Abstract
For the diagnosis and treatment of cancers, it is often assumed that all cells in a tumor are identical. However, solid tumors are composed of cells that differ in cell-type, genotype, and phenotype. Individual cancer cells in tumors regulate their behavior in response to complex internal and external cues. Together, these differences result in heterogeneous cancer cell states that influence tumor growth, metastatic progression, and treatment response. Characterizing the nature and prevalence of heterogeneous cancer cell states is fundamental to understanding why patients diagnosed with the same disease often have variable outcomes.
In this dissertation, we present experimental and bioinformatics approaches to measure heterogenous cancer cell states in breast and lung carcinomas. We coupled laser capture microdissection with sequencing measurements to obtain transcriptomic data from groups of 10 cancer cells in their native context within tumors. This profiling method has improved measurement sensitivity compared to existing single-cell transcriptomic methods, enabling us to deeply interrogate cancer cell transcriptomes. Analyzing 10-cell transcriptomes with an abundance-based dispersion metric, we identified heterogeneously expressed genes that represent different cancer cell states.
To identify early differences between cells that may influence patient outcomes and treatment responses, we profiled five biopsy samples from patients with luminal breast cancer. We detected thousands of heterogeneously expressed genes in individual tumors that comprise many pathways relating to proliferation, immune response, and stress tolerance. Moreover, we identified a recurrent set of genes that are heterogeneously expressed in multiple breast tumors. Genes in this set suggest that breast cancer cells sporadically activate pathways that are known to drive other types of cancer.
To systematically measure the influence of heterotypic interactions on cancer cells, we profiled 3D cell culture and murine models of small cell lung cancer (SCLC). We profiled SCLC cells in isolated 3D cultures and metastatic liver colonies to decode the influence of heterogeneous tumor microenvironments on cancer cell states. We observed a shift in the plasticity of SCLC cells upon liver colonization, and identified an expanded set of heterogenous states that expressed markers of multiple cell-types.
In this dissertation, we interrogated heterogenous cancer cell states within isolated cells, solid tumors, and metastases. The findings presented here provide novel insight into the transcriptional landscapes of breast cancer and lung cancer cells, towards the goal of understanding differential outcomes for patients with these diseases.
