Leptin as a Potential Regulator of FGF21

Fibroblast growth factor 21 (FGF21) is a potent metabolic regulator, predominantly expressed in the liver and white adipose tissue, but also in skeletal muscle and pancreas [1,2]. Pharmacological doses of FGF21 have been shown to exert notable anti-diabetic effects. In particular, FGF21 improves glucose tolerance and insulin sensitivity, and reduces plasma and hepatic triglycerides in ob/ob, db/db, and wild-type mice fed a high-fat diet [3,4,5,6]. It also protects against high glucose induced cellular damage in endothelial cells, thus potentially promoting vascular health in diabetes [7]. Several studies focused on the identification of potential FGF21 mediators [8,9,10]. Among these mediators, adiponectin was shown to mediate part of FGF21 effects [8,9]. Recent evidence also suggests that other factors, such as dual specificity phosphatase 4 or uncoupling protein-1 (UCP1) may mediate FGF21 actions [11,12]. With regard to the regulation of FGF21 expression, very limited information is available as yet [13]. Interestingly, FGF21 expression was shown to be under the control of the stress responsive transcription factor, Signal Transducers and Activators of Transcription 3 (STAT3) [14], which is also known to mediate leptin effects [15]. Subsequently, other studies demonstrated that FGF21 is modulated by key metabolic transcription factors such as PPARα and PPARγ [16,17,18] which are also induced by leptin [19]. Leptin is also known to promote metabolic health by notably increasing fat oxidation in muscle cells [20] and decreasing inflammation [21]. In the present study, we first investigated whether leptin may regulate FGF21 expression in vivo in Wistar rats. Since FGF21 is predominantly produced by the liver [22], potential leptin effects on FGF21 were also investigated in vitro, in human derived hepatocarcinoma HepG2 cells, a cell line known to produce FGF21 although at very low concentrations under basal conditions [23]. This allowed us to investigate the potential signaling pathways involved in the regulation of FGF21 by leptin.

All procedures were performed in accordance with and approved by the Institutional Ethical Committee of Animal Care in Geneva and Cantonal Veterinary Office. Four month-old male Wistar rats were obtained from Charles River (L'Arbresle, France). They were housed under controlled temperature (22°C) and lighting (lights on: 7 AM to 7 PM) with free access to water and food (standard diet, laboratory diet RM3, SDS, Essex, UK). To test the effect of leptin on FGF21 secretion, a cohort of rats was euthanized 30 minutes after a single intraperitoneal injection of human recombinant leptin (2 mg/kg) (PeproTech, Rocky Hill, NJ) or vehicle (saline). This leptin dose was chosen because previous experiments were performed in similar conditions with reliable results. Notably, this dose of leptin was able to decrease food intake in Wistar rats, confirming a response to leptin, which was not the case with lower doses of leptin [24]. Blood and tissues were sampled and stored at -80°C for further analysis.

A commercial ELISA kit was used for the measurement of plasma FGF21 (R&D systems Europe Ltd, Oxon, UK).

Human derived hepatocarcinoma HepG2 cells (American Type Culture Collection, Manassas, VA) were cultured in Dulbecco modified Eagle medium (Gibco®, Life Technologies, Zug, Switzerland) supplemented with 10% fetal bovine serum (FBS, Sigma-Aldrich, Buchs, Switzerland), 1% (v/v) penicillin streptomycin (Sigma-Aldrich,) at 37 °C with 5% CO₂. Equivalent numbers of HepG2 cells starved in DMEM basal serum-free media were treated with 200  ng/mL recombinant leptin (R&D Systems, Abingdon, UK), PBS (control group), and/or a STAT3 inhibitor (10 µmol/L) (Stattic, Sigma-Aldrich) for 1 h.

Total RNA from HepG2 cells was extracted using a single-step extraction with Trizol reagent (Sigma-Aldrich). RNA integrity was assessed by electrophoresis on a 1% agarose gel and concentration was determined by spectrophotometry. A quantity of 1 µg of total RNA was used for RT, using a commercial kit (Takara Bio Europe, St Germain-en-Laye, France). For quantitative PCR, amplification of genes was performed from 50 ng cDNA using the SYBR® green PCR Master Mix (Roche, Rotkreuz, France) and a StepOne™ Real-Time PCR system (Life Technologies), as previously described [25]. Oligonucleotides were used at 300 nmol/L and results were normalized to the expression levels of housekeeping genes (Table 1).

HepG2 cells lysates were prepared in RIPA buffer (150 mmol/L NaCl, 1% (v/v) Triton X-100, 0.1% (w/v) SDS, 0.5% (w/v) Na deoxycholate, 50 mM Tris-HCl) with phosphatase and protease inhibitors (Halt protease and phosphatase inhibitor cocktail, Pierce, Lausanne, Switzerland) as previously described [26]. Samples were resolved on a 4-12% Bis-Tris gel and transferred onto polyvinylidene difluoride membranes. Thereafter, proteins were detected using the following antibodies: anti-pSTAT3 (1:2000, Cell signaling, Allschwil, Switzerland), anti-STAT3 (1:2000, Cell signaling), anti-PPARα (1:2000, Cell signaling) and tubulin (1:4000, Sigma-Aldrich). The bands were visualized by chemiluminescence (Supersignal West Dura substrate, Pierce) on a detection system (GE Healthcare Europe GmbH, Glattbrugg, Switzerland). Densitometric analysis of chemiluminescent signals captured on camera was performed using the Image J software (National Institutes of Health).

Results are expressed as Mean ± SEM. Statistical calculations were carried out with GraphPad Prism 6 (GraphPad Prism, La Jolla, CA). Statistical significance was established at p<0.05 and determined by unpaired Student's t-test or one-way ANOVA.

To study the link between leptin and FGF21, adult male Wistar rats were challenged with an intraperitoneal leptin injection (2 mg/kg). Plasma FGF21 levels were then evaluated 30 minutes later. Leptin treatment significantly increased plasma FGF21 levels compared to the injection of vehicle (saline), confirming a stimulatory effect of leptin on FGF21 release in this model (Fig. 1). The mRNA expression of Fgf21 was thereafter evaluated in different tissues. The liver was the tissue predominantly expressing Fgf21, compared to brown (BAT) and inguinal white (WATi) adipose tissue (Fig. 2). However, under our experimental conditions (i.e. 30 minutes after injection), leptin was unable to increase Fgf21 mRNA levels in the liver or in the other tissues (Fig. 2). This suggests a potential time-dependant effect, or an effect of leptin on FGF21 release rather than expression.

To strengthen our in vivo results, we used an in vitro model, focusing on the liver, which is considered the main site of FGF21 production. HepG2 cells were therefore chosen and treated for 1 h with leptin (200 ng/mL). Leptin treatment enhanced Fgf21 expression under these experimental conditions (Fig. 3A). Interestingly, an almost significant (p=0.07) stimulation of the mRNA expression of Socs3, a direct target of STAT3, was observed (Fig. 3A). As leptin is known to promote activation of STAT3 [27], we assessed whether the leptin-induced increase in Fgf21 expression could be mediated by the STAT3 signaling pathway. To this end, HepG2 cells were treated with leptin (200 ng/mL) in the presence, or in the absence, of a STAT3 inhibitor (Stattic, 10 µmol/L, 1 h). As expected, leptin increased STAT3 phosphorylation whereas the presence of the STAT3 inhibitor in the medium resulted in a significant inhibition of phosphorylation (Fig. 3B). Importantly, the presence of the inhibitor also prevented the leptin-induced Fgf21 expression (Fig. 3C), confirming that the effect of leptin on FGF21 is mediated by STAT3 activation.

As PPARα is also known to regulate Fgf21 gene expression [28], we then tested whether leptin could stimulate Fgf21 expression through the PPARα pathway. We observed that 1 h of leptin treatment increased Pparα mRNA expression in HepG2 cells (Fig. 4A). However, as determined using Western blotting, PPARα protein levels tended to be increased by leptin (Fig. 4B), but this was not significant (p=0.1). These results could potentially suggest a post-transcriptional regulation of PPARα by leptin, which will require further investigations in future studies. Alternatively, it is conceivable that 1 h of leptin treatment was not long enough to observe any significant difference in protein synthesis.

FGF21, a potent metabolic regulator, has been shown to improve glucose and lipid metabolism, as well as to reduce body weight and adipose tissue mass [3,4,5,6,29,30,31]. These favorable effects seem to be mainly mediated by white adipose tissue, since lack of FGF receptor 1 (FGFR1) in this tissue abolishes the metabolic actions of FGF21 [32,33]. FGF21 has also been shown to improve insulin sensitivity and hepatic steatosis in high-fat fed mice [4]. On the opposite, mice lacking Fgf21 develop hepatic steatosis and hepatic insulin resistance when fed a high-fat ketogenic diet [23]. Several studies focused on the identification of FGF21 mediators [8,9,10]. Although adiponectin was described as one of the main mediators of FGF21 action, recent evidence suggests that other factors may be involved as well [11,12]. Nevertheless, limited information is available regarding the regulation of Fgf21 expression itself.

Here, we report that leptin is a potential new regulator of FGF21. Indeed, we observed that leptin administration acutely increased plasma FGF21 plasma levels in Wistar rats. This likely resulted from a release of FGF21 from the liver, as suggested by the literature [3], although we could not find any difference in Fgf21 expression in the liver or other tissues. However, skeletal muscle has also been suggested as a site of FGF21 production [34,35] and could have revealed increased Fgf21 expression after leptin administration. However, we did not sample any skeletal muscle in our rats. Nonetheless, it is also possible that timing of tissue harvesting (30 minutes in our in vivo experimental conditions) might have accounted for our findings, suggesting that a faster or slower regulation by leptin on tissue Fgf21 expression might occur. The rapid effect of leptin on FGF21 levels however suggests an effect on secretion. Notably, leptin is known to rapidly (within minutes) and directly inhibit insulin secretion by the pancreas [36]. FGF21 is also rapidly (within 15 minutes) expressed by the pancreas after acute pancreatic injury [37]. Therefore, it could be hypothesized that leptin might stimulate rapid pancreatic FGF21 expression and release. Unfortunately, we did not sample the pancreas of our leptin-treated rats. Altogether, these considerations therefore warrant further in vivo research.

Nevertheless, our in vitro results obtained in HepG2 cells clearly show an effect of leptin on Fgf21 expression. Moreover, we show that this leptin effect is mediated by STAT3 activation, as not only STAT3 phosphorylation, but also Fgf21 expression, were prevented in the presence of a STAT3 inhibitor. These findings therefore provide a potential signaling pathway by which leptin regulates FGF21. A possible involvement of PPARα activation remains to be elucidated in future studies, but is likely to play a role in the crosstalk between leptin and FGF21. As in vitro studies do not take into account potential interactions between tissues, for example between liver and white adipose tissue, further studies are required in vivo to better clarify the potential regulation of FGF21 by leptin.

The role of FGF21 as a mediator of leptin action may be of significant relevance in various metabolic regulation processes. As an example, impaired leptin signaling in the liver, known to promote hepatic steatosis [38], may be due to impaired hepatic Fgf21 expression. Nevertheless, the peripheral effect of leptin is a subject of debate. Notably, mice with specific leptin receptor deletion in the liver do not recapitulate the metabolic phenotype of db/db mice. Indeed, metabolic effects of leptin and notably its weight-reducing effects are more likely to be the results of the central leptin action on the brain [39]. However, obesity is a leptin resistance state [40]. Moreover, FGF21 levels are also increased in obesity, which also suggest a resistant state [41,42]. Therefore, it could be speculated that co-administration of leptin and FGF21 could reveal a synergistic effect in obesity to overcome this state of hormonal resistance. This may also be true in other situations of hormone resistance, notably insulin resistance, encountered not only in obesity, but also in conditions associated with the metabolic syndrome, such as type 2 diabetes and nonalcoholic fatty liver disease [43]. These considerations warrant further in vivo studies.

Altogether, our results provide new findings regarding a leptin-STAT3-FGF21 axis that may lead to beneficial metabolic effects. These findings warrant further in vivo work to assess whether leptin-induced FGF21 could improve diseases associated with insulin resistance such as obesity, type 2 diabetes and nonalcoholic fatty liver disease.

This work was supported by the Helmut Horten Foundation, the Novartis Foundation for medical-biological research, the Leenaards Foundation and the Raymond Berger Foundation to FR Jornayvaz and by a grant from the Swiss National Science Foundation (Bern, Switzerland), FNRS: 31003A_134919 and 310030_160290/1 to F. Rohner-Jeanrenaud.

The authors have nothing to disclose.

M. Asrih and C. Veyrat-Durebex contributed equally.