A Phase I Trial of the Epigenetic Modulators Vorinostat, in Combination with Azacitidine (azaC) in Patients with the Myelodysplastic Syndrome (MDS) and Acute Myeloid Leukemia (AML): A Study of the New York Cancer Consortium

Background: The hypomethylating agent azacitidine (azaC) which can reverse epigenetic silencing, is the first agent demonstrated to alter the natural history of MDS and improve survival in higher-risk patients (Silverman JCO 2002, Fenaux Blood 2007). AzaC also produces comparable rates of response in patients with non-proliferative AML and appears to affect survival (Silverman JCO 2006). Time to response is slow with single agent azaC, requiring a median of 3 to 4 cycles to initial response and the CR + PR rates ranges from 7 to 27%. Vorinostat, a histone deacetylase inhibitor (HDACI) which inhibits class I and II HDAC, has demonstrated single agent activity in patients with MDS and AML with responses of 25% (Garcia-Manero Blood 2006). In vitro the 2 agents are synergistic in reactivating epigenetically silenced genes. The effect is sequence dependent requiring exposure to the hypomethylating agent first followed by the HDACI. This study was designed to test the safety of the combination and to determine the response rate and effects on the MDS/AML clone.

Methods: In the phase I component eligible patients were entered into one of eight cohorts with the combination of vorinostat and azaC in a 3+3 dose escalating de-escalating design (see table).

Results: As of the data cut for this submission 21 patients are evaluable for toxicity and response. Accrual to all 8 cohorts has been completed, full data will be available for the meeting. Among the 28 patients entered: 20 have MDS and 8 AML with median age 68. Responses among evaluable patients have occurred in 18 of 21 (86%); 9 CR, 2 CRi, (CR+CRi=53%) 7 HI, 2 SD. Median time to response is 2 cycles. Among patients with high risk MDS and AML 10/12 (83%) responded (5CR, 1CRi, 4 HI). Responses including CR occurred, but in 57% of patients the abnormal MDS/AML clone persisted, suggesting a modulating effect of the combination on the clone. A total of 171 cycles have been administered, range 1 to 17 with a mean 5 cycles. Eight patients have come off study for progression (1); relapse (2); co-morbidities (2); or consent withdrawal (3). No grade 3 or 4 non-hematologic toxicities in cycle 1 were observed. Grade 2 fatigue occurred during cycle 1 in 89% of patients in cohorts 2, 3 and 4. Grade 2 anorexia and fatigue occurred in cycle 1 in 31.6% and 57.9% of patients, respectively. The fatigue correlates with the scheduled duration of vorinostat administration with 14 days vorinostat (cohorts 1–4) associated with grade 2 and 7 days (cohorts 5–7) producing grade 1. Some suggestion of cumlative fatigue (grade 3) occurred in cycles 2 and beyond in cohorts 1, 3 and 4 with 14 days of vorinostat. However it did not result in discontinuation of therapy. Correlative biologic studies are underway.

Conclusion: The combination of azaC and vorinostat can be safely combined, is well tolerated in repetitive cycles and is active in both lower and higher risk/AML patients with an OR and CR rate superior to azaC alone. AzaC at a dose of 55 mg/m2 appears to be optimal for combination. The phase II study will utilize an optimization design to further identify the optimal biologic/epigenetic effects among the three vorinostat schedules of 3, 7 or 14 days when combined.