Oncofusion

start / graph / breakpoint-kmt2a-major-bcr

Breakpointestablished confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.Q2-fusion-atlas

KMT2A major breakpoint cluster region (BCR1): intron 7 - exon 13, ~8.3 kb

also called MLL BCR · MLL 8.3 kb breakpoint cluster region · KMT2A BCR1 · MLL exons 5-11 (old numbering)

Almost every leukaemia-causing break in the KMT2A gene happens inside one small stretch of DNA about 8,300 letters long. Where inside that stretch the break lands decides which working parts of the KMT2A protein survive into the fusion protein.

The same thing, for a scientist

The KMT2A major breakpoint cluster region (BCR1) is an ~8.3 kb interval that Meyer et al. map between KMT2A intron 7 and exon 13 (DCAL exon numbering); 3336/3401 (98%) of genomically characterised KMT2A rearrangements break there, versus 47 (1.4%) in the minor BCR (intron 20-exon 24) [E1]. Within BCR1, breaks partition into three functional subregions: A = exon 9 + intron 9 (1761 bp), B = exon 10 + intron 10 (679 bp), C = exon 11 through exon 13 (5026 bp), with an observed mean distribution of A 37.2% / B 19.8% / C 39.5% across 3401 patients [E1]. The functional cut point is the intron 10 / exon 11 boundary, because exons 11-16 encode the PHD1-PHD2-ePHD3 cluster whose inclusion is incompatible with transformation [E1][E2]. The BCR lies 5' of every PHD finger, so all major-BCR in-frame KMT2A fusions retain AT-hooks, the SNL/menin-binding region and CXXC, and lose PHD1-3, the bromodomain, FYRN/FYRC, the TAD and the SET domain [E2][E3].

Go deeper
Why the break is confined to one 8.3 kb window1 deeper

Two independent constraints converge on the same interval.

Constraint 1 - physics and chemistry of the DNA. The BCR carries an unusual density of features that make double-strand breaks likely: intron 9 contains four Alu repetitive elements (three of which are transcriptionally active), and intron 11 contains a DNase I hypersensitive site, an apoptotic cleavage site, an RNA polymerase II binding site and topoisomerase II binding/cleavage sites 1.

Constraint 2 - selection for a functional oncoprotein. Only breaks 5' of the PHD fingers make a transforming protein. When Muntean et al. deliberately added back two or more PHD fingers to MLL-AF9, the fusion lost the ability to immortalise haematopoietic progenitors, showed reduced association with the Hoxa9 locus and failed to upregulate Hoxa9 2. Even if breaks occurred 3' of the PHDs, the resulting clone would not be selected. The observed BCR is therefore the intersection of where DNA breaks easily and which breaks produce a leukaemogenic protein.

The exon-numbering trap

The same region is described in the literature as MLL exons 5-11, exons 8-14, and intron 7 to exon 13, depending on which transcript annotation the authors used. The Meyer/Marschalek series (the largest breakpoint dataset in existence, n = 3401) uses the numbering in which the major BCR runs from intron 7 to exon 13 and in which the three functional subregions are exon 9/intron 9, exon 10/intron 10, and exon 11-13 1. Every exon number in this knowledge graph's KMT2A breakpoint nodes follows the Meyer/Marschalek (DCAL) convention. Cross-reading an older paper without converting is the single most common source of error in this field.

Observed breakpoint distribution (n = 3401)
SubregionContentLengthShare of all breaks
Aexon 9 + intron 91761 bp37.2%
Bexon 10 + intron 10679 bp19.8%
Cexon 11 + intron 11 + exon 12 + intron 12 + exon 135026 bp39.5%

Subregion B is by far the shortest yet absorbs ~20% of breaks, i.e. breaks are not distributed in proportion to length. Meyer et al. explicitly decline to use a length-based random-distribution model for this reason 1.

Outside BCR1: 47 patients (1.4%) broke in the minor BCR (intron 20 - exon 24); 4 upstream of BCR1; 8 between BCR1 and BCR2; 5 downstream of BCR2 1. When only the targeted-NGS subset is considered (~100% discovery rate, versus ~95% for LD-PCR), the BCR1:BCR2 split becomes 94%:6%, i.e. the classical PCR methods under-detect minor-BCR events 1.

What survives on the KMT2A side

Because every BCR1 break is 5' of exon 11, the direct KMT2A fusion protein retains the three AT-hooks, the subnuclear-localisation/repression region containing the menin-binding motifs MBM1 and MBM2, the LEDGF/PSIP1-binding region, and the CXXC CpG-binding domain. It loses PHD1, PHD2, extended PHD3 (exons 11-16), the bromodomain, ePHD4 (exons 17-21), FYRN/FYRC, the transactivation domain, the WIN motif and the SET methyltransferase domain 13. A break within or 3' of exon 11 additionally splits the PHD1-3 cluster itself, which matters mostly for the reciprocal protein (see breakpoint-kmt2a-intron-11).

Facts and figures
geneKMT2A
cytoband11q23.3
genome buildhg38
region namemajor breakpoint cluster region (BCR1)
extent descriptionKMT2A intron 7 to exon 13 (Meyer/Marschalek DCAL exon numbering)
approx length bp8300
population3401 genomically characterised KMT2A-rearranged acute leukaemia patients (DCAL 2003-2022)
share of all kmt2a breaks PCT98
functional cut pointintron 10 / exon 11 boundary
numbering conventionMeyer/Marschalek (DCAL) KMT2A exon numbering; described elsewhere as exons 8-14 or the older exons 5-11
subregions
B679.0B: 679A1,761A: 1761C5,026C: 5026
subregions — the numbers behind the chart
[
  {
    "id": "A",
    "content": "exon 9 + intron 9",
    "length_bp": 1761,
    "share_pct": 37.2
  },
  {
    "id": "B",
    "content": "exon 10 + intron 10",
    "length_bp": 679,
    "share_pct": 19.8
  },
  {
    "id": "C",
    "content": "exon 11 + intron 11 + exon 12 + intron 12 + exon 13",
    "length_bp": 5026,
    "share_pct": 39.5
  }
]
fragility features
[
  "intron 9: four Alu elements, three transcriptionally active",
  "intron 11: DNase I hypersensitive site",
  "intron 11: apoptotic cleavage site",
  "intron 11: RNA polymerase II binding site",
  "intron 11: topoisomerase II binding/cleavage sites"
]
domains retained 5 prime
[
  "AT-hook 1-3",
  "SNL/repression region",
  "MBM1",
  "MBM2",
  "LEDGF/PSIP1-binding region",
  "CXXC"
]
domains lost 5 prime
[
  "PHD1",
  "PHD2",
  "ePHD3",
  "bromodomain",
  "ePHD4",
  "FYRN",
  "FYRC",
  "TAD",
  "WIN motif",
  "SET"
]
domain encoding exons
{
  "PHD1-PHD2-ePHD3": "KMT2A exons 11-16",
  "ePHD4": "KMT2A exons 17-21"
}
exon convention
{
  "frame": "DCAL / Meyer recombinome numbering",
  "offset_to_ensembl_canonical": "+1 (this entry's exon N = Ensembl canonical exon N-1)",
  "established_by": "three EXACT base-pair identities against ENST00000534358 in GRCh38, not by assertion: canonical e8+i8 = 74+1,687 = 1,761 matching Meyer 'exon 9 - intron 9 = 1761'; canonical e9+i9 = 132+547 = 679 matching 'exon 10 - intron 10 = 679'; canonical e10 start to e12 end = 5,026 matching 'exon 11 ... exon 13 = 5026'",
  "so_this_entrys_features_in_GRCh38": {
    "functional cut point, stated here as the intron-10/exon-11 boundary": "canonical intron 9/exon 10 = chr11:118,484,861|118,484,862",
    "this entry's 'exon 11'": "canonical exon 10, chr11:118,484,862-118,484,975 (114 bp). CORRECTED CYCLE 59: this field previously gave the end as 118,484,993, implying a 132 bp exon. Re-derived from Ensembl REST lookup/id/ENST00000534358 expand=1 in GRCh38: the end is 118,484,975 and the width is 114 bp, so the old end overshot by exactly 18 bp. The START was correct, and no conclusion in this entry changes - they all turn on the cut point at chr11:118,484,861|118,484,862, which is unaffected.",
    "the 'exon 11 CTCF site' of PMID 37788669 is NOT this entry's exon 11": "it is CANONICAL exon 11 at chr11:118,488,614-118,488,760, 3,752 bp telomeric of the cut point"
  },
  "worked_damage": "mechanism-h3k9me1-2-ctcf-balance-gates-kmt2a-rearrangement published 'The CTCF site is at the functional cut point of the BCR' by merging this entry's exon 11 with the canonical exon 11. Retracted cycle 57. This field exists so the next entry does not have to make the error to find the offset.",
  "four_frames_are_in_play": [
    "DCAL/recombinome (this entry) = canonical + 1",
    "Ensembl canonical (PMID 37788669, PMID 34405474)",
    "historic, exons 5-11 (breakpoint-kmt2a-bcr-architecture)",
    "MOUSE Kmt2a exons (Canela 2017) - NOT reconciled to human"
  ],
  "caveat": "The offset is exact on three features and is not verified across all 36 exons, so a frame that drifts by more than one elsewhere would not be caught by it. Convert to GRCh38 base positions before combining any two exon-numbered features.",
  "correction_cycle_59": {
    "what_was_wrong": "the stated end of canonical exon 10, 118,484,993 against a measured 118,484,975 - an 18 bp overshoot, and a width of 132 bp against a measured 114 bp",
    "found_by": "Q108, whose lift.py asserts the exon table rather than trusting it",
    "verified_by": "king, re-fetched Ensembl REST lookup/id/ENST00000534358?expand=1 on 2026-09-22",
    "what_is_unaffected": "the cut point (chr11:118,484,861|118,484,862), the +1 offset established on the three Meyer bp identities, and the 3,752 bp distance to canonical exon 11's start - none of which uses this end coordinate",
    "why_it_happened": "this field was written by hand in cycle 57 to convert the entry's own exon numbers into base positions. The three bp identities that established the offset were computed in code and are right; this coordinate was typed alongside them and was not."
  }
}
How this could bite us back
  • Because the BCR lies 5' of every PHD finger, every major-BCR KMT2A fusion protein keeps the menin-binding N-terminus. That is the basis of the menin-inhibitor strategy, but wild-type KMT2A carries exactly the same menin-binding surface, so a menin inhibitor cannot distinguish fusion from wild-type protein by this interface and on-target toxicity in KMT2A-dependent normal tissue is structurally unavoidable rather than an off-target accident.
  • Minor-BCR fusions (1.4-6% of cases) retain PHD1-3 and the bromodomain and therefore carry a chromatin-reading module that major-BCR fusions lack. A therapeutic strategy validated only in major-BCR models may not transfer to them.
Still open
  1. Is subregion B (exon 10 + intron 10, only 679 bp) genuinely hyper-fragile per base pair, or is its ~20% share partly an artefact of primer placement in the LD-PCR assays that generated most of the cohort?
  2. Subregion C (5026 bp) and subregion A (1761 bp) absorb almost identical fractions of breaks (39.5% vs 37.2%), so region A is ~3x more fragile per base. What makes intron 9 so breakable beyond the presence of Alu elements?
  3. Do the ~6% of breakpoints in the minor BCR detectable by NGS but missed by LD-PCR represent a systematically different disease biology that the historical literature has under-sampled?
Every source behind this entry (3)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    The KMT2A major BCR maps between intron 7 and exon 13, contains 98% of characterised breakpoints, and partitions into subregions exon9/intron9 (37.2%), exon10/intron10 (19.8%) and exon11-13 (39.5%) in 3401 patients.

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

    Referral cohort of the Frankfurt Diagnostic Center of Acute Leukemia (DCAL), 2003-2022, accrued from centres that chose to send samples; partner and breakpoint frequencies may be biased by referral patterns and by the pre-screening method used. 2702 of 3401 cases were analysed by LD-PCR (~95% discovery) and only 696 by targeted NGS (~100%), and the BCR1:BCR2 ratio differs between methods, so the 98% figure is method-dependent.

    Evidence that argues against this (1)
    • Rare KMT2A fusions break entirely outside both BCRs - e.g. KMT2A::ARHGEF12 arising from a cryptic 11q deletion, reported outside the known major and minor breakpoint regions - so the BCR is a strong statistical rule, not an absolute one. Assaf N et al. 2021
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    KMT2A translocations are restricted to an 8.3 kb breakpoint region and consistently delete the PHD fingers; adding two or more PHD fingers back into MLL-AF9 abolishes immortalisation of haematopoietic progenitors and prevents Hoxa9 upregulation.

    The PHD fingers of MLL block MLL fusion protein-mediated transformation · Muntean AG, Giannola D, Udager AM, Hess JL · Blood 2008
    PMID 18796627DOI
    What this evidence cannot show

    Retroviral overexpression in murine bone-marrow progenitors with an in vitro immortalisation readout; shows that PHD-retaining fusions are non-transforming in this assay, not that such breaks never occur in patients. Does not address minor-BCR fusions, which do retain PHD1-3.

    Evidence that argues against this (1)
    • Minor-BCR KMT2A::AFDN fusions in T-ALL include the complete PHD1-3, bromodomain and ePHD4 in the direct fusion protein and are nonetheless leukaemogenic in patients, showing the PHD block is context-dependent rather than absolute. Meyer C et al. 2023
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.review

    All frequent MLL/KMT2A fusion proteins retain the nuclear-factor-binding N-terminus (MEN1, LEDGF) that targets them to promoters, and lose the C-terminal PHD/bromodomain/SET modules.

    MLL leukemia and future treatment strategies · Marschalek R · Archiv der Pharmazie 2015
    PMID 25740345DOI
    What this evidence cannot show

    Review; used as a map of the consensus rather than as primary evidence for any single residue boundary.

    Evidence that argues against this (2)
    • The generalisation fails for one of the commonest KMT2A rearrangements of all. The KMT2A partial tandem duplication, present in 11% of acute myeloid leukaemia with a normal karyotype and counted among the most frequent KMT2A rearrangements in the recombinome, duplicates exons 2-6 by joining intron 6 or intron 8 back to intron 1; the resulting protein RETAINS the PHD fingers, the bromodomain, FYRN/FYRC and the SET methyltransferase domain that this item says frequent KMT2A rearrangements lose. All frequent KMT2A fusion proteins is true only if partial tandem duplications are excluded by definition, which the review does not say. Strout MP et al. 1998
    • Losing the C-terminal PHD/bromodomain/SET modules packages together elements with very different functional weight. Inserting PHD1 and PHD2 into MLL-ENL had no measurable effect on stem-cell immortalisation, while inserting PHD3 alone abolished it by restoring CYP33 and HDAC1 recruitment. The functionally decisive loss is PHD3, so a domain bar that treats the whole C-terminal block as one lost unit overstates how much of it matters for transformation. Chen J et al. 2008
Connected to