Advertisement

Bone Morphogenetic Protein 5 Regulates the Number of Keratinocyte Stem Cells from the Skin of Mice

      Understanding keratinocyte stem cell regulation is important in understanding the pathogenesis of wound healing and nonmelanoma skin cancer. We previously used a sensitive and quantitative assay for in vitro keratinocyte colony formation and mapped the keratinocyte stem cell locus (Ksc1) on mouse chromosome 9. Examination of the candidate genes in this locus disclosed a sequence variant in the gene for bone morphogenetic protein 5 (Bmp5). In this report, we used a naturally occurring mouse with a null mutation in this gene to probe stem cell properties in mouse epidermis. We found that the mutant keratinocytes had a significant reduction in the size and number of clonogenic keratinocytes. The mutant mice had a 50% reduction in the number of label-retaining cells when compared with their littermates. Addition of exogenous Bmp5 protein increased the number and size of keratinocyte colonies in the mutant as well as their wild-type littermates. Surprisingly, the mutant mice showed at least a 2-fold increase in skin tumor susceptibility over their littermates. We conclude that a naturally occurring mutation in Bmp5 affects keratinocyte stem cell proliferation, and skin tumor susceptibility, and is a candidate stem cell regulatory gene in the Ksc1 locus.

      Abbreviations

      BMP5
      bone morphogenetic protein 5
      ESC
      epidermal stem cell
      Ksc1
      keratinocyte stem cell locus 1
      LRC
      label-retaining cell
      rhBMP5
      recombinant human BMP5
      TA
      transit-amplifying cell
      TPA
      12-O-tetradecanoylphorbol-13-acetate

      Introduction

      The cutaneous epithelium is a continually renewing tissue that harbors stem cells responsible for differentiation into the lineages of the epidermis under normal condition and in response to wounding. In addition, the behavior of epidermal stem cells (ESCs) has been shown to be associated with a variety of skin diseases including skin tumors.
      Some stem cells from mouse skin can be distinguished according to their slowly cycling behavior in vivo and their ability to form colonies in vitro. Mouse epidermal cells in the basal layer are organized into morphologically defined epidermal proliferative units (
      • Potten C.S.
      Stem cells in epidermis from the back of the mouse.
      ). Epidermal whole mounts of the mouse back skin showed a population of ∼10 basal cells and a hexagonal suprabasal group of maturing, flattened cells. Several laboratories have demonstrated that there is at least one slowly cycling basal cell in the cluster of 3 to 4 basal cells in the central column of suprabasal cells (
      • Bickenbach J.R.
      Identification and behavior of label-retaining cells in oral mucosa and skin.
      ;
      • Mackenzie I.C.
      • Bickenbach J.R.
      Localization of label-retaining cells in mouse epithelia.
      ;
      • Morris R.J.
      • Fischer S.M.
      • Slaga T.J.
      Evidence that the centrally and peripherally located cells in the murine epidermal proliferative unit are two distinct cell populations.
      ;
      • Potten C.S.
      Cell cycles in cell hierarchies.
      ).
      • Cotsarelis G.
      • Sun T.-T.
      • Lavker R.M.
      Label-retaining cells reside in the bulge area of pilosebaceous unit: implications for follicular stem cells, hair cycle, and skin carcinogenesis.
      found that ESCs also resided in the bulge portion of the outer root sheath in both mouse and human hair follicles. Interestingly, this finding is also supported by evidence that epidermal cells from the mouse hair follicles are capable of regenerating the epidermis following irradiation (reviewed by
      • Potten C.S.
      Cell cycles in cell hierarchies.
      ), and they are also capable of regenerating differentiated hair follicles, sebaceous glands, and interfollicular epidermis even after repeated abrasions (
      • Argyris T.S.
      Promotion of epidermal carcinogenesis by repeated damage to mouse skin.
      ).
      In addition, the proliferative characteristic of epidermal cells can be observed in in vitro colony assays that demonstrated that colonies from adult mouse epidermal cells at low densities have an extensive growth potential and can generate a large number of cells that are defined as holoclones (
      • Barrandon Y.
      • Green H.
      Cell size as a determinant of the clone-forming ability of human keratinocytes.
      ). The number of mouse epidermal cell colonies remains essentially constant throughout most of the adult life as well as following carcinogenic initiation; however, it varies among different strains of mice (
      • Morris R.J.
      • Tacker K.C.
      • Fischer S.M.
      • et al.
      Quantitation of primary in vitro clonogenic keratinocytes from normal adult murine epidermis, following initiation, and during promotion of epidermal tumors.
      ;
      • Popova N.V.
      • Tryson K.A.
      • Wu K.Q.
      • et al.
      Evidence that keratinocyte colony number is genetically controlled.
      ).
      Our laboratory has shown that the number of epidermal cell colonies is in fact a genetically definable quantitative trait (
      • Popova N.V.
      • Tryson K.A.
      • Wu K.Q.
      • et al.
      Evidence that keratinocyte colony number is genetically controlled.
      ). To identify genes or factors that regulate stem cell number, a genome-wide linkage analysis with colonies as a phenotype was performed, and several loci showed significant linkage in the propensity to form either larger colonies or smaller colonies (
      • Popova N.V.
      • Teti K.A.
      • Wu K.Q.
      • et al.
      Identification of two keratinocyte stem cell regulatory loci implicated in carcinogenesis.
      ). Through sequencing analysis of the open reading frames in the keratinocyte stem cell locus 1 (Ksc1) locus on chromosome 9 that showed the most significant linkage, we detected a variant in the bone morphogenetic protein 5 (Bmp5) gene (Popova et al., in preparation).
      Bone morphogenetic proteins are members of the transforming growth factor-β family that has a role in proliferation, apoptosis, and differentiation in various tissues and organs including skin (
      • Hogan B.L.
      Bone morphogenetic proteins in development.
      ). Therefore, we chose Bmp5 as a candidate gene potentially regulating ESC number for further analysis. We found a naturally occurring mouse mutation in the Bmp5 gene and we report here that the short-ear mouse, previously characterized for its cartilage and skeletal anomalies by
      • Kingsley D.M.
      • Bland A.E.
      • Grubber J.M.
      • et al.
      The mouse short ear skeletal morphogenesis locus is associated with defects in a bone morphogenetic member of the TGF beta superfamily.
      and
      • King J.A.
      • Marker P.C.
      • Seung K.J.
      • et al.
      BMP5 and the molecular, skeletal, and soft-tissue alterations in short ear mice.
      , has a previously unreported skin stem cell phenotype including a 2-fold increase in skin tumor susceptibility.

      Results

      Bmp5 is expressed in the interfollicular epidermis and the hair follicles of the wild-type mice

      To assess Bmp5 expression patterns, we performed immunohistochemical analysis of Bmp5 on skin sections from all genotypes including wild-type C57BL/6 and BALB/c as low and high controls. As shown in Figure 1, Bmp5 is strongly expressed in the hair follicles, particularly in the hair follicle bulge, and in the interfollicular epidermis in both C57BL/6 and BALB/c mice, the parental strains used for mapping. It seems that Bmp5 is distinctly less expressed in the BALB/c epidermis than that of C57BL/6. However, it has been reported that other bone morphogenetic proteins such as Bmp2 and Bmp4, which are important in the regulation of hair regeneration, showed periodic expression patterns (
      • Plikus M.V.
      • Mayer J.A.
      • de la Cruz D.
      • et al.
      Cyclic dermal BMP signaling regulates stem cell activation during hair regeneration.
      ). Therefore, our data suggest that C57BL/6 and BALB/c may undergo different cyclic regeneration of the hair follicles, and thus showed different levels of Bmp5 expression at 8 weeks of age. The Bmp5 short-ear mice have essentially no immunoreactivity to the Bmp5 antibody when compared with the wild-type littermates, which is consistent with the null mutation in these mice.
      Figure thumbnail gr1
      Figure 1Localization of bone morphogenetic protein 5 (Bmp5) in mouse hair follicles and interfollicular epidermis of wild-type but not mutant mouse skin. Adult mouse skin was collected for paraffin-embedded histological sections for (a) C57BL/6, (b) BALB/c, (c) wild-type, and (d) homozygous short-ear mutant mice at ∼7 to 8 weeks of age, and BMP5 immunohistochemistry was performed. B, hair follicle bulge; D, dermis; E, epidermis; HF, hair follicle; SG, sebaceous gland. Note the intense staining in the hair follicle and in an occasional interfollicular cell of the wild-type epithelium. In the homozygous short-ear mutant skin, Bmp5 staining was consistently absent from hair follicles and interfollicular epidermis. Scale bars=30μm.

      The short-ear mutant mouse shows a decrease in the number and size of ESC colonies

      In Figure 2, we compared the number of colonies between C57BL/6, BALB/c, and all genotypes of the short-ear mutant mice. After 2 weeks of culture, the average number of colonies from BALB/c mice is ∼64% of that from C57BL/c mice (Figure 2b). In addition, the average size of the colonies from BALB/c is ∼36% smaller than that from C57BL/6. We observed no difference in the average numbers of colonies from wild-type and heterozygous groups. However, the short-ear homozygous mouse showed a significant (28%) decrease in colony number. The results indicated that only one copy of functional Bmp5 is sufficient for its function in stem cell number regulation. In addition to a marked reduction in colony number, we also observed that the short-ear mutants also have a reduction in colony size (Figure 2c).
      Figure thumbnail gr2
      Figure 2Quantification of mouse keratinocyte colonies. Keratinocytes were isolated from 7- to 8-week-old mice. (a) The photographs of the colony assays in 60-mm dishes are shown. The (b) number and (c) size of mouse keratinocyte colonies from all genotypes were quantified after 2 weeks of cultivation on irradiated 3T3 fibroblasts as a feeder layer. Note that the number of colonies is significantly less in the homozygous short-ear mutant mice, whereas colony number and size were relatively greater in the heterozygous and wild-type congenic littermates. Average number of colonies±SD. *P-value <0.003, N=3.

      The short-ear mutant mouse shows a decrease in label-retaining cells (LRCs)

      To support the results from colony assays, we also performed label-retention analyses. It is known that LRCs are located in the interfollicular epidermis and bulge region of the hair follicle and are characterized as an important population of ESCs; thus, we decided to investigate whether the mutant mice had different numbers of LRCs than their congenic littermates. At 8 weeks after BrdU labeling, the dorsal skin was collected and analyzed for BrdU-label retention. As shown in Figure 3a–e, we observed that there were significantly fewer LRCs in BALB/c than C57BL/6 and in the mutant mice than the wild-type littermates. However, the general location of LRCs, and the size and number of the hair follicles, remain the same among all genotypes.
      Figure thumbnail gr3
      Figure 3Quantification of label-retaining cells (LRCs) in mouse epidermis. (ad) Immunohistographs for LRCs are shown. Mice received twice-daily (1000 and 1700hours) subcutaneous injections of 50μg per g body weight of BrdU at postnatal days D3, D4, and D5. After 8–10 weeks, dorsal skins were collected and fixed with 10% formalin for at least 24hours. Cells retaining the label are identified as LRCs. (e) The bar graph shows the quantification of labeled cells in the interfollicular epidermis and hair follicles combined. Labeled cells were counted in the same region of the dorsal skin for every 500 cells. Mean±SD. *P-value <0.05. N=3. Scale bars=30μm.

      Exogenously added human recombinant BMP5 protein increases the number and size of stem cell colonies in vitro

      To demonstrate further the functional significance of Bmp5 on ESC colony formation in vitro, we treated primary epidermal cells from C57BL/6 and C57BL/6.Cg-Bmp5se/J heterozygous and homozygous littermates cultured on irradiated 3T3 feeder layers with recombinant human BMP5 (rhBMP5). After 2 weeks of culture, the average number of colonies increased in all groups of mice at both low and high doses of rhBMP5 (Figure 4). As we expected, the homozygous mutant mice showed fewer colonies than the wild type, and the number of colonies increased in the presence of rhBMP5 to the colony number of the wild-type mouse at as low as 0.01μgml–1 concentration. Moreover, the average area of colonies increased with the treatment of rhBMP5 in a similar trend. These results suggest that Bmp5 might have a role in regulating the colony number of keratinocyte stem cells.
      Figure thumbnail gr4
      Figure 4Exogenous recombinant human bone morphogenetic protein 5 (BMP5) increases keratinocyte colony size and number. (a) To determine the functional significance of Bmp5 on keratinocyte colony formation, we treated wild-type and short-ear keratinocytes cultured on 3T3 feeder layers with different concentrations of recombinant human BMP5 (0.01 and 0.1μgml–1). The recombinant peptide was added to the medium that was changed every other day up to 2 weeks of culture. Note that exogenous BMP5 added to culture increased the number and size of colonies in keratinocytes from all three genotypes. (b) The photographs of the colony assays each correspond to the bar graph directly above. Average number of colonies±SD. *P-value <0.05. N=3.

      Bmp5 mutant mice have an increased susceptibility to skin tumors compared with congenic littermates

      Figure 5 demonstrates the increased tumor susceptibility of the short-ear mutant mice (together with the sequence-variant-bearing BALB/c mice) when compared with the heterozygous and congenic littermates. The segregation of tumor number and tumor latency between the mutant and wild-type controls was present from 8 weeks of promotion and continued throughout the experiment. Additionally, the tumor incidence was virtually 100% in all groups of mice. There was an increase in the number of tumors in the short-ear mutant mice; however, we observed that the tumor size was invariable across all genotypes. As shown in Figure 5a, although the difference in tumor number was statistically significant, the error bars appeared large when compared with the few numbers of tumors after 12-O-tetradecanoylphorbol-13-acetate (TPA) treatment, which is commonly the case for tumor-resistant C57BL/6 background. Figure 5b showed the average number per mouse versus weeks after TPA treatment. The mice began to develop tumors after 6–8 weeks.
      Figure thumbnail gr5
      Figure 5Bone morphogenetic protein 5 (Bmp5) regulates the number of tumors in mice in the two-stage skin carcinogenesis model. (a) The bar graphs represent the average number of skin tumor across all genotypes. Mouse skin was initiated with 200nmol of 7,12-dimethylbenz[a]anthracene (DMBA) in 200μl of spectral grade acetone when the mice were 7 weeks of age. After 1 week, there was twice-weekly promotion with 17nmol of 12-O-tetradecanoylphorbol-13-acetate (TPA) in 200μl of acetone. Skin tumors were counted weekly for 20 weeks. Note that Bmp5 status was associated with the number of skin tumors, with short-ear mutant and BALB/c mice developing more skin tumors than heterozygous and wild-type littermates, or C57BL/6J wild-type mice. Wild-type C57BL/6 and BALB/c mice were used, respectively, as low- and high- tumor controls. (b) The marked-line chart shows the appearance of tumors across all genotypes. Mean±SD. *P-value <0.05. N=5.

      Discussion

      We previously performed linkage analysis and identified the Ksc1 locus on chromosome 9 as one of the regions that regulate the small-size phenotype of clonogenic stem cells (
      • Popova N.V.
      • Suleimanian N.E.
      • Stepanova E.A.
      • et al.
      Independent inheritance of genes regulating two subpopulations of mouse clonogenic keratinocyte stem cells.
      ). This observation has led to a hypothesis that the genes that regulate the size and number of clonogenic stem cells might also be involved in skin carcinogenesis. Psl1 (promotion susceptibility locus 1) represents one of the loci previously shown to be associated with resistance or susceptibility to two-stage skin carcinogenesis, and it is overlapped with the Ksc1 locus (
      • Angel J.
      • Beltran L.
      • Minda K.
      • et al.
      Association of murine chromosome 9 locus (Psl1) with susceptibility to mouse skin tumor promotion by 12-O-tetradecanolphorbol-13-acetate.
      ). In addition, D9Mit271 has been linked to skin tumor susceptibility (
      • Mock B.
      • Lowry D.
      • Rehman I.
      • et al.
      Multigenic control of skin tumor susceptibility in SENCAR/Pt mice.
      ), and Skts6 (skin tumor susceptibility locus 6) has been recently mapped on the mouse chromosome 9 (49cM) (
      • Nagase H.
      • Bryson S.
      • Cordell H.
      • et al.
      Distinct genetic loci control development of benign and malignant skin tumors in mice.
      ). These loci are also closely related to the Ksc1 locus. Therefore, genes located in the Ksc1 locus regulating the number of small keratinocyte colonies may also regulate susceptibility or resistance to two-stage skin carcinogenesis.
      Although our use of the naturally occurring mutation in Bmp5 clearly demonstrates its role in skin stem cell regulation, there may be another gene in the locus.
      • Popova N.V.
      • Teti K.A.
      • Wu K.Q.
      • et al.
      Identification of two keratinocyte stem cell regulatory loci implicated in carcinogenesis.
      calculated that skin stem cell number was probably regulated by more than once gene acting together but possibly inherited independently (
      • Popova N.V.
      • Suleimanian N.E.
      • Stepanova E.A.
      • et al.
      Independent inheritance of genes regulating two subpopulations of mouse clonogenic keratinocyte stem cells.
      ). From our data, Bmp5 appears to act differently that other Bmps such as Bmp2, Bmp4, and Bmp6. Bmp2 and Bmp6 were described to inhibit keratinocyte growth and proliferation and induce terminal differentiation in vitro (
      • Gosselet F.P.
      • Magnaldo T.
      • Culerrier R.M.
      • et al.
      BMP2 and BMP6 control p57(Kip2) expression and cell growth arrest/terminal differentiation in normal primary human epidermal keratinocytes.
      ). It would probably be informative to determine whether other Bmps have similar effects of in vitro and in vivo aspects of skin stem cell number as reported in this study. Additionally, it would be interesting to see whether the BALB/c sequence variant could be rescued with Bmp5.
      BMP signaling has been shown to be necessary for maintaining quiescence of adult ESCs in the niche, as Smad1 is phosphorylated and Bmp6 levels are elevated in the hair bulge. Additionally, disruption of BMP signaling through bmpr1a ablation seems to increase expression of the transcription factors known to be important for bulge and progenitor cell maintenance and proliferation, Sox9, Lhx2, and Shh (
      • Kobielak K.
      • Stokes N.
      • de la Cruz J.
      • et al.
      Loss of a quiescent niche but not follicle stem cells in the absence of bone morphogenetic protein signaling.
      ). Thus, it is likely that BMP signaling acts through other pathways in regulating ESC quiescence and proliferation.
      In addition, the BMP signaling pathway was implicated in the initiation of colorectal neoplasia when mutations in BMPR1A were found in patients with juvenile polyposis (
      • Howe J.R.
      • Roth S.
      • Ringold J.C.
      • et al.
      Mutations in the SMAD4/DPC4 gene in juvenile polyposis.
      ). BMP6 has been shown to be absent in benign prostatic hyperplasia but strongly evident in primary tumors with secondary skeletal metastases, suggesting that BMP signaling may promote migration and invasion of prostate cancer cells (
      • Darby S.
      • Cross S.S.
      • Brown N.J.
      • et al.
      BMP-6 over-expression in prostate cancer is associated with increased Id-1 protein and a more invasive phenotype.
      ). The role of BMP signaling in invasiveness and metastasis of cancer cells is not, to our knowledge, previously unreported, as BMPs and SMADs have been shown to regulate epithelial–mesenchymal transition during embryonic development and morphogenesis (
      • Thiery J.P.
      • Sleeman J.P.
      Complex networks orchestrate epithelial-mesenchymal transitions.
      ).
      In our future plans, we would like to investigate keratinocyte cell proliferation and migration especially in the ESC population, which can be sorted by CD34 and α6 integrin. We would also like to explore further the nature of skin tumor development in the short-ear mutant including development of squamous cell carcinomas and metastasis, which will require a longer period of TPA treatment and tumor promotion. Additionally, a recently described Bmp5 dominant negative “cauliflower ear” mouse would be interesting with regard to skin stem cell regulation and tumor promotion (
      • Giggey J.
      • Bauschatz J.
      • Curtain M.
      • et al.
      Cauliflower ear-short ear 7 Jackson, a remutation to Bmp5.
      ).
      Furthermore, our laboratory has had a longstanding hypothesis that skin stem cells are the target cells in skin tumorigenesis (
      • Morris R.J.
      • Tacker K.C.
      • Fischer S.M.
      • et al.
      Quantitation of primary in vitro clonogenic keratinocytes from normal adult murine epidermis, following initiation, and during promotion of epidermal tumors.
      ;
      • Kangsamaksin T.
      • Park H.J.
      • Trempus C.S.
      • et al.
      A perspective on murine keratinocyte stem cells as targets of chemically induced skin cancer.
      ), and hence it is significant that a candidate keratinocyte stem cell regulatory gene should also be a skin tumor susceptibility gene. Nevertheless, the mechanisms through which this occurs remain to be identified.

      Materials and Methods

      Mice and epidermal cell isolation

      All mice were obtained from the Jackson Laboratory (Bar Harbor, MN): C57BL/6J, BALB/cJ, and C57BL/6.Cg-Bmp5se/J. Columbia University Institutional Animal Care and Use Committee approved all animal protocols. Epidermal cells from all groups from the dorsal skin of 50- to 60-day-old female mice were harvested following previously described protocols (
      • Morris R.J.
      • Fischer S.M.
      • Klein-Szanto A.J.
      • et al.
      Subpopulations of primary adult murine epidermal basal cells sedimented on density gradients.
      ). The hair follicles were in the telogen stage during this time that was confirmed by histological sections. BMP5 polyclonal antibody was purchased from BioVision (Mountain View, CA; cat. no. 5675-100) and was used in 1:50 dilutions. Biotinylated secondary antibody was used (Vector Laboratories, Burlingam, CA) in a 1:100 dilution. The slides were washed and incubated in ABC alkaline phosphatase/substrate solutions (Vector Laboratories).

      Cell culture

      Swiss 3T3 fibroblasts obtained at the 119th passage from American Type Culture Collection (Rockville, MD) were irradiated with 50Gy (5,000rads) and used as feeder cells. Epidermal cells harvested from each group were placed in 60-mm dishes (1,000 viable cells per dish) and were cultivated for 2 weeks in Supplemented Williams medium (
      • Morris R.J.
      • Fischer S.M.
      • Klein-Szanto A.J.
      • et al.
      Subpopulations of primary adult murine epidermal basal cells sedimented on density gradients.
      ). The medium was changed three times weekly, beginning the second day after seeding. After 2 weeks of cultivation, dishes were rinsed twice with Dulbecco's phosphate-buffered saline and were fixed in 10% formalin for at least 24hours. All dishes were stained with Rhodamine B. Epidermal cell colonies were stained with a higher density than feeder cells, and colonies >0.5mm were counted.

      Label-retention analysis

      To identify slowly cycling cells, mice received twice-daily (1000 and 1700hours) subcutaneous injections of 50μg per g body weight of BrdU (Sigma-Aldrich, St Louis, MO, cat. no. 858,811) at postnatal days D3, D4, and D5. After 8–10 weeks, dorsal skins were collected and fixed with 10% formalin for at least 24hours. Cells retaining the label are identified as LRCs. Paraffin-embedded skin sections were rehydrated in a series of ethanol solutions and antigen-retrieved in 2N HCl at 37°C for 30minutes and 0.1% trypsin for 15minutes. The slides were then incubated with the rat primary antibody (Accurate Chemical and Scientific, Westbury, NY; cat. no. OBT0030) in a 1:200 dilution, overnight at 4°C, followed by washing and subsequent incubation in biotinylated secondary antibody (Vector Laboratories) in a 1:100 dilution. The slides were washed and incubated in ABC alkaline phosphatase/substrate solutions (Vector Laboratories).

      Skin carcinogenesis experiments

      Skin tumors were initiated with 200nmol of DMBA (7,12-dimethylbenz[a]anthracene) in 200μl of spectral grade acetone when the mice were 7 weeks of age. After 1 week, there was twice-weekly promotion with 17nmol TPA in 200μl of acetone. Skin tumors were counted weekly for 20 weeks. Wild-type C57BL/6 and BALB/c mice were used, respectively, as low and high tumor controls. Additionally, the short-ear congenic and heterozygous mice were used as additional controls.

      Statistical analysis

      Student's t-test in Excel was performed for each experiment. Each experiment, except skin carcinogenesis and LRCs, was performed at least three times.

      ACKNOWLEDGMENTS

      This work was supported in part by the NIH grant AR052713. We thank Tonya Poorman for assistance in preparing the figures and submitting the manuscript.

      REFERENCES

        • Angel J.
        • Beltran L.
        • Minda K.
        • et al.
        Association of murine chromosome 9 locus (Psl1) with susceptibility to mouse skin tumor promotion by 12-O-tetradecanolphorbol-13-acetate.
        Mol Carcinogen. 1997; 20: 162-167
        • Argyris T.S.
        Promotion of epidermal carcinogenesis by repeated damage to mouse skin.
        Am J Ind Med. 1985; 8: 329-337
        • Barrandon Y.
        • Green H.
        Cell size as a determinant of the clone-forming ability of human keratinocytes.
        Proc Natl Acad Sci USA. 1987; 82: 5390-5394
        • Bickenbach J.R.
        Identification and behavior of label-retaining cells in oral mucosa and skin.
        J Dent Res. 1981; 60: 1611-1620
        • Cotsarelis G.
        • Sun T.-T.
        • Lavker R.M.
        Label-retaining cells reside in the bulge area of pilosebaceous unit: implications for follicular stem cells, hair cycle, and skin carcinogenesis.
        Cell. 1990; 61: 1329-1337
        • Darby S.
        • Cross S.S.
        • Brown N.J.
        • et al.
        BMP-6 over-expression in prostate cancer is associated with increased Id-1 protein and a more invasive phenotype.
        J Pathol. 2008; 214: 394-404
        • Giggey J.
        • Bauschatz J.
        • Curtain M.
        • et al.
        Cauliflower ear-short ear 7 Jackson, a remutation to Bmp5.
        (Craniofacial Resource, The Jackson Laboratory www.jax.org/cranio/index.html)
        • Gosselet F.P.
        • Magnaldo T.
        • Culerrier R.M.
        • et al.
        BMP2 and BMP6 control p57(Kip2) expression and cell growth arrest/terminal differentiation in normal primary human epidermal keratinocytes.
        Cell Signal. 2007; 19: 731-739
        • Hogan B.L.
        Bone morphogenetic proteins in development.
        Curr Opin Genet Dev. 1996; 4: 432-438
        • Howe J.R.
        • Roth S.
        • Ringold J.C.
        • et al.
        Mutations in the SMAD4/DPC4 gene in juvenile polyposis.
        Science. 1998; 280: 1086-1088
        • Kangsamaksin T.
        • Park H.J.
        • Trempus C.S.
        • et al.
        A perspective on murine keratinocyte stem cells as targets of chemically induced skin cancer.
        Mol Carcinog. 2007; 46: 579-584
        • Kingsley D.M.
        • Bland A.E.
        • Grubber J.M.
        • et al.
        The mouse short ear skeletal morphogenesis locus is associated with defects in a bone morphogenetic member of the TGF beta superfamily.
        Cell. 1992; 71: 399-410
        • King J.A.
        • Marker P.C.
        • Seung K.J.
        • et al.
        BMP5 and the molecular, skeletal, and soft-tissue alterations in short ear mice.
        Dev Biol. 1994; 166: 112-122
        • Kobielak K.
        • Stokes N.
        • de la Cruz J.
        • et al.
        Loss of a quiescent niche but not follicle stem cells in the absence of bone morphogenetic protein signaling.
        Proc Natl Acad Sci USA. 2006; 104: 10063-10068
        • Mackenzie I.C.
        • Bickenbach J.R.
        Localization of label-retaining cells in mouse epithelia.
        Cell Tissue Res. 1985; 242: 551-556
        • Mock B.
        • Lowry D.
        • Rehman I.
        • et al.
        Multigenic control of skin tumor susceptibility in SENCAR/Pt mice.
        Carcinogenesis. 1998; 19: 1109-1115
        • Morris R.J.
        • Fischer S.M.
        • Klein-Szanto A.J.
        • et al.
        Subpopulations of primary adult murine epidermal basal cells sedimented on density gradients.
        Cell Tissue Kinet. 1990; 23: 587-602
        • Morris R.J.
        • Fischer S.M.
        • Slaga T.J.
        Evidence that the centrally and peripherally located cells in the murine epidermal proliferative unit are two distinct cell populations.
        J Invest Dermatol. 1985; 34: 277-281
        • Morris R.J.
        • Tacker K.C.
        • Fischer S.M.
        • et al.
        Quantitation of primary in vitro clonogenic keratinocytes from normal adult murine epidermis, following initiation, and during promotion of epidermal tumors.
        Cancer Res. 1988; 48: 6285-6290
        • Nagase H.
        • Bryson S.
        • Cordell H.
        • et al.
        Distinct genetic loci control development of benign and malignant skin tumors in mice.
        Nat Genet. 1995; 10: 424-429
        • Plikus M.V.
        • Mayer J.A.
        • de la Cruz D.
        • et al.
        Cyclic dermal BMP signaling regulates stem cell activation during hair regeneration.
        Nature. 2008; 451: 340-344
        • Popova N.V.
        • Ovchinnikov I.
        • Kangsamaksin T.
        • et al.
        Bone morphogenetic protein 5 is a candidate gene in the keratinocyte stem cell locus-1. 2010 (in preparation)
        • Popova N.V.
        • Suleimanian N.E.
        • Stepanova E.A.
        • et al.
        Independent inheritance of genes regulating two subpopulations of mouse clonogenic keratinocyte stem cells.
        J Invest Dermatol Proc. 2004; 9: 253-260
        • Popova N.V.
        • Teti K.A.
        • Wu K.Q.
        • et al.
        Identification of two keratinocyte stem cell regulatory loci implicated in carcinogenesis.
        Carcinogenesis. 2003; 24: 417-425
        • Popova N.V.
        • Tryson K.A.
        • Wu K.Q.
        • et al.
        Evidence that keratinocyte colony number is genetically controlled.
        Exp Dermatol. 2002; 11: 503-508
        • Potten C.S.
        Stem cells in epidermis from the back of the mouse.
        in: Potten C.S. Stem Cells: Their Identification and Characterization. Churchill Livingstone, New York1983: 200-232
        • Potten C.S.
        Cell cycles in cell hierarchies.
        Int J Rad Biol. 1986; 49: 257-278
        • Thiery J.P.
        • Sleeman J.P.
        Complex networks orchestrate epithelial-mesenchymal transitions.
        Nat Rev Mol Cell Biol. 2006; 7: 131-142