Relationship Between 12 Adipocytokines and Distinct Components of the Metabolic Syndrome

Relationship Between 12 Adipocytokines and Distinct Components of the Metabolic Syndrome Abstract Objective Adipose tissue–derived signals potentially link obesity and adipose tissue dysfunction with metabolic and cardiovascular diseases. Although some adipocytokines have been closely related to metabolic and cardiovascular traits, it is unknown which adipocytokine or adipocytokine clusters serve as meaningful markers of metabolic syndrome (MS) components. Therefore, this study investigated the associations of 12 adipocytokines with components of the MS to identify the most relevant cytokines potentially related to specific metabolic profiles. Research Design and Methods Twelve cytokines [adiponectin, adipocyte fatty acid–binding protein (AFABP), angiopoietin-related growth factor, chemerin, fibroblast growth factor (FGF) 19, FGF21, FGF23, insulin-like growth factor-1, interleukin 10, irisin, progranulin, and vaspin] were quantified in a cross-sectional cohort of 1046 subjects. Hypothesis-free cluster analysis, multivariate regression analyses with parameters of the MS, and discriminant analysis were performed to assess associations and the relative importance of each cytokine for reflecting MS and its components. Results Among the studied adipocytokines, adiponectin, AFABP, chemerin, and FGF21 showed the strongest associations with MS and several MS components in discriminant analyses and multiple regression models. For certain metabolic components, these adipocytokines were better discriminators than routine metabolic markers. Other cytokines investigated in the present cohort are less able to distinguish between metabolically healthy and unhealthy subjects. Conclusions Adiponectin, AFABP, chemerin, and FGF21 showed the strongest associations with MS components in a general population, suggesting that adverse adipose tissue function is a major contributor to these metabolic abnormalities. Future prospective studies should address the question whether these adipocytokines can predict the development of metabolic disease states. Obesity is an increasing global health burden, and according to a recent meta-analysis, the number of people with obesity has more than doubled in the last three decades (1, 2). Patients with obesity and its associated comorbidities (i.e., type 2 diabetes, dyslipidemia, and hypertension) have greater mortality and shorter life expectancy (3, 4). Therefore, the metabolic syndrome (MS) as a complex of metabolic risk factors is an increasing public health and clinical problem (4). In recent years, a large number of adipocytokines have been identified as potential links between obesity and other metabolic disease states (5). However, most of the previous studies that aim to evaluate the predictive value of cytokines for metabolic and cardiovascular diseases have the following limitations: Studies investigating adipocytokines often include only a small number of subjects, are performed only in specific disease states, and consider only one rather than a combination of cytokines. For example, we previously demonstrated in a small cohort of 141 patients with obesity that distinct adipocytokine clusters are related to either body fat mass and inflammation or insulin sensitivity/hyperglycemia and lipid metabolism, respectively (6). However, the study investigated a cohort with specific disease states and limited case numbers. Thus, it still remains unclear which adipocytokines are associated with specific facets of the MS (i.e., visceral obesity, hypertension, dyslipidemia, and insulin resistance). The clinical relevance of adipocytokine measurements to estimate the type 2 diabetes risk has been proven for adiponectin (7). However, adipocytokine serum concentrations are still not used in clinical practice to diagnose or predict metabolic and cardiovascular diseases. Adipose tissue probably secretes >600 adipocytokines (8). With the expanding number of identified adipocytokines there is an increasing need to define their function and potential clinical relevance for metabolic diseases. To overcome limitations of previous studies, we aimed to investigate a panel of 12 adipocytokines [i.e., adiponectin, adipocyte fatty acid–binding protein (AFABP), chemerin, progranulin, irisin, fibroblast growth factor (FGF) 19, FGF21, FGF23, vaspin, angiopoietin-related growth factor (AGF), insulin-like growth factor (IGF)-1, and interleukin (IL) 10] in a cross-sectional cohort of 1046 extensively phenotyped subjects not specifically selected for metabolic or cardiovascular disease states. We characterized each protein with a stepwise statistical approach by performing hypothesis-free cluster analysis with metabolic markers and other cytokines in the entire cohort, performing linear regression analysis with components of the MS in a healthy subcohort, and investigating the relative importance of each cytokine for distinguishing components of the MS in all participants. Materials and Methods Study participants The design of the current study has been described previously (9–13). Briefly, all subjects recruited for this study are part of a sample from a self-contained population of Sorbs in eastern Germany and were enrolled in the study between 2005 and 2007. Participants were not specifically selected for metabolic or cardiovascular disease states. For the present analysis, 1046 Sorbs were available. All investigations were performed by trained staff and included standardized questionnaires, anthropometric parameters [body mass index (BMI), waist/hip ratio, body impedance analysis], and a 75-g oral glucose tolerance test. Body composition analysis was performed with the BIA-2000-S device (Data Input GmbH, Darmstadt, Germany) and analyzed with the software Nutri3 (Data Input GmbH). MS and its components, were diagnosed according to the Joint Scientific Statement on Harmonizing the Metabolic Syndrome (4). Homeostasis model assessment of insulin resistance (14) and mean arterial blood pressure (BP) (15) were calculated as previously described. After assessment of serum creatinine, estimated glomerular filtration rate was assessed with the Chronic Kidney Disease Epidemiology Collaboration equation (16). Spot urine specimens were analyzed for urinary albumin and creatinine, and the albumin/creatinine ratio was calculated. A 10-MHz ultrasound sensor (GE Healthcare, Inc., München, Germany) was used to measure the intima-media thickness (IMT) of the common carotid arteries. After three measurements, IMT values on each side were averaged and mean IMT was calculated. All subjects in this study, which was approved by the ethics committee of the University of Leipzig (Reg. No. 088-2005), gave written informed consent before taking part in the study. Assays In all subjects, blood samples were taken in the morning after an overnight fast and were immediately spun and frozen at −80°C until analyses were performed. Serum (adiponectin, AFABP, AGF, chemerin, FGF19, FGF21, irisin, IGF-1, IL10, progranulin, vaspin) and plasma (FGF23) concentrations of all cytokines under investigation were determined with commercially available enzyme-linked immunosorbent assays according to the manufacturers’ instructions (AGF, irisin, progranulin, and vaspin: AdipoGen Inc., Seoul, South Korea; adiponectin, AFABP, chemerin, FGF19, and FGF21: BioVendor Inc., Brno, Czech Republic; FGF23: Immutopics Inc., San Clemente, CA; IGF-1: Liaison, DiaSorin, Saluggia, Italy; IL10: BD Bioscience, Heidelberg, Germany). Fasting insulin was determined with the AutoDELFIA insulin assay (PerkinElmer Life and Analytical Sciences, Turku, Finland). Serum creatinine was measured with the kinetic enzymatic method. Urinary albumin was determined with a turbidimetric assay for the Roche modular automated analyzer (Roche Diagnostics, Mannheim, Germany). Urinary creatinine was quantified by a photometric assay for the Roche/Hitachi cobas c automated analyzer (Roche Diagnostics). Fasting glucose (FG), glucose levels during the oral glucose tolerance test, glycated hemoglobin A1c, total cholesterol, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol, triglycerides (TG), and C-reactive protein were measured by standard laboratory methods in a certified laboratory (University of Leipzig, Institute of Laboratory Medicine). Statistical analysis For statistical analysis, SPSS software version 24.0 (IBM, Armonk, NY) and the statistical software package “R” [www.r-project.org (17)] were used. In a first step, we performed hypothesis-free and cluster analysis to assess proximity of cytokines with anthropometric and metabolic markers. Before cluster analysis, all variables were sex adjusted. Cluster analysis was performed by Ward’s minimum variance method with the function “pvclust” of the statistical software package “R” [www.r-project.org (17)], and 10,000 bootstrapping replications were analyzed. Approximately unbiased (AU) P values and bootstrap probability values of the branching points in our cluster diagrams were calculated as previously described (6). AU P values ≥90% were considered strong evidence for the respective cluster. Next, multiple linear regression analysis was carried out in healthy subjects to identify independent associations of adipocytokines with components of the metabolic syndrome. For this purpose, only participants without a diagnosis or treatment of diabetes mellitus, hypertension, or impaired TG or HDL-C were included in each model, respectively. Thus, all subjects with a diagnosis of diabetes mellitus or on antidiabetic treatment were excluded from the model for FG. For mean arterial BP, all participants with an established diagnosis of hypertension or treatment were excluded. Furthermore, all subjects treated with peroxisome proliferator–activated receptor α-activators were not included in the multivariate model for TG. For HDL-C, all subjects on niacin treatment were excluded before multiple linear regression analysis. Before we performed multiple linear regression analyses, normality of the distribution of residuals was assessed by visual inspection of P-P plots. Homoscedasticity was assessed by visual inspection of residual plots, revealing no signs of strong heteroscedasticity (data not shown). No formal testing was applied to avoid sequential testing situations. To assess the degree of multicollinearity between independent variables, variance inflation factors were calculated for each model. Finally, discriminant analysis was performed to identify the cytokines most relevant for discriminating the MS and its components: impaired FG [FG ≥100 mg/dL (5.56 mmol/L) or treatment], elevated BP (systolic ≥130 or diastolic ≥85 mm Hg or treatment), elevated TG [≥150 mg/dL (1.7 mmol/L) or treatment], low HDL-C [<40 mg/dL (1.0 mmol/L) in men, <50 mg/dL (1.3 mmol/L) in women or treatment], and abdominal obesity (waist circumference ≥94 cm in men, ≥80 cm in women) (4). Components of the MS were transformed to dichotomous variables before analysis according to Alberti et al. (4). Discriminant analysis was then carried out with log-transformed, sex-adjusted, unstandardized residuals. We calculated two models consisting of adipocytokines only (model 1) or all adipocytokines and metabolic markers used for the definition of the MS (model 2). A P value of <0.05 was considered statistically significant in all analyses. Results Baseline characteristics of the entire study cohort Baseline characteristics of the study population are shown in Supplemental Table 1. Median (interquartile range) age was 48.3 (36.5 to 60.2) years and BMI was 26.4 (23.2 to 29.6) kg/m2. In the total cohort, 376 subjects (36.2%) presented with a MS as defined by Alberti et al. (4). Adipocytokine quartiles for each component of the MS are depicted in Table 1. Table 1. Quartiles of Circulating Levels of All Adipocytokines Depending on Components of the MS in the Entire Study Cohort   Quartile  Adiponectin (mg/L)  AFABP (µg/L)  AGF (µg/L)  Chemerin (µg/L)  FGF19 (ng/L)  FGF21 (ng/L)  FGF23 (RU/mL)  IGF-1 (µg/L)  IL10 (ng/L)  Irisin (mg/L)  Progranulin (µg/L)  Vaspin (µg/L)  N    971  1011  828  1009  1004  931  1039  1043  996  1039  1042  1041  FG (N = 1030)  I  18.1  14.0  40.3  108.2  255.0  39.9  70.3  179.5  4.4  0.83  102.9  0.7  II  16.2  15.9  34.7  110.1  240.4  60.0  71.0  160.2  5.0  0.83  106.9  0.5  III  16.4  18.5  38.5  120.2  247.3  80.0  68.6  151.3  4.7  0.76  110.2  0.4  IV  15.5  25.6  43.9  133.5  244.5  119.9  72.5  128.5  5.0  0.77  114.4  0.5  Mean BP (N = 1043)  I  16.5  14.0  36.3  105.7  242.6  41.2  68.2  181.6  5.3  0.81  102.9  0.6  II  17.1  15.9  38.1  112.7  254.7  63.4  71.2  164.9  4.7  0.79  107.8  0.5  III  15.6  19.3  39.1  120.4  217.9  91.2  70.2  145.5  4.7  0.82  108.5  0.4  IV  16.1  23.9  40.3  133.3  253.3  119.7  72.6  129.3  4.8  0.76  112.1  0.5  TG (N = 1044)  I  18.1  13.8  38.8  102.0  243.7  36.5  67.2  169.0  4.9  0.83  103.7  0.4  II  17.6  16.3  38.9  115.4  243.0  58.0  70.1  153.0  4.7  0.81  107.5  0.5  III  16.0  19.7  39.1  124.8  239.6  87.3  72.0  156.7  4.5  0.80  107.9  0.6  IV  14.3  22.3  37.1  127.2  254.2  151.1  72.2  139.2  5.0  0.77  114.3  0.5  HDL-C (N = 1044)  I  13.3  20.7  43.8  122.6  222.2  114.4  72.0  151.1  5.3  0.77  108.5  0.4  II  15.5  17.0  40.5  121.9  255.2  83.2  71.8  162.3  4.9  0.77  109.8  0.5  III  17.1  17.0  36.7  114.5  238.1  57.1  68.8  155.2  4.6  0.81  108.0  0.5  IV  19.7  16.3  35.5  110.4  259.3  51.3  70.2  156.7  4.5  0.83  105.0  0.5  Waist (N = 1039)  I  16.6  10.0  37.8  105.6  267.6  50.8  70.2  180.8  4.7  0.80  107.2  0.5  II  16.5  14.4  35.7  111.9  233.4  63.5  67.1  158.9  4.9  0.81  104.8  0.5  III  16.0  20.3  37.8  121.5  245.7  82.6  71.5  148.2  5.3  0.79  108.1  0.4  IV  16.2  33.2  44.4  133.2  238.1  97.2  73.9  129.8  4.3  0.80  112.3  0.5    Quartile  Adiponectin (mg/L)  AFABP (µg/L)  AGF (µg/L)  Chemerin (µg/L)  FGF19 (ng/L)  FGF21 (ng/L)  FGF23 (RU/mL)  IGF-1 (µg/L)  IL10 (ng/L)  Irisin (mg/L)  Progranulin (µg/L)  Vaspin (µg/L)  N    971  1011  828  1009  1004  931  1039  1043  996  1039  1042  1041  FG (N = 1030)  I  18.1  14.0  40.3  108.2  255.0  39.9  70.3  179.5  4.4  0.83  102.9  0.7  II  16.2  15.9  34.7  110.1  240.4  60.0  71.0  160.2  5.0  0.83  106.9  0.5  III  16.4  18.5  38.5  120.2  247.3  80.0  68.6  151.3  4.7  0.76  110.2  0.4  IV  15.5  25.6  43.9  133.5  244.5  119.9  72.5  128.5  5.0  0.77  114.4  0.5  Mean BP (N = 1043)  I  16.5  14.0  36.3  105.7  242.6  41.2  68.2  181.6  5.3  0.81  102.9  0.6  II  17.1  15.9  38.1  112.7  254.7  63.4  71.2  164.9  4.7  0.79  107.8  0.5  III  15.6  19.3  39.1  120.4  217.9  91.2  70.2  145.5  4.7  0.82  108.5  0.4  IV  16.1  23.9  40.3  133.3  253.3  119.7  72.6  129.3  4.8  0.76  112.1  0.5  TG (N = 1044)  I  18.1  13.8  38.8  102.0  243.7  36.5  67.2  169.0  4.9  0.83  103.7  0.4  II  17.6  16.3  38.9  115.4  243.0  58.0  70.1  153.0  4.7  0.81  107.5  0.5  III  16.0  19.7  39.1  124.8  239.6  87.3  72.0  156.7  4.5  0.80  107.9  0.6  IV  14.3  22.3  37.1  127.2  254.2  151.1  72.2  139.2  5.0  0.77  114.3  0.5  HDL-C (N = 1044)  I  13.3  20.7  43.8  122.6  222.2  114.4  72.0  151.1  5.3  0.77  108.5  0.4  II  15.5  17.0  40.5  121.9  255.2  83.2  71.8  162.3  4.9  0.77  109.8  0.5  III  17.1  17.0  36.7  114.5  238.1  57.1  68.8  155.2  4.6  0.81  108.0  0.5  IV  19.7  16.3  35.5  110.4  259.3  51.3  70.2  156.7  4.5  0.83  105.0  0.5  Waist (N = 1039)  I  16.6  10.0  37.8  105.6  267.6  50.8  70.2  180.8  4.7  0.80  107.2  0.5  II  16.5  14.4  35.7  111.9  233.4  63.5  67.1  158.9  4.9  0.81  104.8  0.5  III  16.0  20.3  37.8  121.5  245.7  82.6  71.5  148.2  5.3  0.79  108.1  0.4  IV  16.2  33.2  44.4  133.2  238.1  97.2  73.9  129.8  4.3  0.80  112.3  0.5  Participants were grouped into quartiles I to IV for each component of the MS, (i.e., FG, mean BP, TG, HDL-C, and waist circumference). For all components of the MS, median circulating levels of all cytokines according to each quartile are depicted. Numbers of subjects included are shown for each metabolic component and adipocytokine. View Large Cluster analysis of all adipocytokines and cardiometabolic markers Unsupervised and sex-adjusted cluster analysis of adipocytokines and anthropometric and cardiometabolic markers revealed four clusters in the entire cohort (Fig. 1): age, mean IMT, and waist/hip ratio (cluster 1); AFABP, BMI, and fat mass (cluster 2); estimated glomerular filtration rate and IGF-1 (cluster 3); and HDL-C and adiponectin (cluster 4), were significantly clustered (AU P values ≥90%). Figure 1. View largeDownload slide Cluster analysis regarding metabolic cytokines and anthropometric and metabolic markers. Before cluster analysis, all variables were sex adjusted. Four clusters of strong correlations could be detected. Cluster analysis was performed by Ward’s minimum variance method, and 10,000 bootstrapping replications were analyzed. AU (numbers in red) P values and bootstrap probability (numbers in green) values of the branching points were calculated. AU P values ≥90% were considered strong evidence for the respective cluster. In the cluster analysis consisting of all variables, 595 subjects were included. ACR, albumin/creatinine ratio; eGFR, estimated glomerular filtration rate; LDL-C, low density lipoprotein cholesterol; HbA1c, glycated hemoglobin A1c; HOMA-IR, homeostasis model assessment of insulin resistance; WHR, waist/hip ratio. Figure 1. View largeDownload slide Cluster analysis regarding metabolic cytokines and anthropometric and metabolic markers. Before cluster analysis, all variables were sex adjusted. Four clusters of strong correlations could be detected. Cluster analysis was performed by Ward’s minimum variance method, and 10,000 bootstrapping replications were analyzed. AU (numbers in red) P values and bootstrap probability (numbers in green) values of the branching points were calculated. AU P values ≥90% were considered strong evidence for the respective cluster. In the cluster analysis consisting of all variables, 595 subjects were included. ACR, albumin/creatinine ratio; eGFR, estimated glomerular filtration rate; LDL-C, low density lipoprotein cholesterol; HbA1c, glycated hemoglobin A1c; HOMA-IR, homeostasis model assessment of insulin resistance; WHR, waist/hip ratio. Linear regression analysis with components of the MS We assessed four different regression models to analyze the associations between the adipocytokines and different components of the MS (Table 2). In each model, only the subgroup of subjects without diagnosed alteration or drug treatment of the respective component of the MS were included (i.e., FG, BP, TG, and HDL-C). Evaluation of the variance inflation factors did not reveal a high degree of multicollinearity (data not shown). In the multivariate model consisting of all cytokines, age, sex, and fat mass, circulating chemerin levels were positively associated with FG (β = 0.109; P = 0.005). Furthermore, serum chemerin (β = 0.141; P = 0.001) and FGF21 (β = 0.092; P = 0.020) were predictors of higher BP. Elevated TG levels were significantly associated with lower adiponectin (β = −0.161; P < 0.001) and irisin (β = −0.099; P = 0.002) concentrations. In contrast, chemerin (β = 0.171; P < 0.001), FGF21 (β = 0.246; P < 0.001), progranulin (β = 0.069; P = 0.037), and vaspin (β = 0.138; P < 0.001) positively correlated with TG. Finally, higher HDL-C was significantly associated with higher adiponectin (β = 0.304; P < 0.001) and IL10 (β = 0.095; P = 0.004), whereas AGF (β = −0.113; P = 0.001), chemerin (β = −0.106; P = 0.004), FGF21 (β = −0.081; P = 0.018), and IGF-1 (β = −0.141; P = 0.001) were predictors of decreased HDL-C, respectively (Table 2). Table 2. Multivariate Linear Regression Analyses Between Components of the MS and Adipocytokines Dependent Variable  Covariates  Independent Variables  Model 1: Single Cytokine  Model 2: All Cytokines  N  r2  β  P  r2  β  P  FG  Age, sex, fat mass  Adiponectin, mg/L  0.346  −0.041  0.180  0.383  −0.032  0.387  570  AFABP, µg/L  0.344  0.070  0.059  −0.005  0.917  AGF, µg/L  0.337  −0.023  0.460  −0.023  0.494  Chemerin, µg/L  0.344  0.076  0.010  0.109  0.005  FGF19, ng/L  0.338  −0.019  0.501  −0.050  0.151  FGF21, ng/L  0.342  −0.002  0.943  0.007  0.836  FGF23, RU/mL  0.341  −0.02  0.461  −0.041  0.229  IGF-1, µg/L  0.340  0.042  0.241  0.072  0.120  IL10, ng/L  0.350  0.003  0.922  −0.034  0.318  Irisin, mg/L  0.346  −0.082  0.003  −0.017  0.616  Progranulin, µg/L  0.342  0.049  0.075  0.034  0.337  Vaspin, µg/L  0.342  −0.011  0.704  −0.031  0.397  Mean BP  Age, sex, fat mass  Adiponectin, mg/L  0.257  −0.044  0.214  0.355  0.015  0.710  491  AFABP, µg/L  0.259  0.052  0.201  0.007  0.888  AGF, µg/L  0.295  −0.022  0.527  0.009  0.804  Chemerin, µg/L  0.270  0.123  <0.001  0.141  0.001  FGF19, ng/L  0.255  0.044  0.162  0.033  0.390  FGF21, ng/L  0.277  0.112  0.001  0.092  0.020  FGF23, RU/mL  0.256  0.002  0.943  −0.035  0.368  IGF-1, µg/L  0.259  −0.045  0.287  −0.026  0.610  IL10, ng/L  0.262  −0.026  0.419  −0.022  0.565  Irisin, mg/L  0.258  0.004  0.893  −0.041  0.277  Progranulin, µg/L  0.263  0.077  0.013  0.047  0.231  Vaspin, µg/L  0.262  0.067  0.045  0.057  0.158  TG (ln)  Age, sex, fat mass  Adiponectin, mg/L  0.228  −0.158  <0.001  0.386  −0.161  <0.001  639  AFABP, µg/L  0.219  0.152  <0.001  0.035  0.436  AGF, µg/L  0.215  −0.041  0.199  −0.020  0.537  Chemerin, µg/L  0.241  0.196  <0.001  0.171  <0.001  FGF19, ng/L  0.209  0.043  0.133  0.004  0.904  FGF21, ng/L  0.277  0.273  <0.001  0.246  <0.001  FGF23, RU/mL  0.210  0.001  0.962  −0.016  0.614  IGF-1, µg/L  0.208  0.003  0.929  0.031  0.478  IL10, ng/L  0.212  −0.032  0.261  −0.042  0.183  Irisin, mg/L  0.212  −0.087  0.002  −0.099  0.002  Progranulin, µg/L  0.218  0.099  <0.001  0.069  0.037  Vaspin, µg/L  0.226  0.145  <0.001  0.138  <0.001  HDL-C  Age, sex, fat mass  Adiponectin, mg/L  0.282  0.261  <0.001  0.366  0.304  <0.001  640  AFABP, µg/L  0.228  −0.058  0.121  0.043  0.351  AGF, µg/L  0.240  −0.104  0.001  −0.113  0.001  Chemerin, µg/L  0.232  −0.076  0.012  −0.106  0.004  FGF19, ng/L  0.231  0.056  0.048  0.038  0.241  FGF21, ng/L  0.223  −0.062  0.037  −0.081  0.018  FGF23, RU/mL  0.233  −0.054  0.054  −0.046  0.158  IGF-1, µg/L  0.242  −0.151  <0.001  −0.141  0.001  IL10, ng/L  0.233  0.074  0.008  0.095  0.004  Irisin, mg/L  0.232  0.041  0.134  0.048  0.141  Progranulin, µg/L  0.230  −0.038  0.172  −0.057  0.092  Vaspin, µg/L  0.230  −0.010  0.740  0.048  0.172  Dependent Variable  Covariates  Independent Variables  Model 1: Single Cytokine  Model 2: All Cytokines  N  r2  β  P  r2  β  P  FG  Age, sex, fat mass  Adiponectin, mg/L  0.346  −0.041  0.180  0.383  −0.032  0.387  570  AFABP, µg/L  0.344  0.070  0.059  −0.005  0.917  AGF, µg/L  0.337  −0.023  0.460  −0.023  0.494  Chemerin, µg/L  0.344  0.076  0.010  0.109  0.005  FGF19, ng/L  0.338  −0.019  0.501  −0.050  0.151  FGF21, ng/L  0.342  −0.002  0.943  0.007  0.836  FGF23, RU/mL  0.341  −0.02  0.461  −0.041  0.229  IGF-1, µg/L  0.340  0.042  0.241  0.072  0.120  IL10, ng/L  0.350  0.003  0.922  −0.034  0.318  Irisin, mg/L  0.346  −0.082  0.003  −0.017  0.616  Progranulin, µg/L  0.342  0.049  0.075  0.034  0.337  Vaspin, µg/L  0.342  −0.011  0.704  −0.031  0.397  Mean BP  Age, sex, fat mass  Adiponectin, mg/L  0.257  −0.044  0.214  0.355  0.015  0.710  491  AFABP, µg/L  0.259  0.052  0.201  0.007  0.888  AGF, µg/L  0.295  −0.022  0.527  0.009  0.804  Chemerin, µg/L  0.270  0.123  <0.001  0.141  0.001  FGF19, ng/L  0.255  0.044  0.162  0.033  0.390  FGF21, ng/L  0.277  0.112  0.001  0.092  0.020  FGF23, RU/mL  0.256  0.002  0.943  −0.035  0.368  IGF-1, µg/L  0.259  −0.045  0.287  −0.026  0.610  IL10, ng/L  0.262  −0.026  0.419  −0.022  0.565  Irisin, mg/L  0.258  0.004  0.893  −0.041  0.277  Progranulin, µg/L  0.263  0.077  0.013  0.047  0.231  Vaspin, µg/L  0.262  0.067  0.045  0.057  0.158  TG (ln)  Age, sex, fat mass  Adiponectin, mg/L  0.228  −0.158  <0.001  0.386  −0.161  <0.001  639  AFABP, µg/L  0.219  0.152  <0.001  0.035  0.436  AGF, µg/L  0.215  −0.041  0.199  −0.020  0.537  Chemerin, µg/L  0.241  0.196  <0.001  0.171  <0.001  FGF19, ng/L  0.209  0.043  0.133  0.004  0.904  FGF21, ng/L  0.277  0.273  <0.001  0.246  <0.001  FGF23, RU/mL  0.210  0.001  0.962  −0.016  0.614  IGF-1, µg/L  0.208  0.003  0.929  0.031  0.478  IL10, ng/L  0.212  −0.032  0.261  −0.042  0.183  Irisin, mg/L  0.212  −0.087  0.002  −0.099  0.002  Progranulin, µg/L  0.218  0.099  <0.001  0.069  0.037  Vaspin, µg/L  0.226  0.145  <0.001  0.138  <0.001  HDL-C  Age, sex, fat mass  Adiponectin, mg/L  0.282  0.261  <0.001  0.366  0.304  <0.001  640  AFABP, µg/L  0.228  −0.058  0.121  0.043  0.351  AGF, µg/L  0.240  −0.104  0.001  −0.113  0.001  Chemerin, µg/L  0.232  −0.076  0.012  −0.106  0.004  FGF19, ng/L  0.231  0.056  0.048  0.038  0.241  FGF21, ng/L  0.223  −0.062  0.037  −0.081  0.018  FGF23, RU/mL  0.233  −0.054  0.054  −0.046  0.158  IGF-1, µg/L  0.242  −0.151  <0.001  −0.141  0.001  IL10, ng/L  0.233  0.074  0.008  0.095  0.004  Irisin, mg/L  0.232  0.041  0.134  0.048  0.141  Progranulin, µg/L  0.230  −0.038  0.172  −0.057  0.092  Vaspin, µg/L  0.230  −0.010  0.740  0.048  0.172  Model 1 is calculated for each adipocytokine separately, adjusted for age, sex, and fat mass. In model 2, all adipocytokines, as well as age, sex, and fat mass, are included as independent variables. The r2 of each model, standardized β, and P values are depicted, and significant predictors for each metabolic component are marked in bold. For each component of the MS, only participants without a diagnosis or treatment of the respective component were included. Numbers of subjects included in model 2 for each metabolic component are depicted. Normal distribution was assessed by visual inspection, and non-normally distributed variables were logarithmically transformed (ln) before testing. View Large Discriminant analysis of the MS and its components by adipocytokines To address the question which adipocytokines are most appropriate to distinguish the presence or absence of a certain component of the MS, as well as the MS itself, we performed sex-adjusted descriptive discriminant analyses of all cytokines. Based on the standardized canonical discriminant coefficients, AFABP, chemerin, and FGF21 were the three most prominent markers for impaired FG (Table 3, model 1). For hypertension, FGF21, IGF-1, and AFABP were most relevant (Table 3, model 1). The adipocytokines FGF21, adiponectin, and AFABP had the highest relative importance for hypertriglyceridemia (Table 3, model 1). Furthermore, adiponectin, FGF21, and FGF23 were the most relevant cytokines for reduced HDL-C levels (Table 3, model 1). Moreover, AFABP, IGF-1, and adiponectin showed highest standardized canonical discriminant coefficients for increased waist circumference (Table 3, model 1). Finally, the adipocytokines AFABP, FGF21, and chemerin were highly relevant to distinguish between the presence or absence of the MS (Table 3, model 1). Table 3. Discriminant Analysis of 12 Adipocytokines and Components of the MS and the MS   Components of MS  MS    FG  Hypertension  TG  HDL-C  Waist        Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  N  648  646  648  636  648  635  648  635  647  635  647  635  Standardized canonical discriminant coefficients                           FG    —    0.127    0.131    0.041    0.200    0.363   Mean BP    0.267    —    0.286    −0.320    0.338    0.283   TG    −0.077    0.151    —    0.466    0.064    0.303   HDL-C    −0.206    0.213    −0.346    —    −0.321    −0.111   Waist    0.498    0.596    0.217    0.598    —    0.486   Adiponectin  0.056  0.175  −0.094  −0.048  −0.313  −0.142  −0.689  −0.457  −0.214  −0.062  −0.208  −0.005   AFABP  0.565  0.200  0.350  0.004  0.305  0.041  0.230  −0.105  0.692  0.516  0.514  0.030   AGF  0.001  −0.060  0.011  −0.019  0.050  −0.019  0.092  0.016  0.096  0.044  0.152  0.043   Chemerin  0.348  0.252  0.260  0.230  0.214  0.098  0.123  0.093  0.098  −0.004  0.325  0.168   FGF19  −0.141  −0.045  −0.168  −0.101  −0.060  −0.006  −0.199  −0.096  −0.126  −0.079  −0.137  −0.029   FGF21  0.293  0.177  0.381  0.270  0.532  0.362  0.323  0.092  0.169  0.019  0.447  0.170   FGF23  −0.252  −0.203  −0.075  0.003  −0.031  −0.018  0.277  0.190  −0.080  −0.049  −0.174  −0.091   IGF-1  −0.141  −0.013  −0.352  −0.154  0.005  0.036  0.024  0.130  −0.333  −0.286  −0.048  0.127   IL10  −0.051  −0.058  −0.051  −0.058  −0.077  −0.065  0.216  0.145  0.050  0.088  −0.063  −0.021   Irisin  0.028  0.065  −0.093  −0.040  −0.295  −0.233  −0.113  −0.011  −0.085  −0.049  −0.110  −0.043   Progranulin  0.132  0.058  0.174  0.131  0.231  0.117  0.093  0.020  0.095  −0.002  0.126  −0.033   Vaspin  0.044  0.070  −0.008  −0.005  0.283  0.257  −0.149  −0.201  −0.045  −0.063  0.093  0.051  Rc2  0.216  0.283  0.183  0.240  0.207  0.266  0.085  0.121  0.328  0.398  0.324  0.543  P  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001    Components of MS  MS    FG  Hypertension  TG  HDL-C  Waist        Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  N  648  646  648  636  648  635  648  635  647  635  647  635  Standardized canonical discriminant coefficients                           FG    —    0.127    0.131    0.041    0.200    0.363   Mean BP    0.267    —    0.286    −0.320    0.338    0.283   TG    −0.077    0.151    —    0.466    0.064    0.303   HDL-C    −0.206    0.213    −0.346    —    −0.321    −0.111   Waist    0.498    0.596    0.217    0.598    —    0.486   Adiponectin  0.056  0.175  −0.094  −0.048  −0.313  −0.142  −0.689  −0.457  −0.214  −0.062  −0.208  −0.005   AFABP  0.565  0.200  0.350  0.004  0.305  0.041  0.230  −0.105  0.692  0.516  0.514  0.030   AGF  0.001  −0.060  0.011  −0.019  0.050  −0.019  0.092  0.016  0.096  0.044  0.152  0.043   Chemerin  0.348  0.252  0.260  0.230  0.214  0.098  0.123  0.093  0.098  −0.004  0.325  0.168   FGF19  −0.141  −0.045  −0.168  −0.101  −0.060  −0.006  −0.199  −0.096  −0.126  −0.079  −0.137  −0.029   FGF21  0.293  0.177  0.381  0.270  0.532  0.362  0.323  0.092  0.169  0.019  0.447  0.170   FGF23  −0.252  −0.203  −0.075  0.003  −0.031  −0.018  0.277  0.190  −0.080  −0.049  −0.174  −0.091   IGF-1  −0.141  −0.013  −0.352  −0.154  0.005  0.036  0.024  0.130  −0.333  −0.286  −0.048  0.127   IL10  −0.051  −0.058  −0.051  −0.058  −0.077  −0.065  0.216  0.145  0.050  0.088  −0.063  −0.021   Irisin  0.028  0.065  −0.093  −0.040  −0.295  −0.233  −0.113  −0.011  −0.085  −0.049  −0.110  −0.043   Progranulin  0.132  0.058  0.174  0.131  0.231  0.117  0.093  0.020  0.095  −0.002  0.126  −0.033   Vaspin  0.044  0.070  −0.008  −0.005  0.283  0.257  −0.149  −0.201  −0.045  −0.063  0.093  0.051  Rc2  0.216  0.283  0.183  0.240  0.207  0.266  0.085  0.121  0.328  0.398  0.324  0.543  P  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  Impaired FG was defined by FG ≥100 mg/dL (5.56 mmol/L) or treatment; hypertension by systolic BP ≥130 or diastolic BP ≥85 mm Hg or treatment; elevated TG by TG ≥150 mg/dL (1.7 mmol/L) or treatment; low HDL-C by <40 mg/dL (1.0 mmol/L) in men, <50 mg/dL (1.3 mmol/L) in women or treatment; and abdominal obesity by waist circumference ≥94 cm in men, ≥80 cm in women (4). Numbers of subjects included in each model, standardized discriminant regression coefficients, squared canonical correlation ( Rc2) test statistic, and P values are depicted. Model 1 includes only adipocytokines; model 2 consists of all adipocytokines, as well as the components of the MS. The three highest standardized discriminant regression coefficients for each metabolic component are marked in bold. Components of the MS were transformed to dichotomous variables before analysis and reference population in all models was the healthy group. All variables were logarithmically transformed and sex adjusted before testing. View Large Waist circumference appeared to be the strongest discriminator for components of the MS (Table 3, model 2). In the ranking of the standardized canonical discriminant coefficients, certain adipocytokines were better discriminators than routine metabolic markers. Thus, FGF21 is the strongest discriminator for elevated TG even when markers used for the definition of the MS were included in the model (Table 3, model 2). In contrast, AGF, FGF19, IL10, irisin, progranulin, and vaspin are less potent to discriminate between metabolically healthy and unhealthy subjects (Table 3, models 1 and 2). Discussion In the current study we investigated the associations of 12 adipocytokines with components of the MS (i.e., impaired FG, hypertension, dyslipidemia, and visceral obesity) and with the MS itself. We used hypothesis-free cluster analysis, multivariate regression models, and discriminant analyses to assess the relations and the relevance of distinct cytokines with different anthropometric and biochemical measures of the MS. In metabolically healthy subjects, the adipocytokines chemerin, FGF21, and adiponectin showed the highest associations as assessed by standardized β with the respective metabolic outcome parameters (i.e., FG, BP, TG, and HDL-C). Thus, in patients not affected by the respective component of the MS, the adipocytokine chemerin is significantly associated with an adverse metabolic status even after adjustment for fat mass. Interestingly, chemerin has been presented as an adipocytokine that influences adipocyte expression of genes involved in glucose and lipid metabolism in vitro (18). Additional studies from our group and others demonstrate that chemerin is also adversely associated with glucose homeostasis (19), body fat mass (6), inflammation (6, 20), TG (21), HDL-C (21), and BP (21) in humans with metabolic disease states, supporting our findings. Importantly, the observed associations between chemerin and cardiometabolic traits are independent of fat mass, and chemerin can distinguish certain metabolic components (i.e., FG and hypertension) besides waist circumference. These results suggest either that chemerin is not exclusively produced by adipose tissue or that impaired adipose tissue function rather than increased fat mass determines increased circulating chemerin. The latter hypothesis is supported by data from age-, sex-, and BMI-matched patients with either insulin-sensitive or insulin-resistant obesity (22). Independently of fat mass, people with insulin-resistant obesity and impaired adipose tissue function had higher serum chemerin concentrations (22). Furthermore, we have recently shown that chemerin serum concentrations closely reflect body weight dynamics in the context of different diet interventions (23). In addition to chemerin, the adipocytokines adiponectin and FGF21 are associated with distinct metabolic parameters in our linear regression analysis. Thus, high adiponectin is a predictor for low TG and higher HDL-C. Interestingly, adiponectin is also significantly clustered with HDL-C in our unbiased cluster analysis of the entire cohort. This finding is in accordance with published data obtained in obese patients undergoing weight loss intervention (24) and healthy patients (25). Treatment with statins and fibrates not only increases HDL-C but also raises adiponectin levels in patients with coronary artery disease and dyslipidemia (26). Furthermore, adiponectin is able to distinguish subjects with impaired HDL-C with similar power to TG and waist circumference. FGF21 is a marker and predictor for an adverse metabolic profile (27, 28). Zhang et al. (27) demonstrated similar regulations in a cohort of 232 Chinese subjects with a high prevalence of overweight or obesity and diabetic patients. This finding is in accordance with the hypothesis of an FGF21 resistance observed in MS and its components, including obesity (29). The results of our linear regression analysis provide further evidence that the observed associations of FGF21 might be causal and independent of fat mass even in a healthy cohort. FGF21 is the strongest discriminator for subjects with elevated TG in our discriminant analysis including all adipocytokines and metabolic markers. In our cohort, AFABP was significantly clustered with fat mass and BMI. This finding is corroborated by our regression models showing no association with facets of the MS when adjusted for fat mass. AFABP has been introduced as an adipocytokine with adverse cardiometabolic effects that is associated with components of the MS [e.g., obesity, insulin resistance, and dyslipidemia, as reviewed in (30)] and renal dysfunction (31). Among all investigated adipocytokines, AFABP also robustly distinguishes the presence or absence of the MS and its components. However, when waist circumference is included in the discriminant analysis, the ability of AFABP to distinguish metabolic components disappears. It should be noted in this context that a recent study by Zachariah et al. (32) demonstrated that AFABP is higher in metabolically healthy obese patients than in lean subjects but does not predict new-onset MS in a longitudinal analysis of the Framingham Heart Study. These data further support our results suggesting that AFABP is a surrogate parameter of fat mass rather than being causally involved in the pathogenesis of the MS. In contrast to adiponectin, AFABP, chemerin, and FGF21, other cytokines in our cohort are associated with markers of the MS in regression models but are not clustered to anthropometric or metabolic markers including irisin and vaspin, as well as AGF, IGF-1, and IL10. It should be noted that these associations are weak in our regression analyses. Furthermore, these adipocytokines have low standardized discriminant regression coefficients in discriminant analyses for components of the MS. Moreover, results of our multivariate linear regression analyses and discriminant analyses cannot be directly compared because different subsets of individuals have been included to meet the distinct questions of interest. Based on our data, AGF, IGF-1, IL10, irisin, progranulin, and vaspin are less important for the diagnosis of the MS and its components as compared with adiponectin, AFABP, chemerin, and FGF21, raising the question of whether clinical relevance can be derived. Limitations of the current study include the cross-sectional design. Therefore, no causality can be derived. Furthermore, the stability of the different adipocytokines in frozen samples could differ, and sample degradation may have contributed to variability within our results. On the other hand, all samples were analyzed in a single laboratory in one batch, and therefore different sample handling and storage should not influence our observed results. Furthermore, Lee et al. (33) suggested that plasma adipocytokine levels in general are stable and single measurements can represent cytokine levels over time in population-based studies. In addition, extensive phenotyping was performed at a high level of standardization by a trained study team. In conclusion, we have demonstrated that some adipocytokines could serve as markers for distinct metabolic disease states in a general population, including adiponectin, chemerin, and FGF21. These adipocytokines might have fat mass–independent effects on markers of lipid and glucose metabolism, as well as hypertension. Future prospective studies should address the question of whether these adipocytokines can predict metabolic disease states, as already shown for adiponectin (7), retinol-binding protein 4 (32), and fetuin-A (32), and should therefore be included in routine clinical measurement for risk stratification. Abbreviations: AFABP adipocyte fatty acid–binding protein AGF angiopoietin-related growth factor AU approximately unbiased BMI body mass index BP blood pressure FG fasting glucose FGF fibroblast growth factor HDL-C high-density lipoprotein cholesterol IGF insulin-like growth factor IL interleukin IMT intima-media thickness MS metabolic syndrome TG triglycerides. Acknowledgments We thank all who participated in the study. We especially thank Beate Gutsmann for her valuable contribution to data collection and excellent technical assistance. Financial Support: This study was supported by grants to A.T. from the Deutsche Forschungsgemeinschaft (DFG; Collaborative Research Centre 1052/1, C01 and Priority Programme 1629 TO 718/2-1), the German Diabetes Association, and the Federal Ministry of Education and Research (BMBF), Germany (Funding number: 01EO1001, Integrated Research and Treatment Center AdiposityDiseases, Postdoctoral Fellow program). Furthermore, T.E. was supported by the Federal Ministry of Education and Research, Germany (Funding number: 01EO1001, Integrated Research and Treatment Center AdiposityDiseases, MetaRot and Postdoctoral Fellow program) and by a Merck Sharp & Dohme grant (MSD Stipendium 2013 Diabetologie). P.K. was supported by the Boehringer Ingelheim Foundation. Author Contributions: T.E., C.G., and A.T. wrote the manuscript and researched data. M. Scholz contributed to statistical analyses. M. Scholz, T.W., D.S., M.F., M.B., M. Stumvoll, and P.K. contributed to the discussion and reviewed and edited the manuscript. T.E. and A.T. are the guarantors of this work and as such had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Disclosure Summary: The authors have nothing to disclose. References 1. Ng M, Fleming T, Robinson M, Thomson B, Graetz N, Margono C, Mullany EC, Biryukov S, Abbafati C, Abera SF, Abraham JP, Abu-Rmeileh NME, Achoki T, AlBuhairan FS, Alemu ZA, Alfonso R, Ali MK, Ali R, Guzman NA, Ammar W, Anwari P, Banerjee A, Barquera S, Basu S, Bennett DA, Bhutta Z, Blore J, Cabral N, Nonato IC, Chang JC, Chowdhury R, Courville KJ, Criqui MH, Cundiff DK, Dabhadkar KC, Dandona L, Davis A, Dayama A, Dharmaratne SD, Ding EL, Durrani AM, Esteghamati A, Farzadfar F, Fay DF, Feigin VL, Flaxman A, Forouzanfar MH, Goto A, Green MA, Gupta R, Hafezi-Nejad N, Hankey GJ, Harewood HC, Havmoeller R, Hay S, Hernandez L, Husseini A, Idrisov BT, Ikeda N, Islami F, Jahangir E, Jassal SK, Jee SH, Jeffreys M, Jonas JB, Kabagambe EK, Khalifa SE, Kengne AP, Khader YS, Khang YH, Kim D, Kimokoti RW, Kinge JM, Kokubo Y, Kosen S, Kwan G, Lai T, Leinsalu M, Li Y, Liang X, Liu S, Logroscino G, Lotufo PA, Lu Y, Ma J, Mainoo NK, Mensah GA, Merriman TR, Mokdad AH, Moschandreas J, Naghavi M, Naheed A, Nand D, Narayan KM, Nelson EL, Neuhouser ML, Nisar MI, Ohkubo T, Oti SO, Pedroza A, Prabhakaran D, Roy N, Sampson U, Seo H, Sepanlou SG, Shibuya K, Shiri R, Shiue I, Singh GM, Singh JA, Skirbekk V, Stapelberg NJ, Sturua L, Sykes BL, Tobias M, Tran BX, Trasande L, Toyoshima H, van de Vijver S, Vasankari TJ, Veerman JL, Velasquez-Melendez G, Vlassov VV, Vollset SE, Vos T, Wang C, Wang X, Weiderpass E, Werdecker A, Wright JL, Yang YC, Yatsuya H, Yoon J, Yoon SJ, Zhao Y, Zhou M, Zhu S, Lopez AD, Murray CJ, Gakidou E. Global, regional, and national prevalence of overweight and obesity in children and adults during 1980–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet . 2014; 384( 9945): 766– 781. Google Scholar CrossRef Search ADS PubMed  2. Afshin A, Forouzanfar MH, Reitsma MB, Sur P, Estep K, Lee A, Marczak L, Mokdad AH, Moradi-Lakeh M, Naghavi M, Salama JS, Vos T, Abate KH, Abbafati C, Ahmed MB, Al-Aly Z, Alkerwi A, Al-Raddadi R, Amare AT, Amberbir A, Amegah AK, Amini E, Amrock SM, Anjana RM, Ärnlöv J, Asayesh H, Banerjee A, Barac A, Baye E, Bennett DA, Beyene AS, Biadgilign S, Biryukov S, Bjertness E, Boneya DJ, Campos-Nonato I, Carrero JJ, Cecilio P, Cercy K, Ciobanu LG, Cornaby L, Damtew SA, Dandona L, Dandona R, Dharmaratne SD, Duncan BB, Eshrati B, Esteghamati A, Feigin VL, Fernandes JC, Fürst T, Gebrehiwot TT, Gold A, Gona PN, Goto A, Habtewold TD, Hadush KT, Hafezi-Nejad N, Hay SI, Horino M, Islami F, Kamal R, Kasaeian A, Katikireddi SV, Kengne AP, Kesavachandran CN, Khader YS, Khang YH, Khubchandani J, Kim D, Kim YJ, Kinfu Y, Kosen S, Ku T, Defo BK, Kumar GA, Larson HJ, Leinsalu M, Liang X, Lim SS, Liu P, Lopez AD, Lozano R, Majeed A, Malekzadeh R, Malta DC, Mazidi M, McAlinden C, McGarvey ST, Mengistu DT, Mensah GA, Mensink GBM, Mezgebe HB, Mirrakhimov EM, Mueller UO, Noubiap JJ, Obermeyer CM, Ogbo FA, Owolabi MO, Patton GC, Pourmalek F, Qorbani M, Rafay A, Rai RK, Ranabhat CL, Reinig N, Safiri S, Salomon JA, Sanabria JR, Santos IS, Sartorius B, Sawhney M, Schmidhuber J, Schutte AE, Schmidt MI, Sepanlou SG, Shamsizadeh M, Sheikhbahaei S, Shin MJ, Shiri R, Shiue I, Roba HS, Silva DAS, Silverberg JI, Singh JA, Stranges S, Swaminathan S, Tabarés-Seisdedos R, Tadese F, Tedla BA, Tegegne BS, Terkawi AS, Thakur JS, Tonelli M, Topor-Madry R, Tyrovolas S, Ukwaja KN, Uthman OA, Vaezghasemi M, Vasankari T, Vlassov VV, Vollset SE, Weiderpass E, Werdecker A, Wesana J, Westerman R, Yano Y, Yonemoto N, Yonga G, Zaidi Z, Zenebe ZM, Zipkin B, Murray CJL; GBD 2015 Obesity Collaborators. Health effects of overweight and obesity in 195 countries over 25 years. N Engl J Med . 2017; 377( 1): 13– 27. Google Scholar CrossRef Search ADS PubMed  3. Di Angelantonio E, Bhupathiraju ShN, Wormser D, Gao P, Kaptoge S, Berrington de Gonzalez A, Cairns BJ, Huxley R, Jackson ChL, Joshy G, Lewington S, Manson JE, Murphy N, Patel AV, Samet JM, Woodward M, Zheng W, Zhou M, Bansal N, Barricarte A, Carter B, Cerhan JR, Smith GD, Fang X, Franco OH, Green J, Halsey J, Hildebrand JS, Jung KJ, Korda RJ, McLerran DF, Moore SC, O’Keeffe LM, Paige E, Ramond A, Reeves GK, Rolland B, Sacerdote C, Sattar N, Sofianopoulou E, Stevens J, Thun M, Ueshima H, Yang L, Yun YD, Willeit P, Banks E, Beral V, Chen ZH, Gapstur SM, Gunter MJ, Hartge P, Jee SH, Lam TH, Peto R, Potter JD, Willett WC, Thompson SG, Danesh J, Hu FB; Global BMI Mortality Collaboration. Body-mass index and all-cause mortality: individual-participant-data meta-analysis of 239 prospective studies in four continents. Lancet . 2016; 388( 10046): 776– 786. Google Scholar CrossRef Search ADS PubMed  4. Alberti KGMM, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato KA, Fruchart J-C, James WPT, Loria CM, Smith SC, Jr; International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; International Association for the Study of Obesity. Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation . 2009; 120( 16): 1640– 1645. Google Scholar CrossRef Search ADS PubMed  5. Fasshauer M, Blüher M. Adipokines in health and disease. Trends Pharmacol Sci . 2015; 36( 7): 461– 470. Google Scholar CrossRef Search ADS PubMed  6. Flehmig G, Scholz M, Klöting N, Fasshauer M, Tönjes A, Stumvoll M, Youn B-S, Blüher M. Identification of adipokine clusters related to parameters of fat mass, insulin sensitivity and inflammation. PLoS One . 2014; 9( 6): e99785. Google Scholar CrossRef Search ADS PubMed  7. Spranger J, Kroke A, Möhlig M, Bergmann MM, Ristow M, Boeing H, Pfeiffer AF. Adiponectin and protection against type 2 diabetes mellitus. Lancet . 2003; 361( 9353): 226– 228. Google Scholar CrossRef Search ADS PubMed  8. Lehr S, Hartwig S, Sell H. Adipokines: a treasure trove for the discovery of biomarkers for metabolic disorders. Proteomics Clin Appl . 2012; 6( 1–2): 91– 101. Google Scholar CrossRef Search ADS PubMed  9. Ebert T, Kralisch S, Wurst U, Scholz M, Stumvoll M, Kovacs P, Fasshauer M, Tönjes A. Association of metabolic parameters and rs726344 in FNDC5 with serum irisin concentrations. Int J Obes . 2016; 40( 2): 260– 265. 10. Ebert T, Kralisch S, Loessner U, Jessnitzer B, Stumvoll M, Fasshauer M, Tönjes A. Relationship between serum levels of angiopoietin-related growth factor and metabolic risk factors. Horm Metab Res . 2014; 46( 10): 685– 690. Google Scholar CrossRef Search ADS PubMed  11. Tönjes A, Zeggini E, Kovacs P, Böttcher Y, Schleinitz D, Dietrich K, Morris AP, Enigk B, Rayner NW, Koriath M, Eszlinger M, Kemppinen A, Prokopenko I, Hoffmann K, Teupser D, Thiery J, Krohn K, McCarthy MI, Stumvoll M. Association of FTO variants with BMI and fat mass in the self-contained population of Sorbs in Germany. Eur J Hum Genet . 2010; 18( 1): 104– 110. Google Scholar CrossRef Search ADS PubMed  12. Rohde K, Keller M, Horstmann A, Liu X, Eichelmann F, Stumvoll M, Villringer A, Kovacs P, Tönjes A, Böttcher Y. Role of genetic variants in ADIPOQ in human eating behavior. Genes Nutr . 2015; 10( 1): 449. Google Scholar CrossRef Search ADS PubMed  13. Kralisch S, Tönjes A, Krause K, Richter J, Lossner U, Kovacs P, Ebert T, Blüher M, Stumvoll M, Fasshauer M. Fibroblast growth factor-21 serum concentrations are associated with metabolic and hepatic markers in humans. J Endocrinol . 2013; 216( 2): 135– 143. Google Scholar CrossRef Search ADS PubMed  14. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia . 1985; 28( 7): 412– 419. Google Scholar CrossRef Search ADS PubMed  15. Crisafulli A, Orrù V, Melis F, Tocco F, Concu A. Hemodynamics during active and passive recovery from a single bout of supramaximal exercise. Eur J Appl Physiol . 2003; 89( 2): 209– 216. Google Scholar CrossRef Search ADS PubMed  16. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF III, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate. Ann Intern Med . 2009; 150( 9): 604– 612. Google Scholar CrossRef Search ADS PubMed  17. Suzuki R, Shimodaira H. Pvclust: an R package for assessing the uncertainty in hierarchical clustering. Bioinformatics . 2006; 22( 12): 1540– 1542. Google Scholar CrossRef Search ADS PubMed  18. Goralski KB, McCarthy TC, Hanniman EA, Zabel BA, Butcher EC, Parlee SD, Muruganandan S, Sinal CJ. Chemerin, a novel adipokine that regulates adipogenesis and adipocyte metabolism. J Biol Chem . 2007; 282( 38): 28175– 28188. Google Scholar CrossRef Search ADS PubMed  19. Tönjes A, Fasshauer M, Kratzsch J, Stumvoll M, Blüher M. Adipokine pattern in subjects with impaired fasting glucose and impaired glucose tolerance in comparison to normal glucose tolerance and diabetes. PLoS One . 2010; 5( 11): e13911. Google Scholar CrossRef Search ADS PubMed  20. Ebert T, Roth I, Richter J, Tönjes A, Kralisch S, Lossner U, Kratzsch J, Blüher M, Stumvoll M, Fasshauer M. Different associations of adipokines in lean and healthy adults. Horm Metab Res . 2014; 46( 1): 41– 47. Google Scholar PubMed  21. Lehrke M, Becker A, Greif M, Stark R, Laubender RP, von Ziegler F, Lebherz C, Tittus J, Reiser M, Becker C, Göke B, Leber AW, Parhofer KG, Broedl UC. Chemerin is associated with markers of inflammation and components of the metabolic syndrome but does not predict coronary atherosclerosis. Eur J Endocrinol . 2009; 161( 2): 339– 344. Google Scholar CrossRef Search ADS PubMed  22. Klöting N, Fasshauer M, Dietrich A, Kovacs P, Schön MR, Kern M, Stumvoll M, Blüher M. Insulin-sensitive obesity. Am J Physiol Endocrinol Metab . 2010; 299( 3): E506– E515. Google Scholar CrossRef Search ADS PubMed  23. Blüher M, Rudich A, Klöting N, Golan R, Henkin Y, Rubin E, Schwarzfuchs D, Gepner Y, Stampfer MJ, Fiedler M, Thiery J, Stumvoll M, Shai I. Two patterns of adipokine and other biomarker dynamics in a long-term weight loss intervention. Diabetes Care . 2012; 35( 2): 342– 349. Google Scholar CrossRef Search ADS PubMed  24. Baratta R, Amato S, Degano C, Farina MG, Patanè G, Vigneri R, Frittitta L. Adiponectin relationship with lipid metabolism is independent of body fat mass: evidence from both cross-sectional and intervention studies. J Clin Endocrinol Metab . 2004; 89( 6): 2665– 2671. Google Scholar CrossRef Search ADS PubMed  25. Cnop M, Havel PJ, Utzschneider KM, Carr DB, Sinha MK, Boyko EJ, Retzlaff BM, Knopp RH, Brunzell JD, Kahn SE. Relationship of adiponectin to body fat distribution, insulin sensitivity and plasma lipoproteins: evidence for independent roles of age and sex. Diabetologia . 2003; 46( 4): 459– 469. Google Scholar CrossRef Search ADS PubMed  26. Nakamura T, Kodama Y, Takano H, Umetani K, Fujioka D, Saito Y, Kawabata K, Obata JE, Kitta Y, Kobayashi T, Mende A, Kugiyama K. Increase in circulating levels of adiponectin after treatment with statin and fibrate in patients with coronary artery disease and hyperlipidemia. Atherosclerosis . 2007; 193( 2): 449– 451. Google Scholar CrossRef Search ADS PubMed  27. Zhang X, Yeung DCY, Karpisek M, Stejskal D, Zhou Z-G, Liu F, Wong RLC, Chow W-S, Tso AWK, Lam KSL, Xu A. Serum FGF21 levels are increased in obesity and are independently associated with the metabolic syndrome in humans. Diabetes . 2008; 57( 5): 1246– 1253. Google Scholar CrossRef Search ADS PubMed  28. Bobbert T, Schwarz F, Fischer-Rosinsky A, Pfeiffer AFH, Möhlig M, Mai K, Spranger J. Fibroblast growth factor 21 predicts the metabolic syndrome and type 2 diabetes in Caucasians. Diabetes Care . 2013; 36( 1): 145– 149. Google Scholar CrossRef Search ADS PubMed  29. Fisher FM, Chui PC, Antonellis PJ, Bina HA, Kharitonenkov A, Flier JS, Maratos-Flier E. Obesity is a fibroblast growth factor 21 (FGF21)-resistant state. Diabetes . 2010; 59( 11): 2781– 2789. Google Scholar CrossRef Search ADS PubMed  30. Kralisch S, Fasshauer M. Adipocyte fatty acid binding protein: a novel adipokine involved in the pathogenesis of metabolic and vascular disease? Diabetologia . 2013; 56( 1): 10– 21. Google Scholar CrossRef Search ADS PubMed  31. Ebert T, Hopf LM, Wurst U, Bachmann A, Kralisch S, Lössner U, Platz M, Kratzsch J, Stolzenburg JU, Dietel A, Grisk O, Beige J, Anders M, Bast I, Klöting N, Blüher M, Stumvoll M, Fasshauer M. Circulating adipocyte fatty acid binding protein is increased in chronic and acute renal dysfunction. Nutr Metab Cardiovasc Dis . 2014; 24( 9): 1027– 1034. Google Scholar CrossRef Search ADS PubMed  32. Zachariah JP, Quiroz R, Nelson KP, Teng Z, Keaney JF, Jr, Sullivan LM, Vasan RS. Prospective relation of circulating adipokines to incident metabolic syndrome: the Framingham Heart Study. J Am Heart Assoc . 2017; 6( 7): e004974. Google Scholar CrossRef Search ADS PubMed  33. Lee S-A, Kallianpur A, Xiang Y-B, Wen W, Cai Q, Liu D, Fazio S, Linton MF, Zheng W, Shu XO. Intra-individual variation of plasma adipokine levels and utility of single measurement of these biomarkers in population-based studies. Cancer Epidemiol Biomarkers Prev . 2007; 16( 11): 2464– 2470. Google Scholar CrossRef Search ADS PubMed  Copyright © 2018 Endocrine Society http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Clinical Endocrinology and Metabolism Oxford University Press

Loading next page...
 
/lp/ou_press/relationship-between-12-adipocytokines-and-distinct-components-of-the-XhvuXmsKxy
Publisher
Endocrine Society
Copyright
Copyright © 2018 Endocrine Society
ISSN
0021-972X
eISSN
1945-7197
D.O.I.
10.1210/jc.2017-02085
Publisher site
See Article on Publisher Site

Abstract

Abstract Objective Adipose tissue–derived signals potentially link obesity and adipose tissue dysfunction with metabolic and cardiovascular diseases. Although some adipocytokines have been closely related to metabolic and cardiovascular traits, it is unknown which adipocytokine or adipocytokine clusters serve as meaningful markers of metabolic syndrome (MS) components. Therefore, this study investigated the associations of 12 adipocytokines with components of the MS to identify the most relevant cytokines potentially related to specific metabolic profiles. Research Design and Methods Twelve cytokines [adiponectin, adipocyte fatty acid–binding protein (AFABP), angiopoietin-related growth factor, chemerin, fibroblast growth factor (FGF) 19, FGF21, FGF23, insulin-like growth factor-1, interleukin 10, irisin, progranulin, and vaspin] were quantified in a cross-sectional cohort of 1046 subjects. Hypothesis-free cluster analysis, multivariate regression analyses with parameters of the MS, and discriminant analysis were performed to assess associations and the relative importance of each cytokine for reflecting MS and its components. Results Among the studied adipocytokines, adiponectin, AFABP, chemerin, and FGF21 showed the strongest associations with MS and several MS components in discriminant analyses and multiple regression models. For certain metabolic components, these adipocytokines were better discriminators than routine metabolic markers. Other cytokines investigated in the present cohort are less able to distinguish between metabolically healthy and unhealthy subjects. Conclusions Adiponectin, AFABP, chemerin, and FGF21 showed the strongest associations with MS components in a general population, suggesting that adverse adipose tissue function is a major contributor to these metabolic abnormalities. Future prospective studies should address the question whether these adipocytokines can predict the development of metabolic disease states. Obesity is an increasing global health burden, and according to a recent meta-analysis, the number of people with obesity has more than doubled in the last three decades (1, 2). Patients with obesity and its associated comorbidities (i.e., type 2 diabetes, dyslipidemia, and hypertension) have greater mortality and shorter life expectancy (3, 4). Therefore, the metabolic syndrome (MS) as a complex of metabolic risk factors is an increasing public health and clinical problem (4). In recent years, a large number of adipocytokines have been identified as potential links between obesity and other metabolic disease states (5). However, most of the previous studies that aim to evaluate the predictive value of cytokines for metabolic and cardiovascular diseases have the following limitations: Studies investigating adipocytokines often include only a small number of subjects, are performed only in specific disease states, and consider only one rather than a combination of cytokines. For example, we previously demonstrated in a small cohort of 141 patients with obesity that distinct adipocytokine clusters are related to either body fat mass and inflammation or insulin sensitivity/hyperglycemia and lipid metabolism, respectively (6). However, the study investigated a cohort with specific disease states and limited case numbers. Thus, it still remains unclear which adipocytokines are associated with specific facets of the MS (i.e., visceral obesity, hypertension, dyslipidemia, and insulin resistance). The clinical relevance of adipocytokine measurements to estimate the type 2 diabetes risk has been proven for adiponectin (7). However, adipocytokine serum concentrations are still not used in clinical practice to diagnose or predict metabolic and cardiovascular diseases. Adipose tissue probably secretes >600 adipocytokines (8). With the expanding number of identified adipocytokines there is an increasing need to define their function and potential clinical relevance for metabolic diseases. To overcome limitations of previous studies, we aimed to investigate a panel of 12 adipocytokines [i.e., adiponectin, adipocyte fatty acid–binding protein (AFABP), chemerin, progranulin, irisin, fibroblast growth factor (FGF) 19, FGF21, FGF23, vaspin, angiopoietin-related growth factor (AGF), insulin-like growth factor (IGF)-1, and interleukin (IL) 10] in a cross-sectional cohort of 1046 extensively phenotyped subjects not specifically selected for metabolic or cardiovascular disease states. We characterized each protein with a stepwise statistical approach by performing hypothesis-free cluster analysis with metabolic markers and other cytokines in the entire cohort, performing linear regression analysis with components of the MS in a healthy subcohort, and investigating the relative importance of each cytokine for distinguishing components of the MS in all participants. Materials and Methods Study participants The design of the current study has been described previously (9–13). Briefly, all subjects recruited for this study are part of a sample from a self-contained population of Sorbs in eastern Germany and were enrolled in the study between 2005 and 2007. Participants were not specifically selected for metabolic or cardiovascular disease states. For the present analysis, 1046 Sorbs were available. All investigations were performed by trained staff and included standardized questionnaires, anthropometric parameters [body mass index (BMI), waist/hip ratio, body impedance analysis], and a 75-g oral glucose tolerance test. Body composition analysis was performed with the BIA-2000-S device (Data Input GmbH, Darmstadt, Germany) and analyzed with the software Nutri3 (Data Input GmbH). MS and its components, were diagnosed according to the Joint Scientific Statement on Harmonizing the Metabolic Syndrome (4). Homeostasis model assessment of insulin resistance (14) and mean arterial blood pressure (BP) (15) were calculated as previously described. After assessment of serum creatinine, estimated glomerular filtration rate was assessed with the Chronic Kidney Disease Epidemiology Collaboration equation (16). Spot urine specimens were analyzed for urinary albumin and creatinine, and the albumin/creatinine ratio was calculated. A 10-MHz ultrasound sensor (GE Healthcare, Inc., München, Germany) was used to measure the intima-media thickness (IMT) of the common carotid arteries. After three measurements, IMT values on each side were averaged and mean IMT was calculated. All subjects in this study, which was approved by the ethics committee of the University of Leipzig (Reg. No. 088-2005), gave written informed consent before taking part in the study. Assays In all subjects, blood samples were taken in the morning after an overnight fast and were immediately spun and frozen at −80°C until analyses were performed. Serum (adiponectin, AFABP, AGF, chemerin, FGF19, FGF21, irisin, IGF-1, IL10, progranulin, vaspin) and plasma (FGF23) concentrations of all cytokines under investigation were determined with commercially available enzyme-linked immunosorbent assays according to the manufacturers’ instructions (AGF, irisin, progranulin, and vaspin: AdipoGen Inc., Seoul, South Korea; adiponectin, AFABP, chemerin, FGF19, and FGF21: BioVendor Inc., Brno, Czech Republic; FGF23: Immutopics Inc., San Clemente, CA; IGF-1: Liaison, DiaSorin, Saluggia, Italy; IL10: BD Bioscience, Heidelberg, Germany). Fasting insulin was determined with the AutoDELFIA insulin assay (PerkinElmer Life and Analytical Sciences, Turku, Finland). Serum creatinine was measured with the kinetic enzymatic method. Urinary albumin was determined with a turbidimetric assay for the Roche modular automated analyzer (Roche Diagnostics, Mannheim, Germany). Urinary creatinine was quantified by a photometric assay for the Roche/Hitachi cobas c automated analyzer (Roche Diagnostics). Fasting glucose (FG), glucose levels during the oral glucose tolerance test, glycated hemoglobin A1c, total cholesterol, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol, triglycerides (TG), and C-reactive protein were measured by standard laboratory methods in a certified laboratory (University of Leipzig, Institute of Laboratory Medicine). Statistical analysis For statistical analysis, SPSS software version 24.0 (IBM, Armonk, NY) and the statistical software package “R” [www.r-project.org (17)] were used. In a first step, we performed hypothesis-free and cluster analysis to assess proximity of cytokines with anthropometric and metabolic markers. Before cluster analysis, all variables were sex adjusted. Cluster analysis was performed by Ward’s minimum variance method with the function “pvclust” of the statistical software package “R” [www.r-project.org (17)], and 10,000 bootstrapping replications were analyzed. Approximately unbiased (AU) P values and bootstrap probability values of the branching points in our cluster diagrams were calculated as previously described (6). AU P values ≥90% were considered strong evidence for the respective cluster. Next, multiple linear regression analysis was carried out in healthy subjects to identify independent associations of adipocytokines with components of the metabolic syndrome. For this purpose, only participants without a diagnosis or treatment of diabetes mellitus, hypertension, or impaired TG or HDL-C were included in each model, respectively. Thus, all subjects with a diagnosis of diabetes mellitus or on antidiabetic treatment were excluded from the model for FG. For mean arterial BP, all participants with an established diagnosis of hypertension or treatment were excluded. Furthermore, all subjects treated with peroxisome proliferator–activated receptor α-activators were not included in the multivariate model for TG. For HDL-C, all subjects on niacin treatment were excluded before multiple linear regression analysis. Before we performed multiple linear regression analyses, normality of the distribution of residuals was assessed by visual inspection of P-P plots. Homoscedasticity was assessed by visual inspection of residual plots, revealing no signs of strong heteroscedasticity (data not shown). No formal testing was applied to avoid sequential testing situations. To assess the degree of multicollinearity between independent variables, variance inflation factors were calculated for each model. Finally, discriminant analysis was performed to identify the cytokines most relevant for discriminating the MS and its components: impaired FG [FG ≥100 mg/dL (5.56 mmol/L) or treatment], elevated BP (systolic ≥130 or diastolic ≥85 mm Hg or treatment), elevated TG [≥150 mg/dL (1.7 mmol/L) or treatment], low HDL-C [<40 mg/dL (1.0 mmol/L) in men, <50 mg/dL (1.3 mmol/L) in women or treatment], and abdominal obesity (waist circumference ≥94 cm in men, ≥80 cm in women) (4). Components of the MS were transformed to dichotomous variables before analysis according to Alberti et al. (4). Discriminant analysis was then carried out with log-transformed, sex-adjusted, unstandardized residuals. We calculated two models consisting of adipocytokines only (model 1) or all adipocytokines and metabolic markers used for the definition of the MS (model 2). A P value of <0.05 was considered statistically significant in all analyses. Results Baseline characteristics of the entire study cohort Baseline characteristics of the study population are shown in Supplemental Table 1. Median (interquartile range) age was 48.3 (36.5 to 60.2) years and BMI was 26.4 (23.2 to 29.6) kg/m2. In the total cohort, 376 subjects (36.2%) presented with a MS as defined by Alberti et al. (4). Adipocytokine quartiles for each component of the MS are depicted in Table 1. Table 1. Quartiles of Circulating Levels of All Adipocytokines Depending on Components of the MS in the Entire Study Cohort   Quartile  Adiponectin (mg/L)  AFABP (µg/L)  AGF (µg/L)  Chemerin (µg/L)  FGF19 (ng/L)  FGF21 (ng/L)  FGF23 (RU/mL)  IGF-1 (µg/L)  IL10 (ng/L)  Irisin (mg/L)  Progranulin (µg/L)  Vaspin (µg/L)  N    971  1011  828  1009  1004  931  1039  1043  996  1039  1042  1041  FG (N = 1030)  I  18.1  14.0  40.3  108.2  255.0  39.9  70.3  179.5  4.4  0.83  102.9  0.7  II  16.2  15.9  34.7  110.1  240.4  60.0  71.0  160.2  5.0  0.83  106.9  0.5  III  16.4  18.5  38.5  120.2  247.3  80.0  68.6  151.3  4.7  0.76  110.2  0.4  IV  15.5  25.6  43.9  133.5  244.5  119.9  72.5  128.5  5.0  0.77  114.4  0.5  Mean BP (N = 1043)  I  16.5  14.0  36.3  105.7  242.6  41.2  68.2  181.6  5.3  0.81  102.9  0.6  II  17.1  15.9  38.1  112.7  254.7  63.4  71.2  164.9  4.7  0.79  107.8  0.5  III  15.6  19.3  39.1  120.4  217.9  91.2  70.2  145.5  4.7  0.82  108.5  0.4  IV  16.1  23.9  40.3  133.3  253.3  119.7  72.6  129.3  4.8  0.76  112.1  0.5  TG (N = 1044)  I  18.1  13.8  38.8  102.0  243.7  36.5  67.2  169.0  4.9  0.83  103.7  0.4  II  17.6  16.3  38.9  115.4  243.0  58.0  70.1  153.0  4.7  0.81  107.5  0.5  III  16.0  19.7  39.1  124.8  239.6  87.3  72.0  156.7  4.5  0.80  107.9  0.6  IV  14.3  22.3  37.1  127.2  254.2  151.1  72.2  139.2  5.0  0.77  114.3  0.5  HDL-C (N = 1044)  I  13.3  20.7  43.8  122.6  222.2  114.4  72.0  151.1  5.3  0.77  108.5  0.4  II  15.5  17.0  40.5  121.9  255.2  83.2  71.8  162.3  4.9  0.77  109.8  0.5  III  17.1  17.0  36.7  114.5  238.1  57.1  68.8  155.2  4.6  0.81  108.0  0.5  IV  19.7  16.3  35.5  110.4  259.3  51.3  70.2  156.7  4.5  0.83  105.0  0.5  Waist (N = 1039)  I  16.6  10.0  37.8  105.6  267.6  50.8  70.2  180.8  4.7  0.80  107.2  0.5  II  16.5  14.4  35.7  111.9  233.4  63.5  67.1  158.9  4.9  0.81  104.8  0.5  III  16.0  20.3  37.8  121.5  245.7  82.6  71.5  148.2  5.3  0.79  108.1  0.4  IV  16.2  33.2  44.4  133.2  238.1  97.2  73.9  129.8  4.3  0.80  112.3  0.5    Quartile  Adiponectin (mg/L)  AFABP (µg/L)  AGF (µg/L)  Chemerin (µg/L)  FGF19 (ng/L)  FGF21 (ng/L)  FGF23 (RU/mL)  IGF-1 (µg/L)  IL10 (ng/L)  Irisin (mg/L)  Progranulin (µg/L)  Vaspin (µg/L)  N    971  1011  828  1009  1004  931  1039  1043  996  1039  1042  1041  FG (N = 1030)  I  18.1  14.0  40.3  108.2  255.0  39.9  70.3  179.5  4.4  0.83  102.9  0.7  II  16.2  15.9  34.7  110.1  240.4  60.0  71.0  160.2  5.0  0.83  106.9  0.5  III  16.4  18.5  38.5  120.2  247.3  80.0  68.6  151.3  4.7  0.76  110.2  0.4  IV  15.5  25.6  43.9  133.5  244.5  119.9  72.5  128.5  5.0  0.77  114.4  0.5  Mean BP (N = 1043)  I  16.5  14.0  36.3  105.7  242.6  41.2  68.2  181.6  5.3  0.81  102.9  0.6  II  17.1  15.9  38.1  112.7  254.7  63.4  71.2  164.9  4.7  0.79  107.8  0.5  III  15.6  19.3  39.1  120.4  217.9  91.2  70.2  145.5  4.7  0.82  108.5  0.4  IV  16.1  23.9  40.3  133.3  253.3  119.7  72.6  129.3  4.8  0.76  112.1  0.5  TG (N = 1044)  I  18.1  13.8  38.8  102.0  243.7  36.5  67.2  169.0  4.9  0.83  103.7  0.4  II  17.6  16.3  38.9  115.4  243.0  58.0  70.1  153.0  4.7  0.81  107.5  0.5  III  16.0  19.7  39.1  124.8  239.6  87.3  72.0  156.7  4.5  0.80  107.9  0.6  IV  14.3  22.3  37.1  127.2  254.2  151.1  72.2  139.2  5.0  0.77  114.3  0.5  HDL-C (N = 1044)  I  13.3  20.7  43.8  122.6  222.2  114.4  72.0  151.1  5.3  0.77  108.5  0.4  II  15.5  17.0  40.5  121.9  255.2  83.2  71.8  162.3  4.9  0.77  109.8  0.5  III  17.1  17.0  36.7  114.5  238.1  57.1  68.8  155.2  4.6  0.81  108.0  0.5  IV  19.7  16.3  35.5  110.4  259.3  51.3  70.2  156.7  4.5  0.83  105.0  0.5  Waist (N = 1039)  I  16.6  10.0  37.8  105.6  267.6  50.8  70.2  180.8  4.7  0.80  107.2  0.5  II  16.5  14.4  35.7  111.9  233.4  63.5  67.1  158.9  4.9  0.81  104.8  0.5  III  16.0  20.3  37.8  121.5  245.7  82.6  71.5  148.2  5.3  0.79  108.1  0.4  IV  16.2  33.2  44.4  133.2  238.1  97.2  73.9  129.8  4.3  0.80  112.3  0.5  Participants were grouped into quartiles I to IV for each component of the MS, (i.e., FG, mean BP, TG, HDL-C, and waist circumference). For all components of the MS, median circulating levels of all cytokines according to each quartile are depicted. Numbers of subjects included are shown for each metabolic component and adipocytokine. View Large Cluster analysis of all adipocytokines and cardiometabolic markers Unsupervised and sex-adjusted cluster analysis of adipocytokines and anthropometric and cardiometabolic markers revealed four clusters in the entire cohort (Fig. 1): age, mean IMT, and waist/hip ratio (cluster 1); AFABP, BMI, and fat mass (cluster 2); estimated glomerular filtration rate and IGF-1 (cluster 3); and HDL-C and adiponectin (cluster 4), were significantly clustered (AU P values ≥90%). Figure 1. View largeDownload slide Cluster analysis regarding metabolic cytokines and anthropometric and metabolic markers. Before cluster analysis, all variables were sex adjusted. Four clusters of strong correlations could be detected. Cluster analysis was performed by Ward’s minimum variance method, and 10,000 bootstrapping replications were analyzed. AU (numbers in red) P values and bootstrap probability (numbers in green) values of the branching points were calculated. AU P values ≥90% were considered strong evidence for the respective cluster. In the cluster analysis consisting of all variables, 595 subjects were included. ACR, albumin/creatinine ratio; eGFR, estimated glomerular filtration rate; LDL-C, low density lipoprotein cholesterol; HbA1c, glycated hemoglobin A1c; HOMA-IR, homeostasis model assessment of insulin resistance; WHR, waist/hip ratio. Figure 1. View largeDownload slide Cluster analysis regarding metabolic cytokines and anthropometric and metabolic markers. Before cluster analysis, all variables were sex adjusted. Four clusters of strong correlations could be detected. Cluster analysis was performed by Ward’s minimum variance method, and 10,000 bootstrapping replications were analyzed. AU (numbers in red) P values and bootstrap probability (numbers in green) values of the branching points were calculated. AU P values ≥90% were considered strong evidence for the respective cluster. In the cluster analysis consisting of all variables, 595 subjects were included. ACR, albumin/creatinine ratio; eGFR, estimated glomerular filtration rate; LDL-C, low density lipoprotein cholesterol; HbA1c, glycated hemoglobin A1c; HOMA-IR, homeostasis model assessment of insulin resistance; WHR, waist/hip ratio. Linear regression analysis with components of the MS We assessed four different regression models to analyze the associations between the adipocytokines and different components of the MS (Table 2). In each model, only the subgroup of subjects without diagnosed alteration or drug treatment of the respective component of the MS were included (i.e., FG, BP, TG, and HDL-C). Evaluation of the variance inflation factors did not reveal a high degree of multicollinearity (data not shown). In the multivariate model consisting of all cytokines, age, sex, and fat mass, circulating chemerin levels were positively associated with FG (β = 0.109; P = 0.005). Furthermore, serum chemerin (β = 0.141; P = 0.001) and FGF21 (β = 0.092; P = 0.020) were predictors of higher BP. Elevated TG levels were significantly associated with lower adiponectin (β = −0.161; P < 0.001) and irisin (β = −0.099; P = 0.002) concentrations. In contrast, chemerin (β = 0.171; P < 0.001), FGF21 (β = 0.246; P < 0.001), progranulin (β = 0.069; P = 0.037), and vaspin (β = 0.138; P < 0.001) positively correlated with TG. Finally, higher HDL-C was significantly associated with higher adiponectin (β = 0.304; P < 0.001) and IL10 (β = 0.095; P = 0.004), whereas AGF (β = −0.113; P = 0.001), chemerin (β = −0.106; P = 0.004), FGF21 (β = −0.081; P = 0.018), and IGF-1 (β = −0.141; P = 0.001) were predictors of decreased HDL-C, respectively (Table 2). Table 2. Multivariate Linear Regression Analyses Between Components of the MS and Adipocytokines Dependent Variable  Covariates  Independent Variables  Model 1: Single Cytokine  Model 2: All Cytokines  N  r2  β  P  r2  β  P  FG  Age, sex, fat mass  Adiponectin, mg/L  0.346  −0.041  0.180  0.383  −0.032  0.387  570  AFABP, µg/L  0.344  0.070  0.059  −0.005  0.917  AGF, µg/L  0.337  −0.023  0.460  −0.023  0.494  Chemerin, µg/L  0.344  0.076  0.010  0.109  0.005  FGF19, ng/L  0.338  −0.019  0.501  −0.050  0.151  FGF21, ng/L  0.342  −0.002  0.943  0.007  0.836  FGF23, RU/mL  0.341  −0.02  0.461  −0.041  0.229  IGF-1, µg/L  0.340  0.042  0.241  0.072  0.120  IL10, ng/L  0.350  0.003  0.922  −0.034  0.318  Irisin, mg/L  0.346  −0.082  0.003  −0.017  0.616  Progranulin, µg/L  0.342  0.049  0.075  0.034  0.337  Vaspin, µg/L  0.342  −0.011  0.704  −0.031  0.397  Mean BP  Age, sex, fat mass  Adiponectin, mg/L  0.257  −0.044  0.214  0.355  0.015  0.710  491  AFABP, µg/L  0.259  0.052  0.201  0.007  0.888  AGF, µg/L  0.295  −0.022  0.527  0.009  0.804  Chemerin, µg/L  0.270  0.123  <0.001  0.141  0.001  FGF19, ng/L  0.255  0.044  0.162  0.033  0.390  FGF21, ng/L  0.277  0.112  0.001  0.092  0.020  FGF23, RU/mL  0.256  0.002  0.943  −0.035  0.368  IGF-1, µg/L  0.259  −0.045  0.287  −0.026  0.610  IL10, ng/L  0.262  −0.026  0.419  −0.022  0.565  Irisin, mg/L  0.258  0.004  0.893  −0.041  0.277  Progranulin, µg/L  0.263  0.077  0.013  0.047  0.231  Vaspin, µg/L  0.262  0.067  0.045  0.057  0.158  TG (ln)  Age, sex, fat mass  Adiponectin, mg/L  0.228  −0.158  <0.001  0.386  −0.161  <0.001  639  AFABP, µg/L  0.219  0.152  <0.001  0.035  0.436  AGF, µg/L  0.215  −0.041  0.199  −0.020  0.537  Chemerin, µg/L  0.241  0.196  <0.001  0.171  <0.001  FGF19, ng/L  0.209  0.043  0.133  0.004  0.904  FGF21, ng/L  0.277  0.273  <0.001  0.246  <0.001  FGF23, RU/mL  0.210  0.001  0.962  −0.016  0.614  IGF-1, µg/L  0.208  0.003  0.929  0.031  0.478  IL10, ng/L  0.212  −0.032  0.261  −0.042  0.183  Irisin, mg/L  0.212  −0.087  0.002  −0.099  0.002  Progranulin, µg/L  0.218  0.099  <0.001  0.069  0.037  Vaspin, µg/L  0.226  0.145  <0.001  0.138  <0.001  HDL-C  Age, sex, fat mass  Adiponectin, mg/L  0.282  0.261  <0.001  0.366  0.304  <0.001  640  AFABP, µg/L  0.228  −0.058  0.121  0.043  0.351  AGF, µg/L  0.240  −0.104  0.001  −0.113  0.001  Chemerin, µg/L  0.232  −0.076  0.012  −0.106  0.004  FGF19, ng/L  0.231  0.056  0.048  0.038  0.241  FGF21, ng/L  0.223  −0.062  0.037  −0.081  0.018  FGF23, RU/mL  0.233  −0.054  0.054  −0.046  0.158  IGF-1, µg/L  0.242  −0.151  <0.001  −0.141  0.001  IL10, ng/L  0.233  0.074  0.008  0.095  0.004  Irisin, mg/L  0.232  0.041  0.134  0.048  0.141  Progranulin, µg/L  0.230  −0.038  0.172  −0.057  0.092  Vaspin, µg/L  0.230  −0.010  0.740  0.048  0.172  Dependent Variable  Covariates  Independent Variables  Model 1: Single Cytokine  Model 2: All Cytokines  N  r2  β  P  r2  β  P  FG  Age, sex, fat mass  Adiponectin, mg/L  0.346  −0.041  0.180  0.383  −0.032  0.387  570  AFABP, µg/L  0.344  0.070  0.059  −0.005  0.917  AGF, µg/L  0.337  −0.023  0.460  −0.023  0.494  Chemerin, µg/L  0.344  0.076  0.010  0.109  0.005  FGF19, ng/L  0.338  −0.019  0.501  −0.050  0.151  FGF21, ng/L  0.342  −0.002  0.943  0.007  0.836  FGF23, RU/mL  0.341  −0.02  0.461  −0.041  0.229  IGF-1, µg/L  0.340  0.042  0.241  0.072  0.120  IL10, ng/L  0.350  0.003  0.922  −0.034  0.318  Irisin, mg/L  0.346  −0.082  0.003  −0.017  0.616  Progranulin, µg/L  0.342  0.049  0.075  0.034  0.337  Vaspin, µg/L  0.342  −0.011  0.704  −0.031  0.397  Mean BP  Age, sex, fat mass  Adiponectin, mg/L  0.257  −0.044  0.214  0.355  0.015  0.710  491  AFABP, µg/L  0.259  0.052  0.201  0.007  0.888  AGF, µg/L  0.295  −0.022  0.527  0.009  0.804  Chemerin, µg/L  0.270  0.123  <0.001  0.141  0.001  FGF19, ng/L  0.255  0.044  0.162  0.033  0.390  FGF21, ng/L  0.277  0.112  0.001  0.092  0.020  FGF23, RU/mL  0.256  0.002  0.943  −0.035  0.368  IGF-1, µg/L  0.259  −0.045  0.287  −0.026  0.610  IL10, ng/L  0.262  −0.026  0.419  −0.022  0.565  Irisin, mg/L  0.258  0.004  0.893  −0.041  0.277  Progranulin, µg/L  0.263  0.077  0.013  0.047  0.231  Vaspin, µg/L  0.262  0.067  0.045  0.057  0.158  TG (ln)  Age, sex, fat mass  Adiponectin, mg/L  0.228  −0.158  <0.001  0.386  −0.161  <0.001  639  AFABP, µg/L  0.219  0.152  <0.001  0.035  0.436  AGF, µg/L  0.215  −0.041  0.199  −0.020  0.537  Chemerin, µg/L  0.241  0.196  <0.001  0.171  <0.001  FGF19, ng/L  0.209  0.043  0.133  0.004  0.904  FGF21, ng/L  0.277  0.273  <0.001  0.246  <0.001  FGF23, RU/mL  0.210  0.001  0.962  −0.016  0.614  IGF-1, µg/L  0.208  0.003  0.929  0.031  0.478  IL10, ng/L  0.212  −0.032  0.261  −0.042  0.183  Irisin, mg/L  0.212  −0.087  0.002  −0.099  0.002  Progranulin, µg/L  0.218  0.099  <0.001  0.069  0.037  Vaspin, µg/L  0.226  0.145  <0.001  0.138  <0.001  HDL-C  Age, sex, fat mass  Adiponectin, mg/L  0.282  0.261  <0.001  0.366  0.304  <0.001  640  AFABP, µg/L  0.228  −0.058  0.121  0.043  0.351  AGF, µg/L  0.240  −0.104  0.001  −0.113  0.001  Chemerin, µg/L  0.232  −0.076  0.012  −0.106  0.004  FGF19, ng/L  0.231  0.056  0.048  0.038  0.241  FGF21, ng/L  0.223  −0.062  0.037  −0.081  0.018  FGF23, RU/mL  0.233  −0.054  0.054  −0.046  0.158  IGF-1, µg/L  0.242  −0.151  <0.001  −0.141  0.001  IL10, ng/L  0.233  0.074  0.008  0.095  0.004  Irisin, mg/L  0.232  0.041  0.134  0.048  0.141  Progranulin, µg/L  0.230  −0.038  0.172  −0.057  0.092  Vaspin, µg/L  0.230  −0.010  0.740  0.048  0.172  Model 1 is calculated for each adipocytokine separately, adjusted for age, sex, and fat mass. In model 2, all adipocytokines, as well as age, sex, and fat mass, are included as independent variables. The r2 of each model, standardized β, and P values are depicted, and significant predictors for each metabolic component are marked in bold. For each component of the MS, only participants without a diagnosis or treatment of the respective component were included. Numbers of subjects included in model 2 for each metabolic component are depicted. Normal distribution was assessed by visual inspection, and non-normally distributed variables were logarithmically transformed (ln) before testing. View Large Discriminant analysis of the MS and its components by adipocytokines To address the question which adipocytokines are most appropriate to distinguish the presence or absence of a certain component of the MS, as well as the MS itself, we performed sex-adjusted descriptive discriminant analyses of all cytokines. Based on the standardized canonical discriminant coefficients, AFABP, chemerin, and FGF21 were the three most prominent markers for impaired FG (Table 3, model 1). For hypertension, FGF21, IGF-1, and AFABP were most relevant (Table 3, model 1). The adipocytokines FGF21, adiponectin, and AFABP had the highest relative importance for hypertriglyceridemia (Table 3, model 1). Furthermore, adiponectin, FGF21, and FGF23 were the most relevant cytokines for reduced HDL-C levels (Table 3, model 1). Moreover, AFABP, IGF-1, and adiponectin showed highest standardized canonical discriminant coefficients for increased waist circumference (Table 3, model 1). Finally, the adipocytokines AFABP, FGF21, and chemerin were highly relevant to distinguish between the presence or absence of the MS (Table 3, model 1). Table 3. Discriminant Analysis of 12 Adipocytokines and Components of the MS and the MS   Components of MS  MS    FG  Hypertension  TG  HDL-C  Waist        Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  N  648  646  648  636  648  635  648  635  647  635  647  635  Standardized canonical discriminant coefficients                           FG    —    0.127    0.131    0.041    0.200    0.363   Mean BP    0.267    —    0.286    −0.320    0.338    0.283   TG    −0.077    0.151    —    0.466    0.064    0.303   HDL-C    −0.206    0.213    −0.346    —    −0.321    −0.111   Waist    0.498    0.596    0.217    0.598    —    0.486   Adiponectin  0.056  0.175  −0.094  −0.048  −0.313  −0.142  −0.689  −0.457  −0.214  −0.062  −0.208  −0.005   AFABP  0.565  0.200  0.350  0.004  0.305  0.041  0.230  −0.105  0.692  0.516  0.514  0.030   AGF  0.001  −0.060  0.011  −0.019  0.050  −0.019  0.092  0.016  0.096  0.044  0.152  0.043   Chemerin  0.348  0.252  0.260  0.230  0.214  0.098  0.123  0.093  0.098  −0.004  0.325  0.168   FGF19  −0.141  −0.045  −0.168  −0.101  −0.060  −0.006  −0.199  −0.096  −0.126  −0.079  −0.137  −0.029   FGF21  0.293  0.177  0.381  0.270  0.532  0.362  0.323  0.092  0.169  0.019  0.447  0.170   FGF23  −0.252  −0.203  −0.075  0.003  −0.031  −0.018  0.277  0.190  −0.080  −0.049  −0.174  −0.091   IGF-1  −0.141  −0.013  −0.352  −0.154  0.005  0.036  0.024  0.130  −0.333  −0.286  −0.048  0.127   IL10  −0.051  −0.058  −0.051  −0.058  −0.077  −0.065  0.216  0.145  0.050  0.088  −0.063  −0.021   Irisin  0.028  0.065  −0.093  −0.040  −0.295  −0.233  −0.113  −0.011  −0.085  −0.049  −0.110  −0.043   Progranulin  0.132  0.058  0.174  0.131  0.231  0.117  0.093  0.020  0.095  −0.002  0.126  −0.033   Vaspin  0.044  0.070  −0.008  −0.005  0.283  0.257  −0.149  −0.201  −0.045  −0.063  0.093  0.051  Rc2  0.216  0.283  0.183  0.240  0.207  0.266  0.085  0.121  0.328  0.398  0.324  0.543  P  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001    Components of MS  MS    FG  Hypertension  TG  HDL-C  Waist        Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  N  648  646  648  636  648  635  648  635  647  635  647  635  Standardized canonical discriminant coefficients                           FG    —    0.127    0.131    0.041    0.200    0.363   Mean BP    0.267    —    0.286    −0.320    0.338    0.283   TG    −0.077    0.151    —    0.466    0.064    0.303   HDL-C    −0.206    0.213    −0.346    —    −0.321    −0.111   Waist    0.498    0.596    0.217    0.598    —    0.486   Adiponectin  0.056  0.175  −0.094  −0.048  −0.313  −0.142  −0.689  −0.457  −0.214  −0.062  −0.208  −0.005   AFABP  0.565  0.200  0.350  0.004  0.305  0.041  0.230  −0.105  0.692  0.516  0.514  0.030   AGF  0.001  −0.060  0.011  −0.019  0.050  −0.019  0.092  0.016  0.096  0.044  0.152  0.043   Chemerin  0.348  0.252  0.260  0.230  0.214  0.098  0.123  0.093  0.098  −0.004  0.325  0.168   FGF19  −0.141  −0.045  −0.168  −0.101  −0.060  −0.006  −0.199  −0.096  −0.126  −0.079  −0.137  −0.029   FGF21  0.293  0.177  0.381  0.270  0.532  0.362  0.323  0.092  0.169  0.019  0.447  0.170   FGF23  −0.252  −0.203  −0.075  0.003  −0.031  −0.018  0.277  0.190  −0.080  −0.049  −0.174  −0.091   IGF-1  −0.141  −0.013  −0.352  −0.154  0.005  0.036  0.024  0.130  −0.333  −0.286  −0.048  0.127   IL10  −0.051  −0.058  −0.051  −0.058  −0.077  −0.065  0.216  0.145  0.050  0.088  −0.063  −0.021   Irisin  0.028  0.065  −0.093  −0.040  −0.295  −0.233  −0.113  −0.011  −0.085  −0.049  −0.110  −0.043   Progranulin  0.132  0.058  0.174  0.131  0.231  0.117  0.093  0.020  0.095  −0.002  0.126  −0.033   Vaspin  0.044  0.070  −0.008  −0.005  0.283  0.257  −0.149  −0.201  −0.045  −0.063  0.093  0.051  Rc2  0.216  0.283  0.183  0.240  0.207  0.266  0.085  0.121  0.328  0.398  0.324  0.543  P  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  <0.001  Impaired FG was defined by FG ≥100 mg/dL (5.56 mmol/L) or treatment; hypertension by systolic BP ≥130 or diastolic BP ≥85 mm Hg or treatment; elevated TG by TG ≥150 mg/dL (1.7 mmol/L) or treatment; low HDL-C by <40 mg/dL (1.0 mmol/L) in men, <50 mg/dL (1.3 mmol/L) in women or treatment; and abdominal obesity by waist circumference ≥94 cm in men, ≥80 cm in women (4). Numbers of subjects included in each model, standardized discriminant regression coefficients, squared canonical correlation ( Rc2) test statistic, and P values are depicted. Model 1 includes only adipocytokines; model 2 consists of all adipocytokines, as well as the components of the MS. The three highest standardized discriminant regression coefficients for each metabolic component are marked in bold. Components of the MS were transformed to dichotomous variables before analysis and reference population in all models was the healthy group. All variables were logarithmically transformed and sex adjusted before testing. View Large Waist circumference appeared to be the strongest discriminator for components of the MS (Table 3, model 2). In the ranking of the standardized canonical discriminant coefficients, certain adipocytokines were better discriminators than routine metabolic markers. Thus, FGF21 is the strongest discriminator for elevated TG even when markers used for the definition of the MS were included in the model (Table 3, model 2). In contrast, AGF, FGF19, IL10, irisin, progranulin, and vaspin are less potent to discriminate between metabolically healthy and unhealthy subjects (Table 3, models 1 and 2). Discussion In the current study we investigated the associations of 12 adipocytokines with components of the MS (i.e., impaired FG, hypertension, dyslipidemia, and visceral obesity) and with the MS itself. We used hypothesis-free cluster analysis, multivariate regression models, and discriminant analyses to assess the relations and the relevance of distinct cytokines with different anthropometric and biochemical measures of the MS. In metabolically healthy subjects, the adipocytokines chemerin, FGF21, and adiponectin showed the highest associations as assessed by standardized β with the respective metabolic outcome parameters (i.e., FG, BP, TG, and HDL-C). Thus, in patients not affected by the respective component of the MS, the adipocytokine chemerin is significantly associated with an adverse metabolic status even after adjustment for fat mass. Interestingly, chemerin has been presented as an adipocytokine that influences adipocyte expression of genes involved in glucose and lipid metabolism in vitro (18). Additional studies from our group and others demonstrate that chemerin is also adversely associated with glucose homeostasis (19), body fat mass (6), inflammation (6, 20), TG (21), HDL-C (21), and BP (21) in humans with metabolic disease states, supporting our findings. Importantly, the observed associations between chemerin and cardiometabolic traits are independent of fat mass, and chemerin can distinguish certain metabolic components (i.e., FG and hypertension) besides waist circumference. These results suggest either that chemerin is not exclusively produced by adipose tissue or that impaired adipose tissue function rather than increased fat mass determines increased circulating chemerin. The latter hypothesis is supported by data from age-, sex-, and BMI-matched patients with either insulin-sensitive or insulin-resistant obesity (22). Independently of fat mass, people with insulin-resistant obesity and impaired adipose tissue function had higher serum chemerin concentrations (22). Furthermore, we have recently shown that chemerin serum concentrations closely reflect body weight dynamics in the context of different diet interventions (23). In addition to chemerin, the adipocytokines adiponectin and FGF21 are associated with distinct metabolic parameters in our linear regression analysis. Thus, high adiponectin is a predictor for low TG and higher HDL-C. Interestingly, adiponectin is also significantly clustered with HDL-C in our unbiased cluster analysis of the entire cohort. This finding is in accordance with published data obtained in obese patients undergoing weight loss intervention (24) and healthy patients (25). Treatment with statins and fibrates not only increases HDL-C but also raises adiponectin levels in patients with coronary artery disease and dyslipidemia (26). Furthermore, adiponectin is able to distinguish subjects with impaired HDL-C with similar power to TG and waist circumference. FGF21 is a marker and predictor for an adverse metabolic profile (27, 28). Zhang et al. (27) demonstrated similar regulations in a cohort of 232 Chinese subjects with a high prevalence of overweight or obesity and diabetic patients. This finding is in accordance with the hypothesis of an FGF21 resistance observed in MS and its components, including obesity (29). The results of our linear regression analysis provide further evidence that the observed associations of FGF21 might be causal and independent of fat mass even in a healthy cohort. FGF21 is the strongest discriminator for subjects with elevated TG in our discriminant analysis including all adipocytokines and metabolic markers. In our cohort, AFABP was significantly clustered with fat mass and BMI. This finding is corroborated by our regression models showing no association with facets of the MS when adjusted for fat mass. AFABP has been introduced as an adipocytokine with adverse cardiometabolic effects that is associated with components of the MS [e.g., obesity, insulin resistance, and dyslipidemia, as reviewed in (30)] and renal dysfunction (31). Among all investigated adipocytokines, AFABP also robustly distinguishes the presence or absence of the MS and its components. However, when waist circumference is included in the discriminant analysis, the ability of AFABP to distinguish metabolic components disappears. It should be noted in this context that a recent study by Zachariah et al. (32) demonstrated that AFABP is higher in metabolically healthy obese patients than in lean subjects but does not predict new-onset MS in a longitudinal analysis of the Framingham Heart Study. These data further support our results suggesting that AFABP is a surrogate parameter of fat mass rather than being causally involved in the pathogenesis of the MS. In contrast to adiponectin, AFABP, chemerin, and FGF21, other cytokines in our cohort are associated with markers of the MS in regression models but are not clustered to anthropometric or metabolic markers including irisin and vaspin, as well as AGF, IGF-1, and IL10. It should be noted that these associations are weak in our regression analyses. Furthermore, these adipocytokines have low standardized discriminant regression coefficients in discriminant analyses for components of the MS. Moreover, results of our multivariate linear regression analyses and discriminant analyses cannot be directly compared because different subsets of individuals have been included to meet the distinct questions of interest. Based on our data, AGF, IGF-1, IL10, irisin, progranulin, and vaspin are less important for the diagnosis of the MS and its components as compared with adiponectin, AFABP, chemerin, and FGF21, raising the question of whether clinical relevance can be derived. Limitations of the current study include the cross-sectional design. Therefore, no causality can be derived. Furthermore, the stability of the different adipocytokines in frozen samples could differ, and sample degradation may have contributed to variability within our results. On the other hand, all samples were analyzed in a single laboratory in one batch, and therefore different sample handling and storage should not influence our observed results. Furthermore, Lee et al. (33) suggested that plasma adipocytokine levels in general are stable and single measurements can represent cytokine levels over time in population-based studies. In addition, extensive phenotyping was performed at a high level of standardization by a trained study team. In conclusion, we have demonstrated that some adipocytokines could serve as markers for distinct metabolic disease states in a general population, including adiponectin, chemerin, and FGF21. These adipocytokines might have fat mass–independent effects on markers of lipid and glucose metabolism, as well as hypertension. Future prospective studies should address the question of whether these adipocytokines can predict metabolic disease states, as already shown for adiponectin (7), retinol-binding protein 4 (32), and fetuin-A (32), and should therefore be included in routine clinical measurement for risk stratification. Abbreviations: AFABP adipocyte fatty acid–binding protein AGF angiopoietin-related growth factor AU approximately unbiased BMI body mass index BP blood pressure FG fasting glucose FGF fibroblast growth factor HDL-C high-density lipoprotein cholesterol IGF insulin-like growth factor IL interleukin IMT intima-media thickness MS metabolic syndrome TG triglycerides. Acknowledgments We thank all who participated in the study. We especially thank Beate Gutsmann for her valuable contribution to data collection and excellent technical assistance. Financial Support: This study was supported by grants to A.T. from the Deutsche Forschungsgemeinschaft (DFG; Collaborative Research Centre 1052/1, C01 and Priority Programme 1629 TO 718/2-1), the German Diabetes Association, and the Federal Ministry of Education and Research (BMBF), Germany (Funding number: 01EO1001, Integrated Research and Treatment Center AdiposityDiseases, Postdoctoral Fellow program). Furthermore, T.E. was supported by the Federal Ministry of Education and Research, Germany (Funding number: 01EO1001, Integrated Research and Treatment Center AdiposityDiseases, MetaRot and Postdoctoral Fellow program) and by a Merck Sharp & Dohme grant (MSD Stipendium 2013 Diabetologie). P.K. was supported by the Boehringer Ingelheim Foundation. Author Contributions: T.E., C.G., and A.T. wrote the manuscript and researched data. M. Scholz contributed to statistical analyses. M. Scholz, T.W., D.S., M.F., M.B., M. Stumvoll, and P.K. contributed to the discussion and reviewed and edited the manuscript. T.E. and A.T. are the guarantors of this work and as such had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Disclosure Summary: The authors have nothing to disclose. References 1. Ng M, Fleming T, Robinson M, Thomson B, Graetz N, Margono C, Mullany EC, Biryukov S, Abbafati C, Abera SF, Abraham JP, Abu-Rmeileh NME, Achoki T, AlBuhairan FS, Alemu ZA, Alfonso R, Ali MK, Ali R, Guzman NA, Ammar W, Anwari P, Banerjee A, Barquera S, Basu S, Bennett DA, Bhutta Z, Blore J, Cabral N, Nonato IC, Chang JC, Chowdhury R, Courville KJ, Criqui MH, Cundiff DK, Dabhadkar KC, Dandona L, Davis A, Dayama A, Dharmaratne SD, Ding EL, Durrani AM, Esteghamati A, Farzadfar F, Fay DF, Feigin VL, Flaxman A, Forouzanfar MH, Goto A, Green MA, Gupta R, Hafezi-Nejad N, Hankey GJ, Harewood HC, Havmoeller R, Hay S, Hernandez L, Husseini A, Idrisov BT, Ikeda N, Islami F, Jahangir E, Jassal SK, Jee SH, Jeffreys M, Jonas JB, Kabagambe EK, Khalifa SE, Kengne AP, Khader YS, Khang YH, Kim D, Kimokoti RW, Kinge JM, Kokubo Y, Kosen S, Kwan G, Lai T, Leinsalu M, Li Y, Liang X, Liu S, Logroscino G, Lotufo PA, Lu Y, Ma J, Mainoo NK, Mensah GA, Merriman TR, Mokdad AH, Moschandreas J, Naghavi M, Naheed A, Nand D, Narayan KM, Nelson EL, Neuhouser ML, Nisar MI, Ohkubo T, Oti SO, Pedroza A, Prabhakaran D, Roy N, Sampson U, Seo H, Sepanlou SG, Shibuya K, Shiri R, Shiue I, Singh GM, Singh JA, Skirbekk V, Stapelberg NJ, Sturua L, Sykes BL, Tobias M, Tran BX, Trasande L, Toyoshima H, van de Vijver S, Vasankari TJ, Veerman JL, Velasquez-Melendez G, Vlassov VV, Vollset SE, Vos T, Wang C, Wang X, Weiderpass E, Werdecker A, Wright JL, Yang YC, Yatsuya H, Yoon J, Yoon SJ, Zhao Y, Zhou M, Zhu S, Lopez AD, Murray CJ, Gakidou E. Global, regional, and national prevalence of overweight and obesity in children and adults during 1980–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet . 2014; 384( 9945): 766– 781. Google Scholar CrossRef Search ADS PubMed  2. Afshin A, Forouzanfar MH, Reitsma MB, Sur P, Estep K, Lee A, Marczak L, Mokdad AH, Moradi-Lakeh M, Naghavi M, Salama JS, Vos T, Abate KH, Abbafati C, Ahmed MB, Al-Aly Z, Alkerwi A, Al-Raddadi R, Amare AT, Amberbir A, Amegah AK, Amini E, Amrock SM, Anjana RM, Ärnlöv J, Asayesh H, Banerjee A, Barac A, Baye E, Bennett DA, Beyene AS, Biadgilign S, Biryukov S, Bjertness E, Boneya DJ, Campos-Nonato I, Carrero JJ, Cecilio P, Cercy K, Ciobanu LG, Cornaby L, Damtew SA, Dandona L, Dandona R, Dharmaratne SD, Duncan BB, Eshrati B, Esteghamati A, Feigin VL, Fernandes JC, Fürst T, Gebrehiwot TT, Gold A, Gona PN, Goto A, Habtewold TD, Hadush KT, Hafezi-Nejad N, Hay SI, Horino M, Islami F, Kamal R, Kasaeian A, Katikireddi SV, Kengne AP, Kesavachandran CN, Khader YS, Khang YH, Khubchandani J, Kim D, Kim YJ, Kinfu Y, Kosen S, Ku T, Defo BK, Kumar GA, Larson HJ, Leinsalu M, Liang X, Lim SS, Liu P, Lopez AD, Lozano R, Majeed A, Malekzadeh R, Malta DC, Mazidi M, McAlinden C, McGarvey ST, Mengistu DT, Mensah GA, Mensink GBM, Mezgebe HB, Mirrakhimov EM, Mueller UO, Noubiap JJ, Obermeyer CM, Ogbo FA, Owolabi MO, Patton GC, Pourmalek F, Qorbani M, Rafay A, Rai RK, Ranabhat CL, Reinig N, Safiri S, Salomon JA, Sanabria JR, Santos IS, Sartorius B, Sawhney M, Schmidhuber J, Schutte AE, Schmidt MI, Sepanlou SG, Shamsizadeh M, Sheikhbahaei S, Shin MJ, Shiri R, Shiue I, Roba HS, Silva DAS, Silverberg JI, Singh JA, Stranges S, Swaminathan S, Tabarés-Seisdedos R, Tadese F, Tedla BA, Tegegne BS, Terkawi AS, Thakur JS, Tonelli M, Topor-Madry R, Tyrovolas S, Ukwaja KN, Uthman OA, Vaezghasemi M, Vasankari T, Vlassov VV, Vollset SE, Weiderpass E, Werdecker A, Wesana J, Westerman R, Yano Y, Yonemoto N, Yonga G, Zaidi Z, Zenebe ZM, Zipkin B, Murray CJL; GBD 2015 Obesity Collaborators. Health effects of overweight and obesity in 195 countries over 25 years. N Engl J Med . 2017; 377( 1): 13– 27. Google Scholar CrossRef Search ADS PubMed  3. Di Angelantonio E, Bhupathiraju ShN, Wormser D, Gao P, Kaptoge S, Berrington de Gonzalez A, Cairns BJ, Huxley R, Jackson ChL, Joshy G, Lewington S, Manson JE, Murphy N, Patel AV, Samet JM, Woodward M, Zheng W, Zhou M, Bansal N, Barricarte A, Carter B, Cerhan JR, Smith GD, Fang X, Franco OH, Green J, Halsey J, Hildebrand JS, Jung KJ, Korda RJ, McLerran DF, Moore SC, O’Keeffe LM, Paige E, Ramond A, Reeves GK, Rolland B, Sacerdote C, Sattar N, Sofianopoulou E, Stevens J, Thun M, Ueshima H, Yang L, Yun YD, Willeit P, Banks E, Beral V, Chen ZH, Gapstur SM, Gunter MJ, Hartge P, Jee SH, Lam TH, Peto R, Potter JD, Willett WC, Thompson SG, Danesh J, Hu FB; Global BMI Mortality Collaboration. Body-mass index and all-cause mortality: individual-participant-data meta-analysis of 239 prospective studies in four continents. Lancet . 2016; 388( 10046): 776– 786. Google Scholar CrossRef Search ADS PubMed  4. Alberti KGMM, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato KA, Fruchart J-C, James WPT, Loria CM, Smith SC, Jr; International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; International Association for the Study of Obesity. Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation . 2009; 120( 16): 1640– 1645. Google Scholar CrossRef Search ADS PubMed  5. Fasshauer M, Blüher M. Adipokines in health and disease. Trends Pharmacol Sci . 2015; 36( 7): 461– 470. Google Scholar CrossRef Search ADS PubMed  6. Flehmig G, Scholz M, Klöting N, Fasshauer M, Tönjes A, Stumvoll M, Youn B-S, Blüher M. Identification of adipokine clusters related to parameters of fat mass, insulin sensitivity and inflammation. PLoS One . 2014; 9( 6): e99785. Google Scholar CrossRef Search ADS PubMed  7. Spranger J, Kroke A, Möhlig M, Bergmann MM, Ristow M, Boeing H, Pfeiffer AF. Adiponectin and protection against type 2 diabetes mellitus. Lancet . 2003; 361( 9353): 226– 228. Google Scholar CrossRef Search ADS PubMed  8. Lehr S, Hartwig S, Sell H. Adipokines: a treasure trove for the discovery of biomarkers for metabolic disorders. Proteomics Clin Appl . 2012; 6( 1–2): 91– 101. Google Scholar CrossRef Search ADS PubMed  9. Ebert T, Kralisch S, Wurst U, Scholz M, Stumvoll M, Kovacs P, Fasshauer M, Tönjes A. Association of metabolic parameters and rs726344 in FNDC5 with serum irisin concentrations. Int J Obes . 2016; 40( 2): 260– 265. 10. Ebert T, Kralisch S, Loessner U, Jessnitzer B, Stumvoll M, Fasshauer M, Tönjes A. Relationship between serum levels of angiopoietin-related growth factor and metabolic risk factors. Horm Metab Res . 2014; 46( 10): 685– 690. Google Scholar CrossRef Search ADS PubMed  11. Tönjes A, Zeggini E, Kovacs P, Böttcher Y, Schleinitz D, Dietrich K, Morris AP, Enigk B, Rayner NW, Koriath M, Eszlinger M, Kemppinen A, Prokopenko I, Hoffmann K, Teupser D, Thiery J, Krohn K, McCarthy MI, Stumvoll M. Association of FTO variants with BMI and fat mass in the self-contained population of Sorbs in Germany. Eur J Hum Genet . 2010; 18( 1): 104– 110. Google Scholar CrossRef Search ADS PubMed  12. Rohde K, Keller M, Horstmann A, Liu X, Eichelmann F, Stumvoll M, Villringer A, Kovacs P, Tönjes A, Böttcher Y. Role of genetic variants in ADIPOQ in human eating behavior. Genes Nutr . 2015; 10( 1): 449. Google Scholar CrossRef Search ADS PubMed  13. Kralisch S, Tönjes A, Krause K, Richter J, Lossner U, Kovacs P, Ebert T, Blüher M, Stumvoll M, Fasshauer M. Fibroblast growth factor-21 serum concentrations are associated with metabolic and hepatic markers in humans. J Endocrinol . 2013; 216( 2): 135– 143. Google Scholar CrossRef Search ADS PubMed  14. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia . 1985; 28( 7): 412– 419. Google Scholar CrossRef Search ADS PubMed  15. Crisafulli A, Orrù V, Melis F, Tocco F, Concu A. Hemodynamics during active and passive recovery from a single bout of supramaximal exercise. Eur J Appl Physiol . 2003; 89( 2): 209– 216. Google Scholar CrossRef Search ADS PubMed  16. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF III, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate. Ann Intern Med . 2009; 150( 9): 604– 612. Google Scholar CrossRef Search ADS PubMed  17. Suzuki R, Shimodaira H. Pvclust: an R package for assessing the uncertainty in hierarchical clustering. Bioinformatics . 2006; 22( 12): 1540– 1542. Google Scholar CrossRef Search ADS PubMed  18. Goralski KB, McCarthy TC, Hanniman EA, Zabel BA, Butcher EC, Parlee SD, Muruganandan S, Sinal CJ. Chemerin, a novel adipokine that regulates adipogenesis and adipocyte metabolism. J Biol Chem . 2007; 282( 38): 28175– 28188. Google Scholar CrossRef Search ADS PubMed  19. Tönjes A, Fasshauer M, Kratzsch J, Stumvoll M, Blüher M. Adipokine pattern in subjects with impaired fasting glucose and impaired glucose tolerance in comparison to normal glucose tolerance and diabetes. PLoS One . 2010; 5( 11): e13911. Google Scholar CrossRef Search ADS PubMed  20. Ebert T, Roth I, Richter J, Tönjes A, Kralisch S, Lossner U, Kratzsch J, Blüher M, Stumvoll M, Fasshauer M. Different associations of adipokines in lean and healthy adults. Horm Metab Res . 2014; 46( 1): 41– 47. Google Scholar PubMed  21. Lehrke M, Becker A, Greif M, Stark R, Laubender RP, von Ziegler F, Lebherz C, Tittus J, Reiser M, Becker C, Göke B, Leber AW, Parhofer KG, Broedl UC. Chemerin is associated with markers of inflammation and components of the metabolic syndrome but does not predict coronary atherosclerosis. Eur J Endocrinol . 2009; 161( 2): 339– 344. Google Scholar CrossRef Search ADS PubMed  22. Klöting N, Fasshauer M, Dietrich A, Kovacs P, Schön MR, Kern M, Stumvoll M, Blüher M. Insulin-sensitive obesity. Am J Physiol Endocrinol Metab . 2010; 299( 3): E506– E515. Google Scholar CrossRef Search ADS PubMed  23. Blüher M, Rudich A, Klöting N, Golan R, Henkin Y, Rubin E, Schwarzfuchs D, Gepner Y, Stampfer MJ, Fiedler M, Thiery J, Stumvoll M, Shai I. Two patterns of adipokine and other biomarker dynamics in a long-term weight loss intervention. Diabetes Care . 2012; 35( 2): 342– 349. Google Scholar CrossRef Search ADS PubMed  24. Baratta R, Amato S, Degano C, Farina MG, Patanè G, Vigneri R, Frittitta L. Adiponectin relationship with lipid metabolism is independent of body fat mass: evidence from both cross-sectional and intervention studies. J Clin Endocrinol Metab . 2004; 89( 6): 2665– 2671. Google Scholar CrossRef Search ADS PubMed  25. Cnop M, Havel PJ, Utzschneider KM, Carr DB, Sinha MK, Boyko EJ, Retzlaff BM, Knopp RH, Brunzell JD, Kahn SE. Relationship of adiponectin to body fat distribution, insulin sensitivity and plasma lipoproteins: evidence for independent roles of age and sex. Diabetologia . 2003; 46( 4): 459– 469. Google Scholar CrossRef Search ADS PubMed  26. Nakamura T, Kodama Y, Takano H, Umetani K, Fujioka D, Saito Y, Kawabata K, Obata JE, Kitta Y, Kobayashi T, Mende A, Kugiyama K. Increase in circulating levels of adiponectin after treatment with statin and fibrate in patients with coronary artery disease and hyperlipidemia. Atherosclerosis . 2007; 193( 2): 449– 451. Google Scholar CrossRef Search ADS PubMed  27. Zhang X, Yeung DCY, Karpisek M, Stejskal D, Zhou Z-G, Liu F, Wong RLC, Chow W-S, Tso AWK, Lam KSL, Xu A. Serum FGF21 levels are increased in obesity and are independently associated with the metabolic syndrome in humans. Diabetes . 2008; 57( 5): 1246– 1253. Google Scholar CrossRef Search ADS PubMed  28. Bobbert T, Schwarz F, Fischer-Rosinsky A, Pfeiffer AFH, Möhlig M, Mai K, Spranger J. Fibroblast growth factor 21 predicts the metabolic syndrome and type 2 diabetes in Caucasians. Diabetes Care . 2013; 36( 1): 145– 149. Google Scholar CrossRef Search ADS PubMed  29. Fisher FM, Chui PC, Antonellis PJ, Bina HA, Kharitonenkov A, Flier JS, Maratos-Flier E. Obesity is a fibroblast growth factor 21 (FGF21)-resistant state. Diabetes . 2010; 59( 11): 2781– 2789. Google Scholar CrossRef Search ADS PubMed  30. Kralisch S, Fasshauer M. Adipocyte fatty acid binding protein: a novel adipokine involved in the pathogenesis of metabolic and vascular disease? Diabetologia . 2013; 56( 1): 10– 21. Google Scholar CrossRef Search ADS PubMed  31. Ebert T, Hopf LM, Wurst U, Bachmann A, Kralisch S, Lössner U, Platz M, Kratzsch J, Stolzenburg JU, Dietel A, Grisk O, Beige J, Anders M, Bast I, Klöting N, Blüher M, Stumvoll M, Fasshauer M. Circulating adipocyte fatty acid binding protein is increased in chronic and acute renal dysfunction. Nutr Metab Cardiovasc Dis . 2014; 24( 9): 1027– 1034. Google Scholar CrossRef Search ADS PubMed  32. Zachariah JP, Quiroz R, Nelson KP, Teng Z, Keaney JF, Jr, Sullivan LM, Vasan RS. Prospective relation of circulating adipokines to incident metabolic syndrome: the Framingham Heart Study. J Am Heart Assoc . 2017; 6( 7): e004974. Google Scholar CrossRef Search ADS PubMed  33. Lee S-A, Kallianpur A, Xiang Y-B, Wen W, Cai Q, Liu D, Fazio S, Linton MF, Zheng W, Shu XO. Intra-individual variation of plasma adipokine levels and utility of single measurement of these biomarkers in population-based studies. Cancer Epidemiol Biomarkers Prev . 2007; 16( 11): 2464– 2470. Google Scholar CrossRef Search ADS PubMed  Copyright © 2018 Endocrine Society

Journal

Journal of Clinical Endocrinology and MetabolismOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

Monthly Plan

  • Read unlimited articles
  • Personalized recommendations
  • No expiration
  • Print 20 pages per month
  • 20% off on PDF purchases
  • Organize your research
  • Get updates on your journals and topic searches

$49/month

Start Free Trial

14-day Free Trial

Best Deal — 39% off

Annual Plan

  • All the features of the Professional Plan, but for 39% off!
  • Billed annually
  • No expiration
  • For the normal price of 10 articles elsewhere, you get one full year of unlimited access to articles.

$588

$360/year

billed annually
Start Free Trial

14-day Free Trial