Inhibition of MMPs supports amoeboid angiogenesis hampering VEGF-targeted therapies via MLC and ERK 1/2 signaling

BACKGROUND: In the past decades studies on anti-tumoral drugs inhibiting matrix metalloproteinase (MMPs) were disappointing. Recently, we demonstrated that mature endothelial cells (ECs) and endothelial colony forming cells (ECFCs) can switch between invasion modes to cope with challenging environments, performing the “amoeboid angiogenesis” in the absence of proteases activity. METHODS: We first set out to investigate by ELISA if the inhibitors of the main protease family involved in angiogenesis were differently expressed during breast cancer progression. We used Marimastat, a broad-spectrum MMP inhibitor, as a means of inducing amoeboid characteristics and studied VEGF role in amoeboid angiogenesis. Thus, we performed invasion and capillary morphogenesis assay, morphological, cell signaling and in vivo mouse studies. RESULTS: Our data showed that TIMP1, TIMP2, alpha2-antiplasmin, PAI-1 and cystatin increase in breast cancer serum of patients with primary cancer and lymph node positive compared to healthy women. In vitro results revealed that the most high-powered protease inhibitors able to induce amoeboid invasion of ECFCs were TIMP1, 2 and 3. Surprisingly, Marimastat promotes ECFC invasion and tubular formation in vitro and in vivo, inducing amoeboid characteristics. We observed that the combination of Marimastat plus VEGF doesn’t boost neither cell invasion nor vessel formation capacity. Moreover, inhibition of VEGF activity with Bevacizumab in the presence of Marimastat confirmed that amoeboid angiogenesis is independent from the stimulus of the main vascular growth factor, VEGF. CONCLUSIONS: We underline the importance to consider the amoeboid mechanism of endothelial and cancer cell invasion, probably responsible for the failure of synthetic metalloproteinase inhibitors as cancer therapy and tumor resistance to VEGF-targeted therapies, to set-up new drugs to be used in cancer therapy.