Oncofusion

start / graph / drug-tazemetostat

Drugmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q41-prc2-arm

Tazemetostat - an approved EZH2 inhibitor, and the drug this disease's own biology predicts will not work

also called tazemetostat · EPZ-6438 · E7438 · Tazverik

Tazemetostat is an approved pill that blocks EZH2, the enzyme that writes the main 'switch off' tag on chromosomes. It works in a lymphoma where EZH2 is overactive - but in KMT2A leukaemia the back-up enzyme EZH1 takes over, and a head-to-head test in leukaemia cells showed a drug that blocks both worked better.

The same thing, for a scientist

Tazemetostat (EPZ-6438) is a first-in-class oral **EZH2-selective** SAM-competitive inhibitor, approved for follicular lymphoma and advanced epithelioid sarcoma [E1][E4]. In relapsed/refractory follicular lymphoma it gave an objective response rate of **69%** in the EZH2-mutant cohort and **35%** in wild-type, with median PFS 13.8 and 11.1 months and a benign safety profile - grade 3+ treatment-related thrombocytopenia, neutropenia and anaemia in 3%, 3% and 2% of 99 patients and no treatment-related deaths [E1]. **In KMT2A-rearranged leukaemia the evidence says the wrong target is being covered.** *Ezh2*-null MLL-AF9 leukaemia survives because EZH1 sustains locus-specific H3K27me3 [E2]; DepMap shows neither catalytic paralogue is a dependency while both are transcribed in every panel line [E3]; and in a direct head-to-head in B-ALL lines the dual EZH1/EZH2 inhibitor DS-3201 inhibited growth **more significantly and strongly than EPZ-6438**, with the dual agent also working in an MLL-AF4 PDX [E5]. A 2026 field review states the problem in the same words: EZH2-selective inhibitor efficacy is “limited by functional redundancy with EZH1”, which valemetostat overcomes [E6]. **The counted absence is clean and is the point of this entry**: `tazemetostat AND (KMT2A OR MLL)` returns exactly **1** PubMed record, and title-checking it finds a pharmacology review, not a study [E7]. **No primary study of tazemetostat in KMT2A-rearranged leukaemia exists.** The entry is written so that if one is ever run and reports a negative, the graph can say the failure was predicted and was about the drug, not about PRC2.

Go deeper
The counted absence, and why it matters more than usual

Run 2026-09-20 7:

queryrecordsafter title check
tazemetostat AND (KMT2A OR MLL)1PMID 35753676, Chemical biology and pharmacology of histone lysine methylation inhibitors - a review, not a study
tazemetostat AND "follicular lymphoma" AND phase 214real trial literature, including the registrational PMID 33035457
valemetostat AND (KMT2A OR MLL)1-

Per Rule 64 the complex form was run and does not rescue it: "Polycomb repressive complex 2" AND (KMT2A OR MLL) AND leukemia returns 47 and title-checking recovers the MLL-AF9 genetics and the response-element literature, no tazemetostat study. Per the subtract-before-you-report rule the repository was grepped: PMID 33035457 appears in this beat's own entries and nowhere else in the graph.

Why this absence is worth an entry rather than a line in a memo. An approved, well-tolerated oral drug against a target that a 2013 paper links to the fusion's own anchor [see mechanism-menin-inhibition-shifts-the-kmt2a-polycomb-balance] is exactly the kind of thing a clinician or a translational group reaches for. The biology says it will fail, and a failure would be read as evidence against PRC2 rather than against EZH2-selectivity. Writing the prediction down before the experiment is the only way the graph can later distinguish the two.

What tazemetostat WOULD be right for, and what it would not

The drug is not useless in this space; it is mis-specified for this disease.

  • Right: a leukaemia or lymphoma whose PRC2 is driven by an EZH2 gain-of-function mutation, where the mutant allele is the whole activity. The 69% versus 35% split in follicular lymphoma is that argument 1.
  • Wrong here: a disease whose PRC2 requirement is complex-level and paralogue-buffered. The three alternatives the graph should carry instead are (i) a dual EZH1/EZH2 inhibitor, which is the class with demonstrated activity 56; (ii) an EED-directed allosteric inhibitor, which is paralogue-agnostic by construction and also covers at least one SAM-competitive resistance mutant [see gene-eed]; (iii) doing nothing, on the grounds that the therapeutic window against normal haematopoiesis has not been established in any KMT2A-rearranged system.

Before any of those, the bite-back in risk-ezh2-inhibition-b-lineage-translocations applies to all three, because they share the H3K27me3-to-H3K27ac exchange that is the reported hazard.

Facts and figures
targetgene-ezh2
modalitysmall molecule, oral, SAM-competitive, EZH2-selective
statusapproved (follicular lymphoma; advanced epithelioid sarcoma). Not approved and not studied in any leukaemia.
predicted status in kmt2a rpredicted ineffective, never tested. EZH1 sustains H3K27me3 when EZH2 is removed (PMID 22396593); DS-3201 beat EPZ-6438 in B-ALL lines (PMID 36280659); the field review names EZH1 redundancy as the limitation (PMID 41443386).
aliases development
[
  "EPZ-6438",
  "E7438"
]
trials
[
  "NCT01897571"
]
key ae
[
  "thrombocytopenia",
  "neutropenia",
  "anaemia"
]
efficacy in approved indication
{
  "orr_ezh2_mutant_fl": 69,
  "orr_ezh2_wildtype_fl": 35,
  "n": 99,
  "pmid": "33035457",
  "title": "Tazemetostat for patients with relapsed or refractory follicular lymphoma: an open-label, single-arm, multicentre, phase 2 trial.",
  "authors": "Morschhauser F, Tilly H, Chaidos A, et al.",
  "venue": "Lancet Oncol",
  "year": 2020,
  "title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
}
counted absence
{
  "query": "tazemetostat AND (KMT2A OR MLL)",
  "records": 1,
  "on_topic_after_title_check": 0,
  "aliases_not_searched": [
    "EPZ-6438",
    "E7438"
  ],
  "date": "2026-09-20"
}
resistance
[
  "EZH1 compensation (intrinsic, pre-existing)",
  "PRC2 gatekeeper mutations",
  "compensatory DNA methylation"
]
better specified alternatives
[
  "drug-valemetostat (dual EZH1/EZH2)",
  "EED-directed allosteric inhibitors (see gene-eed; none tested in leukaemia)"
]
How this could bite us back
  • The most likely harm from this drug in this disease is not toxicity but a misread negative. If tazemetostat is tried and fails, the result will be recorded as 'PRC2 is not a target in KMT2A-rearranged leukaemia' when the biology says the failure is EZH1. Any such trial must carry a locus-resolved H3K27me3 pharmacodynamic or it will be uninterpretable in exactly the way the pinometostat phase 1 was.
  • EZH2 inhibition depletes H3K27me3 and raises H3K27ac, increasing transcription and, in B cells with high AID activity or DNA-repair deficiency, increasing chromosomal translocation frequency (PMID 40825195). The authors found monotherapy safe, but a translocation-defined B-lineage leukaemia in infants is the setting where that conditional hazard most needs an explicit argument.
  • EZH2 loss of function is the causative lesion of a class of myeloid neoplasia. A drug that pharmacologically reproduces it, given to infants who will live for decades, carries a second-malignancy question no follicular-lymphoma trial in adults could have answered.
  • Partial effect is the worst outcome, not no effect. The mouse genetics say Ezh2 loss slows progression without abolishing self-renewal, so a small trial could see a real but insufficient signal and produce an ambiguous result that neither supports nor kills the approach.
Still open
  1. Would tazemetostat plus a menin inhibitor do anything that either does alone? If menin drives EZH2 expression (PMID 23349306), the combination hits the same axis twice and should be redundant; if the bivalency mechanism dominates (PMID 36635503), it could be antagonistic. Nobody has run it and the graph can state the prediction before anyone does.
  2. Does a tazemetostat-class agent deplete H3K27me3 at all in a KMT2A-rearranged line, and at which loci? The EZH1-compensation result is locus-specific, so the informative experiment is a calibrated locus-resolved measurement rather than a bulk blot - and a bulk blot is what a pharmacodynamic endpoint would use.
  3. Does an alias-expanded search (EPZ-6438, E7438) change the counted absence? This entry's own counter-evidence shows the INN-only query misses the one leukaemia experiment that used the compound, so the absence as stated is an over-estimate of how untested the drug is.
Every source behind this entry (7)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    In an open-label single-arm phase 2 trial of 800 mg oral tazemetostat twice daily in 99 patients with relapsed or refractory follicular lymphoma, objective response rate was 69% (95% CI 53-82; 31 of 45) in the EZH2-mutant cohort and 35% (23-49; 19 of 54) in the EZH2 wild-type cohort; median duration of response was 10.9 and 13.0 months, median progression-free survival 13.8 and 11.1 months; treatment-related grade 3 or worse events were thrombocytopenia (3%), neutropenia (3%) and anaemia (2%), serious treatment-related adverse events occurred in 4 of 99 patients and there were no treatment-related deaths. NCT01897571.

    Tazemetostat for patients with relapsed or refractory follicular lymphoma: an open-label, single-arm, multicentre, phase 2 trial. · Morschhauser F, Tilly H, Chaidos A, McKay P, Phillips T, Assouline S, et al. · Lancet Oncol 2020
    PMID 33035457DOI
    What this evidence cannot show

    Abstract read from the eutils record; full text not read. Adults with follicular lymphoma, a mature B-cell malignancy with activating EZH2 mutations in ~20% - a completely different genotype and lineage from KMT2A-rearranged infant ALL. Single-arm, so the 35% wild-type response rate has no control. Funded by Epizyme, the sponsor. Cited here for the drug's approved indication, tolerability and mutation-dependent activity, and for nothing about leukaemia.

    Evidence that argues against this (1)
    • The same trial's strongest signal - 69% versus 35% by EZH2 mutation status - is the reason to expect failure in KMT2A-rearranged leukaemia, which carries no activating EZH2 mutation. A drug whose response rate doubles with a gain-of-function mutation is being asked, in this disease, to work against a wild-type, paralogue-buffered enzyme. Ernst T et al. 2010
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Ezh2-null MLL-AF9 leukaemia maintains self-renewal to the third round of transplantation with locus-specific persistence of H3K27me3 attributed to partial compensation by Ezh1, whereas Eed inactivation ablates PRC2 and is incompatible with leukaemia growth.

    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. Genetic Ezh2 inactivation is not the same experiment as SAM-competitive inhibition - a catalytically inhibited EZH2 remains in the complex and could retain scaffolding functions a null does not, so the genetics is an argument about redundancy rather than a prediction of drug effect. Mouse MLL-AF9, myeloid, no B-lineage or infant arm.

    Evidence that argues against this (1)
    • The paper says Ezh2 is not STRICTLY required and that it 'plays a role in leukemia progression' - Ezh2-null leukaemia failed to accelerate on secondary transplantation. So EZH2-selective inhibition is predicted to be insufficient rather than inert, and a tazemetostat arm could show a real but partial effect that a small trial would be unable to interpret. Neff T et al. 2012
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    In 11 KMT2A-rearranged DepMap leukaemia lines, neither EZH2 (-0.201, difference +0.004 against 43 comparator leukaemia lines) nor EZH1 (-0.100, +0.002) is a dependency, while both are transcribed in 7 of 7 Human Protein Atlas panel lines (EZH2 median 47.2 nTPM, EZH1 13.7) - so an EZH2-selective inhibitor leaves a transcribed H3K27 methyltransferase in every line of this panel.

    What this evidence cannot show

    A single-gene CRISPR knockout cannot see a redundant pair and is not a drug. nTPM is bulk RNA from immortalised lines, not protein and not complex occupancy. 11 index lines, none infant. Raw effects, no z, no permutation null. DepMap release version not stated by the Open Targets API.

    Evidence that argues against this (1)
    • Expression of EZH1 does not by itself establish that EZH1 would cover the loci an EZH2 inhibitor vacates. The compensation reported in Ezh2-null leukaemia was LOCUS-SPECIFIC, so a tazemetostat-treated cell might lose H3K27me3 at some functionally important loci while retaining it elsewhere - a partial effect that neither this analysis nor a bulk western blot would resolve. Neff T et al. 2012
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.review

    Tazemetostat is an EZH2 inhibitor approved for the treatment of follicular lymphoma and advanced epithelioid sarcoma, and is one of over a dozen histone lysine methyltransferase inhibitors advanced to clinical trials alongside DOT1L, menin-MLL and WDR5-MLL inhibitors.

    Chemical biology and pharmacology of histone lysine methylation inhibitors. · Barghout SH, Machado RAC, Barsyte-Lovejoy D · Biochim Biophys Acta Gene Regul Mech 2022
    PMID 35753676DOI
    What this evidence cannot show

    A review, cited only for the approval status and for the fact that it is the SINGLE record returned by `tazemetostat AND (KMT2A OR MLL)` - that is, the one record in that intersection is a map of the field rather than a study. No primary result is taken from it.

    Evidence that argues against this (1)
    • That a review linking tazemetostat and MLL exists while no study does is precisely the failure mode a raw count invites: the count is 1 and reads as 'someone has looked', and title-checking shows nobody has. This is the same shape as the SSRP1/FACT and Mediator absences this project has logged, inverted. NCBI PubMed 2026
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In B-cell ALL cell lines the EZH1/EZH2 dual inhibitor DS-3201 (valemetostat) inhibited growth more significantly and strongly than the EZH2-specific inhibitor EPZ-6438 (tazemetostat), and the dual inhibitor significantly 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. **This is the only published head-to-head involving tazemetostat in a leukaemia context and it is a cell-line growth comparison**, not a matched-exposure or matched-target-engagement comparison - without H3K27me3 depletion matched between arms it cannot separate 'dual coverage' from 'more potent compound'. The MLL-AF4 PDX arm was run with the dual inhibitor only; tazemetostat was not tested in the PDX.

    Evidence that argues against this (1)
    • Because tazemetostat was not tested in the MLL-AF4 PDX, the index-fusion claim in this entry - that tazemetostat would fail there - is an inference from the cell-line comparison plus the mouse genetics, not a measured negative. The honest status is 'predicted to fail and never tested', which is what the counted absence says. NCBI PubMed 2026
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.review

    Therapeutic targeting of H3K27me3 with EZH2-selective inhibitors such as tazemetostat has shown clinical benefit in lymphoma, but their efficacy is limited by functional redundancy with EZH1, which the dual EZH1/2 inhibitor valemetostat overcomes by reactivating tumour-suppressor genes, achieving durable responses in adult T-cell leukaemia/lymphoma and peripheral T-cell lymphoma; resistance can nonetheless emerge through PRC2 gatekeeper mutations and compensatory DNA methylation.

    Polycomb repressive complexes as therapeutic targets in hematologic malignancies. · Yamagishi M, Iwama A, Kitabayashi I · Exp Hematol 2026
    PMID 41443386DOI
    What this evidence cannot show

    A 2026 review, cited for the field's own statement of the EZH1-redundancy limitation and of the resistance routes, not for a primary result. The review does not discuss KMT2A-rearranged leukaemia, so its endorsement of dual inhibition is about ATL and PTCL rather than this disease.

    Evidence that argues against this (1)
    • A review stating a general limitation is not evidence that the limitation applies in a specific disease. The EZH1-redundancy claim is independently supported in this entry by the MLL-AF9 genetics (PMID 22396593), the B-ALL head-to-head (PMID 36280659) and the DepMap/expression panel - which is why the conclusion does not rest on the review. Ito J et al. 2023
  7. E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    PubMed `tazemetostat AND (KMT2A OR MLL)` returns exactly 1 record, and title-checking identifies it as a pharmacology review (PMID 35753676) rather than a study - so no primary study of tazemetostat in KMT2A-rearranged leukaemia has been published.

    What this evidence cannot show

    The symbol set covers the INN and the query translation expands 'tazemetostat' through a Supplementary Concept, but the development codes **EPZ-6438 and E7438 were not searched separately**, and this project's own alias rule says a count keyed to one name can under-count substantially. The complex form was run per Rule 64 and recovers no tazemetostat study. The abstract-level negative does not cover conference abstracts, registry entries or unpublished screening data - an absence of publication, not of experiment.

    Evidence that argues against this (1)
    • The compound IS represented in a leukaemia experiment under its development code: PMID 36280659 tests 'EPZ-6438' head to head against DS-3201 in B-ALL lines. A query on the INN alone therefore misses the single most relevant use of this drug in this disease, which is the alias-precision failure this project has logged for SEPT9/SEPTIN9 and SETD1B/Set1B. Ito J et al. 2023
Connected to