Stratification of breast cancer patients : a proteomic approach

• Main Author: Alburai'Si, Kholoud Mubarak
• Other Authors: Pinder, Sarah Elizabeth ; Parker, Peter Joseph Jacques
• Published: King's College London (University of London) 2016
• Subjects: 616.99
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.689186

Breast cancer represents a heterogeneous collection of different diseases characterized by different pathological and biological features, clinical presentation, clinical behaviour, response to treatment and outcome. In current practice, pathological diagnosis and classification of breast cancer is based mainly on well-established traditional morphologic features. However, morphological features alone do not adequately reveal the molecular heterogeneity and complexity of breast cancer. Still, there are relatively few biomarkers widely used in prognostication in invasive breast cancer and in predicting response to targeted therapies, and even fewer of value in the clinical management of the pre-invasive disease of ductal carcinoma in situ (DCIS). There is therefore an unmet need for biomarkers for better classification, better prediction of prognosis and of prediction of response to therapy for both invasive breast carcinoma and DCIS. The co-expression of a HER2/HER3 combination results in more aggressive tumour growth and is associated with endocrine and chemotherapy resistance, driven not simply by receptor expression but also by signalling via the receptors dimers. Therefore, methods which directly query signalling pathway activation in breast cancer specimens are anticipated to provide important insights into the molecular ‘‘logic’’ that distinguishes cancer from normal tissues and potentially to have an important impact on personalized intervention strategies. The aim of this thesis has been directed at evaluating candidate biomarkers in breast cancer. This has been targeted at examining attributes associated with known functional properties of candidate drivers of disease or resistance to treatment rather than those traditionally based on altered expression of these biomarkers. Specifically, the work was directed at the HER1-3 members of the EGFR family of growth factor receptor in breast cancer. In this project I have in part developed, tested and evaluated two methods, which have the ability to detect protein-protein complexes at a single molecule level and thus allow the study of signalling pathways in situ. The first method is an in-house coincidence detection technology created from two recombinant fusion proteins and the second is a commercially available proximity ligation assay (PLA) method. Both approaches were able to detect the target proteins with high sensitivity and specificity, however the proximity ligation assay was subsequently used here to assess the protein complexes and activation status of the EGFR family in breast cancer patient’s samples. The patient study cohort is derived from a consecutive series of approximately 293 cases of primary operable invasive breast cancers obtained from the Guy’s and St Thomas (King’s Health Partner’s) Breast Cancer Biobank archive presenting between 1990 and 1992. In these cases, 9 different biomarkers were studied for HER family expression level, dimer expression and activation status using proximity ligation assays (PLAs). The relationship between HER activation status, dimer expression and relapse free survival (RFS) was investigated and stratified multivariate regression analysis identified factors influencing patient prognosis. In conclusion, PLA successfully and reproducibly detected HER protein complexes and phosphorylation in vivo. A significant association was identified between high levels of phosphorylated HER2 and reduced recurrence-free survival (RFS) in invasive lobular carcinoma (p = 0.04, HR 0.99, 95% CI: 0.997-1.002). High levels of HER1/HER3 dimers were associated with reduced RFS in T1 (<2cm) breast cancer patients, (p = 0.02, HR 1.84, 95% CI: 1.08-3.13). Similarly, high levels of HER1/HER3 and HER2/HER3 dimers were associated with reduced RFS in breast cancer patients with N1 nodal status (p <0.0001, HR 1.84, 95% CI: 0.58-1.93) and (p <0.0001, HR 0.64, 95% CI: 0.45-0.90) respectively). Work in this thesis demonstrates that in situ detection of HER protein complexes and activation status can be monitored robustly and with specificity in clinical specimens, providing novel prognostic information. This technique was also applied successfully to assess the HER family in a smaller number of DCIS cases. This novel technique and approach could potentially be applied for patient stratification and assist in the selection of more individualized treatment options according to tumour molecular characteristics.