Paralog interdependency between MEF2C and MEF2D sustains protein stability and leukemia maintenance in KMT2A-rearranged AML

Abstract Acute myeloid leukemia (AML) is characterized by lineage infidelity, transcriptional dysregulation, and frequent resistance to standard therapies. Among AML subtypes, those harboring KMT2A (MLL) rearrangements (KMT2A-r) remain particularly aggressive and refractory to current treatments. Transcription factors such as MEF2C play a pivotal role in maintaining leukemic identity and self-renewal, yet are considered “undruggable” due to their structure and nuclear localization. Here, we uncover a previously unrecognized leukemia-specific interdependency between MEF2C and its paralog MEF2D that sustains protein stability and leukemic maintenance through a novel co-stabilization mechanism. We demonstrate that MEF2C and MEF2D are the predominant MEF2 paralogs expressed in KMT2A-r AML, and both are required to maintain leukemic proliferation and block myeloid differentiation. Surprisingly, these transcription factors are not functionally redundant but instead form a heterodimer that stabilizes both proteins. CRISPR/Cas9-mediated knockout of MEF2D leads to rapid post-translational depletion of MEF2C protein across multiple KMT2A-r AML cell models, while MEF2C mRNA levels remain unchanged. Proteasome inhibitor assays showed this degradation is prevented by MG132 treatment, implicating the ubiquitin-proteasome system. To identify the E3 ligase responsible for MEF2C turnover, we performed a CRISPR-based UB-DUB screen using a MEF2C-mCherry/EGFP reporter system in MOLM13 cells. sgRNA enrichment analyses revealed that the CUL2-RBX1 complex and its substrate adaptors ZYG11B and ZER1 are required for MEF2C degradation. ZYG11B and ZER1 are known to recognize substrates bearing N-terminal glycine degrons. Sequence analysis and site-directed mutagenesis confirmed that MEF2C harbors a conserved N-terminal Gly degron that is essential for its recognition and ubiquitylation. Knockout of ZYG11B or ZER1 individually, or in combination, stabilized MEF2C protein, extended its half-life, and reversed the proliferative defect caused by MEF2D loss. We further demonstrated that MEF2D protects MEF2C from degradation by blocking CUL2-ZYG11B/ZER1-mediated ubiquitination. To functionally disrupt the MEF2C–MEF2D complex, we mapped their interaction interface and engineered a competitive peptide (Peptide-54) derived from the MEF2D N-terminus. Expression of Peptide-54 in AML cells selectively interfered with MEF2C–MEF2D binding, resulting in co-degradation of MEF2C, induction of myeloid differentiation, and suppression of leukemic proliferation. In a murine MLL-AF9 AML model, doxycycline-inducible expression of Peptide-54 significantly delayed disease progression, reduced leukemia burden as assessed by bioluminescence imaging, and extended overall survival. To determine whether the therapeutic effect of Peptide-54 depends on the CUL2-ZYG11B/ZER1 degradation axis, we performed genetic deletion of ZYG11B and ZER1 in Peptide-54–expressing AML cells. Loss of these adaptors rescued MEF2C protein levels and reversed the differentiation and proliferation phenotypes induced by Peptide-54, confirming that its effects are mediated through targeted proteolysis. In vivo, mice engrafted with MLL-AF9 AML cells co-expressing Peptide-54 and ZYG11B/ZER1-targeting sgRNAs failed to benefit from peptide induction, whereas mice with intact E3 ligases exhibited marked reductions in disease burden and improved survival. Importantly, Peptide-54 selectively impaired leukemic cells while sparing normal hematopoietic progenitor cell viability, underscoring its therapeutic potential and specificity. In conclusion, our study identifies a novel paralog interdependency between MEF2C and MEF2D that is uniquely required for protein stability and leukemic maintenance in KMT2A-r AML. We show that MEF2D acts as a molecular chaperone to prevent CUL2-ZYG11B/ZER1-mediated degradation of MEF2C. Disrupting this interaction with a rationally designed competitive peptide induces degradation of both paralogs, promotes differentiation, and suppresses AML progression in preclinical models. These findings highlight a potential paradigm of transcription factor co-stabilization and suggest that targeting paralog interactions may represent a therapeutically actionable vulnerability in high-risk leukemia.