Oncofusion

start / graph / exp-crispr-chip-h3k79-partition-in-human-kmt2a-aff1

Experimentmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q-partner-provenance

Repeat the H3K79 partition in a human KMT2A::AFF1 line: menin inhibition and MLLT10 knockout, side by side with a KMT2A::MLLT3 line

also called index-fusion CRISPR-ChIP replication · does the two-unit model exist in t(4;11)

Do the key experiment again, but in human infant-type leukaemia cells instead of mouse myeloid ones, with a mouse-type comparison running alongside. Measure how much of the H3K79 tag a menin drug removes in each.

The same thing, for a scientist

The CRISPR-ChIP partition was established entirely in murine MLL-AF9 [E1]. The index fusion lacks both documented DOT1L interfaces [E2], so the partition's existence there is an open question that two published falsifiers already depend on. This experiment resolves it with existing reagents. **Design.** A matched pair of human lines - one KMT2A::AFF1 (SEM or RS4;11) and one KMT2A::MLLT3 - each Cas9-expressing. Four arms per line: (i) vehicle; (ii) a menin inhibitor to full target engagement, verified by loss of fusion occupancy at a known target rather than by dose alone; (iii) *MLLT10* knockout, two independent guides; (iv) menin inhibitor plus *MLLT10* knockout. **Readout:** calibrated ICeChIP-seq for H3K79me1/me2/me3 giving absolute occupancy, plus fusion ChIP (FLAG-tagged or antibody) and nascent transcription (PRO-seq or TT-seq) in the same samples. **The single number the project wants** is the residual fraction of H3K79me2 at fusion target loci after menin inhibition, in each line. The two-unit model predicts a large drop in the MLLT3 line and, by `hyp-index-fusion-h3k79-is-entirely-endogenous`, little or none in the AFF1 line.

Go deeper
Why calibrated ChIP and not standard ChIP-seq

The quantity at issue is a fraction remaining, and standard ChIP-seq is normalised in a way that destroys exactly that: a genome-wide reduction in a mark is invisible after library normalisation. ICeChIP or spike-in calibration converts the readout to modification density per nucleosome, which is what a residual-fraction claim requires. This graph already holds exp-h3k79-state-resolved-icechip-dose-series, and this experiment is its partner-controlled sibling.

It also matters that H3K79 methylation is removed by dilution through replication rather than by a demethylase, so the time axis is in cell divisions. Menin-inhibitor arms should be sampled at matched division numbers, not matched hours 3.

Controls that decide whether the result is interpretable
  1. Target engagement, verified. A null result in the AFF1 line is only meaningful if the menin inhibitor engaged. Verify by fusion displacement from a canonical target locus in the same chromatin, not by IC50.
  2. Matched division number, because the mark dilutes 3.
  3. Nascent transcription in the same samples, to test the transcription-coupled model directly: if H3K79me2 loss tracks transcription loss with the same slope in both lines, there is no separate fusion-directed pool in either.
  4. A KMT2A::MLLT10 arm if obtainable, because that is the genotype where the two units converge and the partition predicts a qualitatively different response 4.
  5. Two independent MLLT10 guides plus a rescue, since MLLT10 knockout could reduce H3K79me2 by destabilising DOT1L rather than by mis-directing it.
Facts and figures
design2 human cell lines (KMT2A::AFF1 and KMT2A::MLLT3, optionally KMT2A::MLLT10) x 4 arms (vehicle; menin inhibitor at verified full engagement; MLLT10 KO x2 guides; combination), Cas9-expressing
readoutcalibrated ICeChIP-seq for H3K79me1/me2/me3 (absolute modification density), fusion occupancy ChIP, and nascent transcription (PRO-seq or TT-seq) on the same samples
controlsverified target engagement by fusion displacement; matched cell-division number rather than matched hours; two independent MLLT10 guides plus rescue; a KMT2A::MLLT10 arm as the convergence positive control
primary endpointresidual fraction of H3K79me2 at fusion target loci after menin inhibition, per line
cost tiermed
timeline3-4 months including ICeChIP calibration
decision ruleresidual fraction >0.8 in the AFF1 line and <0.4 in the MLLT3 line supports hyp-index-fusion-h3k79-is-entirely-endogenous; comparable residuals in both refute it
How this could bite us back
  • A null result in the AFF1 arm is only as good as the engagement control; without it, the experiment would produce the most consequential finding in this beat as an artefact of underdosing.
  • If the answer is that menin inhibition does not move H3K79me2 in t(4;11) cells, the project loses its mechanistic narrative for the index disease while keeping the drug's clinical activity - an uncomfortable position that is harder to publish than to establish.
Still open
  1. Which menin inhibitor and which exposure should define 'full target engagement' in a B-lymphoid line? The engagement literature is myeloid.
  2. Is a KMT2A::MLLT10 human line available in adequate condition for a calibrated ChIP experiment, or must that arm be murine and therefore a separate comparison?
Every source behind this entry (4)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    Every H3K79me2 sample in the CRISPR-ChIP study is Mus musculus, cell type MLLAF9, including the menin-inhibitor-treated and MLLT10-knockout ChIP-seq arms; the human samples in the study are K562 and carry only Pol II, H3K4me3 and H3K27ac.

    CRISPR-ChIP to identify mechanisms of chromatin regulation [ChIP-seq] (GSE192555) · Gilan O, Dawson MA, et al. (GEO submitters) · NCBI Gene Expression Omnibus 2023
    What this evidence cannot show

    Read on 2026-09-19 from the GEO accession-display text endpoint and enumerated sample by sample. Submitter-entered metadata; bounds deposited sequencing rather than every experiment performed. The menin inhibitor used is not named in GEO, so this experiment cannot match the original compound and should state its own.

    Evidence that argues against this (1)
    • Repeating a mouse experiment in human cells changes many variables at once - species, lineage, fusion, immortalisation history - so a difference in outcome will not by itself identify the fusion as the cause. That is why the matched KMT2A::MLLT3 human line is the load-bearing control rather than a courtesy. Gilan O et al. 2023
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    KMT2A::AFF1 carries neither the AHD (PF17793, on MLLT3 and MLLT1) nor an AF10-type DOT1L-binding leucine-zipper region, so 'the malignant driver directs DOT1L catalysis' has no known molecular referent in the index fusion.

    What this evidence cannot show

    Accessions fetched directly on 2026-09-19 with primary gene names and lengths read back. Pfam absence is absence of a detected match; AFF1 is largely disordered, where short linear motifs are under-detected.

    Evidence that argues against this (1)
    • If the experiment finds a menin-sensitive H3K79 pool in the AFF1 line, the domain argument is not refuted - it simply means the interface is unmapped, and the experiment becomes the starting point for finding it. Leach BI et al. 2013
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Degree of DOT1L recruitment to MLL-AF9 sets the level of H3K79me2 and me3 at target genes and the transformation potential, with structure-guided mutations producing graded effects.

    Degree of recruitment of DOT1L to MLL-AF9 defines level of H3K79 Di- and tri-methylation on target genes and transformation potential. · Kuntimaddi A, Achille NJ, Thorpe J, Lokken AA, Singh R, Hemenway CS, Adli M, Zeleznik-Le NJ, Bushweller JH · Cell Rep 2015
    PMID 25921540DOI
    What this evidence cannot show

    Abstract fetched from PubMed and read on 2026-09-19; full text not read. Entirely MLL-AF9. Cited here for the design principle - that H3K79 methylation state is a graded readout of recruitment, so the assay must be quantitative - rather than for a claim about the index fusion.

    Evidence that argues against this (1)
    • If recruitment sets methylation level, a line with no recruitment interface should show low ectopic H3K79me2, and human MLL-AF4 leukaemias do not. Either the graded relationship is AF9-specific or recruitment is not the only determinant, and this experiment is designed to tell which. Krivtsov AV et al. 2008
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    AF10 is both the endogenous director of progressive H3K79 methylation and a recurrent fusion partner, in KMT2A::MLLT10 and PICALM::MLLT10, and AF10 inactivation impairs transformation by MLL-AF9, MLL-AF6 and NUP98-NSD1.

    AF10 regulates progressive H3K79 methylation and HOX gene expression in diverse AML subtypes. · Deshpande AJ, Deshpande A, Sinha AU, Chen L, Chang J, Cihan A, Fazio M, Chen CW, Zhu N, Koche R, Dzhekieva L, Ibáñez G, Dias S, Banka D, Krivtsov A, Luo M, Roeder RG, Bradner JE, Bernt KM, Armstrong SA · Cancer Cell 2014
    PMID 25464900DOI
    What this evidence cannot show

    Abstract fetched from PubMed and read on 2026-09-19; full text not read. Mouse models; all fusions tested are myeloid. Relevant here because it makes the KMT2A::MLLT10 arm the genotype where the two-unit model must break, which is a positive control for the partition itself.

    Evidence that argues against this (1)
    • If endogenous AF10 is required for transformation by fusions lacking it, MLLT10 knockout will reduce viability in every arm, confounding a chromatin readout with a fitness readout. Short-timepoint sampling and a degron rather than a knockout would mitigate this. Gilan O et al. 2023
Connected to