Oncofusion

start / graph / ana-unstated-system-menin-utx-switch-resolved

Analysisestablished confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.Q18b-unstated-systems

KMT2A-rearranged or NPM1-mutant? Both, in separate arms: WT-MA9 mouse, MOLM13, OCI-AML3, and two patients on AUGMENT-101

also called GSE186711 provenance · Menin-UTX switch system · UTX menin inhibitor resistance genotypes

An entry on menin-inhibitor resistance notes that the paper never says whether its leukaemias were KMT2A-rearranged, NPM1-mutant, or both. The deposited data and full text say both, in clearly separated arms: mouse MLL-AF9 cells and human MOLM13 for the KMT2A side, OCI-AML3 for the NPM1 side, and one patient of each genotype on a menin-inhibitor trial.

The same thing, for a scientist

`resistance-mll34-utx-tumour-suppressor-failure` carries the caveat *"The abstract does not state whether the leukaemias were KMT2A-rearranged, NPM1-mutant or both, which matters for transfer to this project's population"* [E1]. It is both, and the arms are separable. GEO series **GSE186711** holds 72 samples - **48 *Mus musculus*, 24 *Homo sapiens***: 42 carry `cell line: WT-MA9` with `genotype: Stably expressing pMSCV-MLL-AF9-GFP` (some further annotated with lenti-Cas9 and sgUtx or sgChr8), 6 are NIH-3T3 fibroblasts, 6 are `cell line: MOLM13, genotype: Endogenous MLL-AF9 rearrangement`, and 6 are `cell line: OCI-AML3, genotype: DNMT3A;NPM1 mutant`, with `treatment: DMSO` and `treatment: MI-503` at 15 samples each and UTX and Menin ChIP arms [E2]. The recovered full text adds the clinical layer: longitudinal RNA-seq on **two patients** treated with SNDX-5613 on **AUGMENT-101 (NCT04065399)** - patient 1 NPM1c-mutated, patient 2 MLL-rearranged [E3]. So the discovery engine is murine MLL-AF9, the human leukaemia arms are one MLLT3 line and one NPM1c line, and the clinical evidence is n = 1 per genotype. **No KMT2A::AFF1 sample appears anywhere, and no B-lymphoid sample appears anywhere.**

Go deeper
GSE186711 by arm
Cell lineGenotype as depositednSpecies
WT-MA9Stably expressing pMSCV-MLL-AF9-GFP (+/- lenti-Cas9, sgUtx or sgChr8)42mouse
NIH-3T3fibroblasts6mouse
MOLM13Endogenous MLL-AF9 rearrangement6human
OCI-AML3DNMT3A;NPM1 mutant6human

Treatment arms are DMSO (15) versus MI-503 (15), with UTX ChIP (9) and Menin ChIP samples. The full text describes a second, independent genome-wide CRISPR screen in a separately derived MLL-AF9 mouse line using VTP-50469, and a codeletion experiment in an NPM1c-mutant human line - the OCI-AML3 arm 3.

OCI-AML3 is worth naming precisely because the graph's caveat treated NPM1-mutant as a possible alternative to KMT2A-rearranged: OCI-AML3 carries both DNMT3A and NPM1 mutations, so the NPM1 arm is also a DNMT3A arm and the two cannot be separated in it.

What this changes for menin-inhibitor resistance in infant disease

The Menin-UTX switch is one of the mechanistic resistance models this graph carries for menin inhibitors, and menin inhibitors are the project's lead modality in KMT2A::AFF1 infant ALL. The deposition says the switch was established in myeloid cells throughout - mouse MLL-AF9, human MLL-AF9, human NPM1c AML - and validated clinically in two adults on an AML-weighted trial, one of whom is not KMT2A-rearranged at all.

That is not a reason to discard the model. It is a reason for resistance-mll34-utx-tumour-suppressor-failure to say which genotype and lineage each claim came from, because a resistance mechanism that operates through a UTX-dependent complex switch could plausibly depend on the partner: MLL-AF9 and KMT2A::AFF1 assemble different elongation complexes, and the switch is a statement about which complex the cell falls back on.

Facts and figures
scriptnone - GEO accession-display text endpoint plus NCBI efetch db=pmc full text
statusrun
result summaryBoth genotypes, in separate arms. Discovery: murine MLL-AF9 (WT-MA9, 42 samples) plus a second independent MLL-AF9 mouse line. Human lines: MOLM13 (endogenous MLL-AF9) and OCI-AML3 (DNMT3A;NPM1 mutant). Clinical: 2 patients on AUGMENT-101, one NPM1c and one MLL-r. Zero KMT2A::AFF1 samples and zero B-lymphoid samples.
counts
samples_kmt2a_aff10.000samples_kmt2a_aff1: 0samples_b_lymphoid0.000samples_b_lymphoid: 0patients_kmt2a_rearranged1.000patients_kmt2a_rearranged: 1patients_in_clinical_arm2.000patients_in_clinical_arm: 2samples_npm1_mutant6.000samples_npm1_mutant: 6samples_human24.0samples_human: 24samples_mouse48.0samples_mouse: 48samples_mll_af948.0samples_mll_af9: 48samples_total72.0samples_total: 72
inputs
[
  "PMID 36264143",
  "GSE186711",
  "PMC9827117",
  "NCT04065399"
]
counts — the numbers behind the chart
{
  "samples_total": 72,
  "samples_mouse": 48,
  "samples_human": 24,
  "samples_mll_af9": 48,
  "samples_npm1_mutant": 6,
  "samples_kmt2a_aff1": 0,
  "samples_b_lymphoid": 0,
  "patients_in_clinical_arm": 2,
  "patients_kmt2a_rearranged": 1
}
entries affected
[
  "resistance-mll34-utx-tumour-suppressor-failure"
]
import axes
[
  "fusion: KMT2A::MLLT3 and NPM1c, not KMT2A::AFF1",
  "species: mouse dominant",
  "lineage: myeloid throughout",
  "age: adult trial patients, not infants"
]
How this could bite us back
  • If the Menin-UTX switch is the dominant menin-inhibitor escape route in myeloid disease but not in B-lymphoid disease, then a combination designed to block it would add toxicity without benefit in infant KMT2A::AFF1 ALL - and nothing in this deposition can distinguish those cases.
  • UTX/KDM6A is a tumour suppressor; a strategy that depends on its activity being retained is a strategy that fails in exactly the patients whose disease has already lost it, and loss is the commoner event.
  • The clinical corroboration is two patients on an adult trial, one of whom is not KMT2A-rearranged; treating that as validation of a paediatric resistance mechanism overstates it by a wide margin.
Still open
  1. Does the Menin-UTX complex switch occur in a KMT2A::AFF1 pro-B cell? Every deposited sample bearing on it is myeloid, and the fallback complex is precisely what differs between MLLT3 and AFF1 fusions.
  2. Is the switch a resistance mechanism or a differentiation response? The clinical arm is n = 1 per genotype with no comparator, and OCI-AML3 confounds NPM1 with DNMT3A.
Every source behind this entry (3)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    The Menin-MLL inhibitor resistance paper reports a molecular switch between mammalian MLL complexes dictating response to Menin-MLL inhibition, and its abstract does not state whether the leukaemias studied were KMT2A-rearranged, NPM1-mutant or both.

    A Molecular Switch between Mammalian MLL Complexes Dictates Response to Menin-MLL Inhibition. · Soto-Feliciano YM, Sánchez-Rivera FJ, Perner F, Barrows DW, Kastenhuber ER, Ho YJ, Carroll T, Xiong Y, Anand D, Soshnev AA, Gates L, Beytagh MC, Cheon D, Gu S, Liu XS, Krivtsov AV, Meneses M, de Stanchina E, Stone RM, Armstrong SA, Lowe SW, Allis CD · Cancer Discov 2023
    PMID 36264143DOI
    What this evidence cannot show

    Identifiers taken from the NCBI esummary record fetched on 2026-09-19. Full text was read from NCBI efetch db=pmc for PMC9827117 after Europe PMC fullTextXML returned HTTP 500 - one of five papers in this beat recoverable by the NCBI route after the Europe PMC route failed.

    Evidence that argues against this (1)
    • A resistance mechanism defined by complex switching predicts that the fallback complex matters, and MLL-AF9 and KMT2A::AFF1 assemble different elongation complexes, so a switch characterised entirely in MLLT3 and NPM1c myeloid systems may not describe the fallback available to a pro-B blast. Biswas D et al. 2011
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    GEO series GSE186711 contains 72 samples - 48 Mus musculus and 24 Homo sapiens - comprising 42 WT-MA9 samples annotated 'Stably expressing pMSCV-MLL-AF9-GFP', 6 NIH-3T3 fibroblasts, 6 MOLM13 annotated 'Endogenous MLL-AF9 rearrangement' and 6 OCI-AML3 annotated 'DNMT3A;NPM1 mutant', with 15 DMSO and 15 MI-503 samples and UTX and Menin ChIP arms.

    What this evidence cannot show

    Read on 2026-09-19 from the GEO accession-display text endpoint and tallied by characteristic string; a sample may carry several strings so the counts do not partition the series. Submitter-entered. This accession was recovered from the paper's data-availability statement; elink dbfrom=pubmed&db=gds returns no GEO link for this PMID.

    Evidence that argues against this (1)
    • OCI-AML3 carries DNMT3A mutations alongside its NPM1 mutation, so the deposition's NPM1 arm confounds the two lesions and cannot on its own attribute a switch phenotype to NPM1c. Tiacci E et al. 2012
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    The paper's clinical arm is longitudinal RNA-seq on two patients treated with SNDX-5613 on AUGMENT-101 (NCT04065399) - patient 1 NPM1c-mutated and patient 2 MLL-rearranged - and its discovery screens were performed in Cas9-expressing mouse leukaemia cells driven by a human MLL1-AF9 transgene plus an independently derived MLL-AF9 mouse line screened with VTP-50469.

    A Molecular Switch between Mammalian MLL Complexes Dictates Response to Menin-MLL Inhibition. · Soto-Feliciano YM, Sánchez-Rivera FJ, Perner F, Barrows DW, Kastenhuber ER, Ho YJ, Carroll T, Xiong Y, Anand D, Soshnev AA, Gates L, Beytagh MC, Cheon D, Gu S, Liu XS, Krivtsov AV, Meneses M, de Stanchina E, Stone RM, Armstrong SA, Lowe SW, Allis CD · Cancer Discov 2023
    PMID 36264143DOI
    What this evidence cannot show

    Read from the NCBI efetch db=pmc full text on 2026-09-19 and quoted from the figure legend and results text. n = 1 per genotype in the clinical arm, which is a case observation rather than a cohort; the paper presents it as corroboration, not as a frequency estimate, and this entry makes no frequency claim from it. AUGMENT-101 is adult-weighted and neither patient is described as paediatric or KMT2A::AFF1.

    Evidence that argues against this (1)
    • VTP-50469 and MI-503 are tool compounds and SNDX-5613 is the clinical agent; a switch demonstrated with three different menin inhibitors across mouse and human systems is robust to compound, but exposures and target engagement differ enough that a resistance kinetics claim does not transfer between them. Neville D et al. 2026
Connected to