Examining the pathogenesis of breast cancer using a novel agent-based model of mammary ductal epithelium dynamics

The study of the pathogenesis of breast cancer is challenged by the long time-course of the disease process and the multi-factorial nature of generating oncogenic insults. The characterization of the longitudinal pathogenesis of malignant transformation from baseline normal breast duct epithelial dynamics may provide vital insight into the cascading systems failure that leads to breast cancer. To this end, extensive information on the baseline behavior of normal mammary epithelium and breast cancer oncogenesis was integrated into a computational model termed the Ductal Epithelium Agent-Based Model (DEABM). The DEABM is composed of computational agents that behave according to rules established from published cellular and molecular mechanisms concerning breast duct epithelial dynamics and oncogenesis. The DEABM implements DNA damage and repair, cell division, genetic inheritance and simulates the local tissue environment with hormone excretion and receptor signaling. Unrepaired DNA damage impacts the integrity of the genome within individual cells, including a set of eight representative oncogenes and tumor suppressors previously implicated in breast cancer, with subsequent consequences on successive generations of cells. The DEABM reproduced cellular population dynamics seen during the menstrual cycle and pregnancy, and demonstrated the oncogenic effect of known genetic factors associated with breast cancer, namely TP53 and Myc, in simulations spanning ∼40 years of simulated time. Simulations comparing normal to BRCA1-mutant breast tissue demonstrated rates of invasive cancer development similar to published epidemiologic data with respect to both cumulative incidence over time and estrogen-receptor status. Investigation of the modeling of ERα-positive (ER+) tumorigenesis led to a novel hypothesis implicating the transcription factor and tumor suppressor RUNX3. These data suggest that the DEABM can serve as a potentially valuable framework to augment the traditional investigatory workflow for future hypothesis generation and testing of the mechanisms of breast cancer oncogenesis.