Oncofusion

start / graph / hyp-fusion-is-exempt-from-restraint-not-free-of-it

Hypothesisspeculative confidence 1 of 5: Our own inference or analogy. Not shown to be true.Q-subtraction-kmt2a

Hypothesis: the fusion is exempt from KMT2A's regulation, not free of it - the brakes still operate at the same loci, on the other allele

also called exemption not absence · allele-resolved restraint · cis-exempt trans-braked · locus-level not molecule-level restraint

Instead of saying the leukemic protein escaped its controls by deleting them, this says the controls are still running at the same genes on the healthy copy of the gene. For one molecule that is just another way of saying the leukemic protein has nothing to grab onto - but for the gene it sits on, it predicts something, and in one case a drug has already proved it.

The same thing, for a scientist

Both restraint modules in the lost region - the PHD2 CDC34-dependent E3 ligase ceiling and the PHD3/CyP33 activator-to-repressor switch - are encoded by the intact KMT2A allele every patient retains, the fusion requires a prebound wild-type MLL1 complex for stable HoxA9 association, and the wild-type Mll allele is genetically required for MLL-AF9 leukemogenesis [E4]. The hypothesis is that at fusion-occupied loci **the wild-type machinery is present and functioning on the wild-type molecule**, and the fusion's advantage is structural immunity to it rather than its removal from the locus. **Scoped cycle 31:** at the level of a single molecule 'exempt' and 'structurally incapable' are the same statement, and the entry says so. The claim that does work is at the level of the **locus** - that the machinery arrives, acts on the other allele's molecule, and thereby changes what the fusion does. That version is falsifiable, and a second instance of it is now on record in a system nobody framed this way: stabilising wild-type MLL1 by blocking its taspase1 cleavage displaces MLL-AFF1, AFF4 and ELL2 from chromatin in SEM cells and prolongs survival in MLL-AF9 mice, without any agent touching the fusion [E5].

Go deeper
Where the hypothesis comes from

Three observations that do not sit comfortably together under the standard reading:

  1. The fusion needs the wild-type complex. MLL1 fusion proteins lack the PHD fingers and require prebinding of a wild-type MLL1 complex for stable HoxA9 association 1. So a PHD-containing complex is at the locus.
  2. CyP33 binding to PHD3 is required for repression of HOXA9 and HOXC8 2. So the machinery that would repress the locus has a substrate there.
  3. Re-inserting PHD3 in cis into the fusion blocks its transforming activity 3. So the fusion is capable of being braked if the brake is physically attached to it - which is precisely what exemption, rather than absence, predicts.

Put together: the brake exists, its substrate is at the locus, and the fusion becomes brakeable the moment it is given a docking site. The distinguishing feature of the fusion is therefore that the brake cannot reach it, not that the brake is gone.

What this predicts that the standard reading does not
  • P1. Inducing CyP33/PPIE in a KMT2A-rearranged cell that retains a wild-type allele will produce measurable repression at HOXA9 with wild-type KMT2A occupancy falling or changing character, and fusion occupancy unchanged. The removal model predicts either no effect or an effect with no allele asymmetry.
  • P2. In a KMT2A-rearranged cell that has lost the wild-type allele, CyP33 induction will have no effect at HOXA9, and - separately - that cell will be less sensitive to agents acting on wild-type-only interfaces such as MM-401.
  • P3. HDAC1 will be detectable at fusion-occupied HOXA9 in a CyP33-dependent manner, i.e. the repressive effector is recruited to the locus even though it cannot act on the fusion.

P2 is the one with immediate clinical shape: it turns wild-type allele status, which is already readable from existing sequencing, into a candidate response biomarker.

P2 has a problem, added cycle 31. P2 assumes KMT2A-rearranged cells that have lost the wild-type allele exist and can be compared. Thiel et al. found that the wild-type Mll allele is required for MLL-AF9 leukemogenesis and maintenance 4. If that generalises, the comparator cell may not exist - which is not a refutation of the hypothesis but removes its cleanest clinical prediction and its best internal control. P2 should be reframed as a dosage prediction rather than a presence/absence one, and the dosage version is directly supported by 5.

Cycle 31: a second instance, and the distinction the hypothesis has to defend

The second instance. The hypothesis was built on one axis - the PHD3/CyP33/HDAC1 brake. It turns out the same shape appears in a completely separate system that nobody framed this way. Taspase1 cleavage of wild-type MLL1 controls its turnover and therefore its chromatin occupancy. The fusion has no cleavage sites, so the mechanism cannot act on it. Yet when you stabilise the wild-type protein - taspase1 knockout, or CKII inhibition with CX-4945 - the extra wild-type MLL1 competes the fusion off chromatin, lowers target-gene expression in SEM cells, and prolongs survival in an MLL-AF9 mouse 5.

That is exemption in its purest form: a regulatory circuit running on the intact allele, structurally unable to touch the fusion, and still decisive for what the fusion does. It also reverses the therapeutic polarity. The PHD3 version of this hypothesis implies you would want to reattach a brake to the fusion. The taspase1 version says you can win by strengthening the wild-type allele's occupancy and letting competition do the work - and unlike the PHD3 version, somebody has already done it in a mouse.

The distinction the hypothesis has to defend. The obvious objection is that 'exempt' is a dressed-up way of saying 'structurally incapable', which is duller and needs no hypothesis. The objection is partly right and the entry should say so: for the PHD3 axis, exemption is structural incapacity - there is no docking site, end of story, and calling it exemption adds nothing except a reminder to go look at the other allele.

But the two are not the same claim, and they are separable by experiment:

  • 'Structurally incapable' predicts nothing about the locus. It is a statement about one molecule. It makes no prediction about what the brake machinery is doing, where it is, or whether it matters.
  • 'Exempt' predicts that the machinery is present and working at the same loci, on the other molecule, with consequences for the fusion. That is a statement about a system, and it is false if the brake is simply absent from fusion-occupied chromatin.

The discriminating measurement is P3 - HDAC1 arriving at fusion-occupied HOXA9 on CyP33 induction. Under structural incapacity, nothing needs to happen at the locus at all. Under exemption, the effector arrives, acts on the wild-type molecule and cannot reach the fusion. Zhao's ELL2 displacement result 5 is the same logic already confirmed for a different circuit: the perturbation changes what happens at the fusion's loci without touching the fusion.

So the honest position is: exemption is not a different mechanism from structural incapacity; it is a different unit of analysis, and it is testable because the locus, not the molecule, is where the prediction lives. A version of this hypothesis that claims more than that is overreaching.

Facts and figures
statementAt fusion-occupied loci the wild-type KMT2A brake machinery (PHD3/CyP33/HDAC1 and PHD2/CDC34) remains present and functional on the intact allele; the fusion's advantage is structural immunity to that machinery rather than its removal from the locus.
originliterature-gap
prior probability0.45
predictions
[
  "CyP33/PPIE induction in a KMT2A-rearranged cell retaining a wild-type allele represses HOXA9, with allele-asymmetric occupancy changes: wild-type KMT2A occupancy altered, fusion occupancy unchanged",
  "In a KMT2A-rearranged cell that has lost the wild-type allele, CyP33 induction has no effect at HOXA9 and the cell is less sensitive to wild-type-only-interface agents such as MM-401",
  "HDAC1 is recruited to fusion-occupied HOXA9 in a CyP33-dependent manner even though the fusion itself cannot be repressed by it",
  "Wild-type KMT2A allele status stratifies response to any agent acting on a module absent from the fusion",
  "Strengthening the wild-type allele's chromatin occupancy - by blocking taspase1 cleavage, or any other route that stabilises wild-type KMT2A - displaces the fusion from its own target loci and is therapeutic, without any agent ever binding the fusion (already observed for CX-4945 in SEM cells and MLL-AF9 mice)",
  "Wild-type KMT2A abundance has a non-monotonic relationship with fusion chromatin occupancy: too little and the fusion cannot load (Thiel), too much and it is competed off (Zhao). There is therefore an optimum, and leukemic cells should sit near it"
]
experiments
[
  "exp-cyp33-induction-allele-resolved-cutrun",
  "exp-hbm-reinsertion-kmt2a-aff1",
  "exp-wildtype-allele-dosage-vs-fusion-occupancy"
]
analyses
[
  "ana-kmt2a-lost-module-ledger"
]
falsifiers
[
  "CyP33/PPIE induction in a wild-type-allele-retaining KMT2A-rearranged cell produces no repression at HOXA9 and no change in wild-type KMT2A occupancy - the brake is not operating at the locus and the simpler 'removed' model wins",
  "Allele-resolved CUT&RUN shows no wild-type KMT2A at fusion-occupied loci at all, contradicting the prebinding requirement and removing the hypothesis's premise",
  "Wild-type allele loss turns out to be common in KMT2A-rearranged leukemia AND those cases behave identically to allele-retaining cases in every respect, which would make the wild-type allele's brake irrelevant to the disease whether or not it operates",
  "A full-text reading of Martin et al. 2003 shows undimerised truncated MLL IS transforming, which would restore 'loss of restraint is sufficient' and make exemption an unnecessary complication",
  "HDAC1 CUT&RUN shows no CyP33-dependent arrival of the repressive effector at fusion-occupied HOXA9. Exemption predicts the machinery reaches the locus; structural incapacity predicts nothing does, and the duller claim then wins outright.",
  "Titrating wild-type KMT2A abundance produces a flat fusion-occupancy curve. Both the Thiel requirement and the Zhao displacement would then be doing something other than competition at shared loci, and the locus-level framing collapses to a molecule-level one."
]
How this could bite us back
  • If the hypothesis is right, 'restore the brake' therapeutics must act on the fusion molecule itself by induced proximity; agents that act on wild-type KMT2A will do nothing useful, because that allele's brake is already engaged and already being ignored.
  • If wild-type allele status turns out to stratify response, then trials that do not measure it will average two different diseases and may bury a real effect - the same failure mode that has hidden biomarker-defined effects in other programmes.
  • A cell that loses the wild-type allele would, under this model, lose the fusion's own prebinding requirement too - so allele loss might reduce fusion chromatin occupancy rather than increase fitness, making the biomarker's direction non-obvious.
  • If the taspase1 route works by competition, then it is dose-limited in a way menin inhibition is not: you can only displace the fusion in proportion to how much wild-type protein you can stabilise, and a cell that silences or deletes the wild-type allele escapes completely. The graph already records wild-type KMT2A silencing in a related context (`subtraction-kmt2a-ptd-wildtype-silencing`), which is exactly the resistance mechanism this approach invites.
Still open
  1. What fraction of KMT2A-rearranged patients retain a functional wild-type allele? Chen et al. state it is 'most commonly, but not always' retained, and nobody has put a number on it from the large sequencing cohorts that could.
  2. Is exemption structural (no docking site) or kinetic (the fusion turns over too fast, or occupies chromatin too transiently, for the brake to engage)? The two have different therapeutic implications and are distinguishable by degron-controlled residence-time experiments.
  3. Does the reciprocal der(4)/der(11) product, where it is expressed, carry any of the lost modules and therefore reintroduce any brake in cis?
  4. Is the CX-4945 survival benefit in MLL-AF9 mice actually caused by wild-type MLL1 stabilisation? CKII has many substrates, and the cleanest test - CX-4945 in a Kmt2a-null-allele or non-cleavable background - has not been run. If the benefit survives loss of the wild-type allele, the displacement model is wrong and so is this hypothesis's second instance.
  5. Does 'exempt' earn its keep anywhere the PHD3 axis does not? On current evidence it is a useful reframing for two circuits (PHD3/CyP33 and taspase1/CKII) and an empty relabelling for the rest. Naming a third would be the strongest argument for the concept; failing to would be an argument for retiring it.
Every source behind this entry (5)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    MLL1 fusion proteins lack the PHD fingers and require prebinding of a wild-type MLL1 complex, together with CXXC-domain recognition of DNA, for stable HoxA9 association; the essential MLL1 recruitment domain includes the CXXC domain and the PHD fingers.

    Multiple interactions recruit MLL1 and MLL1 fusion proteins to the HOXA9 locus in leukemogenesis · Milne TA, Kim J, Wang GG, Stadler SC, Basrur V, Whitcomb SJ, Wang Z, Ruthenburg AJ, Elenitoba-Johnson KSJ, Roeder RG, Allis CD · Molecular Cell 38(6):853-863 2010
    PMID 20541448DOI
    What this evidence cannot show

    ABSTRACT ONLY. One locus, cell systems, recruitment rather than transformation as the endpoint. 'Prebinding' is the authors' model.

    Evidence that argues against this (1)
    • The prebinding model is the sole basis for this hypothesis's premise that wild-type MLL1 is at fusion-occupied loci, and it is a recruitment result at one locus. A stronger and partly awkward result now exists: the wild-type Mll allele is genetically REQUIRED for MLL-AF9 leukemogenesis and maintenance. That supports the premise but undermines prediction P2, which assumes a comparable KMT2A-rearranged cell lacking the wild-type allele exists to be compared. If wild-type KMT2A is required, the cleanest test of this hypothesis may have no comparator, and P2 must be reframed as a dosage prediction. Thiel AT et al. 2010
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CyP33's PPIase domain isomerises a proline in the MLL1 PHD3-Bromo linker, disrupting the PHD3-Bromo interface and facilitating PHD3 binding to the CyP33 RRM; H3K4me3/2 and CyP33-RRM occupy different PHD3 surfaces and can bind simultaneously; and the MLL1-CyP33 interaction is required for repression of HOXA9 and HOXC8 in vivo.

    Pro isomerization in MLL1 PHD3-bromo cassette connects H3K4me readout to CyP33 and HDAC-mediated repression · Wang Z, Song J, Milne TA, Wang GG, Li H, Allis CD, Patel DJ · Cell 141(7):1183-1194 2010
    PMID 20541251DOI
    What this evidence cannot show

    ABSTRACT ONLY. Structural and cell-line work; 'in vivo' means in cells. The repression experiments were not performed in a KMT2A-rearranged, fusion-driven context, which is exactly the gap this hypothesis proposes to fill.

    Evidence that argues against this (1)
    • The proline-isomerisation mechanism is elegant and it concerns PHD3, which every major-BCR fusion has already lost - the PHD3-bromo cassette lies in exons 14-16, 3' of the breakpoint cluster. So this describes a restraint the fusion is exempt from by deletion rather than by escape, which is a weaker claim than the hypothesis's framing of active exemption. Cell 2010
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Inserting the third PHD finger of MLL into MLL-ENL restores CyP33 and HDAC1 recruitment, mediates CyP33-dependent HOXC8 downregulation, and blocks the fusion's haematopoietic stem cell immortalisation potential in serial plating colony assays; PHD1+PHD2 insertion does not.

    Loss of MLL PHD finger 3 is necessary for MLL-ENL-induced hematopoietic stem cell immortalization · Chen J, Santillan DA, Koonce M, Wei W, Luo R, Thirman MJ, Zeleznik-Le NJ, Diaz MO · Cancer Research 68(15):6199-6207 2008
    PMID 18676843DOI
    What this evidence cannot show

    ABSTRACT ONLY, single study, unreplicated, in vitro, one fusion - see ana-phd3-addback-replication-audit. Confidence lowered to 'moderate' for that reason, not because the citation is doubtful. Its role here is narrow: it shows the fusion CAN be braked when a docking site is attached, which is what makes 'exemption' a coherent alternative to 'absence'.

    Evidence that argues against this (1)
    • The PHD3 add-back result is the entry's evidence that the fusion CAN be braked when given a docking site - but a 60-residue folded insertion into a chimeric protein may reduce its abundance, and the KMT2A half's better-evidenced losses are all degradation determinants, so an insertion effect on protein level is now the leading alternative explanation rather than a remote one. Until an abundance-matched control arm is run, 'the fusion becomes brakeable' cannot be separated from 'the fusion becomes scarcer'. Wang J et al. 2012
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The wild-type Mll allele is required for MLL-AF9-induced leukemogenesis and for maintenance of MLL-AF9-transformed cells; menin recruits both wild-type MLL and the MLL-AF9 fusion to HOX loci, and wild-type MLL controls a distinct component of the fusion-induced histone methylation.

    MLL-AF9-induced leukemogenesis requires coexpression of the wild-type Mll allele · Thiel AT, Blessington P, Zou T, Feather D, Wu X, Yan J, Zhang H, Liu Z, Ernst P, Koretzky GA, Hua X · Cancer Cell 17(2):148-159 2010
    PMID 20159607DOI
    What this evidence cannot show

    ABSTRACT ONLY. Mouse genetics in one fusion; a genetic requirement in a transplantation model, not a human observation. This upgrades the hypothesis's premise from 'wild-type KMT2A is present at the locus' (a recruitment model, [E1]) to 'the leukemia needs it', which is a much stronger foundation - and it also makes prediction P2 harder, because a cell that has lost the wild-type allele may simply not exist.

    Evidence that argues against this (1)
    • Fourth appearance of this mouse result in the beat. For an exemption hypothesis it is double-edged: needing the wild-type allele means the restraint machinery is still present and functioning in the cell, so what is being claimed is exemption of one molecule within a restrained system, not exemption of the system. Cancer Cell 2010
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    A second, independent instance of the exemption pattern: stabilising wild-type MLL1 - by taspase1 knockout, or pharmacologically by inhibiting the CKII phosphorylation that licenses taspase1 cleavage - raises wild-type MLL1 chromatin occupancy and thereby DISPLACES MLL-AFF1, AFF4 and ELL2 from chromatin in SEM cells with reduced target-gene mRNA, while CX-4945 delays progression and prolongs survival in an MLL-AF9 mouse model. The fusion, lacking the cleavage sites, is itself untouched by the perturbation.

    Regulation of MLL/COMPASS stability through its proteolytic cleavage by taspase1 as a possible approach for clinical therapy of leukemia · Zhao Z, Wang L, Volk AG, Birch NW, Stoltz KL, Bartom ET, Marshall SA, Rendleman EJ, Nestler CM, Shilati J, Schiltz GE, Crispino JD, Shilatifard A · Genes & Development 33(1-2):61-74 2019
    PMID 30573454DOI
    What this evidence cannot show

    FULL TEXT read via PMC6317322. The mechanistic experiments are in 293T/MCF7/HCT116; the leukemia arms are SEM/MV4-11/RS4;11 plus an MLL-AF9 transplant. CX-4945 has targets other than CKII, so the survival benefit is not formally attributed to MLL1 stabilisation. Cited here because it is a second, fully independent case of the hypothesis's central shape - a regulatory event that still runs on the wild-type allele, cannot touch the fusion, and nonetheless determines what the fusion does - reached by a group that was not thinking about restraint at all.

    Evidence that argues against this (1)
    • Calling this a second INDEPENDENT instance overstates it. Both instances run through the same taspase1/stability axis and the same paper supplies the pharmacological arm, so the pattern is one mechanism observed twice rather than two mechanisms agreeing. CX-4945 also inhibits CKII broadly, so the pharmacological instance is not a clean second observation. Genes & Development 2019
Connected to