Oncofusion

start / graph / fusion-set-nup214

Fusionstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q58-set-nup214

SET::NUP214 (TAF-Ibeta-NUP214, SET-CAN), cryptic del(9)(q34.11q34.13) - a HOXA-driving fusion of a histone chaperone to a nucleoporin, with no KMT2A and no menin-binding motif

also called SET-NUP214 · SET::NUP214 · TAF_I-NUP214 · TAF-I-NUP214 · TAF-Ibeta-NUP214 · SET-CAN · del(9)(q34.11q34.13)

A small deletion on chromosome 9 glues a histone-handling protein called SET onto a piece of a nuclear pore protein called NUP214, and the resulting hybrid switches on the HOXA genes in T-cell leukaemia. It reaches those genes without containing any part of KMT2A, the gene this project's main drug target works through.

The same thing, for a scientist

**SET::NUP214 arises in human T-lineage acute lymphoblastic leukaemia, mostly in ADULTS, from a cryptic deletion del(9)(q34.11q34.13) or the balanced t(9;9)(q34;q34)** - two steps from this project's index disease of infant B-lineage KMT2A::AFF1, crossing both lineage and age. The chimera joins almost the whole SET/TAF-Ibeta protein (Q01105, 290 aa; dimerisation domain, earmuff domain, acidic C-terminal tail) to the C-terminal FG-repeat region of NUP214 (P35658, 2090 aa), retaining **42 of NUP214's 44 FG repeats** [E5]. At the transcript level the junction is SET exon 7 to NUP214 exon 18, and because the SET genomic breakpoint lies **downstream of the stop codon in SET exon 8**, the mature fusion mRNA requires alternative splicing; the product is a **140 kDa** protein [E3]. In the founding report SET-NUP214 **binds the promoter regions of HOXA9 and HOXA10 specifically** (not HOXA1, HOXA3 or HOXA11), and siRNA knockdown abolishes HOXA expression, blocks proliferation and induces differentiation in **LOUCY but not in other T-ALL lines** [E1]. Incidence estimates: **4.6% (7/152)** in an Italian adult T-ALL series [E2] and **6% (11/196)** in the French GRAALL 2003/2005 trials, where the fusion is **strikingly gamma-delta-lineage-restricted** and associated with corticosteroid and chemotherapy resistance [E4]. **What makes it matter here: the fusion contains no KMT2A, no CXXC domain and no menin-binding motif, and it activates the same HOXA output.** It is the second instance, alongside CALM-AF10, of a test the coordinator wanted - except that it is not as clean as it looks, because SET::NUP214 recruits **wild-type MLL1** into its nuclear bodies [E6][E7], and wild-type MLL1 binds menin.

Go deeper
The two cell lines the whole mechanistic literature runs on

Of 141 human leukaemia/lymphoma cell lines screened, exactly two express SET-NUP214: the T-ALL line LOUCY and the AML line MEGAL 3. Both carry del(9)(q34.11q34.13) by FISH and array copy-number, both show loss of genomic material between SET and NUP214 by quantitative genomic PCR, both express the 140 kDa fusion protein, and both additionally express a TAF-Ialpha-NUP214 mRNA from the alternative first exon 3.

Every mechanistic result in this entry comes from LOUCY, from MEGAL, or from exogenous expression of the cloned fusion in a non-haematopoietic cell. That is the ceiling on the confidence of anything below.

Why this entry says ADULT T-ALL when the founding paper says paediatric

The founding report 1 is titled for paediatric T-ALL and identified the fusion in 3 of 5 HOXA-high paediatric cases. Every subsequent systematic series is adult: 4.6% of 152 Italian adults 2, 6% of 196 French adults on GRAALL 4. Cellosaurus classifies LOUCY itself as adult T-ALL (CVCL_1380, NCIt C9142 'Adult T acute lymphoblastic leukemia') 8.

The brief that commissioned this beat called SET::NUP214 an adult T-ALL entity, and the literature agrees with that as the centre of gravity - but the founding mechanism paper is paediatric, and both ages appear. Rule 73 therefore requires naming the age in each claim, not in a blanket statement here.

What is NOT established, counted
  • No mouse model and no transformation assay found in this beat. A single-centre series reports the rearrangement in individuals who did not develop leukaemia and concludes it is insufficient in isolation 9. Its abstract is not deposited in PubMed and the paper was not readable through any of the five routes, so the claim is carried at weak from the title alone.
  • No published menin or menin-inhibitor experiment. See ana-set-nup214-menin-bounded-absence.
  • No DOT1L-inhibitor experiment. (SET-NUP214 OR TAF-I-NUP214) AND (pinometostat OR EPZ-5676 OR EPZ004777 OR "DOT1L inhibitor" OR "DOT1L inhibition") returns 0 (PubMed esearch, 2026-09-20), against a positive control of 1 for SET-NUP214 AND DOT1L through the identical shape.
  • No DepMap dependency datum. Neither LOUCY nor MEGAL is present in the DepMap CRISPR set served by the Open Targets Platform GraphQL API (1,258 screens carrying a MEN1 gene effect, 8 of them annotated T-Lymphoblastic Leukemia/Lymphoma, none SET-NUP214-positive; checked 2026-09-20).
Facts and figures
gene5SET
gene3NUP214
cytobanddel(9)(q34.11q34.13); also reported as balanced t(9;9)(q34;q34)
uniprot5Q01105
uniprot3P35658
length5 AA290
length3 AA2090
fusion protein kda140
kmt2a contentnone
cxxc domainnone
menin binding motifnone in the fusion itself; but the fusion recruits wild-type MLL1, which carries MBM1/MBM2
diseases
[
  "adult T-ALL",
  "paediatric T-ALL",
  "AML",
  "acute undifferentiated leukaemia",
  "acute megakaryoblastic leukaemia"
]
breakpoints
[
  {
    "id": "bp-set-exon7-nup214-exon18",
    "gene5_exon_last": 7,
    "gene3_exon_first": 18,
    "population": "LOUCY and MEGAL cell lines; the recurrent patient junction",
    "domains_retained_5": [
      "dimerisation domain",
      "earmuff domain",
      "acidic C-terminal region"
    ],
    "domains_lost_5": [],
    "domains_retained_3": [
      "FG repeats (42 of 44)",
      "C-terminal CRM1-binding region"
    ],
    "domains_lost_3": [
      "seven-bladed beta-propeller (aa 41-404)",
      "coiled coil (aa 680-1209)"
    ],
    "phenotype_notes": "The SET genomic breakpoint lies downstream of the stop codon in SET exon 8, so the mature fusion transcript requires alternative splicing. Both LOUCY and MEGAL additionally express a TAF-Ialpha-NUP214 mRNA from the alternative first exon."
  }
]
frequency
{
  "adult_t_all_pct_italian_series": 4.6,
  "adult_t_all_n_italian": "7/152",
  "adult_t_all_pct_graall": 6,
  "adult_t_all_n_graall": "11/196"
}
cell lines
[
  "LOUCY (T-ALL, CVCL_1380, DepMap ACH-000104)",
  "MEGAL (AML)"
]
co occurring lesions
[
  "PHF6 mutation (association P = .002)",
  "NOTCH1 mutation",
  "JAK1 mutation"
]
identifier resolution
{
  "SET": "Q01105, primary gene name SET, 290 aa; gene_exact:SET AND organism_id:9606 AND reviewed:true returns exactly ONE reviewed record, so the SET trap in this project is bibliographic (PubMed) and not a UniProt resolution trap",
  "NUP214": "P35658, primary gene name NUP214, 2090 aa; gene synonyms CAIN, CAN, KIAA0023 - 'CAN' is unusable as a bare search term",
  "MLL_in_E7": "unresolved; see brief rule 66 (MLL2/MLL4 each name two genes)"
}
name list searched
[
  "SET-NUP214",
  "SET::NUP214",
  "TAF_I-NUP214",
  "TAF-I-NUP214",
  "SET-CAN",
  "del(9)(q34.11q34.13)"
]
How this could bite us back
  • The only pharmacological handle demonstrated on this fusion is CRM1 inhibition, and CRM1/XPO1 is flagged common-essential (Open Targets DepMap isEssential = true, mean gene effect -2.14 across 1,208 lines) - the window is against the most depleted gene in that panel, not against a fusion-specific target.
  • SET-NUP214 nuclear bodies reassemble shortly after CRM1 antagonist is washed out, so a reversible CRM1 inhibitor buys occupancy and not a durable state change.
  • Corticosteroid resistance in 91% and chemotherapy resistance in 100% of SET-NUP214-positive adults means any new agent is tested against a background where the backbone is already failing early, which inflates apparent single-agent activity.
  • Because the fusion retains almost all of SET, any agent targeting the fusion by its SET portion also targets a ubiquitous histone chaperone and PP2A inhibitor.
  • If HOXA activation here depends on wild-type MLL1 recruitment, a menin inhibitor might work - and the same logic says the leukaemia can escape by using the CRM1 route without MLL1, which is exactly what the CRM1-first recruitment order predicts.
Still open
  1. Does a menin inhibitor do anything to LOUCY or MEGAL? Nobody has published the experiment, and because SET::NUP214 recruits wild-type MLL1 the answer is not predictable from the fusion's composition.
  2. Is the gamma-delta lineage restriction a property of the fusion or of the cell it arises in? CALM-AF10 shows the same gamma-delta association, and two unrelated lesions restricted to the same precursor is either a shared cell of origin or a shared programme.
  3. Why does SET-NUP214 knockdown abolish HOXA expression in LOUCY and do nothing in other T-ALL lines? The founding paper reports this as a specificity control; it is equally readable as evidence that the fusion needs a permissive cell state.
  4. Does the balanced t(9;9)(q34;q34) form behave like the deletion form? Every mechanistic experiment in the literature uses the deletion-derived transcript.
Every source behind this entry (9)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In PAEDIATRIC human T-ALL, a cryptic recurrent del(9)(q34.11q34.13) produces a conserved SET-NUP214 fusion, also present in the T-ALL cell line LOUCY; by ChIP in LOUCY the fusion is enriched at HOXA9 and HOXA10 promoters but not HOXA1, HOXA3 or HOXA11, and that enrichment is reversed by SET exon 5 siRNA; siRNA knockdown abolishes HOXA expression, inhibits proliferation and induces differentiation in LOUCY but not in other T-ALL lines.

    PMID 18299449DOI
    What this evidence cannot show

    Paediatric T-ALL plus one cell line. The deletion was found in 3 of 5 HOXA-high patients, so the frequency in this paper rests on very small numbers. The knockdown phenotype is demonstrated in LOUCY only; the paper states explicitly that it was not reproduced in other T-ALL lines, which is the correct control and also the ceiling of the result. Full text read via the PMC HTML route (PMC2343598) after Europe PMC fullTextXML returned HTTP 500.

    Evidence that argues against this (1)
    • The same paper's IP experiments FAILED to demonstrate interaction between SET-NUP214 and the INHAT components SET and PP32, so the attractive model that the fusion poisons the INHAT complex by substituting for wild-type SET was tested by the founding authors and not supported. Histone H3 acetylation at the HOXA locus was also unchanged by knockdown within the experiment's time frame. Van Vlierberghe P et al. 2008
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In ADULT human T-ALL, combined interphase FISH plus multiplex PCR and gene-expression profiling across 152 patients estimated the incidence of TAF_I-NUP214 (SET-NUP214) at 4.6% (7/152); the index cases carried cryptic del(9)(q34) with a gene-expression signature of HOXA and NUP214 upregulation, and the same series also contained CALM-AF10 and MLL translocations.

    PMID 20065082DOI
    What this evidence cannot show

    ADULT T-ALL, multicentre Italian series. The 4.6% combines FISH-detected index cases with cases found by PCR and expression profiling in the remainder, so two ascertainment methods share one denominator. Abstract-level reading for this beat; the numbers were checked against the record fetched from PubMed on 2026-09-20.

    Evidence that argues against this (1)
    • ADULT T-ALL IS TWO STEPS FROM THIS GRAPH'S SUBJECT. The index disease is infant B-lineage KMT2A::AFF1, so every use of SET::NUP214 as a comparator carries an age difference and a lineage difference on top of the fusion difference, and no published experiment separates the three. Oncofusion knowledge graph (internal entry) 2026
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Of 141 human leukaemia and lymphoma cell lines screened, only the T-ALL line LOUCY and the AML line MEGAL express the SET(TAF-Ibeta)-NUP214 fusion transcript; both carry del(9)(q34.11q34.13) by FISH and array copy-number, both express a 140 kDa fusion protein, and the SET exon 7 to NUP214 exon 18 transcript requires alternative splicing because the SET genomic breakpoint lies downstream of the stop codon in exon 8, with the NUP214 breakpoint in intron 17/18.

    PMID 19166587DOI
    What this evidence cannot show

    A cell-line survey, not a patient series, and a 2009 panel - a line added to a repository since would not appear. It establishes model availability, not biology.

    Evidence that argues against this (1)
    • A CELL-LINE PANEL MEASURES WHAT HAS BEEN IMMORTALISED, NOT WHAT OCCURS. Finding the fusion in only 2 of 141 lines is consistent with rarity and equally with the lesion being hard to establish in culture, and this graph already records that the available KMT2A-rearranged panel is nine individuals rather than eleven lines. Oncofusion knowledge graph (internal entry) 2026
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In ADULT human T-ALL on the French GRAALL 2003 and 2005 trials, 11 of 196 patients (6%) carried a SET-NUP214 transcript; 10 of the 11 (91%) were TCR-negative and the fusion was strikingly associated with the TCR gamma-delta lineage; SET-NUP214-positive patients showed corticosteroid resistance in 91% versus 44% (P = .003) and chemotherapy resistance in 100% versus 44% (P = .0001), yet overall outcome did not differ from SET-NUP214-negative patients.

    PMID 24449214DOI
    What this evidence cannot show

    ADULT T-ALL, n = 11 positive patients, so the resistance percentages rest on eleven people and the outcome equivalence is underpowered to exclude a real difference. All but one achieved complete remission and nine were allografted, so the outcome comparison is confounded by transplant. The gamma-delta lineage definition used here includes surface-CD3/TCR-negative cases assigned by rearrangement status rather than by surface phenotype.

    Evidence that argues against this (1)
    • 6% IN TWO ADULT TRIAL COHORTS IS A TREATED-POPULATION FREQUENCY. GRAALL 2003/2005 enrolled by protocol eligibility, so the denominator reflects who was recruited and tested rather than incidence, and the same caution this graph applies to the KMT2A recombinome applies here. Oncofusion knowledge graph (internal entry) 2026
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.review

    SET-NUP214 and DEK-NUP214 conserve virtually the entire NUP214 FG domain - 42 of 44 FG repeats - and in both the NUP214 portion is fused C-terminally to almost the entire peptide sequence of the partner; SQSTM1-NUP214 by contrast preserves only 14 FG repeats and correspondingly induces less nuclear accumulation of protein and polyA+ RNA.

    NUP214 in Leukemia: It's More than Transport. · Mendes A, Fahrenkrog B · Cells 2019
    PMID 30669574DOI
    What this evidence cannot show

    A review, cited here as a map of the domain composition rather than as a primary measurement; the underlying FG-repeat counts come from the primary reports it cites. Full text read via Europe PMC (PMC6356203).

    Evidence that argues against this (1)
    • CONSERVATION OF AN FG DOMAIN IS AN ARGUMENT FROM SEQUENCE, NOT A MEASUREMENT OF FUNCTION. 42 of 44 FG repeats retained tells you the module is present; whether the repeats do the same thing inside SET::NUP214 as inside the nucleoporin requires a functional test, and FG-repeat behaviour is notoriously context- and concentration-dependent. Oncofusion knowledge graph (internal entry) 2026
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In human cells, the phase-separated nuclear bodies of NUP98::HOXA9 and SET::NUP214 promote the condensation of wild-type MLL1 together with CRM1; the bodies are robustly associated with MLL1/CRM1 and co-localise with them on chromatin.

    PMID 37516964DOI
    What this evidence cannot show

    ABSTRACT-LEVEL ONLY: this paper is not in PMC (Europe PMC pmcid null) and was not read in full for this beat, so the cell system, the MLL1 antibody and whether MLL1 recruitment was shown for SET::NUP214 specifically rather than inferred from NUP98::HOXA9 are all unresolved here. The 3D-genome half of the result is attributed in the abstract to NUP98::HOXA9, not to SET::NUP214. Phase-separation claims in this graph carry the standing caveat that 1,6-hexanediol and classical FRAP cannot distinguish LLPS from binding to spatially clustered sites.

    Evidence that argues against this (1)
    • PHASE SEPARATION IS THE CLAIM CLASS THIS GRAPH HOLDS TO THE HIGHEST BAR. Nuclear bodies observed under overexpression, with droplet morphology as the readout, do not establish that condensation is the mechanism rather than a consequence; the entry's own standards note requires endogenous tagging, concentration-matched controls and a separation-of-function mutant before a condensate claim carries weight. Oncofusion knowledge graph (internal entry) 2026
  7. E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    SET-NUP214 interacts with MLL via the SET acidic region of SET-NUP214, and SET-NUP214 and MLL cooperatively enhance the promoter activity of the HOXA10 gene; neither the SET region alone nor the NUP214 region alone is sufficient to enhance that promoter.

    PMID 34320268DOI
    What this evidence cannot show

    ABSTRACT-LEVEL ONLY for this beat. A promoter-reporter assay plus interaction mapping; MLL is not resolved in the abstract to a construct or an accession, and the MLL2/MLL4 alias hazard (brief rule 66) means the bare word MLL in a 2021 paper is not self-resolving. Read here as wild-type MLL1/KMT2A on the strength of the HOXA10 context, and recorded as unresolved in data.identifier_resolution.

    Evidence that argues against this (1)
    • THIS IS THE RESULT THAT WEAKENS THE HYPOTHESIS IT IS USUALLY QUOTED TO SUPPORT, AND IT SHOULD BE READ THAT WAY. SET-NUP214 recruiting wild-type MLL1 means the lesion is NOT menin-free, because wild-type MLL1 carries MBM1 and MBM2 - so 'no menin motif in the fusion' does not make the leukaemia menin-independent, and SET::NUP214 is a WEAKER test of separability than CALM-AF10 rather than a second confirmation. Oncofusion knowledge graph (internal entry) 2026
  8. E8established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Cellosaurus record CVCL_1380 (Loucy) is classified NCIt C9142 'Adult T acute lymphoblastic leukemia' and Orphanet_99861 'Precursor T-cell acute lymphoblastic leukemia', is cross-referenced to DepMap ACH-000104, and is derived from peripheral blood.

    Cellosaurus cell line Loucy (CVCL_1380) · Cellosaurus (SIB) 2026
    What this evidence cannot show

    Fetched live 2026-09-20. A cell-line registry's disease assignment is an annotation, not a re-diagnosis of the original patient, and it disagrees with the 'pediatric' framing of the founding paper's title.

    Evidence that argues against this (1)
    • A CELL-LINE ONTOLOGY RECORD IS PROVENANCE, NOT BIOLOGY. Cellosaurus fixes what LOUCY is classified as; it does not establish that LOUCY behaves like patient SET::NUP214 T-ALL, and this graph has already been caught reading a Cellosaurus age field as a patient age when the record contradicted itself. Oncofusion knowledge graph (internal entry) 2026
  9. E9weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    A single-centre series reports that SET-NUP214 rearrangement in isolation is insufficient to induce leukemia.

    PMID 25956045DOI
    What this evidence cannot show

    TITLE ONLY. PubMed carries no abstract for this record and Europe PMC returns no PMCID, so routes 2-4 could not be attempted and the DOI resolves to a paywalled publisher page. This is a 'known but unreadable' item under the project's standing rule: the identifier and what the title asserts are recorded, and nothing has been written about what the paper probably says. n, ascertainment and follow-up are all unknown here.

    Evidence that argues against this (1)
    • A SINGLE-CENTRE SERIES CANNOT ESTABLISH INSUFFICIENCY. 'SET-NUP214 in isolation is insufficient to induce leukaemia' from one centre is compatible with a cooperating lesion being present but untested, with ascertainment favouring co-mutated cases, and with the follow-up being too short. It is already graded weak and should stay there. Oncofusion knowledge graph (internal entry) 2026
Connected to