De novo production of dermal papilla cells during the anagen phase of the hair cycle.

TO THE EDITOR Although keratinocytes are the primary constituents of the hair follicle and generate the hair shaft, mesenchymal cells also have important functions. These include the follicular dermal papilla (DP) and the connective tissue sheath or dermal sheath (CTS). The DP is embedded in the hair bulb during the anagen phase and forms a compact ball during the telogen phase, whereas CTS cells line the outside of the epithelial follicle from the bulge to its base. The central function of the DP in regulating the activity of keratinocytes during hair follicle regeneration and hair shaft morphogenesis has been established by extirpation and grafting studies (Ibrahim and Wright, 1977; Jahoda et al., 1984, 1993; McElwee et al., 2003), and more recently by direct manipulation of gene expression in the DP in vivo (Enshell-Seijffers et al., 2010). A strong correlation between DP size, hair bulb diameter and hair caliber has been noted (Van Scott and Ekel, 1958; Ibrahim and Wright, 1982; Elliott et al., 1999). The CTS is less accessible to experimental manipulation and its function in the intact follicle is more poorly defined. However, the proximal CTS shares properties with the DP that include the capacity to reform the DP in grafting studies (McElwee et al., 2003). DP cells undergo comparatively few divisions and the constituents of a DP are a largely static population when compared with the dynamic changes in the keratinocyte populations that abut them. Modest expansion and contraction of cell numbers in the DP occurs over the course of the hair cycle. Tobin et al. (2003b) quantified these changes for the mouse hair cycle, reporting an increase in DP cell numbers during anagen I–V and a decrease in anagen VI–telogen. They noted that the increase in DP cell number observed in early anagen precedes detectable proliferation in the DP, and suggested that it results from the migration of CTS cells into the DP. We have generated a mouse line, Cor-cre, that expresses cre recombinase in the DP (Enshell-Seijffers et al., 2010). When coupled with a cre-dependent reporter gene, this provides a method to trace the fate of DP cells. The reporter gene contains the sequences encoding yellow fluorescent protein (YFP) in the ubiquitously expressed Rosa26 locus that are separated from a promoter by a transcriptional termination cassette flanked by LoxP sites (Srinivas et al., 2001). When this stop cassette is excised in cells expressing cre-recombinase, YFP is expressed in the cell and its progeny regardless of their position in the tissue or changes in expression of the cell type-specific cre recombinase allele. In Cor-cre/þ ; rYFP/þ mice, YFP is not detected until p3 and effective deletion across the DP population is not complete until p7 (Enshell-Seijffers et al., 2010). By the end of the anagen phase, virtually all DP cells express YFP, while the proximal CTS is variably labeled. Corin is not expressed in catagen or telogen, and although Corin expression returns during early anagen, cre recombinase is not reliably detected until anagen V (data not shown) (Enshell-Seijffers et al., 2008). If all DP cells in a follicle are labeled with YFP at the end of one growth phase, the appearance of unlabeled cells in the subsequent anagen phase would provide evidence of recruitment of new cells to the DP. Mice of the genotype Cor-cre/þ ; rYFP/þ were killed at p13 and the extent of labeling in the DP was determined in tissue sections (Figure 1a). In a survey of follicles from five mice, no unlabeled DP cells were observed in 94±5% of the follicles scored (n1⁄4129) (Figure 2). Although most DP cells were labeled in the remaining follicles, one or two cells were unlabeled. During the catagen phase, a compact ball of YFPþ cells is associated with the regressing epithelial strand. In telogen phase, the descendents of the anagen DP form a compact ball of contiguous labeled cells, surrounded by more elongated cells that are variably labeled (Figure 1b). The variable labeling of these peripheral cells is consistent with the assumption that they derive from the CTS and confirms that most are not descendents of the DP. However, definitive distinction between DP and CTS derivatives at this stage is not possible. In contrast to p13, many unlabeled DP cells are observed during the early anagen phase of the first post-natal hair cycle (Figure 1c and d). Skins in the early anagen phase were harvested (n1⁄45), and follicles in anagen stage III–IV were analyzed (Muller-Rover et al., 2001). Before stage III, distinction between DP and adjacent CTS is open to debate. At stage III, the DP is engulfed by the hair bulb, thereby allowing unambiguous definition of the boundary between DP and CTS with a line between the keratinocytes at the base of the follicle. Overall, 75%±8 of the follicles showed one or more DP cells that did not express YFP (n1⁄4256) (Figure 2). To confirm that this was not a failure of the rYFP reporter transgene, mice in which rYFP had been activated in the germ line were LETTERS TO THE EDITOR