Identification of the Membrane-Type Matrix Metalloproteinase MT1-MMP

Home , Matrix metalloproteinase, Metalloproteinase, Pronase

Journal of Cell Science 110, 589-596 (1997) 589 Printed in Great Britain © The Company of Biologists Limited 1997 JCS9564

Identiﬁcation of the membrane-type matrix metalloproteinase MT1-MMP in osteoclasts

Takuya Sato1, Maria del Carmen Ovejero1, Peng Hou1, Anne-Marie Heegaard1, Masayoshi Kumegawa2, Niels Tækker Foged1 and Jean-Marie Delaissé1,* 1Department of Basic Research, Center for Clinical & Basic Research, Ballerup Byvej 222, DK-2750 Ballerup, Denmark 2The First Department of Oral Anatomy, Meikai University School of Dentistry, Keyakidai 1-1, Sakado, Saitama 350-02, Japan *Author for correspondence

SUMMARY

The osteoclasts are the cells responsible for bone resorp- on bone sections showed that MT1-MMP is expressed also tion. Matrix metalloproteinases (MMPs) appear crucial in osteoclasts in vivo. Antibodies recognizing MT1-MMP for this process. To identify possible MMP expression in reacted with specific plasma membrane areas corre- osteoclasts, we amplified osteoclast cDNA fragments sponding to lamellipodia and podosomes involved, respec- having homology with MMP genes, and used them as a tively, in migratory and attachment activities of the osteo- probe to screen a rabbit osteoclast cDNA library. We clasts. These observations highlight how cells might bring obtained a cDNA of 1,972 bp encoding a polypeptide of MT1-MMP into contact with focal points of the extracel- 582 amino acids that showed more than 92% identity to lular matrix, and are compatible with a role of MT1- human, mouse, and rat membrane-type 1 MMP (MT1- MMP in migratory and attachment activities of the osteo- MMP), a cell surface proteinase believed to trigger cancer clast. cell invasion. By northern blotting, MT1-MMP was found to be highly expressed in purified osteoclasts when compared with alveolar macrophages and bone stromal Key words: Osteoclast, MMP-14, MT1-MMP, Membrane proteinase, cells, as well as with various tissues. In situ hybridization Podosome, Bone invasion, Bone resorption

INTRODUCTION lagenase clearly stained the resorption compartment (Delaissé et al., 1993). Also several other MMPs have recently been The osteoclasts are the cells responsible for the resorption of reported in osteoclasts (Hill et al., 1994), but it is not estab- the extracellular bone matrix. This role is essential for bone lished what might be the main MMP acting in the resorption morphogenesis, repair and maintenance. In order to fulfil this compartment. It is worth noting that all these proteinases are role, the (pre)osteoclast has to recognize, move, and anchor soluble, and it is believed that the sealing ring surrounding the itself to an appropriate site of the bone surface. Further, it has resorption compartment plays an important role for restricting to establish the resorption compartment, which involves a proteolysis therein. series of characteristic events, such as the formation of a tight Recent observations suggest that bone resorption is also seal all around it, the development of the ruffled border, and determined by proteolytic activities that occur before the the secretion of lytic agents. Thus bone resorption relies on a formation of the resorption compartment, and that are series of distinct osteoclastic activities (reviewed by Baron et required for osteoclast migration and anchorage. Indeed, al., 1993). MMPs proved to be indispensable for the recruitment of Because the removal of organic matrix in the resorption osteoclasts to future resorption sites (Blavier and Delaissé, compartment is a striking feature of bone resorption, much 1995), just as they are for many other cell migration effort has been done to identify the proteinases involved in this processes, such as angiogenesis (Fisher et al., 1994), metas- removal. It seems that only cysteine proteinases and matrix- tasis (Tryggvason et al., 1993), wound healing (Salo et al., metalloproteinases (MMPs) are rate limiting for this process 1994), embryo implantation (Librach et al., 1991) or neurite (Everts et al., 1992; Delaissé and Vaes, 1992). Various cysteine outgrowth (Muir, 1994). The molecular mode of action of proteinases have been implicated, but it appeared recently that MMPs in these processes is, however, still unclear. Another cathepsin K might be the key cysteine proteinase (Tezuka et interesting feature is that both osteoclasts and cancer cells al., 1994a; Li et al., 1995; Saneshige et al., 1995; Bossard et exhibit specialized membrane protrusions, called invadopo- al., 1996). High levels of MMP-9 in osteoclasts were repeat- dia or podosomes, establishing close-contact-like adhesions, edly stressed (Wucherpfennig et al., 1994; Reponen et al., and penetrating into the extracellular matrix (Baron et al., 1994; Tezuka et al., 1994b; Okada et al., 1995a; Blavier and 1993; Marchisio et al., 1984; Lakkakorpi and Väänänen, Delaissé, 1995) and antibodies recognizing interstitial col- 1991; Teti et al., 1991; Aubin, 1992). Podosomes of cancer 590 T. Sato and others cells were shown to degrade extracellular matrix molecules Sweden). Single strand cDNA was synthesized from mRNA by and, accordingly, various proteolytic activities were localized using a cDNA synthesis kit (Pharmacia). cDNA fragments were at their surface (Chen et al., 1984; Kelly et al., 1994). It was amplified by polymerase chain reaction (PCR) using degenerate also suggested that podosomes of osteoclasts bear proteinases primers corresponding to the conserved amino acid sequences in the on their surface (Teti et al., 1991; Aubin, 1992). However, cysteine switch (PRCGVPD) and in the catalytic domain neither such proteinases, nor any membrane proteinase have (GDXHFDXXE). The reactions were cycled 45 times through the following steps: 1 minute at 94¡C, 1 minute at 55¡C, 1 minute at ever been reported in osteoclasts. 72¡C. The PCR products were purified after agarose gel elec- The recent discovery of 4 MMPs with a transmembrane trophoresis. The purified cDNA was cloned into a pCRII vector domain (membrane-type MMPs, MT-MMPs) shows promise (Invitrogen, San Diego, CA) according to the instruction manual. for understanding how cell surface proteolysis is involved in The nucleotide sequences were determined from both strands by the invasive processes (Sato et al., 1994a; Will and Hinzmann, dideoxy chain-termination method. Analysis of the sequence data 1995; Takino et al., 1995; Puente et al., 1996). One of these, was performed as described (Devereux et al., 1984). The cDNA MT1-MMP (also called MMP-14), is overexpressed in insert contained in one of the clones shared 90% identity to a portion malignant tumour tissue, is localized on the surface of the (nt 271-648) of human MT1-MMP, and was used as a probe for cancer cells (Sato et al., 1994a), is able to degrade extracellu- screening of a rabbit osteoclast cDNA library. lar matrix molecules either directly (Pei and Weiss, 1996; Imai Isolation of MMP cDNA from osteoclast cDNA library et al., 1996) or indirectly (Sato et al., 1994a; Knaüper et al., A rabbit osteoclast cDNA library (Tezuka et al., 1994b) was screened 1996), and has been considered the trigger of cancer cell with the cDNA probe radiolabeled by using a multiprime DNA invasion (Sato et al., 1994a). labeling system (Amersham International plc., Buckinghamshire, To identify possible MT-MMPs expressed by osteoclasts, we England) and [α-32P]dCTP, as described elsewhere (Tezuka et al., used homology screening for MMPs in a rabbit osteoclast 1992). Isolated clones were converted to the plasmid form according cDNA library. We report here the isolation and the character- to the instruction manual (Stratagene, La Jolla, CA). The nucleotide istics of a cDNA encoding rabbit MT1-MMP, and demonstrate sequences were determined from both strands and the sequence data its expression in osteoclasts. Moreover, immunolocalizations were analyzed as described above. were consistent with an association of MT1-MMP to Amplification of cDNA 3′-ends podosomes and to lamellipodia of osteoclasts, giving insight in cDNA synthesized from rabbit osteoclast poly(A)+ RNA was the way invading cells direct proteolytic activity to specific amplified with the poly(dT) primer 5′-TAGAATTCTTTTTTTT- points of the extracellular matrix. TTTTTTTTTTTT-3′ and a 17 residue primer derived from the rabbit MT1-MMP sequence (1,803-1,819 in Fig. 1) by PCR. A cDNA fragment of 170 bp was cloned and sequenced as described MATERIALS AND METHODS above. Cells and organs for RNA preparation Northern blotting Osteoclasts were purified from long bones of 10-day-old rabbits Total RNA (5 µg) isolated from various organs and cells was blotted (Statens Seruminstitut, Copenhagen, Denmark) as described on nylon membranes after formaldehyde agarose gel electrophoresis, (Tezuka et al., 1992; Sato et al., 1994b; for a detailed protocol, see and hybridized with radioactive probes. A fragment of rabbit MT1- Helfrich et al., 1994). Briefly, the long bones were minced, the bone MMP cDNA (397-783 in Fig. 1), of human MT1-MMP (Sato et al., fragments were agitated, and the cells that were released (‘unfrac- 1994a, nucleotides 1,647-2,889, a kind gift from Drs H. Sato and M. tionated bone cells’) were cultured in plastic dishes in alpha-MEM Seiki, Kanazawa University) and a synthetic oligonucleotide corre- supplemented with 5% FBS. After overnight cultivation, the dishes sponding to 28 S ribosomal RNA were used as probes. cDNA probes were rinsed with phosphate buffered saline (PBS) and treated with were radiolabeled with a multiprime DNA labeling system pronase E (0.001%) and EDTA (0.02%) for 10 minutes, at 37¡C. (Amersham) using [α-32P]dCTP and the oligonucleotide probe was Repeated washings of the dishes resulted in the removal of stromal radiolabeled with 5′-end labeling kit (Amersham) using [γ-32P]ATP. cells, but could not release a cell population of which more than 95% Hybridization was performed as described previously (Sato et al., exhibit the typical characteristics of osteoclasts, i.e. they are multi- 1995) and visualized by a PhosphorImager SF (Molecular Dynamics, nucleated, exhibit high levels of tartrate resistant acid phosphatase Sunnyvale, CA). (TRAP), cathepsin K (OC-2), MMP-9 and vitronectin receptors, respond to calcitonin by contracting and producing cAMP, and In situ hybridization exhibit bone resorptive activity (Tezuka et al., 1992, 1994a,b; Consecutive paraffin sections of metacarpal bones of newborn rabbits Helfrich et al., 1994; Sato et al., 1994b; Kaji et al., 1994, 1996). were prepared as previously described (Blavier and Delaissé, 1995). Because this set of properties is unique to osteoclasts, these cells A fragment of rabbit MT1-MMP cDNA (1-318 in Fig. 1) was were definitively considered osteoclasts, and used in this study for subcloned into the pBluescript (KS) cloning vector (Stratagene) and RNA preparation. sense and antisense digoxygenin-labeled RNA probes were prepared When bone stromal cells were required, the cells removed during by use of a DIG RNA labeling kit (Boehringer Mannheim, Germany). the washings in the above procedure were cultured until confluence Hybridization was performed as described (Tezuka et al., 1994a), and in alpha-MEM containing 10% FBS, and then subcultured 4 times to visualized with a DIG nucleic acid detection kit (Boehringer deplete the cultures of TRAP+ cells. Brains, kidneys, livers, lungs, Mannheim) according to the instruction manual. TRAP staining was calvariae, spleens and alveolar macrophages were also isolated from performed as described (Sato et al., 1995). 10-day-old rabbits. In all cases total RNA was prepared as reported (Tezuka et al., 1992). Immunocytochemistry Unfractionated rabbit bone cells (see above) were seeded on glass Amplification of MMP cDNA fragments coverslips. After 1.5 hours cultivation, the non-adherent cells were Poly(A)+ RNA from purified osteoclasts was prepared using a discarded and the remaining cells were cultured for 0.5 to 3 hours, messenger RNA purification kit (Pharmacia Biotech, Uppsala, fixed and processed for immunocytochemistry (Delaissé et al., MT1-MMP in osteoclasts 591

CCG CTA GGA ATC CAA AGT CGG TGC CTC CGG AAG ACA AAG GCG CCC CCG AGG GAG 54 TGG CGG CGC GAC CCC TAG GCG AGG GCC CCG CCG CGG AAC CGC CCA GCC CGG CTG 108 CCC CGA CGG TCG CGG ACC ATG TCT CCC GCC CCA CGA CCC TCC CGC AGG CTC CTG 162 Met Ser Pro Ala Pro Arg Pro Ser Arg Arg Leu Leu 12 CTC CCC CTG CTC ACA CTC GGC ACC GCA CTC GCC TCC CTC GGC TCG GCC AAA AGC 216 Leu Pro Leu Leu Thr Leu Gly Thr Ala Leu Ala Ser Leu Gly Ser Ala Lys Ser 30 AAC AGC TTC AGC CCC GAA GCC TGG CTG CAG CAG TAT GGC TAC CTG CCT CCA GGG 270 Asn Ser Phe Ser Pro Glu Ala Trp Leu Gln Gln Tyr Gly Tyr Leu Pro Pro Gly 48 GAC CTA CGC ACC CAC ACA CAG CGC TCT CCT CAG TCA CTG TCA GCT GCC ATT GCT 324 Asp Leu Arg Thr His Thr Gln Arg Ser Pro Gln Ser Leu Ser Ala Ala Ile Ala 66 GCC ATG CAG AGG TTC TAC GGT TTG CGA GTG ACA GGC AAG GCC GAT ACA GAC ACC 378 Ala Met Gln Arg Phe Tyr Gly Leu Arg Val Thr Gly Lys Ala Asp Thr Asp Thr 84 ATG AAG GCC ATG AGG CGC CCC CGC TGC GGT GTT CCA GAC AAG TTT GGG GCT GAG 432 Met Lys Ala Met Arg Arg Pro Arg Cys Gly Val Pro Asp Lys Phe Gly Ala Glu 102 ATC AAG GCC AAT GTC CGA AGG AAG CGC TAC GCC ATC CAG GGC CTC AAA TGG CAG 486 Ile Lys Ala Asn Val Arg Arg Lys Arg Tyr Ala Ile Gln Gly Leu Lys Trp Gln 120 CAT AAT GAG ATC ACT TTC TGC ATC CAG AAT TAC ACC CCC AAG GTG GGC GAA TAT 540 His Asn Glu Ile Thr Phe Cys Ile Gln Asn Tyr Thr Pro Lys Val Gly Glu Tyr 138 GCC ACA TTC GAG GCC ATT CGC AAG GCA TTC CGC GTG TGG GAG AGC GCC ACA CCG 594 Ala Thr Phe Glu Ala Ile Arg Lys Ala Phe Arg Val Trp Glu Ser Ala Thr Pro 156 CTG CGC TTC CGC GAG GTG CAC TAT GCC TAC ATC CGC GAT GGC CGT GAG AAG CAG 648 Leu Arg Phe Arg Glu Val His Tyr Ala Tyr Ile Arg Asp Gly Arg Glu Lys Gln 174 GCC GAC ATC ATG ATC TTC TTT GCC GAG GGC TTC CAT GGC GAC AGC ACG CCC TTC 702 Ala Asp Ile Met Ile Phe Phe Ala Glu Gly Phe His Gly Asp Ser Thr Pro Phe 192 GAT GGC GAG GGT GGC TTC CTG GCC CAC GCC TAC TTC CCG GGC CCC AAC ATT GGA 756 Asp Gly Glu Gly Gly Phe Leu Ala His Ala Tyr Phe Pro Gly Pro Asn Ile Gly 210 GGG GAC ACC CAC TTT GAC TCC GCG GAG CCC TGG ACT GTC CGG AAT GAG GAC CTG 810 Gly Asp Thr His Phe Asp Ser Ala Glu Pro Trp Thr Val Arg Asn Glu Asp Leu 228 AAC GGG AAT GAC ATC TTC CTG GTG GCT GTG CAT GAG CTG GGC CAT GCC CTG GGC 864 Asn Gly Asn Asp Ile Phe Leu Val Ala Val His Glu Leu Gly His Ala Leu Gly 246 CTG GAG CAC TCC AAT GAC CCC TCA GCC ATC ATG GCA CCG TTT TAC CAA TGG ATG 918 Leu Glu His Ser Asn Asp Pro Ser Ala Ile Met Ala Pro Phe Tyr Gln Trp Met 264 GAC ACA GAG AAC TTC GTG CTG CCT GAT GAT GAC CGC CGG GGC ATC CAA CAG CTT 972 Asp Thr Glu Asn Phe Val Leu Pro Asp Asp Asp Arg Arg Gly Ile Gln Gln Leu 282 TAT GGG AGC CAG TCG GGG TCC CCC ACA AAG ATG CCT CCT CCA CCC AGG ACA ACC 1026 Tyr Gly Ser Gln Ser Gly Ser Pro Thr Lys Met Pro Pro Pro Pro Arg Thr Thr 300 TCC CGG ACT TTT ATC CCC GAT AAG CCC AGG AAC CCC ACC TAC GGG CCC AAC ATC 1080 Ser Arg Thr Phe Ile Pro Asp Lys Pro Arg Asn Pro Thr Tyr Gly Pro Asn Ile 318 TGT GAC GGG AAC TTT GAC ACT GTG GCC GTG CTC CGA GGA GAG ATG TTT GTC TTC 1134 Cys Asp Gly Asn Phe Asp Thr Val Ala Val Leu Arg Gly Glu Met Phe Val Phe 336 AAG GAG CGC TGG TTC TGG AGG GTG AGG AAC AAC CAA GTG ATG GAC GGC TAC CCA 1188 Lys Glu Arg Trp Phe Trp Arg Val Arg Asn Asn Gln Val Met Asp Gly Tyr Pro 354 ATG CCC ATC GGC CAG TTC TGG CGG GGC CTG CCT GCT TCC ATC AAC ACC GCC TAC 1242 Met Pro Ile Gly Gln Phe Trp Arg Gly Leu Pro Ala Ser Ile Asn Thr Ala Tyr 372 GAG AGG AAG GAT GGC AAA TTC GTC TTC TTC AAA GGA GAT AAG CAC TGG GTG TTT 1296 Glu Arg Lys Asp Gly Lys Phe Val Phe Phe Lys Gly Asp Lys His Trp Val Phe 390 GAC GAG GCT TCC CTG GAG CCT GGC TAC CCC AAG CAC ATC AAG GAG CTG GGC CGA 1350 Asp Glu Ala Ser Leu Glu Pro Gly Tyr Pro Lys His Ile Lys Glu Leu Gly Arg 408 GGG CTT CCC ACC GAC AAG ATC GAT GCC GCT CTC TTC TGG ATG CCC AAT GGA AAG 1404 Gly Leu Pro Thr Asp Lys Ile Asp Ala Ala Leu Phe Trp Met Pro Asn Gly Lys 426 ACC TAC TTC TTC CGG GGA AAC AAG TAC TAC CGA TTC AAC GAG GAG CTC AGG GCA 1458 Thr Tyr Phe Phe Arg Gly Asn Lys Tyr Tyr Arg Phe Asn Glu Glu Leu Arg Ala 444 GTG GAC AGC GAG TAC CCC AAG AAC ATC AAA GTG TGG GAA GGC ATC CCC GAG TCT 1512 Val Asp Ser Glu Tyr Pro Lys Asn Ile Lys Val Trp Glu Gly Ile Pro Glu Ser 462 CCC AGA GGG TCG TTC ATG GGC AGT GAT GAA GTC TTC ACT TAC TTC TAC AAG GGG 1566 Fig. 1. Reconstituted cDNA and deduced Pro Arg Gly Ser Phe Met Gly Ser Asp Glu Val Phe Thr Tyr Phe Tyr Lys Gly 480 amino acid sequence of rabbit MT1-MMP. AAC AAA TAC TGG AAA TTC AAC AAC CAG AAG CTG AAG GTG GAG CCC GGC TAC CCC 1620 The nucleotide sequence data have been Asn Lys Tyr Trp Lys Phe Asn Asn Gln Lys Leu Lys Val Glu Pro Gly Tyr Pro 498 submitted to GenBank and assigned the AAG TCC GCC CTG CGG GAC TGG ATG GGC TGC CCG GCT GGG GGC CGT CCG GAT GAG 1674 Lys Ser Ala Leu Arg Asp Trp Met Gly Cys Pro Ala Gly Gly Arg Pro Asp Glu 516 accession number U73940. The deduced GGG ACT GAG GAA GAG ACG GAG GTG ATC ATC ATC GAG GTG GAC GAG GAG GGC AGC 1728 amino acid sequence begins with the ﬁrst Gly Thr Glu Glu Glu Thr Glu Val Ile Ile Ile Glu Val Asp Glu Glu Gly Ser 534 ATG codon of the cDNA. PRCGVPD, the GGA GCC GTG AGC GCG GCC GCC GTG GTG CTG CCC GTG CTG CTG CTA CTC CTG GTG 1782 most conserved sequence among the MMP Gly Ala Val Ser Ala Ala Ala Val Val Leu Pro Val Leu Leu Leu Leu Leu Val 552 family is boxed. The Zn binding site is CTG GCC GTG GGC CTG GCG GTC TTC TTC TTC AGG CGC CAC GGG ACT CCG AAG CGA 1836 double underlined. The predicted Leu Ala Val Gly Leu Ala Val Phe Phe Phe Arg Arg His Gly Thr Pro Lys Arg 570 transmembrane domain at the C terminus is CTG CTC TAC TGC CAG CGT TCC CTG CTG GAC AAG GTC TGA CCC CCA CCG CTG GCC 1890 underlined. The nucleotides that served as Leu Leu Tyr Cys Gln Arg Ser Leu Leu Asp Lys Val * 582 primer for synthesis of the 3′-end are in CAC CCA CTC CCA CCG CAA GGA CTT TGC TCT TCC GAT TGT ATC CAA TAA AAA ATA 1944 bold. GCA TCA GCA AAA AAA AAA AAA AAA AAA A 1972 592 T. Sato and others

Signal peptide pro-peptide 60 Rabbit MSPAPRPSRRLLLPLLTLGTALASLGSAKSNSFSPEAWLQQYGYLPPGDLRTHTQRSPQS Human MSPAPRPSRCLLLPLLTLGTALASLGSAQSSSFSPEAWLQQYGYLPPGDLRTHTQRSPQS Rat MSPAPRPSRSLLLPLLTLGTTLASLGWAQSSNFSPEAWLQQYGYLPPGDLRTHTQRSPQS Mouse MSPAPRPSRSLLLPLLTLGTALASLGWAQGSNFSPEAWLQQFGYLPRGDLRTHTQRSPQT ********* ********** ***** * ********* **** ************ pro-peptide 120 Rabbit LSAAIAAMQRFYGLRVTGKADTDTMKAMRRPRCGVPDKFGAEIKANVRRKRYAIQGLKWQ Human LSAAIAAMQKFYGLQVTGKADADTMKAMRRPRCGVPDKFGAEIKANVRRKRYAIQGLKWQ Rat LSAAIAAIQRFYGLQVTGKADSDTMKAMRRPRCGVPDKFGTEIKANVRRKRYAIQGLKWQ Mouse LSVDIAAIQKFYGLYVTGKAYSETMKAMRRPRCGVPDKFGTEIKANVRRKRYAIQGLKWQ ** *** * **** ***** ***************** ******************* Catalytic 180 Rabbit HNEITFCIQNYTPKVGEYATFEAIRKAFRVWESATPLRFREVHYAYIRDGREKQADIMIF Human HNEITFCIQNYTPKVGEYATYEAIRKAFRVWESATPLRFREVPYAYIREGHEKQADIMIF Rat HNEITFCIQNYTPKVGEYATFEAIRKAFRVWESATPLRFREVPYAYIREGHEKQADIMIL Mouse HNEITFCIQNYTPKVGEYATFEAIRKAFRVWESATPLRFREVPYAYIREGHEKQADIMIL ******************** ********************* ***** * ******** Catalytic 240 Rabbit FAEGFHGDSTPFDGEGGFLAHAYFPGPNIGGDTHFDSAEPWTVRNEDLNGNDIFLVAVHE Human FAEGFHGDSTPFDGEGGFLAHAYFPGPNIGGDTHFDSAEPWTVRNEDLNGNDIFLVAVHE Rat FAEGFHGDSTPFDGEGGFLAHAYFPGPNIGGDTHFDSAEPWTVQNEDLNGNDIFLVAVHE Mouse FPEGLHGDSTPFDGEGGFLAHAYFPGPNIGGDTHFDSAEPWTVQNEDLNGNDIFLVAVHE * ** ************************************** **************** Catalytic Hinge 300 Rabbit LGHALGLEHSNDPSAIMAPFYQWMDTENFVLPDDDRRGIQQLYGSQSGSPTKMPPPPRTT Human LGHALGLEHSSDPSAIMAPFYQWMDTENFVLPDDDRRGIQQLYGGESGFPTKMPPQPRTT Rat LGHALGLEHSNDPSDIMAPFYQWMDTENFVLPDDDRRGIQQLYGSKSGSPTKMPPQPRTT Mouse LGHALGLEHSNDPSDIMSPFYQWMDTENFVLPDDDRRGIQQLYGSKSGSPTKMPPQPRTT ********** *** ** ************************** ** ****** **** Hinge Hemopexin 360 Rabbit SRTFIPDKPRNPTYGPNICDGNFDTVAVLRGEMFVFKERWFWRVRNNQVMDGYPMPIGQF Human SRPSVPDKPKNPTYGPNICDGNFDTVAMLRGEMFVFKERWFWRVRNNQVMDGYPMPIGQF Rat SRPSVPDKPRNPTYGPNICDGNFDTVAMLRGEMFVFKERWFWRVRNNQVMDGYPMPIGQF Mouse SRPSVPDKPKNPAYGPNICDGNFDTVAMLRGEMFVFKERWLWRVRNNQVMDGYPMPIGQF ** **** ** ************** ************ ********* * ********* Hemopexin 420 Rabbit WRGLPASINTAYERKDGKFVFFKGDKHWVFDEASLEPGYPKHIKELGRGLPTDKIDAALF Human WRGLPASINTAYERKDGKFVFFKGDKHWVFDEASLEPGYPKHIKELGRGLPTDKIDAALF Rat WRGLPASINTAYERKDGKFVFFKGDKHWVFDEASLEPGYPKHIKELGRGLPTDKIDAALF Mouse WRGLPASINTAYERKDGTFVFFKGDKHWVCVEASLEPGYANHIKELVRGLPSDKIDTALF ***************** *********** ******** ***** *** * **** *** Hemopexin 480 Rabbit WMPNGKTYFFRGNKYYRFNEELRAVDSEYPKNIKVWEGIPESPRGSFMGSDEVFTYFYKG Human WMPNGKTYFFRGNKYYRFNEELRAVDSEYPKNIKVWEGIPESPRGSFMGSDEVFTYFYKG Rat WMPNGKTYFFRGNKYYRFNEEFRAVDSEYPKNIKVWEGIPESPRGSFMGSDEVFTYFYKG Mouse WMPNGKTYFFRGNKYYRFNEEFRAVDSEYPKNIKVWEGIPESPRGSFMGSDEVFTYFYKG Fig. 2. Alignment of predicted amino ********************* **************************** * ********* acid sequences of rabbit, human, rat, and Hemopexin mouse MT1-MMPs. The sequences of 540 Rabbit NKYWKFNNQKLKVEPGYPKSALRDWMGCPAGGRPDEGTEEETEVIIIEVDEEGSGAVSAA human, mouse, and rat MT1-MMP were Human NKYWKFNNQKLKVEPGYPKSALRDWMGCPSGGRPDEGTEEETEVIIIEVDEEGGGAVSAA taken from Okada et al. (1995b). The Rat NKYWKFNNQKLKVEPGYPKSALRDWMGCPSGGRPDEGTEEETEVIIIEVDEEGSGAVSAA typical domains are indicated (signal Mouse NKYWKFNNQKLKVEPGYPKSALRDWMGCPSGGRPDEGTEEETEVIIIEVDEEGSGAVSAA peptide, propeptide, catalytic, hinge, ***************************** ******************** * ********* hemopexin-like and transmembrane Transmembrane domain 562 domains). Inserts speciﬁc to MT-MMPs Rabbit AVVLPVLLLLLVLAVGLAVFFFRRHGTPKRLLYCQRSLLDKV are boxed. Amino acids conserved in the Human AVVLPVLLLLLVLAVGLAVFFFRRHGTPRRLLYCQRSLLDKV Rat AVVLPVLLLLLVLAVGLAVFFFRRHGTPKRLLYCQRSLLDKV sequences of all species are indicated by Mouse AVVLPVLLLLLVLAVGLAVFFFRRHGTPKRLLYCQRSLLDKV asterisks. **************************** *************

1993), using either 3% BSA or 10% donkey serum as blocking labeled donkey anti-mouse IgG (Jackson ImmunoResearch Labora- agent. They were incubated for 90 minutes in the presence of 1-3 tories, Inc., West Grove, PA) was used as secondary antibody at 200 µg/ml of the monoclonal MT1-MMP antibody 113-5B7 (Sato et al., times dilution. When simultaneous detections of F-actin were 1994a; Yamada et al., 1995; Fuji Chemical Industries, Ltd, Takaoka, performed, 5 µg/ml fluorescein-labeled phalloidin (Sigma, St Louis, Japan). This antibody was raised against a synthetic peptide from MO) was added during the incubation with the secondary antibody human MT1-MMP (CDGNFDTVAMLRGEM) differing only by 1 (Marchisio et al., 1984). The specimens were observed by epifluo- amino acid from the corresponding rabbit sequence (V at position rescence on an Olympus microscope. Note also that we checked that 10 in rabbit instead of M in human). Controls were performed by the multinucleated cells of this preparation stained strongly for replacing the primary antibody with nonimmune IgGs. Rhodamine TRAP. MT1-MMP in osteoclasts 593 RESULTS 12 3 4 56789 Isolation and characterization of a cDNA encoding A -28S rabbit MT1-MMP To identify possible MMP expression by rabbit osteoclasts, the mRNA isolated from highly purified osteoclasts was reverse- -18S transcribed and amplified by PCR using degenerate primers designed from the conserved regions of MMP genes. Two B -28S major products, 330-340 and 380-390 bp in length, were identified by agarose gel electrophoresis. Since the high expression of MMP-9 mRNA by rabbit osteoclasts has been reported (Tezuka et al., 1994b) and the expected size of MMP-9 cDNA -18S fragments amplified with the degenerate primers used in this PCR is 336 bp, we cloned and sequenced the larger PCR C product. Database searching showed that one of the clones, TS- A3, contained a cDNA insert of 387 bp sharing more than 90% identity with a portion of the human MT1-MMP cDNA Fig. 3. Northern blot analysis of MT1-MMP mRNA levels in various sequence. Screening of the rabbit cDNA library with TS-A3 tissues and cells. Total RNA (5 µg) was separated by formaldehyde yielded one positive clone containing a cDNA insert of 1,842 agarose gel electrophoresis, blotted onto a nylon membrane, and bp. As the 3′-end of this cDNA was not complete, missing hybridized with 32P-labeled probes of rabbit MT-MMP-1 (A), human bases were amplified by rapid amplification of cDNA ends MT-MMP-1 (B) and oligonucleotide for 28 S ribosomal RNA (C). (RACE). Full-length cDNA reconstituted from the isolated RNA sources were as follows: 1, lung; 2, liver; 3, kidney; 4, brain; 5, cDNA and the amplified 3′-end, consisted of 1,972 bp (Fig. 1). spleen; 6, calvaria; 7, purified osteoclasts; 8, bone stromal cells; 9, The sequence contained an open reading frame of 1,746 bp, alveolar macrophages. 18S and 28S indicate the migratory positions of 18 and 28 S ribosomal RNA, respectively. starting at nucleotide position 127, and ending at nucleotide position 1,873, thus predicting a 582-residue polypeptide of 66,044 kDa. The deduced amino acid sequence showed 96%, 92% and 96% identity to MT1-MMP of human, mouse and rat, adult human tissues (Will and Hinzmann, 1995; Takino et al., respectively (based on the sequences reported by Okada et al., 1995). Importantly, we observed a prominent expression of 1995b) (Fig. 2). By sequence alignment with human, rat and MT1-MMP in purified osteoclasts, and lower levels of mouse MT1-MMP, we found no additions or deletions of expression in alveolar macrophages, and in bone stromal cells specific sequences. Comparisons of the predicted amino acid from which the osteoclasts were separated during the purifica- sequence with those of other human MT-MMPs revealed 57% tion procedure. As expected, expression was not detectable in identity to MT2-MMP (Will and Hinzmann, 1995), 52% liver and brain. The size of the major mRNA for MT1-MMP identity to MT3-MMP (Takino et al., 1995), and 38% identity was approximately 3.5 kb. to MT4-MMP (Puente et al., 1996). Identities to other MMPs ranged between 30 and 42%. From these observations, we In situ hybridization concluded that the reconstituted cDNA encodes the rabbit To examine whether MT1-MMP is also expressed in osteo- homolog of MT1-MMP. clasts in vivo, we performed in situ hybridization on sections of metacarpals of newborn rabbits (Fig. 4). The osteoclasts Northern blotting appeared as large multinucleated TRAP+ cells and many were To confirm expression of MT1-MMP in purified osteoclasts typically located against the calcified matrix of the growth and to compare its level with that in other tissues and cells, we plate. Significant levels of MT1-MMP mRNA were found in performed northern blotting using the rabbit or a human MT1- these osteoclasts, and also in other bone cells in accordance MMP probe (Fig. 3). Whatever probe we used, we found the with the above northern blottings and with the recent data of same pattern of distribution as those reported previously for Kinoh et al. (1996).

Fig. 4. In situ hybridization of MT1-MMP in bone sections. Three consecutive sections were stained for TRAP (A), or hybridized with antisense (B) or sense (C) probes respectively. Antisense probe, but not sense probe generated hybridization signals in the osteoclasts attached to the calcified matrix (arrowheads in B). These osteoclasts could also be identified by TRAP staining in the adjacent section (arrowheads in A). Note that significant levels of MT1-MMP mRNA were also found in TRAP- negative cells. Bar, 25 µm. ABC 594 T. Sato and others

ring of bright dots (Fig. 6B) as did the antibody directed against MT1-MMP. We also used phase contrast to detect the podosomes (Marchisio et al., 1984), and found again the same ring of dots (Fig. 6C) that we had seen when staining for MT1- MMP and actin. Thus MT1-MMP appears to be localized on the podosomes of ‘sitting’ osteoclasts. Actin staining was sharper when compared to the more diffuse MT1-MMP staining, probably because the actin dots are due to bundles of actin ﬁlaments in the core of the podosome, while MT1-MMP is likely to be on their surface.

Fig. 5. Double stain fluorescence pictures of an osteoclast with the typical phenotype of a ‘walking’ cell. Osteoclasts were obtained and DISCUSSION processed as explained in Materials and Methods. The pictures were taken adjusting the focus to the level of the contact between the cell This work demonstrates the expression of MT1-MMP in rabbit and the glass-slide. Phalloidin staining shows lamellipodiae rich in osteoclasts. The MT1-MMP identity is obvious from the high actin (B) at the leading edge and retraction fibers at the rear. The sequence similarity of the full length cDNA to human, mouse, MT1-MMP antibody shows immunoreactivity at the leading edge of and rat MT1-MMPs. That it is really expressed by osteoclasts the same cell (A). Bar, 20 µm. is shown by complementary approaches, including northern blotting exhibiting high mRNA levels in purified osteoclasts, and in situ hybridization showing MT1-MMP expression in Immunocytochemistry osteoclasts in vivo. Furthermore, immunolocalizations provide An important property of MT1-MMP in previously investi- evidence compatible with the presence of MT1-MMP protein. gated cells is its localization on their plasma membrane (Sato MT1-MMP expression has been reported in various cell et al., 1994a). To investigate the possible localization of MT1- types (Yamada et al., 1995; Sato and Seiki, 1996; Ailenberg MMP in the plasma membrane of osteoclasts, we performed and Silverman, 1996; Okada et al., 1995b) and its possible role immunocytochemistry. Using an antibody against MT1-MMP, in tissue invasion by cancer (Sato et al., 1994a; Sato and Seiki, we found a prominent fluorescence in all osteoclasts, but no 1996; Yamamoto et al., 1996; Gilles et al., 1996) and endo- fluorescence at all when using nonimmune IgGs. The distrib- thelial cells (Lewalle et al., 1995; Foda et al., 1996) has been ution of the signal was highly dependent on the physiological emphasized. MT1-MMP is indeed able to degrade different state of the osteoclast (Figs 5 and 6). Though some fluores- extracellular matrix proteins, such as fibronectin, vitronectin, cence appeared at the level of the cell body, clear fluorescence laminin, dermatan sulfate proteoglycan, and gelatin (Pei and was also visible at the level of contact between the cells and Weiss, 1996; Imai et al., 1996). MT1-MMP is also able to the glass slides, which is a site allowing the identification of activate extracellular gelatinase-A (proMMP-2) (Sato et al., their plasma membrane. In osteoclasts exhibiting a ‘walking’ 1994a) and interstitial collagenase-3 (proMMP-13) (Knaüper phenotype (Baron et al., 1993), the lamellipodiae at the leading et al., 1996), which in turn may degrade various extracellular edge were illuminated (Fig. 5A). In about half of those exhibit- matrix proteins. Thus it has been speculated that MT1-MMP ing a ‘sitting’ phenotype (Baron et al., 1993), the signals were on the cell surface of cancer cells may degrade extracellular arranged in a ring of small dots at the cell periphery (Fig. 6A). molecules, either directly or indirectly, a process which would The latter pattern is reminiscent of podosomes, which are small then facilitate tissue invasion. extensions of the plasma membrane that become abundant and Also the osteoclasts exhibit invasive activity when they organize in this particular way upon arrest and attachment of migrate and anchor themselves onto the bone surface. Recently the osteoclast (Baron et al., 1993). To investigate whether it was found that MMP activity is indispensable for the MT1-MMP is associated with podosomes, we stained the cells migration of (pre)osteoclasts to the developing marrow cavity simultaneously for actin, a procedure commonly used to of primitive long bones (Blavier and Delaissé, 1995). It is thus identify podosomes. Actin staining revealed indeed the same possible that the MT1-MMP that we identified in the lamel-

Fig. 6. Double stain fluorescence and phase contrast pictures of a typical ‘sitting’ osteoclast. The procedure for obtaining and examining the osteoclast is as indicated in Fig. 5. Immunostaining for MT1-MMP (A) shows a belt of dots at the periphery of the cell. The same belt of dots is revealed upon staining the same cell for actin (B). The latter dots correspond to actin filaments oriented perpendicularly to the substratum, and forming the core of the podosomes (Teti et al., 1991). Again the same belt of podosomes (dark dots) appears when observing the same cell by phase contrast (C). Bar, 50 µm. MT1-MMP in osteoclasts 595 lipodia of osteoclasts, plays a critical role in this migratory Hanning, C. R., Jones, C., Kurdyla, J. T., McNulty, D. E., Drake, F. H., process. Moreover, the development of a ring of podosomes Gowen, M. and Levy, M. A. (1996). Proteolytic activity of human osteoclast has been related to the attachment of the osteoclasts to the bone cathepsin K. Expression, purification, activation and substrate identification. J. Biol. Chem. 271, 12517-12524. surface (Baron et al., 1993). These podosomes are short cell Chen, W. T., Olden, K., Bernard, B. A. and Chu, F. F. (1984). Expression of extensions establishing focal contacts of short duration (Baron transformation-associated protease(s) that degrade fibronectin at cell contact et al., 1993), and they are suspected to bear proteinases (Teti sites. J. Cell Biol. 98, 1546-1555. et al., 1991; Aubin, 1992). Our demonstration of MT1-MMP Delaissé, J. M. and Vaes, G. (1992). Mechanism of mineral solubilization and matrix degradation in osteoclastic bone resorption. In Biology and on these podosomes fits thus the latter speculation, and is com- Physiology of the Osteoclast (ed. B. R. Rifkin and C. V. Gay), pp. 289-314. patible with a role of MT1-MMP in this attachment process. CRC Press, Boca Raton. Other roles are, however, also possible, for instance the acti- Delaissé, J. M., Eeckhout, Y., Neff, L., Francois-Gillet, C., Henriet, P., Su, vation of interstitial collagenase-3 (Knaüper et al., 1996), a Y., Vaes, G. and Baron, R. (1993). (Pro)collagenase (matrix proteinase that has been detected in the resorption compart- metalloproteinase-1) is present in rodent osteoclasts and in the underlying bone-resorbing compartment. J. Cell Sci. 106, 1071-1082. ment of mouse and rat osteoclasts (Delaissé et al., 1993). Devereux, J., Haeberli, P. and Smithies, O. (1984). A comprehensive set of Though there have been many speculations about the sequence analysis programs for the VAX. Nucl. Acids Res. 12, 387-395. involvement of MT1-MMP induced proteolysis in cell invasion Everts, V., Delaissé, J. M., Korper, W., Niehof, A., Vaes, G. and Beertsen, W. (Sato and Seiki, 1996), there is no information on how this pro- (1992). Degradation of collagen in the bone-resorbing compartment underlying the osteoclast involves both cysteine-proteinases and matrix teolysis is controlled and directed to the appropriate sites of the metalloproteinases. J. Cell. Physiol. 150, 221-231. extracellular matrix. The spatial distribution of MT1-MMP Fisher, C., Gilbertson-Beadling, S., Powers, E. A., Petzold, G., Poorman. R. shown in this work gives insights in this mechanism. There is and Mitchell M. A. (1994). Interstitial collagenase is required for indeed a high concentration of MT1-MMP at the tips of spe- angiogenesis in vitro. Dev. Biol. 162, 499-510. cialized membrane protrusions (podosomes and lamellipodia) Foda, H. D., George, S., Conner, C., Drews, M., Tompkins, D. C. and Zucker, S. (1996). Activation of human umbilical vein endothelial cell driven by actin filaments, and used by the cells to extend them- progelatinase A by phorbol myristate acetate: A protein kinase C-dependent selves forward or to anchor. We thus propose that cytoskeletal mechanism involving a membrane-type matrix metalloproteinase. Lab. elements bring MT1-MMP activity into contact with the Invest. 74, 538-545. specific sites of the extracellular matrix that must be altered to Gilles, C., Polette, M., Piette, J., Munaut, C., Thompson, E. W., Birembaut, P. and Foidart, J. M. (1996). High level of MT-MMP expression is allow cell movement or attachment. This might be a general associated with invasiveness of cervical cancer cells. Int. J. Cancer 65, 209- mechanism for cellular control of extracellular proteolysis, 213. since it was already shown that the urokinase receptor of Gyetko, M. R., Todd III, R. F., Wilkinson, C. C. and Sitrin, R. G. (1994). The migrating monocytes is polarized at their leading edge (Gyetko urokinase receptor is required for human monocyte chemotaxis in vitro. J. et al., 1994), and that proteolytic activity is localized on the Clin. Invest. 93, 1380-1387. Helfrich, M., Sato, T., Tezuka, K., Kumegawa, M. and Collin-Osdoby, P. podosomes of cells rendered invasive by transformation by (1994). Isolation of osteoclasts and osteoclast membranes. In Cell Biology: A some oncogenes (Kelly et al., 1994). Laboratory Handbook. pp. 128-141. Academic Press. Although the existence of membrane proteinases in osteo- Hill, P. A., Murphy, G., Docherty, A. J. P., Hembry, R. M., Millican, T. A., clasts had been hypothesized (Teti et al., 1991; Aubin, 1992), Reynolds, J. J. and Meikle, M. C. (1994). The effects of selective inhibitors of matrix metalloproteinases (MMPs) on bone resorption and the such proteinases were never demonstrated. The present identi- identification of MMPs and TIMP-1 in isolated osteoclasts. J. Cell Sci. 107, fication of MT1-MMP in osteoclasts might be relevant to the 3055-3064. invasive activity that these cells exhibit to gain access and Imai, K., Ohuchi, E., Aoki, T., Nomura, H., Fujii, Y., Sato, H., Seiki, M. and anchor themselves to the bone surface. Okada, Y. (1996). Membrane-type matrix metalloproteinase 1 is a gelatinolytic enzyme and is secreted in a complex with tissue inhibitor of The authors are grateful to Dr Ken-ichi Tezuka (Merck Research metalloproteinase 2. Cancer Res. 56, 2707-2710. Kaji, H., Sugimoto, T., Miyauchi, A., Fukase, M., Tezuka, K., Hakeda, Y., Laboratories, West Point, PA) for his helpful advice on molecular Kumegawa, M. and Chihara, K. (1994). Calcitonin inhibits osteopontin cloning, to Dr Kazushi Iwata (Fuji Chemical Industries, Toyama) for mRNA expression in isolated rabbit osteoclasts. Endocrinology 135, 484- the MT1-MMP antibody, and to Lone Bayer, Helle Deleurang and 487. Dorte Larsen for expert technical assistance. Part of this work was Kaji, H., Sugimoto, T., Kanatani, M., Fukase, M., Kumegawa, M. and presented orally at the 18th Annual Meeting of the American Society Chihara, K. (1996). Prostaglandin E2 stimulates osteoclast-like cell for Bone and Mineral Research, Seattle, September 7-11, 1996. formation and bone resorbing activity via osteoblasts: role of cAMP- dependent protein kinase. J. Bone Miner. Res. 11, 62-71. Kelly, T., Mueller S. C., Yeh, Y. and Chen, W. T. (1994). Invadopodia promote proteolysis of a wide variety of extracellular matrix proteins. J. Cell. Physiol. REFERENCES 158, 299-308. Kinoh, H., Sato, H., Tsunezuka, Y., Takino, T., Kawashima, A., Okada, Y. Ailenberg, M. and Silverman, M. (1996). Cellular activation by mesangial and Seiki, M. (1996). MT-MMP, the cell surface activator of proMMP-2 gelatinase A by cytochalasin D is accompanied by enhanced mRNA (pro-gelatinase A) is expressed with its substrate in mouse tissue during expression of both gelatinase A and its membrane-associated gelatinase A embryogenesis. J. Cell Sci. 109, 953-959. activator (MT-MMP). Biochem. J. 313, 879-884. Knaüper, V., Will, H., López-Otín, C., Smith, B., Atkinson, S. J., Stanton, Aubin J. E. (1992). Osteoclast adhesion and resorption: The role of podosomes. H., Hembry, R. M. and Murphy, G. (1996). Cellular mechanisms for J. Bone Miner. Res. 7, 365-368. human procollagenase-3 (MMP-13) activation. J. Biol. Chem. 271, 17124- Baron, R., Chakraborty, M., Chatterjee, D., Horne, W., Lomri, A. and 17131. Ravesloot, J. H. (1993). Biology of the osteoclast. In Physiology and Lakkakorpi, P. T. and Väänänen, H. K. (1991). Kinetics of the osteoclast Pharmacology of Bone (ed. G. R. Mundy and T. J. Martin), pp. 111-147. cytoskeleton during the resorption cycle in vitro. J. Bone Miner. Res. 6, 817- Springer Verlag, Berlin. 826. Blavier, L. and Delaissé, J. M. (1995). Matrix metalloproteinases are Lewalle, J. M., Munaut, C., Pichot, B., Cataldo, D., Baramova, E. and obligatory for the migration of preosteoclasts to the developing marrow Foidart, J. M. (1995). Plasma membrane-dependent activation of gelatinase cavity of primitive long bones. J. Cell Sci. 108, 3649-3659. A in human vascular endothelial cells. J. Cell. Physiol. 165, 475-483. Bossard, M. J., Tomaszek, T. A., Thompson, S. K., Amegadzie, B. Y., Li, Y. P., Alexander, M., Wucherpfennig, A. L., Yelick, P., Chen, W. and 596 T. Sato and others

Stashenko, P. (1995). Cloning and complete coding sequence of a novel Sato, T., Hakeda, Y., Tezuka, Y., Tezuka, K., Matsumoto, K., Nakamura, T., human cathepsin expressed in giant cells of osteoclastomas. J. Bone Miner. Mori, Y., Tanaka, K., Yamaguchi, Y., Yoshizawa, K., Kodama, H. and Res. 10, 1197-1202. Kumegawa, M. (1995). Hepatocyte growth factor is involved in formation of Librach, C. L., Werb, Z., Fitzgerald, M. L., Chiu, K., Corwin, N. M., osteoclast-like cells mediated by clonal stromal cells. J. Cell. Physiol. 164, Esteves, R. A., Grobelny, D., Galardy, R., Damsky, C. H. and Fisher, S. J. 197-204. (1991). 92-KD type IV collagenase mediates invasion of human Sato, H. and Seiki, M. (1996). Membrane-type matrix metalloproteinases cytotrophoblasts. J. Cell Biol. 113, 437-449. (MT-MMPs) in tumor metastasis. J. Biochem. 119, 209-215. Marchisio, P. C., Cirillo, D., Naldini, L., Primavera, M. V., Teti, A. and Takino, T., Sato, H., Shinagawa, A. and Seiki, M. (1995). Identification of the Zambonin-Zallone, A. (1984). Cell-substratum interaction of cultured avian second membrane-type matrix metalloproteinase (MT-MMP-2) gene from a osteoclasts is mediated by specific adhesion structures. J. Cell Biol. 99, 1696- human placenta cDNA library. J. Biol. Chem. 270, 23013-23020. 1705. Teti, A., Marchisio, P. C. and Zambonin Zallone, A. (1991). Clear zone in Muir, D. (1994). Metalloproteinase-dependent neurite outgrowth within a osteoclast function: Role of podosomes in regulation of bone-resorbing synthetic extracellular matrix is induced by nerve growth factor. Exp. Cell activity. Am. J. Physiol. 261, C1-C7. Res. 210, 243-252. Tezuka, K., Sato, T., Kamioka, H., Nijweide, P. J., Tanaka, K., Matsuo, T., Okada, Y., Naka, K., Kawamura, K., Matsumoto, T., Nakanishi, I., Ohta, M., Kurihara, N., Hakeda, Y. and Kumegawa M. (1992). Fujimoto, N., Sato, H. and Seiki, M. (1995a). Localization of matrix- Identification of osteopontin in isolated rabbit osteoclasts. Biochem. Biophys. metalloproteinase 9 (92-kilodalton gelatinase/type IV collagenase = Res. Commun. 186, 911-917. gelatinase B) in osteoclasts: Implications for bone resorption. Lab. Invest. 72, Tezuka, K., Tezuka, Y., Maejima, A., Sato, T., Nemoto, K., Kamioka, H., 311-322. Hakeda, Y. and Kumegawa, M. (1994a). Molecular cloning of a possible Okada, A., Bellocq, J. P., Rouyer, N., Chenard, M. P., Rio, M. C. and cysteine proteinase predominantly expressed in osteoclasts. J. Biol. Chem. Chambon, P., Basset, P. (1995b). Membrane-type matrix metalloproteinase 269, 1106-1109. (MT-MMP) gene is expressed in stromal cells of human colon, breast, and Tezuka, K., Nemoto, K., Tezuka, Y., Sato, T., Ikeda, Y., Kobori, M., head and neck carcinomas. Proc. Nat. Acad. Sci. USA 92, 2730-2734. Kawashima, H., Eguchi, H., Hakeda, Y. and Kumegawa, M. (1994b). Pei, D. and Weiss, S. J. (1996). Transmembrane-deletion mutants of the Identification of matrix metalloproteinase 9 in rabbit osteoclasts. J. Biol. membrane-type matrix metalloproteinase-1 process progelatinase A and Chem. 269, 15006-15009. express intrinsic matrix-degrading activity. J. Biol. Chem. 271, 9135-9140. Tryggvason, K., Hoyhtya, M. and Pyke C. (1993). Type IV collagenases in Puente, X. S., Pendás, A. M., Llano, E., Velasco, G. and López-Otín C. invasive tumors. Breast Cancer Res. Tr. 24, 209-218. (1996). Molecular cloning of a novel membrane-type matrix Will, H. and Hinzmann, B. (1995). cDNA sequence and mRNA tissue metalloproteinase from a human breast carcinoma. Cancer Res. 56, 944-949. distribution of a novel human matrix metalloproteinase with a potential Reponen, P., Sahlberg, C., Munaut, C., Theslef, I. and Tryggvason, K. transmembrane segment. Eur. J. Biochem. 231, 602-608. (1994). Expression of 92-kD type IV collagenase (gelatinase B) in the Wucherpfennig, A. L., Li Y. P., Stettler-Stevenson, W. G., Rosenberg, A. E. osteoclast lineage during mouse development. J. Cell Biol. 124, 1091-1102. and Stashenko, P. (1994). Expression of 92 KD type IV Salo, T., Makela, M., Kylmaniemi, M., Autio-Harmainen, H. and Larjava, collagenase/gelatinase B in human osteoclasts. J. Bone Miner. Res. 9, 549- H. (1994). Expression of matrix metalloproteinase-2 and 9 during early 556. human wound healing. Lab. Invest. 70, 176-182. Yamada, T., Yoshiyama, Y., Sato, H., Seiki, M., Shinagawa, A. and Saneshige, S., Mano, H., Tezuka, K. I., Kakudo, S., Mori, Y., Honda, Y., Takahashi, M. (1995). White matter microglia produce membrane-type Itabashi, A., Yamada, T., Miyata, K., Hakeda, Y., Ishii, J. and matrix metalloprotease, an activator of gelatinase A, in human brain tissues. Kumegawa, M. (1995). Retinoic acid directly stimulates osteoclastic bone Acta Neuropathol. 90, 421-424. resorption and gene expression of cathepsin K/OC-2. Biochem. J. 309, 721- Yamamoto, M., Mohanam, S., Sawaya, R., Fuller, G. N., Seiki, M., Sato, H., 724. Gokaslan, Z. L., Liotta, L. A., Nicolson, G. L. and Rao, J. S. (1996). Sato, H., Takino, T., Okada, Y., Cao, J., Shinagawa, A., Yamamoto, E. and Differential expression of membrane-type matrix metalloproteinase and its Seiki M. (1994a). A matrix metalloproteinase expressed on the surface of correlation with gelatinase A activation in human malignant brain tumors in invasive tumour cells. Nature 370, 61-65. vivo and in vitro. Cancer Res. 56, 384-392. Sato, T., Kamioka, H., Nemoto, K., Tezuka, K., Tanaka, K., Sugimoto, T., Kaji, H., Miyauchi, A., Yoshizawa, K., Yamaguchi, Y., Hakeda, Y. and Kumegawa M. (1994b). A simple method for isolation of rabbit osteoclasts. Dentistry in Japan 31, 3-8. (Received 23 October 1996 - Accepted 3 January 1996)