Oncofusion

start / graph / drug-pinometostat

Drugstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q6b-therapeutics2

Pinometostat (EPZ-5676) — the DOT1L inhibitor, and the most instructive failure in this field

also called EPZ-5676 · DOT1L inhibitor · Epizyme EPZ-5676

Pinometostat blocks DOT1L, an enzyme the KMT2A fusion hijacks to keep cancer genes switched on. It worked beautifully in the laboratory and almost not at all in patients, and understanding why is the single best warning we have about menin inhibitors.

The same thing, for a scientist

Pinometostat is a first-in-class S-adenosylmethionine-competitive inhibitor of DOT1L, the H3K79 methyltransferase recruited by the AEP/DotCom complex to KMT2A-fusion target loci. Despite exemplary target validation, on-target pharmacodynamics (measurable H3K79me2 loss in patient marrow) and a clean safety profile, the adult phase 1 produced only 2 complete remissions in 51 patients, both in t(11;19)/KMT2A::MLLT1. The paediatric phase 1 (NCT02141828, n=18) completed in 2016 and has never posted results or produced an indexed primary publication. The compound was not advanced. The failure was not target validation, dose or safety — it was pharmacology (continuous IV infusion), kinetics (H3K79me2 is a slow, diluting mark) and, most importantly, the adaptive capacity of the leukemia cell.

Go deeper
What was actually shown in patients

The adult first-in-human study enrolled 51 patients across 6 dose-escalation cohorts (n=26) and 2 expansion cohorts (n=25), at 54 and 90 mg/m²/day as a continuous intravenous infusion in 28-day cycles. A maximum tolerated dose was never reached. The most common adverse events were fatigue (39%), nausea (39%), constipation (35%) and febrile neutropenia (35%). Two patients achieved complete remission, both with t(11;19) — i.e. KMT2A::MLLT1 — at 54 mg/m²/day 1.

The authors' own summary is the honest one: proof of concept for targeting DOT1L in KMT2A-rearranged leukemia, with efficacy as a single agent that was modest.

The paediatric study, NCT02141828, enrolled 18 patients aged >3 months to <18 years with relapsed/refractory MLL-rearranged ALL, AML or acute leukemia of ambiguous lineage, requiring >10% marrow blasts. It ran May 2014 to June 2016 at nine sites and is listed as COMPLETED with no results posted 2. We could find no indexed primary publication of its efficacy in PubMed. We therefore make no claim about response rates in children: that number exists only in conference material as far as we can verify, and per this graph's rules we do not launder it.

Why a well-validated target produced a disappointing drug — and why it matters for menin1 deeper

Four distinct failure modes, only one of which is about the target.

1. Delivery. Pinometostat required a continuous intravenous infusion, 21 or 28 days per cycle, because of rapid clearance and extensive oxidative metabolism; faecal excretion accounted for ~79-81% of dose in rat and dog 3. In a relapsed infant with a central line and a marrow full of blasts, a permanent infusion pump is a serious practical ceiling on exposure and on combination design. Menin inhibitors are oral. This failure mode does not transfer.

2. Mark kinetics. H3K79me2 has no known demethylase; it is lost by replication-coupled dilution. So the pharmacodynamic effect lags the pharmacokinetic one by cell divisions, and any interruption of infusion allows recovery. Transcriptional output therefore moves slowly. Menin inhibition evicts the fusion from chromatin within hours, so its kinetics are much more favourable. This failure mode also does not transfer.

3. The leukemia can stop needing the enzyme without stopping needing the protein. This is the transferable lesson and it is the important one. A pediatric KMT2A::AFF1 B-ALL line driven to acquired pinometostat resistance became largely independent of DOT1L-mediated H3K79 methylation while still requiring the physical presence of DOT1L, HOXA9 and the KMT2A::AFF1 fusion itself 4. Resistant cells selectively lost epigenetic regulation and expression of some fusion targets (PROM1/CD133) while others (HOXA9, CDK6) were untouched, and they upregulated myeloid genes including CD33 and LILRB4. In other words the cell kept the oncogenic circuit and rewired the way it was maintained. The direct menin analogue is obvious: menin inhibitors also block a scaffolding interaction, not a cell-autonomous requirement for the fusion protein, and the same rewiring is exactly what 'adaptive non-genetic resistance' means.

4. Monotherapy against a redundant network. An shRNA sensitiser screen in DOT1L-inhibitor-treated KMT2A-rearranged lines found that hitting additional nodes of the MLL complex markedly enhanced killing, with menin-MLL interaction inhibition as the standout combination, suppressing MLL-fusion and MYC target genes more deeply than either alone while sparing normal haematopoiesis and non-KMT2A leukemias 5. The same synergy was later reproduced in primary KMT2A-rearranged ALL with revumenib plus pinometostat 6.

The uncomfortable read-across

If you accept that pinometostat failed mostly for pharmacology and partly for adaptation, menin inhibitors inherit only the adaptation problem — and they inherit it fully. If instead you believe pinometostat failed because blocking one chromatin node is never enough, then menin monotherapy is on the same trajectory and its 2024-2025 single-agent approvals are the beginning of the same story, not a different one. The AUGMENT-101 and KOMET-001 response durations, measured in months, are consistent with either reading. The discriminating experiment is whether frontline combination changes durability, which is precisely what the phase 3 programmes are now testing.

Is DOT1L dead as a target?

Not obviously. Pinometostat sensitised paediatric AML cells to sorafenib irrespective of KMT2A status, implying DOT1L-directed differentiation has combination value beyond the KMT2A subgroup 7. And the revumenib–pinometostat synergy in primary KMT2A-rearranged ALL 6 is a live, cheap, mechanistically coherent combination that nobody is running clinically, largely because no DOT1L inhibitor is in active development. That is a programmatic gap, not a biological verdict.

Facts and figures
targetgene-dot1l
modalitysmall molecule (SAM-competitive methyltransferase inhibitor)
routecontinuous intravenous infusion
statusDevelopment discontinued; never approved. Adult phase 1 completed (51 patients); paediatric phase 1 NCT02141828 completed 2016 with no posted results.
adult phase1 n51
adult phase1 cr n2
adult phase1 cr genotypet(11;19) / KMT2A::MLLT1
pediatric phase1 n18
trials
[
  "NCT01684150",
  "NCT02141828"
]
doses studied mg m2 day
[
  54,
  90
]
key ae
[
  "fatigue 39%",
  "nausea 39%",
  "constipation 35%",
  "febrile neutropenia 35%"
]
resistance
[
  "H3K79me-independent maintenance of the fusion program with retained requirement for DOT1L protein, HOXA9 and KMT2A::AFF1",
  "selective loss of a subset of fusion target genes with myeloid gene upregulation (CD33, LILRB4)"
]
How this could bite us back
  • Adaptive escape: the leukemia can retain the fusion and its target program while dropping its dependence on the specific enzymatic step being drugged. This is the mechanism most likely to be shared with menin inhibitors.
  • Lineage drift: pinometostat-resistant KMT2A::AFF1 ALL cells upregulated myeloid genes (CD33, LILRB4), the same direction as lineage switch under CD19-directed therapy, so epigenetic therapy pressure may itself prime immunotherapy escape.
  • Delivery: continuous IV infusion in an infant with a central line is a real, underrated barrier that limited exposure and effectively prevented combination development.
  • Essential-function risk: DOT1L is required for normal haematopoiesis in mouse conditional knockouts, so a more potent DOT1L inhibitor would not obviously have a wider therapeutic index.
  • Late effects: any epigenetic therapy given to an infant during the period of maximal developmental methylation remodelling carries unknown long-term consequences that a 28-day phase 1 cannot see.
Still open
  1. Would an oral, long-residence-time DOT1L inhibitor combined with a menin inhibitor be curative where either alone is not? The preclinical synergy is strong and there is currently no clinical DOT1L asset to test it with.
  2. Did the paediatric trial NCT02141828 see responses? The result has never been posted or indexed, which means the field is reasoning about DOT1L in children from adult data only.
  3. Are t(11;19)/KMT2A::MLLT1 leukemias genuinely more DOT1L-dependent than KMT2A::AFF1, or were the two responders chance? ENL's YEATS domain is a direct DotCom contact, so there is a mechanistic reason to look.
  4. Does H3K79me2-independent maintenance of the fusion program, as seen in the resistance model, also occur under menin inhibition? If so, the same non-genetic escape route is already waiting for revumenib.
Every source behind this entry (7)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In the adult phase 1, pinometostat given as a continuous IV infusion to 51 patients with advanced acute leukemia reduced H3K79 methylation on target but produced only 2 complete remissions, both in patients with t(11;19); MTD was not reached and single-agent efficacy was modest.

    PMID 29724899DOI
    What this evidence cannot show

    Phase 1 in heavily pretreated adults with advanced disease, a population in which almost nothing works; this is not a fair test of the target in frontline or paediatric disease. The two responders share one fusion partner (MLLT1/ENL), so partner-specific sensitivity cannot be excluded on n=2.

    Evidence that argues against this (1)
    • The 'DOT1L is a weak target' reading has a published alternative explanation: Epizyme's own resistance work notes that the adult trial produced responses followed by relapse, and shows in five MLL-rearranged lines that treatment-emergent resistance is driven by ABCB1/P-glycoprotein efflux upregulation (KOPN-8) or by efflux-independent routes (NOMO-1), with no target-site MEN1-style mutations. On that reading the clinical shortfall is a pharmacology problem - a 24-hour-infusion drug being pumped back out of the cell - not evidence that DOT1L is dispensable. Campbell CT et al. 2017
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.trial_registry

    The paediatric phase 1 of pinometostat (NCT02141828) in relapsed/refractory MLL-rearranged leukemia enrolled 18 patients aged >3 months to <18 years, ran from May 2014 to June 2016, and is listed as completed with no results posted.

    NCT02141828
    What this evidence cannot show

    A registry record establishes design, enrolment and status only. It says nothing about efficacy. Absence of a posted result is not evidence of a negative result, but it is the reason no paediatric response rate is asserted anywhere in this node.

    Evidence that argues against this (1)
    • The registry cannot tell you what this trial found, and the adult data suggest the answer was unremarkable. Raw ClinicalTrials.gov v2 API, retrieved 2026-09-19: NCT02141828 is COMPLETED with an ACTUAL enrolment of 18, primary completion 2016-05, hasResults FALSE, and a last update of 2023-06-15 — ten years after completion with nothing posted. What is publicly known about single-agent pinometostat comes from the adult phase 1 (n=51, six dose-escalation cohorts, 28-day continuous intravenous infusion): pinometostat reduced H3K79 methylation, exposure was approximately dose-proportional, no maximum tolerated dose was reached, and COMPLETE REMISSION OCCURRED IN ONLY TWO PATIENTS, both with t(11;19). The authors describe single-agent efficacy as modest. An unreported paediatric study is not neutral evidence; it sits beside a published adult study with a 2/51 CR rate. Stein EM et al. 2018
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Pinometostat undergoes extensive oxidative metabolism with faecal excretion as the major elimination route (79% rat, 81% dog), and was administered by 24-hour IV infusion, constraining exposure and dosing convenience.

    Metabolism and disposition of the DOT1L inhibitor, pinometostat (EPZ-5676), in rat, dog and human · Waters NJ, Smith SA, Olhava EJ, et al. · Cancer Chemotherapy and Pharmacology 2016
    PMID 26645404DOI
    What this evidence cannot show

    Mass-balance data are principally from rat and dog; human metabolite profiling was comparative rather than a full human ADME study.

    Evidence that argues against this (1)
    • Delivery inconvenience is not a sufficient explanation for pinometostat's clinical ceiling, so citing the infusion schedule as the constraint misattributes the failure. In the adult phase 1, the 28-day continuous intravenous infusion DID deliver the drug: plasma concentrations rose approximately dose-proportionally, reached apparent steady state by 4-8 hours, fell rapidly on cessation, no maximum tolerated dose was reached across six cohorts, and pharmacodynamic reduction of H3K79 methylation was demonstrated during dose escalation. Target engagement was achieved and exposure was not the limiting variable; complete remission still occurred in only 2 of 51 patients. The problem was biological response, not pharmacokinetics. Stein EM et al. 2018
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Paediatric KMT2A::AFF1 B-ALL cells driven to acquired pinometostat resistance became largely independent of DOT1L-mediated H3K79 methylation yet still required the physical presence of DOT1L, HOXA9 and the KMT2A::AFF1 fusion, selectively losing some fusion target genes while upregulating myeloid genes including CD33 and LILRB4.

    Modelling acquired resistance to DOT1L inhibition exhibits the adaptive potential of KMT2A-rearranged acute lymphoblastic leukemia · Schneider P, Crump NT, Arentsen-Peters STCJM, et al. · Experimental Hematology & Oncology 2023
    PMID 37740239DOI
    What this evidence cannot show

    A single cell-line model of acquired resistance generated by prolonged in vitro drug exposure. Cell-line adaptation is not proof that the same route is taken in patients, and no patient samples with acquired pinometostat resistance exist to check against, because the drug was never used long enough clinically.

    Evidence that argues against this (1)
    • An independent, earlier set of treatment-emergent resistance models reached a different conclusion about the dominant route. Across five MLL-rearranged lines (KOPN-8, MOLM-13, MV4-11, NOMO-1, SEM) the primary mechanism in KOPN-8 was increased ABCB1 (P-glycoprotein) drug efflux - a pharmacokinetic escape that says nothing about H3K79 dependence - while NOMO-1 escaped by an efflux-independent route. If efflux is the common route, the 'the leukemia rewires away from H3K79 methylation' interpretation is not the general case. Campbell CT et al. 2017
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    A deep-coverage shRNA sensitiser screen in DOT1L-inhibitor-treated KMT2A-rearranged lines found that combining DOT1L inhibition with inhibition of the MLL-menin interaction markedly enhanced differentiation and killing across KMT2A-rearranged models including primary cells, while sparing normal haematopoiesis and non-KMT2A leukemias.

    Complementary activities of DOT1L and Menin inhibitors in MLL-rearranged leukemia · Dafflon C, Craig VJ, Méreau H, et al. · Leukemia 2017
    PMID 27840424DOI
    What this evidence cannot show

    Preclinical only, using tool menin inhibitors of the MI-2/MI-503 era rather than revumenib-class compounds. 'Sparing normal haematopoiesis' was assessed in short-term assays, not in long-term competitive repopulation.

    Evidence that argues against this (1)
    • The 'sparing normal haematopoiesis' half of this claim is directly contradicted by conditional genetics. Deletion of Dot1l in postnatal mice caused pancytopenia and failure of haematopoietic homeostasis, and Dot1l-deficient cells only minimally reconstituted recipient marrow in competitive transplantation - precisely the long-term assay the screen paper did not run. Sustained DOT1L inhibition, especially in combination, therefore carries a real marrow-failure liability that short-term differentiation assays cannot see. Jo SY et al. 2011
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In primary KMT2A-rearranged ALL, revumenib and pinometostat are significantly synergistic; KMT2A-rearranged ALL shows rapid dose-dependent cell death with revumenib whereas KMT2A-rearranged AML responds more slowly with myeloid differentiation, and acquired revumenib resistance in ALL arises either through MEN1 mutations or independently of them.

    Distinct Responses to Menin Inhibition and Synergy with DOT1L Inhibition in KMT2A-Rearranged Acute Lymphoblastic and Myeloid Leukemia · Adriaanse FRS, Schneider P, Arentsen-Peters STCJM, et al. · International Journal of Molecular Sciences 2024
    PMID 38892207DOI
    What this evidence cannot show

    In vitro primary-sample and cell-line work, not in vivo; synergy metrics in ex vivo leukemia cultures are notoriously sensitive to assay duration. No clinical combination of a menin inhibitor with a DOT1L inhibitor exists, and no DOT1L inhibitor is currently in active clinical development, so this combination is presently untestable in patients.

    Evidence that argues against this (1)
    • In vitro synergy does not price in DOT1L inhibition's on-target immunological cost, which matters most in exactly the patients this combination would treat. DOT1L inhibition with pinometostat reduces the innate antiviral response: knockdown of DOT1L increases influenza and vesicular stomatitis virus replication and lowers the antiviral response, pinometostat reduces Sendai-virus-induced innate signalling, and genes of the host innate response to pathogens (RUBICON, TRIM25, BCL3) are deregulated in pinometostat-treated human lung epithelial cells. The authors conclude that patients on pinometostat may have increased vulnerability to infection and advise close infection follow-up. Layering that onto menin inhibition in infants already rendered neutropenic and B-cell-depleted by Interfant-style chemotherapy and CD19-directed agents compounds a risk that a cell-viability synergy assay cannot see. Marcos-Villar L et al. 2019
  7. E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Pinometostat induces differentiation in paediatric AML cells and sensitises them to sorafenib irrespective of KMT2A rearrangement status, indicating DOT1L-directed activity beyond the KMT2A-rearranged subgroup.

    PMID 32698374DOI
    What this evidence cannot show

    Cell lines plus a limited set of primary paediatric AML samples, in vitro; no in vivo confirmation of the combination and no clinical follow-up.

    Evidence that argues against this (1)
    • The claim that DOT1L activity extends beyond KMT2A-rearranged disease was not borne out clinically: in the adult phase 1 the only two complete remissions occurred in KMT2A-rearranged patients (both t(11;19)), and genetically Dot1l deletion abolished MLL-AF9 transformation but left Hoxa9/Meis1- and E2A-HLF-driven leukemia intact. Both the clinical and the genetic evidence argue the dependency is fusion-specific, and the sorafenib-sensitisation result is an in-vitro combination effect rather than a demonstration of DOT1L dependency in KMT2A-germline AML. Jo SY et al. 2011
Connected to