CCNT2 / 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
also called CCNT2 · cyclin T2 · cyclin-T2 · CycT2 · CycT2a · CycT2b
Cyclin T2 is the alternative partner that can switch on CDK9 in place of cyclin T1. This graph has learned that a paralogue difference only means something if the quiet paralogue is actually being made, so the check was run: cyclin T2 is present in all seven KMT2A-rearranged cell lines tested and the cells still do not need it.
The same thing, for a scientist
CCNT2 (UniProt O60583, 730 aa, HGNC 2q21.3) is the second cyclin partner of CDK9; P-TEFb is defined as CDK9 with cyclin T1 or cyclin T2, and CCNT2 produces the T2a and T2b isoforms [E1]. The question this entry exists to settle is whether the CCNT1/CCNT2 dependency split recorded in `gene-ccnt1` is functional or an artefact of one paralogue being silent. **It is not an artefact.** CCNT2 is detected in all seven KMT2A-rearranged panel lines the Human Protein Atlas carries, median 6.8 nTPM, with no line below 5.3 [E2], while its dependency is near zero (mean Chronos -0.123 across 1,208 lines, rank 5,235 of 17,728) [E3]. The corollary runs the other way too: the same lineage split shows that **CCNT1's dependency is myeloid-driven** - -0.485 in the six KMT2A-rearranged myeloid lines against -0.085 in the four lymphoid ones - so any reading of cyclin T1 as the interesting half of P-TEFb needs that caveat attached [E3].
The control, and what it does and does not prove
| gene | median nTPM (7 KMT2A-r lines) | lines detected | Chronos, all lines | KMT2A-r myeloid | KMT2A-r lymphoid | scan rank |
|---|---|---|---|---|---|---|
| CCNT1 | 16.3 | 7/7 | -0.206 | -0.485 | -0.085 | 561 |
| CCNT2 | 6.8 | 7/7 | -0.123 | -0.169 | -0.003 | 5,235 |
Both cyclins are transcribed in every line. So 'CCNT2 is not required' is a statement about function, not about silence, and the paralogue split passes the control this project applies after an expression artefact overturned a comparable claim 23.
What it does not prove is that CCNT2 is functionally irrelevant. Two proteins that can substitute for each other both score near zero in a single-gene screen. CCNT1's score is only three-fold deeper, and both are shallow. The only clean test is a double knockout, which has not been reported in leukaemia.
A caution about CCNT1's own number
The existing entry gene-ccnt1 treats cyclin T1 as the half of P-TEFb with dependency variance worth pursuing. The lineage split run for this beat qualifies that: CCNT1's KMT2A-rearranged signal is -0.485 in the myeloid half and -0.085 in the lymphoid half, a difference larger than the difference to the comparator arm 3. The KMT2A-rearranged DepMap panel is myeloid-enriched (six of ten lines here), which is the exact confound this project documented for CEBPA and SPI1. This does not refute anything gene-ccnt1 says - that entry's central claim is about a MEN1 correlation, not about KMT2A selectivity - but a cyclin T1 dependency claim built on the pooled arm should be split before it is used.
Literature: almost nothing
Searching (CCNT2[tiab] OR "cyclin T2"[tiab]) AND (leukemia[tiab] OR AML[tiab] OR MLL[tiab] OR KMT2A[tiab]) returns five records on 2026-09-20; title-checking them leaves two miRNA studies in AML (miR-192 and miR-212-5p targeting CCNT2) and three off-topic hits. There is no study of cyclin T2's P-TEFb function in any leukaemia, and none in KMT2A-rearranged disease 4.
| uniprot | O60583 |
|---|---|
| length AA | 730 |
| gene symbol | CCNT2 |
| cytoband | 2q21.3 |
| uniprot verified | O60583 = CCNT2_HUMAN, primary gene name CCNT2, no listed synonyms, 730 aa, cyclin-T2, Homo sapiens; UniProt REST fetched and read back 2026-09-20 |
| paralogue | CCNT1 (O60563, 726 aa) |
| breakpoint side | not applicable - no documented KMT2A contact |
| correction flag for another entry | gene-ccnt1's KMT2A-rearranged dependency is myeloid-driven (-0.485 myeloid vs -0.085 lymphoid); any selectivity reading of CCNT1 should be lineage-split first |
isoforms
[ "cyclin T2a", "cyclin T2b" ]
paralogue split expression control
{
"question": "Is CCNT2's flat dependency an expression artefact?",
"answer": "No - detected in 7/7 KMT2A-rearranged panel lines, median 6.8 nTPM, minimum 5.3",
"source": "Human Protein Atlas rna_celline.tsv, filtered 2026-09-20",
"precedent": "analyses/expression-control-for-paralogue-splits"
}essentiality — the numbers behind the chart
{
"chronos_mean": -0.1228,
"n_lines": 1208,
"chronos_mean_kmt2ar_myeloid": -0.1686,
"chronos_mean_kmt2ar_lymphoid": -0.0032,
"chronos_mean_other_leukemia": -0.0608,
"scan_rank": 5235,
"scan_n_genes": 17728
}expression hpa celline nTPM — the numbers behind the chart
{
"SEM": 6.3,
"RS4;11": 5.3,
"MV4-11": 6.6,
"THP-1": 6.8,
"NOMO-1": 8.2,
"MOLM-13": 7.7,
"OCI-AML-2": 7,
"median": 6.8
}related entries
[ "gene-ccnt1", "gene-cdk9", "protein-ptefb-cdk9-ccnt1" ]
depmap kmt2a selectivity
{
"chronos_mean_kmt2a_rearranged": -0.0933,
"chronos_mean_other_leukaemia": -0.0622,
"difference": -0.0312,
"pan_essentiality_all_lines": -0.1228,
"selectivity_rank": 5235,
"rank_denominator": 17728,
"provenance": "Joined 2026-09-21 from analyses/depmap-kmt2a-selective-scan-full (generated 2026-09-19T10:12:06Z), the project's own genome-wide scan over 11 KMT2A-rearranged and 43 other leukaemia DepMap lines. NOT re-fetched and NOT independent of that scan: every caveat it carries applies here, including that the KMT2A-rearranged arm is a small number of lines drawn from nine individuals, that the rank denominator is 17,728 genes, and that a difference is not a dependency. Added because only 26 of 308 gene entries carried this number, which made any KMT2A-selectivity question answerable for 8% of the corpus.",
"how_to_read": "A negative difference means MORE required in KMT2A-rearranged lines than in other leukaemia lines. A rank near 17,728 is ANTI-selective. Neither is evidence of a fusion-driven dependency on its own: check pan_essentiality_all_lines before calling anything selective, and note that a paralogue-redundant gene can score near zero for reasons that have nothing to do with this disease - see ana-the-paralogue-redundancy-excuse-is-not-tested-and-my-proxy-decided-the-answer."
}- If cyclin T2 can substitute for cyclin T1, a cyclin T1 surface binder is pre-empted by a paralogue that is transcribed in every one of these cells.
- The expression control shows only that CCNT2 mRNA is present; a compensation argument needs protein and a double knockout, neither of which exists.
- Does a CCNT1/CCNT2 double knockout kill KMT2A-rearranged cells more than either alone? Never reported in leukaemia.
- Is cyclin T2-containing P-TEFb captured by the 7SK snRNP on the same terms as cyclin T1-containing P-TEFb? All the structural and biochemical work is on cyclin T1.
- Does the AFF-family cyclin T1-binding sequence bind cyclin T2 at all? If not, cyclin T2 P-TEFb is outside the SEC entirely and outside this disease's mechanism.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
P-TEFb consists of CDK9 with cyclin T, in the forms T1, T2a and T2b, and the tripartite CDK9-cyclin T-AFF1 complex is transferred as a single unit across the P-TEFb network.
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 2014What this evidence cannot show
HeLa co-IP; the paper's focus is AFF1 and Tat, and it does not compare cyclin T1 with cyclin T2 functionally.
Evidence that argues against this (1)
- The founding 7SK papers were written around CDK9/cyclin T1 specifically, and essentially all structural work on the AFF-family contact is with cyclin T1, so whether cyclin T2-containing P-TEFb is captured by 7SK on the same terms is not established. Nguyen VT et al. 2001
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
CCNT2 is detected in all seven KMT2A-rearranged panel lines the Human Protein Atlas carries (SEM 6.3, RS4;11 5.3, MV4-11 6.6, THP-1 6.8, NOMO-1 8.2, MOLM-13 7.7, OCI-AML-2 7.0 nTPM), so the CCNT1/CCNT2 dependency split is not an expression artefact.
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 2026What this evidence cannot show
Bulk poly-A cell-line RNA, not protein. Seven of the project's eleven-line panel; SHI-1 and HB11;19 are not in HPA, and SEMK2 and MOLM-14 are the same patients as SEM and MOLM-13. mRNA presence does not prove the protein is at a level able to substitute for cyclin T1.
Evidence that argues against this (1)
- CCNT2 is the least-expressed member of this beat's gene set in these lines (median 6.8 against CCNT1's 16.3), so 'expressed' and 'expressed enough to compensate' are not the same claim and only the first is supported. Human Protein Atlas 2026
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
CCNT2 has mean Chronos -0.123 across 1,208 lines and ranks 5,235th of 17,728, with the KMT2A-rearranged myeloid half at -0.169 and the lymphoid half at -0.003; CCNT1 in the same query gives -0.485 myeloid against -0.085 lymphoid, so cyclin T1's KMT2A-rearranged dependency is lineage-asymmetric.
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 2026What this evidence cannot show
Six myeloid and four lymphoid lines - too few for the lineage asymmetry to be more than a flag. All values except CCNT1-myeloid are shallow enough that the comparison is between two small numbers.
Evidence that argues against this (1)
- A single-gene CRISPR screen cannot detect a function shared by two paralogues, so a near-zero CCNT2 score is equally consistent with redundancy and with irrelevance; no CCNT1/CCNT2 double knockout has been reported in leukaemia. Lu H et al. 2014
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
PubMed returns five records for (CCNT2[tiab] OR "cyclin T2"[tiab]) AND (leukemia[tiab] OR AML[tiab] OR MLL[tiab] OR KMT2A[tiab]) on 2026-09-20, of which title-checking leaves two miRNA studies in AML and no study of cyclin T2's elongation function in any leukaemia.
HGNC REST records fetched 2026-09-20: CDK13 (HGNC symbol, 7p14.1, previous symbol CDC2L5, aliases CHED/CDC2L/KIAA1791), CDK12 (17q12, previous symbol CRKRS, aliases CRK7/CRKR/KIAA0904), CCNT2 (2q21.3, no aliases), CSNK2A2 (16q21, aliases CSNK2A1 and CK2alpha-prime) · HUGO Gene Nomenclature Committee · Database 2026What this evidence cannot show
A counted absence bounded by its symbol set. HGNC lists no alias or previous symbol for CCNT2, which is unusually clean, but the query would still miss a paper framing the work as 'P-TEFb' without naming the cyclin.
Evidence that argues against this (1)
- This project has documented that complex-level phrasing makes subunits unreachable from their own symbols; a P-TEFb-level paper that discusses both cyclins would not appear in this count. He N et al. 2011
the alternative cyclin partner of CDK9; expressed in all seven panel lines
paralogue split, controlled for expression: both transcribed, neither required
P-TEFb is CDK9 with cyclin T1 or cyclin T2
same proposed annotation for CCNT2, by family inference only - the cited paper does not mention it