Oncofusion

start / graph / drug-etoposide

Drugstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q43-break-repair-machinery

Etoposide (VP-16) - the drug that makes this disease and cures several others, and the only entry in this graph whose main finding is that you must not stop using it

also called etoposide · VP-16 · VP16 · VePesid · etoposide phosphate · Etopophos · epipodophyllotoxin

Etoposide jams the enzyme that untangles DNA while it is mid-cut, so the cut never gets sewn up. That is how it kills cancer cells, and it is also the best-understood reason a child cured of one cancer can develop a KMT2A-rearranged leukaemia years later - which is why the answer is careful dosing and not avoidance.

The same thing, for a scientist

Etoposide (CHEMBL44657, approved 1983; the water-soluble prodrug etoposide phosphate CHEMBL1200645, approved 1996) is an epipodophyllotoxin topoisomerase II poison. It stabilises the TOP2 cleavage complex into a persistent protein-linked double-strand break [see `mechanism-top2-cleavage-at-kmt2a`], and the KMT2A breakpoint cluster region is a favoured site: the same position is cut by etoposide, by doxorubicin and by dietary bioflavonoids, and the cut is reversible until it is not [E4]. Epidemiologically the dose-response is among the steepest in oncology - in a matched case-control study of childhood solid-tumour survivors, >6 g/m2 of epipodophyllotoxin carried a relative risk of secondary leukaemia of **197** (95% CI 19-2058) [E2]. **And the correct reading of that number is not 'avoid etoposide'.** The NCI's own monitoring programme was created because concern about these reports risked 'premature abandonment of these agents', and its first analysis put the 6-year treatment-related AML/MDS rate at low cumulative dose at 3.2% - within the range already reported for alkylator regimens without epipodophyllotoxins [E3]. Etoposide is backbone therapy in paediatric AML, germ-cell tumours, Ewing sarcoma, Hodgkin lymphoma and HLH, and this entry exists to give the graph an accurate causal node, not a reason to withhold a curative drug. Its mechanism target is curated by ChEMBL at the level of the **protein family**, not the isoform [E1].

Go deeper
What it does, and why KMT2A

Etoposide does not inhibit topoisomerase II in the ordinary sense - it is a poison, trapping the covalent enzyme-DNA intermediate so that a transient nick becomes a persistent double-strand break with protein still attached to the 5' end. Removing that protein is TDP2's job [see mechanism-tdp2-repair-of-trapped-top2-complexes], and TDP2 loss makes translocation worse, not better.

Why KMT2A specifically: the BCR is a site where etoposide, doxorubicin and dietary bioflavonoids all induce cleavage at colocalising positions in primary progenitor haematopoietic cells from healthy newborns and adults, and where religation of the induced cleavage site was demonstrated directly 4. Most etoposide-induced breaks at this locus are TOP2B-dependent by genetic depletion [gene-top2b], but no compound - including this one - separates the isoforms 1.

The dose-response, stated precisely1 deeper

Le Deley et al. matched 61 children who developed secondary leukaemia or MDS after a childhood solid tumour with 196 controls within the SFOP. On multivariate analysis only two factors raised risk: the type of first tumour (Hodgkin RR 6.4, osteosarcoma RR 5) and exposure to epipodophyllotoxins and anthracyclines. Risk rose with cumulative etoposide dose: 1.2-6 g/m2 of epipodophyllotoxin or >170 mg/m2 of anthracycline gave a seven-fold higher risk (95% CI 2.6-19); >6 g/m2 of epipodophyllotoxin gave RR 197 (95% CI 19-2058) 2. Alkylating agents and radiotherapy did not raise the risk in this analysis.

Read the confidence interval, not the point estimate. A 95% CI spanning 19 to 2058 is a statement that the risk is large and that its magnitude is not known within two orders of magnitude. The authors' own conclusion is the balanced one and it is worth quoting in substance: the challenge is to minimise mutagenic effects by reducing cumulative doses without losing the therapeutic benefit.

The counterweight the field wrote for itself

In 1993 the NCI's Cancer Therapy Evaluation Program set up a prospective monitoring plan across 12 cooperative-group trials precisely because early case reports 'have led to growing concern among oncologists, which could lead to premature abandonment of these agents at a time when the relationship between cumulative dose of epipodophyllotoxin and risk of treatment-related AML has not been determined' 3. Its first analysis - 207 patients at low cumulative etoposide dose (<1500 mg/m2), three t-AML and one t-MDS - gave a 6-year rate of 3.2% (95% upper bound 7.2%), which the authors note is within the range previously reported for alkylator-based regimens that did not include epipodophyllotoxins.

So the honest pair of statements is: at high cumulative dose the risk is very large; at the doses used in most paediatric protocols it is not clearly different from the alternative agents. Both belong in this entry and neither on its own is usable.

Why this node is filed in the break-repair beat and not the therapeutics beat

Every other drug- entry in this graph is a candidate treatment for KMT2A-rearranged leukaemia. This one is an aetiological agent: the graph holds a patient story about secondary AML, a break-machinery beat and a TDP2 mechanism entry, and until now had no entry for the drug that makes the disease. Nothing here proposes etoposide as therapy for KMT2A-r leukaemia and nothing here proposes withholding it from a child with a curable tumour.

The schema requires data.target; it is set to gene-top2b because that is the isoform to which the KMT2A break has been genetically attributed. Etoposide is not TOP2B-selective - ChEMBL's curated target is a protein family containing both accessions 1 - and the field note in data.target_note says so.

Facts and figures
chembl IDCHEMBL44657
chembl ID prodrugCHEMBL1200645
targetgene-top2b
target noteSchema requires a single target id. gene-top2b is named because the KMT2A break under etoposide has been genetically attributed to TOP2B. Etoposide is NOT TOP2B-selective: ChEMBL's curated mechanism target CHEMBL2094255 is a PROTEIN FAMILY whose components are P11388 (TOP2A) and Q02880 (TOP2B), and no member-level mechanism record exists. Treat any isoform-level statement about this drug as naming, not measurement.
modalitysmall molecule, epipodophyllotoxin, topoisomerase II poison (cleavage-complex stabiliser, not a catalytic inhibitor)
statusApproved 1983 (etoposide) and 1996 (etoposide phosphate). Backbone or component therapy in paediatric AML, germ-cell tumours, Ewing sarcoma, Hodgkin lymphoma, small-cell lung cancer and haemophagocytic lymphohistiocytosis. NOT a proposed therapy for KMT2A-rearranged leukaemia in this graph.
role in this graphaetiological agent, not candidate therapy
mw588.56
trials noteTrial identifiers are taken from the ChEMBL drug_indication records for etoposide phosphate in leukaemia indications and are listed as provenance for the indication, not as trials of this drug in KMT2A-rearranged disease. None was read in this beat.
leukaemogenic risk
{
  "endpoint": "secondary leukaemia or MDS after a childhood solid tumour, any genotype",
  "source": "PMID 12637473 (SFOP case-control, 61 cases / 196 controls)",
  "rr_1p2_to_6_g_per_m2_epipodophyllotoxin_or_over_170mg_per_m2_anthracycline": 7,
  "rr_over_6_g_per_m2_epipodophyllotoxin": 197,
  "ci_for_197": "19 to 2058",
  "low_dose_comparator": "6-year t-AML/MDS rate 3.2% (95% upper bound 7.2%) at <1500 mg/m2 in 207 patients, within the range for alkylator regimens containing no epipodophyllotoxin (PMID 8455202)",
  "latency_and_genotype": "KMT2A-rearranged t-AML after epipodophyllotoxins characteristically has short latency and balanced 11q23 translocations; neither cited study reports genotype-specific risk, so the attribution of the excess risk to KMT2A rearrangement is inference"
}
molecular targets of the break
[
  "TOP2B (genetic attribution at KMT2A, gene-top2b E2)",
  "TOP2A (in-vitro cleavage of patient breakpoint sequence, gene-top2a E3)"
]
key ae
[
  "myelosuppression",
  "mucositis",
  "alopecia",
  "hypotension with rapid infusion",
  "secondary acute myeloid leukaemia with 11q23/KMT2A rearrangement, short latency"
]
resistance
[
  "reduced TOP2 expression lowers the number of cleavage complexes available to trap",
  "efficient TDP2-dependent repair removes the protein adduct and restores a ligatable end (see mechanism-tdp2-repair-of-trapped-top2-complexes)",
  "not enumerated further in this beat; resistance is not the reason this entry exists"
]
trials
[
  "NCT01256398",
  "NCT01423500",
  "NCT01949129",
  "NCT02881086",
  "NCT03983824"
]
How this could bite us back
  • **Etoposide is curative backbone therapy.** It is a component of standard regimens for paediatric AML, germ-cell tumours, Ewing sarcoma, Hodgkin lymphoma and HLH. Nothing in this entry supports withholding it, and an entry that read as 'avoid etoposide' would cost more lives than the leukaemia it describes. The NCI created a monitoring programme in 1993 specifically because premature abandonment was the foreseeable harm.
  • The headline relative risk of 197 comes from the >6 g/m2 stratum with a 95% CI of 19 to 2058 and an endpoint of secondary leukaemia of any genotype. Quoting the point estimate without the interval or the stratum misrepresents the evidence in the direction that does damage.
  • Reducing cumulative etoposide to reduce t-AML risk trades a certain loss of anti-tumour efficacy against an uncertain gain in second-cancer avoidance. The trade is only defensible where a non-inferiority result exists for the specific regimen; for most of the diseases etoposide is used in, it does not.
  • The graph's own break-machinery beat concludes that the KMT2A break is common and leukaemia is rare, and that the rare step is not the cut. If that is right, an etoposide-avoidance strategy targets the abundant step and would be expected to move incidence less than the cleavage data suggest.
  • An etoposide-attributable framing risks being read backwards into de novo infant disease, where there is no etoposide and where the largest epidemiological test of the dietary analogue was null (E4 counter). A therapy-related mechanism applied to a fetus is an extrapolation, not a finding.
Still open
  1. What fraction of the excess secondary-leukaemia risk in the SFOP high-dose stratum is KMT2A-rearranged? Neither cited epidemiological study genotypes its cases, so the drug's most famous causal claim in this disease has never been quantified in the disease itself.
  2. Is the anti-tumour activity of etoposide TOP2A-dependent, TOP2B-dependent or both? Without that answer the 'TOP2B-sparing epipodophyllotoxin' idea cannot be evaluated even in principle, because nobody knows what would be lost.
  3. Does schedule, as distinct from cumulative dose, change KMT2A-rearranged t-AML risk at a fixed total dose? The SFOP study recorded schedule of etoposide administration as a variable and reported only cumulative dose in its multivariate result; the schedule question is the one a protocol committee can act on.
  4. Do the ex-vivo etoposide-sensitive break-prone sites (E6) exist before the first leukaemia, or are they a consequence of it? A remission sample from a patient who already had KMT2A-r AML cannot distinguish susceptibility from sequela.
Every source behind this entry (6)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Etoposide is ChEMBL CHEMBL44657, an approved (max_phase 4, first approval 1983) small molecule of MW 588.56 whose curated mechanism is 'DNA topoisomerase II inhibitor' against CHEMBL2094255, a PROTEIN FAMILY target whose components are P11388 (TOP2A) and Q02880 (TOP2B); the water-soluble prodrug etoposide phosphate is CHEMBL1200645, approved 1996.

    What this evidence cannot show

    A curation fact. ChEMBL records what medicinal chemistry has been done and how a curator filed it; the family-level target reflects both the pharmacology and the curator's choice. It is used here only to establish that no member-level mechanism record exists for either isoform.

    Evidence that argues against this (1)
    • The family-level filing should not be read as saying the isoforms are biologically interchangeable for this drug. Genetic depletion separates them at the KMT2A locus - most etoposide-induced breaks there are TOP2B-dependent - so the enzyme that matters is resolvable even though the compound cannot resolve it. Cowell IG et al. 2012
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In a matched case-control study within the Societe Francaise d'Oncologie Pediatrique (61 children with secondary leukaemia or myelodysplasia after a childhood solid tumour, 196 controls), the risk of secondary leukaemia rose with cumulative epipodophyllotoxin dose: 1.2-6 g/m2 of epipodophyllotoxins or >170 mg/m2 of anthracyclines gave a seven-fold higher risk (95% CI 2.6-19), and >6 g/m2 of epipodophyllotoxins gave a relative risk of 197 (95% CI 19-2058); alkylating agents and radiotherapy did not increase risk.

    PMID 12637473DOI
    What this evidence cannot show

    Case-control within a cooperative group, 61 cases, and the highest dose stratum is small - a 95% CI of 19 to 2058 is a two-order-of-magnitude uncertainty. The endpoint is secondary leukaemia/MDS of any genotype, NOT KMT2A-rearranged leukaemia specifically, so it does not by itself attribute the excess to KMT2A rearrangement. Epipodophyllotoxin and anthracycline exposure are correlated in these regimens and the study treats them as a combined exposure category in its headline result.

    Evidence that argues against this (1)
    • A large relative risk at the top of a dose range is not the same as a large risk at the doses actually used. The NCI CTEP monitoring programme's first analysis, in 207 patients receiving low cumulative etoposide (<1500 mg/m2), gave a 6-year treatment-related AML/MDS rate of 3.2% (95% upper bound 7.2%) - within the range previously reported for alkylator-based regimens containing no epipodophyllotoxin at all. The comparator therefore matters as much as the dose. Smith MA et al. 1993
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The NCI Cancer Therapy Evaluation Program established a prospective monitoring plan across 12 cooperative-group trials explicitly because concern over case reports of treatment-related AML 'could lead to premature abandonment of these agents'; its first analysis found a 6-year treatment-related AML/MDS rate of 3.2% (95% upper confidence bound 7.2%) among 207 patients receiving etoposide at low cumulative dose, within the range previously reported for alkylator-based regimens without epipodophyllotoxins.

    Report of the Cancer Therapy Evaluation Program monitoring plan for secondary acute myeloid leukemia following treatment with epipodophyllotoxins. · Smith MA, Rubinstein L, Cazenave L, Ungerleider RS, Maurer HM, Heyn R, et al. · J Natl Cancer Inst 1993
    PMID 8455202DOI
    What this evidence cannot show

    1993, and it reports only the low-dose stratum - by design, because analysis in each stratum was triggered by the fourth case. It is an interim result from a monitoring plan, not a completed risk estimate, and the moderate- and high-dose estimates it promised are not part of this citation. The 3.2% figure should not be quoted as the risk of etoposide in general.

    Evidence that argues against this (1)
    • The reassuring low-dose figure does not generalise upward: within a decade the SFOP case-control study found the risk rising regularly with cumulative etoposide dose and reaching a relative risk of 197 above 6 g/m2. The two results are compatible - a steep dose-response with a flat low-dose region - but citing either alone misrepresents the drug. Le Deley MC et al. 2003
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Dietary bioflavonoids cause site-specific DNA cleavage in the MLL breakpoint cluster region in primary progenitor haematopoietic cells from healthy newborns and adults, colocalising with the cleavage site induced by etoposide and doxorubicin, with topoisomerase II identified as the target and religation of both the bioflavonoid- and the etoposide-induced cleavage site demonstrated.

    Dietary bioflavonoids induce cleavage in the MLL gene and may contribute to infant leukemia. · Strick R, Strissel PL, Borgers S, Smith SL, Rowley JD · Proc Natl Acad Sci U S A 2000
    PMID 10758153DOI
    What this evidence cannot show

    Cleavage is not translocation. The study demonstrates a shared cleavage position for etoposide and bioflavonoids and reversibility of both, and the authors' extrapolation to maternal ingestion causing infant leukaemia is a proposal in their discussion, not a result. The isoform is not resolved.

    Evidence that argues against this (1)
    • The epidemiological form of the bioflavonoid extrapolation has been tested and was essentially null. In the largest case-control study (Children's Oncology Group; 240 incident infant acute leukaemias 1996-2002, 255 random-digit-dialled controls), there was little evidence of association between a maternal dietary DNA-topoisomerase-II-inhibitor index and infant leukaemia overall or by subtype; the only suggestive signal was in MLL-rearranged AML (Q4 vs Q1 OR 3.2, 95% CI 0.9-11.9, P-trend 0.10) with no signal in MLL-rearranged ALL, and maternal fruit and vegetable intake was inversely associated with risk. Spector LG et al. 2005
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In human haematopoietic stem and progenitor cells, quercetin suppresses both non-homologous end joining and homologous recombination downstream of MLL cleavage via PI3K inhibition, and co-treatment with etoposide and quercetin produces a higher frequency of MLL rearrangements than either single treatment; a tissue-engineered placental barrier established the extent of quercetin transplacental transfer.

    PMID 27813122DOI
    What this evidence cannot show

    Adult/cord human HSPCs in culture at high continuous quercetin concentrations; the placental transfer measurement is an engineered barrier model rather than a pregnancy. The TOP2 and the PI3K arms are separated pharmacologically in the paper, but the second (repair-suppressing) arm is explicitly TopoII-INDEPENDENT, so this is not simply a topoisomerase result.

    Evidence that argues against this (1)
    • The same null epidemiology applies: the COG case-control study of maternal dietary topoisomerase II inhibitor intake found no association with MLL-rearranged infant ALL, and found maternal fruit and vegetable intake inversely associated with risk - the opposite direction to the flavonoid-as-poison prediction. Spector LG et al. 2005
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Low-dose etoposide exposure ex vivo elevated DNA breakage at KMT2A-region and at TOP2-sensitive, CTCF, RAD21 and SMC3 binding sites in cells from KMT2A-rearranged AML patients in remission but not in cells from non-fusion AML patients.

    PMID 34405474DOI
    What this evidence cannot show

    Three KMT2A-rearranged remission patients against three non-fusion AML patients, adult, peripheral blood mononuclear cells cultured ex vivo. n=3 per arm. Author list is abbreviated here because the full byline was not read from the article record; see the primary record for the complete list.

    Evidence that argues against this (1)
    • Because the comparison group is other AML patients rather than healthy donors, the result establishes a difference between AML subtypes and not that KMT2A-rearranged individuals differ from the general population. The authors also report no significant difference in genome-wide median breakage between any of the patient samples, so the effect is locus-specific rather than a global instability phenotype - which cuts against reading it as a heritable fragility trait. Genes Chromosomes Cancer 2021
Connected to