the 01 vol. 268, no. 25. issue of september 5, pp. 19143 ... · the journal 01 biological chemistry...

THE JOURNAL 01 BIOLOGICAL CHEMISTRY (D 1993 by The Amerim Society for Biochemistry and Molecular Biology, lnc.

Vol. 268, No. 25. Issue of September 5, pp. 19143-19161,1993 Printed in U S A .

Interleukin- 18 and Transforming Growth Factor-&pidermal Growth Factor Induce Expression of 2Mr 95,000 Type IV Collagenase/ Gelatinase and Interstitial Fibroblast-type Collagenase by Rat Mucosal Keratinocytes”

(Received for publication, November 17, 1992, and in revised form, March 24, 1993)

J. Guy LyonsSO1, Bente Birkedal-HansenJI , Milton C. Pierson$%, John M. Whitelock%**, and Henning Birkedal-HansenSQSS From the $Department of Oral Bwlogy, the BResearch Center in Oral Bwlogy, and the ((Department of Diugnostit Sciences, University of Alabama School of Dentistry, University of Alabama, Birmingham, Alabama 35294

Rat mucosal keratinocytes serially propagated under permanently serum-free conditions responded to interleukin (IL)-l@/IL-a and to transforming growth factor (TGF)-a/epidermal growth factor (EGF) (as well as to 12-O-tetradecanoylphorbol-13-acetate (TPA)) by up- regulation of M, 96,000 gelatinase (MMP-O)(M, 96K GL) and fibroblast-type collagenase (MMP-1) (FIB- CL), whereas control cells expressed barely detectable levels of either of these enzymes. The cells secreted 8- 10 pg/106 cells/day (M, 96K GL) and 2-3 pg/106 cells/ day (FIB-CL) of enzyme protein for at least 24 h when maximally induced. This level was attained only after a 24-h lag period, and the earliest emergence of enzyme protein in the culture medium required 10-14 h. IL- 1s was by far the most potent cytokine with maximal effect already at lo-’’ M, whereas IL-la, TGF-a, and EGF required 20-100-fold higher concentrations. Pre- treatment of the cells with TPA (lo” M) abolished the subsequent response to IL-lb, TGF-a, and EGF and at the same time resulted in >90% reduction of cytosolic protein kinase C activity. Surprisingly, staurosporine, a potent kinase inhibitor, not only failed to block growth factor/cytokine responses but itself stimulated expression of the enzymes at a magnitude comparable to TPA. The inducing effect of TGF-amGF was down- regulated by 7046% by 10“ M dexamethasone. Dex- amethasone was less effective in ablating the IL-1@ response yielding 60% reduction M. 96K GL and little or no reduction of FIB-CL. Dexamethasone also failed to block the TPA response.

A body of evidence suggests that remodeling of the extracellular matrix is mediated, at least in part, by the collective action of matrix metalloproteinases, a family of nine or more growth factor- and cytokine-regulated, Zn2+-dependent en- dopeptidases that are capable of cleaving most, if not all,

* This work was supported by National Institutes of Health Grants DE06028, DE08228, and DE10631. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduerti-sement” in accord- ance with 18 U.S.C. Section 1734 solely to indicate this fact.

1 Current address: Institut de Chimie Biologique, Facult.4 de MB- decine, 11, Rue Humann, 67085 Strasbourg, France.

** Current address: Dept. Pathology, University of Technology, Sydney, Westbourne St., Gore Hill, NSW 2065, Sydney, Australia.

$$ To whom all correspondence should be addressed: Dept. of Oral Biology, University of Alabama School of Dentistry, SDB Box 54, University of Alabama, Birmingham, AL 35294. Tel.: 205-934-6154; Fax: 205-975-6251.

matrix macromolecules (1). Stromal remodeling at the epithelial/mesenchymal interface is an integral part of wound healing and of invasive epithelial growth whether associated with programmed developmental migration into stromal domains during morpho- and organogenesis or with pathological invasive or expansive growth of carcinomas and cysts. Although the specific communications between epithelial and stromal cells that initiate these events are incompletely understood, both stromal and epithelial cells potentially are capable of driving the process either through inductive influences or through endogenous expression of degradative enzymes. Epi- thelial cells elaborate several defined as well as unidentified growth factors and cytokines capable of inducing expression of tissue-degrading enzymes by stromal cells (2-6). The recent discovery of keratinocyte growth factor, a member of the fibroblast growth factor family expressed by stromal cells, which acta only on keratinocytes, suggests that the commu- nication may well proceed also in the other direction (7, 8). The role of epithelial cells in remodeling of stromal/epithelial interfaces was originally thought of as primarily inductive, mediated by release of one or more growth factors and cytokines (2,9,10). A growing body of evidence, however, suggests that keratinocytes (as well as other cells of epithelial lineage) express several members of the matrix metalloproteinase gene family either constitutively or in response to growth factors/ cytokines or to phorbolesters (11-15). We have observed that rat epithelial cells (mucosal keratinocytes and mammary carcinoma cells) constitutively express an interstitial, fibroblast- type collagenase (MMP-1) (11) and a M, 95,000 gelatinase/ type IV collagenase (MMP-9) in culture (12). Together these two proteinases are capable of degrading a wide range of matrix proteins responsible for the stromal architecture of basement membranes and underlying connective tissue. In this study we have studied growth factors and cytokines that regulate the expression of these two collagen-cleaving proteinases by rat mucosal keratinocytes. Our findings show that the proinflammatory cytokine, IL’-l& and the two epithelial

’ The abbreviations used are: IL, interleukin; ABTS, diammonium 2,2’-azinodi-(3-ethylbenzthiazoline sulfonate); DMEM, Dulbecco’s modified Eagle’s medium; EGF, epidermal growth factor; FIB-CL, fibroblast-type collagenase, MMP-1; IFN, interferon; mAb, monoclonal antibody; mAb,, W,K oJD4 previously identified as mAb GeBC1-ID4 (12); ~ A ~ . c L I I D ~ previously identified as mAb CoBC1- IIDl (12); MMP, matrix metalloproteinase; M, 72K GL, M, 72,000 gelatinase, MMP-8; M. 95K GL, M, 95,000 gelatinase, MMP-9; PBS, phosphate-buffered saline; TGF, transforming growth factor; TNF, tumor necrosis factor; TPA, 12-O-tetradecanoylphorbol-13-acetate; ELISA, enzyme-linked immunosorbent assay.

19143

19144 Regulation of Keratinocyte Matrix Metalloproteinme Expression

growth factors TGF-a and EGF, are potent inducers of M, 95K GL and FIB-CL in keratinocytes.

EXPERIMENTAL PROCEDURES

Materials-Culture media were from Flow (McClean, VA) or Me- diatech (Washington, D. C.). Hews was from Research Organics (Cleveland, OH). SDS-polyacrylamide gel electrophoresis Mr mark- ers, heparin-Sepharose, and CNBr-Sepharose 4B were from Phar- macia LKB Biotechnology Inc. Labtek chamber slides were from Miles Laboratories. Dimethyl sulfoxide was from Aldrich. Coomassie Brilliant Blue G-250 was from Kodak. Electrophoresis chemicals were from Bio-Rad. Immunological reagents were from Southern Biotech- nology (Birmingham, AL). Recombinant human interleukin-la (IL- la) and -18 (IL-lB), recombinant human interferon-7 (IFN-r), and recombinant human tumor necrosis factor-a (TNF-a) were from Genzyme (Boston, MA). Natural murine epidermal growth factor (EGF) was from Collaborative Research (Bedford, MA). Recombinant human platelet-derived growth factor, recombinant human insulin- like growth factors-I and -11, recombinant human basic fibroblast growth factor, and recombinant human transforming growth factor- a (TGF-a) were from Bachem (Torrance, CA). Recombinant human transforming growth factor-81 (TGF-01) was from R & D Systems Inc. (Minneapolis, MN). Triton X-100, Tween 20, dexamethasone, bovine serum albumin, 12-0-tetradecanoylphorbol acetate (TPA), EGTA, EDTA, phenylmethylsulfonyl fluoride, leupeptin, staurosporine, and diammonium 2,2'-azinodi-(3-ethylbenzthiazoline sulfonate) (ABTS) were from Sigma. All other reagents were highest available grade from Sigma.

Cell Culture-A clonal rat keratinocyte cell line (CCL-4) was established from normal rat tongue mucosa as described previously (11, 16) and cultured under defined, permanently serum-free conditions without antibiotics. The medium consisted of a 1:l mixture of Ham's F-12 and Dulbecco's modified Eagle's medium (DMEM) supplemented with insulin (5 pglml), transferrin (1 pg/ml), bovine serum albumin (5 mg/ml), and Hepes (10 m~)(17) . In growth factor/cytokine experiments, the cells were seeded at a density of 25,000 cells/ cm2 (10,000 cells/well) in 96-well plates (0.19 ml/well) allowing 2 days in culture for attachment and growth to subconfluence. One-hundred pl of medium was then replaced with fresh medium containing the factor of interest. TPA, staurosporine, and dexamethasone were added to the medium from lo4 X or lo3 X stock solutions in dimethyl sulfoxide. Concentrations of dimethyl sulfoxide up to 0.1% had no effect on production of M, 95K GL or FIB-CL or on cell morphology.

Early passage rat mucosal keratinocytes were initially established from tongue explants in DMEM, 20% fetal bovine serum as described (11, 16) and then subcultured into Life Technology, Inc. low CaZ+ (0.09 mM) keratinocyte medium supplemented with EGF (5 ng/ml) and bovine pituitary extract (35-50 pglml). When grown to conflu- ence, second passage cells were subcultured into chamber slides, exposed to control or stimulating agents in keratinocyte growth medium as described below, and processed for immunofluorescence.

Isolation of M, 95K GL and FIB-CL-M. 95K GL and FIB-CL were purified from rat mammary carcinoma cells (BC-1) as described previously (12, 18). Conditioned media obtained from serum-free cultures was first passed over a heparin-Sepharose column. M, 95K GL that was not retained on the column was subsequently purified by a single passage over gelatin-Sepharose. The enzyme was retained on this column and eluted with 7.5% dimethyl sulfoxide. FIB-CL was eluted from the heparin-Sepharose column and further purified by Znz+ chelate and molecular sieve chromatography using AcA 44 as described (19). Protein concentrations were determined spectropho- tometrically using Am = 1.29 for both enzymes (12, 19).

Two-site ELZSA-Monoclonal antibodies (mAbs) were produced to M, 95K GL and FIB-CL using techniques describedpreviously (12, 19). Briefly, regional lymph node cells obtained from mice immunized with five injections of 40-80 pg enzyme protein/mouse over a 15-day period were fused with P3-X63Ag8.653 murine myeloma cells. Hy- bridomas that were positive in an ELISA were cloned and recloned by limiting dilution. Isotyping was done by ELISA with isotype- specific second antibodies as described (19). Antibodies were purified from ascites obtained from pristane-primed mice by sequential pre- cipitation with 2.5% caprylic acid and 45% saturated (NH4)S04 (20). Polyclonal antibodies to rat FIB-CL were produced essentially as described (19). Three-kg New Zealand White rabbits were injected at several intradermal sites with 75 pg/kg antigen in Freund's complete adjuvant after collection of preimmune serum. Booster immuniza- tions were conducted every 2 weeks, the first using antigen in Freund's

incomplete adjuvant and subsequent injections using antigen in PBS. The animals were bled (40-50 ml) every 2 weeks starting 1 week after the third injection. Immunoglobulin was isolated by protein A-Seph- arose chromatography. Bound immunoglobulin was eluted with 50 mM sodium acetate buffer, pH 4.0. Two-site ELISAs were performed in 96-well microtitration plates. In preliminary experiments a wide range of catcher/detector combinations for two-site ELISAs were tested as were blocking procedures (non-fat powdered milk or bovine serum albumin supplemented with either Tween 20 or Triton X-100) and varying incubation times and temperatures. Most experiments were performed as described below. For quantification of M, 95K GL, wells were coated with mAb~rM,6~c~VIIC2 for 6 h at 22 "C with 100 pl of a 3 pg/ml solution of purified IgG in sodium borate buffer, pH 8.5, then blocked with borate saline containing 3% bovine serum albumin and 2% Tween 20 and washed twice. The coated wells were incubated overnight at 22 'C with 50 pl of appropriate dilutions of culture media samples or standards in the same buffer followed by three buffer washes. The wells were then incubated with biotinylated mAbMr 9 6 ~ cJD4 (100 p1 of 10 pg/ml in borate saline buffer) for 4 h at 22 "C and washed three times. Biotinylation was with N-hydrox- ysuccinimide biotin in 0.1 M sodium borate buffer, pH 8.8, using 25 pg of ester/mg of mAb for 4 h at 22 "C essentially as described (21). Incubation with the detector antibody was followed by incubation with 50 81 of 1 pg/ml streptavidin-horseradish peroxidase in borate saline buffer for 2 h at 22 "C. The plates were then washed twice with PBS and once with citrate buffer, pH 4.0, and developed by incubation at room temperature for 8-20 min with 0.03% HzOa in citrate buffer containing 0.3 mg/ml ABTS. For quantification of rat FIB-CL, plates were coated with 3 pg/mlmAbpm.cJnC611D2 for 6 h at 22 "C. Detector antibody was 10 pg/ml protein A purified, biotinylated rabbit Ig raised against rat FIB-CL. Incubation was for 4 h at 22 "C.

Other Methods-Cells grown in plastic chamber slides were fixed for 15 min with 3.7% formaldehyde in PBS and permeabilized with 1% Triton X-100 in Tris-buffered saline. Nonspecific IgG binding was blocked by incubation with a mouse mAb specific for shark immunoglobulins (Shark 46; 100 pg/ml). The cells were then stained with bictinylated mAbMr WK OLVIIC~ (50 pg/ml) or ~ A ~ . c L V I I I C ~ . Staining with the first antibody (50 pg/ml immunoglobulin in PBS containing 1% bovine serum albumin) was for 90 min at 22 "C. Development with Texas Red-conjugated streptavidin (1 pg/ml in PBS containing 1% bovine serum albumin) was for 1 h at 22 "C. The slides were mounted in PBS. Gelatin substrate zymography was performed in 7.5% polyacrylamide gels copolymerized with 1 mg/ml gelatin as described (12). Western blots prepared by transfer of electrophoretograms of crude culture medium to nitrocellulose were stained as described (12,19) using a two-layer technique with biotinylated mAb as the first layer and horseradish peroxidase-conjugated streptavidin as the second. Color development was with diaminobenzidine/HzOz/NiClz. For determination of protein kinase C activity, companion cultures were seeded in 10-cm dishes at a density of 25,000 cells/cmz and were permitted to attach and grow to subconfluence for 48 h. The medium was then replaced with fresh medium either with or without 10" M TPA and incubated for another 24 h. The cells were lysed in 5 ml/dish of 1% Triton X-100 in 20 mM Tris-HC1, pH 7.5, 0.5 mM EGTA, 2 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride, 0.1% 2-mercaptoethanol, 10 pM leupeptin. The cells were scraped off the dish, vortexed, and extracted for another 1 h. The extract was then centrifuged at 100,000 X g for 1 h and Ca*+/ phosphatidylserine-dependent protein kinase (protein kinase C) activity in 200-pl aliquots of cytosol was measured by quantification of transfer of 32P radioactivity from [ya2P]ATP to type 111-S histone as described in (22).

RESULTS

Development of n o - s i t e ELISAs for Rat M, 95K GL and FIB-CL-Regulation of metalloproteinase gene expression involves both transcriptional and posttranscriptional events. Since any biologic effects are mediated by the secreted, func- tional gene product, we decided to study growth factor and cytokine regulation of keratinocyte metalloproteinases at the terminal level, i.e. by the emergence of enzyme protein in the extracellular milieu. To this end, the rat homologues of the human M, 92K GL ((Mr 95K in the rat (12)) and FIB-CL were isolated using as source a rat mammary carcinoma cell line (BC-1) which maintains high level constitutive expression

Regulation of Keratinocyte Matrix Metalloproteinuse Expression 19145

2.0

1.6

E 1.2 d d 7

n 0 0.8

0.4 0.2

ENZYME CONCENTRATION (pg/rnl) FIG. 1. Two-site ELISA for quantification of rat M, 96K

GL and FIB-CL. For the M, 95K GL assay, microtiter plates were coated for 6 h a t 22 "C with 3.0 pg/ml mAbM, SSK CLVIIC~ in borate buffer, pH 8.5, and then blocked with borate saline containing 3% bovine serum albumin and 2% Tween 20. The wells were incubated overnight at 22 "C with serial dilutions of purified rat M, 95K GL or FIB-CL ranging from =3 pg/ml to 2 pg/ml in a volume of 100 pl. After washing, the wells were incubated with 100 pl (10 pg/ml) biotinylated dbMrmg6K cJD4 in borate saline buffer for 4 h at 22 "C, then washed and incubated with horse radish peroxidase-conjugated streptavidin (1:1,000) for 2 h at 22 "C and color developed with 0.3 mg/ml ABTS, 0.03% HZ02 in citrate buffer, pH 4.0. The reaction was monitored at 414 nM. The FIB-CL assay was essentially as described for M, 95K GL except that coating was with 100 pl 3.0 pg/ml rnAb~.cJnICGIID2. Detector antibody was 10 pg/ml biotinylated rabbit IgG raised against rat FIB-CL and purified by protein A- Sepharose chromatography. Inset, M, 95K GL and FIB-CL were purified from cultures of BC-1 rat mammary carcinoma cells as described previously (12,18,54). The samples were resolved by SDS- polyacrylamide gel electrophoresis using a 10% gel and stained with Coomassie Blue. Lune 1, M, 95K GL, 1.8 pg; hne 2, FIB-CL, 2.7 pg.

of both enzymes (12,18) (Fig. 1, inset). To permit quantitation of enzyme protein, mAbs were raised against both enzymes and two-site ELISAs developed (Fig. 1). Rat M, 95K GL was quantified by a mAb/mAb catcher/detector combination (mAbM, 95K ~ ~ v I I c 2 / b i o t i n - m A b ~ ~ 95,K ~ L I I D ~ ) with a useful range of 3-300 ng/ml and a lower detection limit of ~ 3 0 0 pg of enzyme protein. For quantification of rat FIB-CL a catcher mAb (mAbF~,.cJnC611DZ) was used in combination with biotinylated rabbit-anti-rat FIB-CL I&. This assay had a useful range of 1-100 ng/ml and a lower detection limit of =IO0 pg of enzyme protein. At enzyme concentrations up to 3 pg/ ml no cross-reactivity between the two proteins was detected (Fig. 1).

Rat Mucosal Keratinocytes Express M, 95K GL and FIB-CL in Response to IL-10, IL-la, TGF-a, and EGF-We have shown previously that clonal derivatives (CCL-4) of rat mucosal keratinocytes retain typical keratinocyte traits and are capable of orderly passage through a characteristic terminal differentiation repertoire (cornified envelope formation, keratin filament assembly, and degradation of nuclear and cytoplasmic structures) (11,16). When cycled between serum-free and serum-supplemented periods these cells maintain significant "basal" expression of the M, 95K GL and FIB-CL as measured by secretion of enzymes to the culture medium (11, 13).* Since several matrix metalloproteinase genes, and the

* J. G. Lyons and H. Birkedal-Hansen, unpublished data.

transcription factor genes that regulate their expression (c- fos), contain serum response elements (23, 24) we sought to reduce basal expression of the enzymes by serially propagating the cells in a defined, serum-free medium supplemented with insulin and transferrin (17). This medium permitted infinite serial propagation of rat mucosal keratinocytes under high Ca" conditions and lowered basal expression of both enzymes to near the detection limit of the assays.

To monitor the response of rat mucosal keratinocytes to growth factors and cytokines, cells were seeded at a density of 25,000/cm2, allowed to attach and grow for 48 h, and then exposed to growth factors/cytokines for 24 h. Under these conditions, EGF (5 X lo-' M), TGF-a (5 X lo-' M), and IL- 10 (lo-'' M), as well as TPA (lo" M), greatly stimulated expression of M, 95K GL and FIB-CL (Fig. 2). M, 95K GL, identified by staining with mAbM, 9 6 ~ G L I D ~ (12), emerged in the culture medium in both monomer and dimer forms, and both were recognized by the antibody (Fig. 2B). The dimer did not dissociate in 1% SDS under nondenaturing conditions but was converted to catalytically active form by exposure to this detergent (Fig. 2A) . FIB-CL identified by staining with mAbFE,-CLIIDl (12) showed less tendency to dimerize and emerged predominantly in monomeric form (Fig. 2C).

To compare the potency of various growth factors and cytokines, cells were exposed to a broad concentration range of stimulants for 20 h and medium concentrations of M, 95K GL and FIB-CL determined by two-site ELISAs. IL-10 was the most potent cytokine with maximal effect at =lo-" M and detectable effect down into the 10-"-10-'2 M range (Fig. 3, upper panels). IL-la required approximately 100-fold higher concentration. Under the same conditions, TNF-a, which induces high level expression of FIB-CL in human skin

- 205k -116k - 84k

- 4%

- 33k

- B

1 2 3 4 5 6 7 0 9 10 11 12 13 14 15 16 17 18 19 21

FIG. 2. Up-regulation of M, 96K GL and FIB-CL by growth factors and cytokines. Rat mucosal keratinocytes were seeded at 25,000 cells/cm2 and allowed to attach and grow to subconfluence for two days. The medium was then replaced by albumin-free medium supplemented with TPA M), EGF (5 X lo-' M), TGF-a (5 X lo-' M), IL-10 (10"O M), staurosporine M), or TGF-a + dexamethasone (lo-' M). The medium was harvested after 24 h and analyzed by gelatin zymography and Western analysis. Panel A, gelatin zymogram. Conditioned culture medium (40 pl) was incubated a t room temperature for 1 h with electrophoretic sample containing 2% SDS under nonreducing conditions and then resolved by SDS- polyacrylamide gel electrophoresis using a 7.5% gel copolymerized with 1 mg/ml gelatin. The gel was washed three times in 1% Triton X-100 in 50 mM Tris-HC1, pH 7.5, 0.2 M NaCl, 5 mM CaC12, then incubated a t 37 "C for 2 h and stained with Coomassie Blue. B, Westem blot. 500-pl aliquota of the samples described in panel A were mixed with 25 mM EDTA, 0.05% Brij 35 and concentrated by a Centricon 10 device and resolved by electrophoresis as described above. The electrophoretograms were transferred to nitrocellulose and stained with biotinylated d b ~ ~ ~ ~ g 6 ~ ~ J D 4 against rat M, 95K GL (1 pg/ml) followed by horseradish peroxidase-conjugated streptavidin. Color was developed using the diaminobenzidine/H202/NiCl~. Panel C, Western blot. The samples described in panel A were analyzed by staining with biotinylated mAbFIB.CLIID1 against rat interstitial collagenase (1 pglml).

19146 Regulation of Keratinocyte Matrix Metalloproteinase Expression

r % - w 4 0 0

900

200

100

-13 -12 -11 -10 -0 d -7

MrSKQL R

3

-13 -12 -11 -10 -e d -7

FIECL 3 3 40=

30

20

10

-10 -0 d -7 d 4 -10 a d -7 a 6

CONCENTRATION (LOGM)

FIG. 3. Doee response of growth factor/cytokine up-regulation of M, 9BK GL and FIB-CL expression. Rat mucosal keratinocytes were seeded in microtiter plates (25,000 cells/cm*) and allowed to attach and grow for 2 days. The medium was then replaced with fresh, albumin-free medium containing the stimulatory com- pounds, and the cultures were incubated for 20 h. The resultant concentrations of M, 95K GL and FIB-CL were determined in the same media samples by two-site ELISAs as described in Fig. 1. Panel A, IL-lP, IL-la, and TNF-a. Panel B, TGF-a and EGF. Panel C, TPA and staurosporine. Each point and bar represent the mean and standard deviation of quadruplicate wells.

and gingival keratinocytes below lo-' M3 was only weakly inductive. TGF-a and EGF, which bind with equal affinity to a common receptor (25), induced expression of both enzymes but required 20-50-fold higher concentrations than IL-18 (2- 5 X lo4 M) (Fig. 3, rniddlepanels). TGF-8, which up-regulates expression of M, 92K GL in human gingival keratinocytes (13), as well as basic fibroblast growth factor, platelet-derived growth factor, and insulin-like growth factors4 and -11 were ineffective over a wide concentration range (10-"-10-7 data not shown).

Previous studies have shown that metalloproteinase transcripts are detectable 3-8 h after the addition of stimulant (26-28) and that transcript levels continue to rise for up to 72 h in response to IL-lD and TNF-a (29). To follow the kinetics of the response at the gene product level, and to measure the magnitude of the response as well, rat mucosal keratinocytes were seeded as described above, and the emergence of immunoreactive enzyme protein following the addition of IL-18 (10"O M), EGF (5 X lo-' M) and TGF-a (5 X 10"' M) was monitored for up to 52 h (Fig. 4). The earliest emergence of enzyme protein was characteristically preceded by a 10-14-h lag period (Fig. 4, insets), and attainment of steady-state secretion rates required approximately 24 h (Fig. 4). During the ensuing linear phase (24-52 h) the cells secreted 8-10 pg/lO%ella/day of M, 95K GL and 2-3 pg/lO%ells/day of FIB-CL. ~~ ~ ~~ ~ ~~ ~ ~

H. Y. Lin and H. Birkedal-Hansen, unpublished data.

Depletion of Cytosolic Protein Kinuse C by TPA Abrogates Growth Factor Responses-The promoter regions of the human FIB-CL gene contain one (30,31), and the M, 92K GL gene two (28), TPA-responsive, AP-1 binding sequences. The AP-1 site is a necessary albeit not sufficient requirement for TPA induction of metalloproteinase gene expression (28,30- 32). This regulatory element appears to be functionally con- served in the rat genes as well since TPA dramatically stimulated expression of M, 95K GL and FIB-CL in the M- lo4 M range (Fig. 2 and Fig. 3, bottom panels). A body of evidence suggests that the AP-1 site is also a necessary requirement for the transcriptional effects of IL-18 and EGF on metalloproteinase gene expression (27, 33, 34). Since the intracellular signaling elicited by these growth factors and cytokines is mediated, at least in part, through protein kinase C-dependent pathways (27, 32, 35-37) and since protein kinase C is activated directly by phorbol esters (38, 39), we examined the effect on metalloproteinase gene expression of pretreatment with TPA which depletes cytosolic protein kinase C and leaves the cells refractory to stimulation through protein kinase C-dependent pathway(s) (40). Exposure to M TPA for 24 h resulted in 90% reduction of cytosolic protein kinase C activity (Fig. 5A) and in 70-95% down-regulation of TGF-cr/EGF, IL-18, and TPA responses for both enzymes (Fig. 5, B and C ) . Attempts to block growth factor induction by staurosporine, a potent protein kinase inhibitor (41, 42), were not successful since staurosporine itself stimulated expression of the enzymes, particularly M, 95K GL, in a relatively narrow concentration range around 10" M (Fig. 2 and Fig. 3, bottom panels). Concentrations of staurosporine which blocked metalloproteinase induction (85 X 10" M) invariably led to cell detachment and cell death. Fig. 5, B and C, also shows that, in contrast to the growth factor and cytokine responses which continued unabated for at least 52 h, the inductive effect of TPA itself was transient, relatively short-lived, and all but played out within the first 24 h.

Dexamethasone Inhibits Growth Factor/Cytokine Induction of Keratinocyte M, 95K GL and FIB-CL-Glucocorticoids repress matrix metalloproteinase gene expression and are thought to mediate anti-inflammatory activity, in part, by down-regulation of these enzymes. In rabbit and human synovial and skin fibroblasts, glucocorticoids reduced IL-1 and EGF responses measured by stromelysin-1 and FIB-CL transcript or medium protein levels by as much as 80-90% (26, 43-45). A sequence partially or completely overlapping the AP-1 site appears to be a target for glucocorticoid repression as the glucocorticoid receptor competes with c-Fos/c-Jun for binding to these overlapping sequences (31,46-48). To deter- mine the effect of glucocorticoids on keratinocyte metalloproteinase expression, subconfluent cultures were exposed to inducing cytokines/growth factors in the presence of increasing concentrations of dexamethasone. TGF-a/EGF stimulation of both FIB-CL and M, 95K GL was greatly inhibited by dexamethasone in a dose-dependent manner (Figs. 1 and 6). Maximal inhibition (75-90%) was achieved at 10-8-10-7 M. Dexamethasone also repressed the effect of IL-18 on M, 95K GL, although less completely (60%), but was ineffective in suppressing FIB-CL expression in the same cultures. Lnter- estingly, dexamethasone was also ineffective in down-regulating the response to TPA. Additional time course experiments showed that the stimulatory effect of TPA starts several hours earlier (6-8 h) than the first noticeable effect of dexamethasone (12 h) (data not shown). We therefore speculate that TPA may already have set in motion a transcriptional/posttranscriptional program that can no longer be aborted by glucocorticoids.

Regulation of Keratirwcyte Matrix Metalloproteinase Expression

Mr 95K GELATINASE COLLAGENASE

19147

0

5 = 1,000 z 0 F

800

2 I-

w 0 600 z 0 0

400

I s z w 200

100

400

m 350 5

E 0 0

250 z 0 m

200 "I z ;D

150 3 z

300 m

e 100 s

'9, 3

50

FIG. 4. Kinetics of M. 9SK GL and FIB-CL secretion in response to growth factors and cytokines. Keratinocytes were seeded as described in Fig. 2 and then exposed to IL-1@ (10"' M), TGF-a (5 X 10- M), EGF (5 X 10- M), or to medium alone (CON). Media aliquots were withdrawn at increasing time intervals from 2 to 52 h after the addition of stimulant and the concentration of M. 95K GL and FIB-CL in quadruplicate wells determined by two-site ELISA as described in Fig. 1. Insets show the emergence of M, 95K GL or FIB-CL during the first 20 h.

CON TPA

1 B Mr95KGL

J CON TPA EGF TGF-CI IL-lp CON TPA EGF TGF-CI IL-1p

FIG. 5. Effect of TPA pretreatment on growth factor/cytokine responses. Subconfluent keratinocyte cultures prepared as described in Fig. 2 were pretreated for 27 h with M TPA and then washed to remove traces of phorbol ester and incubated for another 24 h with medium (DMEM/F-12 containing with insulin, transferrin) supplemented with ECF (5 X lo-' M), TGF-a (5 X lo-' M), IL-18 (10"' M), or TPA M). Panel A , the cytosolic protein kinase C activity of untreated (CON) and TPA-pretreated cultures (TPA) was determined as described by Franklin et al. (22). Panel B, M, 95K GL concentrations were determined by two-site ELISA as described in Fig. 1. Panel C, FIB-CL concentrations were determined by two-site ELISA as described in Fig. 1.

Induction of M, 95K GL and FIB-CL Is Associuted with Morphologic Changes-Induction of metalloproteinase genes in stromal cells is often associated with more extensive shifts in gene expression which result in dramatic changes in cell shape (49). IL-l@, TGF-a, EGF (as well as TPA) had similar, striking effects on the morphology of mucosal keratinocytes (Fig. 7). Unstimulated cells were fully spread with typical epitheloid morphology (Fig. 7A) . Stimulated cells developed more stellate or spindle-shaped appearances with long, thin (stringy) projections across the surface (Fig. 7, C-K) and, particularly in the case of TPA, retraction and cell rounding (Fig. 71). It is of interest to note that staurosporine mimicked the effect of IL-lS and TGF-a/EGF also on cell morphology

(Fig. 7 K ) . Changes of cell shape were accompanied by the emergence of cytoplasmic enzyme protein. Few isolated cells ((5%) expressed the enzymes in unstimulated cultures (Fig. 7B) , but each of the stimulatory agents increased the fraction of immunoreactive cells by >lO-fold and, in most cases, visibly augmented the level of immunoreactivity in individual cells (Fig. 7 ) .

Because of the clonal origin of the cell line described above and because of its ability to propagate serially under permanently serum-free conditions, we asked the question whether the growth factor and cytokine responses of these cells were significantly altered as compared with freshly isolated rat mucosal keratinocytes. Since early passage keratinocytes have

19148 Regulation of Keratinocyte Matrix Metalloproteinase Expression

-10 -9 8 -7 -6 -10 -0 -8 -7 -6

DEXAMETHASONE CONCENTRATION ( l o g [MI)

FIG. 6. Repression of cytokine responses by dexamethasone. Panel A , M, 95K GL. Panel B, FIB-CL. Rat mucosal keratin- ocyt.8~ seeded as described in Fig. 2 were supplemented with EGF (5 X lo-' M), TGF-a (5 X lo-' M), IL-1j3 (10"' M), or TPA (10" M) and with increasing concentrations of dexamethasone. Incubation was for 20 h. Since the various stimulants induce different levels of enzyme secretion, the results are shown relative to induced, dexamethasone- free controls after subtraction of basal expression of uninduced cultures.

Mean and standard deviation of quadruplicate wells are shown.

an absolute requirement for growth factor or serum supple- mentation, it is not possible to establish completely comparable conditions. To address the question, rat mucosal keratinocytes obtained by the explant method using DMEM, 20% fetal bovine serum were subcultured into a low Ca" keratinocyte growth medium supplemented with EGF and bovine pituitary extract and then exposed to IL-1/3 (lo-' M), TGF-a (lo-' M), EGF (lo-' M), and TNF-a (lo-' M) and analyzed by immunofluorescence staining. Our findings showed that third- passage rat mucosal keratinocytes responded to 1L-1/3, TGF- a, EGF (and TPA) by induction of M, 95K GL, whereas the effect of TNF-a was comparatively much weaker. The effect on FIB-CL expression in this cell population was similar in principle but more difficult to ascertain because the "control cells" maintained significant basal expression of this enzyme in culture.

DISCUSSION

The findings of this study show that rat mucosal keratinocytes in culture respond to the proinflammatory cytokines IL- 1/3/IL-la and to the epithelial growth factors EGF and TGF- a by expression of M, 95K GL and FIB-CL as measured by emergence of enzyme protein in the extracellular milieu. IL- 16 was by far the most potent stimulant with maximal induction already at 10"' M. In this capacity it was at least 100- fold more potent than TPA. IL-la and TGF-a/EGF also induced expression of the enzymes but required 20-100-fold higher concentrations (10-8-10-9 M). Since IL-1 concentrations in inflammatory exudates such as rheumatoid synovial fluid and gingival crevicular fluid are frequently in the lo-*- lo-" M range (50-53) our findings suggest that keratinocytes respond to growth factors and cytokines in a physiologically meaningful concentration range. It is of note that emergence of M, 95K GL and FIB-CL in the culture medium is preceded by a considerable lag period. The earliest response was detected after 10-14 h, which is considerably longer than the 3- 8 h required for appearance of transcripts (26-28), and fully 24 h were required for attainment of maximal steady-state secretion rates. During the linear phase, cells induced by IL-

18 and TGF-a/EGF secreted 2-10 pg/106 cells/day of M, 95K GL and FIB-CL, i.e. levels that are comparable to, and per- haps exceed, those attained by stromal cells. By contrast, the effect of TPA (as well as that of staurosporine; data not shown) was transient and rather short lived.

Several studies have suggested that the transcriptional effects of IL-1j3 and EGF (and TGF-a) on metalloproteinase gene expression are mediated by signaling pathways involving protein kinase C (27, 36). TPA pretreatment of rat mucosal keratinocytes lowered cytosolic protein kinase C activity by greater than 90% and all but abolished subsequent responses to IL-lj3 and TGF-a/EGF (as well as to TPA itself). These findings support the notion that these growth factors regulate matrix metalloproteinase expression by rat mucosal keratinocytes through protein kinase C-dependent mechanism(s). Somewhat surprisingly, however, staurosporine, a broad spec- trum protein kinase inhibitor with particular potency against protein kinase C, also induced expression of the enzymes, particularly M, 95K GL. This observation is apparently at variance with findings by Case et al. (36) that 2.5-5.0 X lo-' M staurosporine reduces stromelysin-1 transcript levels in synovial fibroblasts in response to IL-10. A number of factors may possibly account for this discrepancy, as the two studies differ with respect to cell type (keratinocytes uers'sus fibroblasts), method of analysis (medium enzyme concentration uersus steady-state mRNA levels), and gene products analyzed (FIB-CL and M, 95K GL uersw stromelysin-1). Our observation is of particular interest in view of the finding by others that staurosporine possesses tumor promoting activity in mouse skin keratinocytes (54,55), induces expression of urokinase-type plasminogen activator in LLC-PK1 porcine kid- ney epithelial cells (56), and mimics the effect of nerve growth factor in inducing neurite outgrowth in PC12 pheochromo- cytoma cells (57). The induction of degradative enzymes such as matrix metalloproteinases and urokinase-type plasminogen activator are likely to impart aggressiveness to developing tumors and therefore may be directly linked to the tumor promoting activity. It is possible that the staurosporine response is biphasic with inductive effects prevailing at lower concentrations and inhibitory effects at higher, as suggested by other studies (58). Although the staurosporine dose response (Fig. 3) might suggest the existence of such a biphasic response because of the precipitous drop beyond lo-' M, the inherent cytotoxicity of staurosporine, noted previously by others (41), in the case of rat mucosal keratinocytes precluded meaningful analysis at concentrations beyond lo-' M, which coincide with the reported inhibitory range on protein kinases.

The observation that keratinocytes express two matrix metalloproteinases (M, 95 K GL and FIB-CL) with action against stromal and basement membrane components in response to IL-l@/IL-la and TGF-a/EGF is of considerable interest in the context of keratinocyte function in development and in wound healing. Wound healing sites contain several potential sources of IL-1/3 and TGF-a. Wound mac- rophages express each of these cytokines (59), and stimulated keratinocytes are capable of expressing a wide range of growth factors and cytokines including IL-1 (4) and TGF-a (5, 60, 61). A body of evidence suggests that keratinocytes also express the appropriate receptors: basal keratinocytes express EGF receptors in vivo (62,63), and stimulated epithelial cells in culture express receptors for IL-1 as well (64). Studies by Barrandon and Green (65) have shown that human keratinocyte colonies respond to TGF-a by centripetal migration and have suggested that this effect is responsible for epithelial sheet expansion. It is therefore likely that epithelial cell functions in wound healing such as proliferation, migration,

Regulation of Keratinocyte Matrix Metalloproteinase Expression 19149

Control

FIG. 7. Growth factors/cytokines induce morphologic changes and stimulate expression of M, 96K GL. Rat mucosal keratinocytes were seeded in chamber slides and then stimulated for 24 h (TPA, 10 h). The cells were fixed with 3.7% formaldehyde in PBS and permeabilized with 1% Triton X-100. After the blocking of nonspecific binding sites with an irrelevant 100 pg/ml I&, mAb, the cells were stained with 50 pg/ml biotinylated m A b ~ ~ ~ ~ ~ ~ V 1 1 C 2 in PBS containing 1% bovine serum albumin. Stain- ing was developed with 1 pg/ml Texas Red-conjugated streptavidin. Left panels, phase-contrast microscopy; right panels, immunolocalization of MI 95K GL. Panels A and B, control; panels C and D, IL-16 (10"' M); panels E and F, TGF-a (5 X lo-' M); panels G and H, EGF (5 X lo-' M); panels I and J , TPA (lo" M); and panels K and L, staurosporine (lo-' M). Arrowheads point to identical cells in phase-contrast and immunofluorescence microscopy. Magnifi- cation, X 200.

i- - a"-.%,.

I L-1 p

I

TG F-a

I'

Staurosporine

and sheet expansion are regulated by TGF-a. We speculate that epithelial migration a t wound healing sites to cover the denuded connective tissue surface may well be intimately linked to expression of degradative enzymes with activity against stromal macromolecules. Other inflammation-related epithelial proliferative conditions that may require expression of metalloproteinases include the expansive growth of epithelial cysts (for instance radicular cysts surrounding infected teeth) and the "invasive" epithelial apical migration in human periodontal diseases.

It is equally likely that epithelial invasion into mesenchymal domains is dependent on expression of one or more matrix metalloproteinases and that this process is regulated by growth factors and cytokines. A substantial body of evidence suggests that invasiveness of both normal and malignant cells is associated with elevated expression of metalloproteinases

and/or diminished expression of metalloproteinase inhibitors (66-71). Moreover, Turksen et al. (14) recently showed that human keratinocyte raft cultures grown on collagen lattice/ fibroblast matrix respond to TGF-a by induction of FIB-CL throughout the epithelial layer, and of M, 72K GL in the basal layer, and that emergence of these enzymes precedes or coincides with expression of an invasive phenotype. It is often not appreciated that epithelial invasion, although important to tumor progression, is not a tumor-specific trait. Highly invasive but nonmalignant epithelial movements occur during embryonic organo- and morphogenesis and later in life during hair and tooth development, during remodeling of epithelium/ connective tissue interfaces (rete peg formation) and during pocket deepening in progressive human periodontal disease. The observation that normal epithelial cells in culture express FIB-CL (11, 15), MI 72K GL (13), MI 95K GL (13), and

19150 Regulation of Keratinocyte Matr

stromelysin-2' is consistent with the role of epithelial cells also in nonmalignant invasive processes.

At present there is little direct evidence as to the specific role of FIB-CL and M, 95K GL in keratinacyte behavior. The relatively narrow substrate specificity of FIB-CL suggests that this enzyme is involved in the metabolic dissolution of fibrils of type I/III collagen. It is likely that expression of collagenase facilitates the debriding of wound surfaces during epithelial migration. It is more uncertain what the precise role of the M r 95K GL might be. Rat M, 95K GL has some activity against solubilized type IV and V collagens (12), but it remains uncertain whether these collagens serve as substrates for the enzyme in their natural fibrillar and network forms (72). Other possible substrates include collagen chain fragments (gelatin) released from degrading collagen type I fibrils, and elastin, fibronectin, and laminin (73).

Somewhat surprising was our finding that dexamethasone failed to suppress TPA induction of FIB-CL and M, 95K GL as well as IL-10-induction of FIB-CL but was highly effective in down-regulating TGF-a/EGF responses. This observation is at variance with several previous studies on the effect of glucocorticoids on basal and TPA-induced metalloproteinase (FIB-CL, stromelysin-1) expression (31,43,45,48). We were unable to observe any effect of up to M dexamethasone on induction with lo-' M TPA, even under conditions in which TGF-a/EGF responses were inhibited by more than 70%. It is not possible at present to explain adequately the lack of responsiveness of TPA-induced expression of M, 95K GL and FIB-CL to dexamethasone or to explain the differ- ential inhibiting effect of dexamethasone on IL-18-induced M, 95K GL and FIB-CL expression. We speculate that keratinocytes may respond in a radically different manner than fibroblasts and HeLa cells to TPA/dexamethasone and pro- pose that TPA induction within the fiist few hours commits keratinocytes to a terminal differentiation program which then cannot later be aborted even by dexamethasone. In support of this idea are previous findings suggesting that TPA treatment induces keratinacytes to undergo a repertoire of changes characteristic of terminal differentiation (cessation of cell division, expression of transglutaminase activity, and formation of cornified envelopes, destruction of cytoplasmic and nuclear structures) (74-76). It is in keeping with these findings that TPA, in contrast to IL-18 and TGF-a/EGF, induces a transient and short lived burst of expression of metalloproteinases. It is also of note that the earliest effect of TPA (6-8 h) preceded by 4-6 h the first measurable effect of dexamethasone in our system? It is interesting to note that staurosporine, which also induces terminal differentiation in keratinocytes (54), like TPA, gives rise to a transient and short lived burst in M, 95K GL and FIB-CL expression.

Acknowledgments-We acknowledge the assistance of Dr. Chris- topher Franklin, University of Alabama at Birmingham, for conduct- ing the protein kinase C assays, Dr. John Kearney, University of Alabama at Birmingham, for donation of the Shark 46 monoclonal antibody, and Greg Harber and Lori Coward for assistance with the production of monoclonal antibodies and cultivation of rat mucosal keratinocytes.

REFERENCES I. Birkedal-Haneen, H., Moore, W. G. I., Bodden, M. K., Windsor, L. J.,

Birkedal-Hansen, B., DeCarlo, A., and Engler, J. A. (1992) Crit. Reu.

2. Bauer, E. A., Pentland, A. P., Kronberger, A., Wilhelm, S. M., Goldberg, G. I.. Welaus. H. G.. and Eisen, Z. (1988) Ann. N. Y. Acad. Sci 648,

Oml BWL Med. 4, 197-250

174-i79 - . 3. Coffey, R. J. J., Baecom, C. C., Sipes N. J.! Graves-Deal, R., Weissman, B.

4. Kupper, T. S., Ballard, D. W., Chua, A. O., McGuire, J. S., Flood, P. M., E., and M o m , H. L. (1988) Mol. bU. Bwl. 8,3088-3093

L. J. Windsor, unpublished data.

i x Metalloproteinuse Expression Horowitz, M. C., Landon, R., Lightfoot, J., and Gubler, U. (1986) J. Exp.

5. Gottlieb, A. B., Chang C. K., Posnett, D. N., Fanelli, B., and Tam, J. P. Med. 164,2095-2100

(19s) J. ~ z p . Med. i w , 670-675 6. Johnson-Wint, B., and Bauer, E. A. (1985) J. BWL Chem. 260,2080-2085 7. Finch, P. W. Rubin, J. S., Miki, T., Ron, D., and Aaronson, S. A. (1989)

8. Rubin, J. S., Osada, H., Finch, P. W., Taylor, W. G., Rudikoff, S., and Science 266,752-755

9. Johnson-Muller, B., and Gross, J. (1978) Proc. Natl. Acad. Sci. U. S. A. 76, Aaronson, S. A. (1989) Proc. Natl. Acad. Sci. U. S. A. 86,802-806

4417-4421 10. Johnson-Wht, B., and Gross, J. (1984) J. Cell BioL 98,9046 11. Lln, H. Y., Wells, B. R., Taylor, R. E., and Birkedal-Hansen, H. (1987) J. 12. Lyons, J. G., Birkedal-Hansen, B., Moore, W. G. I., O'Grady, R. L., and

13. Salo, T., Lyons, J. G., Rahemtulla, F., Birkedal-Hansen, H., and Larjava,

14. Turksen, K., Choi, Y., and Fuchs, E. (1991) Cell Regul. 2, 613-625 15. Petersen, M. J., Woodley, D. T., Stricklin, G. P., and O'Keefe, E. J. (1987)

16. Birkedal-Hansen, H., Hansen, I. L., Nellemann, K., and Westergaard, J.

17. Stevenson, G. A., Lyons, J. G., Cameron, D. A., and O'Grady, R. L. (1985)

18. Lyons, J . G., Nethery, A., O'Grady, R. L., and Harrop, P. J. (1989) Matrix

19. Birkedal-Hansen, B., Moore, W. G . I., Taylor, R. E., Bhown, A. S., and

20. McKinney, M. M., and Parkinson, A. (1987) J. ImmunoL Methodp 96,

21. Bayer, E. A., and Wilchek, M. (1980) Methods Biochem. A d 26,1-45 22. Franklin, C. C., Turner, J. T., and Kim, H. D. (1989) J. BioL Chem. 264,

23. Treisman, R. (1985) Cell 42,889-902 24. Conca, W., Kaplan, P. B., and Krane, S. M. (1989) J. Clin. Inuest. 83,

BWL Chem. 262,6823-6831

Birkedal-Hansen, H. (1991) Biochemistry 30,1449-1456

H. (1991) J. BwL Chem. 266,11436-11441

J. BWL Chem. 262,835-840

(1981) In Vitro 17,553-562

Biosci. Rep. 6,1071-1077

9, 7-16

Birkedal-Hansen, H. (1988) Biochemistry 27,6751-6758

271-278

6667-6673

1 7 K ! L l ? K 7

25. Carpenter, G. (1984) Cell 37,357-358 26. Saus, J., Quinones, S., Otani, Y., Nagase, H., Harris, E. D., Jr., and

27. McDonnell, S. E., Kerr, L. D., and Matrisian, L. M. (1990) MOL CeU. Biol.

28. Huhtala, P., Tuuttila, A., Chow, L. T., Lohi, J., Keski-Oja, J., and Tryggva-

29. MacNaul, K. L., Chartrain, N., Lark, M., Tocci, M. J., and Hutchinson, N.

30. Angel, P., Imagawa, M., Chiu, R., Stein, B., Imbra, R. J., Rahmsdorf, H. J.,

31. Auble, D. T., and Brinckerhoff, C. E. (1991) Biochemist 30,4629-4635 32. Brenner, D. A., O'Hara, M., Angel, P., Cholkier, M., anTKarin, M. (1989)

33. Lafyatis R Seong-Jin K., Angel, P., Roberts, A. B., Sporn, M. B., Karin,

34. Sirum-Connolly, K., and Brinckerhoff, C. E. (1991) Nucleic Acids Res. 19,

A,"" A,".

Kurkinen, M. (1988) J. BWL Chem. 263,6742-6745

10,42&1-4293

son, K. (1991) J. Bbl. Chem. 266.16485-16490

I. (1990) J. Biol. Chem. 266,17238-17245

Jonat, C., Herrlich, P., and Kann, M. (1987) Cell 49,729-739

Nature 337.661-663

M., A d Wilder, R. i. (1990) Mol. Endoeriml. 4,974-980

X?R-.?Ll 35. Heasley, L. E., and Johnson, G. L. (1989) J. BWL Chem. 264,8646-8652 36. Case, J. P., Lafyatis, R., Kumkumian, G., Remmem, E. F., and Wilder, R.

37. Moscat, J. C., Molloy, C. J., Fleming, T. P., and Aaronson, S. A. (1988)

38. Castagna, M., Takai, Y., Kaibuchi, K., Sano, K., Kikkawa, U., and Nishi-

39. Nishizuka, Y. (1986) Science 309,305-312 40. Collins, M. K., and Rozengurt, E. (1982) J. Cell PhysioL 112,42-50 41. Ruegg, U. T., and Burgess, G. M. (1989) Trends PharmacoL Sci. 10, 218-

220 42. Tamaoki, T., Nomoto, H., Takahashi, I., Kato, Y., Morimoto, M., and

Tomita, F. (1986) Biochem. Biophys. Res. Commun. 136,397-402 43. Clark, S. D., Kobayashi, D. K., and Welgus, H. G. (1987) J. Clin. Inuest.

"- "_ L. (1990) J. Immunol. 146,3755-3761

Mol. Endoerinol. 2,799-895

zuka, Y. (1982) J. BioL Chem. 267,7847-7851

44. Brinckerhoff, C. E., Plucinska, I. M., Sheldon, L. A., and O'Connor, G. T.

45. Fnsch, S. M., and Ruley, H. E. (1987) J. BioL Chem. 262,. 16300-16304 46. Zhann. X.-K.. Dona, J. M., and Chiu, J.-F. (1991) J. BwL Chem. 266.

80,1280-1288

(1986) Biochemistry 26,6378-6384

47.

48. 49. 50.

51.

52.

82&8254 ' - - ---.

Yang-Yen, H. F., Chambard, J. C., Sun, Y. L., Smeal, T., Schmidt, T. J.,

Jonat C Rahmsdorf, H. J., Park, K. K., Cab, A. C. B., Gebel, S., Ponta,

Aggeler, J., Friach, S. M., and Werb, Z. (1984) J. Cell BioL 98,1662-1671 Masada, M. P., Pemon, R., Kenney, J. S., Lee, S. W., Page, R. C., and

di Giovine, F. S., Poole, S., Situnayake, R. D., Wadhwa, M., and Duff, G.

Rooney, M., Symons, J. and Duff, G. W. (1990) Rheumotolo~ 10.217-

Drouin, J., and Karin, M. (1990) CeU 62,1205-1215

H.,'and Herrlich, P. (1990) CeU 62,1189-1204

Allison, A. C. (1990) J. Periodont. Res. 26, 156-163

W. (1990) Rheumatolo 9,259-264

219

and Dieppe, P. A. (1989) Ann. Rheum. Dis. 49,676-681 53. Westacott, C. I., Whicher, J. T., Barnes, I. C., Thompson, D., Swan, A. J.,

54. Sako. T.. Tauber. A. I.. Jeng. A. Y., Yuspa, S. H., and Blumberg, P. M~ ~~ , - . . ".

55. Yoshizawa, S., Fujiki, H., Suguri, H., Su anuma, M., Nakayaau, M., Mat- (1988)'Cancer ices. 46, ~-6-4sS0- -

sushima. R.. and Sueimura, T. (1990) hncer Res. 60,4974-4978 56. Dierks-Ventling, C., Kneael, J., Nagdine, Y., Hemmings, B. A., Pehling,

57. Tischler. A. S.. Ruzicka, L. A., and Perlman, R. L. (1990) J. Neuroehem. G., Fiacher, F., and Fabro, D. (1989) Int. J. Cancer 44.865-870

66, iim-1165 . -. . . . . .

58. Watanabe, M., Tamura, T., Ohashi, M., Hirasawa, N., Ozeki, T., Tsurufuji, S., Fukiki, H., and Ohuchi, K. (1990) Bioehim Biophys. Acta 1047,141- 147

Regulation of Keratinocyte Matrix Metalloproteinase Expression 19151 59. Rappolee, D. A., Mark, D., Banda, M. J., and Werb, Z. (1988) Science 241,

708-712 60. Coffey, R. J., Derynck, R., Wilcox, J. N., Bringman, T. S., Goustin, A. S.,

Moees, H. L., and Pittelkow, M. R. (1987) Nature 328.817-820 61. Pittelkow, M. R., Lindquist, P. B., Abraham, R. T., Graves-Deal, R.,

Derynck, R., and Coffey, R. J:, Jr. (1989) J. BWL Chem 264,5164-5171 62. Nordlund, L., Hormta, M., Saxen, L., and Thesleff, I. (1991) J. P e W n t .

63. Green, M. R., Baeketter, D. A., Couchman, J. R., and F b e ~ , D. A. (1983)

64. Paganelli Parker, K., Sauder, D. N., and Kilian, P. L. (1988) Endoer.

Rea. 26,333338

Deu. Bid. 100,506-512

Metcrbol Immunol. Functions Kemhnocutes 548,346-347 65. Barrandon, Y., and Green, H. (1987) Cell 60,1131-1137 66. DeClerck, Y. A., Perez, N., Shimada, H., Boone, T. C., Langley, K. E., and

Taylor. S. M. (1992) Cancer Res. 52, 701-708

68. 69.

70.

71. 72.

73.

74.

75. 76.

Liotta, L. A., and ktetler-Stevenson, W. (1989) J. NatL Cancer Znst. 81,

Moms, M. A. Sudhalter, J., and Langer, R. (1990) Science 248,1408-1410 Mackay A. e., Hartzler, J. L., Pelma, M. D., and Thorgeirason, U. P.

Senior R Griffin G Fliszar C. J Sha iro S. D., Goldberg, G. I., and

Yusp4 S. H., Ben, T., Hennings, H., and Lichtio, U. (1980) Biochem.

Seinert, P., and Yus a S. H. (1978) Science 200,1491-1493 Remers, J. J., Jr., an8dlaga, T. J. (1983) Cell 32.247-255

. ". W,"

556567

(1990) J. BmL Chem 268,21929-21934

Wel&,'H. G. (i99ij J. Bioi Che;. 268,7670-7875

Bwphys. Res. Corn- 97,700-708

the 01 vol. 268, no. 25. issue of september 5, pp. 19143 ... · the journal 01 biological chemistry...

Documents