Topical application of AP stimulates hair growth
To check the effect of Alpinetin (AP) in promoting hair growth, we closely examined relevant phenotypes after topically applying vehicle control (VC, 65% ethanol treated), positive control (PC, minoxidil treated) or AP on depilated dorsal skin once a day according to experiment design (Fig. 1a). For the skin color of C57BL/6J (C57) mice can indicate the hair follicle cycling stage. As shown in Fig. 1b, skin color change was considerably faster in mice with topical application of AP (D4) compared to the PC- or VC-treated mice (D6). At D8, hair shaft was already observable in AP-treated mice, but not in VC- or PC-treated mice. The skin color of VC- or PC-mice turned white at D18, indicating a transition from anagen to catagen, while that of AP-treated mice did not turn white color until 1 day later (D19).
These data suggested that AP could promote anagen entry and prolong the anagen phase. We next examined whether AP promotes the elongation of hair follicle and shafts. Histological analyses showed AP-treatment significantly increased hair follicle length at D7 (Fig. 1c). The hair shafts at both D13 and D17 in AP-treated mice were significantly longer than those in VC- and PC-treated mice (Fig. 1d). These results indicated that AP could stimulate hair growth in mice.
AP promotes anagen entry of hair cycle
Compared with hair follicles in the telogen phase, hair follicles in anagen phase grow longer and dermal papilla is surrounded by hair matrix cells (Fig. 2a). Therefore, we categorized the hair follicles based on the morphological features in different hair follicle cycles. Given that AP treatment accelerated skin color change and promoted hair shaft elongation, we further examined the skin biopsy histologically in telogen–anagen transition at D4 after depilation (Fig. 2b). Histological analysis showed that while only 17% and 26% of the hair follicles entered the anagen stage in the VC- and PC-treated mice, respectively, all the hair follicles (100%) in the AP group were in the anagen stage (Fig. 2c, d, Additional files 1, 2: Fig. S1a). Notably, there seemed to be two types of hair follicles in the AP group: one type was straight and longer (Fig. 2d, AP#1); the other type was shorter and had a smaller hair follicle in their proximity, which called club hair. (Fig. 2d, AP#2), indicating AP might activate de novo hair follicle growth. P-cadherin protein is a marker indicating hair follicles entering the anagen phase [29]. Consistently, P-cadherin was barely detected in the hair germ of hair follicles in VC-treated mice and only expressed in ~ 50% of the hair follicles in PC-treated mice, while most cells in the hair germ and the bulge area expressed P-cadherin at D4 after AP application (Fig. 2e, Additional files 1, 2: Fig. S1b). Next, we examined cell proliferation in hair follicles using the EdU-incorporation assay. As expected, EdU+ cells were mainly detected in dermal papilla and hair matrix in all three groups (Fig. 2f). We observed additional EdU+ cells located in the bulge area of hair follicles in AP-treated mice, but not in the club hair of AP-treated skin. Remarkably, the EdU+ cells in AP-treated group were much more than those in the other groups, suggesting that AP could facilitate hair cycle entry by promoting proliferation of hair follicle epithelium cells.
AP delays catagen entry by inhibiting cell apoptosis
We further determined whether AP affects anagen–catagen transition (Fig. 3a). As shown in Fig. 2a, compared with the anagen phase, the hair follicle in catagen not only becomes shorter in length, but the dermal papilla also begins to shrink down to the bottom of the hair follicle, while forming an epithelial chain with the bulge, indicating the transition into catagen phase. Histological analyses showed that dermal papilla size reduced in VC- or PC-treated groups at D18 after depilation, indicating hair follicles entering the catagen stage (Fig. 3b, f). On the contrary, all hair follicles following AP treatment remained in the anagen stage at this point. Significantly, most hair follicles in AP-treated mice were still in the late anagen stage at D19. At D22, hair follicles in all the three groups have entered the telogen stage. These results suggested that AP delayed catagen entry as well as shortened catagen duration.
Since the AP group was in the anagen stage at D19, we wanted to know whether cells in hair follicles were proliferative. Surprisingly, cells in the dermal papilla in the AP group, but not those in the VC or PC-treated groups were incorporating EdU at D19 (Fig. 3c), suggesting AP treatment could sustain cell proliferation. On the contrary, cleaved caspase 3, a marker for apoptosis, was expressed in the bulge in the VC or PC-treated groups but was not detected in the AP group at D19 (Fig. 3d), indicating AP hadn’t get into catagen entry. Most hair follicles have entered catagen in VC or PC-treated groups, as the hair follicle retracts, the dermal papilla was pulled upward towards the permanent portion of the hair follicle where stem cells reside. K15+ HFSCs in both hair germ and bulge appear quiescent, no EdU staining was observed in VC group, few staining was observed in PC group, while all hair bulb appeared EdU+ in AP group (Fig. 3e). Taken together, these data showed that AP stimulated hair follicle entering anagen earlier and delayed catagen entry (Fig. 3g).
AP activates hair follicle stem cells
We have shown that AP could extend the anagen stage as well as promote hair follicle cell proliferation and wondered if AP could affect hair follicle stem cells (HFSCs). At D4 after treatment, there were more Lef1+ stem cells in the dermal papilla and K15+ stem cells in the ORS in the AP-treated group compared with the VC- or PC-treated groups, as well as for EdU+ proliferating cells (Fig. 4a). Lgr5+ HFSCs are the pioneer stem cell population that are triggered at the onset of anagen [30]. By using Lgr5EGFP−CreERT2; R26RtdTomato transgenic mice, Lgr5+ HFSCs were labeled 8 days before depilation. Lgr5-EGFP+tdtomato+ cells are HFSCs or progenies that express Lgr5, while Lgr5-EGFP−tdtomato+ cells are hair follicle cells that derived from Lgr5-EGFP+tdtomato+ stem cells, however, not expressing Lgr5 anymore as they become transitional amplification cells, participating in the rapid proliferation of hair follicle in the anagen phase [31]. At D4 after depilation, Lgr5+ (EGFP+; Tomato+) cells were detected in the lower bulge and hair germ of telogen hair follicle in the VC-treated group. In the AP treatment group, Lgr5 + (EGFP + ; Tomato +) cells were present in the expanding bulge, and a small number of transitional amplification cells (EGFP−; Tomato+) derived from Lgr5+ cells were observed in ORS (Fig. 4b). These results indicate that AP promotes the proliferation of Lgr5+ HFSCs (EGFP+; Tomato+) and progenies. Similarly, Gli1creERT2; R26RtdTomato transgenic mice is used to label Gli1+ HFSCs, another marker of HFSCs identified previously [9]. Gli1+ stem cells or progenies were rarely detected in follicles following VC treatment, whereas significant amount of Gli1+ stem cells/progenies (tomato+) were detected in the inner root sheath (IRS) from the upper bulge to the matrix after 6-day AP treatment (Additional files 1, 3: Fig. S2). Fewer hair follicles containing Tomato+ cells were detected in VC group either at D4 or D6 post depilation (Additional file 3: Fig. S2). These results indicated that AP could efficiently stimulate different stem cell populations in hair follicles.
AP functions through Wnt signaling
To explore the underlying molecular mechanism, by using RNA-seq we investigated the transcriptional changes in skin induced by AP treatment. Compared to the VC group, there were 1017 up-regulated and 550 down-regulated genes identified in the AP group (Fig. 5a). Enrichment analysis showed that the significantly up-regulated genes in the AP group were enriched for Gene Ontology (GO) terms including tissue morphogenesis, positive regulation of cell migration, skin development and hair cycle, as well as the Wnt signaling pathway (Fig. 5b). In-depth GO analysis revealed potential key regulators which significantly contributed to the enhanced Wnt signaling pathway and hair regeneration. For instance, Egr1, Wnt11 and Hmga2 are important genes involved in Wnt signaling pathway, whose expression are both upregulated over twofold. Meanwhile, the AP-induced upregulation of Fgf20, Itga6 and Csf1 may be the significant regulators promoting hair regeneration and cycle (Fig. 5c). Tissue-specific marker gene enrichment analysis also showed that a list of epidermis-specific genes are upregulated after AP treatment, in line with the morphological findings that AP treatment facilitates hair follicle proliferation (Fig. 5d).
GSEA analysis revealed that the gene sets of skin development and epidermis development were clearly enriched in the AP-treated group, which was consistent with GO analysis (Fig. 5e). In addition, the set of genes associated with Wnt signaling pathway were also enriched in AP-treated group. Specifically, genes related to epidermal development (Krt10, Krt80, Atp7a, Krt16, Krt77, Itga6) and Wnt signaling pathway (Fzd1, Apcdd1, Fzd6, Sfrp1, Fzd4, Ror1, Lrp6, Lrp5, Reck, Fzd10, Fzd5, Fzd9) were up-regulated in the AP group (Fig. 5f). The expression of Fzd1 and Lef1, the principle receptor and the core transcription factor involved in canonical Wnt pathway, which are important regulators functioned in the induction of primary hair/follicle, were also induced by AP application among the top list and validated by qPCR (Additional files 1, 4: Fig. S3), though not showed in Fig. 5e. These results suggested that AP regulates hair follicle development through the Wnt-Lgr5-Lef1 axis.
AP exhibits no cytotoxicity in keratinocytes and fibroblasts
While AP showed a promising effect to stimulate hair growth, it is critical to test the toxicity of a potential druggable compound. We did the in vitro pioneer cytotoxicity study for AP in primary cultured keratinocytes and fibroblasts, the two main cell types in skin from both human and mice. To examine the cytotoxicity of AP, we treated primary fibroblasts and keratinocytes obtained from human (Fig. 6a) and mice (Fig. 6b) at a wide range of concentration (0.01 μg/ml, 1 μg/ml, 100 μg/ml). Interestingly, AP increased proliferation of human fibroblasts at the high dose (100 μg/ml) and slightly increased proliferation of mouse fibroblasts. However, AP did not affect cell viability in human keratinocytes at all concentrations and slightly increased mouse keratinocytes proliferation at the high dose (100 μg/ml). These findings suggested that AP exposure had no cytotoxicity in the two main types of skin cells.