KMT2A::EPS15 (MLL-AF1p), t(1;11)(p32;q23) — the infant-ALL cytoplasmic partner

This is real experimental data — atoms placed by crystallography or cryo-EM, deposited by the people who did the experiment.

• This fusion supplied the first reported lineage switch from B-ALL to AML after blinatumomab in t(1;11), and the switched AML retained the same IGH/IGL clonal sequences. Any CD19-directed therapy in KMT2A::EPS15 infant ALL should be expected to be able to drive the disease out of the lineage the therapy targets, and the residual 0.14% B-ALL population in that case shows the two states can coexist below routine detection.
• Wild-type EPS15 is a core component of clathrin-mediated endocytosis of EGF and transferrin. Any agent that inhibited EPS15 itself, rather than the fusion's oligomerisation interface, would be expected to impair receptor-mediated endocytosis broadly — a mechanism-based toxicity with no obvious therapeutic window.
• If the coiled coil is the target, the synthetic-dimeriser result is the warning: any module that forces MLL proximity substitutes for it. An inhibitor of the EPS15 coiled-coil interface would be expected to select for rearrangements or mutations that restore oligomerisation by another route, and there are dozens of routes.

1. Is TZ-1 menin-inhibitor sensitive? A KMT2A::EPS15 cell line has existed since 2006 and no menin-inhibitor result has been published in it. This is the cheapest menin experiment available anywhere in this beat and it tests the hypothesis in a partner class that supplies no elongation machinery.
2. Do the B-ALL breakpoints really cluster in EPS15 exons 10-12 while AML breaks at exon 2? Four B-ALL cases support it. If it holds in the recombinome's 36 ALL patients, it would be the first lineage-correlated partner-domain difference in the KMT2A landscape, and it would mean the EH domains are retained in myeloid disease and lost in lymphoid disease.
3. Where exactly is the EPS15 coiled coil? A central mechanistic claim of this field rests on a module with no curated boundary and no structure. A coiled-coil prediction plus a deletion series would fix coordinates that every figure in this graph currently has to omit.
4. Does lineage switch after CD19-directed therapy occur more often with cytoplasmic-partner fusions than with elongation-partner fusions? One EPS15 case is suggestive and nothing more; the recombinome cohorts could answer it.

1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   In Table 1 of the 3401-patient 2023 KMT2A recombinome, EPS15 is the 8th most frequent direct translocation partner with 55 patients — 22 infant ALL, 2 infant AML, 3 infant other, 8 paediatric ALL, 5 paediatric AML, 4 adult ALL, 7 adult AML, 1 adult other, 2 unclassified ALL and 1 unclassified AML; the text reports EPS15 at n = 36 (1.6%) of 2182 ALL patients and n = 15 (1.3%) of 1116 AML patients, and separately reports that 12 of 51 EPS15 cases are spliced fusions in which a truncated 5'-KMT2A was recombined upstream of an intact EPS15.

   The KMT2A recombinome of acute leukemias in 2023. · Meyer C, Larghero P, Almeida Lopes B, Burmeister T, Gröger D, Sutton R, et al. · Leukemia 2023
   PMID 37019990DOI
   ▸What this evidence cannot show

   Read directly from the Europe PMC full-text XML of PMC10169636, with the Table 1 column structure verified from the header. The paper is internally inconsistent about the EPS15 denominator: Table 1 totals 55 patients while the spliced-fusion paragraph refers to '12 out of 51 cases'. Both numbers are reported here and neither is corrected. Referral cohort, ascertainment-weighted, counts are floors.

   ▸Evidence that argues against this (1)

   • The infant-ALL weighting of this partner is unlikely to be fully captured by a compendium built on referred and pre-screened samples: at least one KMT2A::EPS15 case was detected only after molecular/cytogenetic discrepancy prompted FISH, and cases in which a t(1;11) is cryptic or in a complex karyotype would be under-represented in exactly the infant group where the count is highest. Watanabe N et al. 2003
2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   Two t(1;11)(p32;q23) translocations in myeloid leukaemias fuse the HRX (MLL) gene to a novel gene AF-1p on 1p32; the der(11) chromosome expresses the 1368 N-terminal amino acids of HRX, including the AT-hook, snRNP and methyltransferase similarities, fused to almost all of the AF-1p product; the predicted wild-type AF-1p product is a 98 kDa acidic protein with no similarity to the AF-4, AF-9 or ENL products but high similarity to the murine eps15 gene product, a cytoplasmic phosphoprotein — defining a new class of genes fused to HRX in 11q23 abnormalities.

   A novel gene, AF-1p, fused to HRX in t(1;11)(p32;q23), is not related to AF-4, AF-9 nor ENL. · Bernard OA, Mauchauffe M, Mecucci C, Van den Berghe H, Berger R · Oncogene 1994
   PMID 8134107
   ▸What this evidence cannot show

   Two patients, 1994 methods, and both were myeloid — which is not the modal disease for this fusion as later cohorts established. No DOI is registered in the PubMed record for this article (the ArticleIdList carries only the PMID), so the citation rests on PMID and URL. Note the abstract as deposited contains an internal typo, writing 't(1;11)(p32;q11)' in one sentence where the title and the rest of the abstract give q23; the title is taken as correct. Read from the abstract.

   ▸Evidence that argues against this (1)

   • 'Almost all the AF-1p product' is not the architecture in B-lineage disease: in the three reported B-ALL cases the EPS15 breakpoints fall in exons 10-12 rather than exon 2, so the retained partner protein is substantially shorter and lacks the EH domains, and a further therapy-related ALL variant lacking the EH domains has been described independently. Yamamoto K et al. 2021
3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.review

   Eps15 was identified as a substrate of the EGF receptor kinase and has a tripartite structure: an NH2-terminal portion containing three EH domains, a central putative coiled-coil region, and a COOH-terminal domain containing multiple copies of the Asp-Pro-Phe triplet; a pool of Eps15 localises to clathrin-coated pits where it interacts with the clathrin assembly complex AP-2 and with epsin, and perturbing Eps15 or epsin function inhibits receptor-mediated endocytosis of EGF and transferrin; Eps15 also has the potential to transform NIH-3T3 cells and the eps15 gene is rearranged with HRX/ALL/MLL in acute myelogenous leukaemias.

   Epidermal growth factor pathway substrate 15, Eps15. · Salcini AE, Chen H, Iannolo G, De Camilli P, Di Fiore PP · Int J Biochem Cell Biol 1999
   PMID 10481267DOI
   ▸What this evidence cannot show

   A short review, used here as a map of the domain architecture and the endocytic function, not as primary evidence for any single experiment. Note the word the authors use: the coiled-coil region is 'putative'. The NIH-3T3 transformation claim is secondary reporting within the review and is not traced to its primary source in this entry.

   ▸Evidence that argues against this (1)

   • The current UniProt record for P42566 annotates NO coiled-coil feature anywhere in the 896-residue protein — three EH domains, four EF-hands, fifteen DPF triplets at aa 599-827, two UIMs and two disordered regions, but no coiled coil. The module that the leukaemia literature treats as EPS15's transforming element therefore has no curated boundary a figure could draw. UniProt Consortium 2026
4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   MLL can be activated by an alternative mechanism to acquisition of transcriptional effector domains: following fusion with AF1p/Eps15 and GAS7, proteins that normally reside in the cytoplasm, the coiled-coil oligomerisation domains of those proteins are necessary and sufficient for leukaemogenic transformation induced by the respective MLL fusion proteins; homodimerisation of MLL by synthetic dimerisation modules mimics bona fide MLL fusion proteins, resulting in Hox gene activation and enhanced self-renewal of haematopoietic progenitors.

   Dimerization contributes to oncogenic activation of MLL chimeras in acute leukemias. · So CW, Lin M, Ayton PM, Chen EH, Cleary ML · Cancer Cell 2003
   PMID 12957285DOI
   ▸What this evidence cannot show

   Read from the abstract; the full text is subscription-only and cell.com returned HTTP 403, so the residue boundaries of the constructs, the identity of the synthetic dimerisation module, the assay used to demonstrate self-association and the quantitative transformation data are NOT verified here and are asserted nowhere in this entry. Mouse progenitor transduction assays. The authors' own framing is cautious: transformation proceeds 'presumably in part by recruitment of accessory factors through the dimerized MLL moiety'.

   ▸Evidence that argues against this (1)

   • The companion paper in the same journal and year reached a narrower conclusion about how far dimerisation generalises, reporting explicitly that MLL-AF9 is NOT dimerised in vivo even though induced dimerisation of truncated MLL is itself sufficient to immortalise bone marrow and upregulate Hoxa7, Hoxa9 and Meis1. Sufficiency for EPS15 and GAS7 is therefore not evidence that the common fusions work this way. Martin ME et al. 2003
5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

   t(1;11)(p32;q23) was initially shown to fuse KMT2A exon 8 to EPS15 exon 2 in AML, but in B-ALL the EPS15 breakpoints are markedly different, falling in exons 10-12 across three reported B-ALL cases; the index case here was a 75-year-old woman with B-ALL (98.2% lymphoblasts, CD19+ CD22+ CD79a+ CD13+ CD33+, CD10- MPO-, karyotype 46,XX,t(1;11)(p32;q23)) expressing a novel in-frame KMT2A exon 8 to EPS15 exon 12 fusion plus the reciprocal transcript, together with an uncommon FLT3 juxtamembrane-domain mutation (in-frame 4-bp deletion and 10-bp insertion).

   Expression of a novel type of KMT2A/EPS15 fusion transcript in FLT3 mutation-positive B-lymphoblastic leukemia with t(1;11)(p32;q23). · Yamamoto K, Yakushijin K, Mizutani Y, Okuni-Watanabe M, Goto H, Higashime A, et al. · Cancer Genet 2021
   PMID 33647817DOI
   ▸What this evidence cannot show

   A single case plus a comparison against three previously reported B-ALL cases — four B-ALL breakpoints in total against a much larger AML literature, so the lineage-breakpoint correlation is a hypothesis from small numbers, not an established rule. The class I / class II cooperation model invoked for the FLT3 mutation is the authors' interpretation.

   ▸Evidence that argues against this (1)

   • A lineage-specific breakpoint rule is hard to sustain against a therapy-related ACUTE LYMPHOBLASTIC leukaemia case whose MLL-AF1p/Eps15 variant was defined precisely by lacking the EH domains — an architecture the exon 10-12 rule would predict, but reported independently and in a therapy-related rather than de novo context, so the correlation may track breakpoint mechanism rather than lineage. Shinohara A et al. 2010
6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

   A novel MLL-AF1p/Eps15 fusion variant lacking the EH domains was reported in therapy-related acute lymphoblastic leukaemia.

   A novel MLL-AF1p/Eps15 fusion variant in therapy-related acute lymphoblastic leukemia, lacking the EH-domains. · Shinohara A, Ichikawa M, Ueda K, Takahashi T, Hangaishi A, Kurokawa M · Leuk Res 2010
   PMID 19665789DOI
   ▸What this evidence cannot show

   PubMed carries NO abstract for this record, so the claim asserted here is exactly the content of the title and nothing more: no exon numbers, no patient details, no functional data are asserted. This is stated plainly rather than padded out from the title.

   ▸Evidence that argues against this (1)

   • The founding description of this fusion is the opposite — the der(11) joining the MLL N-terminus to ALMOST ALL of the AF-1p product, EH domains included — so EH-domain retention varies between patients and neither architecture can be treated as canonical. Bernard OA et al. 1994
7. E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   TZ-1 is a human acute monocytic leukaemia cell line (FAB AML M5a) established from the ascites of a patient, carrying t(1;11)(p32;q23) with MLL-EPS15 fusion transcript and protein demonstrated by FISH and RT-PCR, and positive by flow cytometry for myeloid and monocytic markers with lymphoid-associated markers; it was the first cell line harbouring t(1;11)(p32;q23) among 37 then-existing leukaemia lines with 11q23/MLL rearrangements.

   Establishment of a new human acute monocytic leukemia cell line TZ-1 with t(1;11)(p32;q23) and fusion gene MLL-EPS15. · Sagawa M, Shimizu T, Shimizu T, Awaya N, Mitsuhashi T, Ikeda Y, et al. · Leukemia 2006
   PMID 16826222DOI
   ▸What this evidence cannot show

   A single cell line, derived from ascites after a year of liquid culture, with the usual caveats about drift and about a line's phenotype representing the patient. Critically for this project, TZ-1 does NOT appear in the DepMap panel used for every dependency claim in this graph, so its existence does not close the dependency gap. It is monocytic while the modal patient with this fusion is an infant with B-ALL.

   ▸Evidence that argues against this (1)

   • TZ-1's monocytic phenotype is atypical for the fusion as a whole: in the 3401-patient recombinome EPS15 is predominantly an infant ALL partner (22 infant ALL of 55 patients, 36 ALL in total against 15 AML), so conclusions drawn in TZ-1 may generalise to the minority disease rather than the majority one. Meyer C et al. 2023
8. E8moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

   Monozygotic infant twins with acute lymphoblastic leukaemia carried an identical, novel MLL-EPS15 translocation together with individual (twin-specific) genomic copy-number alterations.

   Molecular characterization of identical, novel MLL-EPS15 translocation and individual genomic copy number alterations in monozygotic infant twins with acute lymphoblastic leukemia. · Kotecha RS, Ford J, Beesley AH, Anderson D, Cole CH, Kees UR · Haematologica 2012
   PMID 22581003DOI
   ▸What this evidence cannot show

   PubMed carries no abstract for this record, so the claim asserted here is exactly the content of the title: a shared identical translocation plus individual copy-number alterations in monozygotic infant twins. The in-utero-origin and postnatal-second-hit interpretation is the standard reading of concordant twin leukaemia and is stated in the section prose as an inference, not asserted as a finding of this paper.

   ▸Evidence that argues against this (1)

   • Concordant twin leukaemia establishes a prenatal shared clone but not that the fusion alone initiates disease; the same architecture is compatible with a prenatal preleukaemic clone requiring postnatal events, which is precisely what the twin-specific copy-number alterations in this pair suggest and what a documented post-therapy lineage switch in another KMT2A::EPS15 infant demonstrates about clonal plasticity. Du J et al. 2021
9. E9moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

   A 6-month-old girl with t(1;11)(p32;q23) KMT2A/EPS15 B-ALL refractory to ALL-directed chemotherapy experienced lineage switch from B-ALL to AML following blinatumomab; myeloid-directed chemotherapy cleared the AML with a residual 0.14% CD19+ B-ALL population, bridging blinatumomab achieved remission of both, and she relapsed with CD19+ B-ALL one month after allogeneic transplant; ClonoSEQ IGH/IGL next-generation sequencing detected three dominant sequences present in the original B-ALL, the switched AML and the post-transplant relapse, indicating cellular reprogramming and differentiation rather than emergence of an unrelated clone.

   Lineage Switch in an Infant B-Lymphoblastic Leukemia With t(1;11)(p32;q23); KMT2A/EPS15, Following Blinatumomab Therapy. · Du J, Chisholm KM, Tsuchiya K, Leger K, Lee BM, Rutledge JC, et al. · Pediatr Dev Pathol 2021
   PMID 33749383DOI
   ▸What this evidence cannot show

   A single patient. Lineage switch after CD19-directed therapy is documented across KMT2A-rearranged infant ALL generally and is not specific to the EPS15 partner; this case is the first reported for t(1;11) but supplies no evidence that the partner influences switch risk. Clonal identity is inferred from IGH/IGL sequences, which track the B-cell clone and cannot by themselves prove the AML arose from it by reprogramming rather than from a shared earlier progenitor.

   ▸Evidence that argues against this (1)

   • A shared-progenitor account is at least as consistent with the data as reprogramming, and the same fusion's twin case supports it: monozygotic infant twins shared an identical MLL-EPS15 translocation while acquiring individual copy-number changes, which shows the fusion can sit in a prenatal multipotent clone from which either lineage could later emerge. Kotecha RS et al. 2012
10. E10established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Human EPS15 is UniProt P42566 (EPS15_HUMAN), primary gene name EPS15 with AF1P carried as a synonym, 896 aa, with EH domains at aa 15-104, 128-216 and 224-314, EF-hands at 48-83, 160-195, 223-258 and 262-292, fifteen Asp-Pro-Phe triplet repeats spanning aa 599-827, an SH3-binding motif at 768-774, two ubiquitin-interacting motifs at 851-870 and 877-896, disordered regions at 542-606 and 760-848, and NO annotated coiled-coil feature anywhere in the sequence.

    UniProtKB P42566 (EPS15_HUMAN) Epidermal growth factor receptor substrate 15, Homo sapiens, 896 aa · UniProt Consortium · UniProtKB 2026
    ▸What this evidence cannot show

    Canonical isoform. The absence of a coiled-coil annotation is a curation fact, not a demonstration that no coiled coil exists — the primary literature describes a central 'putative' coiled coil and the leukaemia field treats it as real. Resolved by fetching the accession and reading back the PRIMARY gene name (EPS15) and length (896 aa); AF1P appears only as a synonym, which is exactly the situation the project's accession rule exists to catch.

    ▸Evidence that argues against this (1)

    • The functional literature is emphatic that a central coiled-coil region exists and that it is the transforming module of this protein in the MLL fusion, so treating the missing UniProt annotation as evidence of absence would contradict the experiment on which the forced-oligomerisation model rests. So CW et al. 2003
11. E11moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    EPS15 ranks 1155 of 17,728 (6.5th percentile) for KMT2A selectivity in the project's genome-wide DepMap Chronos scan, with a KMT2A-rearranged mean gene effect of -0.170 against -0.070 in other leukaemia lines (difference -0.0997) and a pan-essentiality of -0.068 across 1208 lines; when the KMT2A-rearranged index arm is split by lineage, EPS15 scores -1.24 SD in the 8-line myeloid half and -0.08 SD in the 5-line B-lineage half, which the coordinator's heuristic classifies as a MYELOID-IDENTITY ARTEFACT, the same pattern as CEBPA and SPI1.

    DepMap CRISPR (Chronos) gene effect via the Open Targets Platform GraphQL API: genome-wide KMT2A-selectivity scan (analyses/depmap-kmt2a-selective-scan-full/results.json) plus the lineage-split re-analysis of the deferred KMT2A partners run for this beat (analyses/depmap-deferred-partners-lineage-split/) · Broad Institute DepMap; Open Targets Platform; lineage split run by this project · Database and derived analysis 2026
    ▸What this evidence cannot show

    Decisive limitation: no KMT2A::EPS15 line exists in the DepMap panel — the one published line, TZ-1, is not in it — so this measures wild-type EPS15 in leukaemias driven by other partners. The absolute effect is small (-0.17 in the index arm), a high genome-wide rank on a small difference, and splitting a small signal across 8 and 5 lines is noisy. The myeloid-only pattern also runs opposite to the fusion's own epidemiology, which is 40% infant ALL. DepMap release version not stated by the API.

    ▸Evidence that argues against this (1)

    • A wild-type-partner CRISPR score is weak evidence about a fusion regardless of how it splits: Sept6-null mice develop normally and MLL-SEPT6 disease is unchanged on that background, yet the SEPT6 coiled coil is individually required for transformation within the fusion. Partner-gene dispensability and fusion-module necessity are independent facts, so a null or artefactual EPS15 dependency says nothing about whether the EPS15 coiled coil is required in KMT2A::EPS15. Ono R et al. 2005

• part of →KMT2A / MLL1 — the gene that breaks

  1368 N-terminal residues of HRX/MLL: AT-hook, snRNP and methyltransferase-homology regions

• part of →EPS15 / AF1p — domain architecture, and the one coordinate nobody has published

  almost all of the EPS15 product in myeloid cases; EH domains lost at the exon 10-12 B-ALL junctions

• causes →Infant acute lymphoblastic leukaemia (ALL diagnosed under 1 year of age)

  22 of 55 recombinome patients are infants with ALL, the modal presentation of this fusion

• supports →What is actually conserved across KMT2A partner classes — forced self-association, with two documented exceptions that break the rule

  the EPS15 coiled coil is necessary AND sufficient for transformation, and synthetic dimerisers substitute

• related to →EPS15 (AF1p): the EH domains may or may not be discarded, and either way nobody has shown it matters

  patient-level support for EH-domain dispensability: an EH-less therapy-related variant, and B-ALL breakpoints in exons 10-12

• co-occurs with →FLT3 / CD135 — the strongest non-tether co-dependency, and a target the field may have written off on a badly executed trial

  FLT3 juxtamembrane in-frame 4-bp deletion / 10-bp insertion in a t(1;11) B-ALL case

• analogous to →KMT2A::GAS7 (MLL-GAS7), t(11;17)(q23;p13) — three patients, and the best mouse model in the partner class

  the two partners whose coiled-coil necessity AND sufficiency founded the forced-oligomerisation model, in the same experiment

• ← supportsThe deferred KMT2A partners: what the counts actually are, what was written, what was rejected, and three numbers the previous enumeration got wrong

  55 patients, rank 8, infant-ALL-weighted; the beat's largest partner by patient count

• ← related toGSE252378 fragment BEDs: a complete KMT2A N-terminal / C-terminal antibody set in 27 patient samples across six fusion partners, deposited and until now unopened by this corpus

  2 of 27 are KMT2A::EPS15

• ← related toBetween-patient variation in oncoprotein occupancy is not a fusion-partner effect: across six partners the within-partner range of the amplitude statistic is as wide as the between-partner range of its medians

  two KMT2A::EPS15 samples, the smallest group

• ← related toHOXA occupancy is the partner-invariant: occupied in 17 of 18 non-AFF1 KMT2A-r samples, 2.94x to 29.1x
• ← related toIn infant ALL, IRX1 transcript is enriched in KMT2A::AFF1 (21 of 48) and present but rare outside it (3 of 26) - so the occupancy zero outside AFF1 is not an expression zero

  2 of 3 KMT2A::EPS15 infant ALL cases are IRX1-high at AFF1 magnitude - the counter-example to a t(4;11)-specific reading

• ← related toTwo of three KMT2A::EPS15 infant ALL cases carry IRX1 at KMT2A::AFF1 magnitude, while the only KMT2A::EPS15 patient with occupancy data has no oncoprotein at IRX1 at all

  2 of 3 KMT2A::EPS15 infant ALL cases are IRX1-high at AFF1 magnitude

• ← testsMeasure IRX1 transcript and KMT2A-N oncoprotein occupancy in the SAME KMT2A::EPS15 blast - the one cell where expression-without-occupancy is most likely and has never been paired

  primary t(1;11)(p32;q23) infant ALL blasts