Oncofusion

start / graph / gene-rbbp8

Genestrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q78-interactome-absent

RBBP8 / CtIP - the only unwritten core-complex partner with a DIRECT-interaction record, and the graph already held its mechanism without holding the protein

also called CtIP · CTIP · RBBP8 · CtBP-interacting protein · Retinoblastoma-binding protein 8 · COM1 · SAE2 homolog

CtIP is the enzyme that chews back the ends of a broken DNA strand so the cell can repair it accurately. LEDGF - the protein that anchors the leukaemia fusion to chromatin - is what carries CtIP to the break, and that is the single best-observed interaction among the partners this graph had never written down.

The same thing, for a scientist

RBBP8/CtIP (Q99708, 897 aa) is a DNA endonuclease that initiates 5'-3' resection at double-strand breaks and thereby commits repair to homologous recombination. It reaches this project through **PSIP1/LEDGF**, not through KMT2A. IntAct holds **four PSIP1-RBBP8 records, none spoke-expanded, and one of them is `MI:0407 direct interaction`** measured in vitro [E1] - the only MI:0407 record anywhere in the twelve-partner absence set, and a stronger claim than most of the pairs this graph has already written. All four come from one publication, Daugaard et al. 2012 [E2], **which this graph has cited three times since cycle 1 without ever creating an entry for the partner it is about** - the gap here was in the node inventory, not in the reading. A 2026 paper then places menin itself at the head of the same chain: menin binds gamma-H2A.X through a Val371-Gln400 motif, recruits SETD2, SETD2 writes H3K36me3, LEDGF reads it and delivers CtIP [E3]. That makes CtIP the effector of a **menin-dependent repair pathway**, which is a claim about this project's drug target and not about a peripheral interactor.

Go deeper
What the four IntAct records actually claim

Read through concept-what-a-curated-interaction-row-actually-claims, the PSIP1-RBBP8 pair is unusually clean 1:

IntAct ACtypedetectionhostexpansion
EBI-15994997physical association (MI:0915)anti-bait co-IPHomo sapiensnone
EBI-15995014physical association (MI:0915)anti-tag co-IPHomo sapiensnone
EBI-15995055direct interaction (MI:0407)anti-bait co-IPin vitronone
EBI-15995092physical association (MI:0915)anti-tag co-IPHomo sapiens (reciprocal, RBBP8 as bait A)none

Two things matter. Nothing here is spoke-expanded, so no curator inferred a pair from an n-ary deposition - against a census null in which 70.6% of this interactome's partner pairs are spoke-only 4. And the fourth row is the reciprocal direction, RBBP8 as the A-side bait, which is the control that most rows in this interactome do not have.

What the table cannot show is that all four are one publication. Four records is four curated experiments, not four independent findings.

The interaction is damage-inducible, and that is the part a static interactome table cannot show

Daugaard et al. report that LEDGF is constitutively chromatin-associated through its PWWP domain but binds CtIP only after DNA damage, so the complex is a switch rather than a standing structure 2. An interaction table records the pair and discards the condition. The same paper shows LEDGF depletion impairs CtIP recruitment to breaks and the resection that follows.

The paralogue arm is the useful control: HDGFRP2/HDGFL2, the other PWWP protein, also recruits RBBP8/CtIP to breaks, but reads the marks of silent chromatin rather than active chromatin 5. So CtIP delivery is not a LEDGF-specific idiosyncrasy - it is a shared PWWP-reader function partitioned by chromatin state, and any agent hitting the LEDGF PWWP cage is likely to hit both.

Why this lands on the menin hypothesis rather than beside it

Lian et al. 2026 report menin directly recognising gamma-H2A.X through a Val371-Gln400 motif, with DNA damage sharply increasing the menin-gamma-H2A.X interaction; menin then recruits SETD2, whose H3K36me3 is read by LEDGF, which recruits CtIP as the effector of end resection and activates RPA2/RAD51 3. Men1 ablation gave spontaneous lung carcinogenesis with severe genomic instability, and restoring H3K36me3 pharmacologically rescued the HR defect.

Two consequences this graph should not assume away.

  1. The motif is not obviously the drug pocket. Every clinical menin inhibitor occupies the central MBM cavity that accepts KMT2A residues 4-15 (interface-menin-kmt2a-mbm1). Val371-Gln400 is a different region. So a menin inhibitor might leave the repair function intact - which would be good news - or might not, and nobody has measured it.
  2. 371-390 overlaps the region IntAct records as sufficient for menin-FANCD2 binding, residues 214-390 (interface-menin-fancd2). Two independent DNA-damage partners mapping to overlapping menin surfaces is either a real hub or a coincidence of coarse deletion mapping, and it is answerable with the constructs both papers already made.
In this disease: nothing, and the absence is bounded

No measurement of the LEDGF-CtIP axis exists in a haematopoietic cell, let alone a KMT2A-rearranged one. (LEDGF OR PSIP1)[tiab] AND (CtIP OR RBBP8)[tiab] returns 5 records (2026-09-21, PubMed): the founding paper, the HDGFRP2 paralogue paper, the CRISPR knockout series, an HPV31 replication study, and the 2026 menin scaffold paper 6. None is in a leukaemia system and none involves a KMT2A fusion. Searched in titles and abstracts only; supporting figures not searched (brief rule 80).

Facts and figures
uniprotQ99708
length AA897
gene symbolRBBP8
protein nameDNA endonuclease RBBP8 (CtBP-interacting protein, CtIP)
uniprot resolutionaccession:Q99708, reviewed human, one record; primary gene name read back as RBBP8 (entry name CTIP_HUMAN); length 897 aa checked against UniProt on 2026-09-21. The entry NAME is CTIP_HUMAN while the primary GENE name is RBBP8 - brief rule 65's shape, and the reason to read gene_primary rather than the entry name.
symbol collision checkdata/symbol-collisions.json (74 collision rows + 7 non-gene antigens) greped 2026-09-21 for RBBP8 and CtIP: no row existed, and none is proposed - the aliases are unambiguous.
baitPSIP1 (O75475)
pathway positioneffector of end resection, downstream of menin -> SETD2 -> H3K36me3 -> LEDGF in the 2026 scaffold model (PMID 42578372)
shown for fusionNO - nobody has looked in any KMT2A fusion, in any lineage. The whole LEDGF-CtIP literature is in solid-tumour, cervical and osteosarcoma lines.
fetched2026-09-21
subnuclear localizationDNA-damage-induced nuclear foci at DSBs in S/G2, and ERCC6L2 phase-separated nuclear condensates; diffuse nucleoplasm otherwise. Condition-dependent, not constitutive.
localization shown for fusionNO - nobody has localised CtIP in any KMT2A-rearranged cell, in any lineage. The bounded query below returned 0. This field is distinct from the entry's existing shown_for_fusion, which is about the PSIP1-CtIP INTERACTION; this one is about where the protein is.
interaction with psip1
{
  "source_pmid": "22773103",
  "source_class": "focused mechanistic study",
  "n_intact_records": 4,
  "n_spoke_expanded": 0,
  "n_source_publications": 1,
  "intact_acs": [
    "EBI-15994997",
    "EBI-15995014",
    "EBI-15995055",
    "EBI-15995092"
  ],
  "interaction_types": {
    "MI:0407 direct interaction": 1,
    "MI:0915 physical association": 3
  },
  "detection_methods": [
    "anti bait coimmunoprecipitation",
    "anti tag coimmunoprecipitation"
  ],
  "reciprocal_direction_curated": true,
  "damage_dependence": "binding to LEDGF is DNA-damage-dependent; the interactome row records no condition",
  "shared_pdb_entries": []
}
localization
{
  "subcellular": "Nucleus, and chromosome/chromatin-bound on damage. UniProt Q99708 carries Nucleus (ECO:0000269 PubMed:10764811, 17965729, 23623683) and Chromosome (six ECO:0000269 citations). PubMed:10764811 is titled for the fact - Nuclear localization and cell cycle-specific expression of CtIP - which is the uncommon case where UniProt's coded citation IS the founding localisation paper (contrast brief rule 91).",
  "subnuclear": "Ionising-radiation-induced nuclear FOCI at double-strand breaks in S/G2, and incorporation into ERCC6L2 liquid-liquid-phase-separated nuclear condensates that govern how far resection runs. Steady-state, undamaged: diffuse NUCLEOPLASM (Human Protein Atlas immunofluorescence, main location Nucleoplasm, reliability Supported). No nucleolar, speckle, PML-body or lamina association is reported.",
  "how_shown": "immunofluorescence and cell-cycle fractionation (PMID 10764811); imaging of condensates plus chromatin-recruitment assays (PMID 40913148); antibody immunofluorescence in HPA's non-haematopoietic panel",
  "tissue_of_evidence": "HPA immunofluorescence is done in non-haematopoietic lines (U-2 OS / A-431 / U-251MG class); the DSB-focus and condensate work is in human cancer lines, not leukaemia",
  "damage_dependence": "This is a CONDITIONAL address. The focal subnuclear pattern exists only after DNA damage in S/G2; in G1 or undamaged cells CtIP is diffusely nucleoplasmic. A record that says only 'nuclear foci' without the condition is wrong for half the cell cycle.",
  "node": "localization-condensate-evidence-standards"
}
cell type
{
  "haematopoietic_answer": "EARLY, PROLIFERATING B CELLS. Conditional Ctip deletion in mice blocks early B-cell proliferation and development; naive (resting) B cells tolerate CtIP loss, and so do G1-arrested B cells and V(D)J recombination itself (PMID 31097467).",
  "why_this_matters_here": "This is the index disease's lineage. The graph reached CtIP through a PSIP1/LEDGF interaction curated in solid-tumour and in-vitro systems; the one dedicated haematopoietic genetic experiment is in B-lineage development, which is closer to infant KMT2A::AFF1 than anything in the interaction literature.",
  "expression_atlas": "HPA Blood Atlas: Immune cell enhanced, detected in many immune cell types, highest listed non-classical monocyte 37.0 nTPM; RNA tissue specificity low.",
  "context": "murine conditional knockout (brief rule 73 - this is mouse B-cell development, not human infant KMT2A::AFF1)"
}
cell state
{
  "state": "PROLIFERATION / S-G2, not lineage. CtIP is required in proliferating early B cells and dispensable in G1-arrested and naive B cells (PMID 31097467); the subnuclear focal pattern is likewise S/G2-restricted (PMID 10764811).",
  "consequence_for_this_graph": "Any proposal to combine a menin inhibitor with a genotoxic in KMT2A-r disease inherits a cell-cycle-conditional requirement: a quiescent persister (see cell-state-relapse-initiating-persister) is precisely the state in which CtIP is least needed."
}
localization bounded absence
{
  "date": "2026-09-21",
  "tool": "NCBI E-utilities esearch, db=pubmed",
  "shape": "gene clause AND localisation clause AND KMT2A/MLL-fusion clause, one [tiab] tag per term (brief rule 94), every cytogenetic term quoted (brief rule 95)",
  "querytranslation_fusion_clause": "\"KMT2A\"[Title/Abstract] OR \"MLL\"[Title/Abstract] OR \"MLL1\"[Title/Abstract] OR \"MLL-AF4\"[Title/Abstract] OR \"MLL-AF9\"[Title/Abstract] OR \"MLL-ENL\"[Title/Abstract] OR \"KMT2A-rearranged\"[Title/Abstract] OR \"MLL-rearranged\"[Title/Abstract] OR \"t 4 11\"[Title/Abstract] OR \"t 9 11\"[Title/Abstract] OR \"MLL fusion\"[Title/Abstract]",
  "count": 0,
  "controls": {
    "fusion_clause_alone": 8743,
    "gene_clause_alone": 24007
  },
  "reading": "Zero in TITLES AND ABSTRACTS only; supporting figures were not searched (brief rule 80)."
}
How this could bite us back
  • This is the essential-function liability the LEDGF axis already carries, now with a named effector. An agent that strips LEDGF from chromatin removes CtIP delivery in every cell in a child who is simultaneously receiving etoposide and anthracyclines.
  • The MI:0407 label makes this the strongest-looking row in the absence set, and it is still one paper. A reader ranking partners by interaction type would put RBBP8 above menin-KMT2A pairs supported by several independent groups.
  • If menin inhibition does impair HR, the combination logic inverts in two directions at once: menin inhibitors would sensitise to PARP inhibition and to genotoxics, which is an opportunity, and would also add genomic instability in a largely curable paediatric disease, which is not.
Still open
  1. Does a clinical menin inhibitor (revumenib, ziftomenib, bleximenib) impair the menin -> SETD2 -> LEDGF -> CtIP resection chain? The drugs occupy the MBM cavity and the gamma-H2A.X motif is reported at Val371-Gln400, so the prediction is 'no' - but if it is 'yes', every patient on a menin inhibitor who also receives a genotoxic carries an unmeasured HR defect.
  2. Do menin residues 371-390 mediate both the gamma-H2A.X contact and the FANCD2 contact? The two mappings overlap, neither has a structure, and a co-IP panel with the deletion constructs both papers already made would settle it.
  3. Is CtIP recruitment to breaks altered in a KMT2A::AFF1 blast, where LEDGF is already engaged by the fusion's IBD contact? In principle the fusion and CtIP compete for the same LEDGF molecule; nobody has tested whether they compete in fact.
  4. CtIP's requirement in B cells is a PROLIFERATION requirement, not a lineage one (PMID 31097467). Does the quiescent relapse-initiating persister in KMT2A::AFF1 therefore lose a CtIP dependency that the bulk blast has - i.e. would a resection-directed combination spare exactly the cells that matter?
Every source behind this entry (10)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    IntAct's REST record holds four PSIP1-RBBP8 interaction records, none carrying an expansion method, of which EBI-15995055 is typed MI:0407 direct interaction with host 'In vitro' and EBI-15995092 is the reciprocal direction with RBBP8 as the A-side bait; all four derive from PMID 22773103.

    What this evidence cannot show

    One database on one day, fetched 2026-09-21 through the REST endpoint rather than PSICQUIC (brief rule 93). Four records from ONE publication is not four independent observations - IntAct curates one experiment per record, and a paper contributes as many records as it ran experiments. The MI:0407 call is a curator's reading of an in vitro binding experiment, not an independent structural measurement; no PDB entry contains both chains.

    Evidence that argues against this (1)
    • The census that supplies the spoke-expansion null also warns that ranking partners by row count overstates independent support, because neither of this project's interactome analyses deduplicates rows to publications. Four records here are one paper. EMBL-EBI IntAct 2026
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In human cancer cell lines (not haematopoietic, no KMT2A fusion), LEDGF/p75 promotes homologous recombination: its depletion impairs CtIP recruitment to double-strand breaks and CtIP-dependent end resection, and LEDGF binds CtIP in a DNA-damage-dependent manner while remaining constitutively chromatin-bound through its PWWP domain.

    LEDGF (p75) promotes DNA-end resection and homologous recombination. · Daugaard M, et al. · Nat Struct Mol Biol 2012
    PMID 22773103DOI
    What this evidence cannot show

    Human cancer cell lines only; no haematopoietic cell, no KMT2A fusion, no primary sample (brief rule 73). Europe PMC returns no PMCID for this record, so the cell systems and quantitation were not read from full text - abstract only. This entry does not restate the mechanism, which mechanism-ledgf-dna-end-resection-and-homologous-recombination already holds; it supplies the partner node that mechanism had no target for.

    Evidence that argues against this (1)
    • A later CRISPR knockout series found untreated LEDGF-null cells already carry increased DNA fragmentation with elevated gamma-H2AX and BRCA1, which a pure CtIP-recruitment defect does not predict and which suggests the phenotype is broader than CtIP tethering. Liedtke V et al. 2021
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In mouse Men1-ablated cells and human cell lines (no leukaemia system reported in the abstract), menin directly recognises histone gamma-H2A.X through a Val371-Gln400 motif, DNA damage markedly increases that interaction, and menin then recruits SETD2 to write H3K36me3 which LEDGF reads to recruit CtIP as the end-resection effector, activating the RPA2/RAD51 pathway; Men1 ablation produced spontaneous lung carcinogenesis with severe genomic instability, and pharmacological restoration of H3K36me3 rescued the HR defect.

    PMID 42578372DOI
    What this evidence cannot show

    Abstract read this session; full text not retrieved. Published 2026 and not yet independently replicated; new to this graph (checked with scripts/check-citation-reuse.py). Lung carcinogenesis and cell lines - not haematopoietic, no KMT2A fusion (brief rule 73). The Val371-Gln400 motif is stated in the abstract without the mapping method, and this entry does not assert that it is or is not the menin-inhibitor pocket.

    Evidence that argues against this (1)
    • An independent interface entry in this graph records IntAct's curation of menin residues 214-390 as sufficient for FANCD2 binding from a 2003 co-IP, so the same broad menin surface has already been assigned to a different DNA-damage partner; coarse deletion mapping cannot distinguish a hub from an artefact of construct boundaries. EMBL-EBI IntAct 2026
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Across the six core proteins, 317 of 449 IntAct partner pairs (70.6%, Wilson 66.2-74.6%) rest only on spoke-expanded rows, which is the null against which the PSIP1-RBBP8 pair's four non-spoke records should be read.

    What this evidence cannot show

    The census is this project's own analysis (ana-two-thirds-of-the-curated-interactome-is-a-row-nobody-observed) and its caveats apply, including that pairs are not deduplicated to publications. The endpoint was re-queried on the same day as this entry.

    Evidence that argues against this (1)
    • Re-derived at the same endpoint on the same date, the null moves outside its own stated interval. Complete paginated fetches of all six baits on 2026-09-21 return 1,092 rows carrying only 791 DISTINCT interaction ACs - 301 rows are one interaction returned twice because both partners are core baits. Deduplicated on interaction AC the census is 268 of 452 pairs spoke-only, 59.3%, against a claimed 70.6% whose Wilson interval is 66.2-74.6%. Concatenating without deduplication gives 66.7%; adding a human protein-protein restriction gives 65.3%. So the 11-point swing produced by one bookkeeping choice is three times the sampling width the interval reports, and the census's pair count (452) reproduces while its record count (908 claimed, 791 or 1,092 here) does not. The PSIP1-RBBP8 rows are still non-spoke on any of these readings; the null they are read against is softer than stated. EMBL-EBI IntAct 2026
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In human U2OS and HeLa cells, the structurally related PWWP protein HDGFRP2/HDGFL2 also recruits RBBP8/CtIP to double-strand breaks and supports DNA damage-induced RPA2 phosphorylation, but binds marks of transcriptionally silent chromatin rather than active chromatin.

    PMID 26721387DOI
    What this evidence cannot show

    Osteosarcoma and cervical carcinoma lines; no haematopoietic system and no KMT2A fusion (brief rule 73). Relevant here as the paralogue control on the delivery function, not as evidence about LEDGF itself.

    Evidence that argues against this (2)
    • Loss-of-function genetics do not support a symmetric shared function, and that undercuts the inference that an agent hitting the PWWP cage hits both. Hdgfrp2 knockout mice 'developed normally and were fertile'; only Psip1/Hdgfrp2 DOUBLE knockouts die, at about embryonic day 13.5, and the lesion is a ventricular septal defect - a cardiac-development phenotype, not a repair one. In the integration assay the same asymmetry holds: HRP2-depleted cells support normal HIV-1 infection and defects appear only once Psip1 is already deleted. So HRP2 is a buffer that becomes visible when LEDGF is gone, rather than a co-equal CtIP carrier, and the biochemistry in this item is not matched by a single-mutant phenotype. Wang H et al. 2015
    • The second half of the asymmetry, in the assay where the two PWWP proteins are best compared: HRP2 depletion alone leaves HIV-1 integration efficiency and specificity intact, disrupting Hdgfrp2 in a Psip1 knockout background adds further defects, and ectopic expression of EITHER protein largely restores them - so they are interchangeable when supplied and not equivalent when removed. Double-knockout cells still retain residual targeting to genes, which means neither PWWP reader is the whole of the addressing function this entry treats as partitioned between them. Wang H et al. 2012
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    As of 2026-09-21, PubMed returns 5 records for the query (LEDGF OR PSIP1)[tiab] AND (CtIP OR RBBP8)[tiab], none of which is in a haematopoietic system or involves a KMT2A fusion.

    What this evidence cannot show

    Titles and abstracts only; supporting figures not searched (brief rule 80), and a resection assay in a leukaemia line could sit in a figure of a paper whose abstract names neither protein. Name list tried: LEDGF, PSIP1, CtIP, RBBP8; the alternate strings COM1, SAE2, p75 and DFS70 were not run. Operand control (brief rule 67): each clause alone returns a large literature, so the five is a real intersection rather than a non-composing query.

    Evidence that argues against this (2)
    • The alias gap this caveat flags was measured and is empty, and the field tag is not applying. Re-run 2026-09-21 with COM1, SAE2, p75, DFS70 and 'lens epithelium-derived growth factor' and 'CtBP-interacting protein' added, the count is still 5 and the PMIDs are identical - so the unrun-string worry does not change the answer here. Separately, PubMed's querytranslation for this exact string contains no [Title/Abstract] term at all: a tag written after a parenthesised OR group is dropped and both clauses expand to [All Fields]. These are all-fields counts, so brief rule 80's full-text bound still applies but the 'titles and abstracts only' label is wrong. NCBI PubMed 2026
    • The bound is drawn around the wrong thing. 'None is in a haematopoietic system or involves a KMT2A fusion' is the right test for a partner that reaches this project through the fusion; this one reaches it through MENIN. The most recent of the five, cited at E3 of this same entry, puts the LEDGF-CtIP step downstream of menin's own gamma-H2A.X recognition and SETD2 recruitment - so the axis is relevant to this graph through the drug target regardless of lineage, and a lineage-scoped absence understates rather than bounds it. What is genuinely absent is any measurement of this chain under a clinical menin inhibitor, in any cell type, and that absence is not what this query tested. Lian A et al. 2026
  7. E7established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProt Q99708 places CtIP in the nucleus and on chromosome/chromatin, and the coded experimental citation behind the nuclear call (PMID 10764811) is itself a dedicated CtIP localisation paper rather than an unrelated study - the case brief rule 91 says is the exception, not the rule.

    What this evidence cannot show

    A database comment fetched on one day. UniProt's own note frames the chromosome call as damage-associated (Associates with sites of DNA damage in S/G2 phase), so the chromatin address is conditional, not constitutive. No haematopoietic system appears anywhere in the coded citation set.

    Evidence that argues against this (1)
    • The record reads as though 'nucleus + chromosome' were a fixed property, and it is not. CtIP's chromatin engagement is set by what else is in the cell: in BCR-ABL-positive leukaemia CtIP is driven into MUTAGENIC homologous recombination, i.e. the same protein on the same chromatin produces a different repair outcome under an oncogene. A compartment record that omits the driver context will mislead anyone reasoning about CtIP in a fusion-driven leukaemia, which is precisely this graph's use case. Carcinogenesis 2011
  8. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    Human Protein Atlas immunofluorescence assigns CtIP a single main subcellular location, Nucleoplasm, with reliability Supported, and the Blood Atlas calls RBBP8 Immune cell enhanced and Detected in many immune cell types.

    What this evidence cannot show

    HPA's subcellular panel is antibody immunofluorescence in non-haematopoietic cell lines and reports the UNDAMAGED steady state, so it cannot see the DSB focal pattern; a Nucleoplasm call is therefore compatible with, not contradictory to, the focus literature. Blood Atlas categories are RNA, not protein.

    Evidence that argues against this (1)
    • A 'Supported' reliability score here measures antibody reproducibility across HPA's own lines, not biological completeness, and the argument cuts against using it as a subnuclear record at all: HPA images undamaged fixed cells, so it is structurally blind to the only subnuclear pattern CtIP is known for. Dedicated imaging under damage finds CtIP in phase-separated nuclear condensates that HPA's panel would score as diffuse nucleoplasm. The right reading is that this row is uninformative about subnuclear state, not that it is evidence of a diffuse one. Nat Cell Biol 2025
  9. E9strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CtIP is required for early B-cell proliferation and development in mice and is dispensable in naive B cells, in G1-arrested B cells and for V(D)J recombination, making the requirement a proliferation-state requirement rather than a lineage one.

    PMID 31097467DOI
    What this evidence cannot show

    Mouse conditional knockout; no human data and no KMT2A-fusion context. Read at abstract level. Brief rule 73: this is murine B-cell development, and the index disease is human infant B-lineage KMT2A::AFF1.

    Evidence that argues against this (1)
    • The B-lineage reading is weaker than it looks, and the paper itself supplies the objection: CtIP is dispensable in naive B cells, in G1-arrested B cells and for V(D)J recombination, so what is required is proliferation, not B-ness. Against a lineage interpretation, CtIP is functionally engaged in MYELOID leukaemia as well, where BCR-ABL routes double-strand breaks through it. A pan-proliferative resection factor that happens to have been knocked out in B cells is not a B-lineage protein. Carcinogenesis 2011
  10. E10moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    CtIP is held in ERCC6L2 liquid-liquid-phase-separated nuclear condensates, and disrupting those condensates exposes CtIP to RNF138-mediated ubiquitination and degradation, which sets the extent of DNA end resection.

    PMID 40913148DOI
    What this evidence cannot show

    The condensates described are ERCC6L2's; CtIP is a client whose stability they control, which is a weaker claim than CtIP forms condensates. Non-haematopoietic cancer lines. Abstract-level reading. Judge against localization-condensate-evidence-standards before treating this as a compartment.

    Evidence that argues against this (1)
    • This should not be recorded as 'CtIP forms condensates'. The condensates are ERCC6L2's, built by ERCC6L2's intrinsically disordered regions; CtIP is a client whose stability they control. A protein with no autonomous compartment-forming capacity is exactly what HPA's diffuse nucleoplasm call predicts, and the two observations are consistent rather than in tension - which means the condensate finding adds a regulator, not an address, to CtIP's record. Human Protein Atlas 2026
Connected to