Previously, we reported that, at 3 months after renal transplantation, individuals with CYP3AP1 genotype CYP3AP1*1 (linked to CYP3A5*1 and strongly associated with expression of CYP3A5) required twofold higher doses of tacrolimus to achieve target blood concentrations than individuals with the genotype CYP3AP1*3/*3 (CYP3A5 nonexpressors). This study assesses the relationship between concentration-controlled dosing during the early period after transplantation, the time to achieve target concentrations and genotype in 178 renal transplant recipients (CYP3AP1*1/*3 or *1/*1: n = 53, CYP3AP1*3/*3: n = 125). Patients with CYP3AP1*1/*3 or *1/*1 had lower mean tacrolimus concentrations during the first week (Median 13.5 vs. 18.5 μg/L, p < 0.0001) with significant delay in achieving target concentrations (15–20 μg/L during week 1, then 10–15 μg/L). More CYP3AP1*3/*3 patients had tacrolimus concentrations above target during the first week (73.6% vs. 35.8%, p = 0.003). There was no difference in the rate of biopsy-confirmed acute rejection, but rejection occurred earlier in the CYP3AP1*1/*3 or *1/*1 group (median 7 d vs. 13 d, p = 0.005). In conclusion, an initial dosing regimen for tacrolimus based on knowledge of the CYP3AP1 genotype and subsequently guided by concentration measurements has the potential to increase the proportion of patients achieving target blood concentrations early after transplantation.
Tacrolimus has a narrow therapeutic index and there is significant heterogeneity in the dose required to achieve target blood concentrations (1) which makes planning initial dosing difficult. The therapeutic range for tacrolimus has not been defined tightly, but there is evidence that the incidence of rejection increases with 12-h post-dose whole blood concentrations below 10 μg/L (2,3), and that toxicity increases at concentrations greater than 20 μg/L (4). Tacrolimus is metabolized by the oxidative enzymes cytochrome P450 (or CYP) 3A4 and 3A5. Metabolism in enterocytes acts as a barrier to absorption and hepatic metabolism is the principal route of drug clearance (5).
Both ethnic and genetic factors have been shown to influence the dose-requirement for tacrolimus. Black patients (1,6–8), and non-white South American patients (9) require higher doses to achieve target blood concentrations. In black patients this difference has been shown to be due to lower oral bioavailability (8). Marked heterogeneity in the level of expression of CYP3A4 has led to a search for the underlying genetic determinant. There has been extensive study of single nucleotide polymorphisms (SNP), in particular, CYP3A4*1B but there is some controversy as to whether they are predictive of the level of protein expression (10,11). Individuals with the mutant allele, CYP3A4*1B have been reported to achieve lower ciclosporin blood concentrations in one small study (12) but several other studies failed to identify this association (13–15). One study has identified an association between CYP3A4*1B and tacrolimus dose-requirement, although this difference may have been due to ethnicity as a confounding factor as the difference was not apparent when only white patients were considered (16).
CYP3A4 is thought to have been formed by a gene duplication from CYP3A5, resulting in proteins with very similar sequence and metabolic activity (17), to the extent that CYP3A5 has often been overlooked in studies of tacrolimus metabolism (5). CYP3A5 expression is polymorphic, the enzyme being either present or absent. A SNP in the CYP3A5 gene predicts hepatic and intestinal (18,19) CYP3A5 expression. Individuals with at least one wild-type allele (CYP3A5*1) express CYP3A5 and homozygotes for the mutant allele CYP3A5*3/*3 are nonexpressors.
The wild-type allele (G) at a SNP at position –44 in the CYP3AP1 pseudogene (CYP3AP1*1) is associated with CYP3A5 expression. This is thought to be through tight linkage with CYP3A5*1. The mutant allele CYP3AP1*3 (A) is in tight linkage with CYP3A5*3 with homozygocity predicting nonexpression of CYP3A5 (18). The data in this paper are based on the CYP3AP1 genotype, as in our previous publication (1).
The CYP3A5*1 allele is present in 70–80% of black individuals but only 5–10% of white individuals (1,20,21), and we believe this to be the major determinant of the well-described ethnic differences in tacrolimus dose-requirement (1).
Previously, we demonstrated that at 3 months after renal transplantation, individuals with the CYP3AP1*1 allele (CYP3A5 expressors) required twofold higher doses of tacrolimus than CYP3AP1*3/*3 homozygotes to achieve target blood concentrations (1). This observation has been confirmed with direct determination of the CYP3A5*1 genotype (16,22–24).
We (1), and others (22,25), have also demonstrated a similar but lesser influence on the tacrolimus dose-requirement of a SNP in the multiple drug resistance-1 gene (MDR-1), that encodes the drug efflux pump: P-glycoprotein. In a smaller study, Hesselink et al. were able to detect the influence of the CYP3A5 genotype on tacrolimus dose-requirement but identified no significant association with the MDR-1 genotype (16). Recently, we have reviewed the pharmacogenetics of immunosuppression for organ transplantation (26).
The risk of rejection in patients treated with tacrolimus was greatest in those who failed to achieve tacrolimus exposure equivalent to a trough concentration of 10 μg/L by 2 days after transplantation (3). More recently, data for another calcineurin inhibitor, ciclosporin suggest that achieving target blood concentrations as soon as possible after transplantation is a key factor in the prevention of rejection (27,28). We conducted this study to determine the relationship between the CYP3AP1 genotype and early blood tacrolimus concentrations and subsequent episodes of acute rejection. We report here that there was a significant delay in achieving the target blood tacrolimus concentration in CYP3A5 expressors, in spite of therapeutic drug monitoring.
Patients and Methods
Patients and treatment
All renal transplant recipients (kidney only) transplanted in our centre between 1995 and 2001, treated with tacrolimus, were invited to participate in this retrospective study, including those who had lost their grafts and returned to dialysis. Of the 278 patients transplanted during this period 178 were included in the study. The study was approved by the local research ethics committee and all subjects gave written informed consent. The demographics of the patient population are shown in Table 1. The majority of patients were treated with tacrolimus and prednisolone dual therapy with the addition of azathioprine in 44 cases or mycophenolate mofetil in 26 cases. Our standard steroid dosing regimen was 500 mg of methylprednisolone at the time of surgery, then 20 mg of prednisolone per day, reducing by 5 mg every 2 weeks to a maintenance dose of 5 mg daily.
Table 1. Demographic characteristics of patients at the time of transplantation
CYP3AP1*3/*3 (CYP3A5 nonexpressors)
CYP3AP1*1/*3 or *1/*1 (CYP3A5 expressors)
Age (years, median/range)
Body weight (kg, median/range)
Haemoglobin (g/dL, median/range)
Albumin (g/L, median/range)
Transplant number (1/2/3/4)
Donor age (years, median/range)
Cold ischaemia time for cadaveric kidneys (hours, median/range)
Tacrolimus whole blood concentrations were measured 12 h post-dose, using an immunoassay (Tacrolimus II, Abbott Diagnostics, Abbott Park, IL, USA, performed on an IMx clinical analyser). The same assay was used throughout the course of the study and the laboratory was a member of the International Tacrolimus Proficiency Testing Scheme. Target concentrations were 15–20 μg/L during the first 7 d then 10–15 μg/L until 3 months after transplantation. Our standard protocol for tacrolimus-dosing was to give an initial oral dose of 0.2 mg/kg followed by 0.1 mg/kg twice daily, with dose adjustment based on three times weekly measurement of blood tacrolimus concentrations. The daily tacrolimus dose was adjusted up or down by 20% when results fell outside the target range.
Genomic DNA was extracted from whole blood using a QIAamp DNA Mini Kit (Qiagen, Crawley, UK). Polymerase chain reaction (PCR) followed by restriction fragment length polymorphism analysis (RFLP) was employed for genotyping multiple drug resistance gene 1 (MDR1) and CYP3AP1. PCR primers for CYP3AP1 were designed to amplify a 391 bp fragment of CYP3AP1 (forward primer 5′-GGGGATGGATTTCAAGTATTCTG-3′ and reverse primer 5′-GTCCATCGCCACTTGCCTCT-3′). This was followed by enzymatic digestion of PCR amplification products, which was performed using the AciI endonuclease.
Genotyping of MDR-1 at the exon 26 C3435T SNP was performed using forward: 5′-TGCTGGTCCTGAAGTTGATCTGTGAAC-3′ and reverse 5′-ACATTA[GGCAGTGACTCGATGAAGGCA-3′ primers and the use of MboI endonuclease (29).
Rejection was confirmed, where possible, by biopsy. The data shown are for episodes of biopsy-confirmed rejection of at least Banff 1 in severity (1997 classification) (30).
Differences between genotypes were compared using the Mann–Whitney U-test or chi-square test using SPSS version 11 (SPSS, Chicago, IL, USA).
Mean tacrolimus blood concentration over the first 2 weeks
The mean tacrolimus blood concentration during the first and second weeks was calculated for each patient. The mean tacrolimus blood concentration (μg/L) for each patient was significantly lower for individuals possessing CYP3AP1*1, in spite of therapeutic drug monitoring in both the first [median: 13.5, interquartile range: 9.3–17.5 vs. 18.5 (14.7–21.8), Mann–Whitney U p < 0.0001] and second weeks [11.3 (9.3–14.1) vs. 14.3 (11.6–6.1), p = 0.002] after transplantation (Figure 1). The dose normalized tacrolimus concentration (blood concentration divided by the tacrolimus dose in mg/kg multiplied by 10 to give the effective blood concentration achieved per 0.1 mg/kg daily dose) was significantly lower in individuals possessing CYP3AP1*1 at both 7 and 14 d after transplantation (Table 2).
Proportion of patients who failed to achieve target blood tacrolimus concentrations
The majority of CYP3AP1*3/*3 homozygotes achieved the target concentration within the first 2 weeks, but there was a significant delay for those with the CYP3AP1*1 genotype (Table 3). It could be argued that 15 μg/L is an excessive target tacrolimus concentration during the first week after transplantation (2,3). However, data are also shown here for a target of 10 μg/L which still shows a significant delay in achieving target concentration in individuals with the CYP3AP1*1 genotype (Table 3). There were not significantly more changes in drug dose during the first 3 months after transplantation between the genotypes: 9 ± 3.5 (mean ± SD) for CYP3AP1*3/*3 homozygotes and 12 ± 3.8 for CYP3AP1*1.
Table 3. The impact of the CYP3AP1 genotype on patients achieving target blood tacrolimus concentrations
% of patients with tacrolimus blood concentrations outside the target range
*1/*1 or *1/*3
*1/*1 or *1/*3
All blood tacrolimus concentrations
<15 μg/L (week 1)
<10 μg/L (week 2)
All blood tacrolimus concentrations
All blood tacrolimus concentrations
<15 μg/L (week 1)
<10 μg/L (week 2)
At least one blood tacrolimus concentration
>20 μg/L (week 1)
>15 μg/L (week 2)
The genetic difference was more pronounced when only white patients were studied
Looking only at white patients, the difference was even more marked, with over 70% of patients expressing a CYP3AP1*1 allele failing to reach target concentrations during the first week (Table 3).
Patients with high blood concentrations
A greater proportion of CYP3AP1*3/*3 had at least one blood tacrolimus concentration above the target range of 20 μg/L during the first week, with no significant difference during the second week (Table 3).
The overall proportion of patients experiencing a biopsy-confirmed episode of acute rejection of at least Banff 1 in severity was no different between the groups over the first 3 months [CYP3AP1*3/*3: 52/125 (41.6%) vs. CYP3AP1*1/*3 or *1/*1: 24/53 (45.3%)]. However, the timing of rejection was significantly different, with the mean time of first rejection being 7 d for individuals with a CYP3AP1*1 allele compared with 13 d for CYP3AP1*3/*3, Mann–Whitney U p = 0.005 (Figure 2). There was no difference between the groups in severity of the episodes of rejection. CYP3AP1*3/*3 Banff 1: 28.8%, Banff 2: 8.8%, Banff 3: 4.0%; CYP3AP1*1/*3 or *1/*1; Banff 1: 35.8%, Banff 2: 3.8%, Banff 3: 5.6%. The results were similar when only the white patient population was analysed, with a 41.7% rejection rate at a median of 12 d for the CYP3AP1*3/*3 group and 46.7% at a median of 8 d for the CYP3AP1*1/*3 or *1/*1 group. The rejection rate for the whole population of patients (n = 278) transplanted in our centre during the period of the study was 50.4%.
On the day the first episode of rejection was diagnosed, the tacrolimus blood concentration was significantly lower in the CYP3AP1*1/*3 or *1/*1 group than in the CYP3AP1*3/*3 group (median 9.4 μg/L vs. 14.2 μg/L, Mann–Whitney U p = 0.001). In the rejecting patients 13/24 (54.2%) of CYP3AP1*1/*3 or *1/*1 patients had a blood tacrolimus concentration of less than 10 μg/L compared with only 6/52 (11.5%) in the CYP3AP1*1/*3 or *1/*1 group (chi-square p < 0.0006).
There was no significant difference between the MDR-1 C3435T genotypes in the mean blood tacrolimus concentration during the first 2 weeks after transplantation, the proportion of patients achieving either target or high blood concentrations, or the rate of rejection (data not shown).
We have demonstrated that individuals predicted to express CYP3A5 on the basis of their CYP3AP1 genotype experienced a significant delay in achieving target blood tacrolimus concentrations, in spite of therapeutic drug monitoring. Previously we observed twofold lower dose-normalized tacrolimus concentrations at 3 months after transplantation in individuals with a CYP3AP1*1 allele (1), a finding that was confirmed during the first 2 weeks after transplantation here. The observation that the proportion of CYP3A5 expressors who failed to reach target blood tacrolimus concentrations during the first 2 weeks after transplantation was more pronounced when only white patients were included is intriguing. A possible explanation is that the linkage between CYP3AP1*1 and CYP3A5*1 may be less complete in black patients. It is also possible that the presence of the null allele CYP3A5*6 which is present in 10% of black patients and very few white patients (18,20,21) and can coexist with CYP3A5*1 changing an expressor genotype to a nonexpressor resulting in incorrect assignment of CYP3A5 expressor status in some of the black patients. We did not type for this SNP. We (1) and others (22,25), have shown a significant association between the MDR-1 genotype and tacrolimus dose requirement. We found no significant association between the MDR-1 exon 26 SNP (C3435T) and the mean tacrolimus blood concentration over the first 2 weeks after transplantation, or the time taken to achieve target blood concentrations (data not shown). However, we only typed for the synonymous SNP and not the whole haplotype, which may correlate better with protein expression.
It has been known for some time that some ethnic groups require higher doses of tacrolimus to achieve target blood concentrations (1,6–9). Some centres, including our own (after the period of this study), have adopted the practice of giving twofold higher initial doses of tacrolimus to black patients. However, our data emphasize that skin colour is a crude marker of the genotype influencing drug metabolism. In fact, in our practice, we actually transplant more white CYP3A5 expressors than black patients with this genotype. The other immunosuppressive drugs, ciclosporin and sirolimus are also CYP3A4/A5 and P-gp substrates and the principles discussed here may also be relevant to the prescription of these drugs. An association has been described between ciclosporin oral bioavailability and the CYP3A5 genotype (16,31), but this is controversial (15). The practice of adjusting the initial drug dose on the basis of genotype may be particularly beneficial for drugs with a long half-life, such as sirolimus, for which therapeutic drug monitoring has a long response-time.
Associated with lower tacrolimus blood concentrations early after transplantation, episodes of acute rejection occurred earlier in individuals with CYP3A5*1, but there was no difference in the overall rate of rejection. This study probably did not have sufficient statistical power to detect any difference in the rejection rate.
In conclusion, we have demonstrated that the CYP3AP1*1 genotype that is predictive of CYP3A5 expression identifies patients in whom there will be a delay in achieving target blood tacrolimus concentrations. This suggests a possible pharmacogenetic approach to drug dosing, resulting in the administration of twofold higher starting doses to CYP3A5 expressors. This hypothesis should be tested in a prospective randomized controlled trial with genotyping at the CYP3A5*1 SNP.
We would like to acknowledge the assistance of Dr Sue Snowden in patient recruitment and Michelle Moreton for technical assistance. The renal transplants were performed by Mr Rene Chang and Mr Michael Bewick.