Abstract LB-246: Evaluating the role of admixture in cancer therapy via in vitro drug response and multivariate genome-wide associations

Abstract

Abstract It is well established that the majority of drugs/doses to combat diseases and disorders do not exhibit uniform effects for individuals. Indeed, the relationship between ethnicity, admixture or genetic ancestry and drug response is not fully understood. To explore this relationship, we used the lymphoblastoid cell line model to investigate dose-dependent drug response across a broad group of chemotherapeutic drugs spanning 7 drug families including nucleosides, tyrosine kinase inhibitors and DNA alkylating agents. Cell lines from 589 individuals were tested at 6 different concentrations for 28 chemotherapeutic drugs. Moreover, dose-dependent cytotoxic effects were measured for each drug with samples from two distinct cohorts: Hispanic and non-Hispanic/Caucasian. Genome-wide and phenotypic data from these individuals were used in univariate (IC50) and multivariate analyses to explore the relationship between ethnicity and drug response. Both univariate and multivariate models indicated that the majority of drugs tested showed highly significant correlations between self-reported ethnicity and drug responses. The software package ADMIXTURE was used to estimate genetic ancestry per individual. Using these estimates in lieu of self-reported ethnicity, we found highly significant correlations between percent admixture and drug response for the majority of drugs tested. Genome wide association analyses were carried out on each drug for each cohort independently as well as the combined data. From the association analyses, we were able to generate a number of interesting hypotheses on the role of particular genetic variants in variability of drug responses. 10 out of 28 drugs indicated significant associations in the Hispanic cohort. For the combined association analysis, 9 drugs had significant associations. Among the association findings, for the Hispanic cohort analysis, we were able to recapitulate previous association analysis work (in non-Hispanic/Caucasian) showing the role of variants in the gene, MGMT, and the susceptibility to temozolomide treatment. Remarkably, we were able to confirm the association result independently in a Hispanic and non-Hispanic cohort linking temozolomde response in Lymphoblastoid cells to variants in MGMT. To further understand the biological significance or the clinical impact of all of the association results, additional studies involving gene suppression techniques (e.g., knockdown models) and/or clinical trial gene candidate studies are warranted. Citation Format: John Jack, Tammy M. Havener, Howard L. McLeod, Alison A. Motsinger-Reif, Matthew Foster. Evaluating the role of admixture in cancer therapy via in vitro drug response and multivariate genome-wide associations. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr LB-246. doi:10.1158/1538-7445.AM2015-LB-246