Does revumenib displace JunD, and at what concentration relative to KMT2A? (The yes/no half is already answered)
also called menin JunD co-IP under drug · JunD displacement assay
A menin-blocking drug does knock JunD off menin - that has now been shown with laboratory tool compounds and in leukaemia cells. What is still missing is the same measurement for the three drugs actually given to patients, in an endocrine cell, with the concentration needed for JunD displacement set against the concentration needed to displace KMT2A.
The same thing, for a scientist
Every clinical menin inhibitor is a ligand for the single deep peptide pocket that binds both the MLL1 MBM1 motif and a JunD motif in the same manner [E1], and the deposited structures 3U85 and 3U86 show both peptides bound to the same human menin [E2]. From this the whole endocrine-liability argument infers that a menin inhibitor displaces JunD. **That inference has since been confirmed for tool compounds and, by orthogonal proteomics, in leukaemic cells** - MI-503 and MI-463 abolish menin-JunD co-immunoprecipitation dose-dependently at 30-300 nM, and quantitative mass spectrometry shows menin-MLL inhibitors disrupt menin-JUND in MLL-rearranged cells [counter-evidence to E1, E3]. **What has still not been done is the version that matters for dosing**: revumenib, ziftomenib or bleximenib, in an endocrine-lineage cell, with the JunD-displacement IC50 measured against the KMT2A-displacement IC50 in the same lysate. That ratio, not the existence of displacement, is the number this beat needs, and it is now the sole justification for running the experiment.
Design
Cells. A parathyroid- or islet-lineage human line, plus a KMT2A-rearranged leukaemia line as the positive-control tissue where MLL1 displacement is known to occur. Primary human parathyroid tissue from routine parathyroidectomy would be better and is obtainable from any endocrine surgery service.
Perturbation. Revumenib, ziftomenib and bleximenib across a concentration range spanning the clinically achievable free concentration, 24 and 72 hours.
Primary readout. Immunoprecipitate endogenous menin; blot for JunD and for KMT2A. Report the IC50 for JunD displacement against the IC50 for KMT2A displacement in the same lysate. The ratio of those two numbers is the answer to this beat's central pharmacological question.
Orthogonal readouts.
- NanoBRET or cellular thermal shift for menin-JunD engagement, to avoid the lysis artefacts co-IP is prone to.
- Menin protein level by western, to test whether the drug degrades menin rather than merely occupying it - which would remove all menin functions, not only the pocket-mediated one.
- RNA-seq for AP-1 target genes and for a JunD signature, to show whether displacement has a transcriptional consequence in an endocrine cell.
Biophysical companion. Measure Kd for the menin-JunD peptide by ITC or SPR, alongside the published menin-MBM1 value. This number is currently absent from the literature and is what makes the cellular result interpretable.
What each outcome would mean
| Result | Consequence for this beat |
|---|---|
| JunD displacement IC50 comparable to or below KMT2A displacement IC50 | the endocrine liability is mechanistically live at therapeutic exposure; surveillance is obligatory |
| JunD displacement IC50 an order of magnitude above KMT2A's | the therapeutic window separates the two after all, and the beat should be substantially downgraded |
| JunD not displaced at any achievable concentration | |
| Menin protein degraded rather than merely occupied | the pocket-versus-whole-protein distinction collapses and the risk is larger than the structural argument implies |
Correction (2026-09-19). This section previously nominated the third row as the outcome the beat should most want, because it would cheaply retire the concern. That is withdrawn: the displacement question has been answered in the affirmative twice, so the experiment's value now lies entirely in the first two rows - the potency ratio - and in the degradation-versus-occupancy control. The experiment is still worth running; it is no longer a potential kill shot.
| design | co-immunoprecipitation of endogenous menin with dose titration of revumenib, ziftomenib and bleximenib in an endocrine-lineage human cell line and, ideally, primary human parathyroid tissue, with a KMT2A-rearranged leukaemia line as positive-control tissue |
|---|---|
| readout | IC50 for JunD displacement versus IC50 for KMT2A displacement measured in the same lysate; the ratio is the primary result |
| biophysical companion | Kd of the menin-JunD peptide interaction by ITC or SPR, currently absent from the literature, against the published menin-MBM1 value |
| controls | vehicle; a structurally unrelated cytotoxic; an MBM1 peptide competition arm |
| cost tier | low |
| timeline | one to two weeks for the cellular result; four to six weeks including the biophysics |
| decisive negative outcome | WITHDRAWN 2026-09-19. This field previously read 'no JunD displacement at clinically achievable concentrations would retire this entire beat'. Displacement has since been demonstrated with MI-503/MI-463 by co-IP at 30-300 nM (PMID 33821572) and with menin-MLL inhibitors by quantitative mass spectrometry in leukaemic cells (PMID 42371796). The experiment's remaining decisive output is the JunD:KMT2A displacement IC50 ratio for the three clinical compounds, plus whether the drugs degrade menin rather than occupying it. |
orthogonal readouts
[ "NanoBRET or CETSA menin-JunD engagement", "menin protein level by western to test degradation versus occupancy", "RNA-seq for AP-1 and JunD target signatures in endocrine cells" ]
prior work already published
[
{
"pmid": "33821572",
"what": "MI-503 and MI-463 abolish menin-JunD co-IP dose-dependently in HEK293T at 30/100/300 nM; MI-503 10-100 nM raises Ki67 in primary human islets; 70 mg/kg/day in HFD mice improves glucose tolerance over 12 weeks",
"title": "Menin-regulated Pbk controls high fat diet-induced compensatory beta cell proliferation.",
"authors": "Ma J, Xing B, Cao Y, et al.",
"venue": "EMBO Mol Med",
"year": 2021,
"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."
},
{
"pmid": "42371796",
"what": "quantitative mass spectrometry in MLL-rearranged leukaemic cells: menin-MLL inhibitors disrupt menin-JUND, JUND chromatin occupancy and targets rise, contributing to drug resistance",
"title": "Menin-MLL inhibitors enhance JUND activity in MLLr leukemic cells contributing to tumorigenesis and therapy resistance.",
"authors": "Ozyerli-Goknar E, Nizamuddin S, Fetsch V, et al.",
"venue": "Blood",
"year": 2026,
"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."
}
]- This experiment can no longer kill the beat in a week. Its premise - that displacement under drug was untested - was false, and the published answer is that displacement happens. What remains is a slower, more expensive question about exposure duration in a differentiated endocrine cell.
- A negative co-IP would not fully exonerate the class if the drugs degrade menin rather than occupying it, which is why the western for menin level is not optional.
- Cell-line results in an artificial endocrine model may not predict the human parathyroid, which is why primary tissue is specified even though it is harder to obtain.
- The more consequential bite-back found while checking this entry is therapeutic rather than toxicological: JUND released by a menin inhibitor drives its own target programme and contributes to resistance in leukaemic cells (PMID 42371796). An experiment framed only around endocrine safety will measure the wrong endpoint and miss the efficacy problem sitting in the same mechanism.
- What is the ratio of the JunD-displacement IC50 to the KMT2A-displacement IC50 for revumenib, ziftomenib and bleximenib measured in the same lysate? Displacement itself is established (PMID 33821572, PMID 42371796); the potency ratio - the number that decides whether a therapeutic window exists - has never been reported for any compound.
- Which human endocrine cell line is the right vehicle? Parathyroid lines are notoriously poor; primary tissue from routine parathyroidectomy may be the only faithful system and requires an endocrine surgery collaboration. Note that primary human ISLETS have already been used for this purpose (PMID 33821572), so the islet arm is the cheaper starting point.
- What is the free (unbound) clinically achievable concentration of each menin inhibitor? The labels give total AUC; protein binding data are needed to set the relevant concentration range and to say whether the published 30-300 nM tool-compound window is clinically relevant.
- Does menin-inhibitor treatment reduce menin protein levels in endocrine cells? If it does, the pocket-occupancy framing of this whole beat is the wrong one, and neither published study addresses it.
- Does JUND liberation in an endocrine cell produce the same chromatin-occupancy and target-gene response that it produces in leukaemic cells (PMID 42371796)? If the transcriptional consequence is tissue-specific, the leukaemia resistance result does not transfer to the parathyroid and vice versa.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Human menin contains a single deep peptide-binding pocket that binds short MLL1 and JUND peptides in the same manner but with opposite effects on transcription; several MEN1 missense mutations disrupt the menin-JUND interaction, and menin loss or inactivation causes multiple endocrine neoplasia type 1.
The same pocket in menin binds both MLL and JUND but has opposite effects on transcription · Huang J, Gurung B, Wan B, Matkar S, Veniaminova NA, Wan K, Merchant JL, Hua X, Lei M · Nature 482:542-6 2012What this evidence cannot show
Read as abstract. The paper establishes structural sharing of the pocket and does not test any small molecule. The small-molecule gap this experiment was written to fill has since been partly closed by PMID 33821572 and PMID 42371796; see counter_evidence.
Evidence that argues against this (2)
- The core premise of this experiment - that nobody has run the co-IP under drug - was false when it was written. Ma et al. show MI-503 and MI-463 inhibiting menin-JunD binding in a dose-dependent manner by co-immunoprecipitation in HEK293T at 30, 100 and 300 nM, with marked inhibition at the high concentrations. They further treat primary HUMAN islets with MI-503 at 10-100 nM for 5 days and see dose-dependent Pbk and Ki67 induction, and give a menin inhibitor at 70 mg/kg daily by gavage to high-fat-diet mice with improved fasting glucose and glucose tolerance from week 6 to week 12. The endocrine-lineage arm of this design is therefore partly pre-empted, and its published outcome is proliferative and metabolically beneficial rather than neoplastic. Ma J et al. 2021
- The leukaemia-line positive-control arm has also been run, by a better method than co-IP, and it reports the opposite kind of consequence from the one this experiment was designed to detect. Ozyerli-Goknar et al. used quantitative mass spectrometry in MLL-rearranged leukaemic cells to show menin-MLL inhibitors disrupt menin-JUND interactions; JUND chromatin occupancy increases, target genes are upregulated, and this contributes to RESISTANCE to menin inhibition. JUND inactivation or JNK inhibition synergised with the drug. Quantitative interaction proteomics under drug is strictly more informative than a blot and does not suffer the lysis artefacts this design worries about, so the orthogonal readout section should be reordered to put MS first. Ozyerli-Goknar E et al. 2026
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
PDB 3U86 is the 2.84 A structure of human menin with a JunD peptide (SPGRLFPGAPPTAKK) and PDB 3U85 is the 3.0 A structure of human menin with an MLL1/KMT2A MBM1 peptide (SRWRFPARPGTTGGGGGGGRR); the companion depositions are 3U84 (apo menin) and 3U88 (menin with MLL1 and LEDGF).
PDB 3U86: Crystal structure of human menin in complex with JunD · Huang J, Gurung B, Wan B, Matkar S, Veniaminova NA, Wan K, Merchant JL, Hua X, Lei M · RCSB Protein Data Bank 2012What this evidence cannot show
Entity sequences, methods and resolutions retrieved from the RCSB data API on 2026-09-18. No affinity data are deposited with either structure, which is why the biophysical companion measurement is part of this design.
Evidence that argues against this (1)
- Both structures are peptide complexes, so they establish that the two motifs engage one site and nothing about their relative cellular occupancy. A co-IP experiment designed from them inherits the same gap: it will show whether JunD comes off, not whether enough comes off to matter. PNAS 2003
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Menin binds JunD directly and represses JunD-activated transcription; JunD deprived of menin switches from growth suppressor to growth promoter in immortalised fibroblasts.
Transcription factor JunD, deprived of menin, switches from growth suppressor to growth promoter · Agarwal SK, Novotny EA, Crabtree JS, Weitzman JB, Yaniv M, Burns AL, Chandrasekharappa SC, Collins FS, Spiegel AM, Marx SJ · Proc Natl Acad Sci U S A 100:10770-5 2003What this evidence cannot show
Read as abstract. Provides the reason displacement would matter if it occurs, in a cell system unrelated to endocrine tissue.
Evidence that argues against this (1)
- The growth-suppressor-to-promoter switch was shown in JunD-null immortalised FIBROBLASTS, where menin behaves as a tumour suppressor. In KMT2A-rearranged leukaemia menin is an oncogenic cofactor, and this graph already records that the sign of menin's effect differs between those settings - so the sign of JunD's may too. The design should not assume the fibroblast direction. PNAS 2003
tests the untested pharmacological assumption the entire beat rests on
a negative result would falsify the hypothesis outright in about a week
measures whether clinical menin inhibitors displace JunD and at what concentration relative to KMT2A displacement
same design; FANCD2 and RELA can be blotted on the same membranes
Same missing number in a different tissue: the ratio of the endocrine-disinhibiting concentration to the KMT2A-displacing concentration in the same lysate
the JUND:KMT2A displacement IC50 ratio quantifies how much of the drug window this mechanism consumes