Fig 2: FGFR2 is not required for HFSC specification. a Hematoxylin and eosin (H&E) analysis of E16 skins of Fgfr2 cKO and control mice. b Quantification of hair follicle length, p < 0.001. n = 4. c Quantification of primary hair follicles number per mm of back skin at E16, p = 0.5358. n = 4. d Immunofluorescence staining for SOX9 (red) at E16. e Quantifications of the percentage of SOX9(+) cells per hair follicle in control and Fgfr2 cKO skins at E16, p = 0.1570. n = 4. f Immunofluorescence staining for SOX9 (red) at P0. g Appearance of full-thickness skin engraftment of P0 skins onto Nude mice 35 days post-engraftment (P35). Grafted Fgfr2 cKO, gFgfr2 cKO; grafted control, gControl. h Immunofluorescence staining for HFSC marker SOX9 (red) in P35 gFgfr2 cKO and gControl. i Immunofluorescence staining for HFSC marker CD34 (red) in P35 gFgfr2 cKO and gControl skins. Arrow points to a club hair. Scale = 50 µm in a, d, f. Scale = 20 µm in h, i. The data presented in box plots (b, c, e) shows the median with 25th and 75th percentile borders. Whiskers extend from minimum to maximum. ***p < 0.001; n.s. not significant (the Mann–Whitney test)

Fig 3: CD34 and Sox2 expression in skin tumorsSections of tumors were stained with anti-CD34 A+C or anti-Sox2 B+D. Both chemically induced tumors in haired mice (A+B) (CD34+ 15 of 21 tumors and Sox2+ 7 of 8 tumors) as well as UV-induced tumors in hairless mice (C+D) showed CD34 (6 of 14 tumors) and Sox2 expression (10 of 11 tumors). The CD34 expression in hairless mice was located in the differentiated compartments and more granular (C) compared to the haired mice (A). Scale bar = 100μm.

Fig 4: Pathological examination of the tumors. (A) Histology indicated that the tumors consisted of spindle-shaped cells and collagen fibers, and the former were arranged in a storiform pattern and surrounded by the latter. The arrows a and b indicated the collagen fiber and cell, respectively. (H&E staining; magnification, x200). Immunohistochemistry suggested that the tumor cells were strongly positive for (B) CD68 and (C) XIIIa factor, while they were weakly positive for (D) a-SMA and (E) desmin. (F) The tumor cells were negative for CD34 (magnification, x200; scale bars, 100 µm).

Fig 5: HFSCs are maintained upon PRC2 loss (A) Schematic illustrating expected tdtomato expression in the HF bulge of K15-CrePGR; RosaCAG-tdtomato mice. (B) Schematic showing the experimental strategy to induce K15-CrePGR activity in telogen II (P50) HFs. (C) IF analyses of tdtomato (red) and DAPI (blue) in K15-CrePGR; RosaCAG-tdtomato mice showing efficient targeting of the HF bulge. (D) IF analysis of H3K27me3 (red), CD34 (green) and DAPI (blue) in HFSCs of P70 control and Eed iKO mice. H3K27me3 single channel is shown in gray. Bulge (Bu) and hair germ (HG) regions of the follicle has been outlined in yellow. (E) Fluorescence intensity quantification of H3K27me3 signal (arbitrary units) in control and Eed-null HFSCs P<0.00001, n = 120 cells from 15 HF sections from three independent biological replicates for each group. (F-J) IF analyses of (F) SOX9 (red) and DAPI (blue), (G) NFATC1 (red) and DAPI (blue), (H) activated-CASPASE3 (red), CD34 (green) and DAPI (blue), (I) Ki67 (red), CD34 (green) and DAPI (blue) and (J) LEF1 (red), CD34 (green) and DAPI (blue) in HFs of P70 control and Eed iKO mice. LEF1 single channel is shown in gray. All IF experiments in this figure was conducted on skin sections collected from three animals for each group from two independent litters. Scale bar: 10μm.