect of adherence becomes impossible to analyze because of interference by the pubertal growth spurt. Our data show that over the first two years the change in HSDS gives similar or higher effects than the growth rate parameters of the two mathematical models. Although the cut-offs for adherence were based on the difference in HSDS, a similar cut-off for 0-2 y was found for the slope parameter D. We propose that this outcome measure may be superior to the various indicators of "poor response" reported in the literature [21].
The second condition that has to be met is that sufficient adjustment is made for clinical parameters that influence the growth response to GH treatment. In our study, adherence in the first two years strongly correlated with height gain 0-2 y, even unadjusted for clinical predictors. However, the effect of adherence became considerably stronger when adjustment was made to well established predictors.
While automatic recording of adherence is helpful to give insight into the frequency of suboptimal adherence and its effect on the growth response to GH treatment, for clinical care its most important benefit could be that it gives the clinician a signal to intervene and try to improve adherence. In theory, supportive accountability, including human support, motivational interviewing and communication "bandwidth" could be successful, and potentially successful approaches have been suggested [22,23,24]. However, controlled trials on the efficacy of such programs are lacking.
The limitations of this study include its non-interventional nature, which is associated with a high level of missing data, high inter-patient variability, and the absence of detailed recording of actions performed by health-care providers and carers when poor adherence and/or poor response to treatment was recorded. However, these limitations occur in all surveillance studies [25], whereas the observational nature means that it reflects normal clinical practice. Another limitation is that the cut-off values for defining a high and low adherence were constructed in the same dataset as the correlation analysis. This may have overestimated the correlations. However, similar conclusions could be drawn based on the adjusted correlations between ΔHSDS in the first, second and first two years and adherence in % (linear) and adherence in two categories (high versus low according to the cut-off values that we developed). Also, when we apply the Cutfield cut-off values [9] of high (�86%) versus low adherence (<86%) in the first, second and first two years, similar significance levels were found with the exception of ΔHSDS in the first two years (Adj. B (SE) = 0.23156 (0.12338), p = 0.066). The cut-off value of �86% for a high adherence was chosen by Cutfield et al. [9], because it corresponds to no more than one missed dose a week on average. However, it is to be preferred to take the cut-off value that maximizes the correlation between adherence and height gain, because it provides more insight into the doses that are needed per several days, week or several weeks to have an optimal height gain. Further research is needed to investigate if the Table 2. Results of linear regression analyses with ΔHSDS and the parameters of the growth curves as outcome and adherence as dependent variable with and without adjustment for potential clinical confounders (n = 95). performance of the constructed cut-off values in our study provides similar correlations between high/low adherence and height gain in new datasets compared to our study. Another limitation of our study is that we did not include IGF-I measurements to assess GH status and predict growth response and adherence during GH therapy in our patients. IGF-I measurements were not available for the majority of patients, because ECOS was a surveillance study [26]. For this study, we only selected adherence from the easypod electronic drug-delivery device, growth outcomes and several clinical and background parameters as potential confounders. Strengths include that it is the first study that has used a device with an eHealth platform to report adherence data directly from patients to health-care providers. A number of individual cases from ECOS have been reported [27,28,29,30]; these indicate that direct access to adherence monitoring can make the difference in a patient's management and motivation. For example, a 14% lower adherence (missed 1 injection/week) in the first two years is associated with 14 x 0.00792 = 0.11 SD less height gain, and a 28% lower adherence (missed 2 injections/ week) with 0.22 SDS less height gain.
Outcome
The success of therapy depends mainly on the ability of the patients and their parents to carefully adhere to the recommended treatment regimen. Complex schedules should be avoided. This study shows that each missed injection/week in the first two years resulted in 0.11 SD less height gain, which implies that a daily routine of taking an injection, especially in the second year, is recommended. Although accurate and up-to-date adherence data can be obtained from the easypoddevice, at the moment this data is not directly available for the physician. So complete nonadherence to GH therapy is relatively easy to detect on the basis of growth failure, but suboptimal and/or intermittent adherence are more difficult to assess. Linking the adherence data from the easypod device to the physician can be helpful in the future. At the moment, regularly interviewing the patients and their parents are efficacious means of detecting the degree of adherence [31]. Different strategies can be incorporated to enhance adherence to GH therapy, i.e. providing early patient and parent education and support, medication reminder systems and longer duration of GH prescriptions [32]. A personalized approach seems to be promising, because a previous study on the ECOS data showed that several clinical parameters and background characteristics of the patients are important determinants to predict the level of adherence. Early age of self-administration, weight at start of treatment, and teenage years are associated with a lower adherence to GH treatment [33]. Patient adherence support programs that take into account these factors may improve adherence and subsequent clinical outcomes. Further studies are required to design personalized patient support programs to attain and maintain good adherence to GH treatment.
In conclusion, electronically monitoring adherence enables obtaining reliable information on adherence in a "real-life" situation, and shows a statistically significant effect on the growth response in the first two years of treatment in children with GHD. Each missed injection per week in the first two years, resulted in 0.11 SDS less height gain. The next step should be to develop interventional tools to explore the hypothesis that patient adherence support programs improve adherence and subsequent clinical outcomes.
Supporting information S1 | 2018-11-10T06:29:28.189Z | 2018-10-24T00:00:00.000 | {
"year": 2018,
"sha1": "6d0384c0ea023e89b7c4fa4c53ea30e0bbc67d91",
"oa_license": "CCBY",
"oa_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0206009&type=printable",
"oa_status": "GOLD",
"pdf_src": "PubMedCentral",
"pdf_hash": "6d0384c0ea023e89b7c4fa4c53ea30e0bbc67d91",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
39118411 | pes2o/s2orc | v3-fos-license | Antifungal therapy and surgical drainage for the treatment of a cerebral abscess caused by in a Scedosporium apiospermum renal transplant patient-a case report
the asexual form of , Scedosporium apiospermum, Pseudallescheria boydii is a filamentous, opportunistic fungus which can be found in environmental sources all over the world. It is a human pathogen mostly associated with lung, bone and joint infections and less frequently with infections of the central nervous system (CNS). The latter is generally related to the patient’s immune state, and occurs most frequently in immunocompromised patients. We present the case of a 64-year-old male patient with a background of chronic kidney failure secondary to nephroangiosclerosis and a renal transplantation who presented with left-sided hemiplegia and dysarthria. A brain MRI revealed a hyperintense lesion with ring enhancement at the right paramedian posterior frontal subcortical area with an associated vasogenic edema. A stereotactic biopsy of the lesion revealed the presence of . The patient received a S. apiospermum combined therapy of voriconazole and terbinafine with surgical drainage, which led to temporarily clinical and radiological improvement.
Introduction
Scedosporium apiospermum is a filamentous fungus causing a rare but serious opportunistic infection.It is the asexual form of Pseudallescheria boydii and is found in many environmental sources including soil and fresh water, but most commonly in stagnant or contaminated water 1 .The infection may be acquired by inhaling the microorganism or after traumatic inoculation through the skin 2 .The sites of infection include the lungs, sinuses, bones, skin, joints and notoriously, the central nervous system (CNS) 3 .S. apiospermum can infect the CNS of both healthy 4 and immunocompromised hosts.Cerebral abscess is the most common clinical manifestation of S. apiospermum brain infections, although cases of meningitis and, less frequently, ventriculitis have also been reported 5 .Brain abscesses may be found as one or multiple lesions 6 .The overall mortality rate of patients infected with this pathogen is higher than 70% 7 .Solid organ transplant and its associated immunosuppression are important risk factors for infections with Scedosporium species 8 .Here we present a case of CNS infection caused by S. apiospermum in a patient who had received a kidney transplant and was treated with dual antifungal therapy and surgical drainage.The patient initially responded well to the therapy.
Clinical case
A 64 year-old male patient underwent deceased-donor kidney transplantation following a chronic kidney failure secondary to nephroangiosclerosis.The past medical history was significant for hypertension, hyperlipidemia, peripheral vascular disease, chronic anemia and deep venous thrombosis of right lower extremity.Family history was significant for cardiomyopathy in one brother and diabetes in another.The immunosuppressive medication consisted of tacrolimus 3 mg every 12 hours, prednisone 20 mg daily and mycophenolate mofetil 500 mg every 12 hours.Seventeen days after transplantation, he presented left-sided hemiplegia and dysarthria.A brain MRI was performed, which revealed a hyperintense lesion with ring enhancement at the right paramedian posterior frontal subcortical area with an associated vasogenic edema (Figure 1).
A stereotactic biopsy was performed and tissue examination revealed the presence of a filamentous fungus that was identified as S. apiospermum (Figure 2).The sample was also cultured in Sabouraud's dextrose agar medium at 25°C for a period of 14 days (Figure 3), the culture turned a dark brown color.
The patient started a treatment with voriconazole (6 mg/Kg po q12h for two days, then 4 mg/Kg q12h VO) and terbinafine (250 mg po daily) and subsequently was subjected to surgical drainage by craniotomy in order to remove the infected tissue.
The patient was on terbinafine for 4 months and continued to be on voriconazole for almost a year.At a follow up visit he showed significant recovery from the left-sided palsy and also an absence of dysarthria.The brain MRI follow-up images showed an improvement in the brain lesion.
Unfortunately, 8 months later the patient clinical course was complicated and he eventually died of problems unrelated to fungal CNS disease.
Discussion
Solid organ transplant recipients are highly susceptible to invasive fungal infections 9 .
During the last few decades there has been a marked increase in the number of immunocompromised patients who have suffered Scedosporium infections, the most frequent cases being infections of the CNS 10 .
Solid organ transplant patients are susceptible to invasive fungal infections as their immunity might be compromised due to the use of immunosuppressant drugs 11 .Therefore S. apiospermum should be considered in the differential diagnosis of immunocompromised patients presenting with a brain abscess 12 .Within the nervous system, abscesses may be located in brain hemispheres, the cerebellum, the brain stem or the spinal cord, where they may cause alterations of consciousness levels, signs of meningeal irritation or focal neurological deficits 13 .
If not adequately treated, fungal brain abscesses in immunocompromised patients often result in poor prognosis.The diagnosis of an invasive fungal infection such as S. apiospermum is based on the combination of histopathological, microbiological and clinical findings 14 .As the clinical and histopathological presentations of S. apiospermum infections are similar to those of other fungi such as Aspergillus and Fusarium spp., a culture is necessary for accurate diagnosis.Furthermore, while most species of Aspergillus (except for Aspergillus terreus) are sensitive to amphotericin, S. apiospermum is usually resistant 15 .In addition, PCR techniques are important to diagnose as well as to distinguish between different species 16 .
There are many treatment options described in the literature, but there is an ongoing controversy over which treatment is most suitable for S. apiospermum infections.Voriconazole has emerged as a possible treatment option, since it shows high activity against several species of fungi, including S. apiospermum 17 .Several reports have shown the successful use of voriconazole synergistically combined with terbinafine against S. apiospermum 18 and S. prolificans.Furthermore, many authors recommend the surgical drainage of brain abscesses caused by S. apiospermum 19 .Therefore, a combined antifungal therapy along with an aggressive surgical approach is recommended for therapeutic success 20 .
Conclusion
S. apiospermum infection of the CNS is a rare but it is an extremely serious medical condition.Immediate diagnosis in the event of brain abscess in an immunocompromised patient is crucial and the choice of a suitable medical treatment is a priority.Despite the aggressive surgical treatment and the appropriate anti-fungal therapy used, mortality rates continue to be high.
Open Peer Review
Current Referee Status: No competing interests were disclosed.Competing Interests:
Figure 3 .
Figure 3. S. apiospermum culture in Sabouraud's dextrose agar medium.The culture grew in Sabouraud's dextrose agar medium at 25°C and turned a dark brown color on the 14 th day.
Figure 1 .
Figure 1.Brain MRI T2-weighted image.A hyperintense lesion with ring enhancement at the right paramedian posterior frontal subcortical area with an associated vasogenic edema is shown.
of Pediatrics, Aristotle University of Thessaloniki, Thessaloniki, Greece This is an interesting rare case of brain abscess due to Scedosporium apiospermum in a kidney transplant immunocompromised patient, indeed relatively early (17 days) post-transplantation.The report reads nicely and the Discussion is thoughtful.However, there are a few points that the authors should take into consideration:Have the authors performed an antifungal susceptibility testing?In addition, any molecular identification?It is apiospermum complex now.Please, add ESCMID Guidelines for rare fungi as a reference.This is a document that should be definitely cited and used in clinical practice.Tortorano AM Clin Microbiol Infect.2014 Apr;20 et al.Suppl 3:27-46.Page 2. VO should be po I have read this submission.I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above. | 2017-08-05T12:05:36.793Z | 2014-03-13T00:00:00.000 | {
"year": 2014,
"sha1": "db51396672756f057d11ddb274063a7cd69969e5",
"oa_license": "CCBY",
"oa_url": "https://f1000research.com/articles/3-70/v1/pdf",
"oa_status": "GOLD",
"pdf_src": "ScienceParseMerged",
"pdf_hash": "5db7af7f67bdca98e2ab364661e4dab7cb516aca",
"s2fieldsofstudy": [
"Medicine"
],
"extfieldsofstudy": [
"Medicine"
]
} |
20937501 | pes2o/s2orc | v3-fos-license | Wsp1 Is Downstream of Cin1 and Regulates Vesicle Transport and Actin Cytoskeleton as an Effector of Cdc42 and Rac1 in Cryptococcus neoformans
ABSTRACT Human Wiskott-Aldrich syndrome protein (WASP) is a scaffold linking upstream signals to the actin cytoskeleton. In response to intersectin ITSN1 and Rho GTPase Cdc42, WASP activates the Arp2/3 complex to promote actin polymerization. The human pathogen Cryptococcus neoformans contains the ITSN1 homolog Cin1 and the WASP homolog Wsp1, which share more homology with human proteins than those of other fungi. Here we demonstrate that Cin1, Cdc42/Rac1, and Wsp1 function in an effector pathway similar to that of mammalian models. In the cin1 mutant, expression of the autoactivated Wsp1-B-GBD allele partially suppressed the mutant defect in endocytosis, and expression of the constitutively active CDC42Q61L allele restored normal actin cytoskeleton structures. Similar phenotypic suppression can be obtained by the expression of a Cdc42-green fluorescent protein (GFP)-Wsp1 fusion protein. In addition, Rac1, which was found to exhibit a role in early endocytosis, activates Wsp1 to regulate vacuole fusion. Rac1 interacted with Wsp1 and depended on Wsp1 for its vacuolar membrane localization. Expression of the Wsp1-B-GBD allele restored vacuolar membrane fusion in the rac1 mutant. Collectively, our studies suggest novel ways in which this pathogenic fungus has adapted conserved signaling pathways to control vesicle transport and actin organization, likely benefiting survival within infected hosts.
tion, and the cdc42 mutant strain exhibits a defect in actin repolarization under thermal stress conditions (3).
Mammalian Rac proteins, which share overlapping and complementary functions with Cdc42, regulate mobility and other phenotypes relating to membrane trafficking and the actin cytoskeleton (1,53). In filamentous fungi, such as Penicillium marneffei, the Rac homolog CflB plays a role in hyphal growth and actin organization (6). In the plant-pathogenic fungus Ustilago maydis, disruption of the RAC1 gene affects cell morphology and hyphal growth in the haploid cell (26). Additionally, U. maydis Rac1 plays a role in vacuole morphology, as evidenced by the facts that the rac1 cells had fragmented vacuoles and the expression of a constitutively active RAC1 Q61L allele resulted in growth arrest and in swollen cells containing a single large vacuole (26). Mammalian and plant Rac proteins are thought to function in actin organization through effector WAVE (WASP family verprolin-homologous) proteins, which lack both the WASP homology 1 (WH1) domain and the GBD (9,33,52). At least one Rac homolog, Rac1, has been discovered in C. neoformans. In contrast to Cdc42 and Cdc420, Rac1 appears to affect hyphal differentiation downstream of Ras1, with only a modest effect on high-temperature growth (44). No proteins homologous to WAVE proteins have been found in C. neoformans or any other fungi, suggesting a distinct effector(s) for Rac in fungal actin organization.
We have demonstrated previously that the Cin1 and Wsp1 proteins regulate intracellular trafficking and that both proteins interact with Cdc42 through conserved domains (38,39). Here we present evidence for a Cin1-Cdc42-Wsp1 effector pathway. We also found that Rac1 regulates endocytosis and is required for normal vacuole morphogenesis. A novel Rac1-Wsp1 effector pathway for vacuole morphogenesis is also presented.
MATERIALS AND METHODS
Strains, plasmids, and media. Cryptococcus neoformans var. neoformans (serotype D) strain JEC21 was used as the parental strain. All other strains used are listed in Table S1 in the supplemental material. Important oligonucleotide primers for PCR amplification and DNA plasmids are listed in Tables S2 and S3 in the supplemental material, respectively. All culture media and reagents were prepared as described previously (38,39).
Mutant allele and strain construction. Gene knockout alleles were generated by the split-marker method, and the resulting mutant strains were obtained as described previously (38,39). Auxotrophic and dominant selection marker genes, including URA5 (encoding orotidine monophosphate pyrophosphorylase), HPT (hygromycin phosphate transferase), NEO (neomycin phosphotransferase), and NAT (nourseothricin N-acetyltransferase) were all used in this study. To obtain the rac1 mutant, the 5= and 3= fragments were first amplified with primers PW1197 and PW1200 and with primers PW1198 and PW1201, respectively. The two fragments were digested with BamHI and HindIII and were joined with either the NEO or the HPT gene cassette. The two fragments, amplified with primers PW1196 and PW638 and with primers PW812 and PW1199, respectively, were used to generate the rac1 mutant. The same splitmarker approach was also employed to create wsp1 cdc42 and wsp1 rac1 strains in the cdc42 and rac1 mutants by use of alternate drug resistance markers. A cdc42 rac1 mutant strain was also produced by transforming the rac1 mutant with the cdc42::NAT allele. The WSP1-GBD allele linked to the NAT marker was constructed by PCR overlap (11). The first fragment was amplified with primers PW1265 and PW1188, with pGS778 as the template, and the second fragment was amplified with primers PW1189 and PW1399, with the same plasmid as the template. WSP1-GBD was then amplified with primers PW1266 and PW968 by using the two fragments mentioned above as the template and was inserted into the TOPO TA vector (Invitrogen), resulting in pGS783. pGS783 was then digested with BamHI and XbaI, and the insert was cloned into plasmid pGS324 to generate pGS794.
A similar approach was used to generate a WSP1-B-GBD allele linked to the NEO marker. Fragment 1 was amplified with the M13 forward primer and primer PW1455, with pGS462 as the template. With pGS783 as the template, and with primers PW1266 and PW1167, the target fragment was obtained and was inserted into the TA vector (pGS865), which was then digested with BamHI and XbaI and inserted into pGS277, resulting in pGS1228.
The PCR overlap extension technique was also utilized to generate a CDC42 Q61L allele. The first fragment was amplified with primer PW1499 and reverse primer M13, and the second fragment was amplified with primer PW1498 and forward primer M13 by using pGS950 as the template. The CDC42 Q61L allele was amplified with primers PW1495 and PW1496, cloned into the TOPO TA vector (Invitrogen), and digested with BamHI and BglII to release the CDC42 Q61L allele, which was then inserted into plasmid pGS277, generating pGS1219. To link the pleckstrin homology (PH) domain to CDC42 Q61L , the PH domain was amplified from pGS931 (containing partial CIN1 cDNA [38]) with primers PW1551 and PW1553. The fragment was TA cloned (pGS1261), released by BamHI and XbaI restriction, and inserted into pGS325 to yield pGS1414. The CDC42 Q61L allele with a native CDC42 terminator was released from pGS1403 by BamHI and SpeI restriction and was inserted into pGS1414, yielding pGS1443. pGS690, containing a P GPD1 -Dbl homology (DH) fusion gene, was transformed into the wsp1 mutant (GYS237), resulting in strain GYS566 (see Table S1 in the supplemental material). The rac1 mutants expressing the WSP1-GBD and WSP1-B-GBD alleles were obtained by transforming the mutant with pGS794 and pGS875, respectively. The CDC42-GFP-WSP1 fusion gene was generated by PCR linking CDC42 to GFP-WSP1 by use of plasmids pGS1186, pGS1217, and pGS1236.
The promoter sequence from C. neoformans glyceraldehyde phosphate dehydrogenase was used in most gene fusion or truncation constructs unless stated otherwise.
Fluorescence-tagged protein construction. The green fluorescent protein (GFP)-Cin1 fusion protein was constructed as described previously (38). To produce a GFP-Wsp1 fusion protein, the GFP gene was digested from pGS720 with BamHI and BglII and was fused with the WSP1 gene, which was digested from plasmid pGS869 with BamHI and NotI. The GFP-WSP1 fusion gene was inserted into plasmid pGS324, resulting in pGS1251. The DsRED-WSP1 fusion gene linked to the hygromycin resistance marker gene was constructed similarly, by use of plasmids pGS810, pGS821, and pGS1481. For construction of the DsRED-CDC42 fusion gene, CDC42 was restricted from pGS950 with BamHI and SpeI and was inserted into pGS988, resulting in pGS1005, which contains the NAT marker. The DsRED-RAC1 fusion gene was constructed in a similar manner by use of pGS1011, pGS1035, and pGS1058. Fusion gene constructs were transformed into the respective strains through biolistic transformation, and expression was examined by the ability of each fusion gene to suppress the corresponding mutant's phenotype.
Yeast two-hybrid interaction. An interaction between Cin1 and Wsp1 was previously established using a yeast two-hybrid system (Invitrogen) for Cin1 and Wsp1 (38,39). RAC1 cDNA was digested with EcoRI and BglII from pGS1011 and was then inserted into pGADT7. The gene sequence was verified by DNA sequencing. A yeast strain (Gold) from Invitrogen was used as the host strain, and gene transformation was carried out according to the protocol provided by the manufacturer. Vacuolar fusion following hypotonic stress. To induce vacuole fusion in C. neoformans, we adapted the hypotonic stress method used for S. cerevisiae (20,48). Cells grown in yeast extract-peptone-dextrose (YPD) were stained with FM4-64 either directly or after 20 min of hypotonic treatment (10-fold dilution of YPD in water) and were examined microscopically at the time points indicated in the figure legends.
Microscopy. Cryptococcal cells were prepared and stained with rhodamine-conjugated phalloidin as described previously (38,39). For FM4-64 staining, cells grown in liquid YPD for 14 to 16 h (mid-to latelogarithmic phase) were harvested by centrifugation at 700 ϫ g for 1 min and were resuspended in ice-cold Hanks' buffered salt solution (HBSS; Invitrogen). FM4-64 was added to the cell suspension at a final concentration of 5 g/ml.
To observe fluorescent fusion proteins, cells grown for 12 to 14 h in YPD were harvested and were washed twice with sterile distilled water. Cells were also resuspended in water for subsequent microscopic observations.
Optical and fluorescence microscopic observations were carried out on a Zeiss fluorescence microscope (Axio Imager 2) with a 63ϫ (numerical aperture, 1.4) Plan-Apochromat oil immersion lens (Zeiss). Images were captured with an AxioCam camera (Zeiss) and were adjusted for brightness and contrast only by using Adobe Photoshop CS4.
Wsp1 coresides with and functions downstream of Cin1 in endocytosis.
We have shown previously that both Cin1 and Wsp1 regulate intracellular trafficking and the actin cytoskeleton, which underlie the growth, differentiation, and virulence of C. neoformans (38,39). Cin1 bears a resemblance to the human intersectin protein ITSN, which functions in conjunction with WASP and Rho GTPase Cdc42 protein to regulate intracellular trafficking and maintain the actin cytoskeleton (30). To test the hypothesis that the Cin1, Wsp1, and Cdc42 proteins may function analogously, we examined epistatic relationships among the three proteins. We first validated the previously established interaction between Cin1 and Wsp1 by in vivo colocalization of fluorescencelabeled GFP-Cin1 and DsRed-Wsp1 fusion proteins. Cin1 fluorescence appeared mostly as punctae near the plasma membrane ( Fig. 1A, green), and DsRed-Wsp1 was similarly present as punctae near the plasma membrane. Minimal red and green fluorescence was also observed in the cytosol (Fig. 1A, green and red). Overlay of the two images revealed colocalization of the red and green fluorescent signals, suggesting that Cin1 and Wsp1 occupy similar subcellular domains and could interact with each other in vivo (Fig. 1A, overlay).
In wild-type C. neoformans strains, endocytic uptake of FM4-64 occurs immediately upon exposure to the lipophilic dye under ambient temperatures. FM4-64 is concentrated mostly at the plasma membranes of cells kept on ice (Fig. 1B). Incorporation of this dye into internal ring-like structures, which are early endosomes and endosomes, became apparent within 30 min at room temperature (Fig. 1B). In agreement with previous findings, FM4-64 remained in the plasma membrane and was not internalized at 30 min in either the cin1 or the wsp1 mutant strain (Fig. 1B, upper row), indicating that Cin1 and Wsp1 are similarly required for endocytosis (38,39).
To examine the genetic relationship between Cin1 and Wsp1, we utilized a Wsp1 mutant allele lacking the B domain and the GBD (WSP1-B-GBD). Previous studies of mammalian WASPs indicated that deletion of the B domain and the GBD relieves autoinhibition and results in constitutive WASP activation. The WSP1-B-GBD allele was introduced into the cin1 mutant, with the wild-type WSP1 gene as a control. The cin1 strain overexpressing wild-type WSP1 demonstrated no change in the FM4-64 endocytosis defect (Fig. 1C). In contrast, the cin1 mutant expressing Wsp1-B-GBD showed an FM4-64 staining pattern distinct from that of the cin1 mutant expressing the wild-type Wsp1 allele. FM4-64 was observed both in the plasma membrane and inside the cells as random aggregates at time zero and at 15 and 30 min ( Fig. 1). Staining of the aggregates became progressively intensified as the exposure time increased, and at 60 min, ring-like structures resembling endosomes were observed, indicating partial suppression of the endocytosis defect by autoactivated Wsp1. The incomplete suppression of the endocytosis defect and the failed suppression of the cin1 morphology defect indicated partial phenotypic complementation of the cin1 mutant by the WSP1-B-GBD allele (Fig. 1C).
Cdc42 is an effector of Cin1 in actin cytoskeleton organization. The actin cytoskeleton consists mainly of short-lived actin cables and actin patches. In S. cerevisiae, the actin cable consists of filamentous actin (F-actin), whereas the actin patch is a cortical membrane zone composed of F-actin and actin-binding/regulatory proteins, such as capping proteins (Cap1 and Cap2), actin nucleation proteins (Arp2/3), and regulators (Las17/Bee1) (15,29,54). The actin patch is also enriched at the sites of active polarized cell surface growth, such as the bud, the site of cell-cell separation, and the mating projection (35,40).
Previous studies suggested that C. neoformans maintains actin dynamics similar to those of model yeasts (21,47). In wild-type strains, F-actin is present in polarized patches enriched in the early, emerging daughter cell and at the contractile ring ( Fig. 2A). Consistent with their morphology and cell separation defects, the cin1 and wsp1 mutants display a less defined pattern of F-actin localization ( Fig. 2A). In contrast, the cdc42 mutant has intact actin patches at permissive growth temperatures (30°C), although actin filaments are less apparent in this strain than in the wild type ( Fig. 2A).
To investigate whether Cdc42 plays a conserved role in actin organization by functioning downstream of Cin1 but upstream of Wsp1, we performed the following analysis. A constitutively active Cdc42 allele, CDC42 Q61L , was generated and transformed into the cin1 mutant strain, with the wild type CDC42 allele as a control. The expression of CDC42 Q61L resulted in the formation of abundant actin cables and some actin patches in the cin1 mutant ( Fig. 2A). In contrast, the cin1 mutant expressing the wild type CDC42 allele remained defective in actin cable and patch formation ( Fig. 2A). This finding is consistent with the hypothesis that Cdc42 is an effector of Cin1 in actin organization, particularly actin cable formation.
To examine the relationship between Cdc42 and Wsp1, we expressed the DH (Rho-guanine nucleotide exchange factor [Rho-GEF]) domain of Cin1, which has been shown to have a GEF function in vitro similar to that of Cdc24 in the wsp1 mutant (38). Expression of the Cin1-DH domain restored near-normal actin cable and patch formation to the wsp1 mutant ( Fig. 2A). Interestingly, the constitutively active Cdc42 Q61L protein linked to the PH domain of Cin1, which is thought to promote membrane tethering and Cdc42 activity (16,37), also restored normal actin cable formation and partial actin patch formation to the wsp1 mutant ( Fig. 2A). These findings indicated that Cin1 and Cdc42 do not act exclusively on Wsp1 and that additional effectors of C. neoformans Cdc42 might be present.
Wsp1 interacts with Cdc42 to regulate the actin cytoskeleton downstream of Cin1. Cdc42 was previously shown to interact with Wsp1 in a yeast two-hybrid assay (39). To explore this interaction, DsRed-Cdc42 and GFP-Wsp1 fusion proteins were expressed in a wsp1 cdc42 double mutant, and the patterns of fluorescence were observed. GFP-Wsp1 was localized primarily in punctae or peripheral patches adjacent to the plasma membrane, with some cytoplasmic signals (Fig. 2B). In contrast, DsRed-Cdc42 had more continuous distribution in the plasma membrane and the cytosol (Fig. 2B). Overlay of the two images suggested that the two proteins share certain overlapping distributions (shown as yellow fluorescence) within live cells (Fig. 2B).
Moreover, to further examine the interaction between Wsp1 and Cdc42 and the relevance of that interaction to Cin1, we fused Cdc42 with Wsp1 through a GFP linker. generating a Cdc42-GFP-Wsp1 protein. Overexpression of this fusion protein partially re-stored the growth phenotype and actin cables and patches of the cin1 mutant ( Fig. 2A). Neither Cdc42 nor Wsp1 alone could suppress the phenotype of the cin1 mutant ( Fig. 2A). This resulted in an intriguing proposition: either the Cdc42-GFP-Wsp1construct altered the autoinhibitory conformation of Wsp1, resulting in its activation, or Cdc42 could activate Wsp1 through direct association. Regardless, the finding is consistent with established models staining of the indicated strains. All strains were incubated to the mid-and late-logarithmic phases in YPD prior to incubation in HBSS. WT, wild type. In the cin1 mutant background, the following alleles were expressed using the GPD promoter: the wild-type CDC42 gene, the dominant active CDC42 Q61L allele, the wild-type WSP1 gene, and the CDC42-WSP1 fusion allele. In the wsp1 background, the following genes/domains were overexpressed: the DH domain of the CIN1 gene and the PH domain of CIN1 coexpressed with the CDC42 Q61L allele. DIC, differential interference contrast. (B) GFP-Wsp1 and DsRed-Cdc42 were coexpressed in the wsp1 cdc42 mutant strain. An overlay of the two images suggests that Wsp1 and Cdc42 occupy shared subcellular domains in vivo.
in which Wsp1 is a direct downstream effector of Cdc42 in actin organization.
Wsp1 is required for the localization of Rac1 at the vacuole membrane. Taken together, the cumulative data provide evidence that Wsp1 is a conserved effector of Cdc42. However, the wsp1 mutant lacks both actin patches and actin cables, while the cdc42 mutant appears to have normal actin patches but not actin cables (Fig. 3A), suggesting that the role of Wsp1 in actin patches could be activated by other regulators, such as Rac proteins of the Rho GTPase family. The cryptococcal Rac homolog Rac1 was previously shown to exhibit similar and distinct functions with Cdc42 in the regulation of filamentation and thermal resistance; its other roles were unknown (44). To determine whether Rac1 has any roles in vesicle transport and the actin cytoskeleton, we characterized the function of Rac1 in endocytosis, observed the localization of the DsRed-Rac1 fusion protein, and examined the interaction between Rac1 and Wsp1.
In most cells, the rac1 mutant strain exhibited a normal distribution of actin patches and cables, in contrast to the cdc42 mutant strain, which had normal actin patches but not cables, while the wild-type strain had normal actin patches and cables (Fig. 3A). Intriguingly, no definitive actin patches or cables were observed in the rac1 cdc42 mutant strain, indicating that Rac1 and Cdc42 collectively maintain the actin cytoskeleton (Fig. 3A).
The DsRed-Rac1 fusion protein was enriched in the vacuole membrane: fluorescence occurring as near-continuous rings surrounding the vacuoles was very apparent (Fig. 3B, top). In addition, this distribution was dependent on Wsp1, as evidenced by the fact that in the wsp1 mutant, the DsRed-Rac1 fusion protein no longer appeared in the vacuole membrane, occurring instead as small cytoplasmic punctae (Fig. 3B, bottom). The wsp1 mutant showed a defect in vacuole formation, indicating that Wsp1 is essential for vacuolar biogenesis.
Consistent with these findings, Rac1 interacted with Wsp1 in a yeast two-hybrid screen, as evidenced by the fact that robust cell growth occurred on a stringent selective medium (synthetic dropout medium [SD]-Leu-Trp-His-Ade plus 1 mM 3-aminotriazole ) in the yeast strain cotransformed with the Rac1-and Wsp1-containing vectors (Fig. 3C). Moreover, the distribution of a GFP-Wsp1 fusion protein partially overlapped that of DsRed-Rac1 along the vacuolar membrane in the rac1 wsp1 strain background, indicating possible colocalization of these two proteins in vivo (Fig. 3D).
Wsp1 is an effector of Rac1 in vacuolar morphogenesis. To examine the role of Rac1 in intracellular transport, we examined the rac1 mutant for possible defects in endocytosis and/or vacuole morphology. In comparison to the wild-type strain, which displayed normal endocytosis, and the wsp1 mutant strain, with a severe defect in endocytosis, the rac1 mutant strain showed an intermediate phenotype. Immediately after exposure of the rac1 mutant to FM4-64, the dye was concentrated mostly in the plasma membrane, as in the wild-type and wsp1 strains (Fig. 4A). In addition, bright particulate staining occurred inside the periphery of the plasma membrane in the rac1 and wsp1 mutants. After 15 min of exposure, small, ring-like endosomal structures located near the plasma membrane were observed in some of the rac1 mutant cells. Some of the rings were not complete, indicating that the endosomes had not yet completed invagination or scission from the plasma membrane (Fig. 4A, arrows). At 30 min, some of the FM4-64-stained endosomes/vesicles were internalized, but others remained near the plasma membrane (Fig. 4A, arrows). In contrast, the wild-type strain displayed more-complete internalization of the fluorescent dye at 15 and 30 min. As expected, similar endosomes were not found in the wsp1 strain (Fig. 4A) (39). This finding suggests that Rac1 plays an important role in early steps of endocytosis.
To study the functional consequence of the interaction between Rac1 and Wsp1, we observed vacuole morphology by stain-ing cells of various genetic backgrounds with FM4-64 following hypotonic stress, which was created by switching cells from rich YPD to hypotonic conditions (YPD diluted 10-fold with sterile distilled water). The wild-type strain exhibited a pattern of normal staining by FM4-64: multiple endosomes/vesicles were seen following 30 min of exposure (Fig. 4B). After 20 min of hypotonic stress followed by 5 min of exposure to FM4-64, most of the wildtype cells showed a large, predominant vesicle (vacuole), likely due to membrane fusion of smaller vacuoles (Fig. 4B, WT; fused vacuoles indicated by arrows). The late-appearing and small vacuole structures of the rac1 mutant exhibited no such transition after hypotonic stress (Fig. 4B, rac1). However, expression of the autoactivated Wsp1 (Wsp1-GBD and Wsp1-B-GBD) alleles resulted in suppression of the rac1 defect in vacuole fusion to a large extent. As a result, a single large vacuole occurred in cells after hypotonic stress (Fig. 4B, rac1 WSP1-GBD and rac1 WSP1-B-GBD, arrows). In addition, the effect of the WSP1-B-GBD allele appeared to be greater than that of the WSP1-GBD allele, since the strain expressing Wsp1-B-GBD more closely resembled the wildtype strain (Fig. 4B). This evidence supports a model in which Wsp1 is a novel effector of Rac1 in vacuole morphology and fusion.
DISCUSSION
Intracellular trafficking is an essential cellular process linked to diverse downstream effects. In mammalian cells, these effects include cell surface protrusions and the migration, uptake, and propulsion of macromolecules (12). In S. cerevisiae, C. albicans, and C. neoformans, intracellular transport underlies cellular growth, polarity establishment, and/or virulence (reviewed in reference 47). There is evidence indicating the presence of distinct endocytic machinery in C. neoformans and likely in other basidiomycetous fungi (38,39). Given that intracellular trafficking is a conserved process, studies focused on interactions among the Cin1, Wsp1, Cdc42, and Rac1 proteins should reveal the distinction of the Cin1-mediated pathway(s).
Wsp1 as a conserved target of Cin1 in actin organization. Intracellular trafficking in S. cerevisiae is a highly organized event involving orchestrated functions of many proteins, including the Pan1 endocytic protein, WASP-like Las17/Bee1, Cdc42, and other proteins that link the endocytic cycle to the actin cytoskeleton (14,43,51). Pan1 regulates endocytosis and actin organization through its two N-terminal EH (Eps15 homology) domains and one C-terminal A domain. The EH domains bind the SlaI, End3, Ent1/2, and Yap1801/2 proteins (5,43,50), while the A domain interacts with the Arp2/3 complex. Pan1 does not interact with Las17/Bee1, which lacks the GBD, but interacts through the inhibitory protein SlaI to activate the Arp2/3 complex (13,27,42,50).
We have proposed that C. neoformans evolves an endocytic pathway more similar to the human ITSN1 pathway than that of S. cerevisiae Pan1. In addition to Cin1, which contains one N-terminal EH domain and the coiled-coil region, C. neoformans also contains a WASP homology 2 (WH2) domain, two SRC homology 3 (SH3) domains, and a C-terminal DH-PH domain (38). C. neoformans also contains a WASP homolog, Wsp1, which has a GBD found only in mammalian WASPs. Given that mammalian WASP is a conserved functional partner of ITSN1 in actin regulation, similar domains in Cin1 and Wsp1 might suggest that certain signaling mechanisms are analogous between this fungal pathogen and mammals. strains at 0, 15, and 30 min. In contrast to the wild-type strain, which showed early-appearing endosomes, and the wsp1 mutant, which exhibited none, the rac1 strain showed partial endosome formation at 15 min and more mature endosome formation 30 min after exposure to FM4-64. The numbers of endosomes in the rac1 mutant (indicated by arrows) were lower, and the sizes were smaller, than those in the wild type. DIC, differential interference contrast. (B) Rac1 affected Wsp1-mediated vacuole morphology. Hypotonic stress-induced vacuole fusion was observed in the wild-type strain (arrows) but not in the rac1 mutant. Expression of the autoactivated Wsp1-GBD and Wsp1-B-GBD proteins restored at least partial vacuole fusion in the rac1 mutant (arrows). Hypotonic stress was induced by switching cells to diluted YPD medium (10-fold dilution with sterile distilled water) for 20 min. Cells were photographed 5 min after exposure to FM4-64.
Here we performed protein localization studies and genetic epistasis analysis that support functional interactions between C. neoformans Cin1 and Wasp1 in endocytic trafficking and morphogenesis. Cin1 and Wsp1 likely operate in a common signaling pathway, given shared phenotypic defects in wsp1 and cin1 mutant strains (39). Partial suppression of the cin1 defect in endocytosis by the autoactivated Wsp1 allele further supports a genetic relationship between the two proteins. The partial suppression also suggests that additional activation mechanisms for Wsp1 and/or additional targets of Cin1 in endocytosis might be present.
Wsp1 as one of the conserved effectors of Cdc42 in actin organization. The wsp1 mutant strain failed to display normal actin patches and cables, suggesting a role in the regulation of the actin cytoskeleton. The finding that the autoactivated Wsp1-B-GBD allele partially suppressed the cin1 defect in endocytosis further indicates that Wsp1 functions in a manner similar to that of mammalian WASP and that the GBD indeed contributes to the autoinhibition of Wsp1. Wsp1 interacts with Cdc42 in vitro, and this interaction was corroborated through partial colocalization of the fluorescence-labeled Wsp1 and Cdc42 fusion proteins. Importantly, Wsp1 is required for Cdc42 distribution, indicating that a feedback mechanism likely involving Cin1 is present. In a mechanism still not fully understood, Cdc42-GFP-Wsp1 partially suppressed the cin1 defect in the actin cytoskeleton. Such findings are consistent with a proposed role for Wsp1 as an effector of Cdc42 and Cin1. This is an unprecedented finding, since other known fungal WASP homolog proteins, including S. cerevisiae Las17 and C. albicans Wal1, lack the GBD (24,46) and thus are unlikely to mediate Cdc42 function through similar mechanisms.
Like mammalian Cdc42, which regulates many effects relating to cytoskeleton organization and morphogenesis, fungal Cdc42 has a wide range of functions, ranging from polarity establishment, hyphal differentiation, and cytokinesis to virulence (10,17,32). Distinctively, Cdc42 regulates not only polarity establishment but also thermal tolerance, affecting virulence in C. neoformans. While Cdc42 regulates thermal tolerance downstream of Ras1, its role in polarity establishment might be related to the fact that the cdc42 mutant lacks actin cables. Expression of the constitutively activated Cdc42 Q61l protein was able to restore both actin patches and cables to the wsp1 strain, suggesting that additional effectors of Cdc42 in actin organization might exist, such as formins, which are also involved in the polymerization of actin.
Wsp1 as a novel effector of Rac1 in vacuole morphology. Wsp1 is one of the conserved effectors of Cdc42 in the actin cytoskeleton. The wsp1 mutant exhibits a more severe defect in actin organization than the cdc42 mutant, suggesting that additional regulators of Wsp1 might be present. To identify such regulators, we focused on Rac1, which shows similarity to Cdc42 in its general structure. Rac1 was previously determined to function in hyphal differentiation and high-temperature growth downstream of Ras1 (44). Rac proteins exhibit a wide range of functions in fungi, with the exception of S. cerevisiae, which does not have Rac homologs. The Penicillium marneffei Rac homolog CflB plays a role in hyphal growth and actin organization (6). C. albicans Rac1 has an important role in filamentous growth in cells inside a matrix (4). U. maydis rac1 mutant cells have a defect in vacuole morphology, and expression of a constitutively active RAC1 Q61L allele caused enlarged cells with a single large vacuole (26). Our results corroborated the role of Rac1 in vacuolar morphogenesis. Moreover, we provided the first evidence that Rac1 regulates early endocytosis as well as vacuolar membrane fusion. Earlier studies with mammals have indicated that Rac proteins can bind to and activate WASP (2,41). Through several lines of genetic evidence, we similarly concluded that C. neoformans Rac1 is a novel regulator of Wsp1. First, we demonstrated that Rac1 interacts with Wsp1 in vitro and in vivo. Second, Rac1 and Cdc42 are both required for normal actin cytoskeleton structures. Third, Rac1 enrichment in the vacuole membrane requires a functional Wsp1. Lastly, autoactivated Wsp1 suppresses the defect of the rac1 mutant in vacuole morphology upon induction of vacuolar membrane fusion.
Suppression of the vacuole fusion defect in the rac1 mutant by the Wsp1-GBD and Wsp1-B-GBD alleles was only partial, like the suppression of the cin1 defect by the Wsp1-B-GBD allele. This finding could suggest that additional effectors of Rac1, such as Cin1, might exist and that multiple proteins participate in the regulation of vesicle transport and the actin cytoskeleton. Finally, since Wsp1 is the effector of both Cdc42 and Rac1 in vesicle transport and the actin cytoskeleton, loss of Wsp1 function also negatively affects the distribution, and hence the function, of Cdc42 and Rac1. This may suggest that feedback mechanisms are present, such as that involving Cin1 and Cdc42-Wsp1. No evidence has yet been found to indicate whether the DH domain acts as a GEF for Rac1, but there are Rac-specific GEF homologs, such as DOCK1 (GenBank accession no. XP_573028), in the C. neoformans genome. These proteins remain uncharacterized.
On the basis of these findings, we propose a working model highlighting the interactions between Cin1, Wsp1, Cdc42, and Rac1 and the possible functional consequences of these interactions (summarized in Fig. 5). In this model, Cin1 activates Cdc42 through the DH (Rho-GEF) domain, which, in turn, activates Wsp1 to promote actin polymerization, both of which are required for endocytosis, exocytosis, and other intracellular transport events. In addition to Cdc42, Wsp1 can also be activated by Rac1 for endocytosis and vacuole mor-FIG 5 Proposed model for the Cin1-mediated endocytic pathway, which regulates actin organization through the collective functions of Cin1, Cdc42, Rac1, and Wsp1. Cin1 is a multimodular adaptor protein proposed to regulate endocytosis, exocytosis, and signaling through interactions with several other proteins (47). Cin1 activates Cdc42 through the DH (Rho-GEF) domain; Cdc42, in turn, activates Wsp1 to promote Arp2/3-mediated actin polymerization. In addition to Cdc42, Wsp1 can also be an effector of Rac1 to regulate endocytosis and vacuole morphology. A possible effect of Cin1-DH on Rac1 has not been substantiated. Cin1 also activates Wsp1 through direct binding between the SRC homology 3 (SH3) (Cin1) and proline-rich (PR) (Wsp1) domains. Moreover, Cin1 may promote the extension of the branched actin filaments by binding directly to Arp2/3 homologs through the A domain and to monomeric actin (G-actin) through the WASP homology 2 (WH2) domain.
phology. The possible function of Cin1-DH with respect to Rac1 has not been substantiated. Cin1 could also modulate Wsp1 through direct binding between the SH3 domains and the proline-rich (PR) (Wsp1) domain. Moreover, Cin1 may promote the extension of branched actin filaments by direct binding of Arp2/3 homologs through the A domain and of monomeric actin (G-actin) through the WH2 domain, respectively.
Taken together, the multifunctional roles of Cin1, Wsp1, Cdc42, and Rac1 underscore the importance of intracellular trafficking, actin cytoskeleton organization, and signaling in the physiology and virulence of C. neoformans. Further studies of these proteins and their functions will further facilitate the understanding of intracellular transport in microbial pathogenesis. | 2018-04-03T01:46:36.274Z | 2012-02-10T00:00:00.000 | {
"year": 2012,
"sha1": "00c0821c8adb9ed369fd4da5720f03809778003d",
"oa_license": null,
"oa_url": "https://ec.asm.org/content/eukcell/11/4/471.full.pdf",
"oa_status": "GOLD",
"pdf_src": "Highwire",
"pdf_hash": "d2d518fe0656719fc3e1fa720daf6ab21e5d7d30",
"s2fieldsofstudy": [
"Biology"
],
"extfieldsofstudy": [
"Biology",
"Medicine"
]
} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.