Oncofusion

start / graph / fusion-vim-kmt2a

Fusionstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q92-kmt2a-sarcoma

VIM::KMT2A - the one KMT2A rearrangement that keeps the SET domain and still throws away the menin-binding exon

also called VIM::KMT2A · VIM-KMT2A · vimentin-KMT2A

In a handful of spindle-cell sarcomas the front half of the vimentin gene is joined to almost the whole of the KMT2A gene, which is the reverse of what happens in leukaemia. The joint falls just past KMT2A's first exon, so the resulting protein keeps KMT2A's enzyme end but loses the small piece that menin grips - the piece the menin-inhibitor drugs depend on.

The same thing, for a scientist

VIM::KMT2A joins VIM exon 4 to KMT2A exon 2 (RNA-seq, four published cases across three reports). Translating those exon indices: VIM exons 1-4 encode vimentin aa 1-240 (head, coil 1A, linker 1 and most of coil 1B - 52% of the 466-aa protein), and KMT2A exon 2 begins at codon 145, so the chimera is vimentin aa 1-240 fused to KMT2A aa 145-3969. That retains all three AT-hooks, the CXXC zinc finger, PHD1-3, the bromodomain, FYRN/FYRC and the SET methyltransferase domain - and it discards KMT2A aa 1-144, which is exactly the menin-binding motif (aa 6-25) and LEDGF integrase-binding motif 1 (aa 123-134). IBM2 (aa 147-152) falls 2-7 residues inside exon 2 and is retained. So VIM::KMT2A is a near-full-length, catalytically intact KMT2A carrying an intermediate-filament N-terminal extension and NO menin interface. Morphologically these tumours are uniformly cellular spindle-cell neoplasms without the sclerotic background of the YAP1 arm, CD99- and BCL2-positive with no specific immunoprofile, and both cases in the pan-sarcoma screen carried concurrent CTNNB1, SMARCB1 and ARID1A mutations. Unlike the YAP1 arm, HOXA upregulation WAS reported in the founding VIM::KMT2A case.

Go deeper
The breakpoint, and what it actually encodes

Four published cases, three reports, one breakpoint: VIM exon 4 to KMT2A exon 2 123.

Translated into residues 4:

contribution
VIM exons 1-4vimentin aa 1-240 (exon 1 is UTR-only; exons 2-4 give exactly 720 CDS bp) - head (2-95), coil 1A (96-131), linker 1 (132-153) and most of coil 1B (154-245)
KMT2A exon 2 onwardKMT2A aa 145-3969 - AT-hooks 1-3, CXXC, PHD1-3, bromodomain, FYRN/FYRC, SET
discarded from KMT2Aaa 1-144 (exon 1) = menin-binding motif (6-25) and IBM1 (123-134)
retained by a marginIBM2 (147-152), which begins 3 residues into exon 2

The SET domain is intact. That is unique among the KMT2A rearrangements this graph holds: every leukaemic KMT2A fusion loses SET. The pathologists who described the entity say so themselves - in leukaemia the breakpoints leave the N-terminal DNA-binding domains and lose the SET domain, whereas in these sarcomas the breakpoints leave the DNA-binding domain partially disrupted or lost and the SET domain intact 5.

What this implies for the index fusion, by contrast

KMT2A::AFF1 is a loss-of-SET, gain-of-elongation, menin-dependent molecule. VIM::KMT2A is its mirror on two of those three axes: it keeps SET, gains no elongation partner at all (vimentin is a cytoskeletal structural protein with no transcriptional activity), and cannot bind menin because the motif is in the exon that was cut off. If a KMT2A rearrangement can transform a mesenchymal cell while retaining catalytic activity and lacking the menin interface, then the leukaemic requirement for the menin interface is a property of the leukaemic molecule and its lineage, not of KMT2A rearrangement as such. That is the single most useful thing this rare tumour offers this project.

The HOXA asymmetry between the two arms

The YAP1 arm shows no HOXA9 or MEIS upregulation 5. For the VIM arm the founding report is stated - by the YAP1-arm authors, citing it - to have shown upregulation of KMT2A and of HOXA genes 6. If that holds, the two morphological subtypes of this entity differ precisely on the axis that defines the leukaemia program, and the arm that looks more like a leukaemia fusion (near-full-length KMT2A, AT-hooks and CXXC present) is the arm with the HOXA signal. This is a secondary attribution and should be treated as such: the founding paper's full text could not be read (see E6 caveats), so nobody in this project has seen the figure.

The co-mutations

Both VIM::KMT2A cases in the pan-sarcoma screen carried concurrent CTNNB1, SMARCB1 and ARID1A mutations, with tumour mutational burden of 0.8 and 1.6 mutations/Mb 1. Two of those three are chromatin-remodelling subunits (SMARCB1 and ARID1A are both BAF/SWI-SNF components) and the third is the Wnt pathway's transcriptional effector. That is a striking combination in a tumour whose fusion retains a functioning H3K4 methyltransferase, and it means the fusion is not the sole driver in either case - which is itself a difference from infant KMT2A::AFF1, where the fusion is often the only recurrent lesion. n = 2, so this is a pattern to test, not a finding.

Facts and figures
gene5VIM
gene3KMT2A
architectureconventional 5'-3' fusion, but with KMT2A as the 3' partner - the inverse of every leukaemic KMT2A fusion
cytobandVIM 10p13 / KMT2A 11q23.3 (interchromosomal, unlike the intrachromosomal YAP1 arm)
uniprot5P08670
uniprot3Q03164
length AA 5466
length AA 33969
menin binding motif in fusionfalse
ledgf ibm in fusionIBM1 lost, IBM2 retained by a 3-residue margin - uncertain at residue resolution
cxxc in fusiontrue
set domain in fusiontrue
co mutation notepresent in both cases of the pan-sarcoma screen (n=2); TMB 0.8 and 1.6 mutations/Mb
published case count4
functional work donenone - no construct, no transformation assay, no occupancy, no menin or LEDGF pulldown, no protein ever detected
diseases
[
  "KMT2A-rearranged soft tissue sarcoma, spindle-to-round cell subtype"
]
breakpoints
[
  {
    "id": "bp-vim-ex4-kmt2a-ex2",
    "gene5_exon_last": 4,
    "gene3_exon_first": 2,
    "frequency_pct": 100,
    "population": "all 4 published VIM::KMT2A cases with a reported breakpoint",
    "gene5_residues_retained": [
      1,
      240
    ],
    "gene3_residues_retained": [
      145,
      3969
    ],
    "domains_retained_5": [
      "head (aa 2-95)",
      "coil 1A (96-131)",
      "linker 1 (132-153)",
      "coil 1B partial (154-240)"
    ],
    "domains_lost_5": [
      "coil 1B distal (241-245)",
      "linker 12",
      "coil 2",
      "tail"
    ],
    "domains_retained_3": [
      "AT-hook 1",
      "AT-hook 2",
      "AT-hook 3",
      "CXXC",
      "PHD1",
      "PHD2",
      "PHD3",
      "Bromodomain",
      "FYRN",
      "FYRC",
      "SET",
      "IBM2 (integrase-binding motif 2)"
    ],
    "domains_lost_3": [
      "MBM (menin-binding motif)",
      "IBM1 (integrase-binding motif 1)"
    ],
    "phenotype_notes": "uniformly cellular fascicular spindle cells, no sclerotic background, CD99+ BCL2+ variable S100; HOXA upregulation reported in the founding case (secondary attribution, weak)"
  }
]
co mutations
[
  "CTNNB1",
  "SMARCB1",
  "ARID1A"
]
frequency
{
  "share_of_kmt2a_rearranged_sarcoma_pct": 6,
  "n_of_cohort": "2 of 34 (Massoth 2020)"
}
How this could bite us back
  • A menin inhibitor has no binding site on this fusion either: the motif is in the discarded exon 1. Any KMT2A-sarcoma menin-inhibitor proposal fails on the same coordinate argument in both arms of the entity.
  • If VIM::KMT2A is catalytically active, the therapeutic logic inverts relative to leukaemia: a KMT2A SET-domain or WDR5/WIN inhibitor would be the rational modality here and is exactly the wrong modality in KMT2A::AFF1, where SET is already gone. Getting that backwards in either direction would be a predictable clinical failure.
  • Three co-mutations in two tumours means a single-agent rationale is weak regardless of target: a Wnt-active, BAF-deficient tumour has at least two escape routes from any fusion-directed agent.
  • With four published cases and no cell line, a trial in this entity is not feasible as a stand-alone study; it can only ride on a basket, which means a negative result will be uninterpretable.
Still open
  1. Does VIM::KMT2A retain H3K4 methyltransferase activity? The SET domain and the WIN motif are both in the retained portion, so this is the only KMT2A rearrangement in this graph where the question can be asked, and it has never been asked.
  2. Does the vimentin aa 1-240 extension change where the chimera goes? Vimentin has no human experimental nuclear annotation, so the fusion pairs a cytoplasmic filament segment with a protein that must reach chromatin - an obvious localisation experiment that nobody has run.
  3. Is the HOXA signal in the VIM arm real? One immunoprecipitation-free RT-qPCR panel on one archival case would settle whether the two arms of this entity differ on the leukaemia program.
  4. Do the CTNNB1, SMARCB1 and ARID1A co-mutations recur beyond n = 2, and is the SMARCB1/ARID1A pairing a BAF-dosage lesion of the kind that would matter to a KMT2A fusion retaining catalytic activity?
Every source behind this entry (6)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In a comprehensive genomic profiling database of 14,680 human sarcomas, two spindle-to-round cell sarcomas in men aged 43 and 46 with lower-extremity tumours carried VIM-KMT2A fusions with breakpoints near VIM exon 4 and KMT2A exon 2, and both harboured concurrent mutations in CTNNB1, SMARCB1 and ARID1A with tumour mutational burden of 0.8 and 1.6 mutations/Mb.

    PMID 32461620DOI
    What this evidence cannot show

    n = 2. A commercial profiling cohort enriched for advanced aggressive disease, so neither the 2-case frequency nor the demographics are population estimates. The variant-level detail for CTNNB1, SMARCB1 and ARID1A is in a supplementary table that was not opened for this entry, so the specific alleles, their hotspot status and their zygosity are not recorded here. Read from the open-access full text.

    Evidence that argues against this (1)
    • Three concurrent pathogenic mutations in two tumours with a TMB under 2/Mb means the fusion is not the sole driver in either case, so any phenotype attributed to VIM::KMT2A in these tumours is confounded by a Wnt lesion and two BAF-subunit lesions. No case with an isolated VIM::KMT2A and no co-mutation has been reported. Almohsen SS et al. 2023
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In two further human VIM::KMT2A-rearranged sarcomas arising in the deep soft tissues of adult males, RNA sequencing placed the fusion at VIM exon 4 and KMT2A exon 2 in both, and the tumours were hypercellular fascicular spindle-cell neoplasms with scant delicate collagen, diffusely positive for CD99 and BCL2 and variably positive for S100, lacking the extensively sclerotic background of the KMT2A::YAP1 tumours.

    VIM::KMT2A-rearranged sarcomas: A report of two new cases confirming an entity with distinct histologic features. · Almohsen SS, Griffin AM, Dickson BC, et al. · Genes Chromosomes Cancer 2023
    PMID 36959690DOI
    What this evidence cannot show

    Two case reports; combined with the pan-sarcoma screen and the founding report this entity rests on four published cases in total, which is too few for a morphological definition to be secure. Adult deep soft tissue, not paediatric and not haematopoietic. No functional work: no construct, no transformation assay, no occupancy, no menin or LEDGF pulldown. Read from the abstract.

    Evidence that argues against this (1)
    • The authors themselves state these tumours have no specific immunohistochemical profile and overlap morphologically with other monomorphic spindle cell neoplasms, so cases are unfindable without molecular testing. The four-case series is therefore an ascertainment-limited sample of an unknown larger population, and the uniform breakpoint may reflect uniform assay design as much as uniform biology. Wang Z et al. 2025
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In the founding report of KMT2A fusions in sarcoma, a 30-year-old woman with a femoral and soft-tissue tumour of monomorphic round-to-spindle cells positive for BCOR, WT1 and NKX2-2 carried a VIM (exon 4)-KMT2A (exon 2) fusion confirmed by RT-PCR, Sanger sequencing and FISH, and the authors concluded that the fusion protein structure was different from that in acute leukaemia and suggested a novel oncogenic mechanism.

    KMT2A (MLL) fusions in aggressive sarcomas in young adults. · Yoshida A, Arai Y, Tanzawa Y, et al. · Histopathology 2019
    PMID 31136005DOI
    What this evidence cannot show

    A single case within a two-case report, and the tumour was aggressive with lung metastases and death within a few years. Full text could not be read: Europe PMC returns no PMCID for this PMID, so the fullTextXML, efetch db=pmc and PMC HTML routes are all structurally unavailable, and resolving the Wiley DOI returns HTTP 403. Everything here is from the abstract.

    Evidence that argues against this (1)
    • This case is immunophenotypically a BCOR/WT1/NKX2-2-positive round cell tumour, an immunoprofile shared with BCOR-rearranged sarcoma and with Ewing sarcoma, so a reader could reasonably ask whether the VIM::KMT2A is the driver or a secondary event in a small round cell tumour of another class. A later LDB1::KMT2A case carried both a KMT2A fusion and a pathogenic BCOR frameshift, which makes the question live rather than pedantic. Wang Z et al. 2025
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Computed here (2026-09-21) from live Ensembl and UniProt records: VIM exons 1-4 of ENST00000544301 encode exactly vimentin aa 1-240 (720 CDS bp; exon 1 is untranslated) and KMT2A exon 2 of ENST00000534358 begins at codon 145, so a VIM(ex4)::KMT2A(ex2) transcript encodes vimentin aa 1-240 joined to KMT2A aa 145-3969 - retaining the AT-hooks, CXXC, PHD1-3, bromodomain, FYR domains and SET, and discarding the menin-binding motif (aa 6-25) and IBM1 (aa 123-134).

    What this evidence cannot show

    A coordinate derivation from reported exon indices, not a sequenced cDNA or an observed protein. No VIM::KMT2A protein has ever been detected by any method in any paper read for this beat, so this states what the reported transcript can encode. Exon indices are transcript-specific; the sarcoma reports name RNA-seq breakpoints without always naming the transcript accession, and the frame was taken as in-frame on the founding report's own statement.

    Evidence that argues against this (1)
    • IBM2 (aa 147-152) begins only three residues inside exon 2, so whether the LEDGF interface survives depends on the exact junction sequence rather than on the exon index - and this project's own KMT2A reference sequences differ by 3 residues in total length. The safe statement is that the MENIN motif is lost with certainty and the LEDGF IBM2 status is uncertain at residue resolution. UniProt Consortium 2026
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The pathologists who described the YAP1 arm state explicitly that in KMT2A-rearranged leukaemia KMT2A is the 5' partner with breakpoints between exon 9 and intron 11, retaining the N-terminal DNA-binding domains and losing the SET domain, whereas in these sarcomas KMT2A is presumably the 3' partner with breakpoints in exons 4-5, leaving the DNA-binding domain partially disrupted or totally lost and the SET domain intact.

    PMID 31592798DOI
    What this evidence cannot show

    This is an interpretive statement in a pathology discussion, not a protein measurement, and it describes the YAP1 arm's exon 4-5 breakpoints rather than the VIM arm's exon 2 breakpoint. It is cited here for the polarity and SET-domain point, which applies to both arms. Read from the PMC full text (route 3, efetch db=pmc, after fullTextXML returned HTTP 500).

    Evidence that argues against this (1)
    • 'SET domain intact' in a transcript is not the same as an active methyltransferase in a cell. KMT2A's catalytic activity depends on the WRAD module assembled through the WIN motif at aa 3762-3767 and on the FYRN/FYRC pair, and no study has measured H3K4 methyltransferase activity, WDR5 binding or WRAD assembly on any sarcoma KMT2A fusion protein. Massoth LR et al. 2020
  6. E6weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    SECONDARY ATTRIBUTION: the authors of the YAP1-arm series state that in the previously reported VIM-KMT2A sarcoma, upregulation of KMT2A and of HOXA genes was noted - in contrast to their own three YAP1-KMT2A cases, in which HOXA9 and MEIS were not upregulated.

    PMID 31592798DOI
    What this evidence cannot show

    This is one group reporting another group's result. The founding paper it refers to could not be read: no PMCID exists for that PMID so Europe PMC fullTextXML, efetch db=pmc and the PMC HTML page are all unavailable, and the Wiley DOI returns HTTP 403 - all five routes in this project's unreadable test fail. elink from that PMID to GEO returns zero datasets (checked 2026-09-21), so the expression data are not recoverable from an archive either. Nobody in this project has seen the measurement, and it is recorded at weak confidence for that reason and no other.

    Evidence that argues against this (1)
    • If the HOXA signal in the VIM arm is real, the strongest statement this beat can make - that KMT2A-rearranged sarcoma runs a non-HOXA program - holds only for the YAP1 arm. The two arms would then differ on the leukaemia analogy itself, and the VIM arm, which retains the AT-hooks and CXXC and the SET domain, would be the closer analogue despite also lacking the menin motif. Yoshida A et al. 2019
Connected to