Expression and role of matrix metalloproteinases MMP ... - Springer Link

1 downloads 190 Views 149KB Size Report
Huntington Beach, CA 92649, USA. (Received 15 June 1998; accepted in .... phosphatase (Biogenese, San Ramon, CA). Alkaline phosphatase activity was ...
Clin. Exp. Metastasis, 1998, 16, 721–728

11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111

Expression and role of matrix metalloproteinases MMP-2 and MMP-9 in human spinal column tumors Ziya L. Gokaslan*, Shravan K. Chintala*, Julie E. York*, Venkaiah Boyapati*, Sushma Jasti*, Raymond Sawaya*, Gregory Fuller+, David M. Wildrick*, Garth L. Nicolson‡ and Jasti S. Rao* *Department of Neurosurgery and +Department of Pathology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030 and ‡The Institute for Molecular Medicine, Huntington Beach, CA 92649, USA (Received 15 June 1998; accepted in revised form 10 August 1998)

Matrix metalloproteinases (MMPs) have been implicated in the process of tumor invasion and metastasis formation. Thus, we determined the expression of MMPs in various primary and metastatic spinal tumors in order to assess the role of these enzymes in spinal invasion. MMP expression was examined by immunohistochemical localization, and quantitative evaluation of MMP protein content was determined by enzymelinked immunosorbant assay (ELISA) and Western blotting. MMP enzyme activity was determined by gelatin zymography. Lung carcinomas and melanomas metastatic to the spine were shown to have higher levels of MMP-9 activity than those of breast, thyroid, renal metastases and primary spinal tumors. Immunohistochemical analysis revealed similar difference in expression of MMP-9 in tissue samples. When the tissue samples were subjected to gelatin zymography for examination of MMP-2 and MMP-9 activity and to ELISA and Western blotting for quantitative estimation of protein content, the most striking results were obtained for lung carcinomas and melanomas relative to the other tumors. Lung carcinomas and melanomas metastatic to the spine had considerably higher levels of MMP-9 activity than those of primary spinal tumor or breast, thyroid, and renal carcinoma metastases. Within the metastatic tumor category, neoplasms that are known to be associated with the shortest overall survival rates and most aggressive behavior, such as lung carcinomas and melanomas, had the highest levels of MMP-2 and MMP-9 activity compared to those less aggressive metastatic tumors such as breast, renal cell, and thyroid carcinomas. Our results suggest that MMPs may contribute to the metastases to the spinal column, and overexpression of these enzymes may correlate with enhanced invasive properties of both primary and metastatic spinal tumors. Keywords: immunohistochemistry, MMPs, proteases, spinal metastases, TIMPs

Introduction Malignant tumors of the spine rarely occur as primary neoplasms [1]; rather, they typically arise from hematogenous metastasis of a distant neoplasm Address correspondence to: Ziya L. Gokaslan, Department of Neurosurgery, Box 64, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030, USA. Tel: (713) 792–2400; Fax: (713) 794–4950.

© 1998 Kluwer Academic Publishers

or from local invasion of an adjacent chest wall tumor. Among adults, almost half of all metastatic spinal tumors arise from breast, lung, or prostate cancers [2–4], whereas renal cell and gastrointestinal malignancies each account for about 5% of metastatic spinal tumors. Thyroid cancer accounts for a very small proportion of spinal metastases. Primary malignant tumors of the spine are rare. Chondrosarcoma, a slow-growing, locally aggressive Clinical & Experimental Metastasis Vol 16 No 8 721

Z. L. Gokaslan et al. 11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111

cancer, accounts for about 10% of primary malignant spinal tumors and 1–4% of all malignant bone tumors [1,5,6]. Giant cell tumors of the bone are also rare and represent about 4% of all bone tumors [7]. Tumor metastasis requires the expression of degradative enzymes that degrade the surrounding extracellular matrix (ECM) [8]. Although many proteases can cleave ECM molecules, matrix metalloproteinases (MMPs) cleave fibrillar collagens as well as proteoglycans and other matrix proteins of the ECM [9]. Among metalloproteinases, 72-kDa gelatinase A (MMP-2) and 92-kDa gelatinase B (MMP-9) have frequently been found to be elevated in malignant tumors and associated with metastasis formation. These enzymes are secreted in a latent form and require activation by cleavage of their amino-terminal propeptide at an evolutionarily conserved amino acid sequence, a process that can be mediated by plasmin, trypsin, cathepsins, and organomercurials [10]. MMP expression has been correlated with the invasive and metastatic phenotypes of a number of tumor types, including prostatic adenocarcinoma [11], colorectal carcinoma [12], breast carcinoma [13], squamous cell carcinoma [14], gastric carcinoma [15,16], and pancreatic cancer [17,18]. Inhibition of MMP enzyme expression or activity has been shown to reduce tumor invasiveness and metastasis [9,19,20]. Metastatic potential has also been shown to correlate positively with type IV collagenolytic activity in murine and rat [21] tumor models [21–23]. To date, there are no reports on the expression of MMPs in spinal tumors; however, several recent studies reported the involvement of MMPs in giant-cell tumor of the bone [24–26]. This is the first report on the expression of MMPs in primary and metastatic tumors of the spinal column.

Materials and methods Preparation of tissue specimens Five tissue specimens each of 5 metastatic tumor types (breast, lung, renal cell, and thyroid carcinomas, and melanomas), and 3 types of primary tumors (chordoma, chondrosarcoma, and giant-cell) were obtained from patients undergoing spinal surgery at The University of Texas M. D. Anderson Cancer Center. The histological diagnosis of each sample was confirmed by a neuropathologist. The specimens were flash-frozen in liquid nitrogen immediately after surgical removal and stored at –80oC prior to the determination of MMP activity by gelatin zymography. For immunohistochemical analysis of MMP expression, samples were fixed in 10% formalin and embedded in paraffin. 722 Clinical & Experimental Metastasis Vol 16 No 8

Gelatin zymography The enzymatic activity of MMPs was determined in tumor extracts fractionated by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) as described previously [27]. Briefly, tissue extracts (50 ␮g protein of each) were electrophoresed on 8% SDS-polyacrylamide gels containing gelatin (2 mg/ml) from swine skin (Sigma Chemical Company, St Louis, MO). After electrophoresis, the gels were rinsed twice with 2.5% Triton X-100 (30 min each) and incubated at 37oC overnight in 50 mM Tris-HCl buffer (pH 7.5) containing 0.15 M NaCl, 10 mM CaCl2, and 0.05% NaN3. The gels were rinsed, stained with 0.05% Coomassie Blue and destained in a solution containing 10% isopropanol and 10% acetic acid. Gelatinolytic enzymes were detected as transparent bands on a uniform blue background of Coomassie Blue-stained gels. To quantitate the proteolysis of gelatin, gels were washed with 2.5% Triton X-100 and incubated at 37oC for shorter time intervals than those shown in Figure 1 to ensure a better quantitative estimation. The relative band intensities were determined densitometrically. Enzyme-linked immunosorbent assay Quantitative analysis of the MMP-2 and MMP-9 protein content in spinal tumor extracts (25 ␮g) was performed by enzyme-linked immunosorbant assay (ELISA) using antibodies specific for MMP-2 and MMP-9. Tissue extracts containing MMP-2 and MMP-9 were mixed with phosphate buffer and incubated overnight in 96-well microtiter plates. The wells were washed with phosphate-buffered saline (PBS) and incubated with anti-MMP-2 or antiMMP-9 antibody at 25oC for 3 h. The plates were washed with PBS and incubated with an alkalinephosphatase conjugated secondary antibody, and the color was developed with p-nitrophenol phosphate. The concentration of MMP-2 and MMP-9 in the tissue extracts was determined using a standard curve for MMP-2 and MMP-9. Immunohistochemistry MMP-9 immunolocalization was analyzed in 10% formalin-fixed, paraffin-embedded sections. The appropriate concentrations of primary antibody were determined by titrating the antibody on positive and negative control tissue sections. Paraffin sections (4 ␮M thick) were prepared and mounted on Silane-coated glass slides. MMP-9 expression was detected by incubating the dewaxed, blocked sections with rabbit anti-human MMP-9 antibody (1:200 dilution) in 1% bovine serum albumin in PBS for 1 h at room temperature in a humidified

Expression and role of matrix metalloproteases MMP-2 and MMP-9 11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111

Figure 1. Gelatin zymography of tissue extracts obtained from various primary and metastatic spinal tumor samples. Extracts containing equal amounts of protein (50 ␮g) were fractionated by SDS-PAGE using gels containing denatured gelatin as described in ‘Materials and methods’. Gel sample lanes include extracts from breast carcinoma, melanoma, lung carcinoma, thyroid carcinoma, and renal carcinoma, respectively. Lanes 6–8 include extracts from chordoma, chondrosarcoma, and giant-cell tumor of bone, respectively. Molecular size of major fractions is denoted in kilodaltons (kDa).

chamber. After a brief washing in buffer, the tissue samples were incubated in biotinylated, goat antirabbit secondary antibody and streptavidin-alkaline phosphatase (Biogenese, San Ramon, CA). Alkaline phosphatase activity was visualized by the addition of a substrate solution containing naphthol AS-BI phosphate, levamisole, and fast-red TR, which reacts with alkaline phosphatase to form an intense pink color. The sections were later counterstained with hematoxylin. Positive immunoreactivity of MMP-9 was identified by a pink staining of the sections. Negative control studies used a nonspecific IgG as the primary antibody instead of the MMP-9 antibody.

Results Quantitation of tumor-associated MMP-2 and MMP-9 Gelatin zymography was used to identify and quantitate the amounts of MMP-2 and MMP-9 in bands on SDS-polyacrylamide gels by densitometry (Figure 1). Species of pro-92 kDa, pro-72 kDa, and their active

forms (82 kDa and 62 kDa) were present in tumortissue extracts. The intensity of the Mr 72,000 and 92,000 bands was much higher in melanoma lung cancer samples than in breast cancer, thyroid cancer, chondrosarcoma, giant cell tumor and chordoma samples. In addition, the levels of the active forms of MMP-2 and MMP-9 were also significantly higher in melanoma and lung carcinoma (P < 0.001) than in other metastatic and primary tumors (Figure 1). Chondrosarcoma and giant-cell tumor samples displayed higher levels of MMP-2 and MMP-9 (but levels similar to breast, thyroid, and renal metastases) than did chordoma. For each tumor type, the pro MMP-2 and MMP-9 bands were scanned and quantitated by densitometry in three positions, and the peak areas were averaged (Figure 2). The activity of MMP-9 was ten- to fifteen-fold greater in melanoma and lung cancer (P < 0.001) and two- to three-fold higher in breast cancer, thyroid cancer, renal cell cancer, chondrosarcoma, and giant-cell tumor (P < 0.001) than in chordoma (Figure 2A). The activity of MMP-2 was four- to five-fold greater in melanomas and lung tumors and two- to three-fold higher in breast tumors, thyroid tumors, chondrosarcomas and giant-cell tumors than in renal cell tumors and chordomas (Figure 2B). ELISA Using antibodies specific for MMP-2 and MMP-9, the MMP-2 and MMP-9 content of the spinal tumor tissue extracts was determined by ELISA. Figure 3A shows that the amounts of MMP-9 were approximately sixteen times higher in melanomas and lung carcinomas (P < 0.001) and two to six times greater in breast, thyroid, and renal cell carcinomas, chondrosarcomas, and giant-cell tumors (P < 0.001) than that in chordomas. Figure 3B shows that MMP-2 protein content was five to seven times higher in melanomas and lung cancers and two- to three times higher (P < 0.001) in breast and thyroid carcinomas, chondrosarcomas, and giant-cell tumors than that in renal cell carcinomas and chordomas. Immunohistochemistry To examine the possible role of MMPs in metastatic spinal tumors, immunohistochemical staining for the distribution of MMP-9 protein in tissue sections was performed in both metastatic and primary spinal tumors (Figure 4). Immunohistochemical analysis revealed prominent MMP-9 expression in breast, lung, and melanoma samples, with lower expression found in thyroid and renal cell tumor tissues. Breast Clinical & Experimental Metastasis Vol 16 No 8 723

Z. L. Gokaslan et al. 11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111

Figure 2. Quantitative estimation of MMP-2 (A) and MMP-9 (B) activity obtained from gelatin zymography. Bands on gels as depicted in Figure 1 were scanned on a densitometer, and their relative intensities were plotted as a bar graph. The data are the mean ± S.D. of five different samples in each tumor category. The units of peak areas are arbitrary. The 8 samples are arranged and labeled in the same manner as for lanes 1–8 in Figure 1.

Figure 3. Quantitative estimation of MMP-2 and MMP9 protein content by ELISA. Protein from various primary and metastatic spinal tumors (25 ␮g each) was coated onto wells of 96-well microtiter plates, and the quantity of either MMP-2 or MMP-9 was estimated by incubating the wells with primary antibodies against MMP-9 (A) or MMP-2 (B), respectively, followed by incubation with an alkaline-phosphatase conjugated antibody as described in ‘Materials and methods’. The 8 samples are arranged and labeled as for gel lanes 1–8 in Figure 1.

and lung carcinomas showed diffuse cytoplasmic staining, whereas melanoma samples expressed both membranous and cytoplasmic staining. Among the primary tumors, chondrosarcomas and giant-cell tumors expressed cytoplasmic and membranous staining for MMP-9.

Discussion

724 Clinical & Experimental Metastasis Vol 16 No 8

Tumor invasiveness and metastasis formation are major causes of morbidity and mortality in cancer patients. Progression of a tumor from the benign to the malignant state involves several important

Expression and role of matrix metalloproteases MMP-2 and MMP-9 11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111

Figure 4. Immunohistochemical localization of MMP-9 in spinal tumor samples using an antibody against MMP-9 in chordoma, chondrosarcoma, giant-cell tumor of bone, and in metastases from breast, lung, thyroid, and renal carcinoma, as well as melanoma as described in ‘Materials and methods’.

changes and interactions between the tumor cells and the ECM. Although the precise mechanism of tumor invasion and metastasis is unknown, a number of studies have demonstrated that transformation and tumor progression result in increased production of degradative enzymes. These enzymes degrade the surrounding ECM components at different stages of tumor progression and the metastatic cascade, including invasion through the surrounding connective tissue stroma and penetration of the basement membranes of small blood vessels and lymphatics, allowing cancer cells to disseminate to distant sites.

Recent studies have demonstrated that the degradation and remodeling of the extracellular matrix is caused by the expression of various proteolytic enzymes by tumor cells [28]. Among these proteases are the matrix metalloproteinases, a group of metaldependent collagenases, and these have received attention recently because they appear to contribute significantly to the invasive ability of many cancer types. Numerous studies have shown that overexpression of MMPs is associated with an invasive phenotype [6,29]. Moreover, highly aggressive tumor types such as carcinomas, melanomas, hepatomas, Clinical & Experimental Metastasis Vol 16 No 8 725

Z. L. Gokaslan et al. 11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111

and fibrosarcomas have high levels of MMPs compared to their benign counterparts [30]. A substantial body of evidence indicates that MMPs are necessary for cancer invasion and metastasis [9], and activity of MMP-2 and MMP-9 has been strongly associated with the malignant phenotype [28]. MMPs mediate the degradation of collagenous and possibly other protein components of the ECM and facilitate the invasiveness and angiogenesis of malignant tumors. Although the expression of protease activities of some tumors has been determined in several studies, the expression and distribution of MMPs in metastatic spinal tumors was not known. In this study, we examined the expression of MMPs in various primary and metastatic spinal tumors. Gelatin zymography of spinal extracts showed that both MMP-9 (92 kDa) and MMP-2 (72 kDa) were expressed in high amounts in these tumor samples (Figures 1 and 2). Using specific antibodies against MMP-2 and MMP-9, quantitation of tumor enzyme protein content by ELISA demonstrated significantly higher levels in certain tumors, and this was confirmed by Western blotting of tumor tissue extracts. Because both increased production of MMP protein and increased MMP activity were found, immunohistochemistry was performed in order to try to determine the cellular distributions of these enzymes. MMPs are secreted as proenzymes and can be activated by an autocatalytic mechanism that is induced by a variety of agents, including organomercurials or other proteinases. However, the in vivo mechanism of MMP activation is not clearly understood. Recent studies implicate membrane-type MMPs in the specific activation of other MMPs, and this may allow a focal and localized proteolysis at the cell surface and at the leading edge of an invading tumor. Although MMP-2 and MMP-9 have similar substrate-degrading properties, these two enzymes have different activation mechanisms and appear to be regulated by distinct intracellular signaling mechanisms. There are naturally occurring MMP-inhibitory proteins known as tissue-inhibitors of metalloproteinases (TIMPs) that determine the overall enzymatic activity of these proteinases, and when TIMPs bind to MMPs (at a specific ratio of MMPs to TIMPs), they can inhibit enzymatic activity and block tumor cell invasion and metastasis [31,32]. For example, TIMP-1 is a 28.5-kDa glycoprotein that forms a complex (1:1 stoichiometric ratio) with activated interstitial collagenase, activated stromelysin, and MMP-9 [33], whereas TIMP-2 is a 21-kDa protein that forms a complex with MMP-2. Unlike TIMP-1, TIMP-2 is capable of binding to both latent and activated forms of MMP-2 [30]. 726 Clinical & Experimental Metastasis Vol 16 No 8

Changes in TIMP-1 and TIMP-2 gene expression can inhibit tumor invasiveness and metastasis. Because TIMPs block the activity of MMPs, the net balance between TIMPs and MMPs may be important in preventing invasion and metastasis. Indeed, increased expression of a recombinant TIMP-1 gene in gastric cancer cells reduced the formation of metastases [34]. Moreover, injection of c-Ha-ras-1 transformed cells, which express high levels of TIMP-2 in nude mice, reduces the incidence of metastasis [35]. Recently, Okada and colleagues [36,37] demonstrated that MMP-9 can degrade type I and type IV collagens, which are major components of the bone. They also reported that MMP-9 depolymerizes acidinsoluble polymers of type I collagen and digests collagen fibrils in the demineralized bone [36,37]. Our results demonstrate increased production of MMPs in metastatic spinal tumors, suggesting that MMP expression in these metastatic spinal tumors enhances the ability of the tumor cells to degrade and invade the bone. The underlying mechanism of increased expression of MMPs in metastatic spinal tumors is not known. It is possible that decreased expression of TIMPs and increased production of MMPs contribute to the invasive and metastatic potential of these tumors; however, the procedures used to analyze MMP activity (electrophoresis in SDS-PAGE gels and substrate zymography) result in dissociation of the MMP–TIMP complexes. A further understanding of the regulatory mechanisms of MMPs may provide important information that will be helpful in designing proper therapeutic interventions for the prevention of spinal column metastases. In general, metastatic tumors from lung carcinoma and melanoma are considered to be more aggressive and are known to be associated with a poorer prognosis than carcinomas of renal or thyroid origin [38]. Our data on the levels of metalloproteinase expression suggest a similar trend. We found that the highest levels of MMP-2 and MMP-9 activity were in the more aggressive tumor types, lung carcinomas and melanomas. Thus, increased levels of matrix metalloproteinase activities may not only be important for the breakdown of basement membranes, which precedes distant metastasis formation, but also for local dissolution and invasion of the vertebral body after tumor cells have gained access to the spinal column. Either of these sites is a potential controlling step in the invasive behavior of spinal neoplasms, with potential therapeutic applications for both primary and metastatic tumors. It is likely that other degradative enzymes, such as urokinase-type plasminogen activator, also play an important role in this process.

Expression and role of matrix metalloproteases MMP-2 and MMP-9 11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111

References: 1. Sundaresan N, Galicich JH, Chu FC and Huvos AG, 1979, Spinal chordomas. J Neurosurg, 50, 312–19. 2. Black P, 1979, Spinal metastasis: current status and recommended guidelines for management. Neurosurgery, 5, 726–46. 3. Harrington KD, 1986, Metastatic disease of the spine. J Bone Joint Surg Am, 68, 1110–15. 4. Kostuik JP and Weinstein JN, 1991, Differential diagnosis and surgical treatment of metastastic spine tumors. In: Frymoyer JW, ed: The Adult Spine: Principles and Practice. New York: Raven Press, 861–88. 5. Cummings BJ, Esses S and Harwood AR, 1982, The treatment of chordomas. Cancer Treat Rev, 9, 299–311. 6. Rich TA, Schiller A, Suit HD and Mankin HJ, 1985, Clinical and pathologic review of 48 cases of chordoma. Cancer, 56, 182–7. 7. Komiya S, Inoue A, Nakashima M, et al. 1986, Prognostic factors in giant cell tumor of bone. A modified histological grading system useful as a guide to prognosis. Arch Orthop Trauma Surg, 105, 67–72. 8. Nicholson GL, 1989, Metastastic cell interactions with endothelium, basement membrane and tissue. Curr Opin Cell Biol, 1, 1009–19. 9. Matrisian LM, 1992, The matrix degrading metalloproteinases. Bioessays, 14, 455–63. 10. Nagase H, Ogata Y, Suzuki K, et al. 1991, Substrate specificities and activation mechanisms of matrix metalloproteinases. Biochem Soc Trans, 19, 715–18. 11. Stearns ME and Wang M, 1993, Type IV collagenase (M(r) 72,000) expression in human prostate: benign and malignant tissue. Cancer Res, 53, 878–83. 12. Pyke C, Ralfkiaer E, Tryggvason D and Dano K, 1993, Messenger RNA for two type IV collagenases is located in stromal cells in human colon cancer. Am J Pathol, 142, 359–65. 13. Monteagudo C, Merino MJ, San-Juan J, et al. 1990, Immunohistochemical distribution of type IV collagenase in normal, benign and malignant breast tissue. Am J Pathol, 136, 585–92. 14. Pyke C, Ralfkiaer E, Huhtala P, et al. 1992, Localization of messenger RNA for Mr 72,000 and 92,000 type IV collagenases in human skin cancers by in situ hybridization. Cancer Res, 52, 1336–41. 15. Koshikawa N, Yasumitsu H, Umeda M and Miyazaki K, 1992, Multiple secretion of matrix serine proteinases by human gastric carcinoma cell lines. Cancer Res, 52, 5046–53. 16. Schwartz GK, Wang H, Lampen N, et al, 1994, Defining the invasive phenotype of proximal gastric cancer cells. Cancer, 73, 22–27. 17. Taniguchi S, Iwamura T and Katsuki T, 1992, Correlation between spontaneous metastatic potential and type I collagenolytic activity in a human pancreatic cancer cell line (SUIT-2) and sublines. Clin Exp Metastasis, 10, 259–66. 18. Zucker S, Lysik RM, Wieman J, et al. 1985, Diversity of human pancreatic cancer cell proteinases: role of cell membrane metalloproteinases in collagenolysis and cytolysis. Cancer Res, 45, 6168–78. 19. Liotta LA, Steeg PS and Stetler-Stevenson WG, 1991, Cancer metastasis and angiogenesis: an imbalance of positive and negative regulation. Cell, 64, 327–36. 20. Schultz R, Silberman S, Persky G, et al. 1988, Inhibition

21.

22.

23.

24.

25.

26.

27. 28. 29.

30. 31.

32.

33. 34.

35.

36.

by human recombinant tissue inhibitor of metalloproteinases of human aminion invasion and lung colonization by murine B16-F10 melanoma cells. Cancer Res, 48, 5539–45. Nakajima M, Welch D, Belloni PN and Nicholson GL, 1987, Degradation of basement membrane type IV collagen and lung subendothelial matrix by rat mammary adenocarcinoma cell clones of differing metastatic potentials. Cancer Res, 47, 4869–76. Liotta LA, Tryggvason K and Garbisa S, 1980, Metastatic potential correlates with enzymatic degradation of basement membrane collagen. Nature, 284, 67–68. Turpeenniemi-Hujanen T, Thorgeirsson UP, Hart IR, et al. 1985, Expression of collagenase IV (basement membrane collagenase) activity in murine tumor cell hybrids that differ in metastatic potential. J Natl Cancer Inst, 75, 99–103. Rao VH, Singh RK and Bridge JA, et al. 1997, Regulation of MMP-9 (92 kDa type IV collagenase/gelatinase B) expression in stromal cells of human giant cell tumor of bone. Clin Exp Metastasis, 15, 400–9. Rao VH, Bridge JA and Neff JR, et al. 1995, Expression of 72-kDa and 92-kDa Type IV collagenases from human giant cell tumor of bone. Clin Exp Metastasis, 13, 420–6. Sasaguri Y, Komiya S and Sugama K, et al. 1992, Production of matrix metalloproteinases 2 and 3 (stromelysin) by stromal cells of giant cell tumor of bone. Am J Pathol, 141, 611–21. Rao JS, Steck PA, Mohanam S, et al. 1993, Elevated levels of Mr 92,000 type IV collagenase in human brain tumors. Cancer Res, 53, 2208–11. Liotta LA and Stetler-Stevenson WG, 1991, Tumor invasion and metastasis: an imbalance of positive and negative regulators. Cancer Res, 51, 50543–95. Ballin M, Gomez DE, Sinha CC and Thorgeirsson UP, 1988, RAS Oncogene mediated induction of a 92 Kda metalloproteinase; strong correlation with the malignant phenotype. Biochem Biophy Res Commun, 154, 832–8. Stetler-Stevenson WG, 1990, Type IV collagenases in tumor invasion and metastasis. Cancer Metastasis Rev, 9, 289–303. Testa JE, 1992, Loss of the metastatic phenotype by a human epidermoid carcinoma cell line, Hep-3, is accompanied by increased expression of tissue inhibitor of metalloproteinase 2. Cancer Res, 52, 5597–603. Alvarez O, Carmichael D and DeClerck Y, 1990, Inhibition of collagenolytic activity and metastasis of tumor cells by a recombinant human tissue inhibitor of metalloproteinases. J Natl Cancer Inst, 82, 589–95. Welgus HG, Jeffrey JJ, Eisen AZ, et al. 1985, Human skin fibroblast collagenase: interaction with substrate and inhibitor. Collagen Related Res, 5, 167–79. Tsuchiya Y, Sato H and Endo Y, et al. 1993, Tissue inhibitor of metalloproteinase 1 is a negative regulator of the metastatic ability of a human gastric cancer cell line, KKLS, in the chick embryo. Cancer Res, 53, 1397–1402. DeClerck YA, Perez N, Shimada H, et al. 1992, Inhibition of invasion and metastasis in cells transfected with an inhibitor of metalloproteinases. Cancer Res, 52, 701–8. Okada Y, Gonoji Y and Naka K, et al. 1992, Matrix Clinical & Experimental Metastasis Vol 16 No 8 727

Z. L. Gokaslan et al. 11111 2 3 4 5 6 7 8 9 10111 1 2 3 4 5 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5 6 7 8 9 50111 1 2 3 4111

metalloproteinase 9 (92-KDa gelatinase/Type IV collagenase) from the HT 1080 human fibrosarcoma cells. Purification and activation of the precursor and enzymic properties. J Biol Chem, 267, 21712–19. 37. Okada Y, Naka K and Kawamura K, et al. 1995, Localization of matrix metalloproteinase 9

728 Clinical & Experimental Metastasis Vol 16 No 8

(92-kilodalton gelatinase/type IV collagenase = gelatinase B) in osteoclasts: implications for bone resorption. Lab Invest, 72, 311–22. 38. Lang FF and Sawaya R, 1996, Surgical management of cerebral metastases. Neurosurg Clin N Am, 7, 459–84.