Investigators reported that a distinct promoter sequence arrangement marks high occupancy by the Menin-KMT2A histone-modifying complex. The authors applied interpretable machine learning to genome-scale data and identified a composite architecture: canonical CpG islands bordered by several-kilobase A/T-rich domains (termed AT-flanks). Short statement: CpG islands alone were insufficient to explain selective occupancy. The study team observed that presence and extent of AT-flanks correlated with strong Menin-KMT2A binding at promoters of developmental genes and genes implicated in leukemia. The abstract does not provide effect sizes or the specific classifiers used; those methodological details were not reported in the source abstract.
The authors carried out genetic deletion and chromatin-contact analyses to test the contribution of AT-flanks to complex retention. Data indicated that removal of AT-flanks reduced KMT2A retention at linked CpG islands, which suggests the flanking sequences have an active role in sustaining promoter occupancy rather than being passive sequence context. The study team identified LEDGF as an A/T-rich DNA-associated protein required for Menin engagement at AT-flanks and for sustained KMT2A occupancy at promoters. That is, LEDGF was shown to associate with A/T-rich regions and to be necessary, per the authors, for the Menin complex to engage those flanks effectively. Chromatin-contact analyses reported by the group supported physical linkage between AT-flanks and CpG islands at promoters that maintain Menin-KMT2A binding, implying a spatial mechanism that brings these sequence elements into functional proximity. The abstract does not specify experimental parameters—such as deletion strategy, contact-capture method, or quantitative binding changes—and those particulars were not provided in the source abstract.
Investigators examined the relationship between promoter architecture and sensitivity to pharmacologic disruption of the Menin-KMT2A interaction. Data in the abstract indicate that AT-flank extent associated with differential loss of KMT2A after Menin-KMT2A disruption. In other words, promoters with larger or more pronounced AT-flanks retained or lost KMT2A differently when the Menin interaction was pharmacologically perturbed. The authors point out that this observation provides a genomic basis for variable target sensitivity to recently approved Menin inhibitors in acute leukemia. The abstract notes the clinical relevance — that heterogeneous promoter architectures could underlie why some leukemia-associated targets respond differently to Menin inhibition — but it does not report specific inhibitor names, dosing, cellular models, or quantitative sensitivity metrics; those details were not included in the source abstract.
Beyond the Menin-KMT2A axis, the same composite promoter architecture distinguished promoters with high canonical PRC1 occupancy across diverse cell types. The study team observed that promoters bearing CpG islands plus AT-flanks were not unique to Menin-KMT2A targets; rather, the architecture marked a subset of promoters that recruit different chromatin-regulatory complexes in different contexts. This suggests a single sequence organization can be deployed to achieve distinct chromatin states depending on cellular factors and trans-acting proteins. The abstract does not elaborate on which cell types were profiled or how broadly conserved this association is; such specifics were not provided in the source abstract.
The abstract reports multiple convergent approaches—machine learning, genetic deletion, chromatin-contact analyses and protein identification—but omits many experimental specifics. The source did not report numerical effect sizes, model parameters, sample sizes, or exact methods for chromatin contact mapping and genetic deletion within the abstract. Investigators linked LEDGF to AT-flank engagement, yet the abstract does not detail biochemical interaction assays, binding affinities, or whether LEDGF is sufficient, in isolation, to recruit Menin-KMT2A. The authors also correlate AT-flank extent with variable KMT2A loss after pharmacologic disruption, but the abstract does not disclose which Menin inhibitors were evaluated or the experimental systems used.
Open questions apparent from the abstract include: which molecular determinants beyond A/T content define AT-flank function; whether LEDGF directly bridges Menin-KMT2A to DNA or acts via additional cofactors; how promoter architecture quantitatively predicts inhibitor response in patient-derived samples; and whether this promoter signature can be used prospectively to stratify leukemia patients for Menin-directed therapy. The source abstract states that the promoter architecture provides a genomic basis for heterogeneous sensitivity to Menin inhibitors, but it does not document follow-up clinical validation or regulatory milestones. Those remain unanswered in the provided text and require data beyond what the abstract supplies.