Whether cancer follows a hierarchical (linear) or stochastic (random) model of differentiation has critical clinical implications, yet remains highly controversial. The hierarchical model mirrors normal stem cell biology with cancer stem-like cells (CSC) following an established pattern of asymmetric divisions to give rise to distinct progeny. In this model only CSC can initiate recurrent disease. In the stochastic model, any given tumor cell can asymmetrically divide to give rise to the other cells of the tumor and initiate recurrent disease. The key to determining the pattern of differentiation is characterizing asymmetric division potential. However the asymmetric division potential of primary human cancer cells has not been directly interrogated due to a lack of technology. We hypothesize that cancer cells will follow a hierarchical differentiation pathway, but rare stochastic dedifferentiation events occur allowing some cells to obtain a CSC state. We further hypothesize that factors regulating the asymmetric division of cancer cells, both hierarchical and stochastic, will significantly impact cancer growth and represent therapeutic targets. To address the technology deficiency, we propose SA1: To develop a high-throughput single-cell microfluidic culture device with selective live-cell retrieval. A highly parallel, compact, 1024 microwell design will allow efficient single cell capture, growth, and characterization of asymmetric division of primary human cancer cells. Automated data analysis and selective single-cell retrieval capacity will allow for high-throughput functional and molecular analysis of common and rare events. In order to validate this important new technology we will use ovarian cancer as a model system. We propose: SA2: To directly interrogate the asymmetric division potential of primary ovarian cancer cells. Our preliminary data indicates that ovarian cancer follows a hierarchical differentiation model with rare (~1/3000) putative stochastic ?dedifferentiation? events wherein a progenitor cell gives rise to putative CSC. We will expand our studies of the asymmetric division potential to assess inter-patient variability and determine the impact of environmental stresses on dedifferentiation rates. Critically, using selective cell retrieval we can evaluate the biologic characteristics (chemotherapy resistance, tumor initiation capacity) of single cells derived following an asymmetric division to confirm the biologic implications of differentiation and dedifferentiation events. Finally, we propose SA3: To identify regulators of cancer cell differentiation and drivers of dedifferentiation. In parallel to the functional studies in SA2, mother and daughter cells from asymmetric divisions will be isolated and expression of distinct stem cell related genes assessed via single cell qRT-PCR. Factors differentially expressed between mother and daughter cells will then be evaluated for their impact on CSC growth and differentiation. To identify molecular drivers of the rare dedifferentiation events, we will perform RNASeq comparing clones derived from hierarchical differentiation vs. stochastic dedifferentiation events. Putative molecular drivers of dedifferentiation will then be confirmed both in vitro and in vivo.
Whether cancer follows a hierarchical (linear) or stochastic (random) model of differentiation has critical clinical implications, yet remains highly controversial. This controversy relates to an inability to perform high- throughput analysis of validated single human cancer cell asymmetric division potential and an inability to then extract phenotyped cells for functional characterization. We propose to develop a novel a high-throghput microfluidic culture system with selective single-cell retrieval capacity which will overcome the current technical challenges to studying cellular differentiation. We will confirm the utility of the system by defining the ovarian cancer cell differentiation and identifying and confirming the potential of regulators of ovarian cancer cell differentiation as therapeutic targets.
