Glucocorticoid-induced cell death and proliferation in vitro in childhood acute myeloid leukemia

Glucocorticoids (GC) have significant antileukemic activity in ALL, but their role in AML is less certain. We previously reported that the majority of AML samples was resistant to GC in vitro, and that in 10 out of 28 AML samples prednisolone (and dexamethasone if tested) even induced proliferation in vitro (Kaspers et al., Leukemia 1994). We have extended the number of observations and now studied the clinical relevance of resistance to prednisolone (PRD) and of PRD-induced proliferation in pediatric AML samples taken at initial diagnosis. Using the colorimetric 4 days MTT assay, 126 AML samples were tested successfully. Drug resistance was expressed in LC50 values, the PRD concentration needed to kill 50% of cells. Sensitivity to PRD was arbitrarily defined as LC50 values < 150 ug/ ml, the remainder of the samples as resistant (Kaspers et al., Blood 1998). PRD-induced proliferation was defined as a cell survival greater than that in control wells not containing drug after the 4 days of culture. All patients were treated with PRD during the so-called consolidation course. Results: 15% of the AML samples were relatively sensitive to PRD (although LC50 values were still high compared to most ALL samples, data not shown). PRD sensitivity was more often seen in samples with FAB type Ml (6/14) than in non-Mi samples (13/110, p=0.002; 2 cases not évaluable). PRD-induced AML cell proliferation was seen in 32/126 samples, more often in FAB M5 (13/24) than in non-M5 samples (19/ 102, p<0.001). We reported previously that within AML, FAB M5 samples are relatively in vitro sensitive to most drugs, except PRD (Zwaan et al., Blood 2000). PRD sensitivity or resistance did not significantly correlate with the probability of DPS, although it seemed that in vitro sensitive patients relapsed later. Patients with PRD-induced AML cell proliferation had a lower EPS (20±10% vs 50±7%, p=0.002) and tended to have a lower DPS (30+15% vs 55±8%, p=0.08). In conclusion, the far majority of newly diagnosed pediatric AML samples is in vitro resistant to GC-induced cell death. GC-induced cell proliferation in vitro occurs in about 25% of AML samples (even in about 50% of FAB M5 samples), which phenomenon is associated with a worse clinical outcome.