EXPERIMENTAL CELL RESEARCH ARTICLE NO.
242, 110 –119 (1998)
Hydrocortisone Regulates the Dynamics of Plasminogen Activator and Plasminogen Activator Inhibitor Expression in Cultured Murine Keratinocytes Jenny M. Bator,*,† Rhonna L. Cohen,* and Donald A. Chambers*,†,‡,1 *Center for Molecular Biology of Oral Diseases, †Department of Bioengineering, and ‡Department of Biochemistry and Molecular Biology, University of Illinois at Chicago, Chicago, Illinois 60612
INTRODUCTION The plasminogen activators tPA and uPA, and their inhibitors, PAI-1 and PAI-2, have been associated with epithelial homeostasis and wound healing. In these studies, we investigate the effect of the steroid hormone hydrocortisone, a commonly used therapeutic modality for skin, on PAs/PAIs in serum- and plasminogen-free primary cultures of murine keratinocytes. SDS–PAGE fibrin zymography showed that addition of 1 mM hydrocortisone to cultures significantly reduced tPA fibrinolytic activity in both cell extracts and conditioned medium. uPA activity in conditioned medium was similarly inhibited. Cells were also cultured in the presence of dibutyryl cyclic AMP (dbcAMP). dbcAMP (5 mM) alone enhanced uPA and tPA fibrinolytic activity in conditioned medium, but this increase was diminished in the presence of 1 mM hydrocortisone. Immunoblots revealed a three- to fivefold induction of free PAI-1 by hydrocortisone which was partially blocked by dbcAMP. Northern blots showed that PAI-1 mRNA increased threefold 2 h after addition of hydrocortisone and remained elevated at least 8 h. In contrast, uPA and tPA mRNA were unchanged over the same time course. uPA, tPA, and PAI-1 mRNA increased in the presence of dbcAMP; levels remained elevated at least 8 h. HC suppressed the induction of uPA and tPA by dbcAMP. Studies directed at identifying plasminogen mRNA showed that in this culture system, keratinocytes produce no plasminogen mRNA either in the presence or in the absence of hydrocortisone or dbcAMP. Collectively, these results show that keratinocyte plasminogen activator activity is suppressed by hydrocortisone as a function of increased PAI-1 combined with an attenuation of PA induction by agents that increase intracellular cAMP. These results provide additional information to further define the mechanism by which glucocorticoids inhibit wound healing. © 1998 Academic Press
1 To whom correspondence and reprint requests should be addressed at Department of Biochemistry and Molecular Biology (M/C 536), University of Illinois at Chicago, College of Medicine West, 1819 West Polk Street, Chicago, IL 60612-7334. Fax: (312) 413-1604.
0014-4827/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
Plasminogen activators (PAs) are serine proteases that convert the proenzyme plasminogen to active plasmin, a broad-spectrum proteolytic enzyme that readily degrades fibrin(ogen) as well as extracellular matrix glycoproteins including laminin, vitronectin, fibronectin, and proteoglycans. Two major PAs exist: tissuetype PA (tPA), important in fibrinolysis and thrombolysis, and urokinase-type PA (uPA), often associated with cell migration and extracellular matrix remodeling in both normal and pathological states. The activity of these enzymes is tightly regulated by specific, highaffinity inhibitors, plasminogen activator inhibitor types 1 and 2 (PAI-1 and PAI-2) [1–5]. PAs and PAIs have been associated with epithelial proliferation and differentiation in health and disease [6 –9] as well as in wound healing [10 –12]. Wound healing can be modeled in vitro by using primary cultures of proliferating keratinocytes grown on an artificial substratum . In epidermal cultures, the relative proportion of tPA and uPA varies with the degree of differentiation: uPA predominates in proliferating keratinocytes, while tPA becomes more abundant as the cells differentiate. Proliferating keratinocytes may also express the cell surface receptor for uPA, which is upregulated during wound healing, inflammation, and certain malignant processes [14 –16]. The uPA receptor (uPAR) binds the single-chain, inactive uPA, scuPA, which is rapidly converted to the two-chain, enzymatically active form of uPA. Receptor-bound uPA facilitates localized conversion of plasminogen to plasmin, thus enabling site-specific proteolysis of extracellular matrix components. The PA inhibitors PAI-1 and PAI-2 are also produced by keratinocytes . PAI-1 is the primary physiological inhibitor of PAs, and PA/ PAI-1 complexes are rapidly internalized and degraded. PAI-1 also binds to the somatomedin B binding site on the extracellular matrix adhesive glycoprotein vitronectin . By means of this anchorage to vitronectin, PAI-1 activity is maintained, permitting local-
HC REGULATES KERATINOCYTE PLASMINOGEN ACTIVATION
ized inhibition of plasminogen activation. Thus, spatial regulation of plasmin activity is achieved by site-specific immobilization of both PA and PAI-1. PAI-2 preferentially inhibits uPA , and its function in keratinocytes, as in other cells, is not well understood. Cellular regulation of the complex PA/PAI system is achieved by growth factors, cytokines, and hormones. There is increasing evidence that PA and PAI are regulated in various biological systems by glucocorticoid hormones, including hydrocortisone (HC) and the synthetic analog dexamethasone [19 –28]. Glucocorticoids were shown to inhibit retinoid-induced increase in PA activity in foreskin fibroblasts  and EGF-induced PA increase in epidermoid carcinoma cells . Additionally, glucocorticoids inhibited PA activity both in whole animal models and in in vitro models of the autoimmune skin-blistering disease pemphigus [31– 33]. However, neither the mechanism of inhibition of epidermal PA activity by glucocorticoids nor their effect on the production of PAIs in the epidermis has been addressed. Glucocorticoids exert an effect on most tissues, including regulation of epidermal growth and differentiation and wound healing [36, 37] as well as immune processes . Endogenous hydrocortisone levels are known to increase during times of stress. Short-term use of topical glucocorticoids is effective in treating inflammatory and hyperproliferative disorders of the epidermis [34, 35]. However, chronic systemic administration of exogenous glucocorticoids, as in the treatment of arthritis, can result in undesirable side effects, including marked skin atrophy and delayed wound healing. Yet knowledge of the mechanisms underlying these effects is incomplete. Cyclic AMP (cAMP) also regulates cell growth in many mammalian systems. In keratinocytes cAMP is thought to enhance proliferation  and stimulate wound healing . Imbalances in cyclic nucleotide levels have also been associated with psoriasis, a hyperproliferative disease of the epidermis [41, 42]. cAMP has variable effects on PAs/ PAIs, depending on the cell type [19, 43, 44]. For these reasons we investigated the effects of HC and the cAMP analog dbcAMP on components of the PA/PAI system in a murine keratinocyte culture model of wound healing. MATERIALS AND METHODS Reagents. HC, dbcAMP, fibrinogen, thrombin, trypsin, Triton X-100, Tween 20, bovine serum albumin, and Tris base were from Sigma (St. Louis, MO). Agarose and calcium-free phosphate-buffered saline (PBS) were from Life Technologies, Inc. (Grand Island, NY). Cell culture. Specific pathogen-free male Balb/c mice, 4 – 6 weeks old, were obtained from Harlan (Indianapolis, IN). Mouse skins were cut into several pieces and placed in sterile PBS 1 0.1% trypsin for 1 h at 37°C. After rinsing in PBS 1 10% FBS, epidermal cells were gently scraped away from the underlying dermis, centrifuged at
1000g for 10 min, and resuspended in complete keratinocyte growth medium (KGM, Clonetics, Inc., San Diego, CA; supplements include gentamicin and amphotericin, insulin, epidermal growth factor, bovine pituitary extract, hydrocortisone) 1 0.03 mM Ca21 and 1% FBS. This medium does not contain plasminogen, tPA, uPA, PAI-1, or PAI-2. Cells were plated at 1.5 3 106 viable cells per milliliter into culture dishes. After overnight incubation, medium was removed and replaced with serum-free KGM 1 0.03 mM Ca21. Cultures became confluent after 6 –7 days, after which HC was omitted from the medium and differentiation was induced by increasing the calcium concentration to 1.4 mM . After 48 h in high calcium medium, experimental agents were added as detailed below. At designated times, conditioned medium (CM) was collected and centrifuged to eliminate cellular debris and stored at 270°C until analysis. Immediately after removal of culture medium, culture dishes were placed on ice and extract buffer (0.1 M Tris, 0.1% Triton X-100, pH 8.1, containing 2 mg/ml aprotinin and 2 mg/ml leupeptin) was added. Cell extracts (CE) were frozen at 270°C until analysis, usually within 2 weeks. In other experiments, total RNA was extracted using the RNeasy system (Qiagen, Inc., Chatsworth, CA). Fibrin zymography. Zymography was performed according to a procedure modified from the method of Granelli-Piperno and Reich . Briefly, conditioned medium and cell extract samples were electrophoresed in 10% acrylamide gels under nonreducing conditions. Gels were washed twice in 0.1 M Tris, 2.5% Triton X-100, pH 8.1, for 30 min each to remove SDS and placed on an agarose indicator gel containing fibrinogen, plasminogen, and thrombin in 0.05 M Tris, 0.05% Triton X-100, pH 8.1. Gels were incubated overnight at 4°C to allow penetration of proteins into the fibrin–agarose. Thereafter, gels were incubated in a dry 37°C incubator to induce enzyme activity. Gels were monitored visually for appearance of lysis zones corresponding to uPA and tPA activity and then photographed. Areas of lysis were quantified by scanning densitometry. Immunoanalysis of PAIs. Conditioned medium or cell extract samples were electrophoresed as described above. Gels were rinsed briefly in distilled water and transferred to nitrocellulose (Schleicher and Schuel Inc., Keene, NH). Nitrocellulose blots were incubated in PBS 1 1% BSA for 1.5 h at 37°C or overnight at room temperature with gentle rocking. Blots were then incubated in PBS 1 1% BSA containing 5 mg/ml rabbit anti-rat PAI-1 IgG or goat anti-human PAI-2 IgG (American Diagnostica Inc., Greenwich, CT) for 2 h at 37°C. Blots were washed in PBS 1 0.5% Tween-20 (PBST) three times for 30 min each and then incubated in PBS 11% BSA 1 0.05% Tween-20 containing HRP-conjugated donkey anti-rabbit Ig or HRPconjugated goat rabbit anti-goat Ig (Amersham, Arlington Heights, IL) at 37°C for 45 min. Finally, blots were washed in PBST five times for 15 min each, and immunoreactive bands were visualized using ECL detection reagent (Amersham), followed by autoradiography on Hyperfilm-ECL (Amersham). Band intensity was quantified by scanning densitometry. RNA preparation and Northern blot analysis. RNA samples (10 mg) were denatured, electrophoresed in formaldehyde-containing 1% agarose gels, transferred to Hybond-N1 membranes (Amersham), and fixed by UV irradiation. Blots were prehybridized for 3 h at 42°C and then hybridized overnight at 42°C in the same solution containing approximately 2 3 106 cpm/ml labeled probe. Nylon membrane containing immobilized FTO2B liver cell line RNA was generously donated by Dr. Carolyn Bruzdzinski (University of Illinois at Chicago, Chicago, IL). Following hybridization, blots were washed twice in 23 SSC (0.3 M NaCl, 0.03 M Na citrate) at 42°C for 20 min each and then once in 0.23 SSC at 55°C for 20 min. Blots were exposed to Kodak XAR-5 film (Eastman Kodak, Rochester, NY) at 270°C with intensifying screens. Intensity of mRNA bands was quantified by scanning densitometry of the autoradiographs. Equal sample loading was confirmed by hybridization with either GAPDH or 28S rRNA cDNAs.
BATOR, COHEN, AND CHAMBERS
Probes. cDNA clones for rat uPA, mouse tPA, mouse PAI-1, rat GAPDH, and rat 28S rRNA were generously provided by Dr. Carolyn Bruzdzinski (University of Illinois at Chicago). The cDNA clone for mouse plasminogen was a generous gift from Dr. Sandra J. F. Degen (Children’s Hospital Medical Center, Cincinnati, OH). A 2.3-kb XhoI–HindIII fragment of the mouse PAI-1 cDNA clone (pmPAI-1) and a 2.5-kb XbaI–HindIII fragment of the mouse tPA clone pTAM 2.5-a were prepared. The vectors containing the mouse plasminogen, rat uPA, rat GAPDH, and rat 28S rRNA cDNAs were used intact. All cDNAs were labeled to high specific activity with [a-32P]dCTP (Amersham) using the RadPrime DNA labeling system (Life Technologies, Inc.).
Effect of Hydrocortisone on PA-Mediated Fibrinolysis In all experiments, murine keratinocytes were first allowed to proliferate in KGM containing 0.03 mM Ca21 and all supplements until 90% confluent. Cell differentiation was promoted by increasing the Ca21 concentration to 1.4 mM, whereupon the proliferation rate decreased concomitant with initiation of a differentiated phenotype. HC-free medium was then used during the differentiation stage. Keratinocytes cultured for 48 h after the Ca21 switch secreted uPA, and to a lesser extent, tPA, into the culture medium. To determine the effect of glucocorticoids on PA/PAI production in this system, keratinocytes cultured as above were incubated in the presence (0.001–1.0 mM) or absence of HC for 24 h, beginning 48 h following the Ca21 switch. Zymographic analysis of PA activity in conditioned medium revealed that exposure of cells to HC resulted in a dose-dependent inhibition of uPA-mediated fibrinolysis (Fig. 1A). Maximal inhibition was 60% at 1 mM HC (IC50 5 400 nM). In contrast, enzymatic activity of uPA in cell extract was unaffected by HC at any of the concentrations tested (Fig. 1C). Like uPA, HC inhibited tPA activity in conditioned medium by 75%, with an IC50 of 35 nM (Fig. 1B). However, in contrast to the effects of HC on uPA in cell extracts, tPA-mediated lytic activity was inhibited as much as 95% by 1 mM HC, with an IC50 of 6 nM (Fig. 1D). To confirm that these effects were a specific result of glucocorticoid action on the cells, keratinocytes were incubated with HC in the presence or absence of the glucocorticoid antagonist RU486. When cells were cultured in the presence of both 1 mM HC and 1 mM RU486, the inhibitory effect of HC on uPA and tPA activity in conditioned medium was completely abolished (Fig. 2), whereas RU486 alone had no effect. Effect of Hydrocortisone on PAI Production Total PA activity is a function of the relative amounts of PAs and PAIs present. The marked reduction in PA activity observed in HC-treated cultures
FIG. 1. Effect of HC on PA activity in murine keratinocyte cultures: fibrin zymography. Confluent cells were cultured in keratinocyte growth medium containing 1.4 mM Ca21 with or without HC at the indicated concentrations. Conditioned medium and cell extracts were collected after 24 h. Proteins were separated by SDS–PAGE, and gels were placed on fibrin agarose indicator plates. The presence of plasminogen activators was assessed according to the appearance of bands of lysis. (A) uPA-mediated lysis, conditioned medium; (B) tPA-mediated lysis, conditioned medium; (C) uPA-mediated lysis, cell extracts; (D) tPA-mediated lysis, cell extracts. Values are expressed as percentage of control (lysis in the absence of HC). Results are representative of at least three separate experiments.
may have resulted from decreased PA, increased PAI, or both. To study these alternatives, immunoblot analysis was performed on samples of conditioned medium and cell extract from control and HC-treated (0.001– 1.0 mM) keratinocytes. PAI-1 antigen was not detectable in cell extracts (data not shown), but was detectable in conditioned medium (fivefold concentrated) as a 52-kDa band. Figure 3A shows that treatment of keratinocytes with HC (0.001–1.0 mM) resulted in a dosedependent increase in secreted PAI-1, to about threefold above control. In some experiments, PAI-1 levels increased as much as fivefold (see Fig. 4C). The HCinduced upregulation of PAI-1 in conditioned medium was completely abrogated by 1 mM RU486 (Fig. 3B). HC had little effect on PAI-2 antigen as determined by immunoblotting. PAI-2 was undetectable in conditioned medium in the presence or absence of HC. HC caused small changes in PAI-2 protein in cell extracts (Fig. 3C).
HC REGULATES KERATINOCYTE PLASMINOGEN ACTIVATION
cant effect on levels of free PAI-1. However, dbcAMP partially attenuated HC-induced upregulation of PAI-1. These results suggest a complex interaction between the effector cAMP-mediated signal transduction pathway and the PA/PAI system. FIG. 2. Effect of RU486 on HC-mediated inhibition of PA activity in murine keratinocyte cultures: fibrin zymography. Confluent cells were cultured in keratinocyte growth medium containing 1.4 mM Ca21 (control) or with HC (1 mM), RU486 (1 mM), or both. Samples were collected and analyzed as in Fig. 1.
Effect of Dibutyryl cAMP and HC on PA and PAI Activity To further explore regulation of PA/PAI expression, we examined plasminogen activation in the presence of HC with and without dbcAMP. Figure 4A shows uPA fibrinolytic activity in conditioned medium after cells were cultured in 5 mM dbcAMP 6 1 mM HC for 24 h. uPA-mediated lysis was enhanced twofold with the addition of dbcAMP alone. When HC was added to the culture medium with dbcAMP, uPA activity was reduced to control levels. dbcAMP alone had little effect on tPA activity in conditioned medium, but increased tPA activity in cell extracts (Fig. 4B). When added together, the stimulatory effect of dbcAMP on tPA activity in cell extracts was attenuated by HC. Raising intracellular cAMP decreases PAI-1 in certain cell systems [43, 44], but increases it in others . To determine the effect of increased intracellular cAMP on keratinocyte PAI-1, dbcAMP was added to cell cultures in the presence and absence of HC (Fig. 4C). HC added alone increased PAI-1 levels about fivefold, whereas the addition of dbcAMP had no signifi-
Effect of HC on PA and PAI mRNA The role of HC in PA and PAI gene expression was analyzed by Northern blots of total RNA purified from keratinocytes cultured in the presence or absence of HC. Murine PAI-1 mRNA appeared as a single species about 3.2 kb. Time course experiments revealed that HC upregulated PAI-1 mRNA about two- to threefold over control. Representative experiments are shown in Figs. 5A and 6. PAI-1 mRNA remained elevated at least twofold greater than control for up to 8 h. These results are consistent with the observed increase in PAI-1 protein expression in conditioned medium. HC did not significantly affect uPA or tPA mRNA. These data suggest that the HC-mediated reduction in total PA activity seen by zymography may result from a selective upregulation of PAI-1 gene expression in the absence of a concomitant induction of uPA or tPA expression. Cultures were treated with the protein synthesis inhibitor cycloheximide (CHX), with or without HC (Fig. 6). Cells treated with HC alone showed a threefold increase in PAI-1 mRNA, consistent with the results shown in Fig. 5. CHX alone also increased PAI-1 mRNA levels approximately threefold over control by 1–2 h, but mRNA fell below control levels after 8 h. When CHX was added together with HC, PAI-1 mRNA again increased nearly threefold by 1–2 h, but mRNA
FIG. 3. Effect of HC on PAI-1 and PAI-2 in murine keratinocytes in culture: immunoblot analysis. Confluent cells were cultured in keratinocyte growth medium containing 1.4 mM Ca21 (control) or with HC, RU486, or both, as indicated. Conditioned medium was collected after 24 h. Proteins were separated by SDS–PAGE and transferred to nitrocellulose. Blots were incubated with polyclonal rabbit anti-rat PAI-1 antibody (A, B) or polyclonal goat anti-human PAI-2 antibody (C). Immunoreactivity was detected as described under Materials and Methods. (A, C) Cells cultured in medium alone (control) or with 0.001–1.0 mM HC for 24 h; (B) cells cultured in medium alone (control) or with HC (1 mM), RU486 (1 mM), or both. Values are expressed as fold increase over control. Results are representative of at least three separate experiments.
BATOR, COHEN, AND CHAMBERS
Effect of dbcAMP on PA and PAI mRNA The addition of dbcAMP to keratinocyte cultures resulted in an increase in PAI-1 mRNA similar to that induced by HC (Fig. 5A). The combination of HC and dbcAMP yielded PAI-1 mRNA levels similar to that induced by each agent individually and was not additive. In addition, dbcAMP increased uPA mRNA more than twofold and tPA mRNA threefold by 8 h (Figs. 5B and 5C). The dbcAMP-mediated increases in both uPA and tPA mRNA levels were significantly attenuated by HC. Together, these results suggest that HC alone can decrease PA activity by selective upregulation of PAI-1 gene expression. Furthermore, HC suppresses cAMPmediated increases in uPA and tPA mRNA. Absence of Constitutive or Inducible Plasminogen mRNA in Cultured Murine Keratinocytes Since the epidermis is an avascular organ, the availability of plasminogen to keratinocytes in vivo is thought to be a function of vascular permeability from the underlying dermal vasculature. Thus, several investigators have shown the association of plasminogen with keratinocytes at, or close to, the basement membrane [37, 38]. Since murine keratinocytes in culture express tPA, uPA, PAI-1, and PAI-2, we asked whether keratinocytes in this culture system could express plasminogen mRNA or if HC or dbcAMP could induce its expression. In vivo, plasminogen is produced in the liver. Thus, as a positive control, RNA was isolated from the FTO2B liver cell line and shown to contain plasminogen mRNA (Fig. 7). In contrast, when keratinocytes were similarly assayed, plasminogen mRNA was undetectable in the absence or presence of 1 mM HC or 5 mM dbcAMP (Fig. 7). FIG. 4. Effect of HC and dbcAMP on PA and PAI-1 in murine keratinocytes in culture. Confluent cells were cultured in keratinocyte growth medium containing 1.4 mM Ca21 (control) or with HC, dbcAMP, or both, as indicated. Conditioned medium and cell extracts were collected after 24 h. Proteins were separated by SDS–PAGE and gels were processed for fibrin zymography or immunoblot analysis as described under Materials and Methods. (A) uPA-mediated fibrinolysis in conditioned medium; (B) tPA-mediated fibrinolysis in cell extract; (C) PAI-1 protein expression in conditioned medium. Lane 1, control; lane 2, 1 mM HC; lane 3, 5 mM dbcAMP; lane 4, 1 mM HC 1 5 mM dbcAMP. Values are expressed as fold increase over control. Results are representative of at least three separate experiments.
levels rapidly declined in a pattern similar to that with CHX alone. While the totality of the effect of CHX on cells is not easily assessed, these data could suggest that the stability of PAI-1 mRNA induced by exposure of cells to steroids may be maintained by newly synthesized protein(s).
These investigations demonstrate that the glucocorticoid HC differentially modulates PAs and PAIs in murine keratinocytes in vitro by suppressing plasminogen activation. HC also attenuates cAMP-mediated stimulation of the plasminogen activation system. In our system, keratinocytes produced predominantly uPA. Whereas the enzymatic activity of uPA in conditioned medium was decreased by HC, activity from cell extracts was unaffected. A possible explanation for this observation is that a significant portion of cell-associated PA was bound to the cell surface receptor (uPAR) in the zymogen form (scuPA). uPA is secreted as a zymogen [48, 49] which is not susceptible to inhibition by PAI-1 . Thus, it is likely that scuPA was present on the cell surface prior to treatment of cells with HC, and subsequent exposure to HC, while
HC REGULATES KERATINOCYTE PLASMINOGEN ACTIVATION
FIG. 5. Effect of HC and dbcAMP on PA and PAI-1 mRNA. Confluent cells were cultured in keratinocyte growth medium containing 1.4 mM Ca21 (control) or with HC, dbcAMP, or both, as indicated. At the time points indicated (hours), total RNA was isolated from cells and mRNA was detected by Northern blot analysis as described under Materials and Methods. Each lane contained 10 mg total RNA. (A) PAI-1 mRNA; (B) uPA mRNA; (C) tPA mRNA; (D) GAPDH mRNA. Plots represent the data shown: }, 1 mM HC; ■, 5 mM dbcAMP; Œ, 1 mM HC 1 5 mM dbcAMP. Autoradiographs were analyzed by scanning densitometry, and values were normalized to GAPDH. Values are expressed as fold increase over control. Results are representative of at least three separate experiments.
modulating uPA activity in conditioned medium, apparently had no effect on bound scuPA. Despite a sustained increase in expression of tPA mRNA in response to dbcAMP, we observed only a modest increase in tPA activity in cell extracts and no increase in tPA activity in conditioned medium. Posttranscriptional control mechanisms, including regulation of mRNA stability, translation, posttranslational modifications, and secretion, may account for this. It is also possible that tPA protein did increase, but was complexed with PAI-1 and therefore not readily detectable. dbcAMP increased PAI-1 mRNA, but the amount of free PAI-1 protein did not change. In fibrin zymograms, lysis zones corresponding to PA/PAI complexes appeared weak and variable. Attempts to identify proteins in these zones failed because the anti-PAI-1 antibody available did not recognize murine PAI-1 in complex with PAs. Thus, resolution of this apparent disparity between message and protein awaits further study. Increases in PAI-1 protein and mRNA by glucocorticoids have been demonstrated in rat hepatocytes 
and hepatoma cells , normal and malignant rat osteoblasts , human fibrosarcoma cells , human synovial fibroblasts , human monocytes , and bovine aortic endothelium . Glucocorticoid effects on PA protein and gene regulation, however, are variable. In rat hepatoma cells, dexamethasone induced a transient increase in tPA mRNA . In contrast, dexamethasone enhanced tPA production but decreased uPA production in rat granulosa cells  and in human HT-1080 fibrosarcoma cells . In vivo, cortisol downregulated both tPA and uPA mRNA in the involuting prostate of castrated rats, but PAI-1 mRNA was undetectable . Control of the PA/PAI system thus depends on the species, tissue, and cell type. Our results show that in cultured murine keratinocytes, HC exerts independent regulatory effects on the secreted proteins of the plasminogen system, tPA, uPA, and PAI-1. Glucocorticoids such as HC enter cells passively and bind with high specificity and affinity to cytoplasmic receptors [51, 52]. This complex binds to glucocorticoid response elements (GREs), initiating transcription.
BATOR, COHEN, AND CHAMBERS
FIG. 6. Effect of HC and CHX on PAI-1 mRNA. Confluent cells were cultured in keratinocyte growth medium containing 1.4 mM Ca21 (control) or with 1 mM HC, 0.1 mM CHX, or both, as indicated. Total RNA was isolated from cells at the time points indicated (hours). PAI-1 mRNA levels were detected by Northern blot analysis as described under Materials and Methods. Each lane contained 10 mg total RNA. (A) PAI-1 mRNA; (B) GAPDH mRNA. Plot represents the data shown: F, 0.1 mM CHX; ■, 1 mM HC; Œ, 0.1 mM CHX 1 1 mM HC. Autoradiographs were analyzed by scanning densitometry. GAPDH mRNA appeared to be affected by CHX at later time points; therefore, PAI-1 densitometric values were not normalized to this mRNA. However, equal sample loading can be observed at early time points (t , 4 h) and in all samples treated with HC alone. Values are expressed as fold increase over control. Results are representative of two separate experiments.
GREs have been identified in the rat PAI-1 promoter  and human PAI-1 promoter , and in those studies, increases in PAI-1 mRNA were attributed directly to increased transcription. Although GREs have not yet been identified in the mouse PAI-1 gene, our
FIG. 7. Lack of plasminogen mRNA expression in keratinocytes. Northern blot analysis was performed on total RNA from the plasminogen-expressing FTO2B liver cell line (lanes 1 and 2) or from keratinocytes cultured in KGM alone (lane 3) or with 1 mM HC (lane 4) or 5 mM dbcAMP (lane 5). Blots were hybridized with mouse plasminogen cDNA (top) or rat 28S rRNA cDNA (bottom) as a positive control.
data show a rapid upregulation of PAI-1 mRNA by HC, which supports the existence of such a GRE and its role in increasing PAI-1 mRNA. Increased mRNA levels persisted at least 8 h, but in the presence of CHX, mRNA fell to control levels or below by this time. Thus, it is possible that HC increases transcription of the PAI-1 gene, increases PAI-1 mRNA stability, or both. GREs in tPA or uPA promoter regions have yet to be identified. However, our data showing that HC suppresses cAMP-induced increases in tPA and uPA mRNA suggests that some form of regulatory control exists. It is known that activated glucocorticoid receptor (GR) can interfere with transcriptional activation by other factors, especially AP-1 [55, 56], which is known to be involved in cell proliferation and the induction of many genes. Studies have identified celltype-specific AP-1 binding elements in the human urokinase gene promoter [57, 58]. It has also been shown that the mouse tPA promoter contains a consen-
HC REGULATES KERATINOCYTE PLASMINOGEN ACTIVATION
sus recognition site for AP-1 . The presence of these regions in the uPA and tPA promoter regions provides a basis for a proposed mechanism for inhibition of gene transcription by glucocorticoids. Glucocorticoids also stimulate the production of IkB, a cytoplasmic regulatory binding protein for the transcription factor NF-kB. Increased IkB prevents NF-kB from translocating to the nucleus to activate transcription of a number of genes. It has been reported that inhibition of NF-kB/ Rel A expression by antisense oligonucleotides blocks uPA synthesis in a human ovarian carcinoma cell line , supporting the notion that glucocorticoids may affect uPA expression by this mechanism. For more than a decade, studies have suggested that the PA/PAI system plays a role in epithelial homeostasis, influencing growth, differentiation, wound healing, and extracellular matrix remodeling [6, 7, 9]. The observed inhibition of PA activity by HC is likely a function of the selective upregulation of PAI-1 in the absence of a corresponding increase in plasminogen activators. It seems paradoxical that although cultured keratinocytes synthesize an array of plasminogen-associated molecules (e.g., uPA, tPA, and PAI-1), we find no evidence for either plasminogen protein or plasminogen mRNA. In vivo, plasminogen has been observed only in basal keratinocytes (those keratinocytes closest to the dermal vasculature). Thus, keratinocyte-associated plasminogen is thought to result from internalization of plasma-derived plasminogen binding to the cell surface and not due to production by the keratinocyte itself . This situation suggests that there may be substrates other than plasminogen for PAs. Such an explanation is consistent with the work of others [61–63] and also with the observed axonal secretion of uPA as well as tPA . As a working hypothesis for wound healing and epidermal pathobiologies such as psoriasis, it is likely that uPA, which appears to be more plentiful in cells close to the basal layer, plays a role in cell migration, whereas tPA, found in the more superficial anatomy of the epidermis and the stratum corneum, may relate to the process of terminal differentiation and enucleation. Components of the plasminogen activation system likely act in concert with other cellular molecules, including the uPA receptor and integrins, to direct cell adhesion and migration . In the context of this hypothesis, PAIs serve to modulate PA function in processes such as wound healing, shown to be aberrant in genetically modified knockout mice in which PA-associated molecules have been altered or deleted [66–73]. Our in vitro model was designed to study the effects of HC and dbcAMP on plasminogen activation in activated keratinocytes; thus the responses we observed were independent of contributing effects from other cell types. However, wound healing in vivo is mediated by other cell types in addition to keratinocytes, most notably macrophages and fibroblasts . Macrophages produce several cytokines as well as the rele-
vant components of the plasminogen system, uPA, uPAR, and PAI and thus play a significant role in the repair process. In vitro, glucocorticoids have been shown to inhibit macrophage uPA activity  as well as the production of mitogenic or chemotactic factors such as IL-1, TNF-a, and PGE2 . In vivo, guinea pigs given systemic HC showed macrophage depletion and impaired wound healing . Factors mitogenic for fibroblasts during wound healing include macrophage-mediated cytokines, provisional matrix degradation products, and plasminogen activators tPA and uPA themselves [74, 78]. Glucocorticoid inhibition of PA and PAI expression in both keratinocytes and macrophages may affect fibroblast recruitment, proliferation, and subsequent production of paracrine mitogenic factors including keratinocyte growth factor (KGF). Glucocorticoids are also known to directly suppress collagen production in fibroblasts. Clearly, the interplay between cell types participating in the various stages of wound repair is complex and highly regu-
FIG. 8. (A) The cellular pathways by which hydrocortisone is thought to alter the balance in PA/PAI production by keratinocytes. Hydrocortisone increases PAI-1 directly by binding to the PAI-1 promoter GRE. The glucocorticoid may also decrease PAs by interfering with fos/jun- or NF-kB-mediated transcription of these genes. (B) A proposed mechanism by which hydrocortisone contributes to impaired wound healing. An imbalance in PA/PAI in the epidermis leads to decreased plasminogen activation, delayed or incomplete matrix degradation, and inhibition of keratinocyte migration.
BATOR, COHEN, AND CHAMBERS
lated, and perturbations by agents such as glucocorticoids can have multiple cascading or overlapping effects. In summary, we have shown that the glucocorticoid hydrocortisone affects the plasminogen activator/inhibitor system in murine keratinocytes in vitro by upregulating PAI-1 and downregulating uPA and tPA, as depicted in Fig. 8A. This suppression may be exerted directly, by the binding of hormone/receptor complexes to GREs, or indirectly, as shown by the attenuation of gene induction by other transcription factors such as fos/jun or NF-kB. Plasminogen activators play a role in epithelial homeostasis and wound healing, and steroid hormones are known to suppress wound healing. Figure 8B suggests a mechanistic rationale correlating these two phenomena: steroids suppress the wound healing process in part by altering the balance in the plasminogen activator/inhibitor system in keratinocytes in favor of suppression of plasminogen activation. This suppression likely inhibits proteolytic matrix degradation and reepithelialization by keratinocytes, processes necessary for rapid and efficient wound repair.
The authors thank Dr. Lloyd Graf, Jr., and Dr. Carolyn J. Bruzdzinski for helpful discussion, Dr. Judith M. Brown for initial experiments revealing the absence of plasminogen protein in cultured keratinocytes, and Ms. Diane Rudall for editorial assistance.
REFERENCES 1. 2. 3. 4. 5. 6. 7.
8. 9. 10.
11. 12. 13. 14. 15.
Hart, D. A., and Rehemtulla, A. (1988) Comp. Biochem. Physiol. 90B, 691–708. Mayer, M. (1990) Clin. Biochem. 23, 197–211. Scully, M. F. (1991) Br. J. Haematol. 79, 537–543. Vassalli, J.-D., Sappino, A.-P., and Belin, D. (1991) J. Clin. Invest. 88, 1067–1072. Kramer, M. D., Schaefer B., and Reinartz, J. (1995) Biol. Chem., Hoppe-Seyler 376, 131–141. Brown, J. M. (1982). Plasminogen Activator in Epidermal Homeostasis, Ph.D. thesis, University of Michigan, Ann Arbor, MI. Chambers, D. A., Brown, J. M., and Cohen, R. L. (1983) in Psoriasis: Cell Proliferation (Wright, N. A., and Camplejohn, R. S., Eds.), pp. 131–140, Churchill Livingstone, New York. Jensen, P. J., John, M., and Baird J. (1990) Exp. Cell Res. 187, 162–169. Lazarus, G. S., and Jensen, P. J. (1991) Semin. Thromb. Hemost. 17, 210 –215. Rømer, J., Lund, L. R., Eriksen, J., Ralfkiær, E., Zeheb, R., Gelehrter, T. D., Danø, K., and Kristensen, P. (1991) J. Invest. Dermatol. 97, 803– 811. Quax, P. H. A., Boxman, I. L. A., van Kesteren, C. A. M., Verheijen, J. H., and Ponec, M. (1994) Fibrinolysis 8, 221–228. Scha¨fer, B. M., Maier, K., Eickhoff, U., Todd, R. F., and Kramer, M. D. (1994) Am. J. Pathol. 144, 1269 –1280. Grinnell, F. (1992) J. Cell Sci. 101, 1–5. Blasi, F. (1988) Fibrinolysis 2, 73– 84. Estreicher, A., Wohlwend, A., Belin, D., Schleuning, W.-D., and Vassali, J.-D. (1989) J. Biol. Chem. 264, 1180 –1189.
18. 19. 20. 21.
22. 23. 24.
29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.
Rømer, J., Lund, L. R., Eriksen, J., Pyke, C., Kristensen, P., and Danø, K. (1994) J. Invest. Dermatol. 102, 519 –522. Salonen, E.-M., Vaheri, A., Po¨lla¨nen, J., Stephens, R., Andreasen, P., Mayer, M., Danø, K., Gailit, J., and Ruoslahti, E. (1989) J. Biol. Chem. 264, 6339 – 6343. Thorsen, S., Philips, M., Selmer, J., Lecander, I., and Astedt, B. (1988) Eur. J. Biochem. 175, 33–39. Heaton, J. H., Nebes, V. L., O’Dell, L. G., Morris, S. M., Jr., and Gelehrter T. D. (1989) Mol. Endocrinol. 3, 185–192. Medcalf, R. L., van den Berg, E., and Schleuning, W.-D. (1988) J. Cell. Biol. 106, 971–978. Hamilton, J. A., Cheung, D., Filonzi, E. L., Piccoli, D. S., Wojta, J., Gallichio, M., McGrath, K., and Last, K. (1992) Arthritis Rheum. 35, 1526 –1534. Heaton, J. H., and Gelehrter, T. D. (1989) Mol. Endocrinol. 3, 349 –355. Fukumoto, S., Allan, E. H., Zeheb, R., Gelehrter, T. D., and Martin, T. J. (1992) Endocrinology 130, 797– 804. Andreasen, P. A., Pyke, C., Riccio, A., Kristensen, P., Nielsen, L. S., Lund, L. R., Blasi, F., and Danø K. (1987) Mol. Cell. Biol. 7, 3021–3025. Hamilton, J. A., Whitty, G. A., Wojta, J., Gallichio, M., McGrath, K., and Ianches, G. (1993) Cell. Immunol. 152, 7–17. Blei, F., Wilson, E. L., Mignatti, P., and Rifkin, D. B. (1993) J. Cell. Physiol. 155, 568 –578. Wang, C., and Leung, A. (1989) Endocrinology 124, 1595–1601. Freeman, S. N., Rennie, P. S., Chao, J., Lund, L. R., and Andreasen, P. A. (1990) Biochem. J. 269, 189 –193. Wilson, E. L., and Dowdle, E. B. (1980) Cancer Res. 40, 4817– 4820. Gross, J. L., Kruupp, M. N., Rifkin, D. B., and Lane, M. D. (1983) Proc. Natl. Acad. Sci. USA 80, 2276 –2280. Hashimoto, K., Shafran, K. M., Webber, P. S., Lazarus, G. S., and Singer, K. H. (1983) J. Exp. Med. 157, 259 –272. Hashimoto, K., Singer, K., and Lazarus, G. S. (1984) Br. J. Dermatol. 110, 293–297. Anhalt, G. J., Patel, H. P., Labib, R. S., Diaz, L. A., and Proud, D. (1986) J. Immunol. 136, 113–117. Yohn, J. J., and Weston, W. L. (1990) Curr. Probl. Dermatol. 2, 31– 63. Hepburn, D., Yohn, J. J., and Weston, W. L. (1994) Adv. Dermatol. 9, 225–255. Baxter, J. D., and Forsham, P. H. (1972) Am. J. Med. 53, 573–589. Ponec, M. (1984) Int. J. Dermatol. 23, 11–24. Parrillo, J. E., and Fauci, A. S. (1979) Annu. Rev. Pharmacol. Toxicol. 19, 179 –201. Green, H. (1978) Cell 15, 801– 811. Niimura, M., Yamamaoto, K., Kishimoto, S., Oharra, K., and Ogawa, N. (1990) Jpn. Pharmacol. Ther. 18, 2757–2770. Voorhees, J. J., and Duell, E. A. (1971) Arch. Dermatol. 104, 352–358. Chambers, D. A., Marcelo, C. L., and Voorhees, J. J. (1977) in Clinical Aspects of Cyclic Nucleotides (Volicer, L., Ed.), pp. 407– 427, Spectrum Press, New York. Kooistra, T., Bosma, P. J., Toet, K., Cohen, L. H., Griffioen, J., van den Berg, E., le Clercq, L., and van Hinsbergh, V. W. M. (1991) Arteriosclerosis Thromb. 11, 1042–1052. Higgins, P. J., Ryan, M. P., and Providence, K. M. (1994) J. Cell. Physiol. 159, 187–195.
HC REGULATES KERATINOCYTE PLASMINOGEN ACTIVATION 45. 46. 47. 48. 49. 50.
51. 52. 53.
54. 55. 56.
57. 58. 59. 60.
Granelli-Piperno, A., and Reich, E. (1978) J. Exp. Med. 148, 223–234. Isseroff, R. R., and Rifkin, D. B. (1983) J. Invest. Dermatol. 80, 297–299. Burge, S. M., Marshall, J. M., and Cederholm-Williams, S. A. (1992) Br. J. Dermatol. 126, 35– 41. Skriver, L., Nielsen, L. S., Stephens, R., and Danø, K. (1982) Eur. J. Biochem. 124, 409 – 414. Wun, T.-C., Ossowski, L., and Reich, E. (1982) J. Biol. Chem. 257, 7262–7268. Andreasen, P. A., Christensen, T. H., Huang, J.-Y., Nielsen, L. S., Wilson E. L., and Danø K. (1986) Mol. Cell. Endocrinol. 45, 137–147. Evans, R. M. (1988) Science 240, 889 – 895. Ponec, M. (1993) Curr. Probl. Dermatol. 21, 20 –28. Bruzdzinski, C. J., Riordan-Johnson, M., Nordby, E. C., Suter, S. M., and Gelehrter, T. D. (1990) J. Biol. Chem. 265, 2078 – 2085. van Zonneveld, A.-J., Curriden, S. A., and Loskutoff, D. J. (1988) Proc. Natl. Acad. Sci. USA 85, 5525–5529. Jonat,, C., Rahnsdorf, H. J., Park K. K., Cato, A. C., Gebel, S., Ponta, H., and Herlich, P. (1990) Cell 62, 1189 –1204. Yang-Yen, H. F., Chambard, J. C., Sun, Y. L., Smeal, T., Schmidt, T. J., Drouin, J., and Karin, M. (1990) Cell 62, 1205– 1215. Nerlov, C., Rorth, P., Blasi, F., and Johnsen, M. (1991) Oncogene 6, 1583–1592. Berthelsen, J., Vandekerkhove,, J., and Blasi, F. (1996) J. Biol. Chem. 271, 3822–3830. Feng, P., Ohlsson, M., and Ny, T. (1990) J. Biol. Chem. 265, 2022–2027. Reuning, U., Wilhelm. O., Nishiguchi, T., Guerrini, L., Blasi, F., Graeff, H., and Schmitt, M. (1995) Nucleic Acids Res. 23, 3887– 3893. Werb, Z., Mainardi, C. L., Vater, C. A., and Harris, E. D., Jr. (1977) N. Engl. J. Med. 296, 1017–1023. Quigley, J. P., Gold, L. I., Schwimmer, R., and Sullivan, L. M. (1979) Cell 17, 131–141.
Received October 3, 1997 Revised version received February 23, 1998
63. 64. 65. 66.
73. 74. 75. 76. 77. 78. 79.
Marchina, E., and Barlati, S. (1996) Int. J. Biochem. Cell Biol. 28, 1141–1150. Hayden, S. M., and Seeds, N. W. (1996) J. Neurosci. 16, 2307– 2317. Chapman, H. A. (1997) Curr. Opin. Cell Biol. 9, 714 –724. Carmeliet, P., Schoonjans, L., Kieckens, L., Ream, B., Degan, J., Bronson, R., De Vos, R., van den Oord, J. J., Collen, D., and Mulligan, R. C. (1994) Nature 368, 419 – 424. Carmeliet, P., and Collen, D. (1995) Thromb. Haemost. 74, 429 – 436. Carmeliet, P., Bouche, A., De Clercq, C., Janssen, S., Pollefeyt, S., Wyns, S., Mulligan, R. C., and Collen, D. (1995) Ann. N. Y. Acad. Sci. 748, 367–381. Lijnen, R., Moons, L., Beelen, V., Carmeliet, P., and Collen, D. (1995) Thromb. Haemost. 74, 1126 –1131. Bugge, T. H., Flick, M. J., Danton, M. J. S., Daugherty, C. C., Rømer, J., Danø, K., Carmeliet, P., Collen, D., and Degen, J. L. (1996) Proc. Natl. Acad. Sci. USA 93, 5899 –5904. Bugge, T. H., Kombrinck, K. W., Flick, M. J., Daugherty, C. C., Danton, M. J. S., and Degen, J. L. (1996) Cell 87, 709 –719. Dewerchin, M., Nuffelen, A. V., Wallays, G., Bouche, A., Moons, L., Carmeliet, P., Mulligan, R. C., and Collen, D. (1996) J. Clin. Invest. 97, 870 – 878. Rømer, J., Bugge, T. H., Pyke, C., Lund, L. R., Flick, M. J., Degen, J. L., and Danø, K. (1996) Nature Med. 2, 287–292. Schaffer, C. J., and Nanney, L. B. (1996) Int. Rev. Cytol. 169, 151–181. Hamilton, J. A., Vairo, G., Knight, K. R., and Cocks, B. G. (1991) Blood 77, 616 – 627. Hart, P. H., Whitty, G. A., Burgess, D. R., Croatto, M., and Hamilton, J. A. (1990) Lymphokine Res. 9, 147–153. Liebovich, S. J., and Ross, R. (1975) Am. J. Pathol. 84, 501–513. DePetro, G., Copeta, A., and Barlati, S. (1994) Exp. Cell Res. 213, 286 –294. Hennings, H., Michael, D., Cheng, C., Steinert, P., Holbrook, K. A., and Yuspa, S. H. (1980) Cell 19, 245–254.