The impact of liver transplant recipient and donor genetic variability on tacrolimus exposure and transplant outcome

Abstract

This study investigated the effect of recipient and donor genetic variability on dose‐adjusted steady‐state tacrolimus concentrations (C_ss) and clinical outcomes 3 and 6 months after liver transplant. Twenty‐nine recipients and matched donor blood samples were genotyped for 27 single nucleotide polymorphisms including CYP3A5*3 (rs776746), ABCB1 haplotype and immune genes. Associations between genetic variability and clinical parameters and C_ss and the occurrence of rejection and nephrotoxicity were analysed by multivariate and multinomial logistic regression modelling and Jonckheere–Terpstra tests examined the impact of combined donor/recipient CYP3A5 expression on C_ss. At 3 months post‐transplant modelling revealed an association between tacrolimus C_ss and recipient CASP1 rs580523 genotype (P = 0.005), accounting for 52% C_ss variance. Jonckheere–Terpstra tests revealed that as combined donor/recipient CYP3A5 expression increased, C_ss decreased (P = 0.010 [3 months], 0.018 [6 months]). As this is the first report of CASP1 genetic variability influencing tacrolimus C_ss, further validation in larger cohorts is required.

What is already known about this subject

• Liver transplant outcomes are dependent on appropriate tacrolimus exposure.

• Tacrolimus is metabolised by CYP3A4/5 and is a substrate of P‐glycoprotein (encoded by ABCB1).

• Tacrolimus concentrations are influenced by CYP3A5 recipient and donor genetic variability, but the effects of ABCB1 and cytokine/innate immune variability on tacrolimus concentrations are unknown.

What this study adds

• In our Australian transplant cohort, recipient CASP1 (ICE) rs580523 genetic variability was significantly associated with dose‐adjusted steady‐state tacrolimus concentrations at 3 months post‐transplant. In addition, at 3 and 6 months post‐transplant concentrations decreased as donor/recipient CYP3A5 expression increased.

• Genetic variability was not linked to transplant outcomes.

1. INTRODUCTION

Liver transplant success depends on a fine balance of immunosuppression: too little may result in rejection of the liver graft, whilst too much can lead to nephrotoxicity. Tacrolimus, the main immunosuppressant used in Australian liver transplant recipients, has a narrow therapeutic index and therapeutic drug monitoring (TDM) is used to guide dosing. However, despite TDM, 5‐year survival rates of the transplanted liver and recipient remain suboptimal, at 78 and 83%, respectively.1 It has been proposed that additional use of pharmacogenetics could improve targeting of tacrolimus concentrations to improve outcomes.2

Tacrolimus is metabolised by the cytochrome P450 3A enzyme superfamily (CYP3A), expression of which is influenced by genetic variability and non‐genetic factors (e.g. age, hormones, coadministered medication). Notably, CYP3A5*3 (rs776746, the 6986A > G single nucleotide polymorphism [SNP] in intron 3) carriers, representing 82–95% of Caucasians, do not express any CYP3A5.3, 4 Tacrolimus is also a substrate of the efflux transporter P‐glycoprotein (P‐gp), that together with CYP3A determines tacrolimus oral clearance. P‐gp is encoded by the ABCB1 gene, with significant differences in expression and function as a result of numerous SNPs.5 Polymorphisms that alter activity of both CYP3A5 and P‐gp in different tissues, notably the donor liver, recipient gastrointestinal tract and recipient lymphocytes could contribute to rejection and toxicity. Indeed, a recent meta‐analysis has shown that both recipient's and donor's CYP3A5 expression similarly influence tacrolimus concentrations following liver transplant.6 In contrast, observations relating ABCB1 genetic variability with tacrolimus concentrations remain discordant.2

In addition to metabolism and transport of tacrolimus, the impact of innate immune responses and cytokine expression also play an important role in transplantation through modification of CYP3A and P‐gp expression and/or activity.7 For example, activation of the innate immune receptors Toll‐like receptor (TLR) 2 and TLR4, inhibited expression of CYP3A in animal models.8, 9 Conversely, exposure to tacrolimus in liver transplant recipients decreased the innate immune signal from TLR2 and TLR4 to decrease stimulated cytokine release,10 which in turn impacts on CYP3A and P‐gp expression and activity.11, 12, 13, 14 Therefore, genetic variability that causes expression variability in cytokine and innate immune pathway mediators has the potential to impact on tacrolimus concentrations, and as a consequence alter transplant outcomes both directly via immune effects on the transplanted liver and indirectly via tacrolimus exposure. This is supported by a study in a Chinese cohort of liver transplant recipients where a significant impact of recipient interleukin (IL)‐10 SNPs (rs1800896 [−1082G > A], rs1800871 [−819C > T] and rs1800872 [−592C > A]) on dose‐adjusted tacrolimus concentrations was reported.15, 16 However, this is yet to be fully explored with a wide panel of cytokine and innate immune pathway genetic loci in a liver transplant population.

This study investigated the impact of both recipient and donor genetic variability in CYP3A5, ABCB1 and cytokines/innate immune pathways, on dose‐adjusted trough steady‐state tacrolimus concentrations (C_ss) and the outcomes of biopsy proven acute rejection (BPAR) and nephrotoxicity at 3 and 6 months post‐transplant.

2. METHODS

2.1. Participants

This retrospective cohort study was approved by the Southern Adelaide Human Research Ethics Committee and the Australian Red Cross Blood Service Ethics Committee. Twenty‐nine liver transplant recipients under the care of the SA Liver Transplant Unit during 1994 and 2009 who received tacrolimus treatment for immunosuppression were recruited following written informed consent through the SA Liver Transplant Unit at Flinders Medical Centre, and blood samples from their donor obtained from the Australian Red Cross Blood Service. Tacrolimus C_ss were determined by SA Pathology (measured by Abbott ARCHITECT immunoassay, Abbott Diagnostics, USA) as part of the routine TDM clinical care of patients. Demographic data, drug doses and clinical outcomes of BPAR and nephrotoxicity at 3 and 6 months post‐transplant were obtained from clinical records. BPAR was defined using the Banff score, with biopsies reviewed by an experienced liver histopathologist and nephrotoxicity was defined as a >20% decrease in baseline estimated glomerular filtration rate (eGFR).

2.2. Genotyping

Both recipient and donor DNA were genotyped. Genotyping for the CYP3A5*3 variant allele was conducted using a previously published polymerase chain reaction–restriction fragment length polymorphism assay.17 ABCB1 genotype of the 5 most common SNPs of the ABCB1 gene (A61G, G1199A, C1236T, G2677T and C3435T) in Caucasian populations were determined as previously published.18, 19 In addition, genotypes of 21 SNPs of innate immune pathways and cytokines were determined using a customised Sequenom MassArray (iPLEX GOLD) assay at the Australian Genome Research Facility (Brisbane, Australia)20: IL‐1B (rs16944, rs1143627, rs1143634), IL‐2 (rs2069762), IL‐6 (rs10499563), IL‐10 (rs1800871, rs1800896), IL‐6R (rs8192284/rs2228145), TNFA (rs1800629), TGFB (rs11466314, rs1800469), TLR2 (rs3804100), TLR4 (rs4986790, rs4986791), MD2 (LY96, rs11466004), MYD88 (rs6853), BDNF (rs6265), CRP (rs2794521), CASP1 (ICE, rs554344, rs580253) and OPRM1 (rs1799971).

2.3. Data analysis

ABCB1 haplotypes were determined using PHASE Software (v 2.121, 22). Hardy–Weinberg equilibrium analysis was determined with χ² tests with false‐discovery rate (FDR)‐adjusted P‐values. Allele and haplotype frequencies between recipient and donor groups were compared using odds ratios (95% confidence intervals) and Fisher's exact tests with FDR‐adjusted P‐values to determine significance. The associations between tacrolimus dose‐adjusted C_ss at 3 and 6 months post‐transplant and clinical outcomes of rejection and nephrotoxicity with genetic and clinical parameters were investigated using either general linear models of multivariate regression or multinomial regression. The parameters were treated as either categorical factors (recipient sex; ethnicity (self‐reported); donor and recipient genetics; eGFR; chronic kidney disease (CKD) stage; and the incidences of rejection and nephrotoxicity) or quantitative factors (age; pre‐transplant serum creatinine concentrations; follow‐up serum creatinine concentrations; eGFR and time). Multivariate and multinomial regression models were built with step‐wise addition of factors determined to be significantly associated by FDR‐adjusted P < .05 and χ² tests used to test improvement of the models following the addition of each additional factor. The final models provided individual P‐values for each of the included factors. Analysis was performed using R software.23 In addition, the combined effect of donor/recipient CYP3A5 expression on tacrolimus dose‐adjusted C_ss at 3 and 6 months post‐transplant was examined with Jonckheere–Terpstra tests for trend using SPSS® Statistics software (v 25, IBM®). Data are median (range) unless otherwise stated.

2.4. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY,24 and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18.25, 26

3. RESULTS

3.1. Patient demographics and clinical transplant parameters

Demographics and clinical parameters of the recipients are shown in Table 1; the median age was 54.5 years; 86% were male; 27% (n = 8) recipients experienced BPAR and 14% (n = 4) experienced nephrotoxicity.

Table 1.

Transplant recipient demographic and clinical parameters. Data are median (range) unless otherwise stated

Parameter
Sex, male n (%): female n (%)	25 (86): 4 (14)
Age (y)	55 (22–66)
Ethnicity, n (%)
Caucasian	26 (90)
Black African	1 (3)
Asian	2 (7)
Pre‐transplant serum creatinine concentrations (μmol/L)	93 (46–250)
Pre‐transplant eGFR (mL/min/1.73m 2 )	59 (23–>90)
Follow‐up serum creatinine concentrations (μmol/L)	100 (54–254)
Follow‐up eGFR (mL/min/1.73m 2 )	62 (23–>90)
Follow‐up time (days)	42 (10–137)
3 month dose‐adjusted tacrolimus C ss (ng.mL −1 /mg.day −1 )	1.2 (0.49–3.3)
6 month dose‐adjusted tacrolimus C ss (ng.mL −1 /mg.day −1 ) a	1.5 (0.53–6.9)
CKD stage, n (%)
1: eGFR ≥90 mL/min/1.73m 2	2 (7)
2: eGFR 60–89 mL/min/1.73m 2	16 (55)
3+: eGFR ≤59 mL/min/1.73m 2	11 (38)
BPAR, n (%)	8 (27)
Nephrotoxicity, n (%)	4 (14)

^aData missing for one recipient; BPAR, biopsy‐proven acute rejection; CKD stage, chronic kidney disease stage; C_ss, steady‐state concentration; eGFR, estimated glomerular filtration rate.

3.2. Recipients and donors genetic variability

All genotypes, except for recipient TLR4 rs4986790 and rs4986791 (FDR‐adjusted P‐values of 0.02), conformed with Hardy–Weinberg equilibrium. ABCB1 haplotypes were grouped as zero, one or two copies of the wild‐type haplotype (AGCGC). Table S1 shows the allele and haplotype frequencies in recipients and donors. Two SNPs, MD2 rs11466004 and TGFB rs11466314 were not polymorphic in our cohort and so were removed from subsequent analysis. There were no significant differences in allele or haplotype frequencies (FDR‐adjusted P = .105 to >.999) between recipients and donors.

3.3. Impact of genetic variability on tacrolimus concentrations

Multivariate logistic regression of dose‐adjusted C_ss at 3 months post‐transplant revealed a significant association with recipient CASP1 (ICE) rs580253 (C > T) SNP (FDR‐adjusted P = 0.005, n = 26 as 3 recipients did not have a genotype available) that accounted for 51.6% variability in dose‐adjusted C_ss (Table 2). Recipients with wild‐type genotype (C/C, n = 21) had higher dose‐adjusted C_ss in comparison to recipients with heterozygous genotype (C/T, n = 4), whilst the recipient with homozygote variant genotype had the highest dose‐adjusted C_ss (Table 2). This relationship was not observed at 6 months post‐transplant. No other recipient or donor genetic loci, or any clinical parameters investigated impacted significantly on dose‐adjusted C_ss when examined via modelling.

Table 2.

Impact of the CASP1 rs580253 genotypes on steady‐state dose‐adjusted tacrolimus concentrations (C_ss μg.L⁻¹/mg.day⁻¹) at 3 and 6 months post‐transplant. Data are median (range)

Genotype	3 months	6 months
C/C (n = 21)	1.24 (0.63–2.19)	1.65 (0.53–2.26)
C/T (n = 4)	0.81 (0.49–1.62)	0.78 (0.69–2.25)
T/T (n = 1)	3.26	1.51

With regard to the combined impact of donor and recipient CYP3A5 expression (determined from CYP3A5 genotypes), Jonckheere–Terpstra trend tests revealed that as the presence of the CYP3A5*1 allele and therefore CYP3A5 expression increased, dose‐adjusted C_ss at both 3 and 6 months post‐transplant decreased, P = 0.010 and 0.018, respectively (Figure 1), medians: 3 months — nonexpressor donor + nonexpressor recipient (n = 15) =1.35, nonexpressor donor + expressor recipient (n = 5) =1.01, expressor donor + nonexpressor recipient (n = 8) = 0.97, expressor donor + expressor recipient (n = 1) = 0.49; and 6 months — nonexpressor donor + nonexpressor recipient (n = 15) =1.97, nonexpressor donor + expressor recipient (n = 5) =1.36, expressor donor + nonexpressor recipient (n = 7) =1.05, expressor donor + expressor recipient (n = 1) = 0.73.

Figure 1.

Dose‐adjusted tacrolimus C_ss at 3 (A) and 6 (B) months post‐transplant in liver transplant donor/recipient combinations of differing CYP3A5 expression; −/− donor nonexpressor / recipient nonexpressor (circle), −/+ donor nonexpressor / recipient expressor (square), +/− donor expressor/recipient nonexpressor (triangle), and +/+ donor expressor / recipient expressor (diamond). Lines indicate group medians

3.4. Impact of genetic variability on clinical transplant outcomes

With regard to the clinical transplant outcomes there was no influence of genetic variability or clinical transplant parameters on nephrotoxicity or BPAR episodes in our cohort once P‐values were FDR‐adjusted (P > 0.07).

4. DISCUSSION

Success of liver transplantation has improved through tacrolimus TDM that mitigates pharmacokinetic variability. However, significant host and graft issues remain that limit transplant success. Identifying recipient and/or donor genetic factors that predict these outcomes could further inform individualisation of treatment. This retrospective study examined genetic variability of CYP3A5 and ABCB1, and, for the first time, a number of novel loci in cytokine and innate immune pathways in an attempt to identify loci of clinical importance following liver transplantation.

Of the genetic loci investigated, genotypes for two recipient TLR4 SNPs (rs4986790 and rs4986791) did not conform with Hardy–Weinberg equilibrium. This has been noted previously in the literature for rs4986790.27 The scientific basis for this observation is unclear, but as appropriate controls were employed in all genotyping analyses and this was not observed for donor TLR4 genotypes, this is unlikely to be due to experimental error and is more likely to be a reflection of the small sample size and rare occurrence of these SNPs.

The only significant genetic association revealed by the modelling analysis was between recipient CASP1 (ICE) rs580253 genotype and dose‐adjusted tacrolimus C_ss at 3 months post‐transplant. CASP1 encodes interleukin converting enzyme that mediates the functional conversion of the cytokines IL‐1β and IL‐18. IL‐1β has been previously associated with inhibition of CYP3A mRNA expression and metabolic activity, whilst genetic variability in IL‐18 has been associated with altered tacrolimus metabolism in the 1st week post‐transplant.28 Therefore, modulation of CYP3A activity could explain this association. At present however, the functional impact of this CASP1 SNP on IL‐1β has not been reported, whilst the other CASP1 SNP investigated was not associated with dose‐adjusted C_ss. Consequently, this observation should be considered with caution prior to necessary replication studies in larger cohorts, especially given the current study was limited to a small number of recipients, that the impact of the SNP did not consistently lower dose‐adjusted C_ss, and time dependent nature of the association (i.e. 3 months but not 6 months post‐transplant).

In agreement with previous research,15, 29, 30, 31, 32, 33 Jonckheere–Terpstra tests for trend revealed an association between dose‐adjusted tacrolimus C_ss at 3 and 6 months post‐transplant and combined donor/recipient expression of CYP3A5. This indicates that both liver and gastrointestinal tract of the donor and recipient, respectively, are important metabolising organs in CYP3A5 expressors. In contrast, ABCB1 haplotypes of the recipient and the donor were not related to dose‐adjusted tacrolimus C_ss at 3 or 6 months post‐transplant. This is consistent with prior studies investigating the impact of ABCB1 SNPs on dose‐adjusted tacrolimus C_ss. 29, 32, 34, 35 Despite these findings, ABCB1 genotype may still be an important determinant of dose‐adjusted tacrolimus C_ss in target tissues and cells, given that previous studies have reported an impact on hepatic concentrations.34 Finally, none of the recipient or donor genetic loci investigated were associated with the clinical outcomes of BPAR or nephrotoxicity.

The main limitation of this study was small sample size and consequent power, with fewer subjects meeting the inclusion criteria than predicted at the start of the study. In addition, the impact of co‐administration of CYP3A4 inducers and inhibitors was not accounted for in this study and could also explain a proportion of the interindividual variability in dose‐adjusted tacrolimus C_ss. Despite these limitations a potentially important effect of the innate immune response and confirmatory effect of CYP3A5 expression was identified and this data can contribute to future meta‐analyses examining the impact of donor and recipient genotype on dose‐adjusted tacrolimus C_ss.

In conclusion, we report here that dose‐adjusted tacrolimus C_ss are associated with recipient CASP1 genotype at 3 months, and combined donor/recipient CYP3A5 expression at 3 and 6 months post‐transplant. Future studies are required to validate these observations, in particular the time‐dependent nature of the recipient CASP1 genotype association.

COMPETING INTERESTS

A.W. received an unrestricted grant from Jansen ($10,000 AUD) to assist with study and has received prior grant support/speaking fees/advisory board fees from MSD, AbbVie and Gilead pharmaceutical companies. All other authors have no competing interests to declare.

CONTRIBUTORS

J.R., L.J., A.W. and M.D. all contributed to the study design and conduct, interpretation of outcomes and manuscript preparation. J.T. contributed to the data analysis, interpretation of outcomes and manuscript preparation; and J.K.C. contributed to the study conduct, data analysis, interpretation of outcomes and manuscript preparation.

DATA AVAILABILITY STATEMENT

Data available on request from the authors.

Supporting information

Table S1:

Single nucleotide polymorphism allele and ABCB1 haplotype frequencies (%) in liver transplant recipients and donors with Fisher's exact test pointwise and false‐discovery rate adjusted P‐values comparing recipient and donor frequencies.

Coller JK, Ramachandran J, John L, Tuke J, Wigg A, Doogue M. The impact of liver transplant recipient and donor genetic variability on tacrolimus exposure and transplant outcome. Br J Clin Pharmacol. 2019;85:2170–2175. 10.1111/bcp.14034