In this fusion family, MHC class I is not silenced - it is elevated and load-bearing, holding NK cells off the leukaemia stem cell
also called missing-self in KMT2A-rearranged leukaemia · inhibitory KIR ligands on MLL-rearranged blasts · H2-K1 dependency in MLL-AF9 · NK resistance of KMT2A-rearranged leukaemia
Killer T cells need to see a cell's identity card; a different killer cell, the NK cell, attacks cells that are missing it. Leukaemia driven by these fusions keeps its identity cards up and its NK-activating flags down, which is a good way to hide from NK cells - and in mice, removing the identity card cripples the leukaemia.
The same thing, for a scientist
The intuitive picture of MHC class I in cancer is silencing for T-cell escape. **In this fusion family the best-evidenced picture is the opposite.** Three-quarters of what is documented about class I in KMT2A-rearranged and MLL-fusion leukaemia is about NK cells, not T cells. Chan et al. tested three KMT2A-rearranged lines - **RS4;11, SEM and MV4-11**, all of which have entries in this graph - plus primary blasts, and found all three **resistant to NK lysis**, carrying **inhibitory KIR ligands** and protease inhibitor-9 while expressing **low NKG2D activating ligands and low adhesion molecules**; adding a pan-MHC class I blocking antibody, which breaks the inhibitory KIR-HLA interaction, **further augmented degranulation across all three KIR2DL1, KIR2DL2/3 and KIR3DL1 NK subsets** [E1]. Ghosh et al. then showed the dependency directly: **H2-K1, a classical MHC class I molecule and the murine orthologue of HLA-A, scored among the top in vivo dependencies of MLL-AF9 leukaemia-initiating cells**, its knockout depleted leukaemia from marrow and spleen and extended survival, and **NK-cell depletion rescued the effect entirely while macrophage depletion did not** [E2]. Nakata et al. supply the immunosurveillance context: MLL/ENL leukaemia cells are rejected by NK cells in the absence of adaptive immunity, and NK-resistant leukaemia cells are **selected during leukaemia development** [E3].
The phenotype, as three independent groups describe it
Inhibitory side up. Inhibitory KIR ligands present on all three KMT2A-rearranged lines tested and on primary blasts; blocking them with a pan-class I antibody increases NK degranulation in every KIR subset 1. In the mouse, class I is elevated on c-Kit+ MLL-AF9 leukaemia cells relative to healthy marrow counterparts, and HLA-A is elevated on human AML blasts relative to normal HSCs 2.
Activating side down. Low NKG2D activating ligands and low adhesion molecules on the same lines 1. In the MLL/ENL model, upregulation of NKG2D ligands is what causes NK-mediated elimination - so the baseline low state is the escape state 3.
Plus a granzyme brake. Protease inhibitor-9 (serpin B9) on the same lines 1, which neutralises granzyme B inside the target cell and therefore defeats both NK and T-cell killing downstream of recognition.
That is a coherent, three-layer NK-evasion phenotype documented in the index fusion family, and it is the clearest example of an immune-evasion route that is specific to this disease rather than imported from AML in general.
The clinical counterpart already exists
This is not only preclinical. A child with KMT2A-rearranged infant ALL who relapsed in marrow and CNS eighteen months after HLA-matched cord blood transplant was retransplanted with KIR-ligand-mismatched cord blood under reduced-intensity conditioning and remained in remission at 52 months 4. One patient, one report, and the authors present it as a treatment option rather than a proof - but it is the missing-self logic applied to this exact disease with a durable outcome.
The same logic runs in the opposite direction through the transplant literature in general: KIR-mismatched donors reduce relapse risk in AML, which is why haploidentical transplantation exists as a deliberate strategy 2.
And the newest entrant reads the same way from a third angle: bispecific CD19-CD133 CAR-iNKT cells eradicate medullary and leptomeningeal KMT2A-rearranged leukaemia, and the mechanism the authors identify for their superiority over CAR T cells is CAR-dependent upregulation of NKG2D on the effector and engagement of NKG2D ligands on the KMT2Ar-ALL cells 5. The activating axis that Chan found low and Nakata found decisive is the axis being engineered around.
Where the evidence is thin, named
- Lineage. The H2-K1 dependency is MLL-AF9, myeloid. Chan's lines are B-lymphoid (RS4;11, SEM) and myelomonocytic (MV4-11), which is better, but the functional in vivo dependency has never been tested in a lymphoid model.
- Species and receptor family. Ly49 and KIR are convergently evolved and structurally unrelated; the mouse result transfers as a class-level claim about inhibitory receptors reading class I, not as an HLA-A claim.
- Dose. Nothing in this literature tests whether raising class I further worsens NK control. Every experiment removes or blocks class I. Whether the brake is already saturated is the single most decision-relevant unknown in this beat.
- Age. NK biology in infants differs from adults - NK cells are the dominant lymphocyte subset in early life and KIR repertoires are still maturing - and none of this work is in infant material.
| the unmeasured quantity | whether the inhibitory brake is saturated at baseline class I levels - i.e. whether RAISING class I further worsens NK control. Every published experiment removes or blocks class I; none raises it. |
|---|
phenotype layers
{
"inhibitory_ligands": "inhibitory KIR ligands present on RS4;11, SEM, MV4-11 and primary blasts (PMID 23014531); class I elevated on MLL-AF9 c-Kit+ cells and on human AML blasts versus normal HSC (PMID 39844759)",
"activating_ligands": "low NKG2D ligands and low adhesion molecules (PMID 23014531); NKG2D-ligand engagement nevertheless decisive for CAR-iNKT killing of KMT2Ar-ALL (PMID 40898981); NKG2D-ligand upregulation causes NK elimination in MLL-ENL (PMID 24336127)",
"downstream_brake": "protease inhibitor-9 / serpin B9 on the same lines (PMID 23014531)"
}functional dependency
{
"gene": "H2-k1 (murine class I, orthologue of HLA-A)",
"model": "MLL-AF9 leukaemia-initiating cells",
"effect_of_loss": "no ex vivo growth defect; strong in vivo depletion in marrow and spleen; extended survival",
"rescued_by": "NK-cell depletion (anti-NK1.1)",
"not_rescued_by": "macrophage depletion (clodronate)",
"pmid": "39844759",
"title": "H2-K1 protects murine MLL-AF9 leukemia stem cells from natural killer cell-mediated immune surveillance.",
"authors": "Ghosh S, Rodriguez-Zabala M, Dushime GT, et al.",
"venue": "Haematologica",
"year": 2025,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
}clinical counterpart
{
"case": "KIR-ligand-mismatched cord blood retransplant in relapsed infant KMT2A-r ALL, remission at 52 months",
"pmid": "36706271",
"n": 1,
"title": "Successful Retransplantation With Killer Cell Immunoglobulin-like Receptor Ligand-mismatched Cord Blood in Infant Acute Lymphoblastic Leukemia That Relapsed After Transplantation.",
"authors": "Mizuki K, Honda Y, Asai H, et al.",
"venue": "J Pediatr Hematol Oncol",
"year": 2023,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
}lines named
[ "line-rs4-11-kmt2a-aff1-adult", "line-sem-kmt2a-aff1", "line-mv4-11-kmt2a-aff1-myeloid" ]
- Any intervention that raises MHC class I - a menin inhibitor, an EZH2 inhibitor, interferon - is strengthening the inhibitory ligand for KIR and Ly49 at the same time as it strengthens the T-cell target, and in this fusion family the class I molecule is a demonstrated in vivo dependency.
- NK-resistant variants are selected during leukaemia development in an immunocompetent host, so a diagnosed leukaemia has already passed this filter and NK-directed approaches are attacking a solved problem unless the escape is reversed.
- Protease inhibitor-9 neutralises granzyme B downstream of recognition and therefore defeats NK and T-cell killing alike; raising antigen presentation does nothing about it, and nobody has measured it in this disease since 2012.
- Haploidentical transplantation with KIR mismatch is a deliberate missing-self strategy, and giving a class I-raising drug as post-transplant maintenance - the setting proposed for menin inhibitors - works directly against it. The two most-discussed immunological uses of menin inhibition in this disease are in mechanistic conflict and nobody has said so.
- Is the inhibitory brake saturated? Titrate class I on a KMT2A-rearranged line with interferon or a menin inhibitor and measure NK degranulation across the dose range. If degranulation is already at floor at baseline, raising class I costs nothing on the NK side and the menin-inhibitor immune rationale is clean.
- Does the H2-K1 dependency hold in a KMT2A::AFF1 lymphoid model? Every functional class I result in this fusion family is myeloid, and the index disease is B-lymphoid.
- What are the actual NKG2D ligand levels on KMT2A-rearranged blasts, by modern methods? Two papers give opposite impressions fourteen years apart and the CAR-iNKT strategy depends on the answer.
- How many KMT2A-rearranged children have had KIR-ligand-mismatched grafts, and what happened? A registry query, not an experiment, and it would convert a case report into evidence.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Three MLL-rearranged leukaemia cell lines (RS4;11, SEM and MV4-11) and primary leukaemia blasts were all resistant to NK-cell lysis, had some inhibitory KIR ligands and protease inhibitor-9, and expressed low levels of NKG2D activating ligands and adhesion molecules; treatment with an anti-CD19 (XmAb5574) or anti-CD33 (lintuzumab) antibody made them efficiently killed by NK cells, and adding a pan-MHC class I antibody to block the inhibitory KIR-HLA interaction further augmented degranulation in all of the KIR2DL1, KIR2DL2/3 and KIR3DL1 NK subsets by missing-self recognition; a mouse model showed slower leukaemia progression and longer survival after antibody treatment.
Antibody-dependent cell-mediated cytotoxicity overcomes NK cell resistance in MLL-rearranged leukemia expressing inhibitory KIR ligands but not activating ligands. · Chan WK, Kung Sutherland M, Li Y, Zalevsky J, Schell S, Leung W · Clin Cancer Res 2012What this evidence cannot show
Cell lines plus primary blasts, in vitro degranulation and killing assays with a xenograft survival arm; the three lines are exactly the three KMT2A::AFF1 lines this graph already carries entries for, so the result is narrow in genotype (all t(4;11)) even though it is broad in lineage. 'Some inhibitory KIR ligands' is the authors' own hedge and the HLA typing of the lines is not resolved into KIR ligand groups in the abstract. 2012, and this beat found no replication with modern reagents. Read: abstract only; identifiers fetched 2026-09-20.
Evidence that argues against this (1)
- The paper's own solution is MHC-independent: an antibody against a leukaemia-associated antigen restored NK killing regardless of the class I state, and class I blockade was an add-on. That argues the class I brake is real but not the rate-limiting barrier, and it points to ADCC and CAR approaches rather than to manipulating class I. Chan WK et al. 2012
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
H2-k1, a classical MHC class I molecule described by the authors as the murine orthologue of human HLA-A, scored among the top in vivo dependencies in a CRISPR dropout screen of 961 surface-receptor genes in MLL-AF9 leukaemia-initiating cells; H2-K1 was markedly elevated on c-Kit+ MLL-AF9 leukaemia cells versus healthy bone marrow counterparts and HLA-A was elevated on AML patient blasts versus normal haematopoietic stem cells; H2-k1 disruption did not affect ex vivo growth but strongly depleted leukaemia in bone marrow and spleen and extended survival in vivo; anti-NK1.1 NK-cell depletion accelerated leukaemia and rescued the antileukaemic effect of H2-k1 knockdown, whereas clodronate macrophage depletion did not.
H2-K1 protects murine MLL-AF9 leukemia stem cells from natural killer cell-mediated immune surveillance. · Ghosh S, Rodriguez-Zabala M, Dushime GT, Reinbach K, Ramakrishnan R, Sitnicka E, Jaras M · Haematologica 2025What this evidence cannot show
Murine MLL-AF9 only; the human arm is an expression comparison with no functional test. Published as a Haematologica letter with an empty PubMed abstract field, so all claims come from the full text retrieved via the PMC HTML page (PMC12130766) - the route-4 case in the brief. Ly49 and KIR are convergently evolved inhibitory receptor families, so the transfer to human patients is at the class level. The experiment removes class I; it does not raise it, and says nothing about whether the brake is saturated. Read: full text, 2026-09-20.
Evidence that argues against this (1)
- An in vivo dependency in an immunocompetent mouse is a statement about that mouse's NK compartment. Patients receiving menin inhibitors are frequently lymphopenic, post-chemotherapy or post-transplant on calcineurin inhibitors, and NK function in AML is documented as impaired - so the brake may be pressing against a pedal that is not there. Ghosh S et al. 2025
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
MLL/ENL leukaemia cells transferred into immunocompetent mice without preconditioning did not produce leukaemia, whereas all Rag2-/-gc-/- mice lacking both adaptive cells and NK cells developed leukaemia and 60% of Rag2-/- mice retaining NK cells rejected the leukaemia; NK depletion enhanced engraftment and NK transfer inhibited it; upregulation of NKG2D ligands on MLL/ENL leukaemia cells caused their elimination by NK cells; and leukaemia cells resistant to NK elimination were selected during leukaemia development in Rag2-/- recipients.
In vivo eradication of MLL/ENL leukemia cells by NK cells in the absence of adaptive immunity. · Nakata J, Nakano K, Okumura A, Mizutani Y, Kinoshita H, Iwai M, et al. · Leukemia 2014What this evidence cannot show
Murine MLL-ENL, myeloid, with transfer into non-irradiated hosts - an unusually clean immunosurveillance design and an unusually stringent one, since no conditioning is given. The selection result is the important one for this beat and is reported without molecular characterisation of what was selected. This graph's `import-class-murine-mll-af9-myeloid` records the standing limits of transferring murine MLL-fusion myeloid results to infant human B-ALL. Read: abstract only; identifiers fetched 2026-09-20.
Evidence that argues against this (1)
- The selection of NK-resistant variants during leukaemia development means the NK axis is already a solved problem for an established leukaemia by the time it is diagnosed. Interventions built on NK recognition may therefore be interventions against a barrier the disease crossed long ago, which would explain why NK-directed approaches in AML have been much less successful than the preclinical immunosurveillance literature predicts. Nakata J et al. 2014
- E4weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper
A patient with KMT2A-rearranged infant acute lymphoblastic leukaemia who relapsed in bone marrow and central nervous system eighteen months after HLA-matched cord blood transplantation was retransplanted, after achieving second remission, with killer-cell immunoglobulin-like receptor ligand-mismatched cord blood under reduced-intensity conditioning and remained in remission at 52 months.
Successful Retransplantation With Killer Cell Immunoglobulin-like Receptor Ligand-mismatched Cord Blood in Infant Acute Lymphoblastic Leukemia That Relapsed After Transplantation. · Mizuki K, Honda Y, Asai H, Higuchi N, Morita H, Yabe H, Kusuhara K · J Pediatr Hematol Oncol 2023What this evidence cannot show
A single case report; n=1, no control, and the second transplant differs from the first in conditioning intensity and graft as well as in KIR-ligand matching, so the outcome cannot be attributed to missing-self recognition. It is recorded because it is the only clinical application of the missing-self logic in this exact disease that this beat could find, and because a case report is the correct confidence tier for a single durable remission. Read: abstract only; identifiers fetched 2026-09-20.
Evidence that argues against this (1)
- Publication bias is total for single-case transplant reports: a KIR-ligand-mismatched retransplant that failed would not be published. The relevant denominator - how many KMT2A-rearranged children have received KIR-mismatched grafts and with what outcome - is a registry question that nobody in this beat's searches has answered. Mizuki K et al. 2023
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Bispecific CD19-CD133 CAR-iNKT cells eradicated medullary and leptomeningeal KMT2A-rearranged ALL in vitro and in vivo with sustained remissions and no discernible haematologic toxicity, outperforming monospecific counterparts and dual bispecific CAR T cells; mechanistically the greater potency of CAR-iNKT over CAR T cells was mediated by pronounced CAR-dependent and CAR-antigen-dependent upregulation of the activating receptor NKG2D on the CAR-iNKT cells and its engagement by NKG2D ligands on the KMT2Ar-ALL cells, allowing effective targeting even when CD133 or CD19 is downregulated.
Off-the-shelf dual CAR-iNKT cell immunotherapy eradicates medullary and leptomeningeal high-risk KMT2A-rearranged leukemia. · Ren H, Elliott N, Lye B, Sharif Shohan MU, Cross JW, Field L, et al. · Blood 2026What this evidence cannot show
Preclinical: in vitro assays and xenograft models, no patients. Cited here for the NKG2D-ligand claim on KMT2A-rearranged ALL cells, which is the activating half of the axis this entry describes and is measured in the index disease and lineage - the only such measurement this beat found. PROM1/CD133 as a KMT2A-fusion target is `gene-prom1`'s territory, not this entry's. The NKG2D-ligand finding is a mechanism sub-claim of a therapeutic paper, not the paper's main result. Read: abstract only; identifiers fetched 2026-09-20.
Evidence that argues against this (1)
- This result sits in tension with Chan's, which reports LOW NKG2D activating ligands on KMT2A-rearranged lines. Both can be true - ligands low but not absent, and sufficient once the effector's NKG2D is driven up by CAR engagement - but the two papers are fourteen years apart with different reagents and no direct comparison, and this graph should not state a settled NKG2D-ligand level for this disease. Chan WK et al. 2012
inhibitory KIR/Ly49 ligation by class I holds NK cells off the blast
H2-K1, the murine orthologue of HLA-A, is the molecule shown to be the dependency
NK-resistant with inhibitory KIR ligands, low NKG2D ligands and protease inhibitor-9
NK-resistant with inhibitory KIR ligands, low NKG2D ligands and protease inhibitor-9
- expressed in →MV4-11 (CVCL_0064) — KMT2A::AFF1, yes, but a ten-year-old's acute monocytic leukaemia with FLT3-ITD
NK-resistant with inhibitory KIR ligands, low NKG2D ligands and protease inhibitor-9
missing-self logic and KIR-ligand mismatch; a 52-month remission in a relapsed infant case
the NK brake is the mechanism the risk is built on
route 2 of six
KIR-ligand mismatch is chosen for graft-versus-leukaemia and is the axis a class I-raising drug would blunt