Keratin 15 Expression in Stratified Epithelia: Downregulation in Activated Keratinocytes

Keratin 15 Expression in Stratified Epithelia: Downregulation in Activated Keratinocytes Ahmad Waseem, Bilal Dogan,*1 Nicholas Tidman,* Yasmin Alam, Patricia Purkis,* Sarah Jackson,* Anand Lalli, Michael Machesney,* and Irene M. Leigh* Head and Neck Cancer Research Programme, Divisions of Dentistry, Biochemistry and Molecular Biology, UMDS, Guy’s Hospital, London, U.K.; *ICRF Skin Tumor Biology Unit, The Royal London Hospital, London, U.K.

Keratin 15 (K15) is a type I keratin without a defined type II partner whose expression in epidermal diseases has not been investigated. In this study we have used LHK15, a monoclonal antibody raised against the last 17 amino acids of the K15 polypeptide, to show that K15 is expressed primarily in the basal keratinocytes of stratified tissues, including the fetal epidermis and fetal nail. Although K15 in normal hair follicles was virtually absent from hair bulbs, it was expressed by a subset of keratinocytes in the outer root sheath. By comparison, K14 expression was found throughout the outer root sheath of hair follicles; however, when both K14 alleles were naturally ablated, the expression of K15 was also observed throughout the outer root sheath of the follicles. Expression of K15 mRNA was assessed by in situ hybridization and corroborated the

data from immunostaining. An increase in K15 mRNA and protein expression in hair follicles from the K14 ablated epidermis suggested an upregulation of the K15 gene in the absence of the K14 protein. In organotypical cultures where differentiating keratinocytes expressed markers of activated phenotype, i.e., K6 and K16, expression of K15 was undetectable. The expression of K15 mRNA and protein was also downregulated in two hyperproliferating situations, psoriasis and hypertrophic scars. Because keratinocytes in psoriasis and hypertrophic scars are activated, we conclude that K15 expression is not compatible with keratinocyte activation and the K15 gene is downregulated to maintain the activated phenotype. Key words: hyperproliferation/hypertrophic scar/in situ hybridization/psoriasis. J Invest Dermatol 112:362–369, 1999

K

keratin can form heterotypic complexes with any type II keratin in vitro (Hatzfeld and Franke, 1985; Waseem et al, 1996; Hofman and Franke, 1997) and assemble into filaments in fibroblasts (Lu and Lane, 1990), keratins are often expressed in specific pairs in an epithelial tissue that are unique to that differentiation pathway (Moll et al, 1982; Quinlan et al, 1985). The keratinocytes in the mitotically active basal layer always express K5 and K14 (Nelson and Sun, 1983), and upon commitment to differentiation the basal keratinocytes exit from the cell-cycle and move into the suprabasal compartment (Fuchs, 1993). The commitment to differentiation, which is mostly irreversible (Hotchin et al, 1993), involves major changes in the cytoarchitechture in which keratinocytes are flattened, lose their organelles and nuclei, and their protein contents are cross-linked into a cornified envelope (Reichert et al, 1993; Simon, 1994; Eckert et al, 1997). These changes in keratinocytes are accompanied by a switch in keratin gene expression. The migrating keratinocytes downregulate K5/K14 transcription and activate expression of a new set of keratin pairs that vary among stratified tissues. In cornified epithelia, such as those covering skin and gingivae, the differentiating keratinocytes express K1/K10 (Morgan et al, 1987; Leigh et al, 1993), whereas in noncornified squamous epithelia expression of K4/K13 (Morgan et al, 1987; Sawaf et al, 1991) is induced and in cornea the migrating keratinocytes activate K3/K12 expression (Quinlan et al, 1985). The K6/K16 pair is constitutively expressed in the outer root sheath of hair follicles (Stark et al, 1987; Heid et al, 1988) and in mucosal epithelia of oral cavity, esophagus, and female genital tract (Quinlan et al, 1985); however, in epidermal hyperproliferation, such as in psoriasis (Weiss et al, 1984; Stoler et al, 1988) and during

eratins belong to the family of intermediate filament proteins that are specifically expressed in epithelia. They have a remarkable ability to polymerize into 10 nm filaments without the participation of auxiliary proteins (Coulombe and Fuchs, 1990; Hatzfeld and Weber, 1990; Steinert, 1990). They comprise a total of about 30 genes (including those of hair and nails, the trichocytic keratins) grouped into two types; type I are smaller (40–56.5 kDa) and acidic (pI , 7.0), whereas type II are larger (53–67 kDa) and basic/ neutral (pI ù 7.0) (reviewed in Fuchs and Weber, 1994). The type I keratins include K9–K20 and the type II include K1–K8. The amino acid sequence of keratins is highly conserved in the central rod domain of the polypeptides and forms an α-helical structure (for a review see Albers and Fuchs, 1992). During filament assembly, two keratin polypeptides, one of each type, first form a parallel heterodimer, in which the rod domains assemble into coiled-coil (Hatzfeld and Weber, 1990; Steinert, 1990), which then undergoes further associations with other dimers to produce tetramers (Coulombe, 1993; Stewart, 1993). The association of tetramers produces protofilaments and finally mature filaments. Although any type I

Manuscript received July 23, 1998; revised September 29, 1998; accepted for publication November 25, 1998. Reprint requests to: Dr. Ahmad Waseem, Department of Craniofacial Development, Guy’s Dental School, Floor 28, Guy’s Tower, London Bridge, London SE1 9RT, U.K. 1Current address: GATA Haydarpasa EGt.Hst. Dermatoloji K1.S.Yrd, Istanbul, Turkey.

·

0022-202X/99/$10.50 Copyright © 1999 by The Society for Investigative Dermatology, Inc.

362

VOL. 112, NO. 3 MARCH 1999

wound healing (Mansbridge and Knapp, 1987; Paladini et al, 1996), expression of this pair is activated. K15 is a type I keratin that does not appear to have a natural type II expression partner. Earlier studies using radioactive in situ hybridization have shown that the K15 mRNA is expressed in all layers of stratified epithelia (Leube et al, 1988). The conclusion drawn from these studies was that the K15 expression starts in the basal layer but is independent of the vertical differentiation of migrating keratinocytes. Recent studies, however, have shown that K15 is not expressed in the suprabasal layer but instead is specifically localized in the basal keratinocytes (Lloyd et al, 1995). Despite the fact that K15 is regarded as specific for basal keratinocytes, its distribution in normal and epithelial-related diseases has not been investigated in detail. The studies on K15 reported so far are based on the use of polyclonal antibodies raised in rabbits (Lloyd et al, 1995) or in guinea pigs (Jonkman et al, 1996), and there is no monospecific monoclonal antibody described in the literature against K15. In this study we have produced LHK15, a monospecific antibody against K15, and used this to investigate the expression of K15 in fetal and adult stratified epithelial tissues. We have shown that K15 protein and mRNA are primarily located in the basal keratinocytes of stratified tissues. We have also shown that the K15 gene is upregulated in human subjects where both alleles for K14 have been inactivated. In hyperproliferating conditions, in which keratinocytes are activated, the K15 protein and its mRNA are downregulated, suggesting that its expression may not be compatible with the activated phenotype. MATERIALS AND METHODS Antibodies, cell lines, and tissues Mouse monoclonal anti-involucrin, SY3 (Hudson et al, 1992), was supplied by Dr. F.M. Watt (Imperial Cancer Research Fund, Lincoln’s Inn Fields, London, U.K.), and antitransglutaminase, B.C1 (Thacher and Rice, 1985), was a kind gift from Prof. R.H. Rice (University of California, Davis, CA). Polyclonal antiloricrin antibody (Yuspa et al, 1989), raised in rabbits against a C-terminal peptide derived from murine loricrin, was obtained commercially (Babco, Richmond, CA). The anti-keratin antibodies used in this study were as follows: LHK1 for K1, LHK6b for K6, LHK16 for K16 (Machesney et al, 1998), and LL001 for K14 (Purkis et al, 1990). The keratin antibodies were obtained by culturing hybridomas in 10% fetal calf serum in Dulbecco’s modified Eagle’s medium and the supernatants from confluent cultures were stored in 0.2% (wt/vol) sodium azide until used in immunostaining. Keratinocytes from normal and K14 ablated epidermis were obtained using standard procedures (Lane et al, 1985) and cultured in a mixture of Dulbecco’s modified Eagle’s medium and Ham’s F12 in a ratio of 3:1 plus 10% fetal calf serum. Simple epithelial cell lines, MCF-7 (human breast carcinoma) and HeLa (human cervical adenocarcinoma), were cultured in Dulbecco’s modified Eagle’s medium 1 10% fetal calf serum. Human keratinocyte cell lines A-431, derived from a vulval carcinoma, and TR 146, derived from an epidermal carcinoma, were cultured in Dulbecco’s modified Eagle’s medium 1 10% fetal calf serum, whereas HaCaT, an immortalized epidermal keratinocyte cell line, was cultured using conditions similar to those employed for primary keratinocytes. All tissue samples were frozen soon after surgical removal and stored at –70°C until needed for sectioning. The protocol for the use of human material was approved by the local ethical committee. Production of monoclonal antibody LHK15 The monoclonal antibody LHK15 was raised in mice against a peptide (NH2-[Cys]VEESVDGQVVSSHKREI-COOH) derived from the last 17 amino acids of the K15 polypeptide. The peptide was conjugated to thyroglobulin through an extra cysteine residue at the N-terminus and was used to immunize mice. A slightly shorter peptide containing the last 12 amino acid residues has been shown to elicit K15-specific antibodies in rabbits when coupled to keyhole limpet hemocyanin (Lloyd et al, 1995). BALB/C mice were immunized subcutaneously at four three-weekly intervals with a total of about 1.2 mg of the peptide-thyroglobulin conjugate (50 µg per injection per animal). Tail bleeds were tested by enzyme-linked immunosorbent assay (ELISA) and mice with acceptable level of antibodies (titre 5 1/ 10,000 by ELISA) were boosted intravenously 5 d prior to fusion. Spleen cells were fused with Sp2/O-Ag14 nonproducer myeloma cells in the presence of PEG 6000 (Merck, Leicester, U.K.) as described previously (Waseem and Lane, 1990). Hybridomas were selected in a medium

KERATIN 15 IN STRATIFIED EPITHELIA

363

containing azaserine and hypoxanthine and screened against the peptide in an ELISA. The positive wells were further characterized on sections of normal human skin. The hybridoma with strongest reactivity was cloned twice by limiting dilution and the supernatant collected and stored in 0.2% (wt/vol) sodium azide until used. The isotype of the antibody was determined using a commercially available kit (Amersham International, Amersham, U.K.) and was found to be IgG2a. Production of organotypical cultures Organotypical cultures were generated by growing primary human keratinocytes on a sheet of dead dermis as described previously (Todd et al, 1993). Briefly, pieces of deepidermised dermis (1.5 cm2) were obtained from human skin stored in 80% glycerol (European Skin Bank, The Netherlands) by incubating in PBS for about 10 d. Primary human dermal fibroblasts were allowed to attach overnight on the underside of the de-epidermised dermis. The dermis was turned over the following day and primary keratinocytes (3 3 105 in 500 µl) were seeded into a stainless steel ring placed over the dermis. After the cells had attached, the steel ring was removed and the keratinocytes were allowed to grow in Dulbecco’s modified Eagle’s medium/Ham’s F12 (in a ratio of 3:1) 1 10% fetal calf serum for 4 d to form a monolayer following which the dermis was raised onto a metal grid to form an air–liquid interface. The culture was continued for the next 2 wk after which the dermal plug was removed, mounted onto a cork, and sectioned in a cryostat. Polymerase chain reaction (PCR) and cloning of amplified products The cRNA probes for K14 and K15 were prepared by in vitro transcription of the cDNA fragments cloned in the pGEM-4 vector (Promega, Madison, WI). The K14 fragment was prepared by PCR amplification using a forward (AGGACGGAATTCTCTCCTCCTCCCAGT) and a reverse (GAGCGGGGAAGCTTAGCCTCAGTTCTTGG) primer from the 39 end of K14 cDNA (Marchuk et al, 1985). The K15 cDNA was synthesized by Superscript reverse transcriptase (GibcoBRL/ Life Tech, Glasgow, U.K.) with total RNA purified from A-431 cells and a K15 specific primer (CTTGCTCCAAAGAAGGTGGGG) following the protocol described previously (Waseem et al, 1990, 1998). Using the reverse transcriptase reaction mixture as a source of cDNA, the K15 fragment was amplified using a forward (AGAAATCTGAATTCCTATTGCAGGAGA) and a reverse (CCCTGAAAGCTTAGACCGAGGGACCCT) primer. An EcoR I site in the forward primers and a Hind III site in the reverse primers were introduced to facilitate cloning of the PCR products in the pGEM4 vector (shown by italics in the nucleotide sequence). The PCR amplification was carried out using Expand Long Template PCR system (Boehringer, Mannheim, Germany), which is a mixture of thermostable Taq and Pwo DNA polymerases. The presence of Pwo in the PCR mixture increases the fidelity of amplification due to the proof reading activity of this enzyme. The reaction mixture containing 1 µg cloned cDNA per ml or 5 µl of reverse transcriptase reaction mixture, 350 µM of each dNTP, 1.75 mM of MgCl2, 5 µg per ml each of forward and reverse primers in 50 mM Tris-HCl, pH 9.2 plus 14 mM ammonium sulfate in a total volume of 50 µl was vortexed and covered with mineral oil. The conditions for amplification were identical to those described recently (Machesney et al, 1998). The PCR products were digested with a mixture of EcoR I and Hind III and ligated into the corresponding sites of pGEM-4 (Promega) and used to transform Escherichia coli strain DH5α. The transformants were identified by restriction enzymes and the keratin inserts were confirmed by nucleotide sequencing. The pGEM-K14 construct contained a 148 bp fragment, nucleotides 1335–1483 (Marchuk et al, 1984, 1985), and the pGEM-K15 contained a 127 bp fragment, nucleotides 1437–1563 (Leube et al, 1988). DIG-labeled cRNA probes DIG-labeled cRNA probes were generated using the DIG RNA labeling kit following the method provided by the manufacturer (Boehringer). The purified plasmids, pGEM-K14 and pGEMK15, were linearized with EcoR I (for sense probe) and with Hind III (for the anti-sense probe). The linearized plasmids (1 µg each) were used in in vitro transcription reaction using DIG-UTP and the transcription was catalyzed by SP6 RNA polymerase and by T7 RNA polymerase, respectively, for the sense and anti-sense orientations. Non-radioactive in situ hybridization Tissues sections for in situ hybridization were cut and thaw mounted onto diethylpyrocarbonatetreated 3-aminopropyl triethoxy-silane-coated slides. The sections were fixed in 4% paraformaldehyde and after washing first in PBS and then in PBS/0.1 M glycine, the sections were deproteinated with 1 µg proteinase K per ml at 37°C for 10 min in a Hybaid Omnislide System (Hybaid, Middlesex, U.K.). The reaction was stopped by 4% paraformaldehyde and

364

WASEEM ET AL

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

the sections were acetylated with 0.1 M triethanolamine (pH 8.0)/0.25% acetic anhydride followed by prehybridization for 1 h at 45°C. The sections, first dehydrated by incubating in increasing concentrations of ethanol, were hybridized with 1 ng of the probe in 25 µl of hybridization buffer at 65°C for 7 min and then overnight at 45°C. After washing in 50% formamide/23SCC, the sections were first treated with 50 µg RNase A per ml at 37°C for 30 min followed by two more washes in 50% formamide/0.13SCC at 52°C for 30 min. The sections were treated with 10% sheep serum and incubated with 50 µl of 1:400 dilution of anti-DIG alkaline phosphatase-conjugated antibody (Boehringer). The sections were developed as described before (Machesney et al, 1998) using 5-bromo-4chloro-3-indolyl-phosphate/nitro blue tetrazolium as chromogenic substrate for alkaline phosphatase. Immunocytochemistry The tissue specimens and organotypical cultures of keratinocytes were mounted in cryo-m-bed (Bright Instrument, Cambridge, U.K.) on cork discs and 5 µm sections were cut at –25°C and thaw mounted onto 3-aminopropyl triethoxy-silane-coated slides. To get sections through hair follicles, plucked hairs were embedded into an agarose gel and cut into blocks, which were mounted onto corks and sectioned in a cryostat. The sections were dried in air and incubated with the primary antibody for 1 h at room temperature or at 4°C overnight. The sections were washed in tap water and incubated with either a horseradish peroxidase-conjugated or a fluorescein isothiocyanate-conjugated second antibody (DAKO, Cambridge, U.K.). The fluorescein isothiocyanatelabeled sections were counterstained with 2 µg propidium iodide per ml and mounted in 90% glycerol in PBS containing 25 mg 1,4-diazabicyclo(2,2,2)-octane per ml as anti-fading agent. The sections stained with peroxidase linked antibody were developed with 3-amino-9-ethyl carbizole as substrate. Extraction of keratin polypeptides and their analysis by western blotting was carried out following the method described earlier (Waseem et al, 1996). DNA constructs were confirmed by di-deoxy sequencing using 35S ATP. All blots and micrographs were reproduced and/or assembled using the Adobe PhotoShop and printed using an Epson inkjet color printer.

Figure 1. LHK15 reacts specifically with K15 polypeptide on immunoblots. (a) Detection of K15 in cytoskeletal extract of different human epithelial cells. Keratin polypeptides were extracted from cultured keratinocytes using low and high salt buffers containing Triton X-100, separated on a 12% sodium dodecyl sulfate polyacrylamide gel, transferred onto a nitrocellulose membrane, and probed with LHK15 antibody. Lane 1, A-431 (vulval carcinoma line); lane 2, MCF-7 (breast carcinoma line); lane 3, TR 146 (epidermal carcinoma line); lane 4, HaCaT (immortalized epidermal keratinocyte line); and lane 5, primary human keratinocytes. (b) Detection of K6, K14, and K15 in a keratin extract of HaCaT cell line. Keratin polypeptides isolated from HaCaT cells were analyzed by immunoblotting. Lane 1, LHK6 for K6; lane 2, LL001 for K14; lane 3, LHK15 for K15.

RESULTS AND DISCUSSION

shown). The data suggested that the sequence of the peptide used as immunogen is not conserved in all species and highlights variations in this region of the pig K15 polypeptide.

Reactivity of LHK15 The monoclonal antibody LHK15 reacted strongly with free and conjugated K15 peptide in an ELISA that was the basis for selection of this clone during hybridoma production. In order to demonstrate that LHK15 would react with the peptide when it is a part of the K15 protein, keratin polypeptides extracted from A-431, TR 146, MCF-7, primary keratinocytes, and HaCaT cells were probed with LHK15 on a western blot. As shown in Fig 1(a), the LHK15 antibody reacted with a protein band of about 52 kDa present in the cytoskeletal extract of A-431 and HaCaT cells, but did not react with keratins present in other cell extracts. This clearly suggests that LHK15 is not reacting with K14, which was detected in all the cellular extracts except MCF-7 (data not shown). In order to show that LHK15 did not react with a high molecular weight keratin, such as K6 present in some cell extracts, keratins extracted from HaCaT were probed with K6, K14, and K15 antibodies. As shown in Fig 1(b), the protein bands reactive with K14 and K15 antibodies had similar mobilities but they were distinct from the K6 band. Taken together these data suggested that the LHK15 was able to recognize the peptide when it is a part of the K15 polypeptide. The specificity of the LHK15 was demonstrated by the ability of the K15 peptide to inhibit the binding of the antibody to immobilized peptide in an ELISA. Furthermore, inclusion of 10 µM K15 peptide in the antibody supernatant completely inhibited the immunostaining of basal keratinocytes by LHK15 (not shown); however, the K15 peptide did not compete with the binding of the antibody to K15 polypeptide in a western blot. This may indicate that the peptide, when part of an immobilized K15 polypeptide, has much higher affinity for the antibody than when it is part of a heterotypic complex or when it is a free peptide. Because LHK15 was raised against a peptide derived from human K15, we decided to determine whether this antibody would react with K15 from other species. The reactivity of LHK15 was tested with tissue sections derived from pig, mouse, cow, and rat skin. The antibody reacted with mouse and rat skin samples and cow snout but did not show detectable reactivity with pig skin (not

Expression of K15 in basal keratinocytes of fetal and adult origins Using LHK15 we have determined the distribution of K15 in human tissues and compared it with K14 expression. In human skin taken from different body sites K14 expression was detected continuously in the basal and suprabasal layers. K15 expression, on the other hand, was discontinuous but confined to the basal layer (Fig 2a, b). Similar results have been reported by others using a polyclonal antibody (Lloyd et al, 1995). The discontinuous staining of basal keratinocytes by LHK15 was noticed only in adult skin and not in other stratified tissues (Fig 2d). A pattern of K14 and K15 expression similar to that of skin was observed in other stratified tissues such as palate and vagina (not shown), but not in esophagus and cervix. Although both K14 and K15 were expressed in the basal layer of esophagus, the K15 staining was localized only to a single layer of the basal keratinocytes, whereas K14 was present in more than one layer (Fig 2c, d) but not in different suprabasal layers as seen in skin (see Fig 2a). A staining pattern similar to esophagus was seen in sections derived from cervix (not shown). This difference in the K14 expression (basal versus suprabasal) perhaps represents differentiation-specific lineages, or may be due to the functional differences between epithelia. The data suggest that K15 is a marker of basal keratinocytes and the signals that activate K1 and K10 and other differentiationspecific markers in suprabasal keratinocytes are perhaps responsible for the downregulation of the K15 gene. Amongst the human tissues we analyzed, the expression of K15 was most unusual in hair follicles. Human hairs plucked from normal epidermis with the full repertoire of their follicular structure were used for immunostaining with LL001 and LHK15 antibodies, as has been described previously (Lane et al, 1991). K14 was expressed in keratinocytes of the outer root sheath throughout the hair follicles (see Fig 2e), as has been reported earlier (Lane et al, 1991). K15 on the other hand was absent from the hair bulb but was expressed in keratinocytes of the outer root sheath located above the hair bulb (Fig 2f). A careful comparison of the immunofluorescence using nuclei stained with propidium iodide indicated

VOL. 112, NO. 3 MARCH 1999

KERATIN 15 IN STRATIFIED EPITHELIA

365

that K14 and K15 do not colocalize in the outer root sheath; in fact K14 was localized in the outermost layer of the follicles, whereas K15 was in the next inner layer. It is possible that the K15 positive cells form a distinct layer, either part of Henle’s layer of the inner root sheath or as part of the outer root sheath representing a population of basal keratinocytes. Absence of the K15 in hair bulbs is intriguing and may indicate either the absence of basal keratinocytes in this region or that the basal keratinocytes in the hair bulb are a subset unique from the rest of the population. Keratin expression in adult epidermis is very different from that in fetal skin. The basal keratinocytes of a 15 wk old human foetus express K8, K18, and weakly K4, whereas the periderm and intermediate layers express K4, K14, K13, and K19 (Lane et al, 1985). The K4/K19 expression persists until 20 wk of gestation, following which these keratins are replaced by the adult phenotype of K5/K14 in the basal and K1/K10 in the suprabasal layers (Van Muijen et al, 1987). Among the keratin polypeptides expressed in the fetal epidermis, K4 and K13 are suprabasal in nature and K8, K18, and K19 are perhaps derived from the simple epithelial layer that surrounds the early embryo but is phased out during development. Thus the signals that activate K4 and K13 transcription during skin development are upstream to those required for the activation of K1 and K10 genes. We determined the distribution of K14 and K15 expression in developing epidermis and nails of a 15 wk old human foetus. Interestingly the K15 expression in the fetal skin was localized only in the basal keratinocytes, whereas K14 was in the basal and epibasal layers (Fig 2g, h). Similarly in fetal nail the K14 was distributed homogeneously throughout the basal and suprabasal layers, whereas the K15 expression was concentrated only in the basal keratinocytes (Fig 2i, j). The persistence of K15 in the basal keratinocytes of fetal and adult epidermis may indicate an important role for this keratin in the maintenance of tissue integrity. Furthermore, because the basal layer contains mitotically active cells, some of which may account for the stem cell population (Lane et al, 1991), the expression of K15 in these cells may suggest a role for this keratin in keratinocyte growth and differentiation. In this context it is interesting to note a strong reactivity of LHK15 with keratin filaments of premitotic basal keratinocytes, identified from the rest by virtue of their strong reactivity with α6 integrin (Dr. P. Kaur, personal communication), may indicate expression of K15 predominantly in stem cell population of the basal keratinocytes. K15 gene is transcribed primarily in the basal keratinocytes of stratified epithelia In order to determine how K15 expression is regulated in different stratified squamous epithelia, we determined K15 mRNA in different layers of stratified tissues and compared with that of the K14 mRNA expression. In the tissues we analyzed we detected K15 mRNA only in the basal layer corroborating the expression of K15 protein, which suggests that K15 expression is regulated at the level of transcription without the post-transcriptional constraints described for other keratins (Tyner and Fuchs, 1986). In oral mucosa the expression of K15 mRNA was identical to that of K14, with both mRNA being located only in the basal layer (Fig 3a, b). In palate, however, the K14 expression was in the basal and suprabasal layers, whereas the K15 mRNA was in the basal layer and the next two upper layers (Fig 3c, d). In skin,

Figure 2. Reactivity of the monoclonal antibody LHK15 with basal keratinocytes in stratified squamous epithelia of fetal and adult origins. Comparison of the K14 expression with that of K15 in human stratified epithelia. Unfixed frozen human tissue sections were reacted with LL001 for K14 (a, c, e, g, i) and with LHK15 for K15 (b, d, f, h, j) for 2–4 h at room temperature. After washing, the sections were incubated first with fluoresceine-labeled goat anti-mouse followed by a propidium iodide solution before mounting. The tissues used were normal breast skin (a, b), esophagus (c, d), hair follicles (e, f), fetal epidermis (g, h), and fetal nail (i, j). Discontinuous staining of the basal layer by LHK15 in (b) is shown by arrows. Note the red staining of nuclei with propidium iodide. Scale bar: (a–d, g, h) 50 µm, (e, f, i, j) 100 µm.

366

WASEEM ET AL

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Figure 3. K15 mRNA is primarily expressed in the basal keratinocytes of stratified squamous epithelia. Comparison of the K14 mRNA expression with that of K15 in stratified squamous epithelia. Unfixed frozen human tissue sections were fixed in 4% paraformaldehyde and processed for in situ hybridization using a highly specific cRNA probe for K14 (a, c, e) and K15 (b, d, f) as described in Materials and Methods. The tissues used were oral mucosa (a, b), palate (c, d), and skin (e, f). Expression of K15 mRNA (g) and protein (h) in hair follicles from a patient suffering from the recessive form of epidermolysis bullosa simplex. In this case both K14 alleles were naturally ablated and the K14 protein was undetectable. The separation of epidermis from the dermis due to skin blistering is marked by asterisks. The expression of K15 in the hair bulb that was absent in normal follicles (see Fig 2f) can be clearly seen here. Scale bar: 100 µm.

the K14 expression was in the basal and suprabasal layers, whereas the K15 mRNA was localized only in the basal layer (Fig 3e, f). In general, we have noticed that the K15 mRNA expression was stronger in mucosal tissues than in skin, which is consistent with our immunoblotting data showing higher levels of K15 expression in A-431, a mucosal cell line, compared with HaCaT cells that are epidermal in origin (see Fig 1a, compare lane 1 with lane 4). This may indicate either a higher rate of K15 transcription or an elevated K15 mRNA stability in mucosal tissues compared with that in skin. Because the same probes were employed on various tissues, the differences in K14 and K15 expression represent actual mRNA levels rather than a nonspecific probe reactivity. K15 expression is upregulated in K14 ablated epidermis Because K14 and K15 are the major type I keratins of basal keratinocytes, we decided to determine K15 expression in the absence of K14. Point mutations in the K14 gene have been shown to cause a dominant form of epidermolysis bullosa simplex (Coulombe et al, 1991; Fuchs and Coulombe, 1992), whereas a complete absence of the gene leads to an epidermolysis bullosa simplex that is inherited as an autosomal recessive disorder (Rugg et al, 1994; Chan et al, 1995). We used LHK15 to determine K15 expression in epidermis from an individual in which both K14 alleles were naturally ablated. Although the LHK15 staining in the absence of K14 was stronger compared with normal epidermis, it

was most pronounced in hair follicles. Unlike normal skin where K15 was absent in hair bulbs (Fig 2f), the LHK15 staining in the K14 ablated epidermis was present throughout the hair follicles (Fig 3h), a pattern similar to K14 expression in normal skin (Fig 2e). This apparent increase in the LHK15 staining could either be due to stabilization of the K15 in absence of K14 or be due to an increase in transcription of the K15 gene. We therefore determined the level of K15 mRNA in the K14 ablated hair follicles by in situ hybridization and the results suggested that the K15 mRNA was present throughout the outer root sheath, thus corroborating the K15 protein expression (Fig 3g). The elevated level of K15 mRNA in the absence of K14 protein perhaps indicates transcriptional activation of the K15 gene. Transgenic studies have shown that in the basal keratinocytes of K14 ablated mice, keratin filaments can assemble between K5 and K15 without an upregulation of K15 (Lloyd et al, 1995). These results might suggest that the heterotypic associations between K5 and K15 could stabilize the K15 polypeptide; however, if K15 turnover is reduced by K5 then one would expect an apparent increase in the level of K15. In fact, recent studies (Jonkman et al, 1996) have identified four individuals with a recessive form of epidermolysis bullosa simplex in which both K14 alleles were naturally ablated. Tissue samples from these patients showed that K5 was able to associate with K15 into protofilaments but not into tight intermediate filaments, and that there was a distinct

VOL. 112, NO. 3 MARCH 1999

KERATIN 15 IN STRATIFIED EPITHELIA

367

Figure 4. Downregulation of K15 in primary human keratinocytes undergoing differentiation in an organotypical culture. Primary human keratinocytes were cultured on a layer of human dermal fibroblasts growing on de-epidermised dermis as described in the Materials and Methods. The dermal plug was removed from the medium after 16 d and sectioned in a cryostat. The sections were fixed in a mixture of acetone/methanol (1:1), dried in air, and incubated with primary antibodies for transglutaminase (a), loricrin (b), involucrin (c), K1 (d), K10 (e), K14 (f), K6 (g), K16 (h), and K15 (i). The sections were incubated with fluoresceine-labeled goat anti-mouse and propidium iodide before mounting. The staining of nuclei with propidium iodide is shown by red or yellow. Scale bar: 100 µm.

upregulation of K15 in keratinocytes (Jonkman et al, 1996). The lack of K15 upregulation in the absence of K14 in transgenic mice in not clear. It is possible, however, that the K15 upregulation in humans is a species-specific phenomenon and that in mouse other factors play a role in the filament assembly. Our observation that K15 expression is increased in the absence of K14 might suggest a compensatory role for K15; however, the molecular mechanism of K15 upregulation is not clear. Our in situ hybridization data suggest that the increase in K15 expression in the absence of K14 protein may involve elevated transcription of the K15 gene. There are two possibilities to explain our observations. First, that there is a specific association between K14 protein and the K15 mRNA that increases the turnover of K15 mRNA. This is conceivable because intermediate filament proteins have

been shown to bind nucleic acids (Kuhn et al, 1987; Traub and Shoeman, 1994a, b; Wang et al, 1996; Hartig et al, 1997); however, it seems unlikely that such an interaction would specifically destabilize the K15 mRNA. A second possibility is that the K15 gene is transcriptionally repressed by the K14 protein or by its peptides. This would require the K14 peptides to cross the nuclear membrane barrier to influence gene transcription. In vitro transfection experiments have shown that truncated keratin polypeptides can enter the nucleus (Bader et al, 1991) and it is possible that keratin peptides could cross the nuclear membrane to modulate gene expression. This could be a natural mechanism for keratin homeostasis in keratinocytes; however, it would require the K15 gene to have an element(s) for specific binding of other keratins or their peptides. Although keratin polypeptides have been shown to bind tightly

368

WASEEM ET AL

Figure 5. Expression of the K15 mRNA and protein are downregulated in hyperproliferative conditions. Unfixed frozen tissue sections from psoriatic epidermis (a, b) were immunostained with the monoclonal antibody LL001 (a) and with LHK15 (b). Hypertrophic scar samples (c, d) were probed by in situ hybridization using K14 (c) and K15 (d) cRNA probes as described in Materials and Methods section. Scale bar: 100 µm.

with DNA (Hatzfeld and Weber, 1991), a regulatory element responsive to other keratin polypeptides appears to have escaped detection. Keratinocyte activation downregulates expression of K15 We have studied the expression of K15 in primary human keratinocytes grown on de-epidermised dermis at an air–liquid interface. The differentiation of keratinocytes in this model system was confirmed by the expression of specific markers, such as K1, K10, transglutaminase, loricrin, and involucrin (Fig 4); however, these keratinocytes also expressed K6 and K16, which indicated that under our experimental conditions they had undergone activation. When we tested LHK15 on these cultures we did not see specific reactivity with basal keratinocytes (Fig 4i). This indicated that the lack of K15 expression could have been due to the activation of keratinocytes in skin equivalents. We then studied the expression of K15 in two clinical situations with hyperproliferating epidermis, psoriasis, and hypertrophic scar. We have recently demonstrated that the epidermis of hypertrophic scars contains activated keratinocytes that express K6 and K16 (Machesney et al, 1998), and in that respect their pathology equates to the changes seen in psoriasis (Weiss et al, 1984; Stoler et al, 1988). The expression of K14 and K15 in psoriatic epidermis is shown in Fig 5(a, b). Whereas K14 expression was present throughout the psoriatic epidermis, K15 expression was completely absent. Similar results were obtained when serial sections were analyzed by in situ hybridization (not shown). These results clearly suggest a downregulation of K15 in psoriasis. Similar observations were made with hypertrophic scar epidermis (Fig 5c, d), suggesting

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

that the reduced K15 expression in hyperplasia may be a consequence of keratinocyte activation. It should, however, be noted that in both clinical situations the downregulation of K15 was observed at the level of protein as well as mRNA, which suggested that the transcription of K15 gene was suppressed in activated keratinocytes. The data from the clinical samples and the skin equivalents taken together suggest that K15 expression may not be compatible with keratinocyte activation and that this keratin gene is downregulated to maintain the hyperplastic phenotype. The primary tissue response towards an epidermal injury involves an influx of two types of lymphocytes, Th-1 and Th-2, which produce large number of cytokines to support an immune reaction. Subsequent to this primary response is the activation of keratinocytes, which makes them migratory, produce and respond to growth factors and cytokines, and secrete components of the basement membrane (Kupper, 1990). Activated keratinocytes in psoriasis and many inflammatory diseases overexpress transforming growth factor-α (Elder et al, 1989), interleukin-1 (Ristow, 1987), interleukin-6 (Grossman et al, 1989), and bone morphogenic protein-6 (Blessing et al, 1996), which has a pronounced effect on the expression of keratin genes. The production of interferon-γ by Th-1 lymphocytes is primarily responsible for the activation of K17 in several inflammatory diseases, including psoriasis (De Jong et al, 1991), delayed type hypersensitivity, and contact dermatitis through the induction of STAT family of transcription factors (Komine et al, 1996). The upregulation of K6 and K16 genes in psoriasis is apparently mediated by the action of EGF and TGFα (Jiang et al, 1993). Our observation that K15 is specifically downregulated in hyperproliferating epidermis suggests that the cytokines and growth factors produced by the activated keratinocytes suppress transcription of the K15 gene. Further experiments using the regulatory sequences from the K15 gene will be required to identify the cellular factors involved in suppression of the K15 transcription. In conclusion, we have produced and characterized a highly specific monoclonal antibody against K15 and used it to show that this keratin is specifically expressed in the basal keratinocytes of a variety of cornified and noncornified stratified epithelia, including fetal epidermis and fetal nail. In hair follicles, K15 expression was absent in the hair bulb but specific reactivity was observed in keratinocytes of the outer root sheath clearly distinct from the K14 staining. In situ hybridization studies using a cRNA probe highly specific for K15 also produced reactivity only in the basal keratinocytes, suggesting that the K15 transcription is downregulated as soon as postmitotic cells leave the basal layer. In patients with both K14 alleles ablated, we found an upregulation of K15 most prominently in the hair follicles. In hyperproliferating epidermis, such as in psoriasis and hypertrophic scars, K15 expression, both protein and mRNA, is downregulated. Similar absence of K15 expression was also noted in activated keratinocytes growing in skin equivalent cultures. These data strongly suggest that the activated phenotype is not compatible with the K15 expression.

We are grateful to Dr. Robert Evans, Division of Biochemistry and Molecular Biology, UMDS, and Dr. Kenneth White, School of Biological Sciences, University of North London, for their comments on this manuscript. We would like to thank the Wellcome Trust (grant no. 0378241/1.5 to A.W.) and the Guy’s Special Trustees for their financial support during the course of this study.

REFERENCES Alberts K, Fuchs E: The molecular biology of intermediate filament proteins. Int Rev Cytol 134:243–279, 1992 Bader BL, Magin TM, Freudenmann M, Stumpp S, Franke WW: Intermediate filaments formed de novo from tail-less cytokeratins in the cytoplasm and in the nucleus. J Cell Biol 115:1293–1307, 1991 Blessing M, Schirmacher P, Kaiser S: Overexpression of bone morphogenic protein6 (BMP-6) in the epidermis of transgenic mice: inhibition or stimulation of proliferation depending on the pattern of transgene expression and formation of psoriatic lesions. J Cell Biol 135:227–239, 1996

VOL. 112, NO. 3 MARCH 1999

Chan YM, Anton-Lamprecht I, Yu Q-C, Jackel A, Zabel B, Ernst J-P, Fuchs E: A human keratin 14 ‘‘knockout’’: the absence of K14 leads to severe epidermolysis bullosa simplex and a function for an intermediate filament protein. Genes Develop 8:2574–2587, 1995 Coulombe PA: The cellular and molecular biology of keratins: beginning a new era. Curr Opi Cell Biol 5:17–29, 1993 Coulombe PA, Fuchs E: Elucidating the early stages of keratin filament assembly. J Cell Biol 111:153–169, 1990 Coulombe PA, Hutton ME, Letai A, Hebert A, Paller AS, Fuchs E: Point mutations in human keratin 14 genes of epidermolysis bullosa simplex patients: genetics and functional analyses. Cell 66:1301–1311, 1991 De Jong EMGJ, van Vlijmen IMMJ, van Erp PEJ, Ramaeckers FCS, Troyanovski SM, van de Kerkhof PCM: Keratin 17: a useful marker in anti-psoriatic therapies. Arch Dermatol Res 283:480–482, 1991 Eckert RL, Crish JF, Robinson NA: The epidermal keratinocyte as a model for the study of gene regulation and cell differentiation. Physiol Rev 77:397–424, 1997 Elder JT, Fisher GJ, Lindquist PB, et al: Overexpression of transforming growth factor α in psoriatic epidermis. Science 243:811–814, 1989 Fuchs E: Epidermal differentiation and keratin gene expression. J Cell Sci Suppl 17:197–208, 1993 Fuchs E, Coulombe PA: Of mice and men: genetic skin diseases of keratins. Cell 69:899–902, 1992 Fuchs E, Weber K: Intermediate filaments: structure, dynamics, function, and disease. Annu Rev Biochem 64:345–382, 1994 Grossman RM, Krueger J, Yourish D: Interleukin 6 is expressed in high levels in psoriatic skin and stimulates proliferation of cultured human keratinocytes. Proc Natl Acad Sci USA 86:6367–6371, 1989 Hartig R, Huang Y, Janetzko A, Shoeman R, Grub S, Traub P: Binding of fluorescence and gold-labelled oligonucleotides to cytoplasmic intermediate filaments in epithelial and fibroblast cells. Exp Cell Res 233:169–186, 1997 Hatzfeld M, Franke WW: Pair formation and promiscuity of cytokeratins: formation in vitro of heterotypic complexes and intermediate-sized filaments by homologous and heterologous recombinations of purified polypeptides. J Cell Biol 101:1826–1841, 1985 Hatzfeld M, Weber K: The coiled coil of in vitro assembled keratin filaments is a heterodimer of type I and II keratins: use of site-specific mutagenesis and recombinant protein expression. J Cell Biol 110:1199–1210, 1990 Hatzfeld M, Weber K: Modulation of keratin intermediate filament assembly by single amino acid exchanges in the consensus sequence at the C-terminal end of the rod domain. J Cell Sci 99:351–362, 1991 Heid HW, Moll I, Franke WW: Pattern of expression of trichocytic and epithelial cytokeratins in mammalian tissues. Differentiation 37:137–157, 1988 Hofman I, Franke WW: Heterotypic interactions and filament assembly of type I and type II cytokeratins in vitro: viscometry and determination of relative affinities. Eur J Cell Biol 72:122–132, 1997 Hotchin NA, Kovach NL, Watt FM: Functional down-regulation of alpha 5 beta 1 integrin in keratinocytes is irreversible but commitment to terminal differentiation is not. J Cell Sci 106:1131–1138, 1993 Hudson DL, Weiland KL, Dooley TP, Simon M, Watt FM: Characterisation of eight monoclonal antibodies to involucrin. Hybridoma 11:367–379, 1992 Jiang C-K, Magnaldo T, Ohtsuki M, Freedberg IM, Bernerd F, Blumenberg M: Epidermal growth factor and transforming growth factor α specifically induce the activation- and hyperproliferation- associated keratins 6 and 16. Proc Natl Acad Sci USA 90:6786–6790, 1993 Jonkman MF, Heeres K, Pas HH, et al: Effects of keratin 14 ablation on the clinical and cellular phenotype in a kindred with recessive epidermolysis bullosa simplex. J Invest Dermatol 107:764–769, 1996 Komine M, Freedberg IM, Blumenberg M: Regulation of epidermal expression of keratin K17 in inflammatory skin diseases. J Invest Dermatol 107:569–575, 1996 Kuhn S, Vorgias CE, Traub P: Interaction in vitro of non-epithelial intermediate filament proteins with supercoil plasmid DNA. J Cell Sci 87:453–554, 1987 Kupper TS: Immune and Inflammatory processes in cutaneous tissues. J Clin Invest 86:1783–1789, 1990 Lane EB, Bartek J, Purkis PE, Leigh IM: Keratin antigen in differentiating skin. Ann NY Acad Sci 455:241–258, 1985 Lane EB, Wilson CA, Hughes BR, Leigh IM: Stem cells in hair follicles: cytoskeletal studies. Ann NY Acad Sci 642:197–213, 1991 Leigh IM, Purkis PE, Whitehead P, Lane EB: Monospecific monoclonal antibodies to keratin 1 carboxy terminal (synthetic peptide) and to keratin 10 as marker of epidermal differentiation. Br J Dermatol 129:110–119, 1993 Leube RE, Bader BL, Bosch FX, Zimbelmann R, Achtstaetter T, Franke WW: Molecular characterization and expression of the stratification-related cytokeratins 4 and 15. J Cell Biol 106:1249–1261, 1988 Lloyd C, Yu QC, Cheng J, Turksen K, Degenstein L, Hutton E, Fuchs E: The basal keratin network of stratified squamous epithelia: defining K15 function in the absence of K14. J Cell Biol 129:1329–1344, 1995 Lu X, Lane EB: Retrovirus-mediated transgenic keratin expression in cultured fibroblasts: specific domain functions in keratin stabilisation and filament formation. Cell 62:681–696, 1990 Machesney M, Tidman N, Waseem A, Kirby L, Leigh I: Activated keratinocytes in the epidermis of hypertrophic scars. Am J Pathol 152:1133–1141, 1998 Mansbridge JN, Knapp AM: Changes in keratinocyte maturation during wound healing. J Invest Dermatol 89:253–263, 1987

KERATIN 15 IN STRATIFIED EPITHELIA

369

Marchuk D, McCrohon S, Fuchs E: Complete sequence of a gene encoding a human type I keratin: sequence homologous to enhancer elements in the regulatory region of the gene. Proc Natl Acad Sci USA 82:1609–1613, 1985 Marchuk D, McCrohon S, Fuchs E: Remarkable conservation of structure among intermediate filament genes. Cell 39:491–498, 1984 Moll R, Franke WW, Schiller DL, Geiger B, Krepler R: The catalog of human cytokeratins: patterns of expression in normal epithelia, tumors and cultured cells. Cell 31:11–24, 1982 Morgan PR, Leigh IM, Purkis PE, Gardner ID, Van Muijen GNP, Lane EB: Site variation of keratin expression in human oral epithelia – an immunocytochemical study of individual keratins. Epithelia 1:31–43, 1987 Nelson WG, Sun TT: The 50- and 58-kdalton keratin classes as molecular markers for statified squamous epithelia: cell culture studies. J Cell Biol 97:244–251, 1983 Paladini RD, Takahashi K, Bravo NS, Coulombe PA: Onset of re-epithelialization of keratin filaments in wound edge keratinocytes: defining a potential role for keratin 16. J Cell Biol 132:381–397, 1996 Purkis PE, Steel JB, Mackenzie IC, Nathrath WBJ, Leigh IM, Lane EB: Antibody markers of basal cells in complex epithelia. J Cell Sci 976:39–50, 1990 Quinlan RA, Schiller DL, Hatzfeld M, Achtstatter T, Moll R, Jorcano JL, Magin TM, Franke WW: Patterns of expression and organisation of cytokeratin intermediate filaments. Ann NY Acad Sci 455:282–306, 1985 Reichert U, Michel S, Schmidt R. The cornified envelope: a key structure of terminally differentiating keratinocytes. In: Darmon M, Blumenberg M (eds). Molecular Biology of the Skin: the Keratinocytes. San Diego: Academic Press, 1993, pp. 107–149 Ristow H-J: A major factor contributing to epidermal proliferation in inflammatory skin diseases appears to be interleukin 1 or a related protein. Proc Natl Acad Sci USA 84:1940–1944, 1987 Rugg EL, McLean WHM, Lane EB, et al: A functional ‘‘knockout’’ of human keratin 14. Genes Develop 8:2563–2573, 1994 Sawaf MH, Ouhayoun JP, Forest N: Cytokeratin profiles in oral epithelial: a review and a new classification. J Biol Buccale 19:187–198, 1991 Simon M. The epidermal cornified envelope and its precursors. In: Leigh IM, Lane EB, Watt FM (eds). The Keratinocyte Handbook. Cambridge: Cambridge University Press, 1994, pp. 275–292 Stark HJ, Breitkreutz D, Limat A, Bowden P, Fusenig NE: Keratins of the human hair follicle: ‘‘hyperproliferative’’ keratins consistently expressed in outer root sheath cells in vivo and in vitro. Differentiation 35:236–248, 1987 Steinert PM: The two-chain coiled-coil molecule of native epidermal keratin intermediate filaments is a type I-type II heterodimer. J Biol Chem 265:8766– 8774, 1990 Stewart M: Intermediate filament structure and assembly. Curr Opi Cell Biol 5:3– 11, 1993 Stoler A, Kopan R, Duvic M, Fuchs E: Use of monospecific antisera and cRNA probes to localise the major changes in keratin expression during normal and abnormal epidermal differentiation. J Cell Biol 107:427–446, 1988 Thacher SM, Rice RH: Keratinocyte-specific transglutaminase of cultured human epidermal cells: relation to cross-linked envelope formation and terminal differentiation. Cell 40:685–695, 1985 Todd C, Hewitt SD, Kempenaar J, Noz K, Thody AJ, Ponec M: Co-culture of human melanocytes and keratinocytes in a skin equivalent model: effect of ultraviolet radiation. Arch Dermatol Res 285:455–459, 1993 Traub P, Shoeman RL: Intermediate filament proteins: cytoskeletal elements with gene regulatory function. Int Rev Cytol 154:1–103, 1994a Traub P, Shoeman RL: Intermediate filament and related proteins: potential activator of nucleosomes during transcriptional initiation and elongation. Bioessays 16:349–355, 1994b Tyner AL, Fuchs E: Evidence for post-transcriptional regulation of the keratins expressed during hyperproliferation and malignant transformation in human epidermis. J Cell Biol 103:1945–1955, 1986 Van Muijen GNP, Warnaar SO, Ponec M: Differentiation-related changes of cytokeratin expression in cultured keratinocytes and fetal, newborn, and adult epidermis. Exp Cell Res 171:331–345, 1987 Wang X, Tolestonog G, Shoeman RL, Traub P: Selective binding of specific mouse genomic DNA fragments by mouse vimentin filaments in vitro. DNA Cell Biol 15:209–225, 1996 Waseem A, Alam Y, Dogan B, White KN, Leigh IM, Waseem NH: Isolation, sequence and expression of the gene encoding human keratin 13. Gene 215:269–279, 1998 Waseem A, Alexander CM, Steel JB, Lane EB: Embryonic simple epithelial keratins 8 and 18: chromosomal location emphasizes difference from other keratin pairs. New Biol 2:464–478, 1990 Waseem NH, Lane DP: Monoclonal antibody analysis of the proliferating cell nuclear antigen (PCNA): structural conservation and the detection of a nucleolar form. J Cell Sci 96:121–129, 1990 Waseem A, Lane EB, Harrison D, Waseem N: A keratin antibody recognizing a heterotypic complex: epitope mapping to complementary locations on both components of the complex. Exp Cell Res 223:203–214, 1996 Weiss RA, Eichner R, Sun T-T: Monoclonal antibody analysis of keratin expression in epidermal diseases: a 48- and 56-kdalton keratin as molecular markers for hyperproliferative keratinocytes. J Cell Biol 98:1397–1406, 1984 Yuspa SH, Kilkenny AE, Steinert PM, Roop DR: Expression of murine epidermal differentiation markers is tightly regulated by restricted extracellular calcium concentration in vitro. J Cell Biol 109:1207–1217, 1989