The causes of the incredible variability in tumor growth are poorly understood, confound cancer treatments, and complicate the prognosis of patients with early disease. Theory predicts that the variability is caused by a number of stochastic factors. These stochastic factors include the random accumulation of different mutations, variation in the local environment of the tumor, and differences in properties of the cell of origin (including its level of differentiation, replication potential, and somatic alterations). Mouse models have been used to understand tumorigenesis and to characterize tumorigenic mutations within the natural environment; however, creating these models is technically challenging and tumor growth measurements are imprecise. To overcome these limitations, we previously developed an innovative new method to create thousands of tumors and accurately measure their growth in parallel via DNA barcoding and deep-sequencing (Tuba-seq). Strikingly, Tuba-seq uncovered that isogenic lung tumors within the same mouse will diverge in size by more than one thousand-fold after only a few months of growth. This was unexpected from current models of tumorigenesis. Here, I will characterize the emergence of tumor growth variability and the underlying forces that drive this variability using Tuba-seq. Because different stochastic forces predict different dynamics by which tumor growth variability emerges, I will track the growth of millions of Kras-initiated lung tumors in mice for one year (Aim 1). Next, because we previously found that loss of different tumor suppressors lead to different levels of growth variability, I will track the growth dynamics of Kras-initiated tumors with over twenty different combinations of secondary tumor suppressor loses over time (Aim 2). This will be possible by virtue of our previous work that paired Tuba-seq with CRISPR/Cas9-mediated inactivation of targeted tumor suppressor genes using a high-throughout, multiplexed pool. Lastly, I will transplant thousands of DNA barcoded tumor cells across mice and track their growth in experimental condition designed to uncover the relative contributions of the local tumor environment and finite replicative potential of the cell of origin to growth variability (Aim 3). Collectively, by characterizing the forces governing tumor growth variability, we will improve models of carcinogenesis, which will affect our understanding of the risk factors, genetics, and vulnerabilities of the disease. Additionally, this project will create unique datasets and extensive tumor samples that will be critical for my future independent work modeling tumorigenesis and characterizing the genomic and cellular events that drive tumor growth.
Because tumor growth is unpredictable, it is difficult to provide personalized patient treatments or predict the eventual outcomes of tumors detected at early stages. This project will measure the forces that make tumor growth unpredictable with unprecedented accuracy using a new technology that we developed. Our eventual goal is to develop a mathematical model that forecasts tumor outcomes from early stages.
