Molecular cloning and characterization of canine metalloproteinase-9 gene promoter

Gene 273 (2001) 81±87

www.elsevier.com/locate/gene

Molecular cloning and characterization of canine metalloproteinase-9 gene promoter Sarah E. Campbell*, Lubna Nasir, David J. Argyle, David Bennett Department of Clinical Studies, Faculty of Veterinary Medicine, University of Glasgow, Glasgow, UK Received 21 February 2001; received in revised form 30 May 2001; accepted 14 June 2001 Received by A.J. van Wijnen

Abstract This paper describes the cloning and characterization of the canine matrix metalloproteinase-9 (MMP-9) gene promoter. The 5 0 untranslated region was obtained by genome walking upstream of the canine MMP-9 translation start site using canine genomic DNA as template. A DNA fragment of 1894 bp was isolated and on analysis demonstrated regions of sequence homology with the MMP-9 promoter sequences already determined for other species. In general, conserved regions correlated with DNA binding motifs such as a TATA-like box, AP-1 sites, GC boxes and a nuclear factor-kB binding domain. The DNA promoter fragment was suf®cient to drive basal expression of a luciferase reporter gene in Madin Darby canine kidney (MDCK) cells and to a lesser extent in feline embryonic ®broblast (FEA) cells. Activity of the promoter was enhanced by the treatment of transfected MDCK cells with phorbol 12-myristate 13-acetate but no effect was observed in the FEA cells. Promoter deletion studies revealed that regions of promoter were necessary for induction of reporter gene expression. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Gelatinase B; AP-1; Nuclear factor-kB; Madin Darby canine kidney cells

1. Introduction Mammalian matrix metalloproteinases (MMP) constitute a family of proteolytic enzymes acting together to degrade the components of extra-cellular matrices and have been implicated in many physiological processes including embryonic development and tissue remodelling (Matrisian, 1992; Vu and Werb, 2000). Furthermore, elevated levels of MMPs have been associated with pathological processes such as tumour invasion (Burke, 1999) and in¯ammatory processes such as arthritis (Martel-Pelletier, 1999). To date 14 genetically distinct MMPs have been identi®ed and divided into one of four different groups: collagenases, gelatinases, stromelysins and membrane-type MMPs. Abbreviations: bp, base pair(s); dNTP, deoxyribonucleoside triphosphate; EGF, epidermal growth factor; GCG, Genetics Computer Group; IL-1, interleukin-1; MgCl2, magnesium chloride; MMP, matrix metalloproteinase; PDGF, platelet derived growth factor; PMA, phorbol 12-myristate 13-acetate; RACE, rapid ampli®cation of cDNA ends; RT-PCR, reverse transcription polymerase chain reaction; TGF, transforming growth factor; TNF, tumour necrosis factor; UTR, untranslated region * Corresponding author. Molecular Therapeutics Research Group, Department of Clinical Studies, Faculty of Veterinary Medicine, University of Glasgow, Bearsden Road, Glasgow G61 1HQ, UK. Tel.: 144-141-3306918; fax: 144-141-330-6808. E-mail address: [email protected] (S.E. Campbell).

Of these enzymes the gelatinase group in particular has been intensively studied and shown to comprise two collagenases, MMP-2 (72 kDa, Gelatinase A) and MMP-9 (92 kDa, Gelatinase B), with substrate speci®city for aggrecan, elastin, vitronectin, denatured collagens and intact Type IV, V, VII, X, and XII collagens (Cawston, 1996). Regulation of the MMP-2 and -9 genes is a prerequisite for the fundamental understanding of the role of the gelatinases in both normal and pathological states. The MMP-9 gene has been cloned and sequenced in a number of different species and in some the promoter region has been isolated and analyzed, including human (Huhtala et al., 1991), rabbit (Fini et al., 1994), mouse (Masure et al., 1993) and rat (Eberhardt et al., 2000). Gelatinase B gene expression is inducible, regulated by a number of different agents including growth factors such as epidermal growth factor (EGF), platelet derived growth factor (PDGF), transforming growth factor (TGFb) and in¯ammatory cytokines such as tumour necrosis factor (TNF) and interleukin-1 (IL1). Phorbol myristate acetate (PMA), also shown to enhance the production of MMPs, mimics the action of both these growth factors and cytokines (Borden and Heller, 1997; Bond et al., 1998). The Gelatinase B promoter is unique in its requirement for the nuclear factor-kB (NF-kB) element for induction by in¯ammatory cytokines which

0378-1119/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0378-111 9(01)00573-X

82

S.E. Campbell et al. / Gene 273 (2001) 81±87

activates transcription factor NF-kB via a multi-protein complex (Meyer et al., 1991). The AP-1 DNA binding element also plays a pivotal role in the regulation of MMP expression by growth factors, cytokines and oncogenes (Benbow and Brinckerhoff, 1997). In this study we report the cloning and characterization of canine MMP-9 gene promoter and demonstrate the regions of the promoter required for basal and PMA stimulated expression in Madin Darby canine kidney (MDCK) and feline embryonic ®broblast (FEA) cells. 2. Materials and methods 2.1. Cell lines and reagents All tissue culture reagents were obtained from GibcoBRL unless stated otherwise. MDCK cells maintained in Dulbecco's MEM with gluatamax-1 medium (Cat No. 31966-021) supplemented with 10% foetal calf serum and cipro¯oxacin at 10 mg/ml (Bayer) were sub-cultured at a ratio of 1:9 every 3.5 days. FEA cells were maintained in Dulbecco's MEM medium (Cat No. 42430-025) supplemented with 10% foetal calf serum, 1% 100 mM sodium pyruvate and 1% 200 mM l-glutamine and were sub-cultured at a ratio of 1:4 every 3.5 days. Both cell lines were treated with the antibiotic penstrep and fungizone to prevent contamination. PMA (Sigma) at concentrations of 5 29 and 10 26 M were used to activate transcription from the promoter constructs transfected into MDCK and FEA cells, respectively, for the 24 h prior to the luciferase assays. 2.2. Cloning of canine MMP-9 gene promoter, 5 0 UTR The 5 0 ¯anking region of the canine MMP-9 gene was ampli®ed using the Universal GenomeWalkere Kit (Clontech); two walks were performed to obtain the whole promoter. Canine genomic DNA was prepared from canine whole blood using the phenol/chloroform extraction protocol, digested with four different restriction enzymes (StuI, DraI, PvuII and EcoRV), puri®ed and then ligated to adaptors according to the manufacturer's protocol to create four different canine libraries. Each library was then subjected to two rounds of polymerase chain reaction (PCR) (primary and secondary) using a Perkin-Elmer (PE) 2400 thermocycler. For the ®rst genome walk, the primary round of ampli®cation used antisense gene speci®c primer (GSP1) 5 0 tggaaagaccacaacggtgggcttgtg-3 0 followed by an internal antisense primer (GSP2) 5 0 -atagagcagcagcccagcaccaggaacac-3 0 in the secondary round of PCR. The second genome walk used antisense primers (GSP3) 5 0 -agggagagagttaaggctacaggactc-3 0 and (GSP4) 5 0 -taagtggtcagcctaagggcaagggat3 0 in the primary and secondary reactions, respectively; both primer sequences were based on the 5 0 region of the DNA fragment isolated from the ®rst walk. Sequence information from the promoter fragments enabled the cloning of a 1894 bp promoter fragment by PCR ampli®cation. Canine geno-

mic DNA (150 ng) samples in a total volume of 50 ml were subjected to PCR reactions containing 0.4 mM of both sense primer K9F 5 0 -ggtctgggtgactccaaagccaatgctcat-3 0 and antisense primer K9R 5 0 -ggtgagggtagtggtgtgtctagctactag-3 0 , 0.2 mM of dNTPs, 1.5 mmol/l MgCl2, and 2 units of Taq polymerase (QIAGEN, UK). The resulting products were cloned into pCR w2.1-TOPO vector (Invitrogen) following the manufacturer's instructions. Three individual PCR products were sequenced bi-directionally on an automated ABI 310 genetic analyzer (PE Applied Biosystems, UK). The nucleotide sequences were assembled and compared using the Genetics Computer Group (GCG) DNA sequence analysis package to produce a consensus sequence for the canine MMP-9 promoter. 2.3. Relative semi-quantitative reverse transcription PCR Reverse transcription PCR (RT-PCR) was used to assess endogenous levels of canine MMP-9 promoter activity in MDCK cells. Total RNA was harvested from both untreated and PMA stimulated MDCK cells using RNAzol B (AMS Biotechnology) following the manufacturer's protocol and then treated with DNA-freee (Ambion) to remove contaminating DNA from the RNA preparation. First strand synthesis reactions were performed using 2 mg of total RNA, moloney murine leukaemia virus reverse transcriptase (MMLV RT, GibcoBRL) and random primers. PCR reactions used cDNA samples as template in reactions containing 0.4 mM of both sense (f) and antisense (r) primer pairs (MMP9f, 5 0 -gctggacaaaaccaccctggaggccat-3 0 ; MMP-9r, 5 0 gtcgtcgaagtgggcgtctccctgaat-3 0 ; cyclof, 5 0 -cgtgctctgagtactggagagaaggga-3 0 ; cyclor, 5 0 -ccactcagtcttggcggtgcagatgaa-3 0 ) for amplifying regions of the canine MMP-9 and canine cyclophilin gene, respectively; the latter as an internal control. All reactions contained 0.2 mM dNTPs, 1.5 mmol/l MgCl2 and 2 units of Taq polymerase (GibcoBRL, UK). Samples were subjected to an initial denaturation at 958C for 5 min followed by 35 cycles of ampli®cation, each cycle consisting of a denaturation step of 958C for 1 min, an annealing temperature of 678C followed by an elongation step of 728C for 1 min. A ®nal elongation step of at 728C for 10 min completed the reaction. Samples were removed at 15, 20, 25, 30 and 35 cycles to determine the exponential/ linear phase of the PCR where basal and PMA treated samples could be semi-quanti®ed and compared. All PCR products were analyzed on a 1% agarose gel. 2.4. Determination of MMP-9 transcription initiation sites within both the plasmid vector and endogenous promoters The MMP-9 promoter/luciferase construct (prom1894) was transfected into MDCK cells, stimulated with PMA (5 29 M) and the cells harvested 24 h later. Total RNA was extracted using RNAzol B (AMS Biotechnology) and used as the template for RT-PCR using a GeneRacere Kit (Invitrogen). Reverse primers based on the luciferase and canine MMP-9 genes were used to amplify the region of transcrip-

S.E. Campbell et al. / Gene 273 (2001) 81±87

tion initiation within the plasmid vector and endogenous promoters, respectively. A single PCR product was obtained in each case and cloned into pCR w4-TOPO vector (Invitrogen). Ten positive clones for each were sequenced and the transcriptional start sites were mapped on the promoter sequence. 2.5. Construction of MMP-9/luciferase reporter vectors The largest canine MMP-9 promoter fragment/luciferase vector (prom1894) and the various promoter deletion/luciferase vectors (prom984, prom628, prom534, prom176 and prom102) were engineered using PCR techniques. Speci®c regions of the promoter sequence were chosen, based on motif location, to design primers containing restriction enzyme sites for sub-cloning into the luciferase reporter vector, pGL3-Basic (Promega). The MMP-9/pCR w2.1TOPO vector was used as template in standard PCR protocols. Details of primers and PCR conditions are available from the authors on request. 2.6. Expression of canine MMP-9/luciferase reporter vector constructs Ninety-six well plates were seeded with MDCK cells at a concentration of 6 £ 10 4 cells/ml and incubated overnight at 378C, 5% CO2. Transient transfections were carried out using Transfaste Reagent (Promega) at a 1:1 ratio with DNA (50 ng per well) according to the manufacturer's instructions. Half of the cells were exposed to PMA (5 29 M) for 24 h and all cells were assayed 48 h post-transfection. Dual-Luciferase w Reporter assays were performed according to the manufacturer's protocol (Promega). Experiments were conducted in triplicate for statistical signi®cance, and to ensure reproducibility all transfections were carried out three times. To account for transfection ef®ciency the cells were co-transfected with Renilla luciferase vector (Promega) and the Fire¯y luciferase values were adjusted accordingly. In the experiment all deletion constructs were analyzed together with the promoter-less luciferase vector, pGL3-Basic vector and the positive control vector, pGL3Control, containing the SV40 promoter and enhancer sequences. FEA cells were assayed in the same way. 3. Results and discussion 3.1. Cloning of canine MMP-9 promoter We describe the cloning of the 5 0 ¯anking region of the canine homologue of the MMP-9 gene by genome walking. Two PCR products were generated, the ®rst an approximately 550 bp fragment from the StuI digest library of the ®rst walk and a 1.5 kb product from the PvuII digest of the second walk. A fragment of the promoter (1894 bp) was ampli®ed from canine genomic DNA and the consensus sequence derived from three individual PCR products (Fig. 1).

83

3.2. Analysis of the 5 0 end regulatory region Comparison of the MMP-9 promoter sequence between the different species shows deviation from a common ancestral promoter region that has occurred through mutation to result in species speci®city. However, several conserved sequence motifs throughout the promoter were identi®ed that correlated with potential sites for the binding of transcription regulatory factors as determined by computational analysis of the canine 1894 bp DNA sequence using `MOTIF' (http://motif.genome.ad.jp/). These protein binding domains were thought to be of functional importance as they were found in regions of strong sequence conservation (Fig. 1). A CAAT motif was absent in all cases but a TATA-like box was consistently present, found in the canine sequence at position 251 to 247. Six AP-1 sites, the recognition sequence for members of the Fos and Jun families of transcription factors, were identi®ed within the canine sequence at various positions (273 to 267, 2111 to 2105, 2224 to 2216, 2877 to 2884, 21760 to 21750 and 21888 to 21878). The highly conserved AP-1 site (2111 to 2105) found throughout the species has been shown to be necessary for basal and induced promoter activity in human tumour cells (Sato and Seiki, 1993) with mutation abolishing all promoter activity (Gum et al., 1997). Another important binding domain is the NF-kB site found at position 2554 to 2545. This again is highly conserved throughout the species and acts synergistically with other motifs to participate in MMP-9 expression (Sato and Seiki, 1993), in particular the AP-1 site (Yokoo and Kitamura, 1996). Numerous GC boxes for the binding of transcription factor SP1 can be found at positions 286 to 278, 2516 to 2512, 2944 to 2935 and 21658 to 21648. The most distal site is conserved and present in most of the species. Another important element, the polyoma enhancer activator-3 (PEA3) motif that is recognized by the products of the Ets-1 and Ets-2 protooncogenes, was also found in the 5 0 ¯anking sequence and is known to function synergistically with the AP-1 sites in collagenase promoters (Gutman and Wasylyk, 1990). The microsatellite segment of alternating CA residues d(CA) sequence, identi®ed in the rat, human and mouse promoter, is absent in both the canine and rabbit sequences. The importance of this motif has yet to be determined since con¯icting studies show the requirement of d(CA) for up-regulating expression (Shimajiri et al., 1999) while others show no function at all (Sato and Seiki, 1993). The canine sequence also contains AP-2, SRY, GATA-1, GATA-2 and Lyf-1 sequences but the functional role of each regulatory element and their combined effect remains to be determined. 3.3. Determination of the transcription initiation site Sequence analysis of ten independent clones generated using the RACE technique from the plasmid vector and

84

S.E. Campbell et al. / Gene 273 (2001) 81±87

S.E. Campbell et al. / Gene 273 (2001) 81±87

85

Fig. 2. Relative semi-quantitative RT-PCR to determine the endogenous levels of canine MMP-9 expression in both basal and PMA stimulated cells. A portion of the MMP-9 gene (364 bp) was ampli®ed from cDNA prepared from total RNA isolated from MDCK cells. Samples were removed at ®ve cycle intervals to determine the linear phase of the PCR reaction (cycles 15±35). The basal and PMA induced PCR products over this range can be directly compared visually. A portion of the canine cyclophilin gene (255 bp) was ampli®ed as an internal control for relative RT-PCR. The cyclophilin primers spanned an intron and therefore screened for contaminating genomic DNA (lack of a 474 bp product).

endogenous promoter revealed multiple transcription initiation sites at positions 217, 221, 222, 224, 225, 240 and 255 bp relative to the ATG translation start codon (Fig. 1). However, the regions corresponding to transcription initiation throughout the species (Fig. 1) have been shown to originate from single sites at positions 219, 219, 219 and 224 bp upstream of the start site for translation in the rat (Eberhardt et al., 2000), rabbit (Fini et al., 1994), human (Huhtala et al., 1991) and mouse (Masure et al., 1993) sequences, respectively. 3.4. Endogenous expression of canine MMP-9 in MDCK cells The endogenous level of canine MMP-9 gene expression in MDCK cells was investigated using RT-PCR (Fig. 2). A portion of the canine MMP-9 gene (364 bp) was ampli®ed from cDNA samples prepared from basal and PMA treated MDCK cells. Samples removed at ®ve cycle intervals from 15 to 35 cycles showed that the exponential/linear phase of the MMP-9 gene ampli®cation occurred within this range and could be used to semi-quantify the levels of MMP-9 expression in these cells. Basal expression of the endogenous MMP-9 gene was evident in MDCK cells and treatment with PMA was shown to induce expression of this gene. A portion of the canine cyclophilin gene was also ampli®ed

(255 bp) as an internal control. The primers designed for PCR spanned an intron to control for the presence of contaminating genomic DNA (absence of a 474 bp PCR product). 3.5. Characterization of the canine MMP-9 promoter The canine MMP-9/luciferase reporter constructs as illustrated in Fig. 3 were transiently transfected into both canine MDCK and feline FEA cells (Jarrett et al., 1973). The luciferase values of each promoter deletion construct were directly compared to the SV40 (constitutive promoter) driven luciferase expression (positive control) (Fig. 4). In canine MDCK epithelial cells basal expression of the largest promoter fragment (prom1894) was found to be high, approximately one-third of the value of the positive control, con®rming studies showing that MMP-9 is normally expressed in cells speci®c to the kidney (Yokoo and Kitamura, 1996). Comparison of the basal luciferase expression levels between the various promoter deletions revealed a trend of maximal expression with prom628 containing the NF-kB motif indicating that it may serve to enhance transcription. This trend was observed in three independent experiments but the difference between the promoter constructs was not signi®cant. The promoter activity was not completely lost with the

Fig. 1. The nucleotide sequence of the 1894 bp canine promoter region (Accession number: AF280420) is shown in an alignment with the sequences, where available in GenBank, for other species including rabbit (L36050), human (M68343) mouse (X72794) and rat (AF148065). The numbering system corresponds to the canine sequence with position number 21 immediately upstream of the translation initiation site (ATG) indicated by a large arrow (B). The variable positions of transcription initiation within the canine sequence are shown for both the plasmid vector (.) and endogenous (*) sequences. The equivalent regions of transcribed sequence in the other species are underlined. The various DNA motifs within the canine sequence are highlighted in bold type and labelled throughout, including the TATA-like box, NF-kB domain, AP-1, GC box/SP1, AP-2, PEA3, SRY, Lyf-1, GATA-1 and GATA-2 sites. Where these domains are conserved they are highlighted according in the other species and additional sites unique to each species have been identi®ed in bold type. The d(CA) repeat sequence is indicated by capital letters. The positions of the end points used to construct the luciferase reporter vectors have been shown and are pre-®xed with `prom'.

86

S.E. Campbell et al. / Gene 273 (2001) 81±87

Fig. 3. Illustration of the largest canine MMP-9 promoter fragment (prom1894) and the various promoter deletion constructs (prom984, prom628, prom534, prom176 and prom102) used in the luciferase assays, with each number representing the size of the promoter in bp. The positions of the different protein binding motifs have been shown. SRY, w; AP-2, K; LYF-1, [; PEA3, 1 ; GATA-1, V; GATA-2, S.

removal of the conserved AP-1 site at position 2111 to 105 as described for the human promoter (Sato and Seiki, 1993). Since the canine sequence contains an additional AP-1 site (273 to 267) not found in the human counterpart we speculate that this extra AP-1 site may perhaps be suf®cient to drive basal expression. Similar results have been reported in the rat promoter with mutation of the conserved AP-1 (2111 to 2105) in the rat sequence only reducing basal expression by 25% (Eberhardt et al., 2000). Basal levels of expression from the four largest promoter constructs prom1894, prom984, prom628 and prom534 could be signi®cantly enhanced by the addition of PMA.

Comparison of the PMA luciferase expression levels between the various promoter deletions revealed a trend of maximal expression again with the prom628 construct. This trend was observed in three independent experiments but this time the difference between the promoter constructs was signi®cant. The constructs prom1894, prom984, prom628 and prom534 were signi®cantly higher than the smallest construct prom102 suggesting that the latter construct does not contain any inducible regions. In direct contrast there appeared to be very low levels of MMP-9 expression in feline FEA ®broblast cells where basal expression of the MMP-9 promoter was shown to be approximately one-sixth of the value of the SV40 driven luciferase expression. This was thought to be due to a cell-type dependence since species speci®city could be ruled out by studies showing activity of human promoter in rabbit cells (He, 1996). Cell-type-speci®c expression was thought to reside in the DNA sequence found upstream of 22722 bp in the mouse promoter (Munaut et al., 1999), and since only the 1894 bp of canine promoter DNA had been cloned it was not possible to assess cell-type speci®city in this study. The level of basal expression could not be enhanced with PMA in these cells. 4. Conclusions

Fig. 4. Promoter deletion/luciferase reporter assay results in (a) canine MDCK cells and (b) feline FEA cells. All six MMP-9 promoter deletion constructs were assayed and analyzed together with the promoter-less luciferase vector, pGL3-Basic and the positive control vector, pGL3-Control, containing the SV40 promoter and enhancer sequences. Both basal and PMA stimulated expression is shown. To ensure reproducibility, all transfections were carried out in triplicate three times. The corrected luciferase activity of each construct represents the mean ^ SEM (n ˆ 3). Statistical analysis of the MDCK results showed that basal activity was signi®cantly enhanced with PMA (*P , 0:05 using the Mann±Whitney test) with the four largest constructs. Comparison of the promoter deletion constructs using the Kruskal±Wallis test showed that there was no signi®cant difference in promoter activity between the basal promoter deletions but that the PMA stimulated constructs were signi®cantly different (P , 0:05). A follow-up comparison revealed that the PMA stimulated prom102 construct was signi®cantly lower than the four largest promoter constructs (prom1894, prom984, prom628 and prom534).

The metalloproteinases, in particular the gelatinases, have been shown to be important enzymes in tissue remodelling of many physiological and pathological processes. Understanding the regulation of this canine enzyme may be bene®cial for the advancement in understanding disease processes such as cancer and arthritis. This canine model not only has veterinary implications but may also serve as an animal model for human disease. Acknowledgements We thank Elizabeth Gault for providing technical support and Marian Scott for giving advice on the statistical analysis and acknowledge BSAVA Petsavers for funding this work. References Benbow, U., Brinckerhoff, C.E., 1997. The AP-1 site and MMP gene regulation: what is all the fuss about? Matrix Biol. 15, 519±526. Bond, M., Fabunmi, R.P., Baker, A.H., Newby, A.C., 1998. Synergistic upregulation of metalloproteinase-9 by growth factors and in¯amma-

S.E. Campbell et al. / Gene 273 (2001) 81±87 tory cytokines: an absolute requirement for transcription factor NFkappa B. FEBS Lett. 435, 29±34. Borden, P., Heller, R.A., 1997. Transcriptional control of matrix metalloproteinases and the tissue inhibitors of matrix metalloproteinases. Crit. Rev. Eukaryot. Gene Expr. 7, 159±178. Burke, F., 1999. Cytokines (IFNs, TNF-alpha, IL-2 and IL-12) and animal models of cancer. Cytokines Cell. Mol. Ther. 5, 51±61. Cawston, T.E., 1996. Metalloproteinase inhibitors and the prevention of connective tissue breakdown. Pharmacol. Ther. 70, 163±182. Eberhardt, W., Huwiler, A., Beck, K.-F., Walpen, S., Pfeilschifter, J., 2000. Ampli®cation of IL-1B-induced matrix metalloproteinase-9 expression by superoxide in rat glomerular mesangial cells is mediated by increased activities of NF-kB and activating protein-1 and involves activation of the mitogen-activated protein kinase pathways. J. Immunol. 165, 5788±5797. Fini, M.E., Bartlett, J.D., Matsubara, M., Rinehart, W.B., Mody, M.K., Girard, M.T., Rainville, M., 1994. The rabbit gene for 92-kDa matrix metalloproteinase. Role of AP1 and AP2 in cell type-speci®c transcription. J. Biol. Chem. 269, 28620±28628. Gum, R., Wang, H., Lengyel, E., Juarez, J., Boyd, D., 1997. Regulation of 92 kDa type IV collagenase expression by the jun aminoterminal kinase- and the extracellular signal-regulated kinase-dependent signalling cascades. Oncogene 14, 1481±1493. Gutman, A., Wasylyk, B., 1990. The collagenase gene promoter contains a TPA and oncogene-responsive unit encompassing the PEA3 and AP-1 binding sites. EMBO J. 9, 2241±2246. He, C., 1996. Molecular mechanism of transcriptional activation of human gelatinase B by proximal promoter. Cancer Lett. 106, 185±191. Huhtala, P., Tuuttila, A., Chow, L.T., Lohi, J., Keski-Oja, J., Tryggvason, K., 1991. Complete structure of the human gene for 92-kDa type IV collagenase. Divergent regulation of expression for the 92- and 72-kilodalton enzyme genes in HT-1080 cells. J. Biol. Chem. 266, 16485± 16490.

87

Jarrett, O., Laird, H.M., Hay, D., 1973. Determinants of the host range of feline leukaemia viruses. J. Gen. Virol. 20, 169±175. Martel-Pelletier, J., 1999. Pathophysiology of osteoarthritis. Osteoarthr. Cartilage 7, 371±373. Masure, S., Nys, G., Fiten, P., Van Damme, J., Opdenakker, G., 1993. Mouse gelatinase B. cDNA cloning, regulation of expression and glycosylation in WEHI-3 macrophages and gene organisation. Eur. J. Biochem. 218, 129±141. Matrisian, L.M., 1992. The matrix-degrading metalloproteinases. Bioessays 14, 455±463. Meyer, R., Hatada, E.N., Hohmann, H.P., Haiker, M., Bartsch, C., Rothlisberger, U., Lahm, H.W., Schlaeger, E.J., van Loon, A.P., Scheidereit, C., 1991. Cloning of the DNA-binding subunit of human nuclear factor kappa B: the level of its mRNA is strongly regulated by phorbol ester or tumour necrosis factor alpha. Proc. Natl. Acad. Sci. USA 88, 966±970. Munaut, C., Salonurmi, T., Kontusaari, S., Reponen, P., Morita, T., Foidart, J.M., Tryggvason, K., 1999. Murine matrix metalloproteinase 9 gene. 5 0 -upstream region contains cis-acting elements for expression in osteoclasts and migrating keratinocytes in transgenic mice. J. Biol. Chem. 274, 5588±5596. Sato, H., Seiki, M., 1993. Regulatory mechanism of 92 kDa type IV collagenase gene expression which is associated with invasiveness of tumour cells. Oncogene 8, 395±405. Shimajiri, S., Arima, N., Tanimoto, A., Murata, Y., Hamada, T., Wang, K.Y., Sasaguri, Y., 1999. Shortened microsatellite d(CA)21 sequence down-regulates promoter activity of matrix metalloproteinase 9 gene. FEBS Lett. 455, 70±74. Vu, T.H., Werb, Z., 2000. Matrix metalloproteinases: effectors of development and normal physiology. Genes Dev. 14, 2123±2133. Yokoo, T., Kitamura, M., 1996. Dual regulation of IL-1 beta-mediated matrix metalloproteinase-9 expression in mesangial cells by NFkappa B and AP-1. Am. J. Physiol. 270, F123±F130.