DOT1L/menin inhibitors target replication fork plasticity and create novel vulnerability to PARP inhibitor in MLL-rearranged AML

Abstract Epigenetic deregulation is a major driver for acute myeloid leukemia (AML) including those with Mixed-Lineage Leukemia gene-rearrangement (MLLr). Over the past decades, significant efforts have been made by others and us to identify and target key components associated with oncogenic transcriptional complexes for development of more effective treatments. Among the inhibitors targeting components of MLL complexes, MENIN and DOT1L have successfully made to the clinics. Although promising results from initial clinical trials demonstrate their safeness and potential efficacy, the complete remission rates remained low (10% for DOT1Li and 33% for MENINi) and short-lived with only 25% of AML patients treated with MENINi being MRD free in the 2 years follow up, urging a better understanding of the underlying mechanisms and the development of effective combination therapies. Strikingly, while suppression of key MLL fusion targets such as Hoxa and Meis1 by these epigenetic inhibitors could be detected in most patients, their downregulation did not strictly correlate with or predict treatment response, suggesting that these inhibitors may also act by different mechanisms. To this end, our study aims to better understand the leukemia suppressive functions of DOT1L/MENIN inhibitors and thereby facilitate the rational development of more effective treatments. To identify key pathways commonly affected by inhibitors targeting DOT1L/MENIN complexes, we performed RNA-seq on both primary MLL-AF9 AML derived from mouse hematopoietic stem/progenitor cells (HSPCs) and human MLL-r AML cell line MV4;11 with or without DOT1Li (EPZ004777) or MENINi (MI-503). In addition to the classical Hoxa/Meis1 pathways, both DOT1L and MENIN inhibitions exhibited significant suppression of pathways associated with DNA replication, closely linked to DNA damage and leukemia cell survival. Upon replication stress, two major pathways, namely fork reversal and repriming, can be used to potentially overcome obstacles to DNA synthesis. Fork reversal is associated with slower fork speed and involves key DNA replication/repair proteins such as PARP to stabilize reversed forks during pausing. Conversely repriming requires Primpol to bypass obstacles without slowing fork progression. To assess the impact of DOT1Li and MENINi on DNA replication fork stability and dynamics, DNA fibre assays showed that MLL fusion transformed cells heavily rely on repriming for overcoming replication stress. Strikingly, upon DOT1Li/MENINi treatment, MLL fusion transformed cells were forced to switch to fork reversal resulting in slower fork speed, which may in part account for the inhibitory effects. To further confirm this, we generated a novel MLL-AF9 leukemia model using mouse HPSCs, where Primpol can be inducibly inactivated upon 4-OHT treatment. As a result, we show that inactivation of Primpol phenocopies DOT1Li/MENINi and could drive MLLr-AML to adapt fork reversal as a coping mechanism. Together these results reveal a significant impact of these epigenetic inhibitors on DNA replication fork dynamics and repair mechanisms. Switching from fork repriming to reversal by the epigenetic drugs can lead to a reliance on PARP proteins that stabilize reversed forks during the arrest and protect them from premature restart, thereby creating a novel vulnerability to PARP inhibition. To test this hypothesis, PARPi was used in combination with DOT1Li/MENINi in MLLr-AML. Strikingly, while PARPi alone had no significant impact on fork dynamics, it triggered premature fork restart in DOT1Li treated cells, leading to increase fork speed and subsequent fork collapse, revealing a potential vulnerability created by DOT1Li. Consistently, combination treatment significantly suppressed cell growth of MLLr-AML, which displayed significantly shorter fork length and excessive DNA damage as compared to control and single treatments. Finally, we also tested the efficacy of these novel regimens in vivo. While single treatments with PARPi, DOT1Li, or MENINi exhibited no or moderate effects on disease progression, combo treatments in particular MENINi+PARPi significantly prolonged the disease latency with all animals remaining disease-free at the experimental endpoint. Taken together, these data reveal that DOT1L/MENIN inhibition alter replication fork plasticity and induce a novel vulnerability that can be therapeutically targeted by PARP inhibition for effective leukemia treatment.