Abstract: Wilms' tumor 1 (WT1) protein is highly immunogenic and overexpressed in acute myeloid leukemia (AML), consequently ranked as a promising target for novel immunotherapeutic strategies. Here we report our experience of a phase I/II clinical trial (NCT01051063) of a vaccination strategy based on WT1 recombinant protein (WT1-A10) together with vaccine adjuvant AS01B in five elderly AML patients (median age 69 years, range 63-75) receiving a total of 62 vaccinations (median 18, range 3-20) after standard chemotherapy. Clinical benefit was observed in three patients: one patient achieved measurable residual disease clearance during WT1 vaccination therapy, another patient maintained long-term molecular remission over 59 months after the first vaccination cycle. Interestingly, in one case, we observed a complete clonal switch at AML relapse with loss of WT1 expression, proposing suppression of the original AML clone by WT1-based vaccination therapy. Detected humoral and cellular CD4+ T cell immune responses point to efficient immune stimulation post-vaccination, complementing hints for induced conventional T cell infiltration into the bone marrow and a shift from senescent/exhausted to a more activated T cell profile. Overall, the vaccinations with WT1 recombinant protein had an acceptable safety profile and were thus well tolerated. To conclude, our data provide evidence of potential clinical efficacy of WT1 protein-based vaccination therapy in AML patients, warranting further investigations

Abstract: Despite routine use of DNA-hypomethylating agents (HMAs) in AML/MDS therapy, their mechanisms of action are not yet unraveled. Pleiotropic effects of HMAs include global methylome and transcriptome changes. We asked whether in blasts and T-cells from AML patients HMA-induced in vivo demethylation and remethylation occur randomly or non-randomly, and whether gene demethylation is associated with gene induction. Peripheral blood AML blasts from patients receiving decitabine (20 mg/m2 day 1-5) were serially isolated for methylome analyses (days 0, 8 and 15, n = 28) and methylome-plus-transcriptome analyses (days 0 and 8, n = 23), respectively. T-cells were isolated for methylome analyses (days 0 and 8; n = 16). We noted massive, non-random demethylation at day 8, which was variable between patients. In contrast, T-cells disclosed a thousand-fold lesser, random demethylation, indicating selectivity of the demethylation for the malignant blasts. The integrative analysis of DNA demethylation and transcript induction revealed 87 genes displaying a significant inverse correlation, e.g. the tumor suppressor gene IFI27, whose derepression was validated in two AML cell lines. These results support HMA-induced, non-random early in vivo demethylation events in AML blasts associated with gene induction. Larger patient cohorts are needed to determine whether a demethylation signature may be predictive for response to this treatment