Oncofusion

start / graph / mech-kmt2a-maml2-notch-dominant-negative

Mechanismweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.Q-usp2-maml2-second-bcr

KMT2A::MAML2 keeps MAML2's activation domain and loses its Notch-binding domain, and suppresses Notch output

also called MLL-MAML2 Notch dominant negative · HES1 suppression by MLL-MAML2

MAML2 normally helps the Notch signal switch genes on, using one end to dock onto Notch and the other end to do the switching. The leukaemia fusion throws away the docking end and keeps the switching end, and in the one experiment that has been done it blocked Notch's own target gene instead of helping it.

The same thing, for a scientist

MAML2's N-terminal basic MamL-1 domain (residues 33-87) binds the Notch intracellular domain on CSL/RBPJ; its transcriptional activation domain sits at 609-681 [E2]. KMT2A::MAML2 retains 952 C-terminal MAML2 residues, i.e. approximately 205-1156, so the MamL-1 domain is lost and the activation domain retained - a composition the founding report states explicitly in terms of the deleted NOTCH-binding basic domain [E1]. In a luciferase assay, MLL-MAML2 suppressed HES1 promoter activation by the NOTCH1 intracellular domain [E1]. The mechanism is therefore best described as **loss of the MAML2 protein's targeting with retention of its effector**, producing a dominant-negative effect on Notch transcriptional output while making the effector module available to whatever the KMT2A N-terminus targets - AT-hooks, the menin/LEDGF-binding region and CXXC-directed unmethylated CpG islands. **Two things must be said against it.** First, this rests on one reporter assay on one promoter in a transfected cell, and the fusion protein has never been detected. Second, the disease distribution is T-lineage-biased and T-ALL is the lineage in which Notch is characteristically activated, which is the opposite of what a Notch dominant-negative predicts [E3].

Go deeper
The domain arithmetic1 deeper
quantityvaluesource
MAML2 full length1156 aaUniProt Q8IZL2 2
MAML2 residues in the fusion952 (C-terminal)founding report 1
implied first retained residue~2051156 - 952
MamL-1 domain (binds NOTCH ICD on CSL)33-87InterPro/Pfam PF09596 2
transcriptional activation domain609-681InterPro/Pfam PF20804 2

MamL-1 at 33-87 is well N-terminal of residue 205, so it is lost. The TAD at 609-681 is well within the retained fragment. The arithmetic is done in this graph, combining the source's residue counts with database domain boundaries; neither the founding paper nor UniProt states it in this form, and it is flagged as derived rather than quoted.

What the fusion swaps in for the lost targeting

Losing MamL-1 means the MAML2 effector can no longer be recruited by Notch to CSL/RBPJ sites. But the fusion is not untargeted - it carries the KMT2A N-terminus, which supplies three AT-hooks, the SNL/menin-binding region, the LEDGF/PSIP1-binding region and the CXXC domain that binds unmethylated CpG. So the molecule is a transcriptional activation domain that has been re-addressed from Notch-responsive enhancers to KMT2A-targeted CpG islands. That is the same logical shape as every other KMT2A fusion in this graph - the partner supplies an effector, KMT2A supplies the address - and it means the Notch suppression may be a side effect of sequestration rather than the oncogenic mechanism.

The T-lineage contradiction, stated and not resolved

Reported KMT2A-MAML2 leukaemias skew T-lineage: therapy-related T-ALL in at least three children, T-lymphoblastic lymphoma in one, and two patients whose biallelic inv(11) disease began as T-ALL 3. T-ALL is the disease in which activating NOTCH1 mutations are the single commonest lesion. A fusion whose measured activity is Notch suppression should, naively, be the wrong oncoprotein for that lineage.

Three readings, none of them tested:

  1. The reporter assay is not representative. HES1 in a transfected non-T cell is a long way from endogenous Notch targets in a thymocyte.
  2. Notch suppression is incidental. The oncogenic work is done by the KMT2A moiety and its re-addressed activation domain; the Notch effect is a by-product that the cell tolerates.
  3. Notch suppression is contributory in a specific window. A dominant-negative on Notch could block normal T-lineage differentiation at a stage where Notch drives maturation, creating a differentiation block rather than a proliferation signal.

Reading 3 is the interesting one and is testable: it predicts that the fusion's T-ALLs should be arrested at an early, Notch-dependent thymocyte stage. Nobody has immunophenotyped these cases against that prediction.

The paralogue that makes the same prediction

A KMT2A::MAML3 fusion in type B3 thymoma has been reported; MAML3 and MAML2 are paralogues, their exon 2 fusion sites share similar functional protein domains, and the authors note the same domain appears downstream of the KMT2A::MAML2 fusion protein - leading them to hypothesise that MAML3 fusions likewise produce abnormal Notch signalling 4. That is a second, independent gene converging on the same structural solution, which strengthens the claim that the retained MAML-family module is the selected feature. It is one case, so it strengthens it a little.

Facts and figures
module retainedMAML2 transcriptional activation domain, residues 609-681 (Pfam PF20804)
module lostMAML2 MamL-1 basic domain, residues 33-87 (Pfam PF09596), which binds the NOTCH intracellular domain on CSL/RBPJ
retained fragmentMAML2 residues approximately 205-1156 (952 residues), derived in this graph from the founding report's residue count and the canonical 1156-aa length
measured effectsuppression of NOTCH1-ICD-driven HES1 promoter activation in a luciferase assay
assay limitationsone promoter, one reporter, transfected cell line, overexpression, fusion protein never detected
re addressingthe KMT2A N-terminus supplies AT-hooks, the SNL/menin-binding region, the LEDGF/PSIP1-binding region and CXXC, so the MAML2 activation domain is plausibly re-addressed from Notch-responsive enhancers to KMT2A-targeted unmethylated CpG islands
unresolved tensionthe fusion's leukaemias are T-lineage-biased, and T-ALL is the lineage in which Notch is most often activated rather than suppressed
testable predictionif Notch suppression contributes, the fusion's T-ALLs should be arrested at an early, Notch-dependent thymocyte stage; no one has immunophenotyped the reported cases against that prediction
not measured
[
  "endogenous Notch target genes in any patient sample",
  "fusion protein detection",
  "chromatin occupancy",
  "HOXA9/MEIS1 induction"
]
How this could bite us back
  • The intuitive drug for a Notch-co-activator fusion is a gamma-secretase inhibitor, and that would push Notch output down - the same direction the fusion's only measured activity already points. Treating this genotype as Notch-activated would be an error with a plausible route into a protocol.
  • If the Notch suppression is incidental sequestration rather than mechanism, then the entire Notch framing of this fusion - which the 2008 Leukemia title itself encourages - is a distraction from the KMT2A moiety, which is the part with an approved drug aimed at it.
  • A mechanism resting on one 2007 luciferase experiment should not be load-bearing for anything. It is recorded here at weak confidence precisely so that it does not quietly harden into background knowledge.
Still open
  1. Are endogenous Notch target genes suppressed in a KMT2A::MAML2 patient sample? One RNA-seq comparison against fusion-negative T-ALL would settle in a week what a 2007 reporter assay has carried for nearly twenty years.
  2. Where does the re-addressed MAML2 activation domain actually act? The fusion has CXXC, so a CUT&RUN in any model would show whether the activation domain now sits on CpG islands rather than on RBPJ sites.
  3. Does the retained fragment keep partial Notch binding beyond the annotated MamL-1 boundary? MAML-CSL-Notch structures involve an extended helix, so the clean loss-of-targeting picture may be too clean.
  4. In thymoma, where Notch signalling is important to thymic epithelial and T-cell interactions, does the same fusion suppress Notch output - and is that why it selects B2/B3 histology specifically?
Every source behind this entry (4)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In MLL-MAML2, the N-terminal basic domain of MAML2 including the NOTCH intracellular-domain binding site is deleted, 952 C-terminal MAML2 residues are retained, and the fusion suppressed HES1 promoter activation by the NOTCH1 intracellular domain in a luciferase assay.

    Identification of a novel fusion gene MLL-MAML2 in secondary acute myelogenous leukemia and myelodysplastic syndrome with inv(11)(q21q23). · Nemoto N, Suzukawa K, Shimizu S, Shinagawa A, Takei N, Taki T, Hayashi Y, Kojima H, Kawakami Y, Nagasawa T · Genes, Chromosomes & Cancer 2007
    PMID 17551948DOI
    What this evidence cannot show

    A single reporter assay on one promoter (HES1) in a transfected cell line, with the fusion delivered by overexpression. The fusion protein itself was never detected - the composition is predicted from the chimeric transcript. Two patients, both adults with therapy-related disease. No endogenous Notch target genes, no haematopoietic cell, no dose-response reported in the abstract. Abstract read in full; figures not re-derived.

    Evidence that argues against this (1)
    • Several later KMT2A-MAML2 leukaemias are T-lineage, the lineage in which NOTCH1 activation is the commonest driver lesion. A Notch dominant-negative is a poor fit for that distribution, and no study has reconciled the two. Mariani RA et al. 2020
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    MAML2 is 1156 aa with the MamL-1 Notch-ICD-binding basic domain at residues 33-87 (Pfam PF09596) and the transcriptional activation domain at 609-681 (Pfam PF20804), so a fusion retaining the C-terminal 952 residues loses the former and keeps the latter.

    UniProtKB entry Q8IZL2 (MAML2_HUMAN) and InterPro Pfam domain boundaries for MAML2 · UniProt Consortium; InterPro (EMBL-EBI) · UniProtKB / InterPro 2026
    What this evidence cannot show

    Pfam boundaries are hidden-Markov-model matches, not experimentally determined domain limits, and the ~204 residues lost include unannotated sequence beyond MamL-1 whose function is unknown. The arithmetic assumes the founding report's 952-residue figure refers to the canonical 1156-aa isoform, which it does not state. Accession resolved by reading back the primary gene name and length from the fetched record.

    Evidence that argues against this (1)
    • If MAML2's Notch-binding capacity extends beyond the annotated MamL-1 model boundary - which is likely, since MAML-CSL-Notch structures involve an extended helix - then the retained fragment may keep partial Notch interaction and the clean loss-of-targeting picture is wrong. Nemoto N et al. 2007
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    KMT2A-MAML2 leukaemia is T-lineage-biased: inv(11)(q21q23)/KMT2A-MAML2 is described as a likely recurrent cytogenetic abnormality in T-cell therapy-related ALL associated with paediatric cases.

    Inv(11)(q21q23); KMT2A-MAML2, a Recurrent Genetic Abnormality in T-Cell Therapy-related Acute Lymphoblastic Leukemia. · Mariani RA, Silva M, Caparelli E, Jennings LJ, Yap KL, Leuer KM, Weinstein J, Gong S · Journal of Pediatric Hematology/Oncology 2020
    PMID 31343482DOI
    What this evidence cannot show

    'Recurrent' rests on this single case plus a prior small series, both of them reports of T-lineage disease, so the apparent T bias may partly reflect which cases get written up. The founding cases were myeloid and one reported case is B-ALL, so the bias is a tendency and not a rule. Abstract read in full.

    Evidence that argues against this (1)
    • The founding report describes secondary AML and MDS, and a B-ALL case exists, so the fusion is not T-restricted. Case-report literatures over-represent the surprising, and 'T-ALL with a KMT2A fusion' is more surprising than 'AML with a KMT2A fusion', which biases what reaches print. Nemoto N et al. 2007
  4. E4weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    A KMT2A::MAML3 fusion in type B3 thymoma fuses at MAML3 exon 2, shares similar functional protein domains with the MAML2 fusion site, and the same domain appears downstream of the KMT2A::MAML2 fusion protein, leading the authors to hypothesise a common abnormal-Notch mechanism.

    PMID 40879670DOI
    What this evidence cannot show

    One patient. The domain comparison is a sequence alignment, not a functional experiment, and the Notch conclusion is explicitly framed by the authors as a hypothesis. DNA-based NGS found no pathogenic variants and the fusion was detected only by RNA sequencing. Abstract read in full.

    Evidence that argues against this (1)
    • A single thymoma case cannot corroborate a mechanism measured in a leukaemia reporter assay, and the two tissues differ so completely that domain-level similarity is weak evidence of shared function. Massoth LR et al. 2020
Connected to