Oncofusion

start / graph / fusion-nup98-ddx10

Fusionmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Qn-nup98-family

NUP98::DDX10, inv(11)(p15q22) — the 'helicase' fusion whose helicase activity turns out not to be the point

also called NUP98-DDX10 · inv(11)(p15q22) · NUP98/DDX10

This NUP98 fusion attaches to an RNA-unwinding enzyme, and for nearly thirty years the unwinding was assumed to be what mattered. It is not: a short 24-amino-acid stretch that grabs a nucleolar protein does the work, and the leukaemia depends on that protein instead.

The same thing, for a scientist

**inv(11)(p15q22)** — an **inversion**, not a translocation, because both genes are on chromosome 11 — fuses NUP98 in frame to the putative RNA helicase **DDX10** (Q13206, 875 aa) in de novo **and** therapy-related myeloid malignancies, described in four patients [E1]. DDX10 contributes its helicase machinery: UniProt places the **helicase ATP-binding domain at 100–274**, the **helicase C-terminal domain at 287–448**, the **Q motif at 69–97** and the **DEAD-box motif at 222–225** [E4]. The reciprocal DDX10-NUP98 product is made but *"only NUP98-DDX10 appears to be implicated in tumorigenesis"* [E1]. The original proposed mechanism was aberrant mRNA transport or **ribosome assembly**, DDX10 being a ribosome-assembly factor. **That framing has now been overturned from inside.** The transforming element is **not the helicase activity** but a discrete **24-amino-acid stretch within the DDX10 moiety** that binds **NOL10** (nucleolar protein 10); *Nol10* loss impaired disease progression and improved survival in a mouse NUP98::DDX10 model, and NOL10 acts with the fusion to regulate **serine biosynthesis** and stabilise **ATF4** mRNA [E2]. So the 'helicase class' is a structural description, not a mechanistic one, and this entry names it as such.

Go deeper
Why the class label is misleading and what to say instead

NUP98 partners are conventionally binned as homeodomain, chromatin-reader, methyltransferase and helicase. The helicase bin has exactly one well-studied member, and in that member the helicase function is not what transforms 2.

What the partner actually supplies is a protein-interaction surface — 24 residues that recruit NOL10 — and what the leukaemia depends on is NOL10 and a metabolic output (serine biosynthesis, ATF4 mRNA stability), not unwinding activity 2.

That is worth generalising with care. Across the family, the partner has been assumed to supply a catalytic or reader function: the NSD1 SET domain writes H3K36me2, the KDM5A PHD finger reads H3K4me3. Those assignments are supported by separation-of-function mutants. The DDX10 case shows that the assumption can be wrong when the mutant has not been made, and it is the only NUP98 partner where somebody went looking and found the annotated enzymatic function was a passenger.

The transferable question for every other partner: has a catalytic-dead version of this partner been tested in the fusion, or is the catalytic assignment an inference from the domain annotation?

A metabolic dependency, which is unusual for this family

NOL10 acting with the fusion to regulate serine biosynthesis and stabilise ATF4 mRNA 2 puts NUP98::DDX10 in different therapeutic territory from the rest of the family. Serine-synthesis inhibitors and ATF4/integrated-stress-response modulators exist as tool compounds; menin inhibitors do not obviously act on either.

NUP98::DDX10 was not among the fusions tested in the menin-dependency work, and it was among the five profiled in the AP-MS interactome that failed to recover MLL1 5. So for this partner there is, as far as this pass found, no direct evidence of menin dependency at all — a gap that matters because trial eligibility criteria say 'NUP98 rearrangement' without qualification.

Clinical picture, and the one detailed case

Defined from the start in both de novo and therapy-related myeloid malignancy (AML and MDS) 1, which distinguishes it from NUP98::HOXD13 and NUP98::TOP1, both of which were first seen as therapy-related and only later found de novo.

One documented adult case: a 48-year-old managed with MRD monitoring who relapsed after allogeneic transplant and died; the report states plainly that "the prognosis and indication for allogeneic HSCT are unknown" 3. That is an honest summary of the evidence base for this partner: four founding patients, scattered case reports, one mouse model, and no cohort.

Facts and figures
gene5NUP98
gene3DDX10
cytobandinv(11)(p15q22)
lesion typeinversion, NOT a translocation (both genes on chromosome 11)
partner classlabelled 'helicase', but the helicase activity is not the transforming function
uniprot5P52948
uniprot3Q13206
length AA 51817
length AA 3875
reciprocal productDDX10-NUP98 is generated; the founding paper states only NUP98-DDX10 is implicated in tumorigenesis, without functional testing
actual transforming element24-amino-acid stretch within the DDX10 moiety binding NOL10
downstream dependencyNOL10; serine biosynthesis; ATF4 mRNA stabilisation
menin dependency statusNOT TESTED; this partner was absent from the menin-dependency panel and was among the five fusions whose AP-MS failed to recover MLL1
prognosisunknown; one documented adult case relapsed after transplant and died
diseases
[
  "de novo AML",
  "therapy-related AML",
  "MDS"
]
breakpoints
[
  {
    "id": "bp-nup98-ddx10-inv11",
    "domains_retained_5": [
      "FG repeats 1 (1-156)",
      "GLEBS (157-213)",
      "FG repeats 2 (214-480)"
    ],
    "domains_lost_5": [
      "peptidase S59 (738-880)",
      "Nup96 chain (881-1817)"
    ],
    "domains_retained_3": [
      "helicase ATP-binding (100-274)",
      "helicase C-terminal (287-448)",
      "the 24-residue NOL10-binding stretch"
    ],
    "domains_lost_3": [
      "N-terminal DDX10 sequence upstream of the breakpoint"
    ],
    "phenotype_notes": "The retained helicase domains are not the transforming element; a 24-residue NOL10-binding stretch is."
  }
]
How this could bite us back
  • A partner whose menin dependency has never been tested would nonetheless be enrolled under a partner-blind 'NUP98 rearrangement' criterion, and if it does not respond it dilutes the subgroup that does.
  • If the transforming element is a 24-residue protein-protein interface, that is in principle a better target than a bulk condensate - but it is also a small flat interface of the kind this project repeatedly finds hard to drug.
  • Serine-biosynthesis inhibition as a therapeutic route hits a pathway normal proliferating tissue also needs, so a metabolic strategy for this partner inherits a classic therapeutic-window problem rather than escaping one.
Still open
  1. Is NUP98::DDX10 menin-dependent at all? It was never tested, yet 'NUP98 rearrangement' eligibility criteria would enrol such a patient on a menin-inhibitor trial.
  2. Has a catalytic-dead DDX10 been tested in the fusion? Sufficiency of the 24-residue stretch does not establish that helicase activity is dispensable.
  3. Does the NOL10/serine-biosynthesis/ATF4 axis open a metabolic therapeutic route for this partner that the rest of the family does not share?
  4. For every other NUP98 partner, is the catalytic assignment supported by a separation-of-function mutant, or inferred from the domain annotation? DDX10 is the one case where somebody checked and the annotation was a passenger.
Every source behind this entry (5)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    inv(11)(p15q22) in de novo and therapy-related myeloid malignancies fuses NUP98 in frame to the putative RNA helicase gene DDX10 in four patients; the reciprocal DDX10-NUP98 product is generated but only NUP98-DDX10 appears to be implicated in tumorigenesis.

    PMID 9166830
    What this evidence cannot show

    Four patients, 1997. PubMed holds no DOI for this record so none is given. The paper's own title calls the inversion a 'chromosome translocation', which is loose usage; it is an inversion, both genes being on chromosome 11. The mechanistic proposal (mRNA transport, ribosome assembly) was speculative and has since been superseded.

    Evidence that argues against this (1)
    • The statement that only NUP98-DDX10 is implicated in tumorigenesis was made without functional testing of the reciprocal. In NUP98::TOP1, the reciprocal transcript was likewise reported absent by the founding paper and was subsequently detected and associated with progression to AML, so reciprocal-product claims in this family have a track record of being revised. Iwase S et al. 2003
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The transforming element of NUP98::DDX10 is not the helicase activity but a discrete 24-amino-acid stretch within the DDX10 moiety that binds NOL10; Nol10 loss impaired disease progression and improved survival in a mouse NUP98::DDX10 model, and NOL10 acts with the fusion to regulate serine biosynthesis and stabilise ATF4 mRNA.

    Loss of NOL10 leads to impaired disease progression of NUP98::DDX10 leukemia · Shima Y, et al. · Leukemia 39(6):1368-1379 2025
    PMID 40263434DOI
    What this evidence cannot show

    A single laboratory and a single mouse model; 'impaired disease progression and improved survival' is not abolition. The primary figures were not read in this session - the mechanistic summary comes from a delegated research pass. Whether the 24-residue NOL10-binding stretch is necessary and sufficient in human NUP98::DDX10 patient cells has not been shown.

    Evidence that argues against this (1)
    • The helicase domains are retained in the fusion and no catalytic-dead DDX10 version appears to have been tested, so 'the helicase activity is not the point' is supported by the sufficiency of the 24-residue stretch rather than by a demonstration that catalysis is dispensable. Those are different claims. UniProt Consortium 2026
  3. E3weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    A documented 48-year-old adult with NUP98::DDX10 AML, managed with measurable residual disease monitoring, relapsed after allogeneic transplant and died; the report states that the prognosis and the indication for allogeneic transplant are unknown for this fusion.

    [Acute myeloid leukemia harboring NUP98::DDX10] · Kurita N, et al. · Rinsho Ketsueki 63(10):1397-1401 2022
    PMID 36351646DOI
    What this evidence cannot show

    A single case report in Japanese; the full text was not read in this session and the details come from an abstract-level pass. One patient cannot establish prognosis, and the report's own statement is that prognosis is unknown.

    Evidence that argues against this (1)
    • Recent transplant-heavy paediatric NUP98-rearranged cohorts report 3-year overall survival above 80%, so a single fatal adult post-transplant relapse should not be generalised into a prognosis for this partner in children. Zhang ZX et al. 2026
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    DDX10 is Q13206, 875 aa, with the helicase ATP-binding domain at 100-274, the helicase C-terminal domain at 287-448, the Q motif at 69-97 and the DEAD-box motif at 222-225, and no gene-name synonyms in UniProt.

    What this evidence cannot show

    Primary gene name and length read back from the fetched record. The protein name itself is 'Probable ATP-dependent RNA helicase', so even UniProt hedges the catalytic assignment. Which DDX10 residues the patient fusion retains was not established.

    Evidence that argues against this (1)
    • DDX10 having no synonyms makes it a rare case in this family where alias searching cannot under-count, but it also means older literature indexed under a legacy name would be invisible; the founding-era alias HRH-J8 does not appear in the UniProt synonym list and was not searched. UniProt Consortium 2026
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    NUP98-DDX10 was among the five NUP98 fusions profiled by unbiased AP-MS in HL-60 that failed to recover the reported NSL and MLL1 complex interactions, and it was not among the fusions tested in the menin-dependency work.

    Biomolecular condensation of NUP98 fusion proteins drives leukemogenic gene expression · Terlecki-Zaniewicz S, et al. · Nature Structural & Molecular Biology 28(2):190-201 2021
    PMID 33479542DOI
    What this evidence cannot show

    Absence of NUP98-DDX10 from the menin-dependency study's panel is an absence in one paper's design, not a tested negative; nobody has reported that this fusion is menin-independent. AP-MS not recovering MLL1 applies to all five fusions in that study, not specifically to this one.

    Evidence that argues against this (1)
    • The menin-dependency work tested NUP98-HOXA9, NUP98-JARID1A, NUP98-NSD1 and NUP98-HOXD13 and found the dependency in all of them, so the prior that NUP98-DDX10 shares it is reasonable; treating the untested status as evidence of independence would be wrong in the other direction. Heikamp EB et al. 2022
Connected to