Significant Clinical Activity of the Combination of 5-Azacytidine, Valproic Acid and All-Trans Retinoic (ATRA) Acid in Leukemia: Results of a Phase I/II Study.

5-azacitidine (5-AC), and its analogue 5-aza-2′-deoxycitidine (5-DAC), are DNA hypomethylating agents. Valproic acid (VPA) is a histone deacetylase inhibitor (HDACI). Hypomethylating agents and HDACI restore ATRA sensitivity in resistant cells. The combination of a hypomethylating agent with an HDACI is antileukemic in vitro. Based on this, and on a prior study of 5-DAC and VPA (Garcia-Manero, Blood, in press), we developed a phase I/II study of the combination of 5-AC, VPA and ATRA. Patients (pts) with high risk MDS (≥10% blasts); relapsed/refractory AML; and pts > 60 years with untreated disease and adequate renal, hepatic functions and performance status were eligible. A fixed-dose of 5-AC was used: 75 mg/m² sq daily x 7. ATRA dose was: 45 mg/m² PO daily x 5 starting on day 3 of 5-AC. The phase I followed a classic 3+3 design. The initial dose of VPA was 50 mg/kg PO daily x 7 concomitantly with 5-AC. Cycles were at least 21 days long. The phase II targets a response rate of ≥30% with stopping rules. Cohorts of 10 pts (max 40) are studied. Thirty one pts are evaluable. Median age is 60 years (5–78). All, but 2 pts with MDS, had AML. Median number of prior therapies was 2 (0–5). Twenty six pts (83%) had abnormal cytogenetics. During the phase I, at a VPA dose of 50 mg/kg, 1/6 pts developed grade 3 non-hematological toxicity; at 62.5 mg/kg, 2/7, and at 75 mg/kg, 3/6. The dose limiting toxicity (DLT) was neurotoxicity, confusion and somnolence. The MTD was 50 mg/kg daily x 7. Twelve pts are enrolled in the phase II. Of the 31 evaluable pts, 9 achieved CR (ANC 10 ⁹/L, platelets 100 x 10 ⁹/L, and marrow blasts ≤5%) and 3 a CRp (same criteria as of CR but without complete platelet recovery), overall response (OR)39%. The median number of courses to response was 1 (1–3). Nine responses occurred in 16 untreated pts, OR 56% (Table). Eighteen patients were treated at the MTD with 9 responses (50%). Cytogenetic responses were observed in all responding pts. To assess the hypomethylating effect of 5-AC, the LINE test was used. Median LINE methylation pretreatment was 62.5% (57–67%), declined to 58% by day 7 and returned to baseline by day 0 of next cycle (p<0.001). No significant difference was observed between responders and non responders. Eighteen pts had evidence of histone 3 and 4 acetylation (66%). VPA bound level on day 2 was 146 mcg/ml (95% CI 122–170) in responders and 103 mcg/ml (95% CI 88–118) in non-responders, p< 0.005. Analysis of gene re-expression and cell differentiation are ongoing. In conclusion, the combination of 5-AC, VPA and ATRA is safe. The MTD of VPA is 50mg/kg daily x 7. The DLT is neurotoxicity. The combination has significant clinical activity. An OR of 56% was observed in patients > 60 years with previously untreated AML/MDS. Histone acetylation and transient global hypomethylation are observed. Higher levels of VPA are found in responders. Based on prior CALGB experience (Silverman, ASH 2005), this combination appears to be more active than single agent 5-AC in AML. The study continues to accrue.