Maintenance Therapy with 5-Azacytidine (5-AC) after Allogeneic Stem Cell Transplantation (allo-SCT) for Acute Myelogenous Leukemia (AML) and High-Risk Myelodysplastic Syndrome (MDS): A Dose and Schedule Finding Study.

Background: Relapse is the most frequent cause of treatment failure after allo-SCT in patients (pts) with AML or MDS. Median time to relapse after reduced-intensity allo-SCT is 3–5 months, indicating that if post-transplant interventions are to be proposed, they have to be implemented early. 5-AC is a DNA hypomethylating agent that may induce leukemic cell differentiation and increased immunogenicity, therefore potentially increasing the graft-versus-leukemia effect. Furthermore, lower doses are likely to be better tolerated after allo-SCT and to be as effective as larger doses in inducing hypomethylation. We hypothesized that 5-AC after allo-SCT will result in lower relapse rates, and designed a phase I clinical trial to determine the safest dose and schedule combination of 5-AC used as maintenance therapy after allo-SCT in pts with AML or MDS.

Methods: Pts with AML or high-risk MDS not in first remission are eligible. Donors are HLA-compatible related or unrelated. Conditioning regimen is gemtuzumab ozogamicin 2 mg/m² (day -12), fludarabine, and melphalan 140 mg/m². GVHD prophylaxis is tacrolimus and mini-methotrexate. ATG is administered to recipients of unrelated donor transplants. Three doses of 5-AC are to be studied: 8, 16, and 24 mg/m² daily × 5 starting on day 42 post transplant. Four schedules consisting of 1, 2, 3 or 4 28-day courses of 5-AC are being explored for a total of 3×4 = 12 dose-schedule combinations. The continual reassessment method with toxicity probabilities is used to determine the safest dose schedule combination.

Results: Twelve pts were evaluable. Median age was 55.5 years (range 25–66). Eight pts had AML and 4 had MDS. At the time of the transplant 4 pts were in remission and 8 had refractory disease. Donors were HLA compatible related (n=7) and unrelated (n=5). Based on the Bayesian model, 3 pts (25 %) were assigned to 8 mg/m² × 1 cycle, 2 pts (16%) to 8 mg/m²× 2 cycles, 2 (16%) to 8 mg/m²× 3 cycles and 3 to 16 mg/m²×3 cycles. All evaluable pts have received the assigned 5AC doses, and there has been no major drug-related toxicity (at 8 or 16 mg/m²). Two pts died before receiving 5-AC due to bleeding (n=1, in CR) and bacterial sepsis (n=1) post allo-SCT. With a median follow up of 5 months (2–9) after allo-SCT, none of the pts has relapsed and no drug related induction of GVHD has been observed. All pts were 100% donor chimeras at start of 5AC. To assess the DNA hypomethylating effect of 5-AC, the methylation status of long interspersed nuclear elements (LINE) was analyzed by pyrosequencing and used as a surrogate marker of global DNA methylation in mononuclear cells of 7 patients that have completed 1 cycle at the dose of 8 mg/m². Mean LINE methylation pretreatment was 64.9%(±5), while it was 60.7%(±2.5) by day 5 (p=0.06) (last day of 5-AC administration), returning to baseline by day 1 of the next cycle. Analysis of gene specific methylation is ongoing.

Conclusions: At the dose of 8 mg/m² 5AC is well tolerated and may produce detectable levels of hypomethylation. It is unknown if this effect will translate in reduced relapse rate or if higher doses will be as well tolerated.