The G1 state in KMT2A-rearranged leukaemia — where CDK6 does a job that is not about the cell cycle at all
also called G1 restriction point · CDK6-dependent G1 state · CCND3-CDK6-RB axis · proliferative state
Before a cell commits to dividing it passes a checkpoint in a phase called G1, controlled by a pair of proteins called cyclin D and CDK6. In this leukaemia CDK6 turns out to matter for a different reason: blocking it does not mainly stop cells dividing, it lets them mature — which is the same thing the fusion prevents.
The same thing, for a scientist
This entry defines the cell-cycle state itself, as distinct from the drug class covered by `target-cdk4-6-inhibition-kmt2a-r`. Three facts make the G1 state non-generic in KMT2A-rearranged leukaemia. First, **CDK6 is a direct transcriptional target of the MLL fusion**: CDK6 mRNA is higher in MLL-rearranged than wild-type infant ALL (p < 0.001) and falls when MLL-AF4 or MLL-ENL is knocked down, and knockdown of CDK6 but not the homologous CDK4 inhibits proliferation of SEM [E1]. Second, **the growth inhibition caused by CDK6 depletion is mediated through enhanced myeloid differentiation**, not simply through cycle arrest, which makes CDK6 an effector of the differentiation block rather than a generic proliferation gene [E2]. Third, the D-type cyclin partner is subtype-specific: novel **CCND3 mutations appear in 8.9 percent of paediatric MLL-rearranged AML** and in none of 105 paediatric t(8;21) cases, which instead carry CCND1 and CCND2 mutations [E3]. The state is therefore wired to lineage rather than to proliferation alone, and the graph's own genome-wide scan agrees: CDK6 ranks 13th of 17,728 for selective requirement while CDK4 ranks 16,052nd.
What G1 is, in plain terms
A dividing cell moves through four phases: G1 (grow and decide), S (copy the DNA), G2 (check the copy) and M (divide). The decision to proceed is made at a point late in G1 called the restriction point, after which the cell is committed.
The machinery of that decision is a chain: growth signals raise D-type cyclins (CCND1, CCND2, CCND3), which partner with the kinases CDK4 and CDK6; the resulting complex phosphorylates the retinoblastoma protein RB1, which releases the E2F transcription factors, which turn on the genes needed for S phase. CDK4/6 inhibitors such as palbociclib and abemaciclib break that chain and hold cells in G1.
A cell held in G1 for long enough may slip out of the cycle entirely into G0, which is cell-state-quiescence. That adjacency is the reason this entry and that one have to be read together.
Why CDK6 is not a generic proliferation gene here
The striking result is not that a leukaemia needs a cell-cycle kinase. It is which kinase, and why.
A functional genetic screen found that MLL-AF9-driven AML cells are exceptionally reliant on CDK6 but not on its functional homolog CDK4, that the preferential growth inhibition from CDK6 depletion is mediated through enhanced myeloid differentiation, that CDK6 essentiality extends to alternative MLL fusions and to a mouse MLL-AF9 model, and that it is ascribable to transcriptional activation of CDK6 by mutant MLL. The authors' framing is the one this graph should adopt: CDK6 is a critical effector of MLL fusions that might be targeted to overcome the differentiation block, and cell-cycle regulators can have distinct, non-canonical and non-redundant functions in different contexts 2.
Independently, in infant ALL, CDK6 was shown to be a direct MLL-fusion target with higher expression in MLL-rearranged than wild-type cases, falling on siRNA knockdown of the fusion, and knockdown of CDK6 — but not CDK4 — inhibiting proliferation of the MLL-AF4-positive line SEM, with PD0332991 producing a remarkable G1 arrest and downregulation of pRB1 and EZH2 1.
So CDK6 sits in the same conceptual place as MEIS1 or HOXA9: a gene the fusion turns on, whose product enforces the arrested state. The fact that it happens to be a cell-cycle kinase is close to incidental.
The cyclin partner is subtype-specific, which is a real diagnostic fact
Whole-exome sequencing of 9 and targeted sequencing of 56 paediatric MLL-rearranged AML samples from the Japanese Pediatric Leukemia/Lymphoma Study Group AML-05 study, with 105 paediatric t(8;21) and 30 adult MLL-rearranged samples for comparison, found 2.1 mutations per patient, most often in signalling pathway genes (60.7 percent), then epigenetic regulators (21.4 percent), transcription factors (16.1 percent) and cohesin (8.9 percent).
The cyclin result is the one that matters here. Novel CCND3 mutations were found in 5 of 56 paediatric MLL-rearranged AML (8.9 percent) and 2 of 30 adults (3.3 percent), while paediatric t(8;21) AML carried recurrent CCND1 (2.9 percent) and CCND2 (7.6 percent) mutations and NO CCND3 mutations — a subtype-specific D-type cyclin pattern. Treating MLL-rearranged AML lines with abemaciclib or palbociclib blocked G1-to-S progression and impaired proliferation. Separately, MLL-MLLT3 cases with coexisting mutations had significantly reduced relapse-free and overall survival 3.
That CCND3 is the mutated cyclin in this subtype and not in another, from the same laboratory pipeline, is about as clean a lineage-of-evidence as this kind of claim gets.
The awkward interaction with quiescence, stated as a warning
A CDK4/6 inhibitor holds cells in G1. Cells held in G1 can drop into G0. cell-state-quiescence records that the G0 compartment is where residual disease survives and that differentiating agents act least on transcriptionally disengaged cells.
That gives a concrete and unwelcome prediction for the combination the two entries jointly suggest: a CDK4/6 inhibitor plus a menin inhibitor may be antagonistic, because the first pushes cells toward the state in which the second works least. The counter-argument is equally concrete and comes from the same literature: CDK6 depletion works through differentiation, in the same direction as menin inhibition, so the two might be additive after all.
Neither prediction has been tested. Both are cheap to test. This graph should record that it does not know, rather than assume additivity, because the assumption is the kind that gets into a combination trial design without ever being examined.
The clinical record so far is thin and unencouraging: the one trial restricted to MLL-rearranged acute leukaemia, AMLSG 23-14 / Palbo-AL-1 (NCT02310243), completed in 2019 and has never reported. See target-cdk4-6-inhibition-kmt2a-r and trial-palbo-al-1-amlsg-23-14.
| ontology | GO:0000082 (G1/S transition of mitotic cell cycle) |
|---|---|
| state definition | The G1 phase and its restriction point, governed in this disease by the CCND3-CDK6-RB1-E2F chain, in which CDK6 is a direct transcriptional target of the MLL fusion and acts on differentiation as well as on cycle progression |
| mechanism note | CDK6 depletion inhibits growth through enhanced myeloid differentiation, placing it with HOXA9 and MEIS1 as an effector of the differentiation block rather than as a generic proliferation gene |
| interaction with quiescence | a CDK4/6 inhibitor drives cells into G1 and potentially into G0, which cell-state-quiescence identifies as the compartment least responsive to differentiating agents; whether CDK4/6 plus menin inhibition is additive or antagonistic is untested |
markers
[ "hypophosphorylated RB1", "cyclin D3 (CCND3) protein", "CDK6 protein", "2N DNA content with elevated RNA content (distinguishes G1 from G0)" ]
axis components
[ "CCND3", "CDK6", "RB1", "E2F", "CDKN2A/p16" ]
cdk6 vs cdk4
{
"depmap_rank_cdk6": 13,
"depmap_rank_cdk4": 16052,
"genes_ranked": 17728,
"note": "the class name 'CDK4/6' averages two genes at opposite ends of the genome for this disease; CDK6 sits on the pan-essentiality boundary at -0.632 and should be read with that caveat"
}ccnd3 mutation frequency — the numbers behind the chart
{
"paediatric_kmt2a_r_aml_pct": 8.9,
"adult_kmt2a_r_aml_pct": 3.3,
"paediatric_t821_aml_pct": 0
}atlas refs — the numbers behind the chart
[
{
"label": "CDK6 as a direct MLL-fusion target in infant ALL",
"pmid": "24736461",
"title": "MLL fusion-driven activation of CDK6 potentiates proliferation in MLL-rearranged infant ALL.",
"authors": "van der Linden MH, Willekes M, van Roon E, et al.",
"venue": "Cell Cycle",
"year": 2014,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"label": "CDK6 requirement acting through myeloid differentiation",
"pmid": "24764564",
"title": "Requirement for CDK6 in MLL-rearranged acute myeloid leukemia.",
"authors": "Placke T, Faber K, Nonami A, et al.",
"venue": "Blood",
"year": 2014,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"label": "Subtype-specific CCND3 mutations in MLL-rearranged AML",
"pmid": "30381403",
"title": "Recurrent CCND3 mutations in MLL-rearranged acute myeloid leukemia.",
"authors": "Matsuo H, Yoshida K, Fukumura K, et al.",
"venue": "Blood Adv",
"year": 2018,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
}
]- A CDK4/6 inhibitor holds cells in G1 and may deposit them in G0, which is the compartment that survives therapy; a drug that reduces disease burden while enlarging the dormant fraction would look like a response and behave like a relapse.
- Because the CDK6 effect runs through differentiation rather than through arrest, a cell that finds another route out of the differentiation block will be unaffected by CDK6 inhibition even with the kinase fully inhibited.
- The one trial restricted to MLL-rearranged acute leukaemia completed in 2019 and never reported, so the class's actual performance in this disease is unknown and the silence is itself information.
- CDK6 sits on the pan-essentiality boundary in this project's own scan, so a therapeutic window between leukaemic and normal proliferating cells may be narrow.
- Is CDK4/6 inhibition additive or antagonistic with menin inhibition? One mechanism (differentiation) predicts additive, the other (driving cells toward G0) predicts antagonistic, and nobody has run the combination with a cell-cycle readout.
- Can the differentiation effect of CDK6 loss be separated from its kinase activity? A kinase-dead CDK6 knock-in would settle whether a kinase inhibitor can reproduce the genetic phenotype, and the B-lineage CCND3 result suggests it may not.
- What fraction of a KMT2A-rearranged blast population is actually in cycle at diagnosis and at MRD? The graph assumes a proliferative bulk and a quiescent residuum but has no measured number for either.
- Why is CCND3 the mutated D-type cyclin in KMT2A-rearranged AML and CCND1/CCND2 in t(8;21)? A subtype-specific cyclin preference implies a subtype-specific upstream signal that has not been identified.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
CDK6 is a direct target of the MLL fusion protein in infant ALL: CDK6 mRNA is significantly higher in MLL-rearranged than MLL wild-type infant ALL (p < 0.001), decreasing MLL-AF4 or MLL-ENL fusion mRNA by siRNA downregulates CDK6, knockdown of CDK6 but not the highly homologous CDK4 significantly inhibits proliferation of the MLL-AF4-positive line SEM, and MLL-rearranged lines are sensitive in vitro to the CDK4/6 inhibitor PD0332991 with a remarkable G1 arrest and downregulation of pRB1 and EZH2.
MLL fusion-driven activation of CDK6 potentiates proliferation in MLL-rearranged infant ALL · van der Linden MH, Willekes M, van Roon E, Seslija L, Schneider P, Pieters R, Stam RW · Cell Cycle 2014What this evidence cannot show
Cell-line and primary-sample expression work; the functional validation is essentially in one line (SEM), and 'direct target' is inferred from fusion knockdown rather than from occupancy at the CDK6 locus in this study. Read: abstract only.
Evidence that argues against this (1)
- This project's own genome-wide DepMap scan places CDK6 at rank 13 of 17,728 for selective requirement in KMT2A-rearranged lines but notes it sits on the pan-essentiality boundary at -0.632, the region where genes are more depleted in rearranged lines because that panel is more sensitive rather than because of the fusion. The selectivity of CDK6 is therefore weaker than its rank suggests, and the entry that reported it says so explicitly. Oncofusion Knowledge Graph project analysis 2026
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
A functional genetic screen showed AML cells driven by MLL-AF9 are exceptionally reliant on CDK6 but not its functional homolog CDK4; the preferential growth inhibition induced by CDK6 depletion is mediated through enhanced myeloid differentiation; CDK6 essentiality extends to cells harbouring alternate MLL fusions and to a mouse MLL-AF9 model and is ascribable to transcriptional activation of CDK6 by mutant MLL; and the context-dependent effects of lowering CDK6 are closely phenocopied by a small-molecule CDK6 inhibitor in clinical development.
Requirement for CDK6 in MLL-rearranged acute myeloid leukemia · Placke T, Faber K, Nonami A, Putwain SL, Salih HR, Heidel FH, et al., Frohling S · Blood 2014What this evidence cannot show
Human AML cell lines plus a mouse MLL-AF9 model; 'mediated through enhanced myeloid differentiation' is inferred from phenotype after CDK6 depletion rather than from a separation-of-function CDK6 allele, so the differentiation effect and the kinase effect are not formally separated. Read: abstract only.
Evidence that argues against this (1)
- In B-lineage disease the same axis works through a kinase-independent route, which undercuts the assumption that a CDK4/6 kinase inhibitor reproduces the genetic phenotype. CCND3 is essential for proliferation and survival of B-ALL independent of mutational background, and its anti-apoptotic effect is INDEPENDENT of the kinase activity of the CCND3-CDK4/6 complex, leading those authors to conclude that downregulating CCND3 might be superior to inhibiting CDK4/6 kinase activity. A degrader and an inhibitor of this axis are therefore not interchangeable. Read: abstract only. Ketzer F et al. 2022
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Whole-exome sequencing of 9 and targeted sequencing of 56 paediatric MLL-rearranged AML samples from the JPLSG AML-05 study, with 105 paediatric t(8;21) and 30 adult MLL-rearranged samples, identified 2.1 mutations per patient with signalling-pathway genes most frequent (60.7 percent); novel CCND3 mutations were found in 5 paediatric (8.9 percent) and 2 adult (3.3 percent) MLL-rearranged AML patients, while paediatric t(8;21) AML carried recurrent CCND1 (2.9 percent) and CCND2 (7.6 percent) mutations and no CCND3 mutations; treatment of MLL-rearranged AML lines with abemaciclib and palbociclib blocked G1-to-S progression and impaired proliferation.
Recurrent CCND3 mutations in MLL-rearranged acute myeloid leukemia · Matsuo H, Yoshida K, Fukumura K, Nakatani K, Noguchi Y, Takasaki S, et al., Adachi S · Blood Advances 2018What this evidence cannot show
Five CCND3-mutant cases out of 56 paediatric samples, from a single national study group, so the 8.9 percent estimate has wide confidence limits; the t(8;21) comparison group is larger (n=105) than the index group, which affects the power to detect rare mutations asymmetrically. Read: abstract only.
Evidence that argues against this (1)
- A mutation frequency of 8.9 percent means more than 90 percent of KMT2A-rearranged AML has wild-type CCND3, so the cyclin-mutation finding cannot be the general explanation for the subtype's CDK6 dependence, which is reported across MLL-rearranged models irrespective of CCND3 status and is ascribed instead to direct transcriptional activation of CDK6 by the fusion. The mutation is a cooperating lesion in a minority, not the mechanism. Read: abstract only. Placke T et al. 2014
the same cycle, two different descriptions: a CDK6-dependent G1 restriction point that is pharmacologically addressable, and a population that cannot be held at any phase chemically
- related to →CDK4/6 inhibition in KMT2A-rearranged leukemia — the class was already tried, and the trial never reported
the cell state underneath that drug-class entry
G1 arrest is adjacent to G0, so CDK4/6 inhibition may enlarge the compartment that survives therapy
CDK6 depletion inhibits growth through enhanced myeloid differentiation, making it an effector of the block
CCND3 is the D-type cyclin partner mutated specifically in KMT2A-rearranged AML
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
CDK6 is a direct transcriptional target of MLL-AF4 and MLL-ENL
predicts CDK4/6 inhibition is the wrong combination partner
G1 restriction-point kinase underlying that cell state
USP2 is the specific deubiquitinase for cyclin D1 and stabilises it; USP2 knock-down destabilises cyclin D1 and arrests growth ONLY in cyclin D1-addicted cells (PMID 19917254)
Taspase1 loss down-regulates Cyclins E, A and B, up-regulates p16Ink4a and reduces H3K4me3 at Cyclin E1/E2; MEFs and MLL(nc/nc;2nc/nc) MEFs are both impaired in proliferation (murine; PMID 16951254)
- ← related toHCFC1 (HCF-1) — a founding member of the wild-type KMT2A complex that the fusion cannot bind
IMPERFECT NODE MATCH: HCF-1 is required for progression through G1 as a general property of dividing cells (PMID 10196288); the node is CDK6-framed and the requirement is not, so the edge records the general cell-cycle role rather than an equivalence
the G1 state cyclin D3 maintains is a checkpoint state - only heavy-chain-rearranged pre-B cells keep the protein stable
KAT2B localises to centrosomes in late G1 and around the G1/S transition, and its Rb acetylation controls cell-cycle exit during differentiation
SKP2's compartment is cell-cycle- AND signalling-conditional: it degrades MLL in S phase, and Akt1 phosphorylation of Ser72 inside the 67-73 NLS exports it to the cytoplasm (PMID 19270695) (Q91 2026-09-22)
The compartment question that matters is CDK6's, not CDK4's. CDK4 is LESS required in leukaemia lines than elsewhere (Chronos -0.446 against -0.737) with a POSITIVE KMT2A-r selectivity difference at rank 16,052/17,728, and the one MLL-AF9 result uses palbociclib, which is pan-CDK4/6 and moves both kinases together (PMID 36003783). A staged CDK4 imaging experiment would measure a gene this disease does not need (Q95 2026-09-22)