Chemotherapeutic drug susceptibility associated SNPs are enriched in expression quantitative trait loci

Abstract

Pharmacogenomics has employed candidate gene studies and, more recently, genome-wide association studies (GWAS) in efforts to identify loci associated with drug response and/or toxicity. The advantage of GWAS is the simultaneous, unbiased testing of millions of SNPs; the challenge is that functional information is absent for the vast majority of loci that are implicated. In the present study, we systematically evaluated SNPs associated with chemotherapeutic agent-induced cytotoxicity for six different anticancer agents and evaluated whether these SNPs were disproportionately likely to be within a functional class such as coding (consisting of missense, nonsense, or frameshift polymorphisms), noncoding (such as 3'UTRs or splice sites), or expression quantitative trait loci (eQTLs; indicating that a SNP genotype is associated with the transcript abundance level of a gene). We found that the chemotherapeutic drug susceptibility-associated SNPs are more likely to be eQTLs, and, in fact, more likely to be associated with the transcriptional expression level of multiple genes (n > or = 10) as potential master regulators, than a random set of SNPs in the genome, conditional on minor allele frequency. Furthermore, we observed that this enrichment compared with random expectation is not present for other traditionally important coding and noncoding SNP functional categories. This research therefore has significant implications as a general approach for the identification of genetic predictors of drug response and provides important insights into the likely function of SNPs identified in GWAS analysis of pharmacologic studies.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.

Functional annotation of chemotherapeutic drug susceptibility–associated SNPs. SNPs may fall into more than one functional class. The traditional approach prioritizes coding SNPs (e.g., nonsynonymous polymorphisms) within a gene whereas the eQTL-based approach may yield a set of SNPs throughout the genome that are associated with transcript abundance level for a gene and presumably regulatory, as trans- or cis-acting SNPs. The embedded table indicates the total number of SNPs associated with the IC50 value of each drug at a P ≤ 10−4 level.

Fig. 2.

Distribution of the count of missense polymorphisms in 1,000 simulations, each matching the MAF distribution of the chemotherapeutic drug susceptibility–associated SNPs (chemotherapeutic susceptibility P value threshold of 10−4). The black dot is the observed missense polymorphism count in the chemotherapeutic drug susceptibility–associated SNPs.

Fig. 3.

Distribution of eQTL (expression P value threshold of 10−4) count in 1,000 simulations, each matching the MAF distribution of the chemotherapeutic drug susceptibility–associated SNPs (chemotherapeutic susceptibility P value threshold of 10−4). The black dot is the observed eQTL count in the chemotherapeutic drug susceptibility–associated SNPs.

Fig. 4.

Distribution of master regulator count in 1,000 simulations. A master regulator is defined to be a SNP that predicts the expression of at least 10 genes (eQTL P value threshold of 10−4). The black dot is the observed master regulator count in the chemotherapeutic drug susceptibility–associated SNPs.

Fig. 5.

An example of a master regulator. The genotype of SNP rs1649942, which is associated with sensitivity (cytotoxicity P value threshold of 10−4) for both carboplatin and cisplatin, is associated with the transcriptional expression level of 39 genes throughout the genome (eQTL P value threshold of 10−4). The distance between nodes reflects the strength of the association between the SNP genotype and gene expression levels: the closer the gene is to the SNP, the higher the correlation between the SNP genotype and gene expression.