MCL1 Promotes Fatty Acid β-Oxidation in Therapy-Resistant AML through Cytoplasmic Sequestration of the Transcriptional Repressor IRF2BP2

Despite extensive efforts aimed toward the development of improved molecular therapies targeting acute myeloid leukemia (AML), clinical outcomes remain poor. Notably, targeting BCL2 with venetoclax (Ven) in combination with azacitidine (Aza) has clinically delivered significant responses in newly diagnosed AML patients. Still, both upfront refractory and relapsed diseases are a major obstacle. Of particular importance, a unique subset of Ven/Aza-resistant AML are preferentially reliant on expression of the BCL2 family member MCL1 (5-fold increase in transcript, p<0.001, n=21; 9-fold increase in protein, p<0.01, n=8) and consequently vulnerable to MCL1 genetic or pharmacologic perturbation. Furthermore, our previous studies indicate that MCL1 expression in leukemia stem cells (LSC) is sufficient to metabolically up-regulate fatty acid β-oxidation (FAO), which in turn safeguards the LSC from the cytotoxic effects of Ven/Aza. However, the mechanism by which MCL1 influences FAO remains unknown. To interrogate the underlying metabolic dependency for MCL1 within primary Ven/Aza-resistant AML, we performed co-immunoprecipitation of MCL1 coupled with mass spectrometry analysis. Consistent with reported MCL1 biology, Gene Ontology pathway analyses of the enriched proteins highlight involvement of an array of cellular processes such as apoptosis, oxidative phosphorylation and very long chain fatty acid metabolism (GO: Biological Process, p<0.001). Interestingly, the mass spectrometry analysis revealed Interferon Regulatory Factor 2 Binding Protein 2 (IRF2BP2), a transcriptional repressor recently reported to be highly expressed and functionally relevant in AML blasts, as the top enriched feature (average abundance score of 654, n=3). Whole cell proteomics analysis of Ven/Aza-sensitive and -resistant LSC revealed reduced phosphorylation of IRF2BP2 at Serine 360 (p=0.06, n=8), indicating a potential reduction in the nuclear accumulation of IRF2BP2. Using confocal microscopy, we discovered that primary Ven/Aza-sensitive LSC display robust nuclear IRF2BP2 localization while Ven/Aza-resistant LSC exhibit predominantly cytoplasmic IRF2BP2 expression (75 vs 13 % nuclear localization, p<0.001, n=10). Motif analysis suggests that IRF2BP2 contains several BH3-like alpha helical domains that mediate the interaction with MCL1. Indeed, pharmacologic perturbation of the IRF2BP2-MCL1 interaction, using BH3 mimetic S63845 (an MCL1 inhibitor), induces a rapid re-localization of IRF2BP2 into the nucleus within Ven/Aza-resistant LSC (13 vs 82 % nuclear localization, p<0.001, n=6).

Nuclear transcriptional repressive activity of IRF2BP2 has been reported to largely influence inflammatory and differentiation gene signatures. Overlap between differentially upregulated genes in Ven/Aza-resistant AML with putative IRF2BP2 transcriptional targets (Ellegast et al., 2022) identified 1090 significantly upregulated genes likely to result from the loss of nuclear IRF2BP2 repressive activity. Of particular interest, and functionally relevant to resistant AML metabolism, was acyl-CoA synthetase long chain family member 1 (ACSL1), an essential rate limiting enzyme for FAO (11-fold increase, p<0.001, n=21). We validated the increased ACSL1 transcript expression within Ven/Aza-resistant LSC using quantitative PCR analysis (40-fold increase, p<0.01, n=8). Furthermore, forced IRF2BP2 nuclear localization among Ven/Aza-resistant LSC, using BH3 mimetic S63845, resulted in the transcriptional repression of ACSL1 (2-fold, p<0.01, n=3). Whole cell proteomics and western blotting analyses confirmed elevated ACSL1 protein levels in Ven/Aza-resistant LSC (6-fold, p<0.01, n=8; 5-fold, p<0.01, n=8) consistent with increased FAO. Functionally, pharmacologic inhibition of ACSL1 enzymatic activity impairs Ven/Aza-resistant AML viability (50% decrease, p<0.01, n=3) and colony forming potential (84% decrease, p<0.001, n=3). Collectively, these data provide evidence for a novel mechanism by which MCL1 non-canonically drives IRF2BP2 cytoplasmic sequestration and subsequent activation of ACSL1, promoting fatty acid β-oxidation metabolism. Thus, the unique MCL1-driven loss of IRF2BP2 transcriptional repressive activity represents a critical component defining Ven/Aza-resistant AML which may offer alternative strategies for therapeutic intervention.