Oncofusion

start / graph / risk-partner-blind-menin-inhibitor-eligibility

Riskmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q-usp2-maml2-second-bcr

Menin inhibitors are prescribed on 'KMT2A-rearranged', a category that contains proteins as different as KMT2A::AFF1 and KMT2A::USP2

also called partner-blind enrolment · KMT2A-rearranged as a single category · minor-BCR patients on menin inhibitors

The approved drug for KMT2A leukaemia is given to anyone whose KMT2A gene is rearranged, whatever it is joined to. But some of those rearrangements make a protein that keeps parts the common ones throw away, and nobody has checked whether the drug works on them.

The same thing, for a scientist

Menin-inhibitor eligibility is written on rearrangement status. The registrational study's inclusion criteria read 'acute leukemia harboring KMT2A rearrangement', with phase 2 cohorts defined as R/R ALL/MPAL with KMT2A rearrangement and R/R AML with KMT2A rearrangement - the fusion partner is not specified anywhere in the criteria [E1]. That is a reasonable design decision given that the menin-binding motifs sit 5' of every major-BCR break. **It becomes a risk for the minor-BCR class.** A BCR2 break retains the complete PHD1-3 cluster, bromodomain and ePHD4, modules that reverse KMT2A from activator to repressor when they engage CYP33/PPIE, and that block transformation when added back to MLL-AF9 [E2][E3]. A KMT2A::USP2 protein is therefore not the molecule the menin-inhibitor programme was built on, and **no published work connects any minor-BCR fusion to menin: the USP2/menin and MAML2/menin PubMed intersections are both empty** [E4]. The risk has two halves. Patients with these fusions who receive the drug may not benefit and nobody would detect that, because the partner is usually not determined. And patients whose cytogenetically cryptic KMT2A::USP2 is never found are not eligible at all [E5].

Go deeper
Two failure modes, opposite in direction

Failure mode 1 - treated without evidence. A child with KMT2A::USP2 whose fusion is detected by RNA sequencing is 'KMT2A-rearranged' and is eligible. They receive a drug whose mechanism has never been tested against their fusion protein, which differs structurally from every fusion in the registrational cohorts by retaining an entire chromatin-reader module. If it does not work in that genotype, the signal would be lost in a cohort where 98% of patients are major-BCR: 30 USP2 patients against 3336 BCR1 patients in the largest breakpoint series 2. A subgroup that small cannot declare itself.

Failure mode 2 - not treated at all. Roughly half of KMT2A::USP2 cases cannot be resolved by standard cytogenetic approaches, and karyotype plus break-apart FISH can both read normal 5. A patient whose rearrangement is never found does not meet an inclusion criterion written on rearrangement status, for a drug aimed at a module their fusion demonstrably retains.

The two failure modes select opposite patients and neither is visible in trial data.

Why the minor-BCR class is structurally different, stated carefully1 deeper

The retained modules are not a detail. PHD1-3 plus the bromodomain constitute the unit that, on binding CYP33/PPIE, enables recruitment of a Polycomb repressor complex to CXXC and converts KMT2A from a transcriptional activator into a repressor 2. Adding two or more PHD fingers back to MLL-AF9 abolished immortalisation of haematopoietic progenitors and reduced Hoxa9 locus association and upregulation 3.

So a minor-BCR fusion carries machinery that, in a major-BCR fusion's context, suppresses the HOXA programme. The menin-inhibitor mechanism runs through displacing the fusion from HOXA/MEIS1. If the HOXA programme is not what a minor-BCR fusion drives, a menin inhibitor has nothing to displace. This entry does not assert that - nobody has measured HOXA output in a KMT2A::USP2 cell 4 - but it is the specific, mechanistically grounded reason the category is doing more work than it can bear.

The argument on the other side

Fairness requires stating it. The menin-binding motifs MBM1 and MBM2 lie in the KMT2A N-terminus and are retained by every fusion in both BCRs, so the drug's binding target is present regardless of partner. The registrational cohort 2D was written around 'a genetic mutation expected to lead to HOX/MEIS upregulation', which is a mechanistic rather than a cytogenetic criterion and shows the designers were thinking about the programme, not only the rearrangement 1. And partner-agnostic eligibility is what made the trial feasible at all; a partner-stratified design would have been unenrollable for anything below the top five partners.

The criticism here is therefore not that the design was wrong. It is that the resulting evidence does not extend to the minor-BCR class, and nothing currently in place would reveal that.

What would close this

Three things, in increasing cost.

  1. Report partner identity in menin-inhibitor trial datasets. Sponsors hold this for many patients already. A retrospective look at response by partner, even underpowered, would show whether minor-BCR patients responded at all.
  2. Measure HOXA9 and MEIS1 in a KMT2A::USP2 patient sample. One RNA-seq against KMT2A::AFF1 controls. Banked material exists at DCAL and at the centres that published the cases.
  3. Co-immunoprecipitate menin with a minor-BCR fusion. Direct, unambiguous, and needs a model that does not yet exist - the same obstacle that blocks everything else about this genotype.

Item 2 is cheap, retrospective and answers the question that matters most, and this beat found nobody who has done it.

Facts and figures
kinddelivery
severityModerate at population level (minor-BCR fusions are 1.4-6% of KMT2A-rearranged disease) and potentially high for the individual, because both failure modes end in a patient not receiving an effective therapy - one by receiving an untested one, the other by being ineligible.
mitigationDetermine and report the fusion partner and breakpoint cluster region for every patient receiving a menin inhibitor, and analyse response by BCR class retrospectively in existing sponsor datasets. Where partner determination is not possible, record that as unknown rather than as absent. In parallel, measure HOXA9/MEIS1 in banked minor-BCR patient material, which is the cheapest test of whether the drug's mechanism applies.
why it is undetectable in trialsMinor-BCR patients are at most a few percent of any KMT2A-rearranged cohort, so a complete absence of response in that subgroup would not perturb a trial-level response rate, and no trial reports response by partner.
argument on the other sideMBM1 and MBM2 are retained by every KMT2A fusion in both cluster regions, so the drug's binding target is present regardless of partner, and partner-agnostic eligibility is what made the registrational trial feasible.
affects
[
  "drug-revumenib",
  "drug-ziftomenib",
  "drug-bleximenib",
  "fusion-kmt2a-usp2",
  "fusion-kmt2a-usp8",
  "fusion-kmt2a-maml2",
  "breakpoint-kmt2a-minor-bcr"
]
numbers
{
  "bcr2_share_of_kmt2a_r": "1.4% overall (47/3401), ~6% in the targeted-NGS subset",
  "usp2_patients": 30,
  "bcr1_patients": 3336,
  "menin_publications_on_usp2": 0,
  "menin_publications_on_maml2": 0
}
How this could bite us back
  • Naming this risk could discourage clinicians from offering an approved and effective drug to a patient on the grounds that their partner is unusual. That would be the wrong response: there is no evidence menin inhibitors fail in minor-BCR disease, only an absence of evidence that they work, and an absence of evidence is not a reason to withhold the only targeted option a patient has.
  • Requiring partner determination for eligibility would exclude centres that cannot sequence, which are disproportionately the centres treating infants in lower-income settings. Any move toward partner-stratified eligibility must not become an access barrier.
  • If a retrospective analysis found no response in minor-BCR patients, the numbers would be small enough that the finding could be chance - and publishing it could remove access for a genotype on the strength of a handful of patients.
Still open
  1. How many patients in menin-inhibitor trials carried a minor-BCR fusion, and did they respond? Sponsors may be able to answer this from existing molecular data without enrolling anyone new.
  2. Do minor-BCR fusions drive the HOXA/MEIS1 programme at all? This is the mechanistic question the whole risk rests on, and one RNA-seq on banked material would answer it.
  3. Should breakpoint cluster region, not just rearrangement status, become a reported variable in KMT2A-rearranged trials? It is determinable from the same sequencing that finds the fusion and costs nothing extra.
  4. Does the same concern apply to KMT2A partial tandem duplication and to NUP98-rearranged patients enrolled on the same mechanistic cohort, whose protein architectures also differ from the canonical fusions?
Every source behind this entry (5)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry

    The registrational revumenib study enrols on rearrangement status without specifying the fusion partner: inclusion requires 'acute leukemia harboring KMT2A rearrangement, NUP98 rearrangement, or NPM1 mutation', with phase 2 cohorts 2A (R/R ALL/MPAL with KMT2A rearrangement) and 2B (R/R AML with KMT2A rearrangement), and cohort 2D defined mechanistically as leukaemia with a mutation expected to lead to HOX/MEIS upregulation.

    NCT04065399
    What this evidence cannot show

    Registry text is the sponsor's summary of the protocol, not the protocol, and inclusion criteria are abbreviated as 'Key Inclusion Criteria'. Status recruiting, phase 1/2, estimated enrolment 447, minimum age 30 days, read from the raw ClinicalTrials.gov v2 API on 2026-09-19. The registry does not report how partner identity was recorded for enrolled patients, so it cannot be inferred that partners were unknown - only that they were not an eligibility criterion.

    Evidence that argues against this (1)
    • Partner-agnostic eligibility is defensible on the mechanism: MBM1 and MBM2 lie 5' of every KMT2A breakpoint in both cluster regions, so the drug's binding site is present in every KMT2A fusion protein. Cohort 2D's HOX/MEIS-upregulation wording also shows the design was not purely cytogenetic. The criticism is about what the resulting evidence covers, not about the design being unreasonable. Meyer C et al. 2023
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    A minor-BCR break places the complete PHD1-3 cluster, bromodomain and ePHD4 into the direct fusion protein - modules that on binding CYP33/PPIE enable recruitment of a Polycomb repressor complex to CXXC and reverse KMT2A from activator to repressor - and 47 of 3401 patients (30 of them USP2-fused) carry such a break, against 3336 major-BCR patients.

    The KMT2A recombinome of acute leukemias in 2023. · Meyer C, Larghero P, Almeida Lopes B, Burmeister T, Gröger D, et al., Marschalek R · Leukemia 2023
    PMID 37019990DOI
    What this evidence cannot show

    Single referral centre. The domain-retention claim is positional - inferred from the breakpoint interval and the KMT2A domain map - and has not been confirmed at the protein level in any minor-BCR patient. The CYP33/PPIE repressor-switch description is the authors' summary of prior literature, not a measurement in this study. Counts and coordinates read from the published article in this session.

    Evidence that argues against this (1)
    • Retaining the modules is not the same as using them. A minor-BCR fusion could carry an inert PHD cluster - misfolded, unoccupied or lacking a partner in the relevant cell - in which case it behaves like a major-BCR fusion and the concern evaporates. Nobody has looked. Muntean AG et al. 2008
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Inclusion of two or more PHD fingers in MLL-AF9 blocked immortalisation of haematopoietic progenitors, reduced association with the Hoxa9 locus and suppressed Hoxa9 upregulation - i.e. the retained modules act on precisely the programme menin inhibitors are designed to switch off.

    The PHD fingers of MLL block MLL fusion protein-mediated transformation. · Muntean AG, Giannola D, Udager AM, Hess JL · Blood 2008
    PMID 18796627DOI
    What this evidence cannot show

    Murine retroviral immortalisation assay with MLL-AF9 only; no minor-BCR fusion was tested and no menin inhibitor was used. The extrapolation from 'PHD fingers reduce Hoxa9 association in MLL-AF9' to 'a minor-BCR fusion may not drive HOXA' is an inference made in this entry, not a published finding. Abstract and central claim read in this session.

    Evidence that argues against this (1)
    • Thirty KMT2A::USP2 patients have leukaemia, so whatever the retained PHD fingers do, they do not prevent disease. The programme those leukaemias run might still be HOXA-driven by a route Muntean's add-back construct did not model, in which case menin inhibition would work normally. Meyer C et al. 2019
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    No published work connects either of this beat's fusions to menin: (MLL OR KMT2A) AND USP2 AND menin returns 0 PubMed records, as does the same query with revumenib or MEN1 substituted, and (MLL OR KMT2A) AND MAML2 AND menin also returns 0.

    PubMed E-utilities esearch counts for KMT2A partner / menin intersection queries, run 2026-09-19 · National Center for Biotechnology Information (NCBI), PubMed · PubMed (NCBI E-utilities) 2026
    What this evidence cannot show

    A counted-absence claim and only as good as its symbol set; run over USP2, MAML2, MEN1, menin and revumenib but not over every menin-inhibitor code or trade name, and it would miss any paper not naming both terms in an indexed field. It is a statement about the indexed literature and not about clinical experience.

    Evidence that argues against this (1)
    • Because eligibility is by rearrangement, patients with these fusions have very probably received menin inhibitors already without generating a partner-indexed publication. The literature gap is real; an exposure gap is not established, and sponsor datasets may contain the answer. Syndax Pharmaceuticals (sponsor) 2026
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Approximately half of KMT2A-USP2 and KMT2A-USP8 cases cannot be resolved by standard cytogenetic approaches and are likely under-recognised, with normal karyotypes and normal or atypical KMT2A FISH results reported.

    Cryptic and atypical KMT2A-USP2 and KMT2A-USP8 rearrangements identified by mate pair sequencing in infant and childhood leukemia. · Blackburn PR, Smadbeck JB, Znoyko I, Webley MR, Pitel BA, Vasmatzis G, Xu X, Greipp PT, Hoppman NL, Ketterling RP, Baughn LB, Lindsey KG, Schandl CA, Wolff DJ, Peterson JF · Genes, Chromosomes & Cancer 2020
    PMID 32196814DOI
    What this evidence cannot show

    Three patients selected because they were diagnostically difficult; the 'approximately half' figure summarises prior literature rather than measuring anything in this series. The denominator - all KMT2A::USP2 patients including those FISH did catch - is unknown. Abstract read in full.

    Evidence that argues against this (1)
    • Targeted RNA sequencing now detects the fusion in most laboratories that run it, so the eligibility-exclusion half of this risk is shrinking wherever RNA-based fusion panels are standard. It persists in centres relying on karyotype and FISH, which is much of the world. Li EW et al. 2025
Connected to