Randomized Phase II Trial of Timed Sequential Cytosine Arabinoside with and without the CHK1 Inhibitor MK-8876 in Adults with Relapsed and Refractory Acute Myelogenous Leukemia

Abstract

Background: Cytosine arabinoside (ara-C) remains the backbone of most regimens for the treatment of AML. The incorporation of ara-C into DNA activates checkpoint kinase 1 (Chk1) causing replication fork slowing, the accumulation of cells in S phase, and diminished ara-C cytotoxicity. The selective Chk1 inhibitor MK-8776 abrogates ara-C-induced S-phase arrest, thereby enhancing ara-C-induced apoptosis in vitro.1 A phase I trial demonstrated the safety of MK-8876 in combination with infusional ara-C and correlative studies demonstrated increased ara-C-induced H2Ax phosphorylation in AML cells, consistent with unrepaired DNA damage.2 Herein, we report data from a phase II trial comparing ara-C plus MK8876 with ara-C alone.Methods: Patients with relapsed or refractory AML were randomized to receive either ara-C 2 gm/m2 via continuous infusion (CIV) over 72 hours on days 1 and 10 with MK-8776 100 mg flat dose infused over 30 minutes on days 2, 3, 11, and 12 (combo); or ara-C alone on the same schedule (control). The primary end point was response rate (CR+CRi). Adverse events (AEs) for all patients were reported according to the NCI-CTCAE version 4.0. Correlative studies were performed on blood and marrow samples collected prior to treatment on days 1, 2 and 3 and at one hour after MK-8776 infusion (PB only). Marrow and circulating blasts were analyzed using semi-quantitative flow cytometric determinations of blast cell gH2AX expression, a marker of DNA damage.Results: Thirty-two patients were treated at 4 sites with 14 assigned to the combo arm and 18 to the control arm. The study was stopped early due to lack of drug availability. The median age was 60 (range 29-72) and 17 (53%) were male. Nine (28%) had refractory AML, 16 (50%) were in first relapse, and 7 (22%) were in second relapse. Eleven (34%) had undergone a prior bone marrow transplant (BMT). Risk stratification classified 1 (3%) patient as favorable risk disease, 11 (34%) as intermediate-1, 8 (25%) as intermediate-2, and 11 (34%) as adverse risk.There were 5 (36%) responses (CR/CRi) and 1 (7%) partial response (PR) in the combo arm, and 8 (44%) responses and 1 (6%) PR in the control arm. Median survival did not differ significantly between the two groups (5.9 months in the combo arm vs. 4.2 months in the control arm). Sub-group analysis showed increased responses in the combo arm for the patients in first relapse (50% vs. 40%) and those with a prior BMT (50% vs. 29%). Most non-hematologic grade 3/4 treatment-related AEs were similar in the two groups including febrile neutropenia (64% vs. 61%), infectious complications (50% vs. 50%), and sepsis (21% vs. 17%). Patients in the combo arm experienced more QT prolongation (21% vs. 0%), hypertension (21% vs. 0%), and GI toxicities (21% vs. 0%); while those in the control arm had increased renal (0% vs. 11%) and liver toxicity (7% vs. 17%). Two deaths were attributed to treatment, and both were in the control arm.Eight combo patients had serial PB samples and 2 had serial marrow samples studied by flow cytometry. There was an increase in the percentage of gH2AX+ CD34+ cells on Day 2 directly after MK-8876 infusion (16.9% prior and 36.4% at one hour), which occurred to a lesser extent on Day 3 after MK-8876 infusion (12.7% prior and 15.6% at one hour). Three of eight patients demonstrated a sustained increase in gH2AX+ CD34+ from Day 2 to Day 3, and all three achieved either a PR or CR.Conclusion: This is the first phase II study building on prior work suggesting that inhibition of Chk1 would increase response to ara-C in patients with relapsed or refractory AML. The study confirmed that treatment with the combination is feasible with similar toxicities between the treatment and control groups. The response rates were similar between the two groups with improved responses in patients receiving the combo in first relapse and following BMT. Correlative studies support the finding that responding patients show a sustained increase in DNA damage and this may serve as a biomarker in future studies. Chk1 remains a potential target to improve treatment outcomes with ara-C.