Hypothesis: the PTH rise on menin inhibitors is on-target and exposure-dependent, not renal
also called menin PTH hypothesis · on-target parathyroid hypothesis
We think the rise in parathyroid hormone seen in a third of patients on revumenib is caused by the drug hitting menin in the parathyroid gland itself, not by the kidneys. If so, patients with higher drug levels should show bigger rises, and the effect should fade after the drug stops.
The same thing, for a scientist
The proposal is that intact PTH elevation in 34% of revumenib-treated patients reflects on-target relief of menin-dependent repression in parathyroid chief cells, rather than secondary hyperparathyroidism from renal impairment or phosphate retention. Three discriminating predictions follow, all testable against data that already exist or against samples that are cheap to collect. The rival hypothesis - that this is ordinary secondary hyperparathyroidism in sick patients - is well supported by the same label's 38% rate of rising creatinine and 51% rate of rising phosphate, and is the null this hypothesis has to beat. **Addition, 2026-09-19:** the hypothesis now has a named mechanism it previously lacked. Menin is required for TGF-beta to inhibit parathyroid cell proliferation **and PTH secretion**; menin inactivation abolishes both responses in primary human parathyroid cells, and cells from a MEN1 patient do not respond to TGF-beta at all [E5]. Because the secretory arm does not require growth, an on-target PTH rise is mechanistically available within a 2.5-month exposure without any adenoma - which removes the main objection to the on-target reading. The remaining objection is biochemical and is not removed: the mouse Men1 parathyroid phenotype is hypercalcaemia **with hypophosphataemia** [E6], whereas on revumenib phosphate **rises** in 51%, which is the renal pattern rather than the primary-hyperparathyroid one.
Predictions that separate the two explanations
| Prediction | On-target menin effect | Renal/phosphate secondary effect |
|---|---|---|
| PTH rise versus revumenib AUC | positive, monotonic | none, once creatinine is adjusted for |
| PTH rise in the 2x-exposure dosing arm (without strong CYP3A4 inhibitor) | larger | no difference |
| Corrected calcium in patients with raised PTH | normal or high | low |
| PTH after discontinuation | falls over weeks | falls with renal recovery |
| Ziftomenib, if PTH were measured | also raised | depends on its own renal profile |
| JunD co-immunoprecipitation with menin in a parathyroid line under drug | lost | unchanged |
The third row is the one the published label already half-answers: corrected calcium rose in 15% of patients, which is the wrong direction for pure secondary hyperparathyroidism. What is missing is the cross-tabulation - how often did PTH and calcium rise in the same patient.
Update 2026-09-19 on the ziftomenib row. Ziftomenib is now approved (KOMZIFTI, label revised 11/2025) and its prescribing information does not mention parathyroid, parathyroid hormone or hyperparathyroidism at all, has no Endocrine disorders heading, and lists neither calcium nor phosphate nor PTH in its laboratory-abnormality table [counter-evidence to E1]. That is not a negative result - the analytes appear not to have been selected for labelling - but it means the row cannot be scored, the class-effect prediction has no corroboration, and the revumenib finding stands alone.
A missing row. The one endocrine tissue with a demonstrated in-vivo pharmacological menin-inhibition phenotype is the pancreatic islet, where 12 weeks at 70 mg/kg/day improved glucose tolerance and raised beta-cell proliferation [counter-evidence to E2]. If menin inhibition perturbs endocrine axes at clinical exposure, glucose and insulin should move too, and probably first. That prediction has been added to data.predictions and is cheaper to test than any PTH analysis.
Why this is worth testing rather than assuming
Because the cheap answer settles a disproportionately expensive question. If the PTH rise is renal, the endocrine late-effects concern loses most of its clinical anchor and the beat should shrink to a mechanistic curiosity. If it is exposure-dependent and on-target, then a labelled 34% biochemical effect in the exact axis MEN1 hits first, in a drug given to one-year-olds, becomes a surveillance obligation - and one that FDA's postmarketing requirement 4730-1 does not fully cover, because that requirement names no paediatric endpoint and no off-treatment sampling.
Every discriminating measurement above is either already in a sponsor database (AUC versus PTH; dosing-arm comparison; PTH-calcium cross-tabulation) or costs one assay (JunD co-IP; off-treatment PTH).
Addition, 2026-09-19: the missing mechanism, and the biochemical mismatch that remains
What is new in the hypothesis's favour. Menin is required for TGF-beta signalling in the parathyroid, and the TGF-beta effect it enables is inhibition of both proliferation and PTH secretion; menin antisense abolishes both, and parathyroid cells from a MEN1 patient are unresponsive to TGF-beta 5. This is the first mechanism cited anywhere in this beat that operates on a timescale compatible with the observed exposure. An on-target PTH rise no longer requires anyone to believe a tumour-suppressor phenotype appeared in ten weeks.
What still argues against it. Men1(+/-) mice with parathyroid neoplasia are hypercalcaemic and hypophosphataemic with inappropriately normal PTH 6. On revumenib, phosphate is the single most frequent abnormality and it goes up (51%), with creatinine up in 38% and calcium up in only 15%. Raised PTH with raised phosphate and raised creatinine is the textbook picture of renal secondary hyperparathyroidism. The hypothesis has to explain the phosphate direction and currently does not.
The test is unchanged and is cheap. ana-paediatric-subgroup-and-covariate-reanalysis-menin-inhibitor-labs sets out the exposure-response regression that separates the two, on data that already exists. exp-clinical-menin-inhibitor-on-primary-parathyroid-cells sets out the one bench experiment that has never been done - no menin inhibitor has ever been applied to a parathyroid cell, verified by a title/abstract search returning 0 records against 118 and 167 for the two halves.
| statement | Intact PTH elevation during menin-inhibitor treatment is an on-target consequence of relieving menin-dependent repression in parathyroid chief cells, and is therefore exposure-dependent, accompanied by normal or high calcium rather than low, reversible on discontinuation, and reproducible with a second menin inhibitor if PTH is measured. |
|---|---|
| origin | literature-gap |
| prior probability | 0.25 |
| prior probability note | Lowered from 0.35 to 0.25 on 2026-09-19. Two findings moved it down: the only other approved menin inhibitor's label reports no PTH, calcium or phosphate and has no Endocrine disorders heading, so there is no cross-compound corroboration; and the one endocrine tissue with a demonstrated in-vivo pharmacological phenotype is the islet, where 12 weeks of menin inhibitor improved rather than degraded function. Nothing found refutes the hypothesis; the confounding null simply got stronger. |
predictions
[ "PTH rise correlates positively with revumenib AUC across patients", "PTH rise is larger in the higher-exposure dosing arm (no strong CYP3A4 inhibitor)", "Patients with raised PTH have normal or high corrected calcium, not low", "PTH falls toward baseline within weeks of discontinuation", "Ziftomenib and bleximenib show the same effect when PTH is measured", "A clinical menin inhibitor abolishes menin-JunD co-immunoprecipitation in a parathyroid-derived cell line at clinically achievable concentrations", "Fasting glucose and insulin move before PTH does, because the islet - not the parathyroid - is the endocrine tissue where pharmacological menin inhibition has a demonstrated in-vivo phenotype (PMID 33821572)" ]
experiments
[ "exp-endocrine-surveillance-panel-menin-treated-children", "exp-jund-displacement-coip-under-clinical-menin-inhibitor", "exp-clinical-menin-inhibitor-on-primary-parathyroid-cells" ]
analyses
[ "ana-faers-menin-inhibitor-endocrine-disproportionality", "ana-paediatric-subgroup-and-covariate-reanalysis-menin-inhibitor-labs" ]
falsifiers
[ "PTH elevation shows no relationship to revumenib AUC after adjustment for creatinine and phosphate, which would make it secondary and close the question", "Patients with raised PTH have LOW corrected calcium, the signature of secondary rather than primary hyperparathyroidism", "A clinical menin inhibitor at achievable concentration does not displace JunD from menin in cells, which removes the mechanism entirely", "Bleximenib and ziftomenib, with PTH prospectively measured, show no PTH elevation, making the revumenib finding compound-specific rather than class- and target-related", "The PTH elevation vanishes entirely in the subgroup whose creatinine and phosphate remained stable throughout treatment, which would place the whole effect downstream of renal handling", "PTH rise persists after adjustment for differentiation syndrome, corticosteroid exposure and intercurrent infection - or, conversely, disappears once those covariates are carried, which would place it in the clinical course rather than in the drug (AUGMENT-101 reported grade >= 3 differentiation syndrome in 16.0% and febrile neutropenia in 37.2% of 94 treated patients, PMID 39121437)", "A clinical menin inhibitor produces no change in TGF-beta-suppressed PTH secretion in primary human parathyroid cells at achievable concentrations" ]
- The null hypothesis is strong and probably right for most patients: creatinine rose in 38% and phosphate in 51% of the same population, and both raise PTH by ordinary physiology without any menin involvement.
- There is now a second approved menin inhibitor and it provides no support at all: the KOMZIFTI (ziftomenib) label, revised 11/2025, contains no parathyroid mention, no Endocrine disorders heading and no calcium, phosphate or PTH row in its laboratory-abnormality table. The revumenib finding stands alone, and a single-compound biochemical signal in an unblinded, uncontrolled registration cohort is a weak foundation for a class-level mechanism.
- The FAERS analysis run for this beat found zero endocrine reports across 904 revumenib exposures, which argues against any clinically apparent version of this effect on current exposure durations.
- If this hypothesis is confirmed, the correct response is surveillance and possibly a calcium-sparing schedule, not withholding a drug that produces remissions in children who would otherwise die.
- The hypothesis may be aimed at the wrong gland. The endocrine tissue in which pharmacological menin inhibition has a demonstrated in-vivo phenotype is the pancreatic islet (PMID 33821572), where the effect was compensatory proliferation with improved glucose tolerance over 12 weeks. If the beat's attention follows the evidence rather than the MEN1 penetrance table, the first surveillance analyte is glucose, not PTH.
- This beat argues a pharmacological risk almost entirely from GERMLINE MEN1 penetrance data. Those cohorts measure the cumulative probability of a second somatic hit in a tissue that has carried monoallelic loss since conception, over decades. A treated child has two intact alleles and reversible occupancy of one pocket for weeks. The one genuinely pharmacological observation in the beat is the label's intact PTH increase in 34% — which is a real signal, a laboratory one, and unattributed.
- How often did PTH and corrected calcium rise in the same patient? One cross-tabulation from an existing database would move this hypothesis substantially in one direction or the other.
- Is there an exposure-response relationship between revumenib AUC and PTH? The programme measured AUC in every patient.
- Would a parathyroid-derived or islet-derived cell line show loss of menin-JunD co-immunoprecipitation under clinical concentrations of revumenib? This is a one-week experiment nobody has published.
- Is the effect present with ziftomenib and bleximenib, or specific to revumenib's chemotype?
- Does the revumenib PTH signal survive adjustment for differentiation syndrome and corticosteroid exposure? Differentiation syndrome occurred at grade >= 3 in 16.0% of the AUGMENT-101 phase 2 population (PMID 39121437) and is managed with systemic corticosteroids, which independently perturb calcium, bone turnover and the GH/IGF-1 axis. No published analysis carries that covariate.
- Were PTH, calcium and phosphate measured at all in the ziftomenib programme? The KOMZIFTI label reports neither, and neither a null result nor an absent assay can be distinguished from the label. One question to the sponsor or one FOIA of the clinical review would settle whether prediction 5 has already been answered.
- Why does phosphate rise rather than fall, if the parathyroid effect is primary? Either the renal arm dominates the phosphate column or the on-target effect is secretory-only without the renal phosphate-wasting that established hyperparathyroidism produces.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
The revumenib label reports intact PTH increased in 34%, corrected calcium increased in 15% with 0% grade 3-4, phosphate increased in 51%, phosphate decreased in 25% and creatinine increased in 38% of 241 patients, and lists hyperparathyroidism under Endocrine disorders.
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 DailyMed on 2026-09-18. The label gives no cross-tabulation of PTH with calcium or creatinine, no exposure-response analysis and no off-treatment values, which is why every prediction in this hypothesis is currently untested rather than refuted.
Evidence that argues against this (2)
- The confounding null is stronger than this entry allows, because the registration cohort is a population with frequent acute inflammatory and renal insult. In AUGMENT-101 phase 2 (N = 94 treated, median age 37, range 1.3-75), grade >= 3 adverse events included differentiation syndrome in 16.0% and febrile neutropenia in 37.2%. Differentiation syndrome is managed with systemic corticosteroids per the same label cited in E1, and corticosteroids, sepsis, volume shifts and the label's own 38% creatinine rise and 51% phosphate rise each move PTH by ordinary physiology. Any exposure-response analysis of PTH against revumenib AUC that does not carry differentiation syndrome, corticosteroid exposure and intercurrent infection as covariates will attribute to menin what belongs to the clinical course - and none of the predictions in this hypothesis currently specifies those covariates. Issa GC et al. 2025
- Prediction 5 ('ziftomenib and bleximenib show the same effect when PTH is measured') has a checkable partial answer and it is unfavourable. Ziftomenib was approved in November 2025 as KOMZIFTI for relapsed/refractory NPM1-mutant AML, and its full prescribing information contains NO mention of parathyroid, parathyroid hormone or hyperparathyroidism, has no Endocrine disorders heading in the adverse-reactions section, and its selected laboratory-abnormality table (N approximately 108-110) reports creatinine increased 45% (grade 3-4: 4%) and alkaline phosphatase increased 41% (5%) but lists neither calcium nor phosphate nor PTH. The honest reading is that these analytes were not selected for labelling rather than measured and found normal - but it means the only other approved menin inhibitor provides no corroboration of a class effect, and the revumenib PTH finding currently stands alone. The same label carries no growth-plate or juvenile-animal statement either, which is consistent with its adult-only indication rather than with absence of the finding. Kura Oncology / Kyowa Kirin / US FDA 2025
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Menin binds JunD and MLL1 in the same deep pocket in the same manner with opposite transcriptional effects, several MEN1 missense mutations disrupt the menin-JUND interaction, and menin loss or inactivation causes multiple endocrine neoplasia type 1, of which primary hyperparathyroidism is the first and commonest manifestation.
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. Provides the mechanistic rationale for an on-target parathyroid effect; tests no inhibitor and no endocrine tissue.
Evidence that argues against this (1)
- The one endocrine tissue where pharmacological menin inhibition has actually been characterised in vivo is the pancreatic islet, not the parathyroid, and the reported phenotype runs the opposite way from this hypothesis's implied harm. Ma et al. show MI-503 and MI-463 disrupting menin-JunD binding by co-IP at 30-300 nM, MI-503 at 10-100 nM raising Ki67 in primary human islets, and a menin inhibitor at 70 mg/kg daily by gavage lowering fasting glucose and improving glucose tolerance in high-fat-diet mice from week 6 through week 12. Menin inhibition has in fact been pursued as a candidate diabetes therapy on exactly this basis. Two consequences for this hypothesis: the mechanism it invokes is real and pharmacologically reachable, but the tissue it predicts is not the one with evidence, and the demonstrated endocrine effect at 12 weeks is compensatory proliferation with improved function rather than dysregulation. A parallel prediction - that fasting glucose and insulin should move before PTH does - follows and is not in the current prediction list. Ma J et al. 2021
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Parathyroid hyperplasia and adenoma developed in 77% of 258 systematically screened MEN1 carriers with maximum age-related penetrance around 90%, making the parathyroid the first and commonest MEN1 target organ.
Age-related penetrance of endocrine tumours in multiple endocrine neoplasia type 1 (MEN1): a multicentre study of 258 gene carriers. · Machens A, Schaaf L, Karges W, Frank-Raue K, Bartsch DK, Rothmund M, Schneyer U, Goretzki P, Raue F, Dralle H · Clin Endocrinol (Oxf) 67:613-22 2007What this evidence cannot show
Read as abstract. Establishes that the parathyroid is the right organ to predict; says nothing about drug exposure.
Evidence that argues against this (1)
- Parathyroid penetrance approaching 90% in germline carriers makes the parathyroid the most likely tissue to show something, which is not the same as making it likely to show something under a drug. The hypothesis needs a dose-dependent pharmacological effect; the penetrance data describe the endpoint of a decades-long two-hit process. Clinical Endocrinology 2007
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
Across 904 FAERS reports naming revumenib, no report carried any of 20 prespecified endocrine or bone-mineral MedDRA terms, including hyperparathyroidism and blood parathyroid hormone increased, while differentiation syndrome and QTc prolongation were heavily over-represented.
openFDA drug/event API query of FAERS for menin-inhibitor endocrine and bone-mineral adverse event terms · US Food and Drug Administration (data); analysis run for this graph, analyses/faers-menin-inhibitor-endocrine/run.py · openFDA FAERS, queried 2026-09-18 2026What this evidence cannot show
Run by this agent on 2026-09-18; see analyses/faers-menin-inhibitor-endocrine/. This is the strongest available evidence against the hypothesis producing clinically apparent disease, but it is uninformative about asymptomatic biochemical effects, since the label's own 34% PTH finding generated zero reports.
Evidence that argues against this (1)
- A zero across 904 FAERS reports is close to uninformative here, for three compounding reasons: spontaneous reporting captures a small and unquantified fraction of events; endocrine laboratory drift is exactly the class least likely to be reported spontaneously by an oncology prescriber; and the FDA itself determined that spontaneous postmarketing reports would NOT be sufficient to assess this signal, which is why a postmarketing trial was required instead. The label meanwhile reports intact PTH increased in 34%. The absence in FAERS and the presence on the label are not in conflict - they are measuring different things. FDA Drugs@FDA 2024
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Menin is required for TGF-beta to inhibit parathyroid cell proliferation and PTH secretion; menin inactivation abolishes both, and MEN1 patient parathyroid cells are TGF-beta-unresponsive.
Menin inactivation leads to loss of transforming growth factor beta inhibition of parathyroid cell proliferation and parathyroid hormone secretion. · Sowa H, Kaji H, Kitazawa R, Kitazawa S, Tsukamoto T, Yano S, et al. · Cancer Research 2004What this evidence cannot show
Primary cells from secondary hyperparathyroidism, antisense knockdown not pocket inhibition, single MEN1 patient comparator, 2004 methodology.
Evidence that argues against this (1)
- Knockdown removes the protein; a pocket inhibitor may leave the TGF-beta-relevant surface intact, so the mechanism may not be drug-accessible at all. Ma J et al. 2021
- E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Men1(+/-) mice with parathyroid neoplasia are hypercalcaemic and hypophosphataemic with inappropriately normal PTH - the opposite phosphate direction to the human drug data.
Multiple endocrine neoplasia type 1 knockout mice develop parathyroid, pancreatic, pituitary and adrenal tumours with hypercalcaemia, hypophosphataemia and hypercorticosteronaemia. · Harding B, Lemos MC, Reed AA, Walls GV, Jeyabalan J, Bowl MR, et al. · Endocrine-Related Cancer 2009What this evidence cannot show
Germline heterozygous mouse over 3-21 months; 'inappropriately normal' PTH complicates direct comparison with a human PTH-rise readout.
Evidence that argues against this (1)
- A 2.5-month drug exposure may produce secretory disinhibition without the gland autonomy that drives frank hypercalcaemia and hypophosphataemia, so the mismatch is not decisive. Sowa H et al. 2004
proposes the discriminating measurements that separate on-target from renal explanations
the pharmacovigilance arm of the test, already run
the hypothesis requires that a clinical inhibitor actually displaces JunD, which has never been shown
- depends on →Primary hyperparathyroidism in MEN1: the first and commonest consequence of losing menin
the parathyroid is the predicted organ because it is the first and commonest MEN1 target
a negative result would falsify the hypothesis outright in about a week
Counter-evidence (KOMZIFTI label, NDA 220305, rev. 11/2025): the second approved menin inhibitor's prescribing information contains no parathyroid mention, no Endocrine disorders heading and no calcium, phosphate or PTH row. Prediction 5 (class effect) has no corroboration; the revumenib PTH finding stands alone.
Counter-evidence (Issa 2025, PMID 39121437): grade >=3 differentiation syndrome in 16.0% and febrile neutropenia in 37.2% of 94 AUGMENT-101 patients. Corticosteroids, sepsis and volume shifts move PTH by ordinary physiology; no prediction in this hypothesis carries those covariates.
Counter-evidence (Ma 2021, PMID 33821572): the endocrine tissue with a demonstrated in-vivo pharmacological menin-inhibition phenotype is the pancreatic islet, not the parathyroid, and at 12 weeks the effect was compensatory beta-cell proliferation with IMPROVED glucose tolerance. Prior probability lowered 0.35 -> 0.25.
Supplies the mechanism the hypothesis lacked: menin is required for TGF-beta to restrain PTH SECRETION, not only parathyroid growth, so a PTH rise within a 2.5-month exposure is mechanistically available without any adenoma
Generalises the parathyroid-only hypothesis to a tissue-divergent one, and shares its central discriminator: does the PTH rise track drug exposure after adjustment for creatinine and phosphate