Cancer susceptibility genes such as breast cancer 1 ( BRCA1 ) do not function in isolation — they are parts of networks, and deciphering how these networks operate is an important step in understanding cancer progression. A new study has constructed a network for breast cancer susceptibility using various 'omic' data sets, and identified the hyaluronan-mediated motility receptor gene ( HMMR ) as a new susceptibility locus for the disease.

The authors sought to construct a network around four genes that are known to be associated with breast cancer: BRCA1, BRCA2 , ataxia telangiectasia mutated ( ATM ) and checkpoint homologue ( CHEK2 ). They used published expression data from normal cells to identify 164 other genes, the expression of which correlated with all four of their reference genes. The relevance of these genes for the BRCA-related network was confirmed by the fact that their expression was in many cases upregulated in tumours from carriers of BRCA1 mutations compared with sporadic tumours.

This list was integrated with published functional genomic data about interactions between the genes on the list. The resultant network consisted of 118 genes and proteins, with 321 direct and 545 indirect interactions. One of the most highly connected nodes in the network, and the one with the closest coexpression with BRCA1, was HMMR, which probably has a role in centrosome function.

The authors further explored the role of HMMR using a yeast two-hybrid assay, and found that HMMR interacts with several components of BRCA1 complexes. Using co-immunoprecipitation assays, the association of BRCA1 and HMMR was seen specifically as cells entered mitosis, and the authors also showed that HMMR was ubiquitylated by the BRCA1 complex and was localized to the centrosome. Knocking down either HMMR or BRCA1 caused abnormal increases in centrosome number, but this effect was abolished if the expression of both genes was inhibited simultaneously.

In the light of this functional information, the authors looked for associations between SNPs in HMMR and breast cancer susceptibility. Typing for five SNPs in several independent populations showed that two haplotypes significantly increased susceptibility to breast cancer, and that this effect was independent of BRCA1 or BRCA2 status. Interestingly, expression analysis showed that one of these haplotypes was associated with an increase in HMMR expression whereas the other was associated with a decrease. This implies that any perturbation from a precise level is detrimental.

This study comprised several components — the construction of a network model from published data, the functional molecular characterization of a candidate component of this network, and a candidate-gene population association study to demonstrate its link with cancer susceptibility. It shows the power of network modelling to generate hypotheses for both molecular and population genetics.