Oncofusion

start / graph / gene-brd3

Genemoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q39-acetylation-cycle

BRD3 - named in the founding BET paper's own mechanism and in one title since, and not once separated from BRD4 by any experiment in this disease

also called RING3L · RING3-like protein · ORFX · bromodomain-containing protein 3

BRD3 is a reader protein that grabs onto acetylated histones, and it is a close relative of the much better known BRD4. The landmark paper that made BET inhibitors a treatment idea for MLL leukaemia says the drug displaces BRD3 and BRD4 together - and in the fifteen years since, no experiment in this disease has ever shown which of the two is actually doing the work.

The same thing, for a scientist

BRD3 (Q15059, 726 aa) is a BET-family acetyl-lysine reader with two bromodomains and an extraterminal domain. The brief asked whether the evidence distinguishes BRD2 and BRD3 from BRD4, and **the plain answer is that it does not, in this disease, and this entry says so rather than manufacturing a distinction**. The founding paper states the mode of action of I-BET151 is *"at least in part, due to the inhibition of transcription at key genes (BCL2, C-MYC and CDK6) through the displacement of **BRD3/4**, PAFc and SEC components from chromatin"* [E1]. BRD3 appears there as half of a slash. The only paper in the beat with BRD3 in its title is likewise a pan-BET study: **I-BET151** in human MLL-AF9 models, reporting critical dependence on BRD3/4 - and finding, importantly, that **the majority of I-BET151-responsive genes were not direct MLL-AF9 targets, and that the HOXA cluster, MEIS1 and CDK6 were NOT affected** [E2]. **Every BET result in this disease rests on a pan-BET inhibitor** - I-BET151, JQ1, I-BET762 - which occupy the acetyl-lysine pocket of BRD2, BRD3 and BRD4 alike. This is structurally identical to the CX-4945 problem this graph already records for CK2: **the compound class cannot see the difference, so an attribution to BRD3 is naming and not measurement.** And the resistance literature closes the loop: I-BET-resistant MLL-AF9 clones are cross-resistant to chemically distinct BET inhibitors **and to genetic knockdown of BET proteins** [E3], which means even the genetic experiment was run at the family level.

Go deeper
What would count as separating BRD3 from BRD4, and why nobody has it

To attribute a BET effect in this disease to BRD3 specifically would take one of:

  1. Single-gene genetic perturbation with paralogue controls - BRD3 knockout or degradation with BRD2 and BRD4 intact, and the reciprocal. The resistance paper's knockdowns were of 'BET proteins' as a group 3.
  2. A BRD3-selective ligand. None exists. BET bromodomains are highly conserved in the acetyl-lysine pocket; the selectivity achieved in this chemical class is BD1-versus-BD2, which cuts across paralogues rather than between them.
  3. A degrader with paralogue selectivity. The degraders in the beat's literature - MZ1, dBET1 - are BET-family degraders; MZ1 has reported BRD4 preference but is characterised in this literature as a BRD4 PROTAC used against BET biology generally, and neither paper separates BRD3.

None of the three has been done in KMT2A-rearranged leukaemia. So the correct statement is: BRD3 is present in the mechanism as reported, has never been individually tested, and cannot currently be individually targeted.

This is worth writing down rather than skipping. The granularity rule in this project says one entry per thing, and the temptation with BET is to write one entry for 'BET proteins' because that is how the literature reads. Doing so would hide the fact that the field's own resolution is a limitation of its tools, not a property of the biology - and hiding it is exactly how this graph ended up attributing CK2 results to a subunit.

Two results that should temper BET enthusiasm in this disease

The mechanism does not run through the fusion's own programme. Carretta et al. found that while MYC and BCL2 were consistently downregulated by I-BET151, the majority of responsive genes were not direct MLL-AF9 targets, and the MLL-AF9-specific targets - the HOXA cluster, MEIS1, CDK6 - were not affected 2. Set that beside menin inhibition, which collapses the HOXA/MEIS1 programme and evicts the fusion from chromatin. BET inhibition in this disease looks like a MYC/BCL2 intervention that happens to be effective, not a fusion-directed one - the same distinction mechanism-kat8-mof-requirement-is-myeloid-not-fusion-specific draws for MOF.

The independent replication failed on the in vivo endpoint. The Reproducibility Project: Cancer Biology replicated selected experiments from the founding paper 4. It confirmed selective growth inhibition of MV4;11 but not K-562, and confirmed the BCL2 decrease in MV4;11 but not K-562. It did not find a statistically significant difference in survival in the disseminated xenograft MLL mouse model, where the original reported increased survival; it did find lower median disease burden in all tissues analysed. The replication authors note differing conditioning regimens and I-BET151 doses as possible factors 4.

That is a nuanced outcome and this entry does not overstate it: the cell-intrinsic findings replicated, the survival benefit did not. A graph that cites the 2011 Nature paper for in vivo efficacy should cite the 2017 eLife replication beside it.

Facts and figures
uniprotQ15059
length AA726
gene symbolBRD3
accession resolutiongene_exact:BRD3 AND organism_id:9606 AND reviewed:true returned exactly 1 reviewed row; primary gene name read back as BRD3 (rule 62 applied).
familyBET (bromodomain and extraterminal) family acetyl-lysine reader; two bromodomains plus an ET domain
separated from brd4 by any experiment in this diseasefalse
why notevery BET result in KMT2A-rearranged leukaemia uses a pan-BET inhibitor (I-BET151, JQ1, I-BET762) that occupies the conserved acetyl-lysine pocket of BRD2, BRD3 and BRD4 alike; the one genetic arm knocked down 'BET proteins' as a group
structural analogue in this graphCX-4945 and CK2 - an ATP-competitive/pocket-competitive compound that cannot discriminate paralogues, making every attribution naming rather than measurement
selectivity achieved in the chemical classBD1 versus BD2, which cuts ACROSS paralogues rather than between them
fusion directednesscontested - the founding paper reports displacement of SEC and PAFc components with BRD3/4, while a later study finds most I-BET151-responsive genes are not direct MLL-AF9 targets and that HOXA, MEIS1 and CDK6 are unaffected
replication statuspartial - in vitro selective growth inhibition and BCL2 downregulation replicated; the in vivo survival benefit did not
resistanceemerges from leukaemia stem cells; cross-resistant to chemically distinct BET inhibitors AND to genetic knockdown of BET proteins; driven in part by increased Wnt/beta-catenin signalling; not efflux- or metabolism-mediated
subnuclear localizationPERICENTRIC HETEROCHROMATIN, stress-conditional and NOT paralogue-resolved: under heat shock HSF1 recruits GCN5, TIP60 and p300 to pericentric heterochromatin, and the resulting hyperacetylation directs recruitment of BRD2, BRD3 and BRD4, all three of which are required for satellite III transcription (PMID 28710461) [E_q95_brd3_loc3]. Otherwise BRD3's address is OCCUPANCY, and there the paralogues ARE resolved: mononucleosomes isolated on Brd3 separately from Brd2 and Brd4 show BRD3 and BRD4 most enriched within HOX GENE CLUSTERS (PMID 22897906) [E_q95_brd3_loc2]. No speckle, no PML body, no Polycomb body, and no condensate claim that survives the token check.
localization shown for fusionNO - nobody has looked in any KMT2A fusion, in any lineage: 0 records for the gene clause against a localisation clause AND the fusion clause and 0 against a subnuclear clause AND the fusion clause, from 6 records placing BRD3 and a fusion term in one abstract [E_q95_brd3_loc4]. But the nearest thing to a fusion-relevant positional result exists and is paralogue-resolved: BRD3 and BRD4 are the two BET proteins most enriched on nucleosomes within HOX gene clusters, measured on separately isolated mononucleosomes, while the depletion arm testing whether that enrichment matters was BRD4-only [E_q95_brd3_loc2]. So for the loci this disease turns on, BRD3's POSITION is known and its REQUIREMENT is not.
paralogues
[
  "gene-brd2",
  "gene-brd4"
]
essentiality
{
  "depmap_common_essential": null,
  "note": "Not queried here, and a single-gene BRD3 score would face the same paralogue-buffering interpretive problem as HDAC1/HDAC2."
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": 0.0942,
  "chronos_mean_other_leukaemia": 0.0823,
  "difference": 0.0118,
  "pan_essentiality_all_lines": 0.0885,
  "selectivity_rank": 10699,
  "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."
}
localization
{
  "CONSTITUTIVE": "NUCLEUS AND CHROMOSOME, and this is the one record in this tranche whose citations are both appropriate: 'Nucleus' from a bromodomain-family relocalisation screen (PMID 25593309) and 'Chromosome / Detected on chromatin' from the Brd2-and-Brd3 hyperacetylated-chromatin paper (PMID 18406326) [E_q95_brd3_loc1]. UniProt tissue specificity: 'Ubiquitous'.",
  "WHAT_LOCALISATION_MEANS_FOR_A_BET_PROTEIN": "an acetylation state rather than a place. Brd3-bound chromatin is enriched in H4K5ac, H4K12ac and H3K14ac and depleted of H3K9me2, and BET recruitment to pericentric heterochromatin under heat shock happens only AFTER GCN5, TIP60 and p300 acetylate it [E_q95_brd3_loc2][E_q95_brd3_loc3]. Anything that redistributes histone acetylation redistributes BRD3.",
  "systems": "human cell lines throughout; nothing haematopoietic and nothing leukaemic in the positional literature."
}
cell type
{
  "measured_in": "human cell lines, unnamed in most abstracts; nothing haematopoietic has been imaged or ChIPed for BRD3 [E_q95_brd3_loc2][E_q95_brd3_loc3].",
  "expression": "UniProt: 'Ubiquitous' - no cell-type handle.",
  "in_kmt2a_r": "NOTHING positional. 6 records place BRD3 and a KMT2A/MLL fusion term in one abstract and 0 of them contain a localisation or subnuclear term [E_q95_brd3_loc4]. This entry's existing field separated_from_brd4_by_any_experiment_in_this_disease stands, and the localisation literature does not change it."
}
cell state
{
  "HEAT_SHOCK": "the one documented state that moves BRD3: heat shock redistributes histone acetylation to pericentric heterochromatin and BET proteins follow [E_q95_brd3_loc3].",
  "TRANSCRIPTIONAL_ACTIVITY": "Brd-bound sites correlate with highly expressed genes, so BRD3's chromatin distribution is a readout of the transcriptional state rather than a fixed map [E_q95_brd3_loc2].",
  "not_recorded": "no cell-cycle-phase, damage-induced or differentiation-dependent redistribution of BRD3 specifically; the damage screen that UniProt cites is family-wide."
}
localization bounded absence
{
  "date": "2026-09-22",
  "tool": "NCBI E-utilities esearch, db=pubmed",
  "run_by": "Q95",
  "shape": "one [tiab] tag per term (rule 94); cytogenetic terms quoted (rule 95); querytranslation read back and stored (rule 90)",
  "absence_class": "never-measured",
  "queries": {
    "gene_clause_alone": {
      "query": "BRD3[tiab] OR RING3L[tiab] OR \"ORFX\"[tiab]",
      "count": 654,
      "querytranslation": "\"BRD3\"[Title/Abstract] OR \"RING3L\"[Title/Abstract] OR \"ORFX\"[Title/Abstract]"
    },
    "gene_AND_loc": {
      "query": "(BRD3[tiab] OR RING3L[tiab] OR \"ORFX\"[tiab]) AND (localization[tiab] OR localisation[tiab] OR subcellular[tiab] OR immunofluorescence[tiab] OR nuclear[tiab] OR cytoplasmic[tiab] OR compartment[tiab])",
      "count": 64,
      "querytranslation": "(\"BRD3\"[Title/Abstract] OR \"RING3L\"[Title/Abstract] OR \"ORFX\"[Title/Abstract]) AND (\"localization\"[Title/Abstract] OR \"localisation\"[Title/Abstract] OR \"subcellular\"[Title/Abstract] OR \"immunofluorescence\"[Title/Abstract] OR \"nuclear\"[Title/Abstract] OR \"cytoplasmic\"[Title/Abstract] OR \"compartment\"[Title/Abstract])"
    },
    "gene_AND_subnuclear": {
      "query": "(BRD3[tiab] OR RING3L[tiab] OR \"ORFX\"[tiab]) AND (\"nuclear speckle\"[tiab] OR \"nuclear speckles\"[tiab] OR nucleolus[tiab] OR nucleolar[tiab] OR \"PML body\"[tiab] OR \"PML bodies\"[tiab] OR \"nuclear matrix\"[tiab] OR \"Polycomb bodies\"[tiab] OR \"Polycomb body\"[tiab] OR subnuclear[tiab] OR \"nuclear bodies\"[tiab] OR \"nuclear body\"[tiab] OR chromocent*[tiab] OR heterochromat*[tiab])",
      "count": 7,
      "querytranslation": "(\"BRD3\"[Title/Abstract] OR \"RING3L\"[Title/Abstract] OR \"ORFX\"[Title/Abstract]) AND (\"nuclear speckle\"[Title/Abstract] OR \"nuclear speckles\"[Title/Abstract] OR \"nucleolus\"[Title/Abstract] OR \"nucleolar\"[Title/Abstract] OR \"PML body\"[Title/Abstract] OR \"PML bodies\"[Title/Abstract] OR \"nuclear matrix\"[Title/Abstract] OR \"Polycomb bodies\"[Title/Abstract] OR \"Polycomb body\"[Title/Abstract] OR \"subnuclear\"[Title/Abstract] OR \"nuclear bodies\"[Title/Abstract] OR \"nuclear body\"[Title/Abstract] OR \"chromocent*\"[Title/Abstract] OR \"heterochromat*\"[Title/Abstract])"
    },
    "gene_AND_fusion": {
      "query": "(BRD3[tiab] OR RING3L[tiab] OR \"ORFX\"[tiab]) AND (KMT2A[tiab] OR MLL[tiab] OR \"MLL1\"[tiab] OR \"MLL-AF4\"[tiab] OR \"MLL-AF9\"[tiab] OR \"MLL-ENL\"[tiab] OR \"MLL-AF10\"[tiab] OR \"t(4;11)\"[tiab] OR \"t(11;19)\"[tiab])",
      "count": 6,
      "querytranslation": "(\"BRD3\"[Title/Abstract] OR \"RING3L\"[Title/Abstract] OR \"ORFX\"[Title/Abstract]) AND (\"KMT2A\"[Title/Abstract] OR \"MLL\"[Title/Abstract] OR \"MLL1\"[Title/Abstract] OR \"MLL-AF4\"[Title/Abstract] OR \"MLL-AF9\"[Title/Abstract] OR \"MLL-ENL\"[Title/Abstract] OR \"MLL-AF10\"[Title/Abstract] OR \"t 4 11\"[Title/Abstract] OR \"t 11 19\"[Title/Abstract])"
    },
    "gene_AND_loc_AND_fusion": {
      "query": "(BRD3[tiab] OR RING3L[tiab] OR \"ORFX\"[tiab]) AND (localization[tiab] OR localisation[tiab] OR subcellular[tiab] OR immunofluorescence[tiab] OR nuclear[tiab] OR cytoplasmic[tiab] OR compartment[tiab]) AND (KMT2A[tiab] OR MLL[tiab] OR \"MLL1\"[tiab] OR \"MLL-AF4\"[tiab] OR \"MLL-AF9\"[tiab] OR \"MLL-ENL\"[tiab] OR \"MLL-AF10\"[tiab] OR \"t(4;11)\"[tiab] OR \"t(11;19)\"[tiab])",
      "count": 0,
      "querytranslation": "(\"BRD3\"[Title/Abstract] OR \"RING3L\"[Title/Abstract] OR \"ORFX\"[Title/Abstract]) AND (\"localization\"[Title/Abstract] OR \"localisation\"[Title/Abstract] OR \"subcellular\"[Title/Abstract] OR \"immunofluorescence\"[Title/Abstract] OR \"nuclear\"[Title/Abstract] OR \"cytoplasmic\"[Title/Abstract] OR \"compartment\"[Title/Abstract]) AND (\"KMT2A\"[Title/Abstract] OR \"MLL\"[Title/Abstract] OR \"MLL1\"[Title/Abstract] OR \"MLL-AF4\"[Title/Abstract] OR \"MLL-AF9\"[Title/Abstract] OR \"MLL-ENL\"[Title/Abstract] OR \"MLL-AF10\"[Title/Abstract] OR \"t 4 11\"[Title/Abstract] OR \"t 11 19\"[Title/Abstract])"
    },
    "gene_AND_subnuclear_AND_fusion": {
      "query": "(BRD3[tiab] OR RING3L[tiab] OR \"ORFX\"[tiab]) AND (\"nuclear speckle\"[tiab] OR \"nuclear speckles\"[tiab] OR nucleolus[tiab] OR nucleolar[tiab] OR \"PML body\"[tiab] OR \"PML bodies\"[tiab] OR \"nuclear matrix\"[tiab] OR \"Polycomb bodies\"[tiab] OR \"Polycomb body\"[tiab] OR subnuclear[tiab] OR \"nuclear bodies\"[tiab] OR \"nuclear body\"[tiab] OR chromocent*[tiab] OR heterochromat*[tiab]) AND (KMT2A[tiab] OR MLL[tiab] OR \"MLL1\"[tiab] OR \"MLL-AF4\"[tiab] OR \"MLL-AF9\"[tiab] OR \"MLL-ENL\"[tiab] OR \"MLL-AF10\"[tiab] OR \"t(4;11)\"[tiab] OR \"t(11;19)\"[tiab])",
      "count": 0,
      "querytranslation": "(\"BRD3\"[Title/Abstract] OR \"RING3L\"[Title/Abstract] OR \"ORFX\"[Title/Abstract]) AND (\"nuclear speckle\"[Title/Abstract] OR \"nuclear speckles\"[Title/Abstract] OR \"nucleolus\"[Title/Abstract] OR \"nucleolar\"[Title/Abstract] OR \"PML body\"[Title/Abstract] OR \"PML bodies\"[Title/Abstract] OR \"nuclear matrix\"[Title/Abstract] OR \"Polycomb bodies\"[Title/Abstract] OR \"Polycomb body\"[Title/Abstract] OR \"subnuclear\"[Title/Abstract] OR \"nuclear bodies\"[Title/Abstract] OR \"nuclear body\"[Title/Abstract] OR \"chromocent*\"[Title/Abstract] OR \"heterochromat*\"[Title/Abstract]) AND (\"KMT2A\"[Title/Abstract] OR \"MLL\"[Title/Abstract] OR \"MLL1\"[Title/Abstract] OR \"MLL-AF4\"[Title/Abstract] OR \"MLL-AF9\"[Title/Abstract] OR \"MLL-ENL\"[Title/Abstract] OR \"MLL-AF10\"[Title/Abstract] OR \"t 4 11\"[Title/Abstract] OR \"t 11 19\"[Title/Abstract])"
    },
    "brd3_only_vs_brd4": {
      "query": "BRD3[tiab] NOT BRD4[tiab]",
      "count": 98,
      "querytranslation": "\"BRD3\"[Title/Abstract] NOT \"BRD4\"[Title/Abstract]"
    },
    "brd3_loc": {
      "query": "BRD3[tiab] AND (localization[tiab] OR localisation[tiab] OR immunofluorescence[tiab] OR \"nuclear speckle\"[tiab] OR condensate*[tiab])",
      "count": 9,
      "querytranslation": "\"BRD3\"[Title/Abstract] AND (\"localization\"[Title/Abstract] OR \"localisation\"[Title/Abstract] OR \"immunofluorescence\"[Title/Abstract] OR \"nuclear speckle\"[Title/Abstract] OR \"condensate*\"[Title/Abstract])"
    }
  },
  "paralogue_conflation_measured": {
    "BRD3[tiab]": 654,
    "BRD3[tiab] NOT BRD4[tiab]": 98,
    "co_mention_rate": "85%",
    "reading": "a co-mention rate, not a measure of how many results are BRD4-only; papers naming both can still resolve them, as PMID 22897906 does"
  },
  "title_check": "gene_AND_subnuclear, 7 records: 28710461 (BET at pericentric heterochromatin, on topic), 22897906 (Brd/HP1 nucleosome mapping, on topic), 37269669 (BRD3/BRD4 imaging probes, partly on topic), and four off topic including a NUT midline carcinoma cytology series and an orbivirus protein paper - 2 to 3 of 7 on topic."
}
token hazard q95
{
  "condensate*": {
    "measured": "2026-09-22 by Q95",
    "query": "YAP1[tiab] AND (condensate*[tiab] OR \"phase separation\"[tiab])",
    "records": 15,
    "contamination": "the title-checked top ten include cigarette SMOKE CONDENSATE, waterpipe smoke condensate and an amiodarone/yeast toxicity study",
    "why_it_matters": "smoke condensate is a large toxicology literature; any bounded count built on condensate* is inflated in the FALSE-POSITIVE direction, and this project holds condensate claims on the SEC, PRC1/CBX2, NUP98-FG and ENL",
    "prescription": "use \"biomolecular condensate\" or \"phase separation\" as quoted phrases and title-check every hit, as the registry already prescribes for nucleolar and speckle*",
    "registry_row_proposed": true
  }
}
How this could bite us back
  • Resistance to BET inhibition emerges from leukaemia stem cells and confers cross-resistance to genetic knockdown of the BET proteins themselves. That is the worst shape of resistance available - it cannot be engineered around with better chemistry, because the cells no longer need the target.
  • Attributing any effect to BRD3 on current evidence is naming rather than measurement. If a BRD3 programme were started on this basis it would repeat, in a new gene family, the exact error this project retracted for CK2's catalytic subunits.
  • If the BET mechanism in this disease runs mainly through MYC and BCL2 rather than through the fusion's HOXA/MEIS1 programme, then it is a general leukaemia intervention and its therapeutic window is set accordingly - and BET inhibitors have a well-documented thrombocytopenia and gastrointestinal toxicity profile.
  • The in vivo survival benefit in the founding paper did not replicate independently. A graph that cites the 2011 result for in vivo efficacy without the 2017 replication beside it is overstating the preclinical case, and this entry exists partly to prevent that.
Still open
  1. Does BRD3 matter independently of BRD4 in KMT2A-rearranged leukaemia? Fifteen years after the founding paper, no experiment has asked, and a paralogue-selective degrader would answer it.
  2. Why do the two BET studies disagree about whether the mechanism reaches the fusion's own programme, when both used I-BET151? One found SEC and PAFc displaced; the other found HOXA, MEIS1 and CDK6 unaffected.
  3. Resistance is cross-resistant to genetic BET knockdown, which means it is not a drug-target problem at all. What is the resistant state actually doing, and does the Wnt/beta-catenin finding hold in lymphoid KMT2A-rearranged disease?
  4. Is there any BET result in KMT2A-rearranged ALL as opposed to AML? Almost all of this literature is MLL-AF9 and myeloid.
Every source behind this entry (9)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    MLL fusions, as part of the SEC and the PAF complex, are associated with the BET family of acetyl-lysine recognising chromatin adaptor proteins; the BET inhibitor I-BET151 has profound efficacy against human and murine MLL-fusion leukaemic cell lines, and its mode of action is at least in part inhibition of transcription at BCL2, C-MYC and CDK6 through the displacement of BRD3/4, PAFc and SEC components from chromatin, with survival benefit in mouse models of murine MLL-AF9 and human MLL-AF4 leukaemia.

    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. · Nature 478(7370):529-533 2011
    PMID 21964340DOI
    What this evidence cannot show

    BRD3 and BRD4 are reported together as 'BRD3/4' and were never separated; I-BET151 is a pan-BET inhibitor and cannot discriminate among BRD2, BRD3 and BRD4. Nothing in this paper attributes anything to BRD3 individually and this entry does not read it as doing so. Abstract read; PMCID PMC3679520, full text not fetched in this beat.

    Evidence that argues against this (1)
    • An independent replication under the Reproducibility Project: Cancer Biology confirmed the selective growth inhibition and BCL2 downregulation in MV4;11 but did NOT find a statistically significant survival difference in the disseminated xenograft MLL mouse model, where the original reported increased survival. The in vitro findings replicated and the in vivo survival endpoint did not. Shan X et al. 2017
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Human MLL-AF9 cells critically depend on BRD3/4 as well as on FLT3-ligand-induced pathways for survival; I-BET151 consistently and specifically downregulates MYC and BCL2, but the majority of I-BET151-responsive genes are not direct MLL-AF9 targets, and MLL-AF9-specific targets including the HOXA cluster, MEIS1 and CDK6 are not affected by I-BET151 treatment.

    BRD3/4 inhibition and FLT3-ligand deprivation target pathways that are essential for the survival of human MLL-AF9+ leukemic cells. · Carretta M, Brouwers-Vos AZ, Bosman M, Horton SJ, Martens JHA, Vellenga E, et al. · PLoS One 12(12):e0189102 2017
    PMID 29240787DOI
    What this evidence cannot show

    'BRD3/4' again - I-BET151 is pan-BET and the paper does not separate the paralogues, so despite BRD3 appearing in the title this is not a BRD3-specific result. MLL-AF9, myeloid, human cell lines and primary models. THP1 was used for the ChIP-seq. The negative finding on the HOXA cluster and MEIS1 is the most transferable part and is reported clearly, but it is a transcriptomic comparison rather than a designed test of fusion-directedness. Abstract read; PMCID PMC5730124, full text not fetched.

    Evidence that argues against this (1)
    • The founding paper reports the SEC and PAF complex components being displaced along with BRD3/4, which is a fusion-machinery-directed mechanism, so the two papers differ on how fusion-directed BET inhibition is. Both used the same compound, which suggests the difference lies in what was measured rather than in what happened. Dawson MA et al. 2011
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Primary mouse haematopoietic stem and progenitor cells immortalised with MLL-AF9 generate single-cell clones resistant to I-BET in vitro and in vivo; resistance confers cross-resistance to chemically distinct BET inhibitors such as JQ1 as well as resistance to genetic knockdown of BET proteins; resistance is not mediated by drug efflux or metabolism but emerges from leukaemia stem cells; chromatin-bound BRD4 is globally reduced in resistant cells while Myc expression remains unaltered; and resistance is in part a consequence of increased Wnt/beta-catenin signalling, with negative regulation of that pathway restoring I-BET sensitivity.

    BET inhibitor resistance emerges from leukaemia stem cells. · Fong CY, Gilan O, Lam EY, Rubin AF, Ftouni S, Tyler D, et al. · Nature 525(7570):538-542 2015
    PMID 26367796DOI
    What this evidence cannot show

    The genetic arm knocked down 'BET proteins' as a group, so this study too does not separate BRD3 from BRD4 - which is the point cited here as much as the resistance finding itself. MLL-AF9 mouse models plus human leukaemia cells; myeloid. Single-cell clones selected in vitro may not represent clinical resistance. Abstract read; PMCID PMC6069604, full text not fetched.

    Evidence that argues against this (1)
    • Single-cell clones selected in vitro and passaged in vivo are a strong resistance model but a weak model of clinical resistance, which arises in a treated patient under intermittent exposure. The result is also in MLL-AF9 mouse HSPCs, not the index infant B-lineage fusion. Shan X et al. 2017
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    An independent replication of the founding BET/MLL study found that I-BET151 produced selective growth inhibition in MV4;11 but not K-562 and a statistically significant decrease in BCL2 expression in MV4;11 but not K-562, similar to the original; but did NOT find a statistically significant difference in survival in a disseminated xenograft MLL mouse model where the original study reported increased survival, while finding lower median disease burden in all tissues analysed; differences in conditioning regimens and I-BET151 doses are noted as possible factors.

    PMID 28653617DOI
    What this evidence cannot show

    A Registered Report replication of SELECTED experiments, not of the whole paper, and the replication authors themselves name protocol differences (conditioning regimen, I-BET151 dose) that could explain the in vivo discrepancy. The disease-burden result went the original's way even though the survival endpoint did not, so this is a partial rather than a failed replication and should not be quoted as a refutation. Abstract read; PMCID PMC5487217, full text not fetched.

    Evidence that argues against this (1)
    • A replication that finds the direction of the original effect while differing on magnitude or on which comparisons reach significance is a qualified confirmation, and citing it as support for a BRD3-specific claim overreaches twice: the replication concerned I-BET151, a pan-BET compound, and selective growth inhibition in MV4;11 but not K-562 is a lineage-and-genotype contrast that does not isolate BRD3 from BRD2 or BRD4. Matthews GM et al. 2015
  5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Human BRD3 is UniProt Q15059, primary gene name BRD3, 726 amino acids, primary protein name 'Bromodomain-containing protein 3'; BRD2 is P25440, 801 amino acids.

    UniProtKB Q15059 (BRD3_HUMAN), Bromodomain-containing protein 3 · UniProt Consortium · UniProtKB, retrieved 2026-09-20 2026
    What this evidence cannot show

    Resolved per rule 62: gene_exact:BRD3 AND organism_id:9606 AND reviewed:true returned exactly one reviewed row and the PRIMARY gene name was read back as BRD3. No alias collision was observed with BRD2 or BRD4, unlike the HDAC7/HDAC9 pair in this same beat.

    Evidence that argues against this (1)
    • A reviewed database record fixes the canonical sequence on the date it was read and nothing more. It does not establish which isoform is expressed in KMT2A-rearranged blasts, and this beat's own findings turn on distinctions a canonical record hides - HDAC9's catalytically inactive MITR form, and the nomenclature collision in which HDAC9's protein name reads 'Histone deacetylase 7B'. Accession and length are asserted here so a swap fails the validator, not because they carry biological weight. Vega-Garcia N et al. 2018
  6. E_q95_brd3_loc1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    BRD3's two UniProt location values are both appropriately sourced, which makes this record the exception in this tranche: 'Nucleus' carries ECO:0000269 PubMed 25593309, a bromodomain-wide screen for chromatin recognition of transcription-associated DNA damage, and 'Chromosome' with the note 'Detected on chromatin' carries ECO:0000269 PubMed 18406326, the paper that placed Brd2 and Brd3 on hyperacetylated chromatin. PubMed 25593309 is the paper brief rule 91 lists as one of KMT2A's MISCITATIONS and tranche 5 found to be a CORRECT citation for KAT2A - so this is the third entry in this graph to carry it, and the second where it is appropriate.

    What this evidence cannot show

    A statement about the database record. Appropriateness here rests on the same reasoning tranche 5 used for KAT2A: a screen that measures where bromodomain proteins relocalise is a real localisation measurement FOR A BROMODOMAIN PROTEIN, and BRD3 is one. It is not a BRD3-specific study, and the abstract's named subject is ZMYND8.

    Evidence that argues against this (1)
    • THE PATTERN ACROSS THREE ENTRIES IS NOW STRONG ENOUGH TO STATE AS A RULE ABOUT THE PAPER RATHER THAN ABOUT THE GENES. PubMed 25593309 is a bromodomain-family relocalisation screen. Cited for KMT2A - which has no bromodomain - it is a miscitation, and brief rule 91 records it as such. Cited for KAT2A and for BRD3, both bromodomain proteins, it is exactly on point. So the citation's validity is predictable from one property of the annotated protein, and the audit is per gene not per paper in a way that has a MECHANISM: a family-wide screen is a valid source for every member of the family and for nobody else. That is a sharper statement than tranche 5 could make from two entries, and it also means this graph should expect the same paper to appear correctly on BRD2, BRD4 and other bromodomain entries. Gong F et al. 2015
  7. E_q95_brd3_loc2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    BRD3's compartment answer is OCCUPANCY and it is paralogue-resolved: Brd2 and Brd3 associate preferentially in vivo with hyperacetylated chromatin ALONG THE ENTIRE LENGTHS of transcribed genes, the associated chromatin is significantly enriched in H4K5ac, H4K12ac and H3K14ac and contains relatively little dimethylated H3K9, and both proteins allow RNA polymerase II to transcribe through nucleosomes in a defined system - while the intrinsic histone-chaperone activity and the cyclin D1 requirement in that paper are attributed specifically to BRD2 and not to BRD3.

    PMID 18406326DOI
    What this evidence cannot show

    This is where a protein ACTS along a gene, not which nuclear compartment it occupies, and for a BET protein those are the only two available answers - no BET protein in this tranche has a named subnuclear body. The paralogue attribution matters and is stated: the shared result is chromatin association, the BRD2-only results are the chaperone activity and cyclin D1.

    Evidence that argues against this (1)
    • AND THE MOST PROJECT-RELEVANT OCCUPANCY RESULT SEPARATES THE PARALOGUES IN THE OPPOSITE DIRECTION, WHICH IS AN ARGUMENT AGAINST THIS ENTRY'S OWN FRAMING OF BRD3 AS UNDIFFERENTIATED FROM BRD4. Proteogenomic mapping of mononucleosomes isolated on Brd2, Brd3 and Brd4 separately found that BRD3 AND BRD4 - not Brd2 - are MOST ENRICHED ON NUCLEOSOMES WITHIN HOX GENE CLUSTERS, the exact loci a KMT2A fusion drives. The functional test in the same paper is BRD4-only: expression of those genes falls on BRD4 shRNA depletion and no BRD3 depletion arm is reported. So the physical positioning at HOX is BRD3-resolved and the requirement is not - which is the cleanest statement available of what is and is not known about this paralogue at the loci this disease turns on. source
  8. E_q95_brd3_loc3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    BRD3 has one subnuclear-compartment result and the paralogues are NOT separated in it: under heat shock, HSF1 recruits GCN5, TIP60 and p300 to PERICENTRIC HETEROCHROMATIN producing targeted hyperacetylation there, and the redistribution of histone acetylation toward the pericentric region directs recruitment of the BET proteins BRD2, BRD3 AND BRD4, which are required for satellite III transcription by RNA polymerase II.

    Bromodomain factors of BET family are new essential actors of pericentric heterochromatin transcriptional activation in response to heat shock. · Col E, Hoghoughi N, Dufour S, Penin J, Koskas S, Faure V, Ouzounova M, Hernandez-Vargash H, Reynoird N, Daujat S, Folco E, Vigneron M, Schneider R, Verdel A, Khochbin S, Herceg Z, Caron C, Vourc'h C · Sci Rep 2017
    PMID 28710461DOI
    What this evidence cannot show

    All three BET paralogues are recruited and required in this system, so nothing here is BRD3-specific; it is a stress-induced compartment in non-haematopoietic human cells, and satellite III transcription under heat shock is a specialised phenomenon rather than a general address. It is nonetheless the only named subnuclear structure BRD3 has ever been placed in.

    Evidence that argues against this (1)
    • THE INTERESTING TENSION IS WITH THE MARK PROFILE IN THIS ENTRY'S OTHER OCCUPANCY PAPER, AND IT IS NOT A CONTRADICTION BUT A CONDITION. Brd3-bound chromatin in the unstressed state contains RELATIVELY LITTLE dimethylated H3K9 - it is euchromatic by mark composition - and pericentric heterochromatin is the H3K9me-dense compartment BET proteins would be expected to avoid. The reconciliation the heat-shock paper itself supplies is that the compartment has to be ACETYLATED FIRST by GCN5, TIP60 and p300 before BET proteins arrive, so BRD3's address is not a place but an acetylation state, and a normally excluded compartment becomes accessible when it acquires the mark. That is a better model of what a bromodomain protein's localisation means than any compartment name, and it predicts that anything which redistributes histone acetylation - including a KMT2A fusion's own acetyltransferase partners - redistributes BRD3. LeRoy G et al. 2008
  9. E_q95_brd3_loc4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    Bounded 2026-09-22, with a measurement of the paralogue-conflation problem rather than an assertion of it: BRD3[tiab] returns 654 records and BRD3[tiab] NOT BRD4[tiab] returns 98, so 85% of the BRD3 literature also names BRD4. The gene clause over BRD3, RING3L and ORFX intersects the nine-term KMT2A/MLL fusion clause at 6 records; with a localisation clause the gene clause returns 64 and with the fusion clause added 0; with a subnuclear clause it returns 7 and with the fusion clause added 0; and a BRD3-specific localisation probe (localization, localisation, immunofluorescence, nuclear speckle, condensate*) returns 9.

    What this evidence cannot show

    Titles and abstracts only; supporting figures not searched (rule 80). The 85% figure is a co-mention rate and not a measure of how many results are BRD4-only - a paper naming both may still resolve them, as PubMed 22897906 does. What it bounds is the impossibility of reading the BRD3 literature without deciding, paper by paper, which paralogue each result belongs to.

    Evidence that argues against this (1)
    • A NEW TOKEN CONTAMINATION FOUND WHILE RUNNING THIS, AND IT RUNS IN THE FALSE-POSITIVE DIRECTION, WHICH IS THE DANGEROUS ONE. 'condensate*' is not a safe search token: in this tranche, YAP1[tiab] AND (condensate* OR 'phase separation')[tiab] returned 15 records of which the title-checked top ten included cigarette SMOKE CONDENSATE, waterpipe smoke condensate and an amiodarone-toxicity yeast study - SMOKE CONDENSATE is a large toxicology literature and the wildcard reaches all of it. This matters beyond retrieval: this project holds condensate claims on the SEC, PRC1/CBX2, NUP98-FG and ENL, and any bounded count built on 'condensate*' is inflated by a field with no relation to phase separation. Use 'biomolecular condensate' or 'phase separation' as quoted phrases and title-check every hit, exactly as the registry already prescribes for 'nucleolar' and 'speckle*'. source
Connected to