Advertisement
Cell Stem Cell
This journal offers authors two options (open access or subscription) to publish research

dsRNA Released by Tissue Damage Activates TLR3 to Drive Skin Regeneration

      Highlights

      • dsRNA released after skin injury triggers skin regeneration via TLR3
      • IL-6 and Stat3 signaling, downstream mediators of TLR3, are key to regeneration
      • Loss of TLR3, IL-6Ra, or Stat3 reduces hair neogenesis after wounding
      • TLR3 induces core hair morphogenetic programs and hair follicle markers

      Summary

      Regeneration of skin and hair follicles after wounding—a process known as wound-induced hair neogenesis (WIHN)—is a rare example of adult organogenesis in mammals. As such, WIHN provides a unique model system for deciphering mechanisms underlying mammalian regeneration. Here, we show that dsRNA, which is released from damaged skin, activates Toll-Like Receptor 3 (TLR3) and its downstream effectors IL-6 and STAT3 to promote hair follicle regeneration. Conversely, TLR3-deficient animals fail to initiate WIHN. TLR3 activation promotes expression of hair follicle stem cell markers and induces elements of the core hair morphogenetic program, including ectodysplasin A receptor (EDAR) and the Wnt and Shh pathways. Our results therefore show that dsRNA and TLR3 link the earliest events of mammalian skin wounding to regeneration and suggest potential therapeutic approaches for promoting hair neogenesis.

      Graphical Abstract

      Introduction

      Animals across diverse phyla can regenerate lost structures, a capacity that is considerably more limited in mammals. Several chordate species, including urodele salamanders and teleost fish, can regenerate appendages and solid organs, yet among mammals such adult organogenesis is rarely, if ever, observed. An important exception is wound-induced hair neogenesis (WIHN), a phenomenon in which skin, sebaceous glands, and hair follicles are regenerated following large, full-thickness wounds in mice or rabbits (
      • Breedis C.
      Regeneration of hair follicles and sebaceous glands from the epithelium of scars in the rabbit.
      ,
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ). The complete regeneration observed in WIHN is in marked contrast to the fibrotic scarring that typically results from cutaneous wound healing. Regenerated hair follicles are complex mini-organs with disparate cell types, dedicated neurovascular support, and a distinct stem cell compartment located in the bulge region. These stem cells not only repopulate hair follicles throughout life, but also aid in skin re-epithelialization after wounding, pointing to the potential therapeutic relevance of WIHN (
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ). As WIHN represents a rare example of adult organogenesis in mammals, understanding its mechanisms could aid in efforts to regenerate other structures.
      While originally described in the 1940s, WIHN has recently been characterized in morphogenetic and molecular detail (
      • Breedis C.
      Regeneration of hair follicles and sebaceous glands from the epithelium of scars in the rabbit.
      ,
      • Gay D.
      • Kwon O.
      • Zhang Z.
      • Spata M.
      • Plikus M.V.
      • Holler P.D.
      • Ito M.
      • Yang Z.
      • Treffeisen E.
      • Kim C.D.
      • et al.
      Fgf9 from dermal γδ T cells induces hair follicle neogenesis after wounding.
      ,
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ,
      • Kligman A.M.
      • Strauss J.S.
      The formation of vellus hair follicles from human adult epidermis.
      ,
      • Myung P.S.
      • Takeo M.
      • Ito M.
      • Atit R.P.
      Epithelial Wnt ligand secretion is required for adult hair follicle growth and regeneration.
      ,
      • Nelson A.M.
      • Loy D.E.
      • Lawson J.A.
      • Katseff A.S.
      • Fitzgerald G.A.
      • Garza L.A.
      Prostaglandin D2 inhibits wound-induced hair follicle neogenesis through the receptor, Gpr44.
      ). Following complete excision of skin down to fascia, wounds on the backs of mice heal through initial contracture and then re-epithelialization. Subsequently, hair follicle morphogenesis ensues with recapitulation of events that occur during embryonic hair development. Formation and invagination of epithelial placodes in the epidermis, induction of adjacent dermal papillae, and ultimately, elaboration of distinct hair cell subtypes are observed (
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ). Follicle-associated structures such as sebaceous glands are also regenerated. Regenerated follicles transit through multiple hair cycles, just like neighboring hairs from unwounded skin (
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ). Therefore, WIHN represents functional regeneration rather than mere wound repair through scarring.
      Developmental pathways required for embryonic organogenesis can be reactivated following trauma. In axolotl limb regeneration, for example, Shh signaling is activated at the site of injury in the residual limb, much as it is induced in the zone of polarizing activity during limb development (
      • Torok M.A.
      • Gardiner D.M.
      • Izpisúa-Belmonte J.C.
      • Bryant S.V.
      Sonic hedgehog (shh) expression in developing and regenerating axolotl limbs.
      ). Similarly, during WIHN, signaling pathways utilized in embryonic hair formation reemerge after wounding. Activation of the canonical Wnt pathway is one of the earliest events observed in follicular morphogenesis. Wnt activation occurs around E15 in mice as the undifferentiated epithelium begins to condense into epithelial placodes at sites of future follicle formation (
      • Millar S.E.
      Molecular mechanisms regulating hair follicle development.
      ). Similarly, after cutaneous wounding, the Wnt ligand, Wnt10b, and the Wnt effector, Lef1, are induced after re-epithelialization is complete, but prior to the emergence of new follicles (
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ). Wnt pathway activation is critical for hair morphogenesis during both development and regeneration, as mice deficient in Wnt signaling fail to generate hairs (
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ,
      • Myung P.S.
      • Takeo M.
      • Ito M.
      • Atit R.P.
      Epithelial Wnt ligand secretion is required for adult hair follicle growth and regeneration.
      ). Secondary to Wnt activation during follicular development, Shh signaling is induced in epithelial placodes and underlying dermal papillae. Activation of the Shh pathway contributes to subsequent hair follicle invagination and morphogenesis (
      • St-Jacques B.
      • Dassule H.R.
      • Karavanova I.
      • Botchkarev V.A.
      • Li J.
      • Danielian P.S.
      • McMahon J.A.
      • Lewis P.M.
      • Paus R.
      • McMahon A.P.
      Sonic hedgehog signaling is essential for hair development.
      ). The Shh pathway is similarly induced during adult hair follicle regeneration. Other molecular details of hair regeneration are shared with hair development, including expression of the hair cytokeratin Krt17 and activation of alkaline phosphatase activity in dermal papillae (
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ).
      While downstream morphogenetic events in WIHN parallel those in hair development, the signals triggering reactivation of these programs in adult regeneration are unclear. To initiate regeneration, organisms must first sense a loss of tissue integrity. Candidate signals include molecules liberated from damaged tissues as well as mediators released by infiltrating immune cells. In newts and axolotls, activation of thrombin is a key early event in regeneration. For example, inhibition of thrombin activation abrogates lens regeneration in newts (
      • Imokawa Y.
      • Brockes J.P.
      Selective activation of thrombin is a critical determinant for vertebrate lens regeneration.
      ). Recently, it was demonstrated that FGF9 released from γδ T cells several days after wounding promotes hair regeneration in rodents (
      • Gay D.
      • Kwon O.
      • Zhang Z.
      • Spata M.
      • Plikus M.V.
      • Holler P.D.
      • Ito M.
      • Yang Z.
      • Treffeisen E.
      • Kim C.D.
      • et al.
      Fgf9 from dermal γδ T cells induces hair follicle neogenesis after wounding.
      ). However, the most proximal signals released by damaged keratinocytes to initiate regeneration in the skin remain unknown. Discovery of such damage-associated signals may explain why wound healing during WIHN proceeds through regeneration, whereas most cutaneous wound healing in mammals leads to fibrotic scarring. Identifying these molecules may also suggest therapeutic approaches to promote skin and hair regeneration and reduce fibrosis.
      To identify molecular events that initiate regeneration, we exploited the natural variation in follicle regenerative capacity observed in various mouse strains. Through gene expression screening of healed wounds prior to the onset of follicle regeneration, we identified the pattern recognition receptor, Toll-like Receptor 3 (TLR3), as a critical regulator of cutaneous regeneration. While TLR3 and its ligand, double-stranded RNA (dsRNA), are known to be active during cutaneous wounding, their role in promoting regeneration after re-epithelialization of the skin has not been appreciated (
      • Lin Q.
      • Wang L.
      • Lin Y.
      • Liu X.
      • Ren X.
      • Wen S.
      • Du X.
      • Lu T.
      • Su S.Y.
      • Yang X.
      • et al.
      Toll-like receptor 3 ligand polyinosinic:polycytidylic acid promotes wound healing in human and murine skin.
      ). We identified dsRNA released from damaged cells as a key trigger of the regeneration process through its activation of TLR3. The ensuing damage-induced signaling cascade impedes stratification and maintains keratinocytes in a less differentiated state. Furthermore, TLR3 activation initiates molecular events in the hair morphogenetic program, with activation of canonical Wnt and Shh pathways, and ectodysplasin A receptor (EDAR) resulting in augmented hair follicle neogenesis. Thus, TLR3 activation by dsRNA links damage sensing after wounding to the earliest molecular events in hair regeneration. These results uncover a role for TLR3 as a master regulator of regeneration in the skin.

      Results

      Hair follicle regeneration after wounding recapitulates embryonic follicle development in both morphogenetic and molecular detail. However, the events that dictate whether wound healing proceeds by regeneration or fibrotic scarring remain unclear. In studies characterizing molecular mechanisms of WIHN, we and others observed significant differences in the regenerative capacity of various mouse strains when visualized by confocal scanning laser microscopy (CSLM) (Figure 1A) (
      • Fan C.
      • Luedtke M.A.
      • Prouty S.M.
      • Burrows M.
      • Kollias N.
      • Cotsarelis G.
      Characterization and quantification of wound-induced hair follicle neogenesis using in vivo confocal scanning laser microscopy.
      ,
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ,
      • Nelson A.M.
      • Loy D.E.
      • Lawson J.A.
      • Katseff A.S.
      • Fitzgerald G.A.
      • Garza L.A.
      Prostaglandin D2 inhibits wound-induced hair follicle neogenesis through the receptor, Gpr44.
      ). To identify factors that may initiate regeneration, we compared gene expression profiles from healed wounds of mice with high and low regenerative capacity using C57BL/6 and our mixed background strain of mice (C57BL/6 × FVB × SJL). This analysis was performed at the time of wound closure but before the onset of hair morphogenesis to enrich for upstream factors in the WIHN pathway (Figure S1A). Ingenuity Pathway Analysis identified “viral pattern recognition receptors” and “interferon-signaling” as the most significantly upregulated pathways in highly regenerative mice.
      Figure thumbnail gr1
      Figure 1Tissue Damage and dsRNA activate TLR3 to Promote WIHN
      (A) CSLM images for C57BL6J (low regeneration [LR]) and Mixed B6/FVB/SJL (high regeneration [HR]) strains of mice. Area of WIHN is shown within red box. Original image size is 4 mm2.
      (B) Venn diagram depicting significant overlap between genes associated with high levels of follicle regeneration in mouse skin (in vivo) and human keratinocytes treated with poly (I:C) in vitro, published by
      • Karim R.
      • Meyers C.
      • Backendorf C.
      • Ludigs K.
      • Offringa R.
      • van Ommen G.J.
      • Melief C.J.
      • van der Burg S.H.
      • Boer J.M.
      Human papillomavirus deregulates the response of a cellular network comprising of chemotactic and proinflammatory genes.
      under GEO: GSE21260.
      (C) Mean fold change in TLR3 mRNA in healed scars at WD20-24 in LR versus HR mice, as determined by qRT-PCR and normalized to housekeeping gene β-actin.
      (D) Mean fold change in TLR3 mRNA 4 hr after scratch assay in NHEK, as determined by qRT-PCR and normalized to housekeeping gene RPLP0.
      (E) WIHN levels in WT mice after standard straight cut or “fringe cut” to wound edge, n = 14–15 mice. Area of WIHN is shown within red box. Original image size is 4 mm2.
      (F) Regenerated hair shafts (white, arrows) in healed wounds after single injection of poly (I:C) (500 ng) or control at WD3 and visualized by dissecting microscope at ∼WD58–62.
      (G) Cross-sectional H&E histology through the middle of healed scar at WD22 after single injection of poly (I:C) as in (F). Regenerated hair follicles are marked with arrows. Scale bar represents 500 μm.
      (H) WIHN levels in WT mice after poly (I:C) (500 ng) or PBS control measured by CSLM, n = 10–11 mice.
      (I) WIHN levels in WT mice after RNase III (15 units) or buffer control measured by CSLM, n = 17–19 mice.
      (J) WIHN levels in strain-matched WT control mice and TLR3 KO mice measured by CSLM, n = 6 mice.
      (K) WIHN in TLR3 KO mice after poly (I:C) (500ng) compared with PBS control measured by CSLM, n = 9 mice.
      p < 0.05 by Student’s t test or single factor ANOVA.

      dsRNA Released by Tissue Damage Activates TLR3 to Promote Regeneration

      We focused on the pattern recognition receptor TLR3, which is activated by dsRNA and known to induce interferon signaling (
      • Uematsu S.
      • Akira S.
      Toll-like receptors and Type I interferons.
      ). The gene expression pattern we observed in highly regenerative murine skin wounds showed strong overlap with the pattern obtained from human keratinocytes treated with the synthetic dsRNA mimic, poly (I:C) (Figure 1B) (
      • Karim R.
      • Meyers C.
      • Backendorf C.
      • Ludigs K.
      • Offringa R.
      • van Ommen G.J.
      • Melief C.J.
      • van der Burg S.H.
      • Boer J.M.
      Human papillomavirus deregulates the response of a cellular network comprising of chemotactic and proinflammatory genes.
      ). Strikingly, despite the differences in species and experimental conditions, 25 of the 200 most highly upregulated genes were common to both analyses (Figure S1B). This observed overlap in expression of genes involved in dsRNA-sensing suggested a potentially conserved role for TLR3 in early wound healing responses. Furthermore, expression of TLR3 itself was 3-fold higher in our highly regenerative mouse strain, as observed by qRT-PCR, validating the expression patterns observed in the array analyses (Figure 1C).
      In previous studies, TLR3 mRNA was induced in response to dsRNA released after UVB irradiation (
      • Bernard J.J.
      • Cowing-Zitron C.
      • Nakatsuji T.
      • Muehleisen B.
      • Muto J.
      • Borkowski A.W.
      • Martinez L.
      • Greidinger E.L.
      • Yu B.D.
      • Gallo R.L.
      Ultraviolet radiation damages self noncoding RNA and is detected by TLR3.
      ). This suggested to us that during WIHN, TLR3 may serve as a sensor of tissue damage, consistent with an upstream role in the regeneration process. To test this idea, we examined TLR3 expression following scratching of human keratinocytes in culture. TLR3 expression was nearly 5-fold higher in scratched keratinocytes compared with unmanipulated controls (Figure 1D). Furthermore, we observed a significant increase in the number of regenerated follicles in vivo when we increased the extent of damage during wounding by placing minute perpendicular cuts at the wound edge (Figure 1E). We next explored whether augmenting the natural dsRNA release during wounding could lead to an increase in regeneration. Indeed, a single addition of the dsRNA mimic, poly (I:C), into murine skin wounds led to a greater number of regenerated follicles (Figures 1F–1H). Conversely, addition of the dsRNA-specific endonuclease, RNase III, significantly decreased the number of regenerated follicles (Figure 1I). To confirm that the effects of dsRNA on WIHN are TLR3 dependent, we next examined the extent of regeneration in TLR3 null mice. Strikingly, regeneration was almost completely abolished in these mice compared with strain-matched controls, despite their comparable re-epithelialization kinetics (Figures 1J and S1C). Moreover, the stimulatory effect of dsRNA on WIHN was abrogated in TLR3 null mice, demonstrating the necessity of TLR3 for damage-induced regeneration (Figure 1K). The effects of dsRNA were only observed in the context of regeneration, as poly (I:C) did not affect the hair cycle in normal non-wounded murine skin (Figure S1D). Taken together, these data suggest that TLR3 activation by dsRNA released during wounding initiates regeneration.

      IL-6 and pSTAT3 Are Induced by Tissue Damage and dsRNA

      To examine the mechanism by which TLR3 promotes regeneration, we performed gene expression analysis on mice at 16 to 18 days post-wounding (WD16–18), later than the array above, and at the earliest time points at which regenerated follicles can be detected by CSLM (Figure S2A). Gene expression from healed wound beds of animals with robust regeneration (average 49 hair follicles) was compared with those of animals that failed to regenerate hair follicles (zero hair follicles), revealing upregulation of several interleukins and cytokines in more highly regenerative mice. Of particular interest were interleukin-6 (IL-6) and its pathway components as well as TLR3 itself, which appeared as the top upstream regulator of IL-6 in this analysis (Figures S2B–S2F). Just as we had found for TLR3, mixed strain animals with high regenerative capacity had 3-fold higher levels of IL-6 compared with C57BL/6 mice with poor regeneration (Figure 2A). These data led to the hypothesis that IL-6 may mediate the effects of TLR3 on regeneration. TLR3 has previously been demonstrated to induce IL-6 in a dsRNA-dependent manner (
      • Melkamu T.
      • Kita H.
      • O’Grady S.M.
      TLR3 activation evokes IL-6 secretion, autocrine regulation of Stat3 signaling and TLR2 expression in human bronchial epithelial cells.
      ), and IL-6 is a known activator of regeneration in other contexts, particularly in response to liver damage (
      • Galun E.
      • Rose-John S.
      The regenerative activity of interleukin-6.
      ,
      • Jia C.
      Advances in the regulation of liver regeneration.
      ). Consistent with this, just as TLR3 expression is increased in injured (scratched) keratinocytes in culture, IL-6 mRNA is also increased (Figure 2B). In keratinocytes treated with poly (I:C), we observed a greater than 30-fold induction of IL-6 mRNA (Figure 2C), which is partially mediated through the downstream transcription factor, NF-κB (Figure S2G). This induction is TLR3 dependent, as TLR3 null animals had far less IL-6 mRNA after wounding than strain-matched controls (Figure 2C). Temporally, IL-6 mRNA and protein were sequentially upregulated at early time points following wounding, consistent with a role for this pathway in initiating WIHN (Figure 2D).
      Figure thumbnail gr2
      Figure 2IL-6 and pSTAT3 Are Induced by Tissue Damage and dsRNA
      (A) Mean fold change in IL-6 mRNA in healed scars at WD20-24 in HR versus LR strains of mice, as determined by qRT-PCR and normalized to housekeeping gene β-actin.
      (B) Mean fold change in IL-6 mRNA 24 hr post scratch assay in NHEK, as determined by qRT-PCR and normalized to housekeeping gene, RPLP0.
      (C) Mean fold change in IL-6 mRNA after poly (I:C) addition (20 μg/ml) to NHEK for 6 hr or in strain-matched WT and TLR3 KO mice (n = 3) 6 hr after wounding as determined by qRT-PCR and normalized to housekeeping gene RPLP0 (NHEK) or β-actin (mouse).
      (D) Time course of IL-6 mRNA and protein expression throughout early stage wound healing in WT mice, as determined by qRT-PCR and ELISA, respectively. n = 3 mice per time point.
      (E) IL-6 (middle) and P-STAT3 (right; arrows) immunohistochemistry of healing scars at WD5 (top) and WD8 (bottom) in WT mice. Scale bar represents 50 μm.
      (F) P-STAT 3 levels in NHEKs after poly (I:C) (20ug/mL) for 24 hr compared with vehicle control, as measured by western blot and normalized to STAT3.
      p < 0.05 by Student’s t test or single factor ANOVA.
      IL-6 receptor engagement is known to cause phosphorylation of STAT3 (pSTAT3), leading to its nuclear translocation and transcriptional activation (
      • Heinrich P.C.
      • Behrmann I.
      • Haan S.
      • Hermanns H.M.
      • Müller-Newen G.
      • Schaper F.
      Principles of interleukin (IL)-6-type cytokine signalling and its regulation.
      ). Consistent with a role for TLR3 and IL-6 signaling in WIHN initiation, we observed increased IL-6 and pSTAT3 protein in murine keratinocytes at WD5 and WD8 (Figures 2E and S3), as well as increased pSTAT3 protein expression in human keratinocytes following poly (I:C) treatment (Figure 2F).

      TLR3 Effects on Regeneration Are Mediated by IL-6 and pSTAT3

      To test the functional consequences of IL-6 on follicle regeneration, we injected recombinant IL-6 protein (rmIL-6) into mice following wounding and examined subsequent hair follicle regeneration. Compared with vehicle injected controls, mice receiving IL-6 had a nearly 3-fold increase in the number of regenerated follicles (Figures 3A–3C). The increase in follicle numbers upon activation of the IL-6/STAT3 axis occurred only in the context of WIHN, as dsRNA does not stimulate anagen onset during follicular cycling in unwounded skin (Figure S1D). We also noted a marked reduction in WIHN in mice whose keratinocytes lacked the IL-6 receptor alpha (K14CreERT2-IL-6Rαfl/fl) (Figure 3D), highlighting the importance of the IL-6 signaling pathway during WIHN.
      Figure thumbnail gr3
      Figure 3IL-6 and pSTAT3 Mediate TLR3 Effects on WIHN
      (A) Cross-sectional H&E histology of healed scar at WD22 after a single injection of IL-6 (25 ng) or PBS control at WD7. Regenerated hair follicles are marked with arrows. Scale bar represents 100 μm.
      (B) Regenerated hair shafts (white, arrows) at ∼WD58–62, as visualized by dissecting microscope.
      (C) WIHN in WT mice after single dose of IL-6 (25 ng) compared with PBS control, as measured by CSLM, n = 30 mice.
      (D) WIHN in keratinocyte-specific knockout of IL-6Receptorα compared to control mice measured by CSLM, n = 3–6 mice.
      (E) P-Stat3 levels in the presence of cucurbitacin I (+) compared to control (−) in WT mice as measured by western blot and normalized to Stat3. PC is P-STAT3-positive control cell lysate.
      (F) WIHN levels in WT mice after cucurbitacin I ( 2mg/kg) or control, as measured by CSLM, n = 10–14 mice.
      (G) WIHN in keratinocyte-specific knockout of Stat3 compared with control mice measured by CSLM, n = 10–15 mice.
      (H) WIHN in TLR3 KO mice after single dose of rmIL-6 (500 ng) compared with PBS control measured by CSLM, n = 8–10 mice.
      p < 0.05 by Student’s t test or single factor ANOVA.
      We next blocked the IL-6/STAT3 signaling axis with cucurbitacin I, a highly selective pharmacological inhibitor of STAT3 (
      • Blaskovich M.A.
      • Sun J.
      • Cantor A.
      • Turkson J.
      • Jove R.
      • Sebti S.M.
      Discovery of JSI-124 (cucurbitacin I), a selective Janus kinase/signal transducer and activator of transcription 3 signaling pathway inhibitor with potent antitumor activity against human and murine cancer cells in mice.
      ). Cucurbitacin I strongly suppressed STAT3 phosphorylation in vivo (Figure 3E), and WT mice injected with cucurbitacin I had a greater than 3-fold decrease in the number of regenerated follicles (Figure 3F). Similarly, WIHN in keratinocyte-specific Stat3 KO mice (K5CreERT2-Stat3fl/fl) was significantly decreased compared with control mice (Figure 3G). Remarkably, the previously observed attenuation of WIHN in TLR3 KO mice was rescued by a single injection of recombinant IL-6 protein (Figure 3H), implying that IL-6 functions downstream of TLR3 in follicle regeneration. In aggregate, these data suggest that TLR3 activation during wounding promotes IL-6 production and STAT3 phosphorylation resulting in higher regeneration.

      TLR3 Activation Alters Keratinocyte Differentiation and Induces Markers of Hair Follicle Progenitors

      Previous studies revealed that keratinocytes outside of the bulge region, which ordinarily differentiate into corneocytes during stratification, contribute to regenerated hair follicles during WIHN (
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ,
      • Snippert H.J.
      • Haegebarth A.
      • Kasper M.
      • Jaks V.
      • van Es J.H.
      • Barker N.
      • van de Wetering M.
      • van den Born M.
      • Begthel H.
      • Vries R.G.
      • et al.
      Lgr6 marks stem cells in the hair follicle that generate all cell lineages of the skin.
      ). This finding implies that the normal stratification program is altered during hair follicle regeneration. We hypothesized that TLR3 activation during WIHN may counter stratification of keratinocytes.
      To test this idea, we first injected mice with IL-6 and assessed epidermal thickness as an index of stratification. Ordinarily, stratification causes thickening of the epidermis due to the accumulation of differentiated keratinocytes. However, mice treated with IL-6 had a 2-fold reduction in epidermal thickness compared with strain-matched controls, implying decreased stratification (Figure 4A) because no alterations in apoptosis or proliferation were observed with IL-6 treatment (Figure S4A). Consistent with our findings in vivo, direct TLR3 activation with poly (I:C) in cultured keratinocytes led to a nearly complete loss of expression of markers of keratinocyte differentiation such as keratin 1 (KRT1) and filaggrin (FLG) (Figure 4B). This effect was TLR3 dependent as inhibition of TLR3 through siRNA-mediated depletion or direct small molecule-based antagonism abrogated the loss of KRT1 (Figures 4C, 4D, and S4B). Similarly, keratinocytes treated with IL-6 had a profound decrease in KRT1 expression, an effect that was reversed by the addition of cucurbitacin I (Figure 4E). These data suggest that induction of the TLR3/IL-6 axis during wounding prevents keratinocyte differentiation. Interestingly, the morphology of poly (I:C)-treated keratinocytes was also altered and came to resemble that of keratinocytes in healing human skin wounds (Figure S4C–S4H) (
      • Yan C.
      • Grimm W.A.
      • Garner W.L.
      • Qin L.
      • Travis T.
      • Tan N.
      • Han Y.P.
      Epithelial to mesenchymal transition in human skin wound healing is induced by tumor necrosis factor-alpha through bone morphogenic protein-2.
      ).
      Figure thumbnail gr4
      Figure 4TLR3 Activation Alters Keratinocyte Differentiation and Induces Markers of Hair Follicle Progenitors
      (A) Cross-sectional H&E histology through healed scars treated with IL-6 (25 ng) or control (PBS) at WD7. Scale bar represents 100 μm. Quantification of healed epidermal thickness in healed scars after control or IL-6 addition is shown.
      (B) Mean fold change in KRT1 and FLG mRNA after poly (I:C) (20 μg/ml) addition to NHEK for 24 hr as determined by qRT-PCR and normalized to housekeeping gene, RPLP0.
      (C) Mean fold change in KRT1 mRNA with TLR3-specific or scrambled control siRNA in the presence of poly (I:C) (20 μg/ml) in NHEK as determined by qRT-PCR and normalized as in (B).
      (D) Mean fold change in KRT1 mRNA with TLR3-specific inhibitor or control in the presence of poly (I:C) (20 μg/ml) in NHEK as determined by qRT-PCR and normalized as in (B).
      (E) Mean fold change in KRT1 mRNA after IL-6 (50 ng/ml) +/− cucurbitacin I in NHEK for 24 hr as determined by qRT-PCR and normalized as in (B).
      (F) Mean fold change in Wnt pathway genes, Lgr5 and Lgr6, mRNA after 6 days of continuous exposure to poly (I:C) determined by qRT-PCR and normalized to housekeeping gene, RPLP0.
      (G) Mean fold change in KRT15 mRNA 72 hr after 24 hr of poly (I:C) (20 μg/ml) treatment to NHEK as determined by qRT-PCR normalized to housekeeping gene, RPLP0.
      (H) Flow cytometry analysis of KRT15 protein expression of NHEKs treated as in (G).
      (I) Flow cytometry analysis of CD200 protein expression of NHEKs treated as in (G).
      p < 0.05 by Student’s t test or single factor ANOVA.
      We next examined whether in response to TLR3 activation keratinocytes adopt a less differentiated state that is permissive for subsequent hair follicle differentiation. Following treatment with poly (I:C), keratinocytes expressed markers associated with hair progenitors, including leucine-rich-repeat containing G proteins (LGR) 5 and 6 (Figure 4F) (
      • Tadeu A.M.
      • Horsley V.
      Epithelial stem cells in adult skin.
      ). Induction of LGR6 is particularly notable, as previous lineage tracing experiments demonstrated that LGR6 expressing cells contribute to regenerating hair follicles during WIHN (
      • Snippert H.J.
      • Haegebarth A.
      • Kasper M.
      • Jaks V.
      • van Es J.H.
      • Barker N.
      • van de Wetering M.
      • van den Born M.
      • Begthel H.
      • Vries R.G.
      • et al.
      Lgr6 marks stem cells in the hair follicle that generate all cell lineages of the skin.
      ). Further, we noticed induction of genes associated with hair follicle stem cells upon TLR3 activation. Hair follicle stem cells reside in the bulge region of the follicle and express keratin 15 (KRT15), which is considered the most reliable marker of this population (
      • Liu Y.
      • Lyle S.
      • Yang Z.
      • Cotsarelis G.
      Keratin 15 promoter targets putative epithelial stem cells in the hair follicle bulge.
      ), and CD200, which is associated with hair follicle progenitor cells (
      • Garza L.A.
      • Yang C.C.
      • Zhao T.
      • Blatt H.B.
      • Lee M.
      • He H.
      • Stanton D.C.
      • Carrasco L.
      • Spiegel J.H.
      • Tobias J.W.
      • Cotsarelis G.
      Bald scalp in men with androgenetic alopecia retains hair follicle stem cells but lacks CD200-rich and CD34-positive hair follicle progenitor cells.
      ). We found significantly increased expression of KRT15 mRNA in keratinocytes upon activation of TLR3 with poly (I:C) (Figure 4G). Addition of poly (I:C) nearly doubled the percentage of KRT15-expressing cells and led to a 5-fold increase in number of cells expressing CD200, as assessed by fluorescence-activated cell sorting (FACS) (Figure 4H and 4I).Taken together, these data suggest that TLR3 pathway activation maintains keratinocytes in a less differentiated state and induces the expression of genes associated with hair follicle progenitors.

      Hair Follicle Morphogenetic Pathways Are Induced by TLR3 Signaling

      We investigated whether these TLR3 activated keratinocytes—with their increased expression of hair progenitor markers—would be poised for subsequent activation of the hair follicle morphogenetic program. Core to this program are the Shh and Wnt pathways, which are activated during both embryonic hair follicle formation and in regeneration following wounding (
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ). First, we examined β-catenin translocation to the nucleus, one of the earliest events in canonical Wnt signaling (
      • Barker N.
      The canonical Wnt/beta-catenin signalling pathway.
      ). TLR3 activation with poly (I:C) induced peri-nuclear accumulation and also doubled the amount of nuclear β-catenin in keratinocytes, consistent with activation of the Wnt pathway (Figures 5A and 5B ). In addition, expression of the Wnt ligand, Wnt 7b, and the downstream Wnt effector/target gene, LEF1, were upregulated following poly (I:C) treatment of keratinocytes (Figures 5C–5E) (
      • Barker N.
      • Clevers H.
      Leucine-rich repeat-containing G-protein-coupled receptors as markers of adult stem cells.
      ,
      • Tadeu A.M.
      • Horsley V.
      Epithelial stem cells in adult skin.
      ). Similarly, expression of Shh pathway components SHH and GLI1 was increased following poly (I:C) addition, as was the expression of EDAR, another gene active in skin appendage formation (Figure 5E). These pathways were stably induced for several days despite a transient 24-hr treatment of keratinocytes with poly (I:C), suggesting that TLR3 activation may prime keratinocytes toward a hair follicle or appendage fate. Importantly, the pathway activation we observed is TLR3 dependent, as pretreatment of cells with a specific TLR3 small molecule antagonist markedly reduced the expression of both LEF1 and SHH (Figure 5F). A similar dependence on TLR3 was observed in vivo where TLR3 KO mice had decreased expression of β-catenin, Lef1, Gli2, Shh, and Edar following wounding when compared with strain-matched controls (Figure 5G).
      Figure thumbnail gr5
      Figure 5TLR3 Activation Induces Hair Follicle Morphogenic Program Markers
      (A) β-Catenin immunofluorescence staining in NHEK after 72 hr of 24 hr treatment with poly (I:C) (20 μg/ml) or control.
      (B) Quantitation of nuclear β-catenin to total levels of β-catenin in NHEK as in (A).
      (C) WNT7b immunofluorescence staining (green) after 7 days of continuous poly (I:C) (20 μg/ml) or vehicle control treatment to NHEK. Scale bar represents 50 μm; original magnification is 40×.
      (D) Mean fold change in Wnt7b mRNA after 6 days of continuous exposure to poly (I:C) (20 μg/ml) determined by qRT-PCR and normalized to housekeeping gene, RPLP0.
      (E) Mean fold change in LEF1, GLI1, SHH, and EDAR mRNA after poly (I:C) treatment by qRT-PCR as in (A).
      (F) Mean fold change in LEF1 and SHH mRNA with TLR3-specific inhibitor or control in the presence of poly (I:C) in NHEK as determined by qRT-PCR as in (A).
      (G) Mean fold change in Lef1, Edar, Gli2, Shh, and β-catenin mRNA in TLR3 KO mouse wounds compared with strain-matched control mice as determined by qRT-PCR.
      p < 0.05 by Student’s t test or single factor ANOVA.
      Finally, we explored a mechanism by which TLR3 pathway activation promotes induction of the hair morphogenetic program. Expression of β-catenin and GLI2 mRNA was significantly increased in keratinocytes upon poly (I:C) addition (Figure 6A). Since our data demonstrated that stimulation of TLR3 promotes STAT3 activation, we examined whether STAT3 binding sites are present in the promoters of these Wnt and Shh pathway genes. We identified several consensus binding sites. Next, using ChIP-qPCR, we demonstrated a significant increase in STAT3 occupancy at sites in both the β-catenin and GLI2 promoters upon poly (I:C) treatment in keratinocytes (Figures 6B and 6C). These data suggest that TLR3 pathway activation during wounding may lead to direct transcriptional activation of genes involved in hair follicle morphogenesis.
      Figure thumbnail gr6
      Figure 6TLR3 Activation Increases STAT3 Occupancy of β-Catenin and GLI2 Promoters
      (A) Mean fold change in β-catenin and GLI2 mRNA after continuous exposure to poly (I:C) at indicated time points as determined by qRT-PCR and normalized to housekeeping gene, RPLP0. Fold changes in mRNA expression relative to pre-confluent keratinocytes.
      (B) Relative fold enrichment of STAT3 occupation of β-catenin and GLI2 promoter sites after poly (I:C) treatment of keratinocytes. Negative binding sites are also included below. Data are representative of five independent experiments, p < 0.05.
      (C) Graphical representation of verified positive and negative STAT3 binding on β-catenin and GLI2 after poly (I:C) treatment of keratinocytes. Images were obtained from the ENCODE database.
      p < 0.05 by Student’s t test or single factor ANOVA.

      Discussion

      dsRNA Are Damage-Associated Signals that Promote Regeneration

      While a capacity for regeneration is observed in representatives of almost all animal phyla, its distribution is far from uniform, with some species demonstrating regeneration of multiple body parts while closely related species fail to do so (
      • Brockes J.P.
      • Kumar A.
      • Velloso C.P.
      Regeneration as an evolutionary variable.
      ). Urodele salamanders, for example, are well known to regenerate their limbs, yet among 24 urodele species examined, 4 failed to reconstitute limbs after amputation (
      • Brockes J.P.
      • Kumar A.
      • Velloso C.P.
      Regeneration as an evolutionary variable.
      ). Even within a single species, differences in genetic background can lead to marked differences in regenerative ability. We and others found that the capacity to regenerate skin and hair follicles, a process termed WIHN, varies greatly among different strains of Mus musculus (Figure 1) (
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ,
      • Nelson A.M.
      • Loy D.E.
      • Lawson J.A.
      • Katseff A.S.
      • Fitzgerald G.A.
      • Garza L.A.
      Prostaglandin D2 inhibits wound-induced hair follicle neogenesis through the receptor, Gpr44.
      ). While large variation between strains is challenging and requires very careful strain-matched and transgenic controls for all experiments, it is also an opportunity to explore the causes of this biological variation. We harnessed this variation and examined early time points following wounding to search for early, pivotal events that link tissue damage to regeneration.
      For regeneration to occur, three interrelated events must take place: (1) organisms must sense loss of tissue integrity, (2) precursor cells must be mobilized to reconstitute missing structures, and (3) these cells must be directed along appropriate morphogenetic pathways (
      • Brockes J.P.
      • Kumar A.
      • Velloso C.P.
      Regeneration as an evolutionary variable.
      ). While the latter two processes have been extensively examined in studies of regeneration, less is known about how organisms sense damage and transduce this information to trigger a regenerative response. In hydra, the peptide head activator (HA) is secreted at sites of tissue damage and is required for regeneration (
      • Sánchez Alvarado A.
      Planarian regeneration: its end is its beginning.
      ). In salamanders and newts, an unidentified, thrombin-activated serum factor initiates regeneration of both the limb and lens (
      • Brockes J.P.
      • Kumar A.
      • Velloso C.P.
      Regeneration as an evolutionary variable.
      ,
      • Imokawa Y.
      • Brockes J.P.
      Selective activation of thrombin is a critical determinant for vertebrate lens regeneration.
      ). No such triggers had been discovered in the rare examples of mammalian epimorphic regeneration.
      In the context of WIHN, we identified dsRNA released by damaged cells as early molecular signals triggering regeneration. Several lines of evidence support this: dsRNA responsive pathways are upregulated in mice with a high capacity for regeneration, addition of exogenous dsRNA increases the number of regenerated follicles, and degradation of endogenous dsRNA inhibits regeneration (Figure 1). In previous work, dsRNA was shown to be released upon UV-induced damage to keratinocytes (
      • Bernard J.J.
      • Cowing-Zitron C.
      • Nakatsuji T.
      • Muehleisen B.
      • Muto J.
      • Borkowski A.W.
      • Martinez L.
      • Greidinger E.L.
      • Yu B.D.
      • Gallo R.L.
      Ultraviolet radiation damages self noncoding RNA and is detected by TLR3.
      ). Further, dsRNA has been shown to accelerate re-epithelialization in small wounds of both mice and humans (
      • Lin Q.
      • Wang L.
      • Lin Y.
      • Liu X.
      • Ren X.
      • Wen S.
      • Du X.
      • Lu T.
      • Su S.Y.
      • Yang X.
      • et al.
      Toll-like receptor 3 ligand polyinosinic:polycytidylic acid promotes wound healing in human and murine skin.
      ), suggesting that they play an important, early role in the response to cutaneous wounding. Also, TLR3 has been shown to increase proinflammatory cytokine accumulation after wounding in a manner counterbalanced by cutaneous microflora (
      • Lai Y.
      • Di Nardo A.
      • Nakatsuji T.
      • Leichtle A.
      • Yang Y.
      • Cogen A.L.
      • Wu Z.R.
      • Hooper L.V.
      • Schmidt R.R.
      • von Aulock S.
      • et al.
      Commensal bacteria regulate Toll-like receptor 3-dependent inflammation after skin injury.
      ). While these studies demonstrate that the dsRNA-TLR3 pathway is active early in wound healing, its consequences for events after wound closure have not been examined. Our results demonstrate that dsRNA initiates key events in the regeneration process following re-epithelialization.
      A major receptor for dsRNA in mammalian cells is TLR3. While originally identified for its role in dorsal-ventral patterning in Drosophila and its response to viral pathogens, recent evidence has emerged that TLR3 plays a role in cutaneous wound healing. TLR3-defeicient animals have a decreased inflammatory response to wounding (
      • Lai Y.
      • Di Nardo A.
      • Nakatsuji T.
      • Leichtle A.
      • Yang Y.
      • Cogen A.L.
      • Wu Z.R.
      • Hooper L.V.
      • Schmidt R.R.
      • von Aulock S.
      • et al.
      Commensal bacteria regulate Toll-like receptor 3-dependent inflammation after skin injury.
      ,
      • Lebre M.C.
      • van der Aar A.M.
      • van Baarsen L.
      • van Capel T.M.
      • Schuitemaker J.H.
      • Kapsenberg M.L.
      • de Jong E.C.
      Human keratinocytes express functional Toll-like receptor 3, 4, 5, and 9.
      ). TLR3 is activated by mRNAs released from dying cells, linking its activation to tissue damage (
      • Karikó K.
      • Ni H.
      • Capodici J.
      • Lamphier M.
      • Weissman D.
      mRNA is an endogenous ligand for Toll-like receptor 3.
      ). We find that TLR3 is activated in response to cutaneous wounding in mice as TLR3 mRNA is strongly induced, an effect that can be augmented by administration of exogenous dsRNA. Downstream TLR3 pathway components, including IL-6, are also strongly induced by dsRNA. The activation of IL-6 is critical as it can rescue the defects in hair regeneration observed in TLR3-deficient animals. The early and strong induction of TLR3 and its pathway components upon wounding, coupled with the role of dsRNA in stimulating hair follicle neogenesis, suggest that TLR3 may relay information about tissue damage to activate regeneration. Of note, healed wounds of TLR3 KO mice also have significantly fewer γδ T cells than WT mice (Figure S5A). These cells have been demonstrated to augment WIHN (
      • Gay D.
      • Kwon O.
      • Zhang Z.
      • Spata M.
      • Plikus M.V.
      • Holler P.D.
      • Ito M.
      • Yang Z.
      • Treffeisen E.
      • Kim C.D.
      • et al.
      Fgf9 from dermal γδ T cells induces hair follicle neogenesis after wounding.
      ), although we find WIHN does occur in mice lacking T and B cells, suggesting that they are not absolutely necessary (Figure S5B). Intriguingly, downstream events induced by TLR3—including possibly recruitment of γδ T cells—appear to differ between humans and mice (
      • Lundberg A.M.
      • Drexler S.K.
      • Monaco C.
      • Williams L.M.
      • Sacre S.M.
      • Feldmann M.
      • Foxwell B.M.
      Key differences in TLR3/poly I:C signaling and cytokine induction by human primary cells: a phenomenon absent from murine cell systems.
      ). It will be interesting to examine whether differences in TLR3 responses account for the greater regeneration of skin wounds in mice compared with humans.

      TLR3 Activation Increases Markers of Hair Follicle Progenitors

      In response to damage, organisms must recruit precursor cells to rebuild lost structures. During both physiologic hair cycling and WIHN, KRT15 expressing stem cells of the bulge are mobilized and differentiate into multiple constituent cells of hair follicles. However, during WIHN, cells of the interfollicular epidermis, including those expressing Lgr5 and Lgr6, also contribute to regenerated hair follicles (
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ,
      • Snippert H.J.
      • Haegebarth A.
      • Kasper M.
      • Jaks V.
      • van Es J.H.
      • Barker N.
      • van de Wetering M.
      • van den Born M.
      • Begthel H.
      • Vries R.G.
      • et al.
      Lgr6 marks stem cells in the hair follicle that generate all cell lineages of the skin.
      ,
      • Tadeu A.M.
      • Horsley V.
      Epithelial stem cells in adult skin.
      ). We find an increase in markers of both types of hair progenitor cells upon activation of the TLR3 pathway. In keratinocytes isolated from interfollicular skin, we observed induction of Lgr6 and KRT15 following TLR3 pathway activation with dsRNA.
      These results are in accord with recent findings on the roles of TLR3 and the IL6/pSTAT3 axis in hair follicle and stem cell biology. For example, our observed induction of hair follicle progenitor markers by TLR3 activation is consistent with recent findings on increased pSTAT3 in KRT15-positive hair follicle stem cells during aging (
      • Doles J.
      • Storer M.
      • Cozzuto L.
      • Roma G.
      • Keyes W.M.
      Age-associated inflammation inhibits epidermal stem cell function.
      ). Futhermore, TLR3 has been implicated in the reprogramming of fibroblasts to IPS cells using virally encoded reprogramming factors (
      • Lee J.
      • Sayed N.
      • Hunter A.
      • Au K.F.
      • Wong W.H.
      • Mocarski E.S.
      • Pera R.R.
      • Yakubov E.
      • Cooke J.P.
      Activation of innate immunity is required for efficient nuclear reprogramming.
      ). Activation of TLR3 by dsRNA during wounding may similarly promote the conversion of keratinocytes destined to form stratified epidermis into cells with increased capacity for hair morphogenesis.

      TLR3 Activation Initiates Hair Morphogenesis

      The final event in regeneration is the reactivation of embryonic morphogenetic programs to direct mobilized stem cells to form missing structures. Hair follicle morphogenesis in the developing embryo proceeds through epithelial-mesenchymal crosstalk between the undifferentiated epithelium and the underlying dermis. Our data provide the first physiologic role for TLR3 in Wnt and Shh pathway activation during regeneration, likely through promoting this crosstalk. As with Wnt and Shh signaling, we find that EDAR pathway components are also activated in response to TLR3 signaling in vitro and in vivo. Activation of these appendage specification signals by dsRNA is TLR3 dependent since TLR3 chemical inhibition in vitro or TLR3 gene deletion in vivo blunt Wnt and Shh pathway induction. Finally, the IL-6/STAT3 axis directly links TLR3 and these pathways since dsRNA increases occupancy of STAT3 at the promoters of β-Catenin and Gli2. Given the importance in hair development of in vivo epithelial-mesenchymal crosstalk to amplify EDAR, Wnt, and Shh signaling (
      • Millar S.E.
      Molecular mechanisms regulating hair follicle development.
      ), it is notable that we can detect induction of these pathways with keratinocytes alone. We hypothesize these signals will be enhanced in the presence of competent fibroblasts. These findings for TLR3 initiating morphogenesis are consistent with the original description of Toll receptors as regulators of dorsal ventral patterning in Drosophila (
      • Anderson K.V.
      • Bokla L.
      • Nüsslein-Volhard C.
      Establishment of dorsal-ventral polarity in the Drosophila embryo: the induction of polarity by the Toll gene product.
      ). Together with our findings, this suggests that Toll receptors have an equally important role in tissue specification in addition to their more well-known roles in innate immune activation.
      In summary, we identified the activation of TLR3 by damage induced dsRNA as the linchpin of the regenerative response to murine skin wounds. Current methods where damage induces rejuvenation (laser resurfacing, dermabrasion, chemical peels) might work through similar mechanisms in humans. Strikingly, TLR3 plays a role in all three aspects of regeneration—damage sensing, precursor recruitment, and activation of hair follicle morphogenesis. As such, TLR3 agonists may be powerful therapeutics to decrease fibrosis and promote cutaneous regeneration.

      Experimental Procedures

      WIHN

      All animal protocols are approved by the Johns Hopkins University Animal Care and Use Committee. C57BL/6J, B6;129SF2/J, TLR3 null mice (B6;129S1-Tlr3tm1Flv/J and B6N.129S1-Tlr3tm1Flv/J), Nod-Scid-Gamma (NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ), and IL-6Rαfl/fl (B6;SJL-Il6ratm1.1Drew/J) were obtained from The Jackson Laboratory. K5-Ert2-Cre and K14-Ert2-Cre mice were provided by Pierre Chambon. Stat3fl/fl mice were kindly provided by Cynthia Sears (Johns Hopkins University; Stat3tm2Aki;
      • Takeda K.
      • Kaisho T.
      • Yoshida N.
      • Takeda J.
      • Kishimoto T.
      • Akira S.
      Stat3 activation is responsible for IL-6-dependent T cell proliferation through preventing apoptosis: generation and characterization of T cell-specific Stat3-deficient mice.
      ), and mixed-strain (C57BL/6J × FVB/N × SJL/J) animals were provided by Dr. Jean Richa (University of Pennsylvania).
      A 1-cm2 excisional full-thickness wound to the level of skeletal muscle on the backs of 21-day-old male and female mice was performed as previously described (
      • Ito M.
      • Yang Z.
      • Andl T.
      • Cui C.
      • Kim N.
      • Millar S.E.
      • Cotsarelis G.
      Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
      ,
      • Nelson A.M.
      • Loy D.E.
      • Lawson J.A.
      • Katseff A.S.
      • Fitzgerald G.A.
      • Garza L.A.
      Prostaglandin D2 inhibits wound-induced hair follicle neogenesis through the receptor, Gpr44.
      ). Numbers of regenerated hair follicles were quantified in the re-epithelialized skin by non-invasive CSLM, as published (
      • Fan C.
      • Luedtke M.A.
      • Prouty S.M.
      • Burrows M.
      • Kollias N.
      • Cotsarelis G.
      Characterization and quantification of wound-induced hair follicle neogenesis using in vivo confocal scanning laser microscopy.
      ). For all experiments, 50 μl of “intervention” was injected into healing wound (under scab) or applied topically into open wound as shown in the Table 1.
      Table 1Mouse Strain, Genotype, Number, Intervention, and Schedule Used for WIHN Trials
      ExperimentMouse StrainNumber of MiceInterventionDay of InterventionDay of CSLM Assessment
      High versus low gene expression–early; high versus low gene expression–lateC57 versus C57 × FVB × SJL C57 × FVB × SJL4 per strain; 3 per groupnonewound closure; ∼WD20-24
      Standard WIHN versus fringe CutsC57BL/6J14–15 per group10 cuts per sideWD0∼WD20-24
      Exogeneous poly (I:C) additionC57 × FVB × SJL; B6;129S1-Tlr3tm1Flv/J10–11 per group; 9 per group500 ng Poly IC injected into woundWD3∼WD20-24
      Rnase III additionC57 × FVB × SJL17–19 per group15 units Rnase III injected into woundWD3∼WD20-24
      WIHN in TLR3 KOB6;129S1-Tlr3tm1Flv/J; B6;129SF2/J6 per groupnone∼WD20-24
      Exogeneous IL-6 additionC57BL/6J; B6N.129S1-Tlr3tm1Flv/J30 per group; 8–10 per group25 ng rmIL-6 protein injected into wound; 500 ng rmIL-6 protein injected into woundWD7∼WD20-24
      Importance of IL-6RαK14-ERT2-Cre × IL-6Ralpha fl/fl (both C57BL/6)3–6 per groupintraperitoneal tamoxifin every other dayWD5-WD14∼WD20-24
      Cucurbitacin IC57 × FVB × SJL10–14 per group2 mg/kg cucurbitacin I injected into woundWD7∼WD20-24
      Importance of Stat3 in WIHNK5-ERT2-Cre × Stat3 fl/fl (both C57BL/6)10–15 per groupI.P. tamoxifin every other dayWD0-WD14∼WD20-24
      Role of T and B cellsNOD.Cg-PrkdcscidIl2rgtm1Wjl (NOD/ShiLt)5none∼WD20-24

      Cell Culture

      Neo-natal human epidermal keratinocytes (Lonza) or lab-isolated foreskin keratinocytes were cultured in keratinocyte medium with added supplements (KGM-GOLD). Treatment with recombinant IL-6 protein (50 ng/ml), cucurbitacin I (10–100 nM), poly (I:C) (20 μg/ml), and TLR3 pharmacological inhibitor (80 μM; EMD Millipore) was applied in basal medium containing transferrin, hydrocortisone, and antibiotics for up to 24 hr. After 24 hr, treatment medium was replaced with KGM-GOLD and isolation of RNA as indicated. In some experiments, poly (I:C) was applied for up to a week, with replenishment of poly (I:C) and medium every other day.

      Nucleofection

      Nucleofection with siGENOME SMARTpool Human TLR3, REL-A, and siCONTROL siRNA duplex oligonucleotides (Dharmacon-ThermoFischer Scientific) was performed in NHEK using the Amaxa 4D-Nucleofector according to manufacturer’s instruction. Plated cells were treated with recombinant human IL-6 (50 ng/ml) protein or poly(I:C) (20 μg/ml) for 24 hr. Afterward, treatment medium was removed and replaced with KGM-GOLD complete medium for the duration of the experiment. Levels of appropriate gene expression were assessed by qRT-PCR using inventoried TaqMan reagents in three independent experiments.

      Gene Expression Analysis

      RNA from immediately re-epithelialized skin at ∼12 days after wounding (early stage) or after the earliest time point of hair follicle detection by CSLM (late stage; ∼16 days) was submitted to the JHMI Deep Sequencing & Microarray core for Affymetrix Mouse Exon 1.0ST microarray chips according to manufacturer’s protocols. Raw gene expression signals in the form of Affymetrix CEL files were extracted and normalized with Partek Genomics Suite software using the Robust Multichip Analysis (RMA) algorithm (
      • Irizarry R.A.
      • Bolstad B.M.
      • Collin F.
      • Cope L.M.
      • Hobbs B.
      • Speed T.P.
      Summaries of Affymetrix GeneChip probe level data.
      ). The Student’s t test ANOVA was used to detect genes with significantly different expression. These analyses have been submitted to the GEO database (under GEO: GSE50418 and GEO: GSE50419; http://www.ncbi.nlm.nih.gov/geo/).

      qRT-PCR

      Mouse skin was harvested prior to wounding and throughout wounding as described (
      • Nelson A.M.
      • Loy D.E.
      • Lawson J.A.
      • Katseff A.S.
      • Fitzgerald G.A.
      • Garza L.A.
      Prostaglandin D2 inhibits wound-induced hair follicle neogenesis through the receptor, Gpr44.
      ). RNA was isolated from NHEK with RNeasy Mini Kit (QIAGEN) with DNase I digestion followed by conversion to cDNA using the High Capacity RNA-to-cDNA kit (Life Technologies). qRT-PCR was performed for genes of interest and 18S or ribosomal protein, large P0 (RPLP0) (housekeeping genes) using inventoried TaqMan reagents. Differences in gene expression were assessed by comparative ΔΔCT values with fold change calculations.

      ELISA

      IL-6 protein levels were assayed by ELISA (R&D Systems) from non-wounded and wounded skin or healed mouse scars at times indicated. A minimum of three independent mice was used for each time point.

      Immunohistochemistry, Immunocytochemistry, and Histology

      Immunohistochemistry was performed on formalin-fixed paraffin-embedded mouse skin samples using the avidin-biotin complex method and AEC development (Vector Laboratories). Indicated antibodies were applied overnight. Sections were counterstained with hematoxylin. Images were captured at 40× magnification using a Nikon Optiphot microscope and Nikon Elements F software. Histology was assessed by H&E after IL-6 addition. The epidermal thickness from the basal layer keratinocytes to beginning of stratum corneum in three locations per healed mouse wound in multiple histology sections was measured by ImageJ software.
      Immunocytochemistry was performed on NHEKs plated on plated on collagen-coated coverslips and treated with 20 μg/ml poly (I:C) as above. Fixed cells were incubated with primary antibodies overnight and appropriate Alexa Fluor secondary antibodies and were counterstained using VectaShield DAPI mounting medium (Vector Labs). Slides were imaged at 60× magnification using the Nikon C1si True Spectral Imaging Confocal Laser Scanning Microscope system (Cell Imaging Core Facility, Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins). Cell morphology and beta-catenin nuclear localization were quantified using the CellProfiler image analysis software (http://www.cellprofiler.org) (
      • Carpenter A.E.
      • Jones T.R.
      • Lamprecht M.R.
      • Clarke C.
      • Kang I.H.
      • Friman O.
      • Guertin D.A.
      • Chang J.H.
      • Lindquist R.A.
      • Moffat J.
      • et al.
      CellProfiler: image analysis software for identifying and quantifying cell phenotypes.
      ) from confocal images of nuclei.

      Flow Cytometry

      Keratinocytes and fibroblasts were fixed, permeabilized (BD Cytofix/Cytoperm kit), and stained with antibodies against human vimentin (BD PharMingen clone RV202), KRT15 (Abam clone LHK15) labeled with a chromophore preconjugated to Fab (Zenon mouse IgG labeling kit) or human CD200 (BD Biosciences; 552475). Data were collected on a dual-laser flow cytometer (BD FACSCalibur) followed by FlowJo 10 (TreeStar) software analysis.
      To measure TCRγδ expression, healed wounds (∼WD20) from WT and TLRKO3 mice were minced and digested at 37°C in a buffer containing RPMI 1640, 1.67 collagenase Wunsch units/ml Liberase TL (Roche Life Sciences), and 0.01% DNase (Sigma-Aldrich) for 75 min. Following digestion, samples were washed and filtered (40 μm) to obtain a single-cell suspension. Cells were stained with propidium iodide (Miltenyi Biotec) and TCRγδ (GL3) antibody (Miltenyi Biotec) followed by analysis with MACSQuant cytometer and FlowJo software.

      Chromatin Immunoprecipitation

      Poly (I:C)-treated and control keratinocytes were crosslinked in 1% formaldehyde for 10 min, followed by addition of glycine for 5 min to quench unreacted formaldehyde. Cells were processed with EZ-ChIP Kit (Millipore) according to the manufacturer’s instructions. Cross-linked protein-DNA complexes were captured with rabbit anti-Stat3 or normal rabbit IgG (sc-482X; sc-2027X, SCBT) antibodies. qRT-PCR was performed to determine the relative abundance of the promoter DNA sequence, associated with Stat3. Primers are detailed in Supplemental Experimental Methods. Primers and graphics were designed based on ENCODE data (UCSC Genome Browser).

      Statistical Analysis

      Each experiment was repeated with at least three independent litters of animals or keratinocyte cultures. Data were analyzed using Student’s t test or ANOVA single factor. Statistical significance was considered at p < 0.05.

      Author Contributions

      A.M.N., S.K.R., and L.A.G. designed the studies, analyzed, and interpreted the results and co-wrote the paper. L.S.M. provided assistance in interpreting results and planning experiments. In addition, T.S.R., M.Z.H., A.S.K., A.S.Z., E.C., S.R.R., C.P., D.K., and A.J.W. assisted A.M.N. and S.K.R. in performing laboratory experiments. All members discussed results and helped formulate ongoing conduct of the project.

      Acknowledgments

      The authors thank Conover Talbot Jr. (JHU Microarray Core) for assistance with microarray analysis; Lillian Dasko-Vincent, Cell Imaging Core Facility for assistance with immunocytochemistry analysis; and Dr. Pierre Coulombe and his laboratory for critical discussions. The authors thank Pierre Chambon for use of the K14-Ert2-Cre and K5-Ert2-Cre mouse lines. Research reported in this publication was supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases, part of the NIH, under Award Number F32AR062932 to A.M.N. and R01AR064297 to L.A.G. This work was also supported by the Department of Defense, Armed Forces Institute of Regenerative Medicine, Extremities Regeneration (AFIRM2-ER11), Northrop Grumman Electronic Systems, and Alliance for Veterans Support, Inc. (Veteran/Amputee Skin Regeneration Program Initiative), as well as the Thomas Provost, MD Young Faculty Development Fund of Johns Hopkins Dermatology to L.A.G. Johns Hopkins is the owner of a patent application with A.M.N. and L.A.G. as inventors on the use of dsRNA and pathway to promote regeneration and hair follicle neogenesis.

      Supplemental Information

      References

        • Anderson K.V.
        • Bokla L.
        • Nüsslein-Volhard C.
        Establishment of dorsal-ventral polarity in the Drosophila embryo: the induction of polarity by the Toll gene product.
        Cell. 1985; 42: 791-798
        • Barker N.
        The canonical Wnt/beta-catenin signalling pathway.
        Methods Mol. Biol. 2008; 468: 5-15
        • Barker N.
        • Clevers H.
        Leucine-rich repeat-containing G-protein-coupled receptors as markers of adult stem cells.
        Gastroenterology. 2010; 138: 1681-1696
        • Bernard J.J.
        • Cowing-Zitron C.
        • Nakatsuji T.
        • Muehleisen B.
        • Muto J.
        • Borkowski A.W.
        • Martinez L.
        • Greidinger E.L.
        • Yu B.D.
        • Gallo R.L.
        Ultraviolet radiation damages self noncoding RNA and is detected by TLR3.
        Nat. Med. 2012; 18: 1286-1290
        • Blaskovich M.A.
        • Sun J.
        • Cantor A.
        • Turkson J.
        • Jove R.
        • Sebti S.M.
        Discovery of JSI-124 (cucurbitacin I), a selective Janus kinase/signal transducer and activator of transcription 3 signaling pathway inhibitor with potent antitumor activity against human and murine cancer cells in mice.
        Cancer Res. 2003; 63: 1270-1279
        • Breedis C.
        Regeneration of hair follicles and sebaceous glands from the epithelium of scars in the rabbit.
        Cancer Res. 1954; 14: 575-579
        • Brockes J.P.
        • Kumar A.
        • Velloso C.P.
        Regeneration as an evolutionary variable.
        J. Anat. 2001; 199: 3-11
        • Carpenter A.E.
        • Jones T.R.
        • Lamprecht M.R.
        • Clarke C.
        • Kang I.H.
        • Friman O.
        • Guertin D.A.
        • Chang J.H.
        • Lindquist R.A.
        • Moffat J.
        • et al.
        CellProfiler: image analysis software for identifying and quantifying cell phenotypes.
        Genome Biol. 2006; 7: R100
        • Doles J.
        • Storer M.
        • Cozzuto L.
        • Roma G.
        • Keyes W.M.
        Age-associated inflammation inhibits epidermal stem cell function.
        Genes Dev. 2012; 26: 2144-2153
        • Fan C.
        • Luedtke M.A.
        • Prouty S.M.
        • Burrows M.
        • Kollias N.
        • Cotsarelis G.
        Characterization and quantification of wound-induced hair follicle neogenesis using in vivo confocal scanning laser microscopy.
        Skin Res. Technol. 2011; 17: 387-397
        • Galun E.
        • Rose-John S.
        The regenerative activity of interleukin-6.
        Methods Mol. Biol. 2013; 982: 59-77
        • Garza L.A.
        • Yang C.C.
        • Zhao T.
        • Blatt H.B.
        • Lee M.
        • He H.
        • Stanton D.C.
        • Carrasco L.
        • Spiegel J.H.
        • Tobias J.W.
        • Cotsarelis G.
        Bald scalp in men with androgenetic alopecia retains hair follicle stem cells but lacks CD200-rich and CD34-positive hair follicle progenitor cells.
        J. Clin. Invest. 2011; 121: 613-622
        • Gay D.
        • Kwon O.
        • Zhang Z.
        • Spata M.
        • Plikus M.V.
        • Holler P.D.
        • Ito M.
        • Yang Z.
        • Treffeisen E.
        • Kim C.D.
        • et al.
        Fgf9 from dermal γδ T cells induces hair follicle neogenesis after wounding.
        Nat. Med. 2013; 19: 916-923
        • Heinrich P.C.
        • Behrmann I.
        • Haan S.
        • Hermanns H.M.
        • Müller-Newen G.
        • Schaper F.
        Principles of interleukin (IL)-6-type cytokine signalling and its regulation.
        Biochem. J. 2003; 374: 1-20
        • Imokawa Y.
        • Brockes J.P.
        Selective activation of thrombin is a critical determinant for vertebrate lens regeneration.
        Curr. Biol. 2003; 13: 877-881
        • Irizarry R.A.
        • Bolstad B.M.
        • Collin F.
        • Cope L.M.
        • Hobbs B.
        • Speed T.P.
        Summaries of Affymetrix GeneChip probe level data.
        Nucleic Acids Res. 2003; 31: e15
        • Ito M.
        • Yang Z.
        • Andl T.
        • Cui C.
        • Kim N.
        • Millar S.E.
        • Cotsarelis G.
        Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding.
        Nature. 2007; 447: 316-320
        • Jia C.
        Advances in the regulation of liver regeneration.
        Expert Rev. Gastroenterol. Hepatol. 2011; 5: 105-121
        • Karikó K.
        • Ni H.
        • Capodici J.
        • Lamphier M.
        • Weissman D.
        mRNA is an endogenous ligand for Toll-like receptor 3.
        J. Biol. Chem. 2004; 279: 12542-12550
        • Karim R.
        • Meyers C.
        • Backendorf C.
        • Ludigs K.
        • Offringa R.
        • van Ommen G.J.
        • Melief C.J.
        • van der Burg S.H.
        • Boer J.M.
        Human papillomavirus deregulates the response of a cellular network comprising of chemotactic and proinflammatory genes.
        PLoS ONE. 2011; 6: e17848
        • Kligman A.M.
        • Strauss J.S.
        The formation of vellus hair follicles from human adult epidermis.
        J. Invest. Dermatol. 1956; 27: 19-23
        • Lai Y.
        • Di Nardo A.
        • Nakatsuji T.
        • Leichtle A.
        • Yang Y.
        • Cogen A.L.
        • Wu Z.R.
        • Hooper L.V.
        • Schmidt R.R.
        • von Aulock S.
        • et al.
        Commensal bacteria regulate Toll-like receptor 3-dependent inflammation after skin injury.
        Nat. Med. 2009; 15: 1377-1382
        • Lebre M.C.
        • van der Aar A.M.
        • van Baarsen L.
        • van Capel T.M.
        • Schuitemaker J.H.
        • Kapsenberg M.L.
        • de Jong E.C.
        Human keratinocytes express functional Toll-like receptor 3, 4, 5, and 9.
        J. Invest. Dermatol. 2007; 127: 331-341
        • Lee J.
        • Sayed N.
        • Hunter A.
        • Au K.F.
        • Wong W.H.
        • Mocarski E.S.
        • Pera R.R.
        • Yakubov E.
        • Cooke J.P.
        Activation of innate immunity is required for efficient nuclear reprogramming.
        Cell. 2012; 151: 547-558
        • Lin Q.
        • Wang L.
        • Lin Y.
        • Liu X.
        • Ren X.
        • Wen S.
        • Du X.
        • Lu T.
        • Su S.Y.
        • Yang X.
        • et al.
        Toll-like receptor 3 ligand polyinosinic:polycytidylic acid promotes wound healing in human and murine skin.
        J. Invest. Dermatol. 2012; 132: 2085-2092
        • Liu Y.
        • Lyle S.
        • Yang Z.
        • Cotsarelis G.
        Keratin 15 promoter targets putative epithelial stem cells in the hair follicle bulge.
        J Invest. Dermatol. 2003; 121: 963-968
        • Lundberg A.M.
        • Drexler S.K.
        • Monaco C.
        • Williams L.M.
        • Sacre S.M.
        • Feldmann M.
        • Foxwell B.M.
        Key differences in TLR3/poly I:C signaling and cytokine induction by human primary cells: a phenomenon absent from murine cell systems.
        Blood. 2007; 110: 3245-3252
        • Melkamu T.
        • Kita H.
        • O’Grady S.M.
        TLR3 activation evokes IL-6 secretion, autocrine regulation of Stat3 signaling and TLR2 expression in human bronchial epithelial cells.
        J. Cell Commun. Signal. 2013; 7: 109-118
        • Millar S.E.
        Molecular mechanisms regulating hair follicle development.
        J. Invest. Dermatol. 2002; 118: 216-225
        • Myung P.S.
        • Takeo M.
        • Ito M.
        • Atit R.P.
        Epithelial Wnt ligand secretion is required for adult hair follicle growth and regeneration.
        J. Invest. Dermatol. 2013; 133: 31-41
        • Nelson A.M.
        • Loy D.E.
        • Lawson J.A.
        • Katseff A.S.
        • Fitzgerald G.A.
        • Garza L.A.
        Prostaglandin D2 inhibits wound-induced hair follicle neogenesis through the receptor, Gpr44.
        J. Invest. Dermatol. 2013; 133: 881-889
        • Sánchez Alvarado A.
        Planarian regeneration: its end is its beginning.
        Cell. 2006; 124: 241-245
        • Snippert H.J.
        • Haegebarth A.
        • Kasper M.
        • Jaks V.
        • van Es J.H.
        • Barker N.
        • van de Wetering M.
        • van den Born M.
        • Begthel H.
        • Vries R.G.
        • et al.
        Lgr6 marks stem cells in the hair follicle that generate all cell lineages of the skin.
        Science. 2010; 327: 1385-1389
        • St-Jacques B.
        • Dassule H.R.
        • Karavanova I.
        • Botchkarev V.A.
        • Li J.
        • Danielian P.S.
        • McMahon J.A.
        • Lewis P.M.
        • Paus R.
        • McMahon A.P.
        Sonic hedgehog signaling is essential for hair development.
        Curr. Biol. 1998; 8: 1058-1068
        • Tadeu A.M.
        • Horsley V.
        Epithelial stem cells in adult skin.
        Curr. Top. Dev. Biol. 2014; 107: 109-131
        • Takeda K.
        • Kaisho T.
        • Yoshida N.
        • Takeda J.
        • Kishimoto T.
        • Akira S.
        Stat3 activation is responsible for IL-6-dependent T cell proliferation through preventing apoptosis: generation and characterization of T cell-specific Stat3-deficient mice.
        J. Immunol. 1998; 161: 4652-4660
        • Torok M.A.
        • Gardiner D.M.
        • Izpisúa-Belmonte J.C.
        • Bryant S.V.
        Sonic hedgehog (shh) expression in developing and regenerating axolotl limbs.
        J. Exp. Zool. 1999; 284: 197-206
        • Uematsu S.
        • Akira S.
        Toll-like receptors and Type I interferons.
        J. Biol. Chem. 2007; 282: 15319-15323
        • Yan C.
        • Grimm W.A.
        • Garner W.L.
        • Qin L.
        • Travis T.
        • Tan N.
        • Han Y.P.
        Epithelial to mesenchymal transition in human skin wound healing is induced by tumor necrosis factor-alpha through bone morphogenic protein-2.
        Am. J. Pathol. 2010; 176: 2247-2258