Paper

Cross-validation of technologies for genotyping CYP2D6 and CYP2C19

Wang, Yabing, Bahareh Behroozi Asl, Mors, Ole, Bousman, Chad, Hague, Joshua, Hu, Xiuying, Carr, Michael, Hauser, Joanna, Maier, Wolfgang, Henigsberg, Neven, Yavorskyy, Vasyl, Dernovsek, Mojca Z, Pfeffer, Gerald, Aitchison, Katherine J, Martens, Kristina, Souery, Daniel, Rietschel, Marcella, Beatriz Carvalho Henriques, Cattaneo, Annamaria, Buchner, Amanda

2020

Overview

Abstract Background CYP2D6 and CYP2C19 are cytochrome P450 enzymes involved in the metabolism of many medications from multiple therapeutic classes. Associations between patterns of variants (known as haplotypes) in the genes encoding them (CYP2D6 and CYP2C19) and enzyme activities are well described. The genes in fact comprise 21% of biomarkers in drug labels. Despite this, genotyping is not common, partly attributable to its challenging nature (CYP2D6 having >100 haplotypes, including those with sequence from an adjacent pseudogene, and gene duplications). We cross-validated different methodologies for identifying haplotypes in these genes against each other. Methods Ninety-two samples with a variety of CYP2D6 and CYP2C19 genotypes according to prior AmpliChip CYP450 and TaqMan CYP2C19*17 data were selected from the Genome-based therapeutic drugs for depression (GENDEP) study. Genotyping was performed with TaqMan copy number variant (CNV) and single nucleotide variant (SNV) analysis, the next generation sequencing-based Ion S5 AmpliSeq Pharmacogenomics Panel, PharmacoScan, long-range polymerase chain reaction (L-PCR) followed by amplicon analysis, and Agena for CYP2C19. Variant pattern to haplotype translation was automated. Results The inter-platform concordance for CYP2C19 was high (up to 100% for available data). For CYP2D6, the IonS5-PharmacoScan concordance was 94% for a range of variants tested apart from those with at least one extra copy of a CYP2D6 gene (occurring at a frequency of 3.8%, 33/853), or those with substantial sequence derived from pseudogene, known as hybrids (3%, 26/853). Conclusions Inter-platform concordance for CYP2C19 was high, and, moreover, the Ion S5 and PharmacoScan data were 100% concordant with that from a TaqMan CYP2C19*2 assay. We have also demonstrated feasibility of using an NGS platform for genotyping CYP2D6 and CYP2C19, with automated data interpretation methodology. This points the way to a method of making CYP2D6 and CYP2C19 genotyping more readily accessible. Competing Interest Statement N.H. has participated in research supported by CSF project No. IP-09-2014-2979. D.S. has received grant/research support from Janssen and Lundbeck, and has served as a consultant or on advisory boards for Janssen and Lundbeck. C.A.B. is a member of the Clinical Pharmacogenetics Implementation Consortium and the Pharmacogene Variation Consortium; he has also participated in collaborative research with employees of MyDNA (a pharmacogenetic testing company) but does not have equity, stocks, or options in this company or any other pharmacogenetic companies. K.J.A. is a member of the Clinical Pharmacogenetics Implementation Consortium and the Pharmacogene Variation Consortium, has received two research grants in the last two years from Janssen Inc., Canada (fellowship grants for trainees) and provided consultancy services (unpaid) for HLS Therapeutics. All other authors have nothing to disclose. Footnotes * Updated authors, disclosures, and updates to the paper including added data to tables and supplemental files.

• Publisher: Cold Spring Harbor Laboratory Press,Cold Spring Harbor Laboratory
• Subject: Automation / Copy number / Cost control / CYP2D6 protein / Cytochrome P450 / DNA probes / Enzymatic activity / Gene conversion / Genetics / Genomes / Genotypes / Genotyping / Next-generation sequencing / Pharmacogenomics / Polymerase chain reaction