Cancer mortality is most often the result of metastasis rather than the primary tumour. Previous studies from Kent Hunter's group demonstrated that the genetic background of the host can influence metastatic efficiency. Now, Hunter and colleagues have identified a candidate gene, Sipa1 , with an amino-acid polymorphism that influences this process.

The authors previously used a mouse model of breast cancer to investigate the effect of constitutional genetic polymorphism on metastasis. They expressed the polyoma middle-T transgene in various strains of inbred mice and through quantitative trait genetic mapping showed the presence of a putative metastasis efficiency locus ( Mtes1 ) on mouse chromosome 19. This chromosome region, which is orthologous to human 11q12–13, harbours a known metastasis suppressor gene Brms1. However, this gene has no obvious polymorphisms that influence metastasis and so was discounted from this study.

To identify other potential candidates the authors used a multiple cross-mapping strategy that uses the shared haplotypes in different inbred strains of mice to reduce the number of candidate genes. This reduced the number of potential genes from 500 to 23, which were then prioritized based on their known molecular function. After analysing and discounting several of the genes, the authors found that Sipa1 had a polymorphism that results in an alanine (as found in the DBA mouse strain) to threonine (as found in the FVB mouse strain) substitution in a protein–protein interaction domain known as a PDZ domain. Sipa1 is a mitogen-inducible gene that encodes a GTPase activating protein (GAP) that negatively regulates RAP1 and RAP2 GTPases. Human SIPA1 has recently been found to interact with the water channel aquaporin 2 (AQP2), by its PDZ domain, so the authors used AQP2 to see if the alanine to threonine substitution affected this interaction. They found that it did — the FVB allele bound AQP2 less effectively.

What does this mean biologically? Transient transfection assays demonstrated that the FVB allele is less efficient than the DBA allele at reducing the activity of GTP RAP1. AQP2 inhibits this and does so more effectively with the DBA allele. So, cells expressing the FVB allele will have reduced levels of Rap–GTP activity. Reducing the expression of Sipa1 in cells in vitro indicates that SIPA1 modulates the adhesive properties of cells, consistent with its effect on RAP1, which is known to affect cell–cell interactions. But does Sipa1 influence metastasis? A series of experiments in mouse models showed that RNA inhibition of Sipa1 decreased the numbers of pulmonary metastases from a highly metastatic mammary tumour cell line. Conversely, overexpression of the FVB allele increased the numbers of pulmonary metastases. Analysis of human tumours also demonstrated that overexpression of SIPA1 is associated with metastatic progression.

These results demonstrate that Sipa1, as determined by its overall protein concentration and/or its availability to inactivate RAP1, modulates metastatic progression. The data also predict that homozygotes for the DBA allele would have reduced metastatic capacity because, in the primary tumour, cells are more likely to closely interact with one another. Additional studies are required to verify this and to investigate another potential gene close to Sipa1 that might also contribute to the Mtes1 locus.