Novel pharmacogenetic markers for treatment outcome in azathioprine-treated inflammatory bowel disease

M.A. Smith; A.M. Marinaki; M. Arenas; M. Shobowale-Bakre; C. M. Lewis; A. Ansari; J. Duley; J.D.  Sanderson


Aliment Pharmacol Ther. 2009;30(4):375-384. 

In This Article

Abstract and Introduction


Background: Azathioprine (AZA) pharmacogenetics are complex and much studied. Genetic polymorphism in TPMT is known to influence treatment outcome. Xanthine oxidase/dehydrogenase (XDH) and aldehyde oxidase (AO) compete with TPMT to inactivate AZA.
Aim: To assess whether genetic polymorphism in AOX1, XDH and MOCOS (the product of which activates the essential cofactor for AO and XDH) is associated with AZA treatment outcome in IBD.
Methods: Real-time PCR was conducted for a panel of single nucleotide polymorphism (SNPs) in AOX1, XDH and MOCOS using TaqMan SNP genotyping assays in a prospective cohort of 192 patients receiving AZA for IBD.
Results: Single nucleotide polymorphism AOX1 c.3404A > G (Asn1135Ser, rs55754655) predicted lack of AZA response (P = 0.035, OR 2.54, 95%CI 1.06–6.13) and when combined with TPMT activity, this information allowed stratification of a patient's chance of AZA response, ranging from 86% in patients where both markers were favourable to 33% where they were unfavourable (P < 0.0001). We also demonstrated a weak protective effect against adverse drug reactions (ADRs) from SNPs XDH c.837C > T (P = 0.048, OR 0.23, 95% CI 0.05–1.05) and MOCOS c.2107A > C, (P = 0.058 in recessive model, OR 0.64, 95%CI 0.36–1.15), which was stronger where they coincided (P = 0.019).
Conclusion: These findings have important implications for clinical practice and our understanding of AZA metabolism.


Immunosuppressive drugs have become the mainstay of treatment for inflammatory bowel disease (IBD), with proven efficacy in reducing relapses, permitting steroid withdrawal and closing fistulas.[1,2] Indeed, as many as 60% of patients with Crohn's disease (CD) now receive azathioprine (AZA) or mercaptopurine (MP).[3] This reflects changing goals of treatment in IBD, away from symptom control alone, towards mucosal healing and altered natural history, particularly early in the disease course.[4,5]

As the importance of rapid and effective disease control, and therefore more aggressive use of immunomodulation, has been established, so it has become increasingly urgent to seek pharmacogenetic predictors of response to AZA and 6MP. Ideally, such predictors would prospectively identify patients who will suffer adverse drug reactions (ADRs) or derive no benefit from thiopurines and thus require alternative therapy. They would also assist physicians in dose-optimization and treatment monitoring or facilitate the decision to switch between immunosuppressants.

Azathioprine is a pro-drug with complex metabolism. Once ingested, it is broken down to release MP by both enzymatic[6–8] and non-enzymatic[9] conjugation with glutathione. According to the accepted model, there are then three different pathways competing to act on MP, (Figure 1). The first, hypoxanthine-guanine phosphoribosyltransferase, (HGPRT) is anabolic and constitutes the first step towards the production of the active end-product [thioguanine nucleotides, TGNs]. The other two pathways, thiopurine methyltransferase (TPMT) and xanthine oxidase/dehydrogenase (XDH) both produce metabolites which are thought to be inactive and are the first steps in eliminating thiopurines from the body. An additional pathway, diversion of the TGN path into thio-ITP, is also relevant because it is regulated by a polymorphic enzyme ITP phosphohydrolase, (ITPase, encoded by the gene ITPA).[10] The activity of these various pathways is thought to determine the amount of ingested drug that becomes activated[11,12] and contribute to the inter-patient variability in response to thiopurines.

Figure 1.

The metabolism of the thiopurines drugs. Enzymes targeted in this study are indicated by a yellow star. Thiopurine drugs (shown in circles): AZA: azathioprine, MP: mercaptopurine, 6-TG: 6-thioguanine. Metabolites (shown in bold, active metabolites in boxes): tIMP: thioinosine monophosphate, tXMP: thioxanthine monophosphate, tGMP: thioguanine monosphosphate, tGDP: thioguanine diphosphate, tGTP: thioguanine triphosphate, ITP: inosine triphosphate, tITP: thioinosine triphosphate, 6-MeMP: 6-methylmercaptopurine, 6-Me-tIMP: 6-methyl thioinosine monophosphate, Me-tGMP: methylthioguanine monophosphate, deoxy-tGTP: deoxythioguanine triphosphate, 8-OH AZA: 8-hydroxy azathioprine, 8-OH 6MP: 8-hydroxy mercaptopurine, 8-OH TG: 8-hydroxy thioguanine, 6-MeM-8-OHP: 6 methylmercapto-8-hydroxypurine. Enzymes: TPMT: thiopurine methyltransferase, XDH: xanthine dehydrogenase, PK: phosphokinase, IMPDH: inosine monophosphate dehydrogenase, HGPRT: hypoxanthine guanine phosphoribosyltransferase, GMPS: guanosine monophosphate synthetase.

Much is now known about the role of genetically determined variation in TPMT activity in toxicity and treatment success with AZA. It is estimated that TPMT deficiency is responsible for up to 30% of all ADRs experienced on AZA, but whilst TPMT deficiency strongly predicts the development of myelotoxicity, the most serious ADR of AZA therapy, it fails to account for over 70% of cases of myelotoxicity.[13,14] Furthermore, those with higher TPMT activity are at increased risk of nonresponse to AZA.[15–17] However, it is highly likely that genetic polymorphism in other enzymes involved in thiopurine metabolism also has an impact on each individual's response to AZA therapy (Figure 1).

As the second major contributor to azathioprine breakdown besides TPMT, a logical candidate enzyme for further study is xanthine oxidase/dehydrogenase (XDH).[18,19]XDH is known to be subject to common genetic polymorphism and, although true deficiency (Type 1 Xanthinuria) is rare, there is considerable inter-individual variation in enzyme activity,[20,21] some of which may be attributable to genetic differences.[22] Both increased and decreased XDH activities have been associated with a variety of SNPs in the Japanese population[23] and there is preliminary evidence to suggest that XDH SNPs can affect azathioprine metabolite levels[24] in a pattern attributable to decreased XDH activity. Blocking XDH activity using allopurinol is known to cause severe toxicity with conventional doses of AZA and safe co-prescription of allopurinol requires an AZA dose-reduction of approximately 80%.[25] Indeed, the ability of allopurinol to increase the bioavailability and improve the efficacy of AZA has been well demonstrated in studies of patients failing to respond to full-dose treatment.[26–28] Taken together, these observations suggest that the known inter-individual variability in XDH activity (whether attributable to genetic or other factors),[22,29] could have an impact on an individual's response to AZA.

A molybdenum cofactor[30] is essential for the action of three oxidases, XDH, aldehyde oxidase (AO) and sulphite oxidase. Deficiency of the cofactor is associated with severe neurodegeneration resulting in early infant death caused by the loss of sulphite oxidase activity,[31] coincident with AO and XDH deficiencies. The final step in the specific adaptation of molybdenum cofactor for XDH and AO requires the action of molybdenum cofactor sulfurase (MOCOS). MOCOS deficiency (which results in the deficiencies of both XDH and AO, but not sulphite oxidase) is, in contrast, relatively benign, causing only a predisposition to renal stones (Type II Xanthinuria).[32] The MOCOS gene is also subject to genetic polymorphism, which might also affect AZA metabolism.

The role of AO in human physiology remains unclear. It occurs as a single isoform in humans, is much more widely distributed than XDH and has a broad range of substrates.[33–35] It is therefore thought to have additional functions over and above its contribution to purine catabolism.[36] AO acts on azathioprine, MP and other thiopurine metabolites, contributing to the catabolism of thiopurines[18,34,37] (Figure 1). However, the functional significance of the thiopurine metabolites generated by AO is poorly understood. Despite AO products being found in significant quantities, the role of AO in thiopurine metabolism has been thought to be of minimal clinical significance and has not been examined. There is evidence of varying AO activity between individuals,[38,39] but whether this variability relates to the presence of coding SNPs remains unknown. It is possible that this is at least in part because of the complex regulation of gene expression,[40] but it could equally be attributable to genetic polymorphism.

We hypothesized that genetic polymorphism in XDH, MOCOS and AOX1 contributes to variation in clinical outcome on AZA therapy and set out to examine a well-defined prospective cohort of patients receiving AZA for IBD for the presence of coding SNPs in XDH, MOCOS and AOX1, seeking associations between the presence of such polymorphism and clinical outcome (response, nonresponse or toxicity) on AZA therapy.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.