Cellular and Molecular Life Sciences

, Volume 68, Issue 23, pp 3853–3868

Matrix metalloproteinases in tumorigenesis: an evolving paradigm


  • Hui Hua
    • State Key Laboratory of Biotherapy, Section of Signal Transduction and Molecular Targeted Therapy, West China HospitalSichuan University
  • Minjing Li
    • State Key Laboratory of Biotherapy, Section of Signal Transduction and Molecular Targeted Therapy, West China HospitalSichuan University
  • Ting Luo
    • Cancer Center, West China HospitalSichuan University
  • Yancun Yin
    • State Key Laboratory of Biotherapy, Section of Signal Transduction and Molecular Targeted Therapy, West China HospitalSichuan University
    • State Key Laboratory of Biotherapy, Section of Signal Transduction and Molecular Targeted Therapy, West China HospitalSichuan University

DOI: 10.1007/s00018-011-0763-x

Cite this article as:
Hua, H., Li, M., Luo, T. et al. Cell. Mol. Life Sci. (2011) 68: 3853. doi:10.1007/s00018-011-0763-x


Proteases are crucial for development, tissue remodeling, and tumorigenesis. Matrix metalloproteinases (MMPs) family, in particular, consists of more than 20 members with unique substrates and diverse function. The expression and activity of MMPs in a variety of human cancers have been intensively studied. MMPs have well-recognized roles in the late stage of tumor progression, invasion, and metastasis. However, increasing evidence demonstrates that MMPs are involved earlier in tumorigenesis, e.g., in malignant transformation, angiogenesis, and tumor growth both at the primary and metastatic sites. Recent studies also suggest that MMPs play complex roles in tumor progression. While most MMPs promote tumor progression, some of them may protect the host against tumorigenesis in a context-dependent manner. MMPs have been chosen as promising targets for cancer therapy on the basis of their aberrant up-regulation in malignant tumors and their ability to promote cancer metastasis. Although preclinical studies testing the efficacy of MMP suppression in tumor models were so encouraging, the results of clinical trials in cancer patients have been rather disappointing. Here, we review the complex roles of MMPs and their endogenous inhibitors such as tissue inhibitors of metalloproteinase in tumorigenesis and strategies in suppressing MMPs.


Matrix metalloproteinaseCancer


Tumor development and progression are complex processes that involve oncogenes, tumor suppressor genes, and the tumor microenvironment. Stromal–epithelial interaction plays important roles in the pathogenesis of a variety of tumors [1]. The cross-talk between malignant and surrounding stromal cells (fibroblasts, endothelial cells, and inflammatory cells) may establish fertile soil in which tumor cells grow and metastasize. In addition to the stromal cells within tumors, the extracellular matrix (ECM) also has an important impact on tumor progression [2]. The major components of the ECM include structural proteins such as collagen and elastin; specialized proteins such as fibrillin, fibronectin, and laminin; and proteoglycans. There are diverse types of proteases that control ECM remodeling and drive the dissemination of cancer cells into adjacent tissue. Among them, serine proteases such as urokinase plaminogen activator (uPA)-plasminogen-plasmin system and cysteine cathepsins are involved in the breakdown of multiple ECM proteins during various physiopathological situations, including cell adhesion, invasion, angiogenesis, and metastasis [3]. In addition, matrix metalloproteinases (MMPs) are intensively studied and have been demonstrated to play key roles in inflammation and carcinogenesis [4]. Also, MMPs are involved in cell signaling and are capable of activating specific cell receptors and growth factors or liberating them from the ECM, thereby regulating various cell behaviors, such as cell growth, differentiation, apoptosis, and migration [4, 5].

In light of the critical roles for MMPs in inflammation and tumorigenesis, MMPs appear to be ideal drug targets [5, 6]. The development of MMP inhibitors has been very attractive within the academy and industry. Many pharmaceutical inhibitors have been developed thus far. Several of them have undergone clinical trials in patients with cancer. Although many of them have shown promising effects in preclinical studies, the results from clinical trials with these agents appear to be disappointing in cancer treatment. These outcomes also stimulate intensive studies on the roles of MMPs in different stages of cancer. It has been recognized that MMPs play complex, and often opposing roles in cancer progression, which raises question marks on how to effectively target MMPs for cancer therapy. In this review, we emphasize on the complex roles of MMPs in tumor progression and recent studies on MMP inhibitors in cancer therapy.

MMPs and the endogenous inhibitors: an overview

MMPs are the principal matrix-degrading proteinases that play key roles in embryonic development, organ morphogenesis, and wound healing [5]. Abnormal activation of MMPs has been implicated in numerous diseases including arthritis, atherosclerosis, and tumorigenesis. The MMPs are endopeptidases that can cleave many components of the ECM. So far, more than 20 MMPs have been identified. Since almost every MMP has an alias that may be confusing, the names of individual MMPs and their aliases are listed in Table 1. Typically, the MMPs contribute to ECM turnover in either secreted or membrane-bound forms. Based on the structure and substrate specificity, MMPs can be classified into several groups, namely collagenases, gelatinases, stromelysins, membrane-type MMPs, and non-classified MMPs (Table 1). In general, the structure of the MMPs includes a signal peptide, a propeptide domain, a catalytic domain, and zinc-binding motif. Except MMP-7 and MMP-26, other MMPs also contain hemopexin-like domain. The most extensively studied MMP-2 and MMP-9 share a gelatin-binding domain [7]. MMP-14 (MT1-MMP), -15 (MT2-MMP), -16 (MT3-MMP), and MMP-24 (MT5-MMP) harbor a transmembrane domain [8]. MMP-17 (MT4-MMP), and MMP-25 (MT6-MMP) are a unique set of membrane-anchored MMPs with a glycosyl-phosphatidyl inositol (GPI) anchor [9]. The diversity of structure endows individual MMP with different distribution and substrate specificity.
Table 1

Members of the MMP family





Interstitial collagenase; fibroblast collagenase



72-kDa type IV collagenase; gelatinase A; 72-kDa gelatinase






Matrilysin-1; PUMP1



Neutrophil collagenase; PMNL collagenase



92-kDa type IV collagenase; gelatinase B; 92-kDa gelatinase



Stromelysin-2; transin-2






Macrophage proteinase; macrophage elastase; metalloelastase



Collagenase 3




Membrane-type MMP



Membrane-type MMP



Membrane-type MMP



Membrane-type MMP


MMP-18; Matrix metalloproteinase RASI; RASI-1; stromelysin-4



Enamel metalloproteinase; enamelysin













Membrane-type MMP



Membrane-type MMP


Matrilysin-2; endometase








MMPs not only serve as effectors in physiological processes such as embryo implantation and development, branching morphogenesis, bone remodeling, and angiogenesis but also have additional roles in the reorganization of tissues during pathological conditions such as inflammation, wound healing, and invasion of cancer cells. MMPs are frequently overexpressed in a variety of human tumors [1012]. Ets sites have a critical role in MMP transcription and they often cooperate with AP1 in multiple MMP promoters. Another major cytokine-induced signaling pathway that regulates MMP transcription involves translocation of nuclear factor-κB (NF-κB) family members from the cytoplasm to the nucleus [13]. After being synthesized into proteins, most MMPs require proteolytic cleavage for enzymatic activity after reaching the cell surface or secreting into the extracellular space. A few MMPs, however, are activated intracellularly by a furin-like mechanism [14]. The activities of MMPs can be counteracted by endogenous inhibitors. α2-Macroglobulin is an abundant plasma protein that acts as a main MMPs inhibitor in tissue fluids. α2-Macroglobulin does not inhibit the activation of MMPs, nor does it inhibit MMPs directly. Instead, active MMPs are efficiently captured by α2-macroglobulin in tissue fluids, followed by low-density lipoprotein-receptor-related protein (LRP)-mediated endocytosis and clearance. Interestingly, LRP itself is a target of MT-MMP proteolysis [15]. Thrombospondin (TSP)-2 regulates the clearance of extracellular MMP-2 in a similar way to α2-macroglobulin. Lysosomal degradation of MMP2–TSP2 complexes by LRP is an important mechanism for the regulation of extracellular MMP-2 levels [16]. In contrast, TSP1 binds to pro-MMP2 and pro-MMP9 and directly inhibits their activation [17, 18].

Whereas both α2-macroglobulin and TSP can regulate a broad spectrum of proteases, tissue inhibitors of metalloproteinases (TIMP) are more specific endogenous MMP inhibitors. So far, four TIMPs have been identified, designated as TIMP-1, TIMP-2, TIMP-3, and TIMP-4, respectively. The four TIMPs share common evolutionary structure in their N-terminal regions [17]. Twelve cysteine residues and their relative spacing are highly conserved in all the TIMPs, which form six disulfide bonds that are essential for the native conformations and the MMP-inhibitory activities. The C-terminal regions of the TIMP family are divergent, which may dictate the selectivity of TIMPs to MMPs or confer MMP-independent function. TIMPs inhibit MMPs by forming a strong noncovalent complex with a 1:1 stoichiometry. Although various TIMPs can bind to different MMPs, there is some difference in the inhibitory properties among the TIMPs. TIMP-1 prefers to inhibit MMP-9, stromelysin, and collagenase. Pro-MMP2 prefers to bind to TIMP-2. TIMP-4 is similar to TIMP-2 in that it is a more potent inhibitor of MMP2 than of MMP9 and matrilysin [18]. The type I transmembrane MT-MMPs are relatively well inhibited by TIMP-2, TIMP-3, and TIMP-4, but are poorly inhibited by TIMP-1. In contrast, the GPI-anchored MT-MMPs are inhibited by both TIMP-1 and TIMP-2 [19].

Paradoxically, TIMP-2 and TIMP-3 are involved in the cell surface activation of proMMP-2. TIMP-2 can bind to proMMP-2 via C-terminal interaction, and to MT1-MMP by its N-terminal domain. This dual binding brings proMMP-2 close to the cell surface, where it can be activated by neighboring TIMP-2-free MT1-MMP molecules [2023]. It has been reported that the entire propeptide domain of MT1-MMP is required for the TIMP-2 binding and subsequent proMMP-2 activation [24]. Fully functional TIMP-2 is essential for efficient activation of proMMP-2 both in vitro and in vivo [25]. However, the active 65-kDa MMP-2 can be inhibited by plasma membrane-bound TIMP-2 [26]. These results suggest that the pericellular activity of MMP-2 is tightly regulated by membrane-bound TIMP-2 and surrounding ECM components. In addition, TIMP-3 can bind to proMMP-2 with high affinity and promote MT3-MMP mediated activation of proMMP-2 [27], but does not promote MT1-MMP-mediated activation of proMMP-2 [21]. pro-MMP-2 activation by MT3-MMP on the cell surface involves a ternary complex with proMMP-2 assembled with the catalytic domain of MT3-MMP and TIMP-2 or TIMP-3 [27]. In this case, TIMP-2 or TIMP-3 may act as a scaffold to tether proMMP-2 and MT3-MMP. Although TIMP-4 associates with MMP-2 and MT1-MMP in a manner similar to TIMP-3, it does not support MT1-MMP-mediated activation of proMMP-2 [28], nor does it affect MT3-MMP-mediated activation of pro-MMP-2 [27], suggesting that individual TIMP has a different role in the regulation of pro-MMP-2 processing. MT1-MMP also promotes the activation of pro-collagenase 3 (MMP-13) [29], a potent collagenolytic protease. It remains unknown whether TIMPs regulate the activation of proMMP-13.

Complex roles of MMPs in tumorigenesis

Breakdown of extracellular components

The ECM remodeling is an active event during tumor progression. On one hand, ECM serves as a niche for tumor cells to survive and proliferate. On the other hand, it is a barrier that suppresses the spreading of tumor cells. Degradation of ECM is one of the first steps in tumor invasion and metastasis. MMPs promote cell invasion and motility by pericellular ECM degradation. The expression and activity of MMP-2 and MMP-9, two intensively studied gelatinases, are frequently elevated in human cancer, which correlates with advanced tumor stage, increased metastasis, and poor prognosis. MMP-2 and MMP-9 participate in the degradation of ECM components including the basement membrane, which separates epithelia from stroma [30]. In addition, MMP-1 digests type III collagen more efficiently than collagen types I and II. MMP-3 is known to degrade collagen types III, IV, IX and X, proteoglycans, laminin, elastin, and fibronectin [31]. MT1-MMP, a key MMP that regulates invasion and metastasis, plays a dual role in pathophysiological digestion of the ECM through activation of proMMP-2 and direct cleavage of substrates such as collagen types I, II, and III. Regulated positioning of MT1-MMP to invadopodia, the specialized ECM-degrading membrane protrusions of invasive cells, enables focal degradation of ECM during invasion and metastasis [32].

Release of bioactive molecules

Degradation of structural and specialized components of the ECM by MMPs not only breaks the barrier that restrains tumor cell dissemination but also generates some bioactive fragments. Upon MMP digestion, the ECM also releases biologically active fragments called matrikines, i.e., peptides originating from the fragmentation of matrix proteins and presenting biological activities [33]. For example, cleavage of laminin-5 γ2 chains by MMP-2 and MT1-MMP produces a fragment containing epidermal growth factor (EGF)-like motifs that engages EGFR signaling and larger fragments that engage integrin signaling, leading to cell migration [34, 35]. Cleavage of osteopontin by MMP-9 generates a 5-kDa fragment that promotes tumor cell invasion [36]. Moreover, cleavage of insulin-like growth factor-binding protein (IGFBP) by MMP-7 and MMP-9 leads to release of the bioavailable IGF and activation of IGFR signaling [3739]. Thus, MMPs degrade multiple components of the ECM thereby generating bioactive molecules that promote tumor progression.

Regulation of urokinase plasminogen activator (uPA), growth factor, and cytokine signaling

The cross-talk between MMPs and other proteases also contributes to tumor progression. MMP-9 can regulate the activity of other proteases such as uPA. A recent study reveals that MMP-9 degrades the serpin protease nexin-1, an inhibitor of uPA and the invasion of tumor cells [40]. However, the effects of nexin-1 on tumor metastasis seem to be controversial. Nexin-1 reportedly binds LRP-1 and stimulates extracellular signal-regulated kinase signaling, MMP-9 expression, and metastatic spread of mammary tumors [41]. The inconsistency of these studies is hard to explain. Given that MMP-9 can be upregulated by nexin-1, the degradation of nexin-1 by MMP-9 may represent a negative feedback regulation of nexin-1 activity.

Except for components of the ECM, there are non-ECM substrates for MMPs that include growth factors, kinase, cytokines, chemokines, and receptors. A number of growth factors can stimulate the secretion of MMPs. Conversely, metalloproteinases may release matrix-sequestered growth factors, thereby creating a vicious cycle of autocrine growth. MT1-MMP cleaves HB-EGF and removes the NH(2)-terminal 20 amino acids that are important for binding heparin. The truncated form of HB-EGF is independent of heparin and exhibits enhanced mitogenic activity [42]. Moreover, MT1-MMP degrades the Wnt/planar cell polarity protein-tyrosine kinase-7 (PTK7), an inhibitor of cell invasion. The cleavage of PTK7 by MT1-MMP leads to an increase in cell invasion and migration [43]. Tumor-associated MT1-MMP sheds RANKL and activates src-dependent prostate cancer migration and bone metastasis [44]. Thus, MT1-MMP may play pivotal roles in both the growth and metastasis of tumor cells. Moreover, MMP-13 regulates mammary tumor-induced osteolysis by activating MMP-9 and TGF-β signaling at the tumor–bone interface [45], a common feature of a variety of cancer.

Regulation of angiogenesis

MMPs also have complex roles in angiogenesis. It is known that MMPs can promote endothelial cell migration and trigger angiogenic switch. For example, MMP-9 participates in switching angiogenesis by releasing VEGF from ECM [46]. Also, MMP-2 activity was suggested to be necessary for the switch to angiogenic phenotype in an animal model [47]. Furthermore, MMPs increase the bioavailability of the pro-angiogenic growth factors vascular endothelial growth factor (VEGF), fibroblast growth factor-2 (FGF-2), and TGF-β, which stimulate proliferation and migration of endothelial cells. MT1-MMP regulates VEGF-A expression by promoting VEGFR-2 cell surface localization thereby activating VEGFR-2-Src-Akt-mTOR pathway [48]. MMP-7 degrades soluble VEGF receptor-1 (sVEGFR-1/sFlt-1), an endogenous VEGF inhibitor that sequesters VEGF and blocks its access to VEGF receptors. The degradation of sVEGFR-1 then liberates VEGF from the endogenous trap and allows its access to membrane receptors on endothelial cells [49]. However, MMPs may have adverse effects on angiogenesis. For example, MT1-MMP-mediated endoglin shedding inhibits tumor angiogenesis [50]. The generation of angiostatin and endostatin, two potent angiogenesis inhibitors, also involves MMPs. MMP-7 and MMP-9 hydrolyze plasminogen to generate angiostatin fragments [51]. Endostatin is the C-terminal proteolytic product of the collagen XVIII α1 chain. MMP-7 and MMP-14 can cleave collagen XVIII to generate endostatin-spanning fragment [52, 53]. Hence, we speculate that the effects of MMPs on angiogenesis may be context-dependent. The outcome relies on the availability of specific MMPs and the balance between two opposing effects of individual MMPs on angiogenesis.

Regulation of inflammation

Another notable function for MMPs is their modulatory roles in inflammation. Inflammatory response is a hallmark of cancer. Inflammatory cells, such as macrophages, neutrophils, and mast cells, are components of tumor microenvironment. The tumor-infiltrating inflammatory cells can release cytokines, polypeptide growth factors, and proteases that may regulate stromal remodeling as well as tumor cell invasion. Studies in transgenic mice demonstrate that mast cells, macrophages, and peripheral mononuclear cells amplify neoplastic cell proliferation and angiogenesis largely by release of MMP-9 [54]. The expression of MMPs is regulated by a variety of cytokines and chemokines. Elevated expression and activities of MMPs are observed in acute or chronic inflammatory diseases, such as rheumatoid arthritis and atherosclerosis. Meanwhile, MMPs themselves also manifest as regulators of inflammation. The processing of chemokines may be regulated by MMPs, generating either inactivated fragments or truncated form of chemokines with increased activity (Table 2) [5566]. Therefore, MMPs may exhibit either pro-inflammatory activities or anti-inflammatory activities in a context-dependent manner. For example, MMP-9 potentiates the activation of pro-inflammatory cytokines such as TNF, interleukin-1β (IL1β), IL6, IL8, and chemokines such as CXCL5, CXCL6, and LIX. However, MMP-2 may dampen inflammation by processing monocyte chemoattractant proteins thereby generating chemokine receptor antagonists with anti-inflammatory properties [62]. Rheumatoid arthritis is characterized by the autoimmune inflammation of the joints and degradation of the joint cartilage. MMPs are not only essential for the migration of inflammatory cells to the damaged joints, but also responsible for the cleavage and degradation of structural components of cartilage [67]. In addition, MMPs play key roles in vascular pathologies. MMP-2 and MMP-9 contribute to the proliferation and migration of vascular smooth muscle cells that is responsible for restenosis [68].
Table 2

The processing of chemokines by MMP


Processing MMP




MMP-1, MMP-3




MMP-1, -2, -3, -13, -14
















MMP-8, -9




MMP-8, -9




MMP-8, -9


[60, 61]


MMP-8, -9




MMP-1, -2, -3, -9, -13, -14








MMP-4, -8, -9, -13

Increased activity




Increased activity



MMP-1, -2, -8, -9, -13

Increased activity


Induction of epithelial–mesenchymal transition

Epithelial–mesenchymal transition (EMT) is a key developmental process characterized by loss of cell adhesion, repression of E-cadherin expression, and increased cell mobility [42]. EMT may be essential for numerous developmental processes including mesoderm formation and neural tube formation. During tumor progression, EMT is often activated to promote cancer cell invasion and metastasis. MMPs are important regulators or mediators of EMT. MMP-2 is necessary for the EMT that generates neural crest cells and plays an essential role in producing epithelial–mesenchymal transformations in the avian embryo [69]. In highly invasive human squamous carcinoma cells, TGF-β-induced EMT leads to the upregulation of MMP-2 through a snail and Ets-1-dependent mechanism [70]. However, overexpression of Ets-1 induced the expression of MMP-2 without EMT, which indicates that MMP-2 alone may be not enough for inducing EMT in squamous carcinoma cells [71]. Nevertheless, EMT can be induced by MMP-3 in mammary epithelial cells [72], and this effect may be antagonized by synthetic MMP inhibitor [62]. MMP-3-induced EMT involves activation of Rac1b and induction of Snail, a transcription factor that plays key role in initiating EMT [73]. In this case, MMP-3 induces the expression of Rac1b, which in turn increases the levels of reactive oxygen species, critical stimuli of snail expression and EMT. MMP-3 is also a mediator of wnt1-induced EMT [74]. Moreover, MMP-9 can cleave E-cadherin and promote EMT [65]. The expression of MMP-9 can be stimulated by snail through a MAPK- and PI3K-dependent mechanism [75]. TGF-β is a well-known inducer of EMT. MMP-28 upregulates MT1-MMP and MMP-9, which promote the processing of latent TGF-β complex thereby increasing levels of active TGF-β. Therefore, MMP-28 induces EMT and cell invasion through a TGF-β-dependent mechanism [67]. Recently, it has been demonstrated that MT1-MMP and MT2-MMP are key mediators of snail-driven EMT, tumor angiogenesis, and metastasis [76]. These findings suggest that MMP play a role in cell migration during the EMT and morphogenesis.

Regulation of cell survival

MMPs are also emerging as regulators or mediators of cell survival. Different MMPs may have contrasting effects on cell survival. Even for the same MMP, the effects on cell survival may be cell-type-specific or context-dependent. Upon induction of apoptosis, MMP-1 co-localizes with aggregated mitochondria and accumulates around fragmented nuclei, and inhibits apoptosis [77]. MMP-2 silencing causes cancer cell apoptosis by upregulating Fas/Fas-L and FADD [78]. In contrast, MMP-2 promotes endothelial cell migration but reduces cell viability under hypoxia [79]. MMP-3 stimulates ductal proliferation and branching during puberty [72] but induces apoptosis in anchorage-dependent secretory epithelium during pregnancy [80, 81]. Meanwhile, MMP-3 induces spontaneous mammary carcinogenesis [62]. Upon endoplasmic reticulum stress, MMP-3 is increased and promotes neuronal apoptosis downstream of caspase-12 but upstream of caspase-3 [82]. MMP-7 would cleave FASL and this inhibits or induces apoptosis in a context-dependent manner [8385]. The pro-apoptotic effect of MMP-7 on pancreatic acinar cells is mediated largely by FasL and is responsible for ductal metaplasia [79]. MMP-7 also causes a decrease of FAS at the cell surface, which leads to oxaliplatin-resistance in colon cancer cells [86]. Although MMP-7 cleaves FASL and induces apoptosis, constitutive expression of matrilysin may select for cells with reduced sensitivity to Fas-mediated apoptosis and renders cancer cells more invasive [87]. MMP-9 promotes chronic lymphocytic leukemia B cell survival through its hemopexin domain but not catalytic domain, indicating an enzyme activity-independent action of MMP-9 [82]. In addition to the well-recognized extracellular function, intracellular MMP-9 promotes cell cycle progression and cell growth [88]. However, MMP-9 promotes retinal ganglion cell death after optic nerve ligation [89]. MT1-MMP downregulates the glucose-6-phosphate transporter expression and promotes bone marrow stromal cell death [90]. Active stromelysin-3 (MMP-11), but not the inactive MMP-11, increases MCF-7 survival in three-dimensional Matrigel culture via activation of p42/p44 MAP-kinase and Akt [91]. MMP-15 confers resistance of cancer cells to apoptosis induced by Fas, tumor necrosis factor-related apoptosis-inducing ligand, and serum starvation [92]. Finally, elevation in MMP activity plays a proapoptotic role in β-adrenergic receptor-stimulated apoptosis of adult rat ventricular myocytes [88].

Tumor suppressive roles for MMPs

Although many facets of MMPs action are pro-tumorigenic, some MMPs may also inhibit tumorigenesis. Collagenase-2 (MMP-8) is an MMP mainly produced by neutrophils and associated with many inflammatory conditions. MMP-8 reportedly plays a protective role in cancer through its ability to regulate the inflammatory response induced by carcinogens. Experimental manipulation of the expression levels of this enzyme alters the metastatic behavior of human breast cancer cells [93]. Thus, low levels of MMP-8 expression in tumors may be associated with a poor outcome in patients with cancer. In addition, the tumor suppressive effect of MMP-12 is observed. Overexpression of MMP-12 in colon cancer cells is associated with increased survival [94]. In fact, the complex roles of MMPs in tumor progression remain to be explored. As discussed earlier, several MMPs contribute to the generation of angiogenic inhibitors. MMP-2, MMP-7, MMP-9, and MMP-12 can convert plasminogen into angiostatin [51, 9598]. The generation of another potent antiangiogenic factor, endostatin, may also require MMPs and elastase [99]. Another notable finding is that silencing of MMP-1 in chondrosarcoma cells inhibits distant metastasis but promotes primary tumor growth and angiogenesis [100]. This evidence suggests that temporal or spatial regulation of MMPs may be essential for angiogenesis.

Biological activities of endogenous TIMPs in cancer

The overexpression and elevated activity of MMPs correlate with tumor progression. The balance between activated MMPs and their inhibitors such as TIMPs may define the net activity of MMPs. Down-regulation of TIMPs may result in an increase in the activity of MMPs and the invasive potential of tumor cells. Conversely, tumor invasion and metastasis can be inhibited by up-regulation of TIMPs in tumor cells. Overexpression of TIMP-1 inhibits tumor growth and metastasis of melanoma [101], suppresses human gastric cancer metastasis [102], and prevents oral squamous cell carcinoma progression [99]. Adenoviral transfer of TIMP-3 into HeLa, fibrosarcoma, and melanoma cells inhibits the invasiveness and induces apoptosis [103, 104]. Ectopic expression of TIMP-4 in human breast cancer cells inhibits invasion, metastasis, and tumor growth [102]. In addition to the involvement of MMPs in the late stages of tumorigenesis, there is evidence supporting a role for MMPs at the early stage of carcinogenesis and angiogenesis [46, 51, 62, 105, 106]. Therefore, one can speculate that TIMPs may exert tumor suppressive activities at both the early and late stages of tumorigenesis.

Although accumulating data demonstrate that TIMPs have tumor-suppressive activities, studies also show that TIMPs may play contrasting roles in tumor progression. Indeed, the stimulatory effect on cell growth was initially recognized when TIMP-1 and TIMP-2 were found to be of erythroid-potentiating activities [107, 108]. TIMP-1 and TIMP-2 are also mitogenic for non-erythroid cells, including normal keratinocytes [109], fibroblasts, lung adenocarcinoma cells, and melanoma cells [108, 110]. In addition, TIMPs exert diverse effects on apoptosis. Whereas TIMP-3 can induce apoptosis, TIMP-1, -2, -4 can protect cells from apoptosis. TIMP-1 inhibits apoptosis of B cells and human breast epithelial cells in vitro [111, 112], and rescues mammary epithelial cell apoptosis in transgenic mice [79]. TIMP-2 overexpression protects B16F10 melanoma cells from apoptosis [111]. TIMP-4 also protects human breast cancer cells from apoptosis [113]. The effects of TIMPs on cellular proliferation and apoptosis may be mediated by MMP-independent mechanisms. TIMP-1 and TIMP-2 have been shown to stimulate tyrosine kinase, focal adhesion kinase, and mitogen-activated protein kinase activity in a human osteosarcoma cell line MG-63 and in human breast cancer cell lines [113, 114]. TIMP-1, which can be upregulated by Bcl-2, seems to function downstream of Bcl-2 [112]. TIMP-1 also upregulates Bcl-XL expression in B cells [110]. Although overexpression of TIMP-1 in cancer cells inhibited metastasis, recent studies suggested that TIMP-1 may also promote tumorigenesis and metastasis in some animal models [114116]. It is believed that TIMPs regulate cell growth and apoptosis through paracrine mechanism. Efforts have been placed on the identification of cell surface binding partners for TIMP family members. For the first time, the integrin α3β1 was identified as a cell surface receptor for TIMP-2 [117]. TIMP-2 has been shown to inhibit endothelial cell growth in response to FGF-2, which can be attributed to TIMP-2 binding to endothelial cell surface α3β1 thereby inhibiting FGF2-induced ERK1/2 signaling [118]. TIMP-2 also inhibits VEGF-induced VEGFR-2 phosphorylation and enhances phosphodiesterase activity in endothelial cells [119]. TIMP-3 may bind to VEGFR2 and competitively inhibit VEGF binding and subsequent kinase signaling [120]. In addition, TIMP-1 and TIMP-4 can bind to the tetraspanin CD63 and inhibit both the intrinsic and extrinsic cell death [121, 122].

Strategies for inhibition of MMPs

Synthetic MMP inhibitors

Soon after tumor biologists discovered that MMPs were involved in the invasion and metastasis of cancer cells, intensive efforts have been taken to develop MMP inhibitors (MMPIs) to halt the spreading of cancer cells [5]. Most small-molecule MMPIs are hydroxamate-based in which the hydroxamate acts as a zinc-binding group. Batimastat (BB-94), BB-1101, and marimastat (BB-2516) are the best-known examples of succinyl hydroxamate MMPIs that display efficacy in preclinical studies of cancer therapy [123127]. Batimastat is a potent MMPI with a broad spectrum of substrates and inhibits tumor growth and metastasis in a variety of tumor models [128, 129]. In addition, batimastat potentiated the growth-suppressive effects of cisplatin and demonstrated synergistic anti-tumor effects with docetaxel and captopril [128, 129]. However, there are some unwanted effects that may hinder the effectiveness of batimastat in cancer therapy. Treatment with batimastat may induce stromal uPA expression that can halt the anti-invasive activity [130]. Moreover, batimastat treatment promotes T-cell lymphoma liver metastasis by inducing liver-specific overexpression of MMP-2, -9, and angiogenesis factors [131]. These contradicting outcomes may be attributed to the redundancy of MMP family, temporal, and spatial determinants of MMP action, and compensatory response to MMP inhibition.

In contrast to batimastat, marimastat is orally active, with an absolute bioavailability of 20–50% in preclinical studies [132]. Inhibition of cervical lymph node metastasis by marimastat (BB-2516) was observed in an orthotopic oral squamous cell carcinoma implantation model [133]. Marimastat also inhibited the peritoneal spread of gastric carcinoma in nude mice [134]. These promising preclinical studies led to the initiation of a series of clinical trials for marimastat in treating advanced tumors. However, a phase III trial involving 505 patients with stage III NSCLC randomized to receive either marimastat or placebo failed to show a significant difference in progression-free survival or overall survival between the two groups [135]. Later on, clinical trials involving cancer patients randomized to receive either chemotherapy or combination of marimastat and chemotherapy were initiated. Once again, marimastat failed in clinical trials as part of a combination therapy for pancreatic, gastric, or ovarian carcinoma, and glioma [136]. Bramhall and colleagues conducted a randomized clinical trial involving 414 patients with advanced pancreatic cancer to receive either gemcitabine or marimastat [137]. Marimastat showed inferior efficacy to gecitabine in treating pancreatic cancer. Musculoskeletal toxicity was reported in 39–55% of patients across different dose levels of marimastat. Combination of mariamastat and gemcitabine did not show superior efficacy compared to gemcitabine alone in treating advanced pancreatic cancer [138].

At the very beginning, most hydroxamate-based MMPIs are broad-spectrum inhibitors that lack specificity. Some researchers argue that successful anticancer inhibitor may possess MMP selectivity against the MMP subtype. Later on, new compounds were developed that took advantage of alternative zinc-binding groups, such as carboxylic acids, thiols, and phosphorous-based groups. Prinomastat (AG3340) is an oral active, non-peptide inhibitor of MMPs, including MMP-2, -9, MT1-MMP, and collagenase III [139]. Prinomastat demonstrated anti-tumor activities in a variety of preclinical studies upon oral administration, including cancer of pancreas, breast, lung, melanoma, and glioma models [140143]. The anti-tumor efficacy of prinomastat in chemoresistant human non-small cell lung cancer was also observed either when prinomastat was used as a single agent or when it was combined with chemotherapy. Unfortunately, this compound failed in phase III clinical trials as a combination with taxol and caboplatin for advanced non-small lung cancer and with mitoxantrone for hormone-resistant prostate cancer, due to the drug’s lack of effectiveness in patients with late-stage disease [144, 145]. Patient safety was not a factor in this case. It remains to be known whether such drugs may be more effective in earlier stages of disease, when tumors are smaller and may be more sensitive to these agents. Prinomastat was designed to be highly selective for several specific matrix metalloproteinase enzymes. In addition, the carboxylate BAY12-9566 is a relatively selective inhibitor of gelatinases and stromelysin-1. The clinical trial of Bay 12,9566 was stopped due to more rapid tumor growth in patients treated with the drug. BAY12-9566 proved inferior to gemcitabine in a randomized trial of 277 patients with advanced pancreatic cancer [146]. It is possible that selective inhibition of several MMPs may not be sufficient enough to cause significant inhibition of tumor growth. Furthermore, suppressing several enzymes may also lead to increased production by cancer cells of other enzymes that may compensate the loss of those enzymes, leading to continued tumor growth.

Cis-2-aminocyclohexylcarbamoylphosphonic acid (cis-ACCP) is another selective MMP inhibitor that has been evaluated in vitro and in vivo cancer metastasis models. It significantly inhibits metastasis formation in mice [147]. In addition, S-3304 is a potent, orally active, noncytotoxic inhibitor of MMPs, primarily MMP-2 and MMP-9. S-3304 prolongs survival in mice xenografts and is well tolerated in healthy volunteers [148]. Ro-28-2653, a selective and orally bioavailable MMPI with inhibitory activity against MMPs expressed by tumor and/or stromal cells, is a potent antitumor and antiangiogenic agent [149]. However, it is unknown whether these agents would be effective in a clinical setting.

Pyrimidine dicarboxamides are a class of highly selective MMP-13 inhibitors. These inhibitors do not interact with the catalytic zinc but bind the S1 pocket and extend into an additional S1 side pocket, which is unique to MMP-13 [150]. Compared to the catalytic domain, the binding sites outside the active domain of the MMPs are less conserved and are related to substrate selection [151]. Therefore, targeting the binding sites outside of the catalytic domain represents an alternative approach to achieve MMP- and substrate-specific interference.

The mechanism-based inhibition of selective MMP is an alternative strategy. These inhibitors bind to the active sites of specific MMP and initiate slow binding profile for the onset of inhibition, leading to covalent enzyme modification [152]. In this regard, the specific inhibition of single MMP can be achieved. SB-3CT is a specific gelatinase inhibitor based on this mechanism. The key event in the inhibition of MMP-2 by SB-3CT is enzyme-catalyzed active site ring-opening of the thiirane moiety, giving a stable zinc-thiolate species [153]. It inhibited tumor growth, liver metastasis, and improved survival in an aggressive mouse model of T-cell lymphoma [154]. Also, it reduced prostate cancer growth, osteolysis, and angiogenesis in a bone metastasis model [155].

Antibody-based MMP inhibition

Antibody-based therapy is expanding in controlling diseases including cancer. It remains to be known whether an antibody to a specific MMP will be an effective anti-cancer agent. DX-2400, a highly selective fully human MMP-14 inhibitory antibody discovered using phage display technology, blocked proMMP-2 processing in tumor and stromal cells, inhibited angiogenesis, and slowed tumor progression and metastasis in vivo [156]. The combination of potency, selectivity, and robust in vivo activity shows the potential of a selective MMP-14 inhibitor for the treatment of solid tumors. Certainly, more inhibitory antibodies to individual MMP remain to be developed. The efficacy of antibody-based MMP inhibition, either as a monotherapy or in combination with chemotherapy, in treating human tumors needs to be tested.

Inhibitors of MMPs transcription

Another strategy of inhibiting MMPs is to suppress the expression of MMPs. Chemically modified tetracyclines (CMTs) without antibiotic activities can bind metal ions such as calcium and zinc, and to affect MMP gene transcription [157]. A high-throughput screening system identifies a small molecule 5-methyl-2-(4-methylphenyl)-1H-benzimidazole (MPBD) that can repress MMP-9 expression by antagonizing AP-1 transactivation activity [158]. Consistent with this effect, MPBD inhibits MMP-9-dependent invasion of oral cancer cells, preosteoclast migration, and RANKL-induced osteoclast activity over concentration ranges that repressed MMP-9 expression. In addition, the EGFR inhibitor Gefitinib inhibits MMP-9 and MMP-2 secretion and mRNA expression in HT29 cells [159]. Given that the transcription of MMP is regulated by NF-κB, inhibiting NF-κB activity may reduce the expression of MMP thereby suppressing the intrinsic migration ability and invasive potential of tumor cells [160]. Moreover, another transcription factor, Ets-1, can regulate the transcription of MMP-2. The COX-2 product prostaglandin E2 promotes pancreatic cancer metastasis through an ERK/Ets-1-dependent induction of MMP-2 expression. Rofecoxib, a COX-2 inhibitor, suppressed Ets-1 binding activity and MMP-2 expression, and cellular migratory and invasive potentials [161].

In addition to synthetic inhibitors of MMP transcription, there are natural compounds capable of inhibiting MMP expression. Cannabinoids are natural compounds that down-regulate MMP-2 expression, inhibit tumor angiogenesis, and halt tumor cells spreading [162]. Curcumin, a polyphenol derived from the plant Curcuma longa, inhibited the PMA-induced mRNA expression of MMP-1, -3, -9, and -14 in human astroglioma cells by repressing the DNA-binding and transcriptional activities of AP-1 and the PMA-induced MAP kinase activities, which were involved in modulating the expression of MMPs [163]. These natural compounds may hold promise in cancer chemoprevention.

Stimulators of endogenous MMP inhibitors

Finally, an alternative strategy for inhibiting MMPs is to up-regulate endogenous MMP inhibitors. miRNA is emerging as another regulator of gene expression. A recent study demonstrated that miR-21 could inhibit the expression of RECK and TIMP-3, two tumor suppressors and MMP inhibitors [164]. miR-21 is significantly elevated in glioblastoma and in many other tumors of various origins [165, 166]. This microRNA has been implicated in various aspects of carcinogenesis, including cellular proliferation, apoptosis, and migration. Therefore, targeting this microRNA may increase the expression of RECK and TIMP3 thereby inhibiting MMPs. In addition, 1-alpha-25-dihydroxyvitamin D3 (1, 25-VD) can inhibit the expression of MMP-9 and cathepsins, while it increases the expression of their counterparts, tissue inhibitors of metalloproteinase-1 (TIMP-1) and cathepsin inhibitors [167].

Concluding remarks

ECM homeostasis is critical for tissue architecture that regulates development and cancer. Among diverse proteases, MMPs play important roles in regulating ECM homeostasis. The best-known function of MMPs is to degrade the components of ECM. During tumor progression, the expression and activation of MMPs is upregulated in the tumor infiltrative frontier. Traditionally, MMPs are known as one of the key promotors of tumor metastasis. However, mounting evidence demonstrates that MMPs also play complex roles in malignant transformation and tumor onset (Fig. 1). In addition to ECM, MMPs can target non-matrix proteins such as growth factors and their receptors, chemokines, adhesive molecules. MMPs not only regulate the invasive potential of cancer cells but also regulate cellular proliferation, apoptosis, EMT, angiogenesis, and inflammatory response. Notably, these processes may be temporally and spatially regulated by MMPs. The individual MMP in such a large family may share some common features but also possesses unique substrates or function.
Fig. 1

Complex roles for MMP in cancer. The multifunctional MMPs can not only degrade structural components within the extracellular matrix but also process growth factors and receptors (EGF, TGF-β, sFlt-1, etc.), cytokines and chemokines, and apoptosis-related molecules. Moreover, MMPs are involved in epithelial-mesenchymal transition. Therefore, MMPs have effects on cell growth and survival, angiogenesis, inflammation, invasion, and metastasis. These effects may be context-dependent. While MMPs can stimulate angiogenesis through release of angiogenic factors such as VEGF and bFGF, they are also involved in generating angiogenesis inhibitors such as angiostatin and endostatin

MMPs have been considered as promising targets for cancer therapy on the basis of their aberrant up-regulation in malignant tumors and their ability to degrade components of the ECM. Multiple pharmacological strategies can be taken to target MMPs (Fig. 2). Indeed, preclinical studies testing the efficacy of MMP suppression in tumor models were encouraging. Based on the promising results from preclinical studies on synthetic MMPs inhibitors, the clinical trials on MMPs inhibitors in cancer therapy were conducted. Unfortunately, the results of these trials in cancer patients have been rather disappointing. Recent studies demonstrate that MMPs may have paradoxical roles in tumor progression, with either tumor-promoting effects or tumor suppressive effects in a context-dependent manner. Increase in our knowledge of the molecular mechanisms underlying MMPs action in tumor progression may help understand the failure in previous clinical trials for MMP inhibitors. Meanwhile, these new insights would stimulate the development of novel approaches to target MMPs. Regardless of the clinical outcome, the potential efficacy of MMP inhibitors in cancer therapeutics must be based on the success in inhibiting MMPs. In this respect, imaging activity of specific MMPs in vivo with probes may help evaluate the efficacy of MMP inhibition in certain tumors [168]. Given that members of the MMP family have different roles during cancer progression, the use of broad-spectrum MMP inhibitors may result in unsatisfying clinical outcomes due to the inhibition of tumor suppressive MMPs. Although novel strategies to develop inhibitors of individual MMP are emerging, it is even more difficult to discern what MMPs to target and when to target. Taking into consideration of the redundancy of MMPs, it remains unknown if these specific MMP inhibitors would be successful anticancer agents. In addition, it is possible that inhibition of individual MMP will paradoxically up-regulates other MMPs, leading to a compensatory response that hampers the anticancer effects.
Fig. 2

Physiological and pharmacological regulation of MMP. The inactive pro-MMPs are synthesized from MMPs transcripts. The transcription of MMPs is regulated by transcription factors including Ets1, AP1, and NF-κB. The mechanisms underlying the activation of pro-MMPs are complex. Both MT1-MMP and MT3-MMP are involved in the activation of pro-MMPs. In addition, MMPs may be activated by plasmin and furin-like mechanism. Tissue inhibitors of metalloproteinase (TIMP) are endogenous MMP inhibitors that prevent the activation of pro-MMPs. TSP1 also inhibits the activation of MMPs. Paradoxically, TIMP-2 and TIMP-3 are involved in the activation of pro-MMP as well. The turnover of active MMPs may involve LRP-mediated endocytosis and clearance. α2-Macroglobulin and TSP-2 act as endogenous MMP inhibitors through promoting LRP-mediated endocytosis and clearance of active MMPs. The pharmacological inhibitors of MMPs transcription include 5-methyl-2-(4-methylphenyl)-1H-benzimidazole (MPBD), Gefitinib, NF-κB inhibitors, COX-2 inhibitors, and curcumin. In addition to hydroxamate-based MMP inhibitors (MMPIs), there are exosite MMPIs and mechanism-based MMPIs. Moreover, neutralizing antibodies to individual MMP are emerging as another choice for selective inhibition of MMP

Since many clinical trials were conducted in patients with late-stage tumors, it remains to be known whether MMP inhibitors may have benefits in treating early stage tumors. In addition, adverse effects, including musculoskeletal syndrome, limit the maximal-tolerated dose of some early generation of MMP inhibitors such as marimastat and thereby dampen drug efficacy. These side-effects may be the result of broad-spectrum inhibition of the MMP family. Therefore, toxicity effects must be minimalized to ensure that the drug efficacy would not be limited by the maximum-tolerated dose. Moreover, biomarkers may add substantial value to clinical practice by providing an integrated approach to prediction or monitoring drug response using the genetic makeup of the tumor and the genotype of the patient to guide personalized treatment selection. The genetic background of the host might be an issue that determines the efficacy of MMP inhibitors as anticancer agents. A study in a mouse model of breast cancer demonstrated that the anti-metastatic outcome of MMP-9 inhibition was strain-specific [169]. Further work on the identification of specific genetic factors that affect the efficacy of MMP inhibitors are warranted. These studies may be helpful for the use of tumor genotyping to guide the choice of MMP inhibitors. Other predictive biomarkers for MMP inhibitors need to be addressed in future studies. It already takes decades to learn what roles MMPs have in tumor progression. There is still a long way to go to uncover the complex functions of MMPs in tumorigenesis. Certainly, the development of MMP inhibitors and subsequent pharmacological improvements would benefit from knowledge of MMPs.


The work in the authors’ lab was supported by the National Natural Science Foundation of China (Grants 30900554 and 30973435).

Copyright information

© Springer Basel AG 2011