Identification of proteins and signalling pathways involved in neoadjuvant chemotherapy responsiveness of breast tumours using proteomics

Retrospective studies on male breast cancer (MBC) have suff ered from small numbers of cases available from any one centre; thus a signifi cant problem in eff ectively studying this disease is accruing suffi ciently large numbers to allow comparative analysis of biomarkers associated with response. Using a coordinated multicentre approach, we present the fi rst large-scale study to address the relevance of the expression of hormone receptors in MBC and female breast cancer (FBC) using immunohistochemistry combined with a novel bioinformatics approach. Following ethical approval, 523 archival blocks (260 MBCs and 263 matched FBCs) were obtained retrospectively. Tissue microarrays were constructed and sections stained for ER, ER1, ER2, ER5, total PR, PRA, PRB and AR and typed using CK5/6, CK14, CK18 and CK19 by immunohistochemistry. Following scoring, a range of ordination techniques were conducted on the datasets including hierarchical clustering and principal component analysis (PCA) to determine the diff erential nature of infl uences and interactions between MBC and FBC. Luminal A subgroup (ER + and/or PR + , HER2 -) was the most common phenotype in both sexes. Luminal B (ER + and/or PR + , HER2 + ) was not seen in males, while basal-like tumours (ER -, PR -, HER2 -, CK5/6 + ) were infrequent in both. Hierarchical clustering revealed common clusters between MBC and FBC including total PR-PRA-PRB and ER1/2 clusters. ER occurred on distinct clusters between males and females. AR, ER1, ER2 and ER5 all existed on the same cluster but with a diff erent substructure, particularly around the positioning of AR. ER associated with this cluster in the male but not the female group. PCA confi rmed that in both groups strong infl uences came from PR-PRA-PRB. In MBC strong infl uences additionally came from AR and ER1, ER2 and ER5, whereas in FBC strong infl uences came from ER alone. Our data support the hypothesis that breast cancer is biologically diff erent in male and females, which could have implications for therapy.