Dermal β-catenin activity in response to epidermal Wnt ligands is required for fibroblast proliferation and hair follicle initiation

Abstract

Dermal fibroblasts are required for structural integrity of the skin and for hair follicle development. Uniform Wnt signaling activity is present in dermal fibroblast precursors preceding hair follicle initiation, but the functional requirement of dermal Wnt signaling at early stages of skin differentiation and patterning remains largely uncharacterized. We show in mice that epidermal Wnt ligands are required for uniform dermal Wnt signaling/β-catenin activity and regulate fibroblast cell proliferation and initiation of hair follicle placodes. In the absence of dermal Wnt signaling/β-catenin activity, patterned upregulation of epidermal β-catenin activity and Edar expression are absent. Conversely, forced activation of β-catenin signaling leads to the formation of thickened dermis, enlarged epidermal placodes and dermal condensates that result in prematurely differentiated enlarged hair follicles. These data reveal functional roles for dermal Wnt signaling/β-catenin in fibroblast proliferation and in the epidermal hair follicle initiation program.

Fig. 1.

Wnt signaling reporter activity and Wls protein expression in embryonic mouse skin. (A-C) X-Gal staining of transverse sections of dorsal skin at the forelimb level from Wnt reporter TCF-Lef/lacZ embryos. (D-G) Wls protein expression in embryonic dorsal skin revealed detectable expression in the epidermis, hair follicle placode, hair follicle peg (peg) and dermal condensate (dc). Dashed lines indicate the dermal-epidermal boundary. Dotted lines indicate the dermal condensate boundary.

Fig. 2.

Sustained β-catenin activity in dermal fibroblast progenitors leads to an increase in dermal thickness and to the proliferation of dermal fibroblasts. (A-N) X-Gal staining, Fgfr1, Twist2 and Col1a1 mRNA expression revealed thickened dermis in the mutant compared with control mouse embryos. White dashed lines demarcate the epidermal-dermal boundary and vertical bars indicate the dermis region. Control and mutant skin sections were photographed at the same magnification. (O-Q) Quantification of dermal thickness (O), proliferation index (P) and cell density (Q) in the upper dermis region showed significant increases in mutant embryos compared with controls. Mean values with s.d. hf, hair follicle.

Fig. 3.

Increase in hair follicle placode size and number and accelerated differentiation in stabilized dermal β-catenin embryos. (A-D) Whole-mount in situ hybridization for Dkk4 at E14.5 (A,B) and X-Gal staining of Wnt reporter activity in mouse ventral skin (C,D). Expanded placodes in mutant embryos are indicated by arrows (B); the distance between placodes was decreased in the mutant skin (D, solid black lines). (E,F) Increase in the number of cells in the Dkk4⁺ and Lef1⁺ domains in mutant embryos as compared with the control. (G,H) Quantification of hair follicle area showed enlargement of hair follicle in mutant compared with control skin of age-matched littermate, and enlargement of mutant hair follicle compared with stage-matched wild-type hair follicle. (I-L) K17 and Sox9 showed comparable expression in the outer root sheath (ORS) in control and mutant skin. (M,N) AE13 immunoreactivity was absent in the control ventral skin, but was present in the enlarged follicles of the mutant skin at E17.5. (O,P) Sox2 was present in the dermal condensate and papilla of the control and enlarged mutant hair follicles (arrows). Numbers (1-4) specify developmental stages of hair follicles. Dashed line demarcates the dermal-epidermal boundary. Control and mutant skin were photographed at the same magnification. Error bars indicate s.d.

Fig. 4.

Patterned but expanded expression of hair follicle placode and dermal condensate markers in stabilized dermal β-catenin mutant mouse embryos. (A-D) Compared with the control embryonic skin, patterned expression of Dkk1 in the dermis surrounding the follicle and Dkk4 in the preplacode was expanded in mutant skin. (E-J) Expression of Bmp4, and of Ptch1 in the dermal condensate and Shh in the epidermis was expanded in the mutant compared with the control. (K,L) Wnt reporter activity in the placode was expanded in mutant skin. Dashed line demarcates the dermal-epidermal border, dotted curved lines outline the placode and dermal condensate domain, and dotted vertical lines outline the preplacode domain.

Fig. 5.

Dermal Wnt/β-catenin activity regulates dermal fibroblast proliferation. (A,B) X-Gal staining shows En1Cre lineage-marked cells in the dermis at E12.5 and E13.5. (C-F) X-Gal staining of control (C,E) and mutant (D,F) embryonic mouse dorsolateral skin. (G,H) Hematoxylin and Eosin (H&E) staining of perinatal skin revealed a loose distribution of dermal fibroblasts and a thinner dermal layer in the mutant. (I-L) Twist2 and Fgfr1 mRNA was present in the upper dermis of control but absent in mutant skin. (M-P) Wnt reporter activity showed expression in the upper dermis and patterned upregulation in the placodes and dermal condensates in controls, whereas reporter activity was absent in the epidermis and dermis of mutant skin. (Q-X) K14, PGP9.5, Pecam1 and Col1a1 mRNA expression revealed comparable epidermal basal keratin expression, innervation, vasculature formation and collagen production, respectively, in control and mutant skin. Dashed lines demarcate the epidermal-dermal boundary. Control and mutant skin sections were photographed at the same magnification. (Y) The percentage of BrdU⁺ proliferating dermal fibroblasts in the upper dermis was significantly decreased in the mutant compared with the control. (Z) Cell density in the upper dermis showed no difference between control and mutant at E13.5-14.5, but a significant decrease in the mutant embryos at E16.5. Error bars indicate s.d. dc, dermal condensate; hf, hair follicle.

Fig. 6.

Patterned upregulated expression of hair follicle initiation markers was absent without dermal Wnt signaling/β-catenin activity. (A-F) In situ hybridization for Dkk4, Wnt10b and Edar mRNA revealed patterned upregulated expression in the control (A,C,E) but an absence of expression in the mutant mouse skin (B,D,F). (G,H) Expression of Lef1 in the epidermis and focal upregulation in the hair follicle placode and dermal condensate in the control, whereas patterned expression of Lef1 was absent in the mutant skin. (I-N) Ptc1, Shh and Bmp4 mRNA expression was observed at sites of hair follicle initiation in control skin (I,K,M), but was absent in mutant skin (J,L,N). Dashed lines indicate the dermal-epidermal boundary. Downward arrows indicate placodes in the epidermis and the upward arrows indicate dermal condensates. All images are at the same magnification.

Fig. 7.

Dermal canonical Wnt signaling activity depends on epidermal but not dermal Wnt ligands. (A-C′) Transverse sections at the forelimb level from E14.5-18.5 control and mutant mouse embryos. (D-G′) Axin2, Twist2, Fgfr1 mRNA expression and Lef1 protein expression revealed loss of Wnt responsiveness in the upper dermis of the mutant compared with the control. Note that Lef1 expression remained intact in the basal layer of the epidermis of the control and mutant skin (G,G′). (H) The percentage of BrdU⁺ proliferating dermal fibroblasts was significantly reduced in the mutant compared with the control. (I) Cell density in upper dermis showed no difference between control and mutant at E13.5-14.5, but a significant decrease in the mutant embryos at E16.5 and E18.5. (J-N′) Transverse sections of skin at the forelimb level from control (J-N) and mutant (J′-N′) skin. Skin samples were stained with H&E (J,J′), for Wls protein (K,K′), Twist2 mRNA (L,L′) or Lef1 protein (M,M′), or for β-gal activity (N,N′). Arrows indicate dermal fibroblasts with positive nuclear Lef1 immunoreactivity. White dashed lines demarcate the epidermal-dermal junction. Control and mutant skin sections are at the same magnification. (O) The relative level of Wls mRNA in the dorsal dermis detected by quantitative RT-PCR showed a significant reduction in the mutant embryo compared with control. There was no discernible difference in morphology, Wnt responsiveness or hair follicle formation between control and mutant skin. Error bars indicate s.d. (P) Proposed model for the reciprocal signaling loop whereby Wnt ligands are first secreted from the epidermis and lead to the activation of canonical Wnt signaling/β-catenin pathway in the underlying dermal fibroblasts, which then send signals back to the epidermis. The epidermis then becomes able to respond to epidermal Wnt ligands and to generate patterned upregulated β-catenin activity in the preplacode. Subsequently, β-catenin-expressing epidermal cells in the placode send signals to recruit dermal fibroblast cells to form the dermal condensate. DC, dermal condensate; hf/HF, hair follicle.