Wnt/β-catenin signaling in dermal condensates is required for hair follicle formation

Abstract

Broad dermal Wnt signaling is required for patterned induction of hair follicle placodes and subsequent Wnt signaling in placode stem cells is essential for induction of dermal condensates, cell clusters of precursors for the hair follicle dermal papilla (DP). Progression of hair follicle formation then requires coordinated signal exchange between dermal condensates and placode stem cells. However, it remains unknown whether continued Wnt signaling in DP precursor cells plays a role in this process, largely due to the long-standing inability to specifically target dermal condensates for gene ablation. Here we use the Tbx18(Cre) knockin mouse line to ablate the Wnt-responsive transcription factor β-catenin specifically in these cells at E14.5 during the first wave of guard hair follicle formation. In the absence of β-catenin, canonical Wnt signaling is effectively abolished in these cells. Sox2(+) dermal condensates initiate normally; however by E16.5 guard hair follicle numbers are strongly reduced and by E18.5 most whiskers and guard hair follicles are absent, suggesting that active Wnt signaling in dermal condensates is important for hair follicle formation to proceed after induction. To explore the molecular mechanisms by which Wnt signaling in dermal condensates regulates hair follicle formation, we analyze genome-wide the gene expression changes in embryonic β-catenin null DP precursor cells. We find altered expression of several signaling pathway genes, including Fgfs and Activin, both previously implicated in hair follicle formation. In summary, these data reveal a functional role of Wnt signaling in DP precursors for embryonic hair follicle formation and identify Fgf and Activin signaling as potential effectors of Wnt signaling-regulated events.

Keywords: Dermal papilla cells; Hair follicle morphogenesis; Hair follicle stem cells; Stem cell niche; Wnt signaling.

Figure 1. Active Wnt/β-catenin signaling in placodes and dermal condensates of all three hair follicle waves

(A) Schematic representation of mouse hair development in three consecutive waves. The 1^st wave starts at E14.5 forming guard hair follicles. 2^nd wave awl/auchene hairs form at E16.5. Zigzag hair follicles form at E18.5 in the 3^rd wave. Right: Illustration of spatial distribution of guard (red), awl/auchene (blue) and zigzag (black) hair follicles in backskin. (B) Whole-mount X-Gal staining of Wnt reporter BATGAL embryos at indicated time points. Note blue dots in a typical hair follicle distribution pattern. (C, D) Wnt reporter activity in embryonic skin sections. LacZ is strongly expressed in hair placodes (arrowheads) and DP condensates (arrows). Lower panels: Pseudo-colored image of LacZ staining overlaid with DAPI to highlight nuclei. The dotted line marks basement membrane between epidermis and dermis. (E) At E16.5, LacZ is expressed in DP niches of both 1^st (arrows) and 2^nd (open arrowheads) wave hair follicles. (F) X-Gal stained sections from BATGAL embryos at E18.5. Wnt reporter activity is also detectable in DP precursor cells of 3^rd wave zigzag follicles (open arrows). (G) Quantification of Wnt signaling. BATGAL is expressed in hair placodes and dermal condensates of all follicle types in all three waves. Note that Wnt signaling is detectable in placodes, DP condensates or both, suggesting dynamic Wnt activity in both compartments. Scale bars: 1mm in B; 50μm in C–F.

Figure 2. Embryonic β-catenin ablation in dermal condensates impairs 1^st wave guard hair follicle formation

(A) Tbx18^Cre mediated β-catenin ablation in embryonic DP precursor cells in dermal condensates at E14.5. Cre and floxed lines were triple crossed with Wnt reporter BATGAL mice. Shown are sections of whole-mount X-gal stained embryos confirming strong reduction of Wnt activity in dermal condensates (arrows) of β-catenin cKO. Note that placodes (arrowheads) remained Wnt reporter positive. Bottom: X-gal stained sections and counterstaining with DAPI to highlight nuclei. Dotted line marks basement membrane between epidermis and dermis. (B) Hematoxylin and eosin stained sections of DP conditional β-catenin knockout (cKO) and wild-type (WT) backskin at indicated time points. Induction of 1^st wave guard hair follicles (arrows) is not affected at E14.5, but their downgrowth and formation at E16.5 and E18.5 was strongly diminished. Formation of 2^nd wave (open arrowheads) and 3^rd wave hair follicles (open arrows) was unaffected, consistent with the lack of efficient β-catenin ablation in these follicle types. (C) Quantification of total hair follicle numbers. (D) Quantification of hair follicle types from the three waves at E18.5. Note 1^st wave guard hair follicles are significantly reduced. (E) E14.5 whole-mount immunofluorescence staining for dermal condensate marker SOX2 and placode marker EDAR. (F) Quantification of SOX2-positive dermal condensates and EDAR-positive placodes. (G) E18.5 whole-mount skins of β-catenin heterozygous (HET) control and cKO embryos crossed with Lef1-RFP reporter mice to highlight DPs. Note that large 1^st wave guard hair follicles (arrows) are missing in cKO. (H) Quantification of hair follicle sizes. Largest 1^st wave guard hair follicles are absent in cKO skins. Scale bars: 100μm.

Figure 3. Impaired 1^st wave guard and whisker hair follicle formation, but normal differentiation markers in developed hair follicles during postnatal growth

(A) Image of WT and cKO back skins at P5. Note absence of guard hair shafts in cKO. White arrows highlight outgrowing guard hair shafts in WT. (B) H&E sections of P5 back skins. Note thinner skin and fewer hair follicles in cKO. (C) H&E sections of whisker pads at E18.5. Several whiskers failed to form in cKO embryos (asterisks). (D) Top view of WT and cKO pups at P10. External whiskers are missing in cKO. (E–G) Immunofluorescence stainings of proliferation and differentiation markers in P5 WT and cKO awl hair follicles. (E) Ki67 labeling of proliferating matrix cells and Lef1 distribution was comparable between WT and cKO mice. (F) Unaltered AE15 and AE13 differentiation marker expression. Itgb4 (green) highlights basement membrane. (G) DP marker GFRA1 and AP are unperturbed. Scale bars: 100μm in B; 25μm in G.

Figure 4. Hair shafts are shorter during postnatal hair growth

(A) Hair growth is strongly reduced in β-catenin cKO mice compared to WT. Note that guard hairs are absent from the lower backskin. (B) H&E sections showed thinner skin with fewer follicles in cKO. (C) Overall length of hair shafts from all three waves is shorter in cKO. Note that 2^nd wave auchene hairs were missing. (D) Morphological analysis of hair shafts from all hair types. Awl hair contained two different subtypes in cKO; auchene hair was absent. (E) Quantification of all hair shaft types. Guard hairs were strongly reduced; auchene hairs were absent in cKO, but awl hair numbers were increased.

Figure 5. Ablation of Wnt/β-catenin signaling alters DP precursor gene expression, including signaling molecules Fgf10 and Inhba

(A) Section of E15.5 Sox2^GFP embryo backskin. Sox2 is specifically expressed in DP condensates. Dotted line marks basement membrane separating epidermis and dermis. (B) FACS plots. Single cells from embryonic Sox2^GFP skin preparations were DAPI stained to exclude dead cells. DP precursor cells were sorted as the GFP^hi cell population. (C) Real-time PCR of signature genes in sorted cells. GFP^hi cells were highly enriched for Sox2, Tbx18 and Enpp2. (D) Real-time PCR confirmed 70% β-catenin knockout efficiency in three sorts. (E) Analysis of enriched gene ontology (GO) categories. (F) Heat map analysis of altered gene expression in microarrays from WT and cKO sorted DP condensates. (G) Real-time PCR verification of downregulated genes in cKO. Expression levels were normalized to Gapdh and presented relative to WT. Data were from two independent sorts.