The DAXX requirement in B-lineage KMT2A-rearranged lines is a telomere phenotype, not a fusion-target phenotype - and the two non-KMT2A lines in the deep group are the test
also called the ALT hypothesis for the deep DAXX group · why RCH-ACV and YCUB-2 are in there
Cells with this fusion need DAXX, but so do two cells without it - and nobody can explain why those two are in the same group. The guess is that it has nothing to do with the fusion and everything to do with how these cells keep the ends of their chromosomes from fraying.
The same thing, for a scientist
Thirteen B-lineage ALL lines split cleanly into six DAXX-dependent and seven not, with a 0.473 Chronos gap and nothing in it. All four KMT2A-rearranged lines are in the dependent six; so are RCH-ACV (TCF3::PBX1) and YCUB-2 (TCF3::HLF), while 697 - carrying the *same* TCF3::PBX1 as RCH-ACV - sits on the shallow side, 0.57 away [E1]. Donor age does not sort them (Spearman +0.265, p 0.53) and fusion identity demonstrably does not [E1]. **Hypothesis: the deep group is defined by telomere maintenance state, and DAXX is required there for the same reason its loss causes ALT elsewhere.** ATRX/DAXX inactivation is the canonical human route to alternative lengthening of telomeres: 61% of pancreatic neuroendocrine tumours are ALT-positive and every one carries an ATRX or DAXX lesion [E2]. If any of these six lines maintains telomeres partly or wholly through an ATRX/DAXX-dependent route, removing DAXX is a telomere crisis rather than a transcriptional insult - and the KMT2A association would be a correlate of cell-of-origin or replicative history, not of the fusion. **Prior: low, roughly 0.15.** Acute leukaemias are overwhelmingly telomerase-positive and ALT is rare in haematological malignancy, which is the main argument against. The hypothesis is worth writing because it is the only explanation on offer that accounts for RCH-ACV and YCUB-2, because the assay is cheap, and because a negative result eliminates it in one experiment rather than leaving it as a permanent alternative reading of the dependency data.
The rival hypotheses this competes with, and how each dies
| hypothesis | what would kill it | status |
|---|---|---|
| H1: telomere state (this entry) | C-circle assay negative in all six deep lines | untested |
| H2: the fusion recruits DAXX | no contact in any assay that can see a chaperone hand-off | badly damaged already - DAXX is absent from all 960 rows of the only MLL-fusion AP-MS interactome 3 |
| H3: B-lineage developmental stage | the deep and shallow groups match on immunophenotype | untested; the deep group spans pro-B and pre-B |
| H4: donor age / replicative history | no age gradient | tested and not supported, Spearman +0.265, p 0.53 1 |
| H5: fusion identity per se | the same fusion on both sides of the gap | refuted - TCF3::PBX1 is at -0.802 and -0.232 1 |
| H6: noise | the gap disappears in an independent screen | untestable within DepMap; the 0.473 gap is 3.7x the next largest |
H1 is the only surviving hypothesis that predicts which non-KMT2A lines join the group, which is the property the others lack.
Why a negative result is worth as much as a positive one
If the C-circle assay is negative across all thirteen lines, the telomere explanation is gone and the deep group needs a different one - but the beat gains something concrete anyway. The bite-back on any DAXX-directed intervention in this disease currently rests on an inference from panNET and glioma: that removing DAXX might supply ALT. A negative ALT screen in leukaemia lines does not prove a drug is safe, but it removes the single most specific mechanistic objection on the table and would change how this graph writes about DAXX as a target.
So the experiment discriminates in both directions, which is the property this project's falsifier standard asks for.
| statement | The six DAXX-dependent B-lineage ALL lines - SEM, SEM-K2, RS4;11, HB11;19, RCH-ACV and YCUB-2 - share a telomere maintenance state in which ATRX/DAXX-mediated H3.3 deposition is load-bearing, and their DAXX requirement is a telomere phenotype rather than a consequence of the KMT2A rearrangement. |
|---|---|
| origin | data-pattern |
| prior probability | 0.15 |
| what would make this more likely before any bench work | Published telomere-length or C-circle data may already exist for RS4;11, SEM, Reh, NALM-6 or 697 - these are heavily used lines. A literature and GEO sweep for telomere measurements on any of the thirteen would cost an hour and could support or eliminate the hypothesis before a single assay is run. This has not been done. |
predictions
[ "A C-circle assay or telomere-FISH ultrabright-foci count separates the six deep lines from the seven shallow ones, with the two non-KMT2A deep lines (RCH-ACV, YCUB-2) grouping with the four KMT2A-rearranged ones.", "697 and RCH-ACV, which carry the same TCF3::PBX1 fusion and differ by 0.57 in DAXX Chronos, differ in telomere state.", "DAXX knockdown in a deep line produces telomere dysfunction-induced foci (53BP1 or gamma-H2AX at telomeres) before it produces a transcriptional change at HOXA9 or MEIS1.", "Telomere length in the deep lines is heterogeneous and long-tailed, the classic ALT appearance, rather than uniformly short." ]
experiments
[ "exp-c-circle-and-telomere-fish-across-the-thirteen-b-all-lines" ]
analyses
[ "ana-daxx-lymphoid-dependency-under-three-nulls", "ana-daxx-and-atrx-are-absent-from-the-only-mll-fusion-interactome" ]
falsifiers
[ "A C-circle assay run on all thirteen DepMap B-lineage ALL lines shows no C-circle signal above the telomerase-positive control in ANY of the six deep lines (SEM, SEM-K2, RS4;11, HB11;19, RCH-ACV, YCUB-2), using U2OS as the ALT-positive reference and HeLa as the negative. That kills this hypothesis outright and should be recorded here as REFUTED.", "Telomere restriction fragment length in the six deep lines is within 1.5-fold of the seven shallow lines and shows no heterogeneous long tail. A telomere explanation that predicts nothing about telomere length is not a telomere explanation.", "shRNA or degron depletion of DAXX in SEM produces a HOXA-cluster transcriptional change measurable by qPCR at 48 hours BEFORE any increase in telomere dysfunction-induced foci above 2 per nucleus. Ordering the two readouts is the discriminating measurement: a transcriptional-first result favours the fusion-target hypothesis over this one.", "The deep and shallow groups are separated by immunophenotypic stage (pro-B versus pre-B by CD19/CD10/cytoplasmic mu) at least as cleanly as they are by DAXX Chronos, in which case H3 (developmental stage) explains the grouping without any telomere biology." ]
- If this hypothesis is TRUE it is an argument against the intervention, not for it: a cell whose telomeres depend on ATRX/DAXX is a cell one step from ALT, and a partial pharmacological knockdown is exactly the perturbation that selects for the escape route.
- If it is FALSE, the DAXX dependency loses its only current mechanistic account and reverts to an unexplained correlation in three donors - which is the more likely outcome given the prior and should be stated when the assay is designed, not after it reads out.
- Has anyone measured telomere length or C-circles in RS4;11, SEM, Reh, NALM-6, 697 or RCH-ACV? These are among the most-used B-ALL lines in the world and the measurement may already be published or deposited. A sweep of the literature and GEO for telomere data on the thirteen lines is the free version of this experiment and has not been run.
- If the deep group is not a telomere group, what else could put RCH-ACV and YCUB-2 with the four KMT2A-rearranged lines? Immunophenotypic stage is the next candidate and can be read off Cellosaurus and the DSMZ line descriptions without any new data.
- Does DAXX depletion in a deep line kill before or after it changes transcription? The ORDER of the two readouts is what separates this hypothesis from the fusion-target one, and no depletion time-course exists for DAXX in any leukaemia line.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
Across 13 B-lineage ALL DepMap lines, DAXX Chronos is bimodal with a 0.473 gap between -0.7921 and -0.3191; the six-member deep group contains all four KMT2A-rearranged lines plus RCH-ACV (TCF3::PBX1) and YCUB-2 (TCF3::HLF), donor age gives Spearman +0.265 with permutation p 0.53 across the eight non-KMT2A lines with an age, and 697 carries the same TCF3::PBX1 as RCH-ACV while sitting 0.57 away on the shallow side.
Bimodality of DAXX Chronos across 13 B-lineage ALL DepMap lines, with donor age and fusion identity tested as rival explanations · Oncofusion knowledge graph (internal analysis) 2026What this evidence cannot show
Thirteen cell lines, of which the four index lines are three donors. The -0.70 cut defining the deep group was chosen after seeing the gap. Cellosaurus fusion annotation is curated and a line with no listed fusion is uncurated rather than fusion-negative, so 'the deep group is all fusion-driven' cannot be asserted from this table. Chronos scores from Open Targets dataVersion 26/09 fetched 2026-09-24.
Evidence that argues against this (1)
- A bimodal appearance in thirteen points is not strong evidence of two populations. With n=13 drawn from any moderately skewed unimodal distribution, a gap of this size is not rare, and the 3.7x-next-largest comparison is an informal statistic with no null behind it. The claim this hypothesis actually needs is only that the six deep lines have SOMETHING in common, and the bimodality is a motivation for looking rather than evidence that a grouping exists. Oncofusion knowledge graph (internal analysis) 2026
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Inactivation of ATRX or DAXX is the established human route to alternative lengthening of telomeres: 61% of pancreatic neuroendocrine tumours show ALT-characteristic telomere abnormalities and every one of those carried an ATRX or DAXX mutation or had lost the nuclear protein, with the same correlation holding in central nervous system tumours.
Altered telomeres in tumors with ATRX and DAXX mutations. · Heaphy CM, de Wilde RF, Jiao Y, et al. · Science 2011What this evidence cannot show
Abstract read from efetch 2026-09-24. The direction established is chaperone LOSS to ALT in solid tumours; this hypothesis runs it backwards, proposing that a cell already using an ATRX/DAXX-dependent telomere route would be harmed by removing the chaperone, and no paper shows that. The telomere readout is FISH morphology rather than a C-circle assay.
Evidence that argues against this (1)
- This is the fact that argues the hypothesis is unlikely as well as the fact that motivates it. In panNET and glioma, ATRX/DAXX loss ENABLES ALT - so the cells that need ATRX/DAXX least are the ALT ones, which is the opposite of what this hypothesis requires. For the hypothesis to work, the deep lines must be in a state where the chaperone is load-bearing for telomere integrity without having tipped into ALT, a narrower and less well-documented condition than the published one. Schwartzentruber J et al. 2012
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The competing hypothesis that the fusion recruits DAXX is already damaged by direct evidence: DAXX appears in none of the 960 rows of the only published affinity-purification interactome of MLL fusion proteins, which covered seven distinct fusions.
MLL-fusion-driven leukemia requires SETD2 to safeguard genomic integrity. · Skucha A, Ebner J, Schmöllerl J, et al. · Nature Communications 2018What this evidence cannot show
The table is rank-thresholded at roughly 293 preys per bait, so absence means 'not top-ranked', not 'not detected'; the raw deposit PXD009338 has never been re-searched. An affinity purification is also the method least able to retain a chaperone caught mid-hand-off, so this damages the recruitment hypothesis without eliminating it.
Evidence that argues against this (1)
- A chaperone can be functionally required at a locus without ever co-purifying with the protein that put the locus into its transcriptional state. The fusion could raise H3.3 turnover at its targets simply by driving elongation through them, making DAXX or HIRA rate-limiting downstream with no physical contact at all. On that reading the interactome absence is expected and says nothing against a fusion-driven mechanism - it only rules out DAXX being a member of the fusion complex. Goldberg AD et al. 2010
the hypothesis exists to explain the deep DAXX group that analysis found and could not account for
C-circle assay across the whole DepMap B-ALL panel; a flat result is a pre-specified refutation
generated by the panNET/glioma analogy: ATRX/DAXX loss is the canonical human route to ALT
proposes that the DAXX requirement is a telomere phenotype rather than a fusion-target one
the gene-body alternative has no support in the dependency data: HIRA's lymphoid difference is +0.032, slightly LESS required in the index arm