Background

Pharmacoepigenomics aims to identify genomic regions of epigenetic variability that associates with patient survival, toxicity and/or drug response. DNA methylation variation can act as a mechanism for phenotypic variability. We hypothesize that inter-individual DNA methylation variation detected in blood mononuclear cells (PBMCs) may provide clinical outcome information following treatment of cancer patients.

Method

In a three phase study, we employed a discovery set to identify novel loci demonstrating variable methylation in PBMC DNA at presentation of 29 responding and 30 non-responding ovarian cancer patients enrolled in the SCOTROC1 phase III trial using differential methylation hybridisation of a customised microarray representing 874 genes of interest. As a test set, we next examined PBMC DNA from 474 SCOTROC1 patients at 96 selected loci, to determine whether individual CpG loci at significant genes identified were associated with patient survival, toxicity and/or drug response using Illumina's Veracode array. Of the 21 genes significant in both arrays, we ran a validation phase on DNA derived from another independent set of 256 SCOTROC1 patients at presentation also using the Veracode array and further validation of selected loci using bisulfite pyrosequencing.

Results

In the discovery set, 991/14485 gene regions had significant differential methylation (q<0.05) between poor and good response groups. 29 loci from 96 examined were found to be significantly associated to response, toxicity, or patient survival in the test set. In the validation set, one locus was found to be significant after adjustment for potential confounders showing association between DNA methylation and progression free survival (p= 0.01), and two loci with CA125 response (p= 0.05 and p=0.05)

Conclusion

In a three phase study we have demonstrated that DNA methylation may identify markers of clinical outcome and toxicity in ovarian cancer patients.