Thiopurine methyltransferase metabolizes 6-mercaptopurine, thioguanine and azathioprine, thereby regulating cytotoxicity and clinical response to these thiopurine drugs. In healthy Caucasian populations, 89–94% of individuals have high thiopurine methyltransferase activity, 6–11% intermediate and 0.3% low, resulting from genetic polymorphism. Four variant thiopurine methyltransferase alleles were detected in over 80% of individuals with low or intermediate thiopurine methyltransferase activity. The wild-type allele is defined as TPMT*1 and the mutant alleles are TPMT*2 (G238C), TPMT*3A (G460A and A719G), TPMT*3B (G460A) and TPMT*3C (A719G). The frequency of these alleles in different ethnic groups is not well defined. In this study, DNA from 199 British Caucasian, 99 British South West Asian and 192 Chinese individuals was analysed for the presence of these variant alleles using polymerase chain reaction-restriction fragment length polymorphism and allele-specific polymerase chain reaction based assays. The frequency of individuals with a variant thiopurine methyltransferase genotype was: Caucasians 10.1% (20/199), South West Asians 2.0% (2/99) and Chinese 4.7% (9/192). Two TPMT*2 heterozygotes were identified in the Caucasian population, but this allele was not found in the two Asian populations. TPMT*3A was the only mutant allele found in the South West Asians (two heterozygotes). This was also the most common mutant allele in the Caucasians (16 heterozygotes and one homozygote) but was not found in the Chinese. All mutant alleles identified in the Chinese population were TPMT*3C (nine heterozygotes). This allele was found at a low frequency in the Caucasians (one heterozygote). This suggests that A719G is the oldest mutation, with G460A being acquired later to form the TPMT*3A allele in the Caucasian and South West Asian populations. TPMT*2 appears to be a more recent allele, which has only been detected in Caucasians to date. These ethnic differences may be important in the clinical use of thiopurine drugs. Pharmacogenetics 9:37–42 © 1999 Lippincott Williams & Wilkins

Substantial hematologic toxicity limits the use of azathioprine. To evaluate 1) polymorphic inactivation of azathioprine by thiopurine methyltransferase and 2) clinical toxicity. Prospective cohort study. Two rheumatology units. 67 patients for whom azathioprine was prescribed as second-line therapy for rheumatic disease. Polymerase chain reaction-based assays were used to detect mutations in thiopurine methyltransferase. The primary end point was discontinuation of azathioprine therapy because of toxicity. Six of 67 patients (9%) were heterozygous for mutant thiopurine methyltransferase alleles. Five of the 6 patients discontinued therapy within 1 month of starting treatment because of low leukocyte counts. The sixth patient did not adhere to treatment. Patients with wild-type thiopurine methyltransferase alleles received therapy longer than did patients with mutant alleles (median duration of therapy, 39 weeks [range, 6 to 180 weeks] and 2 weeks [range, 2 to 4 weeks], respectively; P = 0.018). Analysis of thiopurine methyltransferase genotype is a quick way to identify patients at risk for acute toxicity from azathioprine.

Thiopurine methyltransferase (TPMT) catalyses the S-methylation of thiopurine drugs such as 6-mercapto-purine, 6-thioguanine and azathioprine. TPMT activity is inherited as an autosomal co-dominant trait, and several mutations in the TPMT gene have been identified which correlate with a low activity phenotype. Although ethnic differences in TPMT activity have been described, population frequency analysis of TPMT alleles has not been well defined in different ethnic groups. The frequency of four allelic variants of the TPMT gene, TPMT*2, TPMT*3A, TPMT*3B and TPMT*3C were compared in British Caucasian (n = 199) and Ghanaian (n = 217) populations using PCR-RFLP and allele-specific PCR-based assays. TPMT*3C was found in 14.8% of Ghanaians (31 heterozygotes, one homozygote). The TPMT*2, TPMT*3A and TPMT*3B alleles were not detected in any of the Ghanaian samples analysed. In contrast, 10.1% of British subjects had variant alleles, consisting of TPMT*2 (n = 2), TPMT*3A (n = 17) and TPMT*3C (n = 1) alleles. The frequencies of mutant alleles in this study were 5.3 and 7.6% in British Caucasians and Ghanaians, respectively. Among Ghanaian tribes, Ewe subjects had a lower frequency of mutant alleles (5.9%) than Ga (13.2%) or Fanti (11.6%), although this did not reach statistical significance. This study provides the first analysis of TPMT mutant allele frequency in an African population and indicates that, unlike Caucasians, TPMT*3C is the most common allele in African subjects.

Dihydropyrimidine dehydrogenase (DPD) degrades over 80% of administered 5-fluorouracil (5FU), thereby regulating the efficacy of this commonly used anticancer agent. DPD activity is highly variable (8-21-fold) and individuals with reduced activity have a high risk of 5FU toxicity. DPYD encodes DPD protein and 13 different mutations have been reported in DPD-deficient subjects. However, the contribution of these variant genotypes to polymorphic DPD activity in vivo is not clear. The previously described DPYD mutations are contained in 10 exons. These 10 exons were sequenced in a cohort of cancer patients with reduced (n = 23) or normal (n = 14) DPD activity to determine the contribution of each variant allele to low DPD activity in vivo. Eight of the 13 previously defined DPYD mutations (G62A, ΔTCAT295-298, C703T, G1003T, G1156T, ΔC1897, G2657A, and G2983T) were not detected. A previously defined exon 13 mutation (G1601A) was detected in three individuals with reduced DPD activity. An exon 14 splice donor site mutation (intron14 G1A) was detected in a normal DPD activity individual. It was demonstrated that T85C, A1627G and G2194A are common polymorphisms. A novel exonic mutation (T1679G) was detected in a patient with reduced DPD activity and 5FU toxicity. In addition, three novel common polymorphisms were detected in introns 10 and 13. Only three patients did not have any mutations and 30 had multiple DPYD mutations in the regions examined. However, only 17% (4/23) of the patients with a low DPD phenotype have a molecular basis for reduced activity. Although novel DPYD variants have been identified in this study, the 17 DPYD mutations now described do not entirely explain polymorphic DPD activity and toxic response to 5FU. These data emphasize the complex nature of the molecular mechanisms controlling polymorphic DPD activity in vivo.