Experiment: map what stays on chromatin when the menin inhibitor has done everything it can do
Flood the cells with menin inhibitor until no more can bind, then ask where the fusion protein is still sitting on the DNA. Anything still there is a place the drug can never reach.
The same thing, for a scientist
A saturation-occupancy CUT&RUN experiment. The novelty is not the assay but the question: existing studies profile chromatin after menin inhibition to show what is lost, and none report the reproducible residual set at genuinely saturating concentration with the tether dependencies dissected. The design adds three things: a concentration ladder to demonstrate saturation rather than assume it, genetic separation of the three menin-independent tethers, and intersection of the residual set with the non-canonical bivalent menin-target class implicated in MYC bypass.
Design
Saturation ladder. Revumenib across at least six concentrations spanning IC50 to 20x IC90, with parallel target-engagement confirmation (menin displacement by CUT&RUN). Saturation is demonstrated as a plateau in menin loss, not assumed from a single high dose. Repeat with ziftomenib to separate compound from target.
Occupancy panel at plateau. CUT&RUN for the KMT2A N-terminal epitope (present on both wild-type KMT2A and the fusion, so a fusion-specific readout requires a tagged allele or an AFF1 C-terminal antibody run in parallel), menin, LEDGF/PSIP1, AFF1, and RNA Pol II Ser2P.
Tether dissection. In an isogenic KMT2A::AFF1 background: CXXC point mutants that abolish unmethylated-CpG binding; LEDGF/PSIP1 knockout and IBD-contact point mutants; AFF1 C-terminal truncation removing SEC contacts. Each should shrink the residual set if that tether holds it.
Intersection. Overlay the residual set with H2AK119ub and menin co-occupancy to test enrichment for the non-canonical class 2, and with MYC regulatory elements.
Clinical arm. Where banked serial marrows from menin-treated patients exist, low-input CUT&RUN on-drug versus baseline, to ask whether the residual set is present in patients and whether its size tracks depth of response.
Controls, and the specific way this experiment lies to you
The dominant artefact in any residual-signal experiment is background masquerading as biology. Mandatory controls: IgG at matched depth; a spike-in normalisation (the whole point is a quantitative comparison of total signal between conditions, which uncalibrated CUT&RUN cannot do); a KMT2A-germline line as a negative control for the N-terminal antibody; and at minimum three biological replicates per condition, with the residual set defined by reproducibility across replicates rather than by peak calling in one.
A residual set that is not reproducible across two independent cell models should be reported as absent. Pre-specify that rule.
What each outcome means
- Reproducible residual set, enriched for non-canonical bivalent loci. The strongest result. It names the loci a second agent must cover and supports
hyp-apparent-authority-residual-fusion-occupancy. - Reproducible residual set, no enrichment. Real but possibly irrelevant; follow with a functional test before investing further.
- No residual set. Kills this hypothesis and strengthens
hyp-cell-state-plasticity-limits-interface-disruption, which is a genuinely useful outcome because the two make opposite predictions about the same clinical failure. - Residual set that disappears under KAT6/KAT7 inhibition but not under a second menin inhibitor. Would give a mechanistic rationale for a specific combination that is currently justified only empirically 3.
| design | Concentration-ladder CUT&RUN to demonstrated saturation of menin displacement, followed by multi-factor occupancy profiling at plateau, genetic dissection of the three menin-independent tethers in an isogenic KMT2A::AFF1 background, and intersection of the residual set with the non-canonical bivalent menin-target class. Optional clinical arm on banked serial marrows. |
|---|---|
| readout | Reproducible residual peak set for the fusion at saturating drug; its overlap with H2AK119ub/menin bivalent loci and MYC regulatory elements; the change in residual-set size under CXXC, LEDGF-IBD and AFF1-SEC perturbations; JUND occupancy gain as a positive control for redistribution. |
| controls | IgG at matched depth; spike-in normalisation (mandatory, because the comparison is quantitative between conditions); KMT2A-germline negative-control line; three or more biological replicates with the residual set defined by cross-replicate and cross-model reproducibility; second structurally distinct menin inhibitor. |
| cost tier | med |
| timeline | 6-9 months for the cell-line and isogenic arms; the clinical arm depends entirely on sample access and could run in parallel. |
| discriminates | Separates hyp-apparent-authority-residual-fusion-occupancy from hyp-cell-state-plasticity-limits-interface-disruption, which predict opposite chromatin states behind the same clinical failure. |
| prespecified negative rule | A residual set that does not reproduce across two independent KMT2A::AFF1 models is to be reported as absent. This rule must be recorded before the first library is sequenced. |
tests hypotheses
[ "hyp-apparent-authority-residual-fusion-occupancy" ]
collaborators needed
[ "a chromatin group with quantitative spike-in CUT&RUN", "access to banked on-treatment marrows from menin-inhibitor trials" ]
enl arm added 2026 09 19
{
"change": "Add an ENL CUT&RUN arm, and menin, alongside the fusion antibody. Profiling the fusion alone cannot distinguish 'the site is vacated' from 'the site is reoccupied', and the only human sample under revumenib shows exactly the second: fusion and menin undetectable, ENL retained.",
"reason_pmid": "39472576",
"also": "Include a lineage-switch marker panel. The one human observation is confounded by switching, and this experiment should be able to tell the two apart rather than inherit the confound."
}- If saturation in culture requires concentrations far above clinically achievable free drug, then the residual set in patients is larger than the one measured here, and the experiment understates the problem rather than overstating it.
- Is there an antibody or tagging strategy that distinguishes fusion from wild-type KMT2A on chromatin in primary patient cells, without which the clinical arm cannot be interpreted cleanly?
- What concentration in cell culture corresponds to the clinically achievable free-drug exposure, so that 'saturating' does not mean 'unachievable in a patient'?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Menin inhibitors act by evicting menin and MLL1 from chromatin, and resistance mutations work by preventing that eviction, which makes chromatin occupancy the mechanistically correct readout for asking how complete the drug effect is.
MEN1 mutations mediate clinical resistance to menin inhibition · Perner F, Stein EM, Wenge DV, et al. · Nature 2023What this evidence cannot show
Eviction was shown as a qualitative contrast between sensitive and resistant states rather than quantified to saturation.
Evidence that argues against this (2)
- The design plans to score residual sites against candidate menin-independent tethers including LEDGF/PSIP1. That feature should be dropped or reinterpreted: menin is the adaptor that physically links MLL proteins to LEDGF, LEDGF is required for MLL-dependent transcription and transformation, and the LEDGF-engaging surface is formed jointly by menin and MLL1. Once menin is evicted there is no LEDGF tether to score. Scoring residual sites for LEDGF/PWWP features would therefore produce an uninterpretable covariate at best and a spurious positive at worst. Yokoyama A et al. 2008
- Two competing mechanisms for residual fusion retention already have experimental support and should be arms in this experiment rather than post-hoc explanations. First, combined DOT1L and menin inhibition resembles full fusion degradation whereas single-agent treatment has 'more modest effects on MLL::AF9 occupancy and gene expression', with combined inhibition releasing the oncoprotein from chromatin *globally* — implicating H3K79-methylation-dependent retention. Second, KAT6/KAT7 catalytic inhibition rapidly evicts the MLL fusion from chromatin and overcomes both primary and acquired menin-inhibitor resistance, implicating acetylation-dependent retention. Either would mean the residual set is defined by a chromatin mark rather than by DNA-motif or partner-contact features, which changes both the covariates to score and the therapeutic conclusion. Olsen SN et al. 2022
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
A class of non-canonical menin target loci is bivalently co-occupied by active menin and repressive H2AK119ub, is normally downregulated after menin inhibition, and is reactivated (including MYC) when PRC1.1 subunits PCGF1 or BCOR are lost, causing resistance.
Epigenetic regulation of noncanonical menin targets modulates menin inhibitor response in acute myeloid leukemia · Zhou X, Zhang L, Aryal S, et al. · Blood 2024What this evidence cannot show
Defined in AML models; the equivalent map in KMT2A::AFF1 infant B-ALL does not exist and may differ.
Evidence that argues against this (1)
- The non-canonical bivalent loci are derepressed by *loss of a repressor* (PRC1.1 subunits PCGF1 or BCOR), not by retained fusion occupancy. Reading enrichment of the residual set for these loci as evidence of residual fusion authority would conflate two different causes of the same transcriptional output. The experiment needs H2AK119ub and PRC1.1-subunit status measured alongside occupancy in the same cells, or the enrichment prediction is not diagnostic. Cancer Discovery 2025
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Menin-MLL inhibitors disrupt menin-JUND interactions and increase JUND chromatin occupancy, showing that the drug redistributes chromatin binding rather than only subtracting it, and that a chromatin-occupancy readout under drug can reveal consequential gain-of-occupancy events.
Menin-MLL inhibitors enhance JUND activity in MLLr leukemic cells contributing to tumorigenesis and therapy resistance · Ozyerli-Goknar E, Perner F, Armstrong SA, Zeiser R, Timmers HTM, et al. · Blood 2026What this evidence cannot show
Recent and not independently replicated; JUND occupancy should be added to the panel in this experiment for exactly that reason.
Evidence that argues against this (1)
- Redistribution rather than subtraction is the right reading and it complicates the proposed readout: if occupancy moves rather than falling, then a residual-occupancy measurement under drug will find signal whose interpretation depends on which factor is being mapped. The experiment needs JUND and KMT2A occupancy measured together. Nature 2012
drop or reinterpret the LEDGF/PSIP1 scoring feature: menin is the adaptor linking MLL to LEDGF, so no LEDGF tether survives menin eviction
competing mechanism to include as an arm: combined DOT1L and menin inhibition releases MLL::AF9 from chromatin globally whereas single agents have modest effects on occupancy
saturation-ladder CUT&RUN with genetic dissection of the three tethers
- tests →Hypothesis: the ceiling on menin-interface disruption is cell-state plasticity, not target mutation
the same experiment discriminates between the two competing accounts of menin-inhibitor failure
a concrete structural candidate for residual occupancy under saturating inhibitor
- tests →ENL stays after the fusion leaves — an epigenetic lesion that outlives the protein that made it
adding an ENL arm to the saturating-occupancy CUT&RUN would convert an n=1 observation into a model