Pharmacogenomics and IBD

6-Mercaptopurine and its pro-drug azathioprine are effective in inducing and maintaining remission in patients with ulcerative colitis and Crohn's disease. 1,2 The metabolism of these drugs is complex, and practicing clinicians need to become familiar with their clinical pharmacology. 3,4 This article reviews the clinical pharmacology and pharmacogenomics of azathioprine (AZA) and 6-mercaptopurine (6MP) therapy in patients with inflammatory bowel disease (IBD). Parent AZA and 6MP are not therapeutically active. 3 Azathioprine is a pro-drug that is converted to 6MP via a non-enzymatic nucleophilic attack by sulfhydryl-containing compounds such as glutathione present in erythrocytes and other tissues. 3 Then, 6-Mercaptopurine is metabolized by thiopurine methyltransferase (TPMT) to 6-methylmercaptopurine (6MMP) or by xanthine oxidase to 6-thiouric acid. 3,5 Alternatively, 6MP may be activated via a multistep enzymatic pathway beginning with the enzyme hypoxanthine phosphoribosyl transferase followed by inosine triphosphate pyrophosphatase (ITP) and guanosine monophosphate synthetase. 3,5 The active metabolites, the 6-thioguanine nucleotides (6TGN), have a half-life of several days or more. 5 Steady state concentrations of 6TGN occur after 2 to 4 weeks of oral dosing with AZA 2.0 mg/kg/d. 6 The enzyme activities of TPMT and ITP are under genetic control and will be discussed in more detail below. 7,8 Thiopurine methyltransterase enzyme activity (measured by radioimmunoassay) is genetically determined. 7 There is a trimodal distribution of TPMT in the general population: homozygous low activity (< 5.0 U/mL RBC) occurs at a frequency of 0.3%; heterozygous or intermediate activity (5.0–13.7 U/mL RBC) occurs at a frequency of 11.1%; and homozygous high or normal activity (13.8–25.1 U/mL RBC) occurs at a frequency of 88.6%. 7 This trimodal distribution is typical of monogenic inheritance. The TPMT gene is located on the short arm of chromosome 6. 9 At least 10 variant alleles for TPMT have been associated with decreased enzyme activity (*2, *3A, *3B, *3C, *3D, *4, *5, *6, *7, *10). 10,–12 Patients with low or intermediate TPMT enzyme activity divert 6MP away from the 6MMP metabolite and toward 6TGN. Leukopenia has been associated with high concentrations of 6TGN. 13,14 A pretreatment measurement of TPMT activity (phenotype) or genotype could be used in clinical practice to select a customized drug dose with the intent of decreasing the frequency of leukopenia. Two studies have suggested that this strategy might be effective. In one study, TPMT genotypes were prospectively determined in 67 consecutive patients with rheumatologic diseases who were beginning AZA at a dose of 2–3 mg/kg/d. 15 The TPMT genotypes were determined for the *3A, *3C, and *2 variant allele. Six of 67 patients (9%) were heterozygous for TPMT activity, 6 of 6 had the *3A allele. Five out of 6 heterozygous patients with intermediate TPMT activity stopped treatment during the first month due to leukopenia, and 1 patient was not compliant with treatment. The median duration of treatment was 39 weeks (6 to 180 weeks) in the group with wild-type or normal TPMT activity and 2 weeks (range 2–4 weeks) in the group with intermediate TPMT activity. In the second study, the TPMT genotype was determined in 41 patients with Crohn's disease who had developed severe myelosuppression (white blood cell count < 3000 or platelet count < 100,000) during treatment with AZA or 6MP. 12 The TPMT genotypes determined were the *2, *3A, *3B, *3C, *3D, *4, *5, *6, *7, and *10 variant alleles. Seven of 41 patients (17%) had intermediate TPMT activity and 4/41 patients (10%) had low TPMT activity. Those patients with low or intermediate TPMT activity accounted for approximately half of all early leukopenia (within 6 months of starting AZA or 6MP), whereas patients with normal TPMT activity accounted for approximately 50% of the early leukopenia and nearly all of the late leukopenia (after 6 months of therapy). Recently, two additional studies of TPMT determination in patients have reached contrasting conclusions. Seddik and colleagues prospectively determined the TPMT variant alleles in 75 patients with Crohn's disease initiating AZA 100–200 mg/d or 6MP 50–75 mg/d and then followed the patients for the development of leucopenia. 16 Six patients (8%) were heterozygous for TPMT variant alleles, none were homozygous. Leukopenia occurred in 7/75 patients (9%) at time intervals of 1, 4 (n = 3), 5, 8, and 16 months after azathioprine was initiated. Among these 7 patients, only one was heterozygous for the TPMT mutation. In another study Curvers and colleagues determined the TPMT genotypes of 535 patients with IBD treated with AZA. 17 A case control study was performed in which the cases were patients with TPMT variant alleles and the controls were patients with the wild type alleles (2 controls for each case). Wild type alleles were present in 492 patients (90%), 53 patients were heterozygous for variant alleles (10%), and 2 patients were homozygous for variant alleles (0.4%). Azathioprine was discontinued for nonallergic side effects (see below) in 21 patients with the wild type genotype, (4%), 4 patients who were heterozygous for a variant allele (8%), and 1 patient who was homozygous for two variant alleles (50%), P value not significant. The relative risk for developing leukopenia was 1.5 (0.7–3.2) in patients who had one or more variant alleles. The concentrations of 6TGN were significantly greater in patients with one or more variant alleles compared with patients with wild type alleles. 18 In contrast to the previous studies, the findings from these two recent studies suggest that TPMT variants do not account for the majority of leukopenia episodes that occur during treatment with AZA or 6MP for IBD. However, these studies lack sufficient power to determine the predictive value of variant TPMT homozygotes for early severe leukopenia. Based on the earlier studies, some investigators recommend that pre-treatment determination of TPMT activity (phenotype) or genotype be performed in an attempt to predict the occurrence of early leukopenia in patients with IBD treated with AZA. Patients with normal TPMT activity are treated with standard doses of AZA (2.0–3.0 mg/kg/d) or 6MP (1.0-1.5mg/kg/d); that patients with intermediate TPMT enzyme activity have their dose of AZA or 6MP reduced by 50%19,20; patients with low TPMT activity are not treated due to a high mortality from leukopenia and sepsis. 21 The validity of these recommendations in patients with IBD remains a debated topic. Inosine triphosphate pyrophosphatase (ITP) enzyme activity is genetically determined. 22 The ITP gene is located on chromosome 20p. 23 At least 3 variant alleles for ITP have been associated with decreased enzyme activity (94C→A, IVS2+21A→C, 198C→A). 24,25 The variant allele frequencies in 100 healthy controls showed that 10% were heterozygous for 94C→A (24% ITP enzyme activity), 24% were heterozygous for IVS2+21A→C (60% ITP enzyme activity), 0% homozygous for 94C→A (0% ITP enzyme activity), 0% homozygous for IVS2+21A→C (60% ITP enzyme activity), and 2% were compound heterozygotes (10% ITP enzyme activity). 24 Ansari and colleagues determined the variants of ITP (94C→A) and TPMT in 64 patients with IBD with adverse drug reactions to AZA and 71 patients who did not have adverse reactions. 8 Overall, the ITP variant was significantly associated with adverse drug reactions (odds ratio 4.2, CI 1.6– 11.5, P = 0.0033). Variant TPMT genotypes were not significantly associated with adverse drug reactions overall, but did predict side effects in a subset of 14 patients with nausea and vomiting (odds ratio 5.1, CI 1.3–19.5, P = 0.0239). Conversely, ITP variant was significantly associated with flulike symptoms in 11 patients (odds ratio 6.2, CI 1.4–27.4, P = 0.0251) and rash in 6 patients (odds ratio 10.8, CI 1.8–65.9, P = 0.0190). Myelosuppression, pancreatitis and hepatitis were not predicted by ITP or TPMT genotype in this study. These findings suggest that ITP and TPMT variants each contribute to intolerance to azathioprine, and that not all leukopenia, pancreatitis, and hepatitis can be accounted for by variants of these two enzymes. In vitro studies have demonstrated that AZA and 6MP exert their effect by controlling T cell apoptosis through modulation of Rac1 activation upon CD28 costimulation. 26 Apoptosis induction required costimulation with CD28 and was mediated by specific blockade of Rac1 activation through binding of 6TGN to Rac1 instead of guanine triphosphate (GTP). Activation of the Rac1 gene in turn leads to activation of mitogen-activated protein kinase kinase (MAP kinase), NF-kappaB, and bcl-x(L), and finally to a mitochondrial pathway of apoptosis. Thus AZA and 6MP convert a costimulatory signal into an apoptotic signal by modulating Rac1 activity. In vivo, several studies have reported that patients with IBD treated with AZA or 6MP who respond to therapy have higher median concentrations of 6TGN than patients who fail to respond to therapy. 27,–29 One recent study in 93 patients with IBD reported that the median concentration of 6TGN in responding patients was 312 pmol/8 × 108 RBCs compared with a median concentration of 199 in patients who failed to respond. 28 There was no difference in the median concentrations of 6MMP between the 2 patient groups. The breakpoint between the lower two quartiles and the higher two quartiles of 6TGN concentrations was 235 pmol/8 × 108 RBCs. Sixty-five percent of responding patients had an erythrocyte 6TGN concentration > 235 as compared with only 27% of patients failing therapy. Thus, the authors suggested that clinicians should adjust AZA or 6MP doses to achieve 6TGN concentrations > 235 pmol/8 × 108 RBCs. The authors also reported that hepatotoxicity (defined as liver enzymes more than twice normal) occurred in 16 patients, and that the median 6MMP concentrations were 5463 pmol/8 × 108 RBC's in patients with hepatotoxicity compared with only 2213 pmol/8 × 108 RBCs in patients without hepatotoxicity. 28 These findings have not been universally confirmed. 6,30,31 The role of routine determination of 6TGN and 6MMP in patients with IBD treated with AZA or 6MP remains a debated topic. Patients with low or intermediate TPMT activity, patients requiring co-administration of allopurinol, situations in which the compliance of therapy must be determined, and patients who fail to respond to standard doses of AZA and 6MP are all potential candidates for therapeutic monitoring. Low thiopurine methyltransferase activity leads to greater concentrations of 6-thioguanine nucleotides, and thiopurine methyltransferase activity is genetically regulated. Patients with low thiopurine methyltransferase activity are at risk for dose-related toxicity associated with azathioprine or 6-mercaptopurine therapy. Some but not all studies have indicated that therapy drug monitoring with erythrocyte 6-thioguanine nucleotides may be clinically useful.