Joint differentiation-state plasticity and genetic diversity modeling to predict response and improve efficacy of drug combination therapy

点击翻译按钮获取中文摘要

PROJECT SUMMARY Each tumor faces a common set of obstacles arising from tumor-intrinsic dynamics, microenvironment signals and therapeutic interventions. The “canonical” view of tumorigenesis is a landscape within which each tumor navigates a unique path by combining various tactics to overcome these obstacles. A number of the early breakthroughs in cancer treatment directly resulted from coarse demarcations of these paths into distinct subtypes based on “genomic landmarks” such as the BCR-ABL fusion in chronic myeloid leukemia. However, in acute myeloid leukemia (AML), the majority of patients lack actionable mutations and nearly half of AML patients present with normal karyotype. For certain AML driver mutations, molecularly targeted drugs have shown initial clinical promise but not long-term durability; resistance remains a key challenge for targeted therapies, prompting a shift to combination strategies. Intratumor heterogeneity and plasticity as well as the tumor microenvironment play a fundamental role in shaping response to therapy and development of resistance. Recent findings indicate that targeted agents have different efficacies on AML cells at distinct stages of myeloid differentiation and lead to different resistance mechanisms. This observation opens the potential to improve clinical utility by matching targeted drugs to cell phenotypes. A specific, yet clinically important case of this observation is the different efficacies of targeted BCL2 inhibitor venetoclax and MEK inhibitors in AML cells at distinct stages of myeloid differentiation. In our preliminary analysis of the Beat AML cohort, we observe a concordance between this phenotype and the downstream transcriptional impact of BCL2 and MCL1 proteins suggesting a molecular link to a previously described venetoclax resistance mechanism. We hypothesize that differentiation-state plasticity is a general modulator of short-term resistance to a wide range of targeted compounds. A mechanistic understanding of cell differentiation-state plasticity in conjunction with genome-driven cellular processes can improve stratification of patients to therapies and suggest more tailored combinatorial treatments. To that end, I will use the largest assembled dataset of matched multi-omic assembled profiles, drug screens, and clinical information for primary leukemia samples from the Beat AML project at the Knight Cancer Institute. My objective is to perform systematic analysis of bulk genomic, functional ex-vivo assays (drug and CRISPR), conjointly with single-cell transcriptomic, ex-vivo CyTOF and single-cell drug response data in AML primary samples to understand the effects of tumor heterogeneity and plasticity on drug response at an unprecedented detail. The anticipated outcomes are 1) improved patient outcomes achieved by accurately matching therapies to individual patients and 2) nomination of new approaches to mitigate plasticity-based drug resistance.