Oncofusion

start / graph / hyp-menin-inhibition-derepresses-lxr-target-genes-in-patients

Hypothesisspeculative confidence 1 of 5: Our own inference or analogy. Not shown to be true.Q78-interactome-absent

A menin inhibitor is a partial LXR agonist, and no trial in the class has looked

also called menin inhibitor lipid liability · LXR de-repression under revumenib · menin inhibitor metabolic endpoint

Menin switches off the genes that build fat and move cholesterol. If a drug removes menin, those genes should switch on. Every child on a menin inhibitor is running that experiment and nobody is measuring it.

The same thing, for a scientist

Menin is a corepressor of LXR-alpha: it binds LXR-alpha and LXR-beta, represses LXR transcription, and lowers SREBP-1c, FASN and SCD-1 in hepatocytes; independently, pharmacological and genetic menin loss **upregulates the LXR targets ABCG1 and ABCA1 and lowers cellular cholesterol** in colorectal cells and mouse colonic epithelium [E1][E2]. The prediction follows directly: **a clinical menin inhibitor, in any menin-expressing tissue, should de-repress LXR target genes**, and the class already has an approved agent in children. **No pivotal menin-inhibitor readout has published a lipid panel, a hepatic-fat measure or an LXR target-gene measurement** [E3]. The hypothesis is cheap to test on banked material and its falsifiers are unambiguous. Prior probability is set at 0.45 rather than higher because the same tissue holds a menin function of the opposite sign - menin co-activates PPAR-alpha and prevents steatosis [E4] - so the net direction in vivo is genuinely uncertain even if the LXR arm is correct.

Go deeper
Why this is testable now rather than aspirational

Three things already exist:

  1. The drug in patients. Revumenib is approved in relapsed/refractory KMT2A-rearranged acute leukaemia including children; ziftomenib and bleximenib are in registration programmes. Serum has been banked in all of them.
  2. A published pharmacological arm. The colorectal study used menin inhibition, not only Men1 deletion, and read out ABCG1 and ABCA1 by RNA-Seq plus a cholesterol assay 2. The readouts are defined and the direction is stated.
  3. A baseline confound that makes the measurement necessary rather than merely interesting. Children on these protocols receive asparaginase and corticosteroids, both of which independently drive hypertriglyceridaemia and hepatic steatosis. An additive LXR effect would be invisible in routine care precisely because the background is already abnormal - which is the argument for a designed comparison rather than a chart review.
What the hypothesis does NOT claim
  • It does not claim menin represses LXR in a blast. Every measurement is hepatic or intestinal, and LXR-alpha is not the isoform a lymphoid cell mainly expresses - LXR-beta is 1. The blast arm is the weakest part and is written as a separate prediction.
  • It does not claim a clinical harm. It claims an unmeasured pharmacodynamic consequence, which may be neutral, adverse, or even useful - LXR agonism has been pursued in AML on its own merits.
  • It does not claim the effect runs through the drug pocket. Menin inhibitors both occupy the MBM cavity and reduce menin protein, and the two routes predict different magnitudes. Which one operates is prediction 4.
The sibling prediction this beat separated out

The same reasoning applied to the 2026 menin-HR scaffold result gives a second off-programme prediction: if menin recruits SETD2 to gamma-H2A.X and thereby licenses LEDGF-CtIP end resection, menin inhibition could impair homologous recombination in a population receiving etoposide and anthracyclines. That is a different mechanism, a different menin surface and a different readout, so it is written as its own experiment (exp-hr-competence-under-menin-inhibition-in-kmt2a-rearranged-blasts) rather than folded in here. The two share only the shape: a menin function outside chromatin regulation that the drug class has never been asked about.

Facts and figures
statementClinical menin inhibition de-represses LXR target genes in menin-expressing tissues, producing a measurable pharmacodynamic signature - raised SREBF1/SREBP-1c, FASN, SCD-1, ABCA1 and ABCG1, and a change in serum lipids or hepatic fat - in patients receiving revumenib, ziftomenib or bleximenib.
originliterature-gap, found by resolving three MEN1-fished interactome partners to a single publication
prior probability0.45
would not be detected byroutine trial safety reporting, because asparaginase and corticosteroids already produce hypertriglyceridaemia in this population and an additive drug effect would be absorbed into the expected background
predictions
[
  "In banked plasma from a menin-inhibitor trial, on-treatment triglycerides or total cholesterol differ from baseline beyond what the concomitant asparaginase and corticosteroid exposure explains.",
  "In peripheral blood mononuclear cells or blasts sampled on treatment, ABCA1 and ABCG1 transcripts rise relative to the pre-dose sample from the same patient.",
  "In a hepatocyte line treated with revumenib at clinically achievable exposures, SREBF1, FASN and SCD-1 rise and the rise is abolished by an LXR antagonist.",
  "The effect tracks menin DEGRADATION rather than pocket occupancy: agents in this class differ in how much they lower menin protein, and the transcriptional effect should scale with the protein loss rather than with target engagement.",
  "The effect is LXR-beta-dependent in haematopoietic cells and LXR-alpha-dependent in hepatocytes, because those are the isoforms each tissue expresses."
]
experiments
[
  "exp-lxr-target-panel-in-banked-menin-inhibitor-samples"
]
analyses
[
  "ana-the-unwritten-interactome-partners-are-better-observed-than-the-written-ones"
]
falsifiers
[
  "A menin inhibitor at clinically achievable exposure that does NOT raise ABCA1 or ABCG1 in any menin-expressing human cell kills the hypothesis outright.",
  "A published lipid panel from any menin-inhibitor trial showing no on-treatment change beyond the backbone's expected effect kills the clinical arm, whatever the cell data show.",
  "Evidence that the menin-LXR-alpha contact is an overexpression artefact - for example failure to co-immunoprecipitate at endogenous levels in primary hepatocytes - removes the mechanism and leaves nothing to predict from.",
  "If menin's PPAR-alpha co-activation dominates in vivo, the net direction reverses and the hypothesis as stated is wrong even though the LXR arm is right."
]
How this could bite us back
  • The population is infants and young children who will be followed for decades, on a drug that may be given for months. A metabolic pharmacodynamic effect that nobody measures is not a small omission in that setting.
  • The opposite reading is that LXR de-repression is therapeutically useful - LXR agonism has been pursued in AML - so this could be an unrecognised contributor to efficacy rather than a toxicity. The hypothesis is agnostic and the measurement is the same either way.
  • This prediction rests on a single interaction paper plus one independent replication of the direction, neither in blood. Calling it a liability before the cell experiment is done would be the kind of overreach this project's rules exist to prevent.
Still open
  1. Do any of the three menin-inhibitor programmes already hold lipid data in a supplementary safety table? One look at the published supplements would either close this or sharpen it.
  2. Is LXR-beta - the haematopoietic isoform - repressed by menin functionally, or only bound by it? The whole blast arm of this hypothesis depends on an experiment nobody has run.
  3. Does the PPAR-alpha arm cancel the LXR arm in vivo? Both are from the same group in the same tissue and have never been measured together.
Every source behind this entry (4)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In HepG2 hepatoma cells and mouse liver, menin binds LXR-alpha and LXR-beta, inhibits LXR-alpha transcriptional activity and reduces the LXR target genes SREBP-1c, FASN and SCD-1, with the authors concluding menin is a novel corepressor of LXR-alpha and a negative regulator of hepatic lipogenesis.

    PMID 25962847DOI
    What this evidence cannot show

    Abstract read this session; full text is Elsevier and DOI resolution returns HTTP 403. HepG2 and mouse liver, transfection-based transcription assays; no haematopoietic system and no KMT2A fusion (brief rule 73). LXR-beta, the isoform relevant to a blast, has a co-IP but no functional arm in this paper.

    Evidence that argues against this (1)
    • The same group reports menin as a CO-ACTIVATOR of PPAR-alpha preventing liver steatosis, an opposite-sign lipid function in the same tissue, which is why this hypothesis's prior is set below 0.5. Cheng P et al. 2011
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In HT-29 and HCT-15 colorectal cancer lines and in Men1-floxed;Vil1-Cre mouse colonic epithelium, menin inhibition - pharmacological as well as genetic - upregulated the LXR target genes ABCG1 and ABCA1 and reduced total cellular cholesterol, and promoting cholesterol efflux by menin inhibition or LXR activation suppressed CRC growth under cholesterol depletion and with EGFR inhibitors.

    PMID 37627154DOI
    What this evidence cannot show

    Abstract read this session. Colorectal lines and mouse colon; no haematopoietic system, no KMT2A fusion (brief rule 73). The paper does not name which menin inhibitor was used at what exposure, so whether the concentrations correspond to clinical exposures is unknown. The Men1-floxed;Vil1-Cre mice had no distinct baseline phenotype, so the in vivo arm is a transcriptional and biochemical result rather than a phenotype.

    Evidence that argues against this (1)
    • The pharmacological arm - the part that makes this hypothesis testable rather than genetic - may not be measuring menin. Phillips et al. 2019 show that MI-2's antitumour activity is largely independent of its ability to target menin and that its direct molecular target is lanosterol synthase, producing suppressed cholesterol synthesis, the endogenous LXR ligand 24,25-epoxycholesterol, and cholesterol depletion. That is the same readout Nyul et al. report. Their abstract names neither the compound nor the concentration, so until it does, 'menin inhibition upregulated ABCG1 and ABCA1' cannot be separated from 'an MI-series compound inhibited lanosterol synthase', and the hypothesis's own prediction 3 - that an LXR antagonist abolishes the rise - would not distinguish them either. Phillips RE et al. 2019
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    Re-run on 2026-09-21, the menin-inhibitor drug-name list intersected with lipid, cholesterol, triglyceride and steatosis terms returns 6 PubMed records - NOT the zero this item originally recorded. Title-checked: two are on topic and neither was in this graph (Phillips 2019, MI-2 depletes cholesterol and generates the endogenous LXR ligand 24,25-epoxycholesterol by inhibiting lanosterol synthase; Cao 2026, intestinal-epithelium Men1 deletion and the menin inhibitor MI-463 limit lipid absorption through LXR-beta and Ces1g), one is the colorectal paper already cited at E2, and three match only on ordinary-word uses of 'lipid' (lipid nanoparticle, lipid bilayer, lipid droplet). The class HAS been asked about lipid handling, in mice; what is still unpublished is a lipid panel from a human menin-inhibitor trial.

    What this evidence cannot show

    CORRECTED 2026-09-21 by Q80: this item previously asserted 0 records for this query and the identical query string returns 6 today, so the counted absence it rested on was wrong. Two mechanics matter. (1) The field tag does not apply. PubMed's querytranslation for this exact string contains no [Title/Abstract] term at all - a tag written after a parenthesised OR group is dropped and the group expands to [All Fields] - so this was never the title-and-abstract query it was labelled, and is instead an all-fields query, which is BROADER. (2) Half the returns are ordinary-word contamination on 'lipid', which is why the count has to be title-checked and not quoted. What survives after the correction is a narrower and better-bounded absence: no lipid panel, hepatic-fat measure or LXR target-gene measurement from a HUMAN menin-inhibitor trial is reachable by this query. Supplementary trial safety tables were not searched (brief rule 80). Name list tried, per brief rule 72: revumenib, ziftomenib, bleximenib, SNDX-5613, KO-539, VTP-50469, MI-503, and menin inhibitor / inhibitors / inhibition as three separate strings; a widened re-run adding enzomenib, SNDX5613, KO539, VTP50469, MI503, MI-2-2, MI-538, DS-1594, DSP-5336, BMF-219 and BMF219 returns 8. Operand control (brief rule 67): the drug-name clause alone returns 361 and menin[tiab] alone 1,264.

    Evidence that argues against this (2)
    • The absence this item was built on is gone, and the paper that removes it points the other way from the hypothesis. Cao et al. 2026 report that intestinal-epithelium-specific Men1 knockout PROTECTS mice against high-fat-diet obesity - de-repressed Ces1g, more triglyceride hydrolysis, more fatty-acid beta-oxidation - and that the menin inhibitor MI-463 recapitulates it in mouse and in a human gut organoid-on-chip. So in the one whole-animal system where a menin inhibitor's lipid handling has been measured, the direction is metabolically FAVOURABLE, not the unmeasured liability this hypothesis frames. It also relocates the receptor: the intestinal contact is with LXR-BETA and HDAC1, not LXR-alpha. Cao X et al. 2026
    • The instrument was mislabelled, and in the direction that made the zero look stronger than it was. PubMed's querytranslation for the exact string in this citation contains no [Title/Abstract] term: the tag after a parenthesised OR group is dropped and the group expands to [All Fields]. A zero from an all-fields query is a stronger absence than a zero from a tiab query, so the error did not inflate the claim - but it means the entry's own caveat ('titles and abstracts only') described the wrong instrument, and any future re-run comparing against this baseline would be comparing two different queries. The bare form MI-503[tiab] does translate correctly, so the failure is specific to the parenthesised-group form used here. NCBI PubMed 2026
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Menin co-activates PPAR-alpha and prevents liver steatosis - a menin function of the opposite sign on lipid handling in the same tissue as the LXR result.

    PMID 22001208DOI
    What this evidence cannot show

    Abstract only; mouse liver and hepatoma lines. Recorded as the principal reason the net in vivo direction is uncertain, not as a refutation of the LXR arm.

    Evidence that argues against this (1)
    • Menin's nuclear-receptor coactivation is not PPAR-alpha-specific, and the only human germline evidence is a phenotype neither arm predicts: MEN1 mutation carriers show higher HOMA insulin resistance than matched controls (3.0 +/- 2.0 vs 2.0 +/- 1.0) and more impaired fasting glucose (17% vs 6%), in work that describes menin as a coactivator of PPAR-gamma and the vitamin D receptor. Setting this hypothesis's prior below 0.5 on the strength of one opposite-sign receptor understates the problem: menin coactivates several, and the human data are about glucose rather than lipid. van Wijk JP et al. 2012
Connected to