Background

Thiopurines make up a class of immunosuppressive and chemotherapeutic drugs that is used effectively in the treatment of chronic autoimmune inflammatory conditions, hematological malignancies, and prevention of organ transplant rejection. Azathioprine (AZA), 6-mercaptopurine (6-MP) and 6-thioguanine (6-TG) are the thiopurine drugs currently used in clinical practice. The clinical response to thiopurines varies according to the nature of disease, dose and patient metabolism of the drugs.

AZA and 6-MP are currently widely used as steroid sparing agents in chronic autoimmune inflammatory conditions, including pemphigoid, inflammatory bowel disease, and rheumatoid arthritis, among others. AZA and 6-MP are effective in inducing remission in 50 percent to 60 percent of inflammatory bowel disease patients, and permit steroid reduction or withdrawal in up to 65 percent of patients.² Clinical response rates using AZA to treat nonbullous inflammatory dermatoses can be as high as 75 percent.³ However, use of AZA or 6-MP in other chronic inflammatory disorders including lupus and rheumatoid arthritis has been variable, and they are often not the primary drugs of choice.

When used among organ transplant patients, although AZA has been associated with a 5-year renal graft survival ranging from 70 percent to 92 percent,^4,5 use of AZA and 6-MP in transplantation has declined somewhat in favor of other immunosuppressive drugs.

Both 6-MP and 6-TG have been used effectively in treatment of childhood acute lymphoblastic leukemia, with remission rates (5-year relapse free survival) of approximately 80 percent using 6-MP.⁶

Patients with cancer or transplanted organs are clinically more complex, so this review focuses on thiopurine use in autoimmune conditions.

Biochemistry of Thiopurines

AZA and 6-MP are pro-drugs that have no intrinsic biological activity, and require extensive metabolism for activity (Figure 1). After oral administration of AZA or 6-MP, between 27 percent and 83 percent is biologically available. AZA is often used clinically, as it is more stable and soluble than 6-MP. AZA doses are higher because the molecular weight of 6-MP is 55 percent of that of AZA.

Figure 1

Metabolic pathways of thiopurine drugs. Abbreviations: 6-MP = 6-mercaptopurine; 6-tGN = 6-thioguanine nucleotides; 6-tIMP = 6-thiomercaptopurine; 6-TG = 6-thioguanine monophosphate; 6-tGN = deoxy-6-thioguanosine 5′ triphosphate; 6-tXMP = 6-thiooxanthosine; (more...)

In the gut, approximately 90 percent of AZA is converted to 6-MP, a thiopurine analogue of the purine base hypoxanthine, by cleavage of the imidazolyl moiety which is thought to be catalyzed through the action of glutathione transferase.⁷ 6-MP is then enzymatically converted to its active metabolite, deoxy-6-thioguanosine 5′ triphosphate (6-tGN), through successive enzymatic conversions by hypoxanthine-guanine phosphoribosyl transferase (HGPRT) and inosine monophosphate dehydrogenase (IMPDH). Inactivation of 6-MP (and hence AZA) occurs primarily through S-methylation by thiopurine S-methyltransferase (TPMT), and to a minor degree by catabolism, to thiouric acid by xanthine oxidase (XO).

6-TG is converted to its active metabolite (6-tGN) in a single step involving HGPRT, while inactivation occurs through two pathways. The major metabolic pathway involves guanine deaminase (GD) and aldehyde oxidase (AO) to form inactive 6-thiouric acid. Metabolism by TPMT, to form inactive 6-methyl-TG, is a minor contributor to drug inactivation. TPMT also plays a minor role in directly methylating and inactivating 6-tGN.

Incorporation of 6-tGN into DNA triggers cell-cycle arrest and apoptosis through the mismatch repair pathway. Until recently, this was considered the primary mechanism of action.⁸ However, recent evidence has suggested other mechanisms of immunosuppression not directly related to 6-tGN incorporation into DNA. Metabolism by TPMT of 6-thiomercaptopurine (6-tIMP), an intermediate metabolite, to produce 6-methyl –tIMP has been shown to inhibit de novo purine synthesis in lymphocytes, which likely contributes to the immunosuppressive effects of thiopurines.⁹ Furthermore, accumulation of 6-tGN in lymphocytes has been demonstrated to decrease the expression of TRAIL, TNFRS7, and α-4 integrin, effectively decreasing inflammation. Thiopurine drugs have also been shown to induce apoptosis in T-cells through modulation of Rac1 activation upon CD28 costimulation. Rac1 is a GTPase upstream of MEK, NF-κB, and bcl-xL. Upon binding of 6-thio-GTP with Rac1, activation of its downstream mediators is blocked, inducing apoptosis.¹⁰

Thiopurine Toxicity

Thiopurine-based drugs have been associated with various toxic adverse events, including myelosuppression, hepatotoxicity, pancreatitis, and flu-like symptoms, among others. One of the most serious dose-dependent reactions is myelosuppression, which is believed to be caused by increased 6-tGN levels (the active metabolite), either due to overdosing or a low rate of thiopurine metabolism. The most extensively characterized enzyme in the metabolism of thiopurines is TPMT.

TPMT polymorphisms. The TPMT gene is located on chromosome 6 at 6p22.3. It is approximately 27 kb in size and contains 9 exons.^11,12 A nonfunctional TPMT pseudogene has also been identified on chromosome 18 at 18q21.1.¹³ TPMT is widely expressed in many tissues, but TPMT expression in lymphocytes, red blood cells and bone marrow is most relevant clinically for immunosuppression by thiopurine drugs. To date, at least 30 variant (or mutant) alleles of TPMT have been identified, the majority of which have been associated with lower TPMT enzymatic activity or protein expression (Table 2).¹⁴ Several studies have highlighted the importance of thiopurine drug metabolism by TPMT, as lower TPMT may place patients at higher risk of developing drug-related toxicity.^15,16 The four most common alleles (TPMT*2, TPMT*3A, TPMT*3B, and TPMT*3C) seen in Caucasians, Asians, and Africans account for approximately 80 percent to 95 percent of individuals with lower TPMT activity.^17–22 When comparing genotype to phenotype (enzymatic activity), homozygous mutant individuals have very low or absent enzymatic activity while those heterozygous for a mutant allele demonstrate intermediate enzymatic activity, between those of noncarrier and homozygous individuals. The frequency of the common alleles within each ethnic group varies, as does the overall number of individuals with lower TPMT activity. Heterozygous individuals with intermediate enzymatic activity comprise five percent to 15 percent of patients, while approximately 0.3 percent are homozygous, with very low or absent enzymatic activity.^17,18,23

Table 2

TPMT polymorphisms.

TPMT analysis. Analysis of TPMT status can be accomplished through either analysis of the red blood cell TPMT enzymatic activity, or genotyping. Genetic analysis in routine clinical laboratories involves targeting specific TPMT mutations, usually at least three of the four common alleles. Depending on the mutant alleles targeted and the ethnic background of the patient, genotyping can identify up to 95 percent of affected individuals, but it will not identify those patients with rare mutations. Since the frequency of the rare mutations is exceedingly low, the probability of missing patients with rare mutations is also low. The enzymatic assay is currently considered to be the gold standard measurement, since it should identify all patients with reduced enzymatic activity, regardless of mutations. However, the enzymatic assay is technically more challenging to perform. In current clinical practice, both enzymatic testing and genetic analysis are being performed, depending on the laboratory.

Clinical utility and validity of TPMT analyses. Currently, there is no evidence that the presence of one or more mutant TPMT alleles causes disease or places one at increased risk for disease. However, the presence of a mutant allele has been suggested to increase the risk of thiopurine-related drug toxicity, particularly when using AZA or 6-MP (6-TG is not metabolized to as great an extent by TPMT). Therefore, a fraction of patients prescribed thiopurines are at greater risk of developing drug-related toxicity. Until recently, all patients were prescribed a standard starting dose of either AZA or 6-MP. The current starting dose for AZA ranges from 1.0 to 2.5 mg/kg/day and 0.75 to 1.25 mg/kg/day for 6-MP. Patients with either intermediate or low to absent TPMT activity may benefit from a decreased starting dose.

Various clinical guidelines suggest measuring TPMT enzymatic activity or screening for TPMT alleles associated with reduced enzymatic activity before starting patients on thiopurine drugs.^30,31 However, measuring TPMT activity may not lead to reduced drug-related toxicity since regular monitoring is recommended. Complete blood counts, including platelet counts are recommended to be done weekly during the first month, twice monthly for the second and third months of treatment, then monthly or more frequently if dosage alterations or other therapy changes are necessary.³² As such, there is a need to review the current literature regarding the assessment of TPMT status prior to administration of thiopurine drugs, to determine if TPMT testing will reduce drug-related toxicity.

Scope, Topic Development, and the Key Questions

This review of the effectiveness of determining thiopurine methyl transferase (TPMT) enzymatic activity prior to initiation of thiopurine therapy in patients with chronic autoimmune diseases was nominated by the American Association for Clinical Chemistry (AACC), and commissioned by the Agency for Healthcare Research and Quality (AHRQ).

TPMT status can be assessed by direct determination of the TPMT enzymatic activity (phenotyping), or by genotyping TPMT gene coding for the enzyme for common single nucleotide polymorphisms (SNPs), also referred to as variant alleles, coding for the enzyme.

In theory, TPMT status determination before initiating thiopurine therapy may be undertaken in order to address two potential clinical scenarios:

• minimize thiopurine toxicity in up to 15 percent of patients with lower TPMT enzymatic activity, by thiopurine dose reduction or switching to alternative treatment
• optimize clinical responsiveness in patients with abnormally elevated TPMT enzymatic activity, by dose escalation

In the first scenario, thiopurine dose reduction, in order to minimize drug toxicity associated with excessively elevated levels of thioguanine nucleotides, may negatively affect treatment efficacy (i.e. overzealous dose reduction to subtherapeutic thiopurine levels). Hence, an additional concern in tandem with the first scenario is what an optimally effective dose reduction should be, in light of pretreatment knowledge of TPMT status, in patients likely to experience increased thiopurine toxicity. Based on a scoping review of the literature, it was anticipated that this additional concern had not been adequately investigated in primary research. Therefore, the current systematic review of literature does not investigate the relative efficacy or the effectiveness of thiopurine dose adjusted treatment of chronic autoimmune diseases with pretreatment TPMT testing. As such, this review focuses on the equipoise of whether pretreatment determination of the TPMT status (using genotyping and/or phenotyping) mitigates harms associated with thiopurine therapy. In addition, the accuracy of TPMT status determination by genotyping is also investigated, in reference to the enzymatic activity assay, as well as the costs and potential savings associated with TPMT testing.

The second clinical scenario could not be investigated because abnormally high TPMT enzymatic activity has not been investigated in any detail in primary research, so its significance is not yet appreciated.

The analytic framework (Figure 2) depicts the causal pathways forming the basis of the key questions. Since study of myelosuppression in organ transplant and cancer patients poses several potential confounders (namely, concomitant myelosuppressive treatment and the short-term complications induced by the procedure or disease), the eligibility criteria were restricted to populations with chronic autoimmune diseases. As well, since the metabolism of the drug 6-thioguanine does not involve the TPMT enzyme, we focused on the two thiopurine drugs, AZA and 6-MP.

Figure 2

Analytic framework. Abbreviations: KQ = key question; TPMT = thiopurine methyltransferase

Key Questions Addressed in This Report

The following key questions were generated by the University of Ottawa Evidence Based Practice Centre in consultation with American Association for Clinical Chemistry and the Agency for Health Research Quality. Outcomes are considered only in the context of drug toxicity/adverse events, and not of efficacy.

KQ1. In terms of the analytical performance characteristics of enzymatic measurement of TPMT activity and determination of TPMT allelic polymorphisms:

1. What are the preanalytical requirements for enzymatic measurement of TPMT and determination of TPMT allelic polymorphisms? (e.g. specimen types and collection procedures, lab transportation, interference of coadministered drugs, patient preparation and identification etc.)
2. What are the within and between laboratory precision and reproducibility of the available methods of enzymatic measurement of TPMT and determination of TPMT allelic polymorphisms (proficiency testing)?
3. What is the diagnostic sensitivity and specificity of TPMT allelic polymorphism measurement compared to the measurement of TPMT enzymatic activity in correctly identifying chronic autoimmune disease patients eligible for thiopurine therapy with low or absent TPMT enzymatic activity? How do effect modifiers (e.g. underlying disease prevalence and severity, different activity thresholds, Hardy-Weinberg equilibrium, number and types of alleles tested) explain any observed heterogeneity in sensitivity and specificity?
4. Are there any postanalytical requirements specific to measurement of TPMT enzymatic activity or TPMT allelic polymorphism measurement? (E.g. timely reporting of data, reference intervals, immediate or reporting within a time-frame, highlighting of extreme results)

KQ2. Does the measurement of TPMT enzymatic activity or determination of TPMT allelic polymorphisms change the management of patients with chronic autoimmune disease when compared with no determination of TPMT status?

KQ3. In chronic autoimmune disease patients prescribed thiopurine-based drugs (AZA or 6-MP), does the assessment of TPMT status to guide therapy, when compared with no pretreatment assessment, lead to:

1. reduction in rates of mortality, infection, hospitalization, withdrawal due to adverse events (WDAE), serious adverse events (SAE) and improvement in health-related quality of life?
2. reduction in rates of myelotoxicity, liver toxicity, and pancreatitis?
3. In the absence or inconclusiveness of evidence answering key question 3a and/or 3b above, is there an association between TPMT status (as determined by TPMT enzymatic activity and/or TPMT allelic determination) and/or the following amongst chronic autoimmune disease patients treated with thiopurines?
   1. the clinical outcomes of mortality, infections, hospitalization, WDAE, SAE and health-related quality of life?
   2. surrogate outcomes of myelotoxicity, liver toxicity, and pancreatitis?

KQ4. What are the costs of determining TPMT enzyme activity and/or genotyping for patients with chronic autoimmune disease being considered for thiopurine-based therapy (e.g., costs of testing, costs of care, and costs of treating drug-associated complications)?