Project 2: Targeting Cooperative Phenotypes Common in Spatial Heterogeneity

点击翻译按钮获取中文摘要

ABSTRACT Recent studies in primary tumors have found a remarkable degree of intratumor heterogeneity, where a single tumor is comprised of a range of subclones exhibiting a diversity of phenotypes, including molecular profiles, proliferation capacity, and response to therapies. Although heterogeneity is now widely reported, few studies have investigated the heterogeneity of metastatic tumors at the end stage, despite the fact that metastatic cancer is estimated to be responsible for over 90% of cancer deaths. For breast and ovarian cancer, tumors that progress to metastasis are refractory to treatment. Therefore, there is a great need to determine the mechanisms by which subclonal diversity can affect the metastatic phenotype and underlie the difficulties in treatment. Studying metastatic tumors is difficult due to the challenges in collecting patient tissues. While primary tissues are typically obtained through biopsy, this is rarely performed for metastatic sites. To address this difficulty, we have developed both a rapid autopsy strategy where we collect fresh samples of metastatic tumors within hours of patient death, as well as collections of metastatic tumor biopsies in the clinical trial setting prior to and after drug treatment. These collections enable us to profile multiple metastatic sites and investigate the association between metastatic sites and subclonal evolution in an isogenic background. We propose to leverage this unique data set to investigate the relationship between evolution of tumor subclones during metastatic progression and the phenotypic profiles of these tumors. We hypothesize that, despite the diversity in their genetic mutation profiles, metastatic tumors exhibit clonal dynamics that ultimately leads to convergence on more common cooperative phenotypic networks, and that targeting the key dependencies within this network will lead to increased collapse of the metastatic tumor population. To investigate this, we will profile the tumors by whole genome sequencing, whole exome sequencing, and single cell RNA sequencing. This data, coupled with our newly developed algorithms for dissecting subclonal populations using tree reconstruction algorithms, for eliciting phenotypes from gene expression profiles using Bayesian statistics, and for simulating phenotypic evolution using mathematical models from ecology; will enable us to understand (Aim 1) the subclonal heterogeneity that underlies metastatic initiation and progression; (Aim 2) how cooperative functions evolve to a chemo-refractory signaling network, and therapeutic strategies to target it; and (Aim 3) how these dynamics are manifested human tumors in a clinical trial. Our investigations represent the first characterization of the clonal dynamics of a large multisite metastatic cohort, and will provide a new framework for understanding and treating end-stage tumors based on the evolution of cooperative phenotypes. We will develop these models on patient samples and test them in a unique clinical trial, ensuring the physiological, if not clinical, relevance of our findings.