Gene expression profiling integrated into network modelling reveals heterogeneity in the mechanisms of BRCA1 tumorigenesis

Abstract

Background: gene expression profiling has distinguished sporadic breast tumour classes with genetic and clinical differences. Less is known about the molecular classification of familial breast tumours, which are generally considered to be less heterogeneous. Here, we describe molecular signatures that define BRCA1 subclasses depending on the expression of the gene encoding for oestrogen receptor, ESR1. Methods: for this purpose, we have used the Oncochip v2, a cancer-related cDNA microarray to analyze 14 BRCA1-associated breast tumours. Results: signatures were found to be molecularly associated with different biological processes and transcriptional regulatory programs. The signature of ESR1-positive tumours was mainly linked to cell proliferation and regulated by ER, whereas the signature of ESR1-negative tumours was mainly linked to the immune response and possibly regulated by transcription factors of the REL/NFκB family. These signatures were then verified in an independent series of familial and sporadic breast tumours, which revealed a possible prognostic value for each subclass. Over-expression of immune response genes seems to be a common feature of ER-negative sporadic and familial breast cancer and may be associated with good prognosis. Interestingly, the ESR1-negative tumours were substratified into two groups presenting slight differences in the magnitude of the expression of immune response transcripts and REL/NFκB transcription factors, which could be dependent on the type of BRCA1 germline mutation. Conclusion: this study reveals the molecular complexity of BRCA1 breast tumours, which are found to display similarities to sporadic tumours, and suggests possible prognostic implications.