Phase I trial of nelarabine in indolent leukemias

To test whether nelarabine is an effective agent for indolent leukemias and to evaluate whether there is a relationship between cellular pharmacokinetics of the analog triphosphate and clinical responses.

Thirty-five patients with relapsed/refractory leukemias (n = 24, B-cell chronic lymphocytic leukemia and n = 11, T-cell prolymphocytic leukemia) were entered onto three different protocols. For schedule A, patient received nelarabine daily for 5 days, whereas for schedule B, nelarabine was administered on days 1, 3, and 5. Schedule C was similar to schedule B except that fludarabine was also infused. Plasma and cellular pharmacokinetics were studied during the first cycle.

Responses were achieved in 20%, 15%, and 63% of patients receiving schedule A, B, and C, respectively. Histologic category, number of prior therapies, and fludarabine refractoriness did not influence the response rate. The most common nonhematologic toxicity was peripheral neuropathy. Grade 4 neutropenia and thrombocytopenia complicated 23% and 26% of courses respectively, and were significantly more frequent among patients with pre-existing marrow failure. Pharmacokinetics of plasma nelarabine and arabinosylguanine (ara-G) and of cellular ara-G triphosphate (ara-GTP) were similar in the two groups of diagnoses, and the elimination of ara-GTP from leukemia cells was slow (median, > 24 hours). The median peak intracellular concentrations of ara-GTP were significantly different (P = .0003) between responders (440 micromol/L; range, 35 to 1,438 micromol/L; n = 10) and nonresponders (50 micromol/L; range, 22 to 178 micromol/L; n = 15).

Nelarabine is an effective regimen against indolent leukemias, and combining it with fludarabine was most promising. Determination of tumor cell ara-GTP levels may provide a predictive test for response to nelarabine.