Abstract 1872: Identification of novel single-nucleotide polymorphisms associated with anticancer drug effects using the SNPChip384 platform

Abstract

Abstract The identification of genetic factors associated with anticancer drug efficacy is an important challenge for optimising chemotherapy and developing personalised treatments. One of the difficulties lies in the absence of preclinical models for identifying reliable tracks that could be translated to the clinics. We have used in vitro models to identify relationships between the presence of a given polymorphic variation and the cytotoxicity of anticancer drugs: the National Cancer Institute (NCI) and the Japanese Foundation for Cancer Research (JFCR) cell line panels, which had been used for the primary screening of thousands of potential anticancer drugs. In order to discover new SNPs of potential interest, high-throughput screening could be an option, but genome-wide studies cannot be used because of the relatively small number of cell lines (105) and the high false-discovery rate that would result. As an intermediate option, we have established a tool (SNPChip384) allowing the simultaneous determination of 384 SNPs in a single DNA sample. The SNPs have been chosen among DNA repair genes, phase I and II drug metabolism genes, drug transport genes and genes involved in oncogenesis, signal transduction and oxidative stress. Genotyping of the 60 cell lines from the NCI panel and the 45 cell lines from the JFCR panel was carried out using the SNPChip384. The concordance between this multiplexed approach and the individual previous determinations was around 95% (0-3 discordances in the 60 cell lines of the NCI panel and 0-2 discordances in the 45 cell lines of the JFCR panel). We first confirmed the significant relationships we had already observed and published concerning MDM2, MTHFR, CYP1B1, ATIC and ERCC2 gene polymorphisms. In addition, we identified novel SNPs involved in the in vitro cytotoxicity of anticancer drugs. The NCI panel was used for establishing these associations and the JFCR panel for validating them independently. The Benjamini-Hochberg correction for multiple testing was used in order to minimize the false-discovery rate. In the NCI panel, ten gene polymorphisms successfully passed this stringent filter (the p values indicated are those obtained after correction): AHR R554K was significantly associated with irofulven cytotoxicity (p=0.0001); APC Y486Y with SN-38 cytotoxicity (p=0.007); ATP7B K832R with carmustine, busulfan and melphalan cytotoxicity (p=0.004 to 0.036); BRCA1 E1038G with geldanamycin cytotoxicity (p=0.01); CASP9 Q221R with the cytotoxicity of a taxol analogue (p=0.03); CDC25C R70C with vinblastine cytotoxicity (p=0.04); CDH1 with spiromustine cytotoxicity (p=0.02); CYP2E1 F421F with irofulven cytotoxicity (p=0.007); PARP4 G1280R with thioguanine cytotoxicity (p=0.04); TLR4 D299G with doxorubicin cytotoxicity (p=0.03). These associations open new tracks for identifying novel determinants of anticancer drug activity. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 1872. doi:1538-7445.AM2012-1872