Hypothesis: a menin inhibitor must come BEFORE an antigen-directed immunotherapy, not with it or after it - and every published experiment has already assumed this without anyone testing it
also called sequencing of menin inhibitor and immunotherapy · menin inhibitor as antigen priming · pre-treatment versus concurrent versus salvage
If a menin inhibitor works by making leukaemia cells easier for the immune system to see, then giving it first and the immune therapy afterwards should work much better than the other way round. Every laboratory experiment has quietly done it that way, and no trial has ever asked.
The same thing, for a scientist
Three independent groups report that menin inhibition makes leukaemia cells more visible to immune effectors - by raising MHC class I [E1], MHC class II and CIITA [E2], and CLEC12A [E3]. **Every one of them ran the experiment as a pre-treatment.** Sparbier: *pre-treating cells with VTP50469 prior to co-culture with OT-I T cells* [E1]. Hogeling: *pre-treated for 4 days with inhibitor, after which cells were washed and activated T cells were added* [E2]. All three therefore encode an ordering assumption - **antigen first, effector second** - in their design, and none states it as a claim or tests the alternatives. The hypothesis here is that the ordering is not incidental but **rate-limiting**, and that it is **antigen-dependent**: for a target the drug *creates* (CLEC12A, MHC), pre-treatment is necessary; for a target already present (CD19, CD22), pre-treatment is at best neutral and may be harmful, because a menin inhibitor drives myeloid differentiation and lineage drift away from the CD19-positive state (`mechanism-lineage-switch-kmt2a`). **The prediction is that the optimal order inverts depending on which antigen is being targeted**, which is testable and would be missed entirely by a trial design that tests 'combination versus monotherapy'.
The counted absence this rests on
(menin inhibitor OR menin inhibitors OR menin inhibition OR revumenib OR ziftomenib OR bleximenib OR SNDX-5613)[tiab] AND (blinatumomab[tiab] OR "CAR T"[tiab] OR "CAR-T"[tiab] OR immunotherapy[tiab]) returned 45 records on 2026-09-20. Title-checking all 45: every one is a review, a narrative overview, a consensus document or an off-topic paper (melanoma BRAF, breast cancer, pancreatic neuroendocrine DNMT1). Not one is a primary study of a menin-inhibitor plus immunotherapy combination, and not one addresses sequence 4.
Rule 72 was applied: singular, plural and nominalisation of 'inhibitor' were all included, four compound codes were used, and the query returned a healthy count rather than a zero - which is the harder case to interpret, because 45 records looks like a literature. It is not. A count is not a result until the titles are read.
The closest thing to a sequencing statement in the whole set is Fetsch's proposal of menin inhibition as maintenance after allogeneic transplant 5 - which is the opposite order from every in vitro experiment, effector first and drug second, and is proposed without a supporting sequencing experiment.
Why the order should matter, mechanistically
For a created antigen, the kinetics are the argument. Antigen induction under a menin inhibitor is reported over 4 days (bleximenib, primary AML) to 7-10 days (VTP50469 and MI-503 titrations) 12. A CAR product or a T-cell engager given on day 0 meets the pre-treatment phenotype. There is no mechanism by which it could induce the antigen it needs faster than the epigenetic drug does.
For an existing antigen, three effects push the other way:
- Lineage drift. Menin inhibition drives myeloid differentiation;
resistance-lineage-switch-under-menin-inhibitionrecords lineage switch as a resistance route to the drug itself. Pre-treating a CD19-positive leukaemia with an agent that promotes myeloid identity before giving a CD19-directed therapy is pre-treating away the target. - Cytoreduction. The drug kills; a smaller antigen-positive population means less CAR expansion, and CAR expansion is antigen-driven.
- Effector exposure. A menin inhibitor acts on T cells as well as blasts - menin-KMT2A is bound at genes encoding negative regulators of T-cell activation, and the drug raises donor T-cell effector function and reduces exhaustion 5. That argues for concurrent exposure, not for a washout.
The three considerations do not have the same sign, which is exactly why the question needs an experiment rather than an opinion.
The falsifiers, stated plainly
This hypothesis is wrong if any of the following is observed:
- Concurrent dosing matches or beats pre-treatment for a created antigen (CLEC12A, MHC class I) at matched total exposure. That would mean the induction is fast enough relative to effector kinetics that order is immaterial.
- Pre-treatment beats concurrent for CD19-directed therapy too. That would mean the effector-side benefit (reduced T-cell exhaustion, raised effector function) dominates lineage drift, and the ordering rule is 'always drug first' rather than antigen-dependent.
- Antigen induction does not decay after washout. The pre-treatment designs all wash the drug out before adding effectors 2. If the induced phenotype is stable for weeks, then the drug can be stopped before immunotherapy and the ordering question collapses into a simple priming schedule.
- The effect is not antigen-dependent at all because the drug's benefit runs entirely through the T cell rather than the blast. E5's cell-extrinsic arm makes this a live possibility and it would make the ordering rule 'concurrent, always'.
| statement | The optimal order of a menin inhibitor and an antigen-directed immunotherapy depends on whether the drug CREATES the antigen or merely modulates a pre-existing one. For created antigens (CLEC12A, MHC class I and II) a menin-inhibitor lead-in of at least 4-7 days will outperform concurrent or reverse-order dosing. For pre-existing lymphoid antigens (CD19, CD22) a lead-in will be neutral or harmful, because menin inhibition promotes myeloid differentiation and lineage drift away from the target. |
|---|---|
| origin | data-pattern |
| prior probability | 0.6 |
| current evidential state | Every published in vitro design is a pre-treatment; no published study compares schedules; the one clinical proposal (maintenance after allo-HCT) is the reverse order and is untested. |
predictions
[ "In a KMT2A-rearranged xenograft, menin-inhibitor lead-in followed by CLEC12A-directed CAR T cells will produce deeper and more durable responses than concurrent dosing at matched total drug exposure.", "In a CD19-positive KMT2A-rearranged model, menin-inhibitor lead-in followed by blinatumomab will be inferior to concurrent dosing, and the difference will be accounted for by CD19 mean fluorescence intensity and by the myeloid fraction at the time the engager is given.", "Induced antigen will decay measurably within days of drug withdrawal, so a lead-in followed by a washout will lose most of the benefit relative to a lead-in followed by continued dosing.", "The cell-extrinsic benefit to T cells requires the drug to be present during effector engagement, so any schedule that stops the drug before infusion will lose it entirely.", "The ordering effect will be larger for CAR products than for T-cell engagers, because CAR expansion is antigen-driven and a bispecific antibody can be re-dosed against a changing antigen level." ]
experiments
[ "exp-mhc-class-i-and-nk-ligand-timecourse-under-menin-inhibition-kmt2a-r", "exp-surface-marker-panel-under-menin-and-dot1l-inhibition" ]
analyses
[ "ana-immune-evasion-routes-documented-in-kmt2a-r-versus-aml-generally" ]
falsifiers
[ "Concurrent dosing matches or beats a lead-in for a created antigen at matched total exposure - the ordering is then immaterial.", "A lead-in beats concurrent dosing for CD19-directed therapy as well - the rule is then 'drug first, always' and is not antigen-dependent.", "Induced antigen persists for weeks after drug withdrawal - the ordering question collapses into a priming schedule with no scheduling constraint.", "The benefit is shown to run entirely through the effector rather than the target - the rule is then 'concurrent, always'.", "A trial that randomises schedule shows no difference in response rate or duration between arms." ]
- A lead-in schedule delays immunotherapy in a disease that relapses within weeks; the time cost is real and is paid by every patient whether or not the antigen benefit materialises.
- If the induced antigen requires continued drug, then any toxicity that forces a dose hold - differentiation syndrome, QTc prolongation - removes the target while an infused CAR product is still circulating.
- Pre-treating a CD19-positive leukaemia with an agent that promotes myeloid identity is pre-treating in the direction of the escape route this graph documents most thoroughly.
- Menin inhibitors act on the graft as well as the leukaemia, so a lead-in that is washed out before effector transfer may be discarding half the mechanism, and a concurrent schedule that keeps it may worsen graft-versus-host disease.
- How fast does menin-inhibitor-induced antigen decay after withdrawal? Nothing in the literature reports a washout timecourse for CLEC12A or for MHC, and the answer determines every schedule.
- Does a menin inhibitor reduce CD19 on KMT2A-rearranged B-ALL blasts? The lineage-drift argument predicts it and this graph's `exp-surface-marker-panel-under-menin-and-dot1l-inhibition` is the designed experiment; the measurement has not been reported.
- Is the effector-side benefit large enough to require concurrent dosing? If menin inhibition improves T-cell function as much as it improves target visibility, the two considerations conflict and only a head-to-head schedule comparison resolves them.
- Should the trial question be 'combination versus monotherapy' at all? A two-by-two of schedule against combination is barely larger and is the only design that can answer the question this hypothesis poses.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In the only in vivo-relevant functional immunology experiment in the menin-inhibitor class I literature, tumour cells were PRE-TREATED with VTP50469 before co-culture with antigen-specific T cells, and this pre-treatment increased tumour cell killing and T-cell interferon-gamma production; inhibitor exposures in the paper were 7-10 days at 0.5-1 micromolar VTP50469 or 500 nanomolar MI-503 before analysis.
Targeting Menin disrupts the KMT2A/B and polycomb balance to paradoxically activate bivalent genes. · Sparbier CE, Gillespie A, Gomez J, Kumari N, Motazedian A, Chan KL, et al., Dawson MA · Nat Cell Biol 2023What this evidence cannot show
The system is murine SCLC-OVA with OT-I T cells, not leukaemia. The pre-treatment design is a methodological choice, and reading an ordering claim out of it is an interpretation this entry makes explicit rather than a claim the authors make. No concurrent or reverse-order arm is reported, so the design cannot support a comparison. Read: full text via Europe PMC fullTextXML (PMC7614190), 2026-09-20.
Evidence that argues against this (1)
- A pre-treatment design is the standard way to run a co-culture killing assay regardless of any hypothesis about sequencing, because it isolates the target-cell effect from any drug effect on the effector. Inferring a therapeutic ordering rule from an assay convention is exactly the sort of inference this graph's brief warns against, and it is recorded here as a hypothesis for that reason. Sparbier CE et al. 2023
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In the leukaemia arm, AML cell lines and primary AML samples were pre-treated with JNJ-75276617 for 4 days, after which cells were washed and anti-CD3/CD28-activated allogeneic T cells or peripheral blood mononuclear cells were added for a further 3 days; 4 of 8 primary AML samples became more sensitive to allogeneic T cells after this pre-treatment, and in the autologous setting sorted blasts cultured with or without their own T cells showed enhanced T-cell cytotoxicity under the inhibitor, coinciding with increased MHC class I and class II expression.
Bleximenib, the novel menin-KMT2A inhibitor JNJ-75276617, impairs long-term proliferation and immune evasion in acute myeloid leukemia. · Hogeling SM, Le DM, La Rose N, Kwon MC, Wierenga ATJ, Van den Heuvel FAJ, et al., Schuringa JJ · Haematologica 2025What this evidence cannot show
The washout in the allogeneic arm removes the drug before the effectors arrive, so that arm measures a target-cell-intrinsic priming effect only; the autologous arm keeps the drug present and is therefore a concurrent design, and the two are not compared. The four responders were NPM1c/DNMT3A-mutant. Read: full text (PMC12130779), 2026-09-20.
Evidence that argues against this (1)
- The paper contains both designs - washed pre-treatment in the allogeneic arm and continuous exposure in the autologous arm - and reports enhanced killing in both, which is weak evidence that the order is NOT critical. The two arms differ in too many other ways (donor versus autologous T cells, different patients, different ratios) to serve as a sequencing comparison. Hogeling SM et al. 2025
- E3weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.preprint
A 2026 preprint reports that menin inhibition induces robust and uniform CLEC12A expression in KMT2A-rearranged and NPM1-mutated AML and that CLEC12A-directed CAR T cells combined with menin inhibition outperform either monotherapy in vitro and in xenograft models, with the menin inhibitor's role being to create the antigen.
Menin-Inhibition Sensitizes Acute Myeloid Leukemia to CLEC12A-Directed CAR Cell Therapy · Rausch et al. · bioRxiv 2026What this evidence cannot show
Preprint, not peer reviewed. The identifier is taken from this graph's `mechanism-menin-inhibition-induces-clec12a`, which owns this source; this beat did not independently verify the preprint's contents beyond the identifier. The word 'combined' in the abstract does not specify a schedule, so whether the reported combination is concurrent or sequential is not established here and should be read from the preprint's methods before this entry is relied on.
Evidence that argues against this (1)
- A target-creation claim is the strongest case for ordering and also the case with the sharpest failure mode: if the antigen is induced only while the drug is present, then stopping the menin inhibitor - for toxicity, for differentiation syndrome, for a QTc signal - removes the target mid-therapy, and a CAR product cannot be un-infused. Daver N et al. 2026
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
As of 2026-09-20, a PubMed query pairing menin-inhibitor names in singular, plural and nominalised forms with blinatumomab, CAR T or immunotherapy returns 45 records, of which title-checking finds none to be a primary study of a menin-inhibitor plus immunotherapy combination and none to address the sequence or schedule of such a combination; the set is composed of reviews, narrative overviews, consensus documents and off-topic records.
Q54 antigen-presentation beat: bounded counted-absence sweep, PubMed E-utilities esearch with querytranslation read back -- menin inhibitor plus immunotherapy, 45 records, all 45 titles checked · Q54-immune, Oncofusion Knowledge Graph · Oncofusion Knowledge Graph 2026What this evidence cannot show
Bounded absence, and the harder kind: the count is not zero, it is 45, which reads like a literature until the titles are examined. Name list: menin inhibitor / inhibitors / inhibition, revumenib, ziftomenib, bleximenib, SNDX-5613. NOT searched: enzomenib, icovamenib, JNJ-75276617 as a code, VTP-50469, DS-1594, DSP-5336, or the phrase 'menin-MLL'. Immunotherapy terms: blinatumomab, CAR T, CAR-T, immunotherapy - not inotuzumab, not gemtuzumab, not 'bispecific', not 'T-cell engager'. Titles only, not abstracts; a sequencing discussion buried in a review would not be visible.
Evidence that argues against this (1)
- Two preclinical combination studies exist and are cited elsewhere in this entry - the CLEC12A CAR preprint and the bleximenib T-cell cytotoxicity arm - and neither is retrieved by this query, the first because it is a preprint with a non-standard DOI prefix and the second because its abstract does not use any of the immunotherapy terms. The absence is about the sequencing question, which genuinely is unaddressed; it is not an absence of combination work. Hogeling SM et al. 2025
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Menin inhibition directly promoted donor T-cell effector functions with increased TNF-alpha, interferon-gamma, perforin and granzyme A and B production and cytolytic activity and reduced T-cell exhaustion, with menin-KMT2A binding found at genes encoding negative regulators of T-cell activation, and the authors propose menin inhibition as maintenance after allogeneic haematopoietic cell transplantation.
Menin inhibition enhances graft-versus-leukemia effects by T-cell activation and endogenous retrovirus induction in AML. · Fetsch V, Schwobel L, Ozyerli-Goknar E, Stell AV, Punta M, Plenge T, et al., Zeiser R · Blood 2026What this evidence cannot show
AML, murine and xenograft models, abstract only. Cited here because it is the single piece of evidence pushing towards concurrent rather than sequential dosing - the drug acting on the effector requires the effector to be present while the drug is - and because the proposed clinical setting (maintenance after transplant) is the reverse order from every in vitro experiment. This graph's entry for this paper is `cell-state-interferon-herv-mhc2`.
Evidence that argues against this (1)
- A drug that acts on the effector as well as the target argues against a washout, and therefore against the pre-treatment ordering this hypothesis proposes. If the cell-extrinsic arm is the larger effect, the correct schedule is continuous concurrent exposure and the whole antigen-dependent ordering rule is wrong. Fetsch V et al. 2026
the ordering rule only matters if the drug really creates the antigen
the clearest target-creation case, and therefore the clearest case for a lead-in
- related to →Blinatumomab in infant KMT2A-rearranged ALL — the largest single outcome gain in thirty years
the pre-existing-antigen case, where a lead-in is predicted to be neutral or harmful
orthogonality and ordering are different questions about the same combination
a menin-inhibitor lead-in before CD19-directed therapy pre-treats towards the escape lineage
the washout arm supplies the decay constant the ordering hypothesis needs