Menin inhibition restores graft-versus-leukaemia only where MHC class II is limiting - so the post-transplant maintenance rationale may hold in myeloid disease and fail in B-precursor infant ALL
also called MHC-II headroom hypothesis · menin inhibitor GVL lineage restriction · post-transplant menin maintenance mechanism · HERV-MHC-II axis lineage dependence
A 2026 study showed that menin inhibitors make leukaemia cells visible to donor immune cells by switching a display molecule back on. But that molecule is already switched on in the type of cell that infant leukaemia comes from, so the same drug may do nothing for visibility in babies - even though several trials are already testing it after transplant.
The same thing, for a scientist
Fetsch and colleagues report that menin inhibition derepresses human endogenous retroviruses, triggers interferon-stimulated gene expression, induces CIITA and MHC class II, and thereby sensitises KMT2A-rearranged and NPM1-mutated AML cells to donor T-cell killing after allogeneic HCT, enhancing graft-versus-leukaemia in mouse and human xenograft models; it also directly augments donor T-cell effector function and reduces exhaustion [E1]. The mechanism has two separable arms - an antigen-presentation arm on the leukaemia cell and a T-cell-intrinsic arm on the donor graft. **MHC class II is constitutively expressed in the B lineage**, and this graph's index disease, KMT2A::AFF1 infant ALL, is B-precursor. The hypothesis is that the antigen-presentation arm has little or no headroom in B-precursor disease, and that any post-transplant benefit of menin inhibition in infant ALL must therefore come from the T-cell-intrinsic arm alone. This matters because five registered post-transplant menin-maintenance studies exist in this graph's `concept-hsct-in-kmt2a-r-leukemia` and none was designed around either arm, and because graft-versus-leukaemia has never been demonstrated in KMT2A-rearranged infant leukaemia at all [E2].
The two arms, and why they separate1 deeper
| arm | acts on | mechanism | predicted headroom in B-precursor infant ALL |
|---|---|---|---|
| antigen presentation | the leukaemia cell | HERV derepression -> interferon-stimulated genes -> CIITA -> MHC class II | low: MHC-II is constitutive in B lineage |
| T-cell intrinsic | donor T cells | reduced exhaustion; increased TNF-alpha, IFN-gamma, perforin, granzyme A/B; menin-KMT2A displaced from genes encoding negative regulators of T-cell activation | unchanged: the donor T cell is not a leukaemia cell and its lineage is the same regardless of the recipient's disease |
The second arm is lineage-independent by construction. The first is not.
This is falsifiable and cheap to test. Measure surface MHC class II on primary KMT2A::AFF1 infant ALL blasts before and after menin-inhibitor exposure. If baseline expression is already high and does not rise, the antigen-presentation arm has no headroom in this disease.
Why the answer is not obvious in advance
Three reasons the hypothesis could be wrong, and they are the reasons it is worth testing rather than assuming:
- KMT2A-rearranged infant ALL is not a normal B cell. Its transcriptional programme is stem-like and lineage-promiscuous, and this graph already records lineage switch to myeloid under CD19-directed therapy under
risk-lineage-switch-cd19-directed-therapy. A blast that can become myeloid may not have myeloid-normal or B-normal MHC-II regulation. - MHC class II is frequently DOWNREGULATED at relapse after transplant across leukaemias, as an immune-escape mechanism. If that happens in relapsed infant ALL, the headroom is created by the relapse itself, and the drug would be restoring something that was lost rather than inducing something absent.
- The HERV/interferon arm may do more than MHC-II. Interferon-stimulated gene induction changes antigen processing, MHC class I, and chemokine secretion, none of which is B-lineage-constitutive.
So the hypothesis is specifically that the CIITA/MHC-II step is where the lineage restriction sits, not that menin inhibition is inert in B-lineage disease.
What this predicts, and what would kill it
Predictions if true:
- Baseline surface MHC class II on primary KMT2A::AFF1 infant ALL blasts is high and does not increase meaningfully under revumenib or an equivalent menin inhibitor.
- HERV transcript induction and interferon-stimulated gene induction still occur in those blasts (the upstream steps are chromatin-level and lineage-independent), while CIITA and MHC-II do not rise.
- In a B-precursor xenograft GVL model, menin inhibition still enhances donor T-cell killing, but by an amount attributable to the T-cell arm, and an MHC-II-blocking antibody does not abolish it - whereas in the myeloid model it substantially does.
- Relapsed post-transplant infant ALL blasts DO show MHC-II downregulation relative to diagnosis, creating headroom that diagnostic samples lack.
Falsifiers:
- Baseline MHC class II is low on primary KMT2A::AFF1 infant ALL blasts and rises under menin inhibition. The hypothesis is simply wrong and the Fetsch mechanism transfers.
- An MHC-II-blocking antibody abolishes the menin-inhibitor GVL enhancement in a B-precursor model as completely as in the myeloid one.
- HERV and interferon-stimulated gene induction fail to occur in B-precursor blasts, which would mean the mechanism fails upstream of MHC-II and the lineage argument is the wrong explanation for the right prediction.
| statement | The graft-versus-leukaemia enhancement produced by menin inhibition has two separable arms - an MHC class II antigen-presentation arm on the leukaemia cell and a lineage-independent effector arm on donor T cells. Because MHC class II is constitutively expressed in the B lineage, the antigen-presentation arm has little headroom in B-precursor KMT2A::AFF1 infant ALL, and any post-transplant benefit in that disease must come from the T-cell-intrinsic arm alone. |
|---|---|
| origin | literature-gap |
| prior probability | 0.4 |
| why it matters | Five registered post-transplant menin-maintenance studies exist, none designed around either arm of this mechanism, and the only randomised one excludes children entirely. If the antigen-presentation arm does not operate in B-precursor disease, the expected effect size in infant ALL is smaller than the AML data imply and the trials are powered on the wrong assumption. |
predictions
[ "Baseline surface MHC class II on primary KMT2A::AFF1 infant ALL blasts is high and does not rise meaningfully under a menin inhibitor", "HERV transcript and interferon-stimulated gene induction still occur in those blasts, while CIITA and MHC-II do not rise", "In a B-precursor xenograft GVL model, an MHC-II-blocking antibody does not abolish the menin-inhibitor GVL enhancement, whereas in a myeloid model it substantially does", "Post-transplant relapsed infant ALL blasts show MHC-II downregulation relative to diagnosis, creating headroom that diagnostic samples lack" ]
experiments
[ "exp-mhc-ii-headroom-under-menin-inhibition-in-b-precursor-infant-all" ]
analyses
[]
falsifiers
[ "Baseline MHC class II is LOW on primary KMT2A::AFF1 infant ALL blasts and rises under menin inhibition - the hypothesis is wrong and the Fetsch mechanism transfers intact", "An MHC-II-blocking antibody abolishes the menin-inhibitor GVL enhancement in a B-precursor model as completely as in the myeloid model", "HERV and interferon-stimulated gene induction fail to occur in B-precursor blasts at all, which would mean the mechanism fails upstream of MHC-II and the lineage argument is the wrong explanation for the right prediction", "A KMT2A-rearranged subgroup analysis of post-transplant menin maintenance shows equal relapse reduction in lymphoid and myeloid disease" ]
reasons it could be wrong
[ "KMT2A-rearranged infant ALL is lineage-promiscuous and stem-like, and this graph records lineage switch to myeloid under CD19-directed therapy - a blast that can become myeloid may not have B-normal MHC-II regulation", "MHC class II is frequently downregulated at post-transplant relapse across leukaemias as an immune-escape mechanism, which would create headroom that diagnostic samples lack", "The HERV/interferon arm induces more than MHC-II, including antigen processing machinery, MHC class I and chemokines, none of which is B-lineage-constitutive" ]
- If the antigen-presentation arm does not operate in B-precursor disease, five registered maintenance studies are powered on an effect size borrowed from the wrong lineage, and a null result would be read as the drug failing rather than as the mechanism not applying.
- Enhancing graft-versus-leukaemia means enhancing alloreactivity, and the CIBMTR data show alloreactivity is net beneficial only at grade I-II acute GVHD. A drug that makes donor T cells more cytotoxic and less exhausted in a transplanted infant could push the GVHD grade past the point where the relapse protection is abrogated.
- Menin is required for long-term haematopoietic stem cell function under transplantation stress, so a menin inhibitor given while the donor graft is establishing itself targets the function the graft depends on - and the more effective the GVL enhancement, the longer the exposure that is wanted.
- Derepressing endogenous retroviruses to raise interferon signalling is an intervention with no long-term paediatric safety data, in a patient whose immune system is being assembled for the first time under immunosuppression.
- The hypothesis is falsifiable and cheap, which cuts both ways: if nobody runs the flow cytometry experiment before the trials read out, a negative result in infant ALL will be uninterpretable between 'the drug does not work' and 'this arm of the mechanism was never available'.
- What is the baseline surface MHC class II level on primary KMT2A::AFF1 infant ALL blasts, and does it change under menin inhibition? This is a flow cytometry experiment on samples that already exist in biobanks.
- Is MHC class II downregulated at post-transplant relapse in infant ALL? If so the headroom is created by relapse and maintenance started before relapse would be acting on the wrong cell state.
- Does the T-cell-intrinsic arm alone produce a measurable relapse reduction? If it does, the lineage argument is academic and the maintenance rationale survives regardless.
- Do menin inhibitors impair donor haematopoietic reconstitution? Menin is required for long-term HSC function under transplantation stress in mice, which is exactly the stress a fresh graft is under - so the same drug may enhance GVL and impair engraftment simultaneously.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Menin inhibition induced CIITA and MHC class II expression in KMT2A-rearranged and NPM1-mutated AML cells in vitro and in vivo via human endogenous retrovirus derepression and interferon-stimulated gene upregulation, sensitised those cells to T-cell-mediated elimination after allogeneic HCT in mice, enhanced the graft-versus-leukaemia effect in human xenograft models, and separately increased donor T-cell TNF-alpha, interferon-gamma, perforin and granzyme A/B production and cytolytic activity while reducing T-cell exhaustion, with menin-KMT2A binding to genes encoding negative regulators of T-cell activation reduced by menin inhibition.
Menin inhibition enhances graft-versus-leukemia effects by T-cell activation and endogenous retrovirus induction in AML. · Fetsch V, Schwoebel L, Ozyerli-Goknar E, Stell AV, Punta M, Plenge T, Klaus T, Gupta MK, et al. · Blood 2026What this evidence cannot show
AML only - both the KMT2A-rearranged and NPM1-mutated models are myeloid, and no lymphoid model is reported in the abstract. Mouse allogeneic HCT and human xenograft systems, where the alloreactive repertoire and MHC context are not those of a matched paediatric transplant; xenograft GVL models use human T cells in immunodeficient mice. The paper's stated conclusion is a rationale for a clinical trial, not a clinical result. No PMC record was available at retrieval on 2026-09-20, so only the abstract was read and the specific menin inhibitor, doses and MHC-II magnitudes are not asserted here.
Evidence that argues against this (1)
- The graph's index disease is B-precursor, where MHC class II is constitutively expressed, so the antigen-presentation arm of this mechanism may have no headroom in KMT2A::AFF1 infant ALL. That is the entire content of this hypothesis, and it means the strongest published mechanistic support for post-transplant menin maintenance was generated in the lineage where the mechanism has the most room to operate and may not transfer to the one this graph is about. Inukai T et al. 2006
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
MLL-rearranged infant ALL and AML cells are highly resistant to TRAIL - in ALL primarily through low cell-surface death-receptor expression and in AML through decoy-receptor expression - and nearly half of cell lines and most primary cells are insensitive to FasL; the authors state that a graft-versus-leukaemia effect after allogeneic transplant has not been clearly evidenced in these leukaemias.
Resistance of infant leukemia with MLL rearrangement to tumor necrosis factor-related apoptosis-inducing ligand: a possible mechanism for poor sensitivity to antitumor immunity. · Inukai T, Zhang X, Goto M, Hirose K, Uno K, Akahane K, Nemoto A, Goi K, et al. · Leukemia 2006What this evidence cannot show
Entirely in vitro, testing recombinant TRAIL and FasL sensitivity and not perforin/granzyme killing, which is a dominant in vivo cytotoxic route and is the one the 2026 menin study measures. No clinical correlation. Published 2006, before the modern understanding of MHC class II regulation in KMT2A-rearranged disease. Notably, the resistance mechanism reported here is at the level of DEATH RECEPTORS, a different step from the antigen-presentation step this hypothesis concerns - so the two arguments for impaired GVL in this disease are independent and could both be true, or one could be true and the other not.
Evidence that argues against this (1)
- The 2026 study measures exactly the effector functions this one omits and finds them enhanced by menin inhibition: increased donor T-cell perforin and granzyme A/B production and cytolytic activity, with reduced exhaustion. Death-ligand resistance in 2006 assays is fully compatible with an intact and pharmacologically augmentable cytolytic route, so the older paper does not establish that GVL cannot work here. Fetsch V et al. 2026
the class II headroom hypothesis has a class I counterpart with a different mechanism and a different limiting factor
partial: concord on the conclusion, contest on the argument - a constitutively class II-positive myeloid cell still had headroom under the drug, so B-lineage HLA-DR positivity does not by itself establish a ceiling
CIITA-null MV4-11 has lower baseline HLA-DR and no drug response - baseline CIITA level sets the achievable induction
proposes a lineage restriction on the one published route to restoring GVL in this disease
the agent in four of the five registered post-transplant maintenance studies
supplies the mechanism the five post-transplant menin maintenance studies were launched without, and a reason it may not transfer to B-precursor disease
- related to →Lineage switch as antigen escape from CD19-directed therapy in KMT2A-rearranged leukemia
a lineage-promiscuous blast may not have B-normal MHC class II regulation, which is the main reason the hypothesis could be wrong
flow cytometry for surface MHC class II at baseline and under menin inhibition, with interferon-gamma as an in-sample positive control
the HERV/interferon/MHC-II state this hypothesis argues has no headroom in B-precursor disease
the transactivator step where this hypothesis predicts the lineage restriction sits