Association of ABCC2 haplotypes to Mycophenolic Acid Pharmacokinetics in Stable Kidney Transplant Recipients

Abstract

Mycophenolic acid exhibits significant interpatient pharmacokinetic variability attributed to factors including race, sex, concurrent medications, and enterohepatic circulation of the mycophenolic acid glucuronide metabolite to mycophenolic acid. This conversion by enterohepatic circulation is mediated by the multidrug resistance-associated protein 2 (MRP2), encoded by ABCC2. This study investigated ABCC2 haplotype associations to mycophenolic acid pharmacokinetics in 147 stable kidney transplant recipients receiving mycophenolic acid in combination with calcineurin inhibitors.

The role of the ABCC2 genotypes: −24C>T (rs717620); 1249 C>T(rs2273697); and 3972 C>T (rs3740066) were evaluated in prospective, cross-sectional pharmacokinetic studies of stable recipients receiving mycophenolic acid and either tacrolimus or cyclosporine. Haplotype phenotypic associations with mycophenolic acid pharmacokinetic parameters were computed using THESIAS (v. 3.1).

Four ABCC2 haplotypes with estimated frequencies greater than 10% were identified (H1:CGC (Wild-Type); H9:CGT; H2:CAC; H12:TGT). There were no differences in haplotype frequencies by either race or sex. There were significant associations of pharmacokinetic parameters with ABCC2 haplotypes for mycophenolic acid clearance (L/hr), mycophenolic acid AUC_0-12h (mg●hr/L) and the ratio of mycophenolic acid glucuronide to mycophenolic acid AUC_0-12h. The wildtype haplotype ABCC2 CGC had greater mycophenolic acid AUC_0-12h (P=0.017), slower clearance (P=0.013) and lower mycophenolic acid glucuronide to mycophenolic acid AUC_0-12h ratio (P=0.047) compared to the reduced function ABCC2 haplotype, CGT. These differences were most pronounced among patients receiving tacrolimus co-treatment. No phenotypic associations were found with the cyclosporine-mycophenolic acid regimen. Variation in ABCC2 haplotypes contributes to sub-therapeutic mycophenolic acid exposures and influences interpatient variability in pharmacokinetic phenotypes based upon the concurrent calcineurin inhibitor treatment.

Introduction

Mycophenolic acid is commonly prescribed with the calcineurin inhibitors, tacrolimus or cyclosporine, as maintenance immunosuppression in post-kidney transplant recipients.^1,2 Since studies have shown a clinically significant benefit between mycophenolic acid exposure and clinical outcomes, this emphasizes the need for informed therapeutic drug monitoring and individualization of regimens.^3–6 Mycophenolic acid pharmacokinetics is complex with significant intra- and inter-patient variability due to concurrent medications, sex, race, genetics, glomerular filtration rate, kidney allograft function, time post-transplant and gastrointestinal disease.^7,8 Mycophenolic acid is available as either the ester prodrug, mycophenolate mofetil (MMF), or as an enteric-coated sodium salt, mycophenolate sodium (EC-MPS). The two mycophenolic acid formulations have similar efficacy with different pharmacokinetics.^8–11 EC-MPS has an enteric coating and is mainly dissolved and absorbed in the small intestine which may reduce gastrointestinal adverse effects.^2,7,12,13,14

Following oral administration of either formulation, carboxylesterases rapidly hydrolyze to generate the active moiety, mycophenolic acid.¹² Liver and intestinal Uridine 5’-diphospho-glucuronosyl transferases (UGTs) are responsible for the metabolism of mycophenolic acid to its major inactive metabolite, mycophenolic acid 7-O-glucuronide. ¹² Other metabolites include an acyl glucuronide form, which is less pharmacologically active than mycophenolic acid and 6-O-desmethyl-mycophenolic acid. ^2,15

Excretion of both mycophenolic acid glucuronide and the acyl glucuronide into the biliary system is mediated by the multidrug resistance protein 2 encoded by a member of the ATP-binding cassette (ABC) transporter superfamily, ABCC2 (formerly CMOAT). The majority of mycophenolic acid glucuronide that is excreted in the bile, undergoes de-glucuronidation by intestinal gram-negative bacteria to convert back to mycophenolic acid. The reabsorption of mycophenolic acid from the gastrointestinal tract results in the appearance of a secondary mycophenolic acid peak appearing 6-12 hours following oral drug administration and attributed to enterohepatic circulation. This process is impacted by the concurrent calcineurin inhibitors.^8,12 This enterohepatic recycling contributes 10 to 60% to the total mycophenolic acid exposure (AUC_0-12h) over the dosing interval ^7,8 The multidrug resistance protein 2 is also expressed in the renal proximal tubules and contributes in the urinary excretion of primarily mycophenolic acid glucuronide. ^2,16,17 Therefore, the multidrug resistance protein 2 transporter modulates mycophenolic acid and mycophenolic acid glucuronide systemic exposure with immunosuppression post-transplant.

Current immunosuppressant protocols recommend a mycophenolic acid formulation with either cyclosporine A or tacrolimus.² The concurrent administration of the calcineurin inhibitors (tacrolimus or cyclosporine) contributes to patient variability in mycophenolic acid pharmacokinetics.⁸ Cyclosporine interferes with the enterohepatic circulation of the inactive metabolite, mycophenolic acid glucuronide, through inhibition of hepatobiliary transporters that include the multidrug resistance protein 2, and the organic anion transporting polypeptide 1B3 (OAT1B3) and 1B1 (OAT1B1).^{18 12} This inhibition reduces the enterohepatic recycling and eliminates the secondary mycophenolic acid peak between 6 to 12 hours post-dose resulting in reduced area under the concentration-time curve from 0 to 12 hours (AUC_0-12h). ^{7,8 16,19–22} The role of cyclosporine in the reduction of mycophenolic acid glucuronide to mycophenolic acid by enterohepatic recycling has recently been questioned.¹⁶ Studies using multidrug resistance protein 2 deficient rats (Wistar TR⁻) have suggested that the role of cyclosporine in reduced enterohepatic circulation can be attributed to the inhibition of biliary excretion of mycophenolic acid glucuronide by multidrug resistance protein 2.^19,23 In contrast, in vitro studies using MDCKII cells suggest that mycophenolic acid is not inhibited by cyclosporine.²¹ These authors identified both organic anion transporting polypeptide 1B3 and 1B1 inhibition as the likely cause of reduced enterohepatic circulation in the presence of cyclosporine.²¹ Further human studies are necessary to establish the mechanism of this drug-drug interaction. In contrast, tacrolimus does not interfere with enterohepatic circulation and a second peak is evident resulting in increased mycophenolic acid exposure.^24,25 Tacrolimus has no apparent drug-drug interaction with mycophenolic acid in healthy subjects or kidney transplant recipients.^26,27

While the exact role of mycophenolic acid in enterohepatic circulation is unclear, polymorphisms in ABCC2, the gene that encodes multidrug resistance protein 2, have been implicated in clinically relevant changes in the mycophenolic acid pharmacokinetics.^28,29 The three common ABCC2 single nucleotide polymorphisms (SNPs) attributed to influencing drug response are −24C>T (rs717620), 1249G>A (rs374066) and 3972C>T (rs3740066). The wildtype alleles which are: −24C, 1246G, and 3972C for each genotype comprise the haplotype H1[CGC].^29,30 The H2 haplotype (carrying the variant allele in 1249G>A and classified as CAC) has been reported to have higher protein expression and transport activity compared to H1. ^29,30 Haplotypes H9 (3972 C<T; CGT) and H12 ( −24C>T and 3972C>T; TGT) are reported to exhibit lower protein expression and transport activity of MRP2 compared to the wildtype haplotype, H1.^29,30 Naesens, et al. evaluated individual ABCC2 SNPs in 95 white, kidney allograft recipients, who received oral tacrolimus with mycophenolate mofetil and found the variant −24 T allele was associated with higher dose-adjusted mycophenolic acid troughs. Additionally, the ABCC2 −24 C>T and 3972 C>T SNPs exhibit linkage disequilibrium and appeared to not impact mycophenolic acid pharmacokinetics during liver dysfunction.²⁹ Zhang, et al. noted that among 46 Chinese kidney transplant patients receiving cyclosporine, no association of the −24C>T SNP with mycophenolic acid exposures was noted with 249G>A associated with higher acyl mycophenolic acid glucuronide concentrations.³¹ Božina, et al. evaluated recipient and donor genotypes for the ABCC2 −24C>T among 68 kidney transplant patients and provided the first report that donor ABCC2 1249G>A substitution appeared to increase mycophenolic acid clearance.³² Ogasawara, et al. investigated ABCC2 1249G>A polymorphism in 102 stable kidney transplant patients receiving mycophenolic acid and tacrolimus with the outcome of decreased dose-normalized tacrolimus troughs.³³ Conversely, the 3972C>T SNP significantly increased dose-normalized tacrolimus troughs. The authors suggested that ABCC2 haplotypes should be incorporated into genomic-pharmacokinetic investigations rather than evaluation of individual polymorphisms. Thus, use of ABCC2 SNPs alone may explain some of the conflicting associations to mycophenolic acid pharmacokinetic ABCC2 phenotypes.

The objectives of this study were: 1) to investigate the association of ABCC2 haplotypes to steady-state mycophenolic acid and mycophenolic acid glucuronide pharmacokinetics in stable kidney transplant recipients receiving long-term maintenance immunosuppression; 2) to assess the role of ABCC2 haplotypes on mycophenolic acid pharmacokinetics relative to differences in concurrent calcineurin inhibitor therapy with mycophenolic acid; and 3) to evaluate the association of individual ABCC2 polymorphisms to mycophenolic acid pharmacokinetics compared to haplotype associations.

Methods

Study Population

The study was approved by the University at Buffalo institutional review board (IRB# PHP0720608B). Each patient provided written informed consent prior to participation. All research activities adhered to the principles of the Declaration of Helsinki and the United States Federal Policy for the Protection of Human Subjects. This research report was consistent with the Principles of the Declaration of Istanbul and outlined in the ‘Declaration of Istanbul on Organ Trafficking and Transplant Tourism’. The study site was the Nephrology Division at the University at Buffalo (UB_MD) located at Erie County Medical Center in Buffalo, New York.

This study pooled data from three prospective, cross-sectional open-label 12-hour pharmacokinetic studies.^24,25,34 The role of mycophenolic acid pharmacokinetics and calcineurin inhibitors with adverse effects in these studies has been previously published in this journal³⁵. All studies utilized the same study design, patient inclusion/exclusion criteria, screening and enrollment processes, study procedures, and analytical drug assays with comparable quality control outcomes as described below and previously published.^{24,25,34–37} Only kidney transplant recipients from the University at Buffalo Kidney Transplant Clinic were enrolled after collaborating nephrologist confirmed clinical stability of each patient. Each study evaluated the 12-hour pharmacokinetics of mycophenolic acid using serial sampling.^24,35 All patients receiving maintenance calcineurin inhibitor and mycophenolic acid immunosuppression for greater than 3 months and receiving steady state dosing regimens for at least 7 days were enrolled using standardized criteria to verify clinical stability. Ethnicity for two previous generations was verified. For all patients, the inclusion criteria were: 1) minimum 6 months post-kidney transplantation; 2) age 25-70 years; 3) immunosuppressive regimen of either cyclosporine with mycophenolate mofetil or tacrolimus with enteric-coated mycophenolate sodium for greater than 3 months; 4) no change in immunosuppressive doses for more than 7 days prior to study; 5) stable serum creatinine with changes no greater than 0.25 mg/dl in prior 2 clinic visits; 6) leukocyte count no less than 3000 cells/mm³; and hematocrit no less than 30% for 4 weeks prior to study. Exclusion criteria were: 1) serum creatinine ≥ 3.5mg/dl; 2) active infection or acute rejection within 2 weeks; 3) significant diseases limiting participation including gastrointestinal, cardiovascular, hematologic, psychiatric, neurologic or oncologic disease; 4) medication non-adherence; 5) drugs, herbal supplements, or drinks (e.g. grapefruit juice) that are Cytochrome P450 3A (CYP3A) or P-glycoprotein inhibitors or inducers within 4 weeks; 6) drugs that interfere with absorption of calcineurin inhibitor or mycophenolic acid. Comprehensive metabolic panels, fasting lipid panels and complete blood counts were completed with physical exams after enrollment and repeated on study morning after overnight fast. and Adherence assessment and medication history were performed upon enrollment and within the week prior to study. Routine therapeutic drug monitoring for calcineurin inhibitors was ongoing. Targeted trough concentrations were 5-10 ng/ml for tacrolimus and 90-150 ng/ml for cyclosporine. Mycophenolic acid dosing adjustments were based upon clinical response, concurrent calcineurin inhibitor and adverse effects without routine therapeutic drug monitoring. Estimated glomerular filtration rate was calculated using the four-factor Modification of Diet in Kidney Disease equation.³⁸

Study Procedure

All patients were studied at steady-state dosing conditions and received the same dose of brand name immunosuppressive drugs for at least 7 days prior to study. Proton pump inhibiters, histamine₂ receptor antagonists and antacids were discontinued at least 36 hours preceding the study. Patients took the immunosuppressive medications between 5:30 to 6:30 PM and had their evening meal at 8:00 to 8:30 PM preceding the study. Patients fasted and abstained from caffeine and alcohol for 12 hours prior to the study. At 6:00 AM on the morning of the study patients were admitted to the research unit with vital signs documented and placement of an intravenous angiocatheter. The 0-hour samples (~15 ml) were collected prior to mycophenolic acid and calcineurin inhibitor administration to determine immunosuppressive trough concentrations, comprehensive metabolic profile and complete blood count. After 0-hour blood collection at approximately 7AM, the doses of tacrolimus (as Prograf^®; Astellas) and enteric-coated mycophenolate sodium (as Myfortic^®; Novartis) or cyclosporine (as Neoral^®; Novartis) with mycophenolate mofetil (as CellCept^®; Roche) were administered orally. Standardized low fat and low sodium meals were provided 4 hours after the study medication was administered. Patients remained in an upright position throughout the study. Anti-hypertensive drugs were administered after 1.5 hours while insulin, anti-lipidemic and other medications were administered 4 hours after the immunosuppressives.

All blood samples collected were immediately placed on ice, centrifuged, and harvested within 60 minutes and then stored at −70°C until assayed as separate aliquots. The detailed sampling procedure is provided in the recent publication from our group³⁵.

Drug Assay Methodology

A liquid chromatography/mass spectrometry assay was used for the simultaneous analysis of plasma mycophenolic acid and mycophenolic acid glucuronide with 5, 5-diphenylhydantoin and flumethasone as internal standards, respectively³⁶. Assay methodology and sample preparation adhered to previously validated procedures³⁶. Standard curve concentrations ranged from 0.145 µg/mL to 15.5 µg/mL for mycophenolic acid and from 22.1 to 295.2 µg/mL for mycophenolic acid glucuronide. The lowest limit of quantitation was 0.072 µg/mL for mycophenolic acid and 11.1 µg/mL for mycophenolic acid glucuronide. The relative standard deviation for intraday variation for mycophenolic acid ranged from 0.76 to 3.99% and 2.11 to 9.57% for mycophenolic acid glucuronide. The relative standard deviation of interday variation for mycophenolic acid ranged from 2.3 to 6.36% and 3.8 to 8.14% for mycophenolic acid glucuronide.

Pharmacokinetic Analysis

Pharmacokinetic parameters (phenotypes) for mycophenolic acid and mycophenolic acid glucuronide included area under the concentration versus time curve from 0 to 12 hours (AUC_0-12h), 12-hour trough concentration (C_p(12h)). The mycophenolic acid dose equivalent was used (720mg mycophenolic acid = 720mg enteric-coated mycophenolate sodium = 1000mg mycophenolate mofetil). Oral apparent mycophenolic acid clearance was calculated as mycophenolic acid dose equivalent divided by AUC_0-12h as determined by the linear trapezoidal rule using non-compartmental pharmacokinetic methods (Phoenix WINNONLIN Version 6.3. Pharsight Corp, Mountain View, CA). The metabolic ratio of mycophenolic acid glucuronide (MPAG) AUC_0-12h to mycophenolic acid AUC_0-12h was calculated and reflects the relationship between primary metabolite and active drug as a surrogate for enterohepatic circulation.

Enterohepatic circulation has been previously defined.³⁵ The mycophenolic acid AUC from 6-12 hours post-dose (AUC_6-12h) was also used as an objective estimate of enterohepatic circulation.^3,28 The ratio of mycophenolic acid AUC_6-12h to mycophenolic acid AUC_0-12h reflects the relative contribution of enterohepatic circulation to the overall mycophenolic acid exposure.

Statistical Analysis

Continuous data were reported using means and standard deviations and categorical data using frequencies and percentages. Group differences in demographic variables by immunosuppressive regimen were statistically assessed with ANOVA, Chi-Square, or Fishers exact tests when appropriate. Diagnostic plots were used to assess model fit and identify the need for data transformations. Statistical outliers were defined as studentized residuals outside ± 3. Model assumptions were assessed using diagnostic plots and data transformations were applied as necessary. All tests were two-sided with significance level of 0.05 utilized for hypothesis testing with SAS statistical software (version 9.3, SAS Institute, Cary, NC).

Genetic Analysis

All blood samples, stored at −70°C until analysis, provided viable DNA for genotyping and were analyzed at the University of New England Genomics Research Core. Genomic DNA was isolated from 600 µl of PBMCs per manufacturers’ protocol (Wizard^® Genomic DNA Purification, Promega Madison, WI). Personnel with no knowledge of clinical data assayed for ABCC2 variants, ABCC2 −24C>T (rs717620), ABBC2 1249 G>A (rs2272697), ABCC2 3972 C>T (rs3740066). Ten ng of patient genomic DNA was used to characterize each single nucleotide polymorphism (SNPs) using validated TaqMan^® allelic discrimination assays (ThermoFisher Scientific, Applied Biosystems, Foster City, CA) with Bio-Rad Laboratories CFX96 Real-Time Polymerase Chain Reaction Detection System (Hercules, CA). For each SNP assay, duplicate samples were analyzed. All protocols and sample handling were in accordance with published guidelines. Allele frequencies for all SNPs were confirmed to be in Hardy-Weinberg equilibrium when adjusted for race.

Given the known linkage among at least two of the three ABCC2 SNPs, haplotype analysis was conducted. Haplotype analysis provides greater power to detect potential unknown functional variants than SNPs alone.³⁹ ABCC2 haplotype estimation was determined using the THESIAS program (Testing Haplotype Effects In Association Studies, ver 3.1).⁴ THESIAS uses a maximum likelihood algorithm for the simultaneous estimation of haplotype frequencies (single chromosome region) and their association to mycophenolic acid pharmacokinetic parameters. Note that THESIAS is unable to assign specific ABCC2 haplotypes to study participants. Figures 2 and 3 and Table 3 provide mycophenolic acid pharmacokinetic estimations for haplotypic regions. Individual study participants are not able to be represented using this THESIAS computation . Significant associations for pharmacokinetic parameters for all patients were divided by calcineurin inhibitor (tacrolimus and cyclosporine) and reported as phenotypic means with confidence intervals for each haplotype on a single chromosome.

Figure 2.

Mycophenolic acid pharmacokinetic parameters vs ABCC2 haplotypes. Significant pharmacokinetic parameters were reported relative to phenotypic means for each estimated haplotype (single chromosomal region) using computation with the THESIAS platform. THESIAS uses maximum likelihood algorithm for the simultaneous estimation of haplotype frequencies and phenotypic means. Pharmacokinetic values for any individual would be the result of the combination of two haplotypes. Error bars indicate 95% CI. Percentages indicate proportion of individuals assigned to each haplotype. AUC 0-12 hr: Area Under the Concentration Time Curve 0-12 hours, MPA: mycophenolic acid, MPAG: mycophenolic acid glucuronide.

Figure 3.

Sub-analysis by Calcineurin Inhibitor treatment: Mycophenolic acid Pharmacokinetic parameters for Tacrolimus (N=67) and Cyclosporine (N=80) vs ABCC2 CGC and CGT haplotypes. Significant pharmacokinetic parameters were reported relative to phenotypic means and confidence intervals for each haplotype on a single chromosomal region using computation with the THESIAS platform. THESIAS uses maximum likelihood algorithm for the simultaneous estimation of haplotype frequencies and phenotypic means. Note that pharmacokinetic parameters for any individual would be the result of the combination of two haplotypes. With tacrolimus concurrent treatment, reduced mycophenolic acid AUC_0-12hr (P=0.031) and AUC _6-12hr (P=0.0148) with increased clearance (P=0.014) and mycophenolic acid glucuronide AUC_0-12hr to mycophenolic acid AUC_0-12hr ratio (P=0.075) were more evident with the variant CGT haplotype compared to the wildtype CGC. See Table 3. Error bars indicate 95% CI. Percentages indicate proportion of individuals assigned to each haplotype. MPA: mycophenolic acid, MPAG: mycophenolic acid glucuronide.

Table 3. Association of ABCC2 haplotypes estimated phenotypic means for Mycophenolic acid pharmacokinetic parameters. Values for all individuals and subdivided by Calcineurin Inhibitor.

Phenotypic mean estimated from maximum likelihood methods by THESIAS; Phenotype reported as contribution of single haplotype (one chromosome region).

	All (N=147)			Cyclosporine Group (n=80)			Tacrolimus Group (n=67)
Mycophenolic Acid Pharmacokinetic Parameters	Phenotypic Mean (±95% CI) Wildtype Haplotype (CGC)	Phenotypic Mean (±95% CI) Variant Haplotype (CGT)	P value	Phenotypic Mean (±95% CI) Wildtype Haplotype (CGC)	Phenotypic Mean (±95% CI) Variant Haplotype (CGT)	P value	Phenotypic Mean (±95% CI) Wildtype Haplotype (CGC)	Phenotypic Mean (±95% CI) Variant Haplotype (CGT)	P value
CL/F (L/hr)	7.11(5.74 to 8.49)	11.04(8.80 to 13.27)	0.013	10.04 (7.9 to 12.18)	13.68 (10.7 to 16.66)	0.958	4.58 (3.53 tp 5.63)	7.22 (5.7 to 8.73)	0.014
AUC0-12hr (mg●hr/L)	31.06 (27.83 to 34.30)	18.03 (8.40 to 27.66)	0.017	21.11 (17.97 to 24.26)	16.64 (9.6 to 23.65)	0.890	39.96 (34.59 to 45.32)	20.06 (4.08 to 36.0)	0.031
Ratio MPAG AUC0-12hr / MPA AUC0-12hr	12.27 (9.73 to 14.81)	18.96 (13.45 to 24.47)	0.047	19.75 (16.19 to 23.3)	25.7 (17.53 to 33.86)	0.809	5.86 (4.86 to 7.02)	8.58 (6.21 to 10.95)	0.075
C12hr (mg/L)	1.63 (1.39 to 1.87)	1.01 (0.42 to 1.61)	0.078	0.965 (0.633 to 1.3)	0.783 (0.182 to 1.38)	0.597	2.25 (1.87 to 2.63)	1.33 (0.411 to 2.25)	0.081
AUC6-12hr	12.05 (10.6 to 13.5)	5.68 (−0.73 to 12.1)	0.064	6.67 (5.47 to 7.86)	5.11 (2.36 to 7.86)	0.091	17.09 (14.65 to 19.54)	6.37 (−7.57 to 20.31)	0.0148
AUC: area under the concentration time curve, C12hr: trough concentration 12 hours post-dose, CL/F: apparent oral clearance, MPA: mycophenolic acid, MPAG: mycophenolic acid glucuronide. Bonferroni corrected level of significance is 0.01 for multiple testing

Results

This study included 147 clinically stable kidney transplant recipients at 4.0 ± 3.2 years’ post-transplant. Table 1 includes the demographics and clinical characteristics of all patients as well as the results stratified by calcineurin inhibitor. The cyclosporine group included more males and whites, and this group received low dose prednisone more frequently. There was an equal distribution of race and sex in the tacrolimus group. Albumin and liver function tests were within normal range for all patients. Blacks had similar estimated glomerular filtration rate (53.9 ± 15.5 ml/min/1.72m²) compared to whites (50.0 ± 18.4 ml/min/1.72m²; P=0.172).

Table 1.

Demographic and clinical factors by Calcineurin Inhibitor.³³

	All (N=147)	Cyclosporine (n=80)	Tacrolimus (n=67)	P-value Tac vs CsA
Age (years)	50.9 ± 11.0	52.53 ± 10.33	48.96 ± 11.50	0.0494
Sex  Male	104 (70%)	66 (82.50%)	38 (56.72%)	0.0006
Female	43 (30%)	14 (17.50%)	29 (43.28%)
Race   Caucasian	82 (55.8%)	50 (62.50%)	32 (47.76%)	0.0731
African-American	65 (44.2%)	30 (37.50%)	35 (52.24%)
Total Body Weight (kg)	88.84 ± 21.05	90.20 ± 21.80	87.21 ± 20.16	0.3929
Body Mass Index (kg/m2)	30.05 ± 6.15	29.99 ± 6.30	30.11 ± 6.02	0.9066
Time Post-Transplant (years)	3.96 ± 3.23	4.76 ± 3.55	3.01 ± 2.53	0.0023
Estimated glomerular filtration rate (ml/min/1.73m2)	51.72 ± 17.22	48.80 ± 17.95	55.19 ± 15.72	0.0244
Glucose (mg/dl)	113.93 ± 63.17	113.44 ± 58.93	114.51 ± 68.33	0.919
Albumin (g/dl)	4.03 ± 0.38	3.94 ± 0.42	4.13 ± 0.29	0.0032
White blood cells (x103/mm3)	5.74 ± 1.95	6.13 ± 1.90	5.27 ± 1.92	0.0033
Neutrophil number (x103/mm3)	3.67 ± 1.60	3.83 ± 1.60	3.50 ± 1.60	0.224
Lymphocyte number (x103/mm3)	1.33 ± 0.57	1.56 ± 0.64	1.06 ± 0.49	<0.0001
Platelets (x 103 cells/mm3)	21636 ±62.0	229.88 ± 65.03	200.22 ± 54.37	0.0036
Hemoglobin (g/dl)	12.56 ± 1.44	12.78 ± 1.47	12.31 ± 1.37	0.046
Calcineurin Inhibitor Trough Concentration (ng/ml)	NA	129.29 ± 46.47	7.20 ± 1.89	NA
Mycophenolic acid dose (mg)	636.73 ± 207.78	693.00 ± 214.51	569.55 ± 178.90	0.0003
Enterohepatic Recirculation	NA	36 (45.0%)	52 (77.61%)	<0.0001

Data presented as mean ± standard deviation. NA – not applicable

Continuous data was transformed to normality as appropriate and analyzed via two-sided independent samples t-test.

Categorical data was analyzed via Chi-Square or Fisher’s Exact test, where appropriate.

Bonferroni corrected level of significance is 0.003 for multiple testing.

Table 2 provides an abbreviated table of pertinent mycophenolic acid and mycophenolic acid glucuronide pharmacokinetic parameters comparing tacrolimus and cyclosporine groups (for complete table see Meaney et al.³⁵). Based on empiric adjustment, patients in the tacrolimus group received lower mycophenolic acid doses. These patients exhibited 48% slower apparent clearance (P<0.001). The mycophenolic acid therapeutic range of AUC_0-12h from 30 to 60 mg●h/L has been established for mycophenolate mofetil and cyclosporine regimens.^40–42 In this study, 56.3% of patients receiving cyclosporine achieved this target compared to 34.3% of the tacrolimus group.

Table 2.

Selected Mycophenolic Acid Pharmacokinetic Parameters Relative to Concurrent Calcineurin Inhibitor Treatment (see Meaney³³ for complete table)

Mycophenolic Acid Pharmacokinetic Parameters	All (N=147)	Cyclosporine Group (n=80)	Tacrolimus Group (n=67)	P-value Cyclosporine vs Tacrolimus
Mycophenolic acid CL/F (L/hr)	15.99 ± 9.41	20.50 ± 9.99	10.60 ± 4.71	<0.0001
Mycophenolic acid AUC0-12hr (mg●hr/L)	53.63 ± 28.21	40.82 ± 17.60	68.91 ± 30.88	<0.0001
Ratio Mycophenolic acid glucuronide AUC0-12hr to Mycophenolic acid AUC0-12hr	28.61 ± 20.49	40.79 ± 20.38	14.06 ± 6.11	<0.0001
Mycophenolic acid C12hr (mg/L)	3.00 ± 2.07	2.032 ± 1.499	3.694 ± 2.019	<0.0001
Mycophenolic acid AUC6-12hr (mg•hr/L)	19.06 ± 15.17	13.07 ± 7.15	26.21 ± 18.78	<0.0001

AUC_0-12hr : area under the concentration time curve from 0 to 12 hr, C_12hr: trough concentration 12 hours post-dose, CL/F: apparent oral clearance, MPA: mycophenolic acid, MPAG: mycophenolic acid glucuronide. Data displayed as mean ± standard deviation. P-values generated using general linear models.

Bonferroni corrected level of significance is 0.01 for multiple testing.

The ratio of mycophenolic acid glucuronide AUC_0-12h to mycophenolic acid AUC_0-12h (metabolite to active drug exposure) was 2.9-fold greater for the cyclosporine regimen suggesting reduced enterohepatic circulation. More than 75% of patients receiving tacrolimus regimen exhibited enterohepatic circulation (P<0.001) with greater mycophenolic acid AUC_6-12h to mycophenolic acid AUC_0-12h ratio (P=0.036) and reduced metabolite to active drug ratio (mycophenolic acid glucuronide AUC_0-12h to mycophenolic acid AUC_0-12h; P<0.001).

The ABCC2 variants, ABCC2 −24C>T (rs717620), ABBC2 1249 G>A (rs2272697), ABCC2 3972 C>T (rs3740066) were assessed using validated TaqMan allelic discrimination assays. Hardy-Weinberg equilibrium when adjusted for race was confirmed for allele frequencies at each position. There were significant differences between Black and White genotype frequencies for both ABCC2 −24C>T and ABCC2 3972 C>T as determined by Goodness-of-fit test (Figure 1A). Linkage disequilibrium (LD) among the ABCC2 SNPs was found to be significant and ranged from 0.92 (ABCC2 −24 – 3972) to 0.89 (ABCC2 1249 – 3972). THESIAS estimated ABCC2 haplotype frequencies (n=147) are summarized in Figure 1B. Four haplotypes were found to have estimated frequencies greater than 10%; the wildtype, H1:CGC (51.3%) and three variant haplotypes H9:CGT (18.1%), H2:CAC (15.4%) and H12:TGT (13.8%). No individual haplotype assignment to each participant can be made using THESIAS estimation analysis. Despite the differences observed in genotype frequencies, there were no significant differences among the haplotype distributions either by race or sex (Figure 1B).

Figure 1.

A. Genotype frequencies segregated by race for each ABCC2 SNP. There were significant differences in genotype frequencies for ABCC2 −24C>T (rs717620) and ABCC2 3972 C>T (rs3740066). Figure 1 B. ABCC2 Haplotype (−24C>T, 1249G>A, 3972C>T) frequencies among races and sex. There were no significant differences among haplotypes for either race (G=6.94, df=3, ns) or sex (G-3.47, df=3, ns).

The significant associations of pharmacokinetic parameters with ABCC2 haplotypes were observed for mycophenolic acid clearance, mycophenolic acid AUC_0-12hr and the ratio of mycophenolic acid glucuronide AUC_0-12hr to mycophenolic acid AUC_0-12hr (Table 3). Specifically, the H1 haplotype ABCC2 CGC had significantly greater mycophenolic acid AUC_0-12hr, lower clearance and lower mycophenolic acid glucuronide AUC_0-12hr to mycophenolic acid AUC_0-12hr than the ABCC2 haplotype H9:CGT (Figure 2). These differences were most pronounced among patients receiving tacrolimus (Figure 3).

Interestingly, analysis of pharmacokinetic parameters based upon individual ABCC2 SNPs rather than haplotypes found mycophenolic acid AUC_0-12hr as the only significant pharmacokinetic phenotype associated with one SNPs (ABCC2 3972 C>T). Post-hoc analysis indicates that the homozygous CC genotype is significantly different for both the CT heterozygote and TT homozygote (Table 4).

Table 4.

ANOVA Analysis of Mycophenolic Acid Pharmacokinetic parameters for individual ABCC2 SNPs.

		Mycophenolic acid AUC0-12hr (hr●mg/L)		Mycophenolic acid Clearance (L/hr)		Ratio MPAG AUC0-12hr / MPA AUC0-12hr
Genotypes	N	Mean	StdDev	Mean	StdDev	Mean	StdDev
ABCC2 −24C>T (rs717620)
CC	110	55.44	29.30	16.21	10.32	28.94	21.70
CT	31	47.63	24.56	15.72	6.22	27.31	17.36
TT	6	51.38	24.37	13.33	4.52	29.27	13.19
All	147	53.63	28.21	15.99	9.41	28.61	20.49
		P = 0.39		P = 0.75		P = 0.93
ABCC2 1249 G>A (rs2273697)
GG	104	53.28	28.59	16.02	9.62	28.19	21.12
AG	39	54.39	28.08	15.76	9.15	28.73	19.03
AA	4	55.14	25.68	17.41	8.36	38.20	20.31
All	147	53.63	28.21	15.99	9.41	28.61	20.49
		P = 0.97		P = 0.95		P = 0.63
ABCC2 3972 C>T (rs 3740066)
CC	71	59.54	30.80	14.83	8.09	27.16	21.42
CT	55	48.89	26.12	17.13	11.60	29.12	19.62
TT	20	44.30	18.83	17.45	6.42	33.27	19.80
All	146	53.44	28.21	16.05	9.41	28.74	20.50
		P = 0.03		P = 0.31		P = 0.50
		LSD Post-hoc Test
			CC	CT
		CC
		CT	P = 0.034
		TT	P = 0.031	P = 0.53

Bonferroni corrected level of significance is 0.017 for multiple testing.

Discussion

This report describes the association of ABCC2 haplotypes to mycophenolic acid pharmacokinetics in stable kidney transplant recipients receiving concurrent calcineurin inhibitor immunosuppression. This is a follow-up investigation to a recent clinical pharmacology report from our group and the implications of interpatient variability in mycophenolic acid expsoure.³⁵ This report provides genomic insight into ABCC2 haplotypes that contribute to the interpatient variability of mycophenolic acid pharmacokinetic and clinical responses during chronic immunosuppression. The novel pharmacogenomic analysis of ABCC2 haplotype associations to mycophenolic acid pharmacokinetic parameters(phenotypes) includes exposure (AUC_0-12hr), clearance and the ratio of mycophenolic acid glucuronide AUC_0-12hr to mycophenolic acid AUC_0-12hr, as an objective endpoint of enterohepatic circulation. Further pharmacogenomic inter-relationships with mycophenolic acid pharmacokinetics are described in further detail in these recent reviews. ^12,17,14

Interestingly, 51% of the combined group (those receiving either cyclosporine or tacrolimus) were estimated to carry at least one copy of the wildtype ABCC2 haplotype, CGC. The mycophenolic acid AUC_0-12hr associated with this wildtype haplotype ranged from 27.83 to 34.30 mg●hr/L which is at the lower therapeutic range for mycophenolic acid target exposure of 30 to 60 mg●hr/L for all participants.^22,41. In contrast, individuals identified to have at least one copy of the reduced function variant haplotype, CGT (estimated haplotype frequency of 18%) were found to have lower mycophenolic acid AUC_0-12hr, more rapid clearance and a higher ratio of mycophenolic acid glucuronide AUC_0-12hr to mycophenolic acid AUC_0-12hr compared to the wild-type haplotype.³⁰ This variant haplotype, CGT is associated with sub-therapeutic mycophenolic acid AUC_0-12hr (phenotypic mean= 18.03 mg●hr/L) in the combined treatment groups. These findings could impact outcomes from maintenance calcineurin inhibitor and mycophenolic acid regimens where patient non-adherence, empiric dosing regimens, minimization protocols, lack of therapeutic monitoring and drug costs may result in sub-therapeutic mycophenolic acid immunosuppression based upon the variant haplotype, CGT.

Enterohepatic circulation of the mycophenolic acid glucuronide to the active drug, mycophenolic acid contributes to interpatient pharmacokinetic variability and is influenced by the concurrent calcineurin inhibitor and its dosing protocol^43–49 It is postulated that cyclosporine contributes to lower mycophenolic acid exposure by interfering with enterohepatic circulation possibly through inhibition of multiple resistance protein 2 efflux transporter resulting in elimination of the second peak of the active moiety.^19,20,50 This inhibition in enterohepatic circulation of mycophenolic acid glucuronide to mycophenolic acid does not appear to occur with concurrent tacrolimus.^24,25.

Sub-analysis of ABCC2 haplotypes and mycophenolic acid pharmacokinetics stratified by co-treatment with either cyclosporine or tacrolimus further reveal the interaction of ABCC2 haplotypes and calcineurin inhibitors. Specifically, the reduced mycophenolic acid AUC_0-12hr with increased clearance and mycophenolic acid glucuronide AUC_0-12hr to mycophenolic acid AUC_0-12hr ratio were more prominent with the CGT haplotype compared to the wildtype CGC (Figure 2, 3 and Table 3). These findings further underscore the clinical implications of under immunosuppression in the recipients with this ABCC2 variant during co-treatment with tacrolimus and the need to formalize consistent therapeutic drug monitoring of mycophenolic acid exposure during maintenance therapy. The exact mechanisms of the cyclosporine with mycophenolic acid and mycophenolic acid glucuronide pharmacokinetic interaction remains unclear with regards to the ABCC2 haplotype associations and the impact on multiple resistance protein 2 function. ^16,19–21 For instance, the effects of the ABCC2 variant haplotype were reduced in the cyclosporine group where enterohepatic recirculation was decreased and AUC_0-12 hr was sub-therapeutic in wild-type and variant groups (Table 3). In the tacrolimus sub-group, the association of the variant, CGT, exhibited reduced mycophenolic acid AUC_0-12hr (phenotypic mean= 20.1 mg•hr/L) and more rapid clearance compared to the wildtype sub-group. These findings provides objective confirmation of sub-therapeutic mycophenolic acid exposure in these stable recipients compared to suggested AUC_0-12 hr range of 30 to 60 mg•hr/L .^40,41 Since no therapeutic drug monitoring of mycophenolic acid troughs was conducted prior to or during these studies, no opportunity to identify recipients with sub-therapeutic mycophenolic acid exposure was incorporated into clinical assessment and may limit achieving optimal maintenance immunosuppression. These haplotypic differences in mycophenolic acid pharmacokinetics between calcineurin inhibitor co-treatment emphasize the need for individualized mycophenolic acid dosing regimens and therapeutic drug monitoring throughout maintenance immunosuppression, during calcineurin inhibitor conversion and with minimization protocols.^40,41

These findings are among the first to employ haplotypic association studies with the mycophenolic acid pharmacokinetics and provide important clinical and genomic perspectives in long-term stable transplant recipients. Of interest, the relationship of mycophenolic acid pharmacokinetic parameters to underlying genetics of ABCC2 was not well characterized by individual SNP analysis (Table 4). In using individual SNPs, only one significant association was identified for ABCC2 3972 C>T with mycophenolic acid AUC_{0-12 hr} . No significant associations were found in sub-analyses by calcineurin inhibitor. The inclusive utility of evaluating ABCC2 haplotypes and relationship to MPA pharmacokinetic parameters rather than individual SNPs was also noted by Ogasawara et al.³³ Similarly , in an earlier study in stable kidney transplant patients, investigation of tacrolimus pharmacokinetics and ABCB1 haplotypes was found to provide a more comprehensive appraisal of interpatient varability. ⁵¹

This prospective, observational pharmacokinetic-pharmacogenomic 12-hour study does have some limitations. Extending the single 12-hour “snapshot” of mycophenolic acid pharmacokinetics with ABCC2 haplotype estimation in these stable recipients during different post-transplant periods may provide additional insight to customize this immunosuppressive regimen. In addition, the inclusion of ongoing monitoring of mycophenolic acid exposure that incorporates formulation, concurrent calcineurin inhibitor treatment and adverse effects may clarify the specific therapeutic range during maintenance immunosuppression.^{13,14,35,40,41} Systematic quantitation of chronic adverse drug effects was determined in all patients with no ABCC2 haplotype associations found and may be due to the sample size. The use of the THESIAS computational approach to evaluate ABCC2 haplotypes is necessary due to the use of standard genotypic tests such as TaqMan Assays which only determine SNP genotypes individually. The THESIAS computational platform is not able to assign specific haplotypes (chromosomal arrangement of the three SNPs) to specific study participants. Generation of individual assigned haplotypes for each participant would involve considerable effort and expense that requires DNA sequencing. Finally, we have not adjusted the p-values for multiple testing using procedures like the Bonferroni correction since many study endpoints are inter-dependent and all were planned comparisons⁵². However, we have provided the Bonferroni adjusted P-values with each table.

Efflux transporters such as P-glycoprotein as well as multi-drug resistant protein 2 are members of the human super family of adenosine triphosphate (ATP)-binding cassette transporters (ABC) and play an essential role in mycophenolic acid and calcineurin inhibitor pharmacologic responses.^17,53 Tacrolimus and cyclosporine are substrates of P-glycoprotein which contribute to the cellular and pharmacokinetic variability of these immunosuppressants¹⁷. Both ABCB1 and ABCC2 that encode for these efflux transporters have been depicted in Figure 4. The three ABCB1 variants, which encodes P-glycoprotein (P-gp), that have commonly been studied are: rs1128503 (1236 C>T), rs2032582 (2677 G>T/A) and rs1045642 (3435 C>T). To date these ABCB1 variants when examined individually have not been shown to influence mycophenolic acid or mycophenolate glucuronide pharmacokinetics in kidney transplant recipients as summarized in the review by Genvigir et al.¹⁷ and Kagaya et al. ²⁷ This was substantiated in a prospective, randomized, multi-center study by Bouamar et al. that reported no associations of ABCB1 polymorphisms to diarrhea or leukopenia in more than 300 kidney transplant patients receiving mycophenolate mofetil with a calcineurin inhibitor ⁵⁴. Interestingly, Venuto et al. found 147 stable renal transplant recipients receiving mycophenolic acid and calcineurin inhibitor regimen that the sex of the patients and ABCB1 haplotypes were associated with common extrarenal adverse reactions³⁹. Although this report was not related to mycophenolic acid pharmacokinetics, these findings again suggest the utility of examining haplotypes rather than individual polymorphisms³⁹. The potential inter-relationship of these efflux transporters and impact on pharmacokinetics of mycophenolic acid and concurrent calcineurin inhibitor during maintenance immunosuppression require further investigation into the therapeutic exposure (i.e. AUC_0-12hr) achieved and implications for therapeutic drug monitoring.

Figure 4.

Diagram showing Mycophenolic Acid metabolic pathway (from Genvigir et al.¹⁷). The multidrug resistance-associated protein 2 (MRP2) encoded by the gene ABCC2 is responsible for the excretion of MPAG and Ac-MPAG into the bile. Both calcineurin inhibitors, cyclosporine and tacrolimus, are substrates for ABCB1, an efflux transporter that encodes P-glycoprotein. See Discussion. MPA – mycophenolic acid; MPAG – mycophenolic acid glucuronide; MMF – mycophenolic mofetil; AcMPAG – acyl-mycophenolic acid glucuronide; DM-MPA – 6-O-desmethyl mycophenolic acid; GMP – guanosine 5’-monophosphate; IMP – inosine 5’-monophosphate; XMP – xanthosine 5-monophosphate.

Conclusions

Variation in ABCC2 haplotypes contributes to sub-therapeutic mycophenolic acid exposures which impacts the interpatient variability in pharmacokinetic phenotypes when prescribed with different calcineurin inhibitors. Significant differences were seen in associations of the variant H2 to Wild-type ABCC2 haplotype with mycophenolic acid AUC_0-12hr, clearance and the ratio of mycophenolic acid glucuronide AUC_0-12hr to mycophenolic acid AUC_0-12hr. Individual carriers of the ABCC2 variant haplotype, CGT, had significantly lower mycophenolic acid AUC_0-12hr, greater clearance and higher mycophenolic acid glucuronide AUC_0-12hr to mycophenolic acid AUC_0-12hr ratio when compared to wildtype CGC. The differences in these pharmacokinetic parameters were most pronounced among patients receiving tacrolimus as the concurrent calcineurin inhibitor. Interpatient variability in mycophenolic acid pharmacokinetics was more accurately characterized with ABCC2 haplotypes rather than individual SNPs. Using scientific computation to generate the association of ABCC2 haplotypes to mycophenolic acid pharmacokinetic parameters provides more individualized characterization of this immunosuppressive exposure during maintenance therapy in stable transplant recipients. These differences in ABCC2 haplotypic associations to the interpatient variability in mycophenolic acid pharmacokinetics support the clinical utility of long-term therapeutic drug monitoring of this immunosuppressive exposure during transplant maintenance therapy to confirm adequate immunosuppression.

Data Accessibility Statement

Data will not be able to be shared.