Genes And Proteins Involved In Extracellular Matrix Remodeling During Tumor Invasion

The structure of the extracellular matrix is generated by a backbone of fibrillar proteins and proteoglycans. This matrix is in a dynamic state. Structural proteins are synthesized by fibroblasts and other cells embedded in connective tissue. These same cells, together with cells of the immune system and from blood vessels, are responsible for the turnover of the extracellular material. Turnover and remodeling of the ECM are enhanced during inflammation and wound repair, as they are during tumor invasion. In fact, essentially the same mechanisms and enzymes are involved in each situtation.

The key enzymes among the many involved in ECM remodeling are proteases, and among these the plasminogen activator (PA) protease cascade and the matrix metalloproteinases (MMPs). These proteases degrade components of the basement membrane as well as proteins and proteoglycans of connective tissue extracellular matrix. They process other proteases and enzymes as well as growth factors and they liberate 'latent' growth factors from their storage sites in the extracellular matrix. These growth factors then act on several different cell types, incuding or inhibiting proliferation or eliciting apoptosis in some cases. Among the factors that are activated in this fashion are FGFs (fibroblast growth factors), HGF (hepatocyte growth factor), and TGF0s (^6.7). Obviously, proteins on cell surfaces can also be substrates.

The end result of the plasminogen activator cascade (Figure 9.4) is localized activation of plasmin near the surface of specific cells. The inactive precursor of plasmin, plasminogen, is synthesized and secreted mainly in the liver and is present throughout the body. It is specifically cleaved to an active protease by plasminogen activator, abbreviated uPA (for 'urokinase-type plasminogen activator'). The activity of uPA is localized to cell surfaces by its binding to a specific membrane receptor, uPAR ('urokinase receptor'). The uPAR in turn is localized on the cell surface by dynamic interaction with specific integrins. Thus, cell-type dependent expression of the uPAR is one mechanism of directing plasmin activity to a specific location. Plasmin activity is further localized by a dynamic association of the uPAR with integrins on the cell surface.

Figure 9.4 The plasminogen cascade The successive steps of the cascade (1. - 5.) lead to localized activation of plasmin and matrix metalloproteinase. See text for further explanations and cf. also Figure 9.5.

Conversely, the expression of plasminogen activator inhibitor 1 (PAI-1) prevents the cleavage of plasminogen and thus excludes plasmin from the environment of cells that express it. Furthermore, the protein protease inhibitor a2-antiplasmin limits the activity of plasmin in the extracellular fluid in general. Plasmin is best known for its ability to digest fibrin, but it is also capable of digesting a range of ECM proteins as well as of cleaving and activating latent growth factors, pro-MMPs, and pro-uPA.

An even wider range of substrates can be digested by matrix metalloproteinases (MMPs). There are at least 23 members of the MMP family in man (Table 9.2). Between them, they are able to degrade every protein in the extracellular matrix, although each MMP displays preferences for certain substrates. All MMPs have a similar basic structure with a Zn2+ ion essential for catalytic activity. Accordingly, compounds that complex Zn2+, such as hydroxycarbamates or tetracyclines, are inhibitors, although often not specific for one member of the family. More specific inhibitors have been designed and have been tested in clinical trials.

As usual for proteases, MMPs are synthesized as proproteases ('zymogens'), and since most of them are secreted, they contain initially a signal peptide. The membrane-type MMPs differ from all others by containing a transmembrane domain at their C-terminal end. This locates them to the cell surface. Since pro-MMPs, mostly MMP-2, are among their substrates, they contribute to precisely localized protease activity, as would be required at the edges of a migrating cell or the growth cone of a tumor mass.

The MMPs are not only regulated at the level of activation from pro-enzymes, but also substantially at the transcriptional level. Their basal expression is cell type dependent and various factors induce their expression. These (depending again on the cell-type) include interleukins and growth factors of the EGF, FGF and TGFp families, as well as stress and adhesion signals. Intracellularly, these factors act through one of the ERK, JNK, or p38 MAPK cascades which increase transcription by activating AP-1 factors at a conserved binding site in the gene promoters.

As for plasmin, specific as well as general protease inhibitors limit the activities of MMPs. MMPs are confined to local activity by ubiquitous protease inhibitors in blood and extracellular fluids, mainly a2-macroglobulin and a1-antiprotease. Moreover, they are restricted as well as directed by specific inhibitors, the tissue

Table 9.2. Some matrix metalloproteinases involved in human cancers

Family Members Structure*

Characteristic Substrates

strongly variant gelatins, catalytic domain laminin differs from various collagenases in structural pro-domain proteins, collagens

proteoglycan, pro-MMPs, protease inhibitors

type 16, 17

Matrilysin 7 S/P-C/D-HR

membrane- pro-MMPs, associated collagen lacks HS domain gelatin, fibronectin

* The basic structural domains are signal peptide/prodomain (S/P), zinc-binding catalytic domain (CD), hinge region (HR), hemopexin similarity (HS), transmembrane domain (TM)

inhibitors of metalloproteinases ('TIMPs'). There are four members of this family, TIMP-1 through TIMP-4. They bind to the substrate-binding site of the MMPs, but cannot be cleaved by their catalytic centres. TIMPs also block the activation of pro-MMPs. The TIMP inhibitors are produced by a broader range of cell types than the proteases, and at least TIMP-2 is often constitutively expressed. However, the expression of most TIMPs is additionally regulated at the transcriptional level. In general, the same factors that induce MMPs are involved, but additional cytokines such as TNFa as well as glucocorticoids and retinoids are active, at least in the regulation of TIMP-1.

It would be too simplistic to state that the activity of MMPs in a tissue depends on the relative levels of MMPs and TIMPs, for two reasons. (1) TIMPs exert a number of (not too well defined) effects on cells beyond inhibiting MMPs. TIMP-3, in particular, appears to act as a pro-apoptotic factor. It is accordingly down-regulated in many tumor cells, usually by promoter hypermethylation. (2) While antagonism between TIMPs and MMPs is the rule, there are clearly exceptions. One of the better defined ones is that TIMP-2 is a cofactor in the activation of MMP2 by membrane-type MMPs. This example illustrates the likely role of TIMPs in physiological tissue remodeling. MMPs are produced by specific cell types, but TIMPs by most others. This means that MMP activity is limited to specific sites within a tissue 'around' cells that perform the remodeling, while other cells remain protected (Figure 9.5). In fact, even the MMP activity 'around' an actively remodeling cell may not be homogeneously distributed, as particularly evident during cell migration.

The changes in protease and protease inhibitor expression taking place during tumor invasion must be considered on this background. In many human cancers, protease expression and activity are increased during tumor progression and invasion. Typically, mRNA and protein levels of several proteases are enhanced, e.g. of MMP-2, MMP-7, MMP-9, and MMP-14 as well as of uPA. Conversely, TIMPs tend to be down-regulated, most consistently TIMP-3. Enhanced expression of MMPs and of uPA/plasmin are almost generally associated with a worse outcome, i.e. primary tumors with increased expression are more likely to have metastasized than those with low expression.

The relationship between TIMP expression and tumor behavior is not as consistent. Whereas TIMP-3 down-regulation is often found in more aggressive cancers, decreased expression of the other inhibitors is not always predictive of a particular clinical course, and in some cancers the relationship is inverse. For instance, increased expression of TIMP-1 may indicate a worse prognosis in breast cancer. The explanation for this apparent paradox may lie in the question where precisely the proteases and the inhibitors are expressed. In many carcinomas, the increased expression of MMPs is found in the stromal part of the tumor. While tumor cells largely stain negative for many MMP mRNAs and proteins, often expressing only MMP-7, the stromal cells, fibroblasts and particularly invading macrophages and monocytes, express a full range of active proteases. Conceivably, in this situation, the expression of TIMPs may actually protect the carcinoma cells. This type of distribution of protease expression between carcinoma and stromal cells appears to be most pronounced in adenocarcinomas and other cancers arising from glandular structures, whereas other carcinomas, e.g. from squamous epithelia, express a wider range of MMPs themselves.

Thus, tissue remodeling during tumor invasion is far from a one-sided affair and in most cases requires an active contribution by the tumor stroma. This may be part of the explanation why an 'active stroma' gene expression 'signature' in expression profiling studies correlates with the presence of metastases.

Figure 9.5 Mechanisms that localize MMP activity Some cells of a cancer (on the left) invading connective tissue (center) undergo an epithelial-mesenchymal transition (spindle-shaped cell in the center). The cancer cells in this case secrete MMP7 but also TIMP1. Endothelial cells (right) secrete TIMP2 which is (still) sufficient to neutralize MMPs secreted by stromal cells. These are only activated around the leading edge of the migrating cancer cell by membrane-type MMP. In the vessel (outside right) a-macroglobulin and a-antiprotease inactivate any MMP reaching the bloodstream.

Figure 9.5 Mechanisms that localize MMP activity Some cells of a cancer (on the left) invading connective tissue (center) undergo an epithelial-mesenchymal transition (spindle-shaped cell in the center). The cancer cells in this case secrete MMP7 but also TIMP1. Endothelial cells (right) secrete TIMP2 which is (still) sufficient to neutralize MMPs secreted by stromal cells. These are only activated around the leading edge of the migrating cancer cell by membrane-type MMP. In the vessel (outside right) a-macroglobulin and a-antiprotease inactivate any MMP reaching the bloodstream.

10 Ways To Fight Off Cancer

10 Ways To Fight Off Cancer

Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.

Get My Free Ebook

Post a comment