Semaphorin 3F (SEMA3F), one of a family of axonal guidance regulators, was first identified on part of a chromosome commonly deleted in lung cancer, indicating that it might also be a tumour suppressor. Michael Klagsbrun, Diane Bielenberg and colleagues now report that SEMA3F inhibits melanoma metastatic spread and angiogenesis — effectively leaving a tumour no way of progressing and rendering it benign.

Many tumour cell types express SEMA3F, but highly metastatic tumours downregulate it. Klagsbrun and colleagues took a highly metastatic human melanoma cell line (SM) that expressed high levels of the SEMA3F receptor neuropilin 2 (NRP2) and showed that transfection of these cells with SEMA3F decreased adhesion and motility in culture — both necessary features of a metastatic phenotype.

So, does SEMA3F expression inhibit metastasis in vivo? When either SM control cells or SM/SEMA3F cells were injected into nude mice tumorigenicity was 100%, but only those injected with SM cell clones developed large lymph-node and lung metastases. The authors confirmed this lack of metastatic ability of the SM/SEMA3F cells using several approaches; for example, fluorescently labelled tumour cells injected into mice resulted in no occult spontaneous metastases. The authors observed that the SM/SEMA3F tumours were encapsulated and had well-defined borders composed of fibroblasts and collagen matrix, whereas the SM tumours showed massive keratinocyte hyperplasia overlying the tumour and were not encapsulated.

Next the authors compared the tumour vasculature in mouse models. The blood vessels in SM/SEMA3F tumours were smaller, less branched and had less than half the number of vessels than SM tumours. Examining the stained cryosections of the tumours and surrounding skin carefully, the authors noticed that blood vessels seemed to be blocked from invading the tumour. Could SEMA3F be repelling the blood vessels in a similar way to that in which SEMA3F repels growing axons from ganglia? Klagsbrun, Bielenberg and colleagues found that when SM/SEMA3F tumour cells were added to cultures of endothelial cells (EC) expressing NRP2, large zones of cell clearance appeared. If the tumour cells lacked SEMA3F or the ECs lacked NRP2 no clearance was observed. Time-lapse video clearly showed that chemorepulsion of the ECs by the tumour cells was an active process.

So, SEMA3F inhibits metastatic growth in many ways, mediated through its functional receptor NRP2 on both melanoma cells and endothelial cells. The authors are now investigating the expression of SEMA3F in primary melanoma and matched melanoma metastatic tissue from patients.