Characterization of high-grade serous ovarian cancer subtypes via single-cell profiling

点击翻译按钮获取中文摘要

High-grade serous ovarian cancer (HGSOC) subtypes have been identified across multiple studies; however, the biologic basis of these subtypes remains poorly understood. The central hypothesis of this proposal is that differences in the cellular composition of HGSOC tumors drives the expression patterns that characterize at least some of the previously described HGSOC subtypes. New technologies that barcode antibodies and transcripts from individual cells before sequencing can characterize gene expression at single cell resolution and detect cell types, which allows the central hypothesis to be directly tested. These combined advances lay the groundwork to identify the basis, in terms of cell composition and pathway expression, of HGSOC subtypes through two aims. Aim 1: Characterize transcriptomes and selected proteins at single cell resolution, and deconvolve existing tumor gene expression data. Single cell RNA and surface protein abundances will be measured at the single cell level for high-grade serous ovarian cancers, unsupervised analysis will be used to identify cell populations, cell surface proteins will be analyzed to characterize the immune compartment, cell-type marker genes will be defined, and marker genes will be used deconvolve matched bulk RNA-seq samples. This will allow existing data from larger cohorts to be deconvolved allowing survival analyses to be performed on tumors stratified by cell composition. Aim 2: Characterize the transcriptomic profile of cancer cells within HGSOC tumors to identify pathways that are variably expressed within cancer cells. Gene expression within cancer cells will be measured using two complementary approaches: (i) orthotopic patient derived xenografts (PDXs) and (ii) single cell RNAseq. For each pathway, an enrichment score will be generated and pathways with expression levels that vary substantially across the cohort will be identified. Combining the expression levels of genes within variable pathways with cancer cell fraction estimates from existing datasets will enable inference of the extent to which these variable pathways differ between reported subtypes after controlling for cancer cell abundances. The proposal is expected to result in two primary outcomes: 1) an understanding of the extent to which cell composition and pathway expression contribute to HGSOC gene expression subtypes; and 2) estimates of the proportions of cell types in existing studies with public gene expression data. A short-term impact is expected through improved survival predictors of HGSOC subtypes based on variation identified from cell composition and pathway expression and the work is expected to be impactful in the longer-term because determining the biologic basis of subtypes is a key step towards developing treatments that target their specific vulnerabilities.