KMT2A intron 9 / exon 9 breakpoints (BCR subregion A): the HOXA-high, older-patient break
also called MLL intron 9 break · 5' BCR break · KMT2A e9 junction · BCR subregion A
One of the two hot spots inside the KMT2A break region sits in intron 9, near the front of the gene. Leukaemias that break here tend to arise in older children and adults and to switch on the HOXA gene cluster strongly.
The same thing, for a scientist
Breaks in KMT2A exon 9 + intron 9 (1761 bp; BCR subregion A) account for 37.2% of all 3401 genomically characterised KMT2A rearrangements [E1]. Intron 9 contains four Alu repetitive elements, three of which are transcriptionally active, a plausible substrate for Alu-mediated and transcription-associated breakage [E1]. Clinically and biologically this is the 'upstream' break: AML breakpoints preferentially occur in intron 9 and this preference declines only slightly with age; KMT2A::AFDN breaks occur mostly in intron 9; KMT2A::AFF1 breaks shift from intron 11 in infants to introns 9-10 in adults; paediatric KMT2A::MLLT10 and KMT2A::ELL shift towards intron 9 [E1]. In t(4;11) the intron 9-10 group is the HOXA-high, non-infant group, whereas the intron 11-12 group is the HOXA-low, infant group [E2], and the intron 9-10 group has the better reported outcome [E3].
An upstream break makes a longer KMT2A portion but the same functional modules1 deeper
A break in intron 9 means the mRNA's last complete KMT2A exon is exon 9; a break in intron 11 means it is exon 11. Both retain AT-hooks, the SNL/menin-binding region and CXXC, and both lose all PHD fingers, the bromodomain, FYRN/FYRC, TAD, WIN and SET 1. So the difference is not a difference in named domains on the direct fusion; it is (a) ~two exons of additional linker/spacer sequence, and (b) a large difference in what the reciprocal protein looks like, because a break upstream of exon 11 keeps PHD1 structurally intact on the reciprocal product whereas a break within or downstream of exon 11 does not 1.
This is the mechanistic crux of the PI's question. Published evidence that breakpoint position changes biology is real, but the proposed mechanism runs mostly through the reciprocal fusion protein and through the transcriptional signature (HOXA), not through gain or loss of a domain on the direct oncoprotein.
HOXA-high vs HOXA-low
Trentin et al. profiled t(4;11) ALL against normal bone marrow and defined a 186-gene core signature, then split the patients using four discriminators (HOXA expression, age at diagnosis, fusion transcript type, chromosomal breakpoint). The split was clean: infants with low HOXA expression and MLL breakpoints in introns 11/12, versus non-infants with high HOXA expression and breakpoints in introns 9/10 2. Meyer et al. reach the same conclusion in the much larger recombinome cohort, stating that patients with breaks in intron 9 / exon 10 / intron 10 tend to display strong HOXA signatures while breaks within or downstream of exon 11 display a HOXA-low signature 1.
This matters therapeutically because HOXA9/MEIS1 induction is the canonical readout of menin-dependent KMT2A-fusion target-gene activation, and menin-inhibitor response in preclinical models is typically measured as HOXA9/MEIS1 collapse.
Breakpoint distribution by partner and age
From the DCAL 3401-patient cohort 1:
- Mean distribution across all patients: subregion A (exon 9 + intron 9) 37.2%.
- KMT2A::AFF1: infants break predominantly in intron 11; adults shift to introns 9 and 10.
- KMT2A::AFDN: breaks mostly in intron 9.
- KMT2A::MLLT10: paediatric cases shift towards intron 9.
- KMT2A::ELL: paediatric cases prefer intron 9 (with a later shift to intron 11 in older groups).
- KMT2A::MLLT11: shifts towards introns 9-10. KMT2A::MLLT6: shifts towards introns 8-9.
- By lineage: AML patients preferentially display an intron 9 break, decreasing slightly with age; ALL patients switch from an intron 11 preference (infants <6 months) to an intron 9 preference thereafter.
- KMT2A::MLLT1 and KMT2A::SEPTIN6 are the exceptions that prefer intron 11.
In an independent small series of 13 paediatric leukaemias with complex/cryptic karyotypes, intron/exon 9 was also the commonest breakpoint region (3/8 informative cases), ahead of exon/intron 11 and 10 4.
| gene | KMT2A |
|---|---|
| cytoband | 11q23.3 |
| genome build | hg38 |
| region | KMT2A exon 9 + intron 9 (BCR1 subregion A) |
| length bp | 1761 |
| frequency PCT | 37.2 |
| population | all KMT2A-rearranged acute leukaemia, n=3401 (DCAL 2003-2022) |
| exon last kmt2a mrna | 9 |
| transcriptional signature | HOXA-high |
| reported outcome | better than intron 11 breaks in infant AL, but on n=30 breakpoint-typed cases only |
| reciprocal product note | Breaks upstream of exon 11 leave PHD1 structurally intact on the reciprocal partner::KMT2A protein |
| numbering convention | Meyer/Marschalek (DCAL) KMT2A exon numbering |
repair features
[ "4 Alu repeats in intron 9, 3 transcriptionally active" ]
enriched in
[ "AML (all age groups)", "KMT2A::AFDN", "adult KMT2A::AFF1", "paediatric KMT2A::MLLT10", "paediatric KMT2A::ELL", "KMT2A::MLLT11 (introns 9-10)", "KMT2A::MLLT6 (introns 8-9)" ]
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" ]
- If HOXA-high disease is the menin-inhibitor-responsive subset, HOXA-low (intron 11, infant) disease - the largest and worst-prognosis group - may be the least responsive, which is the opposite of what the clinical need requires.
- Prognostic claims based on breakpoint position rest on ~30 typed patients. Building risk stratification on that would be premature and could misassign therapy.
- Is the HOXA-high / intron-9 association causal, or is it entirely explained by age and cell of origin? A breakpoint-matched comparison of infant vs non-infant t(4;11) with the same breakpoint is the missing experiment and has never been published at adequate n.
- If intron 9 breaks give HOXA-high disease, are they MORE menin-inhibitor sensitive (more HOXA to collapse) or LESS (more entrenched program)? No trial has reported menin-inhibitor response stratified by KMT2A breakpoint position.
- Does the Alu content of intron 9 mean these breaks are made by Alu-mediated homologous or microhomology-mediated recombination rather than by topoisomerase-II-associated breakage, which is an intron 11 feature?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Exon 9 + intron 9 (1761 bp) carries 37.2% of all KMT2A breakpoints (n=3401); intron 9 contains four Alu elements (three transcriptionally active); AML and KMT2A::AFDN break preferentially there; KMT2A::AFF1 shifts from intron 11 in infants to introns 9-10 in adults; intron 9/exon 10/intron 10 breaks associate with a HOXA-high signature.
The KMT2A recombinome of acute leukemias in 2023. · Meyer C, Larghero P, Almeida Lopes B, et al. · Leukemia 2023What this evidence cannot show
Single-centre referral cohort; breakpoint-by-partner-by-age subgroup analyses are descriptive chi-square comparisons against the cohort mean distribution, not multivariable models, and no correction for multiple testing across ~13 partners x 3 age classes x several subgroups is reported. The HOXA-signature statement in the Discussion is an interpretation rather than a new measurement in that paper.
Evidence that argues against this (2)
- Counting Alu elements in an intron does not establish that the breaks happen in them. In a KMT2A rearrangement mapped to base-pair resolution by an independent group, every breakpoint region was rich in LINE/L1 and SINE/Alu repeats, yet all breakpoints were exclusively identified OUTSIDE those repetitive sequences. Repeat density marks the neighbourhood rather than the cut site, so four Alu elements in intron 9 is a description of the sequence, not a mechanism for why intron 9 is broken in 37.2% of cases. Cerveira N et al. 2010
- The age at which the intron-9/intron-11 preference is said to switch is a partitioning choice, and it has moved. An earlier genomic analysis of 40 t(4;11) patients from the same laboratory lineage split the cohort at ONE YEAR of age rather than at about six months, and defined two hotspot areas on that split; the fusion junctions showed deletions, inversions and duplications of parental sequence, an error-prone end-joining signature rather than homologous recombination between repeats. Both the cross-over age and the mechanism inferred from it depend on how the cohort was cut. Reichel M et al. 2001
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In paediatric t(4;11) ALL, patients split into infants with low HOXA expression and MLL breakpoints in introns 11/12 versus non-infants with high HOXA expression and MLL breakpoints in introns 9/10.
Two independent gene signatures in pediatric t(4;11) acute lymphoblastic leukemia patients · Trentin L, Giordan M, Dingermann T, Basso G, Te Kronnie G, Marschalek R · European Journal of Haematology 2009What this evidence cannot show
Small microarray cohort (the method was explicitly designed for small cohorts), normalised pairwise against only n=3 normal bone marrow samples; age and breakpoint are strongly confounded in t(4;11) so the study cannot separate 'HOXA-low because infant' from 'HOXA-low because intron 11'. Shares a senior author with the recombinome series, so these are not fully independent lines of evidence.
Evidence that argues against this (1)
- HOXA-positive and HOXA-negative subgroups of KMT2A-rearranged leukaemia have also been attributed primarily to partner gene and to cell of origin rather than to breakpoint position, and the infant/non-infant split alone could explain the signature difference. Meyer C et al. 2023
- E3weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper
In infant acute leukaemia, patients with poor outcome preferentially displayed chromosomal breakpoints within MLL intron 11, implying that upstream (intron 9-10) breaks carry the better outcome.
The distribution of MLL breakpoints correlates with outcome in infant acute leukaemia · Emerenciano M, Meyer C, Mansur MB, Marschalek R, Pombo-de-Oliveira MS · British Journal of Haematology 2013What this evidence cannot show
This is the single most-cited source for 'breakpoint position predicts outcome' and it is thin: of 545 childhood leukaemia cases, the genomic breakpoint location was determined in only 30 KMT2A-rearranged cases. The survival comparison by breakpoint is therefore based on tens of patients, was not the primary endpoint, and is not adjusted for age at diagnosis (<=6 months was itself strongly prognostic, P<0.001) or for partner gene. It has not, to our knowledge, been replicated in a prospective trial cohort.
Evidence that argues against this (1)
- Meyer et al. themselves write that whether the partner-specific breakpoint shifts have an impact on clinical outcome 'is yet unclear', i.e. the largest available dataset declines to confirm a breakpoint-outcome association. Meyer C et al. 2023
- E4weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper
In an independent series of 13 paediatric leukaemias with complex or cryptic karyotypes, the most common KMT2A breakpoint region was intron/exon 9 (3 of 8 informative patients), followed by intron/exon 11 and 10.
Molecular characterization of KMT2A fusion partner genes in 13 cases of pediatric leukemia with complex or cryptic karyotypes · Ney Garcia DR, de Souza MT, de Figueiredo AF, et al. · Hematological Oncology 2016What this evidence cannot show
n = 13 patients, 8 with an assignable breakpoint region; selected for complex/cryptic karyotype so not representative of KMT2A-r leukaemia generally.
Evidence that argues against this (1)
- A 13-patient series typed only inside the classical breakpoint cluster region cannot rank breakpoint regions for KMT2A rearrangements generally, because a whole class of them breaks outside it. The KMT2A partial tandem duplication - found in 11% of acute myeloid leukaemia with normal cytogenetics - fuses intron 6 or intron 8 to intron 1, and in seven of nine sequenced cases the junction was a heteroduplex between homologous Alu elements rather than a discrete break, i.e. Alu-mediated homologous recombination rather than the end-joining assumed for intron 9. Neither the position nor the mechanism of those events is visible to an intron-9-versus-intron-11 tally. Strout MP et al. 1998
Counter-evidence shared by both entries: Alu density marks the neighbourhood, not the cut site - every breakpoint in a base-pair-mapped KMT2A rearrangement fell OUTSIDE the LINE/Alu repeats its region is rich in (Cerveira 2010, PMID 20920256).
Counter-evidence to an Alu-fragility reading: in a base-pair-mapped KMT2A rearrangement, all breakpoints fell OUTSIDE the LINE/Alu repeats their regions are rich in (Cerveira 2010, PMID 20920256), and Alu poly-A polymorphisms across the 8.3 kb BCR occur at equal frequency in t-AML patients and controls (Echlin-Bell 2003, PMID 12665971). Alu-mediated homologous recombination IS documented in KMT2A - but for the partial tandem duplication, joining intron 6/8 to intron 1 (Strout 1998, PMID 9482895).
- ← related toEach partner prefers a different KMT2A exon: partner-specific break position in 3,401 patients
Counter-evidence (Zhang 2004, PMID 15334549): with partner and aetiology both held constant -- six therapy-related t(11;16)/CREBBP cases, all topo-II-inhibitor exposed -- breaks still split between intron 11 near the topo II site and intron 9 in Alu repeats. Intra-partner heterogeneity is not obviously smaller than the inter-partner 'preference' being reported.
- ← related toEach partner prefers a different KMT2A exon: partner-specific break position in 3,401 patients
AFDN, pediatric MLLT10 and adult AFF1 prefer intron 9 or upstream
subregion A, 37.2% of breaks
older, de novo, HOXA-high break class
- ← related toRolf Marschalek & Claus Meyer (DCAL, Goethe-University Frankfurt) — the world's KMT2A breakpoint dataset
nucleotide-resolution breakpoint characterisation by long-distance inverse PCR across 3,401 patients
intron 9 versus intron 11 is the pre-specified binary breakpoint classification the analysis is powered for