Menin does NOT leave chromatin when the fusion is removed: three hours after MLL::AF9 is degraded to 2.5-7.1% of its own chromatin sites, menin still holds 80.0% of its strongest peak calls genome-wide - so menin occupancy is not a readout of fusion occupancy, and a menin inhibitor that displaces menin is doing something fusion loss does not do
also called menin stays when the fusion goes · 80.0% against 2.5% · M2 fires · menin occupancy is not a fusion readout
A drug in clinical use works by prising a protein called menin off the cancer protein it partners with, so we asked the control question nobody had asked: if you simply destroy the cancer protein, does menin fall off by itself? Genome-wide it does not - menin still sits at four-fifths of its strongest spots three hours after the cancer protein has gone from its own, so something other than the fusion is holding it there.
The same thing, for a scientist
**Pre-registered branch M2 fires.** In HCB1 cells three hours after dTAG-VHL 500 nM degrades MLL-AF9-HA-FKBP12: | channel | top-decile summit retention | loss relative to HA | |---|---:|---:| | **HA — the fusion itself** | **2.5% / 7.1%** (two batches) | 1.00 | | **MEN1 — menin** | **80.0%** | **0.21** | | DOT1L | 68.7% | 0.33 | | AFF4 | 92.7% | 0.08 | | CDK9 | 83.6% | 0.17 | | RNA Pol II total | 99.6% | 0.00 | Menin retains **80.0%** of its strongest DMSO peak calls (86.7% under the bin measure) in the same cells, in the same three hours, in which the fusion's own tag falls to **2.5-7.1%**. Scaled to the fusion's own loss, menin loses **0.21** of what the fusion loses. **What this licenses, and it is narrow.** Menin's **chromatin residence genome-wide is not a readout of fusion occupancy on a three-hour timescale.** Any entry in this graph treating a menin ChIP as a proxy for where the fusion is, or for how much fusion has been displaced, is over-reading — the two dissociate completely in this experiment. **What it does NOT license, and the authors' own data set the limit.** Olsen *et al.* report, at their defined **MLL::AF9 target-gene set** and by **normalised read-count fold change**, that *"more than 50% of MENIN and almost 70% of DOT1L was lost at MLL::AF9 target genes, while there were no changes in occupancy at non-MLL::AF9-bound genes"* [E1]. **That is not in conflict with 80% peak-call retention.** A >50% signal drop rarely removes a confident peak call (`U2`: a call is a threshold), and the authors state explicitly that the **non-target majority of the genome does not move** — which is precisely what a genome-wide peak-presence measure is dominated by. **So the honest statement is two-part.** *At MLL::AF9 target genes, menin signal falls by more than half within 3 h (the authors' measurement). Genome-wide, menin's peak calls do not move (this measurement). These are different questions and both answers are true.*
The specificity control no menin-inhibitor experiment can supply
A menin inhibitor could displace menin from chromatin by blocking its MBM pocket — breaking the menin–KMT2A contact while leaving the rest of the complex in place — or by dismantling the complex so that everything leaves. No menin-inhibitor experiment distinguishes these, because in both cases menin comes off.
Degrading the fusion separates them, and the answer is unexpected: removing the fusion does not remove menin. So whatever a menin inhibitor does to menin's genome-wide residence, it is not the same event as the fusion ceasing to exist. The drug removes menin; fusion loss largely does not. That is a genuine asymmetry and it constrains how menin displacement can be used as a pharmacodynamic marker.
Why menin stays — the authors' own proposal, and it points at a node this graph already holds
Olsen et al. offer a mechanism in their discussion: "In contrast to DOT1L, MENIN is thought to be tethered to chromatin via LEDGF (Yokoyama and Cleary, 2008) independently of MLL::AF9. Additional and/or compensatory anchorage by LEDGF could explain why MENIN dissociates from chromatin more slowly upon MLL::AF9 degradation" 12.
That is a testable mechanism, not a hand-wave, and this graph already carries mechanism-menin-ledgf-chromatin-tethering and gene-psip1. It is developed as hyp-q139-ledgf-tethering-is-what-holds-menin-on-chromatin-after-the-fusion-is-gone.
The authors' own concluding sentence is the careful version and should be quoted rather than paraphrased: "our studies indicate that MLL::AF9 contributes to retaining MENIN on chromatin at certain loci, suggesting that there are multiple interactions that maintain MENIN occupancy, a point that is relevant as MENIN-directed drugs continue to be developed" 1. At certain loci is doing real work in that sentence, and it is the reconciliation with the genome-wide null.
The pre-registered confound check, and which way it runs
Checked before the retention figure was read, as pre-registered. The MEN1 dVHL library is 1.6% shallower than its DMSO (16,644,579 vs 16,920,529 fragments) and calls fewer peaks (4,074 vs 4,641), which mechanically triggers the pre-registered X-LIVE suppression rule.
The rule fires and the direction exonerates the finding. Suppression exists to stop a shallow library's false peak loss being read as biology — it protects against a false loss. The finding here is retention, and a shallower library can only push retention down. So 80.0% is a lower bound, and a 1.6% depth difference cannot in any case account for a 12% peak-count difference. Both the mechanical verdict and its direction are recorded in results.json; neither is silently applied.
| script | analyses/what-leaves-when-the-fusion-is-degraded/run.py |
|---|---|
| status | run |
| result summary | MEN1 top-decile summit retention 80.0% (bin measure 86.7%) 3 h after fusion degradation, against 2.5-7.1% for the fusion's own HA tag in the same cells. Loss relative to HA = 0.21. Pre-registered branch M2 fires. Depth confound runs against the finding, making 80.0% a lower bound. |
| preregistration | analyses/what-leaves-when-the-fusion-is-degraded/PREREGISTRATION.md |
inputs
[ "GSE173599 GSM5271682 / GSM5271684" ]
- RESISTANCE AND PHARMACODYNAMICS: if menin's chromatin residence does not track the fusion, then a falling menin ChIP signal under a menin inhibitor is reporting on the drug's pocket occupancy, not on whether the oncoprotein has been disabled. A patient could show full menin displacement with the fusion still resident, and this deposit's ladder shows exactly that - VTP-50469 alone leaves 51.7-76.9% of the fusion's own sites occupied.
- COMPENSATION: multiple anchorages maintaining menin occupancy - which is the authors' own reading - is the structural precondition for resistance that does not require a MEN1 mutation. If LEDGF or another tether holds menin at target loci, a pocket inhibitor may be relieved by redundancy rather than defeated by a binding-site change.
- ESSENTIAL FUNCTION: menin retained genome-wide on 80% of its sites after the oncoprotein is gone is a reminder that most menin on chromatin is doing normal-cell business. A therapeutic that strips menin everywhere is not fusion-selective, whatever the ChIP at target genes shows.
- Is menin's persistence LEDGF-dependent? The decisive experiment is a MEN1 ChIP 3 h after fusion degradation in PSIP1-depleted versus PSIP1-intact HCB1, and it has never been done in any system.
- Does menin persist at the SAME sites it occupied before, or does it redistribute while the total peak count holds? A peak-identity comparison distinguishes retention from turnover and was not attempted here.
- There is no menin ChIP beyond 3 h in this deposit or anywhere else under a degron. Does menin eventually leave at 24 h or 4 d, which would make this a kinetics result rather than an independence result?
- Is the genome-wide null an artefact of menin's peak set being dominated by non-fusion-bound sites? Partitioning menin peaks by fusion co-occupancy would settle it and requires only the deposited HA peak table.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
At MLL::AF9 target genes, more than 50% of MENIN and almost 70% of DOT1L signal is lost 3 h after dTAG-VHL treatment, with no change at non-MLL::AF9-bound genes; MENIN loss is 'more sluggish' than DOT1L's, with robust ChIP-qPCR changes only after 3 hours, and the authors attribute menin's persistence to LEDGF tethering.
MLL::AF9 degradation induces rapid changes in transcriptional elongation and subsequent loss of an active chromatin landscape. · Olsen SN, Godfrey L, Healy JP, Choi YA, Kai Y, Hatton C, Perner F, Haarer EL, Nabet B, Yuan GC, Armstrong SA · Molecular Cell 2022What this evidence cannot show
Read in full text at PMC9044330, not from the abstract. The authors' figures are (a) ChIP-qPCR at a handful of selected loci, in experimental triplicate, and (b) normalised read-count fold change restricted to a pre-defined MLL::AF9 target-gene set at TSS -1kb/+3kb or gene bodies. Neither is a genome-wide peak-presence measure, so their numbers and the 80.0% here are not the same quantity and must not be compared as though they were.
Evidence that argues against this (1)
- This is the strongest counter-reading of the genome-wide null and it comes from the same paper: restricted to the loci where the fusion actually sits, menin DOES lose more than half its signal within three hours. A reader who takes 'menin does not leave' as a locus-level claim will be wrong. Olsen SN et al. 2022
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Menin physically links MLL proteins to LEDGF (PSIP1) on chromatin, and LEDGF is required for MLL-dependent transcription and leukemic transformation - the anchorage the authors invoke to explain why menin outlasts the fusion.
Menin critically links MLL proteins with LEDGF on cancer-associated target genes. · Yokoyama A, Cleary ML · Cancer Cell 2008What this evidence cannot show
Read at abstract level plus the citing sentence in Olsen et al.; the primary paper's experiments are in HeLa, U937 and mouse myeloid progenitors, not in HCB1 or any dTAG system, and it long predates degron technology. It establishes that the menin-LEDGF-MLL link exists, not that LEDGF is what holds menin in place after fusion degradation. That step is a hypothesis, not an evidenced claim.
Evidence that argues against this (1)
- This graph already carries an entry titled mechanism-ledgf-dispensable-for-fusion-chromatin-retention, which if correct weakens the LEDGF-anchorage reading. The two should be adjudicated rather than both left standing. Oncofusion knowledge graph 2026
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
Paired MEN1 ChIP peak tables under DMSO and dTAG-VHL 500 nM at 3 h are deposited as GSM5271682 and GSM5271684 and are the only menin ChIP in the deposit at any timepoint.
GSE173599, samples GSM5271682 (HCB1_ChIP_MEN1_DMSO_3h) and GSM5271684 (HCB1_ChIP_MEN1_dVHL_500nM_3h) · Olsen SN, Armstrong SA, et al. · NCBI Gene Expression Omnibus 2022What this evidence cannot show
n = 1 per arm, no replicate, no spike-in, no noise floor measurable for this pair and none imported. Both libraries are from sequencing batch 1912x2 and are not hg38-mapped, so the pair is internally comparable but not comparable across marks. MACS2 was run with --slocal 1500 and estimated a 710 bp fragment size for the DMSO arm, so calls are broad.
Evidence that argues against this (1)
- There is no second menin degron ChIP anywhere in the public record to replicate this against, so a single-library artefact cannot be excluded by any means available. NCBI Gene Expression Omnibus 2022
MENIN RESIDENCE DOES NOT REQUIRE THE FUSION AT 3 h, GENOME-WIDE: 80.0% of menin's strongest peak calls survive fusion degradation, in cells where the fusion's own tag is at 2.5-7.1%.
The menin result is only interpretable because the positive control passed in the same cells and the same hours: 80.0% menin retention means nothing unless the fusion demonstrably left.
The authors' own proposed explanation for menin outlasting the fusion is LEDGF anchorage, citing Yokoyama & Cleary 2008 (PMID 18598942). This is the first degron-based observation consistent with that tether being fusion-independent.
Sharpens that entry from the other direction: not only does 'eviction' often name menin rather than the fusion, but menin's own residence is NOT a proxy for the fusion's - the two dissociate completely when the fusion is removed.
PSIP1/LEDGF is the compensatory anchorage Olsen et al. invoke to explain why menin dissociates more slowly than DOT1L after fusion degradation. No PSIP1 experiment exists in the deposit; the link is the authors' hypothesis.
THE CLAIM IS ONLY DEFENSIBLE WITH ITS DENOMINATOR: 80% genome-wide retention and the authors' >50% loss at MLL::AF9 target genes are both true, and separating them is what makes the menin result a finding rather than a contradiction of the paper.
Supplies the mechanism for menin's persistence: a redundant tether, against DOT1L's single direct one. The ordering is unanimous across three instruments while every magnitude disagrees.
The observation the hypothesis exists to explain: menin at 80.0% retention while the fusion is at 2.5-7.1% in the same cells and the same three hours.
BOUNDS ITS OWN SIBLING'S CLAIM: because no menin ChIP exists past 3 h, 'menin is fusion-independent' cannot be separated from 'menin has a slow off-rate'. The claim must be read as 'at three hours' and the authors' own word for menin loss is 'sluggish', which is kinetic.
reconciles that entry's genome-wide 80% peak retention with the authors' >50% target-gene loss; both are right on the same files
- ← related toPeak-call retention tracks peak strength, not fusion occupancy, and that inverts the answer
the 80% figure reproduces (83.4% here) and is a saturated measure, so it does not bound the menin loss