A desmosomal cadherin controls multipotent hair follicle stem cell quiescence and orchestrates regeneration through adhesion signaling

Abstract

Stem cells (SCs) are critical to maintain tissue homeostasis. However, it is currently not known whether signaling through cell junctions protects quiescent epithelial SC reservoirs from depletion during disease-inflicted damage. Using the autoimmune model disease pemphigus vulgaris (PV), this study reveals an unprecedented role for a desmosomal cadherin in governing SC quiescence and regeneration through adhesion signaling in the multipotent mouse hair follicle compartment known as the bulge. Autoantibody-mediated, mechanical uncoupling of desmoglein (Dsg) 3 transadhesion activates quiescent bulge SC which lose their multipotency and stemness, become actively cycling, and finally delaminate from their epithelial niche. This then initiates a self-organized regenerative program which restores Dsg3 function and bulge morphology including SC quiescence and multipotency. These profound changes are triggered by the sole loss of functional Dsg3, resemble major signaling events in Dsg3-/- mice, and are driven by SC-relevant EGFR activation and Wnt modulation requiring longitudinal repression of Hedgehog signaling.

Keywords: Cell biology; Molecular biology.

Graphical abstract

Figure 1

Delamination and regeneration in the telogen HFBu (A and B) Immunolabeling and H&E of the telogen HFBu in 8-week-old C57BL/6J mice after a single s.c. injection of mIgG/AK23. Note that Dsg3 labeling is highest in the HFBu and low in HG and IFE. A1/3: telogen HFBu, A2/4: IFE: interfollicular epidermis, dIF: direct immunofluorescence detects bound AK23. (C) Schematic representation of the mouse telogen HF. Dsg3 expression (dark blue) is strongest in the bulge. Specific SC markers CD34, Sox9 and NFATc1 (red, highlighted in this study) are expressed in the BuSC while K6 (brown) is expressed in the epithelial niche. (D) Longitudinal quantification of blistered HFBus (Bb). Top panel: First telogen C57BL/6J mice with single s.c. mIgG/AK23 injection (day 0, P20) or Dsg3+/− or Dsg3−/− mice (day 0, P18). Bottom panel: Second telogen C57BL/6J mice with single s.c. mIgG/AK23 injection (day 0, P56) and average % of Ki-67+ cells per HFBu. Shown are mean ± SEM. Data analyzed by a two-tailed Student’s t test. ∗∗∗p < 0.001. (A and B) Dashed line indicates the HF. White arrows indicate blisters. Scale bars: 25 μm. (A–C): IFE, interfollicular epidermis, SG, sebaceous gland, HG, hair germ, DP, dermal papilla. (A, B, and D) per group, n ≥ 4 mice (except Dsg3 immunodetection mIgG/AK23, n = 3 mice).

Figure 2

AK23 binding activates quiescent BuSCs losing their stemness signature (A–C) Eight-week-old tetOFF H2BGFP mice after a 4-week chase and a single s.c. injection of mIgG/AK23. Label retaining cell (LRC) studies. (A) Experimental outline to assess GFP+ LRCs 1 and 2 days after mIgG/AK23 injection. (B) Immunodetection of GFP+ BuSCs and K6+ epithelial niche cells (red). Dashed line indicates the HF. AK23 treated mice already show fewer LRCs after 1 day. (C) FACS quantification of isolated back skin keratinocytes gated for LIVE/DEAD-/CD31-/CD45-/GFPdim and GFPbri (bright) cells. Graph shows reduced GFPbri LRCs. (D–H) Eight-week-old C57BL/6J mice after a single s.c. injection of mIgG/AK23. (D) Immunodetection using indicated antibodies and BrdU incorporation in the HFBu. Note hair shaft (green) represents unspecific staining. Dashed line indicates the HF. White arrows indicate blisters. (E) Quantification of Sox9 and NFATc1 positive cells per HFBu (>50 HFBus per group on random micrographs). Significances are indicated between the sum of positive HFBus in AK23 compared to mIgG mice. (F) FACS of primary viable LIVE/DEAD-/CD31-/CD45-keratinocytes measured 1 and 2 days after treatment, gated for CD34. mIgG set to 1. (G) RT-qPCR of HFBu markers assessed on back skin extracts relative to mIgG set to 1. (H) Quantification of colony forming efficiency (CFE) of primary keratinocytes isolated 1day after mIgG/AK23 injection from back skin. (A–H) per group, n ≥ 4 mice (except B n = 2), shown as mean ± SEM. Data analyzed by a two-tailed Student’s t test. ∗∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. (B and D) Scale bars: 25μm.

Figure 3

AK23 treated HFBu cells restore multipotency (A and B) Eight-week-old CAG-EGFP mice received 2 s.c. injections of high dose mIgG/AK23. (A) Epidermal reconstitution assay of SC lineages, experimental outline and macroscopic views. White hairs of the donor (CAG-EGFP; red arrows) are intermixed with black hairs of the pocket host (C57BL/6J; blue arrows) 25 and 77 days after grafting. Grafted cells on Swiss nude mice from mIgG and AK23 treated mice invariably grew white hair. Per group CAG-EGFP n = 2 mIgG; n = 3 AK23 mice; Swiss nude mice per group, n = 4 mice. See Figure S3C for complete set of mice. (B) Immunohistochemistry of GFP+ cells on paraffin-embedded skin biopsies shows grafted GFP+ keratinocytes were present in all lineages; in sebaceous glans (SG), HFs including the HFBu, and interfollicular epidermis (IFE). Scale bars: 50μm. (C–F) Eight-week-old C57BL/6J mice received 1 s.c. injection of mIgG/AK23. (C) Quantification of blistered HFBu (from Figure 1C), in gray, and colony forming efficiency (CFE) assay relative to mIgG, in purple. Per group for CFE mIgG or AK23, 1 day n = 1; 2–9 days n = 3 mice. (D) Graph represents quantitative immunoblot analyses of Dsg3 in Triton X-100 insoluble skin lysates (containing desmosomes) relative to mIgG set to 1. (E) RT-qPCR of Cd34 on back skin extracts relative to mIgG. (F) Immunodetection of HFBu marker CD34 and direct immunofluorescence (dIF) at indicated time points. Dashed line indicates the HF. White arrow indicates blister. Scale bars: 25μm. (D–F) per group, n ≥ 4 mice. (A, C, and D) Data analyzed by a two-tailed Student’s t test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. Graphs indicate mean ± SEM.

Figure 4

EGFR activation in the telogen HFBu (A–D) Eight-week-old C57BL/6J mice received a single s.c. injection of mIgG/AK23. (A) Quantification of p-Akt positive cells per HFBu (n = > 50 HFBus per group on random micrographs; significances are indicated between the sum of all p-Akt positive HFBus of AK23 compared to mIgG control) and immunodetection of EGFR and EGFR effectors, insets are lower exposure and zoomed in to show immunofluorescence pattern. (B) Immunoblot analyses from back skin lysates, relative to mIgG set to 1 (gray line). (C and D) EGFR inhibitor Lapatinib (LAP) treatment schemes, quantification of blistered HFBus and H&E. (A, C and D) Dashed line indicates the HF. White arrows indicate blisters. Scale bars: 25μm. (A–D) per group, n ≥ 4 mice (except D n ≥ 3), shown as mean ± SEM. Data analyzed by a two-tailed Student’s t test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 5

Wnt signaling in the telogen HFBu (A) Eight-week-old Axin2-LacZ mice received a single s.c. injection of mIgG/AK23. X-Gal quantification on skin sections of Axin2-LacZ mice using ImageJ. mIgG set to 1. (B–E) Eight-week-old C57BL/6J mice received a single s.c. injection of mIgG/AK23. (B) Immunodetection of Wnt effectors and adherens junction components, insets are lower exposure and zoomed in to show immunofluorescence pattern. Abbreviations: β-Cat (β-catenin), PG (plakoglobin), E-Cad (E-cadherin). (C) Quantification of LEF1 positive cells per HFBu (n = > 50 HFBus per group on random micrographs). Significances are indicated between the sum of all LEF1 positive HFBus of each AK23 time point compared to their respective mIgG controls. (D) RT-qPCR of E-cadherin (Cdh1) on back skin extracts relative to mIgG. (E) Wnt activator BIO treatment scheme, quantification of blistered HFBus and H&E. (A, B, and E) Dashed line indicates the HF. White arrows indicate blisters. Scale bars: 25μm. (A and E) per group, n ≥ 3 mice. (B–D) per group, n ≥ 4 mice. (A, D, and E) shown as mean ± SEM. Data analyzed by a two-tailed Student’s t test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 6

Delay in anagen entry/Shh suppression/Dsg3 governed signaling (A–C) Eight-week-old C57BL/6J mice received a single s.c. injection of mIgG/AK23. (A) Hair plucking: treatment scheme, macroscopic hair regrowth (left panel) and H&E (right panel). Hair re-growth (anagen entry) was delayed in AK23-treated mice. Scale bars: 50μm. (B) RT-qPCR of Bmp2 and Shh on back skin extracts relative to mIgG. (C) Shh agonist SAG treatment scheme, quantification of blistered HFBus and H&E. Dashed line indicates the HF. White arrows indicate blisters. Scale bars: 25μm. (D) First telogen Dsg3+/− or Dsg3−/− mice. RT-qPCR of Shh and Gli1 on back skin extracts relative to Dsg3+/− mice. (E) Schematic representation of Dsg3 protein levels in AK23 exposed HFBus. Relative activity of signaling pathways and nuclear proteins investigated in this study and consequences thereof. (A–D) per group, n ≥ 3 mice. (B, C, and D) shown as mean ± SEM. Data analyzed by a two-tailed Student’s t test. ∗p < 0.05, ∗∗p < 0.01.