ORIGINAL ARTICLE See related commentary on pg 7

Adenovirus-Mediated Wnt10b Overexpression Induces Hair Follicle Regeneration Yu-Hong Li1, Kun Zhang1, Ke Yang1, Ji-Xing Ye2, Yi-Zhan Xing1, Hai-Ying Guo1, Fang Deng1, Xiao-Hua Lian1 and Tian Yang1

Hair follicles periodically undergo regeneration. The balance between activators and inhibitors may determine the time required for telogen hair follicles to reenter anagen. We previously reported that Wnt10b (wingless- type mouse mammary tumor virus integration site family member 10b) could promote the growth of hair follicles in vitro. To unveil the roles of Wnt10b in hair follicle regeneration, we established an in vivo mouse model using intradermal injection. On the basis of this model, we found that Wnt10b could induce the biological switch of hair follicles from telogen to anagen when overexpressed in the skin. The induced hair follicles expressed structure markers and could cycle normally into catagen. Conversely, anagen onset was abrogated by the knockdown of Wnt10b with small interfering RNA (siRNA). The Wnt10b aberrant expression data suggest that it is one of the activators of hair follicle regeneration. The b-catenin is translocated to the nucleus in Wnt10b-induced hair follicles. The biological effects of Wnt10b were abrogated when b-catenin expression was downregulated with siRNA. These data revealed that Wnt10b might induce hair follicle regeneration in vivo via the enhanced activation of the canonical . To our knowledge, our data provide previously unreported insights into the regulation of hair follicle cycling and provide potential therapeutic targets for hair follicle–related diseases. Journal of Investigative Dermatology (2013) 133, 42–48; doi:10.1038/jid.2012.235; published online 26 July 2012

INTRODUCTION Research on the regulation of the hair follicle cycle The hair follicle is a specific mini-organ appendage of the is important and urgently needed. Many signaling pathways skin. Postnatal follicles have characteristic periodic growth are involved in the hair follicle cycle, such as the Wnt stages, namely the anagen, catagen, and telogen stages. (wingless-type mouse mammary tumor virus integration site ), The hair follicle can reenter anagen with certain stimuli in bone morphogenetic protein, fibroblast growth factor, and the telogen stage and thus continue the hair follicle cycle. Sonic hedgehog pathways (Kawano et al., 2005; Rendl et al., The process of the hair follicle reentering anagen is termed 2008; Alvarez-Medina et al., 2009). The Wnt signaling hair follicle regeneration (Alonso and Fuchs, 2006). Because pathway has an important role in hair follicle morphogenesis of its ability to undergo periodic growth, the hair follicle is an and regeneration (Andl et al., 2002). Usually, the Wnt effective model for tissue regeneration research. Hair follicles signaling pathway is divided into a canonical pathway and not only produce hair but are also associated with skin a noncanonical pathway. More attention is paid to the regeneration, pilomatricoma, and basal cell carcinoma canonical pathway, and its signal chain is better elucidated. (Kruger et al., 1999). If the hair follicle cycle is interrupted, Canonical Wnt signals are transduced through the hair follicle–related diseases, such as androgenic alopecia, family receptors and coreceptors to the b-catenin signaling might occur (McDonough and Schwartz, 2011). Much cascade. When b-catenin accumulates in the cytoplasm, research is still needed to understand how to solve these it can be translocated to the nucleus, where it interacts with problems. the transcription factors TCF/LEF to regulate the expressions of target (Katoh, 2007). Hair follicle morphogenesis and regeneration would be significantly impacted if any 1 Department of Cell Biology, The Third Military Medical University, molecule in the canonical signal chain was blocked (Andl Chongqing, China and 2Bioengineering College of Chongqing University, Chongqing, China et al., 2002). Correspondence: Xiao-Hua Lian or Tian Yang, Department of Cell Biology, We previously showed that Wnt10b (Wnt family member The Third Military Medical University, Gaotanyan Street Number 30, 10b) could promote the growth of hair follicles in vitro, Shapingba, Chongqing 400038, China. E-mail: [email protected] or probably via a canonical Wnt signaling pathway (Li et al., [email protected] 2011). Reddy et al. (2001) reported that Wnt10b was Abbreviations: siRNA, small interfering RNA; Wnt10b, wingless-type highly and specifically expressed in the placode of the mouse mammary tumor virus integration site family member 10b; X-gal, 5-bromo-4-chloro-3-indolyl-beta-D-galactopyranoside epidermis in hair follicle morphogenesis. Ouji et al. (2007, Received 19 August 2011; revised 26 May 2012; accepted 7 June 2012; 2008 also reported that Wnt10b might promote published online 26 July 2012 epithelial differentiation and shaft growth). However, the

42 Journal of Investigative Dermatology (2013), Volume 133 & 2013 The Society for Investigative Dermatology Y-H Li et al. Wnt10b Induces Hair Follicle Regeneration

in vivo effects of Wnt10b on the hair follicle cycle are still a Wnt10b d Wnt10b/DAPI unknown. DP It was recently reported that in late telogen, the balance between activators and inhibitors may determine whether a DP hair follicle can reenter anagen (Plikus et al., 2011; Oshimori DP DP and Fuchs, 2012; Wang et al., 2012). Activators and inhibitors may be of intrafollicular or extrafollicular origin, Anagen Anagen whereas stem cells will sum them up (Chen and Chuong, be 2012). Thus, the increase in the quantity of activators in late Wnt10b Wnt10b/DAPI telogen may induce hair follicle regeneration. From the literature and our previous report, we infer that Wnt10b may be one of these activators of hair follicle regeneration. To test this hypothesis, we established an in vivo experimental DP model to study hair follicle changes caused by the over- Catagen Catagen expression or suppression of Wnt10b. On the basis of this model, we studied the signaling pathway that Wnt10b cf Wnt10b Wnt10b/DAPI mediates during this process. We found that Wnt10b could induce the biological switch of hair follicles from telogen to anagen. In addition, we show that Wnt10b might signal through the Wnt/b-catenin signaling pathway in vivo. To our knowledge, our data provide previously unreported evidence DP Telogen demonstrating that Wnt10b is one of the regulators under- Telogen DP lying hair follicle cycling. Figure 1. The expression pattern of Wnt10b (wingless-type mouse mammary tumor virus integration site family member 10b) throughout RESULTS the hair follicle cycle. The expression pattern of Wnt10b in the dorsal skin Wnt10b is periodically expressed in the hair follicle cycle of C57 mice was determined by in situ hybridization (a–c) and immuno- The normal expression of Wnt10b in the hair follicle cycle fluorescence (d–f). Wnt10b was expressed in anagen (a, d) but not in catagen was determined by both immunofluorescence and in situ (b, e) or telogen (c, f). The upper right corner shows a magnification of the hybridization (Figure 1). Wnt10b was expressed in anagen framed labeled area. The dashed line delineates the hair follicle structure. hair follicles and was not expressed in catagen or telogen hair Arrowheads in a and d show the expression of Wnt10b in hair matrix follicles. In anagen, Wnt10b was expressed in hair matrix cells. DAPI, 4’,6-diamidino-2-phenylindole; DP, dermal papilla. Scale bar ¼ 10 mm. cells and the precortex cells in the lower part of hair follicles. Real-time PCR demonstrated that wnt10b expression began to increase either at the end of telogen or at the onset of At 24 hours after overexpression, Wnt10b was expressed anagen (Supplementary Figure S1 online). The characteristic in nearly all of the skin samples. Over time, the expres- expression pattern of Wnt10b suggests that it may be a critical sion became more restricted. At 72 hours after overexpres- protein in hair follicle regeneration. sion, the expression of Wnt10b was gradually restricted We constructed an adenovirus-mediated overexpression to the bulge area and hair bulb (Figure 2, Supplementary model to test this hypothesis. X-gal (5-bromo-4-chloro-3- Figure S4 online), especially the bulge area near the indolyl-beta-D-galactopyranoside) staining and fluorescence sebaceous gland. These results further verify that the in vivo microscopy were used to analyze the adenovirus-mediated model works well. injection model. X-gal staining revealed that the target was expressed in the injection area (causing the formation of Overexpressed Wnt10b induces hair follicle regeneration a wheal around the injection site) 24 hours after injection. After being treated with AdWnt10b, the telogen hair follicle Green fluorescence was observed in the AdWnt10b (Wnt10b proceeded into anagen earlier than expected (Figure 3a–d). mediated by adenovirus) injection area (Supplementary Fourteen days after the injection of 108 pfu ml1 AdWnt10b, Figure S2 online). The mRNA expression of wnt10b increased the hair follicle regeneration ratio of the labeled area was by approximately 14-fold 24 hours after AdWnt10b injection 13:20 (65%). When the AdWnt10b titer was increased compared with control telogen skin. The expression level of to 109 pfu ml1, this ratio increased to 4:5 (80%). Thus, mRNA decreased over time. Seventy-two hours after AdWnt10b had a dose-specific effect on the promotion of AdWnt10b injection, the expression level decreased to the regeneration. level observed in normal early anagen skin (Supplementary To estimate systematically the impact of Wnt10b on the Figures S3 and S1 online). These results indicate that this hair follicle cycle, the treated mice were subjected to both in vivo model can restrictively express target genes tempo- macroscopic and microscopic observation through hema- rally in the injection area. toxylin and eosin staining. Beginning on the ninth day after To confirm the validity of the adenovirus-mediated in vivo treatment, the skin at the injection site gradually turned from model, we determined both the mRNA and protein expres- pink to black. Approximately 56 days after treatment, the sion patterns of Wnt10b in Wnt10b-overexpressed skin. surrounding hair follicles entered anagen. There was no

www.jidonline.org 43 Y-H Li et al. Wnt10b Induces Hair Follicle Regeneration

adWnt10b Wnt10b/DAPI using small interfering RNA (siRNA) in vivo. Real-time PCR demonstrated that the expression of wnt10b was successfully interfered temporally (Supplementary Figure S6 online). At DP the physiological transition stage from telogen to anagen, anagen was delayed in the AdSim10b (siRNA of Wnt10b DP DP mediated by adenovirus) injection area (which could be 24 hours 24 hours observed as a wheal surrounding the injection site; Figure 5a–c). The hematoxylin and eosin staining demonstrated that beWnt10b Wnt10b/DAPI the surrounding hair follicles reentered anagen, whereas the hair follicles in the injection area were still in the phy- siological transition stage from telogen to anagen (Figure 5d). These data further suggest that Wnt10b is an activator of hair SG follicle regeneration. DP 48 hours 48 hours The overexpression of Wnt10b activates the Wnt/b-catenin signaling pathway cfWnt10b Wnt10b/DAPI To observe the signaling pathway transduced by Wnt10b in vivo, the expression patterns of b-catenin were identified in regenerated hair follicles induced by Wnt10b. b-catenin was SG expressed in hair germ cells, outer root sheaths, sebaceous glands, and the epidermis at 24, 48, and 72 hours after 72 hours DP 72 hours Wnt10b overexpression. A portion of b-catenin translocated to the cell nucleus at 48 and 72 hours after Figure 2. The expression pattern of Wnt10b (wingless-type mouse Wnt10b overexpression (Supplementary Figure S7 online). mammary tumor virus integration site family member 10b) after the These data suggest that Wnt10b might activate the canonical overexpression of Wnt10b. Telogen C57 mice were intradermally injected with AdWnt10b, and the expression of Wnt10b was detected by in situ Wnt signaling pathway to induce hair follicle regeneration. hybridization (a–c) and immunofluorescence (d–f)at24(a, d), 48 (b, e), and The AdSimBC (siRNA of b-catenin mediated by adeno- 72 hours (c, f) after injection. The upper right corner shows a magnification of virus) we constructed was an siRNA of b-catenin and could the framed labeled area. The dashed line delineates the hair follicle structure. abrogate hair follicle regeneration induced by hair depilation Arrowheads show the expression of Wnt10b. DAPI, 4’,6-diamidino-2- (Figure 6a–c). In AdSimBC-treated skin, AdWnt10b could no phenylindole; DP, dermal papilla; SG, sebaceous gland. Scale bar ¼ 10 mm. longer induce hair follicle regeneration, which means that the induction effect of Wnt10b was also abrogated (Figure 6d). visible difference between hair follicles at the injection area This abrogative effect further demonstrated that overex- and surrounding hair follicles. The hematoxylin and eosin pressed Wnt10b can activate the canonical Wnt signaling staining revealed that structural changes occurred in the pathway. AdWnt10b-treated hair follicles 3 days after treatment. Some of the hair follicles appeared to be in anagen on the third day, DISCUSSION large melanin granules appeared on the seventh day, the We report here that the intradermal overexpression of regenerated hair follicles erupted out of the epidermis on the Wnt10b could induce the biological switch of telogen to eleventh day, and the treated hair follicles started to enter anagen in vivo. This finding indicates the critical role catagen on the twenty-eighth day (Figure 3e–j). that Wnt10b has in the periodic regeneration of hair follicles. Compared with hair follicles in normal anagen, the The regenerated hair follicles induced by Wnt10b were Wnt10b-induced hair follicles were larger in size but were normal in terms of hair follicle structure, with the exception not distinguishable with regard to length (Supplementary that they were larger in size, and no tumor formation Figure S5 online). To characterize the Wnt10b-induced hair occurred after long-term observation. Thus, Wnt10b may follicles, we also determined the expression patterns of some have potential applications in therapies for hair follicle–re- epidermal and hair follicle structure markers, and compared lated diseases. their expression patterns with those of normal hair follicles at It was reported that in late telogen, the balance between the corresponding stage. The Wnt10b-induced hair follicles activators and inhibitors may determine whether a hair expressed AE15, b-catenin, K6, K10, Sox9, and Wnt10b at follicle can reenter anagen (Wang et al., 2012). Some the corresponding positions (Figure 4). There was no visible activators and inhibitors have been reported (Oshimori and difference between hair follicles in the injection area and Fuchs, 2012). Wnt10b has an important role in the onset of surrounding hair follicles, and there was no evidence of tumor normal hair follicle regeneration, but it is not the only factor development after long-term observation (data not shown). involved in regeneration. For example, it was reported that the intradermal injection of Sonic hedgehog homolog could Knockdown of wnt10b abrogates hair follicle anagen onset induce the growth of premature hair follicles (Sato et al., To further elucidate the role that Wnt10b has in hair follicle 1999). We previously reported that the Wnt10b protein was regeneration, we interfered with the expression of wnt10b specifically expressed in the matrix cells of anagen hair

44 Journal of Investigative Dermatology (2013), Volume 133 Y-H Li et al. Wnt10b Induces Hair Follicle Regeneration

abc d

ef

gh Anagen Telogen

ij

Figure 3. Hair follicles after the overexpression of Wnt10b (wingless-type mouse mammary tumor virus integration site family member 10b) in telogen C57 mice. Telogen C57 mice were intradermally injected with adenovirus and observed under gross observation (a–d) or with hematoxylin and eosin staining (e–i). The observation time points were 3 days (a, e), 5 days (f), 9 days (g, h), 11 days (b), 14 days (c), 28 eight days (i), and 56 days (d) after adenovirus injection. (a, b, d–i) Mice injected with AdWnt10b. (c) From left to right, mice injected with phosphate-buffered saline, AdGFP (adenovirus with green fluorescent protein), AdWnt10b, AdWnt10b, and AdWnt10b. (j) Mice injected with AdGFP. Arrowheads show the injection areas. (e–g, i, j) The right panel shows a magnification of the framed labeled area in the left panel. Scale bar ¼ 100 mm (left panel and h) and 10 mm (right panel).

follicles and was not expressed in the middle of telogen Various studies have reported that the Wnt signaling (Zhang et al., 2009). We also reported here that wnt10b pathway regulates hair follicle morphogenesis and the hair mRNA was expressed in anagen and was not expressed in follicle cycle (Andl et al., 2002; Ito et al., 2007; Schlake and catagen or telogen, which is consistent with the literature Sick, 2007). The knockout of b-catenin or Lef1 resulted in the (Reddy et al., 2001). Reddy et al. (2001) also reported that at inhibition of hair follicle morphogenesis. Wnt signaling may anagen onset, wnt10b mRNA was expressed in follicular promote the growth of hair follicles (Huelsken and Behrens, epithelial cells immediately overlying the dermal papilla, 2002; Beaudoin et al., 2005), and excessive activation of Wnt which are usually known as secondary hair germ cells. Those signaling may even lead to tumorigenesis (Ohyama et al., reports indicated that Wnt10b might be an essential factor in 2006). We report here that Wnt10b-induced hair follicle the onset of the hair follicle cycle. Wnt10b is also important regeneration occurs through the Wnt/b-catenin signaling for maintaining the inductive potential of dermal papilla cells pathway. Other Wnts that signal through the canonical Wnt when these cells are cultured in vitro (Ouji et al., 2012). signaling pathway might also exhibit the induction effects. When we interfered with wnt10b mRNA expression with Our results demonstrate that Wnt10b may be a critical siRNA in the physiological transition from telogen to anagen, activator of hair follicle regeneration. However, when hair follicle regeneration was delayed. This finding further we increased the titer of AdWnt10b to 109 pfu ml1, the suggests that Wnt10b is a critical activator of hair follicle inductive effects of Wnt10b still could not approach 100%. regeneration. This result may be due to the different statuses of the mice.

www.jidonline.org 45 Y-H Li et al. Wnt10b Induces Hair Follicle Regeneration

a AdWnt10b Control b AdWnt10b Control

DP

DP AE15/DAPI AE15/DAPI β-catenin/DAPI β-catenin/DAPI c d

K10/DAPI K10/DAPI K6/DAPI K6/DAPI e f

DP

DP DP DP Wnt 10b/DAPI Wnt 10b/DAPI Sox9/DAPI Sox9/DAPI DP

Figure 4. Expression patterns of structure markers in Wnt10b (wingless-type mouse mammary tumor virus integration site family member 10b)-overexpressed skin. Telogen C57 mice were intradermally injected with AdWnt10b, and expressions of AE15, b-catenin, K6, K10, Wnt10b, and Sox9 were detected by immunohistochemistry 9 days after injection (left panel) and compared with those in hair follicles at the corresponding stage (right panel). The upper right corner shows a magnification of the framed labeled area. The dashed line delineates the hair follicle structure. Arrowheads show the expressions of AE15 (a), b-catenin (b), K10 (c), K6 (d), Wnt10b (e), and Sox9 (f). DAPI, 4’,6-diamidino-2-phenylindole; DP, dermal papilla. Scale bar ¼ 10 mm.

Chen and Chuong (2012) reported that the traditional telogen vector is similar to AdGFP (adenovirus with green fluorescent stage of hair follicles could be separated into two different protein) but contains the murine Wnt10b cDNA (NM_011718.2). periods: competent telogen and refractory telogen. In The AdSim10b vector is a siRNA of Wnt10b that is mediated by refractory telogen, there are more inhibitors, and it is more adenovirus, whereas the AdSimBC vector is a siRNA of b-catenin difficult for the hair follicles to reenter anagen, even if they that is mediated by adenovirus (top—50-cctcacttgcaataatta-30, are stimulated by activators (Plikus et al., 2008, 2011). The bottom—50-ataattattgcaagtgagg-30). After being purified on cesium follicles that did not enter anagen in our experiment might be chloride gradients, vectors were dialyzed into storage buffer, and in the phase of refractory telogen. With regard to the Wnt10b- their titers were determined. induced hair follicles, we did not observe any visible difference between the hair follicles in the injection area In vivo injection of adenovirus and those in the surrounding area 1 year after treatment. This The medical ethics committee of the Third Military Medical result is consistent with the notion that stem cells sum up University approved all described studies. Female C57 BL/6 mice the total activators and inhibitors in the intrafollicular were obtained from the Laboratory Animal Center of the Third and extrafollicular environments (Chen and Chuong, 2012). Military Medical University, Chongqing, China. Mice at postnatal The exogenous Wnt10b used in this study functions as an day 30 (anagen), 35 (anagen), 45 (catagen), and 56 (telogen) were activator and tilts the balance toward anagen reentry. anesthetized with 1% pentobarbital sodium. Back hairs were clipped In summary, we demonstrated that Wnt10b might serve as and 25 ml of adenovirus vector was injected intradermally along the one of the critical activators in hair follicle regeneration in median dorsal line of the skin. The injection was angled toward the physiological cycling and after plucking-induced regenera- head and produced a 1-cm wheal. A circle drawn by a cotton bud tion. We also showed that Wnt10b might induce hair follicle dipped with picric acid labeled the wheal. The skin and hair follicles regeneration via the canonical Wnt signaling pathway. of the mice were observed every day. Mice were killed to observe This work has potential applications in the development of the inner structure and protein changes at 1, 2, 3, 5, 7, 9, 14, 28, and therapies for hair follicle–related diseases. Further studies 56 days post injection. The Wnt10b-treated mice were observed for should aim to elucidate more mechanisms associated with up to 1 year. the regulation of Wnt10b in hair follicles and its therapeutic potential. X-gal staining To estimate the diffusion of adenovirus, AdlacZ (LacZ mediated MATERIALS AND METHODS by adenovirus) was injected as described. The expression of Amplification and purification of adenovirus-mediated vectors b-galactosidase protein was detected by X-gal staining. Dorsal skins Adenovirus vectors were constructed and propagated in HEK293 of labeled areas were gathered and fixed in 4% paraformaldehyde, cells as described previously (Luo et al., 2007). The AdWnt10b rinsed in phosphate-buffered saline (PBS), and stained in freshly

46 Journal of Investigative Dermatology (2013), Volume 133 Y-H Li et al. Wnt10b Induces Hair Follicle Regeneration

a bc a

d Telogen Anagen bc d

Figure 5. Hair follicles after the knockdown of Wnt10b (wingless-type mouse mammary tumor virus integration site family member 10b). C57 mice at the period of transition from telogen to anagen (postnatal day 30) were intradermally injected with AdGFP (adenovirus with green fluorescent protein; a) or AdSim10b (b–d). (a, b) Five days after adenovirus injection. (c, d) Nine days after AdSim10b injection. (c) The dorsal skins of treated Figure 6. AdSimBC abrogates the hair follicle regeneration initiated by the mice were spread out on paper and photographed. (d) The skin in c was overexpression of Wnt10b (wingless-type mouse mammary tumor virus cut and stained with hematoxylin and eosin. Hair follicles in the injection integration site family member 10b). (a–c) Telogen C57 mice were depilated area were still in telogen, whereas follicles in the surrounding area and intradermally injected with AdSimBC. (a) Depilated mice. (b) Seven days around the skin were in anagen. Arrowheads show the injection area after AdSimBC injection. Arrowheads show the injection area (wheal (wheal around the injection site). AdSim10b, siRNA of Wnt10b mediated by around the injection site). The triangle shows the delay of anagen entry adenovirus. Scale bar ¼ 500 mm. resulting from the wound, which is different from the delay of anagen entry resulting from AdSimBC treatment. (c) Nine days after AdSimBC injection. Arrowheads show the skin of the injection area, which turned gray. prepared X-gal staining solution (5 mM ferricyanatum Kalium, (d) A telogen C57 mouse was first injected with AdSimBC and then injected with AdWnt10b 24 hours later at the same injection area. Arrowheads 5mM potassium ferrocyanide, 2 mM magnesium chloride, and show that there were no visible changes in the skin adjacent to the injection 10% DMSO in PBS) at room temperature for 4–48hours. The tissues areas 7 days after AdWnt10b injection. AdSimBC, siRNA of b-catenin were rinsed in PBS to terminate the reaction, embedded in mediated by adenovirus. paraffin, cut into 5-mm-thick sections, and observed under a microscope.

Quantitative reverse-transcriptase PCR sections were stained with hematoxylin (Zhongshan Goldenbridge, Complementary DNA was synthesized from RNA using Rever Tre Beijing, China) for 2 minutes and subsequently rinsed with water. Ace, a complementary DNA synthesis Kit (TOYOBA, Osaka, Japan), The sections were later stained with eosin (Zhongshan Golden- according to the manufacturer’s protocol. TaqMan Gene Expression bridge) for 2 minutes and rinsed with water thereafter. After gradually Assays (Applied Biosystems, Foster City, CA) were used to analyze dehydrating, the sections were mounted with neutral gum (Zhong- the expression of murine Wnt10b (NM_011718.2) according to the shan Goldenbridge) and observed under a microscope. manufacturer’s instructions. The primers for Wnt10b were 50-CTGC GGATGGAAGGGTAGTGG-30 and 50-GGCGTCCCACCCTGTTGT Immunofluorescence TG-30. b-actin was used as an internal control to validate RNA Dorsal skin samples were embedded in paraffin and cut into 5-mm- quality for each sample. Its primers were as follows: 50-ACCCCG thick sections. The primary antibodies were antibodies to the TGCTGCTGACCGAG-30 and 50-TCCCGGCCAGCCAGGTCCA-30. following: b-catenin, K6, K10, Sox9 (all diluted 1:100, Santa Cruz, Santa Cruz County, AZ), Wnt10b (diluted 1:600, Sigma, Spruce, LA), Hematoxylin and eosin staining and AE15 (diluted 1:4, provided by TT Sun, New York University, Dorsal skin samples were fixed in 4% paraformaldehyde, gradually New York City, NY). The cy3-labeled secondary antibody (Beyotime, dehydrated through a graded series of alcohol, embedded in Haimen, Jiangsu, China) was used. The sections were counterstained paraffin, and cut into 5-mm-thick sections. After gradually hydrating, with DAPI ((4’,6-diamidino-2-phenylindole); Beyotime).

www.jidonline.org 47 Y-H Li et al. Wnt10b Induces Hair Follicle Regeneration

In situ hybridization Chen CC, Chuong CM (2012) Multi-layered environmental regulation on Digoxin-labeled probes were synthesized according to the manu- the homeostasis of stem cells: the saga of hair growth and alopecia. facturer’s instructions (Dig RNA labeling kit, Roche, Indianapolis, J Dermatol Sci 66:3–11 IN). Before hybridization, slides were dewaxed and hydrated under Huelsken J, Behrens J (2002) The Wnt signalling pathway. J Cell Sci 115:3977–8 RNase-free conditions, digested in 10 mgml1 proteinase K, refixed Ito M, Yang Z, Andl T et al. (2007) Wnt-dependent de novo hair follicle with fresh 4% paraformaldehyde, and prehybridized in prehybridi- regeneration in adult mouse skin after wounding. Nature 447:316–20 zation solution (in a volume of 50 ml, add 25 ml formamide, 12.5 ml Katoh M (2007) WNT signaling pathway and stem network. 1 20 SSC, 50 ml Tween-20, 500 ml of 10% Chaps, 1 ml of 10 mg ml Clin Cancer Res 13:4042–5 1 transfer RNA, 500 ml of 500 mM EDTA, 250 mlof10mgml heparin, Kawano M, Komi-Kuramochi A, Asada M et al. (2005) Comprehensive 1 g blocking reagent; add diethylpyrocarbonate H2Oto50ml).Slides analysis of FGF and FGFR expression in skin: FGF18 is highly expressed were incubated overnight with Wnt10b and b-catenin probes in hair follicles and capable of inducing anagen from telogen stage hair follicles. J Invest Dermatol 124:877–85 (diluted to 100 ng ml1 with prehybridization solution) at 65 1C. Kruger K, Blume-Peytavi U, Orfanos CE (1999) Basal cell carcinoma possibly After hybridization, the slides were washed with 2 SSC, 0.2 SSC, originates from the outer root sheath and/or the bulge region of the vellus and polybutylene terephthalate (PBT), preblocked with 20% goat hair follicle. Arch Dermatol Res 291:253–9 serum in PBT for 2 hours, and incubated overnight with anti-DIG-AP Li YH, Zhang K, Ye JX et al. (2011) Wnt10b promotes growth of hair antibody (diluted 1:1,000; Roche) at 4 1C. The slides were then follicles via a canonical Wnt signalling pathway. Clin Exp Dermatol 36:534–40 washed with Nacl, Tris, Mgcl2, Tween (in a volume of 50 ml, add Luo J, Deng ZL, Luo X et al. (2007) A protocol for rapid generation of recom- 2.5 ml of 2 M (pH 9.5) Tris-HCl, 1.25 ml of 2 M MgCl2, 1 ml of 5 M binant adenoviruses using the AdEasy system. Nat Protoc 2:1236–47 NaCl, 12 mg Levamisol, and 50 ml Tween-20) and stained with NBT/ McDonough PH, Schwartz RA (2011) Adolescent androgenic alopecia. Cutis BCIP (Roche) according to the manufacturer’s instructions. When the 88:165–8 color had developed, the slides were washed with PBT. Finally, the Ohyama M, Terunuma A, Tock CL et al. (2006) Characterization and isolation slides were mounted with neutral gum. of stem cell-enriched human hair follicle bulge cells. J Clin Invest 116:249–60 Statistical analysis Oshimori N, Fuchs E (2012) Paracrine TGF-beta signaling counterbalances Data are presented as the mean±SD. All statistical analyses were BMP-mediated repression in hair follicle stem cell activation. Cell Stem Cell 10:63–75 performed using Student’s t-test. The single-tailed t-test was used Ouji Y, Ishizaka S, Yoshikawa M (2012) Dermal papilla cells serially cultured for the real-time PCR results, and Po0.01 was considered to be with Wnt-10b sustain their hair follicle induction activity after statistically significant. The two-tailed t-test was used to analyze the transplantation into nude mice. Cell Transplantation; e-pub ahead of results of the hair follicle lengths and widths, and Po0.05 was print 2 April 2012 considered to be statistically significant. Ouji Y, Yoshikawa M, Moriya K et al. (2007) Effects of Wnt-10b on hair shaft growth in hair follicle cultures. Biochem Biophys Res Commun 359:516–22 CONFLICT OF INTEREST The authors state no conflict of interest. Ouji Y, Yoshikawa M, Moriya K et al. (2008) Wnt-10b, uniquely among Wnts, promotes epithelial differentiation and shaft growth. Biochem Biophys Res Commun 367:299–304 ACKNOWLEDGMENTS Plikus MV, Baker RE, Chen CC et al. (2011) Self-organizing and stochastic This work was supported by the National Natural Science Foundation of behaviors during the regeneration of hair stem cells. Science 332:586–9 China (number 81101208 and 30872711). We thank Professor TC He at the University of Chicago for constructing and providing AdWnt10b, AdSim10b, Plikus MV, Mayer JA, de la Cruz D et al. (2008) Cyclic dermal BMP and AdSimBC. We thank Professor Jun Yin at Inner Mongolia University signalling regulates stem cell activation during hair regeneration. for helping with in situ hybridization. We also thank Professor TT Sun at Nature 451:340–4 New York University for providing the AE15 antibody. Reddy S, Andl T, Bagasra A et al. (2001) Characterization of Wnt gene expression in developing and postnatal hair follicles and identification of Wnt5a as a target of Sonic hedgehog in hair follicle morphogenesis. SUPPLEMENTARY MATERIAL Mech Dev 107:69–82 Supplementary material is linked to the online version of the paper at http:// Rendl M, Polak L, Fuchs E (2008) BMP signaling in dermal papilla cells www.nature.com/jid is required for their hair follicle-inductive properties. Genes Dev 22:543–57 REFERENCES Sato N, Leopold PL, Crystal RG (1999) Induction of the hair growth phase in Alonso L, Fuchs E (2006) The hair cycle. J Cell Sci 119:391–3 postnatal mice by localized transient expression of Sonic hedgehog. J Clin Invest 104:855–64 Alvarez-Medina R, Le Dreau G, Ros M et al. (2009) Hedgehog activation is required upstream of Wnt signalling to control neural progenitor Schlake T, Sick S (2007) Canonical WNT signalling controls hair follicle proliferation. Development 136:3301–9 spacing. Cell Adh Migr 1:149–51 Andl T, Reddy ST, Gaddapara T et al. (2002) WNT signals are required for the Wang X, Tredget EE, Wu Y (2012) Dynamic signals for hair follicle initiation of hair follicle development. Dev Cell 2:643–53 development and regeneration. Stem Cells Dev 21:7–18 Beaudoin GM III, Sisk JM, Coulombe PA et al. (2005) Hairless triggers Zhang K, Lian X, Li Y et al. (2009) [Expression profile of wnt10b in cycle of reactivation of hair growth by promoting Wnt signaling. Proc Natl Acad mouse hair follicle]. Acta Academiae Medicinae Militaris Tertiae Sci USA 102:14653–8 31:891–4

48 Journal of Investigative Dermatology (2013), Volume 133