Going forward together: cooperative invasion in melanoma

Early-stage melanomas can be excised by surgery, but once they have begun to invade and metastasize, treatment requires additional therapeutic interventions. Thus, understanding the mechanisms that drive tumor invasion and metastasis is critical for drug development, and to prevent the spread of disease. In a recent study from the Wellbrock and Hurlstone laboratories, melanoma cells with less invasive potential are shown to become invasive with the help of other tumor cells, demonstrating that melanoma cells can work together to promote malignancy (Chapman et al., 2014). These interactions may be important drug targets. One of the most enticing features of the zebrafish system for cancer biology is live in vivo imaging (White et al., 2013). Zebrafish embryos are optically transparent, enabling fluorescent reporter lines to reveal tissue development and cellular movements in amazing detail. Zebrafish are fertilized outside the mother, so that embryos can be imaged from the single cell stage through to adulthood. As vertebrates that share over 70% of the genome with humans, zebrafish develop cancers with many of the pathological features as humans. In addition, zebrafish are amenable to (xeno) transplantation, so that the fundamental aspects of cancer cell biology can be visualized in living animals. Their small size (10 embryos can easily fit into a well of a 24-well plate in 1 ml of solution) means that this is an ideal system to screen and test new compounds. Already, drug activity identified through zebrafish small molecule screens are in clinical trial, including leflunomide for the treatment of melanoma (NCT01611675). Malignant melanoma is comprised of different tumor cell subpopulations, but how these subpopulations relate to each other and contribute to cancer progression is not well understood. Multicellular imaging in cancers has revealed networks of paracrine interactions that contribute to heterogeneity and collective cell movements (Calvo and Sahai, 2011). In fact, some of the first evidence for cooperative interactions in melanoma was shown in primary melanoma explants whereby differential adhesion to matrix-enabled collective migration (Hegerfeldt et al., 2002). In a new study by Chapman et al. (2014), they take advantage of a zebrafish-melanoma xenograft model to show that heterogeneous tumor cell subpopulations can interact with each other to contribute to tumor progression via cooperative invasion. To examine the potential importance of melanoma heterogeneity, Chapman et al. used two melanoma cell lines, both with the common BRAF mutation, but with different invasive potentials (WM266-4 highly invasive; 501mel poorly invasive; both derived from human metastatic melanoma). Zebrafish embryos were injected in the pericardium with either WM266-4 cells, 501mel cells or with an equal ratio of both cell types. In all cases, the melanoma cells formed tumorlike masses. Some cells could be imaged migrating away from the pericardium tumor mass in filelike patterns. Consistent with findings in other assay systems, WM266-4 cells displayed high invasiveness, whereas 501mel cells did not invade. The surprise came in the case of heterogeneous xenografts, where the invasion potential of 501mel cells increased to levels similar to that of WM266-4 cells. This indicates that poorly invading melanoma cells can alter their behavior to actively invading when interacting with other cells in a heterogeneous environment. While the invasion potential of 501mel cells could be stimulated by the WM266-4 cells, closer examination revealed that most of the time WM2664 cells were leading the files of cells. It is well known that degradation of the extracellular matrix (ECM) by matrix metalloproteases is an important feature of cancer invasion. Indeed, the authors discover that WM266-4 cells express significantly higher levels of MMPs than 501mel cells. To test the involvement of the proteolytic degradation of ECM in cooperative invasion, the team immersed the xenografted larvae in water with protease inhibitors. This had no effect on homogenous xenografts of either 501mel cells (in which there is no invasion in any case) or WM266-4 cells (these cells continue to invade). In heterogeneous xenografts, however, the inhibition caused a block of invasion in 501mel cells and a significant decrease in invasion in WM266-4 cells. Supporting these findings, they then depleted the expression of a major regulator of protease-driven invasion, MT1-MMP, in WM266-4 cells, resulting again in the suppression of cooperative invasion, but not affecting the invasion of homogenous WM266-4 cells. Thus, MMPs are critical for invasion of both cell types in the context of heterogenous xenografts, but not in the context of homogenous WM266-4 cells alone. The authors hypothesize that 501mel cells secrete a factor that inhibits the proteolytic-independent invasiveness of WM266-4 cells. Chapman et al. use a system with WM266-4 spheroid cells embedded in collagen and cocultured with either 501mel or autologous cells in transwells, so that they are physically separated each other. They find that the exposure to the 501mel cells modulates the WM266-4 cell response to protease inhibitors: WM266-4 cells shift from a Coverage on: Chapman, A., Fernandez del Ama, L., Ferguson, J., Kamarashev, J., Wellbrock, C. and Hurlstone, A. (2014) Heterogeneous tumor subpopulations cooperate to drive invasion. Cell Rep. 8, 688–695.