Correlations of in vitro drug-induced apoptosis with blast cell Pgp features and with achievement of complete remission (CR and CRp) or clearance of marrow blasts after treatment with gemtuzumab ozogamicin.
(A) In vitro blast cell susceptibility to gemtuzumab ozogamicin, as indicated by drug-induced apoptosis, was greater among samples with lower Pgp surface expression; however, the regression coefficient did not reach statistical significance. (B) Higher levels of drug-induced apoptosis were observed predominantly in blast cell samples with lower baseline DiOC2 efflux ratios. This correlation was statistically significant among samples with baseline dye efflux ratios greater than or equal to 1.21 and sensitivity to CSA inhibition of dye efflux (black circles and solid line). In contrast, no correlation was seen among samples either insensitive to CSA or with baseline efflux ratios less than 1.21 (dotted line). (C) Blast cell samples from patients with CR or CRp were more susceptible to drug in vitro, as indicated by a significantly higher median percentage of gemtuzumab ozogamicin–induced apoptosis (12%), compared with samples from patients who failed to achieve remission (median 4% apoptosis; *P = .03). CSA-sensitive DiOC2 efflux (black circles) was more common among the nonresponder patient samples (46%) than samples from responders (19%). (D) Similarly, the median value for the percentage of gemtuzumab ozogamicin–induced apoptosis among samples from patients who cleared marrow blasts was significantly higher than the median value among the samples from patients with residual leukemia (8% vs 2%; *P = .0009). Of note, a level of apoptosis greater than 7% was observed exclusively in samples from patients with marrow blast clearance, whereas lower levels of apoptosis were observed in samples from both responders and nonresponders. CSA-sensitive DiOC2 efflux was more common in nonresponder samples (61%) compared with samples from patients who achieved 5% or fewer marrow blasts (27%).
