Oncofusion

start / graph / hyp-second-hit-is-developmental-window-not-mutation

Hypothesismoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q1b-genesis2

Hypothesis: the second hit in infant KMT2A-rearranged leukaemia is a fetal developmental window, not a mutation

also called cell-of-origin as second hit · developmental permissiveness · why cord blood fails to transform

The fusion cannot turn ordinary newborn blood stem cells into leukaemia in the laboratory, yet in a fetus it apparently can, almost on its own. The proposal is that the missing ingredient is not another broken gene but the particular state of a cell that only exists for a window during pregnancy.

The same thing, for a scientist

This hypothesis reconciles the two halves of the tension recorded in `concept-one-thing-wrong-versus-quiet-genome`. The statement: transformation by KMT2A::AFF1 requires a target cell in a specific fetal developmental state, and that state — not a cooperating mutation — is the second requirement. It predicts that the repeated failure of MLL-AF4 to transform human cord-blood CD34+ cells [E1] is a wrong-cell result; that the five-fold higher mutation burden in older children with the same fusion [E2] represents mutational substitution for a cell state infants have for free; and that the near-total twin concordance and months-long latency in infants [E3] reflect a permissive window rather than an absent requirement. It is falsifiable in an unusually clean way: if an appropriately staged human fetal progenitor is transformed by the fusion alone, the hypothesis is supported, and if a cooperating mutation is still required in that cell, it is refuted. The hypothesis is not novel in spirit — the cell-of-origin literature in this disease has circled it for two decades — but it has not been written as a falsifiable statement with a specified experiment, and the substitution prediction in particular has never been tested.

Go deeper
The substitution prediction, which is the testable core1 deeper

If fetal cell state and adult mutations are alternative routes to the same permissive configuration, then introducing the specific epigenetic-regulator lesions found in non-infant KMT2A-rearranged leukaemia into an MLL-AF4-expressing postnatal progenitor should permit the transformation that FLT3 activation could not 12. This is a single, well-defined, affordable experiment. It has not been done, and a positive result would be the strongest evidence in the field that the second hit is a state that can be reached by more than one route.

Which state, specifically

The graph already holds candidates: bivalent primed chromatin at HOXA (cell-state-primed-bivalent-chromatin, mechanism-bivalent-primed-chromatin-fusion-targets), the fetal-type CD10-negative pre-pro-B compartment (cell-type-fetal-preprob-cd10-negative), fetal liver HSCs (cell-type-fetal-liver-hsc) and haemogenic endothelium (cell-type-haemogenic-endothelium). The hypothesis does not choose between them; specifying which one is the point of the first experiment.

Why this matters therapeutically, not just biologically

If the permissive state is required only for initiation, then it is gone by the time the child is treated and is irrelevant to therapy. If any residue of it is required for maintenance, it is a target, and it is a target that the low mutation burden otherwise denies this project. The distinction is exactly the initiation-versus-maintenance question that hyp-fusion-continuously-required-not-hit-and-run asks about the fusion, applied instead to the cell. Both should be answered with the same degron and cell-state tools.

Facts and figures
statementTransformation by KMT2A::AFF1 requires the target cell to be in a fetal developmental state whose chromatin configuration is permissive for the fusion's program; this state, and not a cooperating somatic mutation, is the second requirement in infants, and the epigenetic-regulator mutations found in older children are an alternative route to the same permissive configuration.
origindata-pattern
prior probability0.55
predictions
[
  "MLL-AF4 expressed in an appropriately staged human fetal liver haematopoietic progenitor transforms with no cooperating mutation, whereas the same construct in cord-blood CD34+ cells does not.",
  "Introducing an epigenetic-regulator lesion found in non-infant KMT2A-r leukaemia (CHD4, KDM6A or NSD1) into MLL-AF4-expressing cord-blood CD34+ cells permits transformation that FLT3 activation could not.",
  "The permissive state is characterised by bivalent H3K4me3/H3K27me3 at HOXA that is present in fetal progenitors and lost postnatally, and restoring bivalency pharmacologically in a postnatal progenitor partially restores permissiveness.",
  "Infant and older-child KMT2A-r leukaemias converge on the same transcriptional output despite reaching it by different routes."
]
experiments
[
  "exp-fetal-versus-cord-blood-transformation-window"
]
analyses
[]
falsifiers
[
  "An appropriately staged human fetal progenitor expressing MLL-AF4 alone still fails to transform, with adequate follow-up and verified fusion expression.",
  "A single cooperating mutation is found that permits transformation in cord-blood CD34+ cells and is also recurrently present in infant patient samples, which would restore the conventional two-hit account.",
  "Infant and older-child KMT2A-r leukaemias are shown to have fundamentally different transcriptional outputs, refuting convergence on a common permissive state."
]
How this could bite us back
  • If the second requirement is a developmental state that has closed by diagnosis, then nothing about it is druggable in a sick child, and the hypothesis explains the disease without helping the patient.
  • If it is partly maintained, targeting it means targeting a fetal-like state in an infant whose normal haematopoiesis is still closer to that state than an adult's, so the therapeutic window may be very narrow.
Still open
  1. What is the correct staging of the human fetal target cell, and is fetal tissue of that stage actually obtainable for research in the jurisdictions where this lab operates? If not, an iPSC-derived fetal-like progenitor is the fallback and its fidelity becomes the limiting assumption.
  2. Is the permissive state required for maintenance as well as initiation? If any residue persists in the diagnostic blast, it is a target; if not, this hypothesis is biologically important and therapeutically inert.
  3. Do infant and older-child KMT2A-r leukaemias converge transcriptionally? This is answerable today from published cohorts and would be the cheapest first test of the hypothesis.
Every source behind this entry (4)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    MLL-AF4 enhances engraftment and clonogenicity in human cord-blood CD34+ cells but does not immortalise or transform them, and FLT3 activation does not cooperate to do so.

    Ligand-independent FLT3 activation does not cooperate with MLL-AF4 to immortalize/transform cord blood CD34+ cells · Montes R, Ayllon V, Prieto C, Bursen A, Prelle C, Romero-Moya D, et al. · Leukemia 2014
    PMID 24240202DOI
    What this evidence cannot show

    A negative result. It is compatible with the hypothesis here (wrong cell) and equally compatible with the conventional reading (a cooperating mutation is missing). The paper's own conclusion is the conventional one, so this evidence supports the hypothesis only in combination with E2 and E3.

    Evidence that argues against this (1)
    • The authors themselves interpret the failure as requiring additional cooperating oncogenic lesions rather than a different cell, which is the competing hypothesis. Montes R et al. 2014
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Older children with KMT2A-rearranged leukaemia carry five times as many non-silent mutations as infants with the same fusion, including frequent epigenetic-regulator mutations that are nearly absent in infants.

    The landscape of somatic mutations in infant MLL-rearranged acute lymphoblastic leukemias · Andersson AK, Ma J, Wang J, Chen X, Gedman AL, Dang J, et al. · Nature Genetics 2015
    PMID 25730765DOI
    What this evidence cannot show

    20 older children versus 47 KMT2A-rearranged infants; the comparison groups differ in lineage as well as age, so 'age' and 'disease type' are partly confounded.

    Evidence that argues against this (1)
    • Directly weakens the premise the hypothesis rests on. Targeted deep sequencing of 84 KMT2A-rearranged infants found RAS-pathway mutations at 100% frequency and at the heaviest burden in the most refractory integrative cluster, with mutation number predicting prognosis. If infants in one cluster universally carry RAS-pathway lesions, then 'infants carry five times fewer mutations' is an average across clusters and across sequencing depths, and a mutational second hit is not absent - it is unevenly distributed and was under-detected at exome depth. Isobe T et al. 2022
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.review

    Near-total monozygotic twin concordance with short latency in infant KMT2A-rearranged leukaemia limits how much additional postnatal genetic work can be required.

    Leukemia in twins: lessons in natural history · Greaves MF, Maia AT, Wiemels JL, Ford AM · Blood 2003
    PMID 12791663DOI
    What this evidence cannot show

    Review of case series with publication bias toward concordant pairs; a discordant pair with spontaneous clearance was later documented.

    Evidence that argues against this (1)
    • The concordance premise is ascertained, and a population-based count is far lower - though the infant stratum survives. A California registry linkage (1982-2022) covering 255 twin pairs with leukaemia found only 2 concordant pairs: 0.8% overall, estimated monozygotic 2.6%, and for lymphoid leukaemia at ages 1-10 only 0.7% raw. The authors attribute higher published figures to ascertainment bias in purpose-built twin cohorts. In the same dataset, AML diagnosed under one year of age was 25% concordant raw and 100% on the monozygotic estimate, so the specific infant claim is NOT refuted - but the review's case-series basis for it is, and the entry should rest on the infant stratum rather than on 'near-total concordance' as a general twin fact. Nickels EM et al. 2025
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The fusion is already present at birth and disease follows within months, so whatever the second requirement is, it is satisfied either in utero or very early postnatally.

    Backtracking leukemia to birth: identification of clonotypic gene fusion sequences in neonatal blood spots · Gale KB, Ford AM, Repp R, Borkhardt A, Keller C, Eden OB, Greaves MF · Proceedings of the National Academy of Sciences of the USA 1997
    PMID 9391133DOI
    What this evidence cannot show

    Selected for backtrack-positive cases, which biases toward short latency.

    Evidence that argues against this (1)
    • Backtracking is bounded by its detection limit, and this hypothesis reasons from backtrack-positive cases. In the largest series applying one identical backtracking method to both non-infant childhood ALL and AML, assay sensitivity was stated explicitly - detection required 1 to 3 positive cells in the neonatal blood spot - and markers were recovered in 3 of 12 ALL patients and in 0 of 13 AML patients. The authors state the negative result is equally compatible with a genuinely lower rate of prenatal origin OR with a lower pre-leukaemic cell load at birth. A fusion 'already present at birth' is therefore a statement about clones above roughly one cell per blood spot, and the cases that anchor the short-latency argument are exactly the ones whose birth clone was large enough to see. Burjanivova T et al. 2006
Connected to