KMT2A::ABI1 (MLL-ABI1), t(10;11)(p11.2;q23) — the partner that binds an elongation partner
also called MLL-ABI1 · MLL-ABI-1 · KMT2A-ABI1 · t(10;11)(p11.2;q23)
KMT2A here joins ABI1, a protein of the cell's actin machinery that normally helps cells move and change shape. What makes it the most interesting partner in this group is that ABI1 physically binds ENL — one of the classic elongation partners — so this fusion may reach the elongation machinery by a completely different route.
The same thing, for a scientist
KMT2A::ABI1 arises from t(10;11)(p11.2;q23) and was cloned from an AML patient in 1998 [E1]; a second case confirmed recurrence in 2001 [E2] and a congenital monocytic leukaemia case followed [E3]. In the DCAL recombinome it is the 17th most frequent partner, n=8 of 3401, distributed 3 infant AML and 5 paediatric AML with no adult and no lymphoid cases — the most tightly paediatric-myeloid distribution of any partner in this beat [E4]. ABI1 (UniProt Q8IZP0, 508 aa) is an Abl-interactor and WAVE-regulatory-complex subunit with a t-SNARE-coiled-coil-homology domain at aa 45-107, a region at 18-79 required for WASF1 binding, and a C-terminal SH3 domain at 446-505 [E5]. Two features make it the most mechanistically informative of the non-elongation partners. First, it has a coiled-coil-homology domain, so it fits the oligomerisation template. Second, and more interesting, ABI1 was identified in a yeast two-hybrid screen as an ENL-binding protein, with the interaction verified by far-Western, GST pulldown and co-immunoprecipitation and mapped to an internal region of ENL and a composite ABI1 motif including the SH3 domain [E6]. A KMT2A-ABI1 fusion could therefore recruit ENL — and with it the super elongation complex — in trans, through a protein-protein interaction rather than by encoding an elongation module. Nobody has tested this. There is no transformation assay for KMT2A::ABI1.
Position and diagnostic confusion at 10p11-12
ABI1 lies at 10p11.2, very close to MLLT10/AF10 at 10p12 and to NEBL, and the three cannot be distinguished by conventional cytogenetics [E4b]. In a Russian paediatric AML registry study of 28 patients with 11q23.3/10p11-12 rearrangements, G-banding, FISH, RT-PCR and long-distance inverse PCR resolved 25 KMT2A-MLLT10, 2 KMT2A-NEBL and 1 KMT2A-ABI1 [E4b]. So roughly 1 in 28 t(10;11) cases is not the fusion it appears to be — the same kind of impersonation problem that SEPTIN11 creates at 4q21, in the partner region that matters most for AML.
The interesting mechanism: reaching elongation in trans
Garcia-Cuellar and colleagues screened for ENL-binding proteins and recovered ABI1, confirming the interaction in vitro by far-Western and GST pulldown and in vivo by co-immunoprecipitation, and mapping it to an internal region of ENL and a composite ABI1 motif including its SH3 domain 6. Their framing at the time was that no unifying property had been found among the then-16 MLL fusion partners, and that this interaction introduced novel aspects.
Read in 2026, the observation is sharper than they could have made it. If MLL-ABI1 retains the ABI1 SH3 domain, the fusion protein carries an ENL-binding surface. ENL is a YEATS-domain SEC component and one of the canonical elongation partners. That would mean a cytoskeletal partner reaches the elongation machinery not by encoding an elongation module but by binding one — a third mechanism, distinct from both the elongation-hijack and forced-oligomerisation models.
This is a hypothesis, not a finding. Two things would have to be shown: that the patient fusion retains the SH3 domain (the founding case used an alternatively spliced ABI1 transcript, so this is not automatic 1), and that ENL is recruited to fusion-bound chromatin. Neither has been done.
Outcome signal, such as it is
Both MLL-ABI1 patients described up to 2001 were surviving, which the authors read as a better prognosis than MLL-AF10 2, and the congenital case responded well to chemotherapy 3. This is three patients. It is recorded here because it is the only outcome statement that exists for the fusion, and because a favourable-outcome claim on n=3 is exactly the kind of thing that hardens into a fact if it is not labelled.
| gene5 | KMT2A |
|---|---|
| gene3 | ABI1 |
| cytoband | t(10;11)(p11.2;q23) |
| uniprot5 | Q03164 |
| uniprot3 | Q8IZP0 |
| uniprot3 verified | Q8IZP0 = ABI1_HUMAN, PRIMARY gene name ABI1, 508 aa, Homo sapiens (UniProt REST, checked 2026-09-19). SSH3BP1 is the only gene synonym on the record; e3B1 and Nap1BP are protein aliases. |
| prognosis | No estimate. Both patients reported up to 2001 were surviving and a congenital case responded well to chemotherapy; that is three patients and must not be read as a favourable-risk assignment. |
| PDB note | No structure of the fusion. No structure of the ABI1 t-SNARE-homology coiled coil or of an ABI1-ENL complex was identified. |
| mechanism class note | Both classes are listed because this is the one non-elongation partner with a documented route to the elongation machinery: ABI1 binds ENL. Neither class has been demonstrated for the fusion; the entry is explicit that this is an untested hypothesis and not a finding. |
| druggability class | unexplored |
| stage of hematopoiesis | myeloid progenitor; a congenital case implies in utero origin |
| age peak | infancy and childhood only; no adult case in the systematic series |
diseases
[ "infant AML", "paediatric AML", "congenital monocytic leukaemia" ]
frequency
{
"dcal_recombinome_n": 8,
"dcal_recombinome_total": 3401,
"infant_aml_n": 3,
"paediatric_aml_n": 5,
"adult_n": 0,
"all_cases_any_age_n": 0,
"russian_registry_share": "1 of 28 patients with 11q23.3/10p11-12 rearrangements"
}breakpoints
[
{
"id": "bp-kmt2a-abi1-index",
"gene5_exon_last": 9,
"gene3_exon_first": 2,
"population": "indicative only. The founding report describes an in-frame fusion combining MLL AT-hooks and the DNA-methyltransferase-homology (CXXC) region with an alternatively spliced ABI1, but does not give a KMT2A exon number in the modern DCAL convention; the exon numbers here are the cohort-typical major-BCR values and are flagged as not patient-derived.",
"domains_retained_5": [
"MBM1",
"MBM2",
"AT-hook 1",
"AT-hook 2",
"AT-hook 3",
"SNL1/2",
"CXXC"
],
"domains_lost_5": [
"PHD1",
"PHD2",
"PHD3",
"Bromodomain",
"PHD4",
"FYRN",
"TAD",
"FYRC",
"WIN",
"SET"
],
"domains_retained_3": [
"SH3 (aa 446-505)"
],
"domains_lost_3": [
"not determined; the index fusion used an alternatively spliced ABI1 transcript"
],
"phenotype_notes": "Myeloid, paediatric. The retained SH3 domain is the element that would allow ENL binding in trans, and its presence in patient fusions is asserted by the founding report for its own case only. The founding report also describes retained polyproline stretches and a 'homeodomain-homologous region'; neither corresponds to a UniProt-annotated domain on Q8IZP0, so they are recorded here in prose rather than in domains_retained_3, where they would produce an undrawable lane."
}
]PDB IDS
[]
mechanism class
[ "forced-oligomerisation", "chromatin-elongation-hijack" ]
drugs
[]
key interactors
[ "gene-men1", "gene-mllt1" ]
trials
[]
- If ABI1 recruits ENL in trans, then a cytoskeletal partner can produce an elongation hijack without encoding one — which means partner class predicts mechanism less reliably than this graph currently assumes, and a menin-inhibitor-plus-DOT1L rationale might apply to a partner nobody would have selected for it.
- ABI1 is a WAVE-regulatory-complex subunit and therefore has a broad essential cytoskeletal role; its modest DepMap signal in KMT2A-rearranged lines is much more likely to reflect that than the fusion, and treating it as a dependency would be a straightforward confound.
- Molecular typing at 10p11-12 is not routine everywhere. Patients with KMT2A::ABI1 will be reported as KMT2A::MLLT10 and treated on MLLT10 evidence, including any future MLLT10-specific strategy.
- Does the KMT2A::ABI1 fusion protein bind ENL, and does ENL occupy fusion-bound chromatin? This is the single most valuable untried experiment in this beat, because a positive result would make the elongation hijack reachable in trans by a cytoskeletal partner.
- Do patient KMT2A::ABI1 fusions retain the SH3 domain? The founding case used an alternatively spliced transcript and nobody has typed the others.
- Does the t-SNARE coiled-coil-homology domain at aa 45-107 mediate self-association, and is it retained in the fusion?
- How many reported t(10;11) AML cases have never been molecularly typed, and what fraction of 'KMT2A-MLLT10' is actually ABI1 or NEBL?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
ABI-1 on 10p11.2, the human homologue of mouse Abl-interactor 1, was identified fused to MLL in a patient with AML and t(10;11)(p11.2;q23); ABI-1 bears no homology to previously described MLL partners, exhibits sequence similarity to homeotic gene products, contains several polyproline stretches and a C-terminal SH3 domain required for Abl binding, and the MLL-ABI-1 fusion transcript in this patient was formed by an alternatively spliced ABI-1, combining MLL AT-hook motifs and the DNA methyltransferase homology region with the homeodomain-homologous region, polyproline stretches and SH3 domain.
ABI-1, a human homolog to mouse Abl-interactor 1, fuses the MLL gene in acute myeloid leukemia with t(10;11)(p11.2;q23). · Taki T, Shibuya N, Taniwaki M, et al. · Blood 1998What this evidence cannot show
Single index patient. No DOI is registered for this 1998 Blood article; the PMID is the only stable identifier and was fetched from NCBI. The 'homeotic gene similarity' is a 1998 sequence observation and is not part of current ABI1 annotation.
Evidence that argues against this (1)
- The index fusion used an alternatively spliced ABI1 transcript rather than full-length e3B1, so the domain content of the fusion protein is not the same as the domain content of the canonical 508 aa protein, and any inference from the UniProt record to the patient fusion is provisional. UniProt Consortium 2026
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Among six acute leukaemia patients with t(10;11) and MLL rearrangement, five had MLL-AF10 and one had MLL-ABI1, making MLL-ABI1 the second reported case and confirming the fusion as a recurring abnormality; both MLL-ABI1 patients were surviving, suggesting better prognosis than MLL-AF10, and ABI1 expression patterns differed between leukaemia and solid tumour cell lines.
t(10;11)-acute leukemias with MLL-AF10 and MLL-ABI1 chimeric transcripts: specific expression patterns of ABI1 gene in leukemia and solid tumor cell lines. · Shibuya N, Taki T, Mugishima H, et al. · Genes Chromosomes Cancer 2001What this evidence cannot show
Two MLL-ABI1 patients in total. A survival comparison on n=2 versus n=5 carries no statistical weight and should not be repeated as a prognostic statement.
Evidence that argues against this (1)
- A larger registry series of 28 patients with 11q23.3/10p11-12 rearrangements found only one KMT2A-ABI1 among 25 KMT2A-MLLT10, and reports no outcome advantage; the favourable-prognosis impression rests entirely on two surviving index patients. Zerkalenkova E et al. 2019
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
A case of congenital monocytic leukaemia with t(10;11)(p11.2;q23) involving MLL and ABI-1, with the same ABI-1 breakpoint as previously reported infant cases, responded well to chemotherapy; the congenital presentation suggests a relationship with in utero leukemogenesis.
Is t(10;11)(p11.2;q23) involving MLL and ABI-1 genes associated with congenital acute monocytic leukemia? · Morerio C, Rosanda C, Rapella A, et al. · Cancer Genet Cytogenet 2002What this evidence cannot show
Single case report framed as a question in its own title. 'Same breakpoint' is a FISH-level statement, not a sequenced junction.
Evidence that argues against this (1)
- The congenital/monocytic association is not supported by the systematic series, which classifies all eight ABI1 patients simply as AML across infant and paediatric age groups without a monocytic designation. Meyer C et al. 2023
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In 3401 genomically characterised KMT2A-rearranged patients, ABI1 was the 17th most frequent partner with n=8: 3 infant AML and 5 paediatric AML, with no adult and no lymphoid cases in any age class.
The KMT2A recombinome of acute leukemias in 2023. · Meyer C, Larghero P, Almeida Lopes B, et al. · Leukemia 2023What this evidence cannot show
Referral cohort. Eight patients; the complete absence of adult and lymphoid cases is a real pattern in this series but the confidence interval on n=8 is wide.
Evidence that argues against this (1)
- ABI1 lies immediately adjacent to MLLT10 and NEBL at 10p11-12 and is indistinguishable from them cytogenetically, so cases reaching a referral service on karyotype are systematically assigned to MLLT10 unless molecularly typed; 8 is a floor. Zerkalenkova E et al. 2019
- E4bstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Among 28 patients in a Russian paediatric AML registration study carrying rearrangements between 11q23.3 and 10p11-12, characterised by G-banding, FISH, RT-PCR and long-distance inverse PCR, 25 harboured KMT2A-MLLT10, 2 KMT2A-NEBL and 1 KMT2A-ABI1; MLLT10, NEBL and ABI1 lie in such close proximity at 10p11-12 that they cannot be distinguished by conventional cytogenetics.
Acute myeloid leukemia with t(10;11)(p11-12;q23.3): Results of Russian Pediatric AML registration study. · Zerkalenkova E, Lebedeva S, Kazakova A, et al. · Int J Lab Hematol 2019What this evidence cannot show
A single national registry, paediatric AML only. The proportions are specific to that cohort.
Evidence that argues against this (1)
- The larger recombinome places ABI1 (n=8) and NEBL (n=3) at similar absolute rarity against MLLT10 (n=209 in AML), so the 1-in-28 rate here is if anything generous and the general lesson — molecular typing is mandatory at 10p11-12 — is stronger than the specific ratio. Meyer C et al. 2023
- E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
Human ABI1 is UniProt Q8IZP0 (ABI1_HUMAN), primary gene name ABI1, 508 aa, also known as e3B1, Eps8 SH3 domain-binding protein and Nap1-binding protein, with a t-SNARE coiled-coil homology domain at aa 45-107, a region at aa 18-79 required for binding to WASF1, disordered regions at 159-290, 306-375 and 388-421, and an SH3 domain at aa 446-505.
UniProtKB Q8IZP0 (ABI1_HUMAN) Abl interactor 1 / e3B1, Homo sapiens, 508 aa · UniProt Consortium · UniProtKB 2026What this evidence cannot show
Canonical isoform. ABI1 is extensively alternatively spliced and the founding fusion used a non-canonical transcript, so these coordinates describe the reference protein rather than the oncoprotein.
Evidence that argues against this (1)
- The t-SNARE coiled-coil homology domain is a homology-based annotation, not a demonstrated self-association module in ABI1, so reading it as the fusion's oligomerisation domain by analogy with EPS15 or GAS7 would be an assumption; no self-association assay has been performed on MLL-ABI1. So CW et al. 2003
- E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
ABI1 was identified by yeast two-hybrid as an ENL-binding protein and the interaction was verified in vitro by far-Western blot and GST pulldown and in vivo by co-immunoprecipitation, with structure-function analysis mapping it to an internal region of ENL and a composite ABI1 motif including its SH3 domain; ENL is the MLL fusion partner in t(11;19) and ABI1 is itself fused to MLL in t(10;11)(p11.2;q23), and at the time no general unifying property had been detected among the 16 characterised MLL fusion partners.
ENL, the MLL fusion partner in t(11;19), binds to the c-Abl interactor protein 1 (ABI1) that is fused to MLL in t(10;11)+. · Garcia-Cuellar MP, Schreiner SA, Birke M, et al. · Oncogene 2000What this evidence cannot show
Yeast two-hybrid plus biochemistry in cell lines; the co-immunoprecipitation is of the wild-type proteins, not of the MLL-ABI1 fusion protein, and no chromatin or transcription readout was performed. Whether the patient fusion retains the SH3 region is not established.
Evidence that argues against this (1)
- The founding fusion transcript used an alternatively spliced ABI1, so the SH3-containing composite motif that mediates ENL binding may or may not be present in the oncoprotein; the ENL-recruitment model therefore rests on a domain whose retention in patients has never been confirmed. Taki T et al. 1998
- E7weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.dataset
In a genome-wide DepMap CRISPR (Chronos) scan, ABI1 had a mean gene effect of -0.065 across 1208 lines, a KMT2A-rearranged group mean of -0.096 versus +0.003 in other leukaemia lines, and ranked 1,169 of 17,728 for KMT2A selectivity (6.6th percentile) — the best-ranked gene in this beat, but with a difference of only -0.099 Chronos units.
DepMap CRISPR (Chronos) genome-wide KMT2A-selectivity scan of 17,728 protein-coding genes, 11 Cellosaurus-annotated KMT2A-rearranged leukaemia lines vs other leukaemia lines (analyses/depmap-kmt2a-selective-scan-full/results.json, all_results; ABI1 row looked up for this beat) · Broad Institute DepMap; Cellosaurus; Open Targets Platform; scan run by this project · Database and derived analysis 2026What this evidence cannot show
A difference of -0.099 across 11 KMT2A-rearranged lines is not a dependency; MEN1 ranks first in the same scan with -0.87. No KMT2A-ABI1 line exists in DepMap. Reported because a 6.6th-percentile rank invites over-reading and should be on record as weak.
Evidence that argues against this (1)
- ABI1 is a core subunit of the WAVE regulatory complex and its knockout effects are confounded by that essential cytoskeletal role, so even a real selective signal could reflect complex-level biology rather than anything to do with the fusion. UniProt Consortium 2026
TWO NON-ELONGATION PARTNERS whose proposed mechanism is oligomerisation of a cytoplasmic protein rather than recruitment of an elongation complex - the alternative transforming route this graph tracks separately from the SEC/DOT1L account.
- part of →KMT2A / MLL1 — the gene that breaks
retained KMT2A N-terminus
alternatively spliced ABI1 retaining polyproline stretches and SH3
8/3401 recombinome cases, all paediatric or infant AML
- co-occurs with →KMT2A::MLLT10 (MLL-AF10), t(10;11)(p12;q23)
indistinguishable at 10p11-12 by conventional cytogenetics
ABI1 is the one non-elongation partner with a possible trans route to elongation
enumerated and written
also untested
ABI1 (10p11.2) assigned to the high-risk partner group
- ← related toOne karyotype band, two partner genes: why t(11;19) and t(10;11) can cancel a real partner effect
10p11.2/ABI1 versus 10p12/MLLT10: different hazard ratios at the same band