PRC2 - the complex, not the subunit, is the unit every experiment in this disease actually resolves

This is real experimental data — atoms placed by crystallography or cryo-EM, deposited by the people who did the experiment.

• PRC2 is required for normal haematopoietic stem cell maintenance, so a complex-directed agent removes it from the compartment that must rebuild the marrow. The only EED-directed inhibitor taken into patients has thrombocytopenia and febrile neutropenia as dose-limiting toxicities.
• The complex is upstream of tumour-suppressor de-repression, which means its therapeutic effect depends on the restraints being intact. CDKN2A deletion is common in this disease and removes the best-characterised mediator, though not all of the effect.
• Every subunit of PRC2 is shared with normal biology and none is fusion-specific. Unlike the menin-KMT2A interface, there is no PRC2 surface that exists only in the leukaemia, so a PRC2-directed agent has no structural route to selectivity - only a therapeutic-window argument.
• Reported resistance routes to PRC2-directed therapy include PRC2 gatekeeper mutations and compensatory DNA methylation. The second is upstream of every PRC2 binding site and would not be addressed by changing which subunit is targeted; t(4;11) infant ALL is among the most promoter-hypermethylated leukaemia subtypes, so it may be pre-adapted to it.

1. Which PRC2 subtype - PRC2.1 or PRC2.2 - operates at the fusion's repressed targets? The question is answerable with a single MTF2 versus JARID2 CUT&RUN in an index-fusion line and has never been asked; PRC2.2 couples to PRC1 through H2AK119ub1, and this graph already holds a PRC1.1 arm, so the answer would connect two arms that are currently unconnected.
2. Is H3K27me3 excluded from the fusion's H3K79me2 domains, and if so by what? The H3K36 and H3K4 allosteric routes are established and the H3K79 route is not, so the observed anticorrelation in this disease has no mechanism attached to it.
3. Does PRC2 occupancy change after menin inhibition in an index-fusion system? If menin drives EZH2 expression, a menin inhibitor is a partial PRC2 inhibitor, and nobody has measured H3K27me3 as a menin-inhibitor pharmacodynamic.

1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   PRC2 is stimulated upon recognition of its own catalytic product and is more active on dinucleosomes than on H3 tails or single nucleosomes; cryo-EM reconstructions of human PRC2 bound to bifunctional dinucleosomes show a single PRC2 positioning the H3 tails of the activating and the substrate nucleosome to interact with EED and with the EZH2 SET domain respectively, accommodating varying linker-DNA lengths - properties that probably facilitate local H3K27me2/3 spreading.

   Cryo-EM structures of PRC2 simultaneously engaged with two functionally distinct nucleosomes. · Poepsel S, Kasinath V, Nogales E · Nat Struct Mol Biol 2018
   PMID 29379173DOI
   ▸What this evidence cannot show

   Abstract read from the eutils record; full text not read. Reconstituted human PRC2 on designed dinucleosome substrates; the spreading model is stated by the authors as probable rather than demonstrated in cells. No residue numbers or PDB spans are asserted here.

   ▸Evidence that argues against this (1)

   • The spreading model predicts propagation that active chromatin does not permit: H3K36me2/3 allosterically inhibits PRC2 and unmodified H3K36 occupies a critical position in the EZH2-DNA interface, so at a transcribed locus the read-write loop is switched off by the substrate itself. Finogenova K et al. 2020
2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   Inactivation of the essential PRC2 gene Eed led to complete ablation of PRC2 function and was incompatible with MLL-AF9 leukaemia growth, whereas Ezh2-null leukaemias maintained self-renewal to the third round of transplantation with locus-specific persistence of H3K27me3.

   Polycomb repressive complex 2 is required for MLL-AF9 leukemia. · Neff T, Sinha AU, Kluk MJ, Zhu N, Khattab MH, Stein L, et al. · Proc Natl Acad Sci U S A 2012
   PMID 22396593DOI
   ▸What this evidence cannot show

   Abstract read from the eutils record; full text not read. Mouse MLL-AF9, myeloid, no B-lineage or infant arm. Genetic ablation removes scaffolding and non-catalytic functions together with the mark, so it is an upper bound on what a catalytic inhibitor could do.

   ▸Evidence that argues against this (1)

   • A complex-level requirement demonstrated by removing one obligate subunit does not establish that the complex's CATALYTIC output is what matters. PRC2:EZH1 has a catalysis-independent chromatin-compaction activity, so an Eed null and a SET-domain inhibitor are not the same experiment. Grau D et al. 2021
3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   An shRNA screen for chromatin regulators uniquely required in mouse MLL-fusion AML recovered PRC2; suppression of Eed, Suz12 or Ezh1/Ezh2 caused proliferation arrest and differentiation with minimal impact on non-transformed haematopoietic cell lines, partly through direct PRC2 repression of self-renewal-limiting genes including Cdkn2a - leading the authors to conclude that Trithorax and Polycomb group proteins can cooperate with one another to maintain aberrant lineage programs in cancer.

   The Polycomb complex PRC2 supports aberrant self-renewal in a mouse model of MLL-AF9;Nras(G12D) acute myeloid leukemia. · Shi J, Wang E, Zuber J, Rappaport A, Taylor M, Johns C, et al. · Oncogene 2013
   PMID 22469984DOI
   ▸What this evidence cannot show

   Abstract read from the eutils record; full text not read. shRNA, mouse, myeloid, with a co-operating Nras-G12D that most infant KMT2A::AFF1 disease does not carry. 'Uniquely required' is relative to the comparators in that screen. The cooperation conclusion is the authors' interpretation of their own data and is a general statement about Trithorax and Polycomb, not a measurement of a physical interaction.

   ▸Evidence that argues against this (1)

   • Cooperation at the level of a shared requirement is not cooperation at the level of a shared locus. PRC2 repressing Cdkn2a while the fusion activates HOXA is two complexes doing opposite things at different places, which is compatible with the textbook antagonism at any individual gene. The strongest evidence for cooperation AT A LOCUS is the menin-EZH2-C/EBPalpha axis, not this screen. Thiel AT et al. 2013
4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   The EZH1/EZH2 dual inhibitor DS-3201 (valemetostat) inhibited B-ALL cell line growth more strongly than the EZH2-specific inhibitor EPZ-6438, up-regulated CDKN1C and TP53INP1 as direct EZH1/2 targets confirmed by CRISPR knockout, and inhibited growth of a B-ALL patient-derived xenograft harbouring MLL-AF4.

   Dual inhibition of EZH1/2 induces cell cycle arrest of B cell acute lymphoblastic leukemia cells through upregulation of CDKN1C and TP53INP1. · Ito J, Yamagata K, Shinohara H, Shima Y, Katsumoto T, Aikawa Y, et al. · Int J Hematol 2023
   PMID 36280659DOI
   ▸What this evidence cannot show

   Abstract read from the eutils record; full text not read. The head-to-head is a growth comparison without stated matched exposure or matched H3K27me3 depletion, so it cannot separate target coverage from potency. One PDX, no n, dose or age given, not stated to be infant.

   ▸Evidence that argues against this (1)

   • The de-repressed genes here (CDKN1C, TP53INP1) are different from those reported for Eed loss in mouse MLL-AF9 (Cdkn2a, Cdkn2b, Gata2, Egr1). Either the mediator is species- and lineage-specific, or the shared feature is simply 'PRC2 represses growth restraints' - in which case no specific biomarker follows, and patient selection for a PRC2-directed agent has no published basis in this disease. Danis E et al. 2015
5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   Menin binds the Ezh2 promoter and promotes Ezh2 expression; EZH2 interacts with C/EBPalpha and represses its target genes; menin depletion reduces MLL binding at the Ezh2 locus, EZH2 expression, and EZH2 binding and repressive H3K27 methylation at C/EBPalpha targets, inducing pro-differentiation C/EBPalpha target expression - so in contrast to its classical role antagonising trithorax function, EZH2 collaborates with trithorax-associated menin to block MLL-AF9 leukaemia cell differentiation.

   The trithorax protein partner menin acts in tandem with EZH2 to suppress C/EBPα and differentiation in MLL-AF9 leukemia. · Thiel AT, Feng Z, Pant DK, Chodosh LA, Hua X · Haematologica 2013
   PMID 23349306DOI
   ▸What this evidence cannot show

   Abstract read from the eutils record; full text not read. Mouse MLL-AF9 with acute menin depletion; the effector C/EBPalpha is a myeloid differentiation factor with no equivalent role in a pro-B blast. Whether EZH1 substitutes when menin-driven EZH2 falls is not addressed.

   ▸Evidence that argues against this (1)

   • Menin regulating Ezh2 transcriptionally is a different claim from menin and PRC2 acting in one complex, and no physical menin-PRC2 complex is reported here. An indirect transcriptional relationship is more fragile across species, lineage and timescale than a protein-protein one, and the whole cross-lineage extension of this axis rests on it. de Barrios O et al. 2025
6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   PRC2 activity is regulated in part by selective incorporation of one of two catalytic paralogues, EZH1 or EZH2, each with specialised biological functions; a cryo-EM structure of PRC2:EZH1 with the PRC2.2 accessory subunits AEBP2 and JARID2 has been deposited (RCSB 7KSO), and dimeric PRC2:EZH1 bound to a nucleosome undergoes a dramatic conformational change and compacts chromatin more effectively than the monomer.

   Structures of monomeric and dimeric PRC2:EZH1 reveal flexible modules involved in chromatin compaction. · Grau D, Zhang Y, Lee CH, Valencia-Sánchez M, Zhang J, Wang M, et al. · Nat Commun 2021
   PMID 33514705DOI
   ▸What this evidence cannot show

   Abstract read from the eutils record; full text not read. The RCSB entry titles for 7KSO and 7KSR were fetched and read back on 2026-09-20, but polymer entities were not fetched, so no chain-to-UniProt mapping, construct composition or residue span is asserted; RCSB returned no primary-citation PubMed id for either entry, so the pairing with this paper is by subject and deposition rather than by a stored identifier. Reconstituted complexes in vitro; the dimer's abundance in cells is not established.

   ▸Evidence that argues against this (1)

   • The structures are of PRC2:EZH1, not PRC2:EZH2, and the same paper shows the two paralogues differ at the loop that drives nucleosome binding and compaction - so a PRC2.2 architecture solved on EZH1 should not be read as a description of the EZH2 complex that most of the leukaemia literature concerns. Poepsel S et al. 2018
7. E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

   Across 11 KMT2A-rearranged DepMap leukaemia lines against 43 other leukaemia lines, the PRC2 dependency profile is EED -0.458 (difference -0.117, rank 824 of 17,728), SUZ12 -0.482, RBBP4 -1.486 (pan-essentiality -1.347), EZH2 -0.201, EZH1 -0.100, RBBP7 -0.013, with all accessory subunits mid-genome (MTF2 -0.072, PHF19 -0.027, JARID2 -0.022, AEBP2 +0.014, EPOP +0.050) - the signature of a complex whose only non-redundant, non-pan-essential node is EED.

   analyses/prc2-dependency-lineage-split - DepMap (Chronos, via Open Targets) PRC2 panel scored with the index arm split by lineage, plus a Human Protein Atlas cell-line RNA expression control · Q41-prc2-arm · Oncofusion knowledge graph, internal analysis 2026
   ▸What this evidence cannot show

   Single-gene knockouts, so every paralogue pair in the complex (EZH1/EZH2, RBBP4/RBBP7, MTF2/PHF19/PHF1) is systematically under-scored. RBBP4 is floor-compressed at -1.347 pan-essentiality and no selectivity may be read from it. 11 index lines, none infant; the lymphoid half is 4 lines. Raw effects, no z, no permutation null - this is a consistency check against the mouse genetics, not a discovery scan. DepMap release version not stated by the Open Targets API.

   ▸Evidence that argues against this (1)

   • Reading the complex's requirement off single-gene scores is exactly what the mouse genetics warns against: the shRNA screen that established PRC2's requirement in MLL-fusion AML used Eed, Suz12 AND Ezh1/Ezh2 arms, and the Ezh1/Ezh2 arm was a pair. A DepMap profile in which no catalytic subunit scores is fully compatible with PRC2 catalysis being required. Shi J et al. 2013
8. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   A whole-genome CRISPR-Cas9 screen coupled to a cell-surface bivalent MHC class I readout identifies specific roles for MTF2-PRC2.1, PCGF1-PRC1.1 and Menin-KMT2A/B complexes in maintaining bivalency; genetic loss or pharmacological inhibition of menin phenocopies polycomb disruption and de-represses bivalent genes in cancer cells and pluripotent stem cells, and a Menin-INDEPENDENT function of KMT2A/B maintains H3K4me3 and opposes polycomb-mediated repression at bivalent genes, with release of KMT2A from active genes after menin targeting shifting the polycomb/KMT2A balance at bivalent genes toward activation.

   Targeting Menin disrupts the KMT2A/B and polycomb balance to paradoxically activate bivalent genes. · Sparbier CE, Gillespie A, Gomez J, Kumari N, Motazedian A, Chan KL, et al. · Nat Cell Biol 2023
   PMID 36635503DOI
   ▸What this evidence cannot show

   Abstract read from the eutils record; full text not read. The system is cancer cells and pluripotent stem cells with MHC class I as a bivalent-gene reporter - NOT KMT2A-rearranged leukaemia and not the index fusion. 'PRC2.1' rests on MTF2 scoring in a single-gene screen that under-scores the PHF19/PHF1 paralogues, so it means 'MTF2 scored' rather than 'the PRC2.1 module was resolved'. The paper concerns wild-type KMT2A/B, not a fusion. This graph already cited it in gene-pcgf1 and protein-prc1-1-complex before this beat existed.

   ▸Evidence that argues against this (1)

   • This result and the menin-EZH2 axis (PMID 23349306) both end in de-repression on menin inhibition but by OPPOSITE routes: there, menin drives Ezh2 expression so losing menin removes PRC2 activity; here, menin loss releases KMT2A/B from active genes and a menin-independent KMT2A/B function then opposes polycomb at bivalent genes. The graph should carry both and merge neither - and the counted absence this beat first wrote for PRC2 subtypes missed this paper entirely because its abstract does not contain the word leukaemia. Thiel AT et al. 2013
9. E9strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   In t(4;11) infant B-ALL, HDAC7 is epigenetically silenced by EZH2 together with the KMT2A::AFF1 fusion protein, and the menin inhibitor MI-538 restores HDAC7 expression.

   HDAC7 induction combined with standard-of-care chemotherapy provides a therapeutic advantage in t(4;11) infant B-cell acute lymphoblastic leukemia. · de Barrios O, Ocón-Gabarró I, Gusi-Vives M, Collazo O, Meler A, Romecín PA, et al. · Biomark Res 2025
   PMID 40722046DOI
   ▸What this evidence cannot show

   Abstract read from the eutils record; full text not read. The abstract does not state whether EZH2 occupancy or H3K27me3 was measured at HDAC7, and the intervention tested is menin inhibition plus an HDAC inhibitor rather than PRC2 inhibition - so the EZH2 step is named but not pharmacologically perturbed.

   ▸Evidence that argues against this (1)

   • A menin inhibitor restoring HDAC7 is equally consistent with menin acting on HDAC7 directly and EZH2 being a bystander, since displacing the fusion from chromatin would change the locus whatever PRC2 was doing. The causal role of EZH2 requires an EZH1/EZH2 inhibitor arm, which the study does not report. Thiel AT et al. 2013
10. E10strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Cdkn2a inactivation only partially rescues growth of Eed-inactivated MLL-AF9 cells in vitro and in vivo; Gata2, Egr1 and Cdkn2b are also de-repressed on Eed inactivation, and direct binding targets of MLL fusion proteins are negatively enriched in Eed-inactivated Cdkn2a-null MLL-AF9 cells.

    Inactivation of Eed impedes MLL-AF9-mediated leukemogenesis through Cdkn2a-dependent and Cdkn2a-independent mechanisms in a murine model. · Danis E, Yamauchi T, Echanique K, Haladyna J, Kalkur R, Riedel S, et al. · Exp Hematol 2015
    PMID 26118502DOI
    ▸What this evidence cannot show

    Abstract read from the eutils record; full text not read. Mouse MLL-AF9, myeloid. The negative enrichment of MLL-fusion targets is an RNA-level gene-set observation and is not accompanied in the abstract by a demonstration that fusion chromatin occupancy changes - the interpretation a reader is most likely to import.

    ▸Evidence that argues against this (1)

    • If PRC2 loss lowers MLL-fusion target expression, PRC2 is upstream of or permissive for the fusion programme - which contradicts the classical antagonism even more strongly than the menin-EZH2 axis does. But it is equally consistent with the cells having differentiated for other reasons and the fusion programme collapsing as a consequence, and no timecourse distinguishes them. Neff T et al. 2012
11. E11established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    At transcriptionally active genes, di- and tri-methylated H3K36 inhibit PRC2; unmodified H3K36 occupies a critical position in the EZH2-DNA interface, mutation of H3K36 to arginine or alanine inhibits H3K27 methylation by PRC2 on nucleosomes in vitro, and Drosophila H3K36A and H3K36R mutants show reduced H3K27me3 and defective Polycomb repression of HOX genes.

    Structural basis for PRC2 decoding of active histone methylation marks H3K36me2/3. · Finogenova K, Bonnet J, Poepsel S, Schäfer IB, Finkl K, Schmid K, et al. · Elife 2020
    PMID 33211010DOI
    ▸What this evidence cannot show

    Abstract read from the eutils record; full text not read. Cryo-EM plus in vitro methyltransferase assays on reconstituted dinucleosomes, with the in vivo arm in Drosophila. There is no leukaemia, no KMT2A fusion and no mammalian haematopoietic system in it, so the transfer to this disease is by mechanism and not by demonstration. The H3K36A/R mutants remove the residue entirely and therefore cannot separate loss of the inhibitory methyl mark from loss of the unmodified lysine's structural role - which the same paper shows is also required.

    ▸Evidence that argues against this (1)

    • The relationship established is between PRC2 and H3K36, not between PRC2 and H3K79 - this graph's fusion biology is built on H3K79 methylation, for which no analogous allosteric relationship with PRC2 is published, so 'the fusion's H3K79 domain excludes PRC2' is an inference from anticorrelation and not a mechanism. And exclusion is not complete: automated CUT&Tag across diverse KMT2A oncofusions finds a SUBSET of oncofusion-binding sites carrying bivalent H3K4me3/H3K27me3 signatures, with single-cell heterogeneity suggestive of lineage plasticity. PRC2's product is present at some fusion sites even though active marks inhibit PRC2. Janssens DH et al. 2021

• ← part ofEZH2 - the Polycomb enzyme that collaborates with the fusion instead of opposing it, and that no single-agent inhibitor can switch off

  catalytic SET-domain subunit; interchangeable with EZH1

• ← part ofEZH1 - the paralogue that makes the approved EZH2 drug the wrong drug for this disease

  alternative catalytic subunit; PRC2:EZH1 also dimerises and compacts chromatin

• ← part ofEED - the PRC2 subunit this disease actually needs, with drugs in the clinic and zero leukaemia studies

  WD40 subunit; reads H3K27me3 and allosterically stimulates catalysis

• ← part ofSUZ12 - the scaffold that holds PRC2 together and decides which PRC2 it is, required in this disease and studied in it exactly once

  scaffold; VEFS-Box to EZH2, N-terminus binds the PRC2.1/PRC2.2 accessory modules

• ← part ofRBBP4 - the shared histone-binding subunit that every leukaemia line needs, and whose identical-length paralogue cannot replace it

  histone-binding WD40; Rbbp4 mutation disrupts PRC2 assembly

• ← part ofRBBP7 - the silent paralogue that is not silent, and the reason this project's paralogue rule finally produced a positive answer

  paralogous histone-binding subunit; abundantly expressed and dispensable in leukaemia lines

• causes →H3K27me3 - the repressive mark this graph argued about for seventy-seven cycles without ever writing an entry for it

  the complex writes H3K27me1/2/3 and is stimulated by its own product

• depends on →KMT2A-rearranged acute myeloid leukaemia in children and infants

  PRC2 is required in MLL-AF9 leukaemia at the complex level

• related to →Polycomb in KMT2A-r leukemia — core PRC2 required, EZH2 and BMI1 dispensable, and CBX8 acting outside its repressive role

  supplies the PRC2 half of an entry whose id named PRC2 while no PRC2 node existed

• related to →PRC1.1 — the complex this graph has been leaning on, now enumerated subunit by subunit, and the enumeration changes the conclusion

  PRC2.2 (JARID2/AEBP2) reads H2AK119ub1 and couples to PRC1; no leukaemia evidence distinguishes the subtypes here

• ← inhibitsH3K36me2/3 - the mark that physically switches PRC2 off, that LEDGF reads to reach chromatin, and that a comparator fusion uses to keep EZH2 off HOXA

  allosteric inhibition; unmodified H3K36 is required in the EZH2-DNA interface

• ← inhibitsH3K4me3 - the mark KMT2A's own SET domain writes, that the fusion cannot write, that an interface in this graph already binds, and that switches PRC2 off

  active-transcription mark restrains PRC2; solved structurally for H3K36, asserted qualitatively for H3K4me3

• ← regulatesTrithorax against Polycomb at HOXA: the textbook antagonism, and the two ways this disease breaks it

  menin licenses PRC2's repressive output (Route 1) or is opposed by relocated KMT2A (Route 2)

• ← supportsThe PRC2 dependency profile, lineage-split and expression-controlled - and the first time this project's paralogue rule came back positive

  the profile is the signature of a complex whose only non-redundant, non-pan-essential node is EED

• ← depends onBAHCC1 - a BAH-domain H3K27me3 reader that an MLL fusion binds and upregulates in mouse, and the locus carrying the two largest MLL1-N-reactive rises under a menin inhibitor in human RS4;11

  reads the PRC2 product H3K27me3; BAHCC1 occupancy is downstream of PRC2 deposition