Oncofusion

start / graph / drug-azacitidine

Drugstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q81-methylome-axis

Azacitidine - the one hypomethylating agent actually given to infants with KMT2A-rearranged ALL: safe, demonstrably demethylating, and it did not change survival

also called azacitidine · 5-azacytidine · 5-aza-CR · AZA · Vidaza · azacytidine · Onureg (oral) · CC-486

Azacitidine strips methyl marks off DNA and is used in adult blood cancers. It is the only drug of its kind that has been given to infants with KMT2A-rearranged leukaemia in a trial: it was safe and it did strip the marks, and it made no difference to how long the babies lived.

The same thing, for a scientist

**Azacitidine (CHEMBL1489, first approval 2004) is a ribonucleoside analogue - not a deoxyribonucleoside like decitabine - and that difference is not cosmetic: a substantial fraction of azacitidine is incorporated into RNA rather than DNA, so the two agents are pharmacologically distinct and are not interchangeable.** It matters here because azacitidine is the only hypomethylating agent that has been tested prospectively in the index disease. **Children's Oncology Group AALL15P1 (NCT02828358, phase 2, n = 78 enrolled, maximum eligible age 364 days) gave five days of azacitidine immediately before chemotherapy in four post-induction courses to infants with newly diagnosed KMT2A-rearranged ALL** [E1][E2]. **The result is the most informative single datum in this beat.** The treatment was well tolerated - 2 of 31 evaluable patients (6.5%) had a dose-limiting toxicity. **Whole-genome bisulfite sequencing of peripheral blood mononuclear cells showed decreased DNA methylation in 87% of samples tested**, so the pharmacodynamic hypothesis was confirmed. **And event-free survival was similar to prior studies of newly diagnosed infant ALL: the incorporation of azacitidine to enhance cytotoxicity did not affect survival** [E1]. Preclinically the drug looks better than that: single-agent azacitidine significantly prolonged survival in infant KMT2A-rearranged ALL xenografts, and **azacitidine plus venetoclax gave a significant in vivo survival advantage** - the combination the paper actually advances [E3]. Azacitidine has also maintained a second remission in a single infant with relapsed KMT2A-rearranged ALL after cord-blood transplant [E4]. **The gap between a confirmed on-target pharmacodynamic effect and an unchanged survival curve is the single most important thing this graph can say about methylome-directed therapy in this disease**, and it should temper any enthusiasm for adding a hypomethylating agent to a menin inhibitor in infants.

Go deeper
AALL15P1: what was proved, and what was disproved
questionanswerstrength
Can you give azacitidine safely to infants with KMT2A-r ALL?Yes. 2/31 evaluable (6.5%) dose-limiting toxicitydirect, prospective, in the index population
Does it demethylate in those infants?Yes. WGBS of PBMCs showed decreased methylation in 87% of samples tested after 5 daysdirect pharmacodynamic confirmation
Does epigenetic priming improve survival?No. EFS similar to prior newly diagnosed infant ALL studiesthe trial was a pilot for safety and biologic activity, not powered for efficacy

The correct reading is narrow and should not be over-sold in either direction. AALL15P1 was "a groupwide pilot study to test the tolerability and biologic activity" 2 - it is not a randomised efficacy trial and 'EFS similar to prior studies' is a historical comparison, not a controlled one. But it is also not nothing: the drug reached its target, the biomarker moved in 87% of patients, and the curve did not. A mechanism that works and a benefit that does not appear is a harder result to explain away than a mechanism that failed.

Three candidate explanations, all live: the PBMC methylome is not the blast methylome and the WGBS was on PBMCs; priming before chemotherapy is a schedule hypothesis and the wrong schedule can waste a real effect; and hypomethylation may simply not be the axis that limits outcome in this disease.

Azacitidine is not decitabine, and this graph should stop treating them as one drug

They are routinely reported together - decitabine[tiab] AND azacitidine[tiab] returns 647 records, 21% of the decitabine literature - and in this project's own searches they collide constantly. They are different molecules with different fates.

azacitidinedecitabine
sugarribosedeoxyribose
incorporationRNA and DNA (RNA substantially)DNA only
DNMT1 degradationyes (both azacytidine and 5-aza-CdR in the founding study)yes
RNA-directed effectsyes, and they are not a methylation mechanismessentially none
this graph's index-disease evidencea completed phase 2 in infantscell lines and xenografts only

Ghoshal 2005 demonstrated DNMT1 degradation for both agents, so the shared mechanism is real 5. What is not shared is everything azacitidine does to RNA, and a trial that fails is not evidence against the DNA-methylation hypothesis if the drug's dose-limiting biology is elsewhere. The honest position: AALL15P1 tested azacitidine, and it is a weaker test of decitabine than it looks.

The combination question, honestly framed

Corrected inside this beat. The first draft of this entry said "there is no published trial of azacitidine plus a menin inhibitor" and put "registry searching was not performed" in the caveat. The registry was then searched, and it changes the picture: azacitidine plus a menin inhibitor is an active and substantial clinical programme, it is simply unpublished 8.

what existswhereage floor
decitabine + revumenib + venetoclax, R/R AML, CRc 71%, n = 42published (SAVE) 612 years
azacitidine + revumenib + venetoclax, paediatric/young adult R/R AML, phase 1, n = 24 est., St. Juderegistry only, recruiting, NCT061770671 year
azacitidine + ziftomenib + venetoclax, paediatric R/R acute leukaemias, phase 1, MD Andersonregistry only, recruiting, NCT063970272 years
azacitidine + revumenib + venetoclax, newly diagnosed KMT2A-rearranged AMLregistry only, not yet recruiting, NCT07605949not retrieved
azacitidine + revumenib + venetoclax, NPM1-mutated or KMT2A-rearranged AMLregistry only, recruiting, NCT06652438not retrieved
azacitidine + bleximenib + venetoclax, newly diagnosed AMLregistry only, recruiting, NCT06852222not retrieved
azacitidine + ziftomenib + venetoclax or 7+3, untreated NPM1-mutated or KMT2A-rearranged AMLregistry only, recruiting, NCT07007312not retrieved

Three things follow. First, the combination this beat was asked about is being tested right now, and the graph should track it rather than call it absent. Second, every one of these regimens is a TRIPLET containing venetoclax, so even when they read out, no pairwise menin-inhibitor-plus-hypomethylating-agent effect will be attributable. Third, and the reason this matters most here: the lowest age floor found across the whole programme is 1 year (NCT06177067), so infants - the index population of this project - are outside every one of them. Azacitidine plus venetoclax preclinically in infant KMT2A-rearranged ALL 3 is the one combination with index-disease support and no trial.

And AALL15P1 sits across all of it. A hypomethylating agent that demethylated 87% of treated infants without changing survival is a warning that adding it to a menin inhibitor may add toxicity and schedule complexity for a benefit that has already failed to appear once in exactly this population 1.

Facts and figures
chembl IDCHEMBL1489
targetgene-dnmt1
modalitysmall molecule (cytidine RIBOnucleoside analogue; incorporated into RNA as well as DNA)
statusApproved (ChEMBL max_phase 4, first approval 2004). In infants with KMT2A-rearranged ALL it was given investigationally under IND# 133688 / NSC# 102816 on COG AALL15P1.
distinct from decitabineRibonucleoside, substantial RNA incorporation; decitabine is a deoxyribonucleoside incorporated into DNA only. Both degrade DNMT1 selectively over DNMT3A/3B. A null for one is a weak test of the other.
shown for fusionYES, and in the index disease: COG AALL15P1 in infants under one year with newly diagnosed KMT2A-rearranged ALL. Safe (6.5% DLT), pharmacodynamically active (87% of WGBS samples demethylated), and NO survival effect. Also single-agent xenograft survival benefit and a single relapsed-infant case report.
measured on2026-09-21, PubMed esearch, querytranslation inspected
retracted citation flaggedPMID 15899874 is RETRACTED (RetractionIn PMID 35412342). Retained with the retraction stated, demoted to weak, and replaced as support by PMID 41460614.
trials
[
  "NCT02828358",
  "NCT06177067",
  "NCT06397027",
  "NCT07605949",
  "NCT06652438",
  "NCT06852222",
  "NCT07007312"
]
key ae
[
  "dose-limiting toxicity in 2 of 31 evaluable infants (6.5%) on AALL15P1 - the trial's headline safety result",
  "myelosuppression (class effect; not separately quantified in the infant trial abstract)"
]
resistance
[
  "ABC-transporter upregulation proposed as the mechanism of antagonism between hypomethylating agents and methotrexate in a subset of infant KMT2A-rearranged ALL samples"
]
index disease outcome
{
  "trial": "NCT02828358 (COG AALL15P1)",
  "population": "infants <1 year, newly diagnosed KMT2A-rearranged ALL",
  "n_enrolled": 78,
  "n_evaluable_for_dlt": 31,
  "dlt_k": 2,
  "dlt_pct": 6.5,
  "pharmacodynamic_response": "decreased DNA methylation in 87% of PBMC samples tested by whole-genome bisulfite sequencing",
  "efficacy": "event-free survival similar to prior studies of newly diagnosed infant ALL; addition of azacitidine did not impact survival",
  "design_limit": "pilot for tolerability and biologic activity; historical rather than randomised comparison"
}
name list for absence bounding
[
  "azacitidine (3,044 tiab)",
  "5-azacytidine",
  "5-aza-CR",
  "AZA",
  "Vidaza",
  "CC-486",
  "Onureg"
]
menin inhibitor combination programme
{
  "published_with_azacitidine": "none",
  "published_with_decitabine": "SAVE, NCT05360160, minimum age 12 years",
  "registered_with_azacitidine": [
    "NCT06177067",
    "NCT06397027",
    "NCT07605949",
    "NCT06652438",
    "NCT06852222",
    "NCT07007312"
  ],
  "lowest_age_floor_found": "1 year (NCT06177067, St. Jude) - infants under one year are outside every registered menin-inhibitor plus hypomethylating-agent trial found",
  "all_are_triplets_with_venetoclax": true,
  "queried": "ClinicalTrials.gov API v2, 2026-09-21"
}
How this could bite us back
  • The lowest age floor found across the entire registered menin-inhibitor plus hypomethylating-agent programme is 1 year. Infant KMT2A-rearranged ALL - the disease with the worst outcome and the strongest hypermethylation phenotype - is excluded from all of it, and the only trial that did include infants is the one that showed no survival benefit.
  • The definitive fact about this drug in this disease is a negative one: it demethylated 87% of treated infants and the survival curve did not move. Any proposal to add a hypomethylating agent to a menin inhibitor in infants must explain why this time is different, and 'we will use decitabine instead' is a hypothesis, not an answer.
  • Hypomethylating agents were antagonistic with methotrexate in a subset of infant KMT2A-rearranged ALL samples, and methotrexate is in the AALL15P1 backbone and in every Interfant protocol. A priming schedule can therefore subtract as well as add, and the trial cannot distinguish 'no effect' from 'a real effect cancelled by an antagonism'.
  • Global DNA demethylation in a neonate is a whole-organism intervention with a decades-long follow-up horizon and none of that follow-up exists. Imprinted loci, transposable elements and germline-adjacent tissues are all exposed, and this graph already carries hyp-transposable-element-reactivation-kmt2a-r as an open question.
  • The one positive clinical report is n = 1, post-transplant, where graft-versus-leukaemia and azacitidine's immunomodulatory effects are at least as plausible as any methylation mechanism. It should never be cited as efficacy evidence.
Still open
  1. Six azacitidine-plus-menin-inhibitor trials are recruiting and every one of them is a triplet containing venetoclax. When they read out, what will attribute any benefit to the hypomethylating agent rather than to venetoclax? No two-arm design exists anywhere in the programme.
  2. Did the leukaemic blasts demethylate, or only the PBMC pool? WGBS was on PBMCs. Sorted-blast methylomes from banked AALL15P1 samples would distinguish 'the drug did not reach the target cell' from 'the target was hit and it did not matter'.
  3. Was the failure a schedule failure? Priming for five days before chemotherapy is one of many possible schedules, and the preclinical support was for continuous or combination exposure.
  4. Would azacitidine plus venetoclax - the combination the preclinical paper actually recommends and the one with index-disease xenograft support - have done better than priming? Nobody has run it in infants.
  5. Is azacitidine's failure in infants evidence against decitabine in infants? The two differ in RNA incorporation, and decitabine has the MLL-r-selective cytotoxicity and the KMT2A-degradation mechanism that azacitidine has never been tested for.
  6. Does azacitidine lower KMT2A protein as decitabine does? The 2025 ALL paper tested decitabine; the thermal proteome profiling in PMID 36380143 covered both agents and was not examined here.
Every source behind this entry (8)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In infants under one year with newly diagnosed KMT2A-rearranged ALL (COG AALL15P1), five days of azacitidine immediately before chemotherapy in four post-induction courses was well tolerated with dose-limiting toxicity in 2 of 31 evaluable patients (6.5%); whole-genome bisulfite sequencing of peripheral blood mononuclear cells showed decreased DNA methylation in 87% of samples tested; and event-free survival was similar to prior studies of newly diagnosed infant ALL, with the addition of azacitidine not affecting survival.

    PMID 38867582DOI
    What this evidence cannot show

    A pilot study designed for tolerability and biologic activity, not powered for efficacy; the survival comparison is against historical prior studies, not a randomised control. WGBS was performed on PERIPHERAL BLOOD MONONUCLEAR CELLS, not on sorted blasts, so the 87% demethylation figure is a systemic pharmacodynamic readout and not proof that the leukaemic methylome moved. 31 evaluable for DLT out of 78 enrolled. Read from abstract only in this beat; the EFS point estimate and confidence interval were not extracted, so 'similar to prior studies' is quoted rather than recomputed.

    Evidence that argues against this (1)
    • Preclinically, single-agent azacitidine significantly prolonged survival in infant KMT2A-rearranged ALL xenografts and azacitidine plus venetoclax gave a further significant advantage - so the clinical null may be about schedule, partner drug or the priming design rather than about hypomethylation as a strategy. Cheung LC et al. 2023
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry

    AALL15P1 is registered as NCT02828358, a completed phase 2 groupwide pilot of azacitidine added to chemotherapy in infants with ALL and KMT2A rearrangement, actual enrolment 78, maximum eligible age 364 days, started 2017-04-01 and completed 2024-12-31, with no whyStopped field set.

    NCT02828358
    What this evidence cannot show

    Read from the raw ClinicalTrials.gov API v2 on 2026-09-21 rather than the MCP tool, precisely so that whyStopped would be visible; it is empty, and the record shows COMPLETED rather than terminated. The registry lists twenty other drugs and procedures in the intervention list because the azacitidine is added to a full chemotherapy backbone, which is why no single-agent inference is available from this trial.

    Evidence that argues against this (1)
    • The registered intervention list includes methotrexate, and hypomethylating agents were reported antagonistic with methotrexate in a subset of infant KMT2A-rearranged ALL samples - so the backbone this trial primed contains the one agent with a documented negative interaction. Cheung LC et al. 2023
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In infant KMT2A-rearranged ALL, cells are characterised by abundant promoter hypermethylation and high BCL-2 expression; single-agent azacitidine and single-agent decitabine each significantly prolonged in vivo survival in xenografts; hypomethylating agents were additive with most conventional chemotherapeutics but antagonistic with methotrexate in a subset of samples via candidate ABC-transporter upregulation; and adding venetoclax to azacitidine produced a significant in vivo survival advantage.

    PMID 36380143DOI
    What this evidence cannot show

    Xenograft survival prolongation, not cure; the number of xenograft models, the survival statistics and the definition of 'a subset of samples' for the methotrexate antagonism were not extracted here. Read from abstract only. Thermal proteome profiling is cited in the abstract as defining the agents' target-binding landscape, and that dataset was not examined - it is the obvious place to look for whether KMT2A is among the bound proteins.

    Evidence that argues against this (1)
    • A completed prospective trial of azacitidine in exactly this population demethylated 87% of treated infants and did not change survival, which is a stronger form of evidence than xenograft survival prolongation and points the other way. Guest EM et al. 2024
  4. E4weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    In a single infant with KMT2A-rearranged ALL relapsing after unrelated cord blood transplantation, a second remission was maintained for 10 months on azacitidine alone, after which the patient proceeded to a second transplant.

    PMID 28675638DOI
    What this evidence cannot show

    n = 1. A case report, in the post-transplant maintenance setting, where a graft-versus-leukaemia effect and azacitidine's immunomodulatory actions are plausible alternative explanations to any methylation mechanism. Author list, DOI and abstract taken from the PubMed record fetched 2026-09-21 - the first draft of this entry carried a wrong author list and a wrong DOI written from recall, both corrected against the fetched record before this file was finished; the article text beyond the abstract was NOT read. Recorded because a single positive case in the index disease is worth knowing exists and is worth not over-reading.

    Evidence that argues against this (1)
    • A single case cannot be weighed against a prospective trial in 78 infants of the same disease that found no survival effect; the case is included as an existence proof of tolerability in a relapsed infant, not as efficacy evidence. Guest EM et al. 2024
  5. E5weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    Both 5-azacytidine (azacitidine) and 5-aza-deoxycytidine (decitabine) induce selective proteasomal degradation of DNMT1 and not of DNMT3A or DNMT3B, establishing the shared mechanism of the hypomethylating-agent class.

    PMID 15899874DOI
    What this evidence cannot show

    *** THIS CITATION IS RETRACTED. *** PMID 15899874 (Mol Cell Biol 2005) carries PublicationType 'Retracted Publication'; RetractionIn Mol Cell Biol 2022;42(5):e0054621, PMID 35412342, and an earlier ErratumIn at PMID 29712769. Found at the coordinator on 2026-09-21 while verifying this beat's citations, AFTER the entry was written. It is retained here ONLY as the historical source of the KEN-box/BAH/NLS mechanistic dissection and must not be cited as support for anything. THE CLAIM ITSELF SURVIVES ON AN INDEPENDENT, NON-RETRACTED SOURCE: PMID 41460614 (Peng 2025, Cell Oncol) reports decitabine-driven ubiquitin-dependent degradation of DNMT1, in colorectal cancer rather than leukaemia. The paper's detailed mechanism (KEN box, BAH, NLS, ubiquitin-activating enzyme, replication independence, catalytic-mutant insensitivity) is demonstrated for 5-aza-CdR; the abstract states the DNMT1 degradation result for both agents but the mutational dissection is reported for the deoxy compound. Cell lines, 2005, no leukaemia model.

    Evidence that argues against this (2)
    • The retraction is itself the strongest counter-evidence to this item as originally written. A retracted paper cannot support a class-level mechanistic claim, and the detailed KEN-box/BAH/NLS dissection this entry drew on is exactly the content a retraction calls into question. The surviving claim is narrower - that hypomethylating agents deplete DNMT1 - and it needs the independent source, not this one. Peng XM et al. 2025
    • A shared DNMT1 mechanism does not make the two agents equivalent: azacitidine is a ribonucleoside with substantial RNA incorporation, so its dose-limiting biology and its off-methylation effects differ from decitabine's, and a null result for one is a weak test of the other. Cheung LC et al. 2023
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    No trial of a menin inhibitor with a hypomethylating agent has PUBLISHED results in KMT2A-rearranged leukaemia using azacitidine; the only published menin-inhibitor plus hypomethylating-agent combination uses DECITABINE - the SAVE triplet of revumenib, decitabine/cedazuridine and venetoclax in relapsed/refractory AML aged 12 and over, with a 71% composite complete remission rate in 42 patients.

    PMID 42272166DOI
    What this evidence cannot show

    The absence half is bounded to the PUBLISHED literature by one PubMed search run 2026-09-21 - (decitabine OR azacitidine OR 5-azacytidine OR hypomethylating agent) AND (menin OR MEN1) AND (leukemia OR leukaemia), 62 records - with titles skimmed rather than each abstract read, so a cohort reported inside a larger paper would be missed (rule 80). IT IS NOT AN ABSENCE OF TRIALS: a ClinicalTrials.gov search the same day returned at least six active azacitidine-plus-menin-inhibitor studies (E8), and the first draft of this entry implied otherwise before that search was run. The SAVE numbers are adult and adolescent AML in a three-drug regimen, and an Erratum (PMID 42492014) exists and was not retrieved in this beat.

    Evidence that argues against this (1)
    • Six or more azacitidine-plus-menin-inhibitor trials are recruiting, two of them paediatric, so 'no published trial' is a statement about publication lag rather than about clinical interest, and reading it as an absence of activity would be wrong. St. Jude Children's Research Hospital 2026
  7. E7established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Azacitidine is ChEMBL1489, a small molecule with maximum phase 4 and a first approval recorded in 2004, carrying the synonym 5-azacitidine.

    ChEMBL molecule record CHEMBL1489 (AZACITIDINE) · ChEMBL, EMBL-EBI · ChEMBL database 2026
    What this evidence cannot show

    Fetched 2026-09-21. ChEMBL's first_approval is one global year with no indication, formulation or jurisdiction attached; the oral formulation (CC-486/Onureg) is a separate later approval. No regulatory registry was queried in this beat, so nothing here bounds paediatric or infant labelling (rule 75). In AALL15P1 the drug was given under IND# 133688 / NSC# 102816, i.e. investigationally.

    Evidence that argues against this (1)
    • That azacitidine was administered to infants under an IND rather than on label is itself the regulatory fact rule 75 is about, and it is visible in the trial's own registered title while being absent from every pharmacology database record. Children's Oncology Group 2026
  8. E8established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry

    A ClinicalTrials.gov query for hypomethylating agents combined with menin inhibitors returned 11 studies on 2026-09-21, of which at least six pair AZACITIDINE with a menin inhibitor - NCT06177067 (revumenib + azacitidine + venetoclax, paediatric and young adult relapsed/refractory AML, St. Jude, phase 1, estimated n = 24, minimum age 1 YEAR), NCT06397027 (ziftomenib + venetoclax + azacitidine, paediatric relapsed/refractory acute leukaemias, MD Anderson, phase 1, minimum age 2 years), NCT07605949 (newly diagnosed KMT2A-rearranged AML), NCT06652438, NCT06852222 and NCT07007312 - and the lowest age floor found is 1 year, so infants under one year are excluded from all of them.

    NCT06177067
    What this evidence cannot show

    Read from the raw ClinicalTrials.gov API v2 on 2026-09-21 rather than the MCP tool. NCT06177067 and NCT06397027 were retrieved individually with eligibility ages, sponsors, phases and intervention lists and neither has a whyStopped field set; the other four azacitidine studies were read from the search result's brief titles only and their age floors were NOT individually retrieved, so 'the lowest age floor found is 1 year' is bounded by the two records fetched in full plus the adult framing of the remaining four by title. A registry listing is a statement of intent, not of results: none of these has reported.

    Evidence that argues against this (1)
    • A registry record is not evidence of benefit, and several of these are phase 1 dose-finding studies with estimated enrolments of 22-24; counting them as a therapeutic programme overstates what is known, which is nothing about efficacy. Guest EM et al. 2024
Connected to