Oncofusion

start / graph / target-cdk4-6-inhibition-kmt2a-r

Targetmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q-scan-hits

CDK4/6 inhibition in KMT2A-rearranged leukemia — the class was already tried, and the trial never reported

also called palbociclib in MLL-rearranged leukemia · CDK6 targeting · abemaciclib · ribociclib

Drugs that block the cell-cycle enzymes CDK4 and CDK6 are already approved in breast cancer, and laboratory work says KMT2A-fusion leukemia cells depend on CDK6 specifically. A trial of one of these drugs, restricted to patients with this exact leukemia, finished in 2019 and has never published a result.

The same thing, for a scientist

The pre-clinical case for CDK4/6 inhibition in KMT2A-rearranged leukemia is old, consistent and mechanistically specific: CDK6 but not CDK4 is required, the fusion drives CDK6 transcription, and both infant ALL and MLL-AF9 AML models show it [E1][E2]. The class is deliverable in children: the COG AINV18P1 phase 1 established a pediatric recommended phase 2 dose for palbociclib with re-induction chemotherapy [E4]. Yet the one trial ever run in exactly this indication — AMLSG 23-14 / Palbo-AL-1, single-agent palbociclib in MLL-rearranged acute leukemia, 50 adults planned — closed in 2019 with no posted results and no traceable publication [E3]. This is the decisive gap. Before this project spends anything on CDK4/6, that trial's outcome should be obtained, because it is the closest thing to an answer that exists and it is sitting in a German registry record.

Go deeper
Why this entry exists

It exists because an unbiased DepMap scan run by this project put CCND3 at rank 4 of 4,589 for KMT2A-rearranged selectivity, with no prior in this graph 5. Following that hit into the literature produced not a speculative target but a drug class with two decades of pre-clinical support in this disease and a completed, unreported trial.

That is a different situation from a new target. The question is not 'should someone test CDK4/6 inhibitors in KMT2A-rearranged leukemia'. Someone did. The question is what happened.

The pre-clinical case, and its exact shape

The consistent finding across two independent groups is a CDK6-versus-CDK4 discrimination, which is a stronger and more falsifiable claim than a generic cell-cycle dependency.

  • In MLL-rearranged infant ALL, CDK6 is higher than in wild-type ALL (P<0.001), fusion knockdown lowers it, CDK6 knockdown inhibits growth of the KMT2A::AFF1 line SEM while CDK4 knockdown has virtually no effect, and MLL-rearranged lines undergo G1 arrest on PD0332991 (palbociclib) 1.
  • In MLL-AF9 AML, cells are 'exceptionally reliant' on CDK6 but not CDK4; the effect is mediated by enhanced myeloid differentiation; it extends to other MLL fusions and to a mouse model; CDK6 is transcriptionally activated by mutant MLL; and a small-molecule CDK6 inhibitor phenocopies depletion 2.
  • In KMT2A-rearranged AML lines, abemaciclib and palbociclib block G1-to-S and impair proliferation 6.

Both papers put the target downstream of the fusion. CDK6 is an output of the KMT2A-fusion transcriptional program, not a component of the fusion machinery. That places this target in the same class as HOXA9/MEIS1 — real, fusion-driven, and exposed to whatever route lets a cell stop needing the fusion's outputs.

The clinical record, which is thinner than the pre-clinical record suggests
TrialPopulationDesignOutcome
NCT02310243 (AMLSG 23-14 / Palbo-AL-1)Adults, acute leukemia with MLL rearrangementPhase Ib/IIa, single-agent palbociclib, 50 planned, Simon two-stage, 24 German sitesCompletion recorded July 2019; registry status 'unknown'; no results posted 3
NCT03792256 (COG AINV18P1)Children/young adults, R/R B- and T-ALL/LL, unselected for fusionPhase 1, palbociclib + four-drug re-induction12 patients; RP2D 50 mg/m2/day x 21 d; 1 haematologic DLT; 5 CRs 4
NCT03515200 (St. Jude)Children, R/R ALLPhase 1, palbociclib + chemotherapyTerminated, 12 enrolled 7
NCT04996160 (Stanford, RELPALL2)Children, R/R ALLPhase 1Completed with 2 patients 7

The single-agent, fusion-selected trial is the only one that tests the hypothesis this graph cares about, and it is the one with nothing to read. The pediatric trials establish that the class can be given to children — a genuine and non-trivial result, because dose and tolerability in children are the usual barrier — but none of them was selected for KMT2A rearrangement and none can attribute a response to palbociclib.

What would have to be true for this to be worth pursuing

Three conditions, in order of how cheaply they can be checked.

  1. AMLSG 23-14 was not a clear negative. Cost: one email to Ulm, or a conference-abstract search this cycle did not run. If it was a clear single-agent negative, the correct conclusion is that this class does not work alone here and the entry becomes a record of a closed question.
  2. The dependency survives lineage conditioning. The CCND3 signal that generated this entry comes from 11 KMT2A-rearranged lines versus 43 mixed leukemia lines, and CCND3 is separately reported essential in B-ALL regardless of driver 8. This project already has the analysis that would settle it.
  3. There is a combination worth testing. The mechanistic pairing is a menin inhibitor, which pushes toward differentiation, with a CDK4/6 inhibitor, which is itself reported to act through differentiation in MLL-AF9 AML 2. Two agents whose claimed mechanism is the same differentiation program are as likely to be redundant as synergistic, and that is an argument for testing it rather than assuming it.
Correction 2026-09-20: the redundancy question is answered, and the answer is synergy

This entry asked whether menin inhibition changes CDK6 expression, called it a two-week experiment with existing reagents, and judged the menin + CDK4/6 combination 'as likely to be redundant as synergistic'. Both are settled in print and this entry did not cite the paper.

PMID 35017466 (Fiskus et al, Blood Cancer Journal 2022): menin-inhibitor treatment attenuates BCL2 and CDK6 levels, and co-treatment with SNDX-50469 plus abemaciclib induces synergistic lethality in cell lines and patient-derived AML cells with MLL1-rearrangement or mutant NPM1.

So CDK6 does fall under menin inhibition AND the combination is still synergistic - which is the outcome this entry listed as the less likely of the two.

Three caveats that keep this short of settling the clinical question. The system is AML, not infant B-lineage KMT2A::AFF1. The paper's in vivo arm is the venetoclax combination, not the abemaciclib one, so the CDK4/6 result is in vitro and ex vivo only. And abemaciclib inhibits both CDK4 and CDK6, so the synergy cannot be attributed to CDK6 alone - which matters because the lineage-split data say CDK4 and CDK6 run in opposite directions in leukaemia.

Facts and figures
target genegene-ccnd3
effector kinaseCDK6
modalitysmall molecule ATP-competitive kinase inhibitor
approval statusCDK4/6 inhibitors are approved in breast cancer; none is approved in any leukemia
fusion relationshipdownstream output — the KMT2A fusion transcriptionally activates CDK6; the target is not part of the fusion complex
pediatric rp2dpalbociclib 50 mg/m2/day orally for 21 days with four-drug re-induction (COG AINV18P1)
decisive missing factthe outcome of NCT02310243, the only trial ever run in MLL-rearranged acute leukemia with this drug class
entered graph viaana-depmap-kmt2a-selective-scan, CCND3 rank 4 of 4,589
agents
[
  "palbociclib",
  "abemaciclib",
  "ribociclib"
]
trials
[
  "NCT02310243",
  "NCT03792256",
  "NCT03515200",
  "NCT04996160",
  "NCT03740334"
]
depmap selectivity
{
  "CDK6": {
    "rank": 13,
    "of": 17728,
    "difference": -0.5266
  },
  "CDK4": {
    "rank": 16052,
    "of": 17728,
    "difference": 0.0716
  },
  "reference": {
    "MEN1": 1,
    "KMT2A": 11
  },
  "note": "the class name averages two genes at opposite ends of the genome for this disease",
  "source": "analyses/depmap-kmt2a-selective-scan-full/results.json (all_results)"
}
How this could bite us back
  • Myelosuppression. The dose-limiting toxicity of this class is neutropenia, and the one DLT in the pediatric COG trial was haematologic. Adding it to induction in a marrow-failure disease costs something real.
  • Growth plate. Chondrocyte proliferation at the growth plate is CDK4/6-driven and the patients here are infants. This graph already carries a growth-plate risk entry for a different agent. No juvenile-animal palbociclib data were located this cycle, so the risk is unquantified rather than excluded — and an unquantified risk in infants is not a small caveat.
  • Redundancy with the agents this project already favours. If CDK6 inhibition acts through myeloid differentiation, and menin inhibition acts through differentiation, the combination may buy nothing while doubling the toxicity.
  • Cytostatic, not cytotoxic. A G1 arrest leaves cells alive. In a disease whose hazard is front-loaded and whose relapses come fast, a drug that stops division without killing may simply select for the clone that escapes it.
  • The target is a fusion output. Any escape that reduces dependence on the fusion program — including the lineage switch this graph documents as the characteristic escape from CD19-directed therapy — takes the dependency with it.
Still open
  1. What did AMLSG 23-14 / Palbo-AL-1 show? This is the cheapest high-value question in the entry and it is answerable by contacting the Ulm group, whose senior investigators are the authors of the MLL-AF9 CDK6 paper.
  2. Was any AMLSG 23-14 result presented at ASH or EHA without full publication? Conference abstracts were not searched in this cycle and should be.
  3. ANSWERED 2026-09-20, PMID 35017466. Menin-inhibitor treatment ATTENUATES CDK6 levels in MLL1-rearranged and NPM1c AML, and SNDX-50469 plus abemaciclib induces SYNERGISTIC lethality in cell lines and patient-derived AML cells. So CDK6 does fall under menin inhibition and the combination is synergistic rather than redundant. Caveats: AML rather than infant B-lineage KMT2A::AFF1; the paper's in vivo arm is the venetoclax combination, not the abemaciclib one; and abemaciclib is a dual CDK4/6 inhibitor, so the synergy cannot be attributed to CDK6 alone.
  4. Is there a CDK6 dependency in KMT2A::AFF1 infant-derived cells specifically, as opposed to the SEM line, which is from a five-year-old and MLL-AF9 mouse models that carry the published evidence?
Every source behind this entry (9)
  1. 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: higher CDK6 mRNA in MLL-rearranged than wild-type infant ALL (P<0.001), downregulation of CDK6 on siRNA knockdown of MLL-AF4 or MLL-ENL, growth inhibition of the MLL-AF4-positive line SEM on CDK6 but not CDK4 knockdown, and in vitro sensitivity of MLL-rearranged lines to the CDK4/6 inhibitor PD0332991 with 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 13:834-844 2014
    PMID 24736461DOI
    What this evidence cannot show

    Read as abstract only. Cell lines and patient expression; no in vivo or clinical readout. 'Direct target' is the authors' wording; nothing in the abstract demonstrates fusion occupancy at the CDK6 locus.

    Evidence that argues against this (2)
    • The trial that tested this drug in this population has posted no results, twelve years after the pre-clinical case was made. University of Ulm; Pfizer; National Center for Tumor Diseases et al. 2015
    • THE COMBINATION QUESTION IS ANSWERED AND THIS ENTRY GUESSED THE WRONG WAY. It judged menin + CDK4/6 'as likely to be redundant as synergistic'. Menin inhibition attenuates CDK6 levels and the SNDX-50469 plus abemaciclib combination is nonetheless synergistic in MLL1-rearranged and NPM1c AML cell lines and patient-derived cells. In vitro and ex vivo only; abemaciclib is dual CDK4/6. Fiskus W et al. 2022
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    MLL-AF9-driven AML cells are exceptionally reliant on CDK6 but not CDK4; the growth inhibition from CDK6 depletion is mediated through enhanced myeloid differentiation; CDK6 essentiality extends to alternative MLL fusions and to a mouse MLL-AF9 model; CDK6 is transcriptionally activated by mutant MLL; and the 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, Kramer A, Root DE, Barbie DA, Krivtsov AV, Armstrong SA, Hahn WC, Huntly BJ, Sykes SM, Milsom MD, Scholl C, Frohling S · Blood 124:13-23 2014
    PMID 24764564DOI
    What this evidence cannot show

    Read as abstract only. AML and an MLL-AF9 model, not the KMT2A::AFF1 infant biology. The senior authors of this paper are at Ulm and Heidelberg, the same institutions that sponsored NCT02310243, which is presumably how the trial came to exist and is a reason to ask them directly what it showed.

    Evidence that argues against this (1)
    • A differentiation mechanism is claimed for menin inhibition, DOT1L inhibition and LSD1 inhibition in this same disease. If CDK6 inhibition works through the same final common path, it is a candidate for redundancy with the agents this project already favours rather than an orthogonal addition. Maiques-Diaz A et al. 2018
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry

    AMLSG 23-14 / Palbo-AL-1 (NCT02310243) was a phase Ib/IIa open-label study of single-agent palbociclib in adults with acute leukemia carrying an MLL rearrangement, planning 50 patients on a Simon two-stage design across 24 German sites, with maximum tolerated dose, adverse events and overall response rate as primary outcomes and CDK6 target inhibition as a secondary outcome; its completion date is recorded as July 2019, its status is 'unknown', and no results are posted.

    Phase Ib/IIa Study of Palbociclib in MLL-rearranged Acute Leukemias AMLSG 23-14/Palbo-AL-1 · University of Ulm; Pfizer; National Center for Tumor Diseases, Heidelberg · ClinicalTrials.gov 2015
    NCT02310243
    What this evidence cannot show

    Full registry record read via the ClinicalTrials.gov API on 2026-09-18; has_results false. Adults only (>= 18 years). No publication was found in PubMed searches run this cycle, but conference abstracts were not searched and absence of a publication is not evidence of a negative trial.

    Evidence that argues against this (1)
    • Accrual failure is at least as likely an explanation as a negative result: 24 sites were opened for a subtype that forms a small minority of adult acute leukemia, and the design required 21 patients before the first interim decision. University of Ulm 2015
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The Children's Oncology Group AINV18P1 phase 1 trial of palbociclib with standard four-drug re-induction chemotherapy in 12 heavily pretreated children and young adults with relapsed/refractory lymphoblastic leukemia and lymphoma reported one dose-limiting haematologic toxicity at the starting dose, a recommended phase 2 dose of 50 mg/m2/day for 21 days, grade 3/4 non-haematologic toxicity comparable to the chemotherapy platform alone, and five complete responses.

    Palbociclib in combination with chemotherapy in pediatric and young adult patients with relapsed/refractory acute lymphoblastic leukemia and lymphoma: A Children's Oncology Group study (AINV18P1) · Raetz EA, Teachey DT, Minard C, Liu X, Norris RE, Denic KZ, Reid J, Evensen NA, Gore L, Fox E, Loh ML, Weigel BJ, Carroll WL · Pediatric Blood & Cancer 70:e30609 2023
    PMID 37553297DOI
    What this evidence cannot show

    Read as abstract only. n=12, unselected for KMT2A rearrangement, combination design, no infants reported. Establishes a pediatric dose and tolerability, not efficacy, and certainly not fusion-specific efficacy.

    Evidence that argues against this (1)
    • Two other pediatric palbociclib ALL studies failed to deliver: one terminated at 12 patients, one completed with 2. Gruber TA et al. 2021
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    CCND3 ranks 4 of 4,589 protein-coding genes by KMT2A-rearranged selectivity in this project's unbiased DepMap scan (difference -0.623) and is not pan-essential (-0.348 across 1,208 lines), and had no entry anywhere in this graph before this cycle.

    DepMap CRISPR (Chronos) gene effect via the Open Targets Platform GraphQL API, 4,589 genes split by Cellosaurus KMT2A gene-fusion annotation; analyses/depmap-kmt2a-selective-scan/ · Broad Institute DepMap; Cellosaurus (SIB); Open Targets Platform; analysis run by this project · Databases and derived analysis 2026
    What this evidence cannot show

    This project's own unreviewed computation. 11 lines, none infant-derived, 24% of the protein-coding genome, and the comparison is not conditioned on lineage. It is a hypothesis generator; the pre-clinical literature above, which is independent of it, is what carries this entry.

    Evidence that argues against this (1)
  6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Treatment of MLL-rearranged AML cell lines with the CDK4/6 inhibitors abemaciclib and palbociclib blocked G1 to S phase cell-cycle progression and impaired proliferation, in the same study that identified recurrent novel CCND3 mutations in 8.9% of pediatric MLL-rearranged AML.

    Recurrent CCND3 mutations in MLL-rearranged acute myeloid leukemia · Matsuo H, Yoshida K, Fukumura K, Nakatani K, Noguchi Y, Takasaki S, et al. · Blood Advances 2:2879-2889 2018
    PMID 30381403DOI
    What this evidence cannot show

    Read as abstract and key points only; the body is not open through PMC or Europe PMC. The cell lines are not named in the abstract and no IC50 or in vivo data were read. A G1 arrest in a cell line is the weakest form of support this class could have, and it is 2018 evidence for a drug that had already entered a trial in this indication in 2015.

    Evidence that argues against this (1)
    • CCND3 upregulation is itself associated with palbociclib resistance, and the CCND3 anti-apoptotic function is reported to be independent of CDK4/6 kinase activity — so the mutation finding and the inhibitor finding may not be about the same thing. Ketzer F et al. 2022
  7. E7established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry

    A St. Jude phase 1 study of palbociclib with chemotherapy in pediatric relapsed/refractory ALL (NCT03515200) is recorded as terminated with 12 participants, and a Stanford phase 1 study of the same combination (NCT04996160, RELPALL2) is recorded as completed with 2 participants.

    NCT03515200
    What this evidence cannot show

    Registry summary fields read via the ClinicalTrials.gov API on 2026-09-18. The reason for termination is not given in the fields read, and termination in a relapsed pediatric population very often reflects accrual rather than toxicity or futility. Neither study was selected for KMT2A rearrangement.

    Evidence that argues against this (1)
    • The COG study of essentially the same combination completed its dose finding and reported responses, so the class is not undeliverable in children; small single-centre studies in relapsed pediatric ALL routinely under-accrue. Raetz EA et al. 2023
  8. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CCND3 is essential for proliferation and survival of B-ALL independent of the mutational background, and its anti-apoptotic effect is independent of the kinase activity of the CCND3-CDK4/6 complex, leading the authors to conclude that downregulating CCND3 might be superior to inhibiting CDK4/6 kinase activity.

    CCND3 is indispensable for the maintenance of B-cell acute lymphoblastic leukemia · Ketzer F, Abdelrasoul H, Vogel M, Marienfeld R, Muschen M, Jumaa H, Wirth T, Ushmorov A · Oncogenesis 11:1 2022
    PMID 35013097DOI
    What this evidence cannot show

    Read as abstract only. If the kinase-independence claim holds, it severs the inference from a CRISPR knockout of CCND3 to a CDK4/6 inhibitor, which is the inference this whole entry depends on. It is one paper and it has not been independently replicated in anything read here.

    Evidence that argues against this (1)
    • Against a pure kinase-independent reading: two independent groups show that small-molecule CDK4/6 or CDK6 inhibition phenocopies CDK6 depletion in MLL-rearranged cells, which is an argument that the kinase arm does carry real weight in this specific context. Placke T et al. 2014
  9. E_cdk4_vs_cdk6_rankstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    The CDK6-specific dependency this entry describes from the pre-clinical literature is reproduced, unbiased and genome-wide, and the gap between the two kinases is enormous. In a scan of 17,728 genes for selective requirement in KMT2A-rearranged versus other leukaemia lines, **CDK6 ranks 13th** (difference -0.527) while **CDK4 ranks 16,052nd** (difference +0.072, i.e. marginally LESS required in the rearranged lines). For scale, MEN1 ranks 1 and KMT2A 11. The class name 'CDK4/6' averages two genes at opposite ends of the genome for this disease.

    Genome-wide KMT2A-rearranged selectivity scan, 17,728 genes (DepMap Chronos via Open Targets) · Broad Institute DepMap / Open Targets; analyses/depmap-kmt2a-selective-scan-full 2026
    What this evidence cannot show

    Coordinator lookup on 2026-09-19 from an already-persisted scan; no new measurement. Eleven KMT2A-rearranged lines against 43 other leukaemia lines, none infant-derived, and CRISPR knockout rather than pharmacological inhibition. GENE EFFECT IS NOT DRUG EFFECT: every approved agent in this class engages both kinases, so the ranking does not describe what palbociclib, ribociclib or abemaciclib actually do in a patient. CDK4 and CDK6 are also partially redundant in some lineages, and single-gene knockout scores understate a target whose paralogue compensates.

    Evidence that argues against this (1)
    • This does not argue against the class, and reading it that way would be the error. A drug that hits CDK6 hits CDK6 whatever CDK4's rank. What the ranking changes is the RATIONALE and the risk calculus: the case should be written as CDK6, a CDK6-selective agent would be the cleaner instrument if one existed, and the CDK4 component is carrying toxicity for a kinase this disease does not preferentially need - which in children is not an academic distinction. Against all of that, the single-agent selectivity screen is also the wrong test for a drug proposed as a COMBINATION partner after KMT2C/KMT2D loss, and it cannot see a dependency acquired only after menin-inhibitor exposure. Broad Institute DepMap / Open Targets; analyses/depmap-kmt2a-selective-scan-full 2026
Connected to