Oncofusion

start / graph / mechanism-kat8-h4k16ac-sets-the-hsc-quiescence-set-point

Mechanismstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q60-cellstate2

KAT8/MOF sets the quiescence set-point of the blood stem cell - and its catalytic activity is separately required for MLL-AF9 leukaemia, which DepMap also cannot see

also called MOF quiescence set-point · H4K16ac and HSC quiescence

The amount of a chemical tag called H4K16 acetylation decides whether a blood stem cell rests or wakes up, and the enzyme that puts the tag on is KAT8. The same enzyme's activity is needed for MLL-rearranged leukaemia to grow in mice. A large cancer screen detects neither, because KAT8 is needed by almost every cell line ever tested.

The same thing, for a scientist

Differential H4K16 acetylation levels - the mark KAT8/MOF writes - ensure a balance between QUIESCENCE and activation in murine haematopoietic stem cells [E1]. Separately, the histone acetyltransferase ACTIVITY of MOF is required for ADULT but not for early FETAL haematopoiesis [E2], which is a cell-type-and-developmental-stage answer of a kind no cancer dependency screen can produce. And it cannot: KAT8 is depleted below -0.5 in 99.5% of 1,258 screened lines, the strongest floor in this beat's worklist, with a KMT2A-rearranged-minus-all-lines difference of -0.036 that is not a small selectivity but an unmeasurable one [E3]. **SCOPE CORRECTED 2026-09-21 (Q68).** This entry previously read as though MOF's catalytic requirement were a normal-haematopoiesis story that a leukaemia screen happens to be blind to. It is not confined to normal haematopoiesis. Valerio et al. (PMID 28202522, Cancer Res 2017) found MOF in an RNAi screen for druggable chromatin regulators in MLL-rearranged leukaemia; **conditional Mof deletion in a mouse MLL-AF9 model reduced tumour burden and prolonged host survival**, catalytically inactive MOF mutants failed to rescue, a small-molecule MYST inhibitor blocked growth of murine AND HUMAN MLL-AF9 lines, and Mof inactivation also suppressed NUP98-HOXA9-driven AML. So the leukaemia requirement was demonstrated - by conditional deletion against a survival endpoint, not by a 21-day proliferation screen. The right statement is not 'DepMap cannot see it because of the floor' but **'a proliferation screen and an in vivo survival experiment measure different quantities, and only the second can attribute a requirement to a genotype'**.

Go deeper
The floor, and what it means for the cell-type question

This graph's standing brief already names KAT8 as the worked example of floor compression (99.5% of lines depleted, difference +0.004 on the earlier scan). This beat reproduces it from a live fetch 3. The consequence for the beat's question is blunt: for KAT8, DepMap answers 'every cell type', which is the same as answering nothing.

The reason the conditional-knockout evidence is so much more informative is that it varies something the screen holds fixed — developmental stage — and finds a boundary there: adult yes, early fetal no 2.

A quiescence result is rarer than it sounds

Across the sixteen highest-degree core genes in this graph, the bounded cell-state queries run in this beat returned exactly one explicit QUIESCENCE result, and it is this one. Self-renewal and the differentiation block are well populated; quiescence, dormancy, drug tolerance and persistence are almost empty. That asymmetry is itself a finding and is recorded in the beat memo.

Note what it means for therapy: quiescence is the state most often invoked to explain why residual disease survives treatment, and the core interactome has essentially no evidence about it.

Identifier trap, because it nearly hid the evidence

(KAT8 OR MOF OR MYST1 OR hMOF)[tiab] returns 28,480 records, overwhelmingly metal-organic frameworks. Three of the six hits in the leukaemia-scoped cell-state query are MOF nanomaterial papers. A reader seeing only the bare count would conclude KAT8 was one of the best-studied genes in the panel. The on-topic set is reachable only by pairing the gene with H4K16 or H4K16ac, which returns five records containing both papers cited here 4.

Correction, 2026-09-21 (Q68)

What was withdrawn. The previous title's clause 'the one explicit quiescence result in the core set, and DepMap cannot see it', read together with a summary that placed MOF's catalytic requirement entirely in normal haematopoiesis. The quiescence half stands. The implication that the leukaemia side is unmeasured does not.

Why. PMID 28202522 reports conditional Mof deletion reducing tumour burden and prolonging survival in mouse MLL-AF9 leukaemia, rescue failure with catalytically inactive MOF, a MYST inhibitor active against human MLL-AF9 lines, and suppression of NUP98-HOXA9 AML. The catalytic-activity requirement - 'the form a drug would test', as E2's caveat correctly says - has been tested in the leukaemia.

Note the direction this moves the therapeutic argument. E2 establishes that MOF HAT activity is required for ADULT and not early FETAL haematopoiesis, and this entry's third open question asks whether a catalytic MOF inhibitor would therefore SPARE fetal-programme cells and so spare infant blasts. PMID 28202522's MLL-AF9 and NUP98-HOXA9 results argue the opposite way, and the two have never been run in the same system. That tension is now the most interesting thing in this entry and it was invisible while the leukaemia arm was missing.

And a citation-reuse finding, which is the cycle-38 king log's item 1 in action. PMID 28202522 was ALREADY cited as support in six entries, including mechanism-kat8-mof-requirement-is-myeloid-not-fusion-specific in comparative/, which reads the paper more finely than this correction first did: the catalytic-rescue work is MLL-AF9 only, the NUP98-HOXA9 arm is gene-level deletion, the requirement is NOT greater than in lineage-matched leukaemia without the fusion, and ana-kmt2a-lost-module-ledger records that the survival effect is 22-24 to 31 days rather than abolition. So the graph already knew, and better - this entry was the one arguing with it. One grep on the PMID before writing would have found that, and the objection is now written to agree with the sibling entries rather than to restate them.

Facts and figures
genegene-kat8
uniprotQ9H7Z6
evidence class3
cell type
{
  "required_in": "adult murine haematopoiesis",
  "not_required_in": "early fetal murine haematopoiesis"
}
cell state
{
  "maintains": "cell-state-quiescence",
  "mark": "H4K16ac"
}
depmap
{
  "pct_all_lines_depleted": 99.5,
  "mean_all_lines": -1.453,
  "mean_kmt2ar": -1.489,
  "floor_compressed": true
}
alias contamination
{
  "query": "(KAT8 OR MOF OR MYST1 OR hMOF)[tiab]",
  "count": 28480,
  "dominant_false_referent": "metal-organic framework"
}
How this could bite us back
  • KAT8 is pan-essential at 99.5% of screened lines, so systemic inhibition has no plausible therapeutic index; any KAT8-directed idea has to be allele-, complex- or context-selective from the start.
  • If H4K16ac sets the quiescence set-point, then a drug that lowers it could push residual blasts INTO quiescence and out of reach of cycle-dependent chemotherapy.
  • The adult-but-not-fetal requirement means a toxicity prediction from an adult model will not transfer to an infant.
Still open
  1. Is the H4K16ac quiescence set-point altered in KMT2A-rearranged blasts? Nobody has measured the mark in this disease in a sorted quiescent fraction.
  2. Does the adult-versus-fetal split in MOF requirement apply to the fusion-transformed cell, which retains a fetal programme?
  3. Would a catalytic MOF inhibitor spare fetal-programme cells and therefore spare infant blasts — the opposite of the usual therapeutic-window argument?
  4. Does the adult-versus-fetal split in MOF catalytic requirement (PMID 27827827) survive in a fusion-transformed cell that retains a fetal programme, given that MOF HAT activity is required for MLL-AF9 leukaemogenesis (PMID 28202522)? The two results point in opposite therapeutic directions and have never been run in one system.
Every source behind this entry (4)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In mice, differential H4K16 acetylation levels ensure a balance between quiescence and activation in haematopoietic stem cells.

    PMID 34362741DOI
    What this evidence cannot show

    Murine, normal haematopoiesis, not leukaemia and not human (brief rule 73). Abstract read in this beat. The mark is the readout; attribution to KAT8 rests on MOF being its principal writer, which the companion genetic work supports but this entry does not independently verify.

    Evidence that argues against this (1)
    • The quiescence result is read out on the MARK, and attributing it to KAT8 assumes MOF is the dominant writer in that compartment. The companion genetic paper shows the acetyltransferase ACTIVITY matters in adult but not early fetal haematopoiesis, which means the writer-to-mark relationship is itself context-dependent and the attribution does not transfer across developmental stage. Valerio DG et al. 2017
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In mice, the histone acetyltransferase activity of MOF is required for adult but not for early fetal haematopoiesis.

    PMID 27827827DOI
    What this evidence cannot show

    Murine. A catalytic-activity requirement, which is stronger than a protein requirement and is the form a drug would test. Abstract read in this beat.

    Evidence that argues against this (1)
    • PUBLISHED CONFLICTING RESULT that takes the MOF HAT-activity requirement out of normal haematopoiesis and into the leukaemia. Valerio et al. found MOF in an RNAi screen for druggable chromatin-regulator targets in MLL-rearranged leukaemia; conditional deletion of Mof in a mouse MLL-AF9 model REDUCED TUMOUR BURDEN AND PROLONGED SURVIVAL, catalytically inactive MOF mutants failed to rescue so the enzymatic activity is what is required, a small-molecule MYST inhibitor blocked growth of murine AND HUMAN MLL-AF9 lines, and Mof inactivation also suppressed NUP98-HOXA9 leukaemia. So the adult-versus-fetal contrast is not the only place MOF's catalytic activity has been separated: it is required in the leukaemia this graph is about, by the same kind of catalytic-rescue experiment. Two qualifications the graph already carries and this entry did not: the catalytic work (HAT-dead constructs, H4K16ac blots, RNA-seq, gamma-H2AX) is MLL-AF9 only and the NUP98-HOXA9 arm received gene-level Kat8 deletion alone, per `mechanism-kat8-mof-requirement-is-myeloid-not-fusion-specific`; and the in vivo magnitude is a median survival extension of roughly 22-24 to 31 days, not abolition, per `ana-kmt2a-lost-module-ledger`. Valerio DG et al. 2017
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    In DepMap fetched 2026-09-20, KAT8 is depleted below -0.5 in 99.5% of 1,258 screened lines (mean gene effect -1.453), and its KMT2A-rearranged mean of -1.489 differs from the all-lines mean by only -0.036.

    What this evidence cannot show

    Floor compression: with 99.5% of lines depleted there is no dynamic range in which selectivity could appear, so this number is uninterpretable rather than null. Cancer cell lines in culture.

    Evidence that argues against this (1)
    • PUBLISHED CONFLICTING RESULT against the inference that floor compression is WHY nothing appears here. A REQUIREMENT was demonstrated - by a different design against a different endpoint. Valerio et al. conditionally deleted Mof in an MLL-AF9 mouse model and measured tumour burden and host survival rather than proliferation of a cultured line over three weeks, with catalytically inactive MOF failing to rescue; a MYST inhibitor also blocks human MLL-AF9 line growth, which a 99.5%-pan-essential gene effect would have predicted anyway and so cannot discriminate. Two limits, both already recorded elsewhere in this graph and neither in this entry: `mechanism-kat8-mof-requirement-is-myeloid-not-fusion-specific` establishes that the requirement is NOT greater in these fusions than in lineage-matched leukaemia without them, so this is a requirement and not a selectivity; and `ana-kmt2a-lost-module-ledger` records that the in vivo magnitude is a median survival extension from 22-24 days to 31, not abolition, while the colony result is abolished. The correct statement is that a proliferation screen and an in vivo survival experiment measure different quantities - not that either shows KAT8 is a selective target. Valerio DG et al. 2017
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    A PubMed query on KAT8 and its aliases returns 28,480 records, of which the MOF alias contributes overwhelming contamination from metal-organic framework chemistry; pairing the gene with H4K16 or H4K16ac reduces the haematopoietic set to five records.

    NCBI PubMed E-utilities esearch, db=pubmed, queried 2026-09-20 · NCBI Entrez Programming Utilities 2026
    What this evidence cannot show

    A retrieval measurement, not a biological one. Counts are as of 2026-09-20 and will drift.

    Evidence that argues against this (1)
    • METHODOLOGICAL DEFEATER: the count is not merely a retrieval measurement, it is a retrieval measurement whose meaning depends on a triage step the item does not name. Valerio et al.'s paper is titled 'Histone Acetyltransferase Activity of MOF Is Required for MLL-AF9 Leukemogenesis' - a title containing neither H4K16 nor any of the eighteen cell-state terms. It survives an abstract-level pairing with H4K16 (the abstract names 'the H4K16 histone acetyltransferase (HAT) MOF' and H4K16 acetylation) and would be discarded by a title-only triage of the 28,480 records. So the five-record haematopoietic set is a function of whether titles or abstracts were read, and the most important leukaemia paper for this gene sits on the boundary. Recording the count without recording the triage leaves the number uninterpretable. Valerio DG et al. 2017
Connected to