Fibroblast growth factor receptor is a mechanistic link between visceral adiposity and cancer

Abstract

Epidemiological evidence implicates excess adipose tissue in increasing cancer risk. Despite a steeply rising global prevalence of obesity, how adiposity contributes to transformation (stage a non-tumorigenic cell undergoes to become malignant) is unknown. To determine the factors in adipose tissue that stimulate transformation, we used a novel ex vivo system of visceral adipose tissue (VAT)-condition medium-stimulated epithelial cell growth in soft agar. To extend this system in vivo, we used a murine lipectomy model of ultraviolet light B-induced, VAT-promoted skin tumor formation. We found that VAT from mice and obese human donors stimulated growth in soft agar of non-tumorigenic epithelial cells. The difference in VAT activity was associated with fibroblast growth factor-2 (FGF2) levels. Moreover, human and mouse VAT failed to stimulate growth in soft of agar in cells deficient in FGFR-1 (FGF2 receptor). We also demonstrated that circulating levels of FGF2 were associated with non-melanoma tumor formation in vivo. These data implicate FGF2 as a major factor VAT releases to transform epithelial cells-a novel, potential pathway of VAT-enhanced tumorigenesis. Strategies designed to deplete VAT stores of FGF2 or inhibit FGFR-1 in abdominally obese individuals may be important cancer prevention strategies as well as adjuvant therapies for improving outcomes.

Figure 1

MFTF stimulates JB6 P⁺ cell transformation. SKH-1 mice (n=5/group) were fed either a HFD or LFD for 4 weeks. Visceral (parametrial and epididymal) adipose tissue was removed to make a filtered conditioned medium (MFTF). (a) Percentage of clones growing in soft agar (% colony formation) significantly increases in JB6 P⁺ cells cultured with MFTF compared no treatment (control; Cont). No significant change in the percentage of colony formation in soft agar is observed in JB6 P⁻ cells cultured with MFTF. (b) JB6 P⁺ colonies growing in soft agar with HFD MFTF is significantly inhibited with proteinase K, but not with lipase, RNase A or DNase A. (c) JB6 P⁺ colonies growing in soft agar decrease as MFTF is exposed to increasing temperatures for 30 min prior to treating the cells in agar. (d) Protein Profiler angiogenesis array of fat tissue filtrates of LFD-fed mice (top panel) and HFD-fed mice (bottom panel). HFD selectively upregulated protein levels of several key adipokines, hormones and growth factors in MFTF, versus those seen in MFTF of LFD-fed mice (boxed proteins). Dot intensity was analyzed by ‘Image J’ software. Data are labeled as the percent of the control (reference) dots located in the upper left hand corner of the arrays. Data are presented as mean±s.d. of values from triplicate. Statistical significance was determined using a one-way ANOVA (**P<0.01, ***P<0.001).

Figure 2

Lipectomy reduces transformation-stimulating serum growth factors in HFD-fed mice. (a) SKH-1 mice were fed either a HFD or LFD for 2 weeks, and half the mice had their parametrial adipose tissue removed or received a sham operation (n=20/group). After 33 weeks of UVB exposure, serum was isolated. Protein Profiler angiogenesis array of pooled sera of HFD-fed sham-operated mice (top panel) and HFD-fed lipectomized mice (bottom panel). HFD-fed mice that had the surgical removal of parametrial adipose tissue showed a decrease in several pro-inflammatory proteins in the circulation. Boxed proteins were those found to be both reduced with lipectomy and induced in the MFTF with HFD (Figure 1d). Dot intensity was analyzed by ‘ImageJ’ software. Data are labeled as the percent of the control (reference) dots located in the upper left hand corner of the arrays. (b) Proteins found in a were tested for their transforming activity in the soft agar assay. HGF (10 ng/ml) and FGF2 (10 ng/ml) significantly stimulated colony formation in soft agar. 12-O-tetradecanoylphorbol-13-acetate (TPA, 10 ng/ml) was used as a positive control. (c) HGF and FGF2 levels in MFTF from Figures 1a and d were quantified by ELISA. HFD feeding increases the levels of both HGF and FGF2. (d) SKH-1 mice were treated as described in a. Serum was isolated and analyzed for HGF and FGF2 by ELISA. (e) Dose response of HGF and FGF2 on JB6 P⁺ cell transformation. Data are presented as mean±s.d. of values from triplicates and statistical significance was determined using a one-way ANOVA (b–e) followed by a Tukey’s test for multiple comparisons (d) (*P<0.05, **P<0.01, ***P<0.001).

Figure 3

Knockout of FGFR-1 in JB6 P⁺ cells inhibits the effect of mouse fat tissue filtrate on transformation in vitro. (a) HFD MFTF-stimulated JB6 P⁺ cell transformation is partially dependent on FGF2 signaling through the fibroblast growth factor receptor-1 (FGFR-1). Cells were treated with a FGFR-1 neutralizing antibody (Ab) (2 μg/ml) and then treated with MFTF. Growth in soft agar was measured after 14 days. (b) FGFR-1 immunofluorescence (red) in WT JB6 P⁺ cells and Fgfr-1^(−/−) JB6 P⁺ cells. Specific staining of FGFR-1 is localized to the cell membrane in the WT cells as indicated by white arrows. Membrane FGFR-1 is absent in the KO cells. Images were taken at × 40 magnification. Details of generation of the FGFR-1 KO are in Supplementary Figure 2. (c) JB6 P⁺ cells deficient in FGFR-1 fail to grow in soft agar above baseline when cultured with FGF2 and MFTF in soft agar. Percentage of clones growing in soft agar (% colony formation) significantly increases in JB6 P⁺ cells deficient in FGFR-1 cultured with HGF compared no treatment (Untx). (d) Nude mice were subcutaneously inoculated with either WT or FgfR-1 KO JB6 P⁺ cells. A JB6 P⁺ transformed clone (WT-Tr) was injected as a positive control. The following day, FGF2 (200 mg/kg) or vehicle was injected i.p. once per day for 7 consecutive days. Photos show s.c. carcinomas induced by FGF2. (e) Growth rates of s.c. tumors formed by WT or FgfR-1^(−/−) (KO) JB6 P⁺ cells injected into nude mice (n=5) that were either injected with saline (vehicle; Veh) or FGF2. A JB6 P⁺ transformed clone (WT-Tr) was injected as a positive control. The tumor was monitored everyday and tumor volume was recorded on days 5, 12 and 20. Volume of the tumor was calculated using the formula: V=length × width² × 0.5. Tumors from FGF2-treated mice inoculated with WT cells are compared to tumors from Veh-treated mice inoculated with WT cells (P<0.01 at 12 days and P<0.001 at 20 days), and tumors from Veh-treated mice inoculated with WT-Tr cells (P<0.01 at 12 days and P<0.05 at 20 days). Data are presented as mean±s.d. of values from triplicates and statistical significance was determined using a one-way ANOVA followed by a Tukey’s test for multiple comparisons (a) (*P<0.05, **P<0.001, ***P<0.0001).

Figure 4

cMYC activity is required for optimal MFTF-transforming capacity and cMYC protein is stably overexpressed in transformed cells. (a) JB6 P⁺ cells were treated with FGF2 (2.5 ng/ml) for indicated time intervals and protein expression was analyzed by western blot. Phosphorylation of Erk1 and mTOR is upregulated and cMyc increases in a time-dependent manner. Actin was used as a loading control. (b) WT JB6 P⁺ cells and Fgfr-1^(−/−) JB6 P⁺ cells were treated with MFTF (150 μg/ml) for 8 h. Actin was used as a loading control. At this dose and time of MFTF, there was no change in ERK and mTOR activity. cMYC was induced in WT cells but not KO cells. (c) Inhibition of cMyc, mTOR and Erk1 activity by pharmacological inhibitors significantly attenuates MFTF-stimulated JB6 P⁺ colony formation in agar. Data are presented as mean±s.d. of values from triplicates and statistical significance was determined using a one-way ANOVA (*P<0.05, **P<0.001, ***P<0.0001). (d) cMYC immunofluorescence (red) in JB6 P⁺ cells and JB6 P⁺ cells transformed with MFTF. MFTF-transformed JB6 P⁺ cells have a higher expression of cMYC compared with non-transformed cells. Transformed JB6 P⁺ cells are obtained by isolating colonies from soft agar and growing the cells in liquid culture. These cells were passaged several times for 28 days and maintained a high expression of cMYC. Images were taken at × 40 magnification.

Figure 5

Transforming activity of HuFTF is associated with FGF2 levels and is dependent on FGFR-1. Fat tissue filtrates were made from human visceral adipose tissue (HuFTF) from four donors undergoing hysterectomy. (a) Percentage of clones growing in soft agar (% colony formation) significantly increases in JB6 P⁺ and NMuMG cells cultured with HuFTF from donors 1, 2 and 4 compared no treatment (control; Cont). (b) HuFTF with higher concentrations of FGF2 are more potent at stimulating cell transformation. (c) Immunodepletion of FGF2 in HuFTF significantly attenuates JB6 P⁺ colony formation in soft agar. (d) JB6 P⁺ cells deficient in FGFR-1 fail to grow in soft agar above baseline when cultured with FGF2 and HuFTF in soft agar. Percentage of clones growing in soft agar (% colony formation) significantly increases in JB6 P⁺ cells deficient in FGFR-1 cultured with HGF compared no treatment (Untx). Data are presented as mean±s.d. of values from triplicates and statistical significance was determined using a one-way ANOVA (*P<0.05, **P<0.001, ***P<0.0001).