Epidermal stem cells in the isthmus/infundibulum influence hair shaft differentiation: evidence from targeted DLX3 deletion

Distinct stem cell populations are present in the skin that comprises the epidermis and the pilosebaceous unit (hair follicle and sebaceous gland) (Clayton et al., 2007; Mascre et al., 2012). Recent studies have highlighted the strong heterogeneity and compartmentalization of stem cell niches in the skin (Jaks et al., 2010; Page et al., 2013). During telogen (resting phase), hair follicle stem cells are located in the lower part of the hair follicle, in the bulge and hair germ areas, and express markers such as Lgr5, K15 and CD34 (Trempus et al., 2007; Jaks et al., 2010). Upon reentry into anagen (growing phase), components of the new hair are derived from these hair follicle stem cells. Even though they do not contribute to the homeostasis of the epidermis, these cells do participate in regeneration in the context of wound healing (Ito et al., 2005; Levy et al., 2005; Jaks et al., 2008). Recently, epidermal stem cells were identified in the upper part of the hair follicle: in the isthmus that is located near the attachment of the sebaceous gland to the follicle, and in the infundibulum area, which spans the isthmus and the epidermis. These cells express specific markers such as Lrig1 and MST24, and contribute to the formation of the sebaceous gland and to epidermal differentiation in response to injury (Jensen et al., 2009; Page et al., 2013). During anagen, the cells from the isthmus/infundibulum area do not contribute to the formation of the new hair. However, here we show that the deletion of the transcription factor DLX3 in the epidermis and isthmus/infundibulum area, but not in the bulge region, leads to altered hair shaft differentiation without affecting hair growth.

We previously showed that during hair morphogenesis DLX3 is expressed in the hair matrix at the beginning of hair shaft differentiation and subsequently in most layers of the hair follicle except for the outer root sheath. Epithelial deletion of DLX3 during embryogenesis (K14Cre; DLX3^cKO) results in impaired expression of hair keratins and leads to alopecia (Hwang et al., 2008). Absence of DLX3 in the epidermis also results in an IL 17 dependent inflammatory response in the skin (Hwang et al., 2011). During telogen, DLX3 expression is found in the bulge, which generates the new hair shaft in the subsequent anagen stage, as well as in the isthmus/infundibulum area (Figure 1a). DLX3 expression near the collar of the sebaceous gland persists during anagen (Figure 1a). However, DLX3 was not detected in the sebaceous gland (Figure 1b). Co staining with Lrig1 demonstrated that DLX3 expression overlaps with the expression of Lrig1 in epidermal stem cells in the infundibulum (Figure 1c).

Figure 1. DLX3 expression in the infundibulum/isthmus and cre recombinase activity in K14CreERT;R26R^YFP.

(a) Detection of DLX3 in the infundibulum/isthmus during telogen (P20) and first postnatal anagen (P24 and P26). DAPI (blue) marks nuclei. Scale bar: 50 μm.

(b) Absence of DLX3 expression in sebaceous gland shown by immunohistochemistry with anti DLX3 antibody. Scale bar: 25 μm.

(c) Detection of Lrig1 (red), DLX3 (green) and DAPI (blue) in the hair follicle at telogen stage. Scale bar: 25 μm.

(d) Time schedule used for the analysis of cre recombinase activity in K14CreERT;R26R^YFP mice.

(e) Analysis of cre recombinase activity at telogen (P56) and 14 days after depilation (PD14). YFP signal (green) marks cells in which cre recombination occurred. Immunohistochemical analysis of DLX3 (red) distribution was performed on the full anagen stage. Scale bar: 25 μm.

(f) YFP tracing and DLX3 staining on WT and Dlx3;K14 CreERT skin at PD14. Scale bar: 25 μm.

In order to address the function of DLX3 in this subpopulation of isthmus/infundibulum stem cells, we used the inducible K14CreERT mouse line. Using topical tamoxifen treatment conditions (sub optimal dose for 5 consecutive days, Figure 1d) established by tracing the cells undergoing cre recombination after tamoxifen treatment in K14CreERT;R26R^YFP line, we obtained cre recombination in the epidermis and isthmus/infundibulum area, but not in the bulge (P56) (Figure 1e, left panel). To confirm that the bulge cells rarely underwent cre recombination in these conditions, we induced anagen by depilation at P56 and analyzed the distribution of YFP positive cells in fully grown hair 14 days after depilation (PD14). At this stage, YFP positive cells were largely detected in the epidermis and isthmus/infundibulum area, but very few hair follicles exhibited YFP positive cells in the newly formed bulb derived from the bulge (Figure 1e, right panel and f). Therefore, these conditions were used to delete DLX3 in the epidermis and isthmus/infundibulum without affecting its expression in the bulge in the majority of hair follicles.

DLX3^{K14CreERT cKO} were generated and treated as described above for K14CreERT;R26R^YFP mice and specimens were collected 6 days (PD6) and 14 days (PD14) after depilation. The gross appearance showed that at PD6 there was similar progression of hair growth between DLX3^{K14CreERT cKO} and control mice (Figure 2a). However, while the newly grown coat appeared smooth in control mice at PD14, it appeared rough in the depilated area of DLX3^{K14CreERT cKO} mice (Figure 2a). The overall histology of the newly formed hair follicles was not significantly affected at PD6 and PD14 (Figure 2a). Consistent with the lineage analysis using K14CreERT;R26R^YFP mice, DLX3 was deleted in the epidermis and isthmus/infundibulum (at PD6 and PD14), while the expression in the bulb of the newly formed hair follicle was not affected in the vast majority of hair follicles (Figure 2b). In addition, the expression of hair keratins that are known targets of DLX3 (Hwang et al., 2008) was not affected (Figure 2c). However, scanning electron microscopy analysis of the hair shaft at PD14 revealed major structural defects in the cuticle in DLX3^{K14CreERT cKO} mice (Figure 2d). While newly grown hair shafts in the depilated area of control mice exhibited a regular parallel distribution, hairs in the depilated area of DLX3^{K14CreERT cKO} grew in a disorganized pattern, which was consistent with the rough appearance of the coat (Figure 2d). Moreover, higher resolution images revealed that most of these hairs showed different degrees of cuticle defects such as the detachment of cuticle scales, the formation of globular structures at the surface of the cuticle, or the complete absence of cuticle scale pattern that appeared normal in all hairs in control mice (Figure 2d). These results demonstrate that the absence of DLX3 in the isthmus/infundibulum has a deleterious effect on hair shaft formation although the pattern of sebum distribution is comparable between WT and DLX3^{K14CreERT cKO} mice (Figure 2e).

Figure 2. Effects of DLX3 deletion in the infundibulum/isthmus on hair differentiation.

(a) Time schedule of tamoxifen treatment and depilation applied on DLX3^{K14CreERT cKO} mice. Gross appearance of hair pelage (right side: depilated) and hematoxylin/eosin staining of skin sections (depilated). Scale bar: 100 μm.

(b) Immunohistochemical analysis of DLX3 (red) in depilated skin samples. DAPI (blue) marks nuclei. Scale bar: 50 μm.

(c) Detection of type I hair keratins (AE13) and trichohyalin (AE15) in hair follicle bulbs from depilated skin at PD6 and PD14. K17 expression in medulla and outer root sheath from depilated skin at PD14. Scale bar: 50 μm.

(d) Scanning electron microscopy analysis of depilated skin from DLX3^{K14CreERT cKO} and control mice at PD14 showing hair structure. Lower panels show high resolution images of the cuticle in control and DLX3^{K14CreERT cKO} mice. Scale bars: upper panels, 100 μm; lower panels, 25 μm.

(e) Detection of lipid deposit in the WT and DLX3^{K14CreERT cKO} skin sections at telogen (P20). Red indicates Oil Red O staining. Scale bar: 50 μm.

In this study we show that even though DLX3 is still expressed in the hair matrix, its deletion in the isthmus/infundibulum affects the integrity of the hair shaft. Therefore, the DLX3 expressing cells of the isthmus/infundibulum that form a collar surrounding the developing hair shaft play an important role in maintaining the structural integrity of the growing hair. Based on these observations, we propose a model in which the differentiation of hair shaft cells is initiated in the hair matrix, but as the cells move to the surface of the epidermis, the surrounding cells in the isthmus/infundibulum signal to the underlying hair shaft and contribute to the final cuticle structure formation. These signals may be relayed by the adjacent inner root sheath where altered TACE TGFα EGFR signaling was recently shown to cause cuticular abnormalities (Inoue et al., 2011). While the signals from the isthmus/infundibulum to the developing hair shaft remain to be elucidated, our results identify DLX3 as a regulator of these signals in the isthmus/infundibulum.