Kidney allograft donor ABCB1 pharmacogenotype predicts long term vascular fibrous thickening and tubular atrophy – pharmacogenomic analysis of a prospective Australian cohort

Introduction: Tacrolimus is a critical backbone of transplant immunosuppression with a narrow therapeutic window. Adverse effects include graft fibrosis and nephrotoxicity in a dose dependent manner. Recipient genotypes affect tacrolimus pharmacokinetics; this is the first study of kidney donor pharmacogenomic impact.
Methods: The Australian Genomics of Chronic Allograft Dysfunction study is a prospective study at Westmead Hospital, with clinical data, biobanked blood, renal allograft biopsies, collected at 0/1/3/12 months post-transplant. Whole genome sequencing performed on Illumina NovaSeq X Plus platform. SNPs for CYP3A4, CYP3A5, ABCB1 extracted for donor/recipient. Biopsy scores extracted and mapped to Banff 2022 criteria. Variant filtering with minor allele frequency of 0.05. Linkage disequilibrium with pruning for r2>0.8. Cumulative link mixed model analyses were modelled in R using the ordinal package for: death-censored graft loss (DCGL), biopsy confirmed acute rejection (BPAR), change in fibrosis score over time (ci, ct, cv, ah, cg), with alpha of 0.05. Predictor genotypes: Donor, recipient, donor/recipient mismatch (matched, heterozygote, homozygote). Tacrolimus time above therapeutic range (TATR) was calculated using modified Rosendaal method – with time screened tacrolimus trough >12 ng/mL from 0-3 months, >10 ng/mL for 3-6 months, >8 ng/mL for 6 months onwards post-transplant.
Results: 216 donor/recipient pairs were included since study inception in 2012. 7 genes after filtering: ABCB1 (rs1045642, rs2032582, rs1128503), CYP3A5 (rs4646450, rs776746), CYP3A4 (rs2242480, rs35599367). 72 recipients (33.3%) were female. 39 (18%) had BPAR and 26 (12%) had DCGL during follow up.
Univariable analysis: recipient rs1128503 was associated with reduced progression of ah score (OR 0.61, 95%CI 0.39-0.93, p= 0.02); Donor rs2242480 with increased ah progression (OR 2.45, 95%CI 1.12-5.36, p=0.02). Exploratory signal (p<0.1) in other SNPs shown in Table 1.
There was no significant association between genotypes and DCGL or BPAR.
Multivariable correction for time post-transplant, tacrolimus TATR, diabetes, DSAs, HLA mismatch, recipient sex, and BPAR: donor rs1045642 and cv (OR 1.70, 95%CI 1.04-2.78, p = 0.036), recipient rs1128503 and ct (OR 0.57, 95%CI 0.32-0.99, p=0.046), recipient rs2032582 and ct (OR 0.54, 95%CI 0.30-1.00, p=0.049).

Conclusion: Donor pharmacogenomics significantly influence kidney graft fibrosis after adjusting for tacrolimus effects. Donors with rs1045642 (ABCB1 variant) had increasing vascular fibrous thickening. Recipient variants rs1128503 and rs2032582 (ABCB1 variants with no clear association with tacrolimus) are protective for fibrosis and tubular atrophy. Donor pharmacogenotypes are important in long term kidney allograft outcomes.