Nobody knows which way to push H2Bub in this disease, and the only available compound pushes it upward
also called the direction problem · raise or lower H2Bub
Two good arguments point in opposite directions: the ubiquitin tag feeds the leukaemia's key mark, so less of it should help — but the same tag is what lets a blocked cell finish maturing, so more of it should help too. The only drug-like molecules that exist against this pathway raise the tag, so the question is not academic.
The same thing, for a scientist
This beat assembled the H2Bub layer and then discovered it could not say which direction of perturbation would be therapeutic. **Lower is indicated by:** H2Bub licenses DOT1L-mediated H3K79me2/3, the fusion's defining mark; Rnf20 suppression reduces MLL-fusion target expression and inhibits proliferation [E1]; panobinostat's activity in MLL-rearranged ALL involves H2Bub depletion via RNF20/RNF40/WAC suppression, and WAC knockdown alone kills [E2]. **Higher is indicated by:** H2Bub1 rises during differentiation, RNF40 is required for induction of differentiation markers and for resolving H3K4me3/H3K27me3 bivalency on lineage genes [E3] — and KMT2A-rearranged leukaemia is defined by a differentiation block on bivalent, primed target genes, two states this graph already holds as entries. **The asymmetry that makes this urgent:** the only chemical matter in this entire axis is a series of selective, cell-active macrocyclic USP22 inhibitors, and they **raise** H2Bub [E4]. If the 'lower' model is right, that compound class is contraindicated here; if the 'higher' model is right, it is a candidate. The experiment that separates them is one dose-response in existing cell lines.
The two models, and why the graph cannot currently choose
| Lower H2Bub | Raise H2Bub | |
|---|---|---|
| Mechanistic rationale | removes the licence for H3K79me2/3 at fusion targets | resolves bivalency and permits the differentiation the block prevents |
| Best evidence | 1, 2 — both in MLL-rearranged leukaemia models | 3 — human mesenchymal stem cells, not haematopoietic, not leukaemic |
| Available agents | panobinostat (pan-HDAC, indirect, many other effects) | selective macrocyclic USP22 inhibitors 4 |
| DepMap support | none — writers rank 16,229/9,953/7,779 | weak — USP22 rank 226, ATXN7L3 rank 290, but ENY2 flat |
| Predicted failure mode | no therapeutic window; hits H3K4me3 arm and normal KMT2A too | amplifies the fusion's own H3K79me2 programme |
The evidence for 'lower' is in the right disease and the wrong assay for selectivity. The evidence for 'higher' is in the right conceptual place and the wrong cell type entirely. Neither is adequate, and the graph should not pretend otherwise by writing a directional recommendation it cannot support.
A third possibility, worth naming because it would explain the muddle: the direction may be locus-dependent. H2Bub supports elongation at fusion targets and bivalency resolution at differentiation genes, and a global perturbation moves both. If so, no global agent has a coherent effect and the whole layer is untargetable by bulk modulation — which is a finding, not a failure.
| statement | The therapeutic direction of H2Bub modulation in KMT2A-rearranged leukaemia is undetermined: lowering it removes the licence for the fusion's H3K79me2 programme, while raising it is what permits bivalency resolution and differentiation, and the only available selective compounds raise it. |
|---|---|
| origin | literature-gap |
| prior probability | 0.5 |
predictions
[ "If 'lower' is right: USP22 macrocycle treatment of KMT2A-rearranged lines increases H2Bub, increases H3K79me2 at HOXA9/MEIS1, and does not reduce - or increases - proliferation", "If 'raise' is right: the same treatment induces differentiation markers and reduces proliferation despite increased H3K79me2", "If locus-dependence is right: H2Bub rises everywhere, H3K79me2 rises at fusion targets and bivalency resolves at differentiation genes simultaneously, with no coherent net phenotype" ]
experiments
[ "exp-usp22-macrocycle-direction-test-kmt2a" ]
analyses
[ "ana-depmap-h2bub-machinery-kmt2a-selectivity" ]
falsifiers
[ "A dose-response in KMT2A-rearranged lines showing a monotonic, consistent direction of benefit for either raising or lowering H2Bub settles the question and retires this hypothesis", "Demonstration that H2Bub does not change measurably at fusion target genes under either perturbation would show the whole layer is not rate-limiting here" ]
- Getting the direction wrong is not a neutral error here. If the 'lower' model is right, giving a USP22 inhibitor would raise H2Bub and could amplify the fusion's H3K79 methylation programme - an intervention that makes the disease worse.
- The graph's existing gene-usp22 entry proposes USP22 inhibition as a resistance-directed strategy on MEIS1-stability grounds. That rationale and this chromatin rationale are independent and could oppose each other in the same compound.
- Panobinostat is already used in paediatric leukaemia contexts and lowers H2Bub as one of several effects; if the 'raise' model is right, part of its mechanism is working against the differentiation route.
- Is the differentiation argument transferable from mesenchymal stem cells to a B-lymphoid progenitor at all? If not, the 'raise' side of this hypothesis has no evidence in the right cell type and the prior should move.
- Does H2Bub differ between KMT2A-rearranged infant B-ALL blasts and normal age-matched progenitors? A single mass-spectrometry comparison on primary material would ground the whole question and has never been done.
- If the direction is locus-dependent, is there any agent that could act locally rather than globally? Nothing in this beat suggests one.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Suppressing Rnf20 in diverse models of MLL-rearranged leukaemia inhibits proliferation in vitro and in vivo, reduces expression of MLL-fusion target genes resembling Dot1l inhibition, and lowers Dot1l-dependent H3K79 methylation at Hoxa9 and Meis1.
Histone H2B ubiquitin ligase RNF20 is required for MLL-rearranged leukemia. · Wang E, Kawaoka S, Yu M, Shi J, Ni T, Yang W, Zhu J, Roeder RG, Vakoc CR · Proc Natl Acad Sci U S A 2013What this evidence cannot show
Abstract read in full; full text not read. shRNA; no non-MLL comparator described; fusions not specified in the abstract. **Resolved:** the underlying data are **GSE43725, three samples of a single murine RN2 (MLL-AF9/NrasG12D) line, n=1 per hairpin**, with a cooperating NRAS G12D that had not been flagged. The direction being called unknown is unknown from one cell line, not from a survey. Resolved 2026-09-19 by the Q18b unstated-systems sweep from the study's deposition, not from its text; bounded as a statement about the deposited data rather than about the whole study.
Evidence that argues against this (2)
- This is the strongest evidence in the beat that the direction IS known: suppressing Rnf20 across diverse MLL-rearranged models inhibits proliferation in vitro and in vivo. A hypothesis entry titled 'direction is unknown' should state why that result does not settle it — the honest answer being that RNF20 loss removes H2Bub globally rather than at fusion targets, so the phenotype does not localise the requirement. PNAS 2013
- That the direction is unknown is if anything understated by this item: the evidence base is three samples of one mouse line carrying a second oncogenic driver, in the wrong lineage for this project, so neither direction has meaningful support. NCBI Gene Expression Omnibus 2026
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Panobinostat's in vivo anti-leukaemic activity against MLL-rearranged ALL involves depletion of H2B ubiquitination via suppression of the RNF20/RNF40/WAC E3 ligase complex, and knockdown of WAC phenocopies loss of H2B ubiquitination with concomitant cell death induction.
The HDAC inhibitor panobinostat (LBH589) exerts in vivo anti-leukaemic activity against MLL-rearranged acute lymphoblastic leukaemia and involves the RNF20/RNF40/WAC-H2B ubiquitination axis. · Garrido Castro P, van Roon EHJ, Pinhancos SS, Trentin L, Schneider P, Kerstjens M, Te Kronnie G, Heidenreich O, Pieters R, Stam RW · Leukemia 2018What this evidence cannot show
Abstract read in full; full text not read. Panobinostat is a pan-HDAC inhibitor with several epigenetic effects; the abstract does not apportion the in vivo effect between them.
Evidence that argues against this (1)
- Panobinostat's H2Bub depletion is one effect of a pan-HDAC inhibitor among many, so it cannot be used to infer the direction of a requirement. If anything it points the opposite way from the entry's framing: a drug that LOWERS H2Bub is anti-leukaemic, which is a directional claim the entry treats as unavailable. Leukemia 2018
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
H2B monoubiquitination increases significantly during differentiation of human mesenchymal stem cells and lineage-committed precursors; RNF40 is required for induction of differentiation markers and transcriptional reprogramming, dependent on CDK9 and WAC; and RNF40 is required for resolution of the H3K4me3/H3K27me3 bivalent poised state on lineage-specific genes during the transition from inactive to active chromatin.
The histone H2B monoubiquitination regulatory pathway is required for differentiation of multipotent stem cells. · Karpiuk O, Najafova Z, Kramer F, Hennion M, Galonska C, Konig A, Snaidero N, Vogel T, Shchebet A, Begus-Nahrmann Y, Kassem M, Simons M, Shcherbata H, Beissbarth T, Johnsen SA · Mol Cell 2012What this evidence cannot show
Abstract read in full; full text not read. Mesenchymal stem cells and assorted precursors - not haematopoietic, not leukaemic, not KMT2A-rearranged. This is the weakest link in the 'raise' argument and it is load-bearing for it.
Evidence that argues against this (1)
- H2Bub rising during mesenchymal stem cell differentiation is a different lineage and a normal developmental programme. Using it to reason about direction in a leukaemic blast assumes the mark's relationship to differentiation is conserved across cell types, which the entry does not establish. Molecular Cell 2012
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Potent macrocyclic cyclic-peptide inhibitors of USP22, the catalytic subunit of the human SAGA DUB module, were identified from a combinatorial library; the top hit was highly specific against a panel of 44 other human DUBs, entered cells and increased cellular levels of H2B monoubiquitination.
Potent macrocycle inhibitors of the human SAGA deubiquitinating module. · Morgan M, Ikenoue T, Suga H, Wolberger C · Cell Chem Biol 2022What this evidence cannot show
Abstract read in full; full text not read. Tool compounds; no in vivo data, no leukaemia data, no potency value quoted in the abstract. Cyclic peptides with cell entry are unusual and the claim should be checked in the full text before building on it.
Evidence that argues against this (1)
- Genetic ablation of USP22 lowers rather than raises global H2Bub, because USP27X and USP51 then have more access to the shared ATXN7L3/ENY2 adaptors. So the chemical and genetic perturbations of the same protein are reported to move the mark in opposite directions, and which one a patient would experience is not obvious. Atanassov BS et al. 2016
which way to move the mark is undetermined in this disease
- related to →The differentiation block — what it actually is, operationally, in KMT2A::AFF1 leukaemia
raising H2Bub is what permits bivalency resolution and differentiation in the one system where it was measured
dose-response with published selective USP22 macrocycles in KMT2A-rearranged lines
supplies the system the hypothesis said was unspecified: mouse MLL-AF9, not KMT2A::AFF1