Next Article in Journal
Effects of Chronic Oral Probiotic Treatment in Paclitaxel-Induced Neuropathic Pain
Previous Article in Journal
Involvement of microRNA in Solid Cancer: Role and Regulatory Mechanisms
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Biomedicines
Volume 9
Issue 4
10.3390/biomedicines9040345
biomedicines-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

The Effect of FGF21 and Its Genetic Variants on Food and Drug Cravings, Adipokines and Metabolic Traits

by
Sarah Epperlein
1,
Claudia Gebhardt
2,
Kerstin Rohde
2,
Rima Chakaroun
1,
Marie Patt
1,
Imke Schamarek
1,
Susan Kralisch
1,
John T. Heiker
2,
Markus Scholz
3,
Michael Stumvoll
1,4,
Peter Kovacs
1,
Jana Breitfeld
1,* and
Anke Tönjes
1,*
1
Medical Department III–Endocrinology, Nephrology, Rheumatology, University of Leipzig Medical Center, 04103 Leipzig, Germany
2
Helmholtz Institute for Metabolic, Obesity and Vascular Research (HI-MAG) Helmholtz Center Munich at the University of Leipzig and the University of Leipzig Medical Center, 04103 Leipzig, Germany
3
Institute for Medical Informatics, Statistics and Epidemiology, Medical Faculty, University of Leipzig, 04107 Leipzig, Germany
4
Deutsches Zentrum für Diabetesforschung, 85764 Neuherberg, Germany
*
Authors to whom correspondence should be addressed.
Biomedicines 2021, 9(4), 345; https://doi.org/10.3390/biomedicines9040345
Submission received: 3 March 2021 / Revised: 24 March 2021 / Accepted: 25 March 2021 / Published: 29 March 2021
(This article belongs to the Section Gene and Cell Therapy)
Versions Notes

Abstract

:
Fibroblast growth factor 21 (FGF21) is a regulator of addictive behavior. Increasing evidence suggests an impact of FGF21 on eating behavior, food and drug cravings and on other adipokines like insulin-like growth factor 1 (IGF-1) or adiponectin. We investigated the association of serum FGF21 and genetic variants with aspects of food and drug craving and obesity related metabolic parameters including serum adipokine levels. Standardized questionnaires, blood samples and anthropometric data of the Sorbs cohort (n = 1046) were analyzed using SPSS. For genetic analyses, the FGF21-locus ±10 kb was genotyped and analyzed using PLINK. Validation was conducted in a second independent cohort (n = 704). FGF21 was significantly associated with alcohol and coffee consumption, smoking and eating behavior (disinhibition). We confirmed correlations of FGF21 serum levels with IGF-1, adiponectin, pro-enkephalin, adipocyte fatty-acid-binding protein, chemerin and progranulin. FGF21 genetic variants were associated with anthropometric and metabolic parameters, adipokines, food and drug craving while strongest evidence was seen with low-density lipoprotein cholesterol (LDL-C). We highlight the potential role of FGF21 in food and drug cravings and provide new insights regarding the link of FGF21 with other adipokines as well as with metabolic traits, in particular those related to lipid metabolism (LDL-C).

1. Introduction

The human fibroblast growth factor 21 (FGF21) is a protein consisting of 209 amino acids [1] which is encoded by the FGF21 gene on chromosome 19q13.33 [2]. In humans, FGF21 is mainly expressed in the liver, but also to a minor extent in the thyroid gland [2]. An increase of circulating FGF21 can be observed e.g., after acute exercising [3], long-term fasting [4], very low caloric diet [5], excessive intake of carbohydrates [6], in the postprandial period [7], after alcohol consumption [8], and lipid infusion [9], in type 2 diabetes (T2D) [10], metabolic syndrome (MS) and obesity [11,12]. Several studies in mice showed FGF21 to suppress ad libitum alcohol intake [8], to have a glucose and lipid lowering effect [13] as well as an insulin sensitizing [14] effect. Furthermore, FGF21 is thought to induce weight loss in obese mice [15], probably through affecting peripheral tissues, for example by upregulating glucose transporter 1 (GLUT 1) in adipocytes and by increasing peripheral glucose expenditure [16]. In humans, FGF21 regulates glucose uptake in monocytes in vitro, may have an impact on immune response [17] and correlates with impaired glucose tolerance, serving as a predictor for T2D [12,18,19]. Moreover, FGF21 as a potential biomarker of non-alcoholic-fatty-liver disease (NAFLD) may play a role in hepatic lipid metabolism as its serum concentration is elevated in NAFLD and its expression correlates positively with liver steatosis [20]. Furthermore, the relationship of FGF21 with other adipokines has recently been addressed. Exemplarily, FGF21 inhibits insulin-like growth factor 1 (IGF-1) secretion in vitro [21] and its activity [22]. In Chinese children with obesity, high FGF21 serum levels correlated inversely with adiponectin levels implicating a potential FGF21 resistance in obese states [23], since FGF21 treatment of mice results in increased adiponectin secretion [24]. Finally, studies in mice also demonstrated that FGF21 signals to the central nervous system e.g., via the paraventricular nucleus to suppress food intake [25]. Moreover, experimental models suggested a potential role of FGF21 in addictive behavior by showing that suppression of alcohol and sweets consumption by FGF21 administration is accompanied by a lower concentration of dopamine in the central nervous system [26].
Based on the current knowledge, FGF21 has become a promising target for pharmacotherapy. Indeed, FGF21-analogons administered to patients with obesity and T2D resulted in weight loss, lower low-density lipoprotein cholesterol (LDL-C) and triglycerides (TG), higher high-density lipoprotein cholesterol (HDL-C), but also higher blood pressure. Additionally it improved fasting insulin (FI) and adiponectin levels [27,28,29]. Subcutaneous administration of a FGF21/β-klotho complex reduced body weight and the cardio-metabolic risk in humans while data also suggest effects on sweet preference and the consumption of carbohydrate rich foods [30].
In line with the physiological effects of FGF21, genetic variants within or near the FGF21 locus were shown to be associated with changes in eating behavior, consumption of addictive substances and metabolic and anthropometric parameters [31,32,33,34,35]. However, in the majority of the studies, circulating FGF21 levels were not measured and therefore prediction of potential causal mechanisms was not possible [31,33,34]. Thus, one objective of this study was to investigate the association of FGF21 serum concentrations with food and drug craving (eating behavior, alcohol, coffee, smoking). Given the proposed relationship of FGF21 with other adipokines, we tested the association of FGF21 with numerous adipokines, namely chemerin, vaspin, pro-neurotensin (pro-NT), pro-enkephalin (PENK), progranulin, adiponectin, adipocyte fatty-acid-binding protein (AFABP), irisin, angiopoietin-related growth factor (AGF), and IGF-1. Finally, we examined whether common genetic variants in the FGF21 locus were related to circulating FGF21 serum levels, the expression of FGF21 in peripheral blood mononuclear cells (PBMCs), metabolic and anthropometric phenotypes including the measured adipokines and food and drug craving.

2. Materials and Methods

2.1. Subjects

Participants included in the present analyses were part of the Sorbs cohort from Eastern Germany (described in detail elsewhere [21,36,37,38]). Briefly, the Sorbs cohort contains a total of 1046 subjects. Regarding clinical association analyses 780 subjects (330 males and 450 females) were included for which serum FGF21 level were available. For genetic association analyses the entire cohort of n = 1046 was included. Baseline characteristics of the study subjects are shown in Table 1. The study has been approved by the ethic committee of the University of Leipzig (Reg. No.: 088-2005 and for the adipokine genome-wide association study A.Z.: 330-12-24092012).
To validate findings concerning FGF21 and alcohol consumption or smoking, we analyzed these traits in a separate validation cohort from Leipzig (Table 2). This cohort consisted of a subgroup (n = 704) of larger recruitment efforts at the Integrated Research and Treatment Center for Adiposity Diseases (Leipzig, Germany) to assess metabolic disease progression in a wide spectrum of phenotypes, ranging from healthy controls, metabolic disease, lean and overweight subjects with T2D and subjects with morbid obesity. Subjects were screened for standardized inclusion and exclusion criteria. The latter comprised acute or chronic inflammatory or infectious diseases, history of organ transplantation including immunosuppressive treatment, severe kidney failure (estimated glomerular filtration rate (eGFR) 50 mL/(min·1.73 m2)) and drug or alcohol addiction.

2.2. Phenotyping

2.2.1. Sorbs Cohort

Anthropometric data and data of past medical history (smoking, alcohol and coffee consumption) were collected using standardized questionnaires. Body impedance analyses and a 75-g oral glucose tolerance test (OGTT) was performed and insulin resistance was defined by calculating the homeostasis model assessment for insulin resistance (HOMA-IR) as described elsewhere [39].
Blood samples were drawn in the morning after overnight fasting. By using standard laboratory methods LDL-C, HDL-C, TG, γ-glutamyltransferase (γGT), alanine aminotransferase (ALAT), aspartate aminotransferase (ASAT), alkaline phosphatase (AP), fasting glucose (FG), hemoglobin A1c (HbA1c (%)) were determined in a certified laboratory (University of Leipzig Institute of Laboratory Medicine). Serum insulin levels were measured by the AutoDELFIA Insulin assay (PerkinElmer Life and Analytical Sciences, Turku, Finland). Serum concentrations of FGF21 and other adipokines (chemerin, vaspin, pro-NT, PENK, progranulin, adiponectin, AFABP, irisin, AGF, IGF-1) were determined using commercially available enzyme-linked immunosorbent assays (ELISAs) following the manufacturer’s protocol. The limit of detection for serum FGF21 concentrations was 7 ng/L, the intra-assay was 2% and inter-assay 3.3% (BioVendor, Brno, Czech Republic). Quality criteria and characteristics of the laboratory methods, the AutoDELFIA and the applied ELISAs were described in detail recently [21,40]. The mRNA expression of FGF21 was assessed in PBMCs extracted from anti-coagulated blood samples [41] and DNA was extracted and genotyped as described in detail elsewhere [36,37,38]. Briefly, the QIAmp DNA Blood Midi Kit (Qiagen Inc., Valencia, CA, USA) was used to extract genomic DNA following the manufacturer’s instructions and genotyping was performed using the 500K Affymetrix GeneChip and the Affymetrix Genome-Wide Human SNP Array 6.0 (Affymetrix Inc., Santa Clara, CA, USA) by the Microarray Core Facility of the Interdisciplinary Center for Clinical Research (University of Leipzig, Germany) and by ATLAS Bio-labs GmbH (Berlin, Germany). The genotypes have further been determined using the GeneChip Genotyping Analysis Software (GTYPE) and the BRLMM algorithm for the 500K arrays and the Birdseed Algorithm for Genome-Wide Human SNP Array 6.0 (Affymetrix Inc.). For gene expression in PBMCs the blood samples have been collected in VACUTAINER CPTs (Cell Preparation Tubes) containing sodium heparin as anti-coagulant. To extract the PBMCs the manufacturer’s protocol was used (BD, Franklin Lakes, NJ, USA). Further, RNA was extracted using the TRIzol protocol (Thermo Fisher Scientific), followed by DNase I digestion and clean-up using the RNeasy MinElute Cleanup Kit (Qiagen, Hilden, Germany). Illumina GeneChip analyses was done at the microarray core facility of the Interdisciplinary Centre for Clinical Research (ICCR) in Leipzig (Dr. Krohn, Faculty of Medicine, University of Leipzig) and the Illumina Human HT-12 v4 Bead Chip was used according to the manufacturer’s instructions.
Eating behavior was assessed by the German Version of the Three-factor eating questionnaire, referred to as “Fragebogen zum Essverhalten (FEV). The FEV assesses three domains of eating behavior: “disinhibition” is comprised of 16 items and is defined as being unable to control food intake under various circumstances for example due to emotional stress. “Hunger” comprises 14 items and is defined as intense sensing of hunger signs. The domain “restraint” comprised of 21 items, describes the ability of cognitive control of eating behavior to reduce or maintain body weight. Higher scores within a single domain indicate that an individual’s eating behavior is best described by this domain [42,43,44,45].

2.2.2. Validation Cohort from Leipzig

Phenotyping of the validation cohort was performed according to standard operating procedures including the assessment of clinical parameters and anthropometric data as well as the acquisition of biological samples. Dietary assessment was performed using a validated semiquantitative Food Frequency Questionnaire (FFQ) [46]. The layout and format were based on the European Prospective Investigation of Cancer (EPIC)-Norfolk FFQ. Portion size and nutrient composition were derived from national food consumption surveys and food composition databases as previously described [46]. Smoking status was assessed in a dichotomous manner evaluating current active/passive status (yes/no). Whole blood was spun down for serum extraction and serum samples were stored at −80 °C until further processing. FGF21 was measured in serum samples using an ELISA sandwich assay according to manufacturer’s recommendations (BioVendor) in 492 subjects.

2.3. Statistical Association Analyses

2.3.1. Clinical Parameters

In the Sorbs cohort, FGF21 serum measurements were performed for 927 subjects. 147 participants displayed serum concentration below the detection limit. Thus, FGF21 serum levels were available in 780 subjects. Clinical association analyses of serum FGF21 levels were performed using SPSS (IBM Corp. Released 2016. IBM SPSS Statistics for Windows, Version 24.0. IBM Corp., Armonk, NY, USA). Multiple linear regression analyses were performed separately for each dependent variable: disinhibition score, hunger score, restraint score, pack years (number of packs of cigarettes/day × amounts of active smoker years) and FGF21 mRNA expression in PBMCs. For ordinal-scaled parameters (alcohol and coffee) an ordinal regression (generalized linear model) was performed. These traits were categorized as followed: alcohol (1 = no consumption, 2 = occasionally/at parties, 3 = two to three glasses of wine or bottles of beer/week, 4 = one glass of wine or one bottle of beer/day, 5 = two glasses of wine or bottles of beer/day, 6 = three or more/day); coffee (0 = no consumption, 1 = one cup/day, 2 = two cups/day, 3 = three cups/day, 4 = four or more cups/day). To test whether serum FGF21 concentrations were associated with smoking status, binary logistic regression was applied. To analyze the relationships of different adipokines and FGF21, Spearman’s rank correlation coefficient was calculated and partial correlation including adjustments for age, sex and body mass index (BMI) was conducted. All continuous parameters (serum FGF21 concentrations, age, BMI, adipokines, pack years and expression parameters) were ln-transformed before entering statistical analysis.
In the validation cohort, analyses were performed using SPSS. Continuous parameter (serum FGF21 concentrations, age, BMI, alcohol consumption) were ln-transformed before entering statistical analysis. The association between alcohol consumption and FGF21 serum concentrations was investigated using linear regression analysis as the applied score of ‘food alcohol’ is a continuous parameter in contrast to the ‘score alcohol’ in the Sorbs cohort which is ordinal scaled. To test whether FGF21 concentrations significantly differed between smokers and non-smokers binary logistic regression was performed.

2.3.2. Genetic Association Studies

We performed genetic association analyses using PLINK in the Sorbs [47]. With the exception of analysis including serum FGF21 concentrations (n = 780), all subjects were included (n = 1046). The analysis of parameters of lipid metabolism (TG, LDL-C, HDL-C) included only subjects without lipid-lowering medication (n = 678). 15 single nucleotide polymorphisms (SNPs) from the FGF21 gene locus ±10 kb filtered from the genome wide data set using imputed variants based on 1000 Genome information [48] were analyzed. A brief description of the genetic variants is provided in the Supplementary File 1. Inclusion thresholds were set as followed: minor allele frequency 0.01, genotyping rate 0.9 and Hardy-Weinberg-Equilibrium 0.0001. One SNP (rs499765) failed the Hardy-Weinberg test and was excluded from further analyses. For the remaining 14 SNPs linear regression analyses were performed for all traits except for alcohol, coffee and smoking (yes/no) which were analyzed using logistic regression models. Ordinal traits were transformed to binary coded variables before being taken forward to logistic regression analyses using PLINK. Smoking was coded to 1 = no smoker and 2 = smoker; alcohol consumption was coded to 1 = low consumption (scores 1 to 3 from the scales above) and 2 = high consumption (scores 4 to 6 from the scales above); coffee consumptions was coded: 1 = low consumption (scores 0 to 1 from the scales above) and 2 = high consumption (scores 2 to 4 from the scales above). All analyses were adjusted for sex, age and BMI (for analyses of BMI only age and sex were considered). All variables were ln-transformed except alcohol, coffee, scores of eating behavior and sex. Results were condensed into 8 linkage disequilibrium (LD-) groups (LD-R2 0.8) showing values of the representative SNP which we chose to be the one with the most associations to different traits (Table 3), results of all analyzed SNPs are shown in Supplementary Files 2–7.

3. Results

3.1. Association of Serum FGF21 Concentrations with Clinical Traits

3.1.1. Food and Drug Cravings

In the Sorbs, serum FGF21 concentrations were associated with disinhibition (β = −0.152, p = 0.001), which remained significant even after adjustment for sex, age and BMI (β = −0.158, padj = 2.17 × 10−4). Serum FGF21 concentrations significantly associated with increased risk of being a smoker (odds ratio (OR) (95% confidence interval) = 1.34 (1.16; 1.55), p = 7.20 × 10−5 and ORadj (95% confidence interval) = 1.39 (1.18; 1.64), padj = 6.70 × 10−5) and a more intense consumption of alcohol (OR = 1.32 (1.16; 1.49), p = 1.30 × 10−5 and ORadj = 1.28 (1.13; 1.46), padj = 1.51 × 10−4). In line with this, elevated circulating FGF21 level were associated with increased pack years in smokers (β = 0.239, p = 6.80 × 10−5 and βadj = 0.111, padj = 0.030). In contrast FGF21 concentrations were associated with a lower intense of coffee consumption (OR = 0.87 (0.77; 0.99), p = 0.029 and ORadj = 0.80 (0.70; 0.91), padj = 0.001) (Table 4).
We also found a significant association between serum FGF21 concentrations and alcohol consumption in our validation cohort (food alcohol: β = 0.162, p = 3.00 × 10−4, and βadj = 0.164, padj = 1.00 × 10−3) (Table 5). In line with our findings in the Sorbs, higher FGF21 serum concentrations also associated with increased risk of being identified as smoker in the validation cohort (OR = 1.32 (1.07; 1.64), p = 0.011 and ORadj = 1.67 (1.33; 2.16), padj = 4.10 × 10−5) (Table 5).
Beyond that, the association of mRNA expression of FGF21 in PBMCs and serum FGF21 concentrations was significant (β = 0.083, p = 0.022) and independent of age, sex and BMI (p 0.05 for each single variable; β = 0.069, p = 0.067 adjusted for all three parameters).

3.1.2. Adipokines

Among the analyzed adipokines, AFABP, chemerin and IGF-1 showed strongest significant partial correlations with serum FGF21 concentrations (all p 0.001 adjusted for age, sex, BMI): AFABP (β = 0.248), chemerin (β = 0.200) and IGF-1 (β = −0.195). Further, significant correlations of serum FGF21 comprised progranulin (β = 0.098, p = 0.019) and PENK (β = 0.092, p = 0.028), whereas adiponectin, irisin, vaspin, AGF and pro-NT were not significantly correlated with circulating FGF21 concentrations including the adjustments. However, in Spearman’s rank correlation without adjustment adiponectin showed significant correlation with FGF21 (p = −0.092, p = 0.013) (Table 6).

3.2. Association of Genetic Variants in FGF21 with Metabolic Traits

3.2.1. Association of rs838133 with FGF21 Concentrations and Metabolic Traits Related to Obesity

After quality control 14 common variants in the FGF21 locus ±10 kb within eight different LD-groups were selected for further analyses (Table 3). We performed association analyses with the recently described SNP rs838133, whose A-allele has been shown to be associated with changes in food and drug craving and metabolic and anthropometric parameters [31,32,49,50]. In the present study the minor G-allele was significantly associated with lower LDL-C in the Sorbs cohort (β = −0.034, p = 0.005); (LD2 in Table 7, Supplementary File 2).

3.2.2. Effects of Common Genetic Variation within the FGF21 Locus ± 10 kb

Three SNPs from two different LD-groups were significantly associated with serum FGF21 level (Table 8): the SNPs in LD-group 1—rs4015 (β = 0.227, p = 0.010) and in LD-group 4—rs11667321, rs8106205 (β = 0.108, p = 0.037) (for details see Supplementary File 3). Two independent SNPs were significantly associated with height: rs4021 (β = −0.004, p = 0.031, LD-group 3) and rs838143 (β = −0.009, p = 0.003, LD-group 7). The same two independent SNPs were significantly associated with weight: rs4021 (β = −0.009, p = 0.022, LD-group 3) and rs838143 (β = −0.018, p = 0.003, LD-group 7) (in detail see Supplementary File 4).
Two LD-groups showed a significant association with FI (Table 7): rs4015 (β = 0.128, p = 0.002) and rs12611028 (β = 0.120, p = 0.008) from LD-group 1 and rs3826821 (β = 0.111, p = 0.033) from LD-group 6. Furthermore, two of these SNPs were significantly associated with HOMA-IR: rs3826821 (β = 0.135, p = 0.017, LD-group 6) and rs4015 (β = 0.109, p = 0.019; LD-group 1) (Table 7; details provided in Supplementary File 5). Nine SNPs from four different LD-groups showed a significant association with LDL-C: LD-group 2: rs838133 (β = −0.034, p = 0.005) and rs838144 (β = −0.035, p = 0.003); LD-group 4: rs12975033, rs8104897, rs8105137, rs8106205, rs11667321 (all p 0.001); LD-group 5 with rs12975781 (β = −0.032; p = 0.018); and LD-group 8 with rs838136 (β = −0.041, p = 0.005) (Table 7; see Supplementary File 2). In the analyses of liver parameters, we observed a significant association of rs838143 (LD-group 7) with γGT (β = −0.113, p = 0.034) and of rs12975033 (LD-group 4) with ASAT (β = −0.022, p = 0.047) (Table 7; for details see Supplementary File 6). In addition to the association with LDL-C, the LD-group 4 variants (with rs12975033, rs8104897, rs8105137) showed a significant effect on IGF-1 (all p 0.037) (Table 9; see Supplementary File 3). Further associations with adipokines were seen for rs4015 (LD-group 1) which was significantly associated with adiponectin (β = 0.057, p = 0.031) and with AFABP (β = 0.088, p = 0.028), for rs4021 (LD-group 3) which was significantly associated with pro-NT (β = −0.044, p = 0.045), and rs3826821 (LD-group 6) which was significantly associated with PENK (β = 0.066, p = 0.009) (Table 9; Supplementary File 3).
In the category of food and drug craving (Table 10; details provided in Supplementary File 7) rs12975781 (LD-group 5) was significantly associated with “hunger” (β = −0.388, p = 0.023) and rs12611028 (LD-group 1) with alcohol consumption (OR = 1.625, p = 0.020). Furthermore, there was a significant association of rs4021 (OR = 1.292, p = 0.038; LD-group 3) and rs838136 (OR = 1.299, p = 0.030; LD-group 8) with smoking status (Table 10 and Supplementary File 7).

4. Discussion

4.1. Association of FGF21 with Food and Drug Craving

Genetic variants within or near the FGF21 locus have been shown to be associated with eating behavior, consumption of addictive substances and metabolic and anthropometric parameters [31,32,33,34,35]. FGF21 administration in mice suppresses alcohol and sweets consumption, accompanied by a lower concentration of dopamine in the central nervous system implicating its potential role in addictive behavior [26]. In the present work, we investigated the association of serum FGF21 with food and drug craving in humans.
The present study revealed a significant negative association of serum FGF21 concentrations with the eating behavior domain of disinhibition while no significant associations with the domains of hunger or restraint were found. It is of note, that Löffler et al. defined an additional factor based on the Three-Factor-Eating-Questionnaire (German version: FEV) named ‘uncontrolled eating’ which consists of items of the domains of ‘disinhibition’ and ‘hunger’ to show that this factor conveys the strongest positive correlation with BMI [42]. It has to be taken into account that despite the robust evidence for the relationships between FGF21 and BMI as well as disinhibition, the current knowledge does not allow drawing any conclusions with regard to possible mechanistic pathways. Nevertheless, FGF21 is known to be elevated in patients with T2D, obesity and with MS [10,11,21], mostly having a higher BMI. Since FGF21 induces weight loss in obese mice [15] and in diabetic monkeys [51] as well as FGF21 analogues do in humans [27,28], it might be assumed that elevated FGF21 levels in humans may reflect a similar pattern. Whether the effect of FGF21 on obesity is mediated by eating behavior, or changes in FGF21 and the disinhibition domain are driven by obesity (BMI) itself, or whether FGF21 is a confounder of the association between disinhibition and BMI will need to be clarified in further studies focusing on mechanistic pathways underlying changes in FGF21 concentrations. It should also be taken into account, that findings in mice suggested development of FGF-21 resistance in obesity [52].
In the present study, we highlight the role of FGF21 in alcohol consumption. Søberg et al. showed elevated serum FGF21 levels in humans after acute alcohol consumption and a higher basal rate of FGF21 after three days of higher alcohol consumption on events like the German “Oktoberfest” [8]. Studies in mice revealed that administration of FGF21 reduced alcohol intake [8,26]. Further studies conducted in mice provide evidence for a liver-brain axis of FGF21 regulating alcohol intake via β-Klotho [53] and suggest that FGF21 protects the liver from alcohol associated injuries [54]. Taken together, it might be hypothesized that an endogenous increase of FGF21 after alcohol consumption, either acute or chronic, could be part of a negative feedback mechanism to reduce further alcohol consumption.
Finally, we found a significant positive association of serum FGF21 with smoking, and a negative association with coffee consumption. In contrast to our findings a previously conducted study in smokers revealed no significant effect of smoking stop on FGF21 serum concentrations [55] and the existing literature regarding the association of genetic variation in FGF21 and smoking is inconsistent. In a study conducted by Frayling et al. only a trend was found towards an association of the rs838133 A-allele with being former smoker, whereas no effect at all was observed with being current smoker [31]. In contrast, Søberg et al. showed a significant association of the rs838133 A-allele with daily smoking and a trend towards association with occasional smoking [32].
Similarly, existing literature concerning the association between genetic variation in FGF21 and coffee consumption is controversial. Whereas Søberg et al., did not find a significant association of the rs838133 A-allele with coffee consumption [32] which is in line with our present findings, Frayling et al., reported a significant association of the rs838133 A-allele with lower risk of being a coffee drinker and with reduced number of cups of coffee consumed per day [31].

4.2. Association of FGF21 with Adipokines

In the present study we observed that circulating serum FGF21 was significantly correlated with AFABP and chemerin, but also with progranulin and PENK independent of sex, age and BMI. It has to be acknowledged that a correlation of FGF21 with adiponectin seemed to be independent of age, but not sex and BMI (data not shown).
FGF21 is a suppressor of IGF-1 secretion in vitro [21] as well as of growth hormone activity by either decreasing expression of IGF-1 or upregulating its binding protein or the suppressor of cytokine signaling 2 (SOCS2) [22]. Consequently, serum FGF21 concentrations correlated inversely with serum IGF-1 concentrations in humans [21]. Additionally, we have recently shown that certain features of the MS are highly associated with circulating adiponectin, AFABP, chemerin and FGF21 [40]. However, it remains unclear, whether the MS may act as a driver or is simply a consequence of changes in adipokine regulation. Moreover, accumulating evidence points to a direct interaction between individual adipokines. For example, a study in mice conducted by Lin et al. showed that the adiponectin secreted by adipocytes is required for FGF21 to exert its effects on energy metabolism and insulin sensitivity in tissues like liver or skeletal muscle [56]. Another study in mice showed that FGF21 increases adiponectin secretion [24]. It is noteworthy that adiponectin itself plays an important role in mediating insulin sensitivity [57] and is associated with lower risk of T2D [58]. Adiponectin, as it correlates inversely with TG, insulin resistance, C-reactive protein and waist circumference and positively with HDL-C [40,59], is when decreased a potential marker of the MS [60]. A study in Chinese youths demonstrated that obesity is associated with higher FGF21 on the one hand and higher and lower adiponectin levels on the other hand [23], which is also in line with our findings. This point is of relevance when considering a possible FGF21 resistance or insulin resistance in obesity and T2D.

4.3. Effects of Genetic Variation within the FGF21 Locus

Based on previously published associations and pharmacological effects of FGF21, including genetic association studies mainly focusing on rs838133, rs838145 and rs838147 [31,32,33,34], we investigated the association of FGF21 genetic variants (spanning the region of FGF21 ±10kb) with FGF21 serum concentrations, FGF21 expression in PBMCs, metabolic traits, adipokines, as well as food and drug craving. Significant associations were mainly seen with LDL-C with 9 SNPs from 4 LD-clusters, thus being in line with previously published studies reporting genetic association of FGF21 [49] and association of serum FGF21 with LDL-C [12]. Based on the present analyses we identified three variants being associated with FGF21 serum concentrations as well as variants associated with the eating behavior trait ‘hunger’, alcohol consumption and smoking status. Interestingly, genetic variants in the co-receptor of FGF21, β-Klotho, through which FGF21 regulates alcohol intake, have previously been shown to be associated with alcohol consumption [53,61].
Chu et al. and Frayling et al. have previously shown that the rs838133 A-allele is significantly associated with lower protein and fat intake but higher carbohydrate intake in large-scale genome-wide association studies (GWAS) [31,33]. Furthermore, Frayling et al. reported the minor A-allele to be associated with higher alcohol intake, waist-to-hip ratio (WHR), blood pressure, LDL-C, γGT, TG, body fat, AP and coffee consumption [31,49,50]. Søberg et al. found the rs838133 A-allele to be associated with higher candy snacking and nominally associated with higher alcohol consumption and daily smoking [32]. In contrast to the study of Frayling et al. [31], no association of genetic variation with coffee consumption was found [32]. In the present study, a significant association was replicated only between rs838133 and LDL-C. In line with previous findings [49], the A-allele was associated with higher LDL-C in our study as well. Furthermore, we could replicate a trend towards an association of rs838133 with disinhibition (p = 0.053) and a trend with AP (p = 0.077). Diverging results in the present study may be a consequence of the smaller sample size and thereby limited statistical power. Beyond that, different questionnaires were applied to assess alcohol and coffee consumption, eating behavior and smoking habits. Previous genetic studies did not find an association of FGF21 genetic variants and corresponding LD-partners with FGF21 expression in liver [31] or blood [35,62].
Also, we could not find an association of genetic variants with FGF21 expression in PBMCs. However, the present findings show a trend toward an association between serum FGF21 levels and its expression in PBMCs.
In the context of genetic variation, two genetic variants which map outside the here examined FGF21 locus have to be mentioned: rs838145 and rs838147 [32,34,35]. Both variants are in LD (R2 0.8) with rs838133. Rs838133 and rs838145 are known to be associated with higher sugar intake and sweet snacking [32]. Furthermore the rs838145 G-allele or the corresponding alleles of their LD-partner were found to be associated with higher carbohydrate, lower fat intake higher serum concentrations of FGF21 but not with mRNA amount in blood [35,62,63]. Consistently, the A-allele of rs838145 was reported to be associated with higher fat intake [63] and lower alcohol consumption [64]. The rs838147 alleles were found to interact with differences in dietary macronutrient intake and the subjects’ preference for it [34] and the A-allele was found to be significantly associated with higher carbohydrate intake on genome-wide scale [35]. However, we could not replicate any of these associations in our analysis. This may be attributed to the fact that we used the FEV questionnaires instead of measuring macronutrient intake. Furthermore, as mentioned above, the sample sizes were different between the studies, strongly influencing the statistical power to detect effects of genetic variants on measured phenotypes. However, they all showed a significant effect on LDL-C, with rs838145 conveying the strongest association (β = 0.040, p 0.001).

5. Conclusions

In summary, our study provides novel findings that may promote further work towards better understanding of the role of FGF21 in regulation of behavioral and metabolic traits related to obesity. Nevertheless, the interpretation of the present results is limited as no correction for multiple testing was applied and therefore, validation in large-scale association studies is warranted. However, our study clearly suggests a link between FGF21 and eating behavior, alcohol consumption, smoking and coffee consumption, which may comprise a component of food and drug craving. Whether lower dopamine concentration in the central nervous system are implicated in these associations like it has been reported in mice [26] remains to be elucidated in humans. Furthermore, we present new results regarding the relationship between FGF21 and other adipocytokines that may uncover and help to understand complex pathways in the pathophysiology of metabolic diseases. Finally, the role of FGF21 in the pathophysiology of metabolic diseases is supported by our genetic analyses providing evidence for a link between genetic variation within FGF21 and metabolic traits, in particular those related to lipid metabolism.

Supplementary Materials

The following are available online at https://www.mdpi.com/article/10.3390/biomedicines9040345/s1. Supplementary File 1: Brief description of analyzed genetic variants in the FGF21 locus ± 10kb. Supplementary File 2: Association of FGF21 genetic variants with lipid traits. Supplementary File 3: Association of FGF21 genetic variants with serum FGF21, FGF21 gene expression and further adipocytokines. Supplementary File 4: Association of FGF21 genetic variants with anthropometric traits. Supplementary File 5: Association of FGF21 genetic variants with glucose traits. Supplementary File 6: Association of FGF21 genetic variants with liver parameters. Supplementary File 7: Association of FGF21 genetic variants with eating behavior and parameters of food and drug craving.

Author Contributions

Conceptualization, P.K. and A.T.; methodology, S.E.; software, S.E., C.G., M.S. (Markus Scholz); validation, M.P.; resources, M.S. (Michael Stumvoll), R.C.; data curation, K.R., S.K.; writing—original draft preparation, S.E. and J.B.; writing—review and editing, J.T.H., C.G., I.S., M.S. (Markus Scholz); supervision, J.B. and A.T.; project administration, A.T., P.K.; funding acquisition, A.T., P.K., M.S. (Michael Stumvoll), R.C. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by grants from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation—Projektnummer 209933838–SFB 1052 “Obesity mechanisms”; B03, C06, C07); DFG Priority Program “Thyroid Transact” SPP1629 (TO 718/2- 1) and from Deutsches Zentrum für Diabetesforschung (82DZD06D03).

Institutional Review Board Statement

The two studies were conducted to the guidelines of the Declaration of Helsinki. Both studies were approved by the local ethics committee of the University of Leipzig (for the Sorbs cohort Reg.No: 088-2005 and AZ.: 330-12-24092012; for the validation cohort: 047-13-28012013).

Informed Consent Statement

All participants gave written informed consent before taking part in the study.

Data Availability Statement

All data generated or analyzed during this study are included in this published article.

Acknowledgments

We thank all those who participated in the studies and we thank Lucas Massier for his support.

Conflicts of Interest

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

References

  1. Nishimura, T.; Nakatake, Y.; Konishi, M.; Itoh, N. Identification of a novel FGF, FGF-21, preferentially expressed in the liver. Biochim. Biophys. Acta 2000, 1492, 203–206. [Google Scholar] [CrossRef]
  2. National Center for Biotechnology Information (NCBI). U.S. Rockville Pike. 2020. Available online: https://www.ncbi.nlm.nih.gov/gene/26291 (accessed on 15 December 2020).
  3. Tanimura, Y.; Aoi, W.; Takanami, Y.; Kawai, Y.; Mizushima, K.; Naito, Y.; Yoshikawa, T. Acute exercise increases fibroblast growth factor 21 in metabolic organs and circulation. Physiol. Rep. 2016, 4. [Google Scholar] [CrossRef] [Green Version]
  4. Fazeli, P.K.; Lun, M.; Kim, S.M.; Bredella, M.A.; Wright, S.; Zhang, Y.; Lee, H.; Catana, C.; Klibanski, A.; Patwari, P.; et al. FGF21 and the late adaptive response to starvation in humans. J. Clin. Investig. 2015, 125, 4601–4611. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  5. Mraz, M.; Bartlova, M.; Lacinova, Z.; Michalsky, D.; Kasalicky, M.; Haluzikova, D.; Matoulek, M.; Dostalova, I.; Humenanska, V.; Haluzik, M. Serum concentrations and tissue expression of a novel endocrine regulator fibroblast growth factor-21 in patients with type 2 diabetes and obesity. Clin. Endocrinol. 2009, 71, 369–375. [Google Scholar] [CrossRef]
  6. Lundsgaard, A.-M.; Fritzen, A.M.; Sjøberg, K.A.; Myrmel, L.S.; Madsen, L.; Wojtaszewski, J.F.P.; Richter, E.A.; Kiens, B. Circulating FGF21 in humans is potently induced by short term overfeeding of carbohydrates. Mol. Metab. 2017, 6, 22–29. [Google Scholar] [CrossRef]
  7. Samms, R.J.; Lewis, J.E.; Norton, L.; Stephens, F.B.; Gaffney, C.J.; Butterfield, T.; Smith, D.P.; Cheng, C.C.; Perfield, J.W.; Adams, A.C.; et al. FGF21 Is an Insulin-Dependent Postprandial Hormone in Adult Humans. J. Clin. Endocrinol. Metab. 2017, 102, 3806–3813. [Google Scholar] [CrossRef] [Green Version]
  8. Søberg, S.; Andersen, E.S.; Dalsgaard, N.B.; Jarlhelt, I.; Hansen, N.L.; Hoffmann, N.; Vilsbøll, T.; Chenchar, A.; Jensen, M.; Grevengoed, T.J.; et al. FGF21, a liver hormone that inhibits alcohol intake in mice, increases in human circulation after acute alcohol ingestion and sustained binge drinking at Oktoberfest. Mol. Metab. 2018, 11, 96–103. [Google Scholar] [CrossRef]
  9. Mai, K.; Andres, J.; Biedasek, K.; Weicht, J.; Bobbert, T.; Sabath, M.; Meinus, S.; Reinecke, F.; Möhlig, M.; Weickert, M.O.; et al. Free fatty acids link metabolism and regulation of the insulin-sensitizing fibroblast growth factor-21. Diabetes 2009, 58, 1532–1538. [Google Scholar] [CrossRef] [Green Version]
  10. Chavez, A.O.; Molina-Carrion, M.; Abdul-Ghani, M.A.; Folli, F.; Defronzo, R.A.; Tripathy, D. Circulating fibroblast growth factor-21 is elevated in impaired glucose tolerance and type 2 diabetes and correlates with muscle and hepatic insulin resistance. Diabetes Care 2009, 32, 1542–1546. [Google Scholar] [CrossRef] [Green Version]
  11. Zhang, X.; Yeung, D.C.Y.; Karpisek, M.; Stejskal, D.; Zhou, Z.-G.; Liu, F.; Wong, R.L.C.; Chow, W.-S.; Tso, A.W.K.; Lam, K.S.L.; et al. Serum FGF21 levels are increased in obesity and are independently associated with the metabolic syndrome in humans. Diabetes 2008, 57, 1246–1253. [Google Scholar] [CrossRef] [Green Version]
  12. Lee, Y.; Lim, S.; Hong, E.-S.; Kim, J.H.; Moon, M.K.; Chun, E.J.; Choi, S.I.; Kim, Y.-B.; Park, Y.J.; Park, K.S.; et al. Serum FGF21 concentration is associated with hypertriglyceridaemia, hyperinsulinaemia and pericardial fat accumulation, independently of obesity, but not with current coronary artery status. Clin. Endocrinol. 2014, 80, 57–64. [Google Scholar] [CrossRef]
  13. Kharitonenkov, A.; Shiyanova, T.L.; Koester, A.; Ford, A.M.; Micanovic, R.; Galbreath, E.J.; Sandusky, G.E.; Hammond, L.J.; Moyers, J.S.; Owens, R.A.; et al. FGF-21 as a novel metabolic regulator. J. Clin. Investig. 2005, 115, 1627–1635. [Google Scholar] [CrossRef] [Green Version]
  14. Xu, J.; Stanislaus, S.; Chinookoswong, N.; Lau, Y.Y.; Hager, T.; Patel, J.; Ge, H.; Weiszmann, J.; Lu, S.-C.; Graham, M.; et al. Acute glucose-lowering and insulin-sensitizing action of FGF21 in insulin-resistant mouse models--association with liver and adipose tissue effects. Am. J. Physiol. Endocrinol. Metab. 2009, 297, E1105–E1114. [Google Scholar] [CrossRef] [Green Version]
  15. Coskun, T.; Bina, H.A.; Schneider, M.A.; Dunbar, J.D.; Hu, C.C.; Chen, Y.; Moller, D.E.; Kharitonenkov, A. Fibroblast growth factor 21 corrects obesity in mice. Endocrinology 2008, 149, 6018–6027. [Google Scholar] [CrossRef]
  16. Ge, X.; Chen, C.; Hui, X.; Wang, Y.; Lam, K.S.L.; Xu, A. Fibroblast growth factor 21 induces glucose transporter-1 expression through activation of the serum response factor/Ets-like protein-1 in adipocytes. J. Biol. Chem. 2011, 286, 34533–34541. [Google Scholar] [CrossRef] [Green Version]
  17. Wang, N.; Li, J.-Y.; Zhao, T.-T.; Li, S.-M.; Shen, C.-B.; Li, D.-S.; Wang, W.-F. FGF-21 Plays a Crucial Role in the Glucose Uptake of Activated Monocytes. Inflammation 2018, 41, 73–80. [Google Scholar] [CrossRef]
  18. Chen, C.; Cheung, B.M.Y.; Tso, A.W.K.; Wang, Y.; Law, L.S.C.; Ong, K.L.; Wat, N.M.S.; Xu, A.; Lam, K.S.L. High plasma level of fibroblast growth factor 21 is an Independent predictor of type 2 diabetes: A 5.4-year population-based prospective study in Chinese subjects. Diabetes Care 2011, 34, 2113–2115. [Google Scholar] [CrossRef] [Green Version]
  19. Cheng, X.; Zhu, B.; Jiang, F.; Fan, H. Serum FGF-21 levels in type 2 diabetic patients. Endocr. Res. 2011, 36, 142–148. [Google Scholar] [CrossRef]
  20. Li, H.; Fang, Q.; Gao, F.; Fan, J.; Zhou, J.; Wang, X.; Zhang, H.; Pan, X.; Bao, Y.; Xiang, K.; et al. Fibroblast growth factor 21 levels are increased in nonalcoholic fatty liver disease patients and are correlated with hepatic triglyceride. J. Hepatol. 2010, 53, 934–940. [Google Scholar] [CrossRef]
  21. 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, 135–143. [Google Scholar] [CrossRef] [Green Version]
  22. Inagaki, T.; Lin, V.Y.; Goetz, R.; Mohammadi, M.; Mangelsdorf, D.J.; Kliewer, S.A. Inhibition of growth hormone signaling by the fasting-induced hormone FGF21. Cell Metab. 2008, 8, 77–83. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  23. Li, G.; Feng, D.; Qu, X.; Fu, J.; Wang, Y.; Li, L.; Li, L.; Han, L.; Esangbedo, I.C.; Li, M.; et al. Role of adipokines FGF21, leptin and adiponectin in self-concept of youths with obesity. Eur. Neuropsychopharmacol. 2018, 28, 892–902. [Google Scholar] [CrossRef]
  24. Minard, A.Y.; Tan, S.-X.; Yang, P.; Fazakerley, D.J.; Domanova, W.; Parker, B.L.; Humphrey, S.J.; Jothi, R.; Stöckli, J.; James, D.E. mTORC1 Is a Major Regulatory Node in the FGF21 Signaling Network in Adipocytes. Cell Rep. 2016, 17, 29–36. [Google Scholar] [CrossRef] [Green Version]
  25. Santoso, P.; Nakata, M.; Shiizaki, K.; Boyang, Z.; Parmila, K.; Otgon-Uul, Z.; Hashimoto, K.; Satoh, T.; Mori, M.; Kuro-O, M.; et al. Fibroblast growth factor 21, assisted by elevated glucose, activates paraventricular nucleus NUCB2/Nesfatin-1 neurons to produce satiety under fed states. Sci. Rep. 2017, 7, 45819. [Google Scholar] [CrossRef]
  26. Talukdar, S.; Owen, B.M.; Song, P.; Hernandez, G.; Zhang, Y.; Zhou, Y.; Scott, W.T.; Paratala, B.; Turner, T.; Smith, A.; et al. FGF21 Regulates Sweet and Alcohol Preference. Cell Metab. 2016, 23, 344–349. [Google Scholar] [CrossRef] [Green Version]
  27. Talukdar, S.; Zhou, Y.; Li, D.; Rossulek, M.; Dong, J.; Somayaji, V.; Weng, Y.; Clark, R.; Lanba, A.; Owen, B.M.; et al. A Long-Acting FGF21 Molecule, PF-05231023, Decreases Body Weight and Improves Lipid Profile in Non-human Primates and Type 2 Diabetic Subjects. Cell Metab. 2016, 23, 427–440. [Google Scholar] [CrossRef] [Green Version]
  28. Gaich, G.; Chien, J.Y.; Fu, H.; Glass, L.C.; Deeg, M.A.; Holland, W.L.; Kharitonenkov, A.; Bumol, T.; Schilske, H.K.; Moller, D.E. The effects of LY2405319, an FGF21 analog, in obese human subjects with type 2 diabetes. Cell Metab. 2013, 18, 333–340. [Google Scholar] [CrossRef] [Green Version]
  29. Kim, A.M.; Somayaji, V.R.; Dong, J.Q.; Rolph, T.P.; Weng, Y.; Chabot, J.R.; Gropp, K.E.; Talukdar, S.; Calle, R.A. Once-weekly administration of a long-acting fibroblast growth factor 21 analogue modulates lipids, bone turnover markers, blood pressure and body weight differently in obese people with hypertriglyceridaemia and in non-human primates. Diabetes Obes. Metab. 2017, 19, 1762–1772. [Google Scholar] [CrossRef]
  30. Baruch, A.; Wong, C.; Chinn, L.W.; Vaze, A.; Sonoda, J.; Gelzleichter, T.; Chen, S.; Lewin-Koh, N.; Morrow, L.; Dheerendra, S.; et al. Antibody-mediated activation of the FGFR1/Klothoβ complex corrects metabolic dysfunction and alters food preference in obese humans. Proc. Natl. Acad. Sci. USA 2020, 117, 28992–29000. [Google Scholar] [CrossRef]
  31. Frayling, T.M.; Beaumont, R.N.; Jones, S.E.; Yaghootkar, H.; Tuke, M.A.; Ruth, K.S.; Casanova, F.; West, B.; Locke, J.; Sharp, S.; et al. A Common Allele in FGF21 Associated with Sugar Intake Is Associated with Body Shape, Lower Total Body-Fat Percentage, and Higher Blood Pressure. Cell Rep. 2018, 23, 327–336. [Google Scholar] [CrossRef] [Green Version]
  32. Søberg, S.; Sandholt, C.H.; Jespersen, N.Z.; Toft, U.; Madsen, A.L.; von Holstein-Rathlou, S.; Grevengoed, T.J.; Christensen, K.B.; Bredie, W.L.P.; Potthoff, M.J.; et al. FGF21 Is a Sugar-Induced Hormone Associated with Sweet Intake and Preference in Humans. Cell Metab. 2017, 25, 1045–1053.e6. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  33. Chu, A.Y.; Workalemahu, T.; Paynter, N.P.; Rose, L.M.; Giulianini, F.; Tanaka, T.; Ngwa, J.S.; Qi, Q.; Curhan, G.C.; Rimm, E.B.; et al. Novel locus including FGF21 is associated with dietary macronutrient intake. Hum. Mol. Genet. 2013, 22, 1895–1902. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  34. Heianza, Y.; Ma, W.; Huang, T.; Wang, T.; Zheng, Y.; Smith, S.R.; Bray, G.A.; Sacks, F.M.; Qi, L. Macronutrient Intake-Associated FGF21 Genotype Modifies Effects of Weight-Loss Diets on 2-Year Changes of Central Adiposity and Body Composition: The POUNDS Lost Trial. Diabetes Care 2016, 39, 1909–1914. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  35. Tanaka, T.; Ngwa, J.S.; van Rooij, F.J.A.; Zillikens, M.C.; Wojczynski, M.K.; Frazier-Wood, A.C.; Houston, D.K.; Kanoni, S.; Lemaitre, R.N.; Luan, J.; et al. Genome-wide meta-analysis of observational studies shows common genetic variants associated with macronutrient intake. Am. J. Clin. Nutr. 2013, 97, 1395–1402. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  36. Tönjes, A.; Kralisch, S.; Lössner, U.; Kovacs, P.; Blüher, M.; Stumvoll, M.; Fasshauer, M. Metabolic and genetic predictors of circulating adipocyte fatty acid-binding protein. Int. J. Obes. 2012, 36, 766–773. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  37. Tönjes, A.; Koriath, M.; Schleinitz, D.; Dietrich, K.; Böttcher, Y.; Rayner, N.W.; Almgren, P.; Enigk, B.; Richter, O.; Rohm, S.; et al. Genetic variation in GPR133 is associated with height: Genome wide association study in the self-contained population of Sorbs. Hum. Mol. Genet. 2009, 18, 4662–4668. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  38. Tönjes, A.; Zeggini, E.; Kovacs, P.; Böttcher, Y.; Schleinitz, D.; Dietrich, K.; Morris, A.P.; Enigk, B.; Rayner, N.W.; Koriath, M.; et al. Association of FTO variants with BMI and fat mass in the self-contained population of Sorbs in Germany. Eur. J. Hum. Genet. 2010, 18, 104–110. [Google Scholar] [CrossRef] [Green Version]
  39. Matthews, D.R.; Hosker, J.P.; Rudenski, A.S.; Naylor, B.A.; Treacher, D.F.; Turner, R.C. Homeostasis model assessment: Insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985, 28, 412–419. [Google Scholar] [CrossRef] [Green Version]
  40. Ebert, T.; Gebhardt, C.; Scholz, M.; Wohland, T.; Schleinitz, D.; Fasshauer, M.; Blüher, M.; Stumvoll, M.; Kovacs, P.; Tönjes, A. Relationship Between 12 Adipocytokines and Distinct Components of the Metabolic Syndrome. J. Clin. Endocrinol. Metab. 2018, 103, 1015–1023. [Google Scholar] [CrossRef]
  41. Tönjes, A.; Scholz, M.; Breitfeld, J.; Marzi, C.; Grallert, H.; Gross, A.; Ladenvall, C.; Schleinitz, D.; Krause, K.; Kirsten, H.; et al. Genome wide meta-analysis highlights the role of genetic variation in RARRES2 in the regulation of circulating serum chemerin. PLoS Genet. 2014, 10, e1004854. [Google Scholar] [CrossRef]
  42. Loffler, A.; Luck, T.; Then, F.S.; Sikorski, C.; Kovacs, P.; Bottcher, Y.; Breitfeld, J.; Tonjes, A.; Horstmann, A.; Loffler, M.; et al. Eating Behaviour in the General Population: An Analysis of the Factor Structure of the German Version of the Three-Factor-Eating-Questionnaire (TFEQ) and Its Association with the Body Mass Index. PLoS ONE 2015, 10, e0133977. [Google Scholar] [CrossRef]
  43. Pudel, V.; Westhöfer, J. Fragebogen zum Eßverhalten (FEV); Verlag für Psychologie Dr. C.J. Hogrefe: Göttingen, Germany; Toronto, ON, Canada; Zürich, Switzerland, 1989. [Google Scholar]
  44. Löffler, A.; Luck, T.; Then, F.S.; Luck-Sikorski, C.; Pabst, A.; Kovacs, P.; Böttcher, Y.; Breitfeld, J.; Tönjes, A.; Horstmann, A.; et al. Effects of psychological eating behaviour domains on the association between socio-economic status and BMI. Public Health Nutr. 2017, 20, 2706–2712. [Google Scholar] [CrossRef] [Green Version]
  45. Stunkard, A.J.; Messick, S. The three-factor eating questionnaire to measure dietary restraint, disinhibition and hunger. J. Psychosom. Res. 1985, 29, 71–83. [Google Scholar] [CrossRef]
  46. Verger, E.O.; Armstrong, P.; Nielsen, T.; Chakaroun, R.; Aron-Wisnewsky, J.; Gøbel, R.J.; Schütz, T.; Delaere, F.; Gausseres, N.; Clément, K.; et al. Dietary Assessment in the MetaCardis Study: Development and Relative Validity of an Online Food Frequency Questionnaire. J. Acad. Nutr. Diet. 2017, 117, 878–888. [Google Scholar] [CrossRef]
  47. Purcell, S.; Neale, B.; Todd-Brown, K.; Thomas, L.; Ferreira, M.A.R.; Bender, D.; Maller, J.; Sklar, P.; de Bakker, P.I.W.; Daly, M.J.; et al. PLINK: A tool set for whole-genome association and population-based linkage analyses. Am. J. Hum. Genet. 2007, 81, 559–575. [Google Scholar] [CrossRef] [Green Version]
  48. Auton, A.; Brooks, L.D.; Durbin, R.M.; Garrison, E.P.; Kang, H.M.; Korbel, J.O.; Marchini, J.L.; McCarthy, S.; McVean, G.A.; Abecasis, G.R. A global reference for human genetic variation. Nature 2015, 526, 68–74. [Google Scholar] [CrossRef] [Green Version]
  49. Willer, C.J.; Schmidt, E.M.; Sengupta, S.; Peloso, G.M.; Gustafsson, S.; Kanoni, S.; Ganna, A.; Chen, J.; Buchkovich, M.L.; Mora, S.; et al. Discovery and refinement of loci associated with lipid levels. Nat. Genet. 2013, 45, 1274–1283. [Google Scholar] [CrossRef] [Green Version]
  50. Chambers, J.C.; Zhang, W.; Sehmi, J.; Li, X.; Wass, M.N.; van der Harst, P.; Holm, H.; Sanna, S.; Kavousi, M.; Baumeister, S.E.; et al. Genome-wide association study identifies loci influencing concentrations of liver enzymes in plasma. Nat. Genet. 2011, 43, 1131–1138. [Google Scholar] [CrossRef]
  51. Kharitonenkov, A.; Wroblewski, V.J.; Koester, A.; Chen, Y.-F.; Clutinger, C.K.; Tigno, X.T.; Hansen, B.C.; Shanafelt, A.B.; Etgen, G.J. The metabolic state of diabetic monkeys is regulated by fibroblast growth factor-21. Endocrinology 2007, 148, 774–781. [Google Scholar] [CrossRef] [Green Version]
  52. Fisher, F.M.; Chui, P.C.; Antonellis, P.J.; Bina, H.A.; Kharitonenkov, A.; Flier, J.S.; Maratos-Flier, E. Obesity is a fibroblast growth factor 21 (FGF21)-resistant state. Diabetes 2010, 59, 2781–2789. [Google Scholar] [CrossRef] [Green Version]
  53. Schumann, G.; Liu, C.; O’Reilly, P.; Gao, H.; Song, P.; Xu, B.; Ruggeri, B.; Amin, N.; Jia, T.; Preis, S.; et al. KLB is associated with alcohol drinking, and its gene product β-Klotho is necessary for FGF21 regulation of alcohol preference. Proc. Natl. Acad. Sci. USA 2016, 113, 14372–14377. [Google Scholar] [CrossRef] [Green Version]
  54. Desai, B.N.; Singhal, G.; Watanabe, M.; Stevanovic, D.; Lundasen, T.; Fisher, F.M.; Mather, M.L.; Vardeh, H.G.; Douris, N.; Adams, A.C.; et al. Fibroblast growth factor 21 (FGF21) is robustly induced by ethanol and has a protective role in ethanol associated liver injury. Mol. Metab. 2017, 6, 1395–1406. [Google Scholar] [CrossRef]
  55. Kamizono, Y.; Shiga, Y.; Suematsu, Y.; Imaizumi, S.; Tsukahara, H.; Noda, K.; Kuwano, T.; Fujimi, K.; Saku, K.; Miura, S.-I. Impact of cigarette smoking cessation on plasma α-klotho levels. Medicine 2018, 97, e11947. [Google Scholar] [CrossRef]
  56. Lin, Z.; Tian, H.; Lam, K.S.L.; Lin, S.; Hoo, R.C.L.; Konishi, M.; Itoh, N.; Wang, Y.; Bornstein, S.R.; Xu, A.; et al. Adiponectin mediates the metabolic effects of FGF21 on glucose homeostasis and insulin sensitivity in mice. Cell Metab. 2013, 17, 779–789. [Google Scholar] [CrossRef] [Green Version]
  57. Fu, Y.; Luo, N.; Klein, R.L.; Garvey, W.T. Adiponectin promotes adipocyte differentiation, insulin sensitivity, and lipid accumulation. J. Lipid Res. 2005, 46, 1369–1379. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  58. Spranger, J.; Kroke, A.; Möhlig, M.; Bergmann, M.M.; Ristow, M.; Boeing, H.; Pfeiffer, A.F.H. Adiponectin and protection against type 2 diabetes mellitus. Lancet 2003, 361, 226–228. [Google Scholar] [CrossRef]
  59. Wannamethee, S.G.; Tchernova, J.; Whincup, P.; Lowe, G.D.; Rumley, A.; Brown, K.; Cherry, L.; Sattar, N. Associations of adiponectin with metabolic and vascular risk parameters in the British Regional Heart Study reveal stronger links to insulin resistance-related than to coronory heart disease risk-related parameters. Int. J. Obes. 2007, 31, 1089–1098. [Google Scholar] [CrossRef] [Green Version]
  60. Ryo, M.; Nakamura, T.; Kihara, S.; Kumada, M.; Shibazaki, S.; Takahashi, M.; Nagai, M.; Matsuzawa, Y.; Funahashi, T. Adiponectin as a biomarker of the metabolic syndrome. Circ. J. 2004, 68, 975–981. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  61. Clarke, T.-K.; Adams, M.J.; Davies, G.; Howard, D.M.; Hall, L.S.; Padmanabhan, S.; Murray, A.D.; Smith, B.H.; Campbell, A.; Hayward, C.; et al. Genome-wide association study of alcohol consumption and genetic overlap with other health-related traits in UK Biobank (N=112 117). Mol. Psychiatry 2017, 22, 1376–1384. [Google Scholar] [CrossRef] [Green Version]
  62. Fehrmann, R.S.N.; Jansen, R.C.; Veldink, J.H.; Westra, H.-J.; Arends, D.; Bonder, M.J.; Fu, J.; Deelen, P.; Groen, H.J.M.; Smolonska, A.; et al. Trans-eQTLs reveal that independent genetic variants associated with a complex phenotype converge on intermediate genes, with a major role for the HLA. PLoS Genet. 2011, 7, e1002197. [Google Scholar] [CrossRef] [Green Version]
  63. Ding, R.; Huang, T.; Han, J. Diet/lifestyle and risk of diabetes and glycemic traits: A Mendelian randomization study. Lipids Health Dis. 2018, 17, 18. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  64. Evangelou, E.; Gao, H.; Chu, C.; Ntritsos, G.; Blakeley, P.; Butts, A.R.; Pazoki, R.; Suzuki, H.; Koskeridis, F.; Yiorkas, A.M.; et al. New alcohol-related genes suggest shared genetic mechanisms with neuropsychiatric disorders. Nat. Hum. Behav. 2019, 3, 950–961. [Google Scholar] [CrossRef] [PubMed] [Green Version]
Table
Table 1. Baseline characteristics of the Sorbs cohort.
Table 1. Baseline characteristics of the Sorbs cohort.
TotalMales
(Above the Limit of Detection)		Females
(Above the Limit of Detection)		p
(Males vs. Females)
n 1046 (419 m/620 f)330450
Anthropometric traits
Age (years)44 (25)50 (23)51 (24)0.313
BMI (kg/m2)26.0 (5.6)27.0 (4.6)26.9 (7.8)0.287
WHR0.88 (0.17)0.96 (0.12)0.83 (0.11)2.599 × 10−60
Height (cm)172 (13)177 (10)164 (9)7.103 × 10−87
Bodyweight (kg)75.0 (20.0)84.5 (16.0)71.0 (17.0)3.004 × 10−36
SBP (mmHg)134.25 (25.00)142.50 (21.00)133.25 (25.00)1.801 × 10−11
Glucose/insulin traits
T2D (yes/no)114/92141/28661/3890.678
FG (mmol/L)5.29 (0.76)5.54 (0.69)5.28 (0.76)3.608 × 10−10
FI (pmol/L)33.90 (28.10)33.75 (28.60)37.20 (30.70)0.038
HbA1c (%)5.3 (0.5)5.4 (0.6)5.4 (0.6)0.989
HOMA-IR1.13 (1.41)1.45 (1.38)1.49 (1.48)0.561
Lipid traits
LLM (yes/no)120/91946/28456/3940.541
TG (mmol/L)1.06 (0.75)1.35 (0.96)1.02 (0.68)4.830 × 10−11
LDL-C (mmol/L)3.34 (1.26)3.57 (1.22)3.32 (1.27)0.002
HDL-C (mmol/L)1.59 (0.58)1.42 (0.45)1.7 (0.56)3.532 × 10−21
Liver metabolism
γGT (µlat/L)0.38 (0.44)0.50 (0.54)0.26 (0.20)9.515 × 10−36
ALAT (µkat/L)0.41 (0.27)0.49 (0.28)0.32 (0.17)4.911 × 10−40
ASAT (µkat/L)0.45 (0.16)0.50 (0.17)0.41 (0.12)7.342 × 10−29
AP (µkat/L)0.99 (0.41)1.13 (0.36)1.03 (0.47)1.400 × 10−5
Adipokines
FGF21 (ng/L)84.48 (162.88)101.31 (147.90)92.39 (113.66)0.023
Irisin (µg/mL)0.76 (0.39)0.76 (0.36)0.81 (0.38)0.018
Adiponectin(µg/mL)14.79 (8.12)13.68 (6.21)17.40 (6.99)2.730 × 10−17
AFABP4 (µg/L)12.64 (15.38)13.76 (11.57)24.53 (22.36)1.663 × 10−25
Chemerin (ng/mL)113.38 (44.70)118.12 (47.00)124.69 (55.60)0.001
Vaspin (ng/mL)0.52 (1.11)0.36 (0.41)0.59 (0.97)4.501 × 10−18
Progranulin (ng/mL)106.50 (37.47)109.55 (36.36)110.15 (35.03)0.430
AGF (µg/L)37.78 (35.36)39.69 (38.48)37.78 (40.10)0.760
IGF-1 (ng/mL)169.20 (77.45)156.05 (74.85)150.05 (76.90)0.023
pro-NT (pmol/L)113.97 (62.54)109.47 (56.13)116.53 (61.91)0.113
PENK (pmol/L)57.75 (20.12)55.10 (18.69)57.79 (22.23)0.010
Luxury food
Smoker(yes/no)351/688180/15092/3585.436 × 10−23
cigarette (packs/year)5.00 (8.50)8.13 (17.00)2.35 (4.18)8.838 × 10−14
Score alcohol3 (3)4 (2)2 (1)2.690 × 10−49
Score coffee1 (1)2 (1)2 (1)0.911
Eating behavior
Score disinhibition4 (4)3 (3)4 (5)0.003
Score hunger3 (4)3 (4)4 (4)0.014
Score restraint7.0 (7.5)6.0 (5.5)9.0 (8.0)1.871 × 10−9
Data were shown as median (interquartile range) or total number; TG, LDL-C, HDL-C were shown for subjects without lipid lowering medication. Assessment of group-differences were calculated by Mann-Whitney-U test (continuous parameters) or Chi-square-test (discrete variables: T2D and smokers). p-values presented comparison of males vs. females including only subjects with detectable FGF21 serum levels. A p-value 0.05 was considered to be significant and was highlighted in bold. n = number of subjects; m = males; f = females; p = p-value; BMI = body-mass index; WHR = waist-to-hip ratio; SBP = systolic blood pressure; T2D = type 2 diabetes; FG = fasting glucose; FI = fasting insulin; HbA1c = hemoglobin A1c, long term glucose level; HOMA-IR = homeostasis assessment model of insulin resistance; LLM = lipid lowering medication; TG = triglyceride; LDL-C = low density lipoprotein cholesterol; HDL-C = high density lipoprotein cholesterol; γGT = gamma-glutamyltransferase; ALAT = alanine aminotransferase; ASAT = aspartate aminotransferase; AP = alkaline phosphatase; FGF21 = fibroblast growth factor-21; AFABP = adipocyte fatty acid binding protein; AGF = angiopoietin-related growth factor; IGF-1 = insulin like growth factor 1; pro-NT = pro-neurotensin; PENK = pro-enkephalin.
Table
Table 2. Baseline characteristics of the validation cohort.
Table 2. Baseline characteristics of the validation cohort.
TotalMalesFemalesp
n 704387317
Age (years)54 (20)52 (22)57 (16)0.004
BMI (kg/m2)37.8 (13.14)37.8 (15.86)37.4 (9.15)0.155
FGF21 (ng/L)198.84 (207.29)184.73 (196.47)212.98 (224.69)0.005
T2D (yes/no)341/371156/228180/1372.00 × 10−5
Smoker (yes/no)114/51862/29551/2180.608
Continuous normally distributed data are shown as mean with standard deviation. FGF21 is shown as median with interquartile rage; for T2D and smokers, total numbers are shown. Group differences were calculated by Mann-Whitney-U test (continuous parameters) or Chi-square-test (discrete variables: T2D and smokers). p-values presented comparison of males vs. females. A p-value 0.05 was considered to be significant and was highlighted in bold. n = number of subjects; p = p-value; BMI = body-mass index; T2D = type 2 diabetes; FGF21 = fibroblast growth factor-21.
Table
Table 3. LD-groups of the investigated FGF21 SNPs.
Table 3. LD-groups of the investigated FGF21 SNPs.
LD-GroupSNPA1A2MAF
1rs4015TC0.110
rs12611028TC0.095
2rs838133GA0.493
rs838144TC0.483
3rs4021GA0.245
4rs12975033TA0.401
rs8104897AC0.398
rs8105137AG0.398
rs8106205CT0.414
rs11667321GA0.414
5rs12975781TC0.341
6rs3826821AG0.065
7rs838143AG0.081
8rs838136CT0.285
9rs499765GC0.331
The representative SNP of each LD-group is indicated in bold letters. A1 presents the minor allele and A2 the major allele. LD-plot 9 does not pass the Hardy-Weinberg-Equilibrium with p 0.0001 and was therefore excluded from further analysis. LD = linkage disequilibrium; FGF21 = fibroblast growth factor 21; SNP = single nucleotide polymorphism; MAF = minor allele frequency.
Table
Table 4. Association of FGF21 with eating behavior and food and drug craving in the Sorbs.
Table 4. Association of FGF21 with eating behavior and food and drug craving in the Sorbs.
Dependent VariablesFGF21
(Unadjusted)		FGF21
(Adjusted for Age, Sex, BMI)
Linear regression
βpβp n
Disinhibition−0.1520.001−0.1582.170 × 10−4465
Hunger−0.0610.184−0.0520.260465
Restraint0.0660.1530.0450.318465
Packyears0.2396.800 × 10−50.1110.030268
Generalized linear model
ORpORp n
Alcohol consumption1.3151.300 × 10−51.2811.510 × 10−4768
Coffee consumption0.8720.0290.8000.001761
Binary logistic regression
ORpORp n
Smoker (yes/no)1.3417.200 × 10−51.3946.700 × 10−5780
Prior to statistical analysis the variables packyears, serum FGF21, BMI and age were ln-transformed. Assessments were done using linear regression for eating behavior and packyears, generalized linear model for alcohol and coffee consumption and binary logistic regression for smoking status. A p-value 0.05 was considered to be significant and was highlighted in bold. FGF21 = fibroblast growth factor 21; BMI = body-mass-index; OR = odds ratio; β = regression coefficient; p = p-value; n = number of subjects.
Table
Table 5. FGF21 and alcohol consumption and smoking in the validation cohort.
Table 5. FGF21 and alcohol consumption and smoking in the validation cohort.
Dependent
Variables		FGF21
(Unadjusted)		FGF21
(Adjusted for Age, Sex, BMI)
Linear regression
βpβp n
Alcohol
consumption		0.1623.00 × 10−40.1641.00 × 10−3492
Binary logistic regression
ORpORp n
Smoker (yes/no)1.3220.0111.6664.10 × 10−5614
Prior to analysis, the variables serum FGF21, BMI and age were ln-transformed. Alcohol consumption was operationalized by the scale ‘food alcoholic’. Assessments were done using linear regression for alcohol consumption and binary logistic regression for smoking status. A p-value 0.05 was considered to be significant and was highlighted in bold. FGF21 = fibroblast growth factor 21, BMI = body-mass-index; OR = odds ratio; β = regression coefficient; p = p-value; n = number of subjects.
Table
Table 6. Correlation analyses of adipokines and serum FGF21 in the Sorbs cohort.
Table 6. Correlation analyses of adipokines and serum FGF21 in the Sorbs cohort.
Adipokine n Spearman’s Rank CorrelationPartial Correlation
(Adjusted for Age, Sex, BMI)
Spearman’s Rank
Coefficient		pCorrelation
Coefficient		p
Irisin777−0.0480.178−0.0600.150
Adiponectin725−0.0920.013−0.0680.105
AFABP7800.2711.393×10−140.2482.444×10−10
Chemerin7790.2463.294×10−120.2005.586×10−9
Vaspin7790.0290.4240.0670.109
Progranulin7800.1230.0010.0980.019
AGF6280.0270.498−0.0420.322
IGF-1780−0.2564.195×10−13−0.1953.000×10−6
pro-NT7770.0200.5840.0210.614
PENK7770.0210.5590.0920.028
Prior statistical analysis, continuous variables were ln-transformed. Assessments were done using Spearman’s rank correlation and partial correlation with adjustments for sex, age and BMI. A p-value 0.05 was considered to be significant and was highlighted in bold. n = number of subjects; p = p-value; BMI = body-mass-index; FGF21 = fibroblast growth factor 21; AFABP = adipocyte fatty-acid-binding protein; AGF = angiopoietin-related growth factor; IGF-1 = insulin-like growth factor 1; pro-NT = pro-neurotensin; PENK = pro-enkephalin.
Table
Table 7. Association of representative SNPs for each LD-group with glucose and lipid metabolism, and liver parameters in the Sorbs cohort.
Table 7. Association of representative SNPs for each LD-group with glucose and lipid metabolism, and liver parameters in the Sorbs cohort.
LD 1LD 2LD 3LD 4LD 5LD 6LD 7LD 8 n
Fasting
glucose		β−0.0180.005−0.0030.0030.0040.0160.0120.002
s.e.0.0130.0080.0100.0080.0080.0160.0150.009946
p0.1830.4860.7770.7460.6240.3150.4240.857
Fasting
insulin		β0.1280.0300.0490.0180.0300.1110.0440.026
s.e.0.0420.0250.0300.0250.0270.0520.0460.029946
p0.0020.2140.1040.4680.2690.0330.3490.371
HOMA-IRβ0.1090.0470.0440.0300.0360.1350.0660.046
s.e.0.0460.0270.0330.02740.0290.0560.0510.032935
p0.0190.0790.1800.2740.2150.0170.1960.155
HbA1cβ−0.0050.0010.0010.0010.0010.0020.0010.008
s.e.0.0070.0040.0050.0040.0040.0090.0080.005953
p0.4720.8840.8340.8190.8040.8450.8590.086
LDL-Cβ−0.031−0.034−0.021−0.044−0.0320.0160.021−0.041
s.e.0.0220.0120.0150.0120.0130.0250.0230.015845
P0.1500.0050.1654.482 × 10−40.0180.5380.3520.005
HDL-Cβ0.010−0.006−0.001−0.017−0.016−0.0280.0210.010
s.e.0.0170.0100.0120.0100.0110.0210.0190.012845
p0.5780.5110.9620.1000.1290.1780.2630.417
TGβ0.043−0.0310.017−0.030−0.0180.0720.008−0.012
s.e.0.0380.0210.0260.0220.0230.0450.0400.026845
p0.2600.1420.5170.1690.4270.1070.8330.642
γGTβ0.052-0.035−0.027−0.001−0.0070.072−0.113−0.043
s.e.0.0480.0280.0350.0290.0300.0590.0530.034958
p0.2870.2050.4290.9650.8280.2240.0340.198
APβ0.0140.0220.0090.0200.0140.0200.0030.018
s.e.0.0210.0120.0150.0120.0130.0260.0230.015958
p0.5210.0770.5300.1150.3040.4390.9120.210
ALATβ0.006−0.026−0.017−0.025−0.0230.015−0.012−0.014
s.e.0.0300.0170.0210.0170.0190.0360.0330.021958
p0.8500.1330.4250.1450.2170.6850.7240.508
ASATβ0.020−0.020−0.006−0.022−0.0130.007−0.003−0.016
s.e.0.0190.0110.0130.0110.0120.0230.0210.013958
p0.2810.0680.6610.0470.2630.7540.8650.231
The presented associations were given for the representative SNP within each LD-group. Prior to analysis, all variables were ln-transformed except the covariate sex. p-values were adjusted for: sex, age, BMI and a p-value 0.05 was considered to be significant and was highlighted in bold. Linear regression including all subjects was conducted. For lipid parameters only subjects without lipid lowering medication were analyzed. LD = linkage disequilibrium; HOMA-IR = homeostasis assessment model of insulin resistance; HbA1c = hemoglobin A1c; LDL-C = low density lipoprotein cholesterol; HDL-C = high density lipoprotein cholesterol; TG = triglyceride; γGT = gamma-glutamyltransferase; AP = alkaline phosphatase; ALAT = alanine aminotransferase; ASAT = aspartate aminotransferase; BMI = body-mass index; p = p-value; β = regression coefficient; s.e. = standard error; n = number of subjects.
Table
Table 8. Association of representative SNPs for each LD-group with FGF21-parameters and anthropometric parameters in the Sorbs cohort.
Table 8. Association of representative SNPs for each LD-group with FGF21-parameters and anthropometric parameters in the Sorbs cohort.
LD 1LD 2LD 3LD 4LD 5LD 6LD 7LD 8n
FGF21-
serum level		β 0.2270.0680.0480.0810.076−0.1550.0010.079
s.e.0.0880.0510.0630.0520.0560.1080.0980.061724
p0.0100.1800.4530.1210.1730.1500.9890.195
FGF21-
expression		β −0.0011.261 × 10−42.669 × 10−4−0.001−0.001−2.052 × 10−40.0030.001
s.e.0.0020.0010.0010.0010.0010.0020.0020.001938
p0.4240.9000.8300.3130.3570.9230.1650.571
BMIβ −0.010−0.006−0.007−0.005−0.0050.004−0.002−0.001
s.e.0.0120.0070.0080.0070.0070.0140.0130.008958
p0.4300.3580.3900.5170.5230.7850.9010.940
WHRβ 0.0053.225 × 10−40.0040.0010.0024.664 × 10−44.915 × 10−40.003
s.e.0.0050.0030.0040.0030.0030.0060.0060.003957
p0.2830.9120.3150.7240.6220.9400.9290.403
Heightβ −0.002−0.002−0.0040.0010.0010.004−0.009−0.003
s.e.0.0030.0020.0020.0020.0020.0030.0030.002958
p0.5150.2760.0310.7260.4180.2280.0030.069
Weightβ−0.005−0.004−0.0098.199 × 10−50.0020.008−0.018−0.007
s.e0.0050.0030.0040.0030.0030.0070.0060.004958
p0.3230.1600.0220.9800.6470.2260.0030.051
SBPβ −0.002−0.002−0.008−0.0570.0040.020−0.009−4.078 × 10−4
s.e.0.0080.0050.0060.0050.0050.0100.0090.006958
p0.8500.6590.1930.9910.4750.0540.3220.944
The presented associations were given for the representative SNP within each LD-group. Prior analysis, all variables were ln-transformed except the covariate sex. p-values were adjusted for: sex, age, BMI and p-values 0.05 were considered to be significant and highlighted in bold. Association analysis was done using linear regression analysis. LD = linkage disequilibrium; FGF = fibroblast growth factor; BMI = body-mass index; WHR = waist-to-hip ratio; SBP = systolic blood pressure; p = p-value; β = regression coefficient; s.e. = standard error; n = number of subjects.
Table
Table 9. Association of representative SNPs for each LD-group with adipokines in the Sorbs cohort.
Table 9. Association of representative SNPs for each LD-group with adipokines in the Sorbs cohort.
LD 1LD 2LD 3LD 4LD 5LD 6LD 7LD 8 n
IGF-1β−0.014−0.0170.001−0.029−0.0280.0200.0380.004
s.e.0.0230.0130.0170.0140.0150.0290.0260.016958
p0.5620.2080.9550.0340.0570.4910.1440.783
Chemerinβ−0.001−0.014−0.017−0.014−0.0150.010−0.017−0.021
s.e.0.0240.0140.0170.0140.0150.0300.0270.017929
p0.9610.3060.3120.3370.3320.7480.5280.203
Progranulinβ0.0270.0070.0130.0080.0070.0190.0050.006
s.e.0.0190.0110.0140.0110.0120.0230.0210.013957
p0.1560.5130.3360.4720.5410.4240.8010.668
AFABPβ0.0880.0270.0260.0190.0130.0200.0100.029
s.e.0.0400.0230.0290.0240.0250.0500.0450.028930
p0.0280.2450.3640.4160.6120.6900.8270.296
Adiponectinβ0.057−0.0110.014−0.014−0.0080.0140.008−0.004
s.e.0.0260.0150.0190.0150.0160.0320.0290.018928
p0.0310.4820.4580.3430.6370.6600.7860.807
Irisinβ−0.001−0.002−0.0170.002−0.0080.002−0.023−0.004
s.e.0.0250.0140.0180.0150.0160.0310.0270.017954
p0.9840.8780.3360.8690.6240.9390.4010.810
Vaspinβ0.046−0.0300.031−0.044−0.0300.1290.0360.029
s.e.0.0780.0450.0550.0460.0490.0950.0850.054952
p0.5620.5000.5770.3330.5360.1730.6740.586
AGFβ0.0530.065−0,0580.0660.0210.1540.055−0.008
s.e.0.1010.0570.0720.0590.0630.1250.1100.069768
p0.6000.2560.9990.2630.7370.2180.6160.904
pro-NTβ−0.030−0.005−0.044−0.003−0.011−0.030−0.023−0.016
s.e.0.0310.0180.0220.0180.0190.0370.0340.021953
p0.3230.7580.0450.8690.5620.4380.4950.449
PENKβ0.0150.0010.003−0.0060.0040.0660.0290.014
s.e.0.0210.0120.0150.0120.0130.0250.0230.014953
p0.4800.9580.8270.6040.7700.0090.1950.336
The presented associations are given for the representative SNP within each LD-group. Prior to analysis, all variables were ln-transformed except the covariate sex. p-values were adjusted for: sex, age, BMI and a p-value 0.05 was considered to be significant and highlighted in bold. Linear regression including all subjects was conducted. LD = linkage disequilibrium; IGF-1 = insulin like growth factor 1; AFABP = adipocyte fatty acid-binding protein; AGF = angiopoietin-related growth factor; pro-NT = pro-neurotensin; PENK = pro-enkephalin; BMI = body-mass index; p = p-value; β = regression coefficient; s.e. = standard error; n = number of subjects.
Table
Table 10. Association of representative SNPs for each LD-group with food and drug craving in the Sorbs cohort.
Table 10. Association of representative SNPs for each LD-group with food and drug craving in the Sorbs cohort.
LD 1LD 2LD 3LD 4LD 5LD 6LD 7LD 8 n
Disinhibitionβ−0.014−0.2950.124−0.228−0.2420.034−0.1480.014
s.e.0.2590.1520.1900.1540.1670.3140.2870.186572
p0.9570.0530.5140.1410.1480.9150.6080.942
Hungerβ−0.070−0.241−0.065−0.251−0.3880.154−0.027−0.118
s.e.0.2630.1550.1930.1570.1700.3190.2920.190572
p0.7900.1200.7390.1100.0230.6290.9270.533
Restraintβ−0.554−0.0890.017−0.256−0.1760.1760.544−0.178
s.e.0.4470.2640.3290.2670.2900.5430.4970.322572
p0.2160.7360.9580.3390.5440.7460.2740.580
Smoking yes/noOR1.1881.1051.2921.1001.0241.0541.0581.299
s.e.0.1730.1000.1230.1030.1100.2100.1900.121958
p0.3210.3200.0380.3640.8270.8030.7670.030
Cigarettes
(Packyears)		β0.0210.1400.0700.1450.072−0.2040.0510.073
s.e.0.1460.0860.0980.0880.0950.1820.1510.098323
p0.8830.1050.4760.1020.4500.2620.7340.455
CoffeeOR0.9041.0570.9731.0451.0310.9530.9791.030
s.e.0.1600.0920.1140.0940.1010.1940.1750.111931
p0.5300.5510.8080.6380.7630.8040.9030.788
AlcoholOR1.4341.0101.1501.0511.0271.2040.9730.982
s.e.0.1980.1140.1440.1170.1260.2380.2300.142952
p0.0690.9280.3320.6710.8330.4360.9060.900
The presented associations are given for the representative SNP within each LD-group. Prior analysis, all variables were ln-transformed except smoking, alcohol and coffee consumption, scores of eating behavior and sex. p-values were adjusted for: sex, age, BMI and a p-value 0.05 was considered to be significant and highlighted in bold. Linear regression including all subjects was conducted. The variables alcohol, coffee and smoking have been analyzed using ordinal regression and including all subjects. LD = linkage disequilibrium; OR = odds ratio; p = p-value; β = regression coefficient; s.e. = standard error; n = number of subjects.
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Epperlein, S.; Gebhardt, C.; Rohde, K.; Chakaroun, R.; Patt, M.; Schamarek, I.; Kralisch, S.; Heiker, J.T.; Scholz, M.; Stumvoll, M.; et al. The Effect of FGF21 and Its Genetic Variants on Food and Drug Cravings, Adipokines and Metabolic Traits. Biomedicines 2021, 9, 345. https://doi.org/10.3390/biomedicines9040345

AMA Style

Epperlein S, Gebhardt C, Rohde K, Chakaroun R, Patt M, Schamarek I, Kralisch S, Heiker JT, Scholz M, Stumvoll M, et al. The Effect of FGF21 and Its Genetic Variants on Food and Drug Cravings, Adipokines and Metabolic Traits. Biomedicines. 2021; 9(4):345. https://doi.org/10.3390/biomedicines9040345

Chicago/Turabian Style

Epperlein, Sarah, Claudia Gebhardt, Kerstin Rohde, Rima Chakaroun, Marie Patt, Imke Schamarek, Susan Kralisch, John T. Heiker, Markus Scholz, Michael Stumvoll, and et al. 2021. "The Effect of FGF21 and Its Genetic Variants on Food and Drug Cravings, Adipokines and Metabolic Traits" Biomedicines 9, no. 4: 345. https://doi.org/10.3390/biomedicines9040345

APA Style

Epperlein, S., Gebhardt, C., Rohde, K., Chakaroun, R., Patt, M., Schamarek, I., Kralisch, S., Heiker, J. T., Scholz, M., Stumvoll, M., Kovacs, P., Breitfeld, J., & Tönjes, A. (2021). The Effect of FGF21 and Its Genetic Variants on Food and Drug Cravings, Adipokines and Metabolic Traits. Biomedicines, 9(4), 345. https://doi.org/10.3390/biomedicines9040345

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop