Elevating AML Treatment: Enhancing DNMT1 Inhibition Boosts HMA Therapy

Acute myeloid leukemia (AML) is one of the most common leukemias in adults. Still, treatment poses a significant challenge, especially in elderly patients, who are unfit for standard chemotherapy. For these patients, hypomethylating agents (HMA) such as azacytidine (AZA) and decitabine (DAC) are often the only available treatment. Nevertheless, resistance frequently develops, and strategies to overcome these remain elusive. Moreover, the exact mechanisms of HMA action, potential response modulators, and reliable biomarkers remain to be defined.

To identify modifiers to HMA response in AML cells, we conducted an inducible-pooled short-hairpin RNA (shRNA)-based library screen, targeting all epigenetic regulation genes listed in the KEGG pathway. After ten days of treatment and shRNA induction, the gene expression profiles were linked to treatment response. Candidate genes were further validated through Western Blot and growth competition assays using FACS. We further determined the global methylation levels employing EPIC 450K BeadChip and performed functional enrichment analysis (GSEA). Cell cycle and DNA damage responses were compared to cytarabine (CYT) using FACS and high-throughput laser-scanning microscopy.

Our screen identified crucial genes influencing HMA sensitivity, including DNMT1 itself, chromatin modulators, as well as DNA damage response (DDR) genes such as members of the PARP family and BRCA1.

DNMT1 knockdown (kd) significantly increased the efficacy of AZA and DAC, causing higher G2/M phase arrest (nearly 30% vs. 18% in single treatment) with a significant drop in AML cell viability (from 60% to 20% over eight days). This effect was also prominent in global methylation levels. While AZA treatment alone decreased DNA methylation, DNMT1 kd enhanced this effect correlating with cell depletion in a growth competition assay. GSEA analysis showed increased interferon-γ pathways in hypomethylated regions, though Sting kd did not reduce HMA efficacy.

Moreover, all compounds, HMA and CYT, triggered a similar DNA damage (gamma-H2AX) (70% compared to CTR). However, CYT caused larger and brighter gamma-H2AX-foci, indicating more severe damage compared to HMA. This damage could be caused by replication fork termination in all substances. Further analysis of the DDR pathway showed that all compounds activated pATM and pATR, however, with HMAs specifically increasing pATM activity and 53BP1 activation, leading to rapid but error-prone non-homologous end joining (NHEJ) repair. DNMT1 kd further enhanced the pATM-53BP1 pathway activation. Again, DNMT1 kd enhanced the pATM-53BP1 pathway choice.

In summary, our results demonstrate the central role of DNMT1 in HMA response in AML. While increasing HMA concentration leads to cytotoxicity, combining HMAs with additional DNMT1 inhibitors shows promise for enhancing HMA-based treatment. We already could confirm in preliminary experiments the synergistic effects of HMA with other inhibitors targeting DNMT1 and are currently combining the new DNMT1 inhibitor GSK-3484862 with AZA and DAC. Furthermore, we highlighted the induction of the pATM-53BP1 axis and its role in NHEJ repair, suggesting that modulating this pathway could potentially improve treatment outcomes.

Lübbert:Syros Pharmaceuticals: Honoraria; Janssen: Research Funding; AbbVie: Honoraria; Cheplapharm: Research Funding; Otsuka: Honoraria.