The reversibility question: months of partial occupancy against a syndrome built from decades of loss
also called menin inhibitor reversibility · exposure duration endocrine risk · is it reversible
MEN1 syndrome comes from losing menin permanently, in every cell, for decades. A menin inhibitor takes away part of one of menin's jobs for a few months and then stops, and whether that is enough to matter is genuinely unknown.
The same thing, for a scientist
This is the crux and it should be stated as an open question rather than a risk assertion. MEN1 syndrome requires germline monoallelic loss plus a somatic second hit accumulated over years, and even that leaves 93% of carriers disease-free at age 10 [E1]. Menin-inhibitor exposure is partial (competitive occupancy, not deletion), selective (the pocket-binding arm of menin function, with the kinase-uncoupling arms possibly spared), reversible, and short - median 2.5 months in the revumenib registration population, 10% beyond 6 months [E2]. On those grounds the risk should be small. Three things stop that argument from closing: the exposure distribution is lengthening as the drug moves into maintenance and front-line settings, with post-transplant maintenance already running to 199 days and FDA's own postmarketing requirement specifying at least 3 years of treatment [E3][E4]; irreversible findings already exist in the nonclinical package at 2x human exposure [E2]; and nobody has run the restoration experiment that would bound it.
The four differences that argue the risk is small
- Partial versus complete. A competitive ligand occupies a fraction of pockets at a fraction of the time; a MEN1 mutation removes the allele.
- One function versus all functions. Menin represses JunD by direct pocket binding and by uncoupling JunD from ERK and JNK phosphorylation, and the second of these is not obviously pocket-mediated (
mechanism-menin-jund-transcriptional-repression). A pocket ligand may reproduce only part of the MEN1 phenotype. - Reversible versus permanent. Occupancy ends when the drug is cleared; a second hit does not.
- Months versus decades. Median 2.5 months against a syndrome whose earliest recorded manifestation is 9 years of germline haploinsufficiency 12.
Stated together these are a strong case, and this entry's honest position is that they probably win. The counterweights below are why 'probably' is the right word.
The three counterweights
The exposure distribution is not fixed. The 2.5-month median describes salvage therapy in patients who mostly died or went to transplant. In the pivotal trial, seven responders resumed revumenib after allogeneic transplant as maintenance, with duration running to 199 days at data cut-off and five still on drug 3. FDA's postmarketing requirement 4730-1 specifies that all patients should have at least 3 years of treatment 4. A cured child on maintenance is a different exposure from a dying adult on salvage, and the risk argument has to be made against the exposure that will exist, not the one that has existed.
Irreversibility is already on the label. The juvenile-rat femoral growth-plate closure was irreversible, as were the rat pancreatic, testicular and mammary hyperplasia, all at approximately 2x human AUC over 13 weeks - a duration comparable to a human treatment course 2. Whatever one thinks of rodent translation, the claim that 'the exposure is reversible so the effect must be' is not supported by the drug's own nonclinical package.
The bounding experiment has not been done in the relevant tissue - corrected 2026-09-19. This paragraph previously said that no conditional Men1 system with a defined off-then-on window had been located. That is withdrawn; three systems bear on it and none had been cited here.
- Off-then-on exists, in brain. Ulfat et al. deleted Men1 in hippocampal CA1 of a conditional knockout, produced a contextual fear-conditioning deficit, and then recovered the deficit by re-injecting to overexpress MEN1 5. That is a defined off-then-on window with a functional readout - in neurons, on a behavioural endpoint, with no titration of the off-interval.
- Restoration reverses an established endocrine lesion, in vitro. Reconstituting menin in Men1-deficient mouse Leydig tumour cell lines blocks the G0/G1-to-S transition, raises p18INK4C and p27Kip1 and induces apoptosis 6. Growth driven by menin loss is therefore still menin-reversible after transformation, at least in a dish.
- The onset side is bounded. Lines et al. built Men1-flox x RIP2-CreER mice and deleted menin in beta cells at ~3 months with 5 days of tamoxifen; by 2-2.5 months later islet area was already >4.2-fold increased with >2.3-fold increased beta-cell proliferation 7. So in mouse beta cells, complete menin loss produces measurable hyperplasia within roughly ten weeks - which is the same order as the 2.5-month median human exposure, and shorter than the 199-day maintenance exposures and the 3-year postmarketing requirement.
What is still missing, and what the paragraph should have said: no system varies the duration of menin absence and then restores it, in a menin-relevant endocrine tissue, to find the interval past which the lesion no longer regresses. Until that titration exists the threshold between 'harmless' and 'harmful' is unconstrained - but it is now flanked on both sides rather than unconstrained in both directions, and 7 makes the onset side uncomfortably fast relative to real exposures.
| kind | essential-function |
|---|---|
| severity | unknown; the question is whether any effect persists after discontinuation |
| mitigation | off-treatment endocrine sampling at 3, 6 and 12 months after the last dose - the single cheapest measurement that would convert this from an open question to an answered one |
| status of inference | genuinely unresolved; the balance of argument favours low risk |
| confidence note 2026 09 19 | Updated 2026-09-19: the onset side of the reversibility window is now bounded - complete menin loss in mouse beta cells gives >4.2-fold islet area within 2-2.5 months (PMID 28420716), an interval comparable to real human exposures - and off-then-on Men1 designs exist outside endocrine tissue (PMIDs 36552783, 17184987). |
affects
[ "drug-revumenib", "drug-ziftomenib", "drug-bleximenib" ]
arguments risk is small
[ "partial rather than complete loss", "one menin function rather than all", "reversible rather than permanent", "months rather than decades" ]
counterweights
[ "exposure distribution lengthening toward maintenance and front-line use; FDA PMR specifies at least 3 years", "irreversible nonclinical findings at 2x human AUC over 13 weeks", "no conditional-restoration experiment exists to bound the threshold" ]
current exposure — the numbers behind the chart
{
"median_months": 2.5,
"pct_beyond_6_months": 10,
"max_months": 40,
"post_transplant_maintenance_days_observed": 199
}restoration systems located
{
"off_then_on": [
{
"pmid": "36552783",
"tissue": "hippocampal CA1 neurons",
"readout": "contextual fear conditioning",
"titrated_interval": false,
"title": "Selective Menin Deletion in the Hippocampal CA1 Region Leads to Disruption of Contextual Memory in the MEN1 Conditional Knockout Mouse: Behavioral Restoration and Gain of Function following the Reintroduction of MEN1 Gene.",
"authors": "Ulfat AK, Batool S, Iqbal F, et al.",
"venue": "Cells",
"year": 2022,
"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": "17184987",
"tissue": "Men1-deficient Leydig tumour cell lines",
"readout": "cell cycle arrest, apoptosis",
"titrated_interval": false,
"title": "Reconstituted expression of menin in Men1-deficient mouse Leydig tumour cells induces cell cycle arrest and apoptosis.",
"authors": "Hussein N, Casse H, Fontanière S, et al.",
"venue": "Eur J Cancer",
"year": 2007,
"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."
}
],
"onset_bound": [
{
"pmid": "28420716",
"tissue": "mouse pancreatic beta cell",
"finding": "islet area >4.2-fold at 2-2.5 months after deletion",
"title": "A MEN1 pancreatic neuroendocrine tumour mouse model under temporal control.",
"authors": "Lines KE, Vas Nunes RP, Frost M, et al.",
"venue": "Endocr Connect",
"year": 2017,
"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."
}
],
"still_missing": "duration-titrated deletion-then-restoration in a menin-relevant endocrine tissue",
"recorded": "2026-09-19"
}- The honest bottom line favours the drug: partial, reversible, single-function, months-long occupancy is a far weaker perturbation than lifelong germline haploinsufficiency, which itself spares 93% of carriers to age 10. This entry should not be read as a warning.
- But the argument has a moving part. It is an argument about exposure duration, and exposure duration is increasing as the class moves into maintenance and front-line settings, with FDA's own requirement specifying three years.
- And it has an internal inconsistency that must be acknowledged: the drug's own nonclinical package contains irreversible findings after 13 weeks of dosing, which is the same order as a human treatment course. Reversible exposure does not entail reversible effect.
- Post-transplant maintenance is where chronic menin-inhibitor exposure will actually arise, and allogeneic transplantation in children independently causes thyroid dysfunction, gonadal failure, growth impairment and reduced bone mineral density. Any endocrine surveillance built on that population will measure the conditioning regimen unless it is designed not to — which is a design requirement this project had not written down.
- Do the PTH, phosphate and calcium abnormalities on the revumenib label normalise after discontinuation? Not one off-treatment value has been published for any menin inhibitor, and this is the cheapest and most decisive missing measurement in the whole beat.
- What exposure will a cured child actually receive once menin inhibitors move to maintenance and front-line combination? The risk argument depends entirely on this number and it is currently changing.
- Is there a duration threshold in mouse beyond which menin restoration no longer reverses islet or parathyroid hyperplasia? No restoration experiment was located.
- Does the twofold exposure difference between the CYP3A4-inhibitor and non-inhibitor revumenib arms produce a measurable difference in PTH? The data to answer this already exist in the sponsor database.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Age-related penetrance of MEN1 in germline mutation carriers is 7% at age 10 and 52% at age 20, and even in a systematically screened carrier cohort the earliest tumour manifestation across all MEN1 tumour types was 9 years.
Characterization of mutations in patients with multiple endocrine neoplasia type 1 · Bassett JH, Forbes SA, Pannett AA, Lloyd SE, Christie PT, Wooding C, Harding B, Besser GM, Edwards CR, Monson JP, Sampson J, Wass JA, Wheeler MH, Thakker RV · Am J Hum Genet 62:232-44 1998What this evidence cannot show
Read as abstract for the penetrance figures. Family-ascertained kindreds bias penetrance upward; the endpoint is any MEN1 manifestation. The 9-year earliest-manifestation figure comes from the separate Machens cohort and is quoted here as context, not from this source.
Evidence that argues against this (1)
- This is the comparison the entry is built on and it is not like-for-like. Penetrance at 10 and 20 years measures accumulated second-hit probability in tissue carrying germline monoallelic loss; exposure duration measures weeks of reversible occupancy on intact protein in someone with two normal alleles. Putting the two on one axis implies a shared clock that does not exist. American Journal of Human Genetics 1998
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
The revumenib safety population had a median exposure of 2.5 months (range under 1 to 40 months) with 10% exposed beyond 6 months; in the same programme, 13-week dosing in juvenile rats produced an irreversible increase in femur growth plate closure and irreversible hyperplasia in the pancreas, testes and mammary gland at approximately 2 times the human exposure.
REVUFORJ (revumenib) tablets, for oral use - US full prescribing information (NDA 218944) · Syndax Pharmaceuticals / US FDA · DailyMed / Drugs@FDA (NDA 218944) 2025What this evidence cannot show
Read verbatim from the DailyMed full prescribing information sections 6.1, 8.4 and 13.2 on 2026-09-18. The nonclinical irreversibility statements are about rat findings at supratherapeutic exposure over continuous dosing; human reversibility of the observed PTH, phosphate and calcium changes has not been reported at all, in either direction.
Evidence that argues against this (1)
- Median exposure of 2.5 months cuts BOTH ways and the entry uses only one edge. It is far too short for any germline-latency analogy to apply, and it is also too short to have detected a slow endocrine phenotype had one been forming - so it bounds what is known, not what is safe. The range matters more than the median: some patients reached 40 months, and 10% exceeded six, which is where a duration argument should focus. FDA Drugs@FDA 2024
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In the AUGMENT-101 phase II cohort, 14 of the responders received allogeneic transplant and seven resumed revumenib afterwards as maintenance, with post-transplant maintenance duration ranging from 1 to 199 days at data cut-off and five of the seven still on revumenib; the median treatment duration in the efficacy population was 10 weeks.
Menin Inhibition With Revumenib for KMT2A-Rearranged Relapsed or Refractory Acute Leukemia (AUGMENT-101) · Issa GC, Aldoss I, Thirman MJ, DiPersio J, Arellano M, Blachly JS, et al., Stein EM · J Clin Oncol 43:75-84 2024What this evidence cannot show
Read in full text via PubMed Central on 2026-09-18. Maintenance durations are censored at data cut-off and therefore understate eventual exposure; the number of patients is small (seven).
Evidence that argues against this (1)
- Post-transplant maintenance is exactly where chronic exposure will arise, and it is also the most confounded setting imaginable for an endocrine endpoint. Allogeneic transplantation in children causes thyroid dysfunction, gonadal failure, growth impairment and reduced bone mineral density on its own. Any endocrine finding in a maintenance cohort is attributable to the conditioning regimen before it is attributable to the drug, and a surveillance design that ignores this will produce an uninterpretable signal. Clinical Transplantation 2026
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
FDA postmarketing requirement 4730-1 for revumenib requires a clinical trial in which all patients should have at least 3 years of treatment or have discontinued earlier, to determine whether long-term use is associated with serious potential risk of endocrine-related disorders due to chronic menin inhibition.
NDA 218944 approval letter for REVUFORJ (revumenib), including postmarketing requirements 4730-1 to 4730-6 · US Food and Drug Administration, Center for Drug Evaluation and Research · Drugs@FDA, NDA 218944 approval letter 2024What this evidence cannot show
Text read verbatim from the NDA 218944 approval-letter PDF on 2026-09-18. The trial design is not public; whether it enrols children or collects off-treatment samples is unknown.
Evidence that argues against this (1)
- The same regulator granted a full, non-accelerated approval down to age 1 while this question was open. A postmarketing requirement expresses that a question is worth answering, not that the answer is expected to be bad, and it is evidence that the regulatory system judged the benefit-risk favourable for infants on the data in hand. Syndax Pharmaceuticals / US FDA 2025
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In a conditional MEN1 knockout mouse, AAV-driven MEN1 deletion restricted to hippocampal CA1 produced contextual fear-conditioning deficits, and re-injecting to overexpress MEN1 recovered the deficit - a defined deletion-then-restoration window with a functional readout.
Selective Menin Deletion in the Hippocampal CA1 Region Leads to Disruption of Contextual Memory in the MEN1 Conditional Knockout Mouse: Behavioral Restoration and Gain of Function following the Reintroduction of MEN1 Gene. · Ulfat AK, Batool S, Iqbal F, Syed NI · Cells 2022What this evidence cannot show
Neurons, a behavioural endpoint and viral overexpression rather than restoration of the endogenous allele; the off-interval is fixed, not titrated, and nothing about neoplasia is measured. Relevant only as proof that an off-then-on Men1 design has been built at all.
Evidence that argues against this (1)
- The hippocampal restoration result is in neurons on a behavioural readout with no tumour endpoint, so it bears on reversibility of a menin-dependent FUNCTION rather than on regression of a proliferative lesion. Using it in a duration entry implies an off-window generalises across tissue and endpoint, which it has not been shown to do. Cells 2022
- E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Reconstituting menin expression in Men1-deficient mouse Leydig tumour cell lines inhibits growth, blocks the G0/G1-to-S transition, raises p18INK4C and p27Kip1 and increases apoptosis.
Reconstituted expression of menin in Men1-deficient mouse Leydig tumour cells induces cell cycle arrest and apoptosis · Hussein N, Casse H, Fontaniere S, Morera AM, Asensio MJ, Bakeli S, Lu JL, Coste I, Di Clemente N, Bertolino P, Zhang CX · European Journal of Cancer 2007What this evidence cannot show
Cell lines derived from an established tumour, transfected re-expression, in vitro only, and a Leydig rather than parathyroid or islet lineage. It shows the growth program remains menin-responsive after transformation; it does not show an in vivo tumour regresses.
Evidence that argues against this (1)
- Restoration into an already-transformed Leydig line measures whether menin re-expression arrests growth in vitro, not whether a duration of inhibition is safe. If anything it argues the opposite of the entry's reassurance: the cells restored were already tumour cells, so restoration was studied after the event it is being used to bound. European Journal of Cancer 2007
- E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In Men1(L/L)/RIP2-CreER mice given tamoxifen for 5 days at about 3 months of age, pancreatic islets showed loss of menin in all islets, greater than 4.2-fold increased islet area and greater than 2.3-fold increased beta-cell proliferation by 2-2.5 months after deletion.
A MEN1 pancreatic neuroendocrine tumour mouse model under temporal control · Lines KE, Vas Nunes RP, Frost M, Yates CJ, Stevenson M, Thakker RV · Endocrine Connections 2017What this evidence cannot show
Complete biallelic deletion in mouse beta cells, not partial competitive pocket occupancy in human tissue, and hyperplasia is not a tumour. The comparison to human drug exposure is an order-of-magnitude argument, not an equivalence; mouse and human timescales differ. No restoration arm.
Evidence that argues against this (1)
- This is the closest model in the beat to a drug-like perturbation - temporally controlled rather than developmental - and two things limit it. The deletion is still COMPLETE loss of menin in all islets rather than partial occupancy of one pocket, and the induction uses tamoxifen, a selective oestrogen receptor modulator with endocrine activity of its own, which is an awkward confounder in a study whose endpoints are endocrine. Endocrine Connections 2017
2026-09-19: off-then-on Men1 systems exist (PMIDs 36552783, 17184987) and the onset side is bounded at ~10 weeks in mouse beta cells (PMID 28420716)
the titration this risk entry says is missing; both were narrowed 2026-09-19 because off-then-on Men1 designs exist outside endocrine tissue
- depends on →How long does menin have to be gone? Two months in a mouse islet, a decade or more in a human carrier
the drug exposure sits between the two anchors and is bounded by neither
- refutes →The same pocket suppresses JUND: menin inhibition's unmeasured endocrine liability in infants
the strongest counter-argument, stated at full strength: partial reversible single-function occupancy for months is not lifelong biallelic loss
the internal inconsistency: 13 weeks of dosing already produced irreversible findings at 2x human AUC
the regulator's own requirement specifies at least 3 years of treatment, which is the exposure the risk argument must be made against
the off-treatment arm at 3, 6 and 12 months is the only proposed source of reversibility data for this class
would produce the duration threshold the reversibility argument currently asserts without a number
the normal-tissue exposure argument is written with one KMT2 partner in it
The reversibility argument does not reach the growth plate: closure is irreversible in the animal study, so a short reversible exposure hitting an irreversible developmental window is not reversible in the sense that entry assumes
Makes the exposure-class rule explicit and gives it a table; that entry already frames the crux correctly and this one supplies the taxonomy