KMT2A::SEPTIN2 (MLL-SEPT2), t(2;11)(q37;q23) — the therapy-related septin
also called MLL-SEPT2 · KMT2A-SEPT2 · MLL::SEPTIN2 · t(2;11)(q37;q23)
A rare fusion of KMT2A to septin-2 that has turned up almost only in patients whose leukaemia followed earlier chemotherapy. It is the one septin fusion where somebody measured what happens to the normal copies of both genes, and the answer was that both go down.
The same thing, for a scientist
KMT2A::SEPTIN2 arises from t(2;11)(q37;q23) and was defined in a patient with treatment-related AML in whom the translocation was the only cytogenetic abnormality; MLL exon 7 was fused in frame to SEPT2 exon 3, with genomic breakpoints in MLL intron 7 and SEPT2 intron 2, and two sequences with 94.4% homology to the topoisomerase II consensus cleavage site were found in MLL intron 7 and SEPT2 intron 2 [E1]. It is vanishingly rare in systematic series — n=1 of 3401 in the DCAL recombinome, a single paediatric AML [E2] — but the reported cases are consistently therapy-related, including a second fusion variant in therapy-related MDS [E3] and a further t-AML case with a new transcript subtype [E4]. Its distinguishing feature is quantitative: in all three MLL-SEPT2 myeloid neoplasms described at the time, combined wild-type plus fusion SEPT2 expression was reduced 12.8-fold (p=0.007) and combined MLL expression 12.4-fold (p=0.028) relative to normal controls, with the SEPT2 down-regulation significant against every other AML genetic subgroup including other MLL fusions [E5]. SEPTIN2 (UniProt Q15019, 361 aa) is a group-II septin and the septin whose G domain and heterotrimeric complex were structurally solved [E6].
This is real experimental data — atoms placed by crystallography or cryo-EM, deposited by the people who did the experiment.
I have never read one of these pictures before
A protein is one long chain of amino acids that folds up. The ribbon traces that chain from one end to the other; it is not a wire, it is a simplification of where the backbone runs. Drag to rotate, scroll to zoom.
Domains are the stretches of chain that do a particular job, and they are what the coloured blocks above the picture mark out. Click one and the view flies to it. In this project the important question about any domain is usually whether the fusion still has it — the labels say lost where it does not.
Anything in yellow sticks is a bound small molecule, which is usually the drug.
A therapy-related entity
The index case was treatment-related AML with t(2;11)(q37;q23) as the sole cytogenetic abnormality 1; the second reported case was therapy-related MDS with a novel fusion variant 3; a third was therapy-related AML with a new transcript subtype 4; a fourth was t-AML after treatment for osteosarcoma 7. Topoisomerase II consensus-homologous sequences flank both genomic breakpoints 1, which is the standard molecular correlate of an etoposide-associated KMT2A break, and is the same signature reported for the MLL breakpoint in KMT2A::SEPTIN6 8.
The only septin fusion with a dosage measurement
Cerveira and colleagues quantified MLL and SEPT2 expression in 58 AML patients spanning the major genetic subgroups plus all three MLL-SEPT2 cases then in the literature 5. Combined wild-type and fusion SEPT2 expression fell 12.8-fold (p=0.007) and combined MLL expression 12.4-fold (p=0.028) versus normal controls. The SEPT2 reduction was significant against every other subgroup, including other MLL fusions. MLL itself was also under-expressed in the other-MLL-fusion group (p=0.023), so the MLL half of the observation is a general property of KMT2A fusions rather than something specific here.
That makes this the one place in the septin literature where a subtraction reading has a number attached to it. It is still only an expression measurement in three patients, and it does not show that the reduction is required for or contributes to disease.
Why the founding authors thought septins mattered
Cerveira et al. closed their 2006 paper by noting that SEPT2 was the fifth septin gene fused with MLL, making the septin family the most frequently involved gene family among MLL partners — about 10% of all known fusion partners — and that further studies were warranted to understand why 1. Twenty years on, this entry records that the question is still open, and that the reason it is open is that four of the five septin fusions have never been put into a functional model at all.
| gene5 | KMT2A |
|---|---|
| gene3 | SEPTIN2 |
| cytoband | t(2;11)(q37;q23) |
| uniprot5 | Q03164 |
| uniprot3 | Q15019 |
| uniprot3 verified | Q15019 = SEPT2_HUMAN, PRIMARY gene name SEPTIN2, 361 aa, Homo sapiens (UniProt REST, checked 2026-09-19); NEDD5, DIFF6, KIAA0158 and the retired SEPT2 are synonyms. |
| prognosis | No estimate. Predominantly therapy-related presentation, which carries its own poor prognosis independent of the fusion. |
| PDB note | 2QA5 = crystal structure of the SEPT2 G domain, 3.4 A; 2QNR = human septin 2 in complex with GDP, 2.6 A; 2QAG = human septin trimer 2/6/7, 4.0 A. All verified against the RCSB entry API on 2026-09-19. Nothing covers the fusion. |
| mechanism class note | Assigned by analogy. No functional model of KMT2A::SEPTIN2 exists. The only mechanistic datum specific to this fusion is an expression measurement, which is a subtraction observation rather than an addition one. |
| druggability class | unexplored |
| stage of hematopoiesis | myeloid progenitor, typically after cytotoxic therapy |
| age peak | adults with prior cytotoxic exposure; one paediatric AML in the systematic series |
diseases
[ "therapy-related AML", "therapy-related MDS", "paediatric AML" ]
frequency — the numbers behind the chart
{
"dcal_recombinome_n": 1,
"dcal_recombinome_total": 3401,
"paediatric_aml_n": 1,
"published_case_reports_n": 4
}breakpoints
[
{
"id": "bp-kmt2a-intron-7-septin2",
"gene5_exon_last": 7,
"gene3_exon_first": 3,
"population": "the index therapy-related AML case; further reported cases carry different transcript subtypes",
"domains_retained_5": [
"MBM1",
"MBM2",
"AT-hook 1",
"AT-hook 2",
"AT-hook 3",
"SNL1/2"
],
"domains_lost_5": [
"PHD1",
"PHD2",
"PHD3",
"Bromodomain",
"PHD4",
"FYRN",
"TAD",
"FYRC",
"WIN",
"SET"
],
"domains_retained_3": [
"Septin-type G domain (aa 34-306)",
"G1/P-loop motif (aa 44-51)",
"G3 motif (aa 101-104)",
"G4 motif (aa 182-185)",
"dimerisation-important region (aa 260-270)"
],
"domains_lost_3": [
"SEPTIN2 exons 1-2"
],
"phenotype_notes": "CXXC omitted from both arrays for the same reason as the other septin entries. SEPTIN2 has no annotated coiled coil but does carry a UniProt region (aa 260-270) annotated as important for dimerisation, inside the retained G domain."
}
]PDB IDS
[ "2QA5", "2QNR", "2QAG" ]
mechanism class
[ "forced-oligomerisation" ]
drugs
[]
key interactors
[ "gene-men1" ]
trials
[]
- If therapy-related KMT2A-septin leukaemia is a real enrichment, then the population most likely to carry a septin fusion is also the population with prior cytotoxic exposure, a damaged marrow and competing clonal haematopoiesis — the worst setting in which to attribute a treatment failure to partner biology.
- The measured fall in wild-type SEPT2 means any therapeutic logic that depends on the partner protein being present in the leukaemic cell starts from a lower baseline than the normal-tissue assumption.
- Is the 12.8-fold SEPT2 down-regulation a consequence of the fusion, a cause of anything, or an artefact of combining wild-type and fusion transcripts in one assay? The experiment that separates them has not been done.
- Why is a fusion with four independent published cases represented once in a 3401-patient systematic series? If the answer is referral pattern rather than rarity, several septin frequencies in this beat are underestimates.
- Does KMT2A::SEPTIN2 transform? Never tested.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In a patient with treatment-related AML and t(2;11)(q37;q23) as the sole cytogenetic abnormality, FISH showed MLL rearrangement and molecular analysis identified SEPT2 on 2q37 as the partner, with an in-frame fusion of MLL exon 7 to SEPT2 exon 3, genomic breakpoints in MLL intron 7 and SEPT2 intron 2, and two sequences with 94.4% homology to the topoisomerase II consensus cleavage site in those introns; SEPT2 was the fifth septin gene fused with MLL, making septins about 10% of all known MLL fusion partners.
SEPT2 is a new fusion partner of MLL in acute myeloid leukemia with t(2;11)(q37;q23). · Cerveira N, Correia C, Bizarro S, et al. · Oncogene 2006What this evidence cannot show
Single index patient. The topoisomerase II consensus homology is a sequence observation, not a demonstrated cleavage event.
Evidence that argues against this (1)
- The '10% of all known fusion partners' figure was calculated when about 50 partners were characterised; against the 2023 census of 107 in-frame KMT2A fusions the six septin partners are roughly 6% of partner genes and, more importantly, about 1.4% of patients (46 of 3401), so the septin family's prominence is much greater in the partner list than in the patient population. Meyer C et al. 2023
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In 3401 genomically characterised KMT2A-rearranged patients, SEPTIN2 appears once — a single paediatric AML — ranking 72nd of the partner list.
The KMT2A recombinome of acute leukemias in 2023. · Meyer C, Larghero P, Almeida Lopes B, et al. · Leukemia 2023What this evidence cannot show
A single patient in the systematic series. This is the count that decides whether KMT2A::SEPTIN2 is an entity or an anecdote, and it argues for anecdote in the referral population.
Evidence that argues against this (1)
- The systematic series undercounts this fusion relative to the published literature, which contains at least four independently reported MLL-SEPT2 patients; since MLL-SEPT2 is predominantly therapy-related and t(2;11) is a simple visible translocation, the discrepancy probably reflects which patients reach a genomic-breakpoint referral service rather than true rarity. Cerveira N et al. 2009
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
A novel MLL-SEPT2 fusion variant occurs in therapy-related myelodysplastic syndrome.
A novel MLL-SEPT2 fusion variant in therapy-related myelodysplastic syndrome. · Cerveira N, Santos J, Pinheiro M, et al. · Cancer Genet Cytogenet 2008What this evidence cannot show
Single case, title-and-abstract level, from the same group as the founding report, so the case series is not independent.
Evidence that argues against this (1)
- An independent group reported t(2;11)(q37;q23) in therapy-related MDS after treatment for acute promyelocytic leukaemia, confirming the entity is not confined to one centre — but their report also shows how often such cases are published on cytogenetics alone without the fusion transcript being typed. Snijder S et al. 2008
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
A new subtype of MLL-SEPT2 fusion transcript occurs in therapy-related acute myeloid leukaemia with t(2;11)(q37;q23).
A new subtype of MLL-SEPT2 fusion transcript in therapy-related acute myeloid leukemia with t(2;11)(q37;q23): a case report and literature review. · van Binsbergen E, de Weerdt O, Buijs A · Cancer Genet Cytogenet 2007What this evidence cannot show
Single case with literature review, title-and-abstract level; from a group independent of the founding report.
Evidence that argues against this (1)
- Transcript-subtype heterogeneity across the handful of reported cases means an RT-PCR assay designed on one junction will miss others, so the apparent rarity of this fusion is partly an assay artefact. Cerveira N et al. 2006
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In all three MLL-SEPT2 myeloid neoplasms described in the literature at the time, combined wild-type and fusion SEPT2 expression was reduced 12.8-fold (p=0.007) and combined wild-type and fusion MLL expression 12.4-fold (p=0.028) versus normal controls; the SEPT2 down-regulation was statistically significant against all other leukaemia genetic subgroups including other MLL fusions, while MLL under-expression was also seen in the other-MLL-fusion group (p=0.023).
Both SEPT2 and MLL are down-regulated in MLL-SEPT2 therapy-related myeloid neoplasia. · Cerveira N, Santos J, Bizarro S, et al. · BMC Cancer 2009What this evidence cannot show
Three MLL-SEPT2 patients against 58 comparator AML. A combined wild-type-plus-fusion qRT-PCR measurement cannot separate loss of the wild-type allele from low fusion expression, and no functional consequence was tested.
Evidence that argues against this (1)
- A mouse genetic test of the same idea for a different septin partner came out negative: Sept6-null mice are phenotypically normal and losing Sept6 does not alter MLL-SEPT6 disease, so reduced septin dosage is not obviously contributing to leukemogenesis even where it is measurable. Ono R et al. 2005
- E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
Human septin-2 is UniProt Q15019 (SEPT2_HUMAN), primary gene name SEPTIN2, 361 aa, also known as NEDD5, with a septin-type G domain at aa 34-306 (G1/P-loop 44-51, G3 101-104, G4 182-185), a region at aa 260-270 annotated as important for dimerisation, and NO annotated coiled coil; the human SEPT2 G domain and the SEPT2-SEPT6-SEPT7 heterotrimer are the structures that defined septin filament architecture.
UniProtKB Q15019 (SEPT2_HUMAN) Septin-2 / NEDD5, Homo sapiens, 361 aa · UniProt Consortium · UniProtKB 2026What this evidence cannot show
Canonical isoform; coordinates fetched from the UniProt REST record on 2026-09-19 and the primary gene name read back rather than trusting a synonym match. Like SEPTIN9 and unlike SEPTIN6/5/11, SEPTIN2 carries no coiled-coil annotation, so two of the five KMT2A-partner septins lack the element the one functional study found indispensable.
Evidence that argues against this (1)
- SEPTIN2's structural prominence is a reason to be careful, not confident: the structures were solved for the native heterotrimer, and they specifically showed the predicted coiled coils are not required for assembly, which is the opposite of what the one functional KMT2A-septin fusion study found. Sirajuddin M et al. 2007
- E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Therapy-related acute myeloid leukaemia with t(2;11)(q37;q23) has been reported after treatment for osteosarcoma.
Therapy-related acute myeloid leukemia with t(2;11)(q37;q23) after treatment for osteosarcoma. · Bielorai B, Meyer C, Trakhtenbrot L, et al. · Cancer Genet Cytogenet 2010What this evidence cannot show
Single case, title-and-abstract level. The title reports the cytogenetic translocation; this entry does not assert from the title alone which fusion transcript was present.
Evidence that argues against this (1)
- t(2;11)(q37;q23) is not always KMT2A::SEPTIN2 — the 11q23 breakpoint must be shown to disrupt KMT2A and the transcript typed, and reports on cytogenetics alone are exactly how partner assignment goes wrong. Strehl S et al. 2006
- E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The MLL genomic breakpoint in a complex MLL-SEPTIN6 rearrangement was a functional DNA topoisomerase II cleavage site in an in vitro assay.
MLL-SEPTIN6 fusion recurs in novel translocation of chromosomes 3, X, and 11 in infant acute myelomonocytic leukaemia and in t(X;11) in infant acute myeloid leukaemia, and MLL genomic breakpoint in complex MLL-SEPTIN6 rearrangement is a DNA topoisomerase II cleavage site. · Slater DJ, Hilgenfeld E, Rappaport EF, et al. · Oncogene 2002What this evidence cannot show
In vitro cleavage on naked DNA; cited here only to show the topoisomerase II signature is shared across septin partners and is not specific to SEPTIN2.
Evidence that argues against this (1)
- Topoisomerase II cleavage-site signatures are found at KMT2A breakpoints generally, including in de novo infant leukaemia with no drug exposure, so the signature does not by itself mark a fusion as therapy-related. Meyer C et al. 2023
SEPTIN-FAMILY PARTNERS sharing the coiled-coil oligomerisation route. Grouped so that a claim about one septin fusion is visibly a claim about a family with several members here.
- part of →KMT2A / MLL1 — the gene that breaks
retained KMT2A N-terminus
predominantly therapy-related myeloid neoplasia
G domain retained; no coiled coil exists to retain
written on literature count despite recombinome n=1