The metalloprotease-directed shedding of BP 180 (collagen XVII) from human keratinocytes in culture is unaffected by ceramide and cell-matrix interaction

ARTICLE

Collagen XVII, also designated as the 180-kDa bullous pemphigoid (BP) antigen or BP180, is a structural component of hemidesmosomes in epithelial cells [1]. Full length collagen XVII appears as a type II homotrimeric transmembrane protein with glycosylated collagenous carboxy-terminal extracellular domain which extends to the lamina densa of epidermal basement membrane [2] and a globular, disulfide-linked, amino-terminal cytoplasmic domain [3]. Collagen XVII was reported to play a pivotal function in maintaining linkage between the intracellular and the extracellular structural elements and in anchoring the epithelium to the underlying basement membrane. This concept is supported by pathological skin models. For instance, in bullous autoimmune skin diseases of the pemphigoid group, the presence of autoantibodies reactive with BP180 is associated with dermal-epidermal separation, and mouse passive transfer of BP180 antibodies also results in skin blistering [4]. Furthermore, it is now clearly established that several cases of generalized atrophic benign epidermolysis bullosa are associated with mutations of the BP180 gene [5].

Previous investigations have shown that collagen XVII can be isolated from human epidermis or keratinocyte cultures as a distinct molecular species corresponding to the 120-kDa triple-helical collagenous ectodomain of the full-length molecule [6]. This 120 kDa polypeptide has been isolated as an insolubilized form from normal skin [7] and is also present as a soluble form in keratinocytes culture medium [8]. Similarly, many type I or type II transmembrane proteins, including cell adhesion proteins, growth factors, cytokine receptors, or receptor ligands, are subjected to limited proteolysis, giving rise to soluble forms consisting of the entire extracellular domain of the precursor proteins [9, 10]. This process, called shedding, is now considered to represent an important aspect of cell regulation and cell-cell interaction [11, 12]. This selective post-translational proteolysis of transmembrane molecules from the cell surface is catalyzed by secretases or sheddases belonging to the metallo- or serine protease families [13].

Besides acting as matrix-degrading endopeptidases, matrix metalloproteases (MMPs) were recently shown to display other functions at the pericellular environment. Several members of that family can interact with receptors at the surface of normal or malignant cells as alpha_vbeta₃ for MMP₂ [14], or CD44 for MMP-9 [15] and, in keeping with their broad spectrum specificity, can degrade membrane proteins locally [16, 17].

We here demonstrate that BP-180 shedding from human keratinocytes in culture could be totally suppressed by supplementing culture medium with batimastat, a peptide-hydroxamate MMP inhibitor. However, keratinocytes in culture only expressed a low amount of MMP and enhancing their expression and/or activation by either increasing the intracellular level of ceramide or culturing cells on to matrix components had no further influence on BP-180 processing. Since 120 kDa fragment release was also impeded by a specific furin convertase inhibitor, we postulate that a protease belonging to the ADAM family is involved in BP180 shedding.

Material and methods

Cell culture

Human keratinocytes were isolated from normal human skin from healthy adults (age range: 20 years to 70 years). Epidermis was first mechanically separated from dermis and placed overnight at 4° C in 1% trypsin diluted in PBS without calcium and magnesium (Life technologies, Cergy Pontoise, France) but supplemented with 5 mug/ml gentamycin containing. Epidermis was separated from the remaining dermis and mechanically transformed into a cellular suspension. After centrifugation at 1,500 g for 10 min, the pellet was resuspended in 10 ml of keratinocyte-serum free medium (KFM, Life Technologies) supplemented with 20 ng/ml bovine pituitary extract and 0.2 ng/ml epidermal growth factor (r-EGF) (Life technologies). Primary cells were cultured in 75 cm² flasks (Nunc, Roskilde, Denmark).

For some experiments, keratinocytes, at subconfluency, were trypsinized and seeded onto T75 type IV collagen or type I collagen coated flasks (Becton Dickinson, France). Similarly, keratinocytes were seeded onto previously T75 flasks coated with purified laminin-5 (0.66 mug/cm²) (generously provided by Dr P. Rousselle, IBCP, Lyon, France) or mouse laminin-1 (0.33 mug/cm²) (Becton Dickinson).

Treatment of cell cultures

Subconfluent keratinocytes were grown in the presence of 50 mug/ml ascorbic acid (Sigma, Saint-Quentin-Fallavier, France), added to the media every 24 hrs, to allow for effective proline hydroxylation and stabilization of the collagen triple helix [6] and cells were treated with either neutral sphingomyelinase (Smase) (100 or 200 mU/mL from Staphylococcus aureus, Sigma) or 1 nM to 5 muM N-hexanoyl-D-sphingosine (C6-ceramide) (Sigma) for 48 hrs. Alternatively, subconfluent keratinocytes were treated with 100 mU/ml Smase and either with 5 muM batimastat (British Biotechnology, Oxford, UK) for 48 hrs or with a furin convertase inhibitor i.e. decanoyl-Arg-Val-Lys-Arg-chloromelthylketone (Dec-RVKR-CMK, 30 muM, Bachem Biochimie, France) added each day of culture for 48 hrs.

The viability of keratinocytes, following treatments, was assessed by the tryptan blue exclusion technique. After incubation cells were washed twice with phosphate-buffered saline (PBS) and cell proliferation was determined by the crystal violet assay as described [22]. Comparison of data, in quantitative terms, was evaluated by normalizing volume sample for either western blotting or zymographic analyses (see below) as equal cell number.

To assess the collagenous structure of the 120 kDa polypeptide, keratinocyte culture medium was digested with 0.5 U/ml of Clostridium histolyticum collagenase (Boehringer Mannheim, France) at 37° C for 15 min in 25 mM/l TrisHCL buffer (pH 7.6) containing 0.15 M NaCl and 4 mM CaCl₂. The reaction was stopped by adding 2 mM EDTA and medium was then dialyzed and freeze-dried.

Preparation of protein extracts for immunoblotting

For all experiments, the shed ectodomain of collagen XVII was isolated from the whole culture media, corresponding to a subconfluent keratinocyte culture in T75 flasks. Conditioned medium was first centrifuged at 1,500 g for 10 min to remove cellular debris, and the supernatant was then dialyzed against distilled water for 48 hrs at 4° C, and freeze-dried.

For analysis of cell-associated full-length collagen XVII, keratinocytes were extracted for 5 min on ice with 1 ml/75 cm² of a 65 mM TrisHCl (pH 6.8) buffer containing 2% SDS, 200 mM PMSF, 200 mM EDTA, 10 mug/ml pepstatin, 10 mug/ml antipain, 10 mug/ml leupeptin and 10 mug/ml chymostatin (all inhibitors were obtained from Sigma). The cell lysate was then scraped, and the extract was centrifuged at 10,000 g for 10 min at 4° C to remove cellular debris. For extraction of collagen XVII from skin samples, epidermis was separated from dermis at 56° C as previously reported [18] and then extracted as described above.

Immunoblotting for full length and shedded collagen XVII detection

For immunoblotting, proteins from epidermal or keratinocyte extracts and culture medium were separated on SDS-PAGE using 7.5% polyacrylamide gels under reducing conditions and transferred onto nitrocellulose membrane (Biorad, France). The membrane was then incubated for 2 hrs at 20° C with the primary antibody consisting of human sera from bullous pemphigoïd patients with clinically active disease. All sera were previously only shown to react with BP180 by immunoblotting. Human anti BP180 serum was diluted 1:30 (v/v) in 10 mM TrisHCl (pH 7.4) buffer containing 3% bovine serum albumin (Sigma), 0.5 M NaCl and 0.5% Tween 20. After a series of washings, the membrane was incubated at 20° C with horseradish peroxydase anti-human IgG, (Sanofi, France) diluted at 1:1,000 (v/v) in 10 mM Tris HCl (pH 7.4) containing 0.5% bovine serum albumin, 0.5 M NaCl and 0.5% Tween 20. Immunoreactive bands were revealed with diamino-benzidine (Sigma).

MMP-9 and MMP-1 detection by immunoblotting

Proteins from conditioned media were separated by SDS-PAGE using 10% polyacrylamide gels under reducing conditions and transfered onto nitrocellulose membrane. MMP-9 was detected using a human MMP-9 ProBlot™ Western Blotting Kit (France Biochem, Cergy-Pontoise, France). Proteins were labeled with a specific biotinylated mouse monoclonal anti-MMP-9 antibody. Bands were visualized by a streptavidine-peroxydase staining and chimioluminescence. MMP-1 was detected using an anti-MMP-1 antibody (Valbiotech, Paris, France) and an alkaline-phosphatase-linked secondary antibody followed by Sigma-fast revelation (Biorad).

Preparation of keratinocytes plasma membrane extracts for gelatin zymography

Cell layers were washed 3 times with cold 50 mM Tris-HCL (pH 7.6) containing 150 mM NaCl, 2 mM CaCl₂, 2 mM MgCl₂ (TBS), then incubated with TBS containing 1.5% Triton X-114 (Sigma) for 15 min at 4° C and scraped. Cellular extracts were then centrifuged at 10,000 g for 2 min. Supernatant was incubated at 37° C for 5 min and centrifuged at 5,000 g for 1 min to separate the detergent phase from the aqueous one containing gelatinases. This aqueous phase was further incubated with 20 mul gelatin-sepharose beads (Sigma) for 30 min at 4° C under stirring. Following centrifugation (5,000 g), beads were washed 3 times with TBS and gelatinases were directly eluted using 20 mul of electrophoretic sample buffer.

Gelatin and casein zymography analysis

MMP-2 and MMP-9 activities in conditioned media and in membrane extracts were analysed by zymography using gelatin as substrate. Gelatinases were separated under non reducing conditions on SDS-PAGE impregnated with 1 mg/ml gelatin (Sigma). After electrophoresis, gels were washed with 2.5% Triton X-100 for 30 min at 20° C and incubated overnight at 37° C in 50 mM Tris-HCL (pH 7.6) containing 5 mM CaCl₂ and 200 mM NaCl. They were then stained with 2% Coomassie blue G250 (w/v) and destained with methanol/acetic acid/water (40/10/50; v/v/v). Gelatinolytic activity was evidenced as clear bands against a blue background of stained gelatin.

Casein zymography was used for the detection of urokinase-type plasminogen activator (uPa). SDS-PAGE was performed using 12.5% polyacrylamide gels supplemented with alpha-casein (2 mg) and plasminogen (600 mug). After electrophoresis, gels were washed as described for gelatin zymography and incubated at 37° C in a buffer containing 100 mM glycine, 10 mM EDTA (pH 8.3) and further stained and destained.

Results and discussion

Detection of negative and secreted forms of BP180/collagen XVII in keratinocytes

Extracts from human epidermis and keratinocytes as well as the corresponding cell culture media were first analyzed by western blot in order to identify native BP180 and processed forms. As previously reported [6] supplementation of cell culture medium with ascorbate proved to be an absolute requirement for proper collagen XVII synthesis and folding. A major 180 kDa protein species, corresponding to native BP180 was identified in tissue extracts (Fig. 1A, lane 1); a faint immunoreactive band running at a 120 kDa MW could be also discerned consistent with the reported in vivo constitutive shedding of BP180 [6, 7]. Unprocessed collagen XVII could be also visualized in cell extracts, but consistently and unexpectedly a 120 kDa immunoreactive band was also detected (Fig. 1A, lane 2) [7]. It might suggest that the 120 kDa collagen XVII fragment formed by pericellular proteolysis, might be recycled by keratinocytes; alternatively, BP180 might be partly processed intracellularly by furin convertases since this collagen molecule does contain a tribasic furin/PACE cleavage motif [6]. However, because of the presence of two other upper faint bands in cell extracts, it is not excluded that an unspecific processing could be involved. Culture medium from subconfluent keratinocytes contained a major 120 kDa immunoreactive species (Fig. 1A, lane 3) which run at 350 kDa when the sample was unboiled prior to electrophoresis (not shown).

To further ascertain the collagenous nature of this fragment, culture medium was submitted to a mild bacterial collagenase treatment; under these conditions, the 120 kDa species vanished with production of lower immunoreactive M.W. bands (data not shown) [7]. Shedding of BP180 increased with the length of culture but was similar whether keratinocytes originated from a 45 or a 70 year-old healthy individual (Fig. 1B).

Influence of batimastat on BP180 shedding from human keratinocytes in culture

It was suggested that BP180 proteolytic cleavage could be involved in allowing differentiation and/or migration of keratinocytes [7]. It is now well documented that members of the MMP family are directly implicated in those cell phenotypic modulations [19]. For example, MMP-1 is required for directing the migration of keratinocytes on type I collagen [20], cleavage of laminin-5 by MMP-2 was shown to induce the vertical migration of keratinocyte cell lines [21] and we recently reported the involvement of MMP-9 in keratinocyte growth [22]. To first investigate whether MMPs might be involved in BP180 shedding from keratinocytes, cell culture medium was supplemented with batimastat (5 muM) a wide spectrum peptide hydroxamate MMP inhibitor. Under those conditions, processing of BP180 was totally inhibited (Fig. 2).

Modulation of neutral protease expression and BP180 shedding from keratinocytes by ceramides

As a second set of experiments, we tried to modulate MMP expression and activation at aims to modify in parallel to BP180 shedding. For that purpose, we increased the ceramide (CER) level of keratinocytes by treating cells with Smase or truncated ceramides as previously reported [23]. We recently showed that CER could trigger signaling pathways known to be involved in neutral protease expression from cells of epithelial origin [24]. Indeed, when human keratinocytes were treated with Smase or C₂CER, urokinase, pro MMP-1 as well as pro MMP-9 expressions were significantly stimulated (Fig. 3A-C). Interestingly, increased uPA and MMP-9 expression was observed for C₂CER amount as low as 10 nM (Fig. 3A-C). Consistent with our previous reports [22], ceramides did not influence MMP-2 expression (Fig. 3C). However, whatever the conditions, MMPs were identified in conditioned media only as proforms.

It is now established that most MMPs exert their activity in the pericellular environment where they can be activated and bind to different sets of receptors. Keratinocytes plasma membranes were isolated and analysed for MMP content. Except MT1-MMP which was detected only as a 64 kDa proform (not shown), plasma membranes from control keratinocytes were devoid of any detectable MMP-1, MMP-2 or MMP-9 activity. However, when cells were treated with Smase or truncated ceramides, MMP-9 and MMP-2, in their pro and active forms, were found associated to keratinocytes plasma membranes (Fig. 4A). Nevertheless, despite such ceramide-mediated increased MMP activity, the level of BP180 shedding was unchanged (Fig. 4B).

BP180 shedding is influenced by a furin-convertase inhibitor but is insensitive to keratinocyte-matrix interactions

In addition to MMP, batimastat proved to also inhibit ADAMs, a family of transmembrane metalloproteinases, that contain both metalloproteinase and disintegrin domains. It is now recognized that a majority of shedding events implicate ADAM family, particularly TACE (ADAM-17) [25]. Since modulation of MMP expression and/or activation by ceramides did not modify BP180 shedding to any appreciable extent, we suspected that an ADAM could be involved in the processing of that collagen-type in cultured keratinocytes. To address this question either control or Smase-treated cells were cultured in the presence of decanoyl-RVKA-chloro-methylketone, a lipophilic furin/convertase inhibitor, which can inhibit that class of serine proteinase both intra- and extracellularly. When cell culture medium was supplemented with 30 muM of inhibitor, BP180 shedding was inhibited (Fig. 5A) [6]. However, inhibition was not as complete as found with batimastat, and attempts to increase the concentration of the lipophilic inhibitor led to cell toxicity and also lack of specificity. Nevertheless, these data suggested that ADAMs, known to be processed intracellularly by furin convertases, could be also implicated in BP 180 shedding from keratinocytes.

The "disintegrin domain" of ADAMs refers to a region of the molecule analogous to a group of snake venoms proteases known to bind to integrin at the platelet surface [26]. We thus hypothesized that keratinocyte-matrix interactions could interfere with BP180 processing. To investigate such a possibility, keratinocytes were cultured onto types I and IV collagens as laminin-1 and laminin-5, and extent of BP180 shedding was examined. Figure 5B shows that the level of secreted 120 kDa processed form was unchanged whatever the coating condition.

It must be emphasized that, contrary to most shedding events which necessitate PKC activating agents as Phorbol esters, BP180 processing is constitutive and unaffected neither by stress compounds as ceramides nor by keratinocytes interaction with different basement membranes components. We assume that it might play a prominent function in keratinocyte biology, probably by modulating cell differentiation, a possibility under current investigation.

CONCLUSION

Acknowledgements

This work was supported by CNRS (FRE 2260), the Region Champagne-Ardenne (Grant to W.H.) and funds provided by A.R.D. The authors thank Dr Patricia Rousselle (IBCP, Lyon, France) for kindly providing purified laminin-5.

Article accepted on 14/2/01

REFERENCES

1. Nievers MG, Schaapveld RQJ, Sonnenberg A. Biology and function of hemidesmosomes. Matrix Biology 1999; 18: 5-17.

2. Bedane C, McMillan Jr, Balding SD, Bernard P, Prost C, Bonnetblanc JM, Diaz LA, Eady RA, Giudice GJ. Bullous pemphigoid and cicatricial pemphigoid autoantibodies react with ultrastructurally separable epitopes on the BP180 ectodomain: evidence that BP180 spans the lamina lucida. Invest Dermatol 1997; 108: 901-7.

3. Giudice GJ, Emery DJ, Diaz LA. Cloning and primary structural analysis of the bullous pemphigoid autoantigen BP180. J Invest Dermatol 1992; 99: 243-50.

4. Liu Z, Diaz LA, Troy JL, Taylor A, Emery DJ, Fairley JA, Giudice GJ. A passive transfer model of the organ-specific auto-immune disease: bullous pemphigoid, using antibodies generated against the hemidesmosomal antigen BP180. J Clin Invest 1993; 92: 2480-8.

5. Tasanen K, Eble JA, Aumailley M, Schumann H, Baetge J, Tu H, Bruckner P, Bruckner-Tuderman L. Collagen XVII is destabilized by a glycine substitution mutation in the cell adhesion domain Col15*. J Biol Chem 2000; 275: 3093-9.

6. Schäke H, Schumann H, Hammami-Hauasli N, Raghunath M, Bruckner-Tudermann L. Two forms of collagen XVII in keratinocytes: a full-length transmembrane protein and a soluble ectodomain. J Biol Chem 1998; 273: 25937-43.

7. Hirako Y, Usukura J, Uematsu J, Hashimoto T, Kitajima Y, Owaribe K. Cleavage of BP180, a 180-kDa bullous pemphigoid antigen, yields a 120-kDa collagenous extracellular polypeptide. J Biol Chem 1998; 273: 9711-7.

8. Hendri HP, Kloosterhuis GJ, Heeres K, Van Der Meer JB, Jonkman MF. Bullous pemphigoid and linear IgA dermatosis sera recognize a similar 120 kDa keratinocyte collagenous glycoprotein with antigenic cross reactivity to BP180. J Invest Dermatol 1997; 108: 423-9.

9. Müllberg J, Rauch CT, Wolfson MF, Castner B, Fitzner JN, Otten-Evans C, Mohler KM, Cosman D, Black RA. Further evidence for a common mechanism for shedding of cell surface proteins. FEBS Letters 1997; 401: 235-8.

10. Müllberg J, Althoff K, Jostock T, Rose-John S. The importance of shedding of membrane proteins for cytokine biology. Eur Cytokine Netw 2000; 11: 27-37.

11. Peschon JJ, Slack JL, Reddy P, Stocking KL, Sunnarborg SW, Lee DC, Russel WE, Castner BJ, Johnson RS, Fitzner JN, Boyce RW, Nelson N, Kozlosky CJ, Wolfson MF, Rauch CT, Cerretti DP, Paxton RJ, March CJ, Black RA. An essential role for ectodomain shedding in mammalian development. Science 1998; 282: 1281-4.

12. Tokumaru S, Higashimaya S, Endo T, Nakagawa T, Miyagawa J, Yamamori K, Hanakawa Y, Ohmoto H, Yoshino K, Shirakata Y, Matsuzawa Y, Hashimoto K, Taniguchi N. Ectodomain shedding of epidermal growth factor receptor is required for keratinocytes migration in cutaneous wound healing. J Cell Biol 2000; 151: 209-20.

13. Arribas J, Coodly L, Vollmer P, Kishimoto TK, Rose-John S, Massagué J. Diverse cell surface protein ectodomains are shed by a system sensitive to metalloprotease inhibitors. J Biol Chem 1996; 271: 11376-82.

14. Brooks PC, Strömblad S, Sanders LC, Von Schalscha TL, Aimes RT, Stetler-Stevenson WG, Quigley JP, Cheresh DA. Localization of matrix metalloproteinase MMP-2 to the surface of invasive cells by interaction with integrin alphavbeta3. Cell 1996; 85: 683-93.

15. Yu Q, Stamenkovic I. Localization of matrix metalloproteinase 9 to the cell surface provides a mechanism for CD44-mediated tumor invasion. Genes Devel 1999; 13: 35-48.

16. Levi E, Fridman R, Miao HQ, Ma YS, Yayon A, Vlodavsky I. Matrix metalloproteinase 2 releases active soluble ectodomain of fibroblast growth factor receptor 1. Proc Natl Acad Sci 1996; 93: 7069-74.

17. Murphy G, Knäuper V, Atkinson S, Gavrilovic J, Edwards D. Cellular mechanisms for focal proteolysis and the regulation of the microenvironment. Fibrinolysis and Proteolysis 2000; 14: 165-74.

18. Bernard P, Didierjean L, Denis F, Saurat JH, Bonnetblanc JM. Heterogeneous bullous pemphigoid antibodies: detection and characterization by immunoblotting when absent by indirect immunofluorescence. J Invest Dermatol 1989; 92: 171-5.

19. Stamenkovic I. Matrix metalloproteinases in tumor invasion and metastasis. Sem Cancer Biol 2000; 10: 415-33.

20. Sudbeck PD, Parks WC, Welgus HG, Pentland AP. Collagen-mediated induction of keratinocyte collagenase is mediated by tyrosine kinase and protein kinase C activities. J Biol Chem 1994; 269: 30022-9.

21. Giannelli G, Falk-Marzillier J, Schiraldi O, Stetler-Stevenson WG, Quaranta V. Induction of cell migration by matrix metalloprotease-2 cleavage of laminin-5. Science 1997; 277: 225-8.

22. Buisson-Legendre N, Emonard H, Bernard P, Hornebeck W. Relationship between cell-associated matrix metalloproteinase 9 and psoriatic keratinocyte growth. J Invest Dermatol 2000; 115: 213-8.

23. Buisson-Legendre N, Bernard P, Bobichon H, Hemonard H, Schneider C, Maquart FX, Haye B, Hornebeck W. Involvement of the 92-kDa gelatinase (matrix metalloproteinase-9) in the ceramide-mediated inhibition of human keratinocyte growth. Biochem Biophys Res Commun 1999; 260: 634-40.

24. Maziere C, Conte MA, Leborgne L, Levade T, Hornebeck W, Santus R, Maziere JC. UVA radiation stimulates ceramide production: relationship to oxidative stress and potential role in ERK, JNK, and p38 activation. Biochem Bioph Res Com 2001; 281: 289-94.

25. Amour A, Slocombe PM, Webster A, Butler M, Knight CG, Smith BJ, Stephens PE, Shelley C, Hutton M, Knäuper V, Docherty AJP, Murphy G. TNF-alpha converting enzyme (TACE) is inhibited by TIMP-3. FEBS Letters 1998; 435: 39-44.

26. Reddy P, Slack JL, Davis R, Ceretti DP, Kozlovsky CJ, Blanton RC, Shows D, Peschon JJ, Black RA. Functional analysis of the domain structure of tumor necrosis factor-alpha converting enzyme. J Biol Chem 2000; 275: 14608-14.