Menin holds KMT2A and KMT2B, and the graph has been written as if it held one
also called Menin-KMT2A/B complex · menin has two wild-type KMT2 partners · the second menin interface
Menin is the anchor a whole class of leukaemia drugs is built around. It grips two proteins, not one - KMT2A and its close relative KMT2B - and every entry in this graph describes only the first grip.
The same thing, for a scientist
Menin's KMT2 partnership is a two-member set. Quantitative AP-MS across all six human SET1/MLL complexes recovers menin with MLL1 and MLL2/KMT2B and with neither SET1A/B nor MLL3/4, and states that menin binds the two with equal affinity [E1]. Sequence supports the specificity: KMT2B carries an MBM1-like motif at aa 17-24 in which the three residues alanine scanning found critical in KMT2A (Phe9, Pro10, Pro13) are conserved as Phe20, Pro21, Pro24, while KMT2C, KMT2D, SETD1A, SETD1B and KMT2E have nothing comparable [E2][E3]. The field names the unit 'Menin-KMT2A/B' in its screens [E4]. **Three consequences, in decreasing certainty.** (1) *Chemistry*: menin-MLL interaction compounds inhibit menin-KMT2A, menin-KMT2B and menin-fusion, which is stated in the literature and not in this graph [E5]. (2) *Measurement*: KMT2B occupancy under a menin inhibitor has never been measured, because chromatin-grade KMT2B reagents are inadequate, and the group that said so also said its own conclusions are therefore an underestimate [E4]. (3) *Therapeutic contribution*: tested directly in mouse MLL-AF9 and found minor - Men1 deletion and Mll1;Mll2 co-deletion diverge, and Mll1;Mll2-deficient cells are MORE sensitive to the compounds [E5]. **So the mechanism is real, the measurement is missing, and the efficacy inference is largely refuted - but only in a myeloid mouse and never in normal tissue, which is where a two-target drug's toxicity lives.**
What is actually established, at what resolution
| claim | best evidence | resolution |
|---|---|---|
| menin is in KMT2B complexes | quantitative label-free AP-MS, GFP-tagged baits, HeLa 1 | complex stoichiometry, not affinity |
| menin binds KMT2A and KMT2B equally | same 1 | equal recovery in a pulldown; no Kd, ITC or SPR |
| menin is NOT in KMT2C/KMT2D or SETD1A/B complexes | same 1 | negative in a pulldown with isoform-fragment baits |
| KMT2B has an MBM1-like motif | this project's alignment of UniProt sequences 3 | inference, not measurement |
| menin-KMT2A MBM1 Kd approximately 10 nM, Phe9/Pro10/Pro13 critical | NMR plus alanine scan 2 | residue-level, KMT2A only |
| a co-crystal or Kd for menin-KMT2B | none exists | - |
The gap between rows 4 and 6 is the whole of the problem. This graph carries interface-menin-kmt2a-mbm1 and interface-menin-kmt2a-mbm2 at atomic resolution. The equivalent entries for KMT2B cannot be written, because the experiments have not been done.
Why the normal-tissue case survives when the efficacy case does not
5 tested the efficacy version of this mechanism head-on and it mostly failed: in MLL-AF9 cells, Men1 deletion and Mll1;Mll2 co-deletion produce distinct consequences, the compounds phenocopy menin loss rather than Mll1;Mll2 loss, and Mll1;Mll2-deficient cells are more sensitive to the compounds. The authors conclude the predominant action is direct inhibition of the fusion. This beat accepts that and says so in gene-kmt2b.
The normal-tissue version is untouched by it, and it is a different question. In a hepatocyte, an islet beta cell, a neuron or a normal haematopoietic progenitor there is no fusion. The only menin-KMT2 interactions a menin inhibitor can engage there are menin-KMT2A and menin-KMT2B. So whatever on-target non-tumour pharmacology revumenib and its class have, it is by construction KMT2A/B pharmacology, and half of it is attributed to a protein this graph did not have.
This matters here more than it would elsewhere because of what else is in this graph. endocrine-* entries describe menin's tumour-suppressor role in islets and parathyroid; risk-menin-inhibitor-exposure-duration-vs-men1-latency asks how long a child can be exposed before MEN1-type consequences appear; mechanism-acute-men1-deletion-reverses-established-hyperglycaemia describes metabolic effects of losing menin. Every one of those is a normal-tissue argument, and every one is written with one KMT2 partner in it.
And a named paediatric phenotype exists on the other side. Germline KMT2B loss of function causes complex progressive childhood-onset dystonia 6. That is not a prediction about a drug - constitutional haploinsufficiency and partial reversible interface occupancy are different perturbations, and this project's own rule about gnomAD constraint applies with equal force. It is a reason to know whether the drug engages menin-KMT2B at therapeutic exposure, which is one afternoon of biochemistry nobody has done.
The paradoxical-activation arm, and what it does to the two-target picture
4 adds a twist that a simple 'the drug hits two proteins' framing misses. Menin loss or VTP50469 does not just remove KMT2A/B function - it liberates KMT2A/B from active genes, and the liberated enzyme accumulates at bivalent promoters, raising the H3K4me3:H3K27me3 ratio and activating genes, including MHC class I. Combined KMT2A/KMT2B knockout has the opposite effect: it impairs the same activation.
So the drug and the genetic loss of its targets point in opposite directions, which is itself an argument that the compound is not simply a KMT2A/B loss-of-function mimetic. And the redistribution argument is a two-enzyme argument that was measured on one enzyme, for the stated reason that the reagents for the other do not work.
| kind | mechanism |
|---|---|
| drug engagement | menin-MLL interaction compounds inhibit menin-KMT2A and menin-KMT2B as well as menin-fusion (stated, PMID 30315824) |
| therapeutic contribution of kmt2b arm | tested and concluded minor in mouse MLL-AF9; never tested in B-lymphoid disease or in normal tissue |
| chromatin measurement of kmt2b | none anywhere; blocked by inadequate chromatin-grade reagents (stated, PMID 36635503) |
menin kmt2 partners
[ "KMT2A (Q03164, 3969 aa)", "KMT2B (Q9UMN6, 2715 aa)" ]
menin non partners in family
[ "KMT2C", "KMT2D", "SETD1A", "SETD1B", "KMT2E" ]
evidence resolution
{
"kmt2a": "co-crystal, NMR, alanine scan, Kd ~10 nM (MBM1/MBM2)",
"kmt2b": "complex-level AP-MS stoichiometry only; no structure, no Kd, no alanine scan"
}affects
[ "drug-revumenib", "drug-ziftomenib", "drug-bleximenib", "drug-enzomenib" ]
- This is an on-target toxicity argument, not an off-target one, so it cannot be engineered away by making a more selective menin binder. Any compound that occupies the menin MBM pocket occupies it against both partners.
- It is unmeasurable with current reagents on the KMT2B side, so a trial cannot generate the pharmacodynamic data that would settle it, and a negative clinical safety signal will always be confounded with not having looked.
- The paradoxical-activation arm means the drug is not a clean loss-of-function of either partner. A toxicity model built on 'what happens when you lose KMT2B' may be the wrong model even if the two-target premise is right.
- If the KMT2B arm turns out to contribute to efficacy in B-lymphoid disease - untested - then a future more KMT2A-selective agent would be less effective, not more, which inverts the usual selectivity argument.
- What fraction of menin in a normal human tissue is bound to KMT2B rather than KMT2A? The one quantitative dataset is HeLa. Islet, liver and haematopoietic progenitor answers would change how the endocrine entries in this graph should be read.
- Does a clinical menin inhibitor engage menin-KMT2B at therapeutic plasma exposure? An IC50 against a KMT2B MBM peptide, compared against the KMT2A IC50 the developers already have, is a single experiment and it has not been published for any of the four agents.
- Does the paradoxical bivalent-gene activation happen in a fusion-bearing cell? It is demonstrated in K-562 and stem cells where wild-type KMT2A/B are menin's only clients. In a t(4;11) blast the fusion competes, and the redistribution could be smaller, larger or absent.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Label-free quantitative AP-MS across the six human SET1/MLL complexes recovered menin with MLL1 and MLL2 (Q9UMN6, i.e. KMT2B) and not with SET1A/B or MLL3/4, with the authors stating that menin binds Mll1 and Mll2 with equal affinity; menin was 4- to 5-fold less abundant than MLL1/2, so not all MLL1/2 complexes contain menin, and the menin-bound fraction carries WRAD at approximately 1:1/1/1/6.
Quantitative dissection and stoichiometry determination of the human SET1/MLL histone methyltransferase complexes. · van Nuland R, Smits AH, Pallaki P, Jansen PW, Vermeulen M, Timmers HT · Mol Cell Biol 2013What this evidence cannot show
Full text read via the PMC HTML route (PMC3647974) on 2026-09-20 after routes 2 and 3 failed. HeLa; GFP-tagged overexpressed catalytic subunits; the MLL1 bait is a fragment accession (E9PQG7) rather than full-length Q03164. 'Equal affinity' is the authors' word for equal recovered stoichiometry. Nothing here is a KMT2A-rearranged cell and nothing is haematopoietic.
Evidence that argues against this (1)
- Equal recovery in a pulldown from one cell line is compatible with very unequal affinity if the two enzymes differ in abundance, in complex stability or in how well the GFP tag is tolerated. The absence of any direct binding measurement for menin-KMT2B means the 'equal' is the weakest load-bearing word in this entry. Grembecka J et al. 2010
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
MLL binds menin with a Kd of approximately 10 nM through two motifs, MBM1 and MBM2, in the N-terminal 43 amino acids, with MBM1 the high-affinity motif and alanine scanning identifying Phe9, Pro10 and Pro13 as most critical; MBM1 binds menin in an extended conformation and an MBM1 peptide efficiently dissociates the menin-MLL complex.
Molecular basis of the mixed lineage leukemia-menin interaction: implications for targeting mixed lineage leukemias. · Grembecka J, Belcher AM, Hartley T, Cierpicki T · J Biol Chem 2010What this evidence cannot show
Abstract fetched and read 2026-09-20. **KMT2A only; KMT2B appears nowhere in this study.** NMR on an intrinsically disordered fragment out of the context of the full protein and of chromatin.
Evidence that argues against this (1)
- The crystallographic account adds a composite surface formed jointly by menin and MLL1 that binds LEDGF, so an MBM-only description is incomplete for KMT2A and correspondingly incomplete for any KMT2B extrapolation. Huang J et al. 2012
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
Canonical sequences fetched from UniProt 2026-09-20 show KMT2A beginning MAHSCRWRFPARPG (MBM1 at aa 6-13 with F9/P10/P13) and KMT2B containing RGRFPGRP at aa 17-24 (F20/P21/P24), while KMT2C, KMT2D, SETD1A, SETD1B and KMT2E carry no comparable N-terminal motif - matching the measured specificity of menin for the KMT2A/KMT2B pair.
UniProtKB canonical sequences for Q03164 (KMT2A), Q9UMN6 (KMT2B), Q8NEZ4 (KMT2C), O14686 (KMT2D), O15047 (SETD1A), Q9UPS6 (SETD1B), Q8IZD2 (KMT2E); fetched and aligned 2026-09-20 by this project · UniProt Consortium; alignment by this project · Database 2026What this evidence cannot show
**This is an alignment made by this project, by eye, and it is an inference rather than evidence.** The sequences are database facts fetched the same day; the correspondence of F20/P21/P24 to F9/P10/P13 is a reading of them. The Trp of KMT2A's RWRFP is a Gly in KMT2B, at the position adjacent to the critical Phe, and no experiment tells us what that costs. Graded moderate for that reason.
Evidence that argues against this (1)
- Motif conservation predicts binding poorly in this very family: KMT2A, KMT2B, SETD1A and SETD1B all carry WDR5 WIN motifs and separate co-crystals had to be solved because the sequences did not settle the chemistry. A single substitution adjacent to the critical residue is exactly where a sequence argument fails. Bach C et al. 2009
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Whole-genome CRISPR screens identified Menin-KMT2A/B as a maintainer of bivalency alongside MTF2-PRC2.1 and PCGF1-PRC1.1; Menin loss or VTP50469 derepressed bivalent genes by releasing KMT2A from active genes and shifting the polycomb/KMT2A balance, combined KMT2A/KMT2B knockout had the opposite effect on IFN-gamma-induced MHC-I, and the authors state that inadequate commercial reagents for monitoring KMT2B chromatin binding mean their KMT2A-only analyses underestimate the consequences of liberating KMT2A/B.
Targeting Menin disrupts the KMT2A/B and polycomb balance to paradoxically activate bivalent genes. · Sparbier CE, Gillespie A, Gomez J, et al. · Nat Cell Biol 2023What this evidence cannot show
Full text read via Europe PMC fullTextXML (PMC7614190) on 2026-09-20. The bivalency work is in K-562 (BCR::ABL1) and pluripotent stem cells - **no KMT2A-rearranged line**, so the redistribution is of wild-type menin-KMT2A/B and there is no fusion competing for menin.
Evidence that argues against this (1)
- In a fusion-bearing cell the fusion is the dominant menin client, so 'liberating KMT2A/B from active genes' may not describe what a menin inhibitor does in the disease this graph is about. The redistribution mechanism is demonstrated in cells where wild-type KMT2A/B are the only menin partners. Chen Y et al. 2019
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Compounds disrupting the menin-MLL N-terminal interaction also inhibit menin-MLL1, menin-MLL2 and other menin-interacting proteins in biochemical and structural assays; nevertheless Men1 deletion in MLL-AF9 cells produces distinct consequences from Mll1;Mll2 co-deletion, the compounds phenocopy menin loss, Mll1;Mll2-deficient cells are MORE sensitive to the inhibitors, and the authors conclude the predominant action is direct inhibition of the fusion without significant contribution from MLL1/MLL2 inhibition.
Distinct pathways affected by menin versus MLL1/MLL2 in MLL-rearranged acute myeloid leukemia. · Chen Y, Jones KL, Anastassiadis K, et al. · Exp Hematol 2019What this evidence cannot show
Abstract fetched and read 2026-09-20; full text (PMC6472707) not obtained, so compound identities and the magnitude of the sensitivity shift are unrecorded here. Mouse MLL-AF9, myeloid, leukaemia cells only - **there is no normal-tissue arm**, which is precisely where the two-target concern lives. This paper is cited nowhere else in this graph.
Evidence that argues against this (1)
- The conclusion is about therapeutic contribution in leukaemia cells and is silent on toxicity. The first clause of the same sentence - that the compounds do inhibit menin-MLL1 and menin-MLL2 - is what a normal-tissue argument needs, and it is conceded rather than tested. Sparbier CE et al. 2023
- E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Heterozygous germline KMT2B variants cause a complex progressive childhood-onset dystonia in 27 unrelated individuals, with a typical facial appearance, characteristic brain MRI findings and marked benefit from deep brain stimulation - a paediatric neurological phenotype for loss of the second menin partner.
Mutations in the histone methyltransferase gene KMT2B cause complex early-onset dystonia. · Meyer E, Carss KJ, Rankin J, et al. · Nat Genet 2017What this evidence cannot show
Abstract fetched and read 2026-09-20. **Constitutional haploinsufficiency from conception is not a model for partial reversible pharmacological interface disruption in an already-developed child**, exactly as this project's standing brief says about gnomAD constraint. Cited as a watch item, not a prediction.
Evidence that argues against this (1)
- Menin inhibitors have been dosed in children in AUGMENT-101 and related studies without a reported dystonia signal, so whatever menin-KMT2B engagement occurs at clinical exposure has not produced the germline phenotype. Absence of a reported signal in small, short, heavily pretreated cohorts is weak evidence, but it is the evidence that exists. Larson JK et al. 2023
KMT2B keeps the MBM1 core at the contact residues
menin's KMT2 partnership is a two-member set: KMT2A and KMT2B, and neither KMT2C/D nor SETD1A/B
the unmeasured half of the Menin-KMT2A/B complex
the KMT2A interface is solved at residue resolution; the KMT2B equivalent has no structure and no Kd
- related to →The reversibility question: months of partial occupancy against a syndrome built from decades of loss
the normal-tissue exposure argument is written with one KMT2 partner in it
the hypothesis extends the established two-partner mechanism into untested normal tissue