Oncofusion

start / graph / gene-hdac7

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

HDAC7 - the eraser the index fusion silences, where the proposed therapy is to INDUCE the HDAC, and two 2024-2025 papers point in opposite directions

also called HDAC7A · HD7 · HD7a · histone deacetylase 7

HDAC7 is an eraser enzyme that a young B cell needs in order to stay a B cell, and the t(4;11) fusion switches it off. One group therefore proposes putting HDAC7 back, which is the exact opposite of what an HDAC-blocking drug does, while a second group reports that removing HDAC7 kills the same cells.

The same thing, for a scientist

HDAC7 (Q8WUI4, 952 aa) is a class IIa deacetylase and a B-lymphocyte lineage factor. It is the only HDAC in this beat whose biology is written **in the index fusion itself**. de Barrios et al. report that infants with t(4;11) B-ALL show **poor therapeutic response associated with low expression of HDAC7**, that **HDAC7 is epigenetically silenced by EZH2 and by the KMT2A::AFF1 fusion protein**, that the menin inhibitor **MI-538 restores HDAC7 expression** with the effect enhanced by the class I HDAC inhibitor chidamide, that the combination drives leukaemic pro-B cells toward a more differentiated state and impairs proliferation **in an HDAC7-dependent manner**, and that adding MI-538 plus chidamide to standard chemotherapy reduces PDX engraftment and delays relapse [E1]. **So in this paper the therapeutic move is HDAC7 induction, not HDAC inhibition.** Directly against that, Verbeek et al. report that **shRNA knock-down of the class IIA isoforms HDAC4, HDAC5 and HDAC7 induces apoptosis and cell-cycle arrest in KMT2A-rearranged ALL cells** [E2]. Two groups, the same disease, opposite signs on the same gene. This entry does not adjudicate; it records the conflict as the most interesting open question in the layer and sets out in `hyp-the-class-iia-hdac-direction-conflict-is-scaffold-versus-catalysis` what would resolve it. **A separate and consequential finding: `gene_exact:HDAC7` in UniProt returns HDAC9 as its FIRST reviewed row**, because HDAC9 carries HDAC7 and HDAC7B as synonyms and its own protein name reads 'Histone deacetylase 7B'. Accession Q8WUI4 was kept only after the primary name was read back (rule 62).

Go deeper
The index fusion silences it, which is rare and load-bearing

Almost every gene in this beat reaches the disease through a myeloid mouse model. HDAC7 does not. Its evidence is a large infant pro-B-ALL cohort analysed by multiomics, with t(4;11) - KMT2A::AFF1, the index fusion of this graph - as the defining lesion 1.

The chain reported is:

  1. KMT2A::AFF1 and EZH2 epigenetically silence HDAC7 1.
  2. Low HDAC7 is associated with poor therapeutic response in these infants 1.
  3. The menin inhibitor MI-538 restores HDAC7 expression, enhanced by chidamide 1.
  4. Restoration drives differentiation and impairs proliferation in an HDAC7-dependent manner - the rescue experiment that makes the mechanism a mechanism rather than a correlation 1.
  5. The combination raises glucocorticoid sensitivity of PDX cells ex vivo by repressing RUNX2, and with standard chemotherapy reduces engraftment and delays relapse in vivo 1.

Step 3 is the part that should change how this graph reads menin inhibitors. Revumenib, ziftomenib and their siblings are described here almost entirely through HOXA/MEIS1 eviction. This paper supplies a second, downstream, lineage-differentiation consequence of menin inhibition in the lymphoid index disease - de-repression of a silenced B-cell factor - and it is the mechanism by which the menin inhibitor and an HDAC inhibitor become a rational pair rather than two epigenetic drugs put in a tube together.

The contradiction, stated without resolving it
de Barrios et al. 2025 [E1]Verbeek et al. 2024 [E2]
diseaset(4;11) infant B-ALLKMT2A-rearranged ALL
HDAC7 status reportedlow, silenced by KMT2A::AFF1 and EZH2not framed as low
perturbationrestore HDAC7 (MI-538 +/- chidamide)knock down HDAC4/5/7 (shRNA)
resultdifferentiation, reduced proliferation, HDAC7-dependentapoptosis and cell-cycle arrest
in vivoreduced PDX engraftment, delayed relapse, with chemoclass IIA inhibitors failed in vivo (see below)

Both are peer-reviewed, both are in the right disease, and they are not obviously reconcilable. Candidate reconciliations, none tested:

  • Expression level is not activity. HDAC7 is a class IIa enzyme with negligible intrinsic deacetylase activity on acetyl-lysine; class IIa enzymes work largely as recruiters of the NCoR/SMRT-HDAC3 complex and as MEF2-binding repressors. Restoring the protein and removing the protein can then have different consequences from removing its catalytic function, and shRNA removes the scaffold while an inhibitor does not.
  • The cohorts differ. One is explicitly infant t(4;11); the other is KMT2A-rearranged ALL broadly, which includes older children and other partners.
  • Acute versus chronic. shRNA knock-down over days measures maintenance; the induction arm is a differentiation phenotype read alongside chemotherapy.
  • HDAC7 versus its class IIa siblings. Verbeek et al. knocked down HDAC4, HDAC5 and HDAC7 and report the class as a unit; the killing may not be attributable to HDAC7.

That last point is the one this beat keeps meeting. See ana-no-hdac-inhibitor-result-in-this-disease-attributes-to-an-enzyme.

A live instance of rule 62, in this gene

Run 2026-09-20 against the UniProt REST API with gene_primary requested:

gene_exact:HDAC7 AND organism_id:9606 AND reviewed:true   ->  2 reviewed rows
  Q9UKV0  HDAC9_HUMAN  primary=HDAC9  1011 aa   <- returned FIRST
  Q8WUI4  HDAC7_HUMAN  primary=HDAC7   952 aa   <- the one wanted

HDAC9's synonym list is HDAC7 HDAC7B HDRP KIAA0744 MITR and its protein name literally reads 'Histone deacetylase 7B (HD7) (HD7b)' 3.

This is worse than the CSNK2A1/CSNK2A2 case already logged as rule 63, on two counts. The two proteins are paralogues in the same HDAC subclass (class IIa), of similar length (1011 vs 952 aa), with overlapping MEF2-binding biology - so a swapped record reads as correct. And the wrong record's own protein name contains the symbol you searched, so even a human reading it back can talk themselves into it. A script taking row 2 of the TSV writes HDAC9's accession into an HDAC7 entry with no error anywhere.

Both accessions in this beat were kept only after the PRIMARY gene name was read back. The pairing matters here more than usual because HDAC7 and HDAC9 are the two class IIa enzymes with real findings in this disease, and they point in opposite directions - HDAC7 low and to be induced 1, HDAC9 high and to be repressed (gene-hdac9). Swapping them inverts the therapeutic conclusion.

Facts and figures
uniprotQ8WUI4
length AA952
gene symbolHDAC7
accession resolutiongene_exact:HDAC7 AND organism_id:9606 AND reviewed:true returned TWO reviewed rows with HDAC9 (Q9UKV0) FIRST. Q8WUI4 was kept only after its PRIMARY gene name was read back as HDAC7 (rule 62 applied and fired live).
paralogue alias trapHDAC9 (Q9UKV0) carries HDAC7 and HDAC7B as gene synonyms and 'Histone deacetylase 7B (HD7) (HD7b)' inside its own protein name. Same subclass (IIa), similar length, overlapping MEF2 biology - a swap does not look like an error.
classclass IIa histone deacetylase; minimal intrinsic catalytic activity, acts largely by recruiting the NCoR/SMRT-HDAC3 complex and by binding MEF2
status in index diseaseLOW - epigenetically silenced by EZH2 and the KMT2A::AFF1 fusion protein in t(4;11) infant B-ALL
proposed therapeutic directionINDUCTION, not inhibition
inducing agentmenin inhibitor MI-538, enhanced by class I HDAC inhibitor chidamide
downstream of inductiondifferentiation of leukaemic pro-B cells; glucocorticoid sensitisation via RUNX2 repression
opposing reportPMID 39362994 - shRNA knock-down of HDAC4/5/7 induces apoptosis in KMT2A-rearranged ALL cells
conflict unresolvedtrue
in vivo class iia inhibitor resultNEGATIVE - LMK-235, MC1568 and TMP195 all failed at maximum achievable dose in a KMT2A-rearranged ALL xenograft; LMK-235 plus venetoclax did not prolong survival
lineagelymphoid - this is one of the few entries in the acetylation layer whose primary evidence is B-lineage and infant
index fusion evidencetrue
subnuclear localizationPML nuclear bodies, signal-dependently and reversibly (TNF-alpha recruits, PML knockdown releases); recruitment to a PML body is the mechanism by which HDAC7 leaves MEF2 target promoters. Separately chromatin-associated with H3.3 and HIRA.
shown for fusionNO - AND THIS IS THE LOAD-BEARING ABSENCE OF THIS TRANCHE. HDAC7[tiab] AND (KMT2A OR MLL OR "t(4;11)")[tiab] returns 4 records (2026-09-21), three of them squarely in the index disease, and NOT ONE measures where HDAC7 is. HDAC7 is epigenetically silenced by EZH2 and by the KMT2A::AFF1 fusion protein itself in infant t(4;11) pro-B ALL, and the proposed therapy is to INDUCE it with a menin inhibitor plus chidamide. A class IIa HDAC's activity is set by whether it is in the nucleus on MEF2 promoters, sequestered in a PML body, or exported to the cytoplasm - so 'restore HDAC7 expression' is a proposal whose outcome depends entirely on a compartment that has never been measured in a KMT2A-rearranged cell. Zero in titles and abstracts; supporting figures not searched; the two index-disease papers (PMID 40722046, PMID 39362994) were read to abstract level and neither reports fractionation or imaging of HDAC7.
identifier trapAlready registered: gene_exact:HDAC7 returns HDAC9 (Q9UKV0, 1011 aa) FIRST because HDAC9's protein name reads 'Histone deacetylase 7B'; only gene_primary discriminates and the length gap is 59 aa. Re-verified live 2026-09-21: Q8WUI4, gene_primary HDAC7, 952 aa. The registry's stated stake - that a swap INVERTS a therapeutic recommendation, HDAC7 to be induced in t(4;11) and HDAC9 to be repressed - is now compounded by a genuine, non-swap disagreement about HDAC7 itself (see counter_evidence on E_q65_loc3).
essentiality
{
  "depmap_common_essential": null,
  "note": "DepMap not queried in this beat."
}
localization
{
  "subcellular": "BOTH nucleus and cytoplasm, and for a class IIa HDAC the partition IS the regulation. HDAC7 shuttles; nuclear-to-cytoplasmic transport of HDAC7 is a pharmacologically addressable dependency in HDAC7-high small cell lung cancer, where HDAC7 drives XPO1 expression through beta-catenin/TCF/LEF1 and the XPO1 inhibitor selinexor is effective. UniProt Q8WUI4 (HDAC7_HUMAN, Histone deacetylase 7, 952 aa) was read back only to confirm the identifier - and this is the accession the registry warns about, because gene_exact:HDAC7 returns HDAC9 FIRST with a protein name reading 'Histone deacetylase 7B'.",
  "subnuclear": "PML NUCLEAR BODIES, signal-dependently. Endogenous HDAC7 and PML interact and partially colocalise in PML nuclear bodies in human umbilical vein endothelial cells; TNF-alpha treatment RECRUITS HDAC7 to PML NBs and enhances HDAC7-PML association, which dissociates HDAC7 from MEF2 transcription factors and from the promoters of MEF2 target genes such as MMP-10. PML knockdown reverses this. So HDAC7's subnuclear address is a switch: sequestration in a body is how it is taken off chromatin. A second, distinct subnuclear readout is chromatin-level - HDAC7 binds histone H3.3 and interacts with H3.3 and HIRA on chromatin, and HDAC7 inhibition drives H3.3 onto DAXX and spreads H3K9me3 heterochromatin foci.",
  "how_shown": "endogenous co-immunoprecipitation and immunofluorescence colocalisation with PML, with TNF-alpha stimulation, PML knockdown and PML over-expression, plus ChIP at the MMP-10 promoter, in HUVEC (PMID 18463162); siRNA plus imaging of H3K9me3/DAPI heterochromatin foci in glioma cells (PMID 40967435); RNA-seq, cell assays and xenografts with selinexor in SCLC (PMID 39887933)",
  "systems": "human umbilical vein endothelial cells; glioblastoma, glioma stem cells and paediatric diffuse pontine glioma; small cell lung cancer - NONE of them KMT2A-rearranged",
  "node": "localization-pml-nuclear-bodies-pmlrara"
}
localization counted absence
{
  "query": "HDAC7[tiab] AND (KMT2A[tiab] OR MLL[tiab] OR \"t(4;11)\"[tiab])",
  "records": 4,
  "records_with_hdac7_localization": 0,
  "operand_controls": {
    "HDAC7[tiab] AND (localiz* OR nuclear OR cytoplasm* OR shuttl* OR \"nuclear bodies\")[tiab]": 116
  },
  "date": "2026-09-21",
  "tool": "PubMed E-utilities esearch",
  "bound": "zero in titles and abstracts; supporting figures not searched"
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -0.0881,
  "chronos_mean_other_leukaemia": -0.241,
  "difference": 0.1529,
  "pan_essentiality_all_lines": -0.166,
  "selectivity_rank": 17469,
  "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."
}
cell type
{
  "evidence_class_key": "concept-expressed-required-maintains-three-evidence-classes",
  "identifier_note": "gene_exact:HDAC7 returns HDAC9 FIRST and both protein names contain 'HD7' (brief rule 65). This block is HDAC7, Q8WUI4, 952 aa; HDAC9 is Q9UKV0, 1,011 aa. The two point in OPPOSITE therapeutic directions in this disease, so a swap here inverts a recommendation rather than merely naming the wrong protein.",
  "expressed_in": {
    "class": 1,
    "summary": "LYMPHOID-SPECIFIC within the haematopoietic system. HDAC7 is a class IIa HDAC with a lymphoid-restricted expression pattern in blood, and is HIGHLY EXPRESSED IN PRE-B CELLS, where it is dramatically down-regulated during lineage conversion to macrophages [E_q69_ct1], [E_q69_ct2].",
    "pmids": [
      "23696748",
      "27810920"
    ],
    "system_named_in_the_claim": "murine pre-B cells and a murine conditional knockout"
  },
  "required_in": {
    "class": 2,
    "claim": "In MICE, conditional deletion of Hdac7 DRAMATICALLY BLOCKS EARLY B CELL DEVELOPMENT, causes severe lymphopenia in peripheral organs, and produces PRO-B CELL LINEAGE PROMISCUITY; HDAC7 represses myeloid and T-lymphocyte genes in B-cell progenitors through interaction with MEF2C, is recruited to promoters and enhancers of target genes, and its absence increases enrichment of active histone marks [E_q69_ct2].",
    "pmids": [
      "27810920",
      "23696748"
    ],
    "system_named_in_the_claim": "murine conditional knockout in normal B lymphopoiesis - not leukaemia and not human",
    "the_cell_type_is_the_cell_of_origin": "pro-B and pre-B are the compartments this graph models as the origin of infant t(4;11) B-ALL. HDAC7 is one of very few genes in this tranche whose NORMAL cell-type requirement sits in exactly the cell the disease starts in."
  },
  "not_required_in": {
    "claim": "NO KMT2A-selective requirement, and this is the second-deepest anti-selective placement in the tranche. Chronos -0.0881 in the KMT2A-rearranged arm against -0.2410 in other leukaemias, difference +0.1529, rank 17,469 of 17,728. Pan-essentiality -0.1660 - low, but with enough range that the placement is interpretable rather than floor-compressed.",
    "how_this_bears_on_the_live_contradiction_in_this_entry": "the entry records two groups pointing in opposite directions - de Barrios et al. propose INDUCING HDAC7, Verbeek et al. report that knocking down class IIa HDACs including HDAC7 kills KMT2A-rearranged ALL cells. The DepMap scan is a third instrument and it sides with NEITHER: KMT2A-rearranged lines are LESS dependent on HDAC7 than other leukaemia lines are. A CRISPR proliferation screen cannot see a differentiation phenotype, which is what the induction proposal is about, so this does not settle the conflict - but it does mean the knock-down-kills claim has no support from the largest available loss-of-function dataset, and that is the honest thing to say about it."
  },
  "nodes": [
    "cell-type-pro-b",
    "cell-type-pre-b"
  ],
  "context": "murine normal B lymphopoiesis for the cell-type answer; the human infant t(4;11) material in this entry is separate and is cited there"
}
cell state
{
  "evidence_class_key": "concept-expressed-required-maintains-three-evidence-classes",
  "maintains_state": {
    "class": 3,
    "states": [
      "cell-state-pro-b-arrest"
    ],
    "claim": "HDAC7 MAINTAINS B-CELL IDENTITY as an active repressive state. It silences lineage-inappropriate myeloid and T-cell genes in B-cell progenitors through MEF2 binding sites, and its re-expression during pre-B-to-macrophage transdifferentiation BLOCKS induction of the macrophage programme - immune, inflammatory and defence response genes, cytokine production, phagocytosis - and suppresses the acquired macrophage functions themselves [E_q69_ct1]. Loss of HDAC7 in vivo gives pro-B cells a promiscuous, multi-lineage transcriptional state [E_q69_ct2].",
    "pmids": [
      "23696748",
      "27810920"
    ],
    "system_named_in_the_claim": "murine pre-B cells in an induced transdifferentiation system, and a murine conditional knockout"
  },
  "the_state_is_REPRESSIVE_MAINTENANCE_not_a_block": "this needs saying because the graph's dominant cell state is cell-state-differentiation-block. HDAC7 does not hold cells at an immature stage; it prevents them from expressing the WRONG lineage while they mature along the right one. Losing HDAC7 does not deepen the block - it dissolves lineage fidelity. In t(4;11), where HDAC7 is reported LOW, the predicted consequence is lineage infidelity rather than a deeper arrest, and the lineage-switch phenomenon in KMT2A-rearranged leukaemia is exactly that phenotype. No published work makes this connection and it is offered as a lead.",
  "bounded_query": {
    "date": "2026-09-21",
    "tool": "NCBI esearch db=pubmed",
    "query": "HDAC7[ti] AND (transdifferentiation OR \"pre-B\" OR \"B cell identity\" OR myeloid)[tiab]",
    "count": 9,
    "on_topic_after_title_check": 3,
    "the_three": "PMID 23696748 (pre-B to macrophage), PMID 27810920 (conditional knockout, B-cell identity), PMID 33262526 (infant t(4;11) proB-ALL).",
    "bound": "titles and abstracts only; supporting figures not searched (brief rule 80)"
  },
  "an_unreadable_paper_on_this_beat": {
    "pmid": "33262526",
    "what_the_title_asserts": "that HDAC7 is a major contributor in the pathogenesis of infant t(4;11) proB acute lymphoblastic leukaemia.",
    "routes_tried_2026_09_21": "efetch db=pubmed returns an EMPTY abstract; Europe PMC core returns pmcid null, isOpenAccess N and 0 characters of abstract, so routes 2-4 are structurally unavailable; route 5, resolving the DOI to the publisher with a browser User-Agent, returns a 305-character JavaScript challenge stub rather than the article.",
    "status": "known but unreadable. The identifier and the title are recorded; nothing about its contents is written here, per the standing rule.",
    "confidence": "weak"
  },
  "confidence": "strong",
  "caveats": "Both cell-type results are murine. The transdifferentiation result is an INDUCED conversion system (C/EBPalpha-driven pre-B to macrophage) rather than a physiological transition, and the HDAC7 arm of it is a RE-EXPRESSION experiment - a gain-of-function in a system where the gene had been switched off, which is the correct design for that question but is not evidence about endogenous dose. The conditional knockout is the stronger result and is in vivo. Neither study is human, neither involves a KMT2A fusion, and the DepMap arm places HDAC7 in the anti-selective half.",
  "counter_evidence": {
    "note": "The in vivo requirement and the in vitro repression come from the same laboratory (Parra and colleagues on all three HDAC7 papers cited here, including the unreadable infant t(4;11) one). There is no independent replication of the B-cell-identity result by another group in the records retrieved, and the graph should not treat convergence between three papers from one group as three lines of evidence. The one independent instrument available - this project's genome-wide Chronos scan - places HDAC7 17,469th of 17,728 for KMT2A selectivity and does not corroborate a requirement in these cells at all.",
    "citation": {
      "title": "DepMap KMT2A-selective dependency scan (this project, genome-wide, 11 KMT2A-rearranged vs 43 other leukaemia lines)",
      "url": "analyses/depmap-kmt2a-selective-scan-full/results.json",
      "type": "analysis"
    }
  }
}
How this could bite us back
  • The two published directions on this gene are opposite. Any intervention built on either one is built on a literature that contains its own refutation, and a trial designed from the wrong half would be actively harmful rather than merely ineffective.
  • Class IIa HDAC inhibitors have already failed in vivo in this exact disease - three compounds, maximum achievable dose, no anti-leukaemic effect, including in combination with venetoclax. That is a real negative result and it should discipline any enthusiasm about class IIa as a target here.
  • The induction strategy requires a menin inhibitor plus an HDAC inhibitor plus standard chemotherapy, in infants. That is three epigenetically and myelosuppressively active layers stacked on a population in which infection already causes most treatment-related death.
  • HDAC7 is a B-lymphocyte lineage factor. Manipulating it in either direction in an infant acts on a B-cell compartment that is still being built, and the long-term immunological consequences have no data at all.
  • Because HDAC9 carries HDAC7 as an alias and the two point in opposite therapeutic directions, a database lookup error in either direction produces a confident, clean, inverted recommendation.
  • compartment-blind induction - The t(4;11) proposal is to RESTORE HDAC7 with a menin inhibitor plus chidamide. HDAC7 is a class IIa deacetylase whose activity is set by where it is - on MEF2 target promoters, sequestered in a PML nuclear body, or exported to the cytoplasm - and none of those has been measured in a KMT2A-rearranged cell. Induced protein that is exported or PML-sequestered is not the reagent the rationale assumes. Measure the nuclear/cytoplasmic partition and PML colocalisation of HDAC7 in t(4;11) blasts before and after MI-538 before scaling this combination.
  • an opposite published recommendation in the same disease - shRNA knock-down of HDAC4/5/7 induces apoptosis in KMT2A-rearranged ALL and LMK-235 eradicates KMT2A-r lines and primary samples in vitro, although it failed at the maximum achievable dose in vivo. Inducing and inhibiting HDAC7 cannot both be right in the same disease; a trial designed from either paper alone is exposed to the other.
  • The cell type that needs HDAC7 is the cell type being treated. Conditional Hdac7 deletion in mice blocks early B-cell development and causes severe peripheral lymphopenia (PMID 27810920). A class IIa HDAC inhibitor - the Verbeek direction - is therefore predicted to produce B-cell aplasia on target, in infants. The opposite proposal, inducing HDAC7 (de Barrios direction), does not carry that liability but carries the mirror one: forcing a lineage-fidelity repressor in a cell whose lineage is already ambiguous has no published safety precedent.
  • Neither therapeutic direction has support from the largest independent loss-of-function dataset available. In this project's genome-wide Chronos scan HDAC7 is rank 17,469 of 17,728 for KMT2A selectivity - the KMT2A-rearranged arm is LESS depleted (-0.0881) than the comparator leukaemias (-0.2410), with pan-essentiality -0.1660 and no floor compression. A 21-day proliferation screen cannot see a differentiation phenotype and so cannot refute the induction proposal, but it is squarely against the knock-down-kills claim.
Still open
  1. Which direction is right? Restoring HDAC7 and removing HDAC7 are reported as therapeutic in the same disease by two groups within a year of each other, and nothing in either abstract reconciles them.
  2. Is the class IIA killing in PMID 39362994 attributable to HDAC7 at all, or is it carried by HDAC4 or HDAC5? The knockdowns were reported as a class.
  3. Does a clinical menin inhibitor - revumenib, ziftomenib, bleximenib, enzomenib - restore HDAC7 the way the tool compound MI-538 does? If it does, HDAC7 becomes a pharmacodynamic biomarker for menin inhibition in the lymphoid disease, which this graph currently lacks.
  4. Does the scaffolding function of HDAC7 (NCoR/SMRT-HDAC3 recruitment, MEF2 binding) rather than its catalysis explain the direction conflict? An HDAC7 catalytic-dead knock-in would separate them.
  5. Is low HDAC7 prognostic independently of KMT2A rearrangement and of age, the way HDAC9 overexpression was reported to be?
  6. Where is HDAC7 in a t(4;11) blast before and after MI-538 - nuclear on MEF2 promoters, in PML bodies, or cytoplasmic? This is the single measurement most likely to reconcile the two opposite therapeutic proposals in this disease, and it requires one immunofluorescence panel on cells the field already has.
Every source behind this entry (9)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Infants with t(4;11) B-ALL display poor therapeutic response associated with low expression of HDAC7; HDAC7 is epigenetically silenced by EZH2 and the KMT2A::AFF1 fusion protein; the menin inhibitor MI-538 restores HDAC7 expression and the effect is enhanced by the class I HDAC inhibitor chidamide; this treatment drives leukaemic pro-B cells toward a more differentiated state and impairs aberrant proliferation in an HDAC7-dependent manner; it increases glucocorticoid sensitivity of PDX cells ex vivo by repressing RUNX2; and combining MI-538 and chidamide with standard chemotherapy reduces PDX engraftment in vivo and delays relapse.

    HDAC7 induction combined with standard-of-care chemotherapy provides a therapeutic advantage in t(4;11) infant B-cell acute lymphoblastic leukemia. · de Barrios O, Ocon-Gabarro I, Gusi-Vives M, Collazo O, Meler A, Romecin PA, et al. · Biomarker Research 13(1):99 2025
    PMID 40722046DOI
    What this evidence cannot show

    The cohort size is described only as 'a large infant pro-B-ALL cohort' in the abstract; the actual n was not read, because only the abstract was retrieved in this beat (PMCID PMC12305908, full text not fetched). The therapeutic effect is explicitly stated to be only PARTIALLY mediated by HDAC7, so the combination's benefit is not wholly attributable to this gene. MI-538 is a tool menin inhibitor, not a clinical agent, so the result does not transfer automatically to revumenib or ziftomenib. Chidamide is a class I HDAC inhibitor and therefore multi-isoform. The association between low HDAC7 and poor response is correlative in the cohort even though the rescue experiments are causal in cells.

    Evidence that argues against this (1)
    • A second group reports the opposite sign on the same gene in the same disease: shRNA knock-down of the class IIA isoforms HDAC4, HDAC5 and HDAC7 induces apoptosis and cell-cycle arrest in KMT2A-rearranged ALL cells, and the HDAC4/5-selective inhibitor LMK-235 eradicates KMT2A-rearranged ALL cell lines and primary patient samples in vitro. If class IIA HDAC loss kills these cells, restoring HDAC7 as therapy requires an explanation this graph does not currently have. Verbeek TCAI et al. 2024
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    shRNA-mediated knock-down of the class IIA HDAC isoforms HDAC4, HDAC5 and HDAC7 results in apoptosis induction and cell cycle arrest in KMT2A-rearranged ALL cells, and the HDAC4/5-selective small molecule LMK-235 effectively eradicates KMT2A-rearranged ALL cell lines and primary patient samples in vitro; however, in a KMT2A-rearranged ALL xenograft mouse model the maximum achievable dose of LMK-235 was insufficient to induce anti-leukaemic effects in vivo, with similar results for the class IIA inhibitors MC1568 and TMP195, and LMK-235 plus venetoclax did not prolong survival in treated mice.

    Selective inhibition of HDAC class IIA as therapeutic intervention for KMT2A-rearranged acute lymphoblastic leukemia. · Verbeek TCAI, Vrenken KS, Arentsen-Peters STCJM, Castro PG, van de Ven M, van Tellingen O, et al. · Communications Biology 7(1):1257 2024
    PMID 39362994DOI
    What this evidence cannot show

    HDAC4, HDAC5 and HDAC7 were knocked down and reported together as a class, so the killing is not attributed to HDAC7 individually by this paper. The in vivo arm is a clean NEGATIVE: three different class IIA-selective compounds failed at the maximum achievable dose, so the in vitro genetic result did not translate, and the authors attribute this to compound stability and efficiency rather than to the biology. Abstract read; PMCID PMC11450098, full text not fetched in this beat. No infant-only cohort is specified.

    Evidence that argues against this (1)
    • The directly opposed report in the same disease holds that HDAC7 is LOW in t(4;11) infant B-ALL, silenced by the KMT2A::AFF1 fusion and EZH2, and that restoring it is therapeutic. An entry that used only this paper would recommend inhibiting the gene the other paper recommends inducing. de Barrios O et al. 2025
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    A gene_exact UniProt query for HDAC7 in reviewed human records returns two rows, and the first is HDAC9 (Q9UKV0, 1011 aa), whose gene synonyms include HDAC7 and HDAC7B and whose protein name includes 'Histone deacetylase 7B (HD7) (HD7b)'; the intended record is Q8WUI4 (HDAC7_HUMAN, 952 aa).

    What this evidence cannot show

    Row order in a UniProt TSV is not contractual and may change between releases; the finding to carry forward is that the query returns more than one reviewed row and that one of them is the paralogue, not the specific ordering observed on this date. The same query for HDAC1, HDAC2, HDAC3, HDAC5, HDAC9, KAT5, KAT2A, KAT2B, BRD2 and BRD3 returned exactly one reviewed row each on the same day.

    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
  4. E_q65_loc1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Endogenous HDAC7 and PML interact and partially colocalise in PML nuclear bodies in human umbilical vein endothelial cells; TNF-alpha recruits HDAC7 to PML NBs, which dissociates HDAC7 from MEF2 transcription factors and from MEF2 target promoters such as MMP-10, and PML knockdown reverses both the recruitment and the derepression.

    PMID 18463162DOI
    What this evidence cannot show

    Endothelial cells, not haematopoietic and not KMT2A-rearranged (brief rule 73). 'Partially colocalise' is the authors' own wording - this is a signal-inducible pool, not the bulk of cellular HDAC7. Abstract only.

    Evidence that argues against this (1)
    • PML-body sequestration is not the only way HDAC7 leaves the nucleus: in HDAC7-high small cell lung cancer, HDAC7 drives XPO1 expression and nuclear-to-cytoplasmic transport is the druggable step, so at least two independent routes remove HDAC7 from chromatin and no experiment ranks them in any one cell. Qin T et al. 2025
  5. E_q65_loc2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    HDAC7 binds histone H3.3 and interacts with H3.3 and HIRA on chromatin; downregulating HDAC7 inhibits the H3.3-HIRA interaction, increases H3.3 association with DAXX and H3K9me3, and deposits H3.3 on H3K9me3-positive heterochromatin nuclear foci, triggering heterochromatin spreading and DNA replication defects.

    PMID 40967435DOI
    What this evidence cannot show

    Glioblastoma, glioma stem cells and paediatric diffuse pontine glioma - a solid-tumour CNS system. The spatial readout is where H3.3 goes, not where HDAC7 goes. It is recorded here because it couples HDAC7 to HIRA, another entry in this tranche, on the same H3.3 axis. Abstract only.

    Evidence that argues against this (1)
    • The HIRA entry in this graph records the apparently opposite movement in a leukaemia cell: HIRA knockdown in K562 INCREASES H3K9me3 and HP1alpha and compacts chromatin. Two perturbations on one axis producing the same direction of heterochromatin change means the axis is not a simple linear chain and neither result licenses a prediction in KMT2A-rearranged cells. Shirude MB et al. 2026
  6. E_q65_loc3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In infant t(4;11) B-cell acute lymphoblastic leukaemia - the index disease - HDAC7 is epigenetically silenced by EZH2 and by the KMT2A::AFF1 fusion protein, the menin inhibitor MI-538 restores HDAC7 expression with the effect enhanced by the class I HDAC inhibitor chidamide, and the combination drives leukaemic pro-B cells toward differentiation in an HDAC7-dependent manner; NO measurement of HDAC7's subcellular or subnuclear location appears anywhere in this work.

    PMID 40722046DOI
    What this evidence cannot show

    The localisation absence is the point of citing it. Multiomics in an infant pro-B-ALL cohort plus cell lines, murine models and PDX ex vivo - the right disease and the right lineage, which is rare in this graph (brief rule 73). Abstract read in full; the full text was not read, so the absence of a localisation panel is bounded at abstract level and stated as such.

    Evidence that argues against this (1)
    • The opposite therapeutic direction is published in the same disease and the same period: shRNA knock-down of the class IIA isoforms HDAC4, HDAC5 and HDAC7 induces apoptosis and cell-cycle arrest in KMT2A-rearranged ALL cells, and the HDAC4/5-selective inhibitor LMK-235 eradicates KMT2A-rearranged ALL lines and primary patient samples in vitro - although the maximum tolerated dose failed in vivo, as did MC1568 and TMP195. So one group proposes to INDUCE HDAC7 in this disease and another to INHIBIT it. This is not the rule-65 HDAC7/HDAC9 symbol swap; it is a real disagreement about one protein, and a compartment measurement is one of the few things that could reconcile it, because an induced HDAC7 that is exported or PML-sequestered is not the same reagent as one on MEF2 promoters. Verbeek TCAI et al. 2024
  7. E_q69_ct1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In MICE, the histone deacetylase HDAC7 is highly expressed in pre-B cells and dramatically down-regulated during cellular lineage conversion to macrophages; HDAC7 re-expression interferes with acquisition of the macrophage transcriptional programme, blocking induction of key genes for macrophage function including immune, inflammatory and defence response, cellular response to infection, positive regulation of cytokine production and phagocytosis, and suppresses the ability of the converted cells to phagocytose bacteria and to respond to endotoxin; HDAC7 specifically interacts with the transcription factor MEF2C in pre-B cells and is recruited to MEF2 binding sites at the promoters of genes critical for macrophage function.

    HDAC7 is a repressor of myeloid genes whose downregulation is required for transdifferentiation of pre-B cells into macrophages. · Barneda-Zahonero B, Román-González L, Collazo O, Rafati H, Islam AB, Bussmann LH, di Tullio A, De Andres L, Graf T, López-Bigas N, Mahmoudi T, Parra M · PLoS Genet 2013
    PMID 23696748DOI
    What this evidence cannot show

    Abstract read in full; full text not read. Murine pre-B cells in an INDUCED transdifferentiation system, not a physiological transition and not a leukaemia (brief rule 73). The HDAC7 arm is a re-expression (gain-of-function) experiment in cells where the gene had been switched off. No human data and no KMT2A fusion anywhere in the study.

    Evidence that argues against this (1)
    • An induced C/EBP-driven conversion is a strong perturbation and the repressive role it reveals might be an artefact of forcing two programmes into one cell. The in vivo conditional knockout is the necessary control and it does support a requirement - but it comes from the same laboratory, so the two results are not independent replications of each other. Azagra A et al. 2016
  8. E_q69_ct2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In MICE, HDAC7 shows a lymphoid-specific expression pattern within the haematopoietic system, and conditional knockout demonstrates that HDAC7 deletion dramatically blocks early B-cell development and gives rise to severe lymphopenia in peripheral organs while also leading to pro-B-cell lineage promiscuity; HDAC7 represses myeloid and T-lymphocyte genes in B-cell progenitors through interaction with MEF2C, is recruited to promoters and enhancers of target genes, and its absence leads to increased enrichment of active histone marks.

    In vivo conditional deletion of HDAC7 reveals its requirement to establish proper B lymphocyte identity and development. · Azagra A, Román-González L, Collazo O, Rodríguez-Ubreva J, de Yébenes VG, Barneda-Zahonero B, Rodríguez J, Castro de Moura M, Grego-Bessa J, Fernández-Duran I, Islam AB, Esteller M, Ramiro AR, Ballestar E, Parra M · J Exp Med 2016
    PMID 27810920DOI
    What this evidence cannot show

    Abstract read in full; full text not read. Murine conditional knockout in normal B lymphopoiesis - not leukaemia, not human (brief rule 73). 'Lineage promiscuity' is a transcriptional-profile description; the abstract does not report functional conversion of the promiscuous pro-B cells. Same laboratory as PMID 23696748, so the two are one group's programme rather than independent confirmation.

    Evidence that argues against this (1)
    • This project's genome-wide Chronos scan places HDAC7 at rank 17,469 of 17,728 for KMT2A-rearranged selectivity, with the KMT2A-rearranged arm LESS depleted than other leukaemia lines and pan-essentiality of only -0.1660. Whatever HDAC7 does in normal B-cell development, the dependency screens do not show KMT2A-rearranged leukaemia cells needing it, and the entry's therapeutic contradiction has to be read with that in view. source
  9. E_q69_ct3weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    A 2021 Leukemia paper asserts in its TITLE that HDAC7 is a major contributor in the pathogenesis of infant t(4;11) proB acute lymphoblastic leukaemia. This item records the identifier and the title only: the paper could not be read.

    HDAC7 is a major contributor in the pathogenesis of infant t(4;11) proB acute lymphoblastic leukemia. · de Barrios O, Galaras A, Trincado JL, Azagra A, Collazo O, Meler A, Agraz-Doblas A, Bueno C, Ballerini P, Cazzaniga G, Stam RW, Varela I, De Lorenzo P, Valsecchi MG, Hatzis P, Menéndez P, Parra M · Leukemia 2021
    PMID 33262526DOI
    What this evidence cannot show

    KNOWN BUT UNREADABLE, checked 2026-09-21. efetch db=pubmed returns an empty abstract; Europe PMC core returns pmcid null and isOpenAccess N with a zero-length abstract, so the fullTextXML, efetch db=pmc and PMC HTML routes are structurally unavailable; resolving the DOI to the publisher with a browser User-Agent returns a 305-character JavaScript challenge stub, not the article. Nothing about the paper's contents, methods or conclusions is written here, and none of this entry's claims rests on it.

    Evidence that argues against this (1)
    • An unreadable paper cannot corroborate anything. It is listed because a reader would otherwise reasonably expect the most on-title paper in the field to be cited, and its absence should be visible as a reading failure rather than as an absence of literature. The independent instrument that CAN be read - this project's Chronos scan - does not show a KMT2A-rearranged requirement for HDAC7. source
Connected to