HDAC9 - the only HDAC with a KMT2A-rearrangement expression signature across 211 paediatric patients, and it is high, prognostic, and independent of partner and lineage

{
  "n_patients": 211,
  "population": "paediatric acute leukaemia",
  "genes_measured": "HDAC1-11, SIRT1, SIRT7, MEF2C, MEF2D (mRNA)",
  "kmt2a_signature": "high HDAC9 and high MEF2D",
  "independent_of": [
    "age",
    "KMT2A partner gene",
    "lineage"
  ],
  "prognostic": "HDAC9 overexpression adversely prognostic REGARDLESS of KMT2A rearrangement - i.e. not a KMT2A-specific prognostic effect"
}

[
  "drug-venetoclax plus menin inhibitor combination (MI-503 / MEN1i)"
]

[
  "HDRP",
  "MITR",
  "HDAC7B"
]

{
  "depmap_common_essential": null,
  "note": "DepMap not queried in this beat. An expression signature is not a dependency; the KAT8/MOF case in this graph is the worked example of the difference."
}

{
  "chronos_mean_kmt2a_rearranged": 0.1513,
  "chronos_mean_other_leukaemia": 0.1228,
  "difference": 0.0286,
  "pan_essentiality_all_lines": 0.1041,
  "selectivity_rank": 12717,
  "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."
}

{
  "subcellular": "NUCLEUS, BY SIMILARITY ONLY - the weakest localisation record in this tranche. UniProt Q9UKV0 fetched 2026-09-21 carries a single subcellular line, 'Nucleus', at ECO:0000250, which means INFERRED FROM A HOMOLOGUE and is not experimental evidence of any kind. There is no note, no isoform-scoped comment and no ECO:0000269 [E_q77_loc1]. This is the shape Q71 flagged on SMARCD2 and it is worse here, because HDAC9's own founding paper reports isoforms that LACK THE NUCLEAR LOCALIZATION SIGNAL [E_q77_loc2] - so the one thing the record asserts is the thing the primary literature says is isoform-dependent.",
  "subnuclear": "NOT MEASURED IN ANY SYSTEM, AND FOR HDAC9 THE ISOFORM QUESTION COMES FIRST. No study has placed HDAC9 in a named subnuclear compartment. Before that question is even askable, the record has an unresolved isoform problem: the founding paper reports that HDAC9 has SEVERAL ALTERNATIVELY SPLICED ISOFORMS, that one of them is MITR (also called HDRP), which LACKS THE DEACETYLASE CATALYTIC DOMAIN entirely, and that OTHER ISOFORMS LACK THE NUCLEAR LOCALIZATION SIGNAL [E_q77_loc2]. A class IIa study independently shows that the signal-responsive C-terminal nuclear export sequence present in HDAC4, HDAC5 and HDAC7 is ABSENT from MITR [E_q77_loc3]. So HDAC9 produces isoforms that cannot leave the nucleus and isoforms that cannot enter it, and UniProt Q9UKV0 records ONE line - 'Nucleus' - at ECO:0000250, by similarity [E_q77_loc1]. Asking for HDAC9's subnuclear address before resolving which isoform is expressed in a KMT2A-rearranged blast would be answering the wrong question.",
  "explicit_no": "SAY IT PLAINLY: there is no experimental subcellular or subnuclear localisation of HDAC9 in any haematopoietic cell, and none in a KMT2A-rearranged cell. The only dedicated HDAC9 localisation work found is a methods chapter on visualising HDAC9 spatiotemporal subcellular localisation in PRIMARY NEURON cultures [E_q77_loc4]. UNPUBLISHED rather than not-applicable, and unusually cheap to fix: the entry already records HDAC9 as high in 211 paediatric patients and as indispensable for KMT2Ar-AML proliferation, so the cells and the antibodies exist.",
  "how_shown": "nothing experimental in the database [E_q77_loc1]; isoform cloning and characterisation including identification of isoforms lacking the NLS [E_q77_loc2]; NES mapping across class IIa showing MITR lacks the export sequence [E_q77_loc3]; time-lapse imaging protocols in primary neurons [E_q77_loc4]",
  "systems": "HEK/COS-type expression systems for the isoform work, muscle cells for the class IIa NES mapping, and primary rodent neurons for the imaging protocol. NO haematopoietic system of any kind.",
  "shown_for_fusion_localisation": "NO - nobody has looked in any fusion. HDAC9 has never been imaged or fractionated in a KMT2A-rearranged cell, in an MLL-AF9 mouse, or in any other fusion system, despite being the only HDAC in this beat with a patient-level KMT2A-rearrangement expression signature across 211 children.",
  "node": "localization-dot1l-nucleus-measured-cytoplasm-by-similarity",
  "the_isoform_question_has_a_therapeutic_edge": "MITR/HDRP is an HDAC9 isoform with NO catalytic domain, and it is a MEF2-interacting transcriptional repressor - i.e. it represses by scaffolding, not by deacetylating. This entry records high HDAC9 AND high MEF2D as the KMT2A-rearranged signature. If the transcript driving that signature is catalytically dead, then a catalytic HDAC inhibitor cannot touch it, and the venetoclax-plus-menin-inhibitor result at [E2] - which REPRESSES HDAC9 rather than inhibiting it - would be the mechanistically correct way to hit it. This graph already holds the catalytic-versus-scaffold question for class IIa as an open hypothesis; the HDAC9 isoform structure is a concrete, measurable instance of it, and the measurement is an isoform-resolved RT-PCR or a long-read transcript assembly on existing KMT2A-rearranged samples."
}

{
  "evidence_class_key": "concept-expressed-required-maintains-three-evidence-classes",
  "identifier_note": "HDAC9 is Q9UKV0, 1,011 aa. data/symbol-collisions.json was grepped before the first query and THE ROW ALREADY EXISTED under the token HDAC7: gene_exact:HDAC7 returns HDAC9 FIRST because HDAC9's protein name reads 'Histone deacetylase 7B', the registry marks block_autogeneration TRUE, and it notes the two point in OPPOSITE therapeutic directions in this disease. RE-VERIFIED LIVE 2026-09-21: the reviewed rows are Q9UKV0 HDAC9 (1,011 aa, synonyms HDAC7 HDAC7B HDRP KIAA0744 MITR) and Q8WUI4 HDAC7 (952 aa). The registry file was NOT edited, per the dispatch.",
  "expressed_in": {
    "class": 1,
    "summary": "HIGH in KMT2A-rearranged paediatric acute leukaemia and, per this entry's [E1], high REGARDLESS OF AGE, KMT2A PARTNER AND LINEAGE - the only lineage-independent expression claim in this beat. What is NOT known is which ISOFORM is high, and the isoforms differ in whether they carry a catalytic domain or a nuclear localisation signal [E_q77_loc2]."
  },
  "required_in": {
    "class": 2,
    "claim": "HDAC9 is reported INDISPENSABLE FOR KMT2Ar-AML PROLIFERATION in cell lines and a MOLM13 xenograft, and is specifically repressed by venetoclax plus a menin inhibitor rather than by either monotherapy [E2]. THAT CLAIM IS NOT CORROBORATED BY THE LARGEST AVAILABLE LOSS-OF-FUNCTION DATASET: this entry's own DepMap scan block gives a Chronos mean of +0.1513 in the KMT2A-rearranged arm and +0.1228 in other leukaemia lines - POSITIVE in both, i.e. knocking HDAC9 out very slightly FAVOURS growth - with a pan-essentiality of +0.1041 and a selectivity rank of 12,717 of 17,728. Per brief rule 84 the rank is quoted as a rank and not as membership of any half of the distribution; the load-bearing observation is the SIGN, not the rank.",
    "pmids": [
      "38053201"
    ],
    "system_named_in_the_claim": "human KMT2A-rearranged AML cell lines, primary cells ex vivo and a MOLM13 xenograft - myeloid, not the index lymphoid disease",
    "nodes": []
  },
  "bounded_absence_on_cell_type": {
    "date": "2026-09-21",
    "tool": "NCBI esearch db=pubmed",
    "queries_run": [
      "HDAC9[tiab] AND (hematopoi*[tiab] OR leukemia[tiab] OR lymphoma[tiab] OR B cell[tiab] OR T cell[tiab])  ->  42 records",
      "(MLL[tiab] OR KMT2A[tiab]) AND (MITR[tiab] OR HDRP[tiab] OR HDAC7B[tiab])  ->  0 records",
      "(MITR[tiab] OR HDRP[tiab]) AND (deacetylase[tiab] OR MEF2[tiab])  ->  18 records (positive control)"
    ],
    "finding": "No study reports a normal haematopoietic cell type in which HDAC9 is required. The 42 records are dominated by immunology (Treg function, macrophage polarisation), solid tumours and stroke. The nearest relevant biology is a Mef2c/Mef2d/Hdac9 circuit controlling regulatory T cells, which is interesting only because MEF2C and MEF2D are central to this graph - it is not a cell-type requirement in haematopoiesis.",
    "bound": "titles and abstracts only, brief rule 80; aliases HDRP, MITR and HDAC7B were searched, closing the gap this entry's aliases_not_searched field recorded."
  },
  "context": "human KMT2A-rearranged AML lines and xenograft; no normal-haematopoiesis requirement exists"
}

{
  "evidence_class_key": "concept-expressed-required-maintains-three-evidence-classes",
  "maintains_state": {
    "class": 3,
    "states": [
      "cell-state-hypoxic-glycolytic",
      "cell-state-under-menin-inhibition"
    ],
    "claim": "HDAC9 IS A HYPOXIA-RESPONSIVE NODE AND A MARKER OF THE STATE UNDER COMBINED MENIN AND BCL-2 INHIBITION, which is a state claim rather than a target claim. Hypoxia induces HDAC9 through HIF-1A, and hypoxic induction SENSITISES KMT2A-rearranged AML cells to venetoclax plus MI-503; the combination specifically represses HDAC9 where neither monotherapy does [E2]. So HDAC9 sits at the intersection of the marrow hypoxic niche and this graph's leading drug combination, and it arrives there as a downstream consequence.",
    "pmids": [
      "38053201"
    ],
    "system_named_in_the_claim": "human KMT2A-rearranged AML cell lines, primary cells ex vivo and a MOLM13 xenograft under normoxia and hypoxia - myeloid, human, fusion-bearing"
  },
  "why_this_is_a_state_and_not_a_target": "The entry already says HDAC9 is a downstream effector rather than a validated target. The DepMap sign settles the direction for this beat: gene effect is POSITIVE in both KMT2A-rearranged and other leukaemia lines, so removing HDAC9 does not impair growth in 1,000+ human cancer lines. A gene that is high, prognostic, hypoxia-induced and drug-repressed, but dispensable in CRISPR, is a state marker. The hypoxia link makes it a potentially USEFUL one, because the niche-resident fraction is the fraction that survives therapy.",
  "confidence": "moderate",
  "caveats": "The requirement claim and the hypoxia mechanism come from ONE study in myeloid disease and conflict in sign with the genome-wide CRISPR data. The 211-patient signature is mRNA expression, not protein and not isoform-resolved, so it cannot say whether the catalytically dead MITR isoform or full-length HDAC9 is elevated. There is no selective HDAC9 inhibitor. And no localisation measurement exists in any haematopoietic cell, so a state claim about HDAC9 currently has no spatial component at all.",
  "counter_evidence": [
    {
      "note": "The class-level localisation biology undercuts the simplest reading of the HDAC9 signature. MITR/HDRP is an HDAC9 isoform with NO catalytic domain and, per the class IIa NES mapping, no signal-responsive nuclear export sequence either - while other HDAC9 isoforms lack the nuclear localisation signal. High 'HDAC9 mRNA' in 211 patients is therefore compatible with a catalytically dead, constitutively nuclear scaffold, with a cytoplasmic NLS-less species, or with the full-length enzyme, and the three imply completely different interventions. An isoform-resolved measurement in KMT2A-rearranged samples would settle it and has not been done.",
      "citation": {
        "title": "Cloning and characterization of a histone deacetylase, HDAC9.",
        "authors": "Zhou X, Marks PA, Rifkind RA, Richon VM",
        "year": 2001,
        "venue": "Proc Natl Acad Sci U S A",
        "pmid": "11535832",
        "doi": "10.1073/pnas.191375098",
        "type": "paper"
      }
    }
  ]
}