Postdoctoral Researcher University of Florida Gainesville, Florida, United States
Background: CYP2D6 is the most polymorphic drug-metabolizing enzyme and numerous CYP2D6 genetic variants are used to predict its activity. However, much of the inter-person variability in CYP2D6 remains unexplained. We previously found that rs16947 (core SNV delineating CYP2D6*2) increases the splice isoform CYP2D6ΔE6 and reduces CYP2D6 mRNA expression, which is offset in some individuals who also carry an enhancer variant rs5758550. In addition to CYP2D6ΔE6, CYP2D6ΔE3 is a splice isoform carrying an in-frame deletion that encodes a protein missing several active site residues, with likely altered enzymatic activity. Here, we set out to determine whether genetic variants are promoting the expression of CYP2D6∆E3 and if we could better predict CYP2D6 expression with a model that incorporates all CYP2D6 regulatory variants. Methods: In a 228-sample liver cohort we genotyped CYP2D6 SNPs and CNVs. CYP2D6ΔE3 and ΔE6 were measured via fragment analysis and RT-qPCR. The effects of rs1058164 on CYP2D6 splicing were confirmed using whole-gene transfection in cell culture. We tested three models to predict protein expression based on genotype and then associated the results with full-length CYP2D6 protein levels detected via Western Blot. Model 1 follows the CPIC guidelines, model 2 also includes rs5758550 and rs16947, and model 3 further adds rs1058164. Results: The rs1058164 G allele (present in the wild-type CYP2D6 haplotype) increases alternative splicing to generate CYP2D6∆E3 (1.4-2 fold) in the liver and in transfected cells. Model 3 provided the best fit for CYP2D6 expression and also explained more variability in CYP2D6 protein levels (59.2%) than either model 1 (35.6%) or model 2 (42.1%). Conclusion: Our results indicate that incorporating all three regulatory variants may improve the predictive value of CYP2D6 panels.
