Oncofusion

start / graph / resistance-q125-a-polycomb-deficient-cell-is-evicted-normally-and-silenced-less-so-occupancy-scores-it-fully-engaged

Resistancemoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q125-the-h2ak119ub-arm-of-the-pcgf1-knockout

A cell that cannot deposit H2AK119ub is still evicted normally by a menin inhibitor and silences its vacated sites less - so a fusion-occupancy pharmacodynamic marker reads FULL ENGAGEMENT in exactly the genotype that responds worst, and this is now the fourth independent reason this corpus distrusts its own candidate markers

also called the silencing-arm resistance · occupancy reads engaged in a polycomb-deficient cell

If a leukaemia cell has lost part of its silencing machinery, the drug still does its first job - pulling the fusion off the DNA - but the second job, switching those places off, works less well. A test that only checks the first job would say the drug is working perfectly.

The same thing, for a scientist

**Mechanism.** This cycle measured that a menin inhibitor's effect at the fusion's own sites has two separable arms: **eviction**, which Q117 measured, and **polycomb deposition at the vacated sites**, graded by eviction depth, which this cycle measured. **A PCGF1 knockout attenuates the second without touching the first** - H2AK119ub at the most-evicted quartile falls from 1.4450 to 1.1325 and from 1.0151 to 0.8842 in two replicate conventions, and H3K27me3 from 2.0007 to 1.0134, in a clone whose eviction was not measured but whose FLAG-defined site set is the same [E1]. **Clinical shape.** PCGF1 loss is a published resistance mechanism to menin inhibition; **BCOR and BCORL1 loss are recurrent in myeloid leukaemia** and are the same complex. KDM2B, which targets PRC1.1 to unmethylated CpG islands, is the third node [E2]. **Why it defeats the obvious marker.** A pharmacodynamic assay that measures fusion occupancy, menin occupancy or H3K79me2 reads the eviction arm. **All three would be normal in a PRC1.1-deficient cell.** Q124 made the same point for H3K79me2 from the other direction - the mark is unchanged in this knockout - and this cycle supplies the missing half: **the polycomb channel IS changed in the same cells**, so the two arms are separable in fact and not only in principle [E1]. **Standing count.** This corpus now holds four independent reasons to distrust its candidate pharmacodynamic markers: cycle 83's finding that H3K79me2 is unchanged in a resistance genotype; cycle 83's finding that a DOT1L inhibitor moves the MLL1 channel as much as a menin inhibitor does; Q124's DEVIATION 5, that the two drugs move the mid-occupancy floor in opposite directions; **and this.**

Go deeper
What would make this actionable, and what makes it hard

The marker that would work is the silencing arm itself - H2AK119ub or H3K27me3 at a fixed set of deeply-evicted sites - and this cycle's companion hypothesis says why that is harder than it sounds: the effect is graded by eviction depth, which is unmeasurable in a patient sample, so the site set has to be pre-selected in a model system and transferred. That transfer is the partner-import problem this corpus keeps finding, and it is worse here because every measurement in this cycle is murine MLL-AF9 and this graph exists for human KMT2A::AFF1.

And the effect size is attenuation, not abolition. A 20-30% reduction in a chromatin ratio, in a channel whose single-track resolution was measured this cycle at 0.2623 between two replicates of one culture, is not separable per patient. The honest statement is that the resistance mechanism is real and the assay to detect it does not exist at the required resolution.

Facts and figures
kindresistance
severitymechanism-level prediction; effect size in the model is attenuation of 0.13-0.99 in a share ratio, not abolition
mitigationmeasure the silencing arm (H2AK119ub or H3K27me3) at a pre-selected deeply-evicted site set rather than fusion or menin occupancy; and genotype BCOR, BCORL1, KDM2B and PCGF1 before reading an occupancy-based pharmacodynamic marker as response
shown for fusionmurine MLL-AF9 (KMT2A::MLLT3) with an over-expressed FLAG-tagged construct; NOTHING in KMT2A::AFF1, nothing in a human line, nothing in a B-lineage model
epigenetic readoutH2AK119ub and H3K27me3 share ratios at the fusion's most-evicted bins
absence classmeasured-not-analysed
absence class noteThe chromatin half of this mechanism was deposited in GSE260742 and public from 2025-11-23 and is the analysis this cycle ran. The eviction half IN THE KNOCKOUT is MEASURED-NOT-SEARCHED: GSM8123407-412 may be the required FLAG track and their channel is unresolved from the three metadata routes tried. The clinical half is NEVER-MEASURED: no menin-inhibitor trial analysis by BCOR, BCORL1, KDM2B or PCGF1 genotype has been identified.
absence class assigned2026-09-23 (Q125)
affects
[
  "drug-revumenib",
  "drug-ziftomenib",
  "drug-bleximenib",
  "drug-enzomenib"
]
candidate nodes
[
  "gene-pcgf1",
  "gene-bcor",
  "gene-bcorl1",
  "gene-kdm2b",
  "gene-ring1"
]
How this could bite us back
  • The proposed mitigation - measure the silencing arm instead - inherits every problem this cycle documented: the effect is graded by an unmeasurable variable, the single-track resolution is 0.26, the site set is model-derived, and the reagent cannot distinguish PRC1.1 from canonical PRC1. Proposing a better marker is easy and the measurement that would validate it does not exist.
  • And the mechanism cuts both ways for therapy: if silencing is what kills the cell, a polycomb-deficient leukaemia may be more dependent on something else that is also druggable. Nothing here identifies what.
Still open
  1. Do BCOR- or BCORL1-mutant cases in reported menin-inhibitor trials show normal early pharmacodynamics and poorer depth of response? The trials have reported and the genotypes are collected.
  2. Is eviction really normal in the PCGF1 knockout? It is the untested premise of this entire entry and it may be resolvable from six files already in the deposit.
Every source behind this entry (2)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    In murine MLL-AF9 (GSE260742), a PCGF1-knockout clone shows attenuated menin-inhibitor-induced polycomb deposition at the fusion's most-evicted bins - H2AK119ub 1.4450 -> 1.1325 and 1.0151 -> 0.8842 in two replicate conventions, H3K27me3 2.0007 -> 1.0134 - while Q124 measured that H3K79me2's drug response at the same set in the same two genotypes is unchanged, so the eviction-reporting and silencing-reporting channels are separable in the same cells.

    What this evidence cannot show

    NO SPIKE-IN NORMALISATION IS POSSIBLE. Q117 verified zero hits for the entire spike-in vocabulary across all 419,824 characters of GSE260742's 158 sample records and no scale factor is deposited, so every cross-condition number here is a WITHIN-SAMPLE SHARE, which cancels the unknown per-sample scale factor exactly and IS BLIND TO A UNIFORM CHANGE BY CONSTRUCTION. Nothing here measures any mark's abundance; everything measures its distribution. EXPOSURE ASYMMETRY: DMSO 72 h and SGC0946 5 uM 72 h are matched, VTP-50469 500 nM is 48 h, so every VTP ratio is a LOWER BOUND on the 72-hour effect - identical in both genotypes, so the genotype DIFFERENCE is unaffected. REAGENTS: H2AK119ub is Ubiquityl-Histone H2A (Lys119) (D27C4) XP Rabbit mAb, NEB 8240S, and it CANNOT distinguish PRC1.1 from canonical PRC1 - RING1A/B is the ligase in both and PCGF1 is one of six PCGF paralogs - nor deposition from loss of a deubiquitinase. The H3K27me3 and H3K4me3 reagents are NOT RESOLVED and are recorded unresolved rather than guessed; `antibod` occurs zero times in the 158 GEO records. S_FLAG is defined by DYKDDDDK Tag (D6W5B) Rabbit mAb, NEB 14793S, on a RETROVIRALLY OVER-EXPRESSED tagged MLLAF9 construct, so it is the over-expressed molecule's sites and not an endogenous fusion's. THE KNOCKOUT is one CRISPR guide, one clone C1, no second clone, no rescue and NO deposited PCGF1 protein or transcript measurement. SYSTEM: murine MLL-AF9 (KMT2A::MLLT3) myeloid blood cells, NOT the infant B-lineage KMT2A::AFF1 disease this graph is about. EVICTION ITSELF WAS NOT MEASURED IN THE KNOCKOUT: there is no FLAG, MLL1 or MENIN track in the NTvKO arm by title, so 'evicted normally' is assumed rather than measured and is the clause of hyp-q117's prediction 5 this cycle could NOT score. Six files carry ChIP_FLAG filenames and TDF headers while being titled H2AK119ub, so the measurement may exist in the deposit with an unresolved channel. The quartile-level genotype difference is a POST HOC application of thresholds to pre-registered per-quartile numbers.

    Evidence that argues against this (2)
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    PRC1.1 loss is a documented resistance mechanism to menin inhibition, and BCOR and BCORL1 - subunits of the same complex - are recurrently mutated in myeloid leukaemia, so the polycomb-deficient genotype this entry describes occurs in patients and does not require engineering.

    PMID 39158067DOI
    What this evidence cannot show

    Cited for the resistance framing and for the noncanonical-menin-target axis. That study's own system and lesion set are not this one's, and the specific claim that a BCOR- or BCORL1-mutant leukaemia shows attenuated H2AK119ub in-fill under a menin inhibitor is NOT shown anywhere: this cycle tested PCGF1 only, and GSE260742 contains no BCOR knockout.

    Evidence that argues against this (1)
Connected to