Oncofusion

start / graph / exp-prc1-1-capacity-titration-under-menin-inhibition

Experimentmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q52-spine

Titrate PRC1.1 capacity against menin-inhibitor response, and measure H2AK119ub on the same timecourse as H3K79me2

also called PRC1.1 dose titration · H2AK119ub pharmacodynamics

Two papers disagree about whether one Polycomb machine is the menin drug's helper or its off-switch. This experiment dials that machine down a bit at a time instead of deleting it, and watches the drug's effect change - which tells you which story is right.

The same thing, for a scientist

Tests `hyp-prc1-1-is-effector-and-gatekeeper-of-one-activity` by replacing the knockout with a **dose axis**, which is the variable neither published study has. A degron-tagged PCGF1 or RNF2 allele titrated with sub-saturating degrader gives a continuous range of PRC1.1 capacity; menin-inhibitor response is then measured across that range. **The second half is the calibrated mark timecourse**: spike-in-normalised H2AK119ub and H3K79me2 CUT&RUN on one set of samples, which is the measurement that decides whether the PRC1.1 gain follows H3K79me2 decay as [E1] proposes, or follows menin displacement directly.

Go deeper
Design

Arms. (1) degron-PCGF1 titration x revumenib dose matrix; (2) degron-RNF2 titration as a catalytic-subunit control; (3) RYBP and YAF2 single and double depletion, to test whether catalytic rate or locus selection is what matters 3; (4) parental plus drug.

Lines. A KMT2A::AFF1 line (SEM) and a KMT2A::MLLT3 line (MOLM-13), because the effector claim 1 and the resistance claim 2 come from different fusion contexts and were never run side by side.

Mark timecourse. 0, 6, 24, 48, 72, 120 h, spike-in-calibrated CUT&RUN for H2AK119ub, H3K79me2, H3K4me3 and menin, on one sample set. Calibration is not optional: an uncalibrated ChIP cannot distinguish a redistribution from a global change, which is the exact distinction between the two published accounts.

Locus analysis, and it is free. Before any bench work: intersect the drug-silenced locus set of 1 with the MYC-bypass locus set of 2 from their deposited data. If the two sets do not overlap, the two-locus-class model wins with no experiment at all.

Decision rule
observationreading
graded loss of response across the PCGF1 titrationone-activity model; response depth is a function of PRC1.1 capacity
threshold response with MYC de-repression at the thresholdtwo-locus-class model; resistance is a MYC bypass and PRC1.1 is not the effector
H2AK119ub rise lags H3K79me2 decaysupports the DOT1L-loss route of 1
H2AK119ub rise tracks menin displacement insteadPRC1.1 recruitment is menin-proximal and the DOT1L route is incidental
RYBP/YAF2 double depletion phenocopies PCGF1 losscatalysis is the limiting variable, and RYBP expression becomes a candidate response predictor
Facts and figures
designDegron-PCGF1 and degron-RNF2 sub-saturating titration crossed with a revumenib dose matrix in SEM (KMT2A::AFF1) and MOLM-13 (KMT2A::MLLT3), plus RYBP and YAF2 single and double depletion arms; preceded by a free locus-set intersection of the two source studies' deposited data.
readoutSpike-in-calibrated CUT&RUN for H2AK119ub, H3K79me2, H3K4me3 and menin at 0, 6, 24, 48, 72, 120 h on one sample set; RNA-seq; viability and differentiation across the dose matrix; per-titration-point protein quantification.
controlsparental plus drug; degrader without tag; MEN1 M327I resistance-mutant arm to separate target engagement from downstream effect; spike-in calibration mandatory
cost tiermed
timeline6-9 months
free first stepintersect the drug-silenced locus set of PMID 41634341 with the MYC-bypass locus set of PMID 39158067 from deposited data; non-overlap decides the question without bench work
chainmechanism-kmt2a-fusion-causal-chain
tests
[
  "hyp-prc1-1-is-effector-and-gatekeeper-of-one-activity"
]
decision rule
{
  "graded_response": "one-activity model",
  "threshold_with_MYC": "two-locus-class model",
  "h2ak119ub_lags_h3k79me2": "supports the DOT1L-loss route",
  "h2ak119ub_tracks_menin_displacement": "PRC1.1 recruitment is menin-proximal",
  "rybp_yaf2_phenocopy": "catalysis is limiting; RYBP becomes a candidate response predictor"
}
How this could bite us back
  • A positive result argues for stratifying menin-inhibitor trials on BCOR and PCGF1 status, which reduces the eligible population of an already small disease.
  • If PRC1.1 capacity sets response depth, then any co-medication that perturbs polycomb - and EZH2 inhibitors are in development in this disease - could blunt a menin inhibitor, and the combination would fail for a reason the preclinical rationale did not anticipate.
Still open
  1. Are the deposited locus sets of the two source studies in compatible genome builds and cell systems? If not, the free first step costs a reanalysis rather than an intersection.
  2. Can H2AK119ub be measured reliably in primary blasts at the cell numbers a paediatric sample provides?
  3. Is there a small-molecule handle on PRC1.1 catalytic activity that would make the titration pharmacological rather than genetic, and therefore translatable?
Every source behind this entry (4)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Menin inhibition induces PRC1.1-dependent H2AK119ub deposition at a subset of MLL-fusion targets, arising from progressive loss of DOT1L-mediated H3K79 methylation.

    DOT1L provides transcriptional memory through PRC1.1 antagonism. · Neville D, Ferguson DT, Heikamp EB, Lai Z, et al. · Nat Cell Biol 2026
    PMID 41634341DOI
    What this evidence cannot show

    Abstract read 2026-09-20. Supplies the effector arm and the proposed causal route this design times. Not measured in infant KMT2A::AFF1 material, which is why the design carries both fusion contexts.

    Evidence that argues against this (1)
    • If PRC1.1 activity rises genome-wide under DOT1L inhibition but only locally under menin inhibition, a single timecourse in one drug cannot separate a local recruitment event from a global capacity change without spike-in calibration - which is why calibration is written into the design rather than offered as an option. Zhou X et al. 2024
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Loss of PRC1.1 subunits such as PCGF1 or BCOR causes menin-inhibitor resistance through epigenetic reactivation of non-canonical menin targets including MYC.

    PMID 39158067DOI
    What this evidence cannot show

    Abstract read 2026-09-20. Supplies the resistance arm and the knockout-level perturbation this design replaces with a titration. AML models.

    Evidence that argues against this (1)
    • MYC restoration is a general resistance route in this disease, shown independently for BET inhibitors through PRC2 suppression, so a MYC-routed resistance phenotype is weak evidence for which complex is the drug's effector. Rathert P et al. 2015
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    PRC1 H2AK119ub ligase activity is set by catalytic-subunit composition and is highly regulated by RYBP/YAF2-dependent stimulation.

    RYBP stimulates PRC1 to shape chromatin-based communication between Polycomb repressive complexes. · Rose NR, King HW, Blackledge NP, Fursova NA, et al. · Elife 2016
    PMID 27705745DOI
    What this evidence cannot show

    Abstract read 2026-09-20. Mouse ES cells. Justifies arm 3; if RYBP and YAF2 are redundant, that arm returns a negative that does not discriminate, which is stated in the decision rule rather than discovered afterwards.

    Evidence that argues against this (1)
    • RYBP or YAF2 association also determines which accessory subunits are excluded, so depleting them changes complex composition as well as catalytic rate, and arm 3 is not a clean rate-only perturbation. Gao Z et al. 2012
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    VHL-recruiting dTAG molecules provide rapid, direct, dose-controllable degradation of tagged proteins and overcome a limitation of CRBN-recruiting dTAG molecules on recalcitrant targets.

    Rapid and direct control of target protein levels with VHL-recruiting dTAG molecules. · Nabet B, Ferguson FM, Seong BKA, Kuljanin M, et al. · Nat Commun 2020
    PMID 32948771DOI
    What this evidence cannot show

    Abstract read 2026-09-20. Methodological citation supplying the dose axis. Sub-saturating degrader dosing gives a continuous capacity range but a noisy one; each titration point needs its own protein-level quantification rather than an assumed dose-effect curve.

    Evidence that argues against this (1)
    • Degrading a scaffolding subunit at partial occupancy can produce dominant-negative partial complexes rather than proportionally less activity, so a graded phenotype is not by itself proof of a graded capacity. Rose NR et al. 2016
Connected to