Skin adipocyte stem cell self-renewal is regulated by a Pdgfa/Akt signaling axis

Immature and mature adipocytes are depleted with age and hair depilation

Age-related changes to the skin include thinning of the dermis (Luo et al., 2002; Sun et al., 2004; Tchkonia et al., 2010; Tyner et al., 2002), which coincides with a decrease in dWAT and an increase in the papillary and reticular dermis (Figures 2A–D). To evaluate whether loss of CD24+ ASCs might contribute to the degeneration of dWAT in aged mice, we analysed CD24+ ASCs and CD24− preadipocytes in the skin of young (3 weeks old) and aged (18, 24 and 30 month-old) mice using FACS. The fraction of CD24+ ASCs were preferentially lost during aging, reduced by approximately 50% (Figures 2E–G), suggesting that age predominantly impacts CD24+ ASC maintenance.

Open in a separate window

Figure 2

Loss of adipose tissue in different models correlates with loss of APs

A. Whole mounts images of C57/Bl6 skin stained with stained with bodipy to label neutral lipids and TOPRO-3 at different indicated ages. Scale bar is 200μm. B–D. Quantification of the area of the fibroblast-rich dermis (B) or the dermal adipose layer (C) and the number of mature adipocytes (D) normalized to skin length. (n = 3–5 mice for each time point). E–F. FACS plots and quantification of dermal CD24+ adipocyte stem cells (ASCs) (F) and CD24− preadipocytes (G). The percentage of AP cells is shown in each gate (E) (n = 4–7 mice for each time point). H. Whole-mount images of skin with Bodipy to label neutral lipids and TOPRO3 after a single or three serial depilation. Scale bar is 500μm. I–J. Quantification of the area of dermal adipose (I) and number of mature adipocyte (J) in single or serially depilated skin. (n = 4 mice for each time point). K. Flow cytometry plots and quantification of dermal CD24+ adipocyte stem cells (ASCs) (L) and CD24− preadipocytes (M) isolated from single or serially depilated skin. The numbers inside the gates indicate the percentage of AP cells in each gate (K). Statistical analysis was done using One-way ANOVA with Bonferroni post-test for multiple comparisons (B–D and F–G) or Student’s T test (I–J and L–M). Data are mean ± SD. Asterisk indicates significance, *p<0.05, **p<0.01, ***p< 0.001 or **** p<0.0001 calculated with One-way ANOVA with Dunnett’s or Tukey’s post-test for multiple comparisons. See also Figure S2.

Recent work demonstrated that multiple rounds of hair depilation induce serial cycles of hair growth and reduced the numbers of hair follicle stem cells (Keyes et al., 2013). To determine if repeated depilation alters dWAT and adipogenic lineage cells, we analysed dWAT after a single depilation or after 3 rounds of depilation. Similar to alterations of the dermis with age, serial depilation led to a significant reduction of dWAT area and mature adipocyte numbers compared to a single round of depilation (Figures 2H–J). Multiple rounds of depilation also reduced the percentage and proliferation of CD24+ ASCs and CD24− preadipocytes with CD24+ ASCs displaying a more dramatic reduction (Figures 2K–M and S2A–B). Together, these data indicate that CD24+ ASCs are differentially regulated compared to CD24− preadipocytes during hair cycling and with age.

Pdgfa is neccesary to maintain CD24+ ASCs and dWAT

We previously showed that Pdgfa and its receptor Pdgfrα are expressed by APs (Berry and Rodeheffer, 2013; Festa et al., 2011). Analysis of Pdgfa expression in CD24+ ASCs and CD24− preadipocytes showed that Pdgfa levels are significantly elevated between P16–18 (Figure 3A) when CD24+ ASCs are proliferative (Figure 1F–G) and significantly decrease by P32 (Figure 3A). While Pdgfa expression in CD24+ ASCs is not altered during depilation (Figure S2C–E), Pdgfa and Pdgfrα mRNA levels in AP cells from 3 week, 18 month and 30 month old mice revealed a significant downregulation with age when CD24+ ASCs maintenance is reduced (Figure 3B and 2E–G).

Open in a separate window

Figure 3

Pdgfa is required for dermal CD24+ ASC and intradermal adipocyte maintenance

A. Real time PCR expression Pdgfa mRNA in CD24+ ASCs and CD24− preadipocytes isolated from the skin of mice at indicated ages. Data are mean ± SD normalized to P16 old mice expression levels (n=3–4 mice per group). B. Pdgfa and Pdgfrα mRNA expression in AP cells isolated from WT at 3 weeks, 18 and 30 months of age. Data are mean ± SD normalized to 18 month old mice expression levels (n=3–4 mice per group). C. Whole mount confocal images of P32 skin from WT and Pdgfa cKO mice stained with Bodipy to label neutral lipids and TOPRO3 to stain nuclei. Scale bar is 200μm. Bracket indicates dermal white adipose (dWAT) thickness. D–E. Quantification of dermal adipose tissue area (mm²/mm of skin) (D) and number of mature adipocytes per mm of skin (E) in WT and Pdgfa cKO mice at age P32. (n=3–4 mice per genotype). F. Flow cytometry analysis of APs isolated from skin of WT and Pdgfa cKO mice at indicated ages. The percentage of AP cells is indicated in each gate. G–H. Quantification of CD24+ASCs and CD24− preadipocytes from WT and Pdgfa cKO mice at different postnatal days. AP numbers from Pdgfa cKO mice are normalized to the WT control in each age group. (n=3 mice per genotype). I. Flow cytometry histogram of EdU incorporation in CD24+ ASCs from P20 WT and Pdgfa cKO mice after EdU pulses from P17–19. Control is WT mice without EdU pulse. J–K Quantification of the percentage of EdU+ CD24+ adipocyte stem cells (ASCs) (B) or CD24− preadipocytes (C) in WT and Pdgfa cKO. (n = 3 mice for each genotype). Data are mean ± SD. Asterisks indicates significance, *p<0.05, **p<0.01, calculated with One-way ANOVA with Dunnett’s post-test for multiple comparisons or Student’s T test for two groups. See also Figure S3.

To investigate the role of Pdgfa in maintaining APs in the skin, we deleted Pdgfa in dermal mesenchymal cells by crossing mice expressing Cre recombinase driven by the Pdgfrα promoter with Pdgfa^fl mice (Pdgfa cKO) (Andrae et al., 2014). The floxed Pdgfa allele introduces LacZ into exon 4, creating a truncated, biologically inactive protein (Figure S3A) (Andrae et al., 2014). During hair regression and rest (telogen), dWAT area and adipocyte numbers were not different between WT and Pdgfa cKO mice (Figure S3B–D). However, upon hair follicle growth, the expansion of dWAT was attenuated in Pdgfa cKO (Figure 3C) as demonstrated by a significant reduction in dWAT area and a ~20% reduction in mature intradermal adipocytes in Pdgfa cKO mice (Figures 3D–E). Since 20% of adipocytes are generated de novo during the first hair cycle (Figure 1C) and since ASCs generate mature adipocytes (Festa, Ryan, and Jeffery), these data reveal that the adipogenic program is diminished in the dermis of Pdgfa cKO mice.

Since Pdgfa has been suggested to play a role in regulating the hair follicle cycle (Karlsson et al., 1999), we investigated whether hair cycling defects occurred in Pdgfa cKO mice to impact adipocyte regeneration. Hair regression and regrowth occurred with similar timing in Pdgfa cKO mice and WT mice during the first hair cycle (Figure S3E–H). Since Pdgf family members have overlapping roles (Chen et al., 2004; Donovan et al., 2013), we examined the mRNA expression of Pdgf ligands in CD24+ ASCs and CD24− preadipocytes in Pdgfa cKO mice. While Pdgfc was not detected (not shown) and Pdgfd was not altered in APs from Pdgfa cKO mice, Pdgfb was significantly upregulated in both CD24+ ASCs and CD24− preadipocytes in Pdgfa cKO mice at P22 when hair cycling is initiated. Given the somewhat overlapping roles of Pdgfa and Pdgfb (Chen et al., 2004; Donovan et al., 2013), Pdgfb may compensate for the lack of Pdgfa to maintain hair follicle cycling in Pdgfa cKO mice (Figure S3I–J).

To delineate whether Pdgfa promotes AP maintenance, we examined the number of dermal APs in Pdgfa cKO mice before and after hair follicle regeneration. While CD24− preadipocytes were not altered in the absence of Pdgfa signaling, CD24+ ASCs showed a 50% reduction in Pdgfa cKO mice at all time points examined before and after hair regrowth (Figures 3F–H). This decrease in CD24+ ASC numbers was specific to the skin as CD24+ ASCs were similar in vWAT and sWAT depots of WT and Pdgfa cKO mice (Figure S3K), indicating that Pdgfa signaling plays a specific role in maintaining dermal CD24+ ASCs. To determine if Pdgfa regulates proliferation of dermal CD24+ ASCs, we injected WT and Pdgfa cKO mice with EdU during hair follicle involution and at the peak of CD24+ ASC proliferation from P17 to P19 (Figure 1H). At the initiation of hair follicle cycling (P20), both CD24+ ASCs and CD24− preadipocytes displayed substantially reduced EdU incorporation in Pdgfa cKO mice compared to WT mice (Figure 3I–K). These data demonstrate that dermal Pdgfa expression is required for CD24+ ASC proliferation and maintenance in the skin.

Pdgfa acts directly on CD24+ ASCs

To determine if Pdgfa acts directly on AP cells, we performed competitive transplantation experiments to examine the maintenance of WT and Pdgfrα null APs in the dermis. Donor Rosa-CreER; mT/mG; Pdgfrα^fl/fl mice were treated with tamoxifen at P21–P24 (Figure 4A) and 19% of APs were GFP+ and lacked Pdgfrα (Figures 4A–C). We then transplanted 2×10⁵ total dTomato+ and mGFP+ AP cells into the skin of syngeneic WT hosts together, allowing us to simultaneously analyse WT and Pdgfrα null AP maintenance (Figure 4A). After 3 days of transplantation, the percentage of transplanted GFP+ APs recovered from the recipients was reduced to 1.8% (Figure 4D–E). Both populations of GFP+ APs, CD24+ ASC and CD24− preadipocytes, showed a significant decrease in their recovery 3 days after transplantation (Figure 4D–E). Since only 4×10⁴ GFP+ APs were transplanted in these experiments, we determined if loss of transplanted APs was due to the low number of transplanted cells. Transplantation of 4×10⁴ WT APs were maintained in the dermis after 10 days and formed Perilipin+, mature adipocytes (Figure S4A–B). These data suggest that lack of Pdgfrα impacts the maintenance of APs in the skin.

Open in a separate window

Figure 4

Pdgfrα is required for maintenance and proliferation of dermal CD24+ ASCs

A. Schematic of transplantation experiments with Tomato+ (WT) and GFP+ (Pdgfrα KO) APs from tamoxifen-treated Rosa-CreER;Pdgfrα^fl/fl mice into the skin of P15 syngeneic mice. B–E. Flow cytometry analysis and quantification of APs isolated from tamoxifen-treated Rosa-CreER;Pdgfrα^fl/fl mice (B–C, E) or recovered from host mice 3 days after transplantation (D–E). The dot plot in D shows recovered Tomato+ or GFP+ APs overlaid on WT cells (grey). n=3 mice per genotype (C) or 6 transplant experiments (D). F–G. Flow cytometry analysis (F) of EdU incorporation in CD24+ ASCs and quantification in CD24+ ASCs and CD24− preadipocytes (G) from tamoxifen-injected Rosa-CreER;mTmG;Pdgfrα^fl/fl mice after a 6 hr EdU pulse. n=4 mice per genotype. Data are mean ± SD. Asterisks indicates significance, *p<0.05 calculated with One-way ANOVA with Dunnett’s post-test for multiple comparisons. See also Figure S4.

Since CD24+ ASCs generate CD24− preadipocytes (Berry and Rodeheffer, 2013), we determined whether the loss of both subpopulations of APs was primarily due to a defect in CD24+ ASC proliferation. We pulsed Rosa-CreER; mT/mG; Pdgfrα^fl/fl mice with EdU for 6h at P17 or P18. While CD24-preadipocytes incorporated low amounts of EdU in both genotypes, Pdgfrα null CD24+ ASCs incorporated significantly less EdU than WT CD24+ ASCs (Figure 4F–G). These data demonstrate that CD24+ ASCs require intrinsic Pdgf signaling for proper maintenance and proliferation.

Immunofluorescence and tissue staining

Mouse skin was embedded in OCT compound (Tissue-Trek 4583) and frozen on dry ice. OCT blocks were cryosectioned at 14μm and fixed for 10 min with 4% formaldehyde. Skin sections were stained as previously described (Festa et al., 2011). The following antibodies were used: GFP (chicken, Abcam (ab13970), 1:1000), Perilipin A (goat, Abcam (ab61682), 1:1000). Sections were mounted in Prolong Gold anti-fade reagent with 4′6′-diamidino-2-phenylindole (DAPI) (Invitrogen P36935). For skin whole mount staining, skin strips (approximately 1 mm thick) were obtained and fixed in 4% paraformaldehyde for 20 min. Skin strips were then washed with PBS and stained with Bodipy (1:1000, Invitrogen D-3922) and TO-PRO-3 Iodide (1:300, Thermo Fisher Scientific T3605) for 30 min at room temperature. Strips were washed in PBS and mounted using Aqua Poly/Mount (Polysciences 18606). Images were acquired in a LSM 510 Visible Confocal microscope (Zeiss) using the ZEN software. Image analysis and measurements were done using ImageJ software (NIH) and Adobe Photoshop.

Flow activated cell sorting and analysis

FACS analysis of APs was performed as described previously (Festa et al., 2011; Jeffery et al., 2015). Briefly, mouse skin was dissected and digested in collagenase buffer (HBSS containing 3% BSA, Collagenase 1A 1:100, Worthington LS004196, 1.2mM calcium chloride, 0.8mM zinc chloride) for 60 min at 37°C in a shaking water bath. Undigested tissue was separated from released cells through filtration using 70μm filters. Floating mature adipocytes were separated from the stromal vascular fraction (SVF) by centrifugation at 300g for 3 min. To identify adipocyte precursors (APs) SVF was stained in 3% BSA in HBSS with the following antibodies: CD31-PE-Cy7 (1:500, eBioscience 25-0311-82), CD45 APC-eFluor 780 (1:5000, eBioscience 47-0451-82) CD29 Alexa Fluor 700 (1:400, Biolegend 102218), CD34 Brilliant Violet 421 (1:50, Biolegend 119321) Ly-6A/E (Sca1) V500 (1:500, BD Biosciences, 561229) and CD24 PerCP-Cy5.5 (1:200, eBioscience 45-0242-82). For live/dead discrimination, cells were stained with Sytox Orange (1:100,000, Invitrogen S11368). For analysis of proliferation by EdU incorporation, the Click-iT EdU Flow Cytometry Assay Kit (Invitrogen C10419) was used following the manufacturer’s instructions. Detection of phosphorylated Akt in APs was performed as previously described (Jeffery et al., 2015) using pAkt S473 (1:50; Cell Signaling 9271). Samples were sorted or analyzed with a FACS Aria III with DiVA software. Analysis of flow cytometry data was performed using FlowJo Software.

Adipocyte precursor cell culture

FACS-isolated APs were cultured as described (Festa et al., 2011; Rodeheffer et al., 2008). FACS sorted cells were plated on carboxyl-coated 24 well plates (BD Biosciences 354775) in DMEM supplemented with 10% FBS. AT 50% confluency, APs were switched to 0.5% FBS DMEM overnight and treated with Pdgfa (30ng/ml, Affimetryx/eBioscience 14-8989-80) or vehicle 1% BSA in PBS for specific periods of time. For proliferation assays APs were treated with Pdgfa (30ng/ml, Affimetryx/eBioscience 14-8989-80), 5μM EdU, LY294002 (2μM, Selleckchem S1105), Vemurafenib (1μM, Selleckchem, S1267) and Ruxolitinib (5μM, Selleckchem, S1378) or DMSO for 24h.

RNA extraction and Real-Time PCR

APs were either sorted directly into Trizol LS (Invitrogen 10296-028) or lysed in 24 well plates by adding Trizol (Invitrogen 15596-026). RNA extraction and purification was done using the RNeasy mini kit (QIAGEN 74104) following manufacturer instructions. cDNA was generated using equal amounts of total RNA with the Superscript III First-Strand Synthesis System (Invitrogen 18080051) using Oligo dT per the manufacturer’s instructions. Real time PCR was performed as previously described (Festa et al., 2011) using SYBR green I Master mix (Roche 04887352001) on a LightCycler 480 (Roche). Primers for specific genes are listed in the supplemental information section. Results were normalized to β-actin expression as described previously (Festa et al., 2011).

Western Blot

Adipocyte precursors were processed as previously described (Goldstein et al., 2014). Briefly, APs were collected in RIPA buffer supplemented with protease inhibitors. Equal amount of protein lysates were loaded into acrylamide SDS-PAGE gels, blotted into PVDF membranes and developed by chemiluminescence. The following primary antibodies were used: Akt (1:500, Cell Signaling 9272), pAkt Ser473 (1:500, Cell Signaling 9271), PKC (1:500, Abcam ab19031), pPKC βII Ser660 (1:500 Cell Signaling 9371) and β-actin (1:3000, Sigma AC-15).

Competitive transplantation assay

APs cells were isolated by FACS from Rosa-CreER; mT/mG; Pdgfrα^fl/fl injected with tamoxifen for four days as described above. Isolated APs included WT (Tomato+) and Pdgfrα KO (GFP+) that were injected intradermally into the back skin of P14 Rosa-CreER; Pdgfrα^fl/fl male mice. Each recipient mouse was injected with 200,000 APs, at two different sites, one closer to the anterior and another closer to the posterior area. At P17 the skin of recipient mice was analysed for the presence of both WT (Tomato+) and Pdgfrα KO (GFP+) APs using the FACS strategy described previously.

RNA-seq

RNA samples were prepared from primary APs isolated from the back skin of four 6 week old C57/Bl6 mice and treated with Pdgfa for 1 or 2 hours or left untreated as described above. RNA from the different samples was pooled and RNA-seq was performed as described previously (Tadeu et al., 2015). Single-end RNA sequencing was done using an Illumina sequencer at the Yale Center for Genome Analysis. Using the TopHat and Cufflinks suite (Roberts et al., 2012), the raw reads were assembled into a transcriptome, and a list of differentially expressed and regulated genes and transcripts was produced. Gene Ontology analysis was done on the significantly regulated genes (experimental logarithmic ratio < 1, p < 0.05 and q < 0.05) using DAVID (Huang et al., 2009a; 2009b) or the QIAGEN’s Ingenuity® Pathway Analysis software suite (IPA®, QIAGEN Redwood City, www.qiagen.com/ingenuity). RNA-seq data was deposited to the GEO database accession number GSE84370.

We thank Horsley and Rodeheffer lab members for technical assistance, critical reading of the manuscript and scientific discussions. We thank Kaitlyn M. Sabin for technical assistance with section staining and imaging. This project was funded by the NIA through the pilot project grants from the Claude D. Pepper Older Americans Independence Center at Yale (NIA P30AG21342) awarded to V.H. and (NIA P30AG21342) awarded to M.S.R. This work was also supported by grants from NIDDK (R01DK090489) to M.S.R and NIAMS (AR060296) awarded to V.H.