Despite improved treatment strategies, inherent and acquired resistance to HER2-targeted therapies is common and inevitably lead to disease progression and cancer-related deaths. PIK3CA mutations occur in ~40% of HER2+ tumors and have been associated with resistance to treatment. To explore intratumor heterogeneity for HER2 and PIK3CA mutation in HER2+ tumors and the potential role of this in treatment resistance, we have developed a technique, STAR-FISH (Specific To Allele pcR FISH) that allows for the combined detection of point mutations and copy number gain in intact tissue samples at the single cell level. By applying STAR-FISH to HER2+ tumors before and after neoadjuvant therapy, we found that PIK3CA mutation and HER2 amplification do not always co-occur in the same cell but can be present in distinct subpopulations within tumors. Post-treatment residual tumors have higher fraction of mutant PIK3CA cells with a concomitant decrease in the relative ratio of cells with HER2 amplification. Patients with significant differene in cell type diversity between pre and post-treatment samples have shorter disease-specific survival than those with no change. Based on our preliminary data, we hypothesize that 1) not all somatic mutations that have been reported to occur in a given tumor are present in the same tumor cell but some may represent genetically divergent clones within tumors, (2) the combination of mutations within the same cell and within the same tumor is not random, but reflect cooperation among pathways and among clones, and (3) the presence of a combination of mutations within the same cell versus in different clones in the same tumor will have different functional and clinical consequences. We propose three specific aims to test these hypotheses:
Aim 1. Analyze intratumor subclonal and cellular genetic diversity in HER2+ breast tumors.
Aim 2. Develop and characterize experimental models of intratumor heterogeneity in HER2+ breast cancer.
Aim 3. Develop mathematical models to infer the evolution of HER2+ tumors during treatment. Our goal is to translate our findings into future clinical trials.
Intratumor heterogeneity is a major obstacle in the understanding and treatment of cancer, yet it can also be used to predict tumor evolution. The application aims to develop individualized treatment strategies using a combination of predictive mathematical modeling and experimental approaches both in models of HER2+ breast cancer and in clinical samples.
