The "second breakpoint cluster region" of KMT2A (Meyer 2019): the founding report of BCR2, not a third cluster
also called MLL second BCR · KMT2A minor BCR founding report · MLL intron 19 - exon 24 breakpoint region · BCR2 discovery cohort
A 2019 paper titled "a second breakpoint cluster region" is easily read as announcing a brand-new, third place where the KMT2A gene breaks. It is not: it is the paper that first found the region this graph already calls the minor BCR, and reading the title without the paper would count one region twice.
The same thing, for a scientist
Meyer et al. (Leukemia 2019) designed 2688 overlapping capture probes across the whole KMT2A gene and applied targeted NGS to 109 acute-leukaemia patients pre-screened as KMT2A-rearranged whose breakpoints MP-PCR and LDI-PCR could not resolve. Ninety-three of 109 yielded a rearrangement: 67 (72%) broke in the major BCR, 5 (5%) showed only a reciprocal der(TP) product, and 21 (23%) broke outside the major BCR in what the authors call a novel, minor BCR lying between KMT2A intron 19 and exon 24, with a clear preference for introns 21-23 [E1]. Seventeen of the 21 were KMT2A::USP2; the remaining four were balanced translocations (one USP8, two AF4/AFF1, one AF9/MLLT3) [E1]. This is the same region the 2023 recombinome and this graph's breakpoint-kmt2a-minor-bcr call BCR2, where it is mapped slightly more conservatively to intron 20 - exon 24 and carries 47/3401 (1.4%) of all genomically characterised breakpoints [E2]. The correct structural addition to the atlas is therefore not a third cluster but (i) a founding citation the minor-BCR entry does not currently carry, (ii) a 5' boundary discrepancy of roughly one intron between the 2019 and 2023 mappings, and (iii) a hard number on ascertainment: in a cohort deliberately enriched for PCR-unresolvable cases, BCR2 accounts for 23% of breakpoints rather than 1.4%.
What the title says and what the paper says1 deeper
The title is "Human MLL/KMT2A gene exhibits a second breakpoint cluster region for recurrent MLL-USP2 fusions". Read cold, "second" sounds like "one more, in addition to the major and the minor". Read against the paper, "second" means second in total: the major BCR is the first, and this region is the second one ever described. The closing paragraph is explicit - the authors write that they "have identified a minor BCR within the human MLL gene" and that "the discovery of a second, minor BCR extends our knowledge of the MLL recombinome" 1. They use "second" and "minor" interchangeably for one interval.
So the honest finding for this beat is a negative on the dispatch premise: the atlas does not need a new cluster entry. It needs its existing minor-BCR entry to carry the founding citation and the boundary discrepancy.
Why this is more than a pedantic point
The graph's breakpoint atlas is a set of mutually exclusive intervals with percentages attached that sum to roughly 100%. Adding a third cluster that is in fact the second one would have created an interval overlapping breakpoint-kmt2a-minor-bcr across the whole of introns 20-24, with 21 patients and 47 patients both presented as real and separate. Every downstream percentage citing the atlas would then have been wrong, silently, in the direction of inflating the non-major-BCR share.
The boundary discrepancy: intron 19 or intron 20?
| source | 5' boundary | 3' boundary | preferred sub-interval | n in BCR2 |
|---|---|---|---|---|
| Meyer 2019, targeted NGS, n = 93 resolved 1 | KMT2A intron 19 | exon 24 | introns 21-23 | 21 |
| Meyer 2023 recombinome, n = 3401 2 | intron 20 | exon 24 | not stated | 47 |
Same authors, same centre, four years apart, and the 5' boundary moves inward by one intron. Neither paper comments on the change. The most likely explanation is that the 2019 boundary was set by the most 5' breakpoint seen in a 21-patient series while the 2023 mapping is a consensus over 47, but that is an inference and neither paper states it. breakpoint-kmt2a-minor-bcr currently gives only the intron 20 figure; a reader designing a capture panel from this graph would not cover intron 19-20 and could miss a real patient. The graph should carry both boundaries and say which is which, which is what this entry does.
Ascertainment: 1.4% or 23%?1 deeper
Both numbers are correct and they measure different things.
- 1.4% (47/3401) is the share of BCR2 among all genomically characterised KMT2A rearrangements in the DCAL referral series 2.
- 23% (21/93) is the share of BCR2 among patients who were already known to be KMT2A-rearranged and whose breakpoint MP-PCR and LDI-PCR had failed to find 1.
The 2019 cohort is by construction the residue left behind by methods that work well on the major BCR. The 23% is therefore not an incidence, it is a conditional enrichment, and it is the most operationally useful number in the paper: if standard assays say KMT2A-rearranged but cannot name a partner, roughly one in four of those patients is a BCR2 case and roughly one in five is KMT2A::USP2. The 2023 paper makes the same point from the other side - in the targeted-NGS-only subset the BCR1:BCR2 split becomes about 94:6 rather than 98:1.4 2.
Three-prime deletions, which the older methods also missed
The same targeted-NGS run found 3' deletions in 23 of 88 patients (26%), at similar rates in the major and minor breakpoint groups (26.9% and 23.8%). The authors note this is much higher than the 13% and 7% reported in two earlier series they cite 1. That is a second, independent instance of the same methodological bias: a feature's apparent rarity tracked the sensitivity of the assay used to look for it.
What actually is structurally new, and what it does to the atlas
Three things in this paper are additions rather than restatements.
- KMT2A::USP2 is an intrachromosomal inversion, not a translocation. USP2 lies about 1 Mbp telomeric to KMT2A at 11q23.3 and is transcribed towards the centromere, so every KMT2A::USP2 event is an inversion within chromosome 11 1. The atlas's breakpoint entries are written around reciprocal translocation products; an inversion has different reciprocal geometry, and the reciprocal USP2 breakpoints here were scattered over a larger 11q23.3 region involving the upstream gene C2CD2L and the downstream USP2-AS1 1.
- AFF1 and MLLT3 can break in BCR2. Two AF4/AFF1 cases and one AF9/MLLT3 case in this series broke in the minor BCR 1. The minor-BCR entry describes the BCR2 partner spectrum as almost disjoint from BCR1 and names USP2 and SEPTIN6 as exclusive to BCR2, which remains true; but the converse - that the big four elongation partners are confined to BCR1 - is not true and must not be inferred from it.
- A mechanistic proposal for why a PHD-retaining fusion can transform at all (see hyp-uch-domain-rescues-phd-mediated-mll-degradation), which is the standing paradox of the entire minor BCR 13.
Relation to the other atlas entries, stated plainly
breakpoint-kmt2a-major-bcr(BCR1, intron 7 - exon 13): unaffected. Nothing here changes it.breakpoint-kmt2a-minor-bcr(BCR2, intron 20 - exon 24): this entry is its founding report. Same region, not a separate cluster. The two should be read together; this one carries the discovery cohort, the boundary discrepancy and the ascertainment arithmetic, and deliberately does not restate the 2023 partner table.breakpoint-kmt2a-intron-9andbreakpoint-kmt2a-intron-11: unaffected, both sit inside BCR1.breakpoint-kmt2a-bcr-architecture: unaffected as written, but note it describes the fragility features of BCR1 only. No equivalent characterisation of BCR2 fragility has been published, which is a real hole (see open questions).
| region | KMT2A minor breakpoint cluster region (BCR2) |
|---|---|
| exon numbering convention | Meyer/Marschalek (DCAL), the convention used by every KMT2A breakpoint entry in this graph. |
| mechanism class | For KMT2A::USP2, intrachromosomal inversion at 11q23.3 (USP2 lies ~1 Mbp telomeric to KMT2A and is transcribed toward the centromere). For the four balanced cases, reciprocal translocation. |
| relation to existing atlas entries | Same interval as breakpoint-kmt2a-minor-bcr. NOT a third cluster. Does not alter breakpoint-kmt2a-major-bcr, breakpoint-kmt2a-intron-9 or breakpoint-kmt2a-intron-11. |
| not characterised | No scaffold-attachment region, TOP2 consensus, DNase-hypersensitivity or Alu mapping of BCR2 has been published, in contrast to the detailed fragility map of BCR1 held in breakpoint-kmt2a-bcr-architecture. Nothing here should be read as asserting a breakage mechanism for BCR2. |
boundaries reported
[
{
"source_pmid": "30899083",
"year": 2019,
"five_prime": "intron 19",
"three_prime": "exon 24",
"preferred_subinterval": "introns 21-23",
"n_patients": 21,
"cohort": "109 pre-screened KMT2A-r patients unresolved by MP-PCR and LDI-PCR; 93 resolved"
},
{
"source_pmid": "37019990",
"year": 2023,
"five_prime": "intron 20",
"three_prime": "exon 24",
"preferred_subinterval": "not stated",
"n_patients": 47,
"cohort": "3401 genomically characterised KMT2A rearrangements, DCAL 2003-2022"
}
]breakpoint distribution 2019 cohort — the numbers behind the chart
{
"major_bcr": 67,
"reciprocal_der_only": 5,
"minor_bcr": 21,
"total_resolved": 93,
"unresolved_set_aside": 16,
"cohort_total": 109
}minor bcr partners 2019 — the numbers behind the chart
{
"USP2": 17,
"USP8": 1,
"AFF1": 2,
"MLLT3": 1
}three prime deletions
{
"overall": "23/88 (26%)",
"major_bcr_group_pct": 26.9,
"minor_bcr_group_pct": 23.8,
"prior_estimates_cited_by_authors_pct": [
13,
7
]
}- Any MRD assay or capture panel built from the 2023 intron 20 boundary does not cover intron 19-20. If the 2019 boundary is the correct one, a patient breaking there is invisible to the panel and is scored MRD-negative when they are not - and this is exactly the group (BCR2, USP2-fused, 76% MRD-positive at day 33) in whom MRD monitoring matters most.
- Reporting BCR2 prevalence as 1.4% to a diagnostic laboratory understates the yield of looking for it. The decision-relevant conditional number is 23% of PCR-unresolvable KMT2A-r cases.
- The 26% rate of 3' deletions means a FISH break-apart probe can read as 3'-signal loss or as normal, so a negative FISH does not exclude a BCR2 rearrangement - see risk-kmt2a-usp2-cytogenetically-cryptic.
- Is the fragility architecture of BCR2 the same as BCR1? BCR1 has a mapped scaffold-attachment region, TOP2 consensus cluster, DNase I hypersensitive site and Alu load; nobody has looked for any of these in KMT2A introns 19-24, so the field does not know whether BCR2 breaks for the same physical reason or a different one.
- Why do introns 21-23 dominate within BCR2? The 2019 letter states the preference and offers no explanation; a sequence-feature comparison of introns 21-23 against 19-20 and 24 is a cheap, purely computational test.
- Is the intron 19 versus intron 20 boundary difference real, or an artefact of the smaller 2019 series? Resolving it needs the per-patient breakpoint coordinates, which are held by the DCAL group and deposited in neither paper.
- Does the inversion geometry of KMT2A::USP2 - a ~1 Mbp intrachromosomal event with reciprocal breakpoints scattered across C2CD2L and USP2-AS1 - imply a loop-anchor mechanism distinct from the interchromosomal translocations that dominate BCR1?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Targeted NGS across the whole KMT2A gene in 109 pre-screened patients resolved 93 rearrangements, of which 21 (23%) broke in a novel minor BCR between KMT2A intron 19 and exon 24 (preference introns 21-23); 17 were KMT2A::USP2 inversions and four were balanced translocations (one USP8, two AFF1, one MLLT3).
Human MLL/KMT2A gene exhibits a second breakpoint cluster region for recurrent MLL-USP2 fusions. · Meyer C, Lopes BA, Caye-Eude A, Cavé H, Arfeuille C, et al., Marschalek R · Leukemia 2019What this evidence cannot show
A Letter, not a full paper: no abstract is indexed and the methods are compressed. The cohort is deliberately selected - 109 patients whose breakpoints two PCR methods had already failed to resolve - so no population incidence can be read off it. Sixteen of the 109 remained rearrangement-negative and were set aside as 'unknown status' rather than resolved, so the denominator of 93 is itself a survivorship set. Single diagnostic centre (DCAL Frankfurt) with referral bias. Full text read (PMC6756029), not the abstract alone.
Evidence that argues against this (1)
- The same group's later and much larger series moves the 5' boundary of this region from intron 19 to intron 20 and reports its prevalence as 1.4% rather than 23%, without comment on either change. Both revisions go against the 2019 letter, so its boundary and its frequency are the earlier and smaller-sample of the two estimates. Meyer C et al. 2023
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The 2023 KMT2A recombinome maps BCR1 between intron 7 and exon 13 and the minor BCR (BCR2) between intron 20 and exon 24, with 3336/3401 (98%) of breakpoints in BCR1 and 47/3401 (1.4%) in BCR2; USP2 is the ninth most frequent partner overall at 30 patients.
The KMT2A recombinome of acute leukemias in 2023. · Meyer C, Larghero P, Almeida Lopes B, Burmeister T, Gröger D, et al., Marschalek R · Leukemia 2023What this evidence cannot show
Referral cohort accrued 2003-2022 at one diagnostic centre, so the partner spectrum reflects who sends samples there. The BCR1:BCR2 ratio is itself method-dependent (98:1.4 overall against roughly 94:6 in the targeted-NGS subset), so 1.4% is a floor rather than an estimate. Partner counts and BCR coordinates were read from the published article text and Table 1 in this session; the exon numbering is the DCAL convention used throughout this graph.
Evidence that argues against this (1)
- The 2019 letter from the same centre reports the region as starting at intron 19 and as accounting for 23% of breakpoints in a PCR-refractory cohort. Neither figure is wrong, but quoting 1.4% without the ascertainment qualifier understates how often a diagnostic laboratory will actually meet a BCR2 case among its unresolved samples. Meyer C et al. 2019
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Adding PHD fingers back into MLL-AF9 blocks immortalisation of haematopoietic progenitors and reduces Hoxa9 association and upregulation, which is the standing paradox any BCR2 entry has to address.
The PHD fingers of MLL block MLL fusion protein-mediated transformation. · Muntean AG, Giannola D, Udager AM, Hess JL · Blood 2008What this evidence cannot show
Murine retroviral immortalisation assay with MLL-AF9 only; PHD fingers were added back to a major-BCR fusion rather than tested in the context of a real BCR2 fusion protein, and no USP2 or AFDN construct was examined. The readout is immortalisation of progenitors in culture, not leukaemia in an animal. Abstract and the paper's central claim were read in this session; figure-level data were not re-derived.
Evidence that argues against this (1)
- The existence of 47 BCR2 patients, 30 of them USP2-fused, is itself the counter-observation: PHD retention demonstrably does not prevent leukaemia in humans, so the murine block is either partner-specific, dosage-dependent, or counteracted by something the partner supplies. Citing Muntean as a general rule over-reads it. Meyer C et al. 2023
SAME interval - this is the founding report of BCR2, not a third cluster; adds the intron 19 boundary and the 23% conditional ascertainment figure
the other, 98% cluster; unaffected by this entry - 67/93 of the 2019 discovery cohort still broke in BCR1
- part of →KMT2A / MLL1 — the gene that breaks
KMT2A intron 19/20 to exon 24, DCAL exon numbering
the fragility map exists for BCR1 only; no SAR, TOP2 or DNase-HS characterisation of BCR2 has been published
USP2 is the partner that defined the second/minor BCR; 17 of 21 minor-BCR cases in the discovery cohort
the fusion whose recurrence defined the second/minor BCR
the 23% conditional yield is the operational consequence of the cryptic phenotype
records the reading that the 2019 'second BCR' is the existing minor BCR, with the counts behind it