Sensitivity and Cross-Resistance to Deoxynucleoside Analogs in Childhood Acute Leukemia.

Ara-C is a key drug in the treatment of AML and ALL. However, resistance may occur, when it is associated with treatment failure. Resistance might be circumvented by other deoxynucleoside analogs. We previously reported increased resistance to ara-C in childhood leukemia subgroups, e.g. in T-cell ALL and AML FAB type M1 and M2. We have now studied subgroup-related differences in sensitivity to ara-C analogs in 362 untreated childhood leukemia and 32 normal pediatric bone marrow samples; all obtained with informed consent. To analyze group-differences the Mann-Whitney-U test was used. In addition, cross-resistance patterns were determined within a group of 60 AML samples in which all 4 deoxynucleoside analogs had been tested successfully, applying the Pearson correlation test after log-transformation of the data. Statistical significance was set at p<0.01. The 4-days total cell kill colorimetric tetrazolium-based MTT assay was used to determine the cytotoxicity of ara-C, cladribine, fludarabine and gemcitabine, each tested at 6 concentrations in duplicate. After the 4 days, control wells had to contain at least 70% of leukemic blasts and the mean control optical density had to exceed 0.050 arbitrary units for an assay to be considered successful. Samples were all tested in Amsterdam but were provided by several study groups and many centers over a 10 year period. Normal BM mononuclear cell samples were significantly more resistant than leukemic samples to the deoxynucleotide analogs, except for gemcitabine, and in the case of ALL cladribine. Only cladribine was significantly more cytotoxic (1.4-fold) to AML compared to ALL cells, while the other drugs were equally active. Within AML subtypes, both ara-C and cladribine were significantly more cytotoxic (2-fold) in FAB type M5 samples than in M1/M2 and M4 samples, while fludarabine and gemcitabine showed similar cytotoxicity. Within ALL subtypes, T-cell ALL samples were more resistant to cladribine (7-fold, p=0.001) and borderline to ara-C (2-fold, p=0.04) than BCP-ALL samples. Too few (6) T-ALL samples were successfully tested for fludarabine and gemcitabine for statistical analysis. Finally, cross-resistance was observed between all 4 analogs, with the strongest correlations being between the structurally closest related drugs: ara-C vs gemcitabine: Rp=0.67, p<0.0001, and fludarabine vs cladribine: Rp=0.72, p<0.0001. Drugs other than deoxynucleoside analogs were tested against AML samples as well. Cross-resistance was also observed between ara-C and these drugs, although to a lesser extent: vs etoposide: Rp=0.54, p<0.0001; vs daunorubicin: Rp=0.48, p<0.0001; vs vincristine: Rp=0.33, p=0.01; and vs l-asparaginase: Rp=0.30, p=0.02. In conclusion, drug resistance in childhood leukemia can be explained by drug-type specific as well as more general mechanisms, e.g. apoptosis or DNA repair. Important subgroup-related differences in sensitivity to deoxynucleoside analogs exist and are not limited to ara-C. Increased resistance to ara-C in vitro could not be overcome by the substitution of an alternative analog; cladribine, fludarabine or gemcitabine. However, cladribine may have advantages in AML, particularly in FAB type M5.