Fibroblast Growth Factor 1: A Key Regulator of Human Adipogenesis

Unless otherwise stated, all chemicals used were obtained from Sigma-Aldrich (Castle Hill, Australia). Goat anti-human polyclonal FGF-1 neutralizing antibody was purchased from R&D Systems (Minneapolis, MN). Mouse anti-human monoclonal FGF-1 antibody (Sigma) was used in Western blotting procedures. The thiazolidinedione (TZD) compound rosiglitazone was supplied by GlaxoSmithKline (Middlesex, U.K.). The FGF inhibitory peptide AcValTyrMetSerProPheNH₂ was prepared by manual synthesis using Fmoc chemistry on Rink Amide MBHA resin. After washing and drying, the peptide was cleaved using TFA/water/triisopropylsilane (95:2.5:2.5) and purified by reverse-phase high-performance liquid chromatography on a Vydac C-18 column using aqueous acetronitrile containing 0.1% formic acid.

Paired omental (intra-abdominal) and subcutaneous adipose tissue biopsies were obtained from 17 male (average age 63.5 years [range 37–82]; average BMI 27.2 kg/m² [21–39]) and 25 female (average age 51.2 years [31–77]; average BMI 30 kg/m² [20.5–47.9]) patients undergoing elective open-abdominal surgery. None of the patients had diabetes or severe systemic illness, and none were taking medications known to affect adipose tissue mass or metabolism. The protocol was approved by the research ethics committees of the Princess Alexandra Hospital and the Queensland University of Technology. All patients gave their written informed consent.

Preadipocytes, adipocytes, and MVECs were obtained from adipose tissue biopsies as previously reported (4,11,12).

Preadipocytes and MVECs were isolated, cultured, and characterized as previously reported (4). Homogeneous cultures of both cell types were grown for up to 8 weeks and maintained at 37°C with 5% CO₂. Cells were used in experimental work at passages 2 and 3. For control and growth factor experiments, preadipocytes were grown in the same serum-containing (10%) medium as the MVECs, containing 90 μg/μl heparin, 0.014 mol/l HEPES, and 0.15% NaHCO₃ (wt/vol), with or without growth factors (all human): β-endothelial cell growth factor (ECGF) (at 10 ng/ml) (Sigma), FGF-1, FGF-2, or IGF-1 (at 1 ng/ml) (R&D Systems). In some experiments, FGF inhibitors (anti–FGF-1 neutralizing antibody; R&D Systems) (10 μg/ml), the synthetic peptide Ac-ValTyrMetSerProPhe-NH₂ (10 μg/ml), and suramin (20 μmol/l) were used in the presence of FGF-1 to determine the effects on both proliferation and differentiation. Human dermal fibroblasts were isolated and cultured as previously described (13). These cells were used as controls in the coculture studies.

For coculture experiments, preadipocytes were seeded at low cell density with homogeneous MVECs (each cell type plated at approximately one-third confluence at opposite ends of the culture dish) and the mixed cultures allowed to grow together for 2–6 weeks in endothelial cell growth medium containing heparin and 10 ng/ml β-ECGF. Controls included preadipocytes plus human skin fibroblast cocultures and preadipocyte-only cultures grown under the same conditions as the MVEC plus preadipocyte cultures.

3T3-L1 fibroblasts were grown to confluence in serum-containing medium consisting of Dulbecco’s modified Eagle’s medium/HAM’s F12 (1:1) (ICN Biomedical Australasia) containing 4.5 g/l glucose, 10% FCS, 100 IU penicillin, 100 μg/ml streptomycin, and 2 mmol/l glutamine. In some experiments FGF-1, FGF-2, or IGF-1 (all at 1 ng/ml) (R&D Systems) were included in the growth medium. Differentiation was initiated by addition of 250 nmol/l dexamethasone, 0.5 mmol/l of 1-methyl-3-isobutylxanthine, 1 μmol/l insulin, and 0.1 μmol/l rosiglitazone. After 2 days, cells were transferred to serum-containing medium containing 0.5 μmol/l insulin and refed twice weekly. Cells were studied at 13–19 days postdifferentiation. To determine the role of endogenous FGF-1 in 3T3-L1 differentiation, cells were treated with anti–FGF-1 neutralizing antibody (10 μg/ml) for 1 week before and throughout the differentiation period and effects assessed by G3PDH activity assay.

As previously reported, confluent preadipocytes at passage 2 were changed to a chemically defined serum-free medium with or without rosiglitazone for induction of differentiation (4,6,12–14). In some experiments, growth factors were included in the differentiation medium at a concentration of 1 ng/ml.

After 21 days in serum-free differentiation medium, adipogenesis was assessed using a Nile Red assay for triacylglycerol accumulation and an enzyme assay measuring G3PDH activity (4,12,13). The Nile Red results are presented as optical density units at 480/540 nm, and G3PDH results are given as activity in milliunits per milligram of protein (Act mU/mg Pr).

Following various treatments, proliferation of preadipocytes was assessed using an MTS proliferation assay (Promega, Madison, WI) as described previously (4).

FGF-1 expression was determined in whole-cell protein isolated from human preadipocytes treated or not treated with FGF-1, human adipose–derived MVECs, 3T3-L1 preadipocytes (all these cell types were maintained in 10% serum-containing medium before protein extraction), and isolated human adipocytes. Cells were harvested in buffer containing 20 mmol/l HEPES, 1 mmol/l EDTA, phosphatase inhibitors (2 mmol/l Na₃VO₄, 1 mmol/l Na₄P₂O₇, and 10 mmol/l NaF), and protease inhibitors (Complete Mini; Roche Diagnostics, Mannheim, Germany). Whole-cell lysates (15 μg) were resolved by SDS-polyacrylamide gel (8%) electrophoresis, transferred to Immobilon-P PVDF membrane (Millipore, Bedford, MA), and probed with mouse anti-human monoclonal FGF-1 antibody (Sigma). Proteins were detected by enhanced chemiluminsence substrate (Amersham, Little Chalfont, U.K.).

The homogenate was passed through a 21-gauge needle 10 times and a 22-gauge needle 4 times. The samples were cleared by centrifugation at 1,000g for 10 min at 4°C. Protein determination was performed using bicinchoninic acid reagent (Pierce, Rockville, IL). Proteins (15 μg) were resolved on SDS-polyacrylamide gels and transferred to Immobilon-P PVDF membranes (Millipore, Bedford, MA). Membranes were blocked with 2% (wt/vol) BSA/PBST overnight at 4°C, incubated with mouse anti-human monoclonal FGF-1 antibody (Sigma), and probed with horseradish peroxidase–linked secondary antibodies (Amersham) for 1 h at room temperature. Proteins were detected by enhanced chemiluminescence substrate (Amersham).

Total RNA was obtained from both human subcutaneous adipose tissue and preadipocytes (at confluence and following 21 days of differentiation). Total RNA was obtained using the TRIzol reagent extraction procedure (GIBCO/BRL) according to the manufacturer’s instructions. RNA was further purified using an RNAeasy column (Qiagen) as described by the manufacturer. An on-column DNase digest was performed as part of the purification protocol. Total RNA (1 mg) was reverse transcribed using random hexamers and reverse transcription reagents from Applied Biosystems (Victoria). The absence of contaminating genomic DNA was verified by including samples to which no reverse transcriptase was added (no reverse transcriptase controls). For PCR, no template controls confirmed the absence of PCR contamination. Fibroblast growth factor receptor (FGFR)-1 to -4 primer sequences were those used by Tartaglia, Fragale, and Battaglia (15). Thermal cycling conditions were as follows: first denaturing step of 94°C for 8 min; 35 cycles of 94°C for 45 s, 60°C for 30 s, and 72°C for 30 s; and a final extension step of 72°C for 10 min. FGFR cDNA (in pcDNA3.1) was a kind gift from Pamela Maher (Scripps Research Institute, La Jolla, CA) and was used as a positive control for each reaction.

Data were analyzed using repeated-measures ANOVA for differences across experimental groups. Student’s t test was used to evaluate the significance of the difference in mean values between different treatments. Data are expressed as means ± SE.

Human MVECs and preadipocytes were isolated, cultured, and characterized as previously described (4). Once phenotype was established, these two cell types were used in coculture experiments to study involvement of MVECs in human adipogenesis.

At confluence following a 2-week coculture period, all cultures were changed to serum-free medium containing known inducers of adipocyte differentiation, including the PPAR-γ ligand rosiglitazone. In preadipocyte-only cultures, in accordance with our previous observations using standard human differentiation techniques (13), lipid accumulation was apparent in a small number of cells by day 7–10, with 5–25% of cells acquiring lipid by 21 days (Fig. 1A). In marked contrast, triacylglycerol accumulation was apparent as early as day 3–5 in the majority of cocultured preadipocytes, with 70–95% of cells accumulating large amounts of triacylglycerol by day 21 (Fig. 1B). Evidence for the specificity of these effects for human preadipocytes came from observations showing that human skin fibroblasts cultured either alone or in coculture with MVECs, under identical conditions as the preadipocytes, did not accumulate lipid when treated with differentiation medium (data not shown).

To confirm that the lipid-containing morphology of these cells was consistent with triacylglycerol accumulation and biochemical differentiation, Nile Red assay for nonpolar lipid and G3PDH activity assays were performed. The results demonstrate that preadipocytes from the MVEC plus preadipocyte coculture had significantly increased G3PDH activity and triacylglycerol accumulation following 21 days in differentiation medium over preadipocytes cultured alone in both regular serum-containing medium (medium 1) or the same medium with added β-ECGF (medium 2; β-ECGF is required for maintenance of endothelial cells in the coculture conditions). This was the case for preadipocytes from both subcutaneous (G3PDH, Fig. 1C; triacylglycerol, Fig. 1D) and omental (G3PDH, Fig. 1E; triacylglycerol, Fig. 1F) adipose tissue depots. While the rise in G3PDH activity in subcutaneous preadipocytes was greater than that seen in cells from omental tissue, the degree of differentiation of omental cells in coculture was significant. To our knowledge, this represents the first time quantifiable and consistent differentiation that has been demonstrated in intraabdominal preadipocytes following prolonged culture in the presence of serum. After treatment with the same differentiation medium as human preadipocytes, G3PDH activity was negligible in human skin fibroblasts cocultured with adipose-derived MVECs (data not shown).

To determine whether the differentiation potential of human preadipocytes could be further enhanced, coculture with MVECs was extended to a period of 6 weeks before induction of differentiation. Following 21 days in differentiation medium, after the coculture period, both G3PDH activity and triacylglycerol accumulation were once again significantly increased in cocultured preadipocytes over that demonstrated in preadipocytes previously maintained as homogeneous cultures in regular serum-containing growth medium (medium 1) (G3PDH, Fig. 2A). Surprisingly, however, both parameters of differentiation were also elevated to the same high levels as those measured in the cocultured cells in preadipocyte-only cultures maintained for 6 weeks in serum-containing medium supplemented with β-ECGF (medium 2) (Fig. 2A). These results suggest that β-ECGF is acting as an adipogenic agent under these conditions. The presence of heparin in medium 2 was required for maximal β-ECGF activity, but heparin, in the absence of growth factor, had no effect on differentiation of human preadipocytes (data not shown).

β-ECGF is a member of the FGF family and is a precursor molecule of FGF-1 (16). To further investigate the role of growth factors in this system, differentiation of preadipocytes was carried out following a 6-week proliferation period with or without FGF-1, FGF-2, or IGF-1 (at 1 ng/ml). Results demonstrated that pretreatment with FGF-1 and FGF-2 but not IGF-1 resulted in significant increases in G3PDH activity in human preadipocytes, relative to controls, with FGF-1 having the greatest effect in subcutaneous preadipocytes (Fig. 2B). The differentiation potential of these cells could be further enhanced if exposed to FGF-1 during both proliferation and differentiation (Fig. 2C). Intriguingly, preadipocytes treated with FGF-1 only during the differentiation period exhibited a trend to a decrease in both G3PDH activity (Fig. 2C) and lipid accumulation following differentiation compared with cells grown in the absence of FGF-1. The time course of action of FGF-1 suggests that its effect is to prime preadipocytes for subsequent differentiation rather than to act as a direct inducer of differentiation.

The proliferation rate of human preadipocytes is increased in response to FGF-1, and this effect can be blocked by addition of a commercially available neutralizing antibody raised against FGF-1 (Fig. 3). To establish more directly whether FGF-1 may represent an adipogenic factor produced by MVECs, we used the same neutralizing anti–FGF-1 antibody to block any endogenous FGF-1 actions in MVEC-conditioned medium. Proliferation assays were carried out and, in agreement with our previous results (4), preadipocyte proliferation was enhanced in response to medium that had been conditioned by 48-h exposure to adipose-derived MVECs from both subcutaneous and omental sites (Fig. 3). Inclusion of the FGF-1 neutralizing antibody at 10 μg/ml to the MVEC-conditioned medium resulted in almost total abrogation of these proliferative effects (Fig. 3).

Significant, but relatively small, increases in G3PDH activity were demonstrated in FGF-1–treated preadipocytes from both subcutaneous and omental adipose tissue differentiated in the absence of PPAR-γ ligand (Fig. 4). However, the maximal adipogenic effect of either coculture with MVECs or treatment with FGF-1 was only observed when cells were differentiated in the presence of the PPAR-γ ligand rosiglitazone (a TZD) (Fig. 4). Differentiation in the presence of the TZD exhibited depot specificity, in agreement with our previous results and those of others (13,14,17), with the effect being greater in subcutaneous than omental preadipocytes (Fig. 4). The requirement for TZD for maximal differentiation suggests that neither the factor(s) produced by MVECs nor exogenous FGF-1 are acting as PPAR-γ ligands. Further evidence for this came from electromobility shift and reporter construct assays, which showed no endogenous PPAR response element binding or activation of PPAR-γ (data not shown) in response to FGF-1 treatment, and TZD-induced PPAR-γ activity was not further enhanced in the presence of FGF-1 (data not shown).

Using a combination of Western blot analysis and immunofluorescence, robust expression of FGF-1 was demonstrated in adipose-derived MVECs and 3T3-L1 cells. In constrast, expression in human preadipocytes, maintained under the same conditions, was low to undetectable (although this does not rule out the possibility of expression at very low levels). Expression in isolated human adipocytes was more variable than in preadipocytes but, if detectable, was always lower than that seen in either 3T3-L1 cells or human MVECs (Fig. 5). Gabrielsson et al. (18) demonstrated FGF-1 expression in the stromovascular fraction of human adipose tissue, which contains both MVECs and preadipocytes. Our findings suggest that the FGF-1 expression in this adipose tissue fraction is predominantly contributed to by the MVEC component.

FGFR expression in human samples was determined using RT-PCR (Fig. 6). In subcutaneous adipose tissue, expression of FGFR-1 and -2 was detected. Similar expression patterns were demonstrated in differentiated human preadipocytes, with R1 and R2 being present, whereas little, if any, expression of R3 and R4 was detected. In contrast, all four FGFRs were found to be expressed in nondifferentiated human preadipocytes (Fig. 6). Western blotting and immunofluorescence techniques also demonstrated expression of FGFR-1 to -4 in human nondifferentiated preadipocytes (data not shown). In 3T3-L1 cells, expression of all four receptors has been observed by RT-PCR or Western blotting (data not shown).

To determine whether the adipogenic effects of FGF-1 can be blocked in vitro, preadipocytes were treated with a combination of FGF-1 and FGF inhibitors and effects on both proliferation and differentiation assessed. The inhibitors used included the synthetic peptide Ac-ValTyrMetSerProPhe-NH₂ (peptide 1), which specifically binds to, and blocks, FGFRs (19); suramin (a polysulfonated binaphthyl urea that inhibits complex formation among FGF, heparin, and cognate receptor) (20); and a neutralizing antibody specific for FGF-1 (21). All of these compounds are reported to inhibit the mitogenic activity of FGF-1 (19–21). All three compounds demonstrated significant inhibition of FGF-1 proliferative effects (Fig. 3, anti–FGF-1; Fig. 7A, peptide 1 and 7B, suramin), while suramin and anti–FGF-1 neutralizing antibody also demonstrated inhibition of differentiation (Fig. 7C, suramin and 7D, anti–FGF-1). These data support the role of FGF-1 as an adipogenic agent and show that the adipogenic effects can be blocked in vitro.

Investigation of adipogenic effects of growth factors was also carried out using the 3T3-L1 murine model of adipogenesis which, unlike human preadipocytes, display rapid growth and high differentiation capacity. Exposure of 3T3-L1 fibroblasts to FGF-1, FGF-2, or IGF-1, before differentiation, caused no significant increases in either growth, G3PDH activity, or triacylglycerol accumulation (data not shown) compared with the already high levels of both parameters seen in the absence of growth factors (controls). For the FGF-1 and -2 treatments, this was in direct contrast to results in human preadipocytes. Western blot analysis demonstrated expression of FGF-1 in 3T3-L1 fibroblasts but not in human preadipocytes (Fig. 5). Also, treatment of 3T3-L1 cells with anti–FGF-1 neutralizing antibody for 1 week before and throughout differentiation resulted in a significant decrease in G3PDH activity following differentiation compared with controls (Fig. 7E). Hence the expression of FGF-1 by the embryonically derived 3T3-L1 cells may underpin their capacity to differentiate more readily in vitro than human preadipocytes.

In contrast to 3T3-L1 cells, human preadipocytes are slow growing and differentiate inefficiently in vitro. Human preadipocytes also have more defined and specific requirements for induction of differentiation than 3T3-L1 cells, including an absolute requirement for serum-free conditions (6). Following treatment with FGF-1, however, we observe that human preadipocytes display rapid growth characteristics and a high capacity for differentiation that are more similar to 3T3-L1 cells than to non–FGF-1–treated preadipocytes. To determine whether growth factor treatment of human preadipocytes renders them capable of differentiating in the presence of serum, cells were proliferated in the presence of FGF-1, FGF-2, or IGF-1 and induced to differentiate in serum-containing medium. Using a 3T3-L1 differentiation protocol consisting of serum-containing medium (10% vol/vol), insulin, and, for the first 3 days, glucocorticoid and IBMX, FGF-1–treated preadipocytes from both subcutaneous and omental sites developed both morphological (Fig. 8A) and biochemical (Fig. 8B) features of a differentiated phenotype. The effect of FGF-2 treatment during growth also lead to an increased capacity for differentiation in the presence of serum, but this effect, although similar to FGF-1 treatment for omental preadipocytes, was much less in subcutaneous preadipocytes than that measured in FGF-1–treated cells. Preadipocytes treated with IGF-1 were similar in differentiation capacity to controls, with no lipid accumulation and negligible G3PDH activity (Fig. 8B). In our experience, this is the first time preadipocyte differentiation has been observed in cells maintained for an extensive period in, and subsequently differentiated in, the presence of serum.

Knowledge of the cellular mechanisms implicated in the overdevelopment of adipose tissue is a critical issue in the design of pharmaceutical strategies aimed at lessening nutritional obesity. We have previously reported that adipose-derived MVECs produce factors that have a mitogenic effect on human preadipocytes. We now demonstrate that these endothelial cells also secrete factors that significantly increase the differentiation potential of preadipocytes from both subcutaneous and omental adipose tissue depots (subcutaneous more than omental). We present data that indicate that much of the adipogenic effect of MVECs can be attributed to FGF-1 secretion.

FGF-1 is a member of a large family (currently >22 identified) of related growth factors (22). All FGFs act via a dual receptor system comprising four high-affinity tyrosine kinase receptors (FGFR-1 to -4) and a class of low-affinity receptors (the heparan sulfate proteoglycans) (23,24). Multiple layers of regulation control the activity of FGFs, including tissue-specific expression of ligands and receptors, modulation of binding specificity by alternative splicing of FGFRs and by sequence differences between FGFs and FGFRs, and modulation of binding affinity and specificity by heparan sulfate molecules with specific patterns of sulfation (23,25–27). In addition to localization in the plasma membrane, FGFRs are also expressed within the nuclear envelope and matrix (28). Signal transduction in response to FGFs occurs through receptor dimerization and complex formation with heparan sulfate proteoglycans (26). Subsequent phosphorylation at multiple sites on the intracellular domain of the FGFR initiates recruitment and/or phosphorylation of multiple downstream signal transduction molecules and pathways (28,29).

This study demonstrates a novel role of FGF-1 as a key human adipogenic molecule. Culture in the presence of FGF-1 from the time of preadipocyte isolation appears to prime the cells for differentiation. This priming effect is made much more apparent if the cells are subsequently differentiated in the presence of a PPAR-γ ligand, the TZD rosiglitazone. Our results suggest that FGF-1 is not acting as, or promoting production of, a PPAR-γ ligand. It is possible, however, that FGF-1 treatment of preadipocytes may be acting on the PPAR-γ system in other ways. These could include increased expression of PPAR-γ, as well as members of the CCAAT enhancer binding protein family of transcription factors, before induction of differentiation. This would be consistent with the finding that FGF-1 primes preadipocytes for TZD-induced differentiation.

The adipogenic effects of MVECs and FGF-1 were specific for cells obtained from the stromal-vascular compartment of adipose tissue because human skin fibroblasts maintained under identical conditions did not accumulate cytoplasmic lipid and no G3PDH activity was measured in these cells. This is consistent with the premise that preadipocytes are a unique type of mesenchymal cell that are distinctly different from dermal fibroblasts and capable of expressing a developmental commitment that, at terminal differentiation, results in a typical mature fat cell (5).

The proliferative effect of FGF-1 on human preadipocytes was the same in cells from subcutaneous and intra-abdominal sites (data not shown). Human preadipocytes maintained in the presence of FGF-1 also retained their differentiation capacity after prolonged culture involving repeated passaging. FGF-1 also induced differentiation of preadipocytes to a similar degree irrespective of depot, when differentiation was performed in the absence of TZD. In contrast, and consistent with our previous reports and those of others (13,14,30,31), differentiation was greater in subcutaneous compared with omental preadipocytes when a TZD was included. Taken together, these observations suggest that subcutaneous and omental preadipocytes may share similar capacity to differentiate in the absence of TZDs and that subcutaneous preadipocytes are more responsive to the adipogenic effects of TZDs than their omental counterparts. This difference may contribute to the observed remodeling of adipose mass, with increased subcutaneous and decreased omental fat, in humans receiving TZDs (32,33). It is also noteworthy that before this work, omental preadipocytes that have been cultured in the presence of serum have proved refractory to differentiation in vitro (13,14). We have now shown that exposure of these cells to FGF-1 during proliferation results in marked differentiation, thus overcoming the serum-induced inhibition of differentiation. Indeed, in FGF-1–treated omental preadipocytes, the levels of G3PDH activity and intracellular lipid accumulation are comparable to those we have previously reported in subcutaneous preadipocytes (13).

The murine 3T3-L1 cell line is a widely used model system for the study of differentiation. In contrast to human preadipocytes, these cells differentiate with high efficiency in the presence of serum and after high passage number (typically up to 20 in our experience). We report that 3T3-L1 cells, unlike human preadipocytes, express high levels of FGF-1 and that exogenous FGF-1 does not further increase proliferation or differentiation of these cells. We also demonstrate that human preadipocytes treated with FGF-1 resemble 3T3-L1 cells in terms of rapid growth and very high capacity for differentiation. Taken together, these results suggest that an important difference between 3T3-L1 and human preadipocytes is higher endogenous expression of FGF-1 by the 3T3-L1 cells. The key role of FGF-1 in promotion of adipogenesis is demonstrated by the inhibitory action of neutralizing anti–FGF-1 antibodies on both human and 3T3-L1 differentiation.

The role of FGFs in adipogenesis is controversial. Although both positive and negative effects have been reported for different family members, including FGF-1 and -2 (34), in general these growth factors are reported to have a mitogenic effect on preadipocytes while having no effect or even potently inhibiting their differentiation (35–42). In contrast to previous studies, where exposure to FGF-1 was restricted to relatively short periods during either growth or differentiation, we have cultured cells in the presence of FGF-1 for sustained periods (1–2 months). In addition, previous studies have not addressed the effects of sustained culturing in the presence of FGF-1 during the growth period, before the induction of differentiation. We suggest that these differences in protocol underlie the differences in experimental observations. Consistent with this, when preadipocytes were exposed to FGF-1 only during differentiation, as in earlier studies, we observed no significant effect on either lipid accumulation or G3PDH activity, with a trend toward a decrease in these markers of differentiation. Taken together, our results suggest that sustained incubation with FGF-1 during the growth period may “prime” the preadipocytes for subsequent differentiation. FGF-2 treatment of human preadipocytes also resulted in increased adipocyte differentiation, suggesting that this feature may be common to members of the FGF family, although FGF-1 was consistently the more potent adipogenic agent, particularly in cells from subcutaneous depots. Further work is aimed at investigation of the role of a range of growth factors in human adipogenesis, including that of FGF-2 and other members of the FGF family.

Evidence that members of the FGF family may play an in vivo role in adipogenesis is provided by studies demonstrating de novo adipogenesis in mice at the site of implanted FGF-2–impregnated microspheres (43,44). Also, a role for FGFs in human adipose tissue metabolism is suggested by the observation that preadipocytes from massively obese individuals have much higher expression of FGF-2 than the same cells from lean counterparts (45). The fact that human adipose tissue expresses members of the FGF family, including their cognate receptors, provides further support for a metabolic role of these growth factors in human adipose tissue (18,40).

In summary, in this report we describe an important role of adipose-derived MVECs in the paracrine regulation of human adipose tissue expansion and demonstrate a novel role of FGF-1 in human adipogenesis. Our data indicating that disruption of FGF-1 signaling at multiple biochemical steps inhibits preadipocyte replication and/or differentiation support the key adipogenic role of FGF-1 and represent an important area for further research. Future studies will be aimed at establishing the molecular mechanisms by which FGF-1 exerts its adipogenic effects. These data suggest that antagonists of FGF-1 or of FGFRs may present new avenues for the development of novel pharmaceutical antiobesity strategies targeting adipose cell number.

This work was supported by the Diabetes Australia Research Trust, the National Heart Foundation of Australia, the Wellcome Trust, and the Princess Alexandra Hospital Foundation. L.H. holds a Smart State Fellowship.

The authors give special thanks to the volunteers and surgeons at the Princess Alexandra Hospital for providing the tissue for study.