Identification of genetic variants associated with tacrolimus metabolism in kidney transplant recipients by extreme phenotype sampling and next generation sequencing

Abstract

An extreme phenotype sampling (EPS) model with targeted next-generation sequencing (NGS) identified genetic variants associated with tacrolimus (Tac) metabolism in subjects from the Deterioration of Kidney Allograft Function (DeKAF) Genomics cohort which included 1,442 European Americans (EA) and 345 African Americans (AA). This study included 48 subjects separated into 4 groups of 12 (AA high, AA low, EA high, EA low). Groups were selected by the extreme phenotype of dose-normalized Tac trough concentrations after adjusting for common genetic variants and clinical factors. NGS spanned > 3 Mb of 28 genes and identified 18,661 genetic variants (3961 previously unknown). A group of 125 deleterious variants, by SIFT analysis, were associated with Tac troughs in EAs (burden test, p = 0.008), CYB5R2 was associated with Tac troughs in AAs (SKAT, p = 0.00079). In CYB5R2, rs61733057 (increased allele frequency in AAs) was predicted to disrupt protein function by SIFT and PolyPhen2 analysis. The variants merit further validation.

Figure 1:. Extreme Phenotype Sampling (EPS) Model to Detect Genetic Variants Associated with Tacrolimus Metabolism from African American (AA) or European American (EA) Kidney Transplant Recipients.

The graphs represent the mean dose-normalized Tac troughs on the y-axis and the distribution of subjects on the x-axis. The 12 recipients with highest or lowest Tac troughs, after adjusting for clinical variables and common genetic variants, from each group were selected for targeted next generation sequencing (NGS). A. The model used to select AA kidney transplant recipients was adjusted for genetic variants CYP3A5 *3, *6, and *7. The 12 AA subjects with the highest (3.5%) or 12 with the lowest (3.5%) Tac troughs were used for NGS from a cohort of 345 total subjects. B. The model used to select EA kidney transplant recipients was adjusted for genetic variants CYP3A5 *3 and CYP3A4 *22. The 12 EA subjects with the highest (0.8%) or 12 with the lowest (0.8%) dose-normalized Tac troughs were used for NGS from a cohort of 1,443 total subjects.

Figure 2:. Dose Normalized Tac Troughs of Subjects from Extreme Phenotype Sampling (EPS) Model used for Next Generation Sequencing (NGS).

The figure shows natural log transformed Tac dose-normalized troughs over time, in high and low AA or EA Tac groups. Data lines represent smoothed conditional means and gray areas represent 95% confidence intervals. The 12 EA subjects with the highest (0.8%) or 12 with the lowest (0.8%) Tac troughs were used for NGS from a cohort of 1,443 total subjects. The 12 EA subjects with the highest (3.5%) or 12 with the lowest (3.5%) Tac troughs were used for NGS from a cohort of 345 total subjects after adjustment for known genotypes and clinical factors.

Figure 3:. Variant Effect Predictor (VEP) results based on genetic variants identified

A. Predicted consequences of the 18,661 genetic variants identified in this sequencing study. B. Predicted gene expression consequences from coding sequences in the VEP analysis.

Figure 4:. SIFT and PolyPhen2 Results of all 18,661 variants in a Venn diagram

SIFT and PolyPhen2 are bioinformatics analytic tools that predict the affect specific genetic variants may have on protein function. Of the 18,661 variants, 125 were deleterious and 22 were deleterious with low confidence by SIFT while the remaining variants were tolerated. Polyphen2 analysis found 110 of the variants were probably damaging, 63 were possibly damaging while the remaining variants were benign to impacting protein.