Oncofusion

start / graph / concept-neither-t411-product-changes-dna-damage-repair

Explainerstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q24-reciprocal-product

A tested negative: neither fusion product alters DNA double-strand break repair

also called MLL-AF4 DNA repair negative · AF4-MLL and DSB repair · no repair defect

People suspected these fusion proteins might make cells worse at fixing broken DNA, which would explain how more damage accumulates. Somebody tested it properly and the answer was no - both fusions repair DNA exactly like normal cells.

The same thing, for a scientist

MLL-AF4 and AF4-MLL were CRISPR/Cas9 knocked into the AAVS1 safe-harbour locus of HEK293 cells and tested for DNA double-strand break handling. Repair kinetics after etoposide and after ionising radiation were identical to wild type by flow cytometry, gamma-H2AX immunoblot and comet assay; levels of KU86, KU70, LigIII-alpha, WRN, PARP1 and RAD51 were unchanged; functional NHEJ and HR efficiencies were identical irrespective of genotype; and etoposide induced S/G2/M arrest independently of fusion expression [E1]. This is one of very few properly controlled negatives about the reciprocal product, and it is worth recording precisely because negatives of this quality are rare here. It also matters for the graph's framing of the silent mutational landscape of infant t(4;11) [E2]: the low secondary mutation burden is not explained by, and does not require, any repair phenotype of either fusion. The caveat is the usual one and the authors chose it deliberately - HEK293, not a haematopoietic cell - and an AAVS1 knock-in expresses from a heterologous locus rather than from the translocated allele.

Go deeper
Why a good negative deserves an entry

This project's standing rule is that an absence of testing is not a negative result. The converse also holds: a real, controlled negative is information, and it is invisible if nobody writes it down. Three orthogonal assays, a panel of repair proteins, functional NHEJ and HR reporters and a checkpoint readout, all concordant 1. Anyone proposing that t(4;11) blasts are repair-deficient - for instance to justify a PARP or ATR strategy - now has to explain this result rather than discover it.

What it does not exclude

It is a HEK293 knock-in, so it cannot exclude a repair phenotype that requires haematopoietic context, physiological expression level, or the presence of both fusions on their native alleles. And it tests repair of induced breaks, not the generation of the translocation itself, where topoisomerase II chemistry and the breakpoint cluster region are the established mechanism.

Facts and figures
resultno difference from wild type for either fusion
systemHEK293 with AAVS1 knock-in
classcontrolled negative
assays
[
  "flow cytometry repair kinetics",
  "gamma-H2AX immunoblot",
  "comet assay",
  "NHEJ functional reporter",
  "HR functional reporter",
  "cell-cycle checkpoint"
]
proteins unchanged
[
  "KU86",
  "KU70",
  "LigIII-alpha",
  "WRN",
  "PARP1",
  "RAD51"
]
How this could bite us back
  • A synthetic-lethal strategy premised on a repair defect in t(4;11) has no support from the one controlled test of that premise.
  • This graph carries a Fanconi-pathway dependency signal in KMT2A-rearranged lines; that signal has to coexist with a normal induced-break repair phenotype, and reconciling them is unfinished.
Still open
  1. Would the same assays give the same answer in a haematopoietic progenitor carrying an engineered balanced t(4;11)? The CRISPR platform to do this in primary human HSPCs now exists.
  2. Does the double-fusion state, which was not tested here, behave like either single fusion?
  3. Is the very low secondary mutation burden of infant t(4;11) explained by anything, if not by repair?
Every source behind this entry (2)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    MLL-AF4 and AF4-MLL CRISPR/Cas9 knocked into the AAVS1 locus of HEK293 cells showed repair kinetics for etoposide- and irradiation-induced double-strand breaks identical to wild type by flow cytometry, gamma-H2AX immunoblot and comet assay; no differences in KU86, KU70, LigIII-alpha, WRN, PARP1 or RAD51; identical functional NHEJ and homologous recombination efficiency irrespective of genotype; and etoposide-induced S/G2/M arrest independent of fusion expression.

    Expression of MLL-AF4 or AF4-MLL fusions does not impact the efficiency of DNA damage repair. · Castaño J, Herrero AB, Bursen A, et al. · Oncotarget 2016
    PMID 27119507DOI
    What this evidence cannot show

    HEK293 cells with AAVS1 knock-in, so expression is from a heterologous safe-harbour locus in a non-haematopoietic background; single-fusion genotypes rather than the double-fusion state that t(4;11) patients carry; a negative result whose power is not stated.

    Evidence that argues against this (1)
    • The same fusion pair has been shown to change the chromatin-accessibility landscape genome-wide in a comparable model system, so 'no phenotype' in one assay family in this cell type does not generalise to no phenotype. Wilhelm A et al. 2021
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Infant t(4;11) B-ALL shows the most silent mutational landscape reported for any sequenced paediatric cancer, with recurrent mutations found exclusively in KRAS and NRAS, subclonal and frequently lost at relapse.

    PMID 30679323DOI
    What this evidence cannot show

    Whole-genome and whole-exome analysis of 124 infants; a mutation-centric analysis cannot detect structural variation or deletion, so 'silent' is silent for the classes assayed.

    Evidence that argues against this (1)
    • A low mutation burden is consistent with both a normal repair apparatus and with a disease that simply does not need secondary mutations, so the genomic observation cannot by itself confirm the functional repair result. Castaño J et al. 2016
Connected to