Distribution of TPMT risk alleles for thioupurine toxicity in the Israeli population
- First Online:
- Cite this article as:
- Efrati, E., Adler, L., Krivoy, N. et al. Eur J Clin Pharmacol (2009) 65: 257. doi:10.1007/s00228-008-0590-7
- 87 Views
Background and objective
Individuals with intermediate or no thiopurine S-methyltransferase (TPMT) activity are at risk of hematotoxicity when treated with standard doses of thiopurines, thus, pretreatment identification of these individuals is of major importance. The purpose of this study was to determine the frequency and distribution of TPMT polymorphic variants, known to functionally impair TPMT activity, in the highly heterogeneous Israeli population.
TPMT genotyping of individuals representing three major demographic groups in Israel was carried out by PCR restriction fragment length polymorphism and high-resolution melting.
Frequencies of TPMT risk alleles differed significantly among the screened Israeli subpopulations: Druze showed fivefold and twofold higher frequencies than Jews and Moslems, respectively. Specifically, allelic frequencies of TPMT*3A were 0.73% (95% CI 0.34-1.45%), 0.79% (95% CI 0.16-2.39%), and 3.19% (95% CI 1.78-5.58%) in Jews, Moslems, and Druze, respectively. Although not found in Jews, TPMT*3C was found at an allelic frequency of 1.05% (95% CI 0.31-2.76%) and 0.75% (95% CI 0.02-2.84%) in Moslems and Druze. TPMT*2 and TPMT*3B were not detected in any of the Israeli subpopulations studied.
These data indicate that the Israeli population displays a distinct TPMT genetic variability that is comprised of a mix of three major genetically diverse subpopulations, each with its unique TPMT allelic frequency distribution pattern and likelihood of developing an adverse reaction to thiopurine drugs.
Thiopurines (mercaptopurine, thioguanine, and azathioprine) are widely used as immunosuppressants for the treatment of inflammatory and neoplastic diseases . In hematopoietic tissues, thiopurine S-methyltransferase (TPMT) catalyzes the major inactivation route of thiopurines . Individuals who are TPMT-deficient or have intermediate TPMT activity accumulate excessive levels of active thioguanine nucleotides and are at high risk of developing myelosuppression and life-threatening leukopenia when treated with standard doses of thiopurines [3–7]. TPMT enzyme activity is determined by co-dominant genetic polymorphic sequence variants at the TPMT locus on 6p22.3 . Since these variations have a profound effect on thiopurine bioavailability and toxicity, identifying TPMT-deficient individuals has been shown to be helpful in individualizing drug dosage and/or directing drug regimen choices [7, 8].
Distribution of TPMT allele frequencies (%) in various populations
No. of subjects
Despite the wealth of data related to the distribution of TPMT polymorphisms across various populations worldwide, little is currently known about the distribution of these variants in the Israeli population. Israel, with its religious and ethnic diversity, offers a unique opportunity to examine genetic variability in a heterogeneous society that maintains distinct subpopulations due to negligible rates of interethnic marriages. The present study was designed to delineate the relative frequency of the common TPMT polymorphisms in three Israeli subpopulations, Jews, Arab Moslems, and Druze.
Materials and methods
Blood samples were collected from 881 unrelated healthy Israeli volunteers of Jewish, Druze, or Moslem descent, as defined by the donors themselves. Genomic DNA was extracted using ArchivePure DNA Purification kit (5-Prime, USA) according to the manufacturer′s protocol.
DNA samples were assessed for the presence of the following polymorphic changes: c.G238C, c.G460A, and c.A719G. Variants TPMT*2 (c.G238C), TPMT*3A (c.G460A and c.A719G), TPMT*3B (c.G460A), and TPMT*3C (c.A719G) were ascertained by allele-specific PCR followed by restriction fragment length polymorphism analysis (PCR-RFLP). Results were confirmed using probe-free high-resolution melting technology (HRM), whereby genotyping is based on the effect of mutations on amplicon melting.
For PCR-RFLP, three DNA samples were PCR-amplified with Taq DNA Polymerase (Qiagen, Germany) using the following primers: for c.G238C (variant TPMT*2): 5′-TAAGTGTAAATGTATGATTTTATGC-3′ and 5′-CAATTATTTACCCAAATCAAAACAA-3′; for c.G460A (variants TPMT*3A and TPMT*3B): 5′-GAAACGCAGACGTGAGATCC-3′ and 5′-GCCTTACACCCAGGTCTCTG-3′; and for c.A719G (variants TPMT*3A and TPMT*3C): 5′-CCACCATACCCAGCTCATTT-3′ and 5′-CCTCAAAAACATGTCAGTGTGA-3′.
PCR products were subsequently digested with the following restriction enzymes: Hpy188III (c.G238C), MwoI (c.G460A), or AccI (c.A719G) (New England Biolabs, USA) and separated on 3% agarose gels.
For HRM, the primers used were as follows: for c.G238C (variant TPMT*2): 5′-TGTAAATGTATGATTTTATGCAGGTTT-3′ and 5′-TCTGAGTAAGAAAGATTCTGCTCTGT-3′; for c.G460A (variants TPMT*3A and TPMT*3B): 5′-TAGGACAAATATTGGCAAATTTGA-3′ and 5′-TTACCATTTGCGATCACCTGGATTCATGGCAAC-3′; and for c.A719G (variants TPMT*3A and TPMT*3C): 5′-GGTTGATGCTTTTGAAGAACG-3′ and 5′-CATCCATTACATTTTCAGGCTTT-3′. Reactions were carried out using Thermo-Start PCR master mix (ABgene, UK) with the intercalating dye LC-Green (Idaho Technology, USA) in an HRM-enabled real-time PCR, Rotor-Gene 6000 (Corbett Research, UK).
Confidence intervals, χ2 test, and Fisher′s exact test were calculated by standard methods.
The 881 Israelis included in this project comprised 531 Jews, 194 Moslems, and 156 Druze, thus representing the three major demographic groups in Israel. Individuals were screened for the most common TPMT low-activity variants: TPMT*2 (c.G238C), TPMT*3A (c.G460A, c.A719G), TPMT*3B (c.G460A), and TPMT*3C (c.A719G). No discrepancies between RFLP and HRM results were found.
TPMT polymorphism frequencies differed significantly between the subpopulations screened. The highest level of polymorphisms was found in Druze, where, of the 156 individuals examined, 11 (7.1%, 95% CI 3.86-12.31%) were either homo- or heterozygous for the screened TPMT mutations. In Moslems, 7 (3.6%, 95% CI 1.62-7.4%) of 194 individuals examined displayed polymorphisms in TPMT. The lowest levels of polymorphism were found in Jews, where only 8 (1.5%, 95% CI 0.71-3.0%) of the 531 individuals examined were heterozygous for a sequence variant associated with decreased TPMT activity. These differences between Jews and Druze were more significant (P = 0.000205) than between Jews and Moslems (P = 0.078) or between Druze and Moslems (P = 0.147). The overall frequencies of the screened TPMT alleles associated with low enzyme activity were 0.73% in Jews, 1.83% in Moslems, and 3.94% in Druze. TPMT genotypes in Jews and Moslems were in Hardy-Weinberg equilibrium, but in Druze they were not.
Although the clinical relevance of TPMT genetic variants is well established and has been studied in a variety of populations (Table 1 and references therein), limited information is available to date on the distribution of these polymorphisms in the highly heterogeneous Israeli population. This heterogeneity, maintained by extremely low rates of interethnic marriages, is readily apparent in the results presented here, and indicates that, in terms of TPMT genetic variability, the Israeli population cannot be compared to other studied populations and, furthermore, that they should not be considered as a single entity but rather a population comprised of three major genetically diverse subpopulations, each with its own unique TPMT allelic frequency distribution pattern. The present data suggest the need to assess the possibility that these genetic variations may influence the relative risk of the various Israeli populations of developing thiopurine-related, often life-threatening, side effects.
Despite dissimilarities in TPMT allele distribution, a complete lack of TPMT*2 and TPMT*3B (Fig. 1) was evident in all the Israeli subpopulations studied. This is hardly surprising in light of the extremely close affinity observed between Jewish and non-Jewish Middle Eastern populations, supporting the hypothesis of a common Middle Eastern origin . As seen in Table 1, allelic variant TPMT*2 is one of the more common variants associated with decreased TPMT activity in Caucasians, but is virtually nonexistent in populations of African and East/West Asian descent, placing Jews, Moslems, and Druze with the latter, earlier-to-emerge populations. Allelic variant TPMT*3B is much rarer and was previously reported only in Sardinian, Swedish, Slovenian, and Brazilian populations.
Of the subpopulations assessed in this study, Israeli Druze displayed the highest frequency of TPMT risk alleles (3.9%), Moslems had intermediate rates (1.8%), and Jews showed significantly lower frequencies (0.7%). These low rates in Jews were comparable to those of West Asians, Egyptians, Taiwanese, Japanese, and Uygur Chinese, but were 5-10 times lower than Caucasians, Africans, or African Americans (Table 1).
In addition to differences in overall variability between the three major subpopulations examined and despite the fact that only two of the four screened alleles were detected, we identified subpopulation-specific distribution patterns for the various polymorphisms tested. Thus, as seen in Fig. 1, in Jews, TPMT*3A was the only low-activity variant found and all other variants were altogether absent. In Druze, TPMT*3A accounted for the majority of TPMT low-activity alleles (81%) and TPMT*3C for the remaining 19%, whereas in Moslems, TPMT*3A and TPMT*3C were nearly equal (43% and 57%, respectively). It is noteworthy that homozygosity in TPMT allelic variants, known to be associated with markedly reduced enzyme activity [4, 7, 30] was found only in individuals of Druze descent. In fact, the frequency of homozygotes found in this study in the Druze population was nearly fivefold higher than that previously reported in Caucasians (1.6% versus 0.33%) , causing this population to deviate from Hardy-Weinberg equilibrium.
Of the mutant alleles associated with reduced TPMT activity identified to date, variants TPMT*3A and TPMT*3C are by far the most widespread with a distinct worldwide distribution (Table 1). Variant TPMT*3A is a composite of variants TPMT*3B (c.G460A in exon 7) and TPMT*3C (c.A719G in exon 10). The fact that TPMT*3C is present worldwide and TPMT*3A is more restricted to individuals of European descent suggests that TPMT*3C is the most ancestral TPMT variant allele, as confirmed by variation analysis at fast mutating microsatellite markers . Variant TPMT*2 has only been detected in Caucasians suggesting that among the alleles screened for in this study, it is the latest to have emerged. Interestingly, this late variant was nonexistent in all three subpopulations examined herein.
A high correlation between TPMT genotype and phenotype is known to exist [7, 8]. Thus, in Caucasians a trimodal distribution was found, consistent with autosomal co-dominant inheritance, since 89% are wild-type with high TPMT activity, 11% are heterozygous with intermediate activity, and 0.3% are homozygous for TPMT risk variants and have negligible enzyme activity . In Jewish and Korean populations, on the other hand, a nearly unimodal pattern of high enzyme activity was demonstrated [12, 19, 28]. This distribution of activity in the Jewish population is consistent with the genetic findings of this study, which showed that only 1.5% of the individuals examined carried a risk allele. Of the 1.5% affected individuals, all were heterozygotes, presumably with intermediate TPMT activity.
Although TPMT activity may be critical for individuals treated with thiopurines, the physiological substrate of this enzyme remains unknown. Based on the significant homology with TPMT-like genes from bacteria, Krynetski et al.  speculated that TPMT may be involved in methylation of sulfur-, selenium-, or tellurium-containing compounds as a means of detoxification via conversion to relatively nontoxic methylation products. Phenotypically, however, TPMT-deficient individuals are indistinguishable from those who are TPMT-proficient, until challenged with thiopurines . It is therefore unclear whether selection played a role in creating the unique patterns of haplotypic diversity for TPMT deficiency alleles reported in this and previous studies (Table 1). Inadvertent selection for TPMT caused by selective pressure on neighboring genes is also unlikely, since the nearest protein-coding genes are AOF1 and DEK, and in both cases no indications as to their being targets for selection have been found .
In conclusion, each of the non-mixing subpopulations living in Israel presents a unique pattern of TPMT risk alleles. Given that TPMT variants largely determine the safety of treatment with thiopurines [3–7], the present data suggest the need to ascertain population-specific risks associated with the use of these drugs, and consequently to design therapeutic strategies tailored to the genetic features of local Israeli subpopulations.
We wish to thank Prof. Karl Skorecki for helpful comments.