Tissue Regeneration: Hair Follicle as a Model

Shigetoshi Sano, Junji Takeda,* Kunihiko Yoshikawa, and Satoshi ItamiDepartments of Dermatology and *Social and Environmental Medicine, Osaka University Graduate School of Medicine, Osaka, Japan

The Cre-loxP strategy has allowed us to generate themice whose keratinocytes are devoid of Stat3, whichplay a pivotal role in the signal transduction follow-ing the stimulation with various growth factors/cytokines, such as EGF, HGF, or IL-6. Althoughkeratinocyte-speci®c Stat3-disrupted mice were bornnormal with intact skin and the ®rst hair cycle, theyexhibited retardation of wound healing and absence

of the second hair cycle onward, leading to develop-ment of spontaneous skin ulcers and alopecia as theyaged. Thus, analyses of these mice reveal that Stat3in keratinocytes contributes to the regeneration ofepidermis and hair cycle process. Key words: haircycle/keratinocyte-speci®c gene targeting/Stat3/wound heal-ing. Journal of Investigative Dermatology SymposiumProceedings 6:43±48, 2001

The hair follicle is an epidermal derivative thatundergoes cycles of growth, regression, and rest,referred to as anagen, catagen, and telogen, respec-tively. Many growth factors are involved in theprogression of hair cycle through the interaction

between epithelial and mesenchymal cells (Hardy, 1992; Peus andPittelkow, 1996; Stenn et al, 1996). In response to growth factorsfrom dermal papilla cells, keratinocytes in the bulge regionproliferate and migrate into the dermis to generate new follicles(Cotsarelis et al, 1990). During the anagen phase, proteases such ascollagenase appear at the leading edge of hair follicles, which growdownwards (Yuspa et al, 1993). Thus, the anagen process isanalogous to wound healing in many ways (Cotsarelis, 1997).Wounding elicits a complex series of biologic responses to growthfactors (Bennett and Schultz, 1993; Martin, 1997). The healingprocess includes keratinocyte migration, proliferation, differentia-tion, elimination of damaged tissue, and production of extracellularmatrices. Keratinocytes proliferate in vivo and in vitro in response tomany growth factors that are found at the wound edge (Bennettand Schultz, 1993; Martin, 1997). Epidermal growth factor (EGF)promotes keratinocyte migration in vitro through integrin upregu-lation (Chen et al, 1993). Thus, keratinocytes at the wound edgecan pave the wound surface with provisional matrices such as®bronectin and collagen. Also, EGF and hepatocyte growth factor(HGF) stimulate keratinocytes to produce collagenase-1(Dunsmore et al, 1996; Pilcher et al, 1997) or matrix metallopro-teinase (MMP) 9 (McCawley et al, 1998). Keratinocytes at thewound edge migrate along the wound bed after detachment fromthe basement membrane through the degradation of extracellularmatrices by these proteases.

The in vivo role of a given growth factor has been elucidated inpart by analyzing knockout or transgenic mice of the speci®c genefor the growth factor or its speci®c receptor. Different growthfactors could be redundant in that they might share the samereceptors, and in that they might activate the same signalingmolecules. Therefore, it should be more appreciable to disrupt

genes of the signaling molecules at the downstream of such growthfactors to understand their roles.

STAT TRANSDUCE SIGNALS TO THE NUCLEUS WHENACTIVATED BY EXTRACELLULAR SIGNALING

MOLECULES

STAT (signal transducers and activators of transcription) belong to afamily of cytoplasmic proteins that are activated by a large numberof extracellular signaling molecules, including cytokines, growthfactors, and hormones (Darnell, 1997; Leonard and O'Shea, 1998).Upon stimulation with extracellular ligands, STAT are phos-phorylated in the tyrosine residues by kineses, resulting indimerization through reciprocal Src homology 2 (SH2)/phospho-tyrosine interactions. Dimerized STAT translocate to the nucleus,and then activate the speci®c promoters of target genes (Fig 1).STAT are phosphorylated either by receptor tyrosine kinases suchas receptors for EGF, platelet-derived growth factor (PDGF), andcolony-stimulating factor-1 (CSF-1), or by Janus kinases (JAK),which are noncovalently associated with cytokine receptors. Thus,cytokine receptors lacking an intrinsic kinase activity mediatesignals to STAT through the activation of JAK.

Seven STAT have been identi®ed in mammals: Stat1, Stat2,Stat3, Stat4, Stat5A, Stat5B, and Stat6. STAT are differentiallyactivated by various extracellular ligands, leading to transduction ofsignals to activate gene transcription. The biologic role of a STAThas been delineated by investigating the phenotype of knockoutmice of the speci®c STAT gene. Stat3 is expressed in ubiquitouscells and activated by ligands of wider spectrum than those for othermembers of the Stat family (Darnell, 1997; Leonard and O'Shea,1998). They include the interleukin (IL)-6 families that share thecommon receptor, gp130, such as IL-6, IL-11, leukemia inhibitoryfactor (LIF), ciliary neurotrophic factor (CNTF), oncostatin M, andcardiotropin 1 (Akira, 1997), and noncytokine ligands such as EGF,HGF, PDGF, granulocyte colony-stimulating factor (G-CSF), andleptin (Darnell, 1997; Boccaccio et al, 1998). Ablation of the Stat3gene in the germline leads to an early embryonic lethality (Takedaet al, 1997), suggesting that Stat3 plays a fundamental and criticalrole in the development of tissues and the survival of cells. Becauseof the embryonic lethality by systemic gene targeting, theconditional, namely, tissue-speci®c ablation of the Stat3 gene wasrequired to assess the biologic roles in the speci®c tissues (Takeda

1087-0024/01/$15.00 ´ Copyright # 2001 by The Society for Investigative Dermatology, Inc.

43

Manuscript received June 14, 2001; accepted for publication June 14,2001.

Reprint requests to: Dr. Shigetoshi Sano, Department of Dermatology,Osaka University Graduate School of Medicine, 2±2 Yamadaoka, Suita,Osaka, 565±0871, Japan. Email: sano@derma.med.osaka-u.ac.jp

et al, 1998, 1999; Chapman et al, 1999; Sano et al, 1999; Akira,2000; Hirano et al, 2000).

GENERATION OF KERATINOCYTE-SPECIFIC STAT3-DISRUPTED MICE

Stat3 is activated by many growth factors, some of which are alsoinvolved in growth/differentiation of keratinocytes. They includeEGF families (Barrandon and Green, 1987), HGF (Matsumotoet al, 1991), IL-6 (Grossman et al, 1989), and so on. To explore therole of Stat3 in keratinocyte activation, we generated keratinocyte-speci®c Stat3 knockout mice using the Cre-loxP strategy (Sanoet al, 1999). Mice carrying a keratin-5 promoter-driven (K5)-Cretransgene (Tarutani et al, 1997) together with a heterozygous Stat3-null allele (K5-Cre: Stat3+/±) (Takeda et al, 1997) were mated withStat3¯ox/¯ox mice (Takeda et al, 1998). Offspring carrying a ¯oxedStat3 allele and/or K5-Cre transgene (K5-Cre: Stat3¯ox/+, K5-Cre:Stat3¯ox/±, Stat3¯ox/+, Stat3¯ox/±) were used for further analyzes.Allele-speci®c polymerase chain reaction (PCR) was carried out todetermine the genotypes of the mice, and the expression of atruncated Stat3 was veri®ed by western blotting with anti-Stat3antibody (C-20; Santa Cruz Biotechnology, Santa Cruz, CA).Since Stat3+/± mice are devoid of any alterations (Sano et al,1999), they were used as controls in some experiments.

STAT3 IS NOT INVOLVED IN SKIN MORPHOGENESIS

Stat3-disrupted mice were born at the expected ratio according toMendelian's law, and they appeared normal (Fig 2a). No alterationwas found in the histology of their skin, including epidermis andappendages such as hair follicles and sebaceous glands (Fig 2b). This®nding indicated that Stat3 of keratinocytes is dispensable for the

morphogenesis of the skin. The ®rst anagen belongs to amorphogenic event, since it begins 14.5 d postcoitus, at which adowngrowth of the epidermis takes place (Kashiwagi et al, 1997).The hair follicles shown in Fig 2(b) were in mid-anagen stage ofthe ®rst hair cycle with no histologic difference between wild-typeand Stat3-disrupted mice. Furthermore, no difference was foundbetween the two types of mice in the expression of differentiationor proliferation-associated markers of keratinocytes, including K5,K1, K10, involucrin, ®laggrin, Ki67, and K6 (data not shown).

STAT3 IS ESSENTIAL FOR SECOND ANAGEN ONSET

Follicular morphogenesis in the ®rst 16 d of postnatal life isfollowed by a short catagen/telogen period. The second anagenbegins around postnatal day (PND) 21±22, and lasts 9 d (Fig 3a,left); however, Stat3-disrupted mice were arrested in the ®rsttelogen phase (Fig 3a, right). Interestingly, histologic scrutiny ofthe mutant mice at this stage revealed hyperkeratosis, acanthosis,dermal ®brosis, and grains of hair germ-like cells in the dermis(arrowheads in Fig 3a, right), although these changes were somewhattransient. To explore the hair cycle progression ex vivo, wetransplanted the full thickness skin from new born Stat3-disruptedmice onto the back of 9-wk-old-male BALB/C nu/nm mice.About 2 wk after the transplantation, hair emerged from the skingraft derived from wild-type mice. In contrast, no hair wasrecognized in the graft of the mutant mice throughout the durationobserved over 4 mo after the transplantation (Fig 3b). In fact,histologic examination revealed that hair follicles stayed in telogenin the graft of mutant mice, whereas mature anagen follicles were

Figure 1. Activation of the STAT signaling pathway. Binding of anextracellular ligand to the speci®c receptor on the plasma membraneleads to tyrosine phosphorylation of the receptor or a closely associatedJAK. STAT migrates from the cytoplasm to the membrane and targets tothe phosphotyrosine motif of the receptor or JAK through the Srchomology 2 (SH2) domain, and undergoes tyrosine phosphorylation bytyrosine kinase of the receptor or JAK (STAT activation). Finally,activated STAT leaves the receptor complex to form homo/heterodimers through reciprocal SH2±phosphotyrosine interactions, resulting inthe nuclear translocation and hence the transcriptional regulation of thetarget genes. L, ligands; R, receptor; J, JAK; P, phosphorylated tyrosineresidue; N, nucleus.

Figure 2. Gross and histologic appearance of newborn Stat3-disrupted mice. (a) One-week-old littermates; wild-type (left two) andStat3-disrupted mice (right two). The tips of tales were cut forgenotyping. There was no macroscopic change in Stat3-disrupted mice,including body size, activities, and skin texture. (b) Histologic appearanceof the back skin of 1-wk-old wild-type (+/+, left) and Stat3-disruptedmice (±/±, right). There was no histologic alteration in the skin,including epidermis and appendages. Follicles shown were of the ®rstanagen phase without any difference between the two types of mice.Hematoxylin and eosin, scale bar: 240 mm.

44 SANO ET AL JID SYMPOSIUM PROCEEDINGS

recognized in the graft of control mice (Fig 3c). These resultsclearly show that the skin phenotype of Stat3-disrupted mice wasreproduced in the skin graft, in which the anagen progression of thehair follicle was compromised.

STAT3 IS INVOLVED IN WOUND HEALING

Since anagen progression is analogous to wound healing in manyways (Cotsarelis, 1997), we examined the healing process in Stat3-disrupted mice. Wounds in the back skin using a biopsy punchhealed in 10 d in control mice, but Stat3-disrupted mice demon-strated a marked retardation of wound healing, taking 2-fold longerto heal than controls (Sano et al, 1999). As shown in Fig 4(a), thehistologic difference was obvious: the wounds of control mice were

completely covered by the regenerated epidermis overlying scarry®brotic dermis on day 12 after wounding, but at the same timepoint, Stat3-disrupted mice still had ulcers with leading tips of theepidermis being separate (Fig 4a, arrows). It should be noted thatthe wound-associated dermal reaction, such as scarry ®brosis andneovasculization, were as well recognized in the mutant mice as incontrols, suggesting that the retardation of wound healing was notdue to changes in dermal components. Although the woundhealing process requires keratinocyte proliferation, Stat3-disruptedkeratinocytes proliferate at the wound edge as well as normal mice.Immunohistochemical study with antibodies against Ki67, which isa marker associated with proliferative cells, was expressed inepidermal basal cells of Stat3-disrupted as well as control mice(Fig 4b, left panels). This result was consistent with our previousreport that the in vitro proliferative activity of Stat3-disruptedkeratinocytes was not impaired (Sano et al, 1999). Keratinocytes atthe wound edge express K6, which is constitutively expressed infollicular keratinocytes of the outer root sheath but is also expressedin the interfollicular epidermis in response to stressful stimuli. It wasreported that K6a knockout mice showed a delay in reepitheliza-tion from the hair follicles, suggesting that K6a was involved in theactivation of follicular keratinocytes after wounding (Wojcik et al,2000). Epidermis at the wound edge strongly expressed K6 inStat3-disrupted mice as well as control mice (Fig 4b, right panels),indicating that Stat3 is not required for K6 expression in response towounding stress.

STAT3 ACTIVATION IS PREREQUISITE FORKERATINOCYTE MIGRATION

Although the proliferative activity was intact in Stat3-disruptedkeratinocytes, their in vitro migration was severely compromised.Keratinocytes derived from newborn to 5-d-old mice werecultured in MCDB153 medium in dishes that were precoatedwith collagen type I. They were treated with mitomycin C for 2 hbefore the assay to prevent subsequent proliferation. Cell migrationto the cell-free area during the next 24±48 h was evaluated in theabsence or presence of 10 ng per ml of EGF, transforming growthfactor-a, heparin binding EGF-like growth factor (HB-EGF), orHGF. The former three growth factors bind to a shared receptor,EGFR (Coffey et al, 1987; Higashiyama et al, 1991; Massague andPandiella, 1993), and HGF binds to its speci®c receptor, Met(Naldini et al, 1991). These factors used are known to activate Stat3(Darnell, 1997; Boccaccio et al, 1998; Leonard and O'Shea, 1998;Sano et al, 1999) and promote keratinocyte migration (Barrandonand Green, 1987; Matsumoto et al, 1991) (Fig 4c, top panels). Bycontrast, Stat3-disrupted keratinocytes did not migrate in responseto inclusion with any of them (Fig 4c, bottom panels). Takencollectively, it is clear that the retardation of wound healing inStat3-disrupted mice was due to their impaired migration ofkeratinocytes, while their proliferative potential was not com-promised (Fig 4b, left panels, Sano et al, 1999). As the expression ofMMP and integrins, which are known to be involved inkeratinocyte migration, was not impaired in Stat3-disrupted mice(Sano et al, 1999), the downstream molecule(s) of Stat3 responsiblefor keratinocyte migration is not clear at present.

AGE-RELATED SKIN PHENOTYPE IN STAT3-DISRUPTED MICE

Although Stat3-disrupted mice were born normal, they developedskin alterations at 1 mo of age, around which their hair follicleswere arrested in telogen, not entering the second anagen (Fig 3a,right). As they aged, their skin phenotype appeared more severe.They developed diffuse alopecia, eczema-like dermatosis with awrinkly, rough, and scaly surface. Also, there were spontaneousulcers and scratching-induced ulcers (Fig 5a). Histologic examin-ation of the aged Stat3±/± mice revealed that hyperkeratosis, amarked acanthosis, ulcers (Fig 5b, arrow), hypertrophy andhyperplasia of sebaceous glands, ®brosis and in¯ammatory cells inthe dermis (Fig 5b, c). Also, unsynchronized, rather misaligned

Figure 3. Impaired second anagen entry and hair growth inStat3-disrupted mice. (a) Histologic appearance of the back skin ofwild-type (+/+, top) and Stat3-disrupted mice (±/±, bottom) at PND 28.Elongating hair follicles reaching to the adipose tissue were shown inwild-type mice, but the hair follicles of Stat3-disrupted mice stayed intelogen. The mutant mice developed hyperkeratosis, acanthosis, anddermal ®brosis. Arrowheads indicate clusters of hair germ-like cells.Hematoxylin and eosin, scale bar: 200 mm. (b) Gross appearance of theskin graft 2 mo after the implantation of full-thickness skin of newbornwild-type (left) and Stat3-disrupted mice (right). No hair was recognizedin Stat3-disrupted skin. (c) Histologic examination revealed no anagenfollicles in the graft from Stat3-disrupted skin (±/±, right), whereasanagen follicles were in the control graft (+/+, left). Hematoxylin andeosin, scale bar: 320 mm.

VOL. 6, NO. 1 NOVEMBER 2001 REQUIREMENT OF STAT3 FOR SKIN REGENERATION 45

anagen follicles were noted, unlike well-organized follicles that areseen in the second anagen. Therefore, this ®nding suggests thatthere is an alternate, Stat3-independent signaling pathway to formanagen follicles (Sano et al 2000). Epidermis in the mutant mice wascomposed of 6±8 layers of keratinocytes (Fig 5c, bottom), whereasthere were 2±3 layers in controls (Fig 5c, top). In addition,individual cell size appeared to be increased with large nucleicompared with controls (Fig 5c). Mitotic ®gures were frequentlyseen in the basal layer of Stat3-disrupted mice, which is indicativeof a hyperproliferative state in their keratinocytes (Fig 5c, bottom).So far, it is not known whether these changes are primary orsecondary to the absence of Stat3 activation in keratinocytes. It isplausible that Stat3-disrupted mice were subjected to skin ulcersfrom minimal insults, which might be so subtle that they inducedno pathologic change in wild-type mice. The chronic, repeatedulceration in the mutant mice might result in the persistenthyperproliferation of epidermal keratinocytes, which presumablycompensated for their impaired migration. Taken together with the®nding that Stat3-disrupted keratinocytes at the wound edge werepositive for Ki67 comparable with controls (Fig 4b, left panels), it is

likely that intracellular signals towards proliferation differ fromthose towards migration in keratinocytes. It is well known thatmigration does not depend on cell proliferation. Some hours afterthe onset of migration, keratinocytes just back from the woundmargin undergo a proliferative burst that, although not strictlyrequired for migration, provides a pool of extra cells to replacethese lost during the injury (Garlick and Taichman, 1994).Therefore, at least two types of cells may exist at the wound site:proliferating and migrating keratinocytes. Further investigation willbe needed to explain the underlying mechanism for the acanthosisin aged Stat3±/± mice.

STAT3-DEPENDENT BULGE STEM AND/ORTRANSIENT AMPLIFYING CELL ACTIVATION FORANAGEN PROGRESSION AND WOUND HEALING

The anagen phase, a process of follicle downgrowth, is initiated bymesenchymal signals, which are intrinsically derived from dermalpapilla (Fig 6a) (Hardy, 1992; Peus and Pittelkow, 1996; Stennet al, 1996). During late telogen, a slow-cycling stem cell of the

Figure 4. Retarded wound healing and compromised migration of Stat3-disrupted keratinocytes. (a) On day 12 after wounding, controlmice healed completely (+/+, left), but there was still an ulcer in Stat3-disrupted mice (±/±, right). Arrows indicate two leading tips of the epidermisbeneath the crust. Note that no difference was shown in the dermal responses, including scarry ®brosis and neovascularization, between the two typesof mice. Hematoxylin and eosin, scale bar: 500 mm. (b) At the wound edge 4 d after wounding, Ki67, a nuclear antigen speci®c for proliferating cells,was expressed in the epidermal basal layer of Stat3-disrupted as well as control mice (arrowheads in left two panels). K6, which is constitutively expressedin follicular keratinocytes and an inducible keratin of interfollicular epidermis in response to stressful stimuli, was highly expressed in the wound edge ineither type of mice (right two panels). (c) In vitro wounds were introduced in cultured keratinocytes at con¯uency derived from wild-type (top panels) andStat3-disrupted mice (bottom panels), and the keratinocytes were cultured for another 48 h in the presence of indicated growth factors. The controlkeratinocytes migrated by stimulation with 10 ng per ml of EGF, transforming growth factor-a, HB-EGF, or HGF, whereas Stat3-disruptedkeratinocytes (±/±) did not migrate by any of them. Scale bar: 250 mm.

46 SANO ET AL JID SYMPOSIUM PROCEEDINGS

bulge area is activated and divides in response to the mesenchymalsignals, which are paracrine growth factors, and generate twodaughter cells. One of the daughter cells returns to its normallyquiescent, slow-cycling state in the niche of the bulge area.Another daughter cell itself or transient amplifying (TA) cellsmigrate downwards and proliferate to form matrix cells (Cotsareliset al, 1990). Our ®nding that Stat3 activation is a prerequisite forthe second anagen progression (Fig 3a) and plays a critical role inkeratinocyte migration (Fig 4c) suggests that the stem cells and/ortheir progeny (TA cells) migrate in a Stat3-dependent fashionduring the early anagen. The growth factors involved in this processshould be activators of Stat3 (Fig 6b). The most likely candidate isHGF, because HGF was found in the dermal papilla ®broblasts, andits receptor Met in the neighboring keratinocytes (Lindner et al,2000). A highly critical fact is that both HGF and Met expressionpeaked together at the initial stage of the second anagen (Lindneret al, 2000). Like in the anagen progression, wound-associatedmesenchymal signals are required for the healing process. Theyinclude growth factors and cytokines secreted from in¯ammatory

leukocytes, endothelial cells, and ®broblasts at the wound site(Bennett and Schultz, 1993; Martin, 1997), and they stimulate thestem/TA cells in a paracrine fashion (Fig 6a). Thus, our hypothesisis that the anagen onset and wound healing require Stat3 activationin bulge stem/TA cells to perform vertical and horizontal migration,respectively (Fig 6b). Our hypothesis is supported in part by therecent report showing that follicular bulge stem cells were poten-tially bipotent as they gave rise to not only the hair follicle, but alsothe interfollicular epidermis upon wounding (Taylor et al, 2000).We believe that further analyzes of Stat3-disrupted mice willcontribute to elucidating the underlying mechanism of certaindermatoses such as chronic ulcer and alopecia.

We thank Drs K. Takeda and S. Akira for providing the mice carrying ¯oxed Stat3

and Stat3 null alleles. We thank Mrs S. Okamoto for excellent technique for

immunohistochemical studies and Dr. K. Nishida for helping with the illustrations.

This work was supported by grants of the Ministry of Education, Science and

Culture of Japan.

Figure 5. Severe skin phenotype of aged Stat3-disrupted mice. (a) Gross appearance of Stat3-disrupted mouse at 6-mo-old. It underwent diffusealopecia, ulcers, and severe eczema-like change. Close-up view (right) showed diffusely distributed, yellowish scale-crusts, and small ulcers over almostthe entire body. (b) Histologic appearance of the back skin from 6-mo-old Stat3-disrupted (bottom) and control littermates (top). The mutant micedemonstrated histopathologic changes not only in the epidermis but also in the dermis. There were ulcers (arrows) and unsynchronized, ratherdisoriented anagen follicles were found. Hematoxylin and eosin, scale bar: 250 mm. (c) Higher magni®cation. Epidermis of Stat3-disrupted mice, unlikecontrol mice, exhibit hyperkeratosis and a marked acanthosis composed of 6±8 squamous cell layers with a considerable number of mitoses in the basalcells. Hematoxylin and eosin, scale bar: 50 mm.

VOL. 6, NO. 1 NOVEMBER 2001 REQUIREMENT OF STAT3 FOR SKIN REGENERATION 47

REFERENCES

Akira S: IL-6-regulated transcription factors. Int J Biochem Cell Biol 29:1401±1418,1997

Akira S: Roles of STAT3 de®ned by tissue-speci®c gene targeting. Oncogene19:2607±2611, 2000

Barrandon Y, Green H: Cell migration is essential for sustained growth ofkeratinocyte colonies: the roles of transforming growth factor-alpha andepidermal growth factor. Cell 50:1131±1137, 1987

Bennett NT, Schultz GS: Growth factors and wound healing: biochemical propertiesof growth factors and their receptors. Am J Surg 165:728±737, 1993

Boccaccio C, Ando M, Tamagnone L, Bardelli A, Michieli P, Battistini C, ComoglioPM: Induction of epithelial tubules by growth factor HGF depends on theSTAT pathway. Nature 391:285±288, 1998

Chapman RS, Lourenco PC, Tonner E, et al: Suppression of epithelial apoptosis anddelayed mammary gland involution in mice with a conditional knockout ofStat3. Genes Dev 13:2604±2616, 1999

Chen JD, Kim JP, Zhang K, Sarret Y, Wynn KC, Kramer RH, Woodley DT:Epidermal growth factor (EGF) promotes human keratinocyte locomotion oncollagen by increasing the alpha 2 integrin subunit. Exp Cell Res 209:216±223,1993

Coffey RJ Jr, Derynck R, Wilcox JN, Bringman TS, Goustin AS, Moses HL,Pittelkow MR: Production and auto-induction of transforming growth factor-alpha in human keratinocytes. Nature 328:817±820, 1987

Cotsarelis G: The hair follicle: dying for attention. Am J Pathol 151:1505±1509, 1997Cotsarelis G, Sun TT, Lavker RM: Label-retaining cells reside in the bulge area of

pilosebaceous unit: implications for follicular stem cells, hair cycle, and skincarcinogenesis. Cell 61:1329±1337, 1990

Darnell JE Jr: STATs and gene regulation. Science 277:1630±1635, 1997Dunsmore SE, Rubin JS, Kovacs SO, Chedid M, Parks WC, Welgus HG:

Mechanisms of hepatocyte growth factor stimulation of keratinocytemetalloproteinase production. J Biol Chem 271:24576±24582, 1996

Garlick JA, Taichman LB: Fate of human keratinocytes during reepithelialization inan organotypic culture model. Laboratory Invest 70:916±924, 1994

Grossman RM, Krueger J, Yourish D, et al: Interleukin 6 is expressed in high levelsin psoriatic skin and stimulates proliferation of cultured human keratinocytes.Proc Natl Acad Sci USA 86:6367±6371, 1989

Hardy MH: The secret life of the hair follicle. Trends Genet 8:55±61, 1992Higashiyama S, Abraham JA, Miller J, Fiddes JC, Klagsbrun M: A heparin-binding

growth factor secreted by macrophage-like cells that is related to EGF. Science251:936±939, 1991

Hirano T, Ishihara K, Hibi M: Roles of STAT3 in mediating the cell growth,differentiation and survival signals relayed through the IL-6 family of cytokinereceptors. Oncogene 19:2548±2556, 2000

Kashiwagi M, Kuroki T, Huh N: Speci®c inhibition of hair follicle formation byepidermal growth factor in an organ culture of developing mouse skin. Dev Biol189:22±32, 1997

Leonard WJ, O'Shea JJ: Jaks and STATs: biological implications. Annu Rev Immunol16:293±322, 1998

Lindner G, Menrad A, Gherardi E, et al: Involvement of hepatocyte growth factor/scatter factor and met receptor signaling in hair follicle morphogenesis andcycling. FASEB J 14:319±332, 2000

Martin P: Wound healing ± aiming for perfect skin regeneration. Science 276:75±81,1997

Massague J, Pandiella A: Membrane-anchored growth factors. Annu Rev Biochem62:515±541, 1993

Matsumoto K, Hashimoto K, Yoshikawa K, Nakamura T: Marked stimulation ofgrowth and motility of human keratinocytes by hepatocyte growth factor. ExpCell Res 196:114±120, 1991

McCawley LJ, O'Brien P, Hudson LG: Epidermal growth factor (EGF) - and scatterfactor/hepatocyte growth factor (SF/HGF) - mediated keratinocyte migrationis coincident with induction of matrix metalloproteinase (MMP)-9. J CellPhysiol 176:255±265, 1998

Naldini L, Vigna E, Narsimhan RP, Gaudino G, Zarnegar R, Michalopoulos GK,Comoglio PM: Hepatocyte growth factor (HGF) stimulates the tyrosine kinaseactivity of the receptor encoded by the proto-oncogene c-MET. Oncogene6:501±504, 1991

Peus D, Pittelkow MR: Growth factors in hair organ development and the hairgrowth cycle. Dermatol Clin 14:559±572, 1996

Pilcher BK, Dumin JA, Sudbeck BD, Krane SM, Welgus HG, Parks WC: Theactivity of collagenase-1 is required for keratinocyte migration on a type Icollagen matrix. J Cell Biol 137:1445±1457, 1997

Sano S, Itami S, Takeda K, et al: Keratinocyte-speci®c ablation of Stat3 exhibitsimpaired skin remodeling, but does not affect skin morphogenesis. Embo J18:4657±4668, 1999

Sano S, Kira M, Takagi S, Yoshikawa K, Takeda J, Itami S: Two distinct signalingpathways in hair cycle induction: Stat3-dependent and -independent pathways.Proc Natl Acad Sci USA 97:13824±13829, 2000

Stenn KS, Combates NJ, Eilertsen KJ, Gordon JS, Pardinas JR, Parimoo S, ProutySM: Hair follicle growth controls. Dermatol Clin 14:543±558, 1996

Takeda K, Clausen BE, Kaisho T, Tsujimura T, Terada N, Forster I, Akira S:Enhanced Th1 activity and development of chronic enterocolitis in micedevoid of Stat3 in macrophages and neutrophils. Immunity 10:39±49, 1999

Takeda K, Kaisho T, Yoshida N, Takeda J, Kishimoto T, Akira S: Stat3 activation isresponsible for IL-6-dependent T cell proliferation through preventingapoptosis. generation and characterization of T cell- speci®c Stat3-de®cientmice. J Immunol 161:4652±4660, 1998

Takeda K, Noguchi K, Shi W, et al: Targeted disruption of the mouse Stat3 geneleads to early embryonic lethality. Proc Natl Acad Sci USA 94:3801±3804, 1997

Tarutani M, Itami S, Okabe M, et al: Tissue-speci®c knockout of the mouse Pig-agene reveals important roles for GPI-anchored proteins in skin development.Proc Natl Acad Sci USA 94:7400±7405, 1997

Taylor G, Lehrer MS, Jensen PJ, Sun TT, Levker RM: Involvement of follicularstem cells in forming not only the follicles but also the epidermis. Cell102:451±461, 2000

Wojcik SM, Bundman DS, Roop DR: Delayed wound healing in keratin 6aknockout mice. Mol Cell Biol 20:5248±5255, 2000

Yuspa SH, Wang Q, Weinberg WC, Goodman L, Ledbetter S, Dooley T, Lichti U:Regulation of hair follicle development: an in vitro model for hair follicleinvasion of dermis and associated connective tissue remodeling. J InvestDermatol 101:27S±32S, 1993

Figure 6. Stat3-dependent keratinocyte migration towards theanagen progression and wound healing. (a) Either unknown stimuliat the late telogen or wound stimuli elicit the mesenchymal signals fromdermal papilla cells or wound-reactive cells in the dermis, respectively.These signals lead to the bulge activation. B, bulge; DP, dermal papilla;MS, mesenchymal signal. (b) Subsequently, the bulge stem cells and/ortheir progeny (TA cells) migrate in a Stat3-dependent manner: verticalmigration is for anagen progression, and horizontal migration is forwound healing.

48 SANO ET AL JID SYMPOSIUM PROCEEDINGS