Oncofusion

start / graph / hyp-q147-the-irx1-address-requires-the-aff1-partner-not-the-menin-interface

Hypothesisweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.Q147-hoxa-irx1-across-partners

The IRX1 address requires the AFF1 partner; the menin interface explains the shared HOXA address and not this one

Every KMT2A fusion keeps the same front half of the protein, the part that binds menin - and every one of them sits on HOXA. Only the AFF1 version sits on IRX1, so whatever puts it there is not the menin-binding part.

The same thing, for a scientist

**The argument, in three measured steps.** 1. The KMT2A N-terminal fragment - menin interface, LEDGF interface, CXXC domain - is identical in all six fusions in GSE252378. 2. **HOXA9/HOXA10 is fusion-occupied in 17 of 18 non-AFF1 samples**, at 2.94-29.1x. The shared address travels with the shared fragment [E1]. 3. **IRX1 is fusion-occupied in 0 of 18**, and in 3 of 8 KMT2A::AFF1 [E1]. An address that appears in one partner and not the other five **cannot be a property of the fragment they share.** So the menin-interface hypothesis is **supported on HOXA and silent on IRX1**. It is not threatened by this result - it is bounded by it: menin-dependence would predict the partner-invariant address, and the partner-restricted one needs a second explanation. **The competing explanation is not a partner at all** and is written out separately as `hyp-q147-irx1-occupancy-is-a-lineage-switch-state-confounded-with-one-institution`. In GSE252378 the two cannot be separated: all three IRX1-occupied samples are AFF1, from one institution, and lineage-switch-associated. **Published support for the partner reading, from a different assay and a different direction:** Kuhn, Loscher & Marschalek titled their mechanistic study *a novel **t(4;11)-specific** cancer mechanism* and showed IRX1 binding the MLL-AF4 complex at target promoters [E2]. They did not test the other partners, so their 'specific' is where the mechanism was demonstrated; this beat supplies the exclusion.

Go deeper
What it predicts that is checkable now
  • IRX1 occupancy should be absent in KMT2A-AFF4 and KMT2A-AFF3 too, and present in t(4;11) variants - i.e. it should track the AF4/AFF family rather than AFF1 exactly. No deposit currently has the AFF4/AFF3 samples.
  • Menin should be at HOXA in every partner and its presence at IRX1 should be restricted to AFF1. GSE252378 has no menin track for the non-AFF1 samples, so this is unmeasured.
  • A menin inhibitor should clear HOXA in all partners - and the one drug-exposed aliquot, KMT2A-ENL 152985, does lose HOXA (9.69x to 0.97x in its patient-matched pair), along with everything else.
  • The AFF1 C-terminus should be sufficient, which is a construct experiment.
Facts and figures
claimthe AFF1 C-terminus, or something it recruits, is required for oncoprotein occupancy at IRX1; the shared KMT2A N-terminal menin/LEDGF/CXXC module accounts for the partner-invariant HOXA address and not for IRX1
experiments
[
  "exp-q147-swap-the-aff1-c-terminus-onto-kmt2a-af9-and-score-irx1-occupancy",
  "exp-q147-kmt2a-n-cut-and-run-in-an-untreated-non-aff1-lineage-switch-blast",
  "exp-q147-paired-irx1-rna-and-oncoprotein-occupancy-in-non-aff1-kmt2a-r-blasts",
  "exp-q147-irx1-occupancy-under-partial-menin-inhibition-in-kmt2a-aff1"
]
analyses
[
  "ana-q147-irx1-fusion-occupancy-is-confined-to-kmt2a-aff1",
  "ana-q147-the-hoxa-irx1-anticorrelation-does-not-reproduce-in-kmt2a-enl",
  "ana-q147-a73051-is-a-hoxa-negative-irx1-negative-kmt2a-af10-blast"
]
falsifiers
[
  "KMT2A-N occupancy at the IRX1 gene body (chr5:3,595,831-3,601,403) reaching >= 2.0x an average promoter in BOTH N antibodies and above a width-matched cannot-be-occupied floor, in any depth-adequate primary KMT2A-r blast whose partner is not an AFF/AF4 family member. One such sample refutes the partner reading. CURRENT STATUS: 0 of 18 tested in GSE252378 (ENL 6, AFDN 4, AF10 3, AF9 3, EPS15 2), maximum observed 0.269x.",
  "A KMT2A-AF9 construct carrying the AFF1 C-terminus, at matched oncoprotein protein level in a single cell background, failing to acquire IRX1 occupancy while the KMT2A-AFF1 control has it - which would show the AFF1 C-terminus is not sufficient and point back at cell state. NOT YET RUN.",
  "Menin occupancy at the IRX1 promoter in a KMT2A::AFF1 IRX1-high blast being ABSENT while KMT2A-N is present, which would show the IRX1 address is menin-independent within AFF1 too and make the contrast with HOXA a menin-dependence contrast rather than a partner contrast. NOT YET RUN - GSE252378 has no menin track in an IRX1-high sample.",
  "An untreated non-AFF1 infant lineage-switching blast acquiring IRX1 occupancy, which would attribute the address to the cell state rather than the partner. NOT YET RUN - the only such pair in GSE252378 is revumenib-exposed."
]
How this could bite us back
  • If the IRX1 address is partner-driven and menin-independent, then the IRX1-high subgroup - which already has the worse prognosis and the younger age at diagnosis - is the group a menin inhibitor is least likely to help, and current trials do not stratify on it.
  • The hypothesis also predicts that the therapeutic handle is the AFF1 side of the fusion, which has no drug and no obvious pocket. A correct hypothesis here would make the clinical picture worse, not better, which is a reason to test it rather than to prefer the alternative.
  • If instead the cell-state reading is right, a menin inhibitor could DRIVE patients into the IRX1 state - the deposit's own paper shows lineage switching under revumenib - and the drug would be manufacturing the resistant compartment.
Still open
  1. Does KMT2A-AFF4 or KMT2A-AFF3 show IRX1 occupancy? If the address tracks the AF4/AFF family rather than AFF1 exactly, that localises it to the shared AF4 C-terminal homology domain.
  2. Is menin at IRX1 in the IRX1-occupied samples? GSE252378 has no menin track for p318 or p454, so the most direct test of whether this address is menin-dependent has never been run.
Every source behind this entry (2)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    In GSE252378, KMT2A-N occupancy clears the pre-registered floor at HOXA9-HOXA10 in 17 of 18 untreated non-AFF1 samples (2.94-29.1x) and at IRX1 in 0 of 18, while IRX1 is occupied in 3 of 8 KMT2A::AFF1 samples at 2.68-6.35x. The KMT2A N-terminal fragment carrying the menin and LEDGF interfaces and the CXXC domain is common to all six fusions.

    What this evidence cannot show

    One deposit. Partner, institution and lineage-switch state are confounded across the three IRX1-occupied samples, so the partner attribution is the hypothesis and not the observation. The argument from 'shared fragment cannot produce a partner-restricted address' is sound only if the address is truly partner-restricted rather than cohort-restricted. Every number is a WITHIN-SAMPLE share: an interval's mean coverage divided by that same library's mean coverage over all 19,404 hg38 RefSeq promoters, so 1.0 = an average promoter in that library. GSE252378 has no spike-in, so no between-library absolute statement is licensed; the only cross-sample quantity used is a ratio of two loci measured inside one library, where the unknown per-library scale cancels. The occupancy call is absolute, not rank-based: >=2.0x an average promoter in BOTH KMT2A-N antibodies AND above that library's own floor, where the floor is the 95th percentile of eight width-matched windows at loci that cannot be occupied in a leukaemia (ALB, RHO, MYH7, MYOD1, NEUROD1, KRT14, GFAP, INS). IgG is not the gate: q145 measured GSE252378 IgG libraries as 57-94% zero at promoters, and one of them (TB051939) reads 13.1x an average promoter at a locus that cannot be occupied.

    Evidence that argues against this (1)
    • The deposit paper's own model is that oncoprotein LEVEL and cell state redirect the programme, and that lineage-switching samples have reduced levels - which would produce the observed pattern with no partner effect at all, since all three IRX1-occupied samples are lineage-switch-associated. Janssens DH et al. 2024
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Kuhn, Loscher & Marschalek report IRX1 binding the MLL-AF4 complex at target-gene promoters and counteracting its promoter-activating function, and frame the IRX1/HOXA connection as a t(4;11)-specific cancer mechanism in their title.

    The IRX1/HOXA connection: insights into a novel t(4;11)-specific cancer mechanism · Kuhn A, Loscher D, Marschalek R · Oncotarget 7:35341-52 2016
    PMID 27175594DOI
    What this evidence cannot show

    HEK293T constructs plus t(4;11) material; the other KMT2A partners were not tested, so their claim of specificity is about the system studied and not a measured exclusion. Their causal direction is also the reverse of the one this hypothesis proposes - IRX1 acting on the complex rather than the complex being addressed to IRX1 - and both are compatible with the occupancy data.

    Evidence that argues against this (1)
    • If Kuhn's direction is the operative one, then IRX1 protein binding the MLL-AF4 complex could ALSO explain why KMT2A-N is detected at the IRX1 locus, via an IRX1-complex interaction at IRX1's own promoter, without any partner-specific addressing. That reading makes the AFF1 restriction a property of the AF4 complex's IRX1 affinity rather than of an address. Kuhn A et al. 2016
Connected to