Multi‐ancestral origin of intestinal tumors: Impact on growth, progression, and drug efficacy

BACKGROUND: Data are steadily accruing that demonstrate that intestinal tumors are frequently derived from multiple founding cells, resulting in tumors comprised of distinct ancestral clones that might cooperate or alternatively compete, thereby potentially impacting different phases of the disease process. AIM: We sought to determine whether tumors with a multi‐ancestral architecture involving at least two distinct clones show increased tumor number, growth, progression, or resistance to drug intervention. METHODS: Mice carrying the Min allele of Apc were generated that were mosaic with only a subset of cells in the intestinal epithelium expressing an activated form of PI3K, a key regulatory kinase affecting several important cellular processes. These cells were identifiable as they fluoresced green, whereas all other cells fluoresced red. RESULTS: Cell lineage tracing revealed that many intestinal tumors from our mouse model were derived from at least two founding cells, those expressing the activated PI3K (green) and those which did not (red). Heterotypic tumors with a multi‐ancestral architecture as evidenced by a mixture of green and red cells exhibited increased tumor growth and invasiveness. Clonal architecture also had an impact on tumor response to low‐dose aspirin. Aspirin treatment resulted in a greater reduction of heterotypic tumors derived from multiple founding cells as compared to tumors derived from a single founding cell. CONCLUSION: These data indicate that genetically distinct tumor‐founding cells can contribute to early intratumoral heterogeneity. The coevolution of the founding cells and their progeny enhances colon tumor progression and impacts the response to aspirin. These findings are important to a more complete understanding of tumorigenesis with consequences for several distinct models of tumor evolution. They also have practical implications to the clinic. Mouse models with heterogenous tumors are likely better for predicting drug efficacy as compared to models in which the tumors are highly homogeneous. Moreover, understanding how interactions among different populations in a single heterotypic tumor with a multi‐ancestral architecture impact response to a single agent and combination therapies are necessary to fully develop personalized medicine.