FGF11 influences 3T3‐L1 preadipocyte differentiation by modulating the expression of PPARγ regulators

Fibroblast growth factor 11 (FGF11) is a member of the intracellular fibroblast growth factor superfamily. Here, we identified FGF11 as a novel mediator of adipogenesis. During 3T3‐L1 adipocyte differentiation, the expression of FGF11 decreased at the mitotic clonal expansion stage and increased at the terminal differentiation stage. FGF11 knockdown reduced the expression of peroxisome proliferator‐activated receptor gamma (PPARγ), a master regulator of adipogenesis, resulting in the inhibition of adipocyte differentiation. Treatment with the PPARγ agonist rosiglitazone restored the inhibition of adipogenesis caused by FGF11 knockdown. We also report that the expression of the PPARγ regulators CCAAT/enhancer‐binding protein α, sterol regulatory element‐binding protein 1, KLF9, KLF2, GATA binding factor 2, and GATA binding factor 3 was influenced by FGF11. These results suggest that FGF11 indirectly controls the expression of PPARγ through modifying the expression of multiple PPARγ regulators, thereby mediating adipogenesis.

Fibroblast growth factor 11 (FGF11) is a member of the intracellular fibroblast growth factor superfamily. Here, we identified FGF11 as a novel mediator of adipogenesis. During 3T3-L1 adipocyte differentiation, the expression of FGF11 decreased at the mitotic clonal expansion stage and increased at the terminal differentiation stage. FGF11 knockdown reduced the expression of peroxisome proliferator-activated receptor gamma (PPARc), a master regulator of adipogenesis, resulting in the inhibition of adipocyte differentiation. Treatment with the PPARc agonist rosiglitazone restored the inhibition of adipogenesis caused by FGF11 knockdown. We also report that the expression of the PPARc regulators CCAAT/enhancerbinding protein a, sterol regulatory element-binding protein 1, KLF9, KLF2, GATA binding factor 2, and GATA binding factor 3 was influenced by FGF11. These results suggest that FGF11 indirectly controls the expression of PPARc through modifying the expression of multiple PPARc regulators, thereby mediating adipogenesis.
The family of fibroblast growth factors (FGFs) is composed of 22 members and is associated with various biological functions, such as growth, wound healing, repair, differentiation, angiogenesis, embryonic development, and metabolic regulation [1]. FGFs can be classified into canonical FGFs, endocrine FGFs, and intracellular FGFs (iFGFs). Canonical and endocrine FGFs function by binding to FGF receptors (FGFRs) with heparin sulfate proteoglycans or klotho proteins, whereas the iFGFs FGF11-FGF14 act as intracellular molecules in an FGFR-independent manner [2]. iFGFs are mainly expressed in the nervous system, and much of the research concerning them has focused on neuronal development [3]. In contrast to other iFGFs, the function and molecular mechanism of FGF11 activity have not been studied well. Recently, our group reported that the promoter of FGF11 contains a hypoxia response element (HRE), and FGF11 expression is induced under hypoxic conditions [4]. We also identified FGF11 as a stabilizer of HIF-1a and an enhancer of capillary-like endothelial tube formation [5,6]. It was also documented that FGF11 is related to osteoclast-mediated bone resorption, tumorigenesis, and liver regeneration [7]. Transcriptome analysis of human adipose-derived stem cells during adipogenesis suggested that FGF11 is a potential mediator of adipogenesis [8]; however, further study is needed to determine the detailed functions and regulatory mechanisms of FGF11 in this process.
Adipocyte differentiation is an orchestrated process controlled by a cascade of multiple regulators. The 3T3-L1 cell line is a well-established model for studying adipogenesis [9]. Adipocyte differentiation can be divided into two phases: the mitotic clonal expansion (MCE) phase and the terminal differentiation phase [10]. Upon treatment with adipogenic inducers (dexamethasone, IBMX, and insulin, also called DMI), growth-arrested 3T3-L1 preadipocytes re-enter the cell cycle and then undergo 2-3 rounds of mitosis [11]. After entering MCE, the adipogenic genes are expressed according to a precise temporal process, ultimately leading to the terminal differentiation phase. CCAAT/enhancer-binding protein b (C/EBPb), which is temporally increased in the MCE phase, has been identified as an important regulator that promotes MCE and initiates terminal differentiation. In terminal adipocyte differentiation, peroxisome proliferator-activated receptor gamma (PPARc) and CCAAT/ enhancer-binding protein a (C/EBPa), essential adipogenic regulators, are induced by C/EBPb. PPARc and C/EBPa positively regulate each other, and the cooperation of these two proteins drives the expression of genes involved in the adipogenic phenotype, such as morphological changes, lipid accumulation, and insulin sensitivity [12].
Understanding the complexity of adipocyte differentiation is of important relevance to human health, because adipocyte dysfunction contributes to metabolic diseases. In the present study, we found that the expression of FGF11 changed dramatically between the two phases of adipogenesis, and the increase of FGF11 in the terminal differentiation phase was necessary for adipocyte maturation. The expression of PPARc contributed to the effect of FGF11 on adipogenesis, and FGF11 altered the expression of a regulator of PPARc expression. This study provides insight into a novel role for FGF11 in adipogenesis.

Cell culture
3T3-L1 preadipocytes were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% bovine serum and 1% penicillin/streptomycin in a 5% CO 2 incubator (GIBCO, Life Technologies Ltd., Paisley, UK). Two days after the cells had reached confluence, differentiation was induced. Cells were incubated in culture media with DMEM containing 10% FBS, 0.5 lM 3-isobutyl-1-methylxanthine (IBMX), 0.25 lM dexamethasone, and 5 lgÁmL À1 insulin for 2 days. Then, cells were incubated in DMEM with 10% FBS and 1 lgÁmL À1 insulin for the following 2 days and were then further cultured in DMEM with 10% FBS for another 4 days. For assessment of adipocyte differentiation, after fixation with 3.5% formaldehyde and washing with PBS, lipid droplets were stained with 1.5% Oil Red O (ORO) in 60% isopropanol for 1 h and washed with PBS to remove the unbound stain. To quantify the ORO content, DMSO was added to each sample, and the absorbance of each sample was measured at 510 nm.

Transient transfection of siRNAs and transduction with adenoviruses
For FGF11 knockdown, cells were transfected with siR-NAs (50 nM) using Lipofectamine RNAiMAX (Invitrogen) 1 day before the induction of adipogenesis. The next day, the culture medium was replaced with differentiation medium. FGF11 was overexpressed by poly-L-lysine (PLL; P9155; Sigma-Aldrich, St. Louis, MO, USA)-assisted transduction 2 days after the induction of adipogenesis. Briefly, adenoviruses and PLL (2.5 lgÁmL À1 ) were incubated in serum-free medium for 30 min. Then, cells were incubated in this mixture for 1 h, followed by the addition of insulinsupplemented medium. After 1 day, the culture medium was replaced with fresh insulin-supplemented medium.

RNA preparation and real-time RT/PCR
Total RNA was extracted using TRIzol (Invitrogen) according to the manufacturer's instructions, and cDNA was synthesized using MMLV reverse transcriptase and random hexamers (Invitrogen) with 1 lg of total RNA. The levels of each gene transcript were analyzed by realtime PCR with gene-specific primers (Table 1), and 18s rRNA was used as an endogenous control. Real-time PCR was performed using Thunderbird SYBR qPCR Mix (TOYOBO, Osaka, Japan) and a StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA, USA).

Statistics
The data were expressed as the means AE standard error (SE), and the difference between the averages was assessed using the Mann-Whitney U-test. A P-value of ˂ 0.05 on the basis of at least three independent sets of experiments was considered to be statistically significant.

FGF11 is differentially expressed in two phases of adipocyte differentiation
The FGF11 of both mouse and human has two isoforms: isoform 1 and isoform 2. Human FGF11 isoform 1 had 97.3% (219/225) sequence identity with mouse FGF11 isoform 1 (Figs 1A and S1). Unlike FGF11 isoform 1, human FGF11 isoform 2 had different features from mouse FGF11 isoform 2. Because the features and sequences of mouse FGF11 isoform 1 showed similarity with human FGF11 isoform 1, we focused on the roles of FGF11 isoform 1 in this study. We measured the expression levels of total FGF11 and FGF11 isoform 1 in 3T3-L1 using real-time PCR with specific primers during adipogenesis (Fig. 1B). The expression of total FGF11 and FGF11 isoform 1 showed similar patterns. Interestingly, the expression of FGF11 changed depending on the phase of adipogenesis. Compared with FGF11 levels before the initiation of MCE, FGF11 expression decreased below 20% in the MCE stage (day 0-2), but it then increased approximately twofold in the terminal differentiation Table 1. Primers used for real-time RT/PCR.

Primers
Sequence Species  Growth-arrested 3T3-L1 preadipocytes were incubated with differentiation inducers, and total mRNA was harvested daily. The FGF11 mRNA level was determined by real-time RT/PCR and normalized to the 18s rRNA level. The value from harvested cells just before the adipogenic induction (day 0) was set to 1, and the others were calculated relative to this value (mean AE SEM, n = 4). The significance was assessed using the Mann-Whitney U-test. *P < 0.05 vs day 0. stage (day 3-6). The temporal regulation of FGF11 expression during adipogenesis suggests a role for this protein in this process.

FGF11 knockdown inhibits adipocyte differentiation
The expression of FGF11 was obviously elevated in the terminal differentiation stage. To investigate whether the increase of FGF11 is involved in adipogenesis, siRNAs against FGF11 isoform 1 (siFGF11) were transfected into 3T3-L1 preadipocytes 1 day before differentiation induction. Transfection of siFGF11 prevented the elevation of FGF11 and maintained a low level of FGF11 until 6 days after adipogenic stimulation ( Fig. 2A). The siFGF11transfected cells showed remarkably low amounts of lipid droplets stained by ORO compared to control cells (Fig. 2B). The knockdown of FGF11 accompanied a decrease in adipocyte differentiation, indicating the involvement of FGF11 in adipogenesis. During adipogenesis, PPARc is known to be induced beginning at day 2 after adipogenic stimulation, and it plays a pivotal role in terminal adipocyte differentiation [13]. To investigate whether PPARc is involved in the reduced adipocyte differentiation observed in FGF11-depleted cells, we increased the transcriptional activity of PPARc by treatment with a PPARc agonist, rosiglitazone [14] beginning on day 2 after adipogenesis induction. The reduced lipid level induced by FGF11 knockdown was mostly restored by treatment with rosiglitazone, although the restoration was not complete (Fig. 2B), implying that lower PPARc is associated with the inhibition of adipogenesis by FGF11 knockdown.

FGF11 knockdown inhibits the expression of PPARc and C/EBPa
To investigate how FGF11 affects 3T3-L1 adipogenesis, we measured the levels of key regulators of adipogenesis in FGF11-depleted cells and control cells. The expression of PPARc and C/EBPa, essential regulators of terminal adipocyte differentiation, was induced on day 2 at the end of the MCE phase and was increased during terminal differentiation in cells transfected with siNS (Fig. 3A,B). In contrast, the expression of these proteins was greatly reduced in FGF11-deficient cells throughout the adipocyte differentiation phases, consistent with Fig. 2. This result demonstrates that FGF11 is required for the induction and maintenance of PPARc and C/EBPa during adipogenesis.
CCAAT/enhancer-binding protein b is a positive regulator of the expression of PPARc and C/EBPa during adipocyte differentiation. The long C/EBPb isoform [liver-enriched activating protein (LAP)] is a known transcriptional activator, whereas the truncated isoform LIP, or liver-enriched inhibitory protein, has been reported to be a dominant-negative repressor [15]. To determine whether FGF11 induces PPARc and C/EBPa expression by regulating C/EBPb expression, the levels of C/EBPb were examined. Total protein levels of C/EBPb in FGF11-depleted cells were similar to those of the siNS-treated sample during the MCE phase (Fig. 3B), and the ratio of LAP/LIP was not changed by FGF11 knockdown (Fig. 3C). These data imply that the expression of C/EBPb is not associated with FGF11, and the reduced expression of PPARc and C/EBPa is not due to the failure of C/EBPb induction during the MCE phase. b-Catenin is known to suppress adipogenesis by reducing the expression of PPARc and C/EBPa [16,17]. To investigate whether the expression of b-catenin was also influenced by FGF11 expression, the b-catenin protein levels were monitored during adipogenesis. In control cells, the expression of b-catenin was significantly decreased from the baseline levels after induction of MCE and was further reduced during terminal differentiation. However, in FGF11depleted cells, the expression of b-catenin was reduced on day 2, similar to the level observed in the control cells, but it increased during terminal differentiation. These results raise a possibility that b-catenin may be related to the role of FGF11 during terminal differentiation (Fig. 3A,B). However, during adipogenesis, PPARc downregulated b-catenin expression by a mechanism that involved proteasomes [18], indicating that the effect of FGF11 knockdown on b-catenin may be due to an indirect effect of PPARc cross-regulation. Taken together, these data suggest that FGF11 knockdown inhibits adipocyte differentiation by regulating the protein levels of the main factors in adipogenesis, such as PPARc, C/EBPa, and b-catenin.

Expression of FGF11 regulates adipogenesis by modulating the expression of PPARc and C/EBPa
To confirm the effect of FGF11 on 3T3-L1 adipogenesis, we investigated whether the overexpression of FGF11 enhanced adipogenesis by upregulating PPARc and C/EBPa. 3T3-L1 cells were transduced with FGF11-expressing adenoviruses at day 2 after adipogenic stimulation to mimic the temporal expression of FGF11 during adipogenesis. As shown in Fig. 1B, FGF11 mRNA decreased during the first 2 days of the MCE phase and then abruptly increased to a higher level than that observed in the preadipocyte state. When the cells were stained with ORO on day 6 after transduction, significantly higher amounts of lipid droplets were found in cells overexpressing FGF11, clearly demonstrating that FGF11 has an essential role in adipocyte differentiation (Fig. 4A). The levels of adipogenic regulators were analyzed at day 4 after adipogenic stimulation, which was 2 days after transduction and is the point at which the expression of PPARc and C/EBPa reached their maximum levels (Fig. 3A,B). The levels of b-catenin were not changed in cells overexpressing FGF11 (Fig. 4B,C), suggesting that FGF11 is not directly involved in the suppression of b-catenin during terminal adipocyte differentiation. We observed increases in the protein and mRNA levels of both PPARc and C/EBPa upon FGF11 overexpression (Figs 4B,C and 5B). It is known that PPARc and C/EBPa are cooperatively expressed by a cross-regulatory mechanism [19,20]. Furthermore, the level of accumulated lipids was also enhanced by FGF11 overexpression from day 2 after adipogenic stimulation (Fig. 4A). Taken together, these results suggest that the regulation of PPARc expression could be a major target of FGF11 in adipogenesis.
To investigate how FGF11 affects adipogenesis, we assessed the transcript levels of regulatory genes involved in adipocyte differentiation. Among the regulatory genes, we examined the expression of regulators of PPARc, because PPARc is likely to be the major mediator of the effects of FGF11 on adipocyte differentiation. As positive regulators of PPARc [21], the transcripts of KLF9, sterol regulatory element-binding protein 1 (SREBP1), and C/EBPa were reduced by FGF11 knockdown (Fig. 5A). However, the expression levels of GATA binding factor 2 (GATA2), GATA binding factor 3 (GATA3), and KLF2, which are negative regulators of PPARc [21], were upregulated by FGF11 knockdown (Fig. 5A). When FGF11 was overexpressed in 3T3-L1 cells beginning on day 2 of adipogenesis, the expression levels of KLF9, zinc finger protein 423 (ZFP423), nuclear factor I (NFI), and C/EBPa, which are positive regulators of PPARc, were increased, whereas GATA3 transcripts were decreased (Fig. 5B). Contrary to the behavior of the other negative regulators, GATA2 expression was increased by FGF11 overexpression (Fig. 5B). These data could be partly explained by the results of a previous study, in which GATA2 negatively regulates the expression of PPARc in the early stage of adipogenesis, whereas the relationship between GATA2 and PPARc is changed in the mature adipocytes [22]. The transcript analysis of PPARc regulators suggests that FGF11 expression may modulate PPARc transcription by changing the expression of PPARc regulators such as C/EBPa, KLF9, SREBP1, GATA2, GATA3, and KLF2.
To investigate whether intracrine FGF11 activates FGFR during 3T3-L1 preadipocyte differentiation like other secretory FGFs, the expression and phosphorylation of FGFRs in FLAG-FGF11-overexpressed 3T3-L1 cells were monitored. When FLAG-tagged FGF11 was overexpressed by adenoviral expressing system from day 2 to 4 of adipogenesis and in fully differentiated cells, the expression of FGFRs (1-4) and the levels of phosphorylated FGFR 1 and total FGFR were not changed (Fig. 6), which indicates the regulatory mechanism of adipogenesis by FGF11 is different from that of other FGFs.

Discussion
Fibroblast growth factors function as important factors in a variety of biological processes such as embryonic development, tissue maintenance, repair, regeneration, survival, and metabolic regulation. FGF11 belongs to iFGFs group, which show different biochemical and functional characteristics than the typical FGF group. While other iFGFs have received much attention for their roles and molecular mechanisms, FGF11 has not been well studied. However, recent research has revealed novel characteristics of FGF11 that other iFGFs do not share. Here, we found that FGF11 expression exerts a regulatory effect on the adipogenesis of 3T3-L1 cells. This novel function of FGF11 on preadipocyte differentiation was not disclosed in other intracrine FGFs, even though other endocrine or paracrine/autocrine FGFs are implicated in the adipogenesis. FGF1 was suggested as a key human adipogenic factor [23] and has emerged as a potential drug candidate for the treatment of type 2 diabetes mellitus [24,25]. FGF10 is secreted by cultured preadipocytes and plays an important role in adipogenesis in a paracrine or autocrine manner [26] by activating FGFR2b with heparin/heparin sulfate as a cofactor [27]. Endocrine FGFs such as FGF19 and FGF21 act as important metabolic regulators [28]. The incubation of 3T3-L1 with FGF21 stimulates the phosphorylation of signaling molecules and preadipocyte differentiation [29], and FGF21 requires bklotho to form the FGF21-FGFR complex and modulate glucose uptake in 3T3-L1 cells [30]. Since FGF21 has different characteristics from other metabolic regulators, FGF21 gains attention as an agent to treat metabolic disease [31]. To date, most of the research regarding the relationship between FGF and adipogenesis has focused on secretory FGFs. We suggest that endogenous FGF of the intracrine FGF11 can regulate FGFR-independent preadipocyte differentiation (Fig. 6), which is a different regulatory mechanism in adipogenesis from other FGFs.
Fibroblast growth factor 11 has an N-terminal nuclear localization sequence (NLS) [32], but it has been reported that FGF11 is localized in the cytosol as well as the nucleus [5,33]. Therefore, the presence of the NLS in FGF11 isoforms could have a profound impact on their localization and function. Mouse FGF11 isoform 2 does not have the 27 amino acid sequence in the middle region, compared to the two human FGF11 isoforms and mouse FGF11 isoform 2, and the remaining sequences are quite similar. In contrast, only human FGF11 isoform 2 lacks an N-terminal NLS. Unlike FGF11 isoform 2, human FGF11 isoform 1 and mouse FGF11 isoform 1 share high similarity (Fig. S1). Due to these distinct features of FGF11 isoform 2, the roles of the mouse FGF11 isoform 2 might be different from those of the human FGF11 isoform 2. Hence, FGF11 isoform 1 was preferentially selected as the FGF11 isoform for this study.
First, we observed dynamic changes in the expression of FGF11 during adipogenesis. In the MCE phase, FGF11 expression remained at low levels (day 0-2). In the terminal differentiation phase (day 3-6), however, FGF11 expression dramatically increased to approximately twofold higher than the level observed before adipogenic stimulus. To explore whether reduction of FGF11 in the MCE phase is involved in adipogenesis, we overexpressed FGF11 in 3T3-L1 preadipocytes under the MCE phase. Growth-arrested cells were transduced with FGF11-expressing adenoviruses 1 day before differentiation induction, and subsequently, the process of adipocyte differentiation was initiated. Surprisingly, the protein level of exogenous FLAG-tagged FGF11 was extremely low after the stimulation of adipogenesis, even though the protein was overexpressed well at day 0 (Fig. 7). Since FGF11 was overexpressed under the control of the CMV promoter, FGF11 could be regulated post-transcriptionally. The decrease of endogenous FGF11 just after MCE initiation (Fig. 1B) could be forced by a regulatory mechanism. This result raised the possibility that the reduction in FGF11 in the MCE phase is an important regulatory process during adipogenesis.
Further studies will be performed to determine why FGF11 expression is maintained at a low level. When the increase of FGF11 was blocked by FGF11 knockdown in the terminal differentiation phase, the induction of PPARc and C/EBPa, well-known critical adipogenic regulators, was mostly attenuated, resulting in the inhibition of adipogenesis. In contrast, the expression of PPARc and C/EBPa was increased and adipocyte differentiation was enhanced by FGF11 overexpression in the terminal differentiation phase. These results indicate the importance of the increase in FGF11 during the terminal differentiation phase. One day before adipogenic induction, growth-arrested 3T3-L1 preadipocytes were transfected with 50 nM of siFGF11 or siNS. Total RNAs at the indicated days were isolated and subjected to real-time RT/PCR using specific primers as described in Table 1. The values of siNS-treated cells at day 0 were set to 1, and the other values were calculated relative to this value (n = 3, average AE SEM). The significance was assessed using the Mann-Whitney U-test. *P < 0.05 vs siNS at the same time point. †P = 0.05 vs siNS at the same time point. (B) FGF11 was overexpressed using an adenoviral expression system (600 MOI) on day 2 of adipogenesis. Adipocyte differentiation progressed until day 4 of adipogenesis, and total RNAs were prepared for real-time RT/PCR using specific primers. The value of GFP of each was set to 1, and the value of FGF11 of each sample was calculated relative to this value (n = 3, average AE SEM). The significance was assessed using the Mann-Whitney U-test. *P < 0.05 vs GFP each. Taken together, the data suggest that the expression of FGF11 may be a regulatory component involved throughout the entire process of adipogenesis.
Peroxisome proliferator-activated receptor gamma is a major regulator of adipocyte differentiation. Since the expression level of PPARc was lower in FGF11depleted cells than in control cells, we increased the transcriptional activity of PPARc by treatment with a PPARc agonist, rosiglitazone, to determine the involvement of PPARc in the FGF11 effect. The inhibition of adipocyte differentiation by FGF11 knockdown was significantly restored by the addition of rosiglitazone. These results suggest that a low level of PPARc may be a necessary factor for the attenuated adipogenesis induced by FGF11 knockdown. However, adipocyte differentiation was partly recovered. It is likely that the remaining PPARc level is too low for differentiation to be fully recovered. This result coincides with the expression of FABP4 (Fig. 8), which is downstream of PPARc and regulates lipid accumulation [34].
Peroxisome proliferator-activated receptor gamma expression is controlled by multiple positive and negative transcription factors during adipogenesis [21]. Here, we examined the expression levels of several regulators of PPARc expression: C/EBPb, C/EBPa, KLF2/9, ZFP423, NFI, SREBP1, and GATA2/3. We found that several regulators were involved in PPARc expression based on the transcriptional changes of regulators by FGF11 knockdown or overexpression: C/ EBPa, KLF9, SREBP1, GATA2, GATA3, and KLF2. The transcript levels of negative regulators of PPARc  can exert PPARc-dependent regulatory effects on adipogenesis, which means that the FGF11 effect on adipogenesis is attributed to the combined action of these PPARc regulators controlled by FGF11. FGF11 expression could modulate the expression of each PPARc regulator or control a common factor upstream of the PPARc regulators.
In summary, we hypothesized that the expression of FGF11 is regulated precisely during the MCE phase and the terminal differentiation phase. We showed a decrease of PPARc by FGF11 knockdown, an increase of PPARc by FGF11 overexpression, and a recovery of adipogenesis by treatment with a PPARc agonist. These results suggest that the expression of PPARc is a major target of FGF11 during adipogenesis, and the precise regulation of the expression of FGF11 may play an important role in adipogenesis. In addition, we demonstrated that the expression of C/EBPa, KLF9, SREBP1, GATA2, GATA3, and KLF2 was influenced by FGF11 expression, which results in the regulation of PPARc. In conclusion, we suggest that FGF11 is a novel mediator of adipogenesis and indirectly controls the expression of PPARc by modifying the expression of regulators of PPARc expression. The effect of FGF11 on adipogenesis may be an integrated outcome through the expression of many PPARc regulators that are controlled by FGF11.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Fig. S1. Multiple amino acid sequence alignment of mouse and human FGF11 isoforms.