Abstract P5-10-02: Investigating the response and adaptation of estrogen receptor (ER)+ tumors in response to decreased estrogen availability using an agent-based model of the mammary epithelium

Abstract

Abstract Introduction: Breast tumors consist of heterogenous cell populations with dominant features that can change over time. The development of resistance to previously effective therapies is an example of tumor evolution in response to selective pressure, where the differential composition of cells within a tumor may predispose to identifiable patterns of tumor growth and progression. We employ a previously developed computational model of ductal epithelial dynamics (the Duct Epithelial Agent Based Model or DEABM), able to reproduce known rates of tumorigenesis in both wild-type and BRCA1 populations, to investigate the putative population dynamics of ER+ tumors in response to decreased estrogen availability (DEA). Methods: The DEABM is composed of computational agents representing individual cells (luminal and myoepithelial cells, their progenitor cells and fibroblasts) behaving based on rules established from published cellular and molecular mechanisms. Cells implement DNA damage/repair, cell division in response to estrogen/local growth factors, and apoptosis. Unrepaired DNA damage impacts genomic integrity, including eight representative oncogenes and tumor suppressors affecting critical pathways previously implicated in breast cancer (BRCA1, P53, E-cadherin, RUNX3, Myc, TGF-beta, MMP-3 and Telomerase). ER+ tumors were generated in 40 year simulations of wild-type and BRCA1 populations and subjected to DEA (90% reduction in estrogen effect targeted to ER+ cells, which could generally represent endocrine therapy for breast cancer.) Cell populations were characterized by mutation profiles, ER status, and response to estrogen suppression. Simulations were continued until development of DEA resistance, identified by reversal of growth suppression, and the mutations present at that point were analyzed for change. Results: 3500 DEABM simulations over 40 years generated 69 wild-type ER+ tumors and 119 BRCA ER+ tumors. Of these 25% of wild-type ER+ tumors were sensitive to DEA vs. 35% of BRCA1 tumors, consistent with previously reported response rates. The percentage of ER+ cells was higher in initially sensitive tumors than in initially resistant tumors (84% vs. 66%, p = .001). Tumors initially resistant to DEA were more likely to carry mutations in the genes p53, E-cadherin and Myc (p = .001). Sensitive tumors acquired a significant increase in mutations during the interval between responsiveness to DEA and development of resistance (p = .001). Lastly, sensitive BRCA1 carriers were more likely to convert from ER+ to ER- status (p = .001). Discussion: The DEABM generated simulated breast tumors with intra-tumoral heterogeneity that demonstrate varied responsiveness to DEA, similar to what is observed clinically. These results suggest there may be definable patterns of tumor evolution in response to DEA that could potentially guide the development or sequencing of therapeutic regimens. Computational models such as the DEABM can aid in visualizing molecular data in a dynamic form and allow researchers to carry out “thought experiments” concerning tumor behavior and intervention effect. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P5-10-02.