Induction of matrix metalloproteinases in keratinocytes ... - Springer Link

41 downloads 0 Views 102KB Size Report
sue homogenates of cholestea- ... sue extract was similarly inhibited by cytokine antibodies. .... Lowry OH, Rosebrough NNJ, Farr AL, Randall RJ (1951) Pro-.
Eur Arch Otorhinolaryngol (2000) 257 : 425–429

© Springer-Verlag 2000

O TO L O G Y

Marianne Schmidt · Petra Grünsfelder · Florian Hoppe

Induction of matrix metalloproteinases in keratinocytes by cholesteatoma debris and granulation tissue extracts Received: 8 October 1999 / Accepted: 8 March 2000

Abstract Although it is generally accepted that destruction and remodeling of temporal bone associated with middle ear cholesteatoma is mainly caused by the action of osteoclasts, it has been shown that neutral collagenases also play a role in predigesting the osteoid layer and exposing the mineralized bone to osteoclastic activity. Here we show that gelatinase B (matrix metalloproteinase-9) is over-expressed in cholesteatoma compared to external ear canal skin (EACS). Expression of MMP-9 in cholesteatoma mainly occurs in suprabasal layers, and more rarely in basal layers of cholesteatoma epithelium, as well as in inflammatory cells of the perimatrix. We further analyzed the influence of cholesteatoma debris, cholesteatoma granulation tissue, and cholesteatoma components such as keratin, cholesterol and bacterial endotoxin on the expression of MMPs in EACS keratinocytes. We show that cholesteatoma debris and granulation tissue extract both induced the secretion of MMP-9 by EACS keratinocytes, while keratin, bacterial lipopolysaccharide (LPS) or cholesterol did not show any effect. We further performed co-incubation and immunoprecipitation experiments using neutralizing interleukin-1α, EGF, TGF-β, TGF-α, interleukin-6 and TNF-α antibodies. Inhibition of MMP-9 up-regulation by debris or granulation tissue extract could be revealed with diverse cytokine antibodies. The results are discussed with regard to previously published studies. Key words Metalloproteinases · Cholesteatoma · Zymography · Cytokine

Introduction

modeling in middle ear cholesteatoma. Today, it is obvious that bone destruction in middle ear cholesteatoma is caused by the action of osteoclasts. Proteases such as plasminogen activators and matrix metalloproteinases (MMPs), however, have also been mentioned as playing a role in the degradation of the osteoid layer [4] or in the exposure of epithelial cells to the various cytokines of granulation tissue by producing gaps in the basal membrane [15]. MMPs are a family of zinc metalloenzymes secreted as latent proenzymes and activated by proteolytic cleavage. Roles in inflammatory processes with bone remodeling, such as periodontitis among others, have been described for MMP-9 (92 kDa gelatinase) and MMP-2 (72 kDa gelatinase), with both enzymes degrading basement membrane type IV collagens (for review see [3]). In this study we examined the expression of MMPs in cholesteatoma and external ear canal skin (EACS) extracts, using quantitative zymography, and found increased expression of MMP-9 in cholesteatoma. Furthermore we tried to establish the factors responsible for MMP-9 upregulation in EACS keratinocytes.

Materials and methods Cell culture Keratinocytes were prepared from EACS samples. In brief, tissue samples were incubated for 12 h in Dispase (Boehringer Mannheim, Germany; 25 U/ml). Subsequently, epithelial sheets were dissected manually, and dissociated with trypsin (GIBCO, Eggenstein, Germany). Dissociation was stopped with soybean trypsin inhibitor (GIBCO, Eggenstein, Germany; 10 mg/ml), and cells were cultivated in keratinocyte-SFM (serum free medium for keratinocytes; GIBCO, Eggenstein, Germany) at 37 °C and 5% CO2.

During the recent decade it has become evident that proteolytic activity plays a role in the pathology of bone re-

Extracts, samples and antibodies

M. Schmidt () · P. Grünsfelder · F. Hoppe University of Wuerzburg, Department of Otorhinolaryngology, Josef-Schneider-Straße 11, 97080 Wuerzburg, Germany Tel.: +49-931-2012365, Fax: +49-931-2012248 e-mail: [email protected]

Thirty cholesteatoma specimens and eight EACS samples were collected during surgery. Tissue samples for zymographic analysis were shock-frozen and homogenized in RIPA buffer (PBS, containing 1% NP40, 0,5% sodium deoxycholate, 0.1% SDS). Tissue homogenates were centrifuged at 15,000 g for 30 min. The protein content of cell lysates was determined using the method of Lowry [10] (Biorad, Muenchen, Germany).

426 For in vitro experiments debris and granulation tissue were microscopically dissected, shock-frozen and stored at –80 °C as separate samples. Pooled samples were freeze-dried, homogenized with a power homogenizer, and re-dissolved in sterile PBS. Insoluble particles were sedimented by centrifugation. The protein content of homogenates was determined according to the method of Lowry [10]. Human or bovine keratin, Salmonella typhimurium lipopolysaccaride (LPS) and cholesterol were purchased from Sigma (Deisenhofen, Germany). Keratin was dissolved at 100 mg/ml in 8 M-urea and dialyzed overnight into sterile PBS. For immunoprecipitation the following antibodies were used: interleukin 1α, EGF, TGF-α, TNF-α, interleukin-6, TGF-β (R&D systems, Wiesbaden, Germany). For immunoprecipitation controls, total goat immunoglobulins (R&D systems, Wiesbaden, Germany) were used. In vitro assays and immunoprecipitation Keratinocytes were seeded at a density of 6 × 104 cells/well. Cells were incubated with debris or granulation tissue extract, 1, 5 or 10 mg/ml; keratin 1 and 10 mg/ml; LPS 0.1 µg/ml, 10 µg/ml or 100 µg/ml; cholesterol 3.75 or 7.5 µg/well. All constituents were diluted in keratinocyte-SFM. For antibody inhibition experiments debris or granulation tissue samples at a concentration of 5 mg/ml were incubated with 5 µg specific antibody or 5 µg total goat immunoglobulins as control for 3 h and subsequently applied to the cells for 24 h. For immunoprecipitation, samples were incubated with 1 µg antibody or 1 µg goat immunoglobulins for 3 h and subsequently immunoprecipitated overnight at 4 °C with 20 µl protein A/G agarose (Santa Cruz Biotechnology, Calif.). Antibody precipitates were pelleted at 1000 g for 5 min. The samples, immunodepleted of single cytokines, were applied directly to the cells for 24 h. Thereafter the incubation medium was changed and the incubation continued for an additional 24 h. The supernatants of these cells were harvested for zymography. The cells were examined microscopically every 12 h. Protease zymography Twenty microlitres of conditioned media or 14.4 µg cholesteatoma/ EACS protein extracts were subjected to electrophoresis in 10% SDS-polyacrylamide gels under non-reducing conditions [8] containing 1 mg/ml gelatin (Sigma, Deisenhofen, Germany). After electrophoresis gels were renaturated 2 × 30 min in 2.5% Triton X-100 and developed overnight in developing solution (gelatin zymography: 50 mM-Tris HCl, pH 6.8, 0.2 M NaCl, 10 mM-CaCl2, 0.02% Brij 35) at 37 °C. Subsequently they were stained with Coomassie Brilliant Blue, destained and dried. Zymograms of cholesteatoma/skin extracts were evaluated densitometrically (Gel-Pro Analyzer; Media Cybernetics, Silver Spring, Md). Regression analysis was performed using SPSS. The Mann-Witney test was used for statistical evaluation.

Fig. 1 Comparison of six tissue homogenates of cholesteatoma and EACS samples by gelatin-zymography. Molecular weights are indicated in kDa on the left. HT positive control is conditioned culture supernatant of HT-1080 fibrosarcoma cells

Immunolocalization Frozen sections were fixed and stained by the indirect immunoperoxidase method according to standard protocols. The primary antihuman MMP-9 antibody was purchased from R&D systems (Wiesbaden, Germany).

Results Gelatinolytic activity of tissue extracts and immunolocalization in tissue sections Zymographic analysis of cholesteatoma and EACS samples produced lytic bands at 92 kDa (MMP-9), 72 kDa (MMP-2), 66 kDa (activated form of MMP-2), and, in a few samples, 84 kDa (activated form of MMP-9). Higher molecular weight bands at 130 kDa and above 200 kDa were present in all tissue samples, probably representing MMP-9-containing complexes [16]. Representative zymographies for cholesteatoma and EACS samples are shown in Fig. 1. Quantification of lytic activities revealed significantly higher MMP-9 (P = 0.000) values in cholesteatoma samples (n = 30) when compared with EACS samples (n = 8). MMP-2 values did not reveal statistically significant differences in activity in cholesteatoma and EACS samples (P = 0.538). Immunolocalization with a monoclonal MMP-9 antibody demonstrated MMP-9 expression in the epithelium of cholesteatoma (Fig. 2). MMP-9 sublocalization in cholesteatoma was usually confined pericellularly to the suprabasal layers of most specimens. MMP-9 was also expressed in inflammatory cells of the perimatrix (arrowheads). In vitro experiments Incubation of keratinocytes with lysates of cholesteatoma debris or granulation tissue resulted in a dose-dependent up-regulation of MMP-9 (Fig. 3). MMP-9 and MMP-2 basic levels produced by keratinocytes varied somewhat from experiment to experiment. MMP-2 secretion was not essentially altered by debris or granulation tissue extract. No up-regulation occurred with various concentrations of keratin, bacterial endotoxin, or cholesterol (Fig. 3).

427 Fig. 2 Immunohistological staining of cholesteatoma with MMP-9 antibody. Arrowheads see text, bars 100 µm

Fig. 3 a–d Zymographic analysis of cell culture supernatants of stimulated EACS keratinocytes: 20 µl conditioned culture medium was subjected to electrophoresis. Co control, non-treated cells; concentrations indicated in mg/ml for debris, granulation tissue extract and keratin. Concentrations of LPS were 0.1, 10 and 100 µg/ml. Cholesterol was applied at 3.75 or 7.5 µg/well (60,000 cells). Molecular weights are indicated in kDa on the left

MMP-9 up-regulation induced by cholesteatoma debris or granulation tissue extract could be reduced by cytokine antibodies. Up-regulation caused by debris extract could be reversed most effectively by using neutralizing EGF-antibodies. An inhibitory effect was also obtained by blocking of interleukin-1-α, TNF-α and TGF-α and TGFβ activity either by antibody incubation or by immunodepletion (Fig. 4). Incubation with goat immunoglobulins as an immunoprecipitation control for the polyclonal goat antibodies did not show an effect on MMP-9 production (data not shown).

Increased MMP-9 secretion caused by granulation tissue extract was similarly inhibited by cytokine antibodies. As in the debris experiments. the strongest down-regulation of MMP-9 secretion was revealed with EGF neutralizing antibodies (Fig. 4). Like the debris samples the other cytokine antibodies also showed an effect, at least in immunodepletion experiments. In general, immunodepletion was more effective in inhibiting MMP-9 up-regulation by tissue extracts than antibody incubation.

Discussion The role of MMPs in the pathology of middle ear cholesteatoma is still controversial. The expression of diverse metalloproteinases has been examined by Schönermark et al. [15]. These authors confined the expression of MMP-9 and MMP-2 to basal and suprabasal epithelial cell layers of cholesteatoma by immunohistochemical

428

Fig. 4 Zymographic analysis of cell culture medium of stimulated EACS keratinocytes: 20 µl conditioned culture medium was subjected to electrophoresis. Co control, non-treated cells; D/G cells treated with 5 mg/ml of either debris (D) or granulation tissue (G) extract. Antibodies used for experiments are indicated in the top lane. Molecular weights are indicated in kDa on the left. AB cells were incubated with neutralizing antibodies; IP debris or granulation tissue (Gran. T.) extracts were immunoprecipitated with the antibodies indicated in the top lane before incubation with the keratinocytes

staining. This is in agreement with our observations. No staining was detected in middle ear mucosa and tympanic membrane samples [15]. In a more recent study Banerjee et al. [2] performed zymographic analysis of culture supernatants of ten cholesteatoma and four meatal skin samples and found no statistically significant differences between controls and cholesteatoma samples. In contrast to these authors [2], we analyzed tissue homogenates of cholesteatoma and EACS samples for the expression of MMP-9 and MMP-2. In our study cholesteatoma extracts clearly contained higher amounts of MMP-9. These different observations might be explained by the different experimental settings: while Banerjee et al. [2] examined the in vitro activity of resected tissue specimens, our analysis recorded gelatinases produced in intact living tissue. Immunohistological staining mainly confined MMP-9 staining to the basal layers of the cholesteatoma epithelial cells, as described previously [15]. In order to analyze which components of the cholesteatoma tissue induce epithelial cells to produce MMP-9 we incubated cultured EACS keratinocytes with extracts of cholesteatoma debris and granulation tissue. The two tissue extracts had a striking effect on MMP-9 production. The major debris components, keratin, bacterial LPS and cholesterol, however, were not able to induce gelatinase secretion from keratinocytes directly. In contrast, fibroblasts were much more insensitive to the extracts with respect to the induction of MMP-9 secretion (data not shown). The lack of influence of keratin, cholesterol and bacterial endotoxin on MMP-9 secretion in keratinocytes indicates that induction may take place via substances secreted by different cell types present in the granulation tissue. Similar inductive events were described for gingival fibroblasts in peri-

odontitis, in which induction by bacterial endotoxin was initiated by activation of macrophages [7]. In vitro, incubation of keratinocytes with tissue extracts and anticytokine antibodies blocked the effects of cytokines already present in tissue extracts, as well as blocking autoinductive events. In order to exclude autoinductive events as much as possible, we performed immunodepletion of tissue extracts with cytokine antibodies and incubated cells with extracts with a reduced content of certain cytokines but without inhibitory antibodies. The antibody co-incubation and immunodepletion experiments revealed similar results. It is therefore likely, that granulation tissue extracts as well as debris extracts contain considerable amounts of different cytokines which can influence epithelial cells. The strongest inhibition of MMP-9 up-regulation in our in vitro experiments could be revealed with neutralizing EGF-antibodies. EGF is a strong stimulus for MMP-9 production in keratinocytes [11]. Increased expression of EGF in middle ear cholesteatoma has been described for epithelial cells [5, 9], and high EGF levels have been shown to be present in cholesteatoma debris [6]. However, it has to be taken into consideration antibodies that different binding affinities of blocking antibodies might be at least in part responsible for the strong inhibitory effect of EGF in vitro, especially since interleukin-1-α, TNF-α, TGF-α and TGF-β antibodies were also able to reduce MMP-9 secretion. Interestingly, both EGF [12] and MMP-9 [14] are strong markers for wound-healing processes. During wound healing EGF stimulates proliferation of keratinocytes and promotes motility, which has been shown to be coincident with induction of MMP-9 [13]. Increased MMP-9 secretion as a consequence of cytokine secretion in cholesteatoma might therefore be important for keratinocyte migration and granulation tissue remodeling in cholesteatoma pathology. The panel of cytokines and proteases participating in cholesteatoma pathology indicates a process resembling wound healing rather than a neoplastic process, also proposed previously by Albino et al. [1]. Summarizing our results, we demonstrated increased MMP-9 expression in cholesteatoma tissue compared to EACS. MMP-9 up-regulation in vitro is not directly triggered by the major debris components bacterial endotoxin, keratin or cholesterol. Up-regulation is probably caused by various cytokines present in debris and granulation tissue. A possible model for the MMP-9 up-regulating events in cholesteatoma epithelial cells involves an immunological response to bacterial products, stimulating macrophages to invade tissue and to produce cytokines. These cytokines may stimulate epithelial cells to produce proteases and cytokines. Inductive factors in MMP-9 upregulation by cholesteatoma debris and granulation tissue extract are EGF, interleukin-1-α, TNF-α, TGF-α and TGF-β, which probably at least partially originate from the epithelium itself and from underlying stromal cells.

429

References 1. Albino AP, Kimmelmann CP, Parisier SC (1998) Cholesteatoma: a molecular and cellular puzzle. Am J Otol 19:7–19 2. Banerjee AR, James R, Narula AA (1998) Matrix metalloproteinase-2 and matrix metalloproteinase-9 in cholesteatoma and deep meatal skin. Clin Otolaryngol 23:345–347 3. Birkedal-Hansen H (1993) Role of cytokines and inflammatory mediators in tissue destruction. J Periodont Res 28:500–510 4. Chambers TJ, Darby JA, Fuller K (1985) Mammalian collagenase predisposes bone surfaces to osteoclastic resorption. Cell Tissue Res 241:671–675 5. Goto Y (1990) Epidermal growth factor in cholesteatoma – the first report: the localization in the cholesteatoma tissue. Nippon Jibiinkoka Gakkai Kaiho 93:1186–1191 6. Goto Y (1990) Epidermal growth factor in cholesteatoma – the second report: the localization in the horny layer. Nippon Jibiinkoka Gakkai Kaiho 93:1192–1199 7. Hanazawa S, Nakada K, Ohmori Y, Miyoshi T, Amano S, Kitano S (1985) Functional role of interleukin-1 in periodontal disease: induction of interleukin-1 production by Bacteroides gingivalis lipopolysaccharide in peritoneal macrophages from C3H/HeN and C3H/HeJ mice. Infect Immunol 50:262–270 8. Heussen D, Dowdle EB (1980) Electrophoretic analysis of plasminogen activators in polyacrylamide gels containing sodium dodecyl sulfate and copolymerized substrates. Anal Biochem 102:196–202 9. Kojima H, Shiwa M, Kamide Y, Moriyama H (1994) Expression and localization of mRNA for epidermal growth factor and epidermal growth factor receptor in human cholesteatoma. Acta Otolaryngol 114:423–429

10. Lowry OH, Rosebrough NNJ, Farr AL, Randall RJ (1951) Protein measurement with the folin phenol reagent. J Biol Chem 193:265–275 11. Lyons JG, Birkedal-Hansen B, Pierson MC, Whitelock JM, Birkedal-Hansen H (1993) Interleukin-1b and transforming growth factor-a/epidermal growth factor induce expression of Mr 95,000 type IV collagenase/gelatinase and interstitial fibroblast-type collagenase by rat mucosal keratinocytes. J Biol Chem 268:19143–19151 12. Martin P, Hopkinson-Woolley J, McCluskey J (1992) Growth factors and cutanous wound repair. Prog Growth Factor Res 4:25–44 13. McCawley LJ, O’Brien P, Hudson LG (1998) Epidermal growth factor (EGF)- and scatter factor/hepatocyte growth factor (SF/HGF)-mediated keratinocyte migration is coincident with induction of matrix-metalloproteinase (MMP)-9. J Cell Physiol 176:255–265 14. Salo T, Makela M, Kylmaniemi M, Autio-Harmainen H, Larjava H (1994) Expression of matrix metalloproteinase-2 and -9 during early wound healing. Lab Invest 70:176–182 15. Schönermark M, Mester B, Kempf H-G, Bläser J, Tschesche H, Lenarz T (1996) Expression of matrix-metalloproteinases and their inhibitors in human cholesteatomas. Acta Otolaryngol 116:451–456 16. Wysocki AB, Staiano-Coico L, Grinell F (1993) Wound fluid from chronic leg ulcers contains elevated levels of metalloproteinases MMP-2 and MMP-9. J Invest Dermatol 101:64–68