Accelerated differentiation of melanocyte stem cells contributes to the formation of hyperpigmented maculae.

It has been reported that the abnormal regulation of melanocyte stem cells (McSCs) causes hair greying; however, little is known about the role of McSCs in skin hyperpigmentation such as solar lentigines (SLs). To investigate the involvement of McSCs in SLs, the canonical Wnt signalling pathway that triggers the differentiation of McSCs was analysed in UVB‐induced delayed hyperpigmented maculae in mice and human SL lesions. After inducing hyperpigmented maculae on dorsal skin of F1 mice of HR‐1× HR/De, which was formed long after repeated UVB irradiation, the epidermal Wnt1 expression and the number of nuclear β‐catenin‐positive McSCs were increased as compared to non‐irradiated control mice. Furthermore, the expression of dopachrome tautomerase (Dct), a downstream target of β‐catenin, was significantly upregulated in McSCs of UVB‐irradiated mice. The Wnt1 expression and the number of nuclear β‐catenin‐positive McSCs were also higher in human SL lesions than in normal skin. Recombinant Wnt1 protein induced melanocyte‐related genes including Dct in early‐passage normal human melanocytes (NHEMs), an in vitro McSC model. These results demonstrate that the canonical Wnt signalling pathway is activated in SL lesions and strongly suggest that the accelerated differentiation of McSCs is involved in SL pathogenesis.

differentiation; melanocyte stem cell; solar lentigo; Wnt

Solar lentigines (SLs) are characterised by hyperpigmented maculae, commonly seen on sun‐exposed areas of the skin, such as the face and arms [1] . Because of the negative impact that solar lentigines (SLs) have on the quality of life of patients, elucidation of the pathogenic mechanism of SLs is highly anticipated to develop effective treatments. According to previous studies, elevated expressions of major melanogenesis‐stimulating factors, endothelin‐1 (ET‐1) and stem cell factor (SCF) [2] , [3] , a significant increase in the number of melanocytes and rete ridge elongation were observed in SL lesions [1] , [4] , [5] , [6] , [7] ; however, the detailed mechanism of SL pathogenesis has not been fully understood.

Wnt signalling pathways can be divided into three classes: the canonical pathway, which is also called the Wnt/β‐catenin pathway, the non‐canonical pathways, which are composed of planer cell polarity (PCP) and the calcium‐dependent pathway [8] . In mammals, 19 Wnt ligands and 10 frizzled (Fzd) receptors have been identified to date. In the Wnt/β‐catenin signalling pathway, canonical Wnts such as Wnt1, Wnt3a and Wnt10b bind to Fzd and coreceptor low‐density lipoprotein‐related protein (Lrp) 5/6 and block the phosphorylation of cytoplasmic β‐catenin protein by the destruction complex. Non‐phosphorylated β‐catenin does not undergo proteasomal degradation and translocates to the nucleus to regulate gene expression by interaction with transcription factor LEF‐1, causing various cellular events, such as proliferation, differentiation, migration and adhesion [9] . The signalling pathway is inhibited by extracellular antagonists, secreted frizzled‐related proteins and dickkopfs [10] . The secreted frizzled‐related proteins bind directly to Wnt ligands, and dickkopfs bind to Lrp5/6. Wnt5a, one of the non‐canonical Wnt ligands, also inhibits Wnt3a‐induced canonical Wnt signalling by an antagonistic effect [11] .

Wnt signalling regulates melanocyte differentiation in embryonic development [12] , [13] . Recently, it has been reported that Wnt/β‐catenin signalling regulates the differentiation of mouse melanocyte stem cells (McSCs), which exist in the bulge region of adult hair follicles [14] . When the hair cycle progresses from the resting phase (telogen) to growth phase (anagen), the activated signalling pathway triggers the differentiation of McSCs into mature melanocytes located in hair bulbs. Mature melanocytes supply melanins to the surrounding keratinocytes [15] . UVB irradiation also activates Wnt/β‐catenin signalling and then induces the differentiation of mouse McSCs into melanocytes located in the interfollicular epidermis [16] . Although it is not clear that human McSCs differentiate into epidermal melanocytes via Wnt/β‐catenin signalling activated by UVB irradiation, the skin of human and mice responds to UV radiation in a very similar manner (e. g. increased expression of SCF). It is known that differentiation of human melanocytes is inhibited by Dkk1, an inhibitor of Wnt/β‐catenin signalling [17] , [18] ; therefore, the signalling pathway is strongly assumed to be important for both human and mouse McSC differentiation.

There is accumulating evidence about the molecular mechanism of the differentiation of McSCs; however, little is known about the relationship between the differentiation of McSCs and pigmentary disorders such as SLs. To address this issue, we analysed Wnt/β‐catenin signalling in UVB‐induced delayed hyperpigmented maculae in mice. F1 mice of HR‐1× HR/De present with a unique feature that induces delayed hyperpigmented maculae after UVB irradiation [5] , [19] , [20] , [21] , [22] . Based on the results from the above‐described animal model, we further compared the status of Wnt/β‐catenin signalling in healthy human skin and SL lesions and investigated the role of the differentiation of McSCs in the formation of SLs.

F1 hairless mice of HR‐1× HR/De were obtained from Japan SLC (Shizuoka, Japan). All animal experiments were approved by both the Nippon Menard Research Laboratories Subcommittee on Research Animal Care and the Education and Research Center for Animal Models of Human Diseases of Fujita Health University on Research Animal Care. Seven‐week‐old mice were exposed to UVB radiation three times a week by Toshiba FL‐20 SE fluorescent lamps (Toshiba Electric, Tokyo, Japan). The daily radiation was 100 mJ/cm2.

Samples of SLs and normal skin were collected from the faces of 11 subjects of each group (SL group: three males and eight females, average age 63.6 ± 15.3; normal skin group: one male and 10 females, average age 66.7 ± 14.7) at Fujita Health University Hospital and were examined by conventional Fontana‐Masson staining. After ensuring that the patients fully understood the study objective and other related information, written informed consent was obtained from each subject. This study was approved by the Ethics Committee of Fujita Health University.

Epidermal sheets were separated from collected mouse skin by forceps after incubation in 2M NaBr at 37°C for 2 h and were fixed with 4% paraformaldehyde to perform immunostaining. Frozen sections were prepared in a cryostat (Carl Zeiss, Thornwood, NY, USA) from mouse skin samples embedded in OCT compound and were fixed with acetone at −20°C for 10 min. The formalin‐fixed and paraffin‐embedded sections prepared from the skin biopsies were deparaffinised and boiled in Target Retrieval Solution (Dako, Glostrup, Denmark). After washing with PBS, these sheets and sections were blocked with 5% normal donkey serum for 1 h and were then incubated with primary antibodies overnight at 4°C. After washing with phosphate‐buffered saline (PBS), the epidermal sheets or sections were incubated with secondary antibodies conjugated to Alexa488 and Alexa594 (Life Technologies, Carlsbad, CA, USA). 4′,6‐Diamidino‐2‐phenylindole (DAPI; Vector Laboratories, Burlingame, CA, USA) was used for nuclear staining. Detailed information about antibodies is described in Table S1.

To visualise dopa‐positive melanocytes, epidermal sheets were fixed with 10% formalin for 30 min and then incubated in 0.1% L‐dopa in PBS for 3 h at 37°C.

After immunostained specimens were observed by fluorescent microscopy, microscopic images were collected randomly. Relative fluorescent units (RFU) of the images were analysed using NIH Image, and the relative expression levels of Wnt1 or Wnt2 were expressed as RFU per unit area. After counting nuclear β‐catenin+/Kit+ cells and nuclear β‐catenin−/Kit+ cells in mouse hair follicles, the percentages of nuclear β‐catenin+/Kit+ cells in total Kit+ cells were calculated. Similarly, nuclear β‐catenin+/MITF+ cells and nuclear β‐catenin−/MITF+ cells were counted in the human bulge region to calculate the percentages of β‐catenin+/MITF+ cells in total MITF+ cells.

Total RNA were extracted from mouse epidermal sheets using TRIzol Reagent (Life Technologies). cDNA was synthesised by reverse transcription. Real‐time semiquantitative RT‐PCR was performed with the SuperScript III Platinum Two‐Step qRT‐PCR kit (Life Technologies) using the StepOnePlus Real‐time RT‐PCR system (Life Technologies). The primer sequences are shown in Supplementary Table S2. Amplification was normalised to a housekeeping gene, glyceraldehyde‐3‐phosphate dehydrogenase (Gapdh) or 18S ribosomal RNA. Differences between samples were quantified based on the ∆∆Ct method.

Mouse epidermal sheets were lysed on ice with 2% SDS, and the lysates were sonicated. Then, total protein samples were prepared by centrifugation. Proteins were separated by SDS‐PAGE and transferred to PVDF membrane (Millipore, Billerica, MA, USA). The membrane was incubated with 5% skim milk followed by primary antibodies. HRP‐conjugated secondary antibodies were used in combination with ECL Prime detection system (GE healthcare Life Science, Uppsala, Sweden) to visualise immunoreactive bands.

Mouse dorsal skin cut into fine pieces (size: 5 mm2) by scissors was incubated on 200 U/ml dispase overnight at 4°C. Then, the epidermis was separated from the dermis with forceps. After incubation in 0.25% trypsin at 37°C for 10 min, the obtained cell suspension was filtrated and centrifuged, and then, cells were isolated by fluorescence‐activated cell sorting (FACS; FACSAria; BD Biosciences, San Jose, CA, USA). In brief, the cells were stained with biotin‐conjugated anti‐Fzd4 antibodies (R&D Systems, Minneapolis, MN, USA) at 4°C for 30 min, followed by incubation with APC‐conjugated StreptAvidin at 4°C for 15 min. After staining with antibodies, Fzd4+ cells were sorted with FACS.

Normal human epidermal melanocytes were purchased from TOYOBO (Osaka, Japan) and maintained between passages 1 and 3 in Medium 254 containing Human melanocyte growth supplement (Life Technologies). The cells were seeded onto 3.5‐mm plates at a density of 5 × 104 cells/well and were incubated overnight. The cells were then incubated with recombinant human Wnt1 protein (Peprotec, Rocky Hill, NJ, USA) for 8–24 h to analyse intra‐cellular β‐catenin localisation and melanogenesis‐related gene expressions.

Data are presented as the mean ± SD. P < 0.05 was considered significant. Statistical analysis was performed using a t‐test.

F1 hairless mice of HR‐1× HR/De present with a unique feature that induces delayed hyperpigmented maculae after UVB irradiation [5] , [19] . Based on previous studies, mice were irradiated with UVB (100 mJ/cm2) three times a week for 4 weeks without any additional exposure (Figure S1a). The dorsal skin of mice in the UVB‐irradiated group became apparently darker than the non‐irradiated group 4 weeks after the first irradiation (Figure S1b). After cessation of UVB irradiation, the skin colour of the UVB‐irradiated group became similar to the non‐irradiated group at 8 weeks, and hyperpigmented maculae subsequently appeared at 16 weeks and became distinct 36 weeks after the first irradiation (Fig. [NaN] a and Figure S1b). These findings of skin appearance were consistent with the difference in the skin brightness L* value between non‐irradiated and irradiated groups, the number of epidermal melanocytes and the melanin deposition in the epidermis (Fig. [NaN] b–d and Figure S1c,d). At 36 weeks after UVB irradiation, the expressions of melanogenesis‐related genes such as tyrosinase‐related protein 1 (Tyrp1), dopachrome tautomerase (Dct) and tyrosinase (Tyr) were significantly upregulated (Fig. [NaN] e). Solar lentigines are hyperpigmented maculae, which occur on sun‐exposed areas long after UV exposure; and they are histologically characterised by an increase in the number of melanocytes and melanogenesis in the epidermis. These characteristics were similar to the features of hyperpigmented maculae observed on the skin of UVB‐irradiated mice as previously reported [20] , [21] . These results indicated that UVB irradiation induced SL‐like hyperpigmented maculae in the dorsal skin of mice; this mouse model was therefore considered to be suitable for investigation of the SL pathogenic mechanism.

The formation of hyperpigmented maculae in mice was associated with an increased number of epidermal melanocytes; therefore, we hypothesised that the accelerated differentiation of McSCs causes the formation of hyperpigmented maculae. To investigate the involvement of McSCs in the formation of hyperpigmented maculae, we analysed the gene expression levels of Wnt signalling molecules in the epidermis, because recent studies have revealed that the canonical Wnt signalling pathway triggers the differentiation of McSCs [15] , [16] . As compared to the UVB non‐irradiated group, the mRNA expressions of canonical Wnt ligands, Wnt1 and Wnt2, were significantly upregulated, whereas Wnt9a and a non‐canonical Wnt ligand, Wnt5a, were downregulated in the epidermis of the UVB‐irradiated group (Fig. [NaN] a). The reduced expression of Wnt9a was probably not related to the differentiation of McSCs because Wnt9a is not a major Wnt ligand, which is expressed in the epidermis [23] , [24] . The period of increased expression of Wnt1 and Wnt2 was consistent with that of the formation of hyperpigmented maculae, because the mRNA expression levels of Wnt1 and Wnt2 were not increased at 8 weeks after the first irradiation of UVB, when the skin colour of the UVB‐irradiated group became similar to the non‐irradiated group (Fig. [NaN] b). In accordance with the mRNA expressions, elevated expressions of Wnt1 and Wnt2 proteins in the epidermis were also detected by Western blot analysis (Fig. [NaN] c) and immunostaining of skin sections (Fig. [NaN] d–k). The expression pattern of these Wnt ligands was not uniform and was strong in a part (Fig. [NaN] e–g and h–k). Without Wnt ligands, β‐catenin is phosphorylated by the destruction complex and undergoes proteasomal degradation [8] . Phosphorylated β‐catenin, its inactivated form, was also reduced in irradiated group (Fig. [NaN] c). Because of loss of its antagonistic effect, Wnt5a reduction is likely to activate the canonical Wnt signalling pathway [11] . Although the slight reduction in Wnt5a protein expression was detected by Western blotting, no difference in Wnt5a expression was observed between UVB‐irradiated and non‐irradiated groups by immunostaining (Figure S2a–c). No decrease in the gene expressions of Wnt inhibitors, such as secreted frizzled‐related protein 1–5 and dickkopf 1–4, was observed (data not shown). These results suggest that canonical Wnt signalling is activated in mouse delayed hyperpigmented maculae and thus raises the possibility that the differentiation of McSCs is accelerated.

β‐Catenin activated by canonical Wnt molecules does not undergo proteasomal degradation and translocates to the nucleus and then regulates target gene expression. To investigate whether differentiation of McSCs was accelerated, the localisation of β‐catenin in melanocyte lineage cells was analysed by immunostaining. Immature melanocyte lineage cells can be detected as Kit+ cells in the bulge region of hair follicles [25] . As compared to the UVB non‐irradiated group, the number of nuclear located β‐catenin+/Kit+ cells in the total Kit+ cells was larger in the irradiated group as well as the number of Kit+ cells in a hair follicle (Fig. [NaN] a–l). FACS analysis on Fzd4, a cell surface marker of McSCs [26] , also showed that the number of McSCs increased in the UVB‐irradiated group (Fig. [NaN] m). It has been reported that β‐catenin directly induces Dct expression during the differentiation of McSCs [27] . After isolation of McSCs by FACS, the mRNA expression level of Dct was analysed. As a result, real‐time RT‐PCR analysis revealed that the Dct expression was significantly higher in McSCs of the UVB‐irradiated group than the non‐irradiated group (Fig. [NaN] n). To exclude the possibility that the proliferation of residential melanocytes and melanoblasts in epidermis contribute to the formation of SL‐like hyperpigmented maculae, dual immunostaining for Ki67, a cell proliferation marker, and Kit or Tyrp1 was performed on skin sections. Although the number of Kit+ cells and Tyrp1+ cells was increased in the irradiated group as compared to non‐irradiated control group (Fig. [NaN] o–z), Ki67+/Kit+ cells or Ki67+/Tyrp1+ cells were hardly detected; there was no significant difference in the rate of these cells (Fig. 3aa). Therefore, these results strongly suggest that differentiation of McSCs in mouse hyperpigmented maculae is accelerated by the activated canonical Wnt signalling pathway and thus speculates its contribution to the formation of SLs.

Next, we examined whether the observations in the mouse SL model can be applied to human SLs by analysing skin sections obtained from human normal skin or SL lesions. Immunostaining revealed that, although there was no difference in Wnt2 expression level between normal skin and SL sections, which indicates human SL lesions are not completely in agreement with the hyperpigmented maculae in mouse model, Wnt1 expression was markedly upregulated in SL lesions (Fig. [NaN] a–d, Figure S3). Image analysis also confirmed that Wnt1 expression, not Wnt2, was significantly higher in SL lesions than normal skin (Fig. [NaN] e). It has been reported that human McSCs were identified as MITF+ cells in the outer root sheath of the bulge region [14] . To analyse the localisation of β‐catenin in McSCs, dual immunostaining for β‐catenin and MITF was performed on SL and normal skin sections. As a result, it was observed that nuclear β‐catenin+/MITF+ cells were increased in the bulge region of SL lesions (Fig. [NaN] f–z). The rate of nuclear β‐catenin+/MITF+ cells in total MITF+ cells was significantly higher in SL lesions (Fig. 4aa). Therefore, these results demonstrate that the Wnt1 expression and differentiation of McSCs were strongly induced in human SL lesions, as similarly observed in the mouse model of SLs.

Finally, we examined the impact of Wnt1 on early‐passage normal human epidermal melanocytes (NHEMs), an in vitro human McSC model. In a previous study, NHEMs maintained between passages 1–4 were used to examine the effect of TGF‐β1 on the maintenance of McSCs [28] . Wnt1 treatment successfully induced the nuclear translocation of β‐catenin in NHEMs (Figure S4a and b) and significantly upregulated the expressions of melanogenesis‐related genes in NHEMs (Figure S4d). In long‐cultured NHEMs (passage 10), mRNA expression levels of these genes were significantly higher than early‐passage NHEMs (passage 3) without Wnt1 treatment (Figure S4c). Although Wnt1 treatment also enhanced the expressions of melanogenesis‐related genes in long‐cultured NHEMs, the increase rates of the gene expressions were lower than early‐passage NHEMs (Figure S4e). Based on the results, it was considered that early‐passage NHEMs were proper as an in vitro model for McSCs, because long‐cultured NHEMs were thought to lose their immaturity. These results indicate that Wnt1 possibly accelerates the differentiation of McSCs and is probably involved in the formation of SLs.

Enhanced melanogenesis and an increased number of epidermal melanocytes are pathological characteristics of SLs. Previous studies have demonstrated that melanogenesis‐related factors such as ET‐1 and SCF in keratinocytes were upregulated in SL lesions [2] , [3] . These factors are well known to stimulate melanogenesis and the proliferation of melanocytes and are believed to play a central role in the formation of SLs. However, previous studies lacked an analysis of the relationship between SL pathogenesis and the source of melanocytes (i.e. McSCs). In the present study, we analysed the canonical Wnt signalling pathway, which triggers the differentiation of McSCs [15] , [16] , in delayed hyperpigmented maculae in mice and human SL lesions, and demonstrated that the accelerated differentiation of McSCs was involved in the formation of SLs.

The canonical Wnt signalling pathway is closely related to melanocyte biology. In the embryonic developmental stage, Wnt1 and Wnt3a, respectively, regulate the differentiation and proliferation of both neural crest cells and their progeny, melanoblasts [12] . DKK1, an antagonistic protein to Wnt ligands, was highly expressed by fibroblasts in hypopigmented skin such as the sole of the foot and the palm, prevented melanogenesis and decreased the density of melanocytes in the palmoplantar skin [18] . This signalling pathway is also important for McSCs to trigger the differentiation into follicular melanocytes and epidermal melanocytes [15] , [16] . Therefore, abnormalities in the canonical Wnt signalling pathway are likely to cause pigmentary disorders by dysregulation of the development or differentiation of melanocyte lineage cells. In the present study, we examined whether abnormalities in Wnt signalling are involved in the pathogenesis of SLs. The results showed that Wnt1 and Wnt2 were highly expressed and that the numbers of nuclear β‐catenin+ McSCs were increased in delayed hyperpigmented maculae in mice (Figs [NaN] and [NaN] ). Although there was no elevation of Wnt2 expression in human SL lesions, which indicates human SL lesions are not completely in agreement with the hyperpigmented maculae in mouse model, upregulation of Wnt1 and an increased number of nuclear β‐cateninin+ McSCs in the lesions were detected (Fig. [NaN] ). Furthermore, Wnt1 promoted gene expressions related to melanogenesis in an in vitro McSC model (Figure S3). Taken together, the canonical Wnt signalling pathway was activated in SL lesions and thus strongly suggests that accelerated differentiation of McSCs causes an increase in the number of melanocytes in the epidermis, which is a pathological characteristic of SL. As the signalling pathway is also closely related to the melanogenesis in mature melanocytes (Figure S3), it raised the possibility that melanogenesis was activated in the melanocytes [17] , [18] . It has been reported that immature melanoblasts as well as mature melanocytes reside in the human epidermis [29] ; therefore, Wnt1/β‐catenin signalling may induce the differentiation of epidermal melanoblasts. Although the melanocyte lineage, McSCs, melanoblasts and melanocytes, is likely to be involved in the formation of SLs, it is suggested that McSCs, the source of melanocytes, play the most significant role in SLs, considering that SL formation needs long period of time and it persists for a long time. Further study on multiple phases of SL formation will clarify the precise role of McSCs in SLs.

The most important issue to be elucidated is why elevation of Wnt1 expression occurs in SL lesions. SLs are considered to be a result of cumulative UVB exposure and appear on the sun‐exposed skin of middle‐aged and older people [1] . Based on such a phenomenon, it is thought that the combination of cumulative UVB radiation and chronological ageing is likely to cause the formation of SLs. To support this hypothesis, the effect of previous cumulative exposure of UVB should be sustained for a long period until people age. Recently, it has been reported that CpG dinucleotide methylation and post‐translational modification of histone tails are closely involved in cell fate and function in both the developmental stage and the maintenance of adult tissues [30] , [31] . Abnormalities in these epigenetic processes cause various diseases, especially cancers [32] . If cumulative UVB exposure induces an abnormal DNA methylation pattern in epidermal stem cells, it would be a long‐lasting event because adult stem cells are considered to be maintained for a long period in adult tissues from birth onward. In addition to UVB, chronological ageing may also induce other epigenetic abnormalities. We speculate that aberrant activation of canonical Wnt signalling is induced by these epigenetic abnormalities and accelerates the differentiation of McSCs, as seen in SLs. Although it was thought accelerated differentiation of McSCs might cause a reduction in the number of McSCs, we observed increased numbers of McSCs in both mouse hyperpigmented maculae (Fig. [NaN] ) and human SL lesions [33] . Epigenetic abnormalities may also induce such alteration in the number of McSCs. We believe that further investigation of epigenetic alterations between normal skin and SL lesions will provide with more information to understand the pathogenic mechanism of SLs.

To our knowledge, hair greying is the only situation in which involvement of the abnormal regulation of McSCs has been reported. It has been suggested that dysfunction of McSC maintenance caused their ectopic differentiation into mature melanocytes in the bulge area, which in turn resulted in the loss of McSCs [34] . Our findings that the differentiation of McSCs is accelerated in SL lesions imply that the abnormal regulation of McSCs is likely to be involved in other pigmentary disorders characterised by hyperpigmentation as well as depigmentation, such as hair greying. Therefore, we believe that a study focusing on McSC regulation as well as differentiated melanocytes will shed light on the pathogenic mechanism of pigmentary disorders in the future.

Takaaki Yamada, Seiji Hasegawa, Kayoko Matsunaga and Hirohiko Akamatsu designed the research study. Takaaki Yamada, Naoki Yamamoto, Hiroshi Mizutani, Satoru Nakata, Kayoko Matsunaga and Hirohiko Akamatsu wrote the paper. Takaaki Yamada, Yu Inoue, Yasushi Date, Masaru Arima and Masayuki Takahashi performed the research. Takaaki Yamada, Akiko Yagami and Yohei Iwata analysed the data.

The authors state no conflict of interest.

Graph: Analysis of the formation of delayed hyperpigmented maculae in F1 mice of HR‐1× HR/De. (a) Photographs of dorsal skin appearances of UVB non‐irradiated mice (UVB (−)) and irradiated mice (UVB (+)) were taken at 36 weeks after the first UVB irradiation. Scale bar = 5 mm. (b and c) Epidermal sheets were obtained from mouse dorsal skin at 36 weeks after the first UVB irradiation and were immunostained with anti‐Tyrp1 antibody or dopa reaction. Scale bar = 50  μ m. (d) Fontana‐Masson staining was performed on skin sections obtained from mouse dorsal skin at 36 weeks after the first UVB irradiation. Scale bar = 50  μ m. (e) At 36 weeks after UVB irradiation, mRNA expression levels of melanogenesis‐related genes in the epidermis were analysed by real‐time RT‐PCR. ** P  <   0.01 versus non‐irradiated group as determined by t ‐test (mean ± SD).

Graph: Canonical Wnt signalling pathway was activated in solar lentigines (SL) model. (a–c) At 36 weeks after the first UVB irradiation, total mRNA and protein were extracted from epidermal sheets of the mice, and then, real‐time RT‐PCR (a) and Western blotting (c) were performed. At 8 weeks after the first UVB irradiation, mRNA expression levels of Wnt1 and Wnt2 were analysed by real‐time RT‐PCR (b). * P  <   0.05 versus UVB non‐irradiated group as determined by t ‐test (mean ± SD). ND: not detected. (d–k) Immunostaining for Wnt1 (d–g) or Wnt2 (h–k) was performed on skin sections prepared from the dorsal skin of UVB non‐irradiated mice (d and h) and irradiated mice (e and i) at 36 weeks after UVB irradiation. Nuclei were stained with DAPI (blue). Boxed regions in (e) and (i) were magnified in (f and g) and (j and k), respectively. Scale bar = 30  μ m.

Graph: Accelerated differentiation of mouse melanocyte stem cells (McSCs) in solar lentigines (SL) model. (a–j) Localisation of β‐catenin in Kit + cells was analysed by immunostaining of hair follicles of UVB non‐irradiated mice (a–e) and irradiated mice (f–j) at 36 weeks after UVB irradiation. Nuclei were stained with DAPI (blue). Dotted line in (a) and (f) depicts the boundary of hair follicles. Scale bar = 50  μ m. The boxed regions of (a) and (f) are shown enlarged (b–e, g–j), respectively. Dotted line depicts Nuclei of Kit + cells. Scale bar = 10  μ m. Arrowheads indicate Kit + cells. (k and l) After counting nuclear β‐catenin + /Kit + cells and nuclear β‐catenin − /Kit + cells, the percentages of nuclear β‐catenin + /Kit + cells in the total Kit + cells (k) and total Kit + cells in hair follicles were calculated. (m and n) FACS analysis was performed on epidermal cell suspensions obtained from the mice at 36 weeks after UVB irradiation. The percentages of Fzd4 + cells in total epidermal cells are shown (m). After isolation by FACS, mRNA expression levels of Dct in Fzd4 + cells were analysed by real‐time PCR (n). ** P  <   0.01 versus UVB non‐irradiated group as determined by t ‐test (mean ±  SD ). (o–z) Immunostaining for Ki67 and melanocyte markers, Kit (o–t) or Tyrp1 (u–z), was performed on mouse skin sections obtained at 36 weeks after UVB irradiation. Nuclei were stained with DAPI (blue). Arrowheads indicate Kit + or Tyrp1 + cells. Scale bar = 50  μ m. (aa) After counting Ki67 + /Kit + cells and Ki67 + /Tyrp1 + cells, the percentages of Ki67 + cells in total Kit + cells or Tyrp1 + cells were calculated.

Graph: Analysis of Wnt expression and β‐catenin activation in human solar lentigines ( SL ) lesions. (a–d) Wnt1 (a, b) and Wnt2 (c, d) expressions were analysed by immunostaining on skin sections obtained from human normal skin (a, c) and SL lesions (b, d). Nuclei were stained with DAPI (blue). Scale bar = 50  μ m. (e) To compare Wnt1 and Wnt2 expression levels between normal and SL skin sections, relative fluorescent units ( RFU ) were analysed using image analysis software. Relative expression levels of Wnt1 or Wnt2 were expressed as RFU per unit area. * P  <   0.05 versus normal skin as determined by t ‐test. (mean ±  SD ). (f–z) Immunostaining for β‐catenin and MITF was performed on the bulge region of normal skin (f–m) and SL lesions (n–z). Dotted line in (f–i) and (n–q) depicts the boundary of hair follicles and arrowheads indicate MITF+ cells. Scale bar = 20  μ m. The boxed regions of (i) and (q) are shown enlarged (j–m, r–z), respectively. Dotted line depicts Nuclei of MITF+ cells. (aa) Nuclear β‐catenin + / MITF+ cells and nuclear β‐catenin − / MITF+ cells were counted in the human bulge region. The percentages of β‐catenin + / MITF+ cells in total MITF+ cells were calculated. * P  <   0.05 versus normal skin as determined by t ‐test (mean ±  SE ).

Graph: Data S1. Materials and methods. Figure S1. Analysis of the formation of delayed hyperpigmented spots in F1 mice of HR‐1× HR/De. Figure S2. Wnt5a expression in SL model. Figure S3. Wnt1 expression in human normal skin and SL lesions. Figure S4. Effects of Wnt1 treatment on an in vitro human McSC model. Table S1. List of antibodies. Table S2. Primers used in real‐time RT‐PCR analysis.