Identification of prognostic targets in breast cancer using immunohistochemical and in situ hybridisation methodologies

Abstract

Advances in our understanding of the molecular mechanisms underlying breast cancer are leading to improvements in its management. Many of the recent gene-based signatures developed for breast cancer have highlighted the importance of proliferative markers, leading to renewed interest in their role as prognostic and predictive agents. Similarly, the small non-coding RNAs, microRNAs (miRNAs), have demonstrated potential as novel breast cancer biomarkers.

The aims of this work were to identify prognostic targets in breast cancer and to evaluate expression of the miRNA let-7a in tumour tissue with the aim of exploring its role as a potential diagnostic target. Immunohistochemical and in situ hybridisation (ISH) techniques were used to accomplish these aims.

Firstly, a consecutive series of 666 invasive breast cancers was characterised on tissue microarray using an immunohistochemical panel. Clinicopathological data and outcome status were collated for each case. On multivariate analysis, nodal status, ER and Ki67 were independent predictors of disease-free survival (DFS) (p value <0.0001, 0.05 and 0.17 respectively) and nodal status and ER were independent predictors of overall survival (OS) (p value <0.0001 for both). Tumours were categorised into molecular subtypes. It was found that DFS for molecular subtypes was best predicted by a classification system using ER and HER2 and OS by a classification system using ER, PR and HER2. The HER2-overexpressing subtype had the poorest outcome for both survival measures.

Secondly, immunohistochemistry for 10 of the genes analysed in the Oncotype DX Breast Cancer Assay was carried out on a series of 52 patients who had previously had Oncotype DX testing done as part of the TAILORx (Trial Assigning IndividuaLised Options for Treatment (Rx) trial. These variables, when analysed together with routinely-assessed pathological variables, could predict Oncotype DX and TAILORx low, intermediate and high risk categories in up to 85% and 76% of cases respectively. The most useful parameters were nuclear pleomorphism, PR, BAG1 and the proliferative markers Ki67 and survivin.

Finally, ISH was optimised for miRNA evaluation on formalin-fixed paraffin-embedded (FFPE) tissue. A pilot study evaluating let-7a by ISH was carried out on the tumour tissue of 15 patients who had previously had let-7a measured in circulating blood. Expression was shown to be generally downregulated in breast tumour cells compared to benign cells. No correlations were seen between expression in tumour tissue and blood levels.

This work supports the recent publications highlighting the prognostic importance of proliferative activity in breast cancer. In a small pilot project, the risk categories identified by the Oncotype DX assay could be predicted using a combination of traditional pathological variables and immunohistochemical assessment of the relevant proteins in the majority of cases. Finally, ISH was optimised and shown to be a feasible method for miRNA evaluation in FFPE tissue.