The partner's deubiquitinase domain is what lets a PHD-retaining KMT2A fusion survive and transform
also called UCH rescue hypothesis · Marschalek UCH degradation-rescue proposal · why minor-BCR fusions transform
KMT2A fusions that keep a set of protein fingers called PHD domains should not cause cancer, because those fingers normally trigger the protein's destruction. The idea here is that when the partner gene is a de-ubiquitinating enzyme, it removes the destruction tag and rescues the fusion, which would explain why the rare minor-breakpoint fusions work at all.
The same thing, for a scientist
Breaks in the KMT2A minor BCR place the complete PHD1-3 cluster, bromodomain and ePHD4 into the direct fusion protein. PHD2 binds CDC34 and PHD3/bromodomain binds a partner that together mediate polyubiquitination and proteasomal destruction of MLL, and adding two or more PHD fingers back into MLL-AF9 abolishes immortalisation of haematopoietic progenitors [E1][E2]. Yet 47 patients in the DCAL series have minor-BCR fusions and 30 of them are KMT2A::USP2 [E3]. Meyer et al. propose the resolution: the partner in the commonest minor-BCR fusions is a deubiquitinase, and the UCH domain fused to the PHD-containing MLL portion counteracts that degradation mechanism, restoring transformation capacity [E1]. The hypothesis is attractive because it is partner-specific in exactly the right way - USP2 (30 patients) and USP8 (3) are both UCH-domain proteins and both are minor-BCR-restricted - and it makes sharp predictions about protein half-life and about catalytic-dead mutants. **It is also, as stated, untested: no half-life measurement, no ubiquitination assay, no catalytic-dead control and no transformation assay with a real KMT2A::USP2 construct has been published.** And it has a visible problem: two AFF1 and one MLLT3 minor-BCR cases supply no deubiquitinase at all [E1].
The chain of reasoning, and where each link sits on the evidence scale1 deeper
| link | status |
|---|---|
| Minor-BCR breaks retain PHD1-3, the bromodomain and ePHD4 | established from the breakpoint interval and the KMT2A domain map 3 |
| PHD fingers added back to MLL-AF9 block immortalisation and Hoxa9 upregulation | strong, one murine study, MLL-AF9 only 2 |
| PHD2 and PHD3/BD recruit factors that polyubiquitinate and degrade MLL | cited, not read here. Asserted by Meyer et al. citing prior work; this entry carries it only as the authors' framing 1 |
| KMT2A::USP2 and KMT2A::USP8 retain the conserved UCH domain | moderate, a sequence observation in the founding letter 1 |
| The UCH domain deubiquitinates the fusion and rescues it from degradation | speculative. Nobody has measured it 1 |
The entry's confidence is set by the weakest link, which is the last one.
What would make it beautiful
If true, this is a rare case of a fusion partner contributing a catalytic function that acts on the fusion protein itself rather than on the transcriptional machinery. Every other partner class in this graph - elongation (AFF1, MLLT1, MLLT3, ELL), oligomerisation (AFDN, GPHN, septins), or chromatin - acts in trans on the genome. A self-stabilising fusion is a different kind of object, and it would predict that the therapeutic handle is degradation rather than transcription.
The three patients who argue against it
In the same discovery cohort, two KMT2A::AFF1 and one KMT2A::MLLT3 case broke in the minor BCR 1. AFF1 and MLLT3 contribute elongation modules and no deubiquitinase activity of any kind. If PHD retention made a fusion non-viable without a UCH rescue, those three patients should not exist.
Three readings survive this, and the hypothesis is not dead, but it must be stated in the weaker form:
- UCH rescue is sufficient but not necessary. A strong enough elongation partner may overcome the PHD block by a different route, which would also explain why AFF1 and MLLT3 are rare in BCR2 while USP2 dominates it.
- The PHD block is dosage- or context-dependent and the murine MLL-AF9 add-back overstates it.
- The rescue is not about degradation at all and the UCH domain contributes something else entirely - oligomerisation, for instance, which is the conserved core across otherwise unrelated KMT2A partners.
The experiment below distinguishes 1 and 3 from the trivial case where the UCH domain contributes nothing.
Why the answer matters therapeutically
If the fusion depends on its own deubiquitinase activity to persist, then a USP2 inhibitor is a degrader by proxy for this genotype, and the target is the fusion protein's stability rather than its transcriptional output. That is a fundamentally different therapeutic axis from menin inhibition, and it would be the first KMT2A fusion with a partner-specific small-molecule rationale that does not route through menin. If the UCH domain is dispensable, the genotype falls back to whatever the retained KMT2A modules do, and menin inhibition becomes the obvious thing to test - which, remarkably, nobody has.
| statement | In KMT2A fusions that break in the minor BCR and therefore retain the PHD1-3 cluster and bromodomain, a partner-supplied deubiquitinase (UCH) domain rescues the fusion protein from PHD-dependent polyubiquitination and proteasomal degradation, and this rescue is what permits transformation. Stated in the sufficiency form, because three minor-BCR patients have non-deubiquitinase partners. |
|---|---|
| origin | literature-gap |
| prior probability | 0.25 |
predictions
[ "A KMT2A(1-BCR2)::USP2 fusion protein will have a measurably longer half-life than the same KMT2A portion fused to an inert C-terminus, in the same cell, under cycloheximide chase.", "A catalytic-dead UCH mutant (USP2 C276A) fusion will lose that half-life advantage and lose transforming activity in a colony-replating assay, while retaining expression at the mRNA level.", "Proteasome inhibition will rescue the catalytic-dead mutant's protein level but not necessarily its transforming activity.", "Ubiquitin-chain profiling of the wild-type versus catalytic-dead fusion will show reduced K48-linked chains on the wild-type.", "If instead the UCH domain works by oligomerisation rather than catalysis, the catalytic-dead mutant will transform normally and the prediction above fails cleanly." ]
experiments
[ "exp-usp2-uch-catalytic-dead-fusion-transformation" ]
analyses
[ "ana-depmap-usp2-maml2-lineage-split" ]
falsifiers
[ "A catalytic-dead USP2 UCH mutant fused to the same KMT2A portion transforms progenitors as well as the wild-type fusion. This kills the catalysis-based version outright.", "The wild-type KMT2A::USP2 fusion protein shows no half-life advantage over a PHD-retaining KMT2A fusion with an inert partner.", "PHD-dependent degradation of MLL turns out not to occur in the cell type in which this fusion arises, making the rescue unnecessary regardless of whether the UCH domain is active.", "Muntean's PHD block is shown to be caused by loss of Hoxa9 locus association rather than by protein degradation, in which case deubiquitination cannot relieve it.", "A KMT2A::AFDN minor-BCR fusion, which supplies no UCH domain, is shown to transform at the same efficiency as KMT2A::USP2 with the same KMT2A portion - reducing the UCH contribution to zero explanatory value." ]
- If the hypothesis is right, the therapeutic implication is a USP2 catalytic inhibitor - but USP2's best-characterised substrate is MDM2, whose stabilisation lowers p53. A systemic USP2 inhibitor would raise p53 everywhere, which is a plausible benefit and an obvious source of on-target toxicity in proliferating normal tissue.
- If the hypothesis is wrong and the UCH domain is inert in the fusion, then the minor-BCR genotype has no partner-specific handle at all and falls back on the retained KMT2A modules - including a PHD1-3/bromodomain unit that can convert KMT2A into a repressor, which would make transcriptional-activation-based drugs the wrong class entirely.
- Either way, the hypothesis gives a reason to expect KMT2A::USP2 to behave unlike KMT2A::AFF1 under any therapy chosen on the basis of 'KMT2A-rearranged', which is how these patients are currently treated.
- Does the KMT2A::USP2 fusion protein have detectable deubiquitinase activity at all? This is the gating experiment and it has never been done.
- Is the fusion protein its own substrate, or does it deubiquitinate something in trans? An auto-stabilising fusion and a fusion that stabilises MDM2 are different molecules with different drug targets.
- Do the three non-deubiquitinase minor-BCR patients (two AFF1, one MLLT3) have any distinguishing feature - a 3' deletion, a second lesion, a different exact breakpoint - that would explain how they escape the PHD block?
- Does KMT2A::AFDN, the second commonest minor-BCR partner with no UCH domain, behave like USP2 or unlike it in a transformation assay? This single comparison would settle whether the UCH story explains BCR2 or only part of it.
- E1speculative confidence 1 of 5: Our own inference or analogy. Not shown to be true.paper
The authors of the founding minor-BCR report propose that the UCH domain of MLL-USP2 or MLL-USP8 counteracts the well-described PHD-dependent polyubiquitination and proteasomal degradation of MLL, restoring oncogenic transformation capacity; the same cohort also contained two AFF1 and one MLLT3 minor-BCR case with no deubiquitinase contribution.
Human MLL/KMT2A gene exhibits a second breakpoint cluster region for recurrent MLL-USP2 fusions. · Meyer C, Lopes BA, Caye-Eude A, Cavé H, Arfeuille C, et al., Marschalek R · Leukemia 2019What this evidence cannot show
This is a discussion paragraph in a Letter, explicitly hedged ('may change profoundly', 'may now be counteracted'). No experiment in the paper addresses it. The upstream claims about PHD2/CDC34 and PHD3/bromodomain-mediated degradation are cited to earlier literature that was not read in this session. Full text read (PMC6756029).
Evidence that argues against this (1)
- The same letter reports two AFF1 and one MLLT3 minor-BCR case. Those fusions retain the PHD fingers and supply no deubiquitinase, so UCH rescue cannot be necessary for a PHD-retaining fusion to cause leukaemia. The hypothesis must be weakened to sufficiency. Meyer C et al. 2023
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Inclusion of two or more PHD fingers in MLL-AF9 blocked immortalisation of haematopoietic progenitors, reduced association with the Hoxa9 locus and suppressed Hoxa9 upregulation, establishing the block the UCH rescue is supposed to relieve.
The PHD fingers of MLL block MLL fusion protein-mediated transformation. · Muntean AG, Giannola D, Udager AM, Hess JL · Blood 2008What this evidence cannot show
Murine retroviral immortalisation assay, MLL-AF9 only, no in vivo endpoint, no minor-BCR fusion tested and no measurement of fusion-protein half-life - so it establishes a functional block but not the degradation mechanism that this hypothesis assumes underlies it. Abstract and central claim read in this session.
Evidence that argues against this (1)
- Muntean et al. attribute the block to reduced Hoxa9 locus association and reduced Hoxa9 upregulation, i.e. a chromatin-targeting effect, not to protein destabilisation. If the block is targeting rather than degradation, a deubiquitinase rescue would not be expected to relieve it at all, and the hypothesis loses its mechanistic footing. Meyer C et al. 2019
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Forty-seven of 3401 genomically characterised KMT2A rearrangements break in the minor BCR, USP2 accounts for 30 patients (rank 9 overall) and USP8 for 3, and a BCR2 break places the complete PHD1-3 cluster, bromodomain and ePHD4 into the direct fusion protein.
The KMT2A recombinome of acute leukemias in 2023. · Meyer C, Larghero P, Almeida Lopes B, Burmeister T, Gröger D, et al., Marschalek R · Leukemia 2023What this evidence cannot show
Single referral centre, so the enrichment of UCH-domain partners within BCR2 could in principle reflect referral patterns. The domain-retention claim is positional, inferred from the breakpoint interval rather than from protein-level evidence in any patient. Counts read from the published article in this session.
Evidence that argues against this (1)
- USP2 and USP8 together are 33 of 47 BCR2 patients, so UCH-domain partners are common but not universal in BCR2; AFDN alone supplies 9 with no deubiquitinase activity. A correlation of 33/47 is suggestive, not a rule, and the AFDN cases are a standing counterexample the hypothesis does not address. Meyer C et al. 2019
the hypothesis exists to explain how this fusion transforms despite retaining the PHD fingers
applies to every BCR2 fusion, because all of them retain PHD1-3 and the bromodomain
C276A catalytic-dead arm: equivalent transformation falsifies the catalytic form of the hypothesis
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
counterexample: two KMT2A::AFF1 minor-BCR patients retain the PHD fingers with no deubiquitinase partner, so UCH rescue cannot be necessary
- ← supportsKMT2A::USP8, t(11;15)(q23;q21) - three patients, and the second instance of a deubiquitinase partner
second, independent deubiquitinase partner with a different fusion geometry - the shared feature is the UCH domain, not the way the break happened
The 30-patient member of the deubiquitinase pair; its analogy section adds the regulatory-domain reading - the fusion delivers catalysis stripped of USP2's N-terminal MDM4-interaction region - and proposes the USP8 arm that would discriminate catalytic class from partner identity.