Oncofusion

start / graph / mechanism-q143-in-a-patient-the-menin-fusion-association-is-demonstrable-before-the-drug-and-unmeasurable-after-it

Mechanismmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q143-menin-under-drug-in-infants

In a patient, the menin-fusion association is demonstrable before the drug and unmeasurable after it - and the asymmetry is what this pair establishes, not a rate of menin loss

In a baby's leukaemia cells taken before treatment, menin is sitting on the fusion protein's genes and not on the normal gene's. In cells from the same child taken during treatment, nothing is sitting anywhere that can be measured. The useful finding is the first half; the second half is a limit of what can be seen, not a measurement of how much menin left.

The same thing, for a scientist

This patient pair supports one positive mechanistic statement and one negative methodological one, and they should not be run together. **Positive:** in drug-naive primary infant KMT2A::MLLT1 B-ALL, menin's promoter distribution tracks the KMT2A **N-terminal** channel (accessibility-corrected enrichment 2.73x against the fusion's own 3.12x at the top 50 sites) and not the **C-terminal** channel, which reports wild-type KMT2A and sits at background (1.10x). The N/C difference is a fusion-specific readout requiring no tag, and this is its first use in this corpus. **Negative:** in the patient-matched relapse aliquot drawn on revumenib, no channel - menin, fusion N-terminus, wild-type C-terminus, ENL or DOT1L - exceeds a non-specific reference, the sample's own top-ranked promoters are mitochondrial pseudogenes, and the three CUT&Tag channels fail their own replicate check. **The drug arm therefore bounds nothing quantitatively.** The one thing it does establish is that the collapse is **not uniform across channels**: ENL alone redistributes toward the myeloid-program peaks, and that ordering reproduces in the authors' own numbers.

Go deeper
What is established, at what strength
statementstrengthwhy
menin co-localises with the fusion N-terminus in drug-naive primary infant KMT2A::MLLT1moderaten=1 patient, cross-assay, out-of-sample accessibility reference, but with a 22.6-fold gene-set positive control and two reagents per end
the wild-type-reporting C-terminal channel does not rank the oncoprotein targetsstrong1 of 50 vs 40 of 50, p=0.84 vs 2e-49, and the two C reagents agree at 0.935
under revumenib nothing in the sample is above a non-specific referencemoderatethree readings survive (target gone / below resolution / poor library) and this deposit separates none
menin does not outlast the fusion under revumenibweaka ratio of two near-background numbers in a different clone; the falsifier's threshold was not met but neither number is trustworthy alone
ENL alone redistributes toward the GMP-like peaksmoderatereproduced per-replicate and in a second instrument, on n=52 and n=45 peaks
What the pair says about the loose-coupling reading, in both directions

Damaging. The pre-registered falsifier did not fire. There is no menin surplus over the fusion in the one piece of primary infant material that exists, so the 50.3% fusion-independent menin plateau seen after a KMT2A::MLLT3 degron does not reproduce in a patient under a menin inhibitor. If loose coupling is stated as a general property of KMT2A fusions, this is the first evidence against it.

Not damaging, and possibly supporting. A menin inhibitor removes menin through menin's own pocket, so every tether model predicts menin leaving - the experiment has no discriminating power over the tether question by construction. The drug arm is also a different clone, at relapse, with intrinsically lower oncoprotein, and is at background in every channel. Meanwhile the drug-free half of the same patient gives the cleanest primary-material evidence this corpus holds that the menin that matters is fusion-associated: menin with the N channel, not with the C channel. That is the load-bearing half of the loose-coupling picture - the interface is real - and it now has a human infant observation behind it.

The honest summary sentence

In primary infant KMT2A::MLLT1 leukaemia, menin sits where the fusion sits and not where wild-type KMT2A sits; and in the same child on revumenib, neither menin nor the fusion sits anywhere this assay can see. Everything else in this beat is a qualification of one of those two clauses.

Facts and figures
systemone patient, Texas Children's Hospital: 148752 diagnostic infant B-ALL before revumenib; 152985 relapse infant AML on revumenib; KMT2A::MLLT1
positive statementmenin tracks the KMT2A N-terminal (fusion-containing) channel and not the C-terminal (wild-type) channel in drug-naive primary infant material
negative statementthe revumenib arm is at background in every channel and bounds nothing quantitatively
shown for fusionKMT2A::MLLT1 in one primary infant patient; nothing in KMT2A::AFF1
epigenetic readoutprotein occupancy by CUT&RUN and CUT&Tag; no histone mark was measured in this pair
statusrun
loose coupling
{
  "damaging": "the pre-registered falsifier did not fire; no menin surplus over the fusion",
  "not_damaging": "a menin inhibitor cannot discriminate tether models; the drug arm is a different clone at background",
  "supporting": "the drug-free arm gives the first primary-infant evidence that the menin that matters is fusion-associated"
}
How this could bite us back
  • The positive half of this entry will be used to argue that the menin-KMT2A interface is the right target in infants. It is one patient, in the wrong fusion, with a borrowed accessibility reference.
  • The negative half will be used to argue that menin inhibition clears menin completely in vivo. It cannot - the assay returned background in every channel including ones expected to be positive, and a negative from a failed assay is indistinguishable from a negative from a working one.
  • The patient this all comes from RELAPSED on the drug, with the fusion transcript still expressed and ENL still on the myeloid program. Whatever the occupancy numbers say, the clinical outcome says menin inhibition alone was not enough here.
Still open
  1. Does the N-versus-C separation reproduce in the untreated KMT2A::AFF1 samples of this same deposit? That is the index-fusion version and it is computable from files already public.
  2. Is there any material in which menin can be watched while the fusion is removed WITHOUT touching menin, in a patient? No such perturbation exists clinically.
  3. How much of the menin at these promoters is LEDGF-delivered? No LEDGF channel exists in this deposit or, as far as this beat found, in any primary infant sample.
Every source behind this entry (3)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    In drug-naive primary infant KMT2A::MLLT1 B-ALL, menin's promoter enrichment tracks the KMT2A N-terminal channel and not the C-terminal one.

    analyses/menin-under-revumenib-in-an-infant-patient-pair - pre-registered re-analysis of GSE252378 samples 148752 and 152985 · Q143-menin-under-drug-in-infants (Oncofusion Knowledge Graph) · this repository 2026
    What this evidence cannot show

    n = 1 patient, one aliquot; menin is CUT&Tag and KMT2A is CUT&RUN three weeks apart; the accessibility reference is an IgG from a different patient with a different fusion; no spike-in; the patient is a lineage-switching case, a selected population; and the fusion is KMT2A::MLLT1, not the index fusion.

    Evidence that argues against this (1)
    • Menin also reaches MLL target genes through LEDGF/PSIP1, so co-localisation with the N-terminal channel does not establish that the fusion is what placed it there. Yokoyama A et al. 2008
  2. E2weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.dataset

    In the patient-matched relapse aliquot on revumenib, no channel exceeds a non-specific reference and the pre-registered menin-over-fusion threshold of 1.5 is not met (1.21, or 1.45 after genome-wide correction).

    analyses/menin-under-revumenib-in-an-infant-patient-pair - pre-registered re-analysis of GSE252378 samples 148752 and 152985 · Q143-menin-under-drug-in-infants (Oncofusion Knowledge Graph) · this repository 2026
    What this evidence cannot show

    A ratio of two near-background quantities. The drug arm's CUT&Tag replicates agree at 0.11-0.25; drug exposure and clonal replacement are completely confounded; the same intervals in the authors' own table give 0.871 rather than 0.586 for menin.

    Evidence that argues against this (1)
    • The authors' own deposited quantification over the same intervals gives menin 0.871 - a mild reduction rather than a collapse - so the magnitude of the drug-arm effect depends on the denominator chosen and not on the data. Janssens DH 2024
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The paper's own finding for this patient is that the oncoprotein and menin become undetectable while ENL persists on oncoprotein-target loci.

    MLL oncoprotein levels influence leukemia lineage identities · Janssens DH, Duran M, Otto DJ, Wu W, Xu Y, Kirkey D, Mullighan CG, Yi JS, Meshinchi S, Sarthy JF, Ahmad K, Henikoff S · Nature Communications 15:9341 2024
    PMID 39472576DOI
    What this evidence cannot show

    Browser tracks, peak calls and the authors' own quantification in one patient. The ENL 'persistence' reproduces here as a redistribution toward myeloid-program peaks rather than as absolute retention, and ENL's absolute retention is the lowest of the five channels.

    Evidence that argues against this (1)
    • The MLL::ENL transcript is still expressed in the relapse sample by the authors' own RT-PCR, so a protein that is transcribed but chromatin-undetectable is either displaced, degraded or below assay resolution, and the paper does not distinguish these. Janssens DH et al. 2024
Connected to