MT1-MMP: A potent modifier of pericellular microenvironment

Authors


Abstract

Cells are regulated by many different means, and there is more and more evidence emerging that changes in the microenvironment greatly affect cell function. MT1-MMP is a type I transmembrane proteinase which participates in pericellular proteolysis of extracellular matrix (ECM) macromolecules. The enzyme is cellular collagenase essential for skeletal development, cancer invasion, growth, and angiogenesis. MT1-MMP promotes cell invasion and motility by pericellular ECM degradation, shedding of CD44 and syndecan1, and by activating ERK. Thus MT1-MMP is one of the factors that influence the cellular microenvironment and thereby affect cell-signaling pathways and eventually alters cellular behavior. As a proteinase, MT1-MMP is regulated by inhibitors, but it also requires formation of a homo-oligomer complex, localization to migration front of the cells, and internalization to become a “functionally active” cell function modifier. Developing new means to inhibit “functional activity” of MT1-MMP may be a new direction to establish treatments for the diseases that MT1-MMP mediates such as cancer and rheumatoid arthritis. © 2005 Wiley-Liss, Inc.

Multicellular organisms consist of cells and surrounding extracellular matrix (ECM). While cells have a major role in tissue functioning, the ECM is essential for the maintenance of tissue architecture, providing cells scaffolding, survival signals, and a growth factor pool, thereby controlling the correct local cellular environment (Werb, 1997). Because cells are directly attached to ECM through their receptors on the cell surface, changes in the immediate ECM microenvironment (i.e., degradation of ECM or deposition of different ECM) will critically affect cellular behavior and tissue functioning (Sternlicht and Werb, 2001). Under physiological conditions, cells take control to change their microenvironment to suit their requirements. When cells loose its control over their microenvironment, such incorrectly organized microenvironment can in turn affect cells' behavior, and repetition or accumulation of such events may result in diseases including cancer, fibrosis, arthritis, and atherosclerosis. Matrix metalloproteinases (MMPs) are the enzymes, which are responsible for degrading ECM components (Nagase and Woessner, 1999). There are 23 MMPs in human and many of them are being characterized (Brinckerhoff and Matrisian, 2002). MMPs can be further divided into two subgroups based on whether the enzymes are secreted or expressed on the cell surface as a membrane-tethered form: soluble MMPs and membrane-type MMPs (MT-MMPs) (Seiki, 1999). Soluble MMPs are secreted from cells into the extracellular milieu and can diffuse to distal sites. Therefore it is thought that this type of MMPs are useful for the degradation of ECM in a wider area, such as seen in tissue resorption, for example. On the other hand, MT-MMPs are expressed on the cell surface as a type I transmembrane protein (Sato et al., 1994) or a glycosylphosphatidylinositol (GPI)-anchored protein (Itoh et al., 1999; Kojima et al., 2000). As a result they can act only in the pericellular space. The restricted distribution of the MT-MMPs provides these enzymes with a unique functional property in that they can alter the immediate microenvironment surrounding the cells (Seiki, 2002). Recently particular attention has been given to MT1-MMP, because this enzyme affects cellular functions in a variety of ways. In this review, we will focus on MT1-MMP by exploring what this enzyme does to affect cellular function by changing the microenvironment and the many ways in which cells can regulate the activity of this enzyme.

GENERAL PROPERTIES OF MT1-MMP

MT1-MMP shares a common domain structure with other MMP family members, including a pre/propeptide (M1–R111), a catalytic domain (Y112–G285), a hinge region (linker-1) (E286–I318), a hemopexin (Hpx) domain (C319–C508), a stalk (linker-2) region (P509–S538), a transmembrane domain (A539–F562), and a cytoplasmic tail (R563–V582) (Sato et al., 1994; Brinckerhoff and Matrisian, 2002; Itoh and Nagase, 2002; Seiki, 2003; Zucker et al., 2003). Like other MMPs, the enzyme is produced as a zymogen and requires an activation step to remove the propeptide proteolytically (Nagase, 1997; Seiki, 1999). MT1-MMP has a basic amino acid motif of RRKR111 at the end of the propeptide which is cleaved by furin or related proprotein convertases (Sato et al., 1996; Yana and Weiss, 2000). Thus, activation of MT1-MMP takes place during secretion in the Golgi (Mazzone et al., 2004), and the enzyme is expressed on the cell surface as the active form. MT1-MMP can degrade a number of ECM macromolecules including collagens I, II, and III, gelatin, laminins 1 and 5, fibronectin, vitronectin, aggrecan, fibrin, and lumican (d'Ortho et al., 1997; Ohuchi et al., 1997; Hiraoka et al., 1998; Koshikawa et al., 2000; Li et al., 2004). MT1-MMP also activates proMMP-2 (Sato et al., 1994) and proMMP-13 (Knäuper et al., 1996). MT1-MMP is O-glycosylated at Thr291, Thr299, Thr300, and/or Ser301 in linker-1 region, and this modification is shown to be essential for the proMMP-2 activation on the cell surface (Wu et al., 2004). MT1-MMP also cleaves several cell surface proteins such as CD44 (Kajita et al., 2001), transglutaminase (Belkin et al., 2001), low-density lipoprotein receptor related protein (Rozanov et al., 2004), αv integrin (Deryugina et al., 2002a), and syndecan-1 (Endo et al., 2003). Expression of MT1-MMP on the cell surface, together with the MMPs it activates, creates wide proteolytic repertoire in the pericellular space, and allows for efficient proteolytic changes to the cellular microenvironment which can result in modification of cellular function (Fig. 1).

Figure 1.

Biological functions of MT1-MMP which promote cell migration and invasion. MT1-MMP sheds CD44 and syndecan-1, degrades collagens and other matrix components, releases EGF-like repeats from laminin 5, activates proMMP-2 and proMMP-13, upregulates ERK, and upregulates VEGF-A expression through Src. All of these activities result in increased cell migration and invasion in tissue.

MODULATION OF CELL FUNCTION BY MT1-MMP

MT1-MMP promotes cell invasion

When a cell migrates in tissue, the ECM is a physical barrier and needs to be degraded to clear a path (Murphy and Gavrilovic, 1999). MMP-2 degrades type IV collagen, a major component of the basement membrane. As cancer cells need to traverse the basement membrane to achieve invasion and metastasis, the activation of proMMP-2 by MT1-MMP is considered to be a critical step. (Stetler-Stevenson et al., 1993; Seiki, 2003). Over expression of MT1-MMP enhances cellular invasiveness in in vitro invasion experiments with Matrigel, a reconstituted basement membrane (Sato et al., 1994) and in an in vivo lung metastasis assay, which requires degradation of real basement membrane (Tsunezuka et al., 1996). Furthermore, MT1-MMP gene silencing using double stranded RNA interference in cancer cells also showed that downregulation of MT1-MMP only is sufficient to inhibit Matrigel invasion of cancer cells significantly (Ueda et al., 2003). These findings strongly suggest that MT1-MMP is indeed a critical part of the invasion machinery.

MT1-MMP is a critical collagen environment modulator

One of the important functions of MT1-MMP is its collagen degrading activity. Collagen is a major ECM component that plays an important role in maintaining tissue architecture and in forming a stable scaffold for cells. It consists of three alpha chains forming triple helical structure, which is then assembled into fibrils. Collagen is resistant to most proteinases at neutral pH apart from collagenolytic MMPs (i.e., MMP-1, MMP-2, MMP-8, MMP-13, and MT1-MMP) (Visse and Nagase, 2003). All collagenolytic MMPs cleave interstitial collagen at a specific site, 3/4 away from the N-terminus, which then initiates denaturation of triple helical collagen into gelatin, rendering it susceptible to many different proteinases (Visse and Nagase, 2003). Among the collagenases MT1-MMP is unique in that it is a membrane-tethered collagenase (Holmbeck et al., 2004).

The importance of MT1-MMP as a collagenase is particularly highlighted by the phenotype of MT1-MMP gene knockout (KO) mice (Holmbeck et al., 1999; Zhou et al., 2000). The neonatal KO mice are indistinguishable from heterozygous and wild-type littermates, but at 5 days of age growth impairment becomes apparent (Holmbeck et al., 1999; Zhou et al., 2000). KO mice develop severe fibrosis in periskeletal soft tissue, and show delayed ossification of bone, and die around 7–12 weeks of age (Holmbeck et al., 1999; Zhou et al., 2000). Fibroblasts derived from KO mice lack the ability to degrade type I collagen (Holmbeck et al., 1999). Detailed analyses suggest that the bone phenotype of the KO mice may in part result from the lack of cellular collagen degradation by osteocytes (Holmbeck et al., 2005). This suggests that osteocytes were not able to create the right ECM environment for osteocytogenesis. No other MMP gene KO mice have shown such drastic phenotypes and many events including long bone growth, soft tissue organization, molar root formation and eruption, and cartilage remodeling in the skull are found to be MT1-MMP-dependent (Holmbeck et al., 2004). These findings strongly suggest that MT1-MMP is an essential cellular collagenase important for organizing the ECM microenvironment, which cannot be substituted for by any other MMP during development.

MT1-MMP is an important collagenase during migration of cells in a collagen matrix. It has been shown that only MT1-MMP can promote cellular invasiveness into collagen I matrix in epithelial cells and fibroblasts (Hotary et al., 2000; Sabeh et al., 2004). Other soluble collagenase MMPs, such as MMP-1, MMP-2, MMP-8, and MMP-13, cannot promote invasion even though they are secreted as an active form under the experimental conditions used (Hotary et al., 2000; Sabeh et al., 2004). For MT1-MMP to promote collagen invasion, the enzyme needs to be membrane-bound as a soluble form of MT1-MMP lacking the transmembrane, and cytoplasmic domains could not promote invasion (Hotary et al., 2000; Sabeh et al., 2004). Thus, MT1-MMP as a membrane-tethered collagenase plays a unique role.

Interestingly, pericellular collagen degradation is essential not only for invasion, but also for tumor cells to grow within a collagen-based 3-D matrix (Hotary et al., 2003). Again, only MT1-MMP as a collagen-degrading enzyme is found to play a critical role (Hotary et al., 2003). The effect of MT1-MMP is specific to 3-D cultures, suggesting that it is acting through pericellular degradation of collagen matrix. Soluble collagenase MMPs (MMP-1, MMP-2, and MMP-13) cannot substitute for MT1-MMP, and without MT1-MMP tumor cells cannot proliferate (Hotary et al., 2003). Though MT1-MMP was originally thought to be involved in only cancer cell invasion for metastasis, these findings suggest that MT1-MMP may also play a role in promoting tumor progression.

MT1-MMP stimulates cell motility

Initially, the enhanced invasiveness by MT1-MMP expression was thought to be due to enhanced pericellular ECM degradation allowing cells to migrate through. However, it has become evident that “clearing a path” is not the only mechanism. It was found that shedding of CD44 from the cell surface by MT1-MMP enhances cell migration on a hyaluronan-based 2-D matrix (Kajita et al., 2001). CD44 is a widely expressed major hyaluronan receptor, which interacts with the cytoskeleton though its cytoplasmic domain (Naot et al., 1997). The exact mechanism behind the MT1-MMP/CD44-mediated promotion of cell migration is not clear. MT1-MMP may regulate the adhesion properties of lamellipodia through CD44 shedding or it may modulate signals mediated by CD44. A similar observation was made with syndecan-1 (Endo et al., 2003). MT1-MMP sheds syndecan-1 and this shedding is required for efficient cell migration on collagen. Again the underlying mechanism is not clear. Syndecan-1 is known to associate with and stabilize αvβ3 integrin, promoting cell spreading (Beauvais et al., 2004). It is possible that shedding of syndecan-1 associated with αvβ3 affect the integrin signaling. Expression of MT1-MMP is reported to activate extracellular signal-regulated protein kinase (ERK) and ERK activation is shown to be essential for MT1-MMP-dependent cell migration (Gingras et al., 2001; Takino et al., 2004). It is not clear how MT1-MMP activates ERK, but this may be a part of the mechanism(s) of CD44- and/or syndecan 1-shedding mediated cell migration. Elucidating this mechanism may be one of the keys to understand how MT1-MMP promotes cell migration.

Another observation of MT1-MMP-mediated stimulation of cell motility is through the cleavage of laminin 5 (Koshikawa et al., 2000, 2005), a major component of basement membrane. It has been shown that MT1-MMP cleaves the γ2 chain of laminin 5, releasing EGF-like domains, which ligate with the EGF receptor and stimulate epithelial cell motility (Schenk et al., 2003). The cleavage of the γ2 chain of laminin 5 by MT1-MMP may be tissue-specific as in keratinocytes astacin-like metalloproteinases, bone morphogenic protein-1 (BMP-1), and its variant forms (tolloid metalloproteinases), but not MT1-MMP, are reportedly responsible for γ2 chain cleavage (Veitch et al., 2003). However, it is not clear if this cleavage of laminin γ2 chain by BMP-1 affects keratinocyte migration. It has also been shown that the β3 chain of human laminin 5 is cleaved by MT1-MMP (Udayakumar et al., 2003). Interestingly the cleavage product is shown to stimulate prostate cancer cell motility (Udayakumar et al., 2003), but the mechanism remains to be determined.

MT1-MMP promotes angiogenesis

Angeogenesis is a formation of new vessels from existing vessel and provides a mechanism to establish a blood supply. It is playing an important role in wound healing process, tumor growth, and progression of rheumatoid arthritis (Sivakumar et al., 2004). Angiogenesis itself is a process of cellular invasion by endothelial cells. Endothelial cells need to detach from neighboring cells, invade into stromal tissue, proliferate, and generate a tube structure. During this process, endothelial cells need to degrade basal lamina, fibrin, and collagen enriched stromal tissue. Previous reports have shown that MMP-2 (Brooks et al., 1998), MMP-9 (Bergers et al., 2000), and their cognate cell surface receptors such as αvβ3 integrin (Brooks et al., 1998; Silletti et al., 2001) and CD44 (Yu and Stamenkovic, 1999) play a critical role in this process, but Chun et al. (2004) recently reported that plasmin, MMP-2, MMP-9, β3 integrin, and CD44 are not essential for angiogenesis, but MT1-MMP is. The tissues from the mice lacking these genes generated normal neovessels in type I collagen in an ex vivo model, but MT1-MMP KO tissue failed to show angiogenesis (Chun et al., 2004). This observation is specific for collagen matrix as in fibrin matrix MT1-MMP KO mice tissue showed normal neovessel formation. The fact that neovessel formation in fibrin gel is inhibited by TIMP-2 but not by TIMP-1, and the observation that gene expression of MT2-MMP and MT3-MMP were upregulated during tube formation of human umbilical vein endothelial cells in fibrin gel (Lafleur et al., 2002) suggests that these MT-MMPs may compensate for the role of MT1-MMP.

MT1-MMP also promotes angiogenesis by means other than stimulation of endothelial cell invasion. Expression of MT1-MMP in tumors stimulates angiogenesis in vivo by stimulating VEGF synthesis from the tumor cells (Deryugina et al., 2002b; Sounni et al., 2002). MT1-MMP specifically upregulates VEGF-A gene expression, and this is mediated by src tyrosine kinase (Sounni et al., 2004). This event requires proteolytic activity of MT1-MMP and is dependent on the cytoplasmic domain (Sounni et al., 2004). The mechanism behind this is not clear, but MT1-MMP activity on the cell surface presumably modifies the immediate microenvironment, which triggers a cell-signaling pathway to produce VEGF-A.

Taken together, MT1-MMP is an important pericellular microenvironment modifier which affects cell functions and promotes invasion in tissue (Fig. 1).

REGULATION OF MT1-MMP: CONTROLING CELL FUNCTIONS

Inhibition of MT1-MMP activity

Since MT1-MMP is expressed on the cell surface as an active form, inhibition is one of the critical steps to regulate its activity. MT1-MMP is inhibited by endogenous inhibitors TIMP-2, -3, and -4, but not by TIMP-1 (Will et al., 1996; Bigg et al., 2001). This TIMP-1 insensitive nature is common to all transmembrane type MT-MMPs (e.g., MT2-, MT3-, and MT5-MMPs) and allows them to work under conditions where high levels of TIMP-1 are present. Another inhibitor for MT1-MMP is RECK (reversion-inducing-cysteine rich protein with Kazal motifs), a GPI-anchored glycoprotein (Oh et al., 2001). RECK was originally discovered as a tumor invasion suppressor gene with dual functionality: suppressing MMP-9 expression and inhibiting its enzyme activity (Takahashi et al., 1998). Later, it was also found to inhibit the proteolytic activity of MT1-MMP and MMP-2 (Oh et al., 2001). Mice with a disrupted RECK gene are embryonic lethal at E10.5 showing defects in collagen fibrils, the basal lamina, and vascular development, showing a phenotype that may be the result of excess MMP activities. Interestingly this phenotype was partially suppressed by an MMP-2 null mutation (Oh et al., 2001). Some of the phenotype could be due to excess MT1-MMP or MMP-9, but RECK may also have other important biological role besides MMP inhibition. Chondroitin/heparan sulfate proteoglycans, Testican 1 and 3 and a splicing variant of Testican 3, N-Tes, have also been shown to inhibit MT1-MMP (Nakada et al., 2001), although the mechanisms of inhibition are not known.

Proteolytic processing

It has been shown that the 60 kDa active MT1-MMP undergoes further processing to a 44–45 kDa species by MMP-2 or MT1-MMP itself (Lehti et al., 1998; Stanton et al., 1998; Toth et al., 2002). This removes the catalytic domain of MT1-MMP making it inactive, and is a mechanism of downregulation. The proteolytic processing of MT1-MMP may be an indication of how active the enzyme is on the cell surface. A high level of 45 kDa form coincides with high proMMP-2 activation whereas no proMMP-2 activation occurs when 45 kDa form is not detected besides full length mature MT1-MMP on the cell surface (Lehti et al., 1998; Stanton et al., 1998). Thus, MT1-MMP can be “functionally active,” and as a result undergoes further processing, or “functionally inactive” remaining as an intact mature form. This suggests that there may be a step which regulates its functional activity after expression of the mature enzyme on the cell surface, and one of the possible mechanisms to achieve this could be a homo-oligomer complex formation of the enzyme (see below) (Osenkowski et al., 2004).

In some cells, the whole ectodomain of MT1-MMP was shown to be shed (Harayama et al., 1999; Toth et al., 2002, 2004). The cleavage occurs at Val524-Ile bond in the linker-2 region and releases functional MT1-MMP from the cell surface (Toth et al., 2004). Although shed MT1-MMP may act as a soluble proteinase in the extracellular milieu, biological activity of MT1-MMP on cell function would effectively be lost since the enzyme needs to be membrane-tethered to exert these effects (Cao et al., 1995; Hotary et al., 2000; Sabeh et al., 2004). Thus this too can be a part of downregulation mechanism, although it is not clear how frequently and widely this shedding event occurs.

Regulation of MMP-2 activation

As proMMP-2 activation is an important step for cancer cells to invade into basal lamina, the mechanism of activation has been extensively studied. This process is not a simple interaction of proMMP-2 and MT1-MMP, but involves its endogenous inhibitor TIMP-2 (Strongin et al., 1995). MT1-MMP expressed on the cell surface forms a complex with TIMP-2 through the catalytic domain of the enzyme and the N-terminal inhibitory domain of TIMP-2 as an enzyme-inhibitor complex. The exposed C-terminal domain of TIMP-2 has an affinity for the Hpx domain of proMMP-2, and this results in the formation of an MT1-MMP-TIMP-2-proMMP-2 ternary complex (Strongin et al., 1995). Formation of this complex is absolutely essential and the proMMP-2 activation does not occur without TIMP-2 (Butler et al., 1998; Kinoshita et al., 1998; Wang et al., 2000). Since MT1-MMP in this complex is inhibited by TIMP-2, another MT1-MMP free from TIMP-2, is required to carry out the activation of proMMP-2. To arrange another molecule of MT1-MMP next to the ternary complex of MT1-MMP-TIMP-2-proMMP-2, MT1-MMP forms a homo-oligomer complex through its Hpx domains and/or transmembrane/cytoplasmic domains (Itoh et al., 2001; Lehti et al., 2002). In this complex one of the MT1-MMP molecules acts as a receptor and the other acts as an activator, forming a proMMP-2 activation complex. This homo-oligomer complex formation is important in the activation process on the cell surface, because separating two MT1-MMPs by over expressing a catalytic domain deletion mutant of MT1-MMP effectively inhibits proMMP-2 activation (Itoh et al., 2001). Also replacement of the Hpx domain with one derived from another MMP, such as MT4-MMP (Itoh et al., 2001) or MMP-2 (Cao et al., 2004) significantly inhibits the activation, presumably because they do not form a homo-oligomer complex. Interestingly, the homo-oligomer complex formation can be upregulated by the expression of a constitutively active form of Rac1, a small GTPase which generates extensive lamellipodia (Itoh et al., 2001). Under these conditions, MT1-MMP exclusively localizes at the lamellipodia edge, and this is accompanied by enhanced proMMP-2 activation (Itoh et al., 2001). This suggests that homo-oligomer complex formation is a regulated process and may be one of the mechanisms that determines the “functional activity” of MT1-MMP. In test tubes, proMMP-2 can be activated by the recombinant catalytic domain of MT1-MMP (Will et al., 1996), a domain which does not form a homo-oligomer complex. It is plausible that the ability of homo-oligomer complex formation was adopted to facilitate the activation process on the cell surface, and the ancillary domains, including Hpx, transmembrane, and cytoplasmic domains, play a role in this event. It is also possible that other proteins, which interact with MT1-MMP mediate clustering the enzyme. There is a report showing that a Hpx domain-deleted mutant of MT1-MMP activates proMMP-2 (Wang et al., 2004a). This may be attributed to transmembrane/cytoplasmic domain-mediated homo-oligomer complex formation (Lehti et al., 2002). It may be important to evaluate the contribution of each domain to the homo-oligomer formation for proMMP-2 activation.

Cell surface localization of MT1-MMP

When cells migrate in tissue, degradation of ECM barrier is essential, but only in the direction of migration, because ECM is also important scaffolding. To achieve such focal degradation, cells localize their MT1-MMP at lamellipodia, the migration front of the cells (Sato et al., 1997; Itoh et al., 2001; Mori et al., 2002). This localization is achieved by interaction of MT1-MMP with CD44 through the Hpx domain of the enzyme and stem region of CD44 (Mori et al., 2002). CD44 in turn is associated with F-actin through its cytoplasmic domain by interacting with Ezrin/Radixin/Moesin proteins (Naot et al., 1997). This associates MT1-MMP indirectly with F-actin. Over expression of the cytoplasmic domain deletion mutant of CD44 (CD44ΔCP) abrogated MT1-MMP association with F-actin presumably by competing for the interaction of MT1-MMP with full-length CD44 (Mori et al., 2002). This completely abolishes localization of MT1-MMP to the lamellipodia (Mori et al., 2002), suggesting that association with F-actin is critical for localization of both CD44 and MT1-MMP to the cell front. The interaction of MT1-MMP with CD44 through the Hpx domain of the molecule is also required for CD44 shedding by the enzyme (Suenaga et al., 2005).

It has been also shown that MT1-MMP can be co-localized with β1 integrin (Ellerbroek et al., 2001; Wolf et al., 2003). It is not clear whether this is due to a direct/indirect interaction, but these reports suggest that MT1-MMP and β1 integrin are functioning in the same area on the cell surface.

Trafficking and intracellular regulation of MT1-MMP

MT1-MMP is shown to be internalized by clathrin-dependent and caveolae-dependent pathways (Jiang et al., 2001; Uekita et al., 2001; Remacle et al., 2003). Because internalization removes MT1-MMP from the cell surface, this can be a mechanism of downregulation. However, almost paradoxically, the internalization process appears to be essential for the enzyme to promote cell migration (Uekita et al., 2001). Clathrin-dependent internalization of MT1-MMP is mediated through the C-terminal cytoplasmic tail. The μ2 subunit of adaptor protein 2 (AP-2), which mediates incorporation of membrane proteins into clathrin cages, interacts with the LLY573 sequence in the cytoplasmic tail of MT1-MMP (Uekita et al., 2001). When LLY573 was mutated to alanines or deleted, the internalization of MT1-MMP is inhibited, and the mutant MT1-MMPs cannot promote cell migration (Uekita et al., 2001). MT1-MMP is also internalized through caveolae (Remacle et al., 2003). Structural or sequence requirements for this kind of internalization are not known, but it is generally thought that localization to the membrane lipid rafts is essential since caveolae are lipid rafts containing caveolin (Varma and Mayor, 1998). Galvez et al. (2004) have demonstrated that caveolae-mediated internalization of MT1-MMP also plays an important role in MT1-MMP-mediated endothelial cell migration on a collagen substratum.

A part of internalized MT1-MMP was shown to be transported to the CD63 positive lysosomes for degradation (Takino et al., 2003). Interestingly CD63 tetraspanin, a well established component of late endosomal and lysosomal membranes, interacts with MT1-MMP directly through the N-terminus of CD63 and the Hpx domain of MT1-MMP (Takino et al., 2003). Expression of CD63 accelerates internalization and lysosomal degradation of MT1-MMP, and it requires both the Hpx and the cytoplasmic domains of MT1-MMP (Takino et al., 2003). Therefore, CD63 is one of the regulators of MT1-MMP trafficking.

It has been shown that MT1-MMP is also recycled back to the cell surface after the internalization. Wang et al. (2004c) have shown that the C-terminus of the cytoplasmic domain, DKV583, is a critical sequence. Deletion of DKV583 inhibits recycling of MT1-MMP. A similar motif, EWV is present at the C-terminus of MT2-, MT3- and MT5-MMP, and MT3-MMP is shown to be able to co-recycle with MT1-MMP in MDCK cells (Wang et al., 2004c). Mint-3 has been identified as a protein that interacts with the EWV motif of MT5-MMP (Wang et al., 2004b). Mint-3 upregulates surface levels of MT5-MMP, presumably by increasing the recycling frequency. It would be interesting to know if Mint-3 can also regulate trafficking of MT1-MMP.

Another protein which interacts with the cytoplasmic domain of MT1-MMP has been named “MT1-MMP cytoplasmic-binding protein-1” (MTCBP-1) (Uekita et al., 2004). MTCBP-1 does not affect proMMP-2 activation by MT1-MMP but suppresses MT1-MMP-mediated cell migration. Interestingly many cancer cell lines, which express high levels of MT1-MMP express low level of MTCBP-1 and normal fibroblasts which express low levels of MT1-MMP express a high levels of MTCBP-1 (Uekita et al., 2004). This inverse correlation suggests that MTCBP-1 may act as an invasion suppressor.

CONCLUSIONS

Since the discovery of MT1-MMP in 1994 (Sato et al., 1994), the enzyme has been studied extensively and become one of the best characterized MMPs. Particularly significant progress has been made over the last decade in cellular regulation of the enzyme (Fig. 2). MT1-MMP is an essential pericellular collagenolytic enzyme during development, cell invasion, tumor growth, and angiogenesis. The enzyme also modifies cell signaling pathways either by shedding cell surface molecules or by producing functional fragments from ECM components, and still more mechanisms are likely to be revealed. Thus MT1-MMP is indeed a microenvironment modifier and also a cell function modifier. One of the striking facts is that presence of active form of MT1-MMP on the cell surface does not necessary mean that it is functionally active (Itoh and Seiki, 2004). For instance, localization of MT1-MMP at the migration front of the cells and its internalization from the cell surface are essential for the enzyme to promote cell migration. When these requirements are not met, the enzyme looses its activity to promote cell migration even though it is still present on the cell surface as proteolytically active form (Itoh and Seiki, 2004). Such tight control of enzyme function is presumably necessary for this enzyme because MT1-MMP has evolved to be a crucial cell function modifier. Cell migration is a dynamic and a complex process that involves many different molecules, which need to act coordinately to facilitate cell migration. We believe that further understanding of the critical regulatory mechanisms of MT1-MMP-mediated cell migration will give insight into developing new ways to treat diseases what MT1-MMP-mediates such as cancer and rheumatoid arthritis.

Figure 2.

Cellular regulation of MT1-MMP. MT1-MMP is regulated in a number of ways including: transcription, trafficking to cell surface, homophilic complex formation, association with CD44 followed by localization to migration front, autodegradation, and inhibition by inhibitors such as TIMPs, N-Tes, and RECK. The enzyme is also internalized by caveolae-dependent and clathrin-dependent pathways. The internalized enzyme can be recycled back to the cell surface or degraded in CD63-positive lysosomes. MT1-MMP at the migration front can colocalize with β1 integrin, and MTCBP-1 can bind to the cytoplasmic domain of MT1-MMP, suppressing its pro migratory effects.

Acknowledgements

We thank Rob Visse and Hideaki Nagase for critical reading of this manuscript.

Ancillary