Gemtuzumab ozogamicin (GO or Mylotarg®) is increasingly being used in the treatment of AML. GO consists of a cytotoxic drug - calicheamicin which is conjugated to an anti-CD33 antibody. Most AML patients highly express CD33 on their blasts. Studies relating CD33 expression to response to GO have failed to show an association. Although 30–50% of patients respond to Mylotarg®, the causes of primary resistance to this drug remain unclear. Previous reports have studied P-gp status and CD33 expression as a cause of resistance to GO. Another factor that might determine response to GO is cellular drug resistance to calicheamicin.

In this study we examined in vitro resistance to calicheamicin in 90 initially diagnosed and 32 relapsed pediatric AML samples using the 4 day MTT assay (concentration range 0.000004 – 0.4 μg/ml). The LC50 value, the drug concentration at which 50% of the cells is killed by the drug, is used as a measure of sensitivity. In addition to calicheamicin, some samples were also tested successfully for in vitro sensitivity to etoposide, cytarabine, daunorubicin, idarubicin, mitoxantrone, 6-thioguanine, vincristine and L-asparaginase.

The characteristics of the 122 pediatric AML samples included are as follows: 62% boys, median age 9.6 years, median WBC 53.0*109/L, FAB types M0 n=9, M1 n=10, M2 n=18, M3 n=9, M4 n=33, M5 n=24, M7 n=3, unknown n=16. There was a more than 100,000 fold difference in calicheamicin sensitivity between the most sensitive and the most resistant patient samples. FAB M2 samples taken at initial diagnosis (n=13) were significantly more resistant to calicheamicin compared to the other FAB types (Resistance Ratio (RR)=2.5, median LC50 0.033 vs. 0.013 μg/ml, p=0.008). Newly diagnosed AML samples were significantly more sensitive to calicheamicin compared to relapsed AML samples (RR=0.68, median LC50 0.023 vs. 0.034 μg/ml, p=0.042) (although these patients had not been treated with calicheamicin). There was strong cross-resistance between calicheamicin and the anthracyclines idarubicin (Spearmans rho = 0.73, p<0.0001, n=23), daunorubicin (rho=0.61, p<0.0001, n=103) and mitoxantrone (rho=0.52, p=0.039, n=16). In addition, there was moderate cross-resistance with etoposide (rho=0.42, p<0.0001, n=101). No cross-resistance was observed between calicheamicin and cytarabine (rho=0.11, p=0.28, n=106), 6-thioguanine (rho=0.20, p=0.054, n=97), vincristine (rho=0.12, p=0.44, n=46) or L-asparaginase (rho=0.21, p=0.16, n=45). In conclusion, the interpatient differences in calicheamicin sensitivity are the largest differences in in vitro drug sensitivity we have ever observed in pediatric AML. FAB M2 samples are 2.5 fold more resistant to calicheamicin than samples with other FAB types. Initially diagnosed pediatric AML samples are 1.5 fold more sensitive to calicheamicin than relapsed AML samples. There is marked cross-resistance between calicheamicin and the related anthracyclin compounds. Given the large differences in sensitivity to calicheamicin in pediatric AML samples, it is likely that calicheamicin resistance plays a role in resistance to Mylotarg®.

Author notes

Corresponding author