Integrative single cell analyses of inflammation-derived circulating hybrid cells to identify aggressive disease

PROJECT SUMMARY The lifetime risk for acquiring colorectal cancer (CRC) is 7%, with an astounding rate of disease recurrence in 32% of newly diagnosed patients after their “successful” treatment. Patient with recurrent disease have a dismal 14.3% five-year survival. Lack of effective biomarkers hampers early detection of pre-metastatic disease,

impacting overall survival from CRC. We identified a promising disseminated tumor cell—a product of macrophage (MФ) and cancer cell fusion—that harbors genotypic and phenotypic features of both cells of origin. Detectable along the metastatic cascade, hybrid cells can initiate tumor growth, migrate in response to MФ

receptor-ligand chemotaxis, and seed metastatic sites. In peripheral blood, hybrids, named circulating hybrid cells (CHCs) outnumber conventionally defined circulating tumor cells (CTCs) in CRC patients, overcoming the sensitivity of CTC—a primary barrier—to usage as a biomarker for disease. CHCs are phenotypically diverse

and reflect protein expression of the primary tumor. Based on these exciting findings, we propose that hybrid cells subpopulations harbor discrete phenotypes of pre-metastatic cells that can be identified and defined using single cell image-based phenotyping through multiplexed imaging and multimodal integration with –omics. To

this end, we will analyze CHCs derived from early stage and metastatic tumors for image-based phenotyping with single cell gene expression. Utilizing quantitative and advanced image analytics including deep learning approach for image-based cell profiling, we will define inter/subcellular spatial features in single cells to identify

new subpopulations and differentiate discrete phenotypic populations associated with metastatic signatures. In addition, the application of both imaging and genomic technologies to the same specimen independently measures highly dimensional, yet non-orthogonal, sets of cellular features. Multimodal integration of imaging

and single cell data will quantify systems-level biological functions of cellular subpopulation and enhance imaging biomarker panel to gain biomarker specificity and sensitivity for validation in a discrete CRC patient cohort. Our overall goal is to develop a novel tumor biomarker, based upon CHC phenotyping and –omics analyses that can

be used to provide new quantitative insights and develop machine-driven prediction with superior accuracy for identifying risk of metastases in CRC patients to ultimately impact survival.