Fibroblast activation protein restrains adipogenic differentiation and regulates matrix-mediated mTOR signaling

Abstract

Obesity is a risk factor for multiple diseases, including diabetes, cardiovascular disease, and cancer. Within obese adipose tissue, multiple factors contribute to creating a disease-promoting environment, including metabolic dysfunction, inflammation, and fibrosis. Recent evidence points to fibrotic responses, particularly extracellular matrix remodeling, in playing a highly functional role in the pathogenesis of obesity. Fibroblast activation protein plays an essential role in remodeling collagen-rich matrices in the context of fibrosis and cancer. We observed that FAP-null mice have increased weight compared to wild-type controls, and so investigated the role of FAP in regulating diet-induced obesity. Using genetically engineered mouse models and in-vitro cell-derived matrices, we demonstrate that FAP expression by pre-adipocytes restrains adipogenic differentiation. We further show that FAP-mediated matrix remodeling alters lipid metabolism in part by regulating mTOR signaling. The impact of FAP on adipogenic differentiation and mTOR signaling together confers resistance to diet-induced obesity. The critical role of ECM remodeling in regulating obesity offers new potential targets for therapy.

Keywords: Adipogenesis; Collagen; Extracellular matrix; Fibroblast activation protein; Obesity; mTOR.

Figure 1:. FAP^−/− mice display enhanced diet-induced weight gain with minimal changes to systemic metabolism.

A) Timeline for diet-induced obesity model. B) Total body mass at time of euthanasia, 30 weeks of age (N=23-27 mice/group). C) Mass of total subcutaneous fat at 30 weeks of age (N=14-17 mice/group). D) Mass of total abdominal fat at 30 weeks of age (N=3-5 mice/group) E) Blood glucose levels following bolus insulin injection (1U/kg body mass; N=4-6 mice/group). F) Blood glucose levels following bolus glucose injection (p<0.05 between FAP^+/+ LFD and FAP^+/+ HFD by area under the curve; N=4-6 mice/group) G) Serum triglyceride levels at 30 weeks of age (N=6-10 mice/group). Statistical analysis by two-way ANOVA.

Figure 2:. FAP^−/− mice display enhanced adipocyte hypertrophy and obesity-related gene expression.

A) Subcutaneous fat adipocyte hypertrophy measured in H&E stained sections (scale bar =200μm; N=6-10 mice/group, 10 images/mouse). B) GSEA hallmark pathways that are significantly altered by FAP deletion (regardless of diet). C) Heatmap of RNA levels of all collagen genes as measured by RNAseq (N=2-3 mice/group). Statistical analysis by two-way ANOVA.

Figure 3:. FAP is down-regulated on mature adipocytes.

A) ) qRT-PCR analysis of expression of the preadipocyte marker Zfp423 in freshly isolated murine ASCs (FAP^+/+ and FAP^−/−). B) qRT-PCR analysis of FAP expression on primary murine ASCs with and without adipogenic differentiation stimuli. C) Microarray analysis [18] of FAP expression in human adipose progenitors (CD45−CD31−CD34+) and mature (buoyant) adipocytes. D) qRT-PCR analysis of FAP expression on a human pre-adipocyte cell line with and without adipogenic differentiation stimuli. For all of the above: statistical analysis by two-tailed t-test. Each point represents cells harvested from an individual mouse/patient.

Figure 4:. FAP restricts adipogenesis.

A) Experimental schematic of in vitro differentiation. B) qRT-PCR analysis of PPARγ expression in differentiated murine ASCs (FAP^+/+ and FAP^−/−). C) ORO stain for lipid accumulation in differentiated ASCs (scale bar =100μm). D) qRT-PCR analysis of PPARγ expression in differentiated murine ASCs following acute knockdown of FAP via lentiviral shRNA. E) ORO stain for lipid accumulation in ASCs differentiated following shRNA treatment. For all of the above: statistical analysis by two-tailed ratio paired t-test on raw values. Each point represents cells harvested from an individual mouse F) Flow cytometry for surface FAP expression 48 hours post-transduction with lentiviral shRNA. One representative graph of three independent experiments.

Figure 5:. FAP deletion promotes accumulation of collagen in non-fibrillar forms.

A) Total collagen measured by aniline blue in subcutaneous fat (N=3-7 mice/group, 3 images/mouse). Scale bars=200 μm; statistical analysis by two-way ANOVA. B) Fibrillar collagen measured by SFIG in subcutaneous fat (N=3-7 mice/group, 5 images/mouse). Scale bars=200 μm; statistical analysis by two-way ANOVA. C) Fibrillar collagen imaged by picrosirius red stain under circular polarized light, where thin fibers appear green, intermediate fibers red, and thick fibers yellow (N=3-7 mice/group, 5 images/mouse). Scale bars=200 μm; statistical analysis by chi-square test.

Figure 6:. FAP deletion reduces collagen fibrillogenesis.

A) Experimental schematic for CDM generation B) Total collagen in CDMs measured by picrosirius red dye binding colorimetric assay. C) Fibrillar collagen in CDMs measured by SFIG D) qRT-PCR analysis of Collagen I, E) Collagen III, F) Collagen XI, and G) Lysyl oxidase expression in CDM-laying ASCs. For all of the above: statistical analysis by two-tailed t-test. Each point represents cells harvested from an individual mouse.

Figure 7:. Matrix laid by FAP^−/− ASCs enhances lipid accumulation of pre-adipocytes.

A) Experimental design for testing the effects of matrix on pre-adipocyte lipid accumulation. B) qRT-PCR analysis of FAP expression on pre-adipocytes with and without differentiation. C) qRT-PCR analysis of PPARγ expression following differentiation of pre-adipocytes on FAP^+/+ and FAP^−/− CDMs. D) ORO stain for lipid accumulation in pre-adipocytes following differentiation on FAP^+/+ and FAP^−/− CDMs (scale bars =100μm). E) ORO stain for lipid accumulation in pre-adipocytes in the presence or absence of exogenous lipid, on FAP^+/+ and FAP^−/− CDMs. For all of the above: statistical analysis by two-tailed ratio paired t-test. Each point represents CDMs generated from an individual mouse.

Figure 8:. Matrix laid by FAP^−/− ASCs enhances lipid accumulation of pre-adipocytes via FAK and mTOR pathways.

A) Timeline for various inhibitor treatments during preadipocyte differentiation. B) Immunoblot for phospho- vs. total S6K in pre-adipocytes on day four of differentiation on FAP^+/+ and FAP^−/− CDMs, with and without 8 hour Rapamycin treatment. C) Immunoblot for phospho- vs. total FAK and D) S6K in on day six of differentiation on FAP^+/+ and FAP^−/− CDMs, with and without FAK inhibition for two days. For all blots, density of total and phospho bands was first normalized to β-actin loading control before calculating phospho/total ratio. Statistical analysis by two-tailed ratio paired t-test. E) qPCR analysis of mTOR targets ACLY, FAS, and ChREBPβ in pre-adipocytes differentiated on FAP^+/+ and FAP^−/− CDMs. Statistical analysis by two-tailed t-test. F) ORO stain for lipid accumulation on day six of differentiation on FAP^+/+ and FAP^−/− CDMs, with FAK or mTOR inhibition compared to vehicle control. Statistical analysis by two-tailed ratio paired t-test. Each point represents CDMs generated from an individual mouse.