Resistance to chemotherapy is a major problem in the treatment of acute myeloid leukemia (AML). As cytosine-arabinoside (Ara-C) is an important agent in the treatment of AML it is conceivable that leukemic blasts become resistant to Ara-C during the development to relapse/resistant disease. Ara-C is a cytotoxic nucleoside analogue which is phosphorylated intracellularly into its active form Ara-CTP, by the action of three enzymes: deoxycytidine kinase (dCK), deoxycytidine monophosphate (dCMP) kinase and nucleoside diphosphate (NDP) kinase. Ara-CTP inhibits DNA polymerase, and acts by competing with its physiological counterpart, the natural substrate dCTP, for incorporation into nucleic acids. Although several resistance mechanisms are involved in Ara-C metabolism, deoxycytidine kinase (dCK) is of particular interest because it is the rate-limiting enzyme in the phosphorylation process from Ara-C to Ara-CTP. Structural analysis of the dCK gene has revealed mutations which are associated with dCK deficiency and Ara-C resistance in vitro and in vivo. Flasshove et al1 found two silent point mutations (codon 86 and 42), and five mutations resulting in amino acid changes (codon 20, 93, 99, 98, 154) in the dCK cDNA in seven out of 16 adult patients with relapsed and refractory AML. One of them, a point mutation in codon 99 (TAT! TGT) leading to an amino acid substitution from tyrosine to cysteine, was associated with absent dCK activity, whereas the enzyme activity was normal in patients with a point mutation in codon 98 and 20.

We searched for mutations in the dCK gene in a unique set of paired purified bone marrow samples obtained from 30 AML patients at diagnosis and at relapse and/or refractory disease and from one CML patient at first and second blast crisis (10 children, 21 adults). All patients had been treated with chemotherapy according to national protocols including anthracyclines and Ara-C. Initially, we used the procedure as described by Flasshove et al. 1 Briefly, total RNA was isolated and converted to cDNA in a reverse transcriptase reaction. The dCK cDNA was amplified by polymerase chain reaction (PCR) using Taq polymerase and two primers, flanking the dCK coding sequence as described. The single PCR product of 884 base pairs, that could be detected in all samples, was cloned into pCR II (TA cloning kit; Invitrogen, Carlsbad, CA, USA). Both strands of at least three clones from 21 diagnostic samples and 25 samples of relapsed and/or refractory AML patients were completely sequenced. Thirty-three different point mutations were found (12 silent and 21 giving rise to amino acid substitutions) in the diagnostic samples, and 61 mutations were found (15 silent and 46 giving rise to amino acid substitutions) in the relapse samples. Some dCK inserts contained no mutations at all whereas others, derived from the same patient sample, displayed multiple mutations. Mutations found in diagnostic samples were not detected in relapse samples of the same patient, nor was the reverse observed. Both the large number of silent mutations that cannot give rise to altered dCK enzymic activity and the abundant number of randomly found mutations suggest that these were experimental artifacts. Therefore, to adjust our procedure, we used another thermostable DNA polymerase that, in contrast to Taq DNA polymerase, possesses 39 to 59 exonuclease proofreading activity enabling the polymerase to correct nucleotide incorporation errors. Secondly, we used a direct method to sequence the amplified PCR product. We attempted to use high-fidelity Pfu DNA polymerase but were unable to reliably amplify dCK cDNA. These problems …