Experiment: read the endocrine limb and the marrow limb of menin-MLL in the SAME animals at the SAME exposure
also called paired-tissue menin inhibitor pharmacodynamics
Give mice a menin-blocking drug at a dose that works on leukaemia, then measure the drug's effect in the hormone glands and the bone marrow of the same animals. Nobody has done both at once, so nobody knows whether the helpful and the harmful effects happen at the same dose.
The same thing, for a scientist
Dose wild-type mice with revumenib or ziftomenib to a marrow-efficacious exposure for 14 days, then in every animal measure: (i) H3K4me3 CUT&RUN at the CDKN1B and CDKN2C promoters in pituitary, parathyroid and islet; (ii) Cdkn1b and Cdkn2c transcript in the same tissues; (iii) Hoxa9 and Meis1 transcript and menin occupancy in marrow; (iv) serum intact PTH and corrected calcium. The design exists because the two limbs of hyp-one-mark-two-opposite-cell-state-programmes have only ever been measured in different papers, different species and different exposures, and because the bounded literature search behind absence-no-cdk-inhibitor-chromatin-readout-under-a-menin-inhibitor-in-endocrine-tissue returned zero endocrine measurements.
Design
- Animals. Wild-type C57BL/6, n = 8 per arm, both sexes. Arms: vehicle; menin inhibitor at the exposure used in published KMT2A-r efficacy work; menin inhibitor at one third of that exposure. The low arm exists because the hypothesis predicts the endocrine effect tracks module dosage sensitivity, so a dose-response is the discriminating measurement, not a single dose.
- Duration. 14 days. Chosen to sit inside the window in which the revumenib label already reports intact PTH increases in 34% of 241 patients at a median exposure of 2.5 months, so a negative at 14 days is informative about onset rather than about the mechanism.
- Primary readout. H3K4me3 CUT&RUN at Cdkn1b and Cdkn2c promoters in pituitary. Pituitary is primary because it has the highest baseline module expression in the one survey that measured it (p27 approx 12x, MLL approx 17x liver) and therefore the largest predicted effect.
- Paired marrow readout. Hoxa9 and Meis1 transcript and menin CUT&RUN in the same animals, to establish that the exposure was efficacious. Without this the experiment cannot distinguish a negative result from an underdose, which is the failure mode that has made previous single-tissue studies uninformative.
- Serum. Intact PTH and corrected calcium at day 0, 7 and 14, because the clinical signal is biochemical and this is the cheapest bridge between the mouse and the label.
Controls and what each one is for
- Positive chromatin control: H3K4me3 at a housekeeping promoter, expected unchanged, to show the CUT&RUN is not globally depleted.
- Positive drug control: the marrow arm above.
- Antibody control: an H3K4me1 and an H3K4me2 channel alongside me3, because the founding endocrine papers say only 'H3 lysine 4 methylation' and 2005-era antibodies do not discriminate states. If the endocrine effect is on me1 at distal elements, an me3-only design returns a false negative.
- Tissue control: liver, in which mouse MLL, p27 and p18 are low and in which MLL heterozygosity does not move p27 - the predicted null tissue.
What each outcome means
| result | reading |
|---|---|
| endocrine H3K4me3 and transcript both fall, marrow limb moves | hypothesis supported; endocrine CDKN1B becomes a pharmacodynamic biomarker for the liability |
| endocrine unchanged, marrow limb moves | hypothesis falsified as stated; the two limbs dissociate at a single exposure |
| endocrine transcript RISES | the osteosarcoma result generalises; the graph's endocrine risk model has the sign wrong and must be rewritten |
| nothing moves anywhere | underdose; the experiment failed and reports nothing |
| design | 3-arm dose-response in wild-type mice (vehicle, efficacious menin inhibitor exposure, one-third exposure), n=8 per arm, 14 days, both sexes |
|---|---|
| readout | H3K4me1/me2/me3 CUT&RUN at Cdkn1b and Cdkn2c promoters in pituitary, parathyroid and islet; Cdkn1b/Cdkn2c transcript in the same tissues; Hoxa9/Meis1 transcript and menin CUT&RUN in paired marrow; serum intact PTH and corrected calcium at days 0, 7, 14 |
| controls | housekeeping-promoter H3K4me3 (global depletion control); paired marrow arm (exposure control, without which a negative is uninterpretable); three methylation-state channels (the founding papers do not resolve the state); liver as predicted null tissue |
| cost tier | med |
| timeline | about 4 months including CUT&RUN optimisation in parathyroid, which is the hard tissue because of its size |
| tests hypothesis | hyp-one-mark-two-opposite-cell-state-programmes |
| why it has not been done | menin inhibitor pharmacology is developed by haematology groups with marrow readouts and endocrine safety is monitored by serum chemistry, so the chromatin question falls between the two. The bounded literature search found 3 records pairing any named menin inhibitor with parathyroid, pituitary or islet and none of them measures a CDK inhibitor. |
| failure mode to avoid | running the endocrine arm without the paired marrow arm. A negative would then be indistinguishable from an underdose, which is the specific reason this graph cannot currently interpret the existing endocrine literature. |
- A clean negative in mouse would not exonerate the class in humans, because the human clinical signal (intact PTH up in 34% of patients) is already positive and is the thing the mouse would be failing to reproduce.
- Can CUT&RUN be made to work on mouse parathyroid? The gland is tiny and this is the practical obstacle; if not, laser-capture plus low-input CUT&RUN or a parathyroid organoid is the fallback and it changes the cost tier.
- Should the model carry a KMT2A fusion? A wild-type mouse isolates the endocrine question; a fusion-bearing mouse answers the clinically relevant version, in which the fusion is competing for menin. Both are worth doing and the wild-type one is cheaper and comes first.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Menin associates with the p27 and p18 promoters and increases H3 lysine 4 methylation there in vivo in mouse islets, and that methylation is reduced in Men1-mutant islet tumours - establishing that the proposed endocrine readout is measurable in the proposed tissue by the proposed assay.
Menin regulates pancreatic islet growth by promoting histone methylation and expression of genes encoding p27Kip1 and p18INK4c. · Karnik SK, Hughes CM, Gu X, Rozenblatt-Rosen O, McLean GW, Xiong Y, Meyerson M, Kim SK · Proc Natl Acad Sci U S A 2005What this evidence cannot show
Genetic Men1 loss, not drug; islet tumours, not normal tissue; abstract read only. The relevance here is feasibility of the assay in the tissue, not the pharmacology.
Evidence that argues against this (1)
- The proposed readout may be the wrong one. Neither pharmacological study of this axis measured chromatin, and one of them raised p27 while H3K4 methylation fell - so H3K4me3 at CDKN1B and transcript at CDKN1B may not move together, and an experiment scored on the mark alone could return an uninterpretable result. Tian S et al. 2025
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In mouse, p27 and MLL mRNA in pituitary are approximately 12-fold and 17-fold higher than in liver, and MLL heterozygosity significantly lowers p27 in pituitary and p27 and p18 in islets but not in liver - which is the basis for choosing pituitary as the primary tissue and liver as the null tissue.
Haploinsufficient and predominant expression of multiple endocrine neoplasia type 1 (MEN1)-related genes, MLL, p27Kip1 and p18Ink4C in endocrine organs. · Taguchi R, Yamada M, Horiguchi K, Tomaru T, Ozawa A, Shibusawa N, Hashimoto K, Okada S, Satoh T, Mori M · Biochem Biophys Res Commun 2011What this evidence cannot show
Single paper, mouse mRNA, abstract read only; whether parathyroid was in the panel cannot be determined from the abstract, which is a real problem for a design that puts parathyroid in the tissue list.
Evidence that argues against this (1)
- Pituitary was chosen as the primary tissue on a single mouse mRNA survey. If the human module's tissue ranking differs - and human MEN1 is 'Low tissue specificity, Detected in all' with no endocrine bias at all - then the tissue with the largest predicted effect may not be pituitary, and the design would be optimised against a mouse-specific ranking. Human Protein Atlas consortium 2026
the paired-tissue experiment that scores the hypothesis' falsifiers
the clinical molecule the design would dose
the experiment designed to close it