Pharmacogenetic association with adverse drug reactions to azathioprine immunosuppressive therapy following liver transplantation



Azathioprine (AZA) is a thiopurine prodrug commonly used in triple-immunosuppressive therapy following liver transplantation. Approximately 1 in 10 patients suffers side effects in response to the drug, the most problematic being bone marrow toxicity. There is evidence that polymorphisms in the genes encoding thiopurine methyltransferase (TPMT) and inosine triphosphate pyrophosphatase (ITPase) predict adverse drug reactions to AZA therapy. Furthermore, common genetic polymorphisms in the gene encoding methylenetetrahydrofolate reductase (MTHFR) may have an indirect impact on thiopurine drug methylation by influencing levels of the methyl donor S-adenosylmethionine (SAM). The aim of this study was to determine whether polymorphisms in these candidate pharmacogenetic loci predict adverse drug reactions to AZA immunosuppressive therapy in liver transplant patients. A series of 65 liver transplant recipients were recruited to the study from the Liver Transplant Out-Patient clinic at The Royal Infirmary of Edinburgh. Clinical response to AZA was retrospectively correlated against TPMT activity, TPMT*2, *3A, and *3A genotypes, inosine triphosphatase (ITPA) 94C>A and IVS2+21A>C genotypes, and MTHFR 677C>T and 1298A>C genotypes. Variant TPMT, ITPA, and MTHFR genotypes were not significantly associated with adverse drug reactions to AZA, including bone marrow suppression. However, the 2 patients who suffered nodular regenerative hyperplasia (NRH) were both heterozygous for the TPMT*3A mutation. In conclusion, our findings suggest that TPMT, ITPA, and MTHFR genotypes do not predict adverse drug reactions, including bone marrow suppression, in liver transplant patients. However, the possible association between NRH and a heterozygous TPMT genotype should be investigated further. (Liver Transpl 2005;11:826–833.)

The standard immunosuppressive regimen after liver transplantation at the Scottish Liver Transplant Unit in Edinburgh is azathioprine (AZA) (the prodrug of 6-mercaptopurine [6-MP]) combined with tacrolimus and prednisolone.1 As an immunosuppressant, thiopurine drugs are used to treat patients with acute lymphoblastic leukemia, inflammatory bowel disease, rheumatoid arthritis, and certain dermatological conditions.2 Unfortunately, AZA has to be withdrawn in 9% to 23% of cases due to side effects,3, 4 and of these, bone marrow toxicity is of the most concern. The aim of this study was to investigate the association between pharmacogenetic loci implicated in thiopurine drug metabolism and adverse drug reactions to AZA therapy in a consecutive series of liver transplant recipients attending the Scottish Liver Transplant Unit in Edinburgh.

AZA is cleaved in the liver, chiefly by glutathione, to form 6-MP and imidazole. 6-MP is then extensively metabolized (Fig. 1) to form 6-thioguanine nucleotides. Immunosuppression is thought to occur via 2 mechanisms; incorporation of 6-thioguanine nucleotides into ribonucleic acid and deoxyribonucleic acid,5 and specific inhibition of Rac1 activation by 6-thioguanosine triphosphate binding.6 In this study, polymorphisms in 3 pharmacogenetically important genes were considered as candidates to explain adverse drug reactions to thiopurine therapy in liver transplant patients.

Figure 1.

Metabolism of azathioprine (AZA), 6-mercaptopurine (6-MP), and 6-thioguanine. On absorption, AZA is rapidly converted to 6-MP. 6-MP availability is reduced either by methylation to 6-methylmercaptopurine (6MeMP), catalyzed by TPMT or conversion to thiouric acid by xanthine oxidase (XO). Cytotoxic thioinosine or guanine nucleotides are formed via the salvage enzyme hypoxanthine-guanine phosphoribosyl transferase (HGPRT). These include 6-thioinosine monophosphate (6tIMP), 6-thioxanthosine monophosphate (6tXMP), 6-thioguanine mono-, di-, and triphosphate (6tGMP, 6tGDP, and 6tGTP), 6-methylthioguanine (6Me-tG), and 6-deoxythioguaninetriphosphate (6t-dGTP). IMPDH, inosine monophosphate dehydrogenase; GMPS, guanine monophosphate synthase; RNR, ribonucleotide reductase.


Genetic polymorphisms in the thiopurine methyltransferase (TPMT) gene (TPMT*2-*13) are known to be associated with deficient TPMT activity,7–10 impaired methylation of thiopurine drug metabolites, increased formation of cytotoxic 6-thioguanine nucleotides,11–13 and a high risk of adverse drug reactions in patients treated with normal doses of AZA.9, 14–18 Approximately 11% of the Caucasian population has intermediate TPMT activity and are heterozygous for a variant allele, most commonly TPMT*3A or TPMT*3C.19 About 1 in 300 patients is homozygous for variant TPMT alleles and is at high risk of severe, sometimes fatal, bone marrow suppression.20 Bone marrow suppression can also be fatal in heterozygous TPMT patients.21


Inosine triphosphate pyrophosphatase (ITPase) deficiency is a clinically benign condition characterized by the marked and abnormal accumulation of inosine triphosphate (ITP) in erythrocytes. ITP is formed by phosphorylation of inosine monophosphate (IMP). In normal cells, ITPase converts ITP back to IMP so that ITP does not accumulate (Fig. 1). 6-MP is activated through a 6-thioinosine monophosphate intermediate and in ITPase-deficient patients being treated with thiopurine drugs, the metabolite 6-thioinosine monophosphate is predicted to accumulate. The inosine triphosphatase (ITPA) 94C>A polymorphism is associated with approximately 25% red blood cell enzyme activity in carriers,22 and has been reported to be associated with side effects to AZA therapy in a retrospective study of patients with inflammatory bowel disease.23 A second polymorphism in the ITPA gene (IVS2+21A>C) results in approximately 60% wild-type activity in homozygotes,22 but has not been reported to be associated with AZA toxicity.23


S-adenosylmethionine (SAM) is the methyl donor for the methylation reaction catalyzed by TPMT, after which it is converted to S-adenosylhomocysteine. Polymorphisms in enzymes catalyzing S-adenosylhomocysteine recycling may thus indirectly impact on TPMT activity and the capacity to methylate thiopurine drug metabolites. The adenosyl moiety of S-adenosylhomocysteine is subsequently cleaved and homocysteine can be remethylated to methionine. The methyl donor for the folate-dependent remethylation cycle is 5-methyltetrahydrofolic acid, which is formed from 5,10-methylenetetrahydrofolate in a reaction catalyzed by methylenetetrahydrofolate reductase (MTHFR). Two common polymorphisms in the MTHFR gene (the thermolabile 677C>T variant24 and the 1298A>C variant25) are associated with decreased MTHFR activity and may influence the recycling of S-adenosylhomocysteine to SAM26 and hence influence functional TPMT activity. The homozygous MTHFR 677TT genotype occurs in 8% to 10% of the general population27 and is associated with 30% wild-type activity.28 A total of 10% of the general population have the homozygous 1298CC genotype,27 which is associated with approximately 60% wild-type activity in lymphocytes.25 Compound heterozygotes also have significantly decreased enzyme activity.29

In this retrospective study of liver transplant recipients attending the Scottish Liver Transplant Unit, we report the association between TPMT activity and genotype, common polymorphisms in the ITPA and MTHFR genes, and clinical response in patients treated with azathioprine as part of an immunosuppressive regime.


AZA, azathioprine; TPMT, thiopurine methyltransferase; ITPase, inosine triphosphate pyrophosphatase; MTHFR, methylenetetrahydrofolate reductase; SAM, S-adenosylmethionine; ITPA, inosine triphosphatase; NRH, nodular regenerative hyperplasia; 6-MP, 6-mercaptopurine; 6MeMP, 6-methylmercaptopurine; ITP, inosine triphosphate; IMP, inosine monophosphate.

Patients and Methods


A series of 65 liver transplant recipients were recruited to the study from the Liver Transplant Out-Patient Clinic at The Royal Infirmary of Edinburgh between March and May 2003. One week prior to their scheduled appointment, letters and information sheets were sent out to patients asking them to participate in the study. When they arrived at the clinic, all questions regarding the study were answered and written consent was obtained. The study was approved by the Healthy Volunteers / Student Research Ethics Committee on behalf of National Health Service Lothian.

All patients were on similar immunosuppressive regimens. The primary immunosuppressants used were tacrolimus or cyclosporine. Drug dosages were adjusted according to trough blood concentration, time since transplant, history of rejection, and presence of side effects (particularly renal impairment). In all patients, AZA was also started immediately after transplant at a dose of around 1 mg/kg once daily. Prednisolone was used initially as an immunosuppressant in all patients, but the dose was tapered and withdrawn over the 1st 3 months (unless the patient had autoimmune chronic active hepatitis or acute graft rejection).

Clinical Criteria for Defining Side Effects

Adverse drug reactions were defined as those judged severe enough to warrant withdrawal of AZA therapy or for the drug dosage to be reduced (as judged by the consultant physician at the follow-up clinic). Only side effects that were explicitly ascribed to AZA in the clinic letter were recorded. Bone marrow suppression was defined as a white blood cell count of less than 2 × 109/L, or a platelet count of less than 40 × 109/L.

Laboratory Investigations

TPMT Phenotyping

Red blood cell TPMT activity was measured in ethylenediamine tetraacetic acid (EDTA) whole blood (collected after liver transplantation) as the conversion of 6-MP to 6-methylmercaptopurine (6MeMP) using a tandem mass spectrometry method developed in the Purine Research Laboratory. TPMT activity was recorded according to the following reference ranges: deficient ≤10 pmol/hr/mg of hemoglobin, carrier 11–25 pmol/hr/mg of hemoglobin, normal 25–50 pmol/hr/mg of hemoglobin, and high >50 pmol/hr/mg of hemoglobin.

Deuterated [d3]-6MeMP was synthesized by QMX Chemicals (Thaxted, UK). Substrates and buffers used were analytical reagent grade from either Merck-BDH (Poole, UK) or Sigma-Aldrich (Poole, UK). Drabkin's reagent and standard hemoglobin were obtained from Sigma-Aldrich and prepared and stored as instructed by the manufacturer. Stock solutions comprised: 150 mmol/L potassium phosphate buffer, pH 7.5; 10 mmol/L dithiothreitol; 18 mg/mL 6-MP (Sigma, Poole, UK) dissolved in dimethyl sulfoxide; 1 mg/mL paratoluene sulfonate salt of SAM dissolved in 0.005 mol/L sulfuric acid; .1 mol/L zinc sulfate. A stock solution of 10 mmol/L [d3]-6MeMP (QMX Chemicals) was prepared in dimethyl sulfoxide and the concentration confirmed by ultraviolet spectrophotometry. To prepare the internal standard for assay, 30 μL [d3]-6MeMP stock was added to 100 mL acetonitrile (Rathburn Chemicals, Walkerburn, UK). Briefly, 100 μL of a 1:6 hemolysate of washed, packed, red blood cells was pipetted into duplicate wells of a deep well microtiter plate. A total of 50 μL of the assay mix (25 μL phosphate buffer, 5 μL SAM, 15 μL dithiothreitol, 5 μL 6-MP) was added to each well. The microtiter plate was capped and incubated for 4 hours at 37°C in a shaking incubator. The enzyme reaction was stopped by the addition of 40 μL of zinc sulfate solution and 500 μL of the acetonitrile standard. After vortexing, the plate was centrifuged for 10 minutes at 3000g. Hemoglobins were determined using Drabkin's reagent. The product of the reaction, 6MeMP, was detected by tandem mass spectrometry and quantified relative to the internal deuterated 6MeMP standard. Results were expressed as pmol/hr/mg of hemoglobin.

TPMT and ITPA Genotyping

DNA was extracted from blood samples using a QIAamp deoxyribonucleic acid blood kit (Qiagen, Crowley, UK). Restriction endonucleases were obtained from New England Biolabs, Hitchin, UK. All patients were genotyped for TPMT*2, TPMT*3A, TPMT*3C, ITPA 94C>A, and ITPA IVS2+21A>C as previously reported.23

MTHFR Genotyping

The MTHFR 677C>T variant in exon 4 creates a HinfI site, which was amplified using primer 677for 5′-CCCAGCCACTCACTGTTTTAGTTC-3′ and 677rev 5′-CCAAAGTACAACAAACCCCTCAAC-3′. The thermocycler profile consisted of 35 cycles of 94°C for 30 seconds, 48°C for 30 seconds, and 72°C for 30 seconds. The 579-bp PCR product was digested with HinfI at 37°C for at least 3 hours. HinfI recognizes a second nonpolymorphic site in the wild-type sequence that acts as a control for complete digestion. For the wild-type sequence, 2 fragments will always be present with lengths 72 and 507 bp.

The mutation in exon 7 of the MTHFR gene is a substitution at position 1298 from adenine to cytosine [G429A]. This mutation creates a MwoI site amplified using primer 1298for 5′-CATGTGGTGGCACTGCCCTCTG-3′ and the 1298rev mismatch 5′-CGAGAGGTAAAGAACAAAGACTTCAGCGACAC-3′. The 214-bp fragment was digested with MwoI at 60°C for at least 3 hours. The mismatch in the reverse primer creates a MwoI recognition site in the presence of the mutation giving fragments of 180 and 34 bp. Thermocycler profile consisted of 35 cycles of 94°C for 30 seconds, 50°C for 30 seconds, and 72°C for 30 seconds.

Statistical Methods

The significance of the association between adverse drug reactions and polymorphisms in each gene was tested using a 2-sided Fisher's exact test; odds ratios and 95% confidence intervals were calculated from contingency tables.


The demographics of the 65 liver transplant recipients recruited to the study are shown in Table 1. Immunosuppressive regimens included prednisolone, azathioprine, and either tacrolimus (71%) or cyclosporine (29%). A total of 6% of patients were prescribed mycophenolate mofetil.

Table 1. Demographics of AZA-Tolerant and AZA-Intolerant Patients
 AZA-tolerant 38AZA-intolerant 27
 Male63% (24)37% (10)
 Female37% (14)63% (17)
Age, mean ± SD57 ± 1058 ± 13
Reason for transplant
 Primary biliary cirrhosis26% (10)44% (12)
 Alcoholic liver disease24% (9)22% (6)
 Primary sclerosing cholangitis21% (8)0%
 Cryptogenic liver cirrhosis5% (2)11% (3)
 Chronic hepatitis C infection11% (4)4% (1)
 Others13% (5)19% (5)
Number of years since transplant
 Within 1 year34% (13)19% (5)
 1-3 years11% (4)19% (5)
 3-5 years29% (11)11% (5)
 5-7 years5% (2)11% (5)
 7-9 years8% (3)30% (8)
 More than 9 years13% (5)11% (3)

AZA therapy was reduced or withdrawn in 42% (27 of 65) of patients due to side effects. Significantly more males (63%) tolerated AZA than females (P = 0.0468). There was no significant difference in patient age between the 2 groups. Reasons for liver transplantation between tolerant and intolerant groups were similar. Side effects resolved in all intolerant patients after AZA dose reduction or withdrawal. The majority of patients experiencing adverse drug reactions suffered bone marrow suppression (15 patients). Joint pain occurred in 6 patients. Two patients suffered severe fatigue, 2 patients suffered biopsy-proven nodular regenerative hyperplasia (NRH), 1 patient suffered headaches. and 1 patient suffered renal impairment.

Table 2 shows that median TPMT activity was not significantly different between tolerant (median: 34.5 units, range: 18.3-59.8 units) and intolerant groups (median: 35.2 units, range: 13.8-53.4 units, Mann-Whitney rank sum test, P = 0.599). One patient with intermediate TPMT activity had a wild-type TPMT genotype. A heterozygous TPMT genotype was not significantly associated with AZA side effects (dominant model, 2-sided Fisher's exact test, P = 1.000). However, the 2 patients who suffered from NRH were both heterozygous for the TPMT*3A mutation.

Table 2. Phenotype and Genotype Frequencies (Number of Patients) in AZA-Tolerant and AZA-Intolerant Groups*
Side effects groupNumber of patientsMean TPMT activityTPMT*3AMTHFR 677MTHFR 1298ITPA 94ITPA IVS
  • *

    Significance was tested using a dominant model for TPMT and ITPA variants, and a recessive model for MTHFR variants.

 WT  0.87 (33)0.45 (17)0.47 (18)0.84 (32)0.74 (28)
 Hetero  0.13 (5)0.42 (16)0.39 (15)0.16 (6)0.26 (10)
 Homo  00.13 (5)0.13 (5)00
Marrow toxicity1538     
 WT  1.0 (15)0.33 (5)0.33 (5)0.8 (12)0.8 (12)
 Hetero  00.53 (8)0.67 (10)0.2 (3)0.07 (1)
 Homo  00.13 (2)000.13 (2)
 P value  0.30551.00000.30550.70100.7355
Other side effects1231     
 WT  1.0 (6)0.33 (2)0.83 (5)0.83 (5)0.83 (5)
 Hetero  00.67 (4)0.17 (1)0.17 (1)0.17 (1)
 Homo  00000
 P value  1.00001.00001.00001.00001.0000

The ITPA 94C>A mutation, which results in <25% residual activity, did not predict side effects to AZA therapy (dominant model, P = 1.000). Similarly, the ITPA IVS2+21A>C mutation, which is associated with 60% residual enzyme activity, was not significantly associated with AZA intolerance (dominant model, P = 0.7766; recessive model, P = 0.1688).

Neither the MTHFR 677TT genotype, the MTHFR 677T/1298C compound heterozygous genotype, nor the 1298CC variant genotype was significantly associated with AZA intolerance (P = 1.000, P = 1.000, and P = 0.6899, respectively).

No variant genotype was significantly associated with the group of 15 patients suffering bone marrow suppression when compared to AZA-tolerant patients (P = 0.3055). No variant genotype was significantly associated with other side effects (joint pain, fatigue, NRH, renal impairment, headache) when compared to AZA-tolerant patients.

Seven patients were retransplanted, either because of tissue rejection or postoperative complications (Table 3). The variant genotype frequencies in retransplanted patients were not significantly different from AZA-tolerant patients.

Table 3. TPMT Activity and Variant Genotype Frequencies in Patients Who Underwent Retransplantation
Patient groupNumberMean TPMT activity TPMT *3AMTHFR 677MTHFR 1298ITPA 94ITPA IVS
Retransplanted735.1WT0.86 (6)0.57 (4)0.71 (5)0.71 (5)0.86 (6)
   Hetero0.14 (1)0.29 (2)0.14 (1)0.29 (2)0.14 (1)
   Homo00.14 (1)0.14 (1)00


Phenotyping or genotyping for TPMT deficiency in order to predict AZA toxicity is a classic application of pharmacogenetics in clinical medicine and is being increasingly utilized when planning thiopurine treatment for patients with acute lymphoblastic leukemia,30 certain dermatological conditions,31 and inflammatory bowel disease.32–34 Screening patients for TPMT deficiency prior to initiation of therapy is believed to be beneficial from a health economics point of view, given the considerable cost of treating bone marrow failure and other AZA side effects.35, 36

The important findings of this study were that recipient TPMT, ITPA, and MTHFR variant genotypes were not significantly associated with AZA intolerance. Nor were any of the variant genotypes associated with bone marrow suppression. However, the 2 patients who suffered biopsy-proven NRH, which is a significant concern in patients being treated with AZA37 and 6-thioguanine,38 were both heterozygous for the TPMT*3A mutation.

NRH is well described following orthotopic liver transplantation. One previous study reported 9 cases of NRH among liver transplant recipients.39 Six of these patients exhibited features of portal hypertension, whereas 3 were asymptomatic. All patients had elevated serum alkaline phosphatase and gamma glutamyl transferase, and had been taking AZA up until the time of presentation. After withdrawal of AZA, liver function tests improved appreciably in 5 patients and histological improvement was documented in 4 patients on follow-up liver biopsy. NRH of the liver associated with AZA therapy has also been reported in renal transplant recipients,40 inflammatory bowel disease patients,38 and multiple sclerosis patients.41 The TPMT genotype has never been reported in any of these studies. Further studies are needed to confirm the association between NRH and heterozygous TPMT genotype.

ITPase deficiency occurs with polymorphic frequency in Caucasian, African, and Asian populations.42 The enzyme is not essential for life and forms part of a futile cycle that recycles ITP from IMP. ITPase deficiency results in the accumulation of ITP in red blood cells, and deficiency of the enzyme has also been demonstrated in nucleated cells.43 Within the context of thiopurine drug metabolism, 6-MP is activated through a 6-thioinosine monophosphate intermediate. 6-thioinosine triphosphate is formed in red blood cells44 and inheritance of the ITPA 94C>A allele is predicted to result in accumulation of 6-thioinosine triphosphate, a potentially toxic metabolite. This polymorphism has been shown to be associated with adverse drug reactions to AZA therapy in patients with inflammatory bowel disease.23 However, the current study of liver transplant recipients showed no association between ITPase deficiency and AZA intolerance.

Methylated thiopurine metabolites generated by TPMT are thought to have immunosuppressive properties7 and have been implicated in toxicity.37 Any genetic polymorphism that indirectly affects the availability of the methyl donor SAM, and hence the capacity of TPMT catalyzed methylation reactions, may thus influence outcome of thiopurine therapy. The homozygous MTHFR 677TT genotype is associated with 30% residual activity,28 and studies in the MTHFR knockout mouse model have predicted decreased levels of SAM and an altered SAM/S-adenosylhomocysteine ratio.45 There is convincing evidence that the MTHFR 677TT genotype is also associated with deoxyribonucleic acid hypomethylation,45 an increased risk of neural tube defects,46 and a decreased risk of some cancers.47 However, our study showed no association between variant MTHFR genotypes and adverse drug reactions to AZA therapy.

The genotyping and phenotyping performed in this study reflects the transplant recipient's genetic makeup. As the liver is the principal site of 6-MP methylation,48, 49 it is possible that it is in fact the donor genotype that predicts whether or not patients will suffer side effects in response to AZA therapy. Unfortunately, donor liver material was not available for this study.

The series of patients studied was representative of the liver transplant patients seen in Edinburgh, with a similar range of underlying causes of liver disease. Mean and median TPMT activities and genotype frequencies were not significantly different from previously published figures. We are aware that our retrospective study may be biased. First, the number of patients transplanted “within the last year” was overrepresented, which means that many patients may not yet have had the chance to experience AZA side effects. It is also possible that some patients died as a result of AZA side effects and so had no chance of being included in the study. This would lead to overrepresentation of patients who were able to tolerate AZA. Conversely, the reason for our patients' visit to the clinic may have been because of their side effects, leading to overrepresentation of this patient group.

A total of 42% of patients in our study suffered side effects that caused AZA therapy to be reduced or withdrawn. This is higher than has been reported elsewhere and is probably attributable to the differences in defining adverse drug reactions between these studies. Bone marrow suppression and joint pain were the most commonly observed adverse drug reactions in our study. We did not observe some of the other recognized AZA side effects such as nausea and vomiting, pancreatitis, rash, and hepatotoxicity. These may have been overlooked due to the retrospective nature of the study. Alternatively, these side effects may not manifest at the lower AZA doses used in triple immunosuppressive therapy (approximately 1 mg/kg compared to 1.5-2 mg/kg in inflammatory bowel disease2).


Recipient TPMT, ITPA, and MTHFR genotypes do not predict adverse drug reactions to AZA therapy in liver transplant recipients. Further studies with larger sample sizes are necessary to confirm these findings. The finding that both patients with NRH were heterozygous for the TPMT*3A mutation is interesting and warrants further investigation. Further studies are needed to determine whether prospective genotyping of donor livers can predict adverse drug reactions to AZA following liver transplantation.


We thank everyone at the Scottish Liver Transplant Unit and the Royal Infirmary of Edinburgh Clinical Biochemistry Department for their help and support. Thanks also to the patients who agreed to participate in the study.

David Breen recruited patients to the study, collected the clinical data, and wrote the paper. Anthony Marinaki developed the idea for the study, carried out the statistical analysis, and wrote the paper. Monica Arenas carried out the laboratory work. Peter Hayes initiated the study and will act as guarantor for the paper.