Using synthetic oligomers we investigated fresh samples of acute myeloid leukemia (AML) for the presence of mutated ras oncogenes. Our original results showed that five of eight samples contained a mutation in codon 13 of the N-ras gene. In a subsequent study involving 37 samples, we found only one N-ras-13 mutation, and, in addition, mutations in codon 61 of the N-ras gene in four cases and a mutation in codon 12 of the Ki-ras gene in two cases. Amplification of ras genes was not observed. We conclude that in approximately 20% to 25% of AML cases, a mutated ras oncogene is present, predominantly the N-ras gene. The occurrence of mutations does not correlate with the cytological features of the leukemia.

In the present study we demonstrate that Aza-dC in combination with Amsacrine has major antileukaemic properties in patients who have not already received extensive Ara-C therapy. Eight out of 11 patients in their first relapse of acute leukaemia achieved complete remission. Cross resistance between Ara-C and Aza-dC was revealed by the lack of antileukaemic activity in five patients with with Ara-C resistant leukaemia. Combination therapy with Aza-dC/Ams-acrine induced a considerable period of a granulocytopenia (28-35 days), while the toxic effect on erythro- and megakaryopoiesis was comparable to that reported for high dose Ara-C/Amsacrine chemotherapy. Remarkable is the long disappearance time for leukaemic blast cells in bone marrow, i.e. 3-5 weeks in some cases. Analysis of cell membrane markers showed a loss of the early differentiation antigens CD34 and CD33 from leukaemic bone marrow cells after 7 days of Aza-dC treatment, which is suggestive of leukaemic cell differentiation. In the small group of patients tested for DNA hypomethylation no association existed between the degree of hypomethylation and clinical response. Non-haematologic side effects were considerable in patients receiving the highest dosages of Aza-dC and consisted of severe, although usually reversible, gastrointestinal and neurological complications. In comparison with Ara-C, Aza-dC causes less nausea and vomiting and is therefore better tolerated.

Summary Using a Brown Norway rat leukaemia model (BNML), which is a realistic model of human myelocytic leukaemia, we compared the antileukaemic activity, influence on cell cycle kinetics and effect on normal haematopoiesis of 5 aza-2-deoxycytidine (aza-dC) and arabinofuranosyl-cytosine (ara-C). The antileukaemic activity was evaluated by means of a survival study. For aza-dC a dose-response relationship was demonstrated for doses up to 50mgkg-1 (3 times q 12h); a higher dose resulted in only a slight increase in median survival time (MST). For ara-C a weak dose-response relationship was observed. At the maximum dose of aza-dC and ara-C tested, aza-dC induced a 10-day longer survival time than ara-C, which means 2 logs more of leukaemic cell kill for aza-dC.By means of flow cytometric analysis and a 3HTdR uptake study it was shown that aza-dC does not influence the cell cycle kinetics in the first 24 h after exposure, in contrast to ara-C which caused the characteristic G1 /S blockage and synchronization. The influence of aza-dC and ara-C on normal haematopoiesis was evaluated with the CFU-S assay. The dose-response curve for CFU-S did not show a significant difference in stem cell cytotoxicity between aza-dC and ara-C. In the BNML model aza-dC is a much more effective antileukaemic agent than ara-C, while the toxic effect on normal haematopoiesis is comparable to that of ara-C.