Measure the free-versus-7SK-bound P-TEFb split in KMT2A::AFF1 cells - the experiment this whole beat turns on, which has never been done
also called P-TEFb partition assay · 7SK glycerol gradient in KMT2A-r
Spin the contents of leukaemia cell nuclei through a sugar gradient so that heavy particles and light ones separate, then look at which fractions the elongation enzyme is in. It is an old, cheap, unglamorous experiment, and nobody has ever run it on a KMT2A-rearranged leukaemia.
The same thing, for a scientist
Glycerol-gradient sedimentation of nuclear extract separates the large 7SK snRNP-bound form of P-TEFb from the small free form; immunoblotting fractions for CDK9, cyclin T1, HEXIM1, HEXIM2, LARP7 and MEPCE and Northern blotting or RT-qPCR for 7SK gives the partition directly. This is the method by which the reservoir was discovered and by which LARP7's stable-core role was established [E1][E2]. The measurement has never been made in any KMT2A-rearranged cell. The design below adds the two controls that make the result attributable: a lineage-matched comparator arm, because the KMT2A-rearranged panel is myeloid-enriched and the partition may be a lineage property; and a degron arm inside one line, because only removing the fusion from a fixed genetic background can show that the fusion causes any difference found.
Design
Arms.
- KMT2A::AFF1 B-ALL: SEM, RS4;11
- KMT2A::AFF1 AML: MV4-11
- KMT2A::MLLT3 AML: MOLM-13, THP-1 (links directly to the one published HEXIM1 manipulation in a KMT2A-rearranged cell 3)
- Lineage-matched non-KMT2A comparators: a B-ALL pair (for example REH, NALM-6) and an AML pair
- Primary blasts where available, because everything above is a cell line
Readout per arm. 10-30% glycerol gradient on nuclear extract; immunoblot every fraction for CDK9, phospho-CDK9 T186, cyclin T1, cyclin T2, HEXIM1, HEXIM2, LARP7, MEPCE, AFF1 (N-terminal and C-terminal antibodies, to distinguish wild-type from fusion), AFF4 and ELL2; 7SK by Northern or by RT-qPCR on RNA extracted per fraction. Primary outcome: the fraction of total CDK9 signal in the large form.
Attribution arm. A degron-tagged KMT2A::AFF1 line, using the endogenous-tag approach this graph already specifies at exp-endogenous-tag-and-degron-knock-in-kmt2a-aff1. Run the gradient before and 6 h after fusion depletion. If the partition moves, the fusion sets it. If it does not, the partition is a lineage or cell-state property and hyp-kmt2a-aff1-shifts-the-free-versus-sequestered-ptefb-balance is falsified in its causal form.
Perturbation arm (optional, cheap). Treat each line with SAHA or JQ1 and repeat the gradient at 30 and 120 minutes. The size of the acute release is itself informative, because agents that release P-TEFb do almost nothing where the pool is small 4.
Why the antibodies matter more than the gradient
One technical point decides whether this experiment is interpretable. The fusion protein retains only AFF1 residues 347-1210, so an antibody raised against the AFF1 N-terminus detects wild-type AFF1 only and an antibody against the AFF1 C-terminus detects both. Running both lets the gradient answer a second question for free: is wild-type AFF1 still in the 7SK snRNP in these cells, and is the fusion anywhere near it? A KMT2A N-terminal antibody in the same blot gives the fusion's own distribution.
That pair of blots is the direct test of mechanism-kmt2a-aff1-cannot-be-sequestered-in-the-7sk-snrnp, which is currently an argument from residue arithmetic rather than a measurement.
Cost and why it has not been done
Low cost tier: reagents are commercial, the gradient is a day, the blots are standard. The reason it has not been done is not difficulty. It is that the 7SK field works in HeLa, HEK293 and HIV models and the KMT2A field works on chromatin and menin, and (MLL OR KMT2A) AND 7SK returns three PubMed records, none of them a measurement 5. This is a gap of disciplinary adjacency, not of technical reach.
| design | Glycerol-gradient sedimentation (10-30%) of nuclear extract from KMT2A::AFF1 lines (SEM, RS4;11, MV4-11), KMT2A::MLLT3 lines (MOLM-13, THP-1), lineage-matched non-KMT2A comparators and primary blasts where available; immunoblot every fraction for CDK9, phospho-CDK9 T186, cyclin T1, cyclin T2, HEXIM1, HEXIM2, LARP7, MEPCE, AFF1 N-terminal and C-terminal, KMT2A N-terminal, AFF4 and ELL2; 7SK by Northern or per-fraction RT-qPCR. |
|---|---|
| readout | fraction of total CDK9 signal in the large 7SK-containing form; secondary readouts are HEXIM1 per CDK9, phospho-T186 per CDK9, and whether fusion protein is detectable in the large form at all |
| controls | lineage-matched comparator arms because the KMT2A-rearranged panel is myeloid-enriched; a degron arm in one KMT2A::AFF1 line to attribute any difference to the fusion rather than to lineage; an AFF1 N-terminal versus C-terminal antibody pair to separate wild-type AFF1 from the fusion; SAHA or JQ1 time course to size the releasable pool |
| cost tier | low |
| timeline | 6-10 weeks for the cell-line arms; longer if primary blasts are included |
| why not done | the 7SK field works in HeLa, HEK293 and HIV models; (MLL OR KMT2A) AND 7SK returns three PubMed records and none is a measurement |
tests hypotheses
[ "hyp-kmt2a-aff1-shifts-the-free-versus-sequestered-ptefb-balance" ]
related entries
[ "exp-endogenous-tag-and-degron-knock-in-kmt2a-aff1", "exp-five-fraction-blot-attributable-localisation-kmt2a-aff1", "noncoding-rn7sk" ]
- A null result here would remove most of this beat's therapeutic content, and that is the correct outcome to plan for: the honest prior is that the partition is unchanged.
- Is a bulk nuclear partition the right quantity, or does the fusion change a local concentration at its target loci that a gradient would average away?
- Are there validated antibodies that distinguish the KMT2A::AFF1 fusion from wild-type AFF1 by the N-terminal epitope? This graph records that no junction-specific antibody exists for any KMT2A fusion.
- Can 7SK be quantified per fraction reliably by RT-qPCR given the number of RN7SK pseudogenes in the genome?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Glycerol-gradient sedimentation of cell lysate separates the large 7SK-containing P-TEFb complex from the small active form, and this is how the reorganisation of the snRNP after P-TEFb inhibition was demonstrated.
LARP7 is a stable component of the 7SK snRNP while P-TEFb, HEXIM1 and hnRNP A1 are reversibly associated. · Krueger BJ, Jeronimo C, Roy BB, Bouchard A, Barrandon C, Byers SA, Searcey CE, Cooper JJ, Bensaude O, Cohen EA, Coulombe B, Price DH · Nucleic Acids Res 2008What this evidence cannot show
Established in human cell lines; the method's resolution depends on extract preparation, and nuclear versus whole-cell extract changes what is recovered.
Evidence that argues against this (1)
- Gradient fractionation measures a steady state in a lysate and can miss a chromatin-local pool; a bulk partition and a promoter-local concentration are different quantities. Guo C et al. 2020
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
More than half of P-TEFb in HeLa cells is recovered in larger 7SK-containing complexes with very weak kinase activity, establishing the expected shape of the result in a reference cell.
7SK small nuclear RNA binds to and inhibits the activity of CDK9/cyclin T complexes. · Nguyen VT, Kiss T, Michels AA, Bensaude O · Nature 2001What this evidence cannot show
HeLa in 2001; the figure is not resolved to a percentage and should be treated as an order-of-magnitude expectation rather than a benchmark.
Evidence that argues against this (1)
- P-TEFb levels are described as vanishingly low in resting primary CD4+ T cells, so cell state dominates this measurement and a leukaemia blast may sit anywhere on that range. Bartholomeeusen K et al. 2013
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
HEXIM1 has been manipulated in a KMT2A-rearranged cell (MOLM-13, KMT2A::MLLT3) by both knockdown and tetracycline-inducible overexpression, with growth, c-Myc, differentiation and JQ1-response readouts, which makes it the natural positive-control line for a partition assay.
HEXIM1 induction is mechanistically involved in mediating anti-AML activity of BET protein bromodomain antagonist. · Devaraj SG, Fiskus W, Shah B, Qi J, Sun B, Iyer SP, Sharma S, Bradner JE, Bhalla KN · Leukemia 2016What this evidence cannot show
No partition measurement was made in that study; cyclin T1 co-immunoprecipitation with HEXIM1 was assayed, which is a related but coarser readout.
Evidence that argues against this (1)
- A co-IP shows association, not the fraction of the pool; the same paper's JQ1 arm increases cyclin T1 binding to both HEXIM1 and CDK9, which a gradient would resolve and a co-IP cannot. Devaraj SG et al. 2016
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The magnitude of the acute P-TEFb release produced by HDAC inhibitors depends on how much P-TEFb the cell has; in resting primary CD4+ T cells with very low P-TEFb, SAHA has minimal effect until P-TEFb levels are raised.
Histone deacetylase inhibitors (HDACis) that release the positive transcription elongation factor b (P-TEFb) from its inhibitory complex also activate HIV transcription. · Bartholomeeusen K, Fujinaga K, Xiang Y, Peterlin BM · J Biol Chem 2013What this evidence cannot show
An HIV-latency model in T cells; transferring the principle to blasts is an inference.
Evidence that argues against this (1)
- The releasable fraction and the total pool are different quantities, and the T-cell result confounds them. Krueger BJ et al. 2008
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.review
PubMed returns three records for (MLL OR KMT2A) AND 7SK on 2026-09-20 - two reviews and one HIV Tat purification - and none reports a measurement of the P-TEFb partition in a KMT2A-rearranged cell.
CDK9: a signaling hub for transcriptional control. · Bacon CW, D'Orso I · Transcription 2019What this evidence cannot show
A counted absence bounded by the symbol set: PubMed's own MLL/KMT2A expansion and the bare string 7SK. It would miss a paper using only 'RN7SK' or describing the particle without naming 7SK; the complementary queries run for this beat were (MLL OR KMT2A) AND HEXIM1 (2 records) and AND LARP7 and AND MEPCE (1 record each).
Evidence that argues against this (1)
- Complex-level phrasing makes members unreachable from their own symbols in this project's experience, so a study framed at the level of 'P-TEFb equilibrium' in leukaemia might not be retrieved by any of these queries. HUGO Gene Nomenclature Committee 2026
the direct measurement of the partition
the AFF1 N-terminal versus C-terminal antibody pair converts the residue argument into a measurement
the assay that can measure 7SK where poly-A RNA-seq cannot