A menin inhibitor may release RELA from menin's restraint — and released RELA is what keeps MLL fusions on chromatin. Prior 0.20
also called RELA liberation hypothesis · menin inhibitor NF-kB resistance
Menin holds down NF-kB, and NF-kB turns out to be needed to keep the KMT2A fusion protein attached to the genes it controls. So a drug that knocks menin off could accidentally free NF-kB and help the fusion hold on — the drug partly undoing its own job. Nobody has looked, and there is one good structural reason it might not happen at all.
The same thing, for a scientist
**Statement.** In KMT2A-rearranged leukaemia, a clinical menin inhibitor relieves menin's repression of NF-kB; the liberated canonical NF-kB, principally RELA, then sustains fusion retention on chromatin and the leukaemia stem cell programme, blunting the drug's effect and contributing to adaptive, non-genetic resistance. **Prior probability 0.20.** The two halves of the chain are individually solid and were established by different laboratories that never cited each other on this point. Menin binds p50, p52 and p65 and represses p65-mediated transactivation via a mapped central region and a SIRT1 corepressor [E1][E2]. RELA occupies the promoters of crucial MLL target genes, sustains the MLL-dependent LSC programme, and IKK/NF-kB signalling is **required for retention of wild-type and fusion MLL protein on chromatin**, with NF-kB suppression phenocopying loss of the oncoprotein [E3]. The shape is the one this graph already found once: `resistance-jund-liberation-menin`, where a drug that works by occupying menin's pocket liberates a partner that pocket was restraining. **And there is disease-matched precedent for NF-kB as the resistance axis specifically in t(4;11)** [E4]. **The strongest single objection, and it is structural.** JunD is displaceable because it binds the MBM pocket itself (PDB 3U86 against 3U85). **NF-kB is not known to bind that pocket at all** — the only mapping is a linear menin region, 300-381, and menin's repression of p65 is executed by SIRT1 recruitment rather than occlusion [E2]. If the contact does not use the pocket, revumenib will not touch it and the hypothesis is dead at step one. Second objection: **NF-kB is not one thing.** Activating the non-canonical arm SUPPRESSES MLL-AF9 leukaemia, so liberating all three menin-bound subunits at once has an unpredictable net sign [E5].
The chain, link by link, with what is missing at each step
| link | status | evidence |
|---|---|---|
| menin binds RELA | established, one paper, one method (GST pull-down), never replicated in 25 years | 12 |
| menin represses RELA-driven transcription | established in reporter assays and three epithelial systems | 12 |
| that repression is relieved when menin is removed | shown once, in an intestinal crypt line, via proliferation and cyclin D1 | 2 |
| a menin POCKET inhibitor relieves it | never tested; not even structurally plausible yet | — |
| RELA sustains the MLL fusion programme | established, mouse models plus human AML cohorts | 3 |
| the liberated RELA reaches the fusion's targets | never tested | — |
| this contributes to clinical resistance | never tested | — |
Three of seven links have no evidence at all, and the fourth — the pharmacological one — is the load-bearing one. That is why the prior is 0.20 and not higher, despite both endpoints being well supported.
The cheapest way to kill it
One competition experiment. Purified menin plus a RELA N-terminal fragment, titrated with revumenib. If the inhibitor does not compete, the hypothesis is dead and the whole thing costs a week. This is the step to do first, before any cell work, because every downstream prediction depends on it.
If it survives that, the next experiment is already designed for a different partner: exp-jund-displacement-coip-under-clinical-menin-inhibitor. Adding RELA and FANCD2 antibodies to the same membranes costs almost nothing and answers three of this beat's questions in one blot.
The cell-level prediction that would be most convincing is not a co-IP at all: a menin inhibitor should, within hours, increase RELA occupancy or RELA-dependent transcription at MLL target promoters, in the same window in which HOXA9 and MEIS1 output is falling. Two effects in opposite directions at the same loci at the same time would be very hard to explain any other way, and both read-outs are standard.
| statement | A clinical menin inhibitor relieves menin's repression of canonical NF-kB in KMT2A-rearranged leukaemia; the liberated RELA sustains fusion retention on chromatin and the LSC programme, blunting the drug and contributing to adaptive non-genetic resistance. |
|---|---|
| origin | literature-gap |
| prior probability | 0.2 |
| relationship to jund | same architecture, different site. JunD binds the MBM pocket and is therefore certainly displaceable by a clinical inhibitor; NF-kB has no evidence of pocket binding, which is the difference that decides this hypothesis. |
| why nobody found this | the two halves were published eleven years apart in different fields and neither cites the other on this point; PMID 24054986 mentions menin only in its reference list, despite coming from the laboratory that established the menin-MLL requirement |
predictions
[ "revumenib competes with a RELA N-terminal fragment for purified menin in vitro", "a menin inhibitor increases RELA-dependent transcription in a KMT2A-rearranged line within hours, while HOXA9/MEIS1 output is falling", "RELA ChIP occupancy at MLL target promoters rises, or fails to fall, under menin inhibition", "an IKK inhibitor or an NF-kB inhibitor synergises with a menin inhibitor in KMT2A-rearranged but not in non-rearranged lines", "menin-inhibitor-persister cells carry an NF-kB transcriptional signature", "the effect is canonical-specific: p52/RELB output does not rise, or rises with the opposite phenotypic sign" ]
experiments
[ "exp-fancd2-menin-coip-under-menin-inhibition-kmt2a-line" ]
analyses
[ "ana-menin-interactome-enumeration-and-bounded-rejections" ]
falsifiers
[ "revumenib does not compete with RELA for menin in vitro - the cheapest and most decisive kill, because it removes every downstream prediction at once", "menin and RELA do not co-immunoprecipitate at endogenous levels in any KMT2A-rearranged line", "menin inhibition does not change NF-kB reporter output or RELA target gene expression in KMT2A-rearranged cells", "menin ChIP shows no overlap with published RELA occupancy at MLL target promoters", "an NF-kB or IKK inhibitor adds nothing to a menin inhibitor in KMT2A-rearranged lines beyond its single-agent effect", "the menin-RELA repression proves to require SIRT1 recruitment in a way that survives displacement of menin from the MBM pocket" ]
- If true, the mechanism of action contains its own resistance mechanism, and every trial reading only HOXA9/MEIS1 or differentiation endpoints would miss it.
- The obvious countermeasure - adding an NF-kB or IKK inhibitor - stacks two myelosuppressive agents in infants and removes a pathway that normal haematopoiesis and host defence both require.
- If false, the graph must resist the pull of a neat story: two solid endpoints connected by an untested middle is exactly the shape of the four Fanconi explanations this project has already killed.
- Does any menin inhibitor compete with NF-kB for menin? This is the whole hypothesis in one in vitro titration.
- Is there an NF-kB signature in published menin-inhibitor-resistant or persister transcriptomes? This could be answered from existing data without a new experiment.
- Would an IKK inhibitor plus a menin inhibitor be tolerable in infants, given that both suppress marrow?
- Does menin occupancy overlap RELA occupancy at MLL target promoters in any published ChIP dataset pair?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The NF-kB proteins p50, p52 and p65 interact specifically and directly with menin in vitro and in vivo; the N-terminus of the NF-kB proteins is sufficient for binding; amino acids 305-381 of menin are essential; menin represses p65-mediated transactivation on NF-kB sites dose-dependently and specifically, and suppresses PMA-stimulated NF-kB activation.
The tumor suppressor protein menin interacts with NF-kappaB proteins and inhibits NF-kappaB-mediated transactivation. · Heppner C, Bilimoria KY, Agarwal SK, Kester M, Whitty LJ, Guru SC, Chandrasekharappa SC, Collins FS, Spiegel AM, Marx SJ, Burns AL · Oncogene 2001What this evidence cannot show
`established` for the reported system only; no haematopoietic cell appears in it, and a menin partner is not a KMT2A::AFF1 partner. Full text unreadable through all five routes - no PMCID, and the DOI resolves to a 303-character stub. One paper, one laboratory, 2001, never replicated at the protein level.
Evidence that argues against this (1)
- Twenty-five years of menin affinity proteomics has not re-detected any NF-kB subunit; IntAct holds exactly one source publication for all three pairs. Huttlin EL et al. 2021
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Menin's repression of p65 is executed by recruitment of SIRT1, which deacetylates p65 at K310; a SIRT1 inhibitor or siRNA abolishes it. In an intestinal crypt-like cell line, reducing menin raises cyclin D1 promoter activity and proliferation, and blocking NF-kB restores both to the menin-present state - the only published system in which losing menin measurably releases NF-kB with a growth consequence.
The tumor suppressor protein menin inhibits NF-κB-mediated transactivation through recruitment of Sirt1 in hepatocellular carcinoma. · Gang D, Hongwei H, Hedai L, Ming Z, Qian H, Zhijun L · Mol Biol Rep 2013What this evidence cannot show
`moderate` because the claim combines two papers in two unrelated epithelia - hepatocellular carcinoma for the SIRT1 mechanism and a rat intestinal crypt line for the release phenotype - into one statement, and neither was done in a haematopoietic cell. The intestinal paper's own language is 'a correlation between cyclin D1 expression, NF-kB activity and menin expression', not a demonstrated causal chain.
Evidence that argues against this (1)
- The SIRT1 mechanism is the strongest argument AGAINST this hypothesis, not for it: if menin represses p65 by recruiting a deacetylase rather than by occluding a surface, then displacing menin from its MBM pocket need not relieve the repression at all. Murai MJ et al. 2011
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
NF-kB signalling is essential in MLL leukaemia: suppressing it gives robust anti-leukaemia effects phenocopying loss of the functional MLL oncoprotein or its epigenetic cofactors; RELA occupies promoter regions of crucial MLL target genes and sustains the MLL-dependent leukaemia stem cell programme; and IKK/NF-kB signalling is required for retention of wild-type and fusion MLL protein on chromatin and for maintenance of the associated histone modifications.
Epigenetic roles of MLL oncoproteins are dependent on NF-κB. · Kuo HP, Wang Z, Lee DF, Iwasaki M, Duque-Afonso J, Wong SH, Lin CH, Figueroa ME, Su J, Lemischka IR, Cleary ML · Cancer Cell 2013What this evidence cannot show
Mouse MLL-AF1p, MLL-AF10 and MLL-AF9 plus human AML cohorts. Myeloid throughout; no infant B-ALL and no KMT2A::AFF1 model. Full text read from PMC3816582 via efetch db=pmc: the string 'menin' appears twice and both occurrences are in the reference list, so the paper makes no claim about menin at all.
Evidence that argues against this (1)
- If NF-kB is required for fusion retention on chromatin AND menin is required for fusion retention on chromatin, the simplest reading is that they are the same requirement measured twice - in which case removing menin removes the fusion regardless of what NF-kB does, and liberation is irrelevant. Yokoyama A et al. 2008
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In a murine model of t(4;11) acute lymphoblastic leukaemia, leukaemic cells acquired resistance to interferon beta and relapsed; NF-kB activation was identified as the mechanism of that resistance, and inhibiting NF-kB in the resistant cells resensitised them.
NF-kappaB activation mediates resistance to IFN beta in MLL-rearranged acute lymphoblastic leukemia. · Tracey L, Streck CJ, Du Z, Williams RF, Pfeffer LM, Nathwani AC, Davidoff AM · Leukemia 2010What this evidence cannot show
The resistance is to interferon beta delivered by AAV, a therapy that is not used; the relevance is that NF-kB acted as an acquired-resistance axis in t(4;11) disease specifically, not that it would do so against a menin inhibitor. Murine model, no patient data.
Evidence that argues against this (1)
- The resistance here arose against interferon beta, not against a menin inhibitor, so it supports the hypothesis only by analogy. A hypothesis about what menin inhibition liberates needs an experiment in which menin is inhibited. Xiu Y et al. 2018
- E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Stabilisation of NF-kB-inducing kinase activates non-canonical NF-kB signalling, represses canonical NF-kB signalling, and suppresses MLL-AF9-induced AML; canonical NF-kB is constitutively activated in AML stem cells and required for their self-renewal.
Stabilization of NF-κB-Inducing Kinase Suppresses MLL-AF9-Induced Acute Myeloid Leukemia. · Xiu Y, Dong Q, Li Q, Li F, Borcherding N, Zhang W, Boyce B, Xue HH, Zhao C · Cell Rep 2018What this evidence cannot show
Mouse MLL-AF9, myeloid. Cited here as the reason the hypothesis is written on RELA specifically rather than on 'NF-kB': menin binds all three subunits and the two arms have opposite signs in this disease, so a whole-pathway liberation has no predictable direction.
Evidence that argues against this (1)
- This cuts against the hypothesis as much as for it: if stabilising NIK - and so RAISING non-canonical signalling - SUPPRESSES MLL-AF9 AML, then 'more NF-kB reinforces the fusion programme' is false for at least one arm of the pathway. The hypothesis has to name which arm it means. Zhao J et al. 2021
the hypothesis requires that this repression be relievable by a pocket inhibitor
- analogous to →JUND liberation — the drug's mechanism of action and one of its mechanisms of resistance are the same event
same liberation shape; JunD uses the MBM pocket and NF-kB is not known to
- confers resistance to →Revumenib (SNDX-5613, Revuforj) — first approved menin–KMT2A interface inhibitor
proposed adaptive non-genetic resistance route via liberated canonical NF-kB
- depends on →RELA (p65) — menin binds it and represses it, and it is required to keep MLL oncoproteins on chromatin
RELA is the liberated partner the hypothesis is written on
- supports →Adaptive, non-genetic resistance — the leukemia keeps the fusion and rewires how it is maintained
a named candidate mechanism for the adaptive non-genetic resistance class
- ← proposed byRELA (p65) — menin binds it and represses it, and it is required to keep MLL oncoproteins on chromatin
the entry from which the hypothesis was derived
the same axis read backwards: one says menin inhibition self-defeats, this says an IKK inhibitor is the counter