Correspondence

Abstract

Thiopurine methyltransferase (TPMT) degrades 6-thioguanine, 6-mercaptopurine and azathioprine, which are commonly used drugs in the treatment of acute leukaemia and autoimmune disease. A number of mutations in the TPMT gene have been identified that influence the haematological toxicity of these agents (Yates et al, 1997). Approximately 10% of caucasians are heterozygous carriers of null mutations and have increased susceptibility; 1 in 300 patients are homozygous or compound heterozygous and are at risk of severe, life-threatening toxicity. To our knowledge, the significance of TPMT genotype/phenotype status has not been systematically examined in patients with acute non-lymphoblastic leukaemia (ANLL) receiving thioguanine as part of their chemotherapy regime. We describe a 35-year-old man with M2eo-ANLL and an inv16 cytogenetic abnormality who was entered into the current AML-12 trial in the UK. He was randomized to the H-DAT + all-trans retinoic acid (ATRA) schedule, which consists of daunorubicin 50 mg/m2 i.v. d 1, 3 and 5; cytosine arabinoside 200 mg/m2 i.v. 12 hourly d 1–10; thioguanine 100 mg/m2 12 hourly orally on d 1–10 and ATRA 45 mg/m2 orally continuing during and after courses 1 and 2. However, the development of transient respiratory symptoms on d 3 prompted exclusion of the retinoid because of possible ATRA syndrome. Complete remission was achieved after the first course; a d 28 marrow sample was cellular with < 2% blasts and was negative for inv 16 using standard cytogenetics and fluorescence in situ hybridization (FISH) analysis. Haemopoietic regeneration was normal with neutrophil and platelet recovery to 1·0 and 100 × 109/l, respectively, on d 28. After a second course of H-DAT (cytosine arabinoside and thioguanine given on d 1–8), however, haemopoietic regeneration was significantly delayed. A marrow trephine biopsy on d 40 was markedly hypocellular. Neutrophil recovery occurred on d 45, but the platelet count failed to reach 100 × 109/l by d 80, and the patient continued to require red cell transfusions at this time. Cytogenetics/FISH analysis performed at d 56 showed no evidence of relapse. There were no septicaemic episodes or other major complications during the course of the second cycle and platelet-specific antibodies were negative using immunofluorescence and monoclonal antibody-specific immobilization of platelet antigens (MAIPA) assays. Genotyping for TPMT mutations was performed on marrow cells using a reverse transciption polymerase chain reaction/heteroduplex-based screening method recently developed in our laboratory and showed that this patient was heterozygous for the *3A allele. The G460A and A719G mutations were confirmed by direct sequencing. TPMT*3A is the commonest mutant allele identified in caucasian individuals and comprises 85% of the total (Yates et al, 1997). The heterozygous genotype is associated with an approximate 50% reduction in red cell TPMT activity (Coulthard et al, 2000). TPMT genotype has recently been shown to be an important determinant of mercaptopurine toxicity in children with acute lymphoblastic leukaemia receiving maintenance therapy (Relling et al, 1999) and dose-limiting myelosuppression in patients with rheumatoid arthritis treated with azathioprine (Black et al, 1998). The clinical course of our patient raises the possibility that TPMT mutations might also influence thioguanine toxicity in patients with ANLL. Pharmacogenetic factors could be particularly important when this agent is included in regimes that approach maximum haemopoietic tolerance. We would suggest that this might be examined systematically as part of ongoing national trials in ANLL.