Oncofusion

start / graph / target-bet-bromodomain-kmt2a-r

Targetmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q6b-therapeutics2

BET bromodomain inhibition in KMT2A-rearranged leukemia — spectacular preclinically, stalled clinically

also called BRD4 inhibition · I-BET151 · BET inhibitors · bromodomain and extra-terminal

BET proteins are the clips that hold the transcription machinery onto DNA at the genes this leukemia depends on. Drugs that pop those clips off work beautifully in mice and in dishes, but no BET inhibitor has become a leukemia treatment, mainly because of toxicity in blood and gut.

The same thing, for a scientist

BRD4 and its BET family partners read acetylated histones and recruit P-TEFb to super-enhancers, including those driving the KMT2A-fusion transcriptional program. In a preclinical mouse model of KMT2A::AFF1 infant ALL, I-BET151 arrested growth, induced apoptosis, impaired leukemic engraftment, reduced disease burden, resensitised glucocorticoid-resistant cells to prednisolone, and synergised with HDAC inhibitors. Despite twenty years of this kind of data across leukemia, no BET inhibitor is approved for any haematological malignancy. The blocking issue is therapeutic index: BRD4 is DepMap common-essential and the class produces dose-limiting thrombocytopenia and gastrointestinal toxicity.

Go deeper
The preclinical case, specifically in KMT2A::AFF1 infant ALL

I-BET151 in a preclinical mouse model of MLL-AF4 (KMT2A::AFF1) infant ALL 1:

  • arrested growth by blocking cell division and rapidly inducing apoptosis;
  • impaired leukemic engraftment of patient-derived primary samples and lowered disease burden in mice;
  • deregulated the transcriptional program of KMT2A-rearranged ALL;
  • resensitised glucocorticoid-resistant KMT2A-rearranged cells to prednisolone;
  • was more effective combined with HDAC inhibitors.

The glucocorticoid point deserves emphasis because prednisone response is one of the four Interfant risk factors, and a poor prednisone response is a hard prognostic marker. An agent that converts a prednisone poor-responder into a responder would be acting on the exact variable that stratifies these infants.

A second, less obvious dependency: BET recruitment to chromatin in MLL-rearranged AML is sustained by AMPK-dependent acetyl-CoA homeostasis, and combined AMPK and BET inhibition was synergistic in mouse and patient-derived xenograft models 2. That suggests a metabolic route to BET displacement that might have a better therapeutic index than direct bromodomain occupancy.

Why it has not become a drug — and why this matters to the menin programme

BRD4 is DepMap common-essential, with mean Chronos below -0.8 and depletion in more than 92% of cell lines, alongside CDK9, PRMT1, ELL, WDR5, RBBP5, SETD1A, CTR9 and PAF1 3. This is the structural distinction that makes menin an unusually good target and BRD4 an unusually bad one: menin is not common-essential and is selectively required by KMT2A-rearranged and NPM1-mutant cells. A common-essential target has a therapeutic index problem that better chemistry does not fix, because the on-target effect in normal tissue is the toxicity.

That is the whole story in one comparison, and it is the central reasoning of this project: the fusion's partner list contains both kinds of protein, and the programme should spend its effort on the non-essential ones.

BET inhibitors retain a role as combination partners, where lower exposures may suffice. WDR5 WIN-site inhibitors are synergistic with both venetoclax and BET-bromodomain inhibitors in MLL-rearranged cells 4, and BET inhibition is one of the standard proposed partners for overcoming menin-inhibitor resistance. But no BET inhibitor has an approval in leukemia, and this graph does not assert any clinical response rate for the class in KMT2A-rearranged disease, because we could not verify one from a primary source.

Facts and figures
modalitysmall-molecule bromodomain inhibitors; also BET degraders (PROTACs) in development
clinical status in kmt2a rNo BET inhibitor is approved for any haematological malignancy. No verified clinical response rate in KMT2A-rearranged leukemia was found from a primary source.
depmap common essentialtrue
pediatric eligiblefalse
target genes
[
  "BRD4",
  "BRD2",
  "BRD3",
  "BRDT"
]
class dose limiting toxicities
[
  "thrombocytopenia",
  "gastrointestinal toxicity",
  "fatigue"
]
preclinical effects kmt2a r
[
  "growth arrest and apoptosis in KMT2A::AFF1 models",
  "impaired PDX engraftment",
  "resensitisation of glucocorticoid-resistant cells to prednisolone",
  "synergy with HDAC inhibition",
  "synergy with AMPK inhibition",
  "synergy with WDR5 WIN-site inhibition and venetoclax"
]
How this could bite us back
  • Essential-function toxicity is the defining problem: BRD4 is common-essential, so on-target activity in normal haematopoiesis and gut epithelium is the dose-limiting toxicity, and better chemistry does not solve it.
  • Thrombocytopenia from a BET inhibitor in a thrombocytopenic infant with leukemia is additive with the disease itself.
  • Compensation: BET family members are partially redundant and BRD2/BRD3 upregulation is a described escape route from selective BRD4 inhibition.
  • Evidence risk: the size of the preclinical literature on BET inhibitors in KMT2A-rearranged leukemia is wildly out of proportion to the clinical result, which is the classic profile of a target that reads well in models and fails in patients — the same profile DOT1L had.
Still open
  1. Can a BET inhibitor be dosed low enough, intermittently enough, or targeted well enough to exploit the glucocorticoid-resensitisation effect without hitting the thrombocytopenia ceiling? That is the only clinically interesting question left for this class in infant ALL.
  2. Would a BET degrader have a different therapeutic index than a bromodomain occupancy inhibitor, given that the essential function is scaffolding rather than reading?
  3. Does BET inhibition prevent or accelerate adaptive resistance to menin inhibition? Both act on the same super-enhancer machinery, so they might be redundant rather than complementary.
  4. Is the AMPK-acetyl-CoA route a way to displace BET proteins metabolically with a wider window? Nobody has a clinical AMPK inhibitor to test it with.
Every source behind this entry (4)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In a preclinical mouse model of MLL-AF4 (KMT2A::AFF1) infant ALL, the BET inhibitor I-BET151 arrested growth and induced apoptosis, impaired leukemic engraftment of patient-derived samples, lowered disease burden, resensitised glucocorticoid-resistant cells to prednisolone, and was more effective combined with HDAC inhibitors.

    Antileukemic Efficacy of BET Inhibitor in a Preclinical Mouse Model of MLL-AF4+ Infant ALL · Bardini M, Trentin L, Rizzo F, et al. · Molecular Cancer Therapeutics 2018
    PMID 29748211DOI
    What this evidence cannot show

    Preclinical only, using a tool compound (I-BET151) that was never developed clinically. Xenograft disease-burden reduction is not remission. Glucocorticoid resensitisation was shown in vitro and has never been tested in a patient. No BET inhibitor has produced a verified clinical response rate in KMT2A-rearranged leukemia.

    Evidence that argues against this (1)
    • The preclinical-to-clinical gap is documented, not merely suspected. In the first-in-human phase 1 of the BET inhibitor OTX015 in 41 patients with acute leukaemia (36 AML), only 3 patients achieved CR or CRi lasting 2-5 months and 2 more had partial blast clearance; no dose-limiting toxicity was reached until 160 mg/day, but grade 1-2 gastrointestinal, fatigue and skin toxicity from 120 mg/day 'hampered patient compliance', forcing a recommended dose of 80 mg on a 14-on/7-off schedule, and no predictive biomarker of response was identified. Chronic-dosing tolerability, not potency, is what the mouse data fail to predict. Berthon C et al. 2016
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    AMPK maintains the epigenome of MLL-rearranged AML by linking acetyl-CoA homeostasis to BET protein recruitment to chromatin; AMPK deletion reduced acetyl-CoA, histone acetylation and BET occupancy in leukemia-initiating cells, and combined AMPK and BET inhibition synergistically suppressed AML in mouse and patient-derived xenograft models.

    PMID 31697807DOI
    What this evidence cannot show

    Mouse and PDX models of MLL-rearranged AML, not infant ALL. AMPK is itself a broadly important metabolic regulator and systemic AMPK inhibition has no clinical precedent or safety dataset.

    Evidence that argues against this (1)
    • Combination does not rescue BET inhibition in patients either. In the first-in-human study of the pan-BET inhibitor mivebresib, 19 patients received monotherapy (1 CRi, 15 with resistant disease) and 30 received mivebresib plus venetoclax (2 CR, 2 PR, 2 morphologic leukaemia-free state, 12 resistant disease), with serious adverse events in 74% of monotherapy and 88% of combination patients. Pharmacodynamic target engagement was proportional to dose and exposure, so this is a second instance of a BET inhibitor demonstrably hitting its target and not converting that into clinical benefit - the same failure mode as pinometostat. Borthakur G et al. 2021
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    BRD4 is a DepMap common-essential gene, with mean Chronos gene effect below -0.8 and depletion in more than 92% of cell lines, in contrast to MEN1 which is not common-essential (mean Chronos about -0.36) yet is selectively depleted in KMT2A-rearranged and NPM1-mutant leukemia lines.

    What this evidence cannot show

    Cell-line essentiality is an imperfect model of normal-tissue tolerance, and 'common-essential' does not mean 'undruggable' — it means the therapeutic window must come from something other than target selectivity, such as tumour-specific dependency magnitude or drug distribution. The specific values were computed from an API response by another agent in this project rather than read from a DepMap release table; the DepMap release version is unstated. No counter-argument is recorded because none is possible in kind: this is a database readout of a CRISPR screen statistic, not an inferential claim, and no published study disputes the Chronos values. The contestable part is the inference drawn from it - that common-essentiality forecloses a therapeutic window - and that inference is contested by the clinical record, where BET inhibitors reached phase 1 with dose-limiting toxicities that were gastrointestinal and constitutional rather than haematologic (see the counter-evidence on E1 and E2).

    Evidence that argues against this (1)
    • A Chronos essentiality contrast predicts neither the toxicity nor the efficacy that clinical BET inhibition actually produced, so it should not be used as a clinical argument. In the phase 1 dose escalation of the BET inhibitor OTX015 in acute leukaemia (41 patients, 36 with AML, median age 70, two prior lines), NO dose-limiting toxicity occurred until 160 mg/day — the opposite of what 'common essential in more than 92% of cell lines' predicts — yet grade 1-2 gastrointestinal, fatigue and cutaneous effects from 120 mg hampered compliance, so 80 mg on a 14-days-on/7-off schedule was chosen as the recommended dose. Only THREE patients achieved complete remission or CR with incomplete platelet recovery, lasting 2-5 months. The clinical wall was cumulative low-grade intolerance plus insufficient activity at a tolerable exposure, not the acute on-target catastrophe cell-line essentiality implies. Berthon C et al. 2016
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    WDR5 WIN-site inhibitors act in MLL-rearranged cells by suppressing ribosome protein gene transcription, inducing nucleolar stress and activating p53 via MDM4 inactivation, and are synergistic with venetoclax and with BET-bromodomain inhibitors.

    PMID 38682900DOI
    What this evidence cannot show

    In vitro multi-omics in MLL-rearranged cell lines. A p53-dependent mechanism is a serious limitation in relapsed disease where TP53 mutation is enriched, and WDR5 is itself DepMap common-essential.

    Evidence that argues against this (1)
    • Routing the mechanism through p53 activation is a restriction, not only an explanation, and it collides with the co-mutation landscape of the disease this combination would treat. If WDR5 WIN-site inhibitors work by nucleolar stress and MDM4-dependent p53 activation, they are predicted to fail wherever p53 signalling is disabled — and TP53 mutation is one of the alterations that emerges in multiply relapsed KMT2A::AFF1 ALL, appearing alongside RAS-pathway mutations in serially sampled patient-derived xenografts from menin-inhibitor-naive patients whose later samples had already lost menin-inhibitor sensitivity. Combining a p53-dependent agent with a menin inhibitor therefore concentrates the whole strategy on p53-competent, less heavily pretreated leukaemia, which is not the population in which either agent is currently used. Mahdavi L et al. 2026
Connected to