Oncofusion

start / graph / gene-usp7

Proteinmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q50-prc1-myc

USP7 - the enzyme that holds PRC1.1 together, and the only pharmacological handle on this complex that exists today

also called USP7 · HAUSP · herpesvirus-associated ubiquitin-specific protease · ubiquitin carboxyl-terminal hydrolase 7

PRC1.1 is the silencing machine at the centre of this beat and there is no drug that hits it directly. USP7 is a deubiquitinating enzyme that keeps PRC1.1 assembled, so blocking USP7 makes the complex fall apart - which is, at present, the only chemical way anyone has to turn PRC1.1 down, and it comes with a large warning attached.

The same thing, for a scientist

USP7 (Q93009, 1,102 aa, primary gene name USP7, synonym HAUSP) is an integral component of PRC1.1 together with TRIM27 [E1]. USP7 interactome analysis shows PRC1.1 is the **predominant** Polycomb complex co-precipitating with USP7, and USP7 inhibition causes PRC1.1 disassembly, loss of chromatin binding, reduced H2AK119ub and H3K27ac, and diminished transcription of active PRC1.1-controlled loci; USP7 and TRIM27 are reciprocally required for incorporation into the complex, and USP7 inhibitors impair AML proliferation independently of the USP7-MDM2-TP53 axis and delay MLL-AF9-induced leukaemia in human xenografts [E1]. A second group reaches the same axis by orthogonal means: quantitative proteomics/epigenomics places USP7 as a control hub for ncPRC1-mediated H2AK119ub1 deposition, stabilising ncPRC1.6 strongly and ncPRC1.1 less so [E2]. A third, in neuroblastoma, finds PRC1.1 as a DepMap complex-level dependency, USP7 as its top co-dependency, and the USP7 inhibitor XL177A reducing growth with reduced H2AK119Ub on chromatin [E3]. USP7 is also a hit in an ncPRC1 CRISPRi screen in GCB-DLBCL [E4]. **The warning is the point of this entry.** In KMT2A-rearranged AML, PRC1.1 loss is an established menin-inhibitor **resistance** route [E5], so a USP7 inhibitor given with a menin inhibitor is a candidate pharmacological antagonist, and the combination has never been tested.

What it looks like
loading AF-Q93009-F1…
AlphaFold AF-Q93009-F1 ↗
colourshape

This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.

very high (90+)confident (70–90)low (50–70)very low (<50)
I have never read one of these pictures before

A protein is one long chain of amino acids that folds up. The ribbon traces that chain from one end to the other; it is not a wire, it is a simplification of where the backbone runs. Drag to rotate, scroll to zoom.

Domains are the stretches of chain that do a particular job, and they are what the coloured blocks above the picture mark out. Click one and the view flies to it. In this project the important question about any domain is usually whether the fusion still has it — the labels say lost where it does not.

Anything in yellow sticks is a bound small molecule, which is usually the drug.

Go deeper
Why this entry is a warning and not a target dossier

Three independent groups agree that USP7 inhibition disassembles PRC1.1 and lowers H2AK119ub 123. In two of those three settings - AML 1 and neuroblastoma 3 - that is reported as a therapeutic good. In the setting this project actually cares about it points the other way.

The menin-inhibitor mechanism, as Neville et al. state it, is PRC1.1-dependent H2AK119ub deposition at a subset of MLL-fusion targets. Zhou et al. independently show that losing PRC1.1 subunits (PCGF1, BCOR) produces menin-inhibitor resistance by reactivating non-canonical menin targets including MYC 5. A USP7 inhibitor is a pharmacological way to lose PRC1.1 function. Putting those three statements together gives a concrete, testable prediction that nobody has tested: a USP7 inhibitor should antagonise a menin inhibitor in a KMT2A-rearranged cell.

That prediction is falsifiable and cheap - a two-drug matrix in MV4-11 or MOLM-13 with H2AK119ub CUT&RUN at the Zhou non-canonical target set - and it is the sort of combination a clinical programme might otherwise reach for on the strength of 1 alone, because 1's headline is 'USP7 inhibition delays MLL-AF9 leukaemia'. This entry exists so that the graph carries the contradiction before the combination is proposed rather than after.

What is actually established about USP7 in PRC1.1, and the standing confound

Established. USP7 and TRIM27 are integral PRC1.1 components; USP7 is required for complex integrity, and its inhibition causes loss of chromatin binding and reduced H2AK119ub 1. Independent proteomics confirms USP7 as a control hub for ncPRC1-driven H2AK119ub1 2.

The standing confound. USP7's best-known substrate is MDM2, and USP7 inhibition stabilises p53. Every USP7 phenotype in a TP53-wild-type cell is therefore contaminated by a p53 response, and this is not a fringe objection - it is the dominant published mechanism for the target class. Maat et al. address it explicitly, reporting activity 'also independent of the USP7-MDM2-TP53 axis' 1. That phrasing is an exclusion, not a measurement: it establishes that some of the effect survives p53 loss, not how much. Any USP7 result in this disease has to be read with that discount, and infant KMT2A-rearranged ALL is overwhelmingly TP53 wild-type, so the confound is maximal exactly where this project would want to use the drug.

Flavour ambiguity. Sijm et al. report USP7 stabilising ncPRC1.6 strongly and ncPRC1.1 to a lesser extent 2. Maat et al. report PRC1.1 as the predominant Polycomb complex in the USP7 interactome 1. Those are not flatly contradictory - interactome abundance and stabilisation magnitude are different quantities - but they do mean that 'USP7 inhibition is a PRC1.1 intervention' is an over-statement of what has been shown.

Facts and figures
uniprotQ93009
length AA1102
gene primaryUSP7
protein nameUbiquitin carboxyl-terminal hydrolase 7
uniprot entry nameUBP7_HUMAN
alphafold IDAF-Q93009-F1
accession resolutionUniProt REST gene_exact:USP7 AND organism_id:9606 AND reviewed:true, run 2026-09-20; exactly one reviewed row; gene_primary read back as USP7, length 1102 from the same record.
enzyme classdeubiquitinase (ubiquitin-specific protease)
role in prc1 1integral component with TRIM27; required for complex integrity and chromatin binding; USP7 counteracts TRIM27 E3 ligase activity
partner in complexTRIM27 (P14373, 513 aa, reviewed; gene_primary read back 2026-09-20) - not yet a node in this graph
standing confoundUSP7-MDM2-TP53: USP7 inhibition stabilises p53, so every phenotype in a TP53-wild-type cell is confounded; PMID 34113809 excludes but does not quantify this route
untested combination this entry flagsUSP7 inhibitor plus menin inhibitor in a KMT2A-rearranged cell; predicted antagonism; no published test found 2026-09-20
measured in kmt2a aff1 infant b allfalse
gene symbolUSP7
gene symbol notepromoted from gene_primary 2026-09-20: verify-accessions.py reads data.gene_symbol, so a gene_primary-only entry was being length-checked or id-inferred rather than symbol-asserted
subnuclear localizationMICROPUNCTATE NUCLEOPLASM, with a MINORITY pool in PML nuclear bodies (ND10). The founding description is explicit that only SOME of the larger discrete foci colocalise with PML [E_q71_loc1]. USP7 also carries a Chromosome annotation, which is the pool this entry's PRC1.1 argument actually needs and which is not the PML-body pool. [EXPLICIT NEGATIVE] NOT MEASURED IN ANY KMT2A-REARRANGED CELL. UNPUBLISHED rather than not-applicable.
epigenetic readoutH2AK119ub, INDIRECTLY AND BY STABILISING THE WRITER - and this entry already carried the content in prose at E1, E2 and in its sections, with no epigenetics key, which is the fifth PI dimension's measurement failure in one file. USP7 is a deubiquitinase (Q93009, 1,102 aa) and it is NOT a histone deubiquitinase in this graph's argument: its chromatin readout runs through PRC1.1 integrity. USP7 and TRIM27 are integral PRC1.1 components; PRC1.1 is the predominant polycomb complex in the USP7 interactome; USP7 inhibition causes PRC1.1 DISASSEMBLY and loss of chromatin binding with REDUCED H2AK119ub AND REDUCED H3K27ac, impairs AML proliferation in vitro and delays MLL-AF9 leukaemia in vivo [E1]. Independently, proteomics/transcriptomics/epigenomics define USP7 as a control hub for H2AK119ub1 deposition by non-canonical PRC1 - strongly stabilising ncPRC1.6 and, less so, ncPRC1.1 - with the deposition UNCOUPLED FROM H3K27me3 [E2]. So the readout is: LOWER H2AK119ub when USP7 is inhibited, via the ligase complex rather than via the histone. DIRECTION IS DISEASE-DEPENDENT AND THIS GRAPH'S CASE IS THE AWKWARD ONE. In AML [E1] and neuroblastoma [E3] losing H2AK119ub is reported as a therapeutic good; but this graph also holds that loss of PRC1.1 subunits (PCGF1, BCOR) causes MENIN-INHIBITOR RESISTANCE by epigenetic reactivation of non-canonical menin targets including MYC in KMT2A-rearranged AML [E4]. A USP7 inhibitor phenocopies PRC1.1 loss, so it is predicted to ANTAGONISE a menin inhibitor in the index disease while being beneficial in the same disease as a single agent - and the two-drug experiment that would settle it has never been run. AND THERE IS A THIRD DIRECTION PROBLEM ALREADY IN THIS ENTRY: if USP7 is a DOT1L target gene, a DOT1L inhibitor lowers USP7 and therefore lowers H2AK119ub, which runs OPPOSITE to the genome-wide H2AK119ub INCREASE reported after DOT1L inhibition in MLL-fusion leukaemia [E3 counter]. Two of this graph's chromatin arms make opposite predictions about the same mark under the same drug. PROVENANCE NOTE, because it is the third stale-inventory case in this cycle: my dispatch brief and cycle 52's king-log next-step 5 both state 'USP7 has no entry', traced to a gap-marker clause inside gene-trim27's title claiming USP7 was 'named once in this graph and never written'. THIS ENTRY HAS EXISTED SINCE CYCLE 26 with nine evidence items and already named H2AK119ub, H3K27ac and H3K79 in its body. The brief asked me to report the trim27 sentence as an open defect; IT HAD ALREADY BEEN FIXED before I read it - commit 4db6b99, 'A gap-marker in a title outlived its gap, and I quoted it into a next-steps list', and gene-trim27 now states plainly that both halves were false. Recorded here rather than re-reported, because the useful part is the shape: an inventory sentence in prose does not go stale visibly, and two documents quoted it after it stopped being true.
gene synonyms
[
  "HAUSP"
]
tool compounds
[
  "XL177A (covalent USP7 inhibitor, used in the neuroblastoma study PMID 41855552)"
]
direction of effect is disease dependent
{
  "neuroblastoma": "PRC1.1 removal differentiates the tumour - USP7 inhibition helps",
  "aml_mll_af9_untreated": "USP7 inhibition delays leukaemia (PMID 34113809)",
  "kmt2a_r_aml_under_menin_inhibitor": "PRC1.1 loss is a RESISTANCE route (PMID 39158067) - USP7 inhibition predicted to antagonise"
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -0.7937,
  "chronos_mean_other_leukaemia": -0.7528,
  "difference": -0.041,
  "pan_essentiality_all_lines": -0.8466,
  "selectivity_rank": 4242,
  "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
{
  "subcellular": "FOUR COMPARTMENTS, all experimental. UniProt Q93009 (fetched 2026-09-21) annotates Nucleus, Cytoplasm, 'Nucleus, PML body' and Chromosome, every one at ECO:0000269. A USP7 inhibitor reaches all four.",
  "subnuclear": "MICROPUNCTATE NUCLEOPLASM, with a MINORITY pool in PML nuclear bodies (ND10). The founding description is explicit that only SOME of the larger discrete foci colocalise with PML [E_q71_loc1]. USP7 also carries a Chromosome annotation, which is the pool this entry's PRC1.1 argument actually needs and which is not the PML-body pool.",
  "two_pools_two_arguments": "This entry's therapeutic argument is about holding PRC1.1 together on chromatin. The best-imaged subnuclear pool of USP7 is in PML bodies, which is a different compartment with a different partner set and, in leukaemia, an inhibitory relationship [E_q71_ct1]. Do not let the imaged pool stand in for the argued pool.",
  "explicit_no": "NOT MEASURED IN ANY KMT2A-REARRANGED CELL. UNPUBLISHED rather than not-applicable.",
  "how_shown": "immunofluorescence in a herpesvirus ND10 system (PMID 9034339); UniProt Q93009 SUBCELLULAR LOCATION comment, fetched 2026-09-21",
  "systems": "human non-haematopoietic cells for the imaging; BCR-ABL CML for the compartment-conditional activity; no KMT2A-rearranged system",
  "shown_for_fusion_localisation": "no",
  "node": "localization-pml-nuclear-bodies-pmlrara"
}
cell type
{
  "evidence_class_key": "concept-expressed-required-maintains-three-evidence-classes",
  "identifier_note": "USP7 (Q93009, 1102 aa), entry name UBP7_HUMAN, synonym HAUSP. The 1997-2014 literature that carries most of the localisation and leukaemia biology calls it HAUSP, so an absence claim keyed to USP7 alone would miss it - both papers cited here are indexed under HAUSP in their titles or abstracts. USP7 is not USP47, its closest functional partner in some contexts.",
  "expressed_in": {
    "class": 1,
    "summary": "ubiquitous; a broadly required deubiquitinase, pan-essentiality -0.8466 across 1,208 DepMap lines."
  },
  "required_in": {
    "class": 2,
    "claim": "In BCR-ABL CHRONIC MYELOID LEUKAEMIA, USP7/HAUSP's activity toward PTEN is compartment-conditional: BCR-ABL activates HAUSP to exclude PTEN from the nucleus everywhere EXCEPT the leukaemic stem cell compartment, where high PML inhibits it [E_q71_ct1]. This is the only published cell-type-conditional USP7 result in any leukaemia and it is in the wrong disease.",
    "pmids": [
      "24317448"
    ],
    "system_named_in_the_claim": "BCR-ABL-positive chronic myeloid leukaemia - a different fusion and a different lineage",
    "nodes": [
      "cell-state-stem-like-self-renewal"
    ]
  },
  "the_selectivity_and_what_it_does_not_show": {
    "claim": "Rank 4,242 of 17,728, difference -0.0410 - nominally the selective half, but the difference is trivial next to a pan-essentiality of -0.8466. USP7 is required almost everywhere and marginally more so in the KMT2A arm.",
    "floor_check": "NOT floor-compressed: both arms sit near -0.75 to -0.79, well short of the -2 floor, so there is dynamic range and the small difference is a small difference rather than an uninterpretable one. This is the honest middle case between the floor-compressed and the non-dependency categories."
  },
  "context": "BCR-ABL CML for the compartment-conditional requirement; this project's DepMap scan for the selectivity"
}
cell state
{
  "evidence_class_key": "concept-expressed-required-maintains-three-evidence-classes",
  "maintains_state": {
    "class": 3,
    "states": [
      "cell-state-stem-like-self-renewal"
    ],
    "claim": "In BCR-ABL CML, the LEUKAEMIC STEM CELL state is where USP7/HAUSP is switched OFF rather than on: high PML in that compartment inhibits HAUSP activity toward PTEN, so PTEN stays nuclear there while it is excluded elsewhere [E_q71_ct1]. If that compartment behaviour generalises, a USP7 inhibitor would be doing least where the disease-propagating cells are.",
    "pmids": [
      "24317448"
    ],
    "system_named_in_the_claim": "BCR-ABL-positive chronic myeloid leukaemia leukaemic stem cells"
  },
  "confidence": "moderate",
  "caveats": "The only compartment-resolved USP7 result in leukaemia is in a different disease with a different fusion. Nothing has been measured in a KMT2A-rearranged cell, and the generalisation to an infant B-ALL propagating cell is an inference this entry flags rather than makes. Confidence is moderate on system distance, not on execution.",
  "counter_evidence": {
    "note": "Reading USP7 as a stem-cell-compartment target inverts what the CML result actually says. The paper's finding is that HAUSP is INACTIVE toward PTEN in the LSC compartment; the authors' therapeutic framing is about eradicating minimal residual disease by exploiting that, not about inhibiting HAUSP there. An entry that cites this as support for USP7 inhibition in stem-like cells would be arguing against its own citation.",
    "citation": {
      "title": "BCR-ABL disrupts PTEN nuclear-cytoplasmic shuttling through phosphorylation-dependent activation of HAUSP.",
      "authors": "Morotti A, Panuzzo C, Crivellaro S, Pergolizzi B, Familiari U, Berger AH, Saglio G, Pandolfi PP",
      "year": 2014,
      "venue": "Leukemia",
      "pmid": "24317448",
      "doi": "10.1038/leu.2013.370",
      "type": "paper"
    }
  }
}
depmap release backfill
{
  "open_targets_dataversion": "26.06",
  "note": "Added by Q71 per brief rule 83; Open Targets dataVersion fetched live 2026-09-21 as year 26 month 06. Scan denominator for this gene: n_all_lines 1208."
}
How this could bite us back
  • A USP7 inhibitor combined with a menin inhibitor is predicted to be ANTAGONISTIC, because it removes the PRC1.1 machinery through which the menin inhibitor silences its targets. This is the clearest example in this beat of a combination that looks rational from one paper's abstract and is contraindicated by another's.
  • USP7 inhibition stabilises p53, so a single-agent response in a TP53-wild-type patient cannot be attributed to the PRC1.1 mechanism without a TP53-null control - and the p53 route also carries its own toxicity profile.
  • Removing PRC1.1 pharmacologically phenocopies a documented menin-inhibitor resistance lesion, so a USP7 inhibitor could select for resistance even where it produces an initial response.
  • USP7 has many substrates beyond Polycomb; the target's tractability is not the same as the mechanism's selectivity, and a clean chemical probe does not make a selective biological intervention.
  • USP7 has a pan-essentiality of -0.8466 across 1,208 DepMap lines and a KMT2A-versus-other difference of only -0.0410. The available tool compounds therefore remove a broadly required deubiquitinase for a 0.04 Chronos advantage - the therapeutic index is set by the shared requirement, not by the difference.
  • The compartment argument may run backwards. In CML, HAUSP is inhibited by PML precisely in the leukaemic stem cell compartment (PMID 24317448). If the same holds here, a USP7 inhibitor would be least active in the cells that cause relapse.
  • The imaged pool and the argued pool are different. USP7's best-characterised subnuclear location is a minority pool in PML bodies (PMID 9034339), while this entry's PRC1.1 rationale needs the chromatin pool. A localisation-based argument built on the wrong pool would look well-cited and be wrong.
  • The mark and the therapy point in opposite directions inside this graph. A USP7 inhibitor lowers H2AK119ub by disassembling PRC1.1 (E1), and losing PRC1.1 subunits causes menin-inhibitor resistance through reactivation of non-canonical menin targets including MYC (E4). Adding a USP7 inhibitor to a menin inhibitor is therefore predicted to be antagonistic in the exact disease both are aimed at, and no combination data exist.
Still open
  1. Does a USP7 inhibitor antagonise a menin inhibitor in a KMT2A-rearranged cell? The three papers cited here make that prediction jointly and none of them tests it; a two-drug matrix with H2AK119ub CUT&RUN would settle it.
  2. How much of the USP7 anti-leukaemic effect survives TP53 loss quantitatively? The published exclusion is qualitative, and the answer determines whether this target is usable at all in a TP53-wild-type infant disease.
  3. Does acute chemical disassembly of PRC1.1 phenocopy PCGF1 or BCOR knockout transcriptionally? If it does not, the resistance concern above is weaker than it looks; if it does, the combination is contraindicated.
  4. Is USP7 a DOT1L target gene in leukaemia as it appears to be in GCB-DLBCL? If so, a DOT1L inhibitor lowers PRC1.1 integrity by a second, indirect route that runs opposite to the direct H3K79me-PRC1 antagonism.
  5. What fraction of USP7 is chromatin-bound versus PML-body-associated versus free nucleoplasmic in a KMT2A-rearranged cell? The PRC1.1 argument in this entry needs the first number and nobody has measured any of them in this disease.
  6. Does the PML-inhibits-HAUSP relationship documented in CML leukaemic stem cells hold in KMT2A-rearranged cells? One PML immunoblot across sorted fractions would say whether the USP7 rationale survives in the compartment that matters.
  7. How much of the USP7 anti-leukaemic effect runs through LRRK2 rather than PRC1.1? PMID 40907896 reports a PARTIAL rescue by LRRK2 overexpression and does not quantify the fraction. With the TP53 exclusion also qualitative, this entry now carries two unquantified rival mediators for one drug effect, and an epistasis experiment - USP7 inhibition in cells with LRRK2 restored and PCGF1 intact versus deleted - would separate them.
  8. Does a USP7 inhibitor lower H2AK119ub in a KMT2A::AFF1 B-lymphoid cell? Every H2AK119ub measurement supporting this entry is AML, neuroblastoma or non-leukaemic, and the index lineage has none.
  9. Does the SAGA DUB module erase the same H2A ubiquitin that PRC1.1 deposits? gene-usp22, amended in the same tranche, cites the founding characterisation naming histone H2A as well as H2B as substrates of the ATXN7L3/USP22/ENY2 module - so this graph's polycomb arm and its SAGA arm may converge on one mark and nobody has measured the two together.
Every source behind this entry (10)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    USP7 and TRIM27 are integral components of PRC1.1; PRC1.1 is the predominant Polycomb complex co-precipitating with USP7; USP7 inhibition causes PRC1.1 disassembly and loss of chromatin binding with reduced H2AK119ub and H3K27ac, impairs AML proliferation in vitro independently of the USP7-MDM2-TP53 axis, and delays MLL-AF9-induced leukaemia in vivo in human xenografts.

    The USP7-TRIM27 axis mediates non-canonical PRC1.1 function and is a druggable target in leukemia. · Maat H, Atsma TJ, Hogeling SM, Rodriguez Lopez A, Jaques J, Olthuis M, et al. · iScience 2021
    PMID 34113809DOI
    What this evidence cannot show

    The MDM2-TP53 exclusion is an exclusion, not a quantification - 'also independent of' does not say what fraction of the effect is p53-independent. The in vivo result is a delay in a xenograft, not a cure. No lymphoid or infant KMT2A::AFF1 model. USP7 inhibitors available in 2021 had substantial off-target liability, so the chemistry is a probe rather than a drug.

    Evidence that argues against this (1)
    • The direction is incompatible with the menin-inhibitor resistance literature in the same disease: loss of PRC1.1 subunits confers menin-inhibitor RESISTANCE with MYC reactivation, so pharmacologically disassembling PRC1.1 in a KMT2A-rearranged cell is the same lesion as the resistance route. 'USP7 inhibition delays MLL-AF9 leukaemia' and 'PRC1.1 loss rescues MLL-fusion leukaemia from menin inhibition' are both on record and have never been run in the same experiment. Zhou X et al. 2024
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Combined quantitative proteomics, transcriptomics and epigenomics define USP7 as a control hub linking genome regulation and H2AK119ub1 deposition by non-canonical PRC1 complexes: USP7 strongly stabilises ncPRC1.6 and, to a lesser extent, ncPRC1.1, and USP7-dependent changes in H2AK119ub1 are uncoupled from H3K27me3.

    USP7 regulates the ncPRC1 Polycomb axis to stimulate genomic H2AK119ub1 deposition uncoupled from H3K27me3. · Sijm A, Atlasi Y, van der Knaap JA, Wolf van der Meer J, Chalkley GE, Bezstarosti K, et al. · Sci Adv 2022
    PMID 36332031DOI
    What this evidence cannot show

    Not leukaemia and not a KMT2A-rearranged background. The paper's strongest USP7-ncPRC1 link is to PRC1.6, not to the PRC1.1 flavour this project needs; the PRC1.1 stabilisation is explicitly the weaker of the two. USP7 also represses AUTS2, adding a third route to H2AK119ub that is not PRC1.1 at all.

    Evidence that argues against this (1)
    • Maat et al. report PRC1.1 as the PREDOMINANT Polycomb complex in the USP7 interactome, which sits awkwardly against ncPRC1.6 being the more strongly stabilised complex. Interactome abundance and stabilisation magnitude are different measurements, so the two are reconcilable in principle, but 'USP7 inhibition is a PRC1.1-selective intervention' is not supported by either paper on its own. Maat H et al. 2021
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    A CORUM-plus-DepMap complex-level dependency analysis identifies PRC1.1 as selectively enriched in neuroblastoma, USP7 as the deubiquitinase most strongly co-dependent with PRC1.1, and the USP7 inhibitor XL177A as reducing neuroblastoma growth in cells and animals with reduced H2AK119Ub deposition on chromatin and destabilisation of non-canonical PRC1 by proteolysis.

    Inhibition of USP7 Destabilizes the Noncanonical PRC1.1 Complex and Induces Neuroblastoma Differentiation. · Cmarik EA, Wahi A, Chandekar SS, Otchere SK, Sturgis JE, Salhotra S, et al. · Mol Cancer Res 2026
    PMID 41855552DOI
    What this evidence cannot show

    Neuroblastoma, a solid paediatric tumour of the developing peripheral nervous system - a different lineage and a different oncogenic driver. Cited here for the chemistry (XL177A is a covalent USP7 inhibitor with a cleaner profile than earlier tools) and for the replication that USP7 inhibition lowers H2AK119Ub, not for the therapeutic conclusion. DepMap co-dependency is a correlation across cell lines and is subject to the lineage-split and z-versus-effect-size caveats this graph applies to every DepMap number.

    Evidence that argues against this (1)
    • The therapeutic inference does not transfer: in neuroblastoma PRC1.1 is the machine holding the differentiation block, so removing it differentiates the tumour; in KMT2A-rearranged AML under a menin inhibitor, PRC1.1 is the machine the DRUG uses, so removing it produces resistance. The same molecular action has opposite therapeutic signs in the two diseases. Zhou X et al. 2024
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    USP7 was identified alongside KDM2B, RING1 and PCGF1 in an EZH2-inhibition-anchored genome-wide CRISPR interference screen in GCB-DLBCL, and USP7 is a potential direct target of DOT1L whose downregulation was associated with increased EZH2-inhibitor sensitivity.

    DOT1L Shapes ncPRC1-Target Gene Repression to Maintain Germinal Center B Cell Identity of Diffuse Large B cell Lymphoma. · Gobel C, Niccolai R, Gregoricchio S, de Groot MHP, Kneefel SFK, Kreft M, et al. · Blood 2026
    PMID 42213647DOI
    What this evidence cannot show

    GCB-DLBCL rather than leukaemia; screen anchored on EZH2 inhibition, not on menin or DOT1L inhibition. 'Potential direct target of DOT1L' is the authors' own hedged wording. Read at abstract level. Cited as the fourth independent group to place USP7 inside the ncPRC1 axis, which is why it carries weight despite the indirect setting.

    Evidence that argues against this (1)
    • If USP7 is a DOT1L target gene, then a DOT1L inhibitor should lower USP7 and therefore lower PRC1.1 integrity and H2AK119ub - the opposite of the genome-wide H2AK119ub INCREASE that Neville et al. report after DOT1L inhibition in MLL-fusion leukaemia. Either the USP7 route is not operative in leukaemia or it is outweighed by the direct H3K79me-PRC1 antagonism. Neville D et al. 2026
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Loss of PRC1.1 subunits such as PCGF1 or BCOR leads to menin-inhibitor resistance by epigenetic reactivation of non-canonical menin targets including MYC, in human and mouse KMT2A-rearranged AML models.

    Epigenetic regulation of noncanonical menin targets modulates menin inhibitor response in acute myeloid leukemia. · Zhou X, Zhang L, Aryal S, Veasey V, Tajik A, Restelli C, et al. · Blood 2024
    PMID 39158067DOI
    What this evidence cannot show

    Genetic loss of a subunit is not the same lesion as pharmacological destabilisation of the assembled complex - a USP7 inhibitor removes the complex acutely and reversibly, a knockout removes it permanently, and the transcriptional consequences need not match. No patient data. Myeloid only.

    Evidence that argues against this (1)
    • Maat et al. report the opposite therapeutic direction for pharmacological PRC1.1 disruption in the same fusion background - USP7 inhibition delayed MLL-AF9 leukaemia in vivo - so it is not established that acute chemical PRC1.1 disassembly phenocopies subunit knockout. That is precisely the experiment this entry argues has to be done before the combination is contemplated. Maat H et al. 2021
  6. E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProt reviewed human record: USP7 is Q93009, entry name UBP7_HUMAN, primary gene name USP7, synonym HAUSP, 1,102 aa, recommended protein name 'Ubiquitin carboxyl-terminal hydrolase 7'.

    What this evidence cannot show

    Database fact fetched 2026-09-20; exactly one reviewed row returned and gene_primary matched the symbol asked for. The alias HAUSP is historically dominant in the p53/MDM2 literature, so a symbol-only search of older work under 'USP7' will under-retrieve - the alias has to be searched as well.

    Evidence that argues against this (1)
    • The record's own alternative names ('Herpesvirus-associated ubiquitin-specific protease') point at the protein's first-described function, which has nothing to do with Polycomb. A reader checking 'is this the PRC1.1 deubiquitinase?' against the protein name gets no confirmation, which is brief rule 65's failure mode in its benign form: the name is not wrong, it is simply silent about the function under discussion. Maat H et al. 2021
  7. E_q71_loc1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    In HUMAN cells, the 135 kDa cellular protein bound by the herpes simplex virus immediate-early protein Vmw110 was identified as a novel ubiquitin-specific protease (USP7/HAUSP), and it is DISTRIBUTED IN THE NUCLEUS IN A MICROPUNCTATE PATTERN with a limited number of larger discrete foci, SOME OF WHICH CO-LOCALIZE WITH PML IN ND10 - so the PML-body pool is explicitly a subset of a broadly micropunctate nucleoplasmic distribution, not the protein's address.

    PMID 9034339DOI
    What this evidence cannot show

    Abstract read in full; full text not read. Human non-haematopoietic cells in a herpesvirus system - not leukaemia, not KMT2A-rearranged (brief rule 73). The wording matters and is preserved here: 'some of which co-localize' is a minority claim, and this graph has twice had a partial colocalisation read back as an address. UniProt Q93009 independently annotates Nucleus, Cytoplasm, 'Nucleus, PML body' and Chromosome, all at ECO:0000269 (fetched 2026-09-21), which agrees: four compartments, not one.

    Evidence that argues against this (1)
    • The USP7-PML relationship is REGULATORY, not merely positional, and it runs in the direction that limits USP7. In chronic myeloid leukaemia, BCR-ABL phosphorylates HAUSP to activate it and drive PTEN out of the nucleus - but this does NOT happen in the leukaemic stem cell compartment, because high PML there is a potent inhibitor of HAUSP activity toward PTEN. So PML both colocalises with USP7 and suppresses it, and the compartment where a USP7-directed drug would matter most is the compartment where PML has already switched it off. Morotti A et al. 2014
  8. E_q71_ct1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In HUMAN AND MURINE BCR-ABL-POSITIVE CHRONIC MYELOID LEUKAEMIA, USP7/HAUSP activity is cell-type conditional: BCR-ABL physically interacts with and tyrosine-phosphorylates HAUSP to enhance HAUSP-induced deubiquitination of PTEN and favour PTEN's nuclear exclusion, and this PTEN delocalisation DOES NOT OCCUR IN THE LEUKAEMIC STEM CELL COMPARTMENT because of high PML levels there, PML being a potent inhibitor of HAUSP activity toward PTEN.

    BCR-ABL disrupts PTEN nuclear-cytoplasmic shuttling through phosphorylation-dependent activation of HAUSP. · Morotti A, Panuzzo C, Crivellaro S, Pergolizzi B, Familiari U, Berger AH, Saglio G, Pandolfi PP · Leukemia 2014
    PMID 24317448DOI
    What this evidence cannot show

    Abstract read in full; full text not read. BCR-ABL-driven CML - a DIFFERENT fusion, a different lineage and a different disease from KMT2A-rearranged infant B-ALL (brief rule 73). It is cited here because it is the only published result that makes a USP7 claim CONDITIONAL ON A HAEMATOPOIETIC COMPARTMENT, which is exactly the evidence class this dimension asks for; it is NOT evidence about KMT2A. 'Leukaemic stem cell compartment' is an immunophenotype-plus-function definition in a CML system and does not transfer to an infant B-ALL propagating cell.

    Evidence that argues against this (1)
    • This project's own scan places USP7 in the selective half but with a trivial effect: Chronos -0.7937 in the KMT2A-rearranged arm against -0.7528 in other leukaemias, difference just -0.0410, rank 4,242 of 17,728, pan-essentiality -0.8466. USP7 is a BROADLY required gene - more required across all lines than in either leukaemia arm - so the rank describes a near-zero difference on top of a large shared dependency. The Fanconi comparison in this tranche makes the general point: a strong absolute requirement in blood cells is routinely a lineage or housekeeping property, and only a between-arm difference speaks to the fusion. Ceccaldi R et al. 2012
  9. E_k43_lrrk2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    A SECOND PUBLISHED MEDIATOR FOR THE SAME DRUG EFFECT. USP7 is reported as the deubiquitinase that stabilises LRRK2 by preventing its proteasomal degradation, with both proteins highest in AML among the cancers surveyed, USP7 inhibition suppressing AML growth through LRRK2, and the suppression PARTIALLY RESCUED by LRRK2 overexpression. A partial rescue is the strong form of the claim: it says some of the anti-leukaemic effect this entry attributes to PRC1.1 disassembly runs through a substrate that has nothing to do with Polycomb.

    PMID 40907896DOI
    What this evidence cannot show

    Read from the PubMed abstract only; the full text was not retrieved, so the magnitude of the partial rescue and the cell lines used are not established here (rule 80). This does not displace the PRC1.1 account. Two mediators of one drug effect are the normal situation for a deubiquitinase with many substrates, and the entry's existing TP53 caveat already makes the same structural point. What is new is that there are now TWO named published rivals, not one, and neither has been quantified against the Polycomb route. USP7 has a large substrate set, so the existence of a third, fourth and fifth mediator should be the prior. The value of this item is as a bound on attribution confidence, not as a claim that LRRK2 is the real mechanism. LRRK2 arrives in this graph from an unrelated direction - a single protein-array row naming it an AFF4 interaction partner - and the two facts are independent. Do not read them as corroborating each other.

    Evidence that argues against this (1)
    • The PRC1.1 route is supported by a chromatin-level mechanism with H2AK119ub and H3K27ac readouts and an in vivo MLL-AF9 delay, which is a heavier body of evidence than one 2025 report of a partial rescue. If the two are weighed by evidential mass rather than by recency, PRC1.1 remains the better-supported account and LRRK2 is a caveat on it. Maat H et al. 2021
  10. E_q97_epi1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    Repository audit, not a literature claim: gene-usp7 exists on disk at data/graph/nodes/prc1-myc/gene-usp7.json with nine evidence items and already named H2AK119ub, H3K27 and H3K79 in its serialised body before this amendment, so it was counted by the mark-in-body measure of scripts/check-pi-dimensions.py and NOT by the dedicated-field measure - and the two documents asserting that USP7 has no entry (this beat's dispatch brief and cycle-52 king-log next-step 5) were both wrong, both tracing to a gap-marker clause in gene-trim27's title that commit 4db6b99 has since removed.

    What this evidence cannot show

    An internal repository observation, not an external source, recorded as evidence only because it settles two statements that are themselves in the repository and in a king log. Method: glob for gene-usp7.json under data/graph/nodes/*/, load it, count evidence items, run _EPI_MARK over json.dumps of the entry (returns H2AK119ub, H3K27, H3K79) and _EPI_KEY over the keys of data (returned nothing before this amendment). It does NOT establish that this entry is correct, only that it existed and carried mark content. The gene-trim27 clause was checked on disk and found ALREADY CORRECTED, so it is not reported as an open defect.

    Evidence that argues against this (1)
    • The failure this documents is not the wrong sentence but its propagation: a gap-marker written in a title was quoted into a dispatch brief and a king log after it stopped being true, and neither reader globbed for the file. Cycle 52's own log records the same shape twice more - the USP8/TRIM27 premise, and the standing brief's canonical-id list which was wrong on three ids beneath a note saying exactly that. An inventory claim in prose is not an inventory. Oncofusion Knowledge Graph 2026
Connected to