Oncofusion

start / graph / exp-q192-rpa-chip-at-the-kmt2a-breakpoint-cluster-in-a-kmt2a-rearranged-blast

Experimentmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q192-the-rpa-heterotrimer

Repeat the one published RPA-at-the-KMT2A-breakpoint experiment in a haematopoietic cell, with an RPA-dose arm - the protocol is fully specified in a 2010 paper whose every arm used fibroblasts and a CML line, and whose primer sets, dose and controls are all published

also called RPA ChIP at the KMT2A BCR · the Sun 2010 replication with an RPA-dose arm

One paper has ever measured the DNA-coating protein RPA sitting on the part of the KMT2A gene that breaks to cause leukaemia. It used skin-derived cells and a chronic-leukaemia line, and it was published in 2010. The experiment should be repeated in the blood cells this disease actually arises in, with an extra arm that raises and lowers RPA to see whether the amount of RPA changes how often the break turns into a translocation.

The same thing, for a scientist

**Design: a replication plus one discriminating arm.** Sun et al. 2010 supplies the entire protocol - **six primer sets (t2, t4, bbt56, bt56, t56, in14) around the ~20 bp translocation breakpoint hotspot inside the 8.3 kb KMT2A BCR**, a **beta-globin control region at 11p15**, **etoposide 100 uM for 10 min**, ChIP at 0 and 30 minutes of recovery, an ATM-deficient/proficient pair, a **KU55933 pharmacological ATM arm**, and a **bleomycin negative-agent arm** [E1]. **Every arm of that paper was an SV40-transformed fibroblast pair or the CML-derived line BV173. None was a haematopoietic progenitor and none carried a KMT2A fusion** [E1]. **Arm 1 - replication in the right cell.** Run the identical ChIP-qPCR in (a) a KMT2A-rearranged B-lineage line, (b) a KMT2A-germline B-lineage line and (c) human CD34+ progenitors. **Cost: low.** This alone is worth doing, because a negative would show the founding observation is a fibroblast property. **Arm 2 - the RPA-dose arm, which discriminates two models.** Partial RPA1 or RPA2 knockdown to 40-60% of control protein (quantified western, viability matched within 10%) and RPA over-expression, each during the etoposide pulse, reading out **BCR junction frequency by LDI-PCR and split-MLL FISH** [E2, E3]. `hyp-the-kmt2a-breakpoint-cluster-is-an-rpa-loading-site-...` predicts a monotone dose dependence; the authors' own **excessive-RAD51-loading** model predicts **none**, because RPA is not rate-limiting in it [E1]. **Cost: medium.** **Readout hygiene.** Chromatin-bound RPA INTENSITY (the exhaustion readout) and RPA2 FOCI (the resection readout) are different measurements that look alike; both should be recorded and labelled separately [E3].

Go deeper
Why this is cheap: nothing about it needs inventing

This is a rare case where the instrument is fully published and simply has never been pointed at the right cell.

componentalready specified by PMID 21048951
primer setssix, named t2, t4, bbt56, bt56, t56, in14, positioned around the ~20 bp hotspot
control locusbeta-globin, chromosome 11p15
damageetoposide 100 uM, 10 min, DMSO vehicle at 0.1%
timepointsimmediately after removal, and 30 min recovery
negative agentbleomycin 30 ug/ml, 60 min - does not produce the effect, which is a real specificity control
ATM armgenetic pair plus KU55933 10 uM, 2 h pre-treatment
translocation readoutdual-colour break-apart FISH (split signal >1 um) and long-distance inverse PCR on BamHI-digested self-ligated genomic DNA
amplification controlreal-time PCR efficiency of input DNA compared before and after etoposide, all six primer sets

Two things are NOT specified and must be fixed before running: the primer sequences are listed in the paper's Methods as empty placeholders in the PMC rendering, so they have to be recovered from the publisher's version or redesigned against the BCR coordinates; and the antibody clone is not given (named 'anti-RPA34 (LabVision)' in Methods and 'anti-RPA32' in the legends), and LabVision's catalogue has since migrated 1.

What each possible result means, pre-registered

Arm 1.

  • RPA enriches at the BCR after etoposide in a haematopoietic cell, above 1.5x the beta-globin control at t56/bt56 -> the founding observation transfers, and the locus is real in the relevant lineage. Everything downstream becomes worth doing.
  • No enrichment -> the 2010 result is a fibroblast property, hyp-the-kmt2a-breakpoint-cluster-is-an-rpa-loading-site-... is refuted at step one, and the corpus's absence entries about resection factors at this locus recover their generality for RPA too.
  • Enrichment in the KMT2A-rearranged line but not the germline line -> a fusion-dependent effect, which nobody has proposed and which would be the most interesting outcome available. Pre-registered third option: neither - enrichment in both lines with no difference, which is the likeliest result and which answers the transfer question while saying nothing about the fusion.

Arm 2.

  • Junction frequency rises on RPA knockdown and falls on over-expression -> reserve-dependent; the hypothesis survives and the RAD51-excess model is incomplete.
  • No dose dependence in either direction -> the RAD51-excess model is sufficient and the hypothesis is refuted. This is the outcome the prior favours (0.15).
  • Junction frequency rises in BOTH directions -> not a reserve effect; treat as a toxicity artefact and check viability and cell-cycle distribution before interpreting anything.

And the control that must not be omitted: partial RPA knockdown changes cell-cycle distribution, and etoposide damage is S-phase-weighted. Any junction-frequency difference must be matched on S-phase fraction, or the experiment measures how many cells were in S phase rather than how much RPA they had. This corpus has been bitten by exactly this class of composition confound.

Facts and figures
designTwo-arm study. ARM 1 (replication, low cost): ChIP-qPCR for RPA at the KMT2A BCR after etoposide 100 uM/10 min, six primer sets across the ~20 bp hotspot within the 8.3 kb BCR, normalised to a beta-globin control region at 11p15, sampled at 0 and 30 min of recovery, in three cell contexts - a KMT2A-rearranged B-lineage line, a KMT2A-germline B-lineage line, and human CD34+ progenitors. Include the published bleomycin negative-agent arm (30 ug/ml, 60 min) and the KU55933 ATM-inhibitor arm (10 uM, 2 h pre-treatment) as positive and specificity controls. ARM 2 (discriminating, medium cost): repeat Arm 1's best-responding context with partial RPA1 or RPA2 knockdown (titrated to 40-60% of control protein by quantified western, viability within 10% of control) and with RPA over-expression, reading out BCR translocation junction frequency by long-distance inverse PCR junction counts and by dual-colour break-apart FISH split signals in at least 200 cells, at least 3 independent experiments per condition.
readoutPRIMARY, Arm 1: RPA ChIP enrichment at t56/bt56 relative to the beta-globin control locus, fold-change, at 0 and 30 min. PRIMARY, Arm 2: BCR junction frequency by LDI-PCR and split-MLL FISH. SECONDARY, both arms and labelled separately because they look alike: (a) RPA2 FOCUS COUNTS per nucleus - the resection readout; (b) CHROMATIN-BOUND RPA MEDIAN INTENSITY per nucleus - the exhaustion readout. These are different quantities and conflating them would let an exhaustion phenotype be reported as increased resection.
controlsPublished and mandatory: beta-globin control locus; bleomycin as a non-11q23-associated damaging agent (the published negative); DMSO vehicle at 0.1%; input-DNA amplification-efficiency check for all six primer sets before and after etoposide; normal rabbit/mouse IgG ChIP. Added by this design and NOT optional: (1) S-PHASE FRACTION MATCHING - partial RPA knockdown alters cell-cycle distribution and etoposide damage is S-phase-weighted, so any junction-frequency difference must be matched on S-phase fraction or the experiment measures cycling rate rather than RPA dose; (2) viability matched within 10% across all knockdown arms; (3) quantified western for actual knockdown depth, reported as a number not as 'knockdown confirmed'; (4) a KMT2A-germline line as the lineage-matched comparator, so a positive result can be attributed to the locus rather than to the lineage.
cost tierlow for Arm 1, med for Arm 2
timelineArm 1: 2-3 months, of which the first task is reagent recovery (see blockers). Arm 2: a further 4-6 months.
what it discriminatesThe RPA-dose arm separates two models of the same observation. hyp-the-kmt2a-breakpoint-cluster-is-an-rpa-loading-site-... predicts a monotone dependence of junction frequency on RPA level in BOTH directions. The published excessive-RAD51-loading model predicts NO dependence, because RPA is an intermediate and not rate-limiting in it. A single knockdown/over-expression pair therefore has real discriminating power, which is unusual for an experiment this cheap.
why nobody has done itThe 2010 paper is in a general-interest journal, has attracted little follow-up in sixteen years, and sits at the intersection of two fields that rarely meet: therapy-related leukaemia cytogenetics and replication-stress biochemistry. Its cell systems were chosen to manipulate ATM, not to model the disease - which is a reasonable choice for the question they asked and is exactly why the haematopoietic version was never run.
statusplanned
absence classnever-measured
absence class noteNo RPA occupancy measurement at any translocation breakpoint in a haematopoietic cell, and no RPA-dose experiment at any breakpoint in any cell. Cost is nonetheless moderate because the assay is published in full - the class here reflects that nobody has pointed the instrument, not that the instrument needs inventing.
blockers before starting
[
  "PRIMER SEQUENCES: the six BCR primer sets are rendered as empty entries in the PMC text of PMID 21048951. Recover them from the publisher's PDF or from the authors; only redesign as a last resort, because a signal localised to the ~20 bp hotspot can be missed by primers placed elsewhere in the 8.3 kb region, and a redesigned negative would be uninterpretable.",
  "ANTIBODY: no clone is stated (anti-RPA34, LabVision) and that vendor's catalogue has migrated. The original reagent immunoprecipitates both phospho-RPA32 and the unmodified form, which is a property to match deliberately rather than by accident (rule 74).",
  "LDI-PCR PRIMERS: also rendered as bare numbers in the PMC Methods."
]
prereg outcomes
{
  "arm1_transfers": "RPA enrichment above 1.5x beta-globin at t56/bt56 in a haematopoietic cell -> the locus is real in the relevant lineage; proceed to Arm 2",
  "arm1_fails": "no enrichment -> the 2010 result is a fibroblast property; the hypothesis is refuted at step one and the corpus's resection-absence entries recover their generality for RPA",
  "arm1_fusion_dependent": "enrichment in the KMT2A-rearranged line but not the germline line -> a fusion-dependent effect nobody has proposed; the most interesting available outcome",
  "arm1_third_option_neither": "PRE-REGISTERED: enrichment in BOTH lines with no difference between them. This is the likeliest result, it answers the transfer question, and it says nothing about the fusion - and stating it in advance is what stops it being written up as either a positive or a negative for the fusion question.",
  "arm2_reserve_dependent": "junctions up on knockdown and down on over-expression -> reserve-dependent; hypothesis survives",
  "arm2_no_dependence": "no dose dependence -> RAD51-excess model sufficient; hypothesis refuted. THE PRIOR FAVOURS THIS (0.15 for the hypothesis).",
  "arm2_both_directions_up": "junctions up in both directions -> not a reserve effect; treat as toxicity artefact, check viability and S-phase fraction before interpreting"
}
measurements
[
  {
    "quantity": "count",
    "subject": "cell contexts in the published RPA-at-the-BCR experiment that were haematopoietic progenitors or carried a KMT2A fusion",
    "value": 0,
    "units": "cell contexts",
    "n": 3,
    "universe": "the three cell systems used in PMID 21048951: AT5BIVA, 11-4 and BV173, read from the Methods",
    "null": null,
    "note": "BV173 is a CML-derived B-cell line, so one of the three is haematopoietic in origin but carries BCR::ABL1 and not a KMT2A fusion, and it was not used in the ATM comparison. This is the gap the experiment closes."
  },
  {
    "quantity": "threshold",
    "subject": "RPA ChIP enrichment at t56/bt56 versus beta-globin required to call Arm 1 positive",
    "value": 1.5,
    "units": "fold",
    "n": 3,
    "universe": "pre-registered decision threshold for this experiment, at least 3 independent experiments",
    "null": 1,
    "note": "null is no enrichment. The threshold is set in advance because this project has had verdicts track an imported threshold rather than the data; 1.5x is the same figure used in the linked hypothesis's falsifiers, deliberately, so the two cannot disagree."
  }
]
How this could bite us back
  • If Arm 2 shows that lowering RPA raises junction frequency, the immediate clinical implication is a warning about ATR inhibitors near topoisomerase-II poisons - a drug class already moving into MLL-rearranged AML. The harm would fall on surviving normal progenitors and would appear years later, so no conventional efficacy endpoint would see it.
  • A negative Arm 1 is a useful result and must be published internally as such, not filed as a failed experiment: it would establish that the only existing RPA-at-the-KMT2A-locus observation is a fibroblast property, which changes how several entries in this corpus should read.
  • The S-phase confound is the most likely way this experiment produces a wrong positive. Partial RPA knockdown slows cycling, etoposide damage is S-phase-weighted, and an unmatched comparison would report a cell-cycle difference as an RPA-dose effect.
Still open
  1. Can the six BCR primer sequences be recovered from the publisher's version of PMID 21048951, or from the authors? This is the first task and it determines whether the study is a replication or a new experiment that happens to resemble one.
  2. Which KMT2A-rearranged B-lineage line, and is a lineage-matched KMT2A-germline comparator available in the same background? The comparator choice decides whether a positive can be attributed to the locus rather than to the lineage.
  3. Should the menin-inhibitor arm be folded in? The corpus's existing unrun experiment reads out RPA2 under menin inhibition, and a shared cohort would answer both questions at once - but it would also confound the etoposide arm, so the two should probably stay separate.
  4. Why did bleomycin NOT produce BCR loading while etoposide did (PMID 21048951)? That agent specificity is unexplained and is arguably a more tractable question than the dose question, because it points at the lesion type - a topoisomerase-II-linked protein-DNA adduct - rather than at the break.
Every source behind this entry (3)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Sun et al. 2010 specifies the complete protocol for measuring RPA occupancy at the KMT2A breakpoint cluster region: ChIP followed by real-time PCR with six primer sets (t2, t4, bbt56, bt56, t56, in14) around the approximately 20 bp translocation breakpoint hotspot within the 8.3 kb BCR spanning exons 7 to 13, normalised to a beta-globin control region; etoposide at 100 uM for 10 min with 0.1% DMSO vehicle; sampling immediately after exposure and after 30 min recovery; bleomycin at 30 ug/ml for 60 min as a non-11q23-associated agent control; ATM kinase inhibition with 10 uM KU55933 for 2 h; dual-colour break-apart FISH scoring split signals separated by more than 1 um in at least 200 cells; long-distance inverse PCR on BamHI-digested self-ligated genomic DNA to recover junctions; and a real-time PCR amplification-efficiency control on input DNA before and after etoposide for all six primer sets - with every arm performed in the ATM-deficient SV40-transformed fibroblast line AT5BIVA, its chromosome-11-complemented derivative 11-4, or the CML-derived B-cell line BV173.

    ATM modulates the loading of recombination proteins onto a chromosomal translocation breakpoint hotspot. · Sun J, Oma Y, Harata M, Kono K, Shima H, Kinomura A, et al. · PLoS One 5(10):e13554 2010
    PMID 21048951DOI
    What this evidence cannot show

    Methods read in full from PMC 2026-09-24. TWO GAPS THAT BLOCK A LITERAL REPEAT: (1) the PRIMER SEQUENCES are rendered as empty entries in the PMC text - the Methods list 't2 forward:', 't2 reverse:' etc. with no sequence - so they must be recovered from the publisher's PDF or redesigned against the BCR coordinates, and any redesign changes what 'the same primer set' means; (2) the ANTIBODY CLONE is not stated (anti-RPA34, LabVision) and the vendor's catalogue has migrated (rule 74). Also unspecified in the PMC text: the BamHI LDI-PCR primer sequences, listed as bare numbers. The ChIP was formaldehyde 1% for 10 min with sonication to 300-500 bp; antibody amounts are not given.

    Evidence that argues against this (1)
    • A protocol recovered from a paper is not a working protocol, and this one has two missing reagent specifications that are exactly the ones that decide whether a replication is a replication. Redesigned primers at a different position within the 8.3 kb BCR could easily miss a signal localised to the ~20 bp hotspot, and a different RPA antibody may not recognise the same phospho-species - the original immunoprecipitates both phosphorylated and unmodified RPA32, which is a property of that reagent. A failed replication with new reagents would be uninterpretable, so reagent recovery is the first task and not a detail. Sun J et al. 2010
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Partial reduction of RPA accelerated fork breakage and forced elevation of RPA delayed replication catastrophe even in the absence of ATR activity, establishing that RPA level is an experimentally tractable dose variable with a measurable phenotype in both directions and that sub-lethal partial knockdown is achievable.

    ATR prohibits replication catastrophe by preventing global exhaustion of RPA. · Toledo LI, Altmeyer M, Rask MB, Lukas C, Larsen DH, Povlsen LK, Bekker-Jensen S, Mailand N, Bartek J, Lukas J · Cell 155(5):1088-103 2013
    PMID 24267891DOI
    What this evidence cannot show

    Abstract level; full text not opened, so the actual knockdown depth, the over-expression construct and the viability window are NOT known to this beat and the '40-60% of control' target in this experiment's design is this beat's own specification, not the paper's. Human tumour cell lines; no haematopoietic system and no topoisomerase-II poison. The expected_absent_terms field records that KMT2A and MLL do not appear here - the transfer to this locus is the hypothesis's, not the paper's.

    Evidence that argues against this (1)
    • Borrowing a dose-perturbation strategy from a genome-wide-ssDNA regime and applying it to a 10-minute etoposide pulse assumes the perturbation is calibrated for the new regime, and it is not. A knockdown depth that produces a clean phenotype under hydroxyurea-driven fork stalling may produce nothing, or only toxicity, under a brief topoisomerase-II poisoning. The knockdown depth should be titrated in the new system before the junction readout is trusted, and this experiment's stated 40-60% window is a starting guess. Toledo LI et al. 2013
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    This corpus already specifies RPA2 foci and RPA2 phosphorylation as the resection readout in KMT2A-rearranged blasts, in an experiment entry that has not been executed - so an RPA readout in the index disease is designed and unrun rather than unimagined, and the two RPA imaging readouts in play (focus counts for resection, chromatin-bound intensity for exhaustion) are different measurements that must be labelled separately.

    What this evidence cannot show

    A corpus read, not an external source. It establishes what this project has planned, not what anyone has measured. The distinction between focus-count and chromatin-bound-intensity readouts is this beat's own framing, drawn from the fact that the exhaustion work uses single-cell chromatin-bound RPA intensity while the resection literature uses focus counts; no source read by this beat states the distinction in those terms.

    Evidence that argues against this (1)
    • Pointing at an existing unrun experiment entry could make this one look redundant, and the overlap should be stated rather than glossed: the existing entry reads out RPA2 as a resection assay under a menin inhibitor, with no etoposide arm, no locus-level ChIP and no RPA-dose arm. The two experiments share a reagent and answer different questions. If resources allowed only one, the existing one is the better buy because it serves a hypothesis the corpus has invested in more heavily. This repository 2026
Connected to