Oncofusion

start / graph / risk-q139-menin-displacement-is-a-poor-pharmacodynamic-proxy-for-disabling-the-fusion

Riskmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q139-fusion-degron

Menin displacement is a poor pharmacodynamic proxy for disabling the fusion, and the degron shows both halves of why: removing the fusion barely moves menin genome-wide, while the menin inhibitor that moves menin leaves half to three-quarters of the fusion's own chromatin sites occupied

also called menin displacement is not target engagement · the PD proxy fails both ways · 51.7% of the fusion still there

The natural way to check that a menin drug is working is to measure whether menin has come off DNA, and two results here say that is the wrong check. Destroying the cancer protein hardly moves menin at all, while the drug that does move menin leaves most of the cancer protein still attached - so the thing you can measure and the thing you care about come apart in both directions.

The same thing, for a scientist

**The risk.** Using menin chromatin displacement as a readout that a menin inhibitor has disabled the fusion oncoprotein. The degron deposit breaks the link **in both directions within one experiment**: | direction | measurement | consequence | |---|---|---| | **fusion gone, menin stays** | degrading MLL::AF9 leaves menin at **80.0%** of its strongest peak calls at 3 h, while the fusion's own tag is at **2.5-7.1%** [E1] | menin occupancy does not report on fusion occupancy | | **menin moved, fusion stays** | VTP-50469 100 nM for 4 d leaves **51.7%** (batch 200908) and **76.9%** (batch 1912x2) of the fusion's own HA sites occupied [E2] | menin displacement does not imply fusion removal | **Neither direction is a subtle effect.** They are the two largest dissociations in the beat. **Why this is a risk entry and not just an analysis.** Menin displacement is the **cheapest and most natural** pharmacodynamic assay for a menin inhibitor — it measures the drug doing the thing its mechanism names. That is exactly what makes it seductive and what makes the failure mode dangerous: a trial or preclinical study can demonstrate complete, dose-proportional menin displacement and have demonstrated nothing about whether the oncoprotein has been evicted. **The degron supplies the missing anchor.** Against total removal (1.0% retention), the menin inhibitor alone reaches roughly half, the DOT1L inhibitor about a quarter, and only the **combination** approaches removal at 10.7% [E2]. **The readout that tracks the thing you care about is occupancy of the fusion itself**, not of its adaptor — and in a patient that is far harder to measure, which is precisely why the proxy is used.

Go deeper
Mitigation, in order of practicality
  1. Read the fusion, not the adaptor, wherever possible. In tagged preclinical systems this is free. In primary samples it requires an antibody against the fusion's own epitopes or a CUT&RUN against the N-terminal MLL1 region, with the caveat that such an antibody also reads wild-type MLL1.
  2. Use a transcriptional readout as the primary PD marker. The authors show elongation changes within 15 minutes of degradation, well before chromatin changes, so target-gene expression is both faster and closer to the biology than any occupancy measure.
  3. If menin displacement must be used, state what it is a proxy for. It is a proxy for drug-target engagement at the menin pocket, which is a legitimate and useful thing to measure. It is not a proxy for oncoprotein eviction.
  4. Do not read a histone mark as the PD marker. No active mark in this deposit loses peak calls genome-wide within 24 h of the fusion being destroyed entirely, so a mark-based assay would report nothing under a perfect drug.
Facts and figures
kindcompensation
severitymoderate - it does not predict clinical failure, it predicts that a commonly used pharmacodynamic readout will overstate target disablement
mitigationRead fusion occupancy or target-gene transcription as the primary pharmacodynamic marker; use menin displacement only as a statement about pocket engagement and label it as such; do not use a bulk histone mark as a PD readout, because no active mark loses peak calls genome-wide within 24 h even when the fusion is destroyed outright.
affects
[
  "drug-revumenib",
  "drug-ziftomenib"
]
How this could bite us back
  • This risk cuts against the project's own instinct to reach for occupancy data, which is abundant and public, over transcriptional data, which is what the authors show moves first. The graph's eviction thread is large and its elongation thread is smaller, and this entry suggests that ratio is the wrong way round.
  • It also weakens any argument for a menin inhibitor built on displacement magnitude alone. If half the fusion is still on chromatin at a dose that fully engages menin, the therapeutic case has to rest on what the remaining fusion can no longer do - not on how much menin was moved.
Still open
  1. Does menin displacement at fusion target genes specifically - rather than genome-wide - track fusion eviction? If it does, a locus-restricted menin PD assay survives and this risk narrows to bulk readouts only. The deposited HA peak table makes this testable without new data.
  2. Do revumenib and ziftomenib, at clinical exposures, leave as much fusion on chromatin as VTP-50469 100 nM does? Nothing in this deposit or this corpus answers it, and the tool-compound caveat is the most likely way this entry is wrong.
  3. Is there a soluble or circulating correlate of fusion eviction that could serve as a PD marker in patients, where chromatin assays are impractical?
Every source behind this entry (2)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    Degrading MLL::AF9 leaves menin at 80.0% of its strongest genome-wide peak calls at 3 h while the fusion's own HA tag falls to 2.5-7.1%, so menin occupancy does not track fusion occupancy.

    GSE173599, samples GSM5271682 / GSM5271684 (MEN1) and GSM5271662/5, GSM5271707/9 (HA) · Olsen SN, Armstrong SA, et al. · NCBI Gene Expression Omnibus 2022
    What this evidence cannot show

    Genome-wide peak presence, n = 1 per arm, no spike-in, HCB1 (KMT2A::MLLT3, not the index fusion). The authors' target-gene-restricted signal measure shows >50% menin loss at MLL::AF9 target genes at the same timepoint, so the dissociation is a genome-scale statement and is weaker at the loci the fusion occupies.

    Evidence that argues against this (1)
    • If menin displacement at FUSION TARGET GENES does track fusion eviction - which the authors' own >50% figure leaves open - then a locus-restricted menin assay could be a perfectly good PD marker and this risk would apply only to bulk or genome-wide readouts. That distinction has not been tested and is the main thing that could narrow this entry. Olsen SN et al. 2022
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    VTP-50469 100 nM for 4 days leaves 51.7% and 76.9% of the fusion's own HA sites occupied in two independent batches, against 1.0-10.3% under degradation and 10.7% under combined DOT1L plus menin inhibition.

    What this evidence cannot show

    VTP-50469 is a tool compound, not a clinical menin inhibitor; revumenib and ziftomenib were not tested here and this project has previously published a negative that turned out to be a fact about a weak tool compound. The two batches give 51.7% and 76.9%, a 25-point spread, so the magnitude is not established even within this deposit. Four days of drug exposure changes population composition, so a retained peak is not necessarily an unmoved molecule in an unchanged cell. n = 1 per arm.

    Evidence that argues against this (1)
    • The authors report by their own normalised measure that single-agent treatment has 'more modest effects on MLL::AF9 occupancy', which agrees in direction. But they also report that combined inhibition releases the fusion 'globally' with kinetics similar to degradation - so the pharmacological ceiling is set by the combination, not by menin inhibition alone, and a combination PD strategy would face a different and weaker version of this risk. Olsen SN et al. 2022
Connected to