Lineage switch as a menin-inhibitor escape route — documented in a patient on revumenib, and invisible to every resistance-mutation screen
also called menin-inhibitor-associated lineage switch · GMP-like lineage switching under menin inhibition · phenotypic escape from menin inhibition
Some leukaemias escape a drug not by changing their genes but by changing what kind of cell they are, swapping a B-cell identity for a myeloid one while keeping the same underlying fault. This has now been seen in a patient being treated with a menin inhibitor, and no DNA test would have found it.
The same thing, for a scientist
Profiling oncoprotein target sites across 36 KMT2A-rearranged leukaemia samples, including three that underwent lymphoid-to-myeloid lineage switching in response to therapy, showed that oncoprotein genomic enrichment is highly variable and dynamically regulated; in lineage-switching samples oncoprotein levels are **reduced** and the remaining oncoprotein preferentially activates granulocyte-monocyte progenitor (GMP) genes. Critically, one of those samples lineage-switched **during treatment with revumenib**, and in it the oncoprotein and menin were reduced to undetectable levels while ENL, a transcriptional cofactor of the oncoprotein, persisted on numerous oncoprotein-target loci including genes of the GMP-like lineage-switching program. The authors propose that KMT2A oncoproteins promote lineage switching through dynamic chromatin binding and may support resistance to menin inhibitors through the same mechanism [E1][E2]. This graph already holds the clinical shape of the phenomenon from the immunotherapy side: two KMT2A-rearranged lineage switches in its patient arm, one of which **reverted within nine days of stopping the drug with nothing given in between** [E3], which excludes subclonal outgrowth in that case and implicates reprogramming. Lineage switch therefore belongs in the resistance beat, not only in the patient and immunotherapy beats.
The observation that makes this a menin-inhibitor problem
Lineage switch has been a recognised escape route from CD19-directed immunotherapy in KMT2A-rearranged B-ALL for years, and this graph records that thoroughly (risk-lineage-switch-cd19-directed-therapy, mechanism-lineage-switch-kmt2a). What it has not recorded is that the same escape has now been observed under a menin inhibitor.
Janssens and colleagues profiled oncoprotein target sites across 36 representative KMT2A-rearranged leukaemia samples, three of which had undergone lymphoid-to-myeloid lineage switching in response to therapy. One of those three switched during revumenib treatment. In that sample the oncoprotein and menin were reduced to undetectable levels — the drug had worked completely, by its own mechanistic definition — and ENL persisted on numerous oncoprotein-target loci, including genes in the GMP-like lineage-switching program 12.
This is a single sample, and the entry says so. But it is a single sample of a kind the field has never had: a human KMT2A-rearranged leukaemia, profiled at the chromatin level, that changed lineage while on a menin inhibitor.
Why a resistance-mutation screen would have reported nothing
Take the two lineage switches already in this graph's patient arm and ask what a MEN1 resequencing panel would have found in each.
In the KMT2A::AFF4 infant case, a pro-B ALL became an M5-phenotype myeloid leukaemia by day 11 of blinatumomab, and then — with the drug stopped and no cytotoxic therapy substituted — the marrow reverted to roughly 40% lymphoblasts nine days later 3. A clone cannot be selected out and then selected back in nine days with nothing applied in either direction. That is a reversible phenotypic state, and there is no mutation to find.
In the KMT2A::AFDN adult case, a T-lymphoblastic lymphoma relapsed as monoblastic leukaemia carrying the identical t(6;11) and the identical fusion transcript, with the fusion sequence in the original mediastinal cells matching that in blood at the myeloid relapse 4. Same clone, different lineage; again nothing for a panel to report.
The general form of the mechanism is that the fusion is retained and the identity around it is rebuilt. Sequencing the drug target in that setting returns wild type and the patient is recorded as having failed for unknown reasons.
The mechanistic account, and what is odd about it
The Janssens data give a specific and slightly counterintuitive model: lineage switching is associated with reduced oncoprotein levels, and at reduced levels the oncoprotein preferentially activates GMP genes rather than the pro-B program it activates at high levels 1. Fusion-partner-specific binding patterns over the ALL-like and AML-like gene sets are highly correlated with how often each fusion is seen in ALL versus AML, which ties the dose-dependence of the oncoprotein directly to lineage assignment at the population level.
A menin inhibitor lowers oncoprotein function. On this model that is not a lineage-neutral act: it moves the cell down a dose-response curve whose lower end is a GMP-like program. Menin inhibition would then be a lineage-switch pressure by construction — the same thing this graph's hyp-menin-release-lineage-destination hypothesis predicts from a completely different starting point (the fetal-type, latent-myeloid-potential character of the infant t(4;11) blast). Two independent routes to the same prediction, and now one patient observation consistent with it.
The odd part is what persists. If the oncoprotein and menin are undetectable and the GMP program is running anyway, the program is not being maintained by the thing the drug targets. ENL on those loci is the candidate; that is treated as its own entry (resistance-enl-marked-epigenetic-memory) because it is a different claim with different consequences.
The clinical picture is not uniformly bad
A 2026 paediatric letter reports successful management of lineage-switch infant leukaemia with revumenib and haematopoietic cell transplantation 5 — that is, revumenib used against a lineage-switched leukaemia, successfully. Only the title and metadata of that letter were read for this entry, so nothing beyond the fact of the report is claimed.
That cuts in a useful direction. A menin inhibitor targets the fusion, and the fusion is retained through a lineage switch; so unlike a CD19-directed agent, a menin inhibitor is not automatically disarmed by the switch. The concern in this entry is narrower and more specific: that menin inhibition can be among the pressures that provoke a switch, and that a switch occurring on therapy will be scored as progression and treated as drug failure when the correct response may be different.
| kind | resistance |
|---|---|
| mechanism class | phenotypic / cell-identity escape (fusion retained, identity rebuilt) |
| seen in patients | yes — one KMT2A-rearranged leukaemia lineage-switched during revumenib treatment and was chromatin-profiled [E1]; two further KMT2A-rearranged lineage switches under other pressures are recorded in this graph's patient arm |
| seen in models | partially — the mechanism is inferred from primary patient samples rather than from an engineered model; no model of menin-inhibitor-induced lineage switch was identified |
| requires selection | not necessarily; a nine-day reversion with no intervening therapy excludes subclonal selection in at least one documented KMT2A-rearranged switch |
| direction observed | lymphoid to myeloid, with a GMP-like program, in all three switching samples profiled and in both patient cases in this graph |
| molecular correlate | reduced oncoprotein levels; preferential activation of GMP genes at low oncoprotein dose; ENL persisting on oncoprotein-target loci after menin and oncoprotein become undetectable |
| detectability | flow cytometry at every disease assessment, not DNA sequencing. A MEN1 or myeloid panel returns wild type. |
| why a mutation screen misses it | the fusion is retained unchanged and no recurrent driver mutation accompanies the switch; the reversible case proves no genetic event is required at all |
| relationship to existing entries | mechanism-lineage-switch-kmt2a and risk-lineage-switch-cd19-directed-therapy cover the phenomenon and its immunotherapy context. hyp-menin-release-lineage-destination predicts it for menin inhibitors from developmental biology. This entry records it as an observed resistance route under menin inhibition, which none of them does. |
affects
[ "drug-revumenib", "drug-ziftomenib", "drug-bleximenib", "drug-blinatumomab" ]
countermeasures
[
{
"agent": "keep targeting the fusion",
"rationale": "unlike CD19-directed agents, a menin inhibitor targets something the switched clone retains, so the switch does not automatically disarm it",
"tier": "mechanistic argument; one clinical letter reports revumenib plus transplant succeeding against a lineage-switched infant leukaemia"
},
{
"agent": "flow cytometry with a myeloid panel at every on-treatment marrow",
"rationale": "the only assay that detects this; costs nothing and is frequently omitted once a diagnosis of B-ALL is fixed",
"tier": "trial design"
},
{
"agent": "allogeneic transplant",
"rationale": "lineage-agnostic; used in the reported successful case",
"tier": "clinical, anecdotal"
}
]pre emption strategies
[ "prespecify lineage reassessment by flow at fixed timepoints in every paediatric menin-inhibitor trial, and report switches as a named event rather than as progression", "bank material for chromatin profiling at progression, because the mechanism is only visible at that level", "treat menin inhibition as a lineage-switch pressure of the same family as CD19-directed therapy when sequencing the two, rather than assuming only immunotherapy carries that risk" ]
- It defeats the diagnostic pathway, not just the drug. The patient is sequenced, MEN1 is wild type, no resistance mutation is found, and the record shows an unexplained failure — while the answer was visible on a flow panel nobody ran because the diagnosis said B-ALL.
- The drug may be creating the pressure it is meant to relieve. If lineage switching happens at low oncoprotein dose, then partial target engagement — the exact condition of a patient at trough, or on a reduced dose, or with a CYP3A4 interaction — is the condition that favours switching.
- Menin inhibition and CD19-directed therapy may be additive lineage-switch pressures, and paediatric protocols are now combining or sequencing them without any data on that interaction.
- It is invisible to the resistance biology the field is investing in. Base-editor screens, MEN1 resequencing and cell-free DNA panels are all built to catch genetic escape; none of them can see a cell that changed what it is.
- For this project's central hypothesis the implication is sharp: if the leukaemia can keep the fusion and change everything around it, then a perfect drug against the fusion's obligate partner does not close the exit. The interface is the right target and may still not be a sufficient one.
- What is the incidence of lineage switch on menin inhibitors? One profiled sample is not a rate. AUGMENT-101 and the paediatric trials have the flow data to answer this retrospectively and have not reported it.
- Does menin inhibition cause the switch or merely permit it? The dose-dependent oncoprotein model predicts causation; the existence of spontaneous and chemotherapy-associated switches means permissiveness alone would also fit. A serial single-cell experiment in this graph (exp-menin-inhibitor-lineage-destination) is designed to discriminate.
- Is a menin-inhibitor-associated switch reversible on drug withdrawal, as the blinatumomab-associated one was? If it is, stopping the drug becomes a diagnostic manoeuvre with a nine-day readout, which is an unusually cheap experiment on an unusually consequential question.
- Does sequencing CD19-directed therapy before or after a menin inhibitor change the switch rate? Both are lineage-switch pressures and current practice sequences them with no data on their interaction.
- Is the GMP-like destination the same one reached under blinatumomab? If the drug classes drive cells to the same state, one countermeasure covers both; if not, the field has two problems.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Profiling oncoprotein target sites of 36 representative MLL-rearranged leukaemia samples, including three that underwent a lymphoid-to-myeloid lineage-switching event in response to therapy, showed that oncoprotein genomic enrichment is highly variable and dynamically regulated; at high expression the oncoproteins preferentially activate either an ALL program enriched for pro-B-cell genes or an AML program enriched for haematopoietic-stem-cell genes, fusion-partner-specific binding patterns over these gene sets correlate highly with the prevalence of each mutation in ALL versus AML, and in lineage-switching samples oncoprotein levels are reduced and the oncoproteins preferentially activate granulocyte-monocyte progenitor genes; in a sample that lineage switched during treatment with the menin inhibitor revumenib, the oncoprotein and menin were reduced to undetectable levels but ENL persisted on numerous oncoprotein-target loci including genes in the GMP-like lineage-switching program.
MLL oncoprotein levels influence leukemia lineage identities. · Janssens DH, Duran M, Otto DJ, et al. · Nat Commun 2024What this evidence cannot show
Abstract read in full via NCBI efetch; full text not obtained. The revumenib observation rests on ONE sample. The 36-sample cohort is described as broadly representative but its composition (ALL versus AML, paediatric versus adult, partner distribution) was not read. CUT&Tag-style occupancy in primary samples is technically demanding and 'undetectable' is a limit-of-detection statement, not a proof of absence.
Evidence that argues against this (1)
- Paired sequencing across lineage-switch cohorts shows both routes exist — selection of a genetically divergent sister clone in roughly half of informative cases, and reprogramming of the bulk population in the other half — so an individual switch on a menin inhibitor may be subclonal selection rather than drug-induced reprogramming, and a single chromatin profile cannot distinguish them. This graph's own mechanism-lineage-switch-kmt2a entry makes that point and it applies here. Wölfl M et al. 2018
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.preprint
The preprint version of the same work states the conclusion in resistance terms: KMT2A oncoproteins promote lineage-switching events through dynamic chromatin binding and can induce epigenetic lesions, marked by ENL, that support resistance to targeted therapies.
KMT2A oncoproteins induce epigenetic resistance to targeted therapies. · Janssens DH, Duran M, Otto DJ, et al. · bioRxiv 2024What this evidence cannot show
This is the preprint of [E1], not an independent study, and is cited only because its title and closing sentence frame the finding as resistance where the published version frames it as lineage identity. Counting it as a second source would be double counting and this entry does not do so.
Evidence that argues against this (1)
- E1 and E2 are the preprint and published versions of one study. They are not independent support, and this entry's mechanistic claim rests on a single dataset. Blood 2025
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
A girl diagnosed at 5 months with KMT2A::AFF4 pro-B ALL, relapsing after transplant, showed by day 11 of blinatumomab a marrow with 5% lymphoblasts and 80% monocytic cells of M5 AML phenotype that remained at least partially positive for the NG2-directed antibody 7.1; on stopping blinatumomab with no substituted chemotherapy, peripheral blasts cleared within days and a marrow nine days later showed reversion to approximately 40% lymphoblasts with only rare monocytic cells.
Spontaneous reversion of a lineage switch following an initial blinatumomab-induced ALL-to-AML switch in MLL-rearranged infant ALL. · Wölfl M, Rasche M, Eyrich M, et al. · Blood Adv 2018What this evidence cannot show
Single case, and the pressure was blinatumomab, not a menin inhibitor — so this establishes that reversible phenotypic switching occurs in KMT2A-rearranged infant leukaemia, not that it occurs under menin inhibition. Title and PMID verified against NCBI on 2026-09-19; the clinical detail is carried from this graph's existing patient node pt-story-infant-all-kmt2a-aff4-blinatumomab-lineage-switch, which cites the same letter. Title, authors and DOI re-fetched from NCBI on 2026-09-19.
Evidence that argues against this (1)
- This is a single case, and the direction of the literature is that switches are neither rare nor idiosyncratic: in 33 clonally confirmed lineage switches, 79% of patients were in flow-cytometric complete remission of the first leukaemia at the time of the switch, and median survival afterwards was 2.9 months. A case report cannot distinguish an instructive mechanism from a common one. Blood Cancer Journal 2024
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
A 27-year-old man with T-lymphoblastic lymphoma carrying t(6;11)(q27;q23) relapsed six times, and the sixth relapse was monoblastic; marrow cells at the fifth (lymphoid) and sixth (myeloid) relapses carried the same t(6;11)(q27;q23) and the same MLL::MLLT4 (KMT2A::AFDN) fusion transcript, and the fusion sequence in the initial mediastinal cells matched that in peripheral blood at the sixth relapse.
Lineage switch with t(6;11)(q27;q23) from T-cell lymphoblastic lymphoma to acute monoblastic leukemia at relapse. · Higuchi Y, Tokunaga K, Watanabe Y, et al. · Cancer Genet 2016What this evidence cannot show
Single adult case; the pressure was conventional chemotherapy and nelarabine, not a menin inhibitor. Title and PMID verified against NCBI on 2026-09-19; clinical detail carried from this graph's existing patient node pt-story-t-lbl-kmt2a-afdn-lineage-switch-long-survival. Title, authors and DOI re-fetched from NCBI on 2026-09-19.
Evidence that argues against this (1)
- A six-relapse T-lymphoblastic case with a shared t(6;11) is strong evidence of clonal identity across the switch, and weak evidence about menin inhibition, which the patient did not receive. Lineage switch in KMT2A-rearranged disease is well documented after immunotherapy - 64.3% of a 70-case international cohort carried a KMT2A rearrangement - so switching is a property of the disease under several pressures rather than a menin-inhibitor effect. Blood 2025
- E5weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper
A 2026 paediatric letter reports successful management of lineage-switch infant leukaemia with revumenib and haematopoietic cell transplantation.
Successful Management of Lineage Switch Infant Leukemia With Revumenib and Hematopoietic Cell Transplantation. · Granberg RE, Magee JA, Shenoy S, et al. · Pediatr Blood Cancer 2026What this evidence cannot show
This is a letter with no abstract in PubMed and the full text was not obtained, so ONLY the title-level claim is made: revumenib plus transplant was used and the outcome was described as successful. Nothing about the patient's age, fusion partner, direction of switch, whether the switch preceded or followed revumenib, or duration of follow-up is claimed here, because none of it was read. Confidence is deliberately weak.
Evidence that argues against this (1)
- A successful-outcome case letter is the most publication-biased evidence class there is, and the success is confounded: the patient received revumenib AND haematopoietic cell transplantation, so the report cannot attribute the outcome to either. Against it, median survival after lineage switch is 2.9 months in a 33-case series. Blood Cancer Journal 2024
the state-level account of that resistance mechanism: complete target engagement with disease persistence through a cofactor the drug does not touch
the state entered at switch
a menin-inhibitor lead-in before CD19-directed therapy pre-treats towards the escape lineage
- ← related toFlow-cytometric MRD: the fastest readout, and the one a lineage switch walks straight out of
switch documented under revumenib; a phenotype assay cannot track it
a lineage switch under a menin inhibitor requires a lineage-plastic cell to have survived it - the hypothesis restated as an observed clinical event
the target list moves with the lineage the drug selects
the same event under a menin inhibitor rather than chemotherapy
the destination the switch is said to reach is observable in one sample here
- confers resistance to →Revumenib (SNDX-5613, Revuforj) — first approved menin–KMT2A interface inhibitor
a KMT2A-rearranged leukaemia lineage-switched during revumenib treatment with menin and oncoprotein undetectable
- depends on →Lineage switch in KMT2A-rearranged leukemia — the same clone changes what kind of blood cell it is
the underlying phenomenon; this entry adds the menin-inhibitor context
reversion in nine days with no intervening therapy excludes subclonal outgrowth in that case
identical fusion transcript across a T-lymphoid to monoblastic switch
a patient observation consistent with the hypothesis that menin release is not lineage-neutral
serial single-cell CITE-seq under menin inhibition would give the rate and the destination
- related to →Lineage switch as antigen escape from CD19-directed therapy in KMT2A-rearranged leukemia
a second, independent lineage-switch pressure now applied to the same patients, with no data on their interaction
- related to →The differentiation block — what it actually is, operationally, in KMT2A::AFF1 leukaemia
the switch is a change of lineage destination once the block is released, not a return of the block
- ← depends onENL stays after the fusion leaves — an epigenetic lesion that outlives the protein that made it
same sample; the switched state and drug saturation are confounded at n=1
same event, opposite prognostic sign; this hypothesis argues the switch may be transitional rather than escape
entry produced by this enumeration