Oncofusion

start / graph / exp-xpo1-chip-in-kmt2a-rearranged-with-menin-inhibitor

Experimentspeculative confidence 1 of 5: Our own inference or analogy. Not shown to be true.Q61-export-axis

The reciprocal ChIP: is chromatin-bound XPO1 present in a KMT2A-rearranged cell, and does a menin inhibitor move it?

also called XPO1 ChIP in KMT2A-r · reciprocal menin-XPO1 occupancy experiment

Nobody has ever checked whether the export protein XPO1 sits on the HOXA genes in a KMT2A-rearranged leukaemia cell - the disease this whole project is about. One standard experiment, run with and without a menin inhibitor, answers that and also tests whether the two drug targets are on one assembly or two.

The same thing, for a scientist

**The gap.** `ana-xpo1-chromatin-measurements-never-made` records that no XPO1 or CRM1 chromatin-occupancy measurement in a KMT2A-rearranged cell appears in the published titles and abstracts, and that no CRM1 occupancy under a menin inhibitor has been reported at all. A 2026 preprint reports the mirror - menin displaced from chromatin by an XPO1 inhibitor - and no peer-reviewed work reports either direction. **The design.** XPO1 ChIP-seq (or CUT&RUN, which is what the DEK::NUP214 group used and which needs far less material) in a KMT2A-rearranged line, in four arms: 1. vehicle 2. **revumenib or ziftomenib** at an established HOXA-suppressing concentration and time 3. **selinexor** at a concentration below the degradation threshold, to separate displacement from protein loss 4. both Readouts in every arm: **XPO1 occupancy**, **menin occupancy**, **KMT2A-fusion occupancy** (N-terminal antibody), and **HOXA9/MEIS1 expression**. Controls: an isotype/IgG arm, a fusion-negative lineage-matched line, and a **spike-in normalisation** because a common-essential target's inhibitor changes global signal. **Lines.** SEM or RS4;11 (KMT2A::AFF1, B-lineage - this graph's index genotype, and the lineage every other XPO1 experiment in the literature lacks), with MV4;11 (KMT2A::MLLT3, myeloid) as the bridge to the published literature and LOUCY as the positive control for a chromatin-bound XPO1 signal. **Why LOUCY is not optional.** The whole experiment's interpretation depends on the XPO1 antibody producing a locus-restricted signal at all. LOUCY is the one cell line in which that signal is published, with two independent antibodies. **Run it in the same batch or the negative is uninterpretable** - this is rule 67 applied to a wet experiment rather than a query.

Go deeper
What each outcome would mean
XPO1 at HOXA in KMT2A-r?menin inhibitor moves XPO1?reading
no-the XPO1 address is genuinely a KMT2A-FREE mechanism; the three-addresses framing survives intact and the preprint's menin claim needs another explanation
yesnoXPO1 is upstream and platform-like in this genotype too; consistent with the preprint; menin and XPO1 are one assembly with XPO1 as the anchor
yesyesthe two are mutually dependent, which argues for a condensate rather than a layered platform, and favours the C-body reading
yes, but everywhere-the 'platform' is not locus-restricted in this cell type and the address model does not apply here

Every one of those four is publishable and three of them change what this graph says. That is the mark of a well-posed experiment rather than a confirmatory one.

Facts and figures
designXPO1 CUT&RUN (preferred) or ChIP-seq in KMT2A-rearranged lines under four arms - vehicle, menin inhibitor, selinexor below the degradation threshold, and both - with parallel menin, KMT2A-fusion and IgG tracks and spike-in normalisation
readoutXPO1 occupancy at HOXA/HOXB and MEIS1; menin occupancy at the same loci; KMT2A-fusion occupancy; HOXA9 and MEIS1 expression by RT-qPCR
controlsLOUCY in the same batch as the published-positive for a locus-restricted XPO1 signal (rule 67 applied to a wet assay); a fusion-negative lineage-matched line; IgG; a selinexor concentration and time chosen to displace without degrading, verified by XPO1 western in the same lysates
cost tiermed
timeline6-10 weeks including antibody validation
why it is worth doingit is the only experiment that tests whether the XPO1 address exists in this graph's index genotype at all, and it simultaneously tests the reciprocal of the only claim linking XPO1 to the menin interface
cell lines
[
  "SEM (KMT2A::AFF1, B-lineage)",
  "RS4;11 (KMT2A::AFF1, B-lineage)",
  "MV4;11 (KMT2A::MLLT3, myeloid - bridge to the published literature)",
  "LOUCY (positive control)"
]
known failure modes
[
  "selinexor degrades XPO1 and its docked partners at high exposure, so a lost peak can be lost protein - hence the sub-degradation arm and the western in the same lysates",
  "a common-essential target's inhibitor shifts global chromatin signal, so without spike-in normalisation every track moves",
  "the XPO1 antibody may not work for CUT&RUN even if it works for ChIP - the LOUCY arm is what tells you"
]
How this could bite us back
  • A negative result here is genuinely ambiguous unless the LOUCY control works in the same batch, and antibody performance across cell types is not guaranteed.
  • If the answer is that XPO1 is everywhere on chromatin in these cells, the experiment kills the locus-restricted platform model in this genotype and leaves the preprint's menin claim without a mechanism - which is a useful but unwelcome outcome.
  • The experiment is cheap enough that a negative is unlikely to be published, so if it has already been run privately the field would not know.
Still open
  1. Is there an existing KMT2A-rearranged ChIP dataset with a spare chromatin aliquot, so this could be run as an add-on rather than a new study?
  2. Which anti-XPO1 antibodies are CUT&RUN-validated? The published chromatin work is all ChIP, and the reagent question may be the rate-limiting step.
  3. Should the fourth arm use eltanexor rather than selinexor, given that the NPM1c literature says duration rather than potency is what matters?
Every source behind this entry (3)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    No XPO1 or CRM1 chromatin-occupancy measurement in a KMT2A-rearranged cell, and no CRM1 occupancy under a menin inhibitor, appears in PubMed titles and abstracts as of 2026-09-20, with operand controls returning 4, 6, 15 and 16 records respectively.

    What this evidence cannot show

    Titles and abstracts only; supporting figures not searched (rule 80). Full working in `ana-xpo1-chromatin-measurements-never-made`.

    Evidence that argues against this (1)
    • A negative XPO1 ChIP in a KMT2A-rearranged line would not refute the preprint's claim, because XPO1 could stabilise menin without being locus-restricted itself - a diffuse chromatin association gives a flat track and the same biology. The experiment is decisive for the ADDRESS model and only suggestive for the stabilisation model. Uddin MH et al. 2026
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The assay design is available and validated in this exact biology: CUT&RUN detected DEK::NUP214 at FOXC1 and HOXA/HOXB promoters and detected the selective reduction of that binding under XPO1 inhibition, in t(6;9) cells.

    DEK::NUP214 acts as an XPO1-dependent transcriptional activator of essential leukemia genes. · Kaya F, Bewicke-Copley F, Miettinen JJ, et al. · Leukemia 2025
    PMID 40204893DOI
    What this evidence cannot show

    Abstract-level. Cited as a design precedent rather than as evidence about KMT2A-rearranged disease, in which this paper contains nothing. SCOPE CORRECTED 2026-09-21. The abstract of PMID 40204893 was re-read in full. The 57 primary AML samples are the MULTI-OMICS AND DRUG-SCREEN arm only - whole genome sequencing, targeted sequencing, transcriptomics and >500-compound screening - and that arm is what shows the selective t(6;9) response to selinexor and eltanexor. The CUT&RUN is a SEPARATE experiment with no n stated in the abstract, and the key result that DEK::NUP214 binding falls under XPO1 inhibition is stated as occurring 'in t(6;9) cells', not in the 57 primary samples. Attributing the occupancy readout to 57 primary patient samples makes it a much stronger claim than the paper supports. The conflation entered via a hand-back summary and was then repeated in a dispatch brief before being caught by the beat that received it.

    Evidence that argues against this (1)
    • NAMED METHODOLOGICAL DEFEATER for the proposed design's target loci. Jain et al. identified about 3,000 chromosomal loci - mostly ACTIVE PROMOTERS - that are prone to non-specific enrichment in ChIP and appear as 'Phantom Peaks', and found potential Phantom Peaks throughout published profiles of epigenetic regulators. FOXC1 and the HOXA/HOXB promoters in a HOX-driven leukaemia are among the most actively transcribed promoters in the cell, which is precisely where this artefact lives. The design as written has four drug arms and no depletion arm; Jain et al.'s stated requirement is validation with chromatin from cells in which the protein of interest has been depleted. An XPO1-degron or XPO1-C528S arm should be added before the experiment is run, not after it produces a peak at HOXA. Jain D et al. 2015
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    A locus-restricted chromatin-bound CRM1 signal is reproducible in LOUCY with two independent anti-CRM1 antibodies and is absent from K562 and HL60, which makes LOUCY a usable positive control for the antibody and the protocol.

    PMID 31755865DOI
    What this evidence cannot show

    The antibodies are not named in what was read here, so sourcing the same reagents requires the paper's methods section. LOUCY is a T-ALL line and has been in culture for decades.

    Evidence that argues against this (1)
    • NAMED METHODOLOGICAL DEFEATER: antibody-independence is not evidence of specificity, so 'reproducible with two independent anti-CRM1 antibodies' does not make LOUCY a validated positive control. Teytelman et al. found 238 euchromatic 'hyper-ChIPable' loci enriched for Sir2, Sir3, Sir4, Ste12 and Cse4 alike; the correlation between transcription level and ChIP enrichment extended genome-wide, was confirmed by ChIP-qPCR rather than being a sequencing artefact, and was reproduced by immunoprecipitating a nuclear-localised GFP, whose enrichment at these loci increased when the GAL genes were induced. Two antibodies against the same protein do not control for an artefact that a nuclear GFP also produces. The absence of the signal in K562 and HL60 is better evidence than the two antibodies - but K562 and HL60 are also not HOX-high, so expression level and cell identity are confounded in that comparison. Teytelman L et al. 2013
Connected to