Introduction

Fibroblast growth factor 21 (FGF21) is a member of the FGF superfamily and acts as an endocrine hormone. It is produced at high levels in several tissues, including the adipocytes, liver, muscle, and heart1. The concentration of hepatic FGF21 increased under various pathophysiological conditions, including fasting, starvation, ketogenic diet, fatty liver disease, obesity, and mitochondrial dysfunction2,3. FGF21 plays a key role in diverse physiological functions, such as gluconeogenesis, ketogenesis, steatosis, and inflammation1,4. Moreover, FGF21 enhances insulin sensitivity and fatty acid oxidation and attenuates fat accumulation, pro-inflammatory cytokines, reactive oxygen species, and growth hormone resistance by controlling several transcription factors, including peroxisome proliferator-activated receptor α (PPARα), farnesoid X receptor (FXR), activating transcription factor 4 (ATF4), and nuclear factor-kappa B (NF-κB)1,5,6.

B-cell translocation gene 2 (BTG2) belongs to the BTG/Tob protein family, which is characterized by the presence of two homology domains (Box A and B) that are highly conserved among diverse species7. BTGs are anti-proliferative factors and are key modulators of cellular manifestations, such as cell growth, proliferation, differentiation, death, and survival7,8. It is mainly expressed in the liver and has also been detected in diverse tissues, including the kidney, skeletal muscle, lung, pancreas, and intestine9. BTG2 levels are increased by hypoxia, oxidative stress, metabolic changes, and retinoic acid, whereas they are reduced by insulin, estrogen, and growth factors10. Interestingly, we demonstrated previously that BTG2, a transcriptional co-regulator, enhances the expression of several target genes, including insulin and hepcidin10,11.

Kruppel-like factor 15 (KLF15), also known as kidney-enriched KLF (KKLF), is a member of the KLF family of transcription factors. Member of this family contain a zinc finger DNA-binding domain that is highly conserved among diverse species. KLF15 is predominantly expressed in the liver, but there is also significant expression in other tissues, including the kidney, pancreas, muscle, and heart12. KLF15 expression is upregulated in diverse tissues by glucagon and glucocorticoids during starvation or under diabetic conditions, whereas feeding and insulin downregulate KLF15 expression12,13. KLF15 also regulates the transcription of target genes involved in diverse physiological processes, including fibrosis, cardiovascular disease, and the immune response. It is associated with metabolic dysfunction, including cardiac hypertrophy, obesity, inflammation, and diabetes12,14. Gluconeogenic signals are known to modulate FGF21 production via the activation of glucocorticoid receptor (GR) and PPARα in mice subjected to prolonged fasting or starvation15,16. We have previously shown that BTG2 is induced under glucagon and fasting condition7,11. Moreover, KLF15 is known to be upregulated by glucagon and downregulated by insulin12,13. However, the connection of BTG2 and KLF15 in the regulation of FGF21 production is largely unknown.

In the current study, we demonstrate that BTG2 is a key modulator of FGF21 production and reveal a novel molecular mechanism that links KLF15 to the control of FGF21 biosynthesis during starvation.

Results

Gluconeogenic signals elevate FGF21 gene expression and biosynthesis

Previous studies have showed that gluconeogenic signals modulate FGF21 production via the GR and PPARα in mice during prolonged fasting and/or starvation15,16. Based on these findings, we examined the physiological connection between gluconeogenic modulators and hepatic FGF21 metabolism. Notably, the expression of Fgf21 was significantly elevated under prolonged fasting, along with the induction of Btg2 and Klf15 expression when compared with that of the fed state (Fig. 1a). Consistent with the increase in Fgf21, Btg2, and Klf15 mRNA expression, the corresponding protein levels were markedly elevated in the livers of the fasted mice (Fig. 1b). In addition, serum FGF21 concentrations was also increased in the fasted mice compared to the fed state (Fig. 1c). Furthermore, we examined the critical effect of gluconeogenic signaling on the regulation of Fgf21 expression and the secretion of FGF21 in the liver. FSK treatment enhanced the protein and mRNA levels of FGF21, BTG2, and KLF15 (Fig. 1d,e). Likewise, FSK challenge also increased the serum FGF21 concentration relative to that of the control groups (Fig. 1f). Collectively, these findings suggested a potential link between BTG2 and FGF21 biosynthesis in response to gluconeogenic signals.

Figure 1
figure 1

Fasting state and forskolin exposure elevate hepatic FGF21 gene expression and production in the liver. (a) Wild-type (WT) mice were fed ad libitum and fasted for 24 h. Gene expression levels were measured by qPCR analysis with various primers. (b) Tissue extracts were analyzed by Western blot analysis using specific antibodies. (c) Serum FGF21 levels at the indicated conditions. (d) WT mice were injected intraperitoneally with forskolin (FSK, 5 mg/kg body weight) for 6 h, and then assessed by qPCR analysis with gene-specific primers. (e) Tissue extracts were analyzed by Western blot analysis with various antibodies. (f) Serum FGF21 levels in the indicated groups of mice; n = 7 mice per group. *P < 0.05, **P < 0.01 vs. fed mice or untreated control mice.

Full size image

BTG2 elevates FGF21 production via the induction of KLF15

We have attempted to investigate the critical role of BTG2 as a key modulator of FGF21 biosynthesis using an adenoviral delivery system expressing Btg2 (Ad-Btg2) or a control green fluorescent protein (Ad-GFP) in mouse livers. Ad-Btg2 was successfully delivered to the livers of wild-type (WT) mice via tail vein injection. The expression levels of Klf15 and Fgf21 were significantly higher in the Ad-Btg2-infected mice than in the Ad-GFP control mice (Fig. 2a,b). As expected, Ad-Btg2 significantly increased the serum FGF21 level compared to that in the Ad-GFP control mice (Fig. 2c). Next, to determine whether BTG2-mediated induction of FGF21 expression and production can be modulated by KLF15, we assessed the effect of Klf15 on the regulation of FGF21 gene expression and biosynthesis using adenoviral-mediated overexpression of Btg2 (Ad-Btg2) and knockdown of Klf15 (Ad-shKlf15) both in vivo and in vitro. Ad-Btg2 effectively enhanced Klf15 and Fgf21 mRNA levels, while this phenomenon was markedly negated by silencing of Klf15 in mouse livers and hepatocytes (Fig. 2d,f). Similarly, the production of FGF21 increased by Ad-Btg2 was remarkably decreased in Klf15 knockdown hepatocytes and mouse livers (Fig. 2e,g). Interestingly, Ad-shKlf15 slightly reduced basal Fgf21 expression in the hepatocytes and mice relative to the Ad-GFP control group, but not the production of FGF21 (Fig. 2d–g). To further confirm the transcriptional activity of Fgf21 by FSK treatment, Btg2, Klf15, we investigated Fgf21 promoter activity in hepatocytes. Notably, FSK exposure or transiently expressed Btg2 significantly elevated the promoter activity of Fgf21, whereas this stimulation was strikingly blunted in Klf15 knockdown groups compared to the control groups (Fig. 2h). Overall, these results demonstrate that BTG2 acts as a major modulator of FGF21 gene expression and biosynthesis by depending on KLF15 both in vivo and in vitro.

Figure 2
figure 2

BTG2 increases FGF21 gene expression and biosynthesis. (a) WT mice were tail-vein injected with Ad-GFP and Ad-Btg2 for 7 days. qPCR analysis showing Btg2, Klf15, Fgf21 expression in the liver. (b) Tissue extracts were analyzed by Western blot analysis with the indicated antibodies. (c) Serum FGF21 production in the observed mice. (d) WT mice were intravenously injected with Ad-GFP, Ad-Btg2, and Ad-shKlf15 for 7 days. Total RNAs were isolated from the mouse livers, and the expression levels of the various genes were determined by qPCR analysis with specific primers. (e) Serum FGF21 production in the indicated mice; n = 7 mice per group. (f) AML12 cells were infected with Ad-GFP, Ad-Btg2, and Ad-shKlf15 for 36 h, and then analyzed by qPCR with various primers. (g) The culture media in the AML12 cells was harvested for FGF21 secretion analysis as indicated. (h) AML12 cells were transiently transfected with siKlf15 and siScram. After transfection for 36 h, the cells were co-transfected with the indicated reporter gene and Btg2 and subjected to FSK treatment for 6 h. *P < 0.05, **P < 0.01, ***P < 0.001 vs. Ad-GFP, Ad-Btg2, untreated controls, or FSK-treated cells.

Full size image

Gluconeogenic signal-stimulated FGF21 production depends on BTG2

We further investigated the pivotal role of BTG2 in fasting-mediated FGF21 metabolism using lentiviral-mediated knockdown of Btg2 (shBtg2) in mouse livers. The expression of Btg2 was successfully attenuated in the mouse livers. The elevation of Btg2, Klf15, and Fgf21 mRNA and protein levels during fasting were markedly alleviated by endogenous Btg2 knockdown (Fig. 3a,b). As anticipated, the increase of serum FGF21 concentration induced by fasting was strikingly reduced in the Btg2 knockdown mice (Fig. 3c). Moreover, basal FGF21 expression was weakly attenuated in Btg2-silenced mice compared to the control group, but not the secretion of FGF21 (Fig. 3a–c). Next, we further verified whether BTG2 modulates the transcriptional activity of Fgf21 in hepatocytes. As shown in Fig. 3d, Fgf21 promoter activity was enhanced by FSK exposure, and this stimulatory effect of FSK was markedly diminished when Btg2 was silenced. Taken together, these findings suggest that BTG2 plays an important role in modulating FGF21 production during fasting.

Figure 3
figure 3

Elevation of FGF21 production by fasting and forskolin treatment is mediated by BTG2. (a) WT mice were tail-vein injected with lentivirus-shBtg2 (shBtg2). After 7 days, the mice were fasted for 24 h. Total RNAs were isolated from the mouse livers and expression levels were determined by qPCR analysis with various primers. (b) Western blot analysis showing BTG2, KLF15, and FGF21 expression in the mouse livers. (c) Serum levels of FGF21 in the indicated group of mice; n = 7 mice per group. (d) AML12 cells were transfected with siBtg2 and siScram. After transfection for 36 h, the cells were transiently transfected with the indicated reporter genes and subjected to FSK treatment for 6 h. *P < 0.05, **P < 0.01 vs. untreated controls, fasted mice, or forskolin-treated cells.

Full size image

KLF15 is required for fasting-induced FGF21 expression and biosynthesis

We explored the possible effects of KLF15 on FGF21 gene regulation and production in response to gluconeogenic stimuli in mouse livers using Ad-Klf15. As shown in Fig. 4a,b, qPCR and Western blot analysis showed the successful overexpression of Klf15 in mouse livers. Ad-Klf15 significantly elevated Fgf21 gene expression (Fig. 4a,b), and consequently increased serum FGF21 levels compared with those in the control groups (Fig. 4c). We further investigated the direct effect of KLF15 on fasting-induced FGF21 gene expression and biosynthesis using Ad-shKlf15. The expression of Klf15 in the liver of mice was successfully decreased by Ad-shKlf15 delivery. The increase in Fgf21 gene expression induced by fasting was markedly attenuated by Klf15 silencing (Fig. 4d,e). Moreover, the fasting-induced serum FGF21 concentration was dramatically diminished by the knockdown of endogenous Klf15 (Fig. 4f). Notably, basal Fgf21 expression was slightly reduced in the Klf15-silenced mice relative to the control group, but not the production of FGF21 (Fig. 4d–f). Collectively, these observations strongly suggest that gluconeogenic stimuli guide FGF21 gene expression and production and are partially dependent on KLF15.

Figure 4
figure 4

KLF15 regulates fasting-induced FGF21 expression and biosynthesis. (a) WT mice were tail-vein injected with Ad-GFP and Ad-Klf15 for 7 days. Total RNA expression levels were measured by qPCR analysis with gene-specific primers. (b) Tissue extracts were analyzed by Western blot analysis with various antibodies. (c) Serum levels of FGF21 in the observed mice. (d) WT mice were tail-vein injected with Ad-shKlf15 for 7 days. The mice were then fasted for 24 h. Gene expression levels were analyzed using qPCR with various primers. (e) Western blot analysis showing KLF15 and FGF21 expression in the mouse livers. (f) Levels of serum FGF21 in the indicated mice; n = 7 mice per group. *P < 0.05, **P < 0.01, ***P < 0.001 vs. Ad-GFP or fasted mice.

Full size image

BTG2 physically interacts with KLF15 and controls KLF15 occupancy on the Fgf21 promoter

To further elucidate whether BTG2 and KLF15 in the regulation of Fgf21 gene transcription via physically interaction, we carried out co-immunoprecipitation (Co-IP) assays using lysates from fasted and fed mouse livers. Endogenous BTG2 interacted strongly with KLF15 protein in the fasted mice compared to the fed mice (Fig. 5a). We identified the KLF15-binding site on the Fgf21 promoter using in silico analysis. Our reporter gene assay indicated that Btg2 and Klf15 alone increased Fgf21 promoter activity, whereas this stimulation was completely negated in the KLF15-binding site-mutated Fgf21 promoter compared to the control group (Fig. 5b). To further confirm that BTG2 affects the recruitment of KLF15 protein to the Fgf21 promoter, we carried out using a chromatin immunoprecipitation (ChIP) assay with an anti-KLF15 antibody in the mouse livers. As expected, endogenous KLF15 occupancy of the proximal (Pro) region was significantly enhanced by fasting and/or Ad-Btg2 compared to that in the control groups. However, this effect was totally lost in the nonspecific distal (Dis) region of the Fgf21 promoter (Fig. 5c). Overall, these results strongly suggest that BTG2 associates with KLF15 and enhances Fgf21 transcription by recruiting KLF15.

Figure 5
figure 5

BTG2 associates with KLF15 and is recruited on the Fgf21 promoter. (a) WT mice were fed ad libitum and fasted for 24 h. Co-immunoprecipitation (Co-IP) assays with liver extracts indicated the association between BTG2 and KLF15. Protein extracted from the livers was immunoprecipitated with either KLF15 or BTG2 antibodies, and then immunoblotted with the same antibodies. (b) Schematic diagrams of the KLF15 binding site on the WT Fgf21 promoter construct from −1091 to −1085 bp and its mutant form (mt). AML12 cells were transfected with the indicated reporter genes by Btg2 or Klf15 co-transfection and subjected to FSK treatment for 6 h. (c) Chromatin immunoprecipitation (ChIP) assay showing the occupancy of KLF15 on the Fgf21 promoter in the liver. WT mice were infected with Ad-Btg2 for 7 days and fasted for 24 h. Input was the 10% of soluble chromatin. The remaining soluble chromatin was subjected to immunoprecipitation using anti-KLF15 antibodies or IgG as indicated. Purified DNA samples were utilized for qPCR analysis with specific primers binding to the indicated regions (proximal and distal region) on the Fgf21 promoter. *P < 0.05 vs. untreated controls or fasting mice.

Full size image

Discussion

In the current study, we demonstrated that fasting and/or FSK treatment remarkably induced hepatic FGF21 gene transcription, and BTG2 elevated FGF21 biosynthesis via the upregulation of KLF15 expression. Conversely, the stimulatory effect of fasting or FSK exposure on hepatic FGF21 gene expression and production was prominently attenuated by the disruption of Btg2 and Klf15 expression in mouse livers and hepatocytes. Therefore, we propose that the fasting-BTG2-KLF15 signaling network may represent a novel molecular mechanism underlying the modulation of hepatic FGF21 gene expression and its biosynthesis.

Several previous reports have shown that gluconeogenic signals modulate hepatic FGF21 gene expression and biosynthesis in rodents15,16,17. Our previous study revealed that the elevation of BTG2 by glucagon increased hepcidin levels and gluconeogenesis in mouse livers11,18. However, the potent effect of BTG2 in modulating hepatic FGF21 gene expression and FGF21 biosynthesis remains largely unexplored. The results of the current study demonstrate that enhanced BTG2 levels induced by fasting or FSK challenge modulate hepatic FGF21 production by upregulating the expression of Klf15. Fasting and FSK treatment significantly increase the expression of Btg2, Klf15, and Fgf21 in mouse livers and significantly elevate the concentration of serum FGF21 (Fig. 1). Ad-Btg2 significantly increased hepatic FGF21 gene expression and production by stimulating KLF15 expression (Fig. 2), whereas this stimulation was markedly attenuated in Btg2 knockdown cells and mice (Fig. 3). Our findings suggest that BTG2 plays a crucial role in modulating the fasting-mediated induction of hepatic FGF21 gene expression and biosynthesis via the induction of KLF15.

KLF15 modulates the transcription of several target genes that are involved in various physiological processes, such as fibrosis, cardiac hypertrophy, obesity, inflammation, insulin resistance, and diabetes12,13,14,19,20. Based on these findings, we investigated the molecular mechanism underlying fasting-stimulated Fgf21 gene transcription by the BTG2-KLF15 pathway both in vivo and in vitro. First, the increase in FGF21 gene expression induced by fasting was markedly diminished by silencing of Klf15 relative to that in the control groups (Fig. 4). Second, BTG2 strongly interacted with KLF15 in the liver lysates from fasted mice compared to the lysates from fed mice. Third, the endogenous KLF15 occupancy of the proximal region of the Fgf21 promoter was effectively enhanced by fasting and Ad-Btg2 transduction; and the transcriptional activity of Fgf21 was significantly increased by Btg2 and Klf15 in the hepatocytes (Fig. 5). Collectively, our results reveal a connection between Fgf21 transcription and the BTG2-KLF15 axis. However, we cannot rule out the possibility that the detailed molecular mechanism that connects FGF21 gene expression and the BTG2-KLF15 signaling network depends on unknown mechanisms such as an association with transcriptional co-activators and competition with co-repressors, microRNAs, and protein stability to modulate hepatic FGF21 gene transcription.

In conclusion, the present study demonstrate that FGF21 is a novel target of BTG2, and that BTG2 promotes FGF21 biosynthesis by upregulating the expression of Klf15 during fasting or FSK exposure. These findings suggest that the elevation of BTG2 by gluconeogenic signals modulates hepatic FGF21 homeostasis by inducing the expression of Klf15. Therefore, as depicted in Fig. 6, a molecular mechanism involving hepatic FGF21 homeostasis in response to BTG2-KLF15 signaling may provide the basis for the development of novel therapeutic agents for the treatment of metabolic disorders.

Figure 6
figure 6

Schematic model illustrating the regulation of FGF21 biosynthesis by the BTG2-KLF15 signaling network. Fasting and forskolin treatment significantly induce the expression of FGF21 by enhancing the BTG2-KLF15 signaling network, which subsequently increases FGF21 production.

Full size image

Materials and Methods

Animals

We used 8-week-old male C57BL/6 mice (Samtako, Osan, Republic of Korea) for the experiments, as described below11. For the fasting and feeding experiments, we fed the mice ad libitum, and then fasted them for 24 h. For the forskolin (FSK) stimulation experiments, we injected the mice intraperitoneally with FSK (Sigma-Aldrich, St. Louis, MO, USA) at a dose rate of 5 mg/kg of body weight and left them for 6 h. For the Btg2 and Klf15 overexpression experiments, we injected wild-type (WT) mice with adenoviral vectors expressing Btg2 and Klf15 (1 × 109 plaque-forming units, pfu) via their tail veins for 7 days. For the disruption of the Btg2 and Klf15 genes, we injected WT mice with lentivirus short hairpin RNA (shRNA) targeting Btg2 (shBtg2, a single dose of 1 × 109 transducing units, TU/mL) and with Ad-shKlf15 (1 × 109 pfu) through their tail veins for 7 days. At the end of the specified experiments or challenge periods, we euthanized the mice with CO2, and harvested liver tissues and blood samples. All animal experiments and protocols were approved and performed by the Institutional Animal Care and Use Committee (IACUC) of the Kyungpook National University according to the rules and guidelines of the National Institutes of Health (NIH).

Measurement of serum FGF21

We collected the blood samples immediately, and determined serum FGF21 levels using a Quantikine FGF21 enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s instructions21.

Construction of plasmids and DNA

The mouse Fgf21 promoter and Klf15 expression vector were kind gifts from Drs. Hueng-Sik Choi (Chonnam National University, Gwangju, Republic of Korea) and Myung-Shik Lee (Yonsei University College of Medicine, Seoul, Republic of Korea), respectively21,22. The KLF15 response element-mutated Fgf21 promoter was generated using a Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA, USA) with the following primers: forward, 5′-CTAATCCTCCCTTTCCCCAAA-3′, and reverse, 5′-TTTGGGGAAAGGGAGGATTAG-3′. The Btg2 expression vector has been reported previously7. All constructs were confirmed via sequencing analysis.

Culture of hepatocytes and transient transfection assays

We cultured AML-12 cells (immortalized mouse hepatocytes) in Dulbecco’s modified Eagle’s medium (DMEM)/F-12 medium (Gibco-BRL, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (Gibco-BRL), insulin-transferrin-selenium (Gibco-BRL), dexamethasone (40 ng/ml, Sigma-Aldrich), and antibiotics in a humidified atmosphere containing 5% CO2 at 37 °C, according to a previously described method18,23. The transient transfection assays were performed using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) with the indicated reporter plasmid and expression vectors encoding various genes and/or chemicals according to the manufacturer’s protocol24. The luciferase activity was normalized to β-galactosidase activity.

Recombinant adenovirus and small interfering RNA (siRNA)

Adenoviruses encoding full-length Btg2 (Ad-Btg2), green fluorescent protein (GFP), and a lentiviral delivery system for Btg2-targeted shRNA (shBtg2) have been described previously18. We purchased Ad-Klf15 and Ad-shKlf15 from Vector Biolabs (Malvern, PA, USA). The siRNAs for Btg2 (siScram and siBtg2) and Klf15 (siScram and siKlf15) were chemically manufactured (Bioneer Research, Seoul, Republic of Korea) and transfected using Oligofectamine reagent (Invitrogen) in accordance with the manufacturer’s protocol, as previously described7.

Measurement of mRNA

Total RNA was extracted from each sample using the TRIzol method (Invitrogen, Carlsbad, CA, USA), as mentioned previously25. We synthesized complementary DNA (cDNA) using a Maxima® First Strand cDNA synthesis kit (Fermentas, Vilnius, Lithuania), and used a StepOneTM Real-time PCR system (Applied Biosystems, Warrington, UK) for quantitative real-time polymerase chain reaction (qPCR) analysis. We determined Btg2, Klf15, and Fgf21 gene expression levels by qPCR, as described previously11,19,21. All the transcripts were normalized to ribosomal L32 expression.

Western blot analysis

We extracted liver samples for Western blot analysis according to a previously described method11,21. The membranes were probed with BTG2, KLF15, and β-actin (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and FGF21 (Abcam, Cambridge, UK). Immunoreactive proteins were developed using an ECL-Plus Western blot detection kit (GE Healthcare, Piscataway, NJ, USA) in accordance with the manufacturer’s protocol.

Co-immunoprecipitation assay

We immunoprecipitated the total proteins from the liver extract with antibodies against BTG2 and KLF15 (Santa Cruz Biotechnology) using protein A/G PLUS-Agarose (Santa Cruz Biotechnology). The immunoprecipitated proteins were then subjected to Western blot analysis using the corresponding antibodies11. Signals were developed using an ECL-Plus Western blot detection kit (Amersham Bioscience, Piscataway, NJ, USA).

Chromatin immunoprecipitation (ChIP) assay

The ChIP assay was carried out using the anti-KLF15 antibody, as described previously11,18. Briefly, soluble chromatin was subjected to immunoprecipitation using the anti-KLF15 antibody (SantaCruz Biotechnology). After recovering DNA, the purified DNA samples were utilized for qPCR with specific primers covering the (proximal and distal regions) of the Fgf21 promoter. The primers used for PCR were as follows: proximal region, forward 5′-GGGTTCCTCCTAGAAATCCA-3′ and reverse 5′-CAGTCACCCTCACCAACCCC-3′; distal region, forward 5′-GCTGCGCCCTCTCCTCCGCC-3′ and reverse 5′-TGGGAACGT-GCATAGAACCT-3′.

Statistics

Data and statistical analysis were performed using GraphPad Prism 3–5.0 software. The statistical significance between groups was determined by applying the Student’s t-test or the Mann-Whitney U-test, and multiple comparisons were analyzed using one-way analysis of variance (ANOVA) or the Kruskal Wallis test. All data are represented as the mean ± the standard error of the mean (SEM). Differences were considered statistically significant at P-values of <0.05.