Cell Biology: Research & TherapyISSN: 2324-9293

All submissions of the EM system will be redirected to Online Manuscript Submission System. Authors are requested to submit articles directly to Online Manuscript Submission System of respective journal.

Special Issue Article, Cell Biol Res Ther S Vol: 0 Issue: 0

Transforming Growth Factor-Beta and Matrix Metalloproteinases Functional Interplay in Cancer; Implications in Epithelial to Mesenchymal Transition

Jelena Krstic and Juan F. Santibanez*
Laboratory for Experimental Hematology and Stem Cells, Institute for Medical Research, University of Belgrade, Serbia
Corresponding author : Juan F. Santibanez
Laboratory for Experimental Hematology and Stem Cells, Institute for Medical Research, University of Belgrade, Dr Subotića 4, 11129 Belgrade, Serbia
Tel: +381-11-2685-788; Fax: +381-11-2643-691
E-mail: [email protected]
Received: May 20, 2013 Accepted: January 04, 2014 Published: January 08, 2014
Citation: Krstic J, Santibanez JF(2014) Transforming Growth Factor-Beta and Matrix Metalloproteinases Functional Interplay in Cancer; Implications in Epithelial to Mesenchymal Transition. Cell Biol: Res Ther S1:004 doi:10.4172/2324-9293.S1-004


Transforming Growth Factor-Beta and Matrix Metalloproteinases Functional Interplay in Cancer; Implications in Epithelial to Mesenchymal Transition

Transforming growth factor beta (TGF-β), a pleiotropic factor with several different roles in health and disease, has been shown to act differently during tumorigenesis, depending on the stage of tumor progression. In the early stages, TGF-β acts as a tumor suppressing factor, while in advanced stages of tumor it acts as a tumor promoting factor. One of the hallmarks of tumor progression is the capacity of cancer cells to migrate and invade surrounding tissues with subsequent metastasis to different organs. The matrix metalloproteinases (MMPs) family comprises more than 25 members whose main function is the proteolytic degradation of the extracellular matrix. Recent studies have demonstrated that MMPs’ roles in extracellular signaling can be involved in both homeostasis and disease. Due to their importance, MMPs are tightly regulated at transcriptional level, and can also be regulated through activation and inhibition.

Keywords: Tumorigenesis; TGF-β; Matrix metalloproteinases


Tumorigenesis; TGF-β; Matrix metalloproteinases


Transforming growth factor beta (TGF-β) acts as a pleiotropic factor, with different roles in health and disease [1]. This factor has been postulated to have a dual role in tumor progression, since it represses epithelial tumor development in early stages, whereas it stimulates tumor progression in advanced stages [2]. During the tumorigenesis, cancer cells acquire the capacity to migrate and invade surrounding tissues and to metastasize to different organs. Matrix metalloproteinases (MMPs) are a group of mainly extracellular matrix (ECM) proteolytic enzymes which enable cells to migrate and invade the surrounding tissue. Due to their importance, many MMPs are tightly transcriptionally regulated during normal development, but are deregulated in cancer, when their activity and expression are related to its development. TGF-β regulates MMPs expression in cancer cells, while MMPs may activate the latent secreted TGF-β, this way producing a harmful cycle which contributes to the tumor progression. Importantly, both TGF-β and MMPs may induce epithelial to mesenchymal transition processes (EMT) in cancer cells. Tumors might exploit the interactions between TGF-β and MMPs to induce EMT and to increase tumor progression and metastasis. In this paper, the specific roles and the interplay between TGF-β and MMPs in cancer cells, as well as their implications in EMT are reviewed.

Transforming Growth Factor Beta

TGF-β belongs to a large family of structurally related regulatory proteins which comprises more than 40 members expressed in mammals, which also include activins, inhibins, bone morphogenetic proteins (BMP) and growth and differentiation factors. TGF-β has been involved in a plethora of distinct biological processes, including cell growth, differentiation and development, as well as tumorigenesis. Among the TGF-β, mammals express three genetically distinct isoforms (TGF-β1, -2 and -3) with high homology. The corresponding human genes are located on chromosomes 19q13, 1q41 and 14q24, respectively [1,3].
TGF-β initiate signaling by binding to their cell-surface serine/ threonine kinase receptors type I and II (TBRI and TBRII), which form heteromeric complexes in the presence of dimerized ligands. Binding of TGF-β to TBRII leads to the phosphorylation of TBRI, thus activating its kinase domain [4]. After the receptor complex is activated by the ligand, it phosphorylates and stimulates the cytoplasmatic mediators, Smad2 and Smad3 [5] (Figure 1). The phosphorylation of Smad2, 3 releases them from the inner membrane side, where they are specifically retained by Smad anchor for receptor activation (SARA), and they form a heterotrimeric complex with the common Smad4. The activated complex is further translocated to the nucleus where, in collaboration with other transcription factors, it binds and regulates the promoters of different target genes [5]. TGF-β signaling is regulated by the expression of other components of Smads, the inhibitory Smad proteins (Smad6 and Smad7 or I-Smads). TGF-β signaling regulates I-Smads transcription, establishing a negative feedback loop. Mainly, Smad7 antagonizes TGF-β by interacting with TBRI and leading to its degradation. Smad6 preferentially inhibits BMP signaling by disrupting the Smad1–Co-Smad interaction and forming an inactive Smad1–Smad6 complex [1,6].
Figure 1: TGF-β signaling.
TGF-β signaling comprises two groups of intracellular transduction pathways: Smad signaling pathway and Non-Smad signaling pathway. When TGF-β is activated, it binds to its cell surface type II receptor (TBRII) inducing the activation of TGF-β type I receptor (ALK5 or TBRI) and forms a heterotetrameric complex.
1) Smad signaling pathway: Active ALK5 in the complex phosphorylates Smad2, 3 which in turn promotes the release of Smads from the complex with SARA from the inner face of the plasma membrane. Phosphorylated Smads interact with co-Smad4, forming a heteromeric complex to be translocated into the cell nucleus, where it modulates gene expression by interacting with other transcription factors, co-repressors and co-activators.
2) Non-Smad signaling pathways: Active TGF-β-receptor complex interacts with ubiquitin ligase tumor necrosis factor receptor - associated factor 6 (TRAF6) which in turn recruits TGF-β activated kinase 1 (TAK1) to activate p38, JNK and NFkB pathways. In addition, TGF-β binding provokes the phosphorylation of ALK5 at tyrosine residues which enable the formation of Shc-Grb2/SoS complex to activate Ras-Raf1-MEK1, 2-ERK1,2 signaling. On the other hand, receptor activated complexes can activate PI3K, provoking the activation of AKT and the Small Rho GTPases. The activation of Non-Smad signaling pathways in turn initiate transcriptional or nontranscriptional activity to regulate cellular responses.
The TGF-β/receptor/Smads cascade is subjected to posttranslational modification which finely regulates TGF-β signaling, including processes such as phosphorylation/dephosphorylation, sumoylation and ubiquitination which reversibly regulate their stability and availability. Another level of regulation is the internalization and recycling of the ligand–receptor complexes via either lipid rafts/caveolae or clathrin-coated vesicles, which can modulate signaling as well as protein degradation in the proteasome [1,7].
In addition to the canonical Smad2,3 pathway, TGF-β activates several other intracellular signaling pathways, commonly referred to as non-Smad pathways (Figure 2), which include: mitogen-activated protein kinases (MAPKs): ERK1,2, JNK and p38; phosphoinositide 3-kinase (PI3K); AKT1,2 and mTOR which have a role as cell survival mediators; nuclear factor κB (NF-κB); Cyclooxygenase-2 and prostaglandins; the small GTPase proteins: Ras and Rho family (Rho, Rac1 and Cdc42), among others [8,9]. The capacity of TGF-β1 to activate different signal transduction pathways partly explains its pleiotropic capacity to regulate many functions at molecular, biochemical and biological level.
Figure 2: TGF-β processing and MMP-dependent activation.
1) TGF-β is synthesized as a precursor protein. The signal peptide (SP), attached to the TGF-β precursor protein, is cleaved during its transit through the rough endoplasmic reticulum (RER); a protein homodimer is formed and is then cleaved by furin convertase to produce the small latent complex (SLC). In SLC, mature TGF-β remains non-covalently bound to the latencyassociated peptide (LAP), which keeps it in an inactive form. Next, the SLC complex covalently binds to the latent TGF-β binding protein (LTBP), producing the large latent complex (LLC); finally LLC is secreted and stored in the extracellular matrix (ECM) before subsequent activation. 2) Membrane bound MMPs or soluble MMPs can directly degrade ECM and/or may promote the activation of latent TGF-β by proteolytic cleavage within the N-terminal region of the LAP. In turn, active TGF-β, by binding to its cell surface receptors ALK/TBRI-TBRII, triggers the activation of intracellular signal transduction to exert its cellular effects.

The Role of TGF-Β in Cancer

In the early steps of epithelial carcinogenesis, TGF-β operates as a tumor suppressor factor, due to its anti-proliferative and pro-apoptotic roles which counter the effects of local mitogenic stimulation in the injured or stressed epithelium. Conversely, in advanced stages, TGF-β operates as a tumor promoter, as cancer cells become refractory to its growth inhibitory effects by different mechanisms, including modifications in the components of TGF-β signaling, such as inactivating mutations in TBRII and Smad4, and other not fully elucidated alterations [10,11]. Carcinoma cells may utilize the capacity of TGF-β to modulate different events which can be directed at the stimulation of tumor progression and metastasis, as well as at the evasion of the immune system [2].
Cancer cells can produce and secrete elevated levels of TGF-β, which acts as the most potent immunosuppressive cytokine, with specific influence on the local tumor “non-transformed” cells, as well as on distant cells in the host, this way suppressing anti-tumor immune responses and creating an environment of immune tolerance, thus allowing metastatic cancer cells to escape from immune surveillance [12,13], one of the most important defense mechanisms against cancer progression. In addition, TGF-β stimulates monocyte and macrophage chemotaxis to the tumor microenvironment, leading to enhanced tumor invasion, angiogenesis, and metastasis, and to diminished antigen presentation and immunosurveillance towards developing neoplasm [14]. In breast, prostate, pancreatic and renal cancer elevated levels of TGF-β in plasma have been associated with advanced cancer stage, metastases, and poor clinical outcome [15-18]. It is believed that active TGF-β, produced by the tumor and local stroma, contributes to the progression and metastatic potential of the cancer through autocrine and paracrine effects [13]. Elevated serum levels of TGF-β have been observed in myeloma patients, with both malignant cells and bone marrow stromal cells being the source of TGF-β [19]. TGF-β levels also are elevated in non–Hodgkin’s lymphoma and are markedly elevated in high-grade lymphomas, cutaneous T cell lymphomas with a T-regulatory phenotype, as well as in splenic marginal zone lymphomas presenting as myelofibrosis [13]. Several hereditary cancer syndromes with mutations in TGF-β superfamily members are known. The autosomal dominant familial juvenile polyposis syndrome (JPS) is the most common of the hamartomatous syndromes which occurs with an incidence of about one per 100.000 births [20]. Patients present multiple hamartomatous polyps in the gastrointestinal tract, predominantly in the colon, at a young age, and predispose individuals to gastrointestinal tract cancers, including colorectal, gastric, small intestinal and pancreatic cancer, with approximately 70% lifetime risk for colorectal cancer [3]. Germline mutations in different members of the TGF-β superfamily have been described in JPS patients. In 20-25% of cases BMPR1A is mutated, the majority in the kinase domain; 15-20% have Smad4 mutations, predominantly in MH2 domain; in addition, mutations in endoglin gene have been established, but the incidence in unknown [21,22].
The autosomal dominant disorder hereditary nonpolyposis colorectal cancer (HNPCC) is the most common hereditary predisposition for the development of colorectal cancer. HNPC is caused by germline mutations involving DNA mismatch repair system genes, and contributes to microsatellite instability [23]. TBRII gene contains a 10-base pair polyadenine repeat microsatellite sequence, and up to 80% of colon cancer patients with HNPCC presents mutated form of TBRII [24].
The specific response to TGF-β during tumor progression will depend on the stage of carcinogenesis and the responsiveness of the tumor cells, and can be attributed to both independent and interrelated factors including changes in: (1) TGF-βs expression; (2) receptor expression; (3) availability of downstream signaling components; (4) evasion of the immune response; (5) stimulation of inflammation; (6) presence of local and systemic factors (autocrine, endocrine, paracrine, juxtacrine or matricrine interactions); and (7) the recruitment of cell types that lead to an advantage in tumor growth or promote angiogenesis [10].
The importance of the TGF-β signaling pathway in human cancers is evident from the frequent alteration of TGF-β signaling components in hereditary human cancers and sporadic cancers [3]. Several tumors express high levels of the TGF-βs, which correlate with tumor progression and clinical prognosis. As we will further analyze, TGF-β signaling promotes epithelial to mesenchymal transition, a characteristic of invasive and metastatic cells, with constitutive activation of TGF-β or TBRI/ALK5 leading to increased metastases in animal models [11].

Matrix Metalloproteinases

Matrix metalloproteinases, a group of over 25 zinc-dependent endopeptidase enzymes, share a similar structure to each other. The first well known function is the degradation of almost every component of the extracellular matrix which facilitates tissue remodeling, normal and pathological cell migration, and allows malignant cells to invade and move through extracellular matrix barrier. Based on their domain, structure and substrate preference, MMPs are traditionally grouped into: 1) collagenases, including MMP-1/-8/-13; 2) gelatinases, MMP-2 and -9; 3) membranetype MMPs (MT-MMPs); 4) stromelysins, MMP-3, 10 and -11; 5) matrilysins, MMP-7 and -26 and others (Table 1) [25]. All MMPs are synthesized as inactive zymogens or pro-MMPs and, with the exception of the membrane bound MT-MMPs, are secreted into the extracellular environment. All MMPs share at least three conserved domains, an N-terminal signal peptide, which is cleaved during transport through the secretory pathway, followed by the pro-domain of about 80 amino acids and a catalytic domain. Once secreted, pro- MMPs can be activated by the cleavage of the pro-peptide [25]. The proteolytic activity of MMPs is mainly regulated by tissue inhibitors of MMPs (TIMPs). There are 4 different TIMPs (TIMP1, -2, -3, and -4) [26]. TIMPs can inhibit all active MMPs, however, not with the same efficacy. Interestingly, MMPs also have been postulated to have a protective role in cancer. The tumor suppressing role of MMPs may derive from their ability to produce and release natural angiogenic inhibitors, such as angiostatin, endostatin and tumstatin, as a result of degrading extracellular components as plasminogen, collagen XVIII and collagen IV, respectively [27]. In breast and oral cancer patients, MMP8 expression may be a good prognostic marker [28]. In addition to MMP9, other MMPs such as MMP3 (stromelysin 1), MMP11 (stromelysin 3) and MMP19 have been found to play dual roles in cancer and exert pro-tumorigenic or protective roles, depending on the context [27,28].
Table 1: Matrix metalloproteinase subgroups.

MMPs Activate TGF-Β in the Extracellular Matrix Compartment

Several MMPs interact with TGF-β to form a bidirectional regulatory loop associated with cancer (Figure 2). TGF-β needs to be proteolytically activated by MMPs in order to exert its cellular functions, and an important biological activity of TGF-β in tumors is the remodeling of the ECM by regulating the expression of MMPs and their tissue inhibitors TIMPs [29].
TGF-β is synthesized as a homodimeric 75kd precursor protein consisting of three domains: the signal peptide, latency associated peptide (LAP) and the mature TGF-β. Furin–type convertase intracellularly cleaves the precursor to form an inactive complex, where the 25kd dimeric TGF-β remains non-covalently attached to LAP, forming the small latent complex (SLC) [30]. Further on, SLC can bind to the latent TGF-β binding protein (LTBP), creating the large latent complex (LLC), which is then secreted and may remain covalently associated to the ECM.
Activation of TGF-β involves the proteolytic cleavage of LAP, and TGF-β can be activated by thrombospondin-1, plasmin, acidic microenvironment and MMPs, such as MMP2 and 9, and β6 integrin [30]. Soluble or cell-surface MMP9 bound to CD44, MMP2, MMP13 and MMP14 proteolytically activate TGF-β1 [31,32]. Alternatively, MMP2, MMP9 and MMP14 can modulate TGF-β bioactivity by cleaving the LTBP complex, solubilizing ECM-bound TGF-β and increasing its bioavailability [33,34]. Since tumor cells are often resistant to the TGF-β suppressive tumor effects, this suggests that proteolytic activation of TGF-β by MMPs has a tumor-promoting effect by selectively driving stroma-mediated invasion and metastasis of the tumor [29].

Regulation of MMPs Expression by TGF-Β

The expression of MMPs is tightly regulated at the transcriptional level. Although MMP promoters are not fully characterized at the moment, the known promoters revealed several cis-elements which may either activate or repress MMPs gene expression [35,36]. MMP promoters harbor a variety of trans-activators, such as AP-1, PEA3, Sp-1, b-catenin/Tcf-4-lef-1, RARE and NF-κB. Based on the basic promoter composition, three categories of classification for MMPs have been proposed [35] (Table 2 and Figure 3): (1) “TATA and AP-1” group, containing TATA boxes at around −30 bp with AP- 1 sites around −70 bp, which includes MMP1, MMP3, MMP7, MMP9, MMP10, MMP12, MMP13, MMP19 and MMP26; (2) “TATA no AP-1” group, MMP promoters which contain a TATA box, while lacking a proximal AP-1 site, such as MMP-8, -11, and –21; (3) and “no TATA no AP-1” group of promoters, including MMP-2, -14, and -28, which neither harbor TATA boxes nor proximal AP-1 site, therefore, transcription from these promoters starts at multiple sites. In addition, the expression of MMPs in this last group is mainly regulated by SP1 transcription factors. Hence, the expression of these MMPs is partly constitutive, with low modulation by growth factors [37]. According to bioinformatic analysis, MMP20 has been included to group 1, MMP15 and MMP27 to group 2 and MMP16, MMP17, MMP23, MMP24 and MMP25 to group 3 [36].
Figure 3: Examples of MMPs promoter categorization, showing the main transcription factor binding sites. Adapted from [35,36].
Table 2: Categorization of MMPs based on their basic promoter composition.

Transcriptional Regulation of MMPs Expression by TGF-β

At least two different regulatory domains in gene promoters regulated by TGF-β have been described: the TGF-β inhibitory element (TIE) and Smad binding elements (SBEs) [5,38,39]. TIE was first characterized in the repressive role of TGF-β on the MMP3 promoter [40], and is represented by the consensus sequence 5’-GNNTTGGtGa-3’, in which the uppercase letters signify invariance and lowercase letters signify preferred sequence [40]. The SBEs contain a 5’-bp-GTCTG and its palindrome CAGAC within particular promoters, and are recognized by Smad3 through the N-terminal MAD homology domain (MH1) [1,39]. Interestingly, Smad4 recognizes the non-consensus GC-rich motifs. Conversely, a 30-amino-acid insertion at the Smad2’s MH1 domain disables their binding to gene promoters [1,5]. Since MMP1, MMP7, MMP9, MMP13 and MMP14 contain TIE binding sites in their promoters (Figure 3), it is possible that the expression of these MMPs may be modulated by TGF-β [41,42]. In MMP1, the TIE binding site is responsible for the inhibition of gene transcription. Mutation in this element results in a significant increase of basal MMP-1 gene transcription in fibroblasts suggesting that TIE may have a role as an MMP1 constitutive repressor in these cells, which is consistent with the capacity of TGF-β1 to repress MMP1 by Smad3 in dermal fibroblast [43,44]. Similarly, the MMP7 promoter has TGF-β inhibitory elements that negatively regulate its transcription [45]. MMP7 is required for tumor formation in Smad4 deficient intestinal adenocarcinomas, and the ectopic re-expression of Smad4 in colon cancer cells provokes a reduction of MMP7 [46,47]. A putative TIE domain has been described in the MMP9 promoter; however, molecular analysis revealed that this consensus domain was not required for TGF-β induction of MMP9 [48]. Although MMP13 possesses a TIE site in reverse orientation, this site does not specifically bind nuclear proteins from human chondrocytes and does not seem to be implicated in the TGF-β mediated regulation of MMP13 [49]. Similarly, in the MMP14 promoter three sequences with high homology to the TGF-β inhibitory element have been found, but at the moment the role of these domains is not elucidated [50].
Even though TGF-β may regulate MMPs expression by direct binding of Smads to the MMP promoters, the capacity of TGF-β to modulate MMPs is increasing by the interactions with other transcription factors, this way enormously increasing the number of possible interactions within the MMP promoters. In addition, Smads by interacting with the members of AP1 family can alter MMPs expression [51-53], i.e. TGF-β regulates MMP13 gene expression, partly via the AP1 site, partly by interactions of Smad3 with JunB and Runx-2 [54,55].
TGF-β may also directly activate transcription factors implicated in the regulation of MMPs expression. It induces cell signaling which culminates in the trans-activation of AP1, PEA3, NF-κB or SP1 transcription factors to enhance MMP promoters transactivity (Figure 4) [35,56]. AP1 motif appears to be critical in the MMP1 gene response to TGF-β. In dermal fibroblast AP1 complex containing JunB mediates TGF-β repression of MMP1, whereas in epidermal keratinocytes TGF-β induces MMP1 via c-Jun/AP1 complexes [57]. In breast MCF10A cells the binding of ATF2 in AP1 complex as well as SP1 is essential for MMP2 promoter activation by TGF-β [58]. Interestingly, Smad2 is also necessary for TGF-β-induced upregulation of MMP2 in mouse embryonic fibroblast cells [59].
Figure 4: TGF-β signaling integration in transcriptional regulation of MMPs expression.
Active TGF-β, by binding to the receptors, triggers the activation of intracellular signaling pathways, through which it activates different transcription factors or induces the transcription factor complexes to regulate MMPs expression, thus incrementing the protein levels in cancer cells.
In mammary epithelial cells, TGF-β transcriptionaly stimulates MMP10 expression by trans-activating the myocyte enhancer factor (MEF)-2, independent of Smad3. At a molecular level, MMP10 potentially has a central function in the MMP activation cascade because of its ability to cleave several pro-MMPs including MMP1, -7, -8, -9 and -13 [60].
The complexity of transcription factor networks implicated in TGF-β-induced MMPs is increasing. Besides the linear transactivation, interactions between transcription factors and crossregulation are also important for the understanding of the mechanism by which TGF-β regulates the expression of MMPs in cancer cells. This complexity is supported by numerous different transcription factor binding sites in the gene promoters of the MMPs, as shown in Figure 3. TGF-β activates several intracellular signal transduction pathways, converging in the transactivity of the different transcription factors necessary for the regulation of MMPs expression (Figure 4). Novel mechanisms of Smad-transcription factor interactions, as well as new transcription factors implicated in TGF-β-induced MMPs production will most likely be revealed in the near future.

TGF- β Signaling Mediates MMPs Expression in Cancer Cells

TGF-β activates several intracellular signal pathways which may explain its wide role in cancer, as well as its profound impact in the regulation of MMPs. In the next section we will focus on the MAPKs (ERK1,2, JNK and p38) and NF-κB, which, together with the Smads allow a better understanding of the regulation of MMPs by TGF-β in cancer. As mentioned previously, TGF-β induces MMP2 expression by trans-activating ATF2 [58]. This transcription factor has been shown to be a target of Smads and TAK1-p38 MAPK in TGF-β signaling [61]. In fact, in breast epithelial cells p38, but not ERK1,2, is required for the induction of MMP2 by TGF-β [58]. Recently, it was demonstrated that TGF-β enhances the invasiveness of SW1990 cells through the activation of Rac1/reactive oxygen species (ROS)/ NF-κB and the following secretion and activation of MMP2 [62]. MMP9 has been shown to be upregulated by TGF-β through the activation of ERK1,2 in transformed keratinocytes [63]. In addition, the increment of ROS by Rac1 dependent mechanism, mediates the activation of NF-κB by TGF-β to induce MMP9 expression [64]. In breast cancer cells MD-MB-231, TGF-β/ALK5-mediated activation of ERK1,2 contributes to TGF-β-mediated regulation of MMP9, whereas p38, JNK and Smad4 inhibition did not affect MMP9 levels [65]. Furthermore, TAK1/ NF-κB have been essential in the induction of MMP9 by TGF-β and breast and hepatocellular carcinoma cells [66,67].
Further studies showed that PI3K/Akt/Nf-κB signaling pathway was involved in the induction of MMP9 by TGF-β in chronic myeloid leukemia pathogenesis [68], suggesting the complexity of the NF- κB signaling in response to TGF-β, resulting in MMP9 induction in cancer cells. Interestingly, in some non transformed cells, TGF-β produces an inhibition of MMP9 production mediated by NF-κB [48].
MMP13 expression in transformed human epidermal keratinocytes-derived cell lines (HaCaT cell line/A-5) and cutaneous SCC cells (UT-SCC-7) is dependent on TGF-β activation of p38 MAPK, concomitantly with the expression of MMP1 and MMP9 [69].
In pleural malignant mesothelioma TGF-β induces pro-MMP-2 synthesis via the p38 MAPK signaling pathway [70]. In squamous cell carcinoma from the oral cavity, TGF-β enhances MMP2 and MT1-MMP expression, and although p38 and ERK1,2 MAPK were activated, the expression of both MMPs was independent of TGF- β-induced MAPK activation, suggesting that other signaling may be involved in the regulation of these two MMPs. Interestingly, p38 inhibition induces ERK1,2-dependent inhibition of MMP2 activation by inducing TIMP2 expression. These data imply a complex role of MAPKs in the activation of MMP2 in squamous carcinoma cells independent of transcriptional activation of MMP2 and MT1-MMP gene expression [71].
An interesting point, although it was reported in immortalized human keratinocytes HACAT, is that the MT1-MMP-dependent activation of JNK by TGF-β correlated with the increment of MMP9 expression by this growth factor. The attenuation of MT1-MMP by siRNA inhibited JNK activation and the induction of MMP9 in response to TGF-β, concomitantly with the inhibition of cell migration [72]. There is evidence that MMP14 activates a number of intracellular signal pathways, such as MAPKs, during cell migration and tumor invasion [73,74], and data presented by Seomun et al [72] suggested a novel interplay between MMP14 and the activation of JNK and expression of MMP9 induced by TGF-β. However, further studies will be necessary to elucidate the implications on cancer cell migration and the mechanisms involved.

Epithelial Mesenchymal Transition

The discovery that EMT generates cells with many properties of self-renewing stem cells holds the promise of resolving a major problem in cancer biology. EMT is a differentiation process by which epithelial cells undergo transition into mesenchymal cells, and it occurs during embryogenesis and tissue morphogenesis (type 1 EMT); wound healing and tissue fibrosis (type 2 EMT); and cancer progression (type 3 EMT) [75,76]. Many types of cancer cells leaving primary carcinomas appear to rely on the EMT program to facilitate execution of most of the steps of the invasion-metastasis cascade [77,78]. During the EMT, early phenotypic changes involve loss of epithelial cell-cell contacts by downregulation of junction complex members, including claudin-1, ZO-1 and E-cadherin, typical epithelial markers. Interestingly, as E-cadherin plays a critical role in the epithelial homeostasis, its downregulation can lead to decreased expression and/or organization of additional epithelial markers, desmosomal proteins (such as plakoglobin, desmogleins and desmoplakins) [78,79]. Furthermore, epithelial cells lose the apical-basal polarity showing spindle cell phenotype; cytoskeleton is subjected to profound reorganization, the expression of cytokeratins is lost, concomitantly with the expression of mesenchymal vimentin network and rearrangement of actin cytoskeleton. Together with an increase in motile behavior, all these events cooperate to increase tumor cell motility and invasive cell phenotypes [80-82].

The Involvement of TGF-Β and MMPs in EMT

Currently, TGF-β is recognized as a master regulator of EMT, as this growth factor participates in all types of EMT. Tumor cells persistently exposed to TGF-β elicit EMT, which plays a pivotal role in cancer progression [76]. In type 3 EMT, TGF-β may cooperate with several other oncogenic pathways to induce and maintain the mesenchymal phenotype of metastatic tumor cells, allowing the regulation of TGF-β induced genes and downregulation of E-cadherin expression among others [82,83].
TGF-β can induce EMT by activating Smad3 signaling, which, together with Smad4, has been shown to be crucial in EMT promotion [80,84-86]. In contrast with the role of Smad3, Smad2 has been postulated as an inhibitor of EMT, since Smad2 ablation enhances EMT during skin carcinogenesis [87]. Conversely, Smad2 has also been shown to participate in the TGF-β1-induced EMT, since overexpression of constitutively active Smad2 enhances EMT acquisition of carcinoma cells [88]; however, further analyses are necessary to elucidate the specific role of Smad2 in EMT. The capacity of TGF-β to induce EMT also requires cooperation with a number of different intracellular signaling pathways, such as: Ras and Rho GTPases (Rho and Rac1), MAPKs, Wnts and NF-κB and many of them participate in the regulation of MMPs expression [64,89-91]. TGF-β regulates the expression of EMT-involved genes by modulating the expression of transcription factors such Snail and Slug (also named as SNAI1 and SNAI2 in humans). For example, Snail mediates TGF-β- induced EMT by repressing E-cadherin transcription and stimulating the expression of mesenchymal genes vimentin and α-SMA, among others. In turn, Snail promotes collagen-I synthesis and deposition, and may up-regulate the expression of pro-inflammatory interleukins IL-1, -6 and -8 which produce an inflammatory microenvironment supporting the acquisition of EMT of the cancer cells [82,92-94]. During EMT cells acquire mesenchymal and stem cell-like features, increase their motility and invasiveness, as well as become resistant to apoptosis and acquire anchorage-independent growth. Furthermore, up-regulation of serine proteinases and MMPs leads to the degradation of ECM proteins and provides the tumor cells with the capacity to invade surrounding tissues and colonize distant organs [2,27].
New evidence indicates that MMPs may stimulate the EMT. Radisky and Radisky [95], have described three distinct mechanisms by which MMPs can be associated with EMT and tumor progression: firstly, the increased levels of MMPs in tumor microenvironment may directly induce EMT; secondly, cancer cells undergoing EMT can express and produce higher levels of MMPs, which facilitate cell invasion and metastasis; and thirdly, the generation of activated mesenchymal/stromal-like cells by EMT may drive cancer progression through MMP production [95,96]. In further text, we will analyze the direct participation of MMPs in the induction of EMT. One of the first evidence of MMPs-induced EMT comes from experiments using MMP3 [97], which showed that MMP3 directly degraded E-cadherin in mammary epithelial cells leading to EMT. By using recombinant MMP3 to treat the mammary epithelial SCp2 cell line, a display of EMT was obtained, since cells lost the cell-cell interaction concomitantly with the acquisition of the motile phenotype, which was accompanied with the downregulation of cytokeratins and upregulation of vimentin. Also, when cells were stably transfected with MMP3 they produced highly invasive tumors with the expression of mesenchymal phenotype in tumor periphery. Moreover, the specific overexpression of MMP3, in mammary epithelium of mice, resulted in the development of malignant lesions in transgenic animals [98]. Conversely, bitransgenic animals co-expressing MMP3 and TIMP1 did not show tumor development, indicating that active MMP3 was necessary for the promotion of mammary neoplasias displaying EMT phenotype [98]. Several other MMPs have been implicated in the induction of EMT. MMP2 was shown to be necessary and sufficient for the induction of tubular EMT in vitro [99]. Furthermore, MMP9 induces EMT by cleaving the E-cadherin ectodomain, which in turn provokes the dissociation of β-catenin from cell-cell adhesion complexes. Subsequently, β-catenin is translocated to the nucleus leading to cognate gene regulation. Moreover, the disruption of E-cadherin-β-catenin complexes by MMP9 and TGF-β1 results in the increment of the transcription factor SNAI2 which strongly represses E-cadherin expression at transcriptional level [100,101]. In addition, MMP9 cooperates with the transcription factor SNAI1 to induce EMT in A431 cells [102]. This may result in a strong induction and further maintenance of cells in mesenchymal stage. Also, MMP7 appears to cleave E-cadherin leading to tracheal epithelial cell scattering and migration [103]. Similarly, MMP28/epilysin, one on the last MMP members discovered, leads to irreversible EMT in lung carcinoma A549 cells associated with the loss of E-cadherin and enhancement of the invasive phenotype and upregulation of MMP9 and MMP14. Cell surface attached MMP28 activates latent TGF-β complexes, and in turn activated TGF-β cooperates in the induction of EMT and in the upregulation of MMP9 and MMP14 [104]. MMP14 is capable of cleaving E-cadherin in transfected breast cancer cells [105]; similarly, oral squamous cell carcinoma SCC9 cells transfected with MMP14 displayed EMT, concomitantly with the expression of Twist and ZEB, downregulation of E-cadherin and high invasive abilities. Furthermore, transfected cells exhibited cancer stem cells-like characteristics since they have shown low proliferation, self-renewal ability, resistance to chemotherapeutic drugs and apoptosis, and expression of cancer stem cells surface markers [106]. In prostate cancer cells, the overexpression of MMP14 also induced EMT, and EMT phenotypic changes were dependent upon upregulation of Wnt5a. This induction was accompanied with cell scattering and acquisition of migratory mesenchymal-like phenotype, while these changes were abrogated by target inhibition of either catalytic domain or the hemopexin domain of MMP14 [107]. One novel mechanism has been described for the induction of EMT in MCF-7 cancer cells involving IGF-1 [108]. IGF-1 induces the extracellular activation of MMPs in a mechanism dependent of PI3K and MAPK activation. Active MMPs are further capable to release and activate latent TGF-β, which in turn induces EMT. In head and neck squamous carcinoma SCC10A cell line, EGF activation of EGFR promote cell migration and invasion by inducing EMT and MMP9-dependent degradation of E-cadherin as well. EGFR activation triggered ERK1, 2 and PI3k signaling, which mediated the induction of MMP9 and malignant EMT [109]. These are examples of the sophisticated and complex relationships established between different cytokines and signaling pathways which drive tumor progression [27]. Interestingly, other endogenous TGF-β receptor ligands different from the TGF-βs themselves have been found to modulate EMT. Signal peptide-CUB-EGF-like domain containing protein 3 (SCUBE3) is a secreted glycoprotein upregulated in lung cancer that behaves as an endogenous ligand for TβRII. MMP2 and MMP9 cleave SCUBE3 in two major fragments: the N-terminal EGFlike repeat and the C-terminal CUB domain. The CUB fragment (and also the full-length SCUBE3 protein) binds TβRII, activates TGF-β signaling and promotes EMT linked to upregulation of MMP2 and MMP9 and increased tumor cell migration/invasion [110]. This finding constitutes a novel example of a positive TGF-β autocrine regulatory loop linked to tumor progression [27]. It is believed that TGF-β is a potent factor which induces MMPs expression in tumor cells concomitantly with EMT induction. Altogether, TGF-β and MMPs are interacting during the display of EMT and tumor progression (Figure 5), and indicate the importance of mutual autocrine loops in maintaining the mesenchymal features of tumor cells that have undergone EMT.
Figure 5: TGF-β and MMPs system cooperation during the induction of epithelial to mesenchymal transition.
Both TGF-β and MMPs are involved in the induction of EMT. There is mutual cooperation between the two, since TGF-β stimulates the expression of MMPs in cancer cells, and the enhancement of MMPs levels increases the activation of the extracellular matrix associated latent complex, thus exacerbating TGF-β-induced EMT. Meanwhile, the increment in MMPs production also stimulates EMT, and both finally may collaborate to induce the epithelial conversion to the mesenchymal phenotype, thus strengthening the cancer cell invasion and metastasis.


There is a large number of evidence supporting the important role of the TGF-β and MMP family in the course of cancer progression and metastasis. Due to the importance of TGF-β and MMPs in tumorigenesis, they have become attractive targets for cancer therapies.
Targeting TGF-β has already been clinically tested in therapeutic approaches [16,111]. These strategies included small inhibitors of the enzymatic activities of urokinase plasminogen activator (uPA) or TGF-β receptors, specific neutralizing antibodies, peptide inhibitors such as p44, as well as therapeutic approaches to inhibit the expression of TGF-β signaling components at transcriptional level, among others. Huge effort has been made in the last two decades in the design of potent and selective MMPs inhibitors, however, this has been extremely difficult and laborious due to the high similarity of the conformational structures around and within MMPs active sites, which proved to be a big challenge in the use of small molecular inhibitors in clinical trials. Moreover, some MMP members seem to be suppressors of tumor progression, which pointed out the risk of broad spectrum inhibitors use, which can lead to the undesired effect of increasing cell malignancy [95]. The inhibition of MMPs activity is not the only strategy in the treatment of cancer. Another approach for selective MMP targeting with considerable potential lies in RNA interference technology, in which expression of a target gene is reduced by short double stranded RNA. In experimental mouse models targeting MMP9 with siRNAs resulted in the inhibition of tumor initiation and progression [112], and seems to be a promising strategy for future therapeutic selective MMP inhibition [25]. In addition, the development of specific MMP blocking antibodies opened new opportunities in the treatment of cancer, i.e. the mAb REGA-3G12 has been shown to selectively inhibit MMP9 activity without affecting MMP2 [113]. In this review, we attempted to reveal the interplay between TGF-β and MMPs. We believe that the inhibition of the amplification loop operated between TGF-β and MMP system in tumor cells could limit tumor progression and metastasis, impairing tumor dissemination, proliferation and survival. We hope future clinical trials using combined therapies which target TGF-β and MMPs could increase the success of cancer treatment. Moreover, TGF-β and MMPs induce EMT, which enhances tumor cells migration and invasion, and at the same time enhances the population of cancer associated fibroblasts [114], which may open new avenues for the treatment of cancer. By regulating TGF-β and MMPs it could be possible to control the positive tumor microenvironment and cancer cells-stroma cells interaction. The key future challenge is, in one way, to develop highly specific MMP inhibitors (for the activity and/or expression) usable in clinic, and a combined therapy targeting the TGF-β-MMPs loop to control the development of malignant EMT resulting in cancer progression. Elucidating the complex interplay and roles of TGF-β and MMPs in cancer is critical for the understanding of their participation in the initiation, progression and tumor metastasis, and could eventually uncover potential combinatory therapeutic targets for the treatment of cancer.


We apologize to those colleagues whose work, although relevant to the issues dealt within this review, has not been included due to space limitations. The work was supported by Ministry of education; science and technological development of the Republic of Serbia (grant 175062).



Track Your Manuscript