Measure the clearance clock directly: SNAP-tag H3.3 pulse-chase at KMT2A::AFF1 target genes, with and without DOT1L inhibition, in cycling and arrested cells
also called H3.3 turnover pulse-chase at HOXA · the experiment that would give the number the dosing argument needs
Tag the replacement histone so you can tell new copies from old, then watch how fast new ones appear at the genes this fusion controls - in dividing cells and in cells forced to stop dividing. That number is the one the whole drug-timing argument depends on, and nobody has measured it.
The same thing, for a scientist
**The number this experiment produces:** the rate of replication-independent H3.3 deposition at KMT2A::AFF1-occupied loci, and the rate of H3K79me2 loss at the same loci under DOT1L inhibition, measured in the same cells. `mechanism-h3k79-turnover-and-dilution` names 'cell divisions of continuous DOT1L inhibition required to clear H3K79me2 from HOXA9' as the unmeasured number that IS the dosing schedule; this measures its two components separately. **Design.** SNAP-tagged H3.3 knock-in in SEM and RS4;11 (KMT2A::AFF1) plus a KMT2A-wild-type B-ALL comparator. Quench-chase-pulse labelling separates old from new histone. Read out by (a) ChIP of the new-histone label at fusion-occupied loci versus **expression-matched non-occupied loci in the same cells** - the control that makes the whole thing interpretable - and (b) calibrated ChIP for H3K79me1/2/3 in parallel, so mark loss and histone replacement are measured on the same chromatin. **Four arms:** vehicle; pinometostat; cycling; and **cell-cycle-arrested** (serum starvation or a CDK4/6 inhibitor, not a DNA-damaging agent). The arrested arm is the one that has never been run and is the only arm that speaks to `cell-state-quiescence` and `cell-state-relapse-initiating-persister`. **A fifth arm if resources allow: a DOT1L degrader against pinometostat.** `mechanism-nucleosome-turnover-clearance-clock-machinery` derives a falsifiable prediction that these should differ, because yeast Dot1p has intrinsic histone chaperone activity at long transcribed genes [E2].
Why the control arm is the experiment
The rival hypothesis in hyp-kmt2a-aff1-alters-h3-3-deposition-at-fusion-targets is that the fusion sits on loci that were already high-turnover. An experiment that measures turnover only at HOXA cannot distinguish the two. The comparison that can is:
| locus set | fusion-occupied | expression level | prediction if fusion sets the clock |
|---|---|---|---|
| HOXA9/A10, MEIS1, PROM1 | yes | high | fastest turnover |
| expression-matched non-occupied | no | matched | slower |
| fusion-occupied, low-expressed | yes | low | tests whether occupancy or output drives it |
| housekeeping | no | high | reference |
Building that matched set requires the fusion ChIP and the RNA-seq from the same cells. Both exist for SEM in the public domain; this graph should assemble the matched locus set as an analysis before the wet experiment is commissioned, because if no expression-matched non-occupied set exists in a KMT2A-rearranged line, the experiment cannot be interpreted and should not be run.
What could go wrong, stated before it does
- SNAP-tag H3.3 is an over-expressed or knock-in tagged variant. The founding H3.3 localisation study used zinc-finger-nuclease tagging for exactly this reason; a transgene reports on the transgene.
- Cell-cycle arrest is not quiescence. A serum-starved or palbociclib-arrested cell line is a model of non-division, not a model of a relapse-initiating persister. State that in any conclusion.
- ChIP of a labelled histone at a single locus has poor dynamic range. Calibrated ChIP with spike-in is mandatory; this project already requires it for the H3K79 series.
- Cell lines are not the disease. DepMap contains no infant KMT2A::AFF1 line and this project has documented that at least ten infant-derived lines exist unscreened. SEM and RS4;11 are the available compromise and are paediatric/adolescent rather than infant.
- A null result is ambiguous between 'no difference' and 'not enough resolution'. Power the locus comparison before running it.
| design | SNAP-tag H3.3 knock-in quench-chase-pulse labelling in SEM and RS4;11 (KMT2A::AFF1) and a KMT2A-wild-type B-ALL comparator; new-histone ChIP at fusion-occupied versus expression-matched non-occupied loci; parallel spike-in-calibrated ChIP for H3K79me1/2/3 on the same chromatin. Arms: vehicle / pinometostat x cycling / cell-cycle-arrested (serum starvation or CDK4/6 inhibition). Optional fifth arm: DOT1L degrader versus SAM-competitive inhibitor. |
|---|---|
| readout | replication-independent H3.3 deposition rate per locus class, and H3K79me1/2/3 loss rate per locus class, in the same cells; primary comparison is fusion-occupied versus expression-matched non-occupied |
| controls | expression-matched non-occupied loci built from fusion ChIP plus RNA-seq in the same line (must be assembled first); spike-in calibration for every ChIP; a KMT2A-wild-type B-ALL line; arrest verified by EdU and DNA content, not assumed |
| cost tier | high |
| timeline | 9-15 months including knock-in line generation |
| prerequisite analysis | assemble the expression-matched non-occupied locus set from public SEM fusion ChIP and RNA-seq; if no such set exists the experiment is uninterpretable and should not be run |
tests
[ "hyp-kmt2a-aff1-alters-h3-3-deposition-at-fusion-targets", "mechanism-nucleosome-turnover-clearance-clock-machinery", "mechanism-h3k79-turnover-and-dilution" ]
- A positive result - the fusion drives turnover - would mean that any drug which reduces fusion activity also slows clearance of H3K79 methylation, making menin-inhibitor plus DOT1L-inhibitor combinations antagonistic on this axis.
- A negative result in the arrested arm would confirm the harshest version of the quiescence problem: the compartment that causes relapse is the compartment a turnover-limited drug cannot reach, and no scheduling change fixes it.
- The experiment measures a rate in a cell line. Even a clean answer would not transfer to an infant patient's marrow without an in-vivo version that nobody has the reagents for.
- Does a published SEM or RS4;11 dataset contain both KMT2A::AFF1 ChIP and matched RNA-seq of sufficient quality to build the expression-matched control locus set? If not, that is the first piece of work.
- Is SNAP-tag pulse-chase sensitive enough at single-locus resolution, or does this need a genome-wide method (CATCH-IT) with the locus comparison done computationally?
- Which arrest method least perturbs chromatin metabolism? Serum starvation, CDK4/6 inhibition and contact inhibition all have chromatin effects of their own and none is quiescence.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Nucleosome turnover rates determine H3K79 methylation valency and account for the asymmetric distribution of H3K79me toward the transcriptional unit, so measuring turnover and mark loss on the same chromatin is the design that tests the relationship.
Nucleosome Turnover Regulates Histone Methylation Patterns over the Genome. · Chory EJ, Calarco JP, Hathaway NA, et al. · Molecular Cell 2019What this evidence cannot show
Abstract read from efetch 2026-09-20 with identifiers re-fetched for this entry; full text not read. Chemical-induced proximity system, not leukaemia, no KMT2A fusion. Cited as the design rationale, not as a result about this disease.
Evidence that argues against this (1)
- If enzyme recruitment alone sets the mark - as graded weakening of the DOT1L-AF9 interaction suggests - then this experiment will find mark loss uncoupled from turnover and the clearance-clock framing will need replacing rather than refining. Kuntimaddi A et al. 2015
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Yeast Dot1p displays intrinsic histone chaperone activity, regulates nucleosome dynamics via histone exchange, and is particularly important for histone exchange and chromatin accessibility in the transcribed regions of long genes, through a methylation-independent function.
Dot1 regulates nucleosome dynamics by its inherent histone chaperone activity in yeast. · Lee S, Oh S, Jeong K, et al. · Nature Communications 2018What this evidence cannot show
Abstract read from efetch 2026-09-20; full text not read (PMC5770421 exists). **Yeast and in vitro.** No chaperone activity has been demonstrated for human DOT1L. The degrader-versus-inhibitor arm is powered by this transfer holding and should be described as exploratory in any protocol.
Evidence that argues against this (1)
- Human DOT1L's occupancy at fusion targets is set by tethering through AF9/ENL and AF4-family partners with no yeast counterpart, so an intrinsic activity measured in yeast may be swamped at a fusion-occupied locus. Kuntimaddi A et al. 2015
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Pinometostat was administered by continuous intravenous infusion, reduced H3K79 methylation in patients and produced 2 complete remissions in 51 adults with acute leukaemia, without reaching a maximum tolerated dose.
The DOT1L inhibitor pinometostat reduces H3K79 methylation and has modest clinical activity in adult acute leukemia. · Stein EM, Garcia-Manero G, Rizzieri DA, et al. · Blood 2018What this evidence cannot show
Identifiers and title fetched from efetch 2026-09-20 rather than copied from the sibling entries in this graph that also carry this paper; abstract read, full text not read. Adults, not infants. Cited here to establish that the clinical schedule question is live and was answered empirically rather than from a measured clearance rate.
Evidence that argues against this (1)
- Modelled acquired resistance to DOT1L inhibition in KMT2A-rearranged ALL shows the leukaemia adapts transcriptionally under sustained exposure, so a longer schedule derived from a clearance measurement is not automatically better and may select the adaptation that defeats it. Schneider P et al. 2023
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Replication-independent histone turnover is substantial in postmitotic cells: H2B-GFP pulse-chase in adult mouse cardiomyocytes gives a histone half-life of approximately two weeks, and rapid-turnover regions correspond to cis-regulatory domains of heart genes.
Replication-Independent Histone Turnover Underlines the Epigenetic Homeostasis in Adult Heart. · Li Y, Ai S, Yu X, et al. · Circulation Research 2019What this evidence cannot show
Abstract read from efetch 2026-09-20; full text not read (no PMCID). Mouse cardiomyocytes; H2B rather than H3, and H2A-H2B dimers exchange more readily than H3-H4 tetramers. This is the justification for including an arrested arm at all - it says the measurement will not simply return zero - and is not a prediction of the value.
Evidence that argues against this (1)
- A cardiomyocyte is transcriptionally and metabolically nothing like a quiescent leukaemia-initiating cell, and the H2B half-life is an upper bound on H3 replacement speed. The arrested arm could still return a rate too slow to measure in the available window. Arrieta A et al. 2022
measures H3.3 deposition and H3K79me loss on the same chromatin
supplies the unmeasured number the clearance-clock argument rests on
the turnover arm (b) that series proposes, run as a standalone measurement
the experiment that would close this entry's first open question
a DEK arm would separate a global H3.3 redistribution from a local change at fusion targets
the literature bound for the experiment that would produce the number