Hypothesis: what breaks KMT2A in a fetus is TOP2B doing its normal transcription-coupled job in a non-cycling progenitor, with no poison involved and repair capacity as the rate-limiting variable
also called fetal TOP2B hypothesis · unpoisoned TOP2B break · isoform hypothesis for de novo infant KMT2A-r
Therapy-related KMT2A leukaemia happens because a drug jams a DNA-cutting enzyme. Infants get the same disease before birth with no drug anywhere, and this hypothesis says the answer is the same enzyme doing its ordinary job in a cell that is not dividing - so the thing that differs in the rare fetus is not the cutting, but how well the cut is cleaned up.
The same thing, for a scientist
**The claim.** In the human fetal haematopoietic progenitor, double-strand breaks in the KMT2A breakpoint cluster region are made by **TOP2B** acting in its normal transcription-coupled capacity - not by a poison, not by TOP2A, and not by replication - and the rare progression from break to translocation is governed by the cell's adduct-removal and end-joining capacity rather than by the break rate. **Why this is worth stating rather than assuming.** The graph's `mechanism-top2-cleavage-at-kmt2a` calls the TOP2 account 'a well-supported mechanism for therapy-related KMT2A-r leukaemia and an unproven extrapolation for de novo infant KMT2A::AFF1', and its open questions ask whether TOP2B occupancy at the BCR is higher in fetal liver progenitors. Nobody has stated the fetal version as a hypothesis with falsifiers. This does. **What supports it.** TOP2B is the isoform whose depletion removes most etoposide-induced KMT2A breakage [`gene-top2b` E2]; TOP2B, unlike TOP2A, is expressed in post-mitotic and non-proliferating cells in human fetal tissue [`gene-top2b` E6, E7]; the conversion of a trapped TOP2 complex into an irreversible break is about two-fold faster at transcribed loci and requires TDP2 and the proteasome [`gene-tdp2` E2 counter]; and KMT2A rearrangement bands appear in cytokine-stimulated normal CD34+ cells **with no genotoxin at all** and more at day 7 than in etoposide-treated cells [`exp-human-cd34-kmt2a-breakage-assay`]. **What argues against it.** The graph's break-machinery beat concludes that breaking KMT2A is common and leukaemia is rare, and that the rare step is not the cut - which, if right, makes the identity of the cutting enzyme less important than this hypothesis assumes. Umbilical cord blood from preleukaemic-fusion-positive newborns shows no excess endogenous DNA damage [`gene-h2ax` E2]. Direct mapping of etoposide-induced breaks across the MLL BCR placed them at a Pol II binding site and explicitly **not** at a TOP2 consensus sequence [E4]. And a locus-level negative exists for at least one route: apoptotic cleavage within the KMT2A BCR is not blocked by the catalytic TOP2 inhibitor merbarone. **Prior: 0.25.** Most of that is the chance that TOP2B is one contributing route among several, which is the field's own synthesis position, rather than the chance that it is the route.
The three links, and which one is actually missing
- TOP2B is present and catalytically engaged in a human fetal liver HSC or LMPP. Not measured. The two fetal expression studies this graph now cites cover brain, kidney, liver, gut, lung and placenta, and neither resolves haematopoiesis - the fetal liver signal is attributed to hepatocytes [
gene-top2bE6]. - TOP2B is engaged at the KMT2A BCR in that cell. Not measured anywhere, in any cell type, without a poison.
- The rate-limiting step is downstream of the cut. Partially supported: KMT2A rearrangement occurs in untreated normal CD34+ cells, 49% of healthy adults carry detectable KMT2A rearrangements in peripheral lymphocytes with no expressed fusion transcript, and TDP2 loss increases rather than decreases translocation.
Link 2 is the whole hypothesis and it has never been attempted. The reason is technical rather than conceptual: TOP2 engagement is normally measured by trapping it with a poison, which is precisely the perturbation this hypothesis says is absent in the fetus. The experiment that dissolves that circularity is isoform-resolved degradation plus poison-free break mapping, written up as exp-isoform-degron-break-map-kmt2a-fetal-progenitor.
What this hypothesis is NOT claiming
It is not claiming that TOP2B inhibition is a preventive strategy. hyp-block-kmt2a-break-machinery-upstream argues at length that the prevention idea is probably wrong - the window is fetal, the base rate of breakage is enormous, the arithmetic is roughly one in 38,000 pregnancies to prevent one case at 100% efficacy, and every candidate target is essential. TOP2B is a worse target than most: it is required for post-mitotic cell survival.
It is also not claiming that the de novo and therapy-related breaks are at different positions. That dichotomy is recorded as contested in mechanism-top2-cleavage-at-kmt2a, and this hypothesis is agnostic about it - indeed it is more comfortable if the positions are the same, since the same enzyme would be making both.
Its value is diagnostic and explanatory: if the fetal break is TOP2B-made, then the RAG and AID routes this graph holds in concept-programmed-dna-breaks-vdj-and-csr are not needed to explain it, the infant age peak becomes a question about fetal TOP2B expression and fetal repair capacity, and the junction-class distribution becomes predictive.
| statement | In the human fetal haematopoietic progenitor, double-strand breaks in the KMT2A breakpoint cluster region are generated by TOP2B acting in its normal transcription-coupled capacity without any pharmacological poison, and the rare progression from break to translocation is governed by adduct removal and end-joining capacity rather than by break rate. |
|---|---|
| origin | literature-gap |
| prior probability | 0.25 |
| distinguishing alternative | The break-frequency-plus-selection model this graph already holds: KMT2A breaks constantly by several routes (apoptotic nuclease, replication-transcription conflict, RAG/AID side reactions, TOP2), the identity of the nuclease is not rate-limiting, and what is rare is a fetal progenitor in which the resulting fusion confers self-renewal. Under that model every prediction above can be true and the hypothesis still explains nothing about incidence. |
| what it would change if true | The infant age peak becomes a question about fetal TOP2B expression and fetal adduct-removal capacity, both measurable. The junction-class distribution becomes an aetiological readout. And the graph stops needing RAG or AID to explain the fetal break, which frees concept-programmed-dna-breaks-vdj-and-csr to be about lineage rather than about origin. |
predictions
[ "TOP2B protein is detectable and chromatin-associated in human fetal liver HSC/LMPP, and TOP2A is low or absent in the non-cycling fraction", "TOP2B, not TOP2A, occupies the KMT2A BCR in a fetal-like human progenitor in the absence of any poison", "Acute TOP2B degradation reduces poison-free DSB formation at the KMT2A BCR; acute TOP2A degradation does not", "Poison-free breaks at the BCR fall at transcribed positions and correlate with Pol II occupancy rather than with a TOP2 consensus sequence", "TDP2 or proteasome inhibition raises the persistence of those breaks without raising their initial rate", "Junction classes in de novo infant cases include the 4 bp perfect-homology subclass at a rate comparable to therapy-related cases, because the same enzyme and the same adduct-removal step made both" ]
experiments
[ "exp-isoform-degron-break-map-kmt2a-fetal-progenitor", "exp-human-cd34-kmt2a-breakage-assay" ]
analyses
[ "ana-junction-microhomology-signature-mining", "ana-dsb-mapping-assays-at-the-kmt2a-locus" ]
falsifiers
[ "TOP2B protein is absent from human fetal liver HSC/LMPP - kills the hypothesis at link 1 for the cost of one antibody", "Acute TOP2B degradation leaves poison-free DSB formation at the KMT2A BCR unchanged while TOP2A degradation reduces it - the isoform assignment is inverted", "No poison-free DSBs are detectable at the KMT2A BCR above background in any haematopoietic progenitor at the sensitivity of END-seq - the fetal break needs a trigger after all, and the hypothesis is about the wrong steady state", "Degrading both isoforms together leaves KMT2A rearrangement frequency in untreated cytokine-stimulated CD34+ cells unchanged - TOP2 is not the source of the unpoisoned break this graph already observes", "De novo infant junctions show a 4 bp perfect-homology fraction indistinguishable from zero while therapy-related junctions do not - the two aetiologies use different end chemistry" ]
- This hypothesis names an enzyme whose inhibition is not available, not selective, and would be catastrophic in a fetus. Nothing follows from it therapeutically and it should not be read as a prevention argument; `hyp-block-kmt2a-break-machinery-upstream` sets out why that whole class of argument is probably wrong.
- Its most likely failure mode is being half-right in a way that is unfalsifiable: TOP2B contributes some fraction of fetal KMT2A breaks alongside apoptotic nucleases, replication-transcription conflict and programmed-break side reactions. An apportionment is what the field needs and no experiment in this entry delivers one.
- If the hypothesis is true it strengthens rather than weakens the case for not changing etoposide practice, because it says the fetal break happens without any drug at all - which is an argument that the drug is not the whole story in either setting.
- Is there any assay for TOP2 engagement that does not require trapping the enzyme with a poison? The circularity at the centre of this hypothesis is methodological, and a poison-free engagement assay would be more valuable than the hypothesis.
- Does the human fetal liver HSC transcribe KMT2A at all, and at what rate relative to an adult bone marrow HSC? Transcription-coupled anything requires transcription, and nobody in this graph has cited a measurement of KMT2A transcriptional output across developmental stages.
- If TOP2B makes the break in the fetus and also in the etoposide-exposed patient, why is the infant disease lymphoid and the therapy-related disease predominantly myeloid? The hypothesis pushes that difference entirely downstream of the break, into cell of origin - which is where this graph's genesis beat already puts it, and which makes the prediction that the two should share junction chemistry.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Most etoposide-induced chromosome breaks in the MLL/KMT2A locus and the overall genotoxicity of etoposide depend on topoisomerase II-beta, but TOP2A and TOP2B occupancy and etoposide-induced cleavage data indicate that factors other than local topoisomerase II concentration determine breakpoint clustering.
Model for MLL translocations in therapy-related leukemia involving topoisomerase IIβ-mediated DNA strand breaks and gene proximity. · Cowell IG, Sondka Z, Smith K, Lee KC, Manville CM, Sidorczuk-Lesthuruge M, et al. · Proc Natl Acad Sci U S A 2012What this evidence cannot show
Human transformed cell lines with pharmacological etoposide. The isoform attribution is made in the presence of a poison, which is exactly the condition this hypothesis says does not obtain in the fetus, so the transfer is an assumption.
Evidence that argues against this (1)
- Fusion formation at the recurrent hotspots requires BOTH TOP2 isoforms and transcriptional elongation, so a TOP2B-only account of the productive event is narrower than the data even in the therapy-related setting the isoform claim came from. Gothe HJ et al. 2019
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In human foetal tissues, topoisomerase II alpha mRNA is expressed most strongly in zones of proliferation while topoisomerase II beta is expressed in a broader range of cell types; in the early human foetal telencephalon TOP2A tracks KI67 and the G2/M fraction while TOP2B is expressed in post-mitotic compartments with KI67 co-expression rare.
Distinct expression patterns for type II topoisomerases IIA and IIB in the early foetal human telencephalon. · Harkin LF, Gerrelli D, Gold Diaz DC, Santos C, Alzu'bi A, Austin CA, et al. · J Anat 2016What this evidence cannot show
Human fetal brain and, in the companion in-situ study, kidney, gut, liver, lung and placenta. Haematopoiesis is resolved in neither. Applying the isoform-versus-cell-cycle relationship to a fetal liver HSC is the central untested assumption of this hypothesis and is stated as such.
Evidence that argues against this (1)
- Transcript abundance does not determine which isoform is catalytically engaged: HSF1 has been reported to stimulate the catalytic engagement of TOP2B over TOP2A, so a signalling state can invert the isoform balance independently of expression. An expression argument is therefore the weakest form the isoform claim can take. Konada L et al. 2026
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
TOP2 chromatin localisation and etoposide trapping are transcription-independent but require pre-existing cohesin binding, while conversion of trapped cleavage complexes into irreversible TOP2-linked double-strand breaks is accelerated about two-fold at transcribed loci and depends on proteasomal degradation and TDP2 phosphodiesterase activity; cohesin binding and transcriptional activity alone predict the kinetics of TOP2-induced DSBs.
Topoisomerase II-Induced Chromosome Breakage and Translocation Is Determined by Chromosome Architecture and Transcriptional Activity. · Canela A, Maman Y, Huang SN, Wutz G, Tang W, Zagnoli-Vieira G, et al. · Mol Cell 2019What this evidence cannot show
Cultured human cells with etoposide throughout. The two-step separation is the most useful part for this hypothesis - it says transcription controls the conversion rather than the cutting - but every step was measured under a poison.
Evidence that argues against this (1)
- If the conversion step requires proteasomal degradation and TDP2, then in the absence of a poison there may be no persistent cleavage complex to convert at all, and the fetal break would have to come from somewhere else. This is the strongest single argument against the hypothesis and the experiment is designed to address it by mapping breaks without a poison. Gómez-Herreros F et al. 2017
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Direct mapping of etoposide-induced double-strand breaks across the MLL breakpoint cluster region placed them at the RNA polymerase II binding site and the transcription initiation region of a gene-internal MLL promoter and explicitly not at a nearby topoisomerase II consensus sequence, leading the authors to attribute MLL's instability to transcriptional processes.
Transcription linked to recombination: a gene-internal promoter coincides with the recombination hot spot II of the human MLL gene. · Scharf S, Zech J, Bursen A, Schraets D, Oliver PL, Kliem S, et al. · Oncogene 2007What this evidence cannot show
Cell lines and a 2007 mapping resolution well below what END-seq or sBLISS would give today. The authors' preferred interpretation is transcription rather than TOP2 - which is a partial ally of this hypothesis (transcription-coupled) and an opponent of its isoform half (the breaks are not at a TOP2 consensus).
Evidence that argues against this (1)
- TOP2 does not require a strong consensus sequence to cleave, and TOP2B occupancy at transcribed loci is directed by cohesin and by Pol II rather than by sequence, so the absence of breaks at a consensus site is weak evidence against TOP2 involvement. The finding does argue against the specific 'TOP2 consensus site drives breakpoint position' version of the model, which this hypothesis does not hold. Canela A et al. 2017
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Umbilical cord blood samples positive for AML preleukaemic fusion genes including KMT2A-MLLT3 (screen-positive rate 1.17%) accumulated the same level of endogenous DNA damage by gamma-H2AX/53BP1 focus formation, the same reactive oxygen species level and the same apoptosis as fusion-negative controls, and 50 cGy of gamma-rays did not induce these fusions in sorted and expanded HSPC subsets.
Induction of AML Preleukemic Fusion Genes in HSPCs and DNA Damage Response in Preleukemic Fusion Gene Positive Samples. · Kosik P, Durdik M, Skorvaga M, Klimova D, Kochanova D, Cerna Z, et al. · Antioxidants (Basel) 2021What this evidence cannot show
Cord blood at birth, not fetal liver during the window when the break occurs; focus counting is a population average that would miss a small high-damage subpopulation; RT-qPCR fusion screening is artefact-prone.
Evidence that argues against this (1)
- This result is a genuine argument against the hypothesis's weakest reading - that fusion-positive fetuses are globally more damage-prone - and the hypothesis does not make that claim. It claims a locus-specific enzymatic route, which a genome-wide focus count cannot see either way. The distinction is real but it is also convenient, and a reader should weigh it as such. Škorvaga M et al. 2018
- E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In human haematopoietic stem and progenitor cells, quercetin acts through both TopoII-dependent and TopoII-independent routes - suppressing NHEJ and HR downstream of MLL cleavage via PI3K inhibition - and etoposide plus quercetin produces more MLL rearrangements than either alone; a tissue-engineered placental barrier established the extent of quercetin transplacental transfer to the fetal compartment.
Quercetin alters the DNA damage response in human hematopoietic stem and progenitor cells via TopoII- and PI3K-dependent mechanisms synergizing in leukemogenic rearrangements. · Biechonski S, Gourevich D, Rall M, Aqaqe N, Yassin M, Zipin-Roitman A, et al. · Int J Cancer 2017What this evidence cannot show
High continuous concentrations in cultured human HSPCs; the placental transfer measurement is an engineered barrier model. Cited here for the structural point - that a maternally ingested compound can reach fetal HSPCs and that its dominant genotoxic arm in that setting was repair suppression rather than topoisomerase poisoning - and not as support for a dietary aetiology.
Evidence that argues against this (1)
- The epidemiological form of the dietary hypothesis has been tested at scale and was essentially null: in 240 incident infant acute leukaemias against 255 controls there was no association between a maternal dietary topoisomerase-II-inhibitor index and MLL-rearranged ALL, and maternal fruit and vegetable intake was inversely associated with risk. Spector LG et al. 2005
the degron experiment written to falsify it
names TOP2B as the fetal cutting enzyme; the isoform claim is the hypothesis
states the fetal version of a mechanism that entry calls an unproven extrapolation
- related to →Programmed DNA breaks - V(D)J and class switching - and the verdict on whether either one breaks KMT2A
if true, RAG and AID are not needed to explain the fetal break and become questions about lineage instead
the cell in which link 1 and link 2 are both unmeasured
cord-blood gamma-H2AX/53BP1 foci show no excess endogenous damage in fusion-positive newborns
identifies an enzyme without proposing a prevention strategy; that entry sets out why the class of argument is probably wrong