A PPARγ transcriptional cascade directs adipose progenitor cell-niche interaction and niche expansion

Abstract

Adipose progenitor cells (APCs) reside in a vascular niche, located within the perivascular compartment of adipose tissue blood vessels. Yet, the signals and mechanisms that govern adipose vascular niche formation and APC niche interaction are unknown. Here we show that the assembly and maintenance of the adipose vascular niche is controlled by PPARγ acting within APCs. PPARγ triggers a molecular hierarchy that induces vascular sprouting, APC vessel niche affinity and APC vessel occupancy. Mechanistically, PPARγ transcriptionally activates PDGFRβ and VEGF. APC expression and activation of PDGFRβ promotes the recruitment and retention of APCs to the niche. Pharmacologically, targeting PDGFRβ disrupts APC niche contact thus blocking adipose tissue expansion. Moreover, enhanced APC expression of VEGF stimulates endothelial cell proliferation and expands the adipose niche. Consequently, APC niche communication and retention are boosted by VEGF thereby impairing adipogenesis. Our data indicate that APCs direct adipose tissue niche expansion via a PPARγ-initiated PDGFRβ and VEGF transcriptional axis.

Figure 1. PPARγ regulates APC–blood vessel residency.

(a) Illustration of genetic alleles used to generate: AdipoTrak (AT)-control (PPARγ^tTA; TRE-H2B-GFP; TRE-Cre); Sox-PPARγ-LOF (Sox2-Cre; PPARγ^f/tTA; TRE-H2B-GFP); AT-PPARγ-LOF (PPARγ^f/tTA; TRE-Cre; TRE-H2B-GFP); AT-PPARγ-GOF (PPARγ^tTA; TRE-H2B-GFP; TRE-PPARγ); and AT-PPARγ-Rescue (PPARγ^f/tTA; TRE-PPARγ; TRE-Cre; TRE-H2B-GFP). Experiments were performed three times on six mice per group. (b) Representative H&E images from mice described in a. (c) Representative GFP images of freshly isolated subcutaneous IGW depots from mice described in a. Scale bar 10 mm. (d) Representative images of CD31 (endothelial marker) and SMA (mural cell marker) and APC-GFP immunostaining from subcutaneous IGW depots from mice described in a. (e) Representative images of SVPs from subcutaneous adipose depots from mice described in a. Locality of APCs was assessed by GFP fluorescence 12 h after isolation. DAPI was used to visual nuclei and cell number (n=9). Scale bars 100 μm.

Figure 2. PPARγ is required for WAT niche expansion.

(a) Representative sections from AT-control, Sox-PPARγ-LOF, AT-PPARγ-LOF, AT-PPARγ-GOF and AT-PPARγ-Rescue stained for endothelial marker CD31. (b,c) Quantitative RT–PCR analysis of endothelial and mural cell (nichegenic) markers (b) and vasculogenic genes (c) from total SV cells isolated from the mice described in a. (d–g) Representative images of vascular sprouts of subcutaneous IGW explants from the mice described in a, d. Sprout length (e), branching points (f) and GFP progenitor cell occupancy (g) were quantified. Scale bars 100 μm. Data are means±s.e.m. Experiments were performed three times on eight mice per group. *P<0.05 mutant compared to control levels unpaired t-test, two-tailed.

Figure 3. PPARγ transcriptionally controls WAT niche expansion and APC niche interaction.

(a) Microarray heat map of AT-GFP+ (adipose progenitor) cells isolated from 2-month-old AT-control and AT-PPARγ-LOF mice (n=6 mice per group in triplicate). (b) GFP− and GFP+ cells were FACS isolated from AT-GFP control mice and adipocytes were isolated by floatation. PDGFRβ mRNA expression was measured. (c) GFP+ cells were FACS isolated from AT-Control, Sox-PPARγ-LOF, AT-PPARγ-LOF, SMA-PPARγ-LOF and AT-PPARγ-GOF and AT-PPARγ-Rescue or SV cells were isolated from SMA-PPARγ-GOF and PDGFRβ mRNA expression was assessed. (d) Control, Sox-PPARγ-LOF, AT-PPARγ-LOF, AT-PPARγ-GOF and AT-PPARγ-Rescue (n=10 per group) were administered normal chow or rosiglitazone (0.0075% diet) for 7 days. Subsequently, GFP+ cells were FACS isolated and PDGFRβ mRNA expression was measured. (e) GFP+ cells were FACS isolated from AT-GFP control mice (n=8) and treated with the denoted concentrations of rosiglitazone for 4 h. mRNA expression of denoted genes were measured. (f) GFP+ cells isolated from AT-control mice (n=8) were pretreated with cyclohexamide (10 μg ml⁻¹) for 15 min and then treated with 1 μM rosiglitazone for 4 h and PDGFRβ mRNA expression was measured. (g) GFP+ cells were FACS isolated from AT-control mice and cultured. Cells were then treated with vehicle (dimethylsulfoxide (DMSO)) or 1 μM rosiglitazone for 4 h and then ChIP-qPCR analysis was performed to assess PPARγ occupancy. (h) Subcutaneous IGW explants were excised from AT-control mice and encased in Matrigel. Explants were treated with vehicle (5% DMSO), PDGF-B (10 ng ml⁻¹), SU16F (5 μM) or both for 5 days and vascular sprouting was assessed. Scale bar 100 μm. Data are expressed as means±s.e.m. *P<0.05 unpaired t-test, two tailed: GFP+ and adipocytes compared to GFP− SV cells. **P<0.05 unpaired t-test, two tailed: mutants compared to AT-control mice. #P<0.01 unpaired t-test, two tailed: rosiglitazone treated compared to chow treated. §P<0.01 unpaired t-test, two tailed: vehicle compared to IgG control.

Figure 4. PDGFRβ regulates APC niche interaction and WAT niche expansion.

(a) Illustration of genetic alleles used to generate AT- and SMA-Cre^ERT2-PDGFRβ loss (AT-PDGFRβ-LOF or SMA-PDGFRβ-LOF) and constitutively active PDGFRβ (AT-PDGFRβ-CA or SMA-PDGFRβ-CA) mice. Experiments were performed three times on 10 mice per group. (b) Representative H&E images of IGW adipose depots from mice described in a. (c) Representative histological sections from mice described in a stained for GFP, CD31 and SMA. (d) Representative images of SVPs isolated from mice described in a. (e) Representative images of vascular sprouts of subcutaneous IGW explants from the mice described in a. (f,g) Sprout length and branching points were quantified from explants described in e. (h) mRNA expression of vasculogenic genes from mice described in a. Data are expressed as means±s.e.m. Scale bars 100 μm. *P<0.01 unpaired t-test, two tailed: mutant compared to control levels. #P<0.001 unpaired t-test, two tailed: mutant PDGF-B treated compared to control PDGF-B treated.

Figure 5. PDGFRβ restores APC niche interaction and expansion in AT-PPARγ-LOF mice.

(a) Illustration of genetic alleles used to generate AT-control, AT-PPARγ-LOF and AT-PPARγ-LOF-PDGFRβ-CA mice. At 2 months of age, mice were analysed. Experiments were performed three times on 10 mice per group. (b) Representative images of subcutaneous IGW depots stained for CD31, SMA and APC-GFP from mice described in a. (c) Representative image of SVPs isolated from mice described in a and visualized for GFP locality. (d) Representative images of vascular sprouts from subcutaneous IGW WAT explants excised from mice described in a. (e) Quantification of APC-GFP number per 100-micron SVP. (f) Quantification of sprout length of explants described in d. (g) Quantitative RT–PCR analysis of vasculogenic gene expression from mice described in a. Data are expressed as means±s.e.m. Scale bars 100 μm. *P<0.05 unpaired t-test, two tailed: mutant compared to control levels.

Figure 6. Pharmacologically blocking PDGFRβ disrupts APC niche interaction.

(a,b) One-month-old AT-GFP male mice were administered vehicle (5% dimethylsulfoxide (DMSO)) or imatinib (50 μg per mouse) by IP four times a week for 4 weeks. Experiments were performed three times on 8 mice per group. Body weight (a) and food intake (b) were measured. (c) Fat content of mice described in a before and end of treatment regime. (d,e) Representative adipose tissue images (d) and weights (e) from mice described in a. (f) Representative images of H&E staining from subcutaneous IGW depots from mice described in a. Scale bar 100 μm. (g) Sections from subcutaneous IGW depots from mice described in a were stained with CD31 and SMA and visualized for AT-GFP. Scale bar=100 μm. (h) Quantification of distance of AT-GFP+ cells away from CD31/SMA+ blood vessels from sections described in g. (i,j) SVPs were isolated from mice described in a and cultured. GFP locality was visualized 12 h later (j) and AT-GFP number was quantified (i). Scale bar=100 μm. (k–m) Subcutaneous IGW depots from mice described in a were excised and encased in Matrigel. Explants were continually treated with vehicle or imatinib ex vivo for 5 days. Vascular sprouts were then quantified for sprout length (k), branching (l) and progenitor occupancy (m). *P<0.01 Imatinib treated compared to vehicle (DMSO) treated. Data are expressed as means±s.e.m. Scale bars 100 μm.

Figure 7. VEGF is a transcriptional target of PPARγ in APCs.

(a) GFP− and GFP+ cells were FACS isolated from AT-GFP control mice and adipocytes were isolated by floatation (n=6). VEGF mRNA expression was measured. (b) AT-GFP+ cells were FACS isolated from AT-GFP control mice (n=6) and treated with vehicle or VEGF for 8 h and then treated with BrdU for 12 h. BrdU incorporation was monitored. (c) AT-GFP+ cells were FACS isolated from AT-GFP control mice (n=6) and treated with vehicle or VEGF for 8 h and then plated in transwell migration chambers. Migration was monitored 12 h later and quantified. (d–f) Total SV cells were isolated from AT-GFP control mice (n=6). Cells were treated with vehicle or VEGF for 12 h and then mRNA was harvested or cells were imaged for AT-GFP and stained for CD31. (g) Cells and staining described in d were quantified for AT-GFP and CD31 co-localization. (h) AT-GFP+ cells were FACS isolated from AT-control, SMA-PPARγ-LOF and AT-PPARγ-GOF and AT-PPARγ-Rescue or SV cells were isolated from SMA-PPARγ-GOF and VEGF mRNA expression was assessed (n=6 per group). (i) AT-GFP+ cells were FACS isolated from AT-control mice (n=6) and treated with denoted concentrations of rosiglitazone for 4 h. mRNA expression of VEGF and Angptl4 were measured. (j) AT-GFP+ cells isolated from AT-control mice (n=6) and were pretreated with cyclohexamide (10 μg ml⁻¹) for 15 min and then treated with 1 μM rosiglitazone for 4 h and PDGFRβ mRNA expression was measured. (k) AT-GFP+ cells were FACS isolated from AT-control mice (n=6). Cells were then treated with vehicle (dimethylsulfoxide (DMSO)) or 1 μM rosiglitazone for 4 h and then ChIP-qPCR analysis was performed to assess PPARγ occupancy. Data are expressed as means±s.e.m. Scale bar 100 μm. *P<0.05 unpaired t-test, two tailed: GFP+ and adipocytes compared to GFP− SV cells. P<0.01 unpaired t-test, two tailed: VEGF treated compared to vehicle treated cells. **P<0.02 unpaired t-test, two tailed: mutant mice compared to control mice. §P<0.001 unpaired t-test, two tailed: vehicle compared to IgG control. #P<0.001 rosiglitazone treated compared to vehicle (DMSO) treated.

Figure 8. VEGF stimulates niche expansion but blocks fat formation.

(a) Representative images of H&E staining of AT-control (PPARγ^tTA; TRE-H2B-GFP) and AT-VEGF (PPARγ^tTA; TRE-VEGF; TRE-H2B-GFP) at 3 months of age. (b) Representative images of CD31 immunostaining of IGW adipose depots from AT-control and AT-VEGF mice. (c) AT-control and AT-VEGF IGW adipose depots were sectioned and stained for GFP, CD31 and SMA. (d) Representative images of SVPs isolated from 3-month old AT-control and AT-VEGF mice and analysed for GFP locality. (e) Representative images of vascular sprouts from subcutaneous IGW adipose explants from 3-month old AT-control and AT-VEGF mice. (f,g) Quantitative RT–PCR analysis of niche and vasculogenic genes from 2-month old AT-control and AT-VEGF mice. (h,i) Total SV cells were isolated from AT-control and AT-VEGF mice (n=6). Cells were cultured for 7 days and then examined for vascular assembly by examining CD31 staining and GFP+ cell co-localization. Scale bars 100 μm. Data are expressed as means±s.e.m. Experiments were performed three times on 6–10 mice per group. Scale bars 100 μm. *P<0.05 unpaired t-test, two tailed: mutant compared to control levels.

Figure 9. VEGF stimulates niche expansion from adult APCs.

(a) Diagram of experimental paradigm. At P60, SMA-rtTA control or SMA-rtTA; TRE-VEGF (SMArtTA-VEGF) mice were administered Dox for 2 weeks. BrdU was administered 24 h before final analysis. Experiments were performed three times on seven mice per group. (b) Representative images of H&E staining of SMArtTA-control and SMArtTA-VEGF mice described in a. (c) Quantitative RT–PCR analysis of CD31 mRNA expression from denoted tissues. (d–f) Total SV cells were isolated from subcutaneous IGW depots mice described in a. Cells were stained and examined for CD31 (d), SMA (e) or BrdU and CD31-positive cells by flow cytometry. (g–i) Quantitative RT–PCR analysis of VEGF mRNA expression (g), nichegenic (h) and vasculogenic (i) genes from mice described in a. Scale bar 50 μm. Data are expressed as means±s.e.m. *P<0.01 unpaired t-test, two tailed: mutant compared to control levels.