Oncofusion

start / graph / mechanism-set-nup214-second-kmt2a-free-hoxa-fusion

Mechanismstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q55-calm-af10

SET::NUP214 is the second KMT2A-free HOXA fusion, and it converges on the same two proteins CALM-AF10 uses: CRM1 and DOT1L - and so does NPM1c, which IS menin-dependent (corrected 2026-09-21)

also called SET-NUP214 · TAF-I-NUP214 · TAF_I-NUP214

A second T-cell leukaemia fusion, SET-NUP214, also switches on the HOXA genes without involving KMT2A. It sits on the HOXA genes together with the same two proteins CALM-AF10 depends on, which suggests they are a shared route to the same output.

The same thing, for a scientist

**SET::NUP214** arises from a cryptic **del(9)(q34.11q34.13)** and defines a third HOXA-high T-ALL subgroup alongside KMT2A-rearranged and CALM-AF10 cases [E1]. **The convergence is specific, not generic: SET-NUP214 binds in the promoter regions of specific HOXA genes, where it interacts with CRM1 and DOT1L** [E1] - the identical pair that CALM-AF10 depends on. siRNA knockdown of SET-NUP214 abolishes HOXA expression, inhibits proliferation and induces differentiation in **LOUCY** but not in other T-ALL lines [E1]. Estimated incidence in adult T-ALL is **4.6% (7/152)**, found via combined interphase FISH with a gene-expression signature of HOXA and NUP214 upregulation; notably the same T-ALL series also contains **CALM-AF10 and MLL translocations**, so all three HOXA-high mechanisms sit side by side in one cohort [E2]. **All of this is ADULT T-ALL, which is two steps from this project's index disease of infant B-lineage KMT2A::AFF1** - species is human but lineage and age both differ.

Go deeper
What the shared element actually is - three different claims, routinely blurred

CALM-AF10 and SET::NUP214 converge, and it matters which convergence is being asserted:

  1. The enzyme (DOT1L). Both engage DOT1L at HOXA 1. In CALM-AF10 the engagement is through a defined protein-protein interface (AF10 OM-LZ to the DOT1L coiled coil); in SET-NUP214 the interaction is reported but the interface is not mapped in this beat's reading. Shared at the level of 'both use DOT1L', not 'both use the same interface'.
  2. The exportin (CRM1). Both are reported to act with CRM1 at HOXA chromatin 13, and in CALM-AF10 the CRM1 route is supported by a CRM1-AF10 chimera that causes leukaemia 3. This is the strongest of the three convergences and it is a nucleoporin/exportin claim, not a chromatin-reader claim. NUP214 is itself the nucleoporin CRM1 docks against, and mutating the CRM1 residues that contact NUP214 abolishes CRM1-AF10 leukaemogenesis 3 - so the two fusions touch the same physical contact from opposite sides.
  3. The locus (HOXA). Both activate HOXA. This is the weakest claim, because many things activate HOXA and convergence on an output says nothing about mechanism.

For the menin question: SET::NUP214 is a second KMT2A-free instance of HOXA activation with no menin-binding motif, and SET-NUP214 AND menin returns 0 in PubMed (2026-09-20, same shape and controls as the CALM-AF10 absence). If both are menin-independent while both are HOXA-high, the separation this beat proposes is supported twice from two unrelated lesions.

A query trap, recorded because it will be hit again

Searching SET as a token collects SMYD2 ('SET- and MYND-domain containing') and Caenorhabditis elegans SET-2, among a large amount of ordinary English. Use "SET oncoprotein", TAF-I, I2PP2A, SET-NUP214, SET::NUP214 or TAF_I-NUP214, and title-check. The fusion is also published under TAF_I-NUP214 and TAF-I-NUP214, which are different strings from SET-NUP214 and return partly different sets.

CORRECTED 2026-09-20 by the dedicated beat: this is not a second menin-free instance

Three corrections from Q58, all verified before applying:

  1. It recruits wild-type MLL1. SET::NUP214 binds MLL through the SET acidic region and the two cooperatively enhance the HOXA10 promoter (PMID 34320268); its nuclear bodies condense MLL1 in association with CRM1 on chromatin (PMID 37516964). Wild-type MLL1 carries MBM1/MBM2, so 'no menin-binding motif in the fusion' does not make the lesion menin-free — and this is a weaker test than CALM-AF10, not a second instance. Both papers were read abstract-only; neither is in PMC.
  2. The DOT1L question is closed and the answer is CONVERGENT. Anti-NUP214 IP blotted for AF10 is negative and no OM-LZ region is present in SET-NUP214 (PMID 18299449, read in full). CALM-AF10 and KMT2A::MLLT10 share one defined interface; SET::NUP214 reaches the same enzyme by another route. The enzyme is the constraint, not the contact.
  3. The output is wider than recorded — HOXA, HOXB and HOXC across three papers, not HOXA/MEIS1.

See ana-the-crm1-address-is-not-menin-independent-npm1c-proves-it.

CORRECTION, 2026-09-21: the third protein at the same address breaks the inference

This entry treats the convergence of CALM-AF10 and SET::NUP214 on CRM1 and DOT1L as strengthening the KMT2A-free / menin-independent comparison arm. It does not, because a third protein sits at the same address and is menin-dependent. Chromatin-bound CRM1 recruits SET-Nup214 AND NPM1c onto HOX clusters, and NPM1-mutant AML responds to menin inhibition: the first-in-human phase 1 of revumenib enrolled relapsed or refractory acute leukaemia with KMT2A rearrangement OR NPM1 mutation and reported a 30% CR/CRh rate in the efficacy analysis population, establishing menin inhibition as a therapy for both.

So 'KMT2A-free, CRM1-addressed, HOXA-driving' describes a lesion that is menin-inhibitor-sensitive as accurately as it describes these two. Convergence on CRM1 is therefore not evidence of menin independence, and this mechanism should be read as establishing a shared ADDRESS rather than a shared exemption. The same applies to the DOT1L half: SET::NUP214 binds wild-type MLL through the SET acidic region, so its route to HOXA runs through the protein menin holds.

Facts and figures
mechanism classcomparator fusion
fusionSET::NUP214 (TAF_I-NUP214)
lesioncryptic del(9)(q34.11q34.13)
diseaseT-ALL; paediatric in the founding report, 4.6% estimated incidence in adult T-ALL
cell lineLOUCY
kmt2a contentnone
menin binding motifnone IN THE FUSION - but it recruits and condenses WILD-TYPE MLL1, which carries MBM1/MBM2 (PMID 34320268, 37516964). Corrected 2026-09-20.
menin testedfalse
menin literature count0
why recorded hereCORRECTED 2026-09-20: recorded as a THIRD KMT2A-free HOXA-driving lesion, but NOT as a second independent instance of menin-independence. It recruits wild-type MLL1, so the composition argument that made it look menin-free does not hold, and it is a WEAKER test than CALM-AF10 rather than a confirmation.
query trapA bare `SET` token collects SMYD2 (SET- and MYND-domain) and C. elegans SET-2; use SET-NUP214, SET::NUP214, TAF-I-NUP214, TAF_I-NUP214, 'SET oncoprotein' or I2PP2A and title-check.
shared with calm af10
[
  "DOT1L at HOXA promoters (shared enzyme, interface not mapped for SET-NUP214)",
  "CRM1 at HOXA chromatin (strongest convergence; NUP214 is CRM1's docking nucleoporin)",
  "HOXA/MEIS output (weakest convergence, since many lesions reach it)"
]
How this could bite us back
  • SET::NUP214 knockdown affects LOUCY and not other T-ALL lines, so its dependency is demonstrated in one line - the same n=1 problem that limits the CALM-AF10 DepMap datum.
  • The convergence argument is easy to overstate: 'both use DOT1L and CRM1' is three claims of very different strength, and only the CRM1 one has orthogonal support.
Still open
  1. Is SET::NUP214 menin-inhibitor sensitive? Zero published records. Running it alongside CALM-AF10 in the same experiment would give two independent instances of the same separation for little extra cost.
  2. Is the CRM1 convergence the real shared element, or coincidence? CALM-AF10 needs CRM1's NUP214 contact and SET::NUP214 contains NUP214; that is either the deepest point of the analogy or a pun on a shared protein name.
  3. Does SET::NUP214 use the same DOT1L surface as AF10 does? The AF10 OM-LZ interface is crystallographically defined; the SET-NUP214/DOT1L interaction is not mapped.
Every source behind this entry (4)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    A cryptic recurrent deletion del(9)(q34.11q34.13) produces a conserved SET-NUP214 fusion, present in the T-ALL cell line LOUCY; SET-NUP214 binds in the promoter regions of specific HOXA genes where it interacts with CRM1 and DOT1L; siRNA knockdown abolishes HOXA expression, inhibits proliferation and induces differentiation in LOUCY but not in other T-ALL lines; and MLL-rearranged, CALM-AF10 and inv(7)(p15q34) patients form one HOXA-high T-ALL subgroup.

    PMID 18299449DOI
    What this evidence cannot show

    **Paediatric T-ALL** plus the LOUCY cell line; identification rested on 3 of 5 HOXA-high patients, so the deletion frequency here is from small numbers. The CRM1 and DOT1L interactions are reported as interactions at HOXA promoters; no interface is mapped.

    Evidence that argues against this (2)
    • The dependency demonstration is in ONE cell line: SET-NUP214 knockdown abolished HOXA expression and induced differentiation in LOUCY but not in other T-ALL lines. That is the same n=1 limitation that constrains the CALM-AF10 DepMap datum, and it means the convergence argument rests on one instance per fusion. Van Vlierberghe P et al. 2008
    • THIS LESION IS NOT MENIN-FREE BY COMPOSITION. SET::NUP214 interacts with MLL via the SET acidic region and cooperatively enhances the HOXA10 promoter, and its phase-separated nuclear bodies condense MLL1 with CRM1 on chromatin. Wild-type MLL1 carries the menin-binding motifs, so recording this fusion as a second KMT2A-free instance of menin-independence was wrong; it is a weaker test than CALM-AF10. Cigdem S et al. 2021
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In adult 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 two index cases carried cryptic del(9)(q34) with a gene-expression signature of HOXA and NUP214 upregulation and TAF_I, FNBP1, C9orf78 and USP20 downregulation; the same series also contained CALM-AF10 and MLL translocations.

    PMID 20065082DOI
    What this evidence cannot show

    **ADULT** T-ALL, and a multicentre Italian series. Transferring anything from here to infant B-lineage KMT2A::AFF1 crosses lineage and age simultaneously. The 4.6% combines 2 FISH-detected index cases with 5 found by PCR/expression across a further 129 patients, so ascertainment differs between the two parts of the denominator.

    Evidence that argues against this (1)
    • This series is ADULT T-ALL while the founding SET-NUP214 report is paediatric, and the 4.6% figure combines two FISH-detected index cases with five found by PCR and expression profiling in a further 129 patients - two ascertainment methods in one denominator. Gorello P et al. 2010
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CRM1 binds HOXA chromatin in a pattern resembling CALM-AF10 binding, a CRM1-AF10 chimera causes leukaemia with HOXA activation, and mutating the CRM1 motifs that contact NUP214 abolishes self-renewal and prevents leukaemia over 300 days.

    PMID 32733011DOI
    What this evidence cannot show

    Murine. Cited here for the CRM1-NUP214 contact, which is the physical point at which the CALM-AF10 and SET::NUP214 stories touch.

    Evidence that argues against this (1)
    • PUBLISHED CONFLICTING RESULT about what convergence on CRM1 shows. Chromatin-bound CRM1 recruits SET-Nup214 AND NPM1c onto HOX clusters - and NPM1-mutant AML is menin-inhibitor responsive: the first-in-human phase 1 of revumenib enrolled relapsed or refractory acute leukaemia with KMT2A rearrangement OR NPM1 mutation and reported a 30% CR/CRh rate in the efficacy analysis population, establishing menin inhibition as a therapy for both. So the CRM1 address is shared by a lesion that is demonstrably MENIN-DEPENDENT. Two KMT2A-free fusions converging on CRM1 is therefore not evidence that the CRM1 route bypasses menin; the third protein at the same address is the counter-example. Issa GC et al. 2023
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    PubMed returns 0 records for `SET-NUP214 AND menin` and for `menin AND (SET-NUP214 OR SET::NUP214)` as of 2026-09-20, against 2,237 for `menin` alone.

    What this evidence cannot show

    Bounded absence with the operand control. Name list: SET-NUP214, SET::NUP214, TAF-I-NUP214, TAF_I-NUP214, against menin / MEN1 / menin inhibitor(s)/inhibition. A bare `SET` search is unusable because it collects SMYD2 and C. elegans SET-2. Not searched here: the German/Dutch T-ALL literature under other spellings.

    Evidence that argues against this (1)
    • PUBLISHED CONFLICTING RESULT against reading this zero as support. The class of lesion - a nucleoporin fusion with no menin-binding motif of its own, driving a HOX programme - has been tested in the NUP98 family and is menin-dependent: VTP50469 is active at an IC50 similar to KMT2A-rearranged and NPM1c cells, evicts MLL1 and the fusion from chromatin, and prolongs survival in NUP98-NSD1 and NUP98-JARID1A xenografts. A counted zero for SET-NUP214 AND menin therefore records that the experiment has not been done in this fusion, in a class where the analogous experiment came out positive. It is an argument for running it, not an argument about the answer. Heikamp EB et al. 2022
Connected to