Oncofusion

start / graph / drug-venetoclax

Drugmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q6b-therapeutics2

Venetoclax in KMT2A-rearranged leukemia — the apoptotic partner every other drug needs

also called ABT-199 · Venclexta · BCL-2 inhibitor

Venetoclax removes the brake that stops cancer cells from killing themselves. On its own it rarely cures anything, but added to almost any other drug in KMT2A-rearranged leukemia it converts a growth-arresting effect into actual cell death.

The same thing, for a scientist

Venetoclax is a selective BH3-mimetic BCL-2 inhibitor. KMT2A-rearranged infant ALL is characterised by high BCL-2 expression, and KMT2A-rearranged AML by BCL-2 dependency, making it a natural combination partner rather than a single agent. In children with relapsed/refractory AML, venetoclax with high-dose cytarabine ± idarubicin gave CR/CRi in 14 of 20 (70%, 95% CI 46-88) at the recommended phase 2 dose. It is now randomised in frontline infant KMT2A-rearranged ALL (NCT06317662) and paired with menin inhibitors in at least six trials. Critically, venetoclax resensitises PRC1.1-deficient, menin-inhibitor-resistant leukemia, giving it a mechanistically specific role in the resistance problem rather than a generic additive one.

Go deeper
The paediatric clinical data1 deeper

A three-centre phase 1 dose-escalation study enrolled 38 patients aged 3-22 years (median 10) with relapsed/refractory AML or acute leukemia of ambiguous lineage; 36 received combination therapy. The recommended phase 2 dose was venetoclax 360 mg/m² (max 600 mg) with cytarabine 1000 mg/m² x8 doses, ± idarubicin 12 mg/m² once 1.

  • Overall response in 24/35 (69%) evaluable after cycle 1.
  • At the RP2D: CR/CRi in 14/20 = 70% (95% CI 46-88), plus 2 partial responses.
  • Grade 3-4: febrile neutropenia 66%, bloodstream infection 16%, invasive fungal infection 16%. One treatment-related death (colitis and sepsis).

That 16% invasive fungal infection rate is the number to hold onto. Venetoclax is metabolised by CYP3A4, azole antifungals are strong CYP3A4 inhibitors, and the resulting dose-reduction dance is the practical toxicity story of every venetoclax combination — and it collides directly with menin inhibitors, which are themselves given with a strong CYP3A inhibitor (cobicistat) in the revumenib programme.

Preclinical basis specifically in infant KMT2A-rearranged ALL

KMT2A-rearranged infant ALL cells show promoter hypermethylation and high BCL-2 expression. Venetoclax was additive with hypomethylating agents and with conventional chemotherapy in vitro, and azacitidine plus venetoclax gave a significant in vivo survival advantage in KMT2A-rearranged infant ALL xenografts 2. A caution from the same work: hypomethylating agents were antagonistic with methotrexate in a subset of samples, apparently via ABC transporter upregulation — combination is not automatically additive and the specific pairing matters.

BCL-2 alone may not be enough. Co-inhibition of BCL-2 and MCL1 enhanced glucocorticoid activity in preclinical models of infant KMT2A-rearranged ALL 3, which is consistent with MCL1 being the standard escape route from BH3 mimetics and with CDK9 inhibitors (which deplete MCL1) synergising with venetoclax in the same disease.

Why venetoclax is the right partner for a menin inhibitor, mechanistically and not just empirically

Menin inhibitors are differentiation agents. In KMT2A-rearranged AML they act slowly and push cells toward myeloid maturation rather than killing them 4. A differentiating cell that does not die is a cell that can reacquire a stem program, and that is one plausible substrate for adaptive resistance.

The specific link: loss of PRC1.1 subunits (PCGF1, BCOR) causes menin-inhibitor resistance through epigenetic reactivation of non-canonical menin targets including MYC; those resistant cells show reduced monocytic signatures and are susceptible to BCL-2 inhibition, and venetoclax overcomes menin-inhibitor resistance in PRC1.1-deficient cells 5. So venetoclax is not simply an additive cytotoxic here — it covers a named, mechanistically characterised resistance state.

This pairing is now in the clinic at scale: the ziftomenib phase 3 KOMET-017 programme (NCT07007312, n=1300) uses venetoclax+azacitidine as the non-intensive backbone; the revumenib phase 3 NCT06652438 (n=448) adds revumenib to azacitidine+venetoclax; and paediatric phase 1 trials NCT06177067 (St. Jude) and NCT06397027 (MD Anderson) test revumenib or ziftomenib with azacitidine and venetoclax in children 6.

The infant question, which is actually being randomised

NCT06317662 (NCI/COG, recruiting, n=153, 115 sites) gives blinatumomab plus chemotherapy to all newly diagnosed infants with ALL and randomises the KMT2A-rearranged patients to the addition of venetoclax 6. This is the single most important venetoclax question in this disease and it is enrolling now. It is also, notably, testing venetoclax on top of an immunotherapy backbone rather than as a chemotherapy sensitiser, which is a different hypothesis than the preclinical work supports.

Facts and figures
targetBCL-2 (BH3-mimetic; frees BIM to trigger mitochondrial apoptosis)
chembl IDCHEMBL3137309
modalityoral small molecule
statusFDA/EMA approved in CLL and in AML combinations for adults; paediatric use is off-label or protocol-based, no paediatric indication
pediatric eligibletrue
pediatric rp2dvenetoclax 360 mg/m² (max 600 mg) daily with cytarabine 1000 mg/m² x8 ± idarubicin 12 mg/m²
trials
[
  "NCT03194932",
  "NCT06317662",
  "NCT07007312",
  "NCT06652438",
  "NCT06177067",
  "NCT06397027",
  "NCT06284486",
  "NCT06448013"
]
key results
{
  "pediatric_r_r_aml_cr_cri_pct": 70,
  "n": 20,
  "ci95": [
    46,
    88
  ],
  "orr_pct": 69,
  "orr_n": 35
}
key ae
[
  "febrile neutropenia 66%",
  "invasive fungal infection 16%",
  "bloodstream infection 16%",
  "tumour lysis syndrome",
  "CYP3A4 interaction with azoles"
]
resistance
[
  "MCL1 upregulation",
  "BCL-xL dependence",
  "loss of BAX/BAK",
  "monocytic differentiation state"
]
How this could bite us back
  • Resistance: MCL1 upregulation is the canonical escape from BCL-2 inhibition and is already implicated in this disease; adding an MCL1 inhibitor introduces a cardiac toxicity risk that stopped several adult programmes.
  • Essential-function toxicity: BCL-2 is required for normal lymphocyte and renal development; infants are the population in which chronic BCL-2 blockade has the least safety data and the most developmental exposure.
  • Drug-drug interaction as a delivery problem: venetoclax plus azole prophylaxis plus a CYP3A-boosted menin inhibitor in a neonate is a pharmacokinetic problem with no paediatric dataset, and dose errors here are lethal in both directions.
  • Infection: a 16% invasive fungal infection rate in the paediatric combination study is the real limiting toxicity, and it is worse in infants.
  • Tumour lysis in a high-count infant with hyperleukocytosis is a specific, foreseeable early hazard of a rapidly apoptotic combination.
Still open
  1. Does venetoclax add anything on top of blinatumomab in frontline infant KMT2A-rearranged ALL? NCT06317662 is the randomised test and it is enrolling now.
  2. Is the menin-inhibitor plus venetoclax combination additive or genuinely resistance-preventing? The PRC1.1/MYC mechanism predicts the latter, but only in a specific genotype; measuring PCGF1/BCOR status in patients on combination trials would answer it and nobody is reporting it.
  3. Can venetoclax and a menin inhibitor be co-dosed with azole prophylaxis in an infant? Both are CYP3A-sensitive, revumenib is deliberately boosted with cobicistat, and the resulting three-way interaction has no paediatric pharmacokinetic data.
  4. Is high BCL-2 expression in infant KMT2A-rearranged ALL actually predictive of venetoclax response, or merely present? BH3 profiling of primary infant samples would test functional dependency rather than expression.
Every source behind this entry (6)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In a phase 1 dose-escalation study of 38 children and young adults aged 2-22 years with relapsed/refractory AML or acute leukemia of ambiguous lineage, venetoclax with cytarabine ± idarubicin gave overall responses in 24/35 (69%) and, at the recommended phase 2 dose of venetoclax 360 mg/m², CR or CRi in 14/20 (70%, 95% CI 46-88), with grade 3-4 febrile neutropenia in 66%, bloodstream infection in 16% and invasive fungal infection in 16%.

    PMID 32171069DOI
    What this evidence cannot show

    Phase 1, three centres, n=36 treated, median follow-up only 7.1 months, and the population is relapsed/refractory AML rather than KMT2A-rearranged infant ALL specifically. Minimum age was 2 years, so this study says nothing directly about infants. The 70% CR/CRi rate comes from 20 patients and its confidence interval (46-88) is correspondingly wide.

    Evidence that argues against this (1)
    • The high response rate does not transfer to the states KMT2A-rearranged disease tends to occupy. In paired patient samples, monocytic AML subclones - which lose BCL-2 and depend on MCL-1 instead - are intrinsically resistant to venetoclax-based therapy and are selected for at relapse, and monocytic differentiation is exactly the phenotype KMT2A-rearranged myeloid disease and menin-inhibitor treatment produce. A phase 1 response rate in a mixed R/R paediatric AML population therefore over-reads onto the KMT2A-rearranged monocytic subset. Pei S et al. 2020
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    KMT2A-rearranged infant ALL cells show promoter hypermethylation and high BCL-2 expression; venetoclax was additive with hypomethylating and conventional agents in vitro, and azacitidine plus venetoclax gave a significant in vivo survival advantage in KMT2A-rearranged infant ALL xenografts, while hypomethylating agents were antagonistic with methotrexate in a subset of samples via ABC transporter upregulation.

    PMID 36380143DOI
    What this evidence cannot show

    Cell lines and xenografts only; xenograft survival advantage is not remission and mice were not cured. The antagonism with methotrexate is a genuine warning that in vitro additivity does not generalise across combinations, and methotrexate is a backbone drug in Interfant protocols.

    Evidence that argues against this (1)
    • The azacitidine half of this preclinical hypothesis has now been tested clinically in exactly this population and did not move outcome. Children's Oncology Group trial AALL15P1 gave five days of azacitidine immediately before chemotherapy in four post-induction courses to infants with newly diagnosed KMT2A-rearranged ALL. It was well tolerated (dose-limiting toxicity in 2 of 31 evaluable patients, 6.5%), and whole-genome bisulfite sequencing of peripheral blood mononuclear cells showed decreased DNA methylation in 87% of samples after the five days — so the pharmacodynamic premise was confirmed — yet EVENT-FREE SURVIVAL WAS SIMILAR TO PRIOR STUDIES of newly diagnosed infant ALL. Demonstrated hypomethylation plus unchanged EFS is the sharpest possible counter to a preclinical survival advantage: the mechanism worked and the outcome did not follow. Guest EM et al. 2024
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Adding MCL1 inhibition to venetoclax and dexamethasone was tested in preclinical models of infant KMT2A-rearranged ALL. The MCL1 inhibitor S63845 was selectively active in KMT2A-rearranged lines (IC50 9-218 nM, versus 232 nM to >1.5 microM in KMT2A-wildtype lines), the four most sensitive lines were all infant-derived, and the triple combination matched induction chemotherapy in infant patient-derived xenograft models.

    PMID 42246204DOI
    What this evidence cannot show

    Read in full from PMC13461956 on 2026-09-19. This item previously stated that no abstract was available and carried the title alone at weak confidence; that was wrong. PubMed and the Europe PMC core record both return an empty abstract for this PMID, but the paper is open access and the full text is retrievable through the {PMCID}/fullTextXML endpoint, which the original check did not try. See `concept-known-but-unreadable-literature` for the corrected rule. Remaining limits are the ordinary ones: these are cell lines and PDX models, 'matched induction chemotherapy' is a preclinical survival comparison and not a clinical equivalence claim, and no infant has received this combination.

    Evidence that argues against this (1)
    • MCL1 inhibition has a clinical liability that preclinical co-inhibition models do not display. In the first-in-human phase 1 of the MCL-1 inhibitor AZD5991 (monotherapy n=61; AZD5991 + venetoclax n=17 in AML/MDS), asymptomatic troponin I or T elevation occurred in 8 patients (10.3%), and post-hoc analysis found elevated troponin T in 54 of 65 patients after any dose at or after cycle 1, with no association to cardiovascular risk factors. Four deaths occurred from adverse events, including a cardiac arrest. Across the entire study only THREE objective responses were seen (two on monotherapy, one on the venetoclax combination). The authors' stated conclusion is a high incidence of troponin elevation with a low overall response rate. Adding MCL1 blockade to BCL-2 blockade and a glucocorticoid in infants therefore imports a cardiac-biomarker signal into a population with no cardiac reserve to spend. Desai P et al. 2024
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    KMT2A-rearranged ALL shows rapid dose-dependent cell death with revumenib whereas KMT2A-rearranged AML responds more slowly and undergoes myeloid differentiation, indicating that menin inhibition is a differentiation agent in the myeloid setting.

    Distinct Responses to Menin Inhibition and Synergy with DOT1L Inhibition in KMT2A-Rearranged Acute Lymphoblastic and Myeloid Leukemia · Adriaanse FRS, Schneider P, Arentsen-Peters STCJM, et al. · International Journal of Molecular Sciences 2024
    PMID 38892207DOI
    What this evidence cannot show

    In vitro primary-sample work; the ALL-versus-AML kinetic difference has not been confirmed in patients, where response assessment is confounded by concurrent chemotherapy and by differentiation syndrome.

    Evidence that argues against this (1)
    • The lymphoid-versus-myeloid dichotomy of menin-inhibitor mechanism does not survive contact with pretreated disease, so it should not be used to predict clinical behaviour. In serially sampled patient-derived xenografts from patients with multiply relapsed KMT2A::AFF1 ALL, the LATER samples were markedly less menin-inhibitor sensitive than earlier samples from the same patient despite no prior menin-inhibitor exposure — and transcriptomic analysis showed sustained on-target suppression of the canonical menin-inhibitor target genes in those resistant cells. 'Rapid dose-dependent cell death in KMT2A-r ALL' is therefore a property of treatment-naive or lightly treated lymphoid cells, not of the lineage, and the same on-target transcriptional response accompanies both outcomes. Mahdavi L et al. 2026
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Loss of PRC1.1 subunits PCGF1 or BCOR causes menin-inhibitor resistance via epigenetic reactivation of non-canonical menin targets including MYC; these resistant cells show reduced monocytic signatures, are susceptible to BCL-2 inhibition, and venetoclax overcomes their menin-inhibitor resistance.

    Epigenetic regulation of noncanonical menin targets modulates menin inhibitor response in acute myeloid leukemia · See PubMed record PMID 39158067 for the full author list (not transcribed here) · Blood 2024
    PMID 39158067DOI
    What this evidence cannot show

    Chromatin-focused CRISPR screens plus human and mouse KMT2A-rearranged AML models; this is a defined genetic resistance state, not necessarily the dominant clinical one. No patient data demonstrate that PRC1.1-deficient relapses respond to venetoclax.

    Evidence that argues against this (1)
    • This proposes venetoclax as the answer to menin-inhibitor resistance, but the two mechanisms may be self-defeating in sequence. Menin inhibitors act in myeloid disease by driving monocytic differentiation, and monocytic AML subclones lose BCL-2 dependence, switch to MCL-1 and are resistant to venetoclax-based therapy, being selectively enriched at relapse in patient-paired samples. The PRC1.1-loss resistance state described here is reported to have a REDUCED monocytic signature, so the venetoclax rescue may apply only to that minority state and fail in the differentiation-driven majority. Pei S et al. 2020
  6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.trial_registry

    Venetoclax is being combined with menin inhibitors in registered trials including ziftomenib phase 3 KOMET-017 (NCT07007312, n=1300), revumenib phase 3 NCT06652438 (n=448), paediatric NCT06177067 and NCT06397027, and is randomised in frontline infant KMT2A-rearranged ALL in NCT06317662 (n=153, recruiting).

    NCT06317662
    What this evidence cannot show

    Registry data as of 2026-09-18. None of these trials has reported. Registration establishes intent, not benefit.

    Evidence that argues against this (1)
    • None of these registered trials is designed to show that adding venetoclax to a menin inhibitor improves survival in infants. Checked against the raw ClinicalTrials.gov v2 API on 2026-09-19: NCT06317662 (NCI, phase 2, estimated n=153, maximum age 365 days, recruiting since 2025-06-05) has as its PRIMARY outcomes the incidence of dose-limiting toxicities in the safety and expansion phases and the MRD-negative remission rate AT THE END OF INDUCTION, with primary completion 2028-12-31 — event-free and overall survival are not primary endpoints. The two adult phase 3s have a minimum age of 18 years with primary completion 2031-11 (NCT07007312, estimated n=1300) and 2029-12-13 (NCT06652438, estimated n=448). An MRD endpoint at end of induction is a surrogate whose relationship to survival in infant KMT2A-r ALL is itself an open question, so the trial map does not license a survival claim in either direction. National Cancer Institute (NCI) 2025
Connected to