br Results br Discussion Although the concept

Results
Discussion Although the concept that BLIMP1 is a marker of sebocyte progenitor cells has become established in the literature (Beck and Blanpain, 2012; Blanpain and Fuchs, 2014; Niemann and Horsley, 2012; Solanas and Benitah, 2013; Zhang et al., 2011), our studies suggest that BLIMP1 is primarily a marker of terminal differentiation in SG, IFE, and HF and that BLIMP1+ cells do not divide in undamaged postnatal epidermis. Nevertheless, genetic ablation of BLIMP1 confirms its importance in epidermal homeostasis, and clonogenic assays demonstrate that BLIMP1+ sebocytes, whereas nonproliferative in vivo, can divide in vitro. We showed that specific deletion of Blimp1 caused defects in differentiation in multiple epidermal compartments, namely the IFE, SG, HF infundibulum, and junctional zone. Hyperplasia was not restricted to the SG and was more pronounced in the IFE and infundibulum. The aberrant IFE differentiation was suggestive of a barrier defect, which would explain the previously reported inflammatory skin phenotype (Chiang et al., 2013). The two previously published studies on epidermal deletion of Blimp1 in early development (Horsley et al., 2006; Magnúsdóttir et al., 2007) used a K14Cre mouse described by Vasioukhin et al. (1999) to target the epidermal basal layer but used different conditional NMS-873 of Blimp1: Horsley et al. (2006) used an exon 5 floxed Blimp1 strain (Prdm1tm2Masu; Ohinata et al., 2005), whereas Magnúsdóttir et al. (2007) used a floxed Blimp1 strain with loxP sites between exons 6–8 (Prdm1tm1Clme; Figure 3A) (Shapiro-Shelef et al., 2003). Both studies reported SG enlargement (Horsley et al., 2006; Magnúsdóttir et al., 2007), a phenotype that we also observed on epidermal deletion of Blimp1 in adult mice. In addition, Magnúsdóttir et al. (2007), Chiang et al. (2013), and the present study found thickening of the IFE and HF infundibulum, regardless of whether Blimp1 was deleted in the embryo or adult. Potential reasons for the divergent observations regarding whether or not Blimp1 has a selective role in the SG include differences in the genetic background of the mice, animal husbandry (diet, health status, and pathogens), or the presence of truncated BLIMP1 protein that is undetectable with current antibodies (Bikoff et al., 2009). Differences in Cre expression over time could also contribute, as suggested for other transgenic Cre lines (Kang et al., 2014). BLIMP1 is a transcriptional repressor of c-Myc (Lin et al., 1997), and forced activation of MYC resulted in downregulation of BLIMP1 and increased proliferation within the SG (Cottle et al., 2013). We have previously observed that androgen receptor signaling modulates epidermal responses to MYC activation (Cottle et al., 2013), and consistent with this, stimulation of sebocyte differentiation was accompanied by an increase in the number of BLIMP1+ cells. BLIMP1 repression of c-Myc may also contribute to the transition of IFE cells out of the granular layer (Honma et al., 2006). Consistent with BLIMP1 being associated with terminal differentiation, BLIMP1 expression was downregulated in human SG tumors. Another negative regulator of BLIMP1 is miR-125b, which is upregulated in stem cells and progenitors in the HF and SG (Zhang et al., 2011). For lineage-tracing experiments, we crossed Blimp1Cre transgenic mice with two different loxP-STOP-loxP fluorescent reporter strains. We found no evidence that BLIMP1+ cells gave rise to differentiated sebocytes or indeed any labeled progeny. Instead, subpopulations of Lgr6- and Lrig1-expressing cells founded the sebocyte lineage, as reported previously (Jensen et al., 2009; Page et al., 2013; Snippert et al., 2010). Although we used the same Blimp1Cre line as in the earlier lineage-tracing experiments (Horsley et al., 2006; Ohinata et al., 2005), we utilized different reporter lines, transgenic CAGcatEGFP and gene-trap Rosa26tdTomato (used in our study) compared to gene-trap Rosa26EYFP (used in the study by Horsley et al., 2006), which have been shown to be more sensitive (Duffield and Humphreys, 2011; Kawamoto et al., 2000; Kretzschmar and Watt, 2012; Madisen et al., 2010). The fluorescent lineage tracers we used are expressed more strongly upon recombination, due to expression driven from the CAG promoter rather than the weaker Rosa26 promoter used by Horsley et al. (2006; Kawamoto et al., 2000; Soriano, 1999; Srinivas et al., 2001). EGFP and tdTomato also have the spectral advantage over EYFP when distinguishing true epifluorescence from highly autofluorescent structures such as the lipid-rich SG and sebaceous duct.