Targeted inhibition of human collagenase-3 (MMP-13 ... - Nature

Oncogene (2004) 23, 5111–5123

& 2004 Nature Publishing Group All rights reserved 0950-9232/04 $30.00 www.nature.com/onc

ORIGINAL PAPERS

Targeted inhibition of human collagenase-3 (MMP-13) expression inhibits squamous cell carcinoma growth in vivo Risto Ala-aho1,2,3, Matti Ahonen1,2,3, Sarah J George4, Jari Heikkila¨5, Reidar Greńman6, Markku Kallajoki7 and Veli-Matti Ka¨ha¨ri*,1,2,3 1 Department of Medical Biochemistry and Molecular Biology, University of Turku, Turku, Finland; 2Department of Dermatology, University of Turku, FIN-20520 Turku, Finland; 3Centre for Biotechnology, University of Turku and A˚bo Akademi University, Finland; 4Bristol Heart Institute, Bristol Royal Infirmary, Bristol BS2 8HW, UK; 5Department of Biochemistry and Pharmacy, A˚bo Akademi University, FIN-20520 Turku, Finland; 6Department of Otorhinolaryngology – Head and Neck Surgery, Turku University Central Hospital, FIN-20520 Turku, Finland; 7Department of Pathology, Turku University Central Hospital, FIN-20520 Turku, Finland

Squamous cell carcinomas (SCCs) of the head and neck are characterized by a high tendency for local invasion and metastasis to lymph nodes. Collagenase-3 (MMP-13) is specifically expressed by tumor cells in SCCs of the head and neck and its expression correlates with their invasion capacity. To specifically examine the role of MMP-13 in the growth and invasion of SCC, we constructed a hammerhead ribozyme targeted against human MMP-13 mRNA. The anti-MMP-13 ribozyme effectively cleaved MMP-13 transcripts in vitro. Adenoviral delivery of the anti-MMP-13 ribozyme to cutaneous metastatic SCC cells in culture resulted in potent and specific inhibition of the production of proMMP-13 and markedly suppressed invasion of SCC cells through Matrigel. In addition, adenoviral delivery of anti-MMP-13 ribozyme promoted apoptosis in SCC cells within 72 h. Intratumoral injection of anti-MMP-13 ribozyme coding adenovirus into human SCC xenografts established in SCID mice potently suppressed tumor growth, inhibited MMP-13 expression and gelatinolytic activity and reduced the number of proliferating cells within the tumors. These results provide evidence for an important role for MMP-13 in SCC growth and invasion and identify MMP-13 as a promising target for ribozyme-based therapy of SCC in vivo. Oncogene (2004) 23, 5111–5123. doi:10.1038/sj.onc.1207678 Published online 19 April 2004 Keywords: collagenase; squamous cell carcinoma; invasion; ribozyme

*Correspondence: V-M Ka¨ha¨ri, Department of Medical Biochemistry and Molecular Biology, University of Turku, Kiinamyllynkatu 10, FIN-20520 Turku, Finland; E-mail: veli-matti.kahari@utu.fi Received 9 September 2003; revised 19 January 2004; accepted 17 February 2004; published online 19 April 2004

Introduction Invasion and metastasis of malignant tumors involves detachment of cancer cells from primary tumor, degradation of structural barriers, such as basement membrane and collagenous extracellular matrix (ECM), and migration of cells through degraded matrix. Matrix metalloproteinases (MMPs) are a family of zincdependent neutral endopeptidases collectively capable of degrading essentially all ECM components, and they obviously play an important role in tumor invasion and tumor-induced angiogenesis (Overall and Lo´pez-Otı´ n, 2002; Vihinen and Ka¨ha¨ri, 2002). In all, 24 human members of the MMP gene family are known, and in addition to ECM substrates they cleave cell surface molecules and other pericellular proteins and activate other proteinases, thereby regulating cell behavior in several ways (Overall and Lo´pez-Otı´ n, 2002; Vihinen and Ka¨ha¨ri, 2002). Fibrillar collagens are the most abundant structural components of the human connective tissues, and it is conceivable that the ability to degrade them is crucial for invasion and metastasis of neoplastic cells. Collagenase-1 (MMP-1), collagenase-2 (MMP-8), and collagenase-3 (MMP-13) are the principal neutral proteinases responsible for cleavage of native fibrillar collagens of type I, II, III, and V. MMP-13 also cleaves several other ECM components, including type IV, X, and XIV collagens, large tenascin C, fibronectin, aggrecan, versican, and fibrillin-1 (Fosang et al., 1996; Knaüper et al., 1996; 1997a; Ashworth et al., 1999), as well as non-ECM components, such as chemokines macrophage chemotactic protein-3 and stromal cell-derived factor-1 (McQuibban et al., 2000, 2001). The physiologic expression of MMP-13 is limited to situations in which rapid remodeling of collagenous ECM is required, for example, fetal bone development (Johansson et al., 1997b) and fetal skin and adult gingival wound repair (Ravanti et al., 1999b; 2001). The expression of MMP-13 has also been detected in invasive neoplastic tumors, that is, breast carcinomas (Freije et al., 1994), squamous cell carcinomas (SCCs) of

Anti-MMP-13 ribozyme inhibits tumor growth R Ala-aho et al

5112

the head and neck (Airola et al., 1997; Johansson et al., 1997a; Cazorla et al., 1998), vulva (Johansson et al., 1999), and esophagus (Etoh et al., 2000), in chondrosarcomas (Urı´ a et al., 1998), primary and metastatic melanomas (Airola et al., 1999; Nikkola et al., 2001), and urothelial carcinomas (Bostro¨m et al., 2000). In SCCs of the head and neck, vulva, and esophagus, MMP-13 is expressed by cancer cells at the invading edge of the tumor and the levels of expression correlate with the invasion capacity of the tumors (Airola et al., 1997; Cazorla et al., 1998; Johansson et al., 1997a, 1999; Etoh et al., 2000). However, no expression of MMP-13 is noted in premalignant tumors in human skin (Airola et al., 1997), or by normal epidermal keratinocytes in culture or in vivo (Johansson et al., 1997c; Vaalamo et al., 1997), indicating that MMP-13 expression serves as a marker for transformation of squamous epithelial cells. As MMP-13 is not expressed in most adult human tissues under normal conditions, downregulating MMP13 expression may serve as an important strategy for cancer therapy. Antisense oligonucleotides and catalytic RNAs such as ribozymes are capable of specifically modulating gene expression and they have demonstrated utility in attenuating gene expression by cells in culture and in vivo (Santiago and Khachigian, 2001). Ribozymes have drawn attention as novel therapeutic agents that can suppress the expression of deleterious proteins by catalysing cleavage of the corresponding mRNAs (Lewin and Hauswirth, 2001). Compared to traditional antisense techniques, ribozymes are site specific and their catalytic potential makes them more efficient in suppressing specific gene expression. In this study, we have designed an antisense hammerhead ribozyme, which specifically cleaves human MMP13 transcript and constructed a recombinant adenovirus encoding this ribozyme. Our results show that adenoviral delivery of this ribozyme to human SCC cells specifically inhibits MMP-13 expression, invasion, and growth of SCC cells, and promotes their apoptosis. In addition, inhibition of MMP-13 expression suppresses growth of SCC xenografts in SCID mice. Together, these observations provide evidence for an important role for MMP-13 in SCC growth and invasion and identify MMP-13 as a promising target for ribozymebased therapy of SCC in vivo.

Results Characterization of human MMP-13 antisense ribozyme An antisense human MMP-13 hammerhead ribozyme was designed to specifically cleave human MMP-13 mRNA between nucleotides 716 and 717 (Figure 1a). Binding arms complementary to the target mRNA, flanking the catalytic ribozyme structure are 9 and 8 nucleotides in length in the 50 and 30 ends, respectively. The search for human genome sequences with this ribozyme sequence revealed no homology to any other known human or mouse genomic sequences. As a Oncogene

control, we designed a corresponding sense ribozyme containing the hammerhead catalytic loop, but unable to anneal to MMP-13 or any other known human or mouse mRNA (Figure 1a). The MMP-13 antisense and sense sequences containing a catalytic ribozyme loop were cloned into pCI-neo vector, transcribed in vitro, and tested for their ability to cleave human MMP-13 mRNA. The cleavage of MMP-13 mRNA by antisense ribozyme resulted in the generation of fragments of 706 and 736 nucleotides in length, as expected (Figure 1b and c). After 60 min incubation, 50% of target RNA was cleaved and in 8 h all MMP-13 RNA was cleaved into two fragments. No cleavage of MMP-13 mRNA was seen with sense ribozyme (Figure 1a and c). The cleavage activity of three other potential ribozymes was also tested. Of these, one did not cleave human MMP-13 mRNA at all and the cleavage activity of two other ribozymes was less potent than that of the ribozyme shown above (data not shown). Adenoviral delivery of MMP-13 antisense ribozyme inhibits MMP-13 expression and invasion of SCC cells SCCs of the head and neck are malignant tumors with high invasion capacity and they express high levels of MMP-13 (Airola et al., 1997; Johansson et al., 1997a; Cazorla et al., 1998). To examine the role of MMP-13 in SCC growth and invasion, we constructed a recombinant adenovirus (RAdMMP-13ASRz) encoding human MMP-13 antisense ribozyme. Adenovirus-mediated expression of anti-MMP-13 ribozyme resulted in potent inhibition of MMP-13 production by human cutaneous SCC cell line UT-SCC-7 within 24 h, whereas production of MMP-1, 92- (MMP-9), and 72-kDa gelatinases (MMP-2) was not markedly altered (Figure 2a). Infection of cells with control adenovirus RAdMMP13senseRz encoding human MMP-13 sense ribozyme had no effect on MMP-13 production. Marked inhibition of MMP-13 production by ras-transformed HaCaT keratinocytes (Figure 2b) and HaCaT cells (data not shown) was also noted in response to MMP-13 antisense ribozyme, whereas production of MMP-1 and 92- and 72-kDa gelatinases by these cells was not altered. MMP-13 cleaves components of basement membranes (Knaüper et al., 1996, 1997a), and the expression of MMP-13 enhances invasion of malignant cells through reconstituted basement membrane, Matrigel (Ala-aho et al., 2002b). As RAdMMP-13ASRz potently inhibits the expression of MMP-13, we examined its effect on the invasion of SCC cells through Matrigel. As shown in Figure 2c, invasion of UT-SCC-7 cells was significantly (by 80%) inhibited by MMP-13 antisense ribozyme, whereas infection with empty control virus RAdpCA3 and with RAdMMP-13senseRz did not significantly alter the invasion capacity of these cells. Anti-MMP-13 ribozyme inhibits growth and promotes apoptosis of squamous carcinoma cells To test the effect of RAdMMP-13ASRz on SCC growth in vitro, we transduced UT-SCC-7 cells with RAdMMP-


5113

Figure 1 The structure of anti-MMP-13 ribozyme and in vitro cleavage of MMP-13 RNA by antisense ribozyme. (a) The MMP-13 antisense ribozyme targets human MMP-13 mRNA between nucleotides þ 707 and þ 724. The predicted cleavage site is between nucleotides þ 716 and þ 717. The flanking vector-generated sequences are not shown. Control sense hammerhead ribozyme contains a catalytic loop of hammerhead ribozyme flanked by the corresponding sequence of MMP-13 mRNA in sense orientation. (b) The expected cleavage fragments of human MMP-13 transcript by anti-MMP-13 ribozyme. (c) In vitro cleavage of human MMP-13 mRNA by ribozyme. Human MMP-13 mRNA generated by in vitro transcription was incubated with antisense ribozyme for different periods of time (0–8 h) or with sense ribozyme for 8 h, fractionated on a 5% polyacrylamide gel containing 7 M urea, and visualized by ethidium bromide. The size of uncleaved MMP-13 mRNA and specific cleavage fragments (in nucleotides) are indicated at the left side of the panel

13ASRz and determined the number of viable cells at different time points. Interestingly, infection with the MMP-13 antisense ribozyme virus reduced the number of viable UT-SCC-7 cells significantly (by 35%) at 96 h after the infection, while the sense control adenovirus had no effect on cell viability in comparison with uninfected control cells (Figure 3a). Similar results were obtained with HaCaT keratinocytes (Figure 3a). To determine the effect of anti-MMP-13 ribozyme on cell proliferation, UT-SCC-7 cells (2 104) were plated on culture dishes, transduced by RAdMMP-13ASRz and RAdMMP-13senseRz and the number of cells was counted every 24 h beginning at day 2. RAdMMP13ASRz inhibited the growth of UT-SCC-7 cells, first noted at 72 h of incubation, as compared to uninfected and control virus-infected cells, and at 120 h the

difference in cell number was even more pronounced (47 and 38%, respectively; Figure 3b). Similar results were also obtained with HaCaT cells, in which RAdMMP-13ASRz reduced the number of cells by 28 and 32%, as compared to uninfected and sense ribozyme virus infected cells, respectively (Figure 3b). The effect of anti-MMP-13 ribozyme on cell proliferation was also examined by determining the incorporation of bromodeoxyuridine (BrdU) into cells at different time points following transduction with RAdMMP-13ASRz and the sense ribozyme control virus. The level of BrdU incorporation determined at the 96 h time point was reduced by 52 and 43% in RAdMMP-13ASRz-infected UT-SCC-7 cells and HaCaT cells, respectively, as compared to the corresponding uninfected cells (Figure 3c). Oncogene


5114

To further elucidate the mechanism of inhibitory effect of anti-MMP-13 ribozyme on cell growth and viability, we used TUNEL staining to detect apoptotic cells in UT-SCC-7 cultures after infection with MMP-13 antisense and sense ribozyme adenoviruses. Interestingly, 15% of cells were identified as TUNEL positive 72 h after the adenoviral delivery of anti-MMP-13

ribozyme, whereas no TUNEL-positive cells were detected in uninfected or RAdMMP-13senseRz-infected cultures (Figure 3d). The adenovirus-infected cells were also stained with Hoechst to identify apoptotic cells. Slight alterations in nuclear morphology were detectable in RAdMMP13ASRz-infected cultures already at 72 h, and the release of apoptotic bodies was seen in RAdMMP13ASRz-infected SCC cells 96 h after adenoviral infection (Figure 3d). No alterations in nuclear morphology were detected in control sense adenovirus and uninfected cultures (Figure 3d). Anti-MMP-13 ribozyme inhibits growth of squamous cell carcinoma in vivo To examine whether MMP-13 also plays a role in SCC cell survival and invasion in vivo, we infected UT-SCC-7 cells with RAdMMP-13ASRz or RAdMMP-13senseRz for 6 h. After an 18 h incubation in fresh culture medium, UT-SCC-7 cells (5 106) were inoculated subcutaneously into the back of SCID mice and tumor size was measured twice a week. Interestingly, no tumor was detected in one out of five mice in the RAdMMP13ASRz group, whereas all mice in groups infected with control virus and uninfected control cells developed tumors. As shown in Figure 4a, the growth of SCC xenografts established with RAdMMP-13ASRz-infected cells was significantly slower (8–15 days delayed) than growth of tumors established with uninfected cells and control virus RAdMMP-13senseRz-infected cells (Figure 4a). Next, we examined the effect of antisense MMP-13 ribozyme on the growth of pre-established human SCC xenografts. UT-SCC-7 cells (5 106) were inoculated subcutaneously into the back of SCID mice and tumor size was measured at the time of adenoviral injections. In the first experiment, implanted SCC tumors reached a size of 100 mm3 in approximately 6 weeks after tumor cell injection and recombinant adenoviruses were administered intratumorally twice a week for 4 weeks starting on day 41. Injection of tumors with RAdMMP13ASRz (1 109 pfu) resulted in significant inhibition in Figure 2 Adenovirus-mediated expression of MMP-13 antisense ribozyme inhibits MMP-13 expression and invasion of squamous carcinoma cells. (a) Human cutaneous SCC cells (UT-SCC-7) and (b) ras-transformed HaCaT keratinocytes were infected with recombinant adenoviruses expressing human MMP-13 antisense ribozyme (RAdMMP-13ASRz) or sense control ribozyme (RAdMMP-13senseRz) for 6 h and incubated for the indicated periods of time. Production of MMP-13 and MMP-1 was determined by Western blot analysis and the levels of 92 and 72 kDa gelatinases were analysed by gelatin zymography of the conditioned media. Uninfected cells (control) were incubated and analysed in parallel. (c) UT-SCC-7 cells were infected with RAdMMP-13ASRz, with empty control virus RAdpCA3, or with RAdMMP-13senseRz for 6 h, incubated for 24 h, and seeded on top of cell culture inserts pre-coated with 25 mg Matrigel. The number of invaded cells was determined after 24 h. Mean þ s.e.m. of two experiments performed in duplicate are shown. Statistical significance against uninfected control cells was determined by Student’s t-test: *Po0.05

Oncogene


5115

Figure 3 Anti-MMP-13 ribozyme promotes apoptosis of squamous carcinoma cells. (a) UT-SCC-7 cells (left panel) and HaCaT cells (right panel) (1.5 104) were plated on a 96-well plate, infected with adenoviruses harboring human MMP-13 antisense ribozyme (RAdMMP-13ASRz) or sense control ribozyme (RAdMMP-13senseRz), and the number of viable cells was determined at different time points by MTT assay. The means þ s.d. are shown (n ¼ 4). *RAdMMP-13ASRz vs uninfected (control) and RAdMMP13senseRz-infected cells; Po0.002 by Student’s t-test. (b) 2 104 UT-SCC-7 cells (left panel) and HaCaT cells (right panel) were seeded onto plates and infected with recombinant adenoviruses as above, and the number of cells was counted at different time points. The results represent mean7s.d. (n ¼ 3). *RAdMMP-13ASRz vs uninfected (control) cells, Po0.002; RAdMMP-13ASRz vs RAdMMP13senseRz, Po0.05. (c) UT-SCC-7 cells (left panel) and HaCaT cells (right panel) were infected with recombinant adenoviruses as above and cell proliferation was determined at different time points by BrdU incorporation. (d) UT-SCC-7 cells in culture were infected with RAdMMP-13ASRZ and RAdMMP-13SenseRZ adenoviruses as above, and apoptotic cells were detected by TUNEL staining 72 h after infection (upper panels) and by Hoechst staining 72 and 96 h after infection (middle and lower panels, respectively) Oncogene


5116

al dose was escalated by injecting the human SCC xenografts three times a week with the same adenoviral dose (1 109 pfu) initiated at day 36 when the tumors reached approximately the size of 100 mm3. Again, SCC growth was significantly (by 50%) inhibited by RAdMMP-13ASRz, whereas RAdMMP-13senseRz had no effect on tumor growth compared to PBStreated tumors (Figure 4c). To estimate the in vivo adenoviral transduction efficiency, SCC xenografts were injected once with b-galactosidase-coding adenovirus and the infected cells were identified by X-gal staining of the tumors 24 h later. Quantitation of the b-galactosidase-expressing cells revealed that 3–5% of cells within tumors expressed this adenovirally delivered gene after a single adenoviral injection (data not shown). Anti-MMP-13 ribozyme inhibits expression of MMP-13 and gelatinolytic activity in squamous cell carcinomas in vivo

Figure 4 Adenovirus-mediated delivery of anti-MMP-13 ribozyme inhibits squamous cell carcinoma growth in vivo. (a) UT-SCC-7 cells in culture were infected with recombinant adenoviruses expressing anti-MMP-13 ribozyme or MMP-13 sense control ribozyme at MOI 700 for 6 h. After an additional 18 h incubation, cells (5 106) were implanted subcutaneously in the back of SCID/SCID mice and tumor size was measured once a week. Statistical significance between RAdMMP-13ASRz-infected and RAdMMP-13senseRz- or PBS-injected groups were determined by Student’s t-test: *Po0.01 (n ¼ 5 in each group). (b) Subcutaneous SCC tumors were established by injecting 5 106 UT-SCC-7 cells in the back of SCID mice. The tumors were injected with RAdMMP-13ASRz and RAdMMP-13senseRz (1 109 pfu) or with PBS twice a week starting on day 41, and the size of tumors was measured at the time of injections. Statistical significance in the tumor size between RAdMMP-13ASRz-injected and RAdMMP-13senseRz- or PBS-injected groups was determined by Student’s t-test: *Po0.05 (n ¼ 6). (c) Subcutaneous SCC tumors were established as in (b) and injected with RAdMMP-13ASRz and RAdMMP-13senseRz (1 109 pfu) or with PBS three times a week from day 36. Statistical significance in tumor size between RAdMMP-13ASRz- and RAdMMP-13senseRz-treated groups: *Po0.05, **Po0.01 (n ¼ 6)

tumor growth already after two injections (at day 50), and in the end of the experiment (day 67) tumors in RAdMMP-13ASRz group were 62% smaller than in PBS-injected control group (Figure 4b). In addition, no significant growth was noted in RAdMMP-13ASRz injected tumors after initiation of the adenoviral injections. RAdMMP-13senseRz control virus had no effect on tumor growth (Figure 4b). Next, the adenovirOncogene

To examine the effect of adenovirally delivered antisense MMP-13 ribozyme on SCCs, the xenografts were harvested in the end of experiments. To verify the inhibitory effect of adenovirally expressed anti-MMP-13 ribozyme on MMP-13 expression, RT–PCR was performed with the RNA samples isolated from tumors injected with adenoviruses three times a week. Tumors injected with RAdMMP-13ASRz showed reduction in human MMP13 mRNA levels, as compared to RAdMMP-13senseRzor PBS-injected tumors (Figure 5a). To verify the inhibitory effect of adenovirally delivered MMP-13 antisense ribozyme on MMP activity in vivo, tumor sections were studied by in situ gelatin zymography. Marked gelatinase activity was observed in PBS- and RAdMMP-13senseRz-injected tumors, whereas potent reduction in gelatinolytic activity was observed in RAdMMP-13ASRz-injected tumors (Figure 5b). Addition of MMP inhibitor BB-94 entirely blocked the gelatinase activity in tumor tissue, confirming that gelatin degradation is due to MMP activity. Anti-MMP-13 ribozyme reduces proliferation of squamous carcinoma cells in vivo To examine the effect of MMP-13 ribozyme on growth potential of SCC in vivo, proliferating cells in tumor sections were identified by immunohistochemistry using Ki67 as a marker. The Ki67-positive cells were detected near the margin of the tumors (Figure 6a, upper panels). The area of Ki67-positive cells appeared equal in PBSand RAdMMP-13senseRz-injected tumors (Figure 6a). Interestingly, the Ki67-positive tumor cell area was clearly smaller in SCCs injected with anti-MMP-13 ribozyme adenovirus (Figure 6a). Determination of the relative number of Ki67-positive cells in SCC xenografts revealed that the area occupied by proliferating cells in tumors injected with anti-MMP-13 ribozyme virus was 30% lower than in the control groups (Figure 6b). In addition, a correlation was noted between the relative number of proliferating cells and the volume of the


5117

Figure 5 Adenovirus-mediated delivery of anti-MMP-13 ribozyme inhibits MMP-13 expression and gelatinolytic activity in squamous cell carcinomas in vivo. Subcutaneous SCC tumors were established by injecting 5 106 UT-SCC-7 cells in the back of SCID mice. The tumors were injected with recombinant adenoviruses expressing human MMP-13 antisense ribozyme (RAdMMP-13ASRz) or sense control ribozyme (RAdMMP-13senseRz) (1 109 pfu) three times a week starting on day 36 (Figure 4c) and analysed 20 days later. (a) Total RNA was isolated from adenovirally infected tumor tissue and RT–PCR was performed to determine the expression level of human MMP-13 and GAPDH mRNA. The amplification products were fractionated on agarose gel and visualized by ethidium bromide. (b) Gelatinolytic activity in tumors was determined with in situ gelatinase zymography. Gelatinase acitivity is noted as white areas of gelatin degradation in PBS- and RAdMMP-13senseRz-injected tumors (upper panels). The hematoxylin and eosin staining of the parallel tissue sections is shown (lower panels). Gelatinolytic activity of a tumor section of the RAdMMPsenseRzinjected tumor was measured in the presence of MMP inhibitor (BB-94)

tumors in the end of the experiment (Figure 6b). No histological differences between RAdMMP-13ASRz-, RAdMMP-13senseRz- and PBS-injected tumors were detected: all tumors showed similar well-differentiated SCC phenotype (Figure 6a). Weigert-van Gieson staining showed similar collagen layer surrounding the xenografts (Figure 6a, middle panels). The same tumors were examined for apoptotic cells by immunohistochemical staining for activated caspase-3. No differences in the number of activated caspase-3-positive cells detected in the center of tumors were found between RAdMMP-13ASRz-, RAdMMP-13 senseRZ- and PBSinjected groups (Figure 6a, lower panels). Discussion There is a considerable amount of evidence indicating that overexpression of several MMPs correlates with

increased invasion capacity and poor prognosis in malignant tumors (Overall and Lo´pez-Otı´ n, 2002; Vihinen and Ka¨ha¨ri, 2002). Several broad-range MMP inhibitors have shown efficacy against the growth and invasion of malignant tumors in preclinical studies, but the outcome of the clinical trials with these inhibitors in patients with different types of tumors has not been encouraging in general (Overall and Lo´pez-Otı´ n, 2002; Vihinen and Ka¨ha¨ri, 2002). Different MMPs are overexpressed in distinct tumors, and therefore the appropriate targets for therapeutic intervention may vary in each type of malignant tumor (Nelson et al., 2000; Overall and Lo´pez-Otı´ n, 2002; Vihinen and Ka¨ha¨ri, 2002). In this context, specific inhibition of the overexpression of particular MMP by antisense ribozyme may serve as a mode for effective targeted therapy in cancer and in other pathological conditions. Accordingly, a ribozyme targeted to rat MMP-9 has been shown to inhibit metastasis of rat sarcomas (Hua and Oncogene


5118

Figure 6 Anti-MMP-13 ribozyme suppresses proliferation of squamous cell carcinoma. Subcutaneous SCCs were established by injecting 5 106 UT-SCC-7 cells in the back of SCID mice. The tumors were injected with recombinant adenoviruses expressing human MMP-13 antisense ribozyme (RAdMMP-13ASRz) or sense control ribozyme (RAdMMP-13senseRz) (1 109 pfu) three times a week starting on day 36 (Figure 4c) and analysed 20 days later. (a) SCC tumors were immunostained for Ki67 as a marker for proliferating cells (upper panels). Weigert-van Gieson staining showed identical collagen layer surrounding the xenografts (red) (middle panels). Immunohistochemical staining for cleaved caspase-3 showed identical number of positive tumors cells in each group (lower panels). (b) The relative area of Ki67-positive cells were determined in four distinct fields at 4 magnification from all tumor sections using Soft Imaging System’s analySISs program. The average relative number of proliferating cells was compared to the average tumor sizes in each group

Muschel, 1996) and a ribozyme against rat MMP-3 inhibits MMP-3 mRNA expression in articular cartilage explants (Jarvis et al., 2000). Human collagenase-3 (MMP-13) is expressed by malignantly transformed epidermal keratinocytes, for example, squamous carcinoma cells in culture and in vivo (Airola et al., 1997; Cazorla et al., 1998; Johansson et al., 1997a, 1999; Nelson et al., 2000), but it is not expressed by normal epidermal keratinocytes in Oncogene

culture or in vivo (Johansson et al., 1997c; Vaalamo et al., 1997). In addition, no expression of MMP-13 is noted in premalignant tumors or in normal human skin (Airola et al., 1997), indicating that MMP-13 expression serves as a marker for malignant transformation of squamous epithelial cells, and suggesting an important role for MMP-13 in the invasion of SCC cells. The inhibition of MMP-13 expression in transformed human epidermal keratinocytes by interferon-g or p53 tumor


5119

suppressor markedly reduces their invasion capacity (Ala-aho et al., 2000, 2002a). In addition, expression of MMP-13 by human fibrosarcoma (HT-1080) cell line increases its invasion capacity through type I collagen and Matrigel (Ala-aho et al., 2002b). Together, these features suggest MMP-13 as a potential therapeutic target in SCCs and other malignant tumors expressing MMP-13. In the present study, we have designed a human MMP-13 antisense ribozyme and tested the effect of adenoviral delivery of this ribozyme on the expression of MMP-13 and invasion capacity of human SCC cells in culture and in vivo. Based on homology search, the MMP-13 antisense ribozyme does not recognize any other human or mouse transcript. The anti-MMP-13 ribozyme specifically cleaves the human MMP-13 transcript in a cell-free system and adenovirus-mediated expression of the ribozyme results in potent inhibition of MMP-13 expression by transformed keratinocyte lines in culture. In addition, our results show that reduced expression of MMP-13 suppresses growth of human cutaneous SCC xenografts in SCID mice and promotes apoptosis and suppresses invasion capacity of SCC cells in culture. Marked inhibition of MMP-13 production in SCC cells was noted within 24 h after adenoviral delivery of anti-MMP-13 ribozyme, whereas no reduction in cell viability was detected during the first 48 h. Furthermore, no apoptotic cells were detected within the first 48 h after adenoviral delivery of MMP-13 antisense ribozyme, indicating that anti-MMP-13 ribozyme inhibits MMP-13 gene expression and invasion of SCC cells independently of its ability to induce apoptosis. In addition, anti-MMP-13 ribozyme inhibits SCC cell invasion through Matrigel within the first 24 h after adenoviral transduction, indicating that invasion is suppressed due to inhibition of MMP-13 expression rather than reduction in cell viability. The appearance of apoptotic SCC cells was detected 72 h after adenoviral delivery of anti-MMP-13 ribozyme and marked inhibition of cell growth and viability was detected 96 and 120 h postinfection. Together, these data show that suppression of MMP-13 expression by antisense ribozyme results in inhibition of SCC cell proliferation and in apoptosis, suggesting a role for MMP-13 in SCC cell proliferation and survival. These findings are in accordance with our previous observations showing that, in SCC cells treated with interferon-g or infected with p53 adenovirus, downregulation of MMP-13 expression precedes the onset of cell death (Ala-aho et al., 2000, 2002a). Whether the promoting effect of MMP-13 on cell growth and survival involves proteolytic processing of growth-promoting factors or cleavage of proapoptotic proteins is not known at present. In our ex vivo experiment, SCC cell implantation and growth was delayed in SCID mice by adenoviral delivery of anti-MMP-13 ribozyme into SCC cells. Interestingly, one of five mice injected with RAdMMP-13ASRz-infected cells generated no tumor. Since adenoviral genome does not integrate into host cell genome, the adenovirus-mediated gene delivery

results in relatively short-term expression of the ribozyme and subsequent suppression of MMP-13 expression. Nevertheless, these observations provide evidence that MMP-13 plays an important role in the early stage of SCC growth. The injection of human SCC xenografts with RAdMMP-13ASRz resulted in suppression of tumor growth. However, escalation of the adenoviral dose in intratumoral injections from twice a week to three times a week did not further increase the inhibitory effect of anti-MMP-13 ribozyme on SCC growth. This may be due to limited transduction efficiency, as the estimated efficiency of adenoviral transduction into SCC cells by single intratumoral injection is 3–5%. However, as MMP-13 is a secreted proteinase, even this limited transduction efficiency appears to be sufficient to reduce the overall proteolytic capacity within the tumor as a result of repeated injections. Together, these observations show that MMP-13 plays a role in SCC cell implantation and subsequent tumor growth in vivo. Interestingly, the inhibition of tumor growth was associated with reduction in the number of Ki67positive cells in RAdMMP-13ASRz-injected tumors and the relative number of Ki67-positive cells correlated with tumor size. These results clearly show that antiMMP-13 ribozyme inhibits SCC cell survival and growth in vivo. Detection of apoptotic activated caspase-3-positive cells in the xenografts in the end of the experiment revealed no differences between antiMMP-13 ribozyme-injected and control tumors. This is most likely due to the fact that a subpopulation of tumor cells undergoes apoptosis and is eliminated after each adenoviral injection during the entire 20-day experiment and this can be detected as reduction in the relative number of proliferating cells in the end of the experiment. Interestingly, adenoviral delivery of anti-MMP-13 ribozyme potently inhibited gelatinolytic activity in SCC xenografts in vivo, although MMP-13 is not as potent a gelatinase as MMP-2 and MMP-9, the production of which was not inhibited by anti-MMP13 ribozyme. There are two possible explanations for this. First, MMP-13 has been shown to activate latent MMPs with gelatinolytic activity, especially MMP-9 and MMP-2 (Knaüper et al., 1997b; Murphy and Knaüper, 1997). Therefore, the reduction in gelatinolytic activity may represent reduction in activation of MMP-2 and MMP-9 produced by SCC cells in vivo. Secondly, reduced gelatinolytic activity in anti-MMP-13 ribozyme adenovirus-injected tumors is also likely to represent reduction in viable cell number, as demonstrated by reduction in tumor size and relative number of proliferating cells in these tumors. The results reported here demonstrate for the first time the therapeutic potential of ribozyme-based specific targeting of MMP-13 expression in inhibition of SCC growth in vivo. The viral vector-based approach may have clinical utility in the locoregional therapy of MMP13-positive tumors in patients with limited treatment options available. However, for successful therapeutic application, improved delivery approaches will be Oncogene


5120

required to mediate high-level expression of ribozyme. One potential viral vector type for ribozyme delivery is adenoassociated virus, which allows long-term expression of transgene in infected cells in vivo (Hernandez et al., 1999). Another approach is the direct delivery of ribozyme molecules to tissues using nuclease-resistant ribozymes to overcome the limitations due to the short half-life of RNA. The nuclease-resistant chemically synthesized ribozymes can be administered subcutaneously or intravenously, have excellent specificity, and are well tolerated (Usman and Blatt, 2000). The nuclease-resistant ribozyme targeted against VEGF receptor mRNA has been shown to decrease lung metastases in mouse model in a dose-dependent manner (Pavco et al., 2000). Nevertheless, it is conceivable that MMP-13 targeted therapy should be feasible both in malignant tumors, in which tumor cells are positive for MMP-13, such as head and neck SCCs, as well as in therapy of tumors with stromal cells positive for MMP13, for example, breast carcinomas (Nielsen et al., 2001). In addition to invasive malignant tumors, expression of MMP-13 has also been detected in inflammatory conditions characterized by destruction of normal collagenous tissue architecture in osteoarthritic cartilage, rheumatoid synovium, chronic cutaneous ulcers, intestinal ulcerations, chronic periodontitis, atherosclerosis, and aortic aneurysms (Reboul et al., 1996; Lindy et al., 1997; Mao et al., 1999; Sukhova et al., 1999; Uitto et al., 1998; Vaalamo et al., 1998). As MMP13 is not expressed in most normal adult human tissues, targeted downregulation of MMP-13 expression by interfering with MMP-13 mRNA stability and translation may serve as a novel and important strategy for therapy of cancer and inflammatory diseases involving untimely expression of this potentially destructive MMP.

Materials and methods Cell cultures Human SCC cell line UT-SCC-7, established from metastasis of cutaneous SCC (Johansson et al., 1997a), was cultured in DMEM supplemented with 6 mmol/l glutamine, nonessential amino acids, and 10% fetal calf serum (FCS). HaCaT cells, an immortalized human keratinocyte cell line (Boukamp et al., 1988), and A-5 cells, a ras-transformed tumorigenic HaCaT cell line (Boukamp et al., 1990) (both kindly provided by Dr Norbert E Fusenig, Deutsche Krebsforschungszentrum, Heidelberg, Germany), were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% FCS. Design of MMP-13 ribozyme Altogether, four distinct antisense hammerhead ribozymes targeted to cleave human MMP-13 mRNA were designed and tested for their ability to cleave human MMP-13 mRNA in vitro. Of these, the anti-MMP-13 ribozyme targeted to nucleotides 707–724 in the coding region of the human MMP-13 mRNA sequence (Freije et al., 1994) with the cleavage site targeted between the nucleotides 716 and 717 was selected for further studies (Figure 1a). A corresponding Oncogene

hammerhead ribozyme control was designed in sense orientation to the same nucleotides. The following oligonucleotides were used for cloning of MMP-13 ribozyme expression vectors. The flanking restriction enzyme cleavage sites are underlined. Oligonucleotide MMP-13 AS Rz rev MMP-13 AS Rz frw MMP-13 senseRz rev MMP-13 senseRz frw

Sequence 50 -TCTAGATCCTTAGGTTTCGT CCTCACGGACTCATCAGTTGA CCACGAATTC-30 50 -GAATTCGTGGTCAACTGAT GAGTCCGTGAGGACGAAACC TAAGGATCTAGA-30 50 -TCTAGAGGAATCCATTCGTC CTCACGGACTCATCAGAACTG GTGGAATTC-30 50 -GAATTCCACCAGTTCTGATG AGTCCGTGAGGACGAATGGAT TCCTCTAGA-30

Equal amounts of reverse (rev) and forward (frw) oligonucleotides were heated to 801C and allowed to cool to room temperature and anneal to form double-stranded DNA molecules encoding MMP-13 antisense and MMP-13 sense ribozymes, which were then subcloned into pCI-neo vector (Promega). Antisense and sense MMP-13 ribozymes were generated by in vitro transcription of linearized pCIneo-ribozyme vectors using T7 RNA polymerase. MMP-13 mRNA was transcribed from linearized pCI-MMP13neo plasmid (Ala-aho et al., 2002b), resulting in an RNA molecule of 1442 nucleotides in length (Figure 1b). Ribozyme and the target MMP-13 RNA were heated to 801C in the presence of reaction buffer (50 mM Tris–HCl, pH 7.5, 1 mM EDTA, and 50 mM NaCl), and allowed to cool to room temperature. DTT (at the final concentration 10 mM), RNase inhibitor (10 U), and MgCl2 (20 mM) were added and the reactions were incubated at 371C for different periods of time. Reactions were stopped by addition of 5 RNA loading buffer and products were fractionated on a 5% polyacrylamide gel containing 7 M urea, and stained with 10 mg/ml ethidium bromide. Construction of adenoviruses coding for MMP-13 antisense and sense ribozymes Replication-deficient (E1/E3) adenoviruses harboring MMP-13 antisense and sense ribozyme coding sequences were constructed, as previously described (Ala-aho et al., 2002b). The corresponding double-stranded DNA molecules were subcloned into pCA3 shuttle vector under the control of CMV IE promoter. Adenoviral genomic plasmid pBHG10 and the shuttle vectors containing the ribozyme coding region were co-transfected into HEK293 cells (all from Microbix Biosystems, Toronto, ON, Canada). When plaques were visible, cells were harvested in PBS containing 10% glycerol, viruses were released with freon extraction and subjected to plaque purification. Positive recombinants were identified by PCR and sequenced with pCA3 vector-specific primers using recombinant clone viral DNA as template. Positive clones of adenoviruses (RAdMMP-13ASRz and RAdMMP-13senseRz) were chosen to generate high titer preparation (Ala-aho et al., 2002b). Determination of viral titer was conducted as described previously (Lu et al., 1998). Adenoviral cell infections The multiplicity of infection (MOI) for obtaining maximal infection efficiency of UT-SCC-7 cells has been determined


5121 previously (Ala-aho et al., 2002a). The MOI for obtaining maximal infection efficiency of HaCaT and A-5 cells was determined as previously (Ala-aho et al., 2002a) using recombinant adenovirus RAdLacZ, which contains the Escherichia coli b-galactosidase gene (lacZ) under the control of CVM IE promoter (Wilkinson and Akrigg, 1992) (kindly provided by Dr Gavin WG Wilkinson, University of Cardiff, Wales). In the experiments, cells were infected with adenoviruses at MOI 700 for UT-SCC-7 cells, or at MOI 500 for HaCaT and A-5 cells, which gives 100% transduction efficiency in these cells, incubated for 6 h in DMEM with 0.5% FCS. The medium was changed and incubations were continued for 24 h prior to invasion assays or 24–96 h prior to determination of MMP production or cell viability. Assay of MMP-13 and MMP-1 production The production of MMP-13 and MMP-1 was determined by Western blot analysis (Ala-aho et al., 2000) using monoclonal antibody (181-15A12) against human MMP-13 (Calbiochem, San Diego, CA) (1 : 100) and rabbit polyclonal antibody against human MMP-1 (kindly provided by Dr H BirkedalHansen, NIDR, Bethesda, MD, USA) (1 : 5000), followed by detection of specifically bound primary antibodies with peroxidase-conjugated secondary antibodies and visualized by enhanced chemiluminescence (ECL; Amersham). Gelatin zymography Aliquots of conditioned media were fractionated on 10% SDS–PAGE containing 1 mg/ml gelatin (G-9382; Sigma) and 0.5 mg/ml 2-methoxy-2,4-diphenyl-3(2H)-furanone (Fluka 645989). The gels were washed for 30 min in 50 mM Tris, 0.02% NaN3 and 2.5% Triton X-100, pH 7.5, and for 30 min in the same buffer supplemented with 5 mM CaCl2 and 1 mM ZnCl2. The gels were then incubated in 50 mM Tris, 0.02% NaN3, 5 mM CaCl2 and 1 mM ZnCl2 for 24 h at 371C, fixed in 50% methanol/7% acetic acid, stained with 0.2% Coomassie Blue G250, and photographed (Ala-aho et al., 2000). Invasion assays Cell culture inserts (Falcon 3097, Becton Dickinson) with 8.0 mm pore size were coated with 25 mg of reconstituted basement membrane (Matrigel, Becton Dickinson), as described previously (Ala-aho et al., 2000). For invasions, cells infected with antisense and sense MMP-13 ribozyme adenoviruses (2 105/chamber) suspended in DMEM containing 0.1% BSA were seeded on top of the gel in the upper chamber in a final volume of 200 ml, with DMEM (700 ml) containing 10% FCS as chemoattractant in the lower chamber. After 24 h, cells on the upper surface were gently removed with a cotton bud and the invaded cells on the lower surface were fixed in 2% paraformaldehyde, stained with 0.1% crystal violet, and counted.

in DMEM for different periods of time. Cells were trypsinized and counted in each time point. For determination of proliferating cell number, DNA synthesis in RAdMMP-13ASRz- and RAdMMP-13senseRZinfected, and uninfected control UT-SCC-7 and HaCaT cells was analysed by measuring 5-bromo-2-deoxyuridine (BrdU) incorporation in the colorimetric Cell Proliferation ELISA assay (Roche Diagnostics, Mannheim, Germany). The assay was performed according to the manufacturer’s instructions. Briefly, the cell cultures on 96-well plates were labeled with BrdU-labeling reagent (10 ml/well) for 5 h after each time point. The cells were fixed and incubated with peroxidase-labeled anti-BrdU for 90 min. After three washes with PBS, 100 ml/well tetramethyl-benzidine (TMB) was added and the plates were incubated until color development (15 min). Absorbance values were measured at 405 nm using an ELISA reader. Determination of apoptotic cells Adenovirus-infected SCC cells were cultured in serum-free DMEM on glass slides for different periods of times, washed with PBS, fixed with ice-cold methanol, and washed with PBS. To detect apoptotic cells, the TUNEL reaction was performed using the In Situ Cell Death Detection Kit (Roche) according to the manufacturer’s instruction. TUNEL-positive cells from each slide were counted under a fluorescent microscope. In parallel cultures, the nuclei of SCC cells were stained with Hoechst-33342 (10 mg/ml), and analysed by fluorescence microscopy for detection of apoptotic cells. Growth of human SCC xenografts in SCID/SCID mice All experiments with mice were performed according to institutional animal care guidelines and with the permission of the animal test review board of the University of Turku, Finland. Male severe combined immunodeficiency (SCID/ SCID) mice, 6–8 weeks old, were used in all experiments. In ex vivo experiments, UT-SCC-7 cells were infected as described above at MOI 700 for 6 h, washed with PBS, incubated in fresh medium for 18 h and detached with trypsin. Trypsin was neutralized with 10% FCS in DMEM and cells (5 106/ mouse) in 100 ml of PBS were injected subcutaneously to the back of SCID mice (n ¼ 5 for each experimental group). Tumor size was measured twice a week and calculated as length width2 0.5. For intratumoral injection of recombinant adenoviruses, tumors were established by injecting UT-SCC-7 cells (5 106) subcutaneously to the back of mice and allowed to grow to the size of 100 mm3. The adenoviruses (1 109 pfu) in 0.1 ml PBS were then injected intratumorally 2–3 times a week for 3 weeks. Tumor size was measured before each injection and calculated as above. Infection efficiency was determined by single injection of RAdLacZ into tumor, followed by fixation and staining for b-galactosidase activity (Ahonen et al., 2002). RNA analysis by RT–PCR

Determination of viable and proliferating cell number For cell viability assays, 1.5 104 cells were seeded on 96-well plates and infected with adenovirus RAdMMP-13ASRz or with the corresponding sense adenovirus RAdMMP-13senseRz (MOI 700) for 6 h. The cells were incubated for different periods of time and the number of viable cells was determined by CellTiter 96t AQueous Non-Radioactive Cell Proliferation Assay (Promega, Madison, WI, USA). In addition, UT-SCC-7 and HaCaT cells were seeded on 35 mm plates (2 104 cells/plate) in triplicate, infected with recombinant adenoviruses as above, and cultured in 0.5% FCS

Total RNA was isolated from SCC xenografts using RNeasy kit (Qiagen, Chatsworth, CA, USA) and the expression of MMP-13 mRNA was determined by RT–PCR. Aliquots of total RNA (100 ng) were reverse transcribed into cDNA and a 300 bp fragment of human MMP-13 cDNA corresponding to nucleotides 534–833 (Freije et al., 1994) was amplified by PCR using human MMP-13-specific primers, as described previously (Ravanti et al., 1999a). As control, forward oligonucleotide (50 -CCCATGGCAAATTCCATGGCA-30 ) and reverse oligonucleotide (50 -TCTAGACGGCAGGTCAGGTC-30 ) were used to amplify glyceraldehyde-3-phosphate dehydrogenase Oncogene


5122 (GAPDH) with 40 cycles of denaturation at 941C, annealing at 661C, and extension at 721C. The PCR products were subjected to electrophoresis on a 2% agarose gel and visualized by ethidium bromide. Immunohistochemistry Tumors were fixed overnight in phosphate-buffered 10% formalin and embedded in paraffin. Serial sections of 5 mm were deparaffinized and processed with citrate buffer in a microwave oven. Proliferating cells were detected immunohistochemically on paraffin-embedded sections using monoclonal antibody against human Ki67 (MIB-1; DAKO, Denmark). Negative control sections were incubated without primary antibody. The area of Ki67-positive cells was determined in four distinct fields at 4 magnification from all tumor sections using Soft Imaging System’s analySISs program. Mayer’s hematoxylin was used as counterstain in all immunostainings. For detection of apoptotic cells, tissue sections were immunostained using rabbit polyclonal antibodies against cleaved caspase-3 (Asp-175; Cell Signaling Technology, Beverly, MA, USA), as described previously (Ahonen et al., 2003). Collagen in peritumoral area was detected by Weigert-van Gieson staining. In situ gelatin zymography Pieces of tumors were mounted into Tissue-Tek and flashfrozen in liquid isopentane. Gelatinase activity was detected by

gelatin in situ zymography as previously (George et al., 2000). Briefly, 7 mm frozen sections (4 sections/sample) were applied to glass slides and coated with LM-1 photographic emulsion (Amersham International, UK) diluted 1 : 2 with incubation medium (50 mM Tris–HCl, 50 mM NaCl, 10 mM CaCl2, 0.05% Brij 35, pH 7.6). After incubation overnight at 371C in a humidified box, slides were developed in the light with Kodak D-19 developer (Kodak, Bridgend, Wales, UK) and fixed using Kodak Unifix solution. In addition, gelatinase zymography for a tumor sample injected with RAdMMP-13senseRz was performed with 500 nM of the MMP inhibitor BB-94 (Pfizer, Sandwich, UK). Gelatinolytic activity was identified as white areas of lysis on the black background.

Abbreviations ECM, extracellular matrix; MMP, matrix metalloproteinase; SCC, squamous cell carcinoma; BrdU, 5-bromo-2-deoxyuridine. Acknowledgements We thank Ms Sari Pitka¨nen, Ms Marjo Hakkarainen, and Ms Marita Potila for skillful technical assistance. This study was supported by grants from the Academy of Finland (project 45996), Sigrid Juse´lius Foundation, the Cancer Foundation of Finland, and Turku University Central Hospital (EVO13336), and by research contract with Finnish Life and Pension Insurance Companies.

References Ahonen M, Ala-aho R, Baker AH, George SJ, Greńman R, Saarialho-Kere U and Ka¨ha¨ri V-M. (2002). Mol. Ther., 5, 705–715. Ahonen M, Poukkula M, Baker AH, Kashiwagi M, Nagase H, Eriksson JE and Ka¨ha¨ri V-M. (2003). Oncogene, 22, 2121–2134. Airola K, Johansson N, Kariniemi A-L, Ka¨ha¨ri V-M and Saarialho-Kere U. (1997). J. Invest. Dermatol., 109, 225–231. Airola K, Karonen T, Vaalamo M, Lehti K, Lohi J, Kariniemi AL, Keski-Oja J and Saarialho-Kere UK. (1999). Br. J. Cancer, 80, 733–743. Ala-aho R, Greńman R, Seth P and Ka¨ha¨ri V-M. (2002a). Oncogene, 21, 1187–1195. Ala-aho R, Johansson N, Baker AH and Ka¨ha¨ri V-M. (2002b). Int. J. Cancer, 97, 283–289. Ala-aho R, Johansson N, Greńman R, Fusenig NE, Lo´pezOtı´ n C and Ka¨ha¨ri V-M. (2000). Oncogene, 19, 248–257. Ashworth JL, Murphy G, Rock MJ, Sherratt MJ, Shapiro SD, Shuttleworth CA and Kielty CM. (1999). Biochem. J., 340, 171–181. Bostro¨m PJ, Ravanti L, Reunanen N, Aaltonen V, So¨derstro¨m K-O, Ka¨ha¨ri V-M and Laato M. (2000). Int. J. Cancer, 88, 417–423. Boukamp P, Petrussevska RT, Breitkreutz D, Hornung J, Markham A and Fusenig NE. (1988). J. Cell. Biol., 106, 761–771. Boukamp P, Stanbridge EJ, Foo DY, Cerutti PA and Fusenig NE. (1990). Cancer Res., 50, 2840–2847. Cazorla M, Hernandez L, Nadal A, Balbı´ n M, Lo´pez JM, Vizoso F, Fernandez PL, Iwata K, Cardesa A, Lo´pez-Otı´ n C and Campo E. (1998). J. Pathol., 186, 144–150. Etoh T, Inoue H, Yoshikawa Y, Barnard GF, Kitano S and Mori M. (2000). Gut, 47, 50–56. Fosang AJ, Last K, Knaüper V, Murphy G and Neame PJ. (1996). FEBS Lett., 380, 17–20. Oncogene

Freije JM, Die´z-Itza I, Balbı´ n M, Sańchez LM, Blasco R, Tolivia J and Lo´pez-Otı´ n C. (1994). J. Biol. Chem., 269, 16766–16773. George SJ, Lloyd CT, Angelini GD, Newby AC and Baker AH. (2000). Circulation, 101, 296–304. Hernandez YJ, Wang J, Kearns WG, Loiler S, Poirier A and Flotte TR. (1999). J. Virol., 73, 8549–8558. Hua J and Muschel RJ. (1996). Cancer Res., 56, 5279–5284. Jarvis TC, Bouhana KS, Lesch ME, Brown SA, Parry TJ, Schrier DJ, Hunt III SW, Pavco PA and Flory CM. (2000). J. Immunol., 165, 493–498. Johansson N, Airola K, Greńman R, Kariniemi A-L, Saarialho-Kere U and Ka¨ha¨ri V-M. (1997a). Am. J. Pathol., 151, 499–508. Johansson N, Saarialho-Kere U, Airola K, Herva R, Nissinen L, Westermarck J, Vuorio E, Heino J and Ka¨ha¨ri V-M. (1997b). Dev. Dyn., 208, 387–395. Johansson N, Vaalamo M, Greńman S, Hietanen S, Klemi P, Saarialho-Kere U and Ka¨ha¨ri V-M. (1999). Am. J. Pathol., 154, 469–480. Johansson N, Westermarck J, Leppa¨ S, Ha¨kkinen L, Koivisto L, Lo´pez-Otı´ n C, Peltonen J, Heino J and Ka¨ha¨ri V-M. (1997c). Cell Growth Differ., 8, 243–250. Knaüper V, Cowell S, Smith B, Lo´pez-Otı´ n C, O’Shea M, Morris H, Zardi L and Murphy G. (1997a). J. Biol. Chem., 272, 7608–7616. Knaüper V, Lo´pez-Otı´ n C, Smith B, Knight G and Murphy G. (1996). J. Biol. Chem., 271, 1544–1550. Knaüper V, Smith B, Lo´pez-Otı´ n C and Murphy G. (1997b). Eur. J. Biochem., 248, 369–373. Lewin AS and Hauswirth WW. (2001). Trends Mol. Med., 7, 221–228. Lindy O, Konttinen YT, Sorsa T, Ding Y, Santavirta S, Ceponis A and Lo´pez-Otı´ n C. (1997). Arthritis Rheum., 40, 1391–1399.


5123 Lu Y, Zhang Y and Steiner MS. (1998). DNA Cell Biol., 17, 643–645. Mao D, Lee JK, VanVickle SJ and Thompson RW. (1999). Biochem. Biophys. Res. Commun., 261, 904–910. McQuibban GA, Butler GS, Gong JH, Bendall L, Power C, Clark-Lewis I and Overall CM. (2001). J. Biol. Chem., 276, 43503–43508. McQuibban GA, Gong JH, Tam EM, McCulloch CA, ClarkLewis I and Overall CM. (2000). Science, 289, 1202–1206. Murphy G and Knaüper V. (1997). Matrix Biol., 15, 511–518. Nelson AR, Fingleton B, Rothenberg ML and Matrisian LM. (2000). J. Clin. Oncol., 18, 1135–1149. Nielsen BS, Rank F, Lo´pez JM, Balbin M, Vizoso F, Lund LR, Dan K and Lo´pez-Otı´ n C. (2001). Cancer Res., 61, 7091–7100. Nikkola J, Vihinen P, Vlaykova T, Hahka-Kemppinen M, Ka¨ha¨ri V-M and Pyrho¨nen S. (2001). Melanoma Res., 11, 157–166. Overall CM and Lo´pez-Otı´ n C. (2002). Nat. Rev. Cancer, 2, 657–672. Pavco PA, Bouhana KS, Gallegos AM, Agrawal A, Blanchard KS, Grimm SL, Jensen KL, Andrews LE, Wincott FE, Pitot PA, Tressler RJ, Cushman C, Reynolds MA and Parry TJ. (2000). Clin. Cancer Res., 6, 2094–2103. Ravanti L, Ha¨kkinen L, Larjava H, Saarialho-Kere U, Foschi M, Han J and Ka¨ha¨ri V-M. (1999b). J. Biol. Chem., 274, 37292–37300.

Ravanti L, Heino J, Lo´pez-Otı´ n C and Ka¨ha¨ri VM. (1999a). J. Biol. Chem., 274, 2446–2455. Ravanti L, Toriseva M, Penttinen R, Crombleholme T, Foschi M, Han J and Ka¨ha¨ri V-M. (2001). FASEB J., 15, 1098–1100. Reboul P, Pelletier J-P, Tardif G, Cloutier J-M and MartelPelletier J. (1996). J. Clin. Invest., 97, 2011–2019. Santiago FS and Khachigian LM. (2001). J. Mol. Med., 79, 695–706. Sukhova GK, Scho¨nbeck U, Rabkin E, Schoen FJ, Poole AR, Billinghurst RC and Libby P. (1999). Circulation, 99, 2503–2509. Uitto V-J, Airola K, Vaalamo M, Johansson N, Putnins EE, Firth JD, Salonen J, Lo´pez-Otı´ n C, Saarialho-Kere U and Ka¨ha¨ri V-M. (1998). Am. J. Pathol., 152, 1489–1499. Urı´ a JA, Balbı´ n M, Lo´pez JM, Alvarez J, Vizoso F, Takigawa M and Lo´pez-Otı´ n C. (1998). Am. J. Pathol., 153, 91–101. Usman N and Blatt LM. (2000). J. Clin. Invest., 106, 1197–1202. Vaalamo M, Karjalainen-Lindsberg M-L, Puolakkainen P, Kere J and Saarialho-Kere U. (1998). Am. J. Pathol., 152, 1005–1014. Vaalamo M, Mattila L, Johansson N, Kariniemi A-L, Karjalainen-Lindsberg M-L, Ka¨ha¨ri V-M and SaarialhoKere U. (1997). J. Invest. Dermatol., 109, 96–101. Vihinen P and Ka¨ha¨ri V-M. (2002). Int. J. Cancer, 99, 157–166. Wilkinson GW and Akrigg A. (1992). Nucleic Acids Res., 20, 2233–2239.

Oncogene