CDK9 - the kinase the whole SEC story runs through, which this graph held only as half of a two-protein summary entry

This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.

• CDK9 is depleted in 99.3% of DepMap lines. Any therapeutic index has to come from scheduling, exposure or a differential dependence on MYC and MCL-1 turnover, not from target selectivity.
• Enitociclib reduced healthy-donor PBMC and CD34+ viability dose-dependently in the same paper that reports the synergy, so the normal-marrow cost of the combination is real and unquantified.
• The short half-lives of MCL-1 and MYC are what make transient CDK9 inhibition look selective; a blast that raises BCL-2 or stabilises MYC escapes without ever altering CDK9.
• CDK9 inhibitors have reached phase 3 repeatedly over fifteen years with no phase 4. Treating a four-cell-line synergy signal as a route to the clinic repeats that history.
• If the fusion reaches P-TEFb only in trans, then inhibiting CDK9 suppresses the fusion's output and every other gene's output by exactly the same mechanism, with no fusion-specific step to exploit.
• No CDK9 inhibitor in this graph distinguishes the 42-kDa nucleoplasmic isoform from the 55-kDa nucleolar one (PMID 15452830), and they share the catalytic domain. An on-target nucleolar effect - ribosome biogenesis - would therefore be inseparable from the intended elongation effect by any measurement that scores CDK9 activity rather than compartment, which is a candidate source of the narrow therapeutic index these agents have shown.
• If a CDK9 inhibitor works partly by lowering H2B monoubiquitination, its on-target toxicity is broader than the pause-release framing implies: H2Bub1 rises during normal differentiation of human mesenchymal and lineage-committed progenitors and is required for resolving poised chromatin at lineage genes, so the same mechanism that damages the leukaemia would be expected to impair normal differentiation - in an infant, and on top of a disease already defined by a differentiation block.

1. Is CDK9 T186 phosphorylation - the tag that decides capture by 7SK - different in KMT2A-rearranged blasts than in other leukaemia blasts? Nobody has blotted it.
2. Does the fusion's chromatin occupancy require CDK9 catalytic activity or only its physical presence? An analogue-sensitive or kinase-dead CDK9 would separate scaffolding from kinase.
3. Since the AFF-family contact is on cyclin T1 rather than on CDK9, would a cyclin T1 surface binder phenocopy an ATP-site CDK9 inhibitor with a different toxicity profile?
4. What is the 42-kDa to 55-kDa CDK9 isoform ratio in a KMT2A-rearranged blast, and does it differ between the myeloid and the B-lineage form of the disease? The isoforms sit in different subnuclear compartments and are differentially expressed across normal monocytes, macrophages and lymphocytes, so this is a one-Western-blot question that no paper has answered.
5. Enitociclib lowers MCL-1 and potentiates venetoclax in KMT2A-rearranged cells, and this graph records infant KMT2A-rearranged ALL as BCL-2-dependent rather than BCL-XL-co-dependent. Does BH3 profiling under CDK9 inhibition move the priming measurement in the direction that prediction implies?
6. Does a CDK9 inhibitor lower H2BK120ub in a KMT2A-rearranged cell, and does H3K79me2 at fusion targets fall with it? CDK9 activity maintains global and gene-associated H2Bub1, H2Bub is the trans-histone input to H3K79 methylation in this graph's own mechanism entry, and RNF20 is required for MLL-rearranged leukaemia - yet no CDK9 inhibitor has been paired with a histone-mark readout in this disease. This is an immunoblot panel and a ChIP, not a programme.
7. Is the enitociclib/menin-inhibitor synergy a pause-release effect or an H2Bub effect? The two mechanisms predict different partners: pause release predicts additivity with BET inhibition, an H2Bub route predicts interaction with DOT1L inhibition and with RNF20/RNF40 loss. The current explanation is the pause one by default, not by test.
8. Is UBE2A actually a CDK9 substrate? The source calls it 'probably' the relevant Rad6 homolog target, and it is the step that would make the phospho-to-ubiquitin route direct rather than inferential.

1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   The AFF4 N-terminal recognition region binds P-TEFb by meandering over the surface of cyclin T1 and makes no stable contacts with the CDK9 kinase subunit; Tat is positioned to contact AFF4 directly and enhances P-TEFb affinity for AFF4.

   The AFF4 scaffold binds human P-TEFb adjacent to HIV Tat. · Schulze-Gahmen U, Upton H, Birnberg A, Bao K, Chou S, Krogan NJ, Zhou Q, Alber T · Elife 2013
   PMID 23471103DOI
   ▸What this evidence cannot show

   A 2.94 A crystal of recognition fragments rather than of an intact SEC, and the AFF-family chain solved is AFF4, not AFF1. No leukaemia experiment anywhere in the paper.

   ▸Evidence that argues against this (1)

   • A structure of isolated recognition regions does not establish that the same geometry holds in a chromatin-bound SEC carrying ELL2, ENL and a KMT2A fusion; this graph's structural beat notes that essentially every deposit covering this interface is an HIV structural-virology paper, so the contact has only been solved in a Tat context. Lu H et al. 2014
2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   Phosphorylation of CDK9 on the conserved T-loop residue Thr-186 is required for the 7SK-P-TEFb interaction, and the stress-induced disruption of that interaction is caused by dephosphorylation of P-TEFb rather than by a change in 7SK.

   Phosphorylated positive transcription elongation factor b (P-TEFb) is tagged for inhibition through association with 7SK snRNA. · Chen R, Yang Z, Zhou Q · J Biol Chem 2004
   PMID 14627702DOI
   ▸What this evidence cannot show

   A reconstituted in vitro system plus in vivo binding assays in HeLa; the responsible kinase was not identified in that paper, and no leukaemia cell was used.

   ▸Evidence that argues against this (1)

   • Later work shows the released CDK9/cyclin T1 heterodimer dissociates into monomers under stress and that T186 autophosphorylation follows rebinding to Brd4 or the SEC, which makes the phosphorylation state partly a consequence of complex membership rather than purely a cause of it. Zhou K et al. 2022
3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

   CDK9 has mean Chronos -1.536 across 1,208 DepMap lines with 99.3% depleted; within the KMT2A-rearranged panel the myeloid half means -1.826 and the lymphoid half -1.555 against -1.580 in other leukaemia lines, and it ranks 1,136th of 17,728 on index-minus-comparator difference.

   DepMap CRISPR (Chronos) gene effect via the Open Targets Platform GraphQL API, re-queried for this beat on 2026-09-20 and split by lineage inside the KMT2A-rearranged arm; genome-scale ranks read from this repository's analyses/depmap-kmt2a-selective-scan-full/results.json (17,728 genes, 1,208 lines) · Broad Institute DepMap; Open Targets Platform; Oncofusion KG internal scan · Databases 2026
   ▸What this evidence cannot show

   Six myeloid and four lymphoid lines. A difference of -0.100 sits inside the +/-0.15 noise band this project uses at that n, and at 99.3% depletion the measurement is floor-compressed. The DepMap release version is not stated by the API.

   ▸Evidence that argues against this (1)

   • A pan-essential CRISPR score is not a pharmacological claim. In a matched panel, enitociclib killed KMT2A-rearranged lines at lower concentrations than KMT2A-wild-type lines, so partial, reversible, timed inhibition can behave differently from a complete knockout. Sharma R et al. 2025
4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   The CDK9 inhibitor enitociclib synergised with the menin-KMT2A inhibitor MI-463 by Chou-Talalay combination index in all four KMT2A-rearranged cell lines tested, and the combination reduced HOXA9 protein by roughly 75-85% at 24 hours.

   Targeting CDK9 inhibits the growth of KMT2A-rearranged infant leukemia and demonstrates synergy with menin inhibition. · Sharma R, Zhang C, Shah R, Frigault MM, Johnson AJ, Mangum DS, Maese LD, Lacayo NJ, Narendran A · Blood Neoplasia 2025
   PMID 40979068DOI
   ▸What this evidence cannot show

   In vitro only; four KMT2A-rearranged lines (MV4-11, THP-1, KOPN-8, INL2) against two KMT2A-wild-type lines. MI-463 is a tool compound, not a clinical menin inhibitor. Two authors are employees of Vincerx Pharma, the enitociclib sponsor. The individual combination index values appear only in a figure panel this reader could not read.

   ▸Evidence that argues against this (1)

   • The same paper reports dose-dependent loss of viability in healthy-donor PBMCs and CD34+ cells, so the combination's therapeutic window is asserted from a two-line wild-type comparator rather than demonstrated against normal haematopoiesis. Sharma R et al. 2025
5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   AFF1's association with the 7SK snRNP requires its N-terminal cyclin T1-binding sequence: deleting AFF1 residues 1-308 abolishes binding to P-TEFb, HEXIM1 and LARP7 while leaving the ELL2 and ENL contacts intact, and AFF1 1-77 or 38-72 alone is sufficient to join the particle.

   AFF1 is a ubiquitous P-TEFb partner to enable Tat extraction of P-TEFb from 7SK snRNP and formation of SECs for HIV transactivation. · Lu H, Li Z, Xue Y, Schulze-Gahmen U, Johnson JR, Krogan NJ, Alber T, Zhou Q · Proc Natl Acad Sci U S A 2014
   PMID 24367103DOI
   ▸What this evidence cannot show

   Co-immunoprecipitation from HeLa nuclear extracts using overexpressed deletion mutants, in an HIV transactivation study with no leukaemia cell and no KMT2A fusion anywhere in it.

   ▸Evidence that argues against this (1)

   • Overexpressed fragments can join a complex that endogenous full-length protein occupies at different stoichiometry, and a deletion series cannot exclude a weak second contact below co-IP sensitivity. Schulze-Gahmen U et al. 2013
6. E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   CDK9 is detected in all seven KMT2A-rearranged panel lines the Human Protein Atlas carries, median 49.1 nTPM.

   Human Protein Atlas cell-line RNA nTPM, rna_celline.tsv (24,315,372 rows, 1,206 cell lines), downloaded and filtered for this beat on 2026-09-20 · Human Protein Atlas · Database 2026
   ▸What this evidence cannot show

   Bulk poly-A cell-line RNA, not protein and not patient blasts. RS4;11 returns 196.8 nTPM against 36-64 in the other six and is a high outlier for several unrelated genes in the same pull, which looks like a property of that line's library rather than biology.

   ▸Evidence that argues against this (1)

   • HPA cell-line nTPM is poly-A selected and is therefore the wrong instrument for the non-polyadenylated Pol III transcript that regulates this protein: RN7SK returns zero rows in the entire 24.3-million-row table. Human Protein Atlas 2026
7. E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   The ACTIVE pool of CDK9 has a defined subnuclear address: phospho-Thr186 CDK9 is punctate throughout the non-nucleolar nucleoplasm and co-localises with Cyclin T1 almost exclusively within nuclear speckle domains, together with hyperphosphorylated RNA polymerase II; Brd4 and HEXIM1 both contact P-TEFb at or very near speckles, and speckle residence does not require Thr186 phosphorylation or kinase activity.

   T-loop phosphorylated Cdk9 localizes to nuclear speckle domains which may serve as sites of active P-TEFb function and exchange between the Brd4 and 7SK/HEXIM1 regulatory complexes. · Dow EC, Liu H, Rice AP · J Cell Physiol 2010
   PMID 20201073DOI
   ▸What this evidence cannot show

   Not a haematopoietic system and not a KMT2A-rearranged one. Part of the analysis uses transiently expressed kinase-defective mutants, so an overexpression artefact cannot be excluded for that arm, though the phospho-Thr186 staining is of endogenous protein. 'Nuclear speckle' here is an immunofluorescence co-localisation call at deconvolution resolution, not a purified compartment.

   ▸Evidence that argues against this (1)

   • The compartment depends on which CDK9 you mean. Epitope-tagged CDK9 42k localises diffusely in the nucleoplasm while CDK9 55k accumulates in the NUCLEOLUS, and the two isoforms are differentially expressed across primary monocytes, macrophages and lymphocytes with differentially regulated promoters. A CDK9 antibody or a CDK9 inhibitor that does not discriminate the isoforms is acting on two proteins in two compartments, and no CDK9 claim in this graph states which isoform it means. Liu H et al. 2005
8. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   In KMT2A-rearranged infant leukaemia cells, the clinical-stage CDK9 inhibitor enitociclib inhibited growth with an on-target effect, significantly decreased MYC and MCL-1 protein, reduced the growth advantage conferred by the bone marrow microenvironment, synergised with menin inhibitors to decrease HOXA9 protein, and potentiated venetoclax cytotoxicity in relatively venetoclax-resistant KMT2A-rearranged cells.

   Targeting CDK9 inhibits the growth of KMT2A-rearranged infant leukemia and demonstrates synergy with menin inhibition. · Sharma R, Zhang C, Shah R, Frigault MM, Johnson AJ, Mangum DS, Maese LD, Lacayo NJ, Narendran A · Blood Neoplasia 2025
   PMID 40979068DOI
   ▸What this evidence cannot show

   In vitro only - the paper states measurable IN VITRO antitumour activity and proposes clinical exploration; there is no in vivo or patient arm. The readouts are growth, protein level and drug synergy, not a differentiation or identity readout, so this is a dependency result and not a cell-state result. Enitociclib does not separate the 42-kDa and 55-kDa CDK9 isoforms. Read: abstract only.

   ▸Evidence that argues against this (1)

   • CDK9 is pan-essential and floor-compressed, so a growth effect in KMT2A-rearranged cells is expected in almost any cell line: DepMap Chronos mean -1.536 across 1,208 lines with 99.3 percent depleted, and the KMT2A-selectivity scan places it at rank 1,136 of 17,728 with a difference of only -0.100 between the 11 KMT2A-rearranged lines and 43 other leukaemia lines. Without a non-KMT2A comparator arm the paper cannot show that this dependency is anything other than the general one. Oncofusion knowledge graph, analyses/depmap-kmt2a-selective-scan-full; DepMap Chronos via Open Targets; 11 KMT2A-rearranged versus 43 other leukaemia lines, 17,728 genes 2026
9. E9strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   CDK9 expression is cell-type-differential in the haematopoietic system at the isoform level: the 42-kDa and 55-kDa isoforms are produced from separately regulated promoters and are differentially expressed across primary monocytes, macrophages and lymphocytes, with epitope-tagged 42k diffusely nucleoplasmic and 55k accumulating in the nucleolus.

   Differential localization and expression of the Cdk9 42k and 55k isoforms. · Liu H, Herrmann CH · J Cell Physiol 2005
   PMID 15452830DOI
   ▸What this evidence cannot show

   Primary human monocytes, macrophages and lymphocytes plus epitope-tagged transient expression for the localisation half - so the compartment assignment rests on overexpressed tagged protein. Not a leukaemia and not a KMT2A-rearranged system. One gene, one accession: accession-plus-length checking gives zero protection against confusing the two isoforms, which is why this is a registry row. Read: abstract only. Re-fetched from PubMed for this entry rather than copied from the sibling localisation evidence item.

   ▸Evidence that argues against this (1)

   • The cell-type expression difference is measured in normal myeloid and lymphoid cells, and nobody has measured the 42k/55k ratio in a KMT2A-rearranged blast. Since every mechanistic account of this fusion ends at P-TEFb and every available inhibitor hits both isoforms, the quantity that matters for the index disease is the one quantity not measured - and the KMT2A-rearranged infant-leukaemia pharmacology reports no isoform breakdown at all. Sharma R et al. 2025
10. E10strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CDK9 DIRECTS A HISTONE MARK WITHOUT DEPOSITING ONE: CDK9 activity is essential for maintaining global and gene-associated levels of histone H2B monoubiquitination (H2Bub1), and CDK9 knockdown consistently causes inefficient recognition of the correct replication-dependent histone mRNA 3'-end cleavage site with read-through of RNA polymerase II to an alternative downstream polyadenylation signal, so CDK9 integrates transcriptional phosphorylation with chromatin modification.

    CDK9 directs H2B monoubiquitination and controls replication-dependent histone mRNA 3'-end processing. · Pirngruber J, Shchebet A, Schreiber L, et al. · EMBO Rep 2009
    PMID 19575011DOI
    ▸What this evidence cannot show

    Abstract read from the eutils record; full text not read in this tranche. Human somatic cell lines in a replication-dependent histone-gene context, NOT leukaemia and NOT a KMT2A-fusion system - rule 73 applies to any downstream use. The perturbations are CDK9 knockdown and kinase inhibition, which cannot distinguish a direct requirement for CDK9 substrate phosphorylation from a general consequence of shutting down elongation, since H2Bub1 is co-transcriptional and would fall if transcription fell. 'Global and gene-associated' is a bulk plus ChIP statement, not a genome-wide map. Not cited anywhere in this corpus before 2026-09-21, verified with scripts/check-citation-reuse.py.

    ▸Evidence that argues against this (1)

    • The requirement establishes a dependency, not a selective mechanism, and this project has already recorded why that distinction bites for H2Bub. RNF20, the H2B ubiquitin ligase, is required for MLL-rearranged leukaemia - so if a CDK9 inhibitor lowers H2Bub1, it lands on a pathway the leukaemia needs, but this graph's own hyp-h2bub-required-but-not-selective and hyp-h2bub-direction-is-unknown-in-kmt2a hold that neither the direction nor the selectivity of that requirement in KMT2A leukaemia is established. An H2Bub-mediated account of CDK9-inhibitor activity is therefore a hypothesis with a known hole in it, not an explanation. Wang E et al. 2013
11. E11weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    CDK9 is described as guiding a network of chromatin modifications including H2B monoubiquitination (H2Bub1), H3 lysine 4 trimethylation (H3K4me3) and H3K36 trimethylation, in a manner dependent not only on phosphorylation of the RNA polymerase II C-terminal domain but also on other CDK9 targets including SUPT5H, NELF-E and probably the human Rad6 homolog UBE2A, with a proposed model in which CDK9 controls co-transcriptional replication-dependent histone mRNA 3'-end processing in an H2Bub1- and H3K4me3-dependent manner.

    Insights into the function of the human P-TEFb component CDK9 in the regulation of chromatin modifications and co-transcriptional mRNA processing. · Pirngruber J, Shchebet A, Johnsen SA · Cell Cycle 2009
    PMID 19844166DOI
    ▸What this evidence cannot show

    Marked WEAK deliberately, and the reason is the document type: this is a Cell Cycle Extra View that mixes review of the group's previous work with new data ('our previous work and the work presented here'), so it cannot be treated as a primary source for the H3K4me3 and H3K36me3 arms, and the brief's quality bar admits a review only as a map. The UBE2A substrate is 'probably', i.e. not demonstrated in the source. The H3K4me3 and H3K36me3 claims are the ones with no primary citation attached here and should not be repeated without one. Human cells, not leukaemia, not a fusion system. Not cited anywhere in this corpus before 2026-09-21.

    ▸Evidence that argues against this (1)

    • The narrower primary paper from the same group supports only the H2Bub1 arm - 'CDK9 activity is essential for maintaining global and gene-associated levels of H2Bub1' - and the H3K4me3 and H3K36me3 claims appear in the Extra View without primary support cited there. Because H3K4me3 and H3K36me3 are both co-transcriptionally deposited, a fall in either after CDK9 inhibition is the expected consequence of reduced elongation and is not evidence of CDK9 'guiding' those marks specifically. The honest reading is one established mark and two plausible ones. Pirngruber J et al. 2009
12. E12strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    INDEPENDENT SUPPORT FOR THE CDK9 REQUIREMENT, from a different system and a different laboratory's question: H2B monoubiquitination increases during differentiation of human mesenchymal stem cells and lineage-committed precursors, the H2B ubiquitin ligase RNF40 is required for induction of differentiation markers and transcriptional reprogramming, and that function is dependent upon CDK9 and the WAC adaptor protein, which are required for H2B monoubiquitination; RNF40 is also required for resolution of the H3K4me3/H3K27me3 bivalent poised state on lineage-specific genes during the transition to an active chromatin conformation.

    The histone H2B monoubiquitination regulatory pathway is required for differentiation of multipotent stem cells. · Karpiuk O, Najafova Z, Kramer F, et al. · Mol Cell 2012
    PMID 22681891DOI
    ▸What this evidence cannot show

    Abstract read from the eutils record; already cited twice in this corpus (gene-rnf40, hyp-h2bub-direction-is-unknown-in-kmt2a), so it is recorded on this entry for the CDK9/WAC requirement specifically, which is what those entries do not carry. Human mesenchymal stem cells and lineage-committed precursors - NOT leukaemia and NOT a KMT2A-fusion system. The CDK9 and WAC dependence is stated in the abstract as a requirement for H2B monoubiquitination without the experimental detail, so the strength of that specific sub-claim cannot be graded from what was read. The bivalency resolution is an hMSC differentiation result and must not be imported as evidence that bivalent domains exist in KMT2A-rearranged cells - this graph records that as an open gap.

    ▸Evidence that argues against this (1)

    • If CDK9 licenses H2Bub1 and H2Bub1 feeds H3K4 methylation, the naive prediction is that shutting down the elongation kinase lowers H3K4me3 - and the best-calibrated measurement in this disease goes the other way. By ICeChIP-seq in MLL-rearranged cells, depleting H3K79me2 with pinometostat produces PARADOXICAL INCREASES of H3K4me3 alongside loss of H3K27me3. That is a different perturbation, but it demonstrates that the marks in this system do not move together in the direction a linear crosstalk model predicts, so a CDK9-inhibitor prediction built on the crosstalk chain should be treated as a hypothesis to test rather than an expectation. Richter WF et al. 2021

• ← related toMEN1 and CDK9 each carry exactly one genome-wide-significant in-gene common variant and PSIP1 two, so GWAS says almost nothing about on-target liability - yet the revumenib label lists hyperparathyroidism, the organ MEN1 loss famously affects

  near-empty GWAS locus

• ← related toAcross 21 stopped trials of menin, DOT1L, BET, CDK9 and LSD1 inhibitors in acute leukaemia, 12 (57%) were stopped for sponsor or enrolment reasons, and the only two risk-benefit stops are one molecule - failure here has mostly been structural

  CDK9 programmes ended by strategy and enrolment

• ← related toA quarter of this graph's gene entries have a UniProt localization that nobody measured - and the subnuclear vocabulary UniProt does carry avoids the genes we care about most

  P-TEFb catalytic subunit whose UniProt location values carry no evidence code

• ← recruitsH4K16ac - the one mark rule 80 actually rescued in this beat: elevated at MLL-AF9 targets in a supplementary figure, and the reason MLL1's own enzyme turns out to be dispensable

  P-TEFb recruitment falls with promoter H4K16Ac at MLL1 target genes

• part of →P-TEFb (CDK9 + cyclin T1) — the deepest dependency and the clearest therapeutic-index wall in the interactome

  catalytic subunit of the P-TEFb module the graph previously held as a single entry

• binds →CCNT1 / cyclin T1 — the regulatory half of P-TEFb, and the one part of the elongation core with any dependency variance

  obligate cyclin partner; the AFF-family contact is on the cyclin, not on the kinase

• part of →The 7SK snRNP - a stable RNA core with a reversibly docked kinase, which is why the reservoir is a switch and not a sink

  reversibly captured when CDK9 T186 is phosphorylated

• regulates →Plain-English: what transcription elongation is, and why a traffic light explains this leukemia

  phosphorylates the Pol II CTD and the pause factors to license elongation

• related to →CDK9 inhibition in KMT2A-rearranged infant leukemia — the most direct hit on the hijacked machinery, and the narrowest window

  the target dossier this entry supplies the gene-level record for

• ← inhibitsRN7SK / 7SK snRNA - the 332-nucleotide brake on elongation, and the one molecule in this beat that every dataset this project uses is blind to

  sequesters P-TEFb in a low-activity particle; requires HEXIM to inhibit

• ← inhibitsHEXIM1 - the brake pad, the only part of the reservoir anyone has deliberately raised in a KMT2A-rearranged cell, and the result was that the leukaemia grew less

  inhibits P-TEFb through a 7SK-mediated contact on cyclin T1

• ← inhibitsHEXIM2 - the paralogue that holds the bound P-TEFb pool constant when HEXIM1 is removed, which is why single-gene experiments on this axis read flat

  inhibits P-TEFb kinase activity in conjunction with 7SK

• ← bindsCCNT2 / cyclin T2 - expressed in every KMT2A-rearranged line and needed by none of them, so the cyclin T split is real rather than an expression artefact

  the alternative cyclin partner of CDK9; expressed in all seven panel lines

• ← related toCDK12 - a different kinase with a different cyclin, and in KMT2A-rearranged lines it is an anti-dependency

  a transcriptional CDK the CDK9 story does not transfer to: different cyclin, different gene set, anti-selective in KMT2A-rearranged lines

• ← related toCDK13 - rank 171 of 17,728, equally deep in both lineage halves, and essentially unstudied in leukaemia

  the other transcriptional CDK; unlike CDK9 its KMT2A-rearranged difference is lineage-consistent rather than floor-compressed

• ← related toCSNK2A2 - CK2's other catalytic subunit, expressed in every KMT2A-rearranged line and, if anything, slightly favourable to lose

  both are kinases this beat separated by lineage-split dependency; CK2 sits on the LEDGF/taspase1 axis rather than on P-TEFb

• ← regulatesHow P-TEFb gets out of the reservoir: a dephosphorylation, a dissociation, and a reassembly on whichever scaffold is waiting

  T186 phosphorylation is the capture tag; dephosphorylation releases

• ← inhibitsHitting the same axis at two points was predicted to be sub-additive; the one published test found synergy in all four KMT2A-rearranged lines

  enitociclib, CDK9 inhibitor arm of the combination

• ← refutesThe transcriptional CDK with a real KMT2A-rearranged selectivity signal is CDK13, not CDK9 - and nobody has looked

  argues CDK9's depth in this arm is floor compression rather than selectivity

• ← related toTBXT / brachyury - the notochord master regulator, and one of only two of 102 transcription-factor baits whose neighbourhood contained DOT1L

  chordoma's brachyury addiction is suppressed by CDK7/12/13 and CDK9 inhibitors - a disease analogy to P-TEFb dependence, not evidence for the proximity row

• ← testsThree hours after the fusion is degraded to 2.5-7.1% of its own sites, NOT ONE of the eight complex components profiled beside it falls below 68% retention genome-wide - menin 80.0%, DOT1L 68.7%, CDK9 83.6%, AFF4 92.7%, RNA Pol II 99.6%, pS2 98.3%, pS5 86.8% - so nothing in the complex is held on chromatin by the fusion at genome scale

  CDK9 retains 83.6% of its strongest peak calls genome-wide 3 h after fusion degradation, loss relative to the fusion 0.17 - the SEC catalytic subunit stays where it is at genome scale.

• ← related toCDK9 peak widths more than double after fusion degradation, firing the pre-registered re-score trigger

  CDK9 peak and signal readouts in this deposit are both weaker than MEN1 and DOT1L and are not load-bearing

• ← related toAFF4-CDK15: AFF4 is prey in a systematic CMGC-kinase TAP interactome - a kinase adjacent to the complex CDK9 works in, and nobody has followed it up

  a second, orphan CDK on the scaffold whose established kinase is CDK9 - and the null (how many CMGC kinases recover AFF4) is named and unrun

• related to →Nuclear speckles — the compartment everyone assumes KMT2A::AFF1 uses, on evidence that skips AFF1 itself

  the active phospho-T186 CDK9/Cyclin T1 pool is a speckle resident; the speckle record for this disease had no CDK9 evidence before

• related to →The nucleolus — the one compartment with direct published images of a KMT2A fusion inside it

  the 55-kDa CDK9 isoform is nucleolar, so the nucleolar record for this disease has an unexamined P-TEFb component

• expressed in →Monocyte — the graph's most-discussed undefined cell identity, and the confound sitting under its own DepMap haem signal

  42k and 55k CDK9 isoforms differentially expressed across primary monocytes, macrophages and lymphocytes from separately regulated promoters; expression, not dependency

• regulates →The apoptotic priming state — KMT2A-rearranged infant ALL is the one childhood ALL subgroup that is BCL-2-dependent rather than BCL-XL-co-dependent, and BH3 profiling is how that was measured

  enitociclib lowers MCL-1 protein and potentiates venetoclax in relatively venetoclax-resistant KMT2A-rearranged infant leukaemia cells - removing the MCL-1 arm of the priming axis

• ← recruitsCCNT1 / cyclin T1 — the regulatory half of P-TEFb, and the one part of the elongation core with any dependency variance

  cyclin T1 is exclusively nuclear while CDK9 shuttles and is actively exported; enforced cyclin T1 expression drives CDK9 nuclear, so the subunit ratio sets where the kinase is

• ← related toThe binary ladder for AFF1: does the index fusion's partner touch P-TEFb without AFF4, and does it touch AFF4 at all?

  purified CDK9/cyclin T1 as the analyte, with the AFF4 fragment as the same-session positive control

• regulates →H2BK120ub (H2Bub1) — the ubiquitin tag on histone H2B that licenses H3K79 methylation, and the mark this graph was missing

  REQUIRED FOR, NOT A WRITER OF: CDK9 activity is essential for maintaining global and gene-associated H2Bub1 levels, acting through non-histone substrates (Pol II CTD, SPT5/SUPT5H, NELF-E, probably UBE2A) and the RNF20/RNF40-WAC ligase. Human somatic cells, not a fusion system.

• related to →The trans-histone crosstalk H2Bub → DOT1L → H3K79me: modulatory for me1, effectively required for me2/me3

  THE UNJOINED LINK: if CDK9 licenses H2BK120ub and H2BK120ub is the trans-histone input to DOT1L, then a CDK9 inhibitor is an indirect H3K79 agent in KMT2A-rearranged cells. No H2Bub1 or H3K79me2 readout under any CDK9 inhibitor exists in this disease.

• related to →H3K4me3 - the mark KMT2A's own SET domain writes, that the fusion cannot write, that an interface in this graph already binds, and that switches PRC2 off

  WEAK AND SOURCE-LIMITED: CDK9 is described as guiding H3K4me3 alongside H2Bub1 and H3K36me3, but only in an Extra View that mixes review with new data and cites no primary source for the H3K4me3 arm. Because H3K4me3 is co-transcriptional, a fall after CDK9 inhibition is also the expected consequence of reduced elongation.

• related to →RNF40 (BRE1B) — the obligate partner of RNF20, and the one of the pair with a documented link to differentiation and bivalency resolution

  RNF40-dependent H2Bub1 induction during human mesenchymal stem cell differentiation is dependent upon CDK9 and the WAC adaptor, which are required for H2B monoubiquitination - independent support for the CDK9 requirement from a different system.