Wound-Induced Hair Neogenesis: A Portal to the Development of New Therapies for Hair Loss and Wound Regeneration

Abstract

Adult mammals retain the remarkable ability to regenerate hair follicles after wounding. Wound-induced hair neogenesis (WIHN) in many ways recapitulates embryogenesis. The origin of the stem cells that give rise to a nascent hair follicle after wounding and the role of mesenchymal cells and signaling pathways responsible for this regenerative phenomenon are slowly being elucidated. WIHN provides a potential therapeutic window for manipulating cell fate by the introduction of factors during the wound healing process to enhance hair follicle formation.

Scar formation, characterized by fibrosis, often comes at the expense of regeneration. Regenerative capacity also diminishes when organisms develop increased immune capabilities. For example, clawed frogs (Xenopus spp.) are able to regenerate their limbs as tadpoles, but they lose this ability after metamorphosis (Lin et al. 2013). Similarly in humans, there is decreased regenerative potential in adult tissue when compared to that in fetal tissue in most cases (Julier et al. 2017). However, there is some emerging evidence that scarring may diminish with advanced age (Nishiguchi et al. 2018).

The loss of skin appendages, such as hair follicles, in adults was previously thought to be permanent following their complete destruction (Needham 1953). However, regeneration of hair follicles and sebaceous glands in adult rabbits following full-thickness excision down to the panniculus carnosus was documented as early as 1954 (Breedis 1954). This finding was recapitulated in human adults when new vellus hair follicles were noted in cheeks after dermabrasion for removal of acne scars (Kligman and Strauss 1956). More recently, this phenomenon of de novo hair follicle regeneration was confirmed and studied more rigorously by characterizing that (1) new hair follicles (Ito et al. 2007), as well as adipocytes (Plikus et al. 2017), form in the center of large (>1 cm² in 3-wk-old mice and >2.5 cm² in mice older than 7 wk of age) full-thickness wounds, but not in smaller wounds; (2) the regenerative process for these follicles mimic stages of hair follicle development during embryogenesis; and (3) newly formed follicles express markers of follicular differentiation and are able to progress fully through a normal hair cycle (Ito et al. 2007). On post-wound days 12–15, prior to appearance of new hair follicles, increased expression of Wnt10b and Shh, was identified. Sebaceous glands associated with newly formed hair follicles were observed, but melanocytes and arrector pili muscle were not.

De novo hair follicle regeneration after wounding, or wound-induced hair neogenesis (WIHN), is remarkable in its regenerative property in an innately suppressive tissue environment. Thus, obtaining a better mechanistic insight of WIHN may yield therapeutic approaches that strike a favorable balance between scar formation and regeneration.

ROLE OF DISTINCT CELLULAR POPULATIONS

Epithelial Stem Cell Populations

Epithelial stem cells possess long-term self-renewing capabilities, and they are ultimately responsible for maintaining the hair follicle and interfollicular epidermis (IFE) during homeostasis (Cotsarelis 2006; Blanpain and Fuchs 2014). In symmetric cell division, a stem cell divides to generate two additional stem cells or two cells destined to differentiate. In asymmetric cell division, an additional stem cell and a cell destined to differentiate are generated after mitotic division instead. Most mammalian stem cells can divide via symmetric or asymmetric division, and this decision is influenced by both developmental and environmental cues (Morrison and Kimble 2006).

Epidermal stratification in utero is thought to be driven by asymmetric cell division dependent on polarity. A study of transgenic mice expressing keratin 14 (K14)-centrin coupled to green fluorescent protein (GFP) revealed that 92% of divisions were symmetric in a predominantly single-layered epithelium on embryonic day (E) 12.5. However, >70% of divisions from E15.5 were asymmetric, noted by mitotic spindles orthogonal to the basement membrane, following increased stratification of the epithelium (Lechler and Fuchs 2005). The end product of asymmetric division was a proliferative basal cell and a committed suprabasal cell. Keratin 5 (K5) and keratin 14 were found in mitotically active basal layer in the epidermis (Alam et al. 2011). Thus, squamous epithelium stratification is driven by asymmetric cell division.

In a type of pulse-chase experiment designed to identify quiescent stem cell populations, nucleoside analogs (e.g., bromodeoxyuridine or BrdU) are administered to neonatal mice. They are then followed during a “chase” period when rapidly cycling cells dilute out their label, and only quiescent cells, so called “label-retaining” cells (LRCs), retain label. Prior work using this method identified LRCs, presumed to be stem cells, in the IFE basal layer (Potten 1974; Braun and Watt 2004). LRCs in the IFE were previously thought to represent the sole stem cell population in this region, and non-LRCs were thought to be progenitor cells. More recently, Sada and colleagues demonstrated that both LRCs and non-LRCs act as distinct stem cell populations and are comprised of ∼35% and ∼65% of the basal layer, respectively. These LRC and non-LRC populations were molecularly distinct, as demonstrated by respective up-regulations of Distal-Less Homeobox 1 (DLX1) or Solute Carrier Family 1 Member 3 (SLC1A3). Both populations exhibited regenerative capability following injury (Sada et al. 2016).

The identification of stem cells in a hair follicle was perhaps more challenging because of its cyclical nature and structural complexity. A hair follicle cycles through anagen (growing phase), catagen (regressing or transition phase), and telogen (resting phase). In humans, anagen lasts for several years, catagen lasts for ∼2 wk, and telogen for a few months (Kligman 1959). In mice, these phases last 1–3 wk, 2 d, and 2 wk, respectively (Müller-Röver et al. 2001). During the anagen phase, transiently amplifying progenitor (matrix) cells in the bulb divide to form the companion layer, inner root sheath (IRS), and hair shaft. These matrix cells are in contact with the dermal papilla, the follicular mesenchyme that lies at the center of the bulb (Oliver and Jahoda 1988).

Cotsarelis et al. first identified LRCs in the hair follicle bulge, where the arrector pili muscle attaches to the follicle (Cotsarelis et al. 1990). The bulge marks the lower end of the isthmus, and keratinocytes below this region either undergo apoptosis or differentiate during the catagen phase. Pulse-chase experiments showed that bulge cells transform into secondary germ cells during the end of catagen, but secondary hair germ cells were also able to transform into bulge cells following a depilation-induced injury (Ito et al. 2004). RAC1, a small GTPase of the Rho family, was noted to be required for hair follicle development, but not for IFE development. Therefore, IFE and follicular epithelium were posited to be replenished from different stem cell lineages distinguishable by their requirement of RAC1 (Castilho et al. 2007).

An increasing cohort of stem cells that exhibit multilineage differentiation has been identified during the last three decades, and they reside in different compartments of the hair follicle and the IFE. This is a robust area of research and the exact relationship between these stem cells in different niches seems quite complex. Notable markers of stem cell niches in the IFE, sebaceous gland, infundibulum, isthmus, bulge, and secondary hair germ are listed on Table 1. It was initially thought that Lgr6 was a specific marker of self-renewing cells in the isthmus that contribute to the basal progenitor cells in the hair follicle, IFE, and the sebaceous gland (Snippert et al. 2010), but a more recent study by Füllgrabe, Kasper, and colleagues provided evidence that each of these compartments appear to house their own resident long-term self-renewing Lgr6⁺ cells (Füllgrabe et al. 2015). Although the hair follicle and the IFE are predictably replenished by distinct stem cell niches during homeostasis, there appears to be a greater degree of plasticity during injury (Rompolas et al. 2013). Laser ablation of the bulge or the hair germ resulted in hair regeneration. Following bulge ablation, the secondary hair germ encompassed the dermal papilla and regrew the bulge by day 7 (Rompolas et al. 2013). Thus, the bulge and the secondary hair germ were mutually dispensable for hair follicle regeneration induced by injury. Furthermore, transplantation of colonies grown from sorted bulge stem cells was able to regrow the IFE and the entire hair follicle in hairless mice (Blanpain et al. 2004; Morris et al. 2004; Claudinot et al. 2005; Blanpain and Fuchs 2014). Interestingly, a connection between the epithelium and the mesenchymal dermal papilla remained as a requirement for normal hair follicle regeneration after injury (Rompolas et al. 2013). This review will primarily focus on the role of different stem cell populations in the setting of WIHN (Fig. 1) because their roles in normal hair cycling are reviewed elsewhere (Cotsarelis 2006; Daszczuk et al. 2020).

Table 1.

Markers of stem cell populations derived from the hair follicle and the interfollicular epidermis (IFE)

Compartment	Markers
IFE	Cytokeratins K5/K14 (Alam et al. 2011) Surface markers Lgr6 (Liao and Nguyen 2014)
Infundibulum	Surface markers Lrig1 (Jensen et al. 2009; Wang et al. 2013)
Isthmus	Surface markers Lgr6 (Snippert et al. 2010; Füllgrabe et al. 2015) Lrig1 (Jensen et al. 2009; Wang et al. 2013) MTS24/Plet1 (Nijhof et al. 2006; Raymond et al. 2010) Transcription factors Gli1 (Brownell et al. 2011)
Bulge	Cytokeratin K15 (Lyle et al. 1998; Liu et al. 2003; Morris et al. 2004) K19 (Michel et al. 1996) Surface markers CD34 (Trempus et al. 2003) in murine cells, but not human cells CD200 (Ohyama et al. 2006; Garza et al. 2011) Lgr5 (Jaks et al. 2008) Transcription factors FOXC1 (Lay et al. 2016) Gli1 (Brownell et al. 2011) SOX9 (Nowak et al. 2008) Lhx2 (Mardaryev et al. 2011) TCF3 (Merrill et al. 2001) TCF4 (Mardaryev et al. 2011) NFATc1 (Horsley et al. 2008) Other markers S100A4/S100A6 (Ito and Kizawa 2001; Ito et al. 2004)
Sebaceous gland	Transcription factors Blimp1 (Horsley et al. 2006) Surface markers Lgr6 (Page et al. 2013; Füllgrabe et al. 2015) Lrig1 (Jensen et al. 2009; Wang et al. 2013)
Secondary hair germ	Cytokeratin K15 (Morris et al. 2004) K19 (Ghali et al. 2004) Surface markers CD200 (Garza et al. 2011) Lgr5 (Jaks et al. 2008) Transcription factors Gli1 (Brownell et al. 2011) Lef1 (Greco et al. 2009)

(FOXC1) Forkhead box C1, (K5) keratin 5, (LGR) leucine-rich repeat containing G protein–coupled receptor, (Lhx2) Lim-homeodomain transcription factor, (LRIG1) leucine-rich repeats and immunoglobulin-like domain protein 1, (NFATc1) nuclear factor of activated T cells c1, (Plet1) placenta-expressed transcript 1, (SOX9) SRY-box transcription factor 9, (TCF3) transcription factor 3, (TCF4) transcription f actor 4.

Figure 1.

De novo hair follicle regeneration from various stem cell niches. After wounding, bulge stem cells rapidly shunt to the area of depletion, but their contribution decreases over time. Lgr6⁺ stem cells from the isthmus generate long-lasting clones in the basal interfollicular epidermis (IFE) layer and contribute to new hair follicles. IFE and infundibular stem cells also significantly contribute to wound-induced hair neogenesis (WIHN). Sm22 and Sma-positive myofibroblasts up-regulate expression of adipogenic commitment factors, such as ZFP423 and pCEBPb, and differentiate into follicle-associated adipocytes during WIHN.

Initial studies demonstrated that hair follicles typically only form in the center of large full-thickness wounds and not in small wounds. After wounding, bulge stem cells rapidly shunted to the area of depletion, but their contribution markedly dwindled over time (Ito et al. 2007). The role of K15⁺ bulge cells was also studied using transgenic mice expressing herpes simplex virus thymidine kinase paired to a K15 promoter. Administration of ganciclovir generated a toxic analog that selectively ablated K15⁺ bulge cells (Ito et al. 2005). Although hair follicles were destroyed, the IFE survived. The resulting histology demonstrated scarring alopecia.

Although the bulge cells did not contribute to epidermal renewal of the IFE during homeostasis, they did migrate centripetally within days after an excisional wound (4 mm) was made on the back of 7- to 8-wk-old mice. These bulge-derived cells localized to the basal layer after excisional wounding, and cells derived from the bulge showed positive staining for epidermal differentiation markers (e.g., K10 and loricrin) 8 days after excisional wounding. Labeled bulge-derived cells constituted 26% of the wound epithelium on post-wound day 8, but only 3.5% on post-wound day 20 (Ito et al. 2005). From this study, it was posited that (1) the transient nature of bulge-derived cells in the wound suggests that dermal signals alone are not sufficient to convert bulge-derived cells into long-lasting epidermal stem cells, and (2) there was a functional difference between bulge stem cells and IFE stem cells because their relative contributions to the regenerated epithelium diverged over time.

Lgr6⁺ hair follicle stem cells in the isthmus have also been noted to play a significant role in wound healing. Lgr6 is a known early placode marker during murine embryogenesis. Lgr6⁺ stem cells generated long-lasting (>3-mo) clones in the basal IFE layer after a 1-cm² full-thickness excision of back skin, but their contribution greatly diminished over time. Approximately 10% of hair follicles at the center of the wound that formed as a result of WIHN were derived from Lgr6⁺ stem cells (Snippert et al. 2010).

In contrast, the contribution of bulge-derived stem cells may increase over time in smaller wounds. Using lineage tracing, Wong and colleagues (Vagnozzi et al. 2015) quantified the relative contributions of stems cells from the bulge, the hair follicle, and IFE using K15:YFP mice, Shh:YFP mice, and K14:YFP mice, respectively. This study examined the wound healing response from a relatively small wound (0.25 cm² full-thickness wound from 7.5-wk-old mice) and no new hair follicles were noted in the center. In post-wound basal keratinocytes, the contribution from the IFE was the greatest (>50%) at day 3, but the contribution of stem cells from the hair follicle and the bulge increased over time. On post-injury day 50, the hair follicle and the bulge contributed >60% and >20% of the wound basal layer, respectively. This is in agreement with a prior study by Morgan and colleagues who demonstrated that the majority of long-lasting epidermal progenitors were derived from the hair follicle after full-thickness wounding (4-mm punch) (Levy et al. 2007). However, the precise cellular composition and kinetics in the setting of WIHN, which only seems to occur in relatively large wounds, need additional study.

γδ T Cells Up-Regulate Wnt Signaling through Secretion of Fibroblast Growth Factor 9 (FGF9)

Wnt signaling is heavily involved in hair follicle formation during embryogenesis and β-catenin expression is required in dermal condensates for induction of hair follicle placodes (Tsai et al. 2014). Wnt ligands bind to Frizzled receptors. In the absence of Wnt, β-catenin undergoes phosphorylation by glycogen synthase kinase 3β (GSK3β) and is degraded. Accumulation of β-catenin is followed by its translocation into the nucleus, where LEF-1/TCF transcription factors activate Wnt-responsive elements (Cadigan and Waterman 2012). RAC1, which is required for hair follicle development, interacts with β-catenin, and genetic ablation of Rac1 was noted to phenocopy β-catenin deletion by causing limb truncation (Wu et al. 2008). Canonical Wnt signaling is required for proliferation during the anagen phase (Choi et al. 2013). Epithelial deletion of β-catenin caused rapid regression of anagen hair follicles and inhibited matrix proliferation. Furthermore, hair follicle stem cells lacking β-catenin were unable to enter anagen following plucking (Choi et al. 2013), although they quickly regained this potential when β-catenin was restored.

Wnt signaling is also critical in WIHN. Wnt signal blockade via expression of Dkk1 (Dickkopf WNT signaling pathway inhibitor 1), a Wnt antagonist, prevented hair follicle formation after wounding. In the same study, overexpression of Wnt7a in KRT14-Wnt7a transgenic mice increased the number of hair follicles that regenerated (Ito et al. 2007). These studies demonstrated that by manipulating pathways necessary for embryonic hair follicle formation during wound healing, the number of new hair follicles can be modulated.

New hair follicle placodes are observed at least 14 d after wounding (Fig. 2). The inductive signals leading up to this period were investigated, and Fgf9 was found to be differentially expressed (Gay et al. 2013). FGF9 is an endogenous ligand of FGFR2 and FGFR3 on wound dermal fibroblasts, and ligand binding triggers up-regulation of Wnt2a and LEF1 (Colvin et al. 1999; Yin et al. 2008). During WIHN, an increase in dermal γδ T cells was identified between day 9 and day 12. Dermal γδ T cells were determined to be the sole T-cell source of FGF9, and transgenic mice that selectively lacked FGF9 in γδ T cells regenerated fewer follicles (Gay et al. 2013). Injection of an adenovirus vector containing an Fgf9 construct only partially rescued this phenotype, implying that there may be other WIHN mediators regulated by γδ T cells. This study also highlighted a notable difference between WIHN and embryogenesis; during WIHN, dermal γδ T cells and fibroblasts generate FGFs and Wnts, but these signals that induce hair follicle placode formation in utero originate from the epidermis (Chen et al. 2012). Human γδ T cells also show FGF9 expression (Workalemahu et al. 2004). However, humans have a relative paucity of γδ T cells compared to mice. This may partially account for the relative inability in humans to regenerate hair follicles after wounding (Gay et al. 2013).

Figure 2.

Key signaling pathways involved in wound-induced hair neogenesis (WIHN). Known timeline of signaling pathways is summarized. New hair follicle placodes are observed at least 14 d after wounding. TLR3, which binds dsRNA released from tissue damage, up-regulates the expression of IL-6 and STAT3. The peak of IL-6 expression is seen on PWD3. An increase in dermal γδ T cells occurs between PWD9 and 12, and a subsequent increase in FGF9 is observed after. FGF9 is an endogenous ligand of FGFR2 and FGFR3, and ligand binding triggers up-regulation of Wnt2a and LEF1. CX₃CR1 expression was also acutely up-regulated in macrophages during the first 3 d following wounding, and they up-regulate the expression of TGF-β and TNF-α. The Hh pathway gradually increases from PWD3 to PWD7. Pathways that promote or inhibit WIHN, as well as known cellular sources of relevant pathway members, are shown in the table portion. (CX₃CR1) C-X₃-C motif chemokine receptor 1, (DKK1) Dickkopf WNT signaling pathway inhibitor 1, (dsRNA) double-strand RNA, (FGF9) fibroblast growth factor 9, (IL) interleukin, (SHH) Sonic Hedgehog, (TGF-β1) transforming growth factor β1, (PWD) post-wound day, (TLR3) Toll-like receptor 3, (STAT3) signal transducer and activator of transcription 3, (TNF-α) tumor necrosis factor α, (TRPA1) transient receptor potential cation channel subfamily A member 1, (Twist1) Twist family BHLH transcription factor 1, (Wnt7a) Wnt family member 7A. *Deletion or inhibition down-regulates WIHN.

During WIHN, melanocytes migrate out of the hair follicle but do not reach the region of hair follicle neogenesis (Ito et al. 2007). As a result, newly formed hair follicles lack melanocytes, and the resulting hair lacks pigment. A subsequent study by Ito and colleagues demonstrated that constitutive Wnt pathway activation in melanocytes increased the number of epidermal melanocytes during WIHN, and the converse was also shown to be true through the use of Dkk1 (Sun et al. 2018). In the same study, selective loss of epithelial Wnt ligands in K14-Wls cKO mice resulted in decreased melanocyte migration into the epidermis. Furthermore, a week-long treatment of 7-wk-old second telogen mice with lithium chloride, an activator of Wnt signaling, was shown to regenerate pigmented hair (Yuriguchi et al. 2016).

Macrophages Generate Transforming Growth Factor-β1 (TGF-β1) and Up-Regulate Noncanonical Wnt Signaling

C-X₃-C motif chemokine receptor 1 (CX₃CR1) is found on certain cells of the myeloid lineage, such as macrophages and plasmacytoid dendritic cells (Bar-On et al. 2010; Zhang et al. 2012). CX₃CL1 (fractalkine), the ligand for CX₃CR1, is the only known member of the CX₃C chemokine subfamily, and it can be expressed either in a soluble form or a membrane-bound form. A study by Ishida, Gao, and Murphy demonstrated that CX₃CL1 and CX₃CR1 were both acutely induced within 3 d in mice after a full-thickness small wound formation, and the loss of CX₃CR1 significantly impaired macrophage recruitment and delayed closure (Ishida et al. 2008).

During WIHN, CX₃CR1 expression was also acutely up-regulated in macrophages during the first 3 d following wounding (Rahmani et al. 2018). WIHN was not observed in mice lacking CX₃CR1. CX₃CR1⁺ cells were more densely populated at the wound edge and were shown to generate TGF-β and tumor necrosis factor-α (TNF-α), which directly stimulated Lgr5⁺ stem cells through AKT/β-catenin signaling (Wang et al. 2017; Rahmani et al. 2018). Serial injections of TGF-β1 from post-wound days 7–10 increased the number of hair follicles, and selective inhibition of TGF-β1 from post-wound days 5–15 decreased the number of hair follicles (Rahmani et al. 2018). Akt, which is part of the phosphatidylinositol 3-kinase (PI3K) pathway, was previously shown to phosphorylate β-catenin at Ser552 and activate the Wnt pathway in a noncanonical manner (Fang et al. 2007).

Distinct Lineages of Fibroblasts and Their Role in WIHN

Interaction between the dermis and the epidermis is needed for hair follicle placode induction both in utero and in WIHN. In E16.5 mice, lineage restriction occurs and fibroblasts expressing Lrig1, Blimp1, or CD26 generate the papillary dermis, which houses the dermal papilla. On the other hand, fibroblasts expressing delta-like homolog 1 (Dlk1) generate the reticular dermis and the subcutaneous fat (Driskell et al. 2013).

A study by Driskell, Watt, and colleagues discovered that increased epidermal β-catenin expression prior to inducing an 8-mm, full-thickness wound increased the number of hair follicles and recruited Lrig1⁺ fibroblasts. Thus, increased recruitment of papillary dermal fibroblasts as a result of increased epidermal Wnt signaling induced WIHN in a relatively small wound (Driskell et al. 2013). Similarly, a study by Myung, Takeo, Ito, and Atit also demonstrated that a conditional knockout of Wntless, a gene needed for active secretion of Wnt ligands, in K14cKO mice abrogated dermal papilla formation during WIHN (Myung et al. 2013). Thus, epidermal Wnt ligand is directly required for dermal Wnt activation, and losing this interaction abrogates hair follicle initiation during WIHN.

The role of different fibroblast lineages in wound healing was further studied by Driskell and colleagues (Rognoni et al. 2016). When a 2-mm, full-thickness wound was induced in postnatal day 2 (P2) mice, all hair follicle stages (anagen, catagen, and telogen) were noted in the wound after 7 d. In this study, hair follicles that were further along the hair cycle were observed closer to the edges of the wound. As expected, the ability to make new hair follicles after wounding diminished with age. Dermal cellular density was compared between mice in different age groups, and an overall decrease was found with increasing age. However, the papillary dermal density decreased more relative to that of deeper dermis. The authors postulated that this decrease in the papillary dermal density, which was not due to an increase in apoptosis, may correlate with age-dependent decrease in de novo hair follicle regeneration seen in older mice. Surprisingly, this study identified stronger Wnt signaling in the wound beds of P21 and P50 mice when comparing it to that of P2 mice. Furthermore, deletion of β-catenin in dermal fibroblasts increased the number of new follicles in both P4 wounds and P50 wounds (Rognoni et al. 2016). Interestingly, a study by Mastrogiannaki, Watt, and colleagues (Mastrogiannaki et al. 2016) demonstrated that constitutive β-catenin stabilization in postnatal skin fibroblasts promoted fibrosis at the expense of the adipose layer. Therefore, the timing and level of Wnt pathway expression is critical in promoting hair follicle regeneration, as well as regulating fibrosis.

SIGNALING PATHWAYS INVOLVED IN WIHN

Multiple signaling pathways that are typically up-regulated during embryonic development are involved in WIHN. Known regulators of WIHN are summarized in Figure 2. For discussion regarding Wnt signaling and the TGF-β signaling, please refer to the subsections γδ T Cells Up-Regulate Wnt Signaling through Secretion of Fibroblast Growth Factor 9 (FGF9) and Macrophages Generate Transforming Growth Factor-β1 (TGF-β1) and Up-Regulate Noncanonical Wnt Signaling.

Hedgehog Signaling

The mammalian Hedgehog (Hh) signaling is a highly conserved pathway required for hair follicle morphogenesis. The canonical Hh signaling (Carballo et al. 2018) involves binding of a Hedgehog ligand—Sonic Hedgehog, Desert Hedgehog, or Indian Hedgehog—to a negative regulatory receptor, Ptch. Ptch constitutively binds and inhibits Smoothened (Smo), a transmembrane G-protein receptor, and this interaction prevents the ciliary accumulation of Smo. The ligand binding of Ptch causes the accumulation of Smo, which subsequently activates glioma-associated oncogene (Gli) transcription factors, Gli1, Gli2, and Gli3.

Abnormal Hh signaling is implicated in basal cell carcinoma (BCC) pathogenesis (Rudin 2012; Burness 2015). Smo receptor antagonists, vismodegib and sonidegib, that directly target Hh signaling are currently approved by the U.S. Food and Drug Administration (FDA) for the treatment of advanced and metastatic BCCs. Both drugs are associated with hair loss.

Hh signaling also plays an important role in hair follicle development. In Shh^−/− embryos, only rudimentary hair follicles form (St-Jacques et al. 1998; Chiang et al. 1999). In a study by Ito and colleagues, increased Gli1 expression was observed at the site of new hair follicles in the center of large wounds (Lim et al. 2018). Dermal papilla and hair germ did not form in large wounds that underwent epidermal deletion of Shh through tamoxifen induction on post-wound days 3–21 in K14-CreER;Shh^fl/fl mice. The administration of tamoxifen to Pdgfra-CreER;Smo^fl/fl mice also had the same result. Constitutive Shh overexpression in the epidermis increased hair follicle formation, and there was an up-regulation in Hh signaling in both the epidermis and the dermis. No alteration in collagen density was observable on electron microscopy after epithelial Hh overexpression, but BCC-like structures were observed. Furthermore, selective overexpression of Smo in dermal myofibroblasts in the dermis using an SM22 promoter induced hair germ cells directly above Smo-active dermal papilla. Single-cell RNA sequencing showed increased expression of dermal papilla signature genes in Hh-active fibroblasts. Furthermore, selective up-regulation of Wnt signaling in myofibroblasts induced WIHN after small wound formation. This study highlighted that Hh activation is required for WIHN.

Toll-Like Receptor 3 (TLR3) Activates Wnt, Hh, and Ectodysplasin a Receptor (EDAR) Signaling

TLR3 binds double-strand RNA (dsRNA), which may serve as either damage-associated molecular pattern (DAMP) or pathogen-associated molecular pattern (PAMP), and up-regulates innate immunity upon activation (Jounai et al. 2012). In unstimulated cells, TLR3 is expressed on the endoplasmic reticulum. Subsequently, it is trafficked to endosomes after stimulation with dsRNA. TLR3 has been characterized on many different cell types, both immune (e.g., dendritic cells, macrophages, and neutrophils) and nonimmune (epithelial cells, hepatocytes, and Schwann cells) (Cavassani et al. 2008; Perales-Linares and Navas-Martin 2013).

TLR3 has been implicated as a detector of acute tissue damage in WIHN. In a study by Garza and colleagues, placement of minute cuts at the wound edge significantly increased the number of hair follicles at the center of the wound (Nelson et al. 2015). Similarly, addition of dsRNA mimic poly(I:C) also increased the number of hair follicles, and this response was abolished in TLR3 null mice. TLR3 was shown to up-regulate the expression of interleukin-6 (IL-6) and phosphorylation of signal transducer and activator of transcription 3 (STAT3) in epidermal keratinocytes. Pharmacologic inhibition using cucurbitacin I, a known inhibitor of STAT3 phosphorylation, and conditional knockout of STAT3 using K5CreERT2-STAT3^fl/fl mice both significantly reduced the number of de novo hair follicles. Poly(I:C)-induced activation of TLR3 induced Wnt signaling, as evidenced by (1) increased nuclear translocation of β-catenin in vitro using normal human epidermal keratinocytes (NHEKs), and (2) increased expression of Wnt 7b and LEF1. SHH and GLI1 were also noted to be up-regulated. Thus, epidermal keratinocytes likely play a major role in mediating the effects of poly(I:C) in the setting of WIHN. A follow-up study (Zhu et al. 2017) of noncoding dsRNA also confirmed TLR3-dependent activation of WIHN, as well as Wnt7b up-regulation. In this study, celecoxib reduced the number of hair follicles that form through WIHN, and PGE₂ reversed this celecoxib-induced inhibition. Therefore, inhibition of PGD₂ or up-regulation of PGE₂ would likely pose a more selective therapeutic strategy to up-regulate WIHN, rather than targeting an enzyme, such as cyclooxygenase (COX)-2, that lies more upstream in the arachidonic acid metabolism pathway.

Interestingly, EDAR expression was also decreased in a TLR3-dependent manner (Nelson et al. 2015, 3). EDAR is a member of the TNF receptor superfamily, and its mutation causes hypohidrotic ectodermal dysplasia, noted by a triad of defective dentition, diminished sweating and hypotrichosis (Botchkarev and Fessing 2005). During embryonic hair follicle development, EDAR signaling is required to maintain Wnt signaling during placode development (Zhang et al. 2009). Because EDAR signaling has not been extensively characterized in WIHN, this may warrant further investigation. Interestingly, TLR3 null mice were also noted to have fewer γδ T cells, which serve as primary producers of FGF9.

Role of Arachidonic Acid Derivatives in WIHN

COX catalyzes the rate-limiting step in generating arachidonic acid metabolites that include prostaglandins (PGs), thromboxanes, and prostacyclins. COX-1, the dominant isoform, is constitutively active in most cells as a housekeeping enzyme. On the other hand, COX-2 (also known as prostaglandin-endoperoxide synthase 2 and prostaglandin G/H synthase) is undetectable in most healthy adult tissue skin, but it is up-regulated during an inflammatory response (Lee et al. 2003; Attiq et al. 2018). Celecoxib, a selective COX-2 inhibitor, has been on the U.S. market since 1998 to treat inflammatory conditions such as rheumatoid arthritis, ankylosing spondylitis, and osteoarthritis. It was previously shown to significantly lower the incidence of nonmelanoma skin cancers with no significant increase in serious adverse events after 11 mo of treatment (Elmets et al. 2010).

PGs are known mediators of hair growth. PGD₂ synthase, an enzyme downstream of COX-2, was found to be elevated in human scalp samples from men with androgenetic alopecia (Garza et al. 2012). PGD₂ was also found to be elevated and PGE₂ was diminished when measured with ultra-high performance liquid chromatography–mass spectrometry. PTGDR (DP-1) and GPR44 (DP-2) are two known canonical receptors of PGD₂. Mice deficient in GPR44 were resistant to the inhibitory effects of PGD₂, but those deficient in PTGDR were susceptible. Therefore, GPR44 was concluded to be the key target receptor of PGD₂, which inhibits hair growth. On the other hand, PGF_2α stimulates telogen to anagen transition, and PGE₂ stimulates anagen follicles (Sasaki et al. 2005).

After 1-cm² full-thickness excision of P21 mice, effects of different PGs on WIHN were studied. This study utilized different mouse strains: C57Bl/6J, FVB/N, and mixed (C57Bl/6J × FVB/N × SJL/J) (Nelson et al. 2013). The mixed strain had 20 follicles in the wound center after 12 d, and C57Bl/6J had five follicles. The Ptgds mRNA level, as well as the PGD₂ level measured by mass spectrometry, inversely correlated with the number of hair follicles. Furthermore, topical application of PGD₂ significantly inhibited WIHN in wild-type mice, but not in Gpr44 null or heterozygous mice.

Adipocyte Regeneration by Activation of Bone Morphogenetic Protein (BMP)- ZFP423 Pathway

During the majority of the hair cycle, the bulb is housed in subcutaneous adipose tissue. Newly generated hair follicles that form during WIHN also form adipose tissue around them, and these adipocytes appear indistinguishable from normal adipocytes. First adipocytes were noted on post-wound day 23 (Plikus et al. 2017). Interestingly, myofibroblasts, which appear on post-wound 5 and increase in number by post-wound day 12, down-regulated their expression of smooth muscle actin by post-wound day 17 and up-regulated expression of adipogenic commitment factors, such as ZFP423 and pCEBPb, on post-wound days 21–24. BMP signaling, which induces adipogenic lineage commitment in vitro, was also up-regulated; endogenous BMP ligand genes, Bmp4 and Bmp7 were up-regulated by day 21. Deletion of BMP receptor BMPR1A in myofibroblasts abrogated de novo adipocyte formation (Plikus et al. 2017). Wnt signaling and BMP signaling regulate each other, and the overall expression of these pathways oscillate throughout the hair cycle (Daszczuk et al. 2020). Although K14-Wnt7a mice formed an increased number of hair follicles after wounding, they were noted to lack fat regeneration because Wnts are inhibitors of adipocyte differentiation.

WOUND ENVIRONMENT: A WINDOW OF OPPORTUNITY FOR REGENERATION

Emergence of embryonic signaling pathways and induction of cellular plasticity are hallmarks of wounding and, in particular, WIHN. Thus, WIHN holds significant translational potential, and its potential clinical applications will be discussed.

Regeneration of hair follicles and adnexal structures would be highly beneficial for the treatment of various scarring conditions. A cutaneous scar that forms as a result of injury or surgical intervention never regains its original tensile strength and may possess a different contour and color from those of its surrounding area. Furthermore, abnormal wound healing may give rise to a hypertrophic scar or a keloid that may cause significant morbidity and mortality. Current management modalities for scar revision include dermabrasion, ablative laser therapy, serial intralesional corticosteroid injections, and surgical revision (Marshall et al. 2018). Treatment options for keloidal or hypertrophic scars include intralesional corticosteroid injections, surgical excision, and radiation treatment (Davis et al. 2013). However, these modalities are either invasive or lack high-grade evidence. Surgical excision alone for keloid treatment also has a recurrence rate of 45%–100% (Berman and Flores 1997).

After myofibroblasts were identified as adipocyte precursors during WIHN, human-derived keloidal scar cells were treated with BMP4, and they were converted to adipocytes. Coculturing of human-derived keloidal scar cells with human scalp hair follicles yielded a similar result with up-regulation of adipocyte markers. Induced differentiation from myofibroblasts to adipocytes could also theoretically be beneficial for other scarring conditions, such as scleroderma and nephrogenic systemic fibrosis.

Recently, application of imiquimod, a TLR7 agonist that also activates transient receptor potential cation channel subfamily A member 1 (TRPA1), was found to reduce scar formation and up-regulate WIHN 15-fold (Wei et al. 2020). TRPA1-deficient mice did not exhibit this effect. TRPA1, located on peripheral afferent cutaneous neurons, was also found to up-regulate IL-17 production via recruitment of dermal γδ T cells. Thus, targeting TRPA1 may represent a strategy to mitigate scar formation and promote WIHN in the future.

Treatment of alopecia, particularly scarring (cicatricial) alopecia, is another potential application of WIHN. Scarring alopecia may be primary or secondary. Primary scarring alopecia may arise as a result of lymphocytic (e.g., chronic cutaneous lupus erythematosus, lichen planopilaris, and central centrifugal cicatricial alopecia), neutrophilic (e.g., folliculitis decalvans and dissecting cellulitis), or mixed (e.g., erosive pustular dermatosis and folliculitis necrotica) infiltration (Olsen et al. 2003). Many of these conditions are challenging to treat, especially during later stages, and a significant portion of cases remain refractory to all available treatment modalities. Thus, induced formation of de novo hair follicles through WIHN in scarring alopecia may represent a novel treatment strategy, especially in chronic or treatment-resistant scenarios.

Although it was initially noted that hair follicles only form in the center of large wounds, a prior study demonstrated that follicle regeneration is also possible in smaller wounds when there is increased β-catenin expression in the epidermis prior to wounding (Driskell et al. 2013). Furthermore, a study by Garza and colleagues illustrated that introducing minute cuts at the wound edge during the initial wounding process also significantly increased the number of hair follicles at the center of the wound (Nelson et al. 2015). Even just the presence of a wound bed induced using a carbon dioxide laser was previously shown to enhance follicular unit transplantation for the treatment of cicatricial alopecia (Kwon 2007). Inhibition of PGD₂ synthesis or suppression of GPR44, a key target receptor for PGD₂, prior to wounding may also increase de novo hair regeneration. Up-regulation of Wnt or Hh signaling may also represent other paths to augment WIHN, but this must be carefully balanced against their suppression of adipocyte formation and induction of basal cell carcinoma, respectively. More recently, optimizing the wound stiffness to be between 5 and 15 kPa was found to be crucial for WIHN (Harn et al. 2021). Manipulating biomechanical characteristics, such as wound stiffness, may therefore represent an alternative strategy to up-regulate WIHN. Combining ways to increase de novo hair follicle formation may therefore significantly decrease the wound size necessary to achieve hair follicle regeneration in scarring alopecia.

CONCLUDING REMARKS

WIHN is a remarkable example of inducible regeneration in an adult organism. Furthermore, it provides an excellent animal model to study regeneration because it recapitulates signaling pathways only typically seen during embryogenesis. Although the hair follicle and the IFE are replenished by distinct stem cell niches during homeostasis, an increasing body of evidence points to significant cellular plasticity between stem cell populations after wounding. Contributions by different stem cell reservoirs also differ depending on the size of the wound (e.g., dynamic contribution of bulge-derived cells that changes over time in small vs. large wounds), and WIHN only typically occurs in larger wounds. Further studies are warranted to delineate why such differences exist.

Molecular pathways involved in WIHN are also slowly being elucidated. Wnt signaling, which is required in dermal condensates for induction of hair follicle placodes in utero, was the first crucial signaling pathway described in WIHN, and Wnt inhibition abrogates WIHN without impacting wound closure. Binding of TLR3 by dsRNA, an acute damage marker, up-regulates the expression of IL-6 and STAT3. Furthermore, TLR3 activation induces Wnt signaling. Another key activity in WIHN is the recruitment of dermal γδ T cells that occurs between PWD9 and PWD12, and these cells regulate Wnt signaling through secretion of FGF9. The mammalian Hh signaling, which gradually increases from PWD3 to PWD7, is also required for WIHN.

The WIHN phenomenon raises the possibility of developing new therapies for alopecia and scarring. Wounding creates a window of opportunity to push healing cells toward a hair follicle fate. By understanding the signaling pathways important for hair follicle development and regeneration, the opportunity exists to induce hair follicle formation after wounding by activating these pathways in cells that are now receptive or competent in response to wounding. The timing, dose, and delivery of therapeutics requires additional study, but potential translational applications include treatment of androgenetic alopecia, as well as scarring conditions such as keloids and scarring alopecia, as well as acute or chronic wounds.