2.2. Gelatinases and their inhibitors

2.2.1. Matrix metalloproteinase gene family

The matrix metalloproteinase gene family is a class of metalloproteinases capable of extracellular matrix degradation. Nowadays, 24 members of this enzyme family are known to exist (Woessner 1991, 1999). They have several structural similarities in their protein structure, and they all need a zinc atom for their catalytic action (Tryggvasson et al. 1992). MMPs play a role in a huge field of physiological processes. They are assumed to be participated in collagenolyses in, for example, blastocyst implantation, embryonic development, nerve growth, endometrial cycling, mammary gland morphogenesis, bone remodeling, wound healing angiogenesis, and apoptosis. Among pathologic processes, they are assumed to induce tissue destruction in, for example, rheumatoid arthritis, dilated cardiomyopathy, aortic aneurysm, gastric ulcer, fibrotic lung disease, and cancer invasion and metastasis. (Höyhtyä et al. 1990, Turpeenniemi-Hujanen et al. 1992, Parks & Mecham 1998).

The exact role of each individual MMP in each individual process is far from clear. Indirect evidence may support the coexistence of a specific MMP in a specific process, but this does not necessarily imply a causal relationship. Indeed, it may be that MMPs can replace each other in many processes. This assumption is supported by the fact that most MMP knockout mice do not have a sharply defined phenotype, and depletion of one specific MMP has not led to the death of an organism. Also, the specific substrates for each MMP are not clear. Because there are 100 known macromolecular components of the extracellular matrix, it will still take a huge amount of research to clarify precisely which component is a substrate for each specific MMP.

The present work focuses on MMP-2 and MMP-9 (gelatinases A and B), which are the main enzymes able to degrade type IV collagen and gelatin, the product of collagen degradation after lysis with collagenases. Type IV collagen is an integral part of basement membranes. Because the disruption of basement membrane is a key step in malignant transformation, these two MMPs are assumpted to play a key role during metastazing process. Allthough their role is widely studied in solid malignancies little is known about thei role in hematological malignancies which differ a lot of solid tumors by their basic biological features.

2.2.2. Matrix metalloproteinase-2

Many different types of tissue express continuously MMP-2. Unlike most other MMPs, MMP-2 activity has little regulation at the transcriptional level, and most of the regulation takes place at the post-transcriptional level and during enzyme activation and inactivation. This has been verified in experiments showing only minimal variation in the MMP-2 gene transcription, but several agents seems to regulate the gene transcription of the other MMPs (Brown et al. 1990, Overall & Sodek 1990). This is further supported by the fact that the MMP-2 promoter region differs markedly from the promotors of the other MMPs (Matrisian 1994). On the other hand, TGF-α1, for example, can increase the MMP-2 mRNA half-life from 46 to 150 hours at the post-transcriptional level (Overall et al. 1991).

MMP-2 is synthesized and secreted as an inactive proenzyme, which is stable in physiological conditions. After the secretion, proMMP-2 is assumed to be bound to its specific inhibitor, the tissue inhibitor of matrix metalloproteinase-2 (TIMP-2). It is possible that there are several pathways leading to the activation of the proenzyme, but physiologically the most important pathway is thought to be activation by membrane-type metalloproteinases-1 (MT1-MMP). (Brown et al. 1990, Ward et al. 1991, Murphy et al. 1994) MT1-MMP is capable of binding to TIMP-2, bringing the complexed proMMP-2 in proximity to the active site of the MT-MMP enzyme. This leads to the cleavage of two specific propeptides from proMMP-2 and the generation of an active 62 kDa MMP-2 enzyme, which is still bound to TIMP-2 (Bergman et al. 1995, Strongin et al. 1995) (Figure 1). This membrane-bound activation of pro-MMP-2 allows the collagenolytic activity to be focused into the immediate vicinity of the invasive cell and prevents excessive matrix degradation (Brown et al. 1990). This has important implications for the biologic and pathologic differences between the two gelatinases (Table 5).

The binding of a second TIMP-2 molecule or TIMP-1 to the catalytic site of MMP-2 leads to the inactivation of the enzyme. TIMP-2 thus plays a dual role in the regulation of MMP-2 activity. On the other hand, it is essential for the activation of the proenzyme, but it also has an integral role in the inactivation of the active form of the enzyme. Other TIMPs do not seem to be capable of pro-MMP-2 activation, and TIMP-3 and TIMP-4, but not TIMP-1, are even able to inhibit the activation process (Groft et al. 2001, Hernandez-Barrantes et al. 2001, Pepper 2001).

MMP-2 was first discovered in studies on basement membrane disruption. Lately it has been shown to be able to degrade type IV collagen, which is an integral part of basement membrane (Liotta et al. 1979). Metastatic potential of a solid tumor has been shown to correlate with it´s ability to degrade basement membranes (Liotta et al. 1980). Indeed it is not a surprise that in several solid tumor MMP-2 expression correlates with invasive phenotype and poor prognosis (garzetti et al. 1996, Gohji et al. 1996, Väisänen et al. 1996, 1998, Talvensaari-Mattila et al. 1998, Westerlund et al. 1999, Jäälinoja et al. 2000). Other matrix macromolecules suspectible to lysis by MMP-2 are listed in table 5.

Although many tumours have been shown to express MMP-2 protein, not all of them have been shown to express MMP-2 mRNA (Höyhtyä et al.1994). Indeed, many of these tumors have the mRNA localized in stromal cells (Autio-Harmainen et al. 1993, Polette et al. 1994). Tumors may induce stromal cells to synthesize MMP-2, and the secreted protein is recruited by receptors on the surface of neoplastic cells. On the other hand, stromal cells can also induce the MMP-2 synthesis of tumor cells, indicating the presence of complicated interactions between the malignant tumor and the surrounding extracellular matrix. (Westerlund et al. 1997, Kossakowska et al. 1999a)

Along with enzymatic functions, MMP-2 and its degradation products also have regulatory functions. MMP-2 or some of its degradation products decrease cell adhesion, increase migration by cleaving the Ln-5 integrin receptor, thus triggering cell motility, and may serve as chemotactic stimuli (Calof et al. 1994). MMP-2 may induce mesangial cells into an inflammatory phenotype, and neutralization of MMP-2 with the retroviral antisense technique leads to a nearly complete elimination of these cells from the cell cycle (Turck et al. 1996). MMP-2 may also play a role in the cessation of an inflammatory response (McQuibban et al. 2002).

In the start on neovascularization process MMP-2 and other MMPs lyse the matrix surrounding endothelial cells thus enabling the invasion of new vascular structures into the tissues (Pepper 2000). Indeed decreased tumor growth and new vessel formation has been detected in MMP-2 knock-out mice (Itoh et al. 1998).

2.2.3. Matrix metalloproteinase-9

MMP-9 is another metalloproteinase capable of basement membrane degradation in vivo. Unlike MMP-2, which is constitutively expressed by many cells, MMP-9 expression normally only occurs in trophoblasts, osteoclasts, and leukocytes and their precursors (Borregaard et al. 1995, Harvey et al. 1995, Janowska- Wieczorek et al. 1995, Witty et al. 1996). While MMP-2 expression has only slight control at the transcriptional level, MMP-9 transcription can be highly induced by a wide range of agents. These agents include growth factors, cytokines, cell-cell and cell-ECM adhesion molecules, and agents altering cell shape. (Dong et al. 2001, Martin et al. 2001) Along with the differences between the quantities of MMP-2 and MMP-9 synthesis induction, there also exist qualitative differences. For example, TGF-α1 strongly up-regulates MMP-9 mRNA expression while simultaneously down-regulating MMP-2 expression. (Thompson et al. 2001) These differences suggest that these two enzymes have different biological functions.

Similarly to MMP-2, MMP-9 is also synthesized as a precursor, which is bound to TIMP-1 (Murphy et al. 1989, Moll et al. 1990). However, in cell cytosol, the enzyme can be stored in either a latent or an active form, which is in contrast to MMP-2, which can be stored only in a latent form (Nguyen et al. 2001). The activation of proMMP-9 is a complex process, which is regulated by interaction with TIMP and other MMPs (Kolkenbrock et al. 1995, Orgel et al. 1995). Numerous enzymes have been suggested to be capable of proMMP-9 activation. These include MMP-2, leukocyte elastase, tissue kallikrein (Menashi et al. 1994, Ferry et al. 1997), stromelysin, collagenase-1 (Kolkenbrock et al. 1995a,b), and trypsin (Bu & Pourmotabbed 1996). MMP-9 has several active metabolites with molecular weights of 82, 67, 49, 41.5 and 40 kDa.

All TIMPs can inactivate MMP-9, but TIMP-2 seems to have the highest specific activity (Howard et al. 1991).

MMP-9 participates in the invasion of cells through matrix barriers and collagenolysis during invasion and tumor progression by degrading the matrix macromolecules. There are numerous reports demonstrating the ability of MMP-9 to cleave type IV collagen in vitro. The in vivo situation, however, is not equally clear. In addition to type IV collagen, MMP-9 is able to cleave the type V and XI collagens (Pourmotabbed et al. 1994). To a lesser degree, it also has activity against aggrecan (Fosang et al. 1992) and elastin (Senior et al. 1991), but not against type I collagen (Murphy et al. 1982). The known substrates of MMP-9 are presented in table 5.

Physiologically, MMP-9 participates in trophoblast implantation, bone development, wound healing, and inflammatory processes, probably by enabling inflammatory cells to invade into the inflammatory focus and by participating in the regulation of inflammatory responses (Borregaard et al. 1995, Harvey et al. 1995, Janowska-Wieczorec et al. 1995, Goetzl et al. 1996,Witty et al. 1996, Sheu et al. 2001).

Although there are physiologically only a few cell types expressing MMP-9, there are a wide range of tumors showing MMP-9 expression either in the tumor cells or in the normal cells surrounding the tumor (Pyke et al. 1992, Canete-Soler et al. 1994, Soini et al. 1994, Ashida et al. 1996, , Iwata et al. 1996). Many animal studies suggest that MMP-9 (along with MMP-2) has a critical role in tumor invasion (Sier et al.1996). For example, the human osteosarcoma cell line up-regulates MMP-9 expression in response to TNF-α and becomes more invasive in vitro. Treatment of these cells with TNF-α prior to injection into nude mice results in an increased number of lung metastases in a dose-dependent manner (Kawashima et al. 1994). In ICAM-deficient nude mice, lymphomas are not able to disseminate before they attain the capability of continuous MMP-9 expression (Lalancette et al. 2000).

2.2.4. Tissue inhibitor of metalloproteinases-1

Tissue inhibitors of metalloproteinases share the common capability to inactivate metalloproteinase enzymes. Four members of this protein family, TIMP-1, TIMP-2, TIMP-3, and TIMP-4, have been discovered (Vincenti 2001).

TIMP-1 plays a role during the activation of MMP-9 (Kolkenbrock et al. 1995). It also has a capability to inactivate the active forms of both MMP-2 and MMP-9. In accordance with these findings, TIMPs have been shown to be able to inhibit tumor dissemination in rats and mice (Eccles et al. 1996, Parker et al. 2000). In TIMP-1 transgenic mice, the metastatic potential of intradermally injected lymphoma cells has an inverse correlation with TIMP-1 expression (Kruger et al. 1997).

However, in contrast to its anti-invasive property, strong TIMP-1 expression is associated with an adverse outcome in several tumor models, including lymphomas, colorectal cancer, and lung cancer (Kossakowska et al. 1993, Ylisirniö et al. 1999, Stevenson & Brummer 2000, Yoshika et al. 2001). This probably reflects the fact that it actually has a much wider role in regulating the basic biological functions apart from simply inactivating MMPs. Indeed, it has been shown to be an antiapoptotic and growth-stimulating factor for several normal tissues and malignant cell lines (Hayakawa et al. 1990, 1992, Guedez et al. 1998a,b, 2001a). It is also antiangiogenic and decreases insulin-like growth factor II signaling (Martin et al. 1999, Brew et al. 2000). In accordance with these findings, Guedez et al. have shown that induction of TIMP-1 expression in Burkitt´s lymphoma cells implanted in nude mice first induces acclerated tumor growth due to the inhibited apoptosis. After that, however, the tumor spontaneously regresses due to the impaired new vessel formation. (Guedez et al. 2001a)

2.2.5. Gelatinases and angiogenesis

Induction of the host’s vasculature to grow into the tumor to supply it is a phenomenon integral to cancer progression. The magnitude of angiogenesis is dependent on the balance between the positive and negative regulatory factors around endothelial cells, called ”angiogenic switch”. This, in turn, includes numerous regulatory proteins, such as vascular endothelial cell growth factor (VEGF), fibroblast growth factor (FGF), and angiostatin, and activation and inactivation of proteolytic enzymes, including MMPs, and serine proteases and their inhibitors. (Pepper 2001)

It has been evident since the classic experiments of Clark & Clark 1939 that the new vessel formation is accompanied by excessive matrix degradation (Clark & Clark 1939). This is a critical step initiating the process leading to increased vascular permeability, vascular wall disassembly, basement membrane degradation, and endothelial cell migration and invasion. In fact, both MMP-2 and MMP-9 have been shown to be capable of inducing tube formation in endothelial cell culture (Schapner et al. 1993, Chandrasekar et al. 2000). However, at the end of neovascularization process, the capability to switch off proteolytic activity is critical to the stabilization of the new vessel because uncontrolled proteolysis at this step is associated with regression of the newly formed vessels (Zhu et al. 2000, Davis et al. 2001).

Resting endothelial cells do not express detectable amounts of MMPs or TIMPs. However, after growth stimulation in such situations as wound healing, rheumatoid arthritis, and various malignant tumors, they appear to express MMP-1, -2, -7, -9, as well as TIMP-1 and -2. (Pepper 2000). Moreover, various angiogenic regulators, including VEGF and bFGF, are able to induce MMP expression while simultaneously decreasing TIMP expression (Unemori et al. 1992, Cornelius et al. 1995). MT-MMP seems to be critical for extracellular matrix fibrinolysis and the activation of MMP-2 (Hiraoka et al. 1998). The integral role of gelatinases in neovascularization has also been verified in knockout animal tumor models (Itoh et al. 1997, Vu et al. 1998, Zhou et al. 2000). This is further supported by the fact that MMP inhibitors are also potent angiogenesis inhibitors (Guedez et al. 2001, Hajitou et al. 2001)

Along with their role in endothelial cell invasion, gelatinases also have a role in the regulation of neovascularization. MMP-9 releases VEGF from its matrix stores, which phenomenon seems to be critical for the beginning of the process and cannot be substituted with any other enzymes (Bergers et al. 2000). MMP-9 is also able to convert plasminogen into angiostatin, a potent angiogenesis inhibitor (Sang 1998). Many connective tissue degradation products also carry along with them inhibitory or stimulatory activities in neovascularization (Colorado et al. 2000, Kamphaus et al. 2000, Maeshima Y et al. 2000).