Oncofusion

start / graph / hyp-kmt2a-aff1-shifts-the-free-versus-sequestered-ptefb-balance

Hypothesisspeculative confidence 1 of 5: Our own inference or analogy. Not shown to be true.Q35-ptefb-reservoir

KMT2A::AFF1 cells run with more free P-TEFb than other leukaemias - and if they do not, this graph's CDK9 story is about recruitment only

also called altered P-TEFb partition hypothesis · free P-TEFb in KMT2A-r

If the fusion makes the cell hold less of its elongation enzyme in reserve, then drugs that block the enzyme should work differently in these cells than in others. Nobody has ever measured the reserve in a KMT2A-rearranged leukaemia, so the question is completely open.

The same thing, for a scientist

**Statement.** In KMT2A-rearranged blasts a larger fraction of nuclear P-TEFb is in the free, active pool and a smaller fraction is in the 7SK snRNP than in lineage- and proliferation-matched non-KMT2A leukaemia cells. **Origin: literature gap plus a weak data pattern.** The gap is total - no measurement of the P-TEFb partition exists in any KMT2A-rearranged cell; `(MLL OR KMT2A) AND 7SK` returns three PubMed records, none of them a measurement. The data pattern is that KMT2A-rearranged lines are *less* dependent on HEXIM1 than other leukaemia lines in both lineage halves (+0.130 difference, myeloid -0.208 and lymphoid -0.267 against -0.385), which is what one would expect if the brake were already partly disengaged - but +0.13 on ten lines is at the edge of noise. **Prior probability 0.3**, and deliberately not higher, because this beat's central structural finding cuts against the most obvious mechanism: the fusion cannot enter the reservoir at all, since it lacks AFF1 1-346 including the cyclin T1-binding sequence, so it cannot be draining the reservoir by competing for P-TEFb inside it. Two mechanisms remain available: HEXIM1 protein is lower because autophagy degrades it in MLL-fusion AML; or SEC-scaffold concentration draws P-TEFb out of the soluble HEXIM1-bound pool. **The experiment that settles it is a glycerol gradient and has never been run in this disease.**

Go deeper
What each outcome would mean, fixed before anyone runs it
resultinterpretationconsequence
Free P-TEFb fraction higher in KMT2A-rthe fusion changes the partition, not just the recruitmentCDK9 inhibitors are acting on an enlarged pool; combination with a reservoir-filling agent becomes rational; HEXIM1 protein becomes a candidate biomarker
Free fraction the samethe fusion recruits whatever free P-TEFb existsthis graph's elongation-hijack account is a recruitment story only, and nothing about the reservoir is disease-specific; the beat's therapeutic content collapses to the CDK9/menin combination
Free fraction lowerthe fusion sequesters more, perhaps by stabilising the particlewould invert the whole picture and would predict that releasing P-TEFb is harmful, making HDAC and BET inhibitors' acute release phase the relevant one

The third outcome is not idle. HDAC and BET inhibitors, which are active in this disease, acutely release P-TEFb rather than sequestering it, and their activity is minimal in cells with a small P-TEFb pool. If KMT2A-rearranged blasts had a small free pool, that would be a coherent explanation for BET-inhibitor sensitivity and would predict the opposite therapeutic direction from the one the brief assumed.

Why this is worth doing even though the fusion is not a reservoir client

The structural finding in mechanism-kmt2a-aff1-cannot-be-sequestered-in-the-7sk-snrnp removes the simplest mechanism but not the hypothesis. A cell's P-TEFb partition is set by HEXIM1 and HEXIM2 protein, by CDK9 T186 phosphatase and kinase activity, by 7SK abundance, and by how much scaffold is competing for the free pool. The fusion plausibly changes at least two of those - it occupies AFF4 and ELL2, the same subunits the HEXIM1 autoinduction loop requires, and MLL-fusion AML degrades HEXIM1 by autophagy. Neither route needs the fusion to touch 7SK.

Facts and figures
statementIn KMT2A-rearranged blasts a larger fraction of nuclear P-TEFb is free and a smaller fraction is 7SK-bound than in lineage- and proliferation-matched non-KMT2A leukaemia cells.
originliterature-gap
prior probability0.3
measurement barrier7SK is a non-polyadenylated Pol III transcript and returns zero rows in the Human Protein Atlas cell-line table across all 1,206 lines, so no existing RNA resource in this project can measure it; total-RNA or fractionation methods are required
predictions
[
  "Glycerol-gradient fractionation of KMT2A::AFF1 lines (SEM, RS4;11) shows a lower proportion of CDK9 in the large 7SK-containing form than matched non-KMT2A B-ALL lines.",
  "HEXIM1 protein, normalised to CDK9, is lower in KMT2A-rearranged blasts.",
  "CDK9 T186 phosphorylation, the capture tag, is lower in KMT2A-rearranged blasts.",
  "The acute free-P-TEFb surge produced by HMBA, SAHA or JQ1 is smaller in KMT2A-rearranged cells because less is left to release."
]
experiments
[
  "exp-glycerol-gradient-ptefb-partition-kmt2a-aff1"
]
analyses
[
  "ana-ptefb-reservoir-enumeration"
]
falsifiers
[
  "Glycerol-gradient fractionation shows the same proportion of CDK9 in the large 7SK-bound form in KMT2A-rearranged and matched comparator lines.",
  "HEXIM1 protein per CDK9 is equal or higher in KMT2A-rearranged blasts than in matched comparators.",
  "Degron depletion of the fusion in a KMT2A::AFF1 line does not change the partition, which would show the partition is a lineage property rather than a fusion consequence.",
  "The acute P-TEFb release produced by SAHA or JQ1 is the same size in KMT2A-rearranged and comparator cells."
]
How this could bite us back
  • If the partition is unchanged, the reservoir is not a disease-specific handle and every intervention on it is a global elongation brake with no window.
  • If the partition is shifted towards free P-TEFb, the corresponding therapy is to refill the reservoir - which is the intervention with the worst therapeutic index in this beat, because it brakes elongation in every proliferating cell.
Still open
  1. Is the partition a property of the fusion or of the lineage? Only a degron experiment inside one cell separates them.
  2. Does the partition change with menin-inhibitor treatment, as the fusion leaves chromatin?
  3. Is the relevant pool nuclear-total or local? A chromatin-associated free P-TEFb concentration at fusion target loci could differ from the bulk partition and would need imaging, not a gradient.
Every source behind this entry (4)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    In HeLa cells more than half of P-TEFb is sequestered in a 7SK-containing particle with very weak kinase activity, establishing that a substantial reserve exists and can change.

    7SK small nuclear RNA binds to and inhibits the activity of CDK9/cyclin T complexes. · Nguyen VT, Kiss T, Michels AA, Bensaude O · Nature 2001
    PMID 11713533DOI
    What this evidence cannot show

    HeLa only, 2001, gradient fractionation. 'More than half' is not resolved to a number, and the fraction may differ entirely in a haematopoietic cell.

    Evidence that argues against this (1)
    • In resting primary CD4+ T cells P-TEFb levels are described as vanishingly low, so the size of the pool is highly cell-state dependent and a HeLa figure should not be assumed anywhere else. Bartholomeeusen K et al. 2013
  2. E2weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.database

    KMT2A-rearranged lines are less dependent on HEXIM1 than other leukaemia lines, consistently in both lineage halves: myeloid -0.208 and lymphoid -0.267 against -0.385 in other leukaemia lines, ranking HEXIM1 17,307th of 17,728 on index-minus-comparator difference (+0.130).

    What this evidence cannot show

    Six myeloid and four lymphoid lines; +0.130 is at the edge of the noise band this project uses at that n. A knockout of a brake is a poor proxy for a change in a partition. This is offered as the weak data pattern that motivated the hypothesis, not as support for it.

    Evidence that argues against this (1)
    • In MOLM-13, which is in this arm, HEXIM1 knockdown measurably increases growth and c-Myc, so the flat CRISPR difference does not mean HEXIM1 is doing nothing in these cells - which weakens the inference in both directions. Devaraj SG et al. 2016
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    The KMT2A::AFF1 fusion protein begins at AFF1 residue 347 and therefore lacks the AFF1 cyclin T1-binding sequence (1-308) that is necessary and sufficient for AFF1's incorporation into the 7SK snRNP, so the fusion cannot deplete the reservoir by entering it.

    What this evidence cannot show

    A residue-position argument combined with a HeLa deletion series. It removes one candidate mechanism; it does not bear on the partition itself.

    Evidence that argues against this (1)
    • The fusion heterodimerises with AFF4 and wild-type AFF1 through the retained CHD, so a fusion-containing complex could carry a cyclin T1-binding sequence contributed in trans and could in principle still compete for reservoir-bound P-TEFb at the complex level. Lu H et al. 2014
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Autophagy degrades HEXIM1 in AML cells expressing MLL-fusion proteins or mutant NPM1, providing a mechanism by which the brake could be thinned without the fusion touching the reservoir.

    Autophagy mediates proteolysis of NPM1 and HEXIM1 and sensitivity to BET inhibition in AML cells. · Huang M, Garcia JS, Thomas D, Zhu L, Nguyen LX, Chan SM, Majeti R, Medeiros BC, Mitchell BS · Oncotarget 2016
    PMID 27732946DOI
    What this evidence cannot show

    AML cell lines and primary AML; the MLL-fusion arm is not resolved to a partner, no B-lineage or infant cell, and no 7SK or partition measurement.

    Evidence that argues against this (1)
    • The paper's own confounder: JQ1 and I-BET-151 both inhibit autophagy and raise HEXIM1, so the drug used to probe the mechanism moves the same marker. Huang M et al. 2016
Connected to