Abstract
Rather than a traditional growth factor, fibroblast growth factor-21 (FGF21) is considered to be a metabolic hormone. In the current study, we investigated serum FGF21 levels in the self-contained population of Sorbs. Serum FGF21 concentrations were quantified by ELISA and correlated with IGF1 as well as metabolic, renal, hepatic, inflammatory, and cardiovascular parameters in 913 Sorbs from Germany. Moreover, human IGF1 protein secretion was investigated in FGF21-stimulated HepG2 cells. Median FGF21 serum concentrations were 2.1-fold higher in subjects with type 2 diabetes mellitus (141.8 ng/l) compared with controls (66.7 ng/l). Furthermore, nondiabetic subjects with FGF21 levels below the detection limit of the ELISA showed a more beneficial metabolic profile compared with subjects with measurable FGF21. Moreover, FGF21 was significantly lower in female compared with male subjects after adjustment for age and BMI. In multiple regression analyses, circulating FGF21 concentrations remained independently and positively associated with gender, systolic blood pressure, triglycerides, and γ glutamyl transferase whereas a negative association was observed with IGF1 in nondiabetic subjects. Notably, FGF21 significantly inhibited IGF1 secretion into HepG2 cell culture supernatants in preliminary in vitro experiments. FGF21 serum concentrations are associated with facets of the metabolic syndrome, hepatocellular function, as well as GH status.
Introduction
Fibroblast growth factor-21 (FGF21) is a member of the FGF family including 22 members (Beenken & Mohammadi 2009). FGF21 is mainly produced by the liver but also by other tissues including white adipose tissue (WAT), skeletal muscle, and pancreatic β cells (Beenken & Mohammadi 2009). FGF21 is regulated by nutritional status and influences glucose and lipid metabolism by central and peripheral mechanisms (Kralisch & Fasshauer 2011). Thus, the protein induces glucose uptake (Kharitonenkov et al. 2005) and decreases glucose and lipid concentrations in obese mice (Kharitonenkov et al. 2005) and diabetic monkeys (Kharitonenkov et al. 2007). Furthermore, it increases energy expenditure resulting in weight loss (Coskun et al. 2008) and upregulates fatty acid oxidation (Inagaki et al. 2007). These results suggest that FGF21 regulates energy balance and improves glucose homeostasis. Furthermore, FGF21 is also considered to be a starvation signal and is additionally involved in the process of ‘browning’ of WAT (Cantó & Auwerx 2012).
At the cellular level, FGF21 mediates its effects through cell surface receptors composed of the classic four FGF receptor (FGFR) isotypes 1–4 (Zhang et al. 2006). FGFRs are tyrosine kinases that are complexed selectively and non-covalently with an essential co-receptor β-Klotho, forming an FGF21-β-Klotho–FGFR complex (Ogawa et al. 2007). Expression of FGF21 is controlled by various transcriptional activators such as peroxisome proliferator-activated receptor (PPAR) α in the liver (Badman et al. 2007) and PPARγ in the adipocytes (Wang et al. 2008).
Several clinical studies investigated the role of FGF21 in human metabolic disease. Thus, high baseline FGF21 serum levels predicted the risk to develop type 2 diabetes mellitus (T2DM) over 5.4 years (odds ratio 1.792; P<0.01) together with waist circumference and fasting plasma glucose levels in 1.200 nondiabetic Chinese subjects (Chen et al. 2011). Recent data indicate that higher FGF21 serum levels are associated with abnormal fasting glucose and insulin resistance in the Baltimore longitudinal study of aging (Semba et al. 2012). Furthermore, a study by Lin et al. (2012) suggested that FGF21 potentially regulates insulin secretion in humans. In contrast to the study by Semba et al. (2012), FGF21 levels were inversely correlated with changes in serum glucose during an oral glucose tolerance test with lowest FGF21 concentrations after 1 h and a peak at 3 h (Lin et al. 2012). This acute response appears to be abolished in patients with impaired glucose tolerance and T2DM (Lin et al. 2012). FGF21 concentrations were increased in obese children in comparison with lean controls and were related to leptin and free fatty acids (FFAs; Reinehr et al. 2012). Furthermore, FGF21 is considered as a potential marker of nonalcoholic fatty liver disease (NAFLD; Dushay et al. 2010, Li et al. 2010, Yilmaz et al. 2010). Moreover, a correlation of serum FGF21 levels with GH and insulin-like growth factor-1 (IGF1) concentrations is discussed (Inagaki et al. 2008, Wu et al. 2012). In agreement with this hypothesis, FGF21 levels are negatively associated with IGF1 independent of body fat and insulin resistance in patients with anorexia nervosa (Fazeli et al. 2010).
To date, no study with sufficient statistical power has included gender, renal function, metabolic, hepatic, and vascular risk markers combined in its analysis of FGF21 serum concentration. Therefore, the aim of our cross-sectional study was to assess the complex interaction of these clinical parameters, as well as GH status, with FGF21 levels in a large well-characterized sample.
Materials and methods
Subjects
All subjects are part of a sample from an extensively phenotyped self-contained population of Sorbs from Eastern Germany described in more detail recently (Tonjes et al. 2009, 2010, 2012). At present, about 1000 Sorbs are enrolled in the study. Nine hundred and thirteen subjects were available for the present analyses. To exclude any secondary effects, associations with glucose and insulin levels were assessed only in the subgroup of nondiabetic subjects (n=812) defined according to the American Diabetes Association criteria (American Diabetes Association 2010). The study was approved by the local ethics committee and all subjects gave written informed consent before taking part in the study.
Phenotyping
Phenotyping included collection of anthropometric data (weight, height, BMI, waist-to-hip ratio (WHR), waist-to-height ratio (WHtR), and body impedance analysis (BIA)) and a 75 g oral glucose tolerance test. Furthermore, homeostasis model assessment of insulin resistance (HOMA-IR), Stumvoll index, and quantitative insulin sensitivity check (Quicky) index were calculated as described previously (Matthews et al. 1985, Stumvoll et al. 2001, Chu et al. 2003). Renal function was assessed as glomerular filtration rate (GFR) estimated by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation (Levey et al. 2009). Ultrasound sensor (10 MHz; GE Healthcare, Inc., München, Germany) was used to measure the intima-media thickness (IMT) of the common carotid arteries. After three measurements, IMT values at each side were averaged. BIA was performed using BIA-2000-S (Data Input GmbH, Darmstadt, Germany) and evaluated using the software Nutri3 (Data Input GmbH).
Assays
Blood samples were taken after an overnight fast. Serum insulin was measured with AutoDELFIA Insulin assay (PerkinElmer Life and Analytical Sciences, Turku, Finland). The assay sensitivity was 3.0 pmol/l. The interassay coefficient of variance (%) was between 3.37 and 6.13% for concentrations between 71 and 867 pmol/l. FGF21 serum concentrations were determined with a commercially available ELISA (Biovendor, Modrice, Czech Republic) according to the manufacturer's instructions. The sensitivity of FGF21 ELISA was 7 ng/l. The degree of precision of ELISA system in terms of intra-assay coefficient of variance (%) was between 2.0 and 20.8% and that of interassay was between 2.9 and 12.0%. Spike recovery and linearity were in a range of 94.3–105.6 and 101.0–109.9% respectively. Furthermore, the ELISA was specific for human FGF21 and did not cross-react with human FGF19, human FGF23, bovine, cat, dog, goat, hamster, horse, monkey mouse pig, rat, and sheep sera. All serum samples had two freeze–thaw cycles before quantification of FGF21 serum concentrations. According to the manufacturer's instructions, no decline of human FGF21 was observed in serum and plasma samples even after five freeze–thaw cycles. Each commercial FGF21 ELISA included two quality controls (high and low). For all ELISA assessments performed in our study, FGF21 values for these two quality controls were within the range given by the manufacturer. Serum creatinine, total, HDL, LDL cholesterol, triglycerides (TGs), urea, albumin, alanine aminotransferase (ALAT), aspartate aminotransferase (ASAT), alkaline phosphatase (aP), gamma glutamyl transferase (γGT), TSH, and C-reactive protein (CRP) were measured by standard laboratory methods in a certified laboratory. Serum IGF1 concentrations were measured by commercially available automated two-site chemiluminescent immunometric assays (Immulite 2000, Siemens Healthcare Diagnostics GmbH, Bad Nauheim, Germany). The intra-assay coefficient of variance (%) was between 3.1 and 4.4% and that of interassay was between 5.7 and 6.6%.
Hepatocyte culture and analysis of IGF1 protein secretion
Human HepG2 cells (American Type Culture Collection, Rockville, MD, USA) were cultured in DMEM–high-glucose medium (DMEM-H, PAA, Pasching, Austria) supplemented with 10% fetal bovine serum (PAA). Equivalent numbers of HepG2 cells starved in DMEM-H for 24 h were treated with 250 ng/ml recombinant FGF21 (R&D Systems, Wiesbaden, Germany; treatment group) or PBS (control group) for 24 h. Quantification of human IGF1 protein secretion into HepG2 cell culture supernatants was performed with a commercially available ELISA (R&D Systems) according to the manufacturer's instructions. The sensitivity of the ELISA was 0.026 ng/ml. The degree of precision of the ELISA system in terms of intra-assay coefficient of variance (%) was between 3.5 and 4.3% and that of interassay was between 7.5 and 8.3%. Spike recovery and linearity were in a range of 97–102 and 92–108% respectively.
Statistical analysis
PASW software version 18.0.1 (SPSS, Chicago, IL, USA) was used in all statistical analyses. Distribution was tested for normality using Shapiro–Wilk W test and non-normally distributed parameters were naturally logarithmically transformed before analyses. For patients with no measurable FGF21 concentration, we used half (3.5 ng/l) of the sensitivity of the FGF21 ELISA (7 ng/l) in group comparisons between controls and T2DM. Differences between two groups were assessed by the Mann–Whitney U test for continuous parameters and by the χ2 test for gender followed by Bonferroni adjustment for multiple testing. Analyses were adjusted for age and gender and in a second model additionally for BMI to exclude that differences were mainly driven by body weight variation. Correlations were performed using the Spearman's rank correlation method followed by Bonferroni adjustment for multiple testing. Multivariate regression analyses were performed including all parameters with highly significant correlations in the univariate analysis (P≤0.001) as covariates. In case of parameters strongly related to each other, one representative covariate was included in the model. A P value of <0.05 was considered as statistically significant in all analyses.
Results
Study subjects
Baseline characteristics of nondiabetic and diabetic subjects are summarized in Table 1. All continuous variables are given as median (interquartile range). Serum FGF21 levels were significantly increased in subjects with T2DM (141.8 (174.4) ng/l; n=101) compared with nondiabetic controls (66.7 (117.8) ng/l; n=812) after adjustment for gender and age (P<0.001). To exclude any secondary effects, subsequent analyses were performed only in nondiabetic subjects (Table 1). In nondiabetic subjects, circulating FGF21 levels were below the limit of detection of the FGF21 ELISA in 142 subjects. Interestingly, subjects with FGF21 concentrations below the limit of detection showed a more beneficial metabolic profile compared with subjects with measurable FGF21 (Table 1). Thus, the group with non-detectable FGF21 showed significantly lower blood pressure and TG, higher HDL cholesterol levels, as well as better insulin sensitivity (Table 1). These differences remained significant after adjustment for age, gender, and BMI (Table 1). FGF21 was significantly lower in female (median 85.0 (107.8) ng/l, n=385) compared with male (median 92.6 (133.2) ng/l, n=285) subjects after adjustment for age and BMI (P=0.029).
Baseline characteristics
| Non-type 2 diabetes mellitus | |||||
|---|---|---|---|---|---|
| Subjects (above the limit of detection) | Subjects (below the limit of detection) | P1 | Type 2 diabetes mellitus | P2 | |
| n | 670 | 142 | 101 | ||
| FGF21 (ng/l) | 88.9 (117.7) | 3.5 (0) | <0.001a,b | 141.8 (174.4) | <0.001a,b |
| Age (years) | 47.7 (22.1) | 39.6 (19.6) | <0.001b | 65.8 (16.8) | <0.001b |
| Gender (f/m) | 385/285 | 100/42 | 0.005 | 59/42 | 0.064 |
| SBP (mmHg) | 135.0 (22) | 127.3 (19) | 0.001a | 151.0 (26) | <0.001a,b |
| DBP (mmHg) | 83.3 (13) | 78.5 (11) | 0.001a | 86.0 (14) | 0.424 |
| Height (m) | 1.69 (0.13) | 1.69 (0.13) | 0.338 | 1.65 (0.16) | 0.554 |
| Weight (kg) | 76 (20) | 69 (17) | <0.001a,b | 82 (21) | <0.001b |
| BMI (kg/m2) | 26.5 (5.8) | 24.0 (4.0) | <0.001a,b | 30.3 (7.4) | <0.001b |
| Fat mass (%) | 20.1 (10.7) | 16.4 (6.5) | <0.001a,b | 25.3 (14.1) | 0.001 |
| Waist circumference (cm) | 91 (19) | 82 (14) | <0.001a,b | 101 (17) | <0.001b |
| Hip circumference (cm) | 102 (10) | 99 (8) | <0.001a,b | 107 (12) | 0.013a |
| WHR | 0.87 (0.14) | 0.82 (0.11) | <0.001a,b | 0.96 (0.13) | <0.001a,b |
| WtHR | 0.54 (0.11) | 0.48 (0.08) | <0.001a,b | 0.60 (0.11) | <0.001b |
| Glucose 0 h (mmol/l) | 5.32 (0.65) | 5.00 (0.46) | <0.001a,b | 7.36 (2.49) | <0.001a,b |
| Glucose 30 min (mmol/l) | 8.52 (2.31) | 7.95 (1.57) | 0.007a | 11.74 (2.04) | <0.001a,b |
| Glucose 120 min (mmol/l) | 5.22 (2.32) | 4.78 (1.68) | 0.256 | 11.91 (6.27) | <0.001a,b |
| Insulin 0 h (pmol/l) | 33.8 (27.6) | 25.6 (19.3) | <0.001a,b | 52.5 (45.5) | <0.001a,b |
| Insulin 30 min (pmol/l) | 252 (172) | 223 (173) | 0.012a | 219 (153) | 0.190 |
| Insulin 120 min (pmol/l) | 125 (159) | 118 (117) | 0.673 | 279 (292) | <0.001a,b |
| HOMA-IR | 1.09 (1.00) | 0.81 (0.61) | <0.001a,b | 2.42 (2.44) | <0.001a,b |
| Stumvoll index | 0.10 (0.04) | 0.12 (0.02) | <0.001a,b | 0.07 (0.04) | <0.001a,b |
| Quicki index | 0.44 (0.07) | 0.47 (0.07) | <0.001a,b | 0.39 (0.06) | <0.001a,b |
| Cholesterol (mmol/l) | 5.34 (1.30) | 5.14 (1.24) | 0.188 | 5.25 (1.35) | <0.001a,b |
| HDL cholesterol (mmol/l) | 1.58 (0.53) | 1.79 (0.57) | <0.001a,b | 1.33 (0.48) | <0.001a,b |
| LDL cholesterol (mmol/l) | 3.41 (1.30) | 3.11 (1.12) | 0.085 | 3.35 (1.15) | <0.001a,b |
| TG (mmol/l) | 1.13 (0.74) | 0.78 (0.47) | <0.001a,b | 1.64 (1.00) | <0.001a,b |
| Creatinine serum (μmol/l) | 71 (18) | 68 (14) | 0.764 | 69 (25) | 0.175 |
| GFR (ml/min per 1.73 m2) | 97.9 (21.2) | 103.9 (18.8) | 0.587 | 90.2 (25.0) | 0.294 |
| Urea (mmol/l) | 4.7 (1.7) | 4.6 (1.7) | 0.002a | 5.6 (2.0) | 0.004a |
| Protein serum (g/l) | 72.8 (5.3) | 72.6 (5.5) | 0.648 | 72.3 (5.2) | 0.641 |
| Creatinine urine (μmol/l) | 12 129 (8713) | 10 967 (8852) | 0.103 | 10 842 (8581) | 0.718 |
| Protein urine (mg/l) | 80.8 (82.0) | 66.1 (72.0) | 0.219 | 92.5 (72.1) | 0.477 |
| Albumin urine (mg/l) | 8.1 (8.9) | 7.1 (6.2) | 0.248 | 14.2 (28.2) | 0.493 |
| ALAT (μkat/l) | 0.37 (0.25) | 0.29 (0.14) | <0.001a,b | 0.46 (0.26) | 0.008 |
| ASAT (μkat/l) | 0.44 (0.15) | 0.40 (0.12) | 0.021a | 0.48 (0.16) | 0.346 |
| aP (μkat/l) | 1.05 (0.42) | 0.90 (0.34) | <0.001a,b | 1.20 (0.44) | 0.002a |
| γGT (μkat/l) | 0.32 (0.29) | 0.22 (0.12) | <0.001a,b | 0.51 (0.52) | 0.003 |
| IMT right (mm) | 0.63 (0.20) | 0.57 (0.10) | 0.070 | 0.83 (0.27) | 0.001a |
| IMT left (mm) | 0.63 (0.17) | 0.57 (0.13) | 0.278 | 0.82 (0.24) | 0.003a |
| IGF1 (ng/ml) | 159.9 (79.3) | 176.7 (77.6) | 0.291 | 120.0 (59.7) | 0.993 |
| TSH (mU/l) | 1.52 (1.2) | 1.7 (1.3) | 0.297 | 1.33 (1.0) | 0.049a |
| CRP (mg/l) | 1.2 (1.9) | 0.5 (1.3) | 0.016a | 2.5 (3.5) | <0.001b |
ALAT, alanine aminotransferase; aP, alkaline phosphatase; ASAT, aspartate aminotransferase; CRP, C-reactive protein; DBP, diastolic blood pressure; FFA, free fatty acids; FG, fasting glucose; FGF21, fibroblast growth factor-21; FI, fasting insulin; GFR, glomerular filtration rate; γGT, gamma glutamyl transferase; IGF, insulin-like growth factor; IMT, intima-media thickness; SBP, systolic blood pressure; TG, triglyceride; WHR, waist-to-hip ratio; WtHR, waist-to-height ratio; values for median (interquartile range) or total number are shown. Differences between groups were assessed by the Mann–Whitney U test for continuous parameters and by the χ2 test for gender. Age and gender distribution were tested without adjustment. All other analyses were adjusted for age and gender and in a second model additionally for BMI to exclude that differences were mainly driven by body weight variation. P values are presented for model 1 (adjustment for age and gender).
Significant differences after additional adjustment for BMI (model 2).
Differences (model 1) remaining significant after Bonferroni adjustment for multiple testing. P1 indicates P value of comparison between nondiabetic subjects with and without detectable FGF21 levels; P2 indicates P value of comparison between nondiabetic and diabetic subjects.
Univariate correlations
Univariate correlations were performed only in subjects with detectable FGF21 concentrations (n=670). Here, serum FGF21 levels were positively associated with age, blood pressure, parameters of obesity (weight, BMI, fat mass, waist circumference, hip circumference, WHR, and WtHR), parameters of insulin and glucose metabolism (fasting glucose, fasting insulin, and HOMA-IR), dyslipidemia (total cholesterol, LDL-cholesterol, and TG), liver function (ALAT, ASAT, aP, and γGT), and IMT (Table 2). Moreover, serum FGF21 concentrations were negatively associated with Stumvoll index, Quicki index, HDL cholesterol, as well as with renal function (GFR) and IGF1 (Table 2). By contrast, FGF21 levels did not show an association with height, serum creatinine, urine protein, urine albumin, thyroid function (TSH), and CRP (Table 2).
Univariate correlations
| r | P | |
|---|---|---|
| Age (years) | 0.200 | <0.001a,b |
| SBP (mmHg) | 0.277 | <0.001a,b |
| DBP (mmHg) | 0.225 | <0.001a,b |
| Height (m) | −0.037 | 0.345 |
| Weight (kg) | 0.131 | 0.001a,b |
| BMI (kg/m2) | 0.177 | <0.001a,b |
| Fat mass (%) | 0.122 | 0.002a |
| Waist circumference (cm) | 0.241 | <0.001a,b |
| Hip circumference (cm) | 0.088 | 0.022a |
| WHR | 0.249 | <0.001a,b |
| WtHR | 0.246 | <0.001a,b |
| Glucose 0 h (mmol/l) | 0.152 | <0.001a,b |
| Glucose 30 min (mmol/l) | 0.206 | <0.001a,b |
| Glucose 120 min (mmol/l) | 0.103 | 0.009a |
| Insulin 0 h (pmol/l) | 0.130 | 0.001a,b |
| Insulin 30 min (pmol/l) | 0.107 | 0.006a |
| Insulin 120 min (pmol/l) | 0.096 | 0.014a |
| HOMA-IR | 0.148 | <0.001a,b |
| Stumvoll index | −0.201 | <0.001a,b |
| Quicki index | −0.148 | <0.001a,b |
| Cholesterol (mmol/l) | 0.099 | 0.011a |
| HDL cholesterol (mmol/l) | −0.148 | <0.001a,b |
| LDL cholesterol (mmol/l) | 0.104 | 0.007a |
| TG (mmol/l) | 0.327 | <0.001a,b |
| Creatinine serum (μmol/l) | 0.063 | 0.106 |
| GFR (ml/min per 1.73 m2) | −0.131 | 0.001a,b |
| Urea (mmol/l) | 0.000 | 0.996 |
| Protein serum (g/l) | −0.002 | 0.950 |
| Creatinine urine (μmol/l) | 0.016 | 0.685 |
| Protein urine (mg/l) | 0.011 | 0.780 |
| Albumin urine (mg/l) | 0.045 | 0.249 |
| ALAT (μkat/l) | 0.142 | <0.001a,b |
| ASAT (μkat/l) | 0.153 | <0.001a,b |
| aP (μkat/l) | 0.152 | <0.001a,b |
| γGT (μkat/l) | 0.271 | <0.001a,b |
| IMT right (mm) | 0.221 | <0.001a,b |
| IMT left (mm) | 0.219 | <0.001a,b |
| IGF1 (ng/ml) | −0.226 | <0.001a,b |
| TSH (mU/l) | 0.016 | 0.704 |
| CRP (mg/l) | 0.067 | 0.084 |
Univariate correlations with serum FGF21 concentrations in all subjects with measurable FGF21 (n=670); r and P values are given.
Significant correlation as assessed by Spearman's correlation method.
Correlations remaining significant after Bonferroni adjustment for multiple testing. Abbreviations are indicated in Table 1.
Multivariate correlations
Multivariate regression analysis revealed that gender, SBP, TG, γGT, and IGF1 remained independently associated with circulating FGF21 levels after adjustment for age, WHR, Stumvoll index, HDL cholesterol, GFR, and right IMT (P<0.05) (Table 3). This multivariate linear regression model explained 20% of the FGF21 serum variation. The associations between FGF21 serum levels, IGF1, and γGT are depicted in Fig. 1. Here, highest FGF21 levels were observed in the subgroup with lowest IGF1 and highest γGT levels (Fig. 1).
Multivariate linear regression analysis (dependent variable: FGF21)
| Independent variable | β | P |
|---|---|---|
| Age (years) | −0.106 | 0.181 |
| Gender | −0.114 | 0.035a |
| SBP (mmHg) | 0.113 | 0.011a |
| WHR | 0.112 | 0.060 |
| Stumvoll index | 0.041 | 0.375 |
| HDL cholesterol (mmol/l) | −0.010 | 0.822 |
| TG (mmol/l) | 0.223 | <0.001a |
| GFR (ml/min) | 0.020 | 0.694 |
| IMT right (mm) | −0.005 | 0.925 |
| γGT (μkat/l) | 0.170 | <0.001a |
| IGF1 (ng/ml) | −0.170 | <0.001a |
Multivariate linear regression analysis between FGF21 (dependent variable; n=670 nondiabetic subjects with measurable FGF21) and age, gender, SBP, WHR, Stumvoll index, HDL cholesterol, TG, GFR, IMT right, γGT, and IGF1. β-Coefficients and P values are given. Abbreviations are indicated in Table 1.
Significant correlation.
Relationship between serum FGF21, IGF1, and γGT. FGF21 serum levels are depicted depending on IGF1 (low: 0–125 ng/ml; medium: 125–200 ng/ml; and high: >200 ng/ml) and γGT (low: 0–0.25 μkat/l; medium: 0.25–0.5 μkat/l; and high: >0.5 μkat/l) values. Only nondiabetic subjects with FGF21 values above the detection limit of the FGF21 ELISA were included (total n=670).
Citation: Journal of Endocrinology 216, 2; 10.1530/JOE-12-0367
FGF21 is a suppressor of IGF1 protein secretion in vitro
As FGF21 was negatively and independently associated with IGF1 in our study population (Table 3 and Fig. 1), we elucidated in preliminary experiments whether FGF21 might directly influence IGF1 expression in a human hepatocyte model in vitro. Therefore, serum-starved human HepG2 cells were treated with 250 ng/ml recombinant FGF21 for 24 h and IGF1 protein secretion into supernatants was quantified by ELISA. Interestingly, in four independent experiments, FGF21 significantly inhibited IGF1 secretion into the HepG2 cell culture supernatants from 0.089 ng/ml (basal) to 0.062 ng/ml (P<0.05) (Fig. 2). Total protein content in the supernatants was not significantly different between control and FGF21-treated cells (data not shown).
FGF21 downregulates IGF1 protein secretion. Human HepG2 cells were serum deprived overnight before FGF21 (250 ng/ml) was added. After 24 h, IGF1 protein levels were determined in the cell supernatants as described in Materials and Methods section. Results are the mean±s.e.m. of four independent experiments. *Significant (P<0.05) regulation.
Citation: Journal of Endocrinology 216, 2; 10.1530/JOE-12-0367
Discussion
In the current study, we demonstrate for the first time that FGF21 is negatively and independently associated with IGF1 in unselected nondiabetic subjects. A negative association between FGF21 and IGF1 has also been observed in patients with anorexia nervosa independent from body fat and insulin resistance (Fazeli et al. 2010). It is interesting to note in this context that FGF21 inhibits IGF1 production in a human hepatocyte cell model in vitro in our study. However, it needs to be emphasized that dose–response, time-course, and signaling experiments are necessary to more thoroughly establish the link between FGF21 and IGF1 secretion from hepatocytes. Our results suggest that FGF21 directly and negatively impacts IGF1 production. This hypothesis is further supported by recent independent data from Inagaki et al. (2008). The authors demonstrate convincingly that FGF21 reduces levels of active form of STAT5, a major mediator of GH action, in the liver of mice in vivo. FGF21-mediated suppression of STAT5 is accompanied by decreases in the expression of its target genes including IGF1 (Inagaki et al. 2008). Furthermore, FGF21 induces hepatic expression of IGF1 binding protein 1 and suppressor of cytokine signaling 2, both of which impair GH signaling (Inagaki et al. 2008). In agreement with a GH signaling-suppressive effect, chronic FGF21 treatment significantly inhibits growth in mice (Inagaki et al. 2008). Furthermore, these findings are in accordance with the investigations in FGF21 knockout animals (Kubicky et al. 2012). Here, FGF21-deficient mice exhibit greater body and tibial growth during starvation compared with WT littermates, and daily injections of recombinant human FGF21 in a subgroup of food-restricted knockout mice prevent this phenotype (Kubicky et al. 2012). Interestingly, GH binding and GH receptor expression are reduced in the liver and in the growth plate of food-restricted compared with ad libitum-fed WT mice whereas they are similar between food-restricted and ad libitum-fed FGF21-deficient animals. Taking these results into consideration, our present data are in accordance with the hypothesis that FGF21 and IGF1 are not only significantly and negatively associated but that FGF21 directly downregulates IGF1 as well as impairs GH signaling in the liver.
In the current study, we demonstrate that FGF21 serum levels are positively associated with hepatic enzymes including ASAT, ALAT, aP, and γGT. Furthermore, the association between FGF21 and γGT remains independent in multivariate analysis. As γGT is a surrogate parameter of fatty liver degeneration, our results are in accordance with the hypothesis that circulating FGF21 is positively associated with metabolic liver disease including NAFLD (Tyynismaa et al. 2010). In agreement with this notion, FGF21 upregulation has also been demonstrated in recent independent studies using selected NAFLD patient populations (Dushay et al. 2010, Li et al. 2010, Yan et al. 2011, Shen et al. 2012). One potential mechanism contributing to FGF21 upregulation in NAFLD is hepatic FGF21 resistance (Fisher et al. 2010, Shen et al. 2012). In an elegant study, Fisher et al. (2010) demonstrate that FGF21 signaling is attenuated in liver and WAT of diet-induced obese mice, with a fourfold increase in FGF21 serum levels compared with control animals. It needs to be elucidated in future studies whether FGF21 can be used as a biomarker for NAFLD and whether this cytokine is a causal factor of this condition. One limitation of the currently available clinical studies is the lack of liver biopsies to reliably differentiate NAFLD from nonalcoholic steatohepatitis.
In this study, we demonstrate that FGF21 is positively and independently correlated with facets of the metabolic syndrome including obesity, insulin resistance, hypertension, and dyslipidemia in nondiabetic subjects. In agreement with this notion, circulating FGF21 is increased in T2DM compared with nondiabetic controls in our study population. Furthermore, nondiabetic subjects with FGF21 levels below the detection limit of the FGF21 ELISA show a more beneficial metabolic profile compared with subjects with detectable FGF21 concentrations. These data strongly suggest that FGF21 increases with deteriorating metabolic control. Our data are in accordance with recent findings from independent groups. Thus, circulating FGF21 levels in obese subjects are significantly higher compared with lean individuals (Zhang et al. 2008). Furthermore, serum FGF21 correlates positively with adiposity, fasting insulin, and TG and negatively with HDL-cholesterol after adjusting for age and BMI similar to our results (Zhang et al. 2008). Moreover, FGF21 is independently associated with insulin resistance and increased in T2DM in adults from the Baltimore longitudinal study of aging (Semba et al. 2012). In addition, Chen et al. (2011) demonstrate an independent association between serum FGF21 levels on one hand and TG, LDL-cholesterol, and SBP on the other hand in Chinese individuals using multiple linear regression analysis. Consistent with these findings, FGF21 significantly and positively correlates with markers of insulin resistance and dyslipidemia in univariate and multivariate analyses in gestational diabetes mellitus patients in a study from our group (Stein et al. 2010). Mechanistically, elevated FFAs might contribute to FGF21 upregulation when facets of the metabolic disease are present (Mai et al. 2010). To date, the physiological significance for increased FGF21 in metabolic disease is unclear. Paradoxical upregulation of FGF21 might be a compensatory mechanism to improve glucose metabolism when obesity, hypertension, insulin resistance, and an adverse lipid profile are present. Alternatively, the metabolic syndrome might cause resistance to FGF21 leading to compensatory upregulation of this antidiabetic cytokine as proposed by Fisher et al. (2010). Clearly, more work is needed to better elucidate the pathophysiological significance of FGF21 upregulation in metabolic disease states.
Despite the independent association between FGF21 and metabolic parameters, FGF21 is not independently correlated with IMT, a surrogate parameter of atherosclerosis. To our knowledge, IMT has not been included in studies on circulating FGF21 so far. In contrast to our findings, FGF21 concentrations are increased in coronary heart disease patients (Lin et al. 2010). However, coronary heart disease and control groups were not matched for blood pressure, HOMA-IR, and inflammation in this study (Lin et al. 2010). Taking these findings into consideration, it is tempting to speculate that FGF21 is an independent predictor of metabolic but not vascular disease.
Interestingly, the current study population does not show an independent association of circulating FGF21 with renal function assessed by serum creatinine and GFR. These findings are in contrast to a recent report suggesting that FGF21 serum levels are progressively increased from the early to end stages of chronic kidney disease and are associated with renal function in a Chinese cohort (Lin et al. 2011). Furthermore, median FGF21 concentrations are >15-fold higher in hemodialysis patients compared with subjects with a GFR >50 ml/min (Stein et al. 2009). Moreover, serum creatinine positively and GFR negatively predict circulating FGF21 in multiple regression analyses in control subjects (Stein et al. 2009). Similarly, Han et al. (2010) demonstrated that FGF21 levels were eightfold higher in 72 nondiabetic patients receiving peritoneal dialysis compared with controls. Our findings are in agreement with the hypothesis that circulating FGF21 is increased only in more severe stages of renal dysfunction.
Recent studies indicate that FGF21 might mediate glucagon effects. Thus, hepatic FGF21 expression is upregulated by hepatic glucagon receptor activation in a manner that is further augmented by fatty acids. Furthermore, FGF21 preserves β-cell function and survival by stimulating p44/42 mitogen-activated protein and Akt signaling pathways. Moreover, FGF21 lowers glucagon in mice (Iglesias et al. 2012). Unfortunately, glucagon could not be quantified within our study, and, therefore, the relationship between FGF21 and glucagon could not be determined.
Strengths of our study include the high number of subjects, the broad range of phenotypical features, as well as the genetically homogeneous, self-contained population of Sorbs. A limitation of the study is the cross-sectional design and, therefore, causality cannot be established. Furthermore, our clinical study by necessity only measures circulating FGF21. Therefore, our study does not provide information on whether local FGF21 production in liver and adipose tissue is associated with circulating FGF21, as well as with facets of the metabolic syndrome.
Taken together, we demonstrate that FGF21 serum concentrations are associated with facets of the metabolic syndrome, hepatocellular function, as well as GH status. Additional studies are necessary to better elucidate the physiological significance of these findings.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
Financial support was received from the Federal Ministry of Education and Research (BMBF), Germany, FKZ: 01EO1001 (IFB AdiposityDiseases, project K7-9 to M F, K7-37 and K403 to A T, K7-35 and K7-36 to M S), the German Research Council (KFO-152 to M B, M F, and M S), the Interdisciplinary Center for Clinical Research, University of Leipzig (B25 to M F; B27 to A T, P K and M S), the Deutsche Hochdruckliga e.V. (to M F), the Boehringer Ingelheim foundation (to P K), and the German Diabetes Association (to A T and P K).
Acknowledgements
The authors would like to thank Joachim Thiery (Institute of Laboratory Medicine, Clinical Chemistry and Molecular Diagnostics, University of Leipzig) for clinical chemistry services. Furthermore, they also thank Ayman Arafat (Charite Berlin) for performing IGF1 measurements. They thank Beate Enigk (Interdisciplinary Centre for Clinical Research, University of Leipzig) for excellent technical assistance in the laboratory.
References
American Diabetes Association 2010 Diagnosis and classification of diabetes mellitus.
Badman MK, Pissios P, Kennedy AR, Koukos G, Flier JS & Maratos-Flier E 2007 Hepatic fibroblast growth factor 21 is regulated by PPARα and is a key mediator of hepatic lipid metabolism in ketotic states. Cell Metabolism 5 426–437. (doi:10.1016/j.cmet.2007.05.002)
Beenken A & Mohammadi M 2009 The FGF family: biology, pathophysiology and therapy. Nature Reviews. Drug Discovery 8 235–253. (doi:10.1038/nrd2792)
Cantó C & Auwerx J 2012 FGF21 takes a fat bite. Science 336 675–676. (doi:10.1126/science.1222646)
Chen C, Cheung BMY, Tso AWK, Wang Y, Law LSC, Ong KL, Wat NMS, Xu A & Lam KSL 2011 High plasma level of fibroblast growth factor 21 is an independent predictor of type 2 diabetes. Diabetes Care 34 2113–2115. (doi:10.2337/dc11-0294)
Chu JW, Abbasi F, Beatty GW, Khalili M, Koch J, Rosen A, Schmidt JM, Stansell JD & Reaven GM 2003 Methods for quantifying insulin resistance in human immunodeficiency virus-positive patients. Metabolism 52 858–861. (doi:10.1016/S0026-0495(03)00056-8)
Coskun T, Bina HA, Schneider MA, Dunbar JD, Hu CC, Chen Y, Moller DE & Kharitonenkov A 2008 Fibroblast growth factor 21 corrects obesity in mice. Endocrinology 149 6018–6027. (doi:10.1210/en.2008-0816)
Dushay J, Chui PC, Gopalakrishnan GS, Varela-Rey M, Crawley M, Fisher FM, Badman MK, Martinez-Chantar ML & Maratos-Flier E 2010 Increased fibroblast growth factor 21 in obesity and nonalcoholic fatty liver disease. Gastroenterology 139 456–463. (doi:10.1053/j.gastro.2010.04.054)
Fazeli PK, Misra M, Goldstein M, Miller KK & Klibanski A 2010 Fibroblast growth factor-21 may mediate growth hormone resistance in anorexia nervosa. Journal of Clinical Endocrinology and Metabolism 95 369–374. (doi:10.1210/jc.2009-1730)
Fisher FM, Chui PC, Antonellis PJ, Bina HA, Kharitonenkov A, Flier JS & Maratos-Flier E 2010 Obesity is a fibroblast growth factor 21 (FGF21)-resistant state. Diabetes 59 2781–2789. (doi:10.2337/db10-0193)
Han SH, Choi SH, Cho BJ, Lee Y, Lim S, Park YJ, Moon MK, Lee HK, Kang SW & Han DS et al. 2010 Serum fibroblast growth factor-21 concentration is associated with residual renal function and insulin resistance in end-stage renal disease patients receiving long-term peritoneal dialysis. Metabolism 59 1656–1662. (doi:10.1016/j.metabol.2010.03.018)
Iglesias P, Selgas R, Romero S & Díez JJ 2012 MECHANISMS IN ENDOCRINOLOGY: Biological role, clinical significance, and therapeutic possibilities of the recently discovered metabolic hormone fibroblastic growth factor 21. European Journal of Endocrinology 167 301–309. (doi:10.1530/EJE-12-0357)
Inagaki T, Dutchak P, Zhao G, Ding X, Gautron L, Parameswara V, Li Y, Goetz R, Mohammadi M & Esser V et al. 2007 Endocrine regulation of the fasting response by PPARα-mediated induction of fibroblast growth factor 21. Cell Metabolism 5 415–425. (doi:10.1016/j.cmet.2007.05.003)
Inagaki T, Lin VY, Goetz R, Mohammadi M, Mangelsdorf DJ & Kliewer SA 2008 Inhibition of growth hormone signaling by the fasting-induced hormone FGF21. Cell Metabolism 8 77–83. (doi:10.1016/j.cmet.2008.05.006)
Kharitonenkov A, Shiyanova TL, Koester A, Ford AM, Micanovic R, Galbreath EJ, Sandusky GE, Hammond LJ, Moyers JS & Owens RA et al. 2005 FGF-21 as a novel metabolic regulator. Journal of Clinical Investigation 115 1627–1635. (doi:10.1172/JCI23606)
Kharitonenkov A, Wroblewski VJ, Koester A, Chen YF, Clutinger CK, Tigno XT, Hansen BC, Shanafelt AB & Etgen GJ 2007 The metabolic state of diabetic monkeys is regulated by fibroblast growth factor-21. Endocrinology 148 774–781. (doi:10.1210/en.2006-1168)
Kralisch S & Fasshauer M 2011 Fibroblast growth factor 21: effects on carbohydrate and lipid metabolism in health and disease. Current Opinion in Clinical Nutrition & Metabolic Care 14 354–359. (doi:10.1097/MCO.0b013e328346a326)
Kubicky RA, Wu S, Kharitonenkov A & De Luca F 2012 Role of fibroblast growth factor 21 (FGF21) in undernutrition-related attenuation of growth in mice. Endocrinology 153 2287–2295. (doi:10.1210/en.2011-1909)
Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF, Feldman HI, Kusek JW, Eggers P, Van Lente F & Greene T et al. 2009 A new equation to estimate glomerular filtration rate. Annals of Internal Medicine 150 604–612.
Li H, Fang Q, Gao F, Fan J, Zhou J, Wang X, Zhang H, Pan X, Bao Y & Xiang K et al. 2010 Fibroblast growth factor 21 levels are increased in nonalcoholic fatty liver disease patients and are correlated with hepatic triglyceride. Journal of Hepatology 53 934–940. (doi:10.1016/j.jhep.2010.05.018)
Lin Z, Wu Z, Yin X, Liu Y, Yan X, Lin S, Xiao J, Wang X, Feng W & Li X 2010 Serum levels of FGF-21 are increased in coronary heart disease patients and are independently associated with adverse lipid profile. PLoS ONE 5 e15534. (doi:10.1371/journal.pone.0015534)
Lin Z, Zhou Z, Liu Y, Gong Q, Yan X, Xiao J, Wang X, Lin S, Feng W & Li X 2011 Circulating FGF21 levels are progressively increased from the early to end stages of chronic kidney diseases and are associated with renal function in Chinese. PLoS ONE 6 e18398. (doi:10.1371/journal.pone.0018398)
Lin Z, Gong Q, Wu C, Yu J, Lu T, Pan X, Lin S & Li X 2012 Dynamic change of serum FGF21 levels in response to glucose challenge in human. Journal of Clinical Endocrinology and Metabolism 97 E1224–E1228. (doi:10.1210/jc.2012-1132)
Mai K, Bobbert T, Groth C, Assmann A, Meinus S, Kraatz J, Andres J, Arafat AM, Pfeiffer AF & Mohlig M et al. 2010 Physiological modulation of circulating FGF21: relevance of free fatty acids and insulin. American Journal of Physiology. Endocrinology and Metabolism 299 E126–E130. (doi:10.1152/ajpendo.00020.2010)
Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF & Turner RC 1985 Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 28 412–419. (doi:10.1007/BF00280883)
Ogawa Y, Kurosu H, Yamamoto M, Nandi A, Rosenblatt KP, Goetz R, Eliseenkova AV, Mohammadi M & Kuro-o M 2007 βKlotho is required for metabolic activity of fibroblast growth factor 21. PNAS 104 7432–7437. (doi:10.1073/pnas.0701600104)
Reinehr T, Woelfle J, Wunsch R & Roth CL 2012 Fibroblast growth factor 21 (FGF-21) and its relation to obesity, metabolic syndrome, and nonalcoholic fatty liver in children: a longitudinal analysis. Journal of Clinical Endocrinology and Metabolism 97 2143–2150. (doi:10.1210/jc.2012-1221)
Semba RD, Sun K, Egan JM, Crasto C, Carlson OD & Ferrucci L 2012 Relationship of serum fibroblast growth factor 21 with abnormal glucose metabolism and insulin resistance: the Baltimore longitudinal study of aging. Journal of Clinical Endocrinology and Metabolism 97 1375–1382. (doi:10.1210/jc.2011-2823)
Shen J, Chan HL-Y, Wong GL-H, Choi PC-L, Chan AW-H, Chan H-Y, Chim AM-L, Yeung DK-W, Chan FK-L & Woo J et al. 2012 Non-invasive diagnosis of non-alcoholic steatohepatitis by combined serum biomarkers. Journal of Hepatology 56 1363–1370. (doi:10.1016/j.jhep.2011.12.025)
Stein S, Bachmann A, Lossner U, Kratzsch J, Bluher M, Stumvoll M & Fasshauer M 2009 Serum levels of the adipokine FGF21 depend on renal function. Diabetes Care 32 126–128. (doi:10.2337/dc08-1054)
Stein S, Stepan H, Kratzsch J, Verlohren M, Verlohren HJ, Drynda K, Lossner U, Bluher M, Stumvoll M & Fasshauer M 2010 Serum fibroblast growth factor 21 levels in gestational diabetes mellitus in relation to insulin resistance and dyslipidemia. Metabolism 59 33–37. (doi:10.1016/j.metabol.2009.07.003)
Stumvoll M, Van Haeften T, Fritsche A & Gerich J 2001 Oral glucose tolerance test indexes for insulin sensitivity and secretion based on various availabilities of sampling times. Diabetes Care 24 796–797. (doi:10.2337/diacare.24.4.796)
Tonjes A, Koriath M, Schleinitz D, Dietrich K, Bottcher Y, Rayner NW, Almgren P, Enigk B, Richter O & Rohm S et al. 2009 Genetic variation in GPR133 is associated with height: genome wide association study in the self-contained population of Sorbs. Human Molecular Genetics 18 4662–4668. (doi:10.1093/hmg/ddp423)
Tonjes A, Zeggini E, Kovacs P, Bottcher Y, Schleinitz D, Dietrich K, Morris AP, Enigk B, Rayner NW & Koriath M et al. 2010 Association of FTO variants with BMI and fat mass in the self-contained population of Sorbs in Germany. European Journal of Human Genetics 18 104–110. (doi:10.1038/ejhg.2009.107)
Tonjes A, Kralisch S, Lossner U, Kovacs P, Bluher M, Stumvoll M & Fasshauer M 2012 Metabolic and genetic predictors of circulating adipocyte fatty acid-binding protein. International Journal of Obesity 36 766–773. (doi:10.1038/ijo.2011.162)
Tyynismaa H, Raivio T, Hakkarainen A, Ortega-Alonso A, Lundbom N, Kaprio J, Rissanen A, Suomalainen A & Pietilainen KH 2010 Liver fat but not other adiposity measures influence circulating FGF21 levels in healthy young adult twins. Journal of Clinical Endocrinology and Metabolism 96 E351–E355. (doi:10.1210/jc.2010-1326)
Wang H, Qiang L & Farmer SR 2008 Identification of a domain within peroxisome proliferator-activated receptor gamma regulating expression of a group of genes containing fibroblast growth factor 21 that are selectively repressed by SIRT1 in adipocytes. Molecular and Cellular Biology 28 188–200. (doi:10.1128/MCB.00992-07)
Wu S, Levenson A, Kharitonenkov A & De Luca F 2012 Fibroblast growth factor 21 (FGF21) inhibits chondrocyte function and growth hormone (GH) action directly at the growth plate. Journal of Biological Chemistry 287 26060–26067. (doi:10.1074/jbc.M112.343707)
Yan H, Xia M, Chang X, Xu Q, Bian H, Zeng M, Rao S, Yao X, Tu Y & Jia W et al. 2011 Circulating fibroblast growth factor 21 levels are closely associated with hepatic fat content: a cross-sectional study. PLoS ONE 6 e24895. (doi:10.1371/journal.pone.0024895)
Yilmaz Y, Eren F, Yonal O, Kurt R, Aktas B, Celikel CA, Ozdogan O, Imeryuz N, Kalayci C & Avsar E 2010 Increased serum FGF21 levels in patients with nonalcoholic fatty liver disease. European Journal of Clinical Investigation 40 887–892. (doi:10.1111/j.1365-2362.2010.02338.x)
Zhang X, Ibrahimi OA, Olsen SK, Umemori H, Mohammadi M & Ornitz DM 2006 Receptor specificity of the fibroblast growth factor family. The complete mammalian FGF family. Journal of Biological Chemistry 281 15694–15700. (doi:10.1074/jbc.M601252200)
Zhang X, Yeung DC, Karpisek M, Stejskal D, Zhou ZG, Liu F, Wong RL, Chow WS, Tso AW & Lam KS et al. 2008 Serum FGF21 levels are increased in obesity and are independently associated with the metabolic syndrome in humans. Diabetes 57 1246–1253. (doi:10.2337/db07-1476)
(S Kralisch and A Tönjes contributed equally to this work)
