Oncofusion

start / graph / gene-brd4

Proteinstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q5c-interactome2

BRD4 — the best-drugged protein in this beat, and the one whose clinical result is already in and disappointing

also called BRD4 · bromodomain-containing protein 4 · HUNK1 · BET family

BRD4 reads acetyl marks on chromatin and was one of the great hopes for KMT2A-fusion leukemia after two 2011 Nature papers. Drugs that block it were built, taken into patients with leukemia, and largely did not work.

The same thing, for a scientist

BRD4 (O60885, 1,362 aa) is a BET-family acetyl-lysine reader with tandem bromodomains (58-164, 348-457) and an extraterminal (NET) domain (600-682). It is not a SEC or AEP subunit; quantitative chemoproteomics placed BET proteins in association with MLL fusions alongside SEC and PAFc on chromatin. It is DepMap common-essential (-1.01, 94.7% of lines) and maximally constrained (LOEUF 0.171). The BET inhibitor OTX015 produced only 3 complete responses among 41 patients with acute leukaemia in a phase 1 trial, and no BET-inhibitor trial restricted to KMT2A-rearranged disease has been identified.

What it looks like
BRD41,362 aa
Bromodomain 1Bromodomain 1 · residues 58–164Bromodomain 2Bromodomain 2 · residues 348–457ExtrExtraterminal (NET) domain · residues 600–68205001,000
loading AF-O60885-F1…
AlphaFold AF-O60885-F1 ↗
colourshape

This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.

very high (90+)confident (70–90)low (50–70)very low (<50)
I have never read one of these pictures before

A protein is one long chain of amino acids that folds up. The ribbon traces that chain from one end to the other; it is not a wire, it is a simplification of where the backbone runs. Drag to rotate, scroll to zoom.

Domains are the stretches of chain that do a particular job, and they are what the coloured blocks above the picture mark out. Click one and the view flies to it. In this project the important question about any domain is usually whether the fusion still has it — the labels say lost where it does not.

Anything in yellow sticks is a bound small molecule, which is usually the drug.

Go deeper
The preclinical case, and exactly how strong it is

Two papers in the same 2011 issue of Nature established BET inhibition as a strategy in this disease. Dawson and colleagues showed that BET proteins associate with MLL fusions as part of SEC and PAFc on chromatin and that inhibiting BET recruitment is an effective treatment for MLL-fusion leukaemia 1. Zuber and colleagues identified Brd4 in an RNAi screen as a therapeutic target in AML 2.

Two things about that evidence deserve emphasis. BRD4 is not a subunit of SEC or AEP; it is co-recruited as an acetyl-lysine adaptor. And Zuber's work used shRNA rather than knockout, in AML generally rather than KMT2A-rearranged disease specifically.

What happened in patients

The BET inhibitor OTX015/birabresib was taken into a phase 1 dose-escalation study in acute leukaemia. Among 41 patients (36 with AML, median age 70) there were 3 complete or incomplete-count complete responses lasting 2–5 months and 2 partial blast clearances. The recommended phase 2 dose of 80 mg/day on a 14-on/7-off schedule was limited by gastrointestinal toxicity and fatigue at higher doses, and no predictive biomarker was identified 3. The registry record for the study, NCT01713582, is completed with 141 participants across AML, ALL, DLBCL and myeloma 4.

This is a clear clinical disappointment relative to the 2011 preclinical promise, and it is the most important fact in this entry.

One caution that cuts the other way: no BET-inhibitor trial restricted to KMT2A-rearranged leukemia has been identified 4. Every trial has been all-comer AML or mixed haematologic malignancy, in largely elderly populations. So the clinical record does not test the specific hypothesis that KMT2A-rearranged disease is BET-dependent. This graph should not say 'BET inhibitors failed in KMT2A-rearranged leukemia'; it should say 'BET inhibitors performed poorly in unselected acute leukaemia and were never tested in the selected population'.

Why the therapeutic index was always going to be the problem

BRD4 is DepMap common-essential: mean Chronos -1.015, leukemia -1.046, depleted in 94.7% of 1,258 lines 5. gnomAD gives pLI 1.0 and LOEUF 0.171, with 16 observed loss-of-function variants against 142.1 expected — among the most constrained genes in this beat 5. Leukemia-restricted correlation with MEN1 is 0.177, crossing zero 6.

That profile was visible before the clinical trials and predicts what they found: a compound with real on-target pharmacology, real toxicity, and a narrow window. It is the same conclusion this graph reached for CDK9, PRMT1, ELL and the rest of the hijacked machinery, arrived at here by an expensive route.

BRD4 has 618 PDB entries — more than the rest of this beat combined — including the canonical JQ1 co-crystal at 1.60 Å (3MXF) 7. Chemical tractability was never the limiting factor. This is the clearest case in the graph that druggability and therapeutic window are independent properties.

Facts and figures
uniprotO60885
gene symbolBRD4
ensemblENSG00000141867
length AA1362
alphafold IDAF-O60885-F1
PDB count618
roleacetyl-lysine reader co-recruited with SEC and PAFc on chromatin; not a subunit of either complex
druggabilityMaximally tractable chemically and clinically disappointing. 618 PDB entries, multiple clinical-stage inhibitors, and a narrow therapeutic window driven by pan-essentiality. No ChEMBL count quoted; a target-level count here would conflate BD1, BD2 and the NET domain.
subnuclear localizationmitotic chromosomes, at the transcription start sites of M/G1 genes, increasing at telophase - the one compartmental address BRD4 has. No nuclear-body assignment; the interphase address is genomic rather than spatial.
shown for fusionPARTIAL. BRD4's chromatin occupancy and its DISPLACEMENT from chromatin are measured in MLL-fusion leukaemia (PMID 21964340, human and murine MLL-AF9 and MLL-AF4), so the chromatin-bound compartment is shown for the fusion context. NO for the mitotic-bookmarking localisation, which has never been demonstrated in a KMT2A-rearranged cell - and that is the localisation the whole BET-inhibitor rationale would need if the drug's effect is partly on mitotic memory.
domains
Bromodomain 158.0Bromodomain 1: 58Bromodomain 2348.0Bromodomain 2: 348Extraterminal (NET) domain600.0Extraterminal (NET) domain: 600
PDB verified
[
  "3MXF",
  "2OSS",
  "4LYS"
]
domains — the numbers behind the chart
[
  {
    "name": "Bromodomain 1",
    "start": 58,
    "end": 164,
    "pfam": null
  },
  {
    "name": "Bromodomain 2",
    "start": 348,
    "end": 457,
    "pfam": null
  },
  {
    "name": "Extraterminal (NET) domain",
    "start": 600,
    "end": 682,
    "pfam": null
  }
]
clinical
{
  "compound": "OTX015 / birabresib / MK-8628",
  "nct": "NCT01713582",
  "status": "completed, n = 141 across AML/ALL/DLBCL/MM",
  "acute_leukaemia_cohort": "41 patients (36 AML, median age 70); 3 CR/CRi lasting 2-5 months, 2 partial blast clearances",
  "rp2d": "80 mg/day, 14 days on / 7 off; limited by GI toxicity and fatigue at >=120 mg",
  "biomarker": "none identified",
  "kmt2a_restricted_trial": "none identified"
}
essentiality
{
  "depmap_common_essential": true,
  "chronos_mean": -1.0148,
  "chronos_mean_leukemia_lines": -1.0456,
  "n_lines": 1258,
  "fraction_lines_depleted": 0.947,
  "gnomad_pLI": 1,
  "gnomad_loeuf": 0.171,
  "gnomad_obs_lof": 16,
  "gnomad_exp_lof": 142.1,
  "mouse_ko": "Not established; no Brd4 knockout citation verified and none asserted",
  "hpa_ntpm": {
    "liver": 31,
    "heart_muscle": 28.5,
    "spleen": 27,
    "testis": 23.6,
    "bone_marrow": 23,
    "lymph_node": 19.4,
    "thymus": 15.8,
    "naive_B_cell": 4.3,
    "classical_monocyte": 2.2
  },
  "hpa_specificity": "Low tissue specificity"
}
codependency men1
{
  "r_leukemia_lines": 0.1768,
  "ci_leukemia": [
    -0.1193,
    0.4244
  ],
  "crosses_zero": true,
  "p_permutation": 0.19696
}
cell type
{
  "evidence_class_key": "concept-expressed-required-maintains-three-evidence-classes",
  "expressed_in": {
    "class": 1,
    "summary": "ubiquitous"
  },
  "required_in": {
    "class": 2,
    "floor_compressed": true,
    "pct_all_lines_depleted": 94.8,
    "all_1258_lines_mean_gene_effect": -1.015,
    "kmt2a_rearranged_lines_mean_gene_effect": -1.132,
    "myeloid_kmt2ar_mean": -1.128,
    "blineage_kmt2ar_mean": -1.138,
    "verdict": "PAN-ESSENTIAL: depleted below -0.5 in 94.8% of 1,258 screened lines, KMT2A-rearranged arm -0.117 relative to all lines and identical between the two lineages. The BET-inhibitor rationale in KMT2A-rearranged leukaemia does not draw support from selectivity in this measurement.",
    "source": "DepMap via Open Targets fetched 2026-09-20; ana-q60-cell-type-versus-fusion"
  },
  "context": "wild-type protein housekeeping",
  "graph_links": []
}
cell state
{
  "maintains": {
    "class": 3,
    "states": [],
    "claim": "not assessed in this beat"
  },
  "note": "BRD4 was carried as a reference-axis gene in the Q60 DepMap panel, not as a worklist gene, so only the class-2 measurement was made. No bounded cell-state query was run for BRD4 and no absence is claimed."
}
localization
{
  "subcellular": "nucleus, chromatin-bound - and distinctively, RETAINED on chromatin through mitosis when most transcription factors dissociate. Brd4 remains bound to the transcription start sites of M/G1 genes throughout mitosis, with binding increasing at telophase alongside H3 and H4 acetylation, and this mark is required for prompt transcription in daughter cells.",
  "subnuclear": "mitotic chromosomes are the one well-defined subnuclear (strictly, sub-chromosomal) address BRD4 has: it marks the TSSs of a defined subset of genes on the condensed mitotic chromosome and does not bind genes expressed later in the cycle. In interphase the published address is genomic (super-enhancers, TSSs) rather than compartmental, and the condensate literature on BRD4 is subject to this graph's own condensate evidence standards.",
  "how_shown": "mitotic-chromatin ChIP plus imaging across the cell cycle, with Brd4 knockdown showing loss of de novo M/G1 transcription (PMID 19812244); displacement from chromatin in the index disease by BRD3/4 ChIP under I-BET151 (PMID 21964340)",
  "systems": "the mitotic bookmarking work is in non-haematopoietic mammalian cells; the chromatin-displacement work is in human and murine MLL-fusion leukaemia including MLL-AF9 and MLL-AF4 models",
  "node": "localization-nuclear-matrix-mitotic-scaffold-kmt2a"
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -1.1315,
  "chronos_mean_other_leukaemia": -1.0769,
  "difference": -0.0546,
  "pan_essentiality_all_lines": -1.0148,
  "selectivity_rank": 3122,
  "rank_denominator": 17728,
  "provenance": "Joined 2026-09-21 from analyses/depmap-kmt2a-selective-scan-full (generated 2026-09-19T10:12:06Z), the project's own genome-wide scan over 11 KMT2A-rearranged and 43 other leukaemia DepMap lines. NOT re-fetched and NOT independent of that scan: every caveat it carries applies here, including that the KMT2A-rearranged arm is a small number of lines drawn from nine individuals, that the rank denominator is 17,728 genes, and that a difference is not a dependency. Added because only 26 of 308 gene entries carried this number, which made any KMT2A-selectivity question answerable for 8% of the corpus.",
  "how_to_read": "A negative difference means MORE required in KMT2A-rearranged lines than in other leukaemia lines. A rank near 17,728 is ANTI-selective. Neither is evidence of a fusion-driven dependency on its own: check pan_essentiality_all_lines before calling anything selective, and note that a paralogue-redundant gene can score near zero for reasons that have nothing to do with this disease - see ana-the-paralogue-redundancy-excuse-is-not-tested-and-my-proxy-decided-the-answer."
}
How this could bite us back
  • Essential-function toxicity: required by 94.7% of cell lines, with GI toxicity and fatigue as clinical dose-limiters and no biomarker to select patients.
  • The clinical result is already in and it is weak, which is a much more expensive form of bite-back than a preclinical concern.
  • Trial-design trap: because BET inhibitors were tested in unselected elderly AML, the field has a negative result attached to a hypothesis it never tested, which makes it harder to fund the trial that would.
  • The mitotic-bookmarking function is a resistance argument nobody has tested in this disease. If BRD4's role in KMT2A-rearranged leukaemia includes carrying transcriptional memory across mitosis (PMID 19812244), then a BET inhibitor given to a slowly cycling leukaemic stem cell engages a different amount of its target than one given to bulk blasts, and the mitotic pool has never been measured in a KMT2A-rearranged cell.
  • The cell has an endogenous BRD4-displacement pathway. JNK activation releases Brd4 from mitotic chromosomes and that release is PROTECTIVE against drug-induced mitotic stress (PMID 22567088), so pharmacological displacement may phenocopy a survival response rather than a lethal one.
Still open
  1. Would a BET inhibitor perform differently in a KMT2A-rearranged-selected population? The 2011 preclinical case was specifically about MLL fusions and the clinical testing was specifically not, so the hypothesis the field got excited about has never actually been tested.
  2. Does BD2-selective or NET-domain-directed chemistry widen the window that pan-BET inhibition could not? BRD4 knockout is more severe than bromodomain occupancy, which suggests the window may be modality-dependent.
  3. Is there value in BET inhibition as a combination partner rather than a single agent in KMT2A-r disease, given that its single-agent responses were short?
  4. This graph's target-bet-bromodomain-kmt2a-r entry should be reconciled with the OTX015 clinical result recorded here; the two should not carry different levels of optimism.
Every source behind this entry (9)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    BET proteins associate with MLL fusion proteins on chromatin as part of the Super Elongation Complex and PAF complex, and inhibition of BET recruitment to chromatin is an effective treatment for MLL-fusion leukaemia in preclinical models.

    Inhibition of BET recruitment to chromatin as an effective treatment for MLL-fusion leukaemia. · Dawson MA, Prinjha RK, Dittmann A, Giotopoulos G, Bantscheff M, Chan WI, et al., Huntly BJP, Kouzarides T · Nature 2011
    PMID 21964340DOI
    What this evidence cannot show

    BRD4 is co-recruited as an acetyl-lysine adaptor, not as a subunit of SEC or AEP, so 'BET is part of the fusion complex' overstates it. Preclinical models only. The subsequent clinical results in acute leukaemia were weak.

    Evidence that argues against this (1)
    • In patients, the BET inhibitor OTX015 produced only 3 complete responses among 41 patients with acute leukaemia, with no predictive biomarker identified. Berthon C et al. 2016
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    An RNAi screen identified Brd4 as a therapeutic target in acute myeloid leukaemia.

    RNAi screen identifies Brd4 as a therapeutic target in acute myeloid leukaemia. · Zuber J, Shi J, Wang E, Rappaport AR, Herrmann H, Sison EA, et al., Lowe SW, Vakoc CR · Nature 2011
    PMID 21814200DOI
    What this evidence cannot show

    shRNA knockdown rather than knockout, in AML generally rather than KMT2A-rearranged disease specifically. RNAi screens are subject to off-target seed effects, which was a recognised issue in screens of this vintage.

    Evidence that argues against this (1)
    • An RNAi screen identifying Brd4 is a 2011-era murine result whose reproducibility question is not whether BRD4 matters — it is pan-essential — but whether anything in the screen was selective for the fusion. Oncofusion knowledge graph (internal analysis) 2026
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In a phase 1 dose-escalation study of the BET inhibitor OTX015 in 41 patients with acute leukaemia (36 AML, median age 70), there were 3 complete or incomplete-count complete responses lasting 2-5 months and 2 partial blast clearances, with a recommended phase 2 dose of 80 mg/day 14-on/7-off limited by gastrointestinal toxicity and fatigue, and no predictive biomarker identified.

    Bromodomain inhibitor OTX015 in patients with acute leukaemia: a dose-escalation, phase 1 study. · Berthon C, Raffoux E, Thomas X, Vey N, Gomez-Roca C, Yee K, et al. · Lancet Haematology 2016
    PMID 27063977DOI
    What this evidence cannot show

    A phase 1 dose-escalation study in a heavily pretreated, elderly, unselected population is not designed to detect efficacy and a low response rate is not by itself a refutation of the target. The population is the opposite of the one this project cares about in almost every respect: adult, myeloid, unselected for KMT2A rearrangement.

    Evidence that argues against this (1)
    • Three responses in 41 patients with a median age of 70 is the clinical answer to the RNAi screen, and it is the number this entry should lead with. BET inhibition has been tested in acute leukaemia and did not deliver. Blood 2018
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.trial_registry

    NCT01713582, the phase 1 dose-finding study of OTX015/MK-8628 in haematological malignancies, is completed with 141 participants across AML, ALL, DLBCL and myeloma; no BET-inhibitor trial restricted to KMT2A-rearranged leukemia was identified.

    NCT01713582
    What this evidence cannot show

    The absence of a KMT2A-restricted BET trial is the result of a bounded registry search and could be wrong. Registry records also lag and do not capture investigator-initiated studies reliably.

    Evidence that argues against this (1)
    • A completed trial with 141 participants across four diseases leaves very few per disease, and 'no BET-inhibitor trial registered in KMT2A-rearranged disease' is an absence measured through a registry search whose query formulation is not stated. Journal of Clinical Oncology 2025
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    BRD4 is DepMap common-essential with mean Chronos -1.015 across 1,258 lines and depletion in 94.7% of lines; gnomAD reports pLI 1.0 and LOEUF 0.171 with 16 observed against 142.1 expected loss-of-function variants; HPA reports low tissue specificity with bone marrow 23.0 nTPM.

    What this evidence cannot show

    DepMap release version not stated by the API. CRISPR knockout of BRD4 is more severe than reversible bromodomain occupancy by a BET inhibitor, which leaves the NET domain and scaffolding functions intact, so Chronos overstates what a drug does.

    Evidence that argues against this (1)
    • Pan-essentiality is the finding here and it argues against the target rather than for it. A gene depleted in more than 90% of all cell lines is required by the patient's normal marrow as much as by the leukaemia, so its appearance in a KMT2A complex is a statement about the complex and not a therapeutic opening. BRD4 at 94.7% of lines depleted with pLI 1.0 is the clearest case in this beat. gnomAD, Broad Institute 2026
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    Across 54 DepMap leukemia lines BRD4 correlates with MEN1 at r = 0.177, confidence interval -0.119 to 0.424, crossing zero.

    Leukemia-restricted co-dependency with MEN1, bootstrap and permutation testing; analyses/wrad-discrimination/ in this project · Broad Institute DepMap; Open Targets Platform; analysis run by this project · Database and derived analysis 2026
    What this evidence cannot show

    Our own computation in 54 lines. A pan-essential gene has little variance to correlate, so this null is uninformative about biology.

    Evidence that argues against this (1)
    • A correlation of 0.177 with an interval crossing zero is a null result, and the entry should say so plainly. It is also a co-dependency measure over 54 lines, which for a pan-essential gene has almost no dynamic range to detect anything. Oncofusion knowledge graph (internal analysis) 2026
  7. E7established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    BRD4 has 618 PDB entries, including the first bromodomain in complex with JQ1 at 1.60 A (3MXF) and the apo first bromodomain at 1.35 A (2OSS).

    What this evidence cannot show

    The 618 figure is the count of PDB cross-references in the UniProt record and includes many near-duplicate fragment and inhibitor complexes; the overwhelming majority cover bromodomain 1 alone, not full-length BRD4 or the NET domain.

    Evidence that argues against this (1)
    • Counting PDB entries measures how tractable a protein was to CRYSTALLISE, which correlates with being small, ordered and well-behaved rather than with being druggable. What decides design value is whether the deposited density covers the site that matters, at a resolution that supports modelling, in the relevant state — and this graph has repeatedly found structures described in the direction that overstates what is resolved. 618 entries makes BRD4 one of the most-crystallised proteins in biology, which measures the field's investment in bromodomains after JQ1 and not this protein's relevance to a KMT2A fusion. ChEMBL, EMBL-EBI 2026
  8. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Brd4 is retained on mitotic chromatin at the transcription start sites of M/G1 genes throughout mitosis, with occupancy increasing at telophase alongside histone H3 and H4 acetylation, and it does not bind genes expressed later in the cycle; Brd4 knockdown impairs the de novo M/G1 transcription that follows, so the mitotic chromosome association is a functional mark and not a passive one.

    Brd4 marks select genes on mitotic chromatin and directs postmitotic transcription. · Dey A, Nishiyama A, Karpova T, McNally J, Ozato K · Mol Biol Cell 2009
    PMID 19812244DOI
    What this evidence cannot show

    Not a leukaemia system and not a KMT2A-rearranged one. The gene set is cell-cycle genes, not the HOXA/MEIS1 programme this graph cares about, so nothing here says the fusion's targets are bookmarked. Mitotic chromatin ChIP requires synchronisation, and synchronisation artefacts are the standard objection to this class of experiment.

    Evidence that argues against this (1)
    • The mitotic chromosome association is conditional, not constitutive: activating JNK triggers release of Brd4 from mitotic chromosomes, and that release mediates protection from drug-induced mitotic stress. So 'BRD4 is on mitotic chromosomes' is a statement about a signalling state, and any therapeutic argument resting on displacing BRD4 from mitotic chromatin has to contend with the cell already having a pathway that does it. Nishiyama A et al. 2012
  9. E9strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In MLL-fusion leukaemia, the mode of action of the BET inhibitor I-BET151 is at least in part the DISPLACEMENT of BRD3/4, PAF1 complex and super elongation complex components FROM CHROMATIN at key genes including BCL2, MYC and CDK6 - making chromatin residence of BRD4 a measured, drug-modifiable localisation in the index disease.

    Inhibition of BET recruitment to chromatin as an effective treatment for MLL-fusion leukaemia. · Dawson MA, Prinjha RK, Dittmann A, Giotopoulos G, Bantscheff M, Chan WI, Robson SC, Chung CW, Hopf C, Savitski MM, Huthmacher C, Gudgin E, Lugo D, Beinke S, Chapman TD, Roberts EJ, Soden PE, Auger KR, Mirguet O, Doehner K, Delwel R, Burnett AK, Jeffrey P, Drewes G, Lee K, Huntly BJ, Kouzarides T · Nature 2011
    PMID 21964340DOI
    What this evidence cannot show

    ChIP occupancy, not imaging: this bounds where BRD4 is on the genome, not where it is in the nucleus. The systems are human MLL-fusion leukaemia cell lines and murine MLL-AF9 plus human MLL-AF4 models - the right disease, with the usual caveat that the murine myeloid arm dominates. This is the paper the whole BET thread in this graph rests on, and the clinical BET programme it launched has largely disappointed, so the displacement result should not be read as a validated therapeutic mechanism.

    Evidence that argues against this (1)
    • Displacement of BRD4 from chromatin is not the same as removing it from the nucleus, and it is reversible on drug withdrawal. Brd4's own regulated release from mitotic chromosomes by JNK shows the cell has endogenous machinery for exactly this manoeuvre, and that in one setting release is PROTECTIVE rather than lethal - which is a mechanistic reason a BET inhibitor might be cytostatic rather than cytotoxic. Nishiyama A et al. 2012
Connected to