Nature 425, 139 - 141 (11 September 2003); doi:10.1038/425139a
Cardiovascular biology: Signalling silenced
EDWARD M. CONWAY AND PETER CARMELIET
Edward M. Conway and Peter Carmeliet are at the Centre for Transgene Technology and Gene Therapy, Flanders Interuniversity Institute for Biotechnology, University of Leuven, B-3000 Leuven, Belgium.
e-mail: [email protected]
The mechanism by which the TIMP-2 protein inhibits blood-vessel formation has been uncovered — and it is not as expected. The finding has implications for treating a cancer by cutting off its blood supply.
Tumours and inflamed tissues cannot grow without blood vessels to supply them with oxygen. So interfering with the formation of new blood vessels has been proposed as an attractive means of combating both cancer and inflammation1. One class of proteins that stimulates blood-vessel formation (angiogenesis) is the matrix metalloproteinases (MMPs), which are kept in check by molecules called tissue inhibitors of MMPs (TIMPs). TIMPs suppress both angiogenesis and tumour growth, suggesting that synthetic compounds mimicking the MMP-inhibiting effects of these proteins might prove useful in treating cancer. But such efforts have generally been disappointing1-4. This could be explained by the fact that TIMPs also have other effects, and, writing in Cell, Seo et al.5 reveal a further, surprising function for one key TIMP. They show that it prevents angiogenesis by blocking a cellular signalling pathway, and that this ability has nothing to do with its MMP-inhibiting capacity.
The idea that blocking angiogenesis might prove beneficial for treating cancer is supported by recent evidence that an inhibitor of a prime angiogenic molecule — vascular endothelial growth factor (VEGF) — prolongs survival of cancer patients (see ref. 6 for a comment on this finding). One mechanism by which VEGF promotes angiogenesis is by stimulating protein-degrading (proteolytic) enzymes, including MMPs1-4. When first discovered, MMPs were thought to work mainly by breaking through the dense thicket of fibres that surrounds cells, thereby allowing cancer cells and blood vessels to invade surrounding tissues. The TIMPs, which inhibit the proteolytic activity of MMPs, suppress the migration of endothelial cells (which line blood vessels) and angiogenesis1-4. Together, these findings provided the rationale for treating cancer with synthetic compounds that, like TIMPs, block MMP activity — a formidable effort in which the pharmaceutical industry has invested at least a billion dollars over the past 20 years.
But matters are rarely straightforward when it comes to living organisms. For starters, MMPs have a much broader spectrum of targets than originally believed, including other proteinases, proteinase inhibitors, latent growth factors, cell-surface receptors and clotting factors1-4. Moreover, MMPs have diverse — and often opposing — effects on angiogenesis. For instance, they can stimulate angiogenesis by cleaving matrix components and releasing growth factors that are sequestered in the matrix (such as VEGF), exposed on cell membranes or bound to precursors. But they suppress angiogenesis by processes that include activating cell-death pathways, generating angiogenesis inhibitors, and inactivating receptors for angiogenic molecules.
The dogma that TIMPs suppress angiogenesis only by blocking MMP activity has also been challenged, for example by the finding that a synthetic MMP inhibitor, batimastat, does not impair endothelial-cell growth7. And some unexpected roles for TIMPs have been discovered: for instance, TIMP-1 actually stimulates VEGF-driven angiogenesis, and TIMP-3 blocks angiogenesis by preventing VEGF from binding to its receptor8. All of this may help to explain why synthetic MMP inhibitors have not lived up to expectations, and have been plagued with undesirable side effects.
The TIMP studied by Seo et al.5 is TIMP-2, which is perhaps the most biologically relevant angiogenesis inhibitor of this family, as it most consistently blocks endothelial-cell migration and proliferation. How does it do this? More than a decade ago, Stetler-Stevenson et al.9 identified this protein as an inhibitor of MMP-mediated proteolysis; they later showed10 that transferring the TIMP-2 gene into tumours inhibits angiogenesis. At the time, it seemed likely that it does this through its effects on MMPs. But later studies showed that TIMP-2 can downregulate VEGF expression11. And recently, Fernandez et al.12 found that TIMP-2 blocks the growth of endothelial cells in an MMP-independent way. This inhibition requires a peptide at the carboxy-terminal end of TIMP-2, which was superior to an amino-terminal TIMP-2 fragment (which blocks MMP activity) in inhibiting angiogenesis.
Now, Stetler-Stevenson's group (Seo et al.) brings further insight into the MMP-independent activity of TIMP-2. The authors show that TIMP-2 silences two growth-factor receptors — one that detects VEGF, and one for fibroblast growth factor-2 (FGF-2). They also offer clues to how this happens.
Many angiogenic factors, such as VEGF and FGF-2, bind to receptors on endothelial cells, causing the receptors to self-activate by adding phosphate groups to tyrosine amino acids in their intracellular portion13. This leads to a sequence of events that promotes endothelial-cell proliferation and migration, and hence angiogenesis. This process is partly regulated by enzymes that remove phosphate groups, such as SHP-1 and other 'protein tyrosine phosphatases', which prevent the receptors from transmitting further signals14.
Seo et al. show in vitro that a modified form of TIMP-2, which can no longer interact with MMPs, causes SHP-1 and other unidentified protein tyrosine phosphatases to associate with VEGF and FGF-2 receptors. This occurs in a remarkably indirect way: TIMP-2 binds to a neighbouring receptor on endothelial cells, called 31 integrin, with which phosphatases usually associate. This binding causes the phosphatases to move from the integrin to the VEGF and FGF-2 receptors (Fig. 1). The cells are thereby rendered resistant to VEGF and FGF-2 — an effect that Seo et al. also observed in a mouse model of angiogenesis. The authors further show that inhibitors of protein tyrosine phosphatases largely block the anti-angiogenesis effects of TIMP-2. The discovery of such receptor crosstalk reveals a high degree of coordination between proteinase inhibitors, integrins and growth-factor receptors. It also provides insight into how signalling from VEGF receptors is terminated — a poorly understood process.
Figure 1 Blocking blood-vessel growth: how TIMP-2 does it. Full legend
High resolution image and legend (35k)
What are the medical implications of these exciting findings? As mentioned above, the development of synthetic MMP inhibitors for cancer treatment has been based, at least in part, on the observation that TIMP-2 and other TIMPs inhibit tumour growth in animal models1-4. As a consequence, several MMP inhibitors have been extensively evaluated in clinical trials, but with disappointing results. But these inhibitors were selected for their TIMP-like ability to block the proteolytic activity of MMPs — and the findings of Seo et al.5 imply that the anti-angiogenic function of TIMP-2 stems primarily from its MMP-independent activity.
This conclusion needs to be confirmed. And it must be reconciled with gene-inactivation studies15 showing that TIMP-2's main function in vivo is to activate pro-angiogenic MMP-2, and with the finding that, for certain cancers, high levels of TIMP-2 predict poor (not, as one might expect for an angiogenesis inhibitor, good) prognosis. If Seo and colleagues' findings prove correct, however, it might be bad news for the strategy of using synthetic MMP inhibitors in the clinic. The good news, though, is that we now have other ideas for angiogenesis inhibitors: it might, for instance, be worth investigating analogues or fragments of TIMP-2 that induce phosphatases to silence the angiogenic growth-factor receptors.
References 1. Overall, C. M. & Lopez-Otin, C. Nature Rev. Cancer 2, 657-672 (2002). | Article | PubMed | ISI | ChemPort |
2. Brinckerhoff, C. E. & Matrisian, L. M. Nature Rev. Mol. Cell Biol. 3, 207-214 (2002). | Article | PubMed | ISI | ChemPort |
3. Egeblad, M. & Werb, Z. Nature Rev. Cancer 2, 161-174 (2002). | Article | PubMed | ISI | ChemPort |
4. Coussens, L. M., Fingleton, B. & Matrisian, L. M. Science 295, 2387-2392 (2002). | Article | PubMed | ISI | ChemPort |
5. Seo, D. W. et al. Cell 114, 171-180 (2003). | PubMed | ISI | ChemPort |
6. McCarthy, M. Lancet 361, 1959 (2003). | Article | PubMed | ISI |
7. Murphy, A. N., Unsworth, E. J. & Stetler-Stevenson, W. G. J. Cell Physiol. 157, 351-358 (1993). | PubMed | ISI | ChemPort |
8. Qi, J. H. et al. Nature Med. 9, 407-415 (2003). | Article | PubMed | ISI | ChemPort |
9. Stetler-Stevenson, W. G., Krutzsch, H. C. & Liotta, L. A. Matrix 1 (Suppl.), 299-306 (1992). | PubMed | ChemPort |
10. Valente, P. et al. Int. J. Cancer 75, 246-253 (1998). | Article | PubMed | ISI | ChemPort |
11. Hajitou, A. et al. Cancer Res. 61, 3450-3457 (2001). | PubMed | ISI | ChemPort |
12. Fernandez, C. A., Butterfield, C., Jackson, G. & Moses, M. A. J. Biol. Chem. published online 4 August 2003 (doi:10.1074/jbc.M306176200).
13. Schlessinger, J. Cell 103, 211-225 (2000). | PubMed | ISI | ChemPort |
14. Ostman, A. & Bohmer, F. D. Trends Cell Biol. 11, 258-266 (2001). | Article | PubMed | ISI | ChemPort |
15. Wang, Z., Juttermann, R. & Soloway, P. D. J. Biol. Chem. 275, 26411-26415 (2000). | Article | PubMed | ISI | ChemPort |