Can all BRCA1-associated breast tumors be considered as basal-phenotype cases?

1729

In the last few years a great amount of publications have came out regarding the use of expression microarrays for the characterization of breast cancer. Based of expression profiling, the “molecular portraits” of breast cancer have been defined and sporadic breast tumors classified into, at least, five subtypes (Perou et al. 2000), one of them the so called basal phenotype, over expressing a cluster of basal epithelium genes. Little is known about the expression profiling of hereditary tumors, but tumors with mutations in the high susceptibility breast cancer gene BRCA1 have been suggested to present a basal phenotype, although their specific molecular signature has not been established. The main goal of this study is to determine the expression profile of BRCA1 tumors using expression microarrays. For this purpose we have analyzed a series of tumors from fourteen BRCA1 mutation carriers that had been fully screened for mutations in this gene. RNA from the BRCA1 positive tumors were hybridized onto a cDNA microarray with around twenty-seven thousand clones, corresponding to nine thousand cancer related genes. For the data analysis GEPAS suite, CLUSTER and TREE-VIEW software were used. An unsupervised classification was performed in order to evaluate the organization of the samples with the whole set of genes present in the microarray, discriminating the tumors into two main branches. Tumors in one of them (65%) seemed to over express the typical markers of basal-phenotype (KRT5, KRT7, CDH3) as previously described, whereas the other group did not. A supervised hierarchical clustering comparing samples included in each of the two branches was performed, obtaining a signature of around two hundred genes with an adjusted p value < 0.05 The altered pathways were analyzed by KEGG’s bioinformatics tool, finding that cell cycle genes were upregulated in the basal-phenotype branch, in contrast to insulin signaling and MAPK pathways that presented a downregulation within the same group. This signature included several genes already proposed to be relevant to sporadic basal breast malignances. Our results confirm that the majority of BRCA1 tumors have a basal phenotype, but also point out the existence of a non-negligible (35%) percentage of cases that do not share this phenotype. We aim to define the whole molecular signature associated to the BRCA1-basal tumors.
 *Sponsored by Marie Curie 6^th Framework Programme and the Spanish Ministry of Health, Grant number FIS PI061090