Onset of Re-epithelialization After Skin Injury Correlates with a Reorganization of Keratin Filaments in Wound Edge Keratin ytes: Defining a Potential Role for Keratin 16 Rudolph

Injury to stratified epithelia causes a strong induction of keratins 6 (K6) and 16 (K16) in post-mitotic keratinocytes located at the wound edge. We show that induction of K6 and K16 occurs within 6 h after injury to human epidermis. Their subsequent accumulation in keratinocytes correlates with the profound reorganization of keratin filaments from a pan-cytoplasmic distribution to one in which filaments are aggregated in a juxtanuclear location, opposite to the direction of cell migration. This filament reorganization coincides with additional cytoarchitectural changes and the onset of re-epithelialization after 18 h post-injury. By following the assembly of K6 and K16 in vitro and in cultured cells, we find that relative to K5 and K14, a well-characterized keratin pair that is constitutively expressed in epidermis, K6 and K16 polymerize into short 10-nm filaments that accumulate near the nucleus, a property arising from K16. Forced expression of human K16 in skin keratinocytes of transgenic mice causes a retraction of keratin filaments from the cell periphery, often in a polarized fashion. These results imply that K16 may not have a primary structural function akin to epidermal keratins. Rather, they suggest that in the context of epidermal wound healing, the function of K16 could be to promote a reorganization of the cytoplasmic array of keratin filaments, an event that precedes the onset of keratinocyte migration into the wound site. ~ ] ~ typical stratified squamous epithelium, the epidermis shows a polarity in its functional organization and morphology. This polarity reflects the maintenance of an optimal balance between proliferation and differentiation, such that this self-renewing tissue maintains an architecture that is ideal for a barrier function. Progenitor cells reside in the basal layer, where keratinocytes are mitotically active, have a low columnar shape, appear relatively undifferentiated, and express specific markers such as the type II keratin K5 and type I keratins K14 (Nelson and Sun, 1983). Upon commitment to differentiation, basal keratinocytes exit the cell cycle and migrate upward to the suprabasal compartment (Fuchs, 1993). This commitment triggers a specific program of gene expression during which suprabasal keratinocytes steadily progress towards a cytoarchitecture in which they are completely flattened, have lost their organelles and nucleus, and have most of their protein content covalently cross-linked via the action of transglutaminases (Rice and Green, 1979). It is believed that commitment to terminal differentiation, which is accompanied by a switch in keratin gene expression from Address correspondence to Dr. Pierre A. Coulombe, Department of Biological Chemistry, Johns Hopkins University School of Medicine, 725 N. Wolfe St., Baltimore, MD 21205. Tel.: (410) 614-0510. Fax: (410) 955-5759. K5/K14 to K1 (type II) and K10 (type I) (e.g., Fuchs and Green, 1980), is irreversible (Hotchin et al., 1993). Early studies of the re-epithelialization of full-thickness epidermal wounds in human, mouse, rabbit, dog, and pig skin established that keratinocyte migration and proliferation are initiated at 16-24 h following injury (Winter, 1962; Viziam et al., 1964; Odland and Ross, 1968; Croft and Tarin, 1970; Krawczyk, 1971; Winstanley, 1975; Ortonne et al., 1981; Mansbridge and Knapp, 1987). Cell migration occurs in the form of a stratified epithelial sheet, and in its early phase is unaffected by inhibitors of cell mitosis (e.g., Matoltsy, 1955; see Bereiter-Hahn, 1984). However, a transient burst of enhanced mitotic activity is necessary to supply the cells required to sustain re-epithelialization, and is carried out by a distinct subpopulation of epidermal keratinocytes located behind the migrating epithelium (see Bereiter-Hahn, 1984; Clark, 1993). Ultrastructural studies revealed major cytoarchitectural changes in keratinocytes located at the wound edge, coinciding with the onset of a migratory behavior. These changes include cellular hypertrophy, a fragmentation of keratin filaments and their retraction from the cytoplasmic periphery, and alterations in the number and structure of desmosomes, correlating with the widening of the intercellular space between keratinocytes (e.g., Odland and Ross, 1968; Gabbiani et al., © The Rockefeller University Press, 0021-9525/96/02/381/17 $2.00 The Journal of Cell Biology, Volume 132, Number 3, February 1996 381-397 381 on July 0, 2017 jcb.rress.org D ow nladed fom 1978; see Bereiter-Hahn, 1984, and Clark, 1993). Taken together, these observations suggest that during the re-epithelialization of skin wounds, epidermal keratinocytes located at the wound edge deviate from their program of terminal differentiation. The cellular and molecular mechanisms responsible for the cytoarchitectural changes, keratinocyte migration, and hyperproliferation after skin injury remain largely unknown. Keratins, the epithelial-specific intermediate filament proteins, are the major differentiation-specific proteins in epidermis. The >30 known keratin genes and their encoded proteins (40-70 kD MW) are classified as type I (acidic, numbered K9-K20) and type II sequences (basic, numbered K1-K8) (Fuchs and Weber, 1994). Assembly of keratin filaments begins with the formation of a type I-type II coiled-coil heterodimer (see Coulombe, 1993), and this strict requirement calls for the coexpression of at least one type I and one type II keratin gene in epithelial cells. Many keratin genes, such as K5/K14 and K1/K10, are regulated in a differentiation-specific and pairwise fashion (see O'Guin et al., 1990; Fuchs, 1993). In normal epidermis, the function of the keratin filament network is to provide the physical strength that is necessary to maintain cellular integrity in response to a normal load of mechanical stress. In transgenic mice as well as human subjects, mutations or the complete absence of a keratin protein in the epidermis results in defective 10-nm filament structure, mechanical stress-induced cytolysis and blistering (see Fuchs and Coulombe, 1992; Chan et al., 1994; Rugg et al., 1994; Lloyd et al., 1995, and references therein). The location of epidermal tissue cleavage, and thus the cell layer(s) affected by trauma, are determined by the pattern of expression of the mutation-bearing keratin. Mutations affecting specific keratin genes underlie several inheritable skin blistering disorders such as epidermolysis bullosa simplex, epidermolytic hyperkeratosis, and palmoplantar keratoderma (see Coulombe, 1993; Fuchs et al., 1994; McLean and Lane, 1995). Injury to the skin significantly alters keratin gene expression in keratinocytes located near the wound edge. Under such conditions, an induction of K6 (type II), K16 and K17 (type I) occurs in the differentiating layers of epidermis (e.g., Weiss et al., 1984; O'Guin et al., 1990). K6 and K16 proteins have been biochemicaUy detected at 8-10 h in the wounded tissue (Tyner and Fuchs, 1986; Mansbridge and Knapp, 1987; de Mare et al., 1990). Subsequent to this induction, the differentiation-specific keratins K1 and K10 appear down-regulated (e.g., Mansbridge and Knapp, 1987; Coulombe et al., 1991). In addition to wound healing in skin, K6 and K16 are also expressed in stratified epithelia undergoing chronic hyperproliferation or abnormal differentiation, including cancer (Moll et al., 1983; Weiss et al., 1984; Stoler et al., 1988; Schermer et al., 1989). In such hyperproliferative disorders, as in regenerating stratified epithelia, abundant expression of K6 and K16 is often associated with alterations in keratinocyte differentiation. Consistent with this, overexpression of a wild-type human K16 gene in transgenic mice causes the reorganization of the IF network in keratinocytes of the hair follicle outer root sheath and epidermis, leading to aberrant keratinization and hyperproliferation in these tissues (Takahashi et al., 1994). Yet, K6 and K16 are constitutively expressed in a variety of epithelial tissues under normal conditions (e.g., Moll et al., 1982, 1983), without apparent consequences for their differentiation. Thus, the role(s) that K6 and K16 may play during wound healing, as well as the consequences of their expression in chronic hyperproliferative diseases affecting stratified epithelia, remain unclear. We investigated the consequences of K6 and K16 induction in keratinocytes located at the wound edge after injury to epidermis. We characterized the time-course of K6 and K16 induction in wounded human skin, and correlated it with alterations in keratinocyte cytoarchitecture. We found that induction of K6 and K16 proteins occurs within 6 h at the wound edge, and that their subsequent accumulation correlates with a polarized reorganization of keratin filaments in suprabasal keratinocytes, followed by alterations in their shape and cell--cell adhesion. These changes coincide with the onset of re-epithelialization, which begins after 18 h post-injury. To determine whether K6 and K16 could play a direct role in these phenomena, we investigated their assembly properties in vitro as well as in vivo. We show that unlike K5, K6, and K14, human K16 features unusual assembly properties, in that it promotes the formation of short 10-nm filaments that are localized preferentially near the nucleus in transfected cells as well as in skin keratinocytes of transgenic mice. Our data suggest that the alterations in keratin expression, and in particular the induction of K16, could play a role in promoting the reorganization of keratin filaments that occurs in spinous keratinocytes before the onset of re-epithelialization after injury to epidermis. Materials and Methods Human Skin Wound Healing Studies Studies involving human subjects were reviewed and approved by the Joint Committee on Clinical Investigation at the Johns Hopkins University School of Medicine. All experiments were performed under sterile conditions. Small incisions (6 mm long, 2 mm deep) were made on the inside arm of healthy volunteers (26-35 years of age). This site was selected because of the unusually low density of hair follicles. For sampling, 4-mm punch biopsies (Acu-Punch; Acuderm Inc., Ft. Lauderdale, FL) were performed under local anesthesia at either 6, 12, 18, or 30 h after wounding. Each biopsy was divided into two pieces across the wound: one-third of the sample was processed for routine electron microscopy (Coulombe et al., 1989) while the remaining two-thirds was embedded in TBS tissue freezing medium (Triangle Biomedical Sciences, Durham, NC), frozen in liquid nitrogen, and stored at -20°C until further processing. For electron microscopy, large size samples (1 mm × 2 mm) were embedded in epoxy resin (Coulombe et al., 1989) to optimize orientation. After curing, the blocks were trimmed (0.5 mm 2) so as to obtain thin sections from wound edge tissue. For indirect immunofluorescence, 5-~m-thick sections were made from the same (frozen) biopsies, without further trimming, so that skin tissue extended ~1.5 mm laterally from the wound site. The primary antibodies used for immunostaining included rabbit polyclonal antisera directed against human K16 (Takahashi et al., 1994), K6 (Stoler et al., 1988), and filaggrin (Dale et al., 1985); a guinea pig polyclonal antisera directed against human K5 (Lersch et al., 1989); and mouse monoclonal antibodies directed against human K17 and K10/Kll (Sigma Chem. Co., St. Louis, MO). Bound primary antibodies were revealed using goat secondary antibodies conjugated to FITC or rhodamine (KPL Laboratories Inc.). Nonwounded skin tissue was used as a control in these experiments. Production of Human Recombinant K6b and K16 cDNAs encoding K6b and K16 were cloned by applying a coupled reverse transcription-polymerase chain reaction (RT-PCR) as previously reported The Journal of Cell Biology, Volume 132, 1996 382 on July 0, 2017 jcb.rress.org D ow nladed fom (Paladini et al., 1995; Takahashi et al., 1995). Oligonucleotides primers were designed from the published sequences of the human K6b (Tyner et al., 1985) and K16 (Rosenberg et al., 1988) genes and applied on total RNA isolated from cultured primary human epidermal keratinocytes. To facilitate the subcloning of the cDNA clones into either plasmid pET-8c (Studier et al., 1990) or pGEM5zf (Promega Corp., Madison, WI), each primer had a 10-nt extension on its 5' end that contained a restriction enzyme recognition sequence. Recombinant clones were grown, and the entire cDNA inserts subjected to sequencing. The amino acid sequence predicted from the K6b cDNA clone selected corresponds to that of the previously cloned gene (see Takahashi et al., 1995). In the K16 cDNA clone, the first codon was altered (Thr--~Ala) as a result of the subcloning strategy (see Paladini et al., 1995). An identical change in the human K14 cDNA was shown to have no apparent effect on its assembly behavior in vitro (Coulombe and Fuchs, 1990). Keratin Expression, Purification, and Immunological Analyses We used an Escherichia coli expression system based on the phage T7 RNA polymerase gene (Studier et al., 1990) to generate mg quantities of recombinant human epidermal keratins as described (Coulombe and Fuchs, 1990). Plasmids pET-K5 and pET-K14 (Coulombe and Fuchs, 1990), as well as pET-K6b and pET-K16 (this study) were individually transformed into E. coli strain BL21 (DE3), grown to ODr00 of N0.5, and recombinant keratin expression was induced by adding isopropyl-13-Dthiogalactopyranoside to 1 mM and carried out for 5 h. Inclusion bodies were isolated from lysed bacterial pellets and solubilized in a buffer containing 6.5 M urea, 50 mM Tris-HCl, 2 mM DTr , 1 mM EGTA, 1 mM PMSF, pH 8.1 (Q buffer). Recombinant keratins were purified to nearhomogeneity by chromatography in Q buffer on a Pharmacia FPLC Mono Q anion-exchange column operated at 0.5 ml .min-k Proteins of interest were eluted with a 0-200 mM linear gradient of guanidine-HCl over a 15ml vol, and 0.5-ml fractions were collected and analyzed by 10% SDSPAGE. Native human keratins were isolated from cultured SCC-13 cells, a squamous skin carcinoma cell line (Wu et al., 1982), using the high-salt extraction method (Lowthert et al., 1995). The final pellet was solubilized in Q buffer, and subjected to Mono Q chromatography as described above. Protein concentration was determined by the Bradford assay (Bradford, 1976) using reagents purchased from Bio-Rad Labs. (Richmond, CA). For immunoblot analyses, known quantities of recombinant and native human keratins were electrophoresed, electroblotted to nitrocellulose, and the blots incubated with diluted primary antisera prepared in blocking buffer (Tris-buffered saline with 0.5% Tween 20 and 5% dry milk). Bound primary antibodies were revealed by alkaline phosphataseconjugated secondary antibodies as recommended by the manufacturer (Bio-Rad Labs.). Chemical Cross-Linking Mono Q fractions containing heterotypic keratin complexes were used for chemical cross-linking as previously described (Coulombe and Fuchs, 1990). Purified recombinant type I and type II keratins were mixed in a ~45:55 molar ratio at a final concentration of 750 ixg.m1-1 and resubjected to the anion-exchange chromatography protocol described above. Mono Q fractions containing heterotypic complexes were dialyzed overnight against 25 mM sodium phosphate, 10 mM 13-ME, containing either 6.5 or 8 M urea at pH 7.4, to remove Tris ions, which interfere with the cross-linking agents. Protein concentration was adjusted to 200 i.~g-m1-1. Chemical cross-linking was performed by adding BS3 (bis-(sulfosuccinimidyl) suberate; 10 mM) for 1 h at 12°C (Geisler et al., 1992). In some experiments, glutaraldehyde was used at similar concentrations (see Coulombe and Fuchs, 1990). Cross-linked products (3 i.Lg total protein) were resolved on a 3-17.5% gradient SDS-PAGE, and stained with Coomassie blue. The apparent molecular mass of cross-linked species was calculated from a standard curve established with proteins of known molecular mass values. In Vitro Keratin Filament Assembly, Negative Staining, and Electron Microscopy Mono Q fractions containing heterotypic keratin complexes were used for in vitro polymerization assays as previously described (Coulombe and Fuchs, 1990). Polymerization was achieved by extensive dialysis of 0.25-ml samples at 200 ixg.m1-1 against 5 mM Tris-HC1, 10 mM 13-ME, pH 7.4. Dialysis was carried out at 4°C for 16-24 h for optimal results. In some experiments, the polymerization buffer was modified in terms of its ionic strength (50 mM Tris-HCl), pH (7.0), and presence of salt (150 mM NaC1). Polymerized keratins were adsorbed onto glow-discharged carboncoated 400 mesh grids (Ted PeUa Inc., Redding, CA), negatively stained with 1% aqueous uranyl acetate/0.025% tylose, and visualized on a Zeiss EM10A electron microscope operated at 60 kV. Micrographs were recorded at a nominal magnification of 31,500x, and the magnification was calibrated using a carbon grating replica (Ernest Fullam no. 10021). For filament width determination, micrographs were printed (magnification: 120,000×) and a total of 60 filaments were sampled for each type l-type II combination considered in this study (10 randomly sampled filaments per each of three micrographs for each of two distinct assembly experiments). For the determination of polymerization efficiency, final assemblies (80-ul aliquots, corresponding to ~20 ug proteins) were centrifuged at 100,000 g for 40 min at 4°C, and supernatant and pellet fractions were analyzed by SDS-PAGE, Coomassie-blue staining, and gel scanning densitometry (MCID; Imaging Research Inc.). Transient Expression of Keratin cDNAs in Cultured Cells In Vitro Keratin cDNAs were subcloned from pET vectors into the GW1-CMV expression plasmid, featuring a cytomegalovirus promoter and a SV-40 polyadenylation signal. In the case of K16, an 11-kb genomic DNA fragment containing the entire functional gene (Rosenberg et al., 1988) was also subcloned in this expression vector. Transient transfection assays were performed in BHK-21 cells, a hamster kidney cell line (see Quinlan and Franke, 1982), and in PtK2 ceils, a rat kangaroo kidney epithelial cell line (Franke et al., 1978). All transfections were done on subconfluent cells grown on 22-mm 2 glass coverslips using the calcium phosphate precipitation method (see Letai et al., 1992). At 24, 36, 48, or 72 h posttransfection, cells were fixed with absolute methanol for 20 rain at -20°C and processed for morphological analysis. For indirect double-immunofluorescence studies, transfected keratins were detected with combinations of the primary and secondary antibodies described above. In addition, we used the mouse monoclonal antibody L2A1, which recognizes K8-K18 complexes (Chou et al., 1993), and the V9 Mouse monoclonal antibody directed against vimentin (Sigma Chem. Co.). As routine controls in all labeling protocols, mock-transfected cells were processed in parallel with the relevant antisera. Quantitation of Epidermal Keratins in Transfected PtK2 Cells PtK2 cells were seeded on 100-mm plates that contained one 22-mm 2 glass coverslip. A plate was transfected with either control CMV plasmid, CMV-K14 cDNA, or CMV-K16 cDNA exactly as described above. The amounts of DNA and the volume of calcium phosphate precipitates were scaled up on a per surface area basis to maintain conditions similar to those used for the experiments described above. At 72 h posttransfection, the glass coverslip was removed and processed for immunofluorescence staining to determine transfection efficiency. The remaining cells on the 100-mm dish were recovered by scraping and a Triton X-100/high salt insoluble extract was prepared as described (Lowthert et al., 1995). The final pellets were solubilized in Q buffer and protein concentrations determined as described above. For SDS-PAGE/immunoblot analyses, known amounts of purified recombinant K14 or K16 (5 ng; 10 ng; 25 ng; 50 ng) were coelectrophoresed with 1.5 ~g of extracts prepared from transfected cells (control CMV, K14, and K16), and blotted onto nitrocellulose. Blots were incubated with the polyclonal anti-K14 or anti-K16 antiserum, and bound primary antibodies were revealed by enhanced chemiluminescence as per the manufacturer's instructions (Amersham Corp., Arlington Heights, IL). To allow for a direct comparison of the K14 and K16 blots, the primary antibody dilutions were adjusted in preliminary experiments using serially diluted keratin standards.