HDAC5 - the eraser in an existing entry's own id, and the one HDAC in this graph that erases a mark on a fusion partner protein rather than on a histone

[
  "LMK-235 (HDAC4/5)",
  "MC1568 (class IIA)",
  "TMP195 (class IIA)"
]

{
  "depmap_common_essential": null,
  "note": "DepMap not queried in this beat. As with HDAC1/HDAC2, a single-gene score is weak evidence for a class whose members were perturbed together in the only functional study."
}

{
  "chronos_mean_kmt2a_rearranged": -0.0246,
  "chronos_mean_other_leukaemia": -0.1215,
  "difference": 0.0969,
  "pan_essentiality_all_lines": -0.0786,
  "selectivity_rank": 16829,
  "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": "SHUTTLES BETWEEN NUCLEUS AND CYTOPLASM, AND THE UNIPROT RECORD HAS NO EXPERIMENTAL CODE FOR ANY OF IT. Q9UQL6 fetched 2026-09-21 annotates 'Nucleus' and 'Cytoplasm' with EMPTY evidence arrays, and an UNCODED note that states the whole mechanism: 'Shuttles between the nucleus and the cytoplasm. In muscle cells, it shuttles into the cytoplasm during myocyte differentiation. The export to cytoplasm depends on the interaction with a 14-3-3 chaperone protein and is due to its phosphorylation at Ser-259 and Ser-498 by AMPK, CaMK1 and SIK1' [E_q77_loc1]. THE UNCODED NOTE IS THE MOST INFORMATIVE THING IN THE RECORD - it names the residues, the chaperone and three kinases - and an ECO:0000269 filter discards all of it. Brief rule 91 item 2, on a protein whose entire biology in this graph is a location switch.",
  "subnuclear": "NO SUBNUCLEAR COMPARTMENT HAS EVER BEEN MEASURED FOR HDAC5 - and for this protein that is a smaller gap than it looks, because the regulated variable is one level up. What is established, repeatedly and mechanistically, is the NUCLEUS-VERSUS-CYTOPLASM partition: a 14-3-3-dependent export driven by phosphorylation at Ser-259 and Ser-498 [E_q77_loc2], a signal-responsive nuclear export sequence at the extreme C-terminus [E_q77_loc3], and at least three distinct kinases throwing the switch - CaMK in muscle [E_q77_loc2], CaMKII in the AF9 system this entry already holds at [E1], and CK2 under DNA damage [E_q77_loc4]. Within the nucleus nobody has asked where it goes. UNPUBLISHED rather than not-applicable: HDAC5 is a MEF2 corepressor that must reach MEF2 target promoters, and the obvious question - is the nuclear pool at promoters, in a body, or diffuse - has no answer.",
  "how_shown": "yeast two-hybrid, 14-3-3 binding assays and CaMK-stimulated export in muscle cells [E_q77_loc2]; deletion and mutation mapping of a C-terminal signal-responsive NES [E_q77_loc3]; CK2-dependent shuttling with Ku70 deacetylation and DNA-end-resection readouts in pancreatic cancer models [E_q77_loc4]; and, already in this entry, CaMKII phosphorylation at S259/S498 with an S259A/S498A phospho-defective mutant as the control [E1]",
  "systems": "skeletal muscle and C2C12 myocytes for the founding work; pancreatic ductal adenocarcinoma lines, mouse models and patient-derived organoids for the CK2/Ku70 work. NO haematopoietic localisation system and nothing KMT2A-rearranged - the haematopoietic HDAC5 literature is functional, not spatial.",
  "shown_for_fusion_localisation": "PARTIAL, and it is the most fusion-proximal localisation result in the acetylation beat - but the fusion in question is KMT2A::MLLT3 and the protein whose compartment matters is AF9, not HDAC5. This entry's [E1] shows that CaMKII-driven NUCLEAR EXPORT OF HDAC5 is what allows AF9 K339 acetylation to accumulate after ionising radiation, with an HDAC5(S259A,S498A) mutant blocking the change. So HDAC5's compartment is a documented upstream determinant of a fusion partner's modification state. NOTHING has been measured in KMT2A::AFF1, and K339 is absent from the fusion protein in about 92% of KMT2A::MLLT3 patients, so the axis operates on the wild-type allele.",
  "node": "localization-af9-cytoplasmic-pool-and-hsp90-dependent-nuclear-residence",
  "the_eraser_is_regulated_by_address_not_by_activity": "This entry already states the principle in its data block - 'the eraser is regulated by LOCATION, not by catalysis' - and the localisation audit turns it into a bounded, sourced claim with a residue-level mechanism and three independent kinase inputs. The therapeutic consequence is sharp and is not in the literature: a catalytic HDAC5 inhibitor and a kinase that exports HDAC5 produce the same substrate outcome by different means, but ONLY THE KINASE ROUTE IS ALREADY WIRED INTO THE CELL. The class IIA in vivo failure recorded at [E2] - LMK-235, MC1568 and TMP195 all negative at maximum achievable dose in a KMT2A-rearranged ALL xenograft - is a failure of the catalytic route. Nobody has tested whether forcing HDAC5 export, or blocking it, does anything in these cells."
}

{
  "evidence_class_key": "concept-expressed-required-maintains-three-evidence-classes",
  "identifier_note": "HDAC5 is Q9UQL6, 1,122 aa, one clean reviewed row. data/symbol-collisions.json was grepped for HDAC5 before the first query and has NO row; the adjacent HDAC7 row DOES exist and is the family warning - gene_exact:HDAC7 returns HDAC9 first because HDAC9's protein name reads 'Histone deacetylase 7B'. Within class IIa the lengths are HDAC4 1,084, HDAC5 1,122, HDAC7 952 and HDAC9 1,011, so HDAC4 and HDAC5 are only 38 residues apart - which matters here because the only selective tool compound in this entry, LMK-235, is an HDAC4/5 inhibitor and cannot separate them either.",
  "expressed_in": {
    "class": 1,
    "summary": "of the eleven classic HDAC family members, six are expressed in human erythroid cells - HDAC1, 2, 3, 5, 6 and 7 - and HDAC5 IS THE MOST SIGNIFICANTLY UPREGULATED during terminal erythroid differentiation [E_q77_ct1]. That is a specific, human, lineage-resolved expression statement, which is rare in this beat."
  },
  "required_in": {
    "class": 2,
    "claim": "IN HUMAN CD34+ CELLS, HDAC5 IS REQUIRED FOR TERMINAL ERYTHROID DIFFERENTIATION: shRNA or siRNA knockdown followed by erythroid culture increases apoptosis, DECREASES CHROMATIN CONDENSATION and IMPAIRS ENUCLEATION of erythroblasts, with activation of p53 in association with increased p53 ACETYLATION [E_q77_ct1]. Separately and in the same cell source, HDAC5 acts as a NEGATIVE regulator of stem-cell homing: specific HDAC5 inhibition strongly upregulates surface CXCR4 on human cord-blood HSCs and progenitors, enhancing SDF-1/CXCR4 chemotaxis, bone-marrow homing, SCID-repopulating-cell frequency and long-term and secondary engraftment in NSG mice, through raised nuclear acetylated p65 and NF-kB-dependent CXCR4 transcription [E_q77_ct2].",
    "pmids": [
      "34036344",
      "30013077"
    ],
    "system_named_in_the_claim": "HUMAN CD34+ cells and human cord-blood HSC/HPC in culture and in NSG xenografts - human primary haematopoietic material in both arms, and no fusion in either",
    "nodes": [
      "cell-type-erythroid-progenitor",
      "cell-type-hsc-bone-marrow"
    ],
    "why_both_substrates_are_non_histone": "The erythroid requirement runs through p53 acetylation and the homing effect runs through p65 acetylation. Together with the AF9 K339 result this entry already holds, ALL THREE of HDAC5's mechanistically characterised substrates in relevant cells are NON-HISTONE. This entry's title claim - that HDAC5 is the one HDAC here erasing a mark on a protein rather than on a histone - generalises further than the entry originally claimed."
  },
  "context": "human primary haematopoietic cells in both arms; no fusion, no leukaemia"
}

{
  "evidence_class_key": "concept-expressed-required-maintains-three-evidence-classes",
  "maintains_state": {
    "class": 3,
    "states": [
      "cell-state-stem-like-self-renewal",
      "cell-state-differentiation-block"
    ],
    "claim": "HDAC5 SETS A DIFFERENTIATION-VERSUS-HOMING TRADE-OFF IN HUMAN HAEMATOPOIETIC CELLS, AND THE TWO ARMS POINT IN OPPOSITE THERAPEUTIC DIRECTIONS. Losing HDAC5 BLOCKS terminal erythroid differentiation - failed chromatin condensation, failed enucleation, apoptosis, acetylated p53 up [E_q77_ct1]. Inhibiting HDAC5 IMPROVES stem-cell function - CXCR4 up, homing up, long-term and secondary engraftment up [E_q77_ct2]. So the same perturbation makes progenitors better at being stem cells and worse at finishing differentiation, which is the shape of the state this graph calls a differentiation block.",
    "pmids": [
      "34036344",
      "30013077"
    ],
    "system_named_in_the_claim": "human CD34+ erythroid cultures and human cord-blood HSC xenografts - no fusion in either"
  },
  "what_this_predicts_for_the_class_iia_proposal": "This entry records that class IIA knockdown kills KMT2A-rearranged ALL cells in vitro and that three selective compounds then failed in vivo. The human cell-type data adds a prediction the xenograft could not have shown, because NSG mice have no human erythropoiesis to lose: an HDAC5 inhibitor that did reach its target would be expected to impair terminal erythroid differentiation AND to mobilise or retain stem/progenitor cells. In an infant, an enucleation defect is a transfusion-dependent anaemia.",
  "confidence": "strong",
  "caveats": "Both arms are knockdown or pharmacological inhibition in normal human cells, not genetic knockout and not leukaemia. The homing study uses 'specific HDAC5 inhibition' whose selectivity within class IIa is the standing problem for this family - LMK-235 is HDAC4/5 and the class IIA compounds are not isoform-selective, so an HDAC4 contribution cannot be excluded in either study. The erythroid mechanism rests on p53 acetylation, which is an association in the paper rather than a demonstrated requirement. No fusion anywhere.",
  "counter_evidence": [
    {
      "note": "The two human results disagree about whether losing HDAC5 is good or bad for a haematopoietic cell, and the disagreement is not resolvable from the papers: knockdown impairs erythroid differentiation and increases apoptosis, while inhibition improves HSC homing and engraftment. Both used human CD34+ or cord-blood cells. The most likely reconciliation is cell-stage - HDAC5 is most highly expressed at TERMINAL erythroid differentiation and the homing work is in uncommitted HSC/HPC - but nobody has run both readouts in one experiment, and a graph that quotes either alone would misrepresent the toxicity profile of an HDAC5-directed agent.",
      "citation": {
        "title": "Neutralizing negative epigenetic regulation by HDAC5 enhances human haematopoietic stem cell homing and engraftment.",
        "authors": "Huang X, Guo B, Liu S, Wan J, Broxmeyer HE",
        "year": 2018,
        "venue": "Nat Commun",
        "pmid": "30013077",
        "doi": "10.1038/s41467-018-05178-5",
        "type": "paper"
      }
    }
  ]
}