The hypoxic, glycolytic state — the one state that appears independently in leukaemia-initiating cells, in measurable residual disease, and in KMT2A-rearranged B-ALL as a class
also called hypoxia response state · HIF-active state · glycolytic state · niche-imposed hypoxia · pseudohypoxia
The bone marrow is a low-oxygen place, and leukaemia cells switch on a survival programme in response. Three separate studies, using different methods, found that same programme switched on in the cells that start the leukaemia, in the cells left behind after chemotherapy, and in KMT2A-rearranged disease as a whole — and blocking it made chemotherapy work better.
The same thing, for a scientist
The hypoxia response is the single most repeatedly rediscovered state in this disease, and the reason it belongs in this graph as its own entry is convergence rather than any one result. (1) In MLL-rearranged B-ALL, quantitative xenotransplantation showed that **inhibiting hypoxia and glycolysis impairs leukaemia-initiating cells** and gives therapeutic benefit, while OXPHOS inhibition does the opposite [E1]. (2) In a 161,986-cell single-cell atlas of paediatric B-ALL sampled at diagnosis, residual disease and relapse, differential analysis **highlighted activation of the hypoxia pathway in residual cells, in resistant clones, and in B-ALL with MLL rearrangement** as a class, and hypoxia-pathway inhibition sensitised leukaemic cells to chemotherapy in vitro and in vivo [E2]. (3) In KMT2A-rearranged AML, hypoxia induces HDAC9, HDAC9 is indispensable for proliferation, and venetoclax plus a menin inhibitor represses HIF-1A and HDAC9 as a specific consequence of the combination [E3]. Three lineages of evidence, three assay families, one state. **The counter-case is genetic and it is serious**: conditional deletion of Hif-1a in a mouse MLL-AF9 AML accelerated disease after chemotherapy, and Hif-1a/Hif-2a double deletion accelerated leukaemogenesis while being dispensable for maintenance [E4]. So the state is real and repeatedly measured; whether HIF is the right handle on it is contested, and the contradiction falls along a myeloid-versus-lymphoid line that nobody has tested directly.
Why convergence matters more than any single result here
Each of the three supporting studies has a weakness that the other two do not share.
- Morris et al. used functional limiting-dilution xenotransplantation, the only endpoint that corresponds to relapse, but their pharmacology is a class of agents not individually named in the abstract 1.
- Zhang et al. used paired clinical sampling — diagnosis, residual, relapse in the same patients — with 161,986 single-cell transcriptomes and matched B-cell-receptor sequencing, which no mouse study can supply, but their hypoxia finding is differential expression rather than an oxygen measurement 2.
- Ling et al. supplied mechanism and a drug combination, HIF-1A to HDAC9, in KMT2A-rearranged AML specifically, but in myeloid cells and with a small-molecule combination rather than genetics 3.
A reader can dismiss any one of them. Dismissing all three requires an argument that applies to three different designs, and none is obvious.
Note also that Zhang's hypoxia signal appears in three separate contrasts within one dataset — residual cells versus diagnosis, resistant clones versus sensitive, and MLL-rearranged versus other B-ALL. That is the closest thing this literature has to an internal replication.
The genetic counter-case, and why it does not simply cancel the pharmacology
Two mouse genetics papers point the other way, both in MLL-AF9 myeloid disease.
Vukovic et al. found that deleting Hif-2a accelerated LSC development and shortened AML latency initiated by Mll-AF9, that co-deleting Hif-1a potentiated that acceleration, that established LSCs lacking either or both propagated AML with wild-type latency, and that pharmacological HIF inhibition or HIF-2a CRISPR knockout in human MLL-AF9 leukaemic cells had no effect 4. Velasco-Hernandez et al. found that conditional Hif-1a deletion produced faster disease progression after chemotherapy and warned explicitly against combining HIF inhibition with chemotherapy in AML.
Three possible reconciliations, all live:
- HIF is not the state. 'Hypoxia pathway activation' in a transcriptomic contrast can be driven by MYC, mTOR or metabolic pseudohypoxia without HIF being rate-limiting. Deleting HIF then does not remove the state.
- Initiation versus maintenance. Vukovic's own framing: HIFs suppress AML development and are dispensable once LSCs exist. A drug is given to established disease, so the initiation result does not forbid the therapeutic one.
- Lineage. Every genetic counter-result is myeloid; Morris and Zhang are lymphoid. The disease this graph most cares about — infant KMT2A::AFF1 ALL — has no HIF genetics at all.
The third is the most testable and the least tested.
Where this already sits in the graph, and what this entry adds
mechanism-niche-chemoprotection-kmt2a records the Ling result correctly, including the counter-intuitive detail that hypoxia induction sensitised KMT2A-rearranged AML to venetoclax plus MI-503 rather than protecting it, and rightly lists hypoxic protection as contradicted in that specific setting. This entry does not dispute that; it defines the state rather than the protection mechanism, and it adds the lymphoid arm (E1, E2) that the niche entry does not carry.
The distinction is worth keeping sharp, because 'hypoxia protects leukaemia from drugs' and 'the hypoxia response programme is a dependency of leukaemia-initiating cells' are different claims with different evidence and opposite therapeutic implications. The first is contradicted in KMT2A-rearranged AML. The second is supported in KMT2A-rearranged B-ALL.
| state definition | Transcriptional and metabolic activation of the hypoxia response programme, with glycolytic rather than oxidative energy production; scored by hypoxia-signature enrichment, HIF-1A/HIF-2A target expression, extracellular acidification rate, and pimonidazole or equivalent hypoxia probes |
|---|---|
| unresolved | every genetic counter-result is myeloid and every supportive functional result in this lesion is lymphoid; no HIF genetics exists in KMT2A-rearranged B-ALL |
| distinct from | mechanism-niche-chemoprotection-kmt2a, which concerns whether the niche PROTECTS against drugs; this entry concerns whether the hypoxia programme is itself a dependency |
independent observations — the numbers behind the chart
[
{
"setting": "MLL-r B-ALL leukaemia-initiating cells, limiting-dilution xenotransplantation",
"direction": "hypoxia/glycolysis inhibition impairs LICs",
"pmid": "35476984",
"title": "Hypoxic, glycolytic metabolism is a vulnerability of B-acute lymphoblastic leukemia-initiating cells.",
"authors": "Morris V, Wang D, Li Z, et al.",
"venue": "Cell Rep",
"year": 2022,
"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."
},
{
"setting": "paediatric B-ALL residual cells, resistant clones, and MLL-rearranged class, 161,986 single cells",
"direction": "hypoxia pathway activated; inhibition sensitises to chemotherapy",
"pmid": "35145224",
"title": "Elucidating minimal residual disease of paediatric B-cell acute lymphoblastic leukaemia by single-cell analysis.",
"authors": "Zhang Y, Wang S, Zhang J, et al.",
"venue": "Nat Cell Biol",
"year": 2022,
"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."
},
{
"setting": "KMT2A-rearranged AML, venetoclax + MI-503",
"direction": "combination represses HIF-1A and its target HDAC9; hypoxia induction SENSITISES to the combination",
"pmid": "38053201",
"title": "Repressing HIF-1α-induced HDAC9 contributes to the synergistic effect of venetoclax and MENIN inhibitor in KMT2Ar AML.",
"authors": "Ling Q, Zhou Y, Qian Y, et al.",
"venue": "Biomark Res",
"year": 2023,
"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."
},
{
"setting": "CNS-derived leukaemia-propagating cells, murine KMT2A-AFF1 mixed-lineage model",
"direction": "hypoxia pathway enriched in CNS versus bone marrow LPCs",
"pmid": "41371538",
"title": "Mixed-lineage leukemia cells undergo unique adaptations in the CNS niche.",
"authors": "Duguid A, Malouf C, Nitsche L, et al.",
"venue": "Exp Hematol",
"year": 2026,
"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."
}
]genetic counter observations
[
{
"model": "mouse MLL-AF9 AML, conditional Hif-1a deletion + chemotherapy",
"direction": "FASTER progression after chemotherapy",
"pmid": "30595549",
"title": "Hif-1α Deletion May Lead to Adverse Treatment Effect in a Mouse Model of MLL-AF9-Driven AML.",
"authors": "Velasco-Hernandez T, Soneji S, Hidalgo I, et al.",
"venue": "Stem Cell Reports",
"year": 2019,
"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."
},
{
"model": "mouse Mll-AF9, Hif-2a and Hif-1a/Hif-2a deletion",
"direction": "accelerated LSC development and shortened latency; dispensable for maintenance; HIF inhibition ineffective in human MLL-AF9 cells",
"pmid": "26642852",
"title": "Hif-1α and Hif-2α synergize to suppress AML development but are dispensable for disease maintenance.",
"authors": "Vukovic M, Guitart AV, Sepulveda C, et al.",
"venue": "J Exp Med",
"year": 2015,
"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."
}
]atlas refs — the numbers behind the chart
[
{
"label": "Hypoxia/glycolysis inhibition impairs MLL-r B-ALL leukaemia-initiating cells",
"pmid": "35476984",
"title": "Hypoxic, glycolytic metabolism is a vulnerability of B-acute lymphoblastic leukemia-initiating cells.",
"authors": "Morris V, Wang D, Li Z, et al.",
"venue": "Cell Rep",
"year": 2022,
"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": "Hypoxia pathway activated in MRD, resistant clones and MLL-rearranged B-ALL",
"pmid": "35145224",
"title": "Elucidating minimal residual disease of paediatric B-cell acute lymphoblastic leukaemia by single-cell analysis.",
"authors": "Zhang Y, Wang S, Zhang J, et al.",
"venue": "Nat Cell Biol",
"year": 2022,
"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": "Hif-1a deletion worsens chemotherapy outcome in MLL-AF9 AML",
"pmid": "30595549",
"title": "Hif-1α Deletion May Lead to Adverse Treatment Effect in a Mouse Model of MLL-AF9-Driven AML.",
"authors": "Velasco-Hernandez T, Soneji S, Hidalgo I, et al.",
"venue": "Stem Cell Reports",
"year": 2019,
"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."
}
]- The strongest genetic evidence in this disease says that removing the HIF pathway before chemotherapy makes AML worse, so a HIF inhibitor added to induction is a combination with a published preclinical warning against it.
- If hypoxia induction sensitises KMT2A-rearranged AML to venetoclax plus a menin inhibitor, then a hypoxia-pathway inhibitor given alongside that combination could antagonise it — a plausible negative interaction between two agents both justified by this same state.
- Hypoxia-directed agents act on normal haematopoietic stem cells, which occupy the same low-oxygen niche and depend on it for quiescence; an infant's marrow reserve is the thing least able to absorb that.
- A hypoxia signature scored from bulk or single-cell RNA is a correlate. Turning it into a target requires showing dependence, and the only dependence experiments in this lesion are genetic and negative.
- Does HIF loss in KMT2A-rearranged B-ALL behave like HIF loss in KMT2A-rearranged AML? Every genetic experiment is myeloid and every supportive functional result is lymphoid. A conditional Hif-1a/Hif-2a deletion in a KMT2A::AFF1 lymphoid model would resolve the contradiction in one experiment.
- Is the hypoxia signature in residual cells a response to low oxygen, or pseudohypoxia driven by metabolic or oncogenic signalling? These predict different drugs: a HIF inhibitor for the first, a metabolic or mTOR-directed agent for the second.
- Does menin inhibition raise or lower hypoxia-pathway activity in KMT2A-rearranged cells? One study reports that venetoclax plus a menin inhibitor represses HIF-1A, but menin inhibition alone has never been profiled this way, and the answer determines whether a hypoxia-directed agent is a partner or an antagonist.
- Is the hypoxic state a marrow phenomenon or does it follow the cell? CNS-derived leukaemia-propagating cells in a murine KMT2A-AFF1 model are enriched for hypoxia pathways relative to marrow-derived cells, which suggests the CNS sanctuary imposes a MORE hypoxic state than the marrow does — the opposite of the usual assumption.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In MLL-rearranged B-ALL studied by single-cell transcriptomics and quantitative xenotransplantation, inhibiting hypoxia and glycolysis impaired leukaemia-initiating cells and provided therapeutic benefit in xenotransplantation systems, whereas inhibiting oxidative phosphorylation blunted blast proliferation but promoted LIC emergence.
Hypoxic, glycolytic metabolism is a vulnerability of B-acute lymphoblastic leukemia-initiating cells · Morris V, Wang D, Li Z, Marion W, Hughes T, Sousa P, et al., Rowe RG · Cell Rep 2022What this evidence cannot show
Murine xenograft niche, so the oxygen environment imposing the state is not a human marrow. The abstract does not name the hypoxia- and glycolysis-directed agents or separate their contributions, so 'inhibiting hypoxia and glycolysis' is a compound intervention. Infant versus older KMT2A-rearranged B-ALL is not distinguished and fusion partner is not reported. Read: abstract only.
Evidence that argues against this (1)
- Genetics in the myeloid form of the same lesion points the opposite way. Deletion of Hif-2a accelerated leukaemic stem cell development and shortened AML latency initiated by Mll-AF9 and its effectors Meis1 and Hoxa9; co-deletion of Hif-1a potentiated that acceleration; established LSCs lacking Hif-2a or both HIFs propagated AML with wild-type latency; and pharmacological HIF inhibition or HIF-2a CRISPR knockout in human MLL-AF9 leukaemic cells had no impact on their function. If HIF loss is if anything leukaemogenic and dispensable for maintenance, then a therapeutic benefit from 'inhibiting hypoxia' is either lineage-specific or is not acting through HIF. Read: abstract only. Vukovic M et al. 2015
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In 161,986 single-cell transcriptomes combined with B-cell-receptor sequencing from paediatric B-ALL sampled at diagnosis, residual disease and relapse, leukaemic cells at relapse shifted to poorly differentiated states while changes in residual cells were more complex, and differential analysis highlighted activation of the hypoxia pathway in residual cells, in resistant clones and in B-ALL with MLL rearrangement; in vitro and in vivo models showed that inhibition of the hypoxia pathway sensitised leukaemic cells to chemotherapy.
Elucidating minimal residual disease of paediatric B-cell acute lymphoblastic leukaemia by single-cell analysis · Zhang Y, Wang S, Zhang J, Liu C, Li X, Guo W, et al., Cheng T · Nat Cell Biol 2022What this evidence cannot show
Paediatric B-ALL broadly; the MLL-rearranged group is a subgroup of the cohort and its size, age distribution and fusion partners are not given in the abstract, so this is not an infant KMT2A-rearranged study. A hypoxia signature in transcriptomes is not a measurement of oxygen tension and can arise from pseudohypoxic signalling. Residual-disease sampling is by definition of rare cells, with the selection and dissociation biases that implies. Read: abstract only.
Evidence that argues against this (1)
- Conditional genetic deletion of Hif-1a in a mouse MLL-AF9 AML model combined with a standard chemotherapy regimen produced FASTER disease progression after chemotherapy, and the authors state their results 'challenge the general notion of cancer stem cell sensitization by inhibition of the HIF pathway' and 'warrant caution when applying HIF inhibition in combination with chemotherapy in AML'. That is precisely the combination E2's pharmacology proposes, tested genetically in the myeloid arm of the same lesion, with the opposite outcome. Read: abstract only. Velasco-Hernandez T et al. 2019
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In KMT2A-rearranged AML, venetoclax plus a menin inhibitor was synergistic in cell lines, primary cells and a MOLM13 xenograft; RNA sequencing identified HDAC9 as repressed specifically by the combination rather than by either monotherapy; HDAC9 was indispensable for KMT2Ar-AML proliferation; hypoxia induced HDAC9 expression; the combination inhibited the hypoxia pathway, especially HIF-1A, and the HDAC9 downregulation was partially dependent on HIF-1A repression; and hypoxia induction sensitised KMT2Ar-AML to venetoclax plus MI-503.
Repressing HIF-1alpha-induced HDAC9 contributes to the synergistic effect of venetoclax and MENIN inhibitor in KMT2Ar AML · Ling Q, Zhou Y, Qian Y, Qian J, Zhang Y, Wang J, et al., Zhang X · Biomark Res 2023What this evidence cannot show
Myeloid only; no lymphoid arm, and infant ALL is the population this graph is built around. The hypoxia link is established by an induction experiment plus a transcriptional readout rather than by measuring oxygen tension in vivo. 'MENIN inhibitor' is MI-503, a tool compound, not a clinical agent, so the combination is not the clinical revumenib-plus-venetoclax question. The finding that hypoxia induction SENSITISES rather than protects is the opposite of the usual niche-protection expectation and has not been independently reproduced. Read: abstract only.
Evidence that argues against this (1)
- If hypoxia sensitises KMT2A-rearranged AML to venetoclax plus a menin inhibitor, then a strategy of inhibiting the hypoxia pathway to impair leukaemia-initiating cells would, in this setting, be expected to ANTAGONISE that same drug combination. The two results are not formally contradictory - one concerns LIC frequency in B-ALL and the other drug synergy in AML - but they cannot both be turned into the same clinical instruction, and no study has measured hypoxia-pathway inhibition and menin inhibition together in either lineage. Read: abstract only. Morris V et al. 2022
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Conditional Hif-1a deletion in a mouse model of MLL-AF9-driven AML, combined with a standard chemotherapy regimen, produced faster disease progression after chemotherapy, leading the authors to challenge the notion that HIF pathway inhibition sensitises cancer stem cells and to warn against combining HIF inhibition with chemotherapy in AML.
Hif-1alpha Deletion May Lead to Adverse Treatment Effect in a Mouse Model of MLL-AF9-Driven AML · Velasco-Hernandez T, Soneji S, Hidalgo I, Erlandsson E, Cammenga J, Bryder D · Stem Cell Reports 2019What this evidence cannot show
Mouse MLL-AF9 AML, a single model and a single genotype; complete conditional deletion from the outset is not the same perturbation as partial pharmacological inhibition of an established disease, and Hif-1a deletion does not remove HIF-2a, which the double-knockout work shows matters. No lymphoid arm. Read: abstract only.
Evidence that argues against this (1)
- A genetic deletion and a drug are not interchangeable here in a way that specifically weakens this counter-argument. HIF-1-alpha has HIF-independent and non-transcriptional functions and its complete constitutive absence during leukaemogenesis allows compensation by HIF-2-alpha; pharmacological inhibition in established disease is a partial, acute perturbation of a different system. The single-cell paired clinical data showing hypoxia-pathway activation in residual cells of actual patients is a measurement in humans that no mouse deletion can overturn, even though it establishes association rather than dependence. Read: abstract only. Zhang Y et al. 2022
- expressed in →KMT2A-rearranged acute lymphoblastic leukaemia
hypoxia pathway activated in MLL-rearranged B-ALL as a class and in residual cells and resistant clones across 161,986 single cells
the two halves of one metabolic choice; the same study found opposite LIC effects from inhibiting each
distinct claims kept apart: niche chemoprotection concerns whether hypoxia shields cells from drugs (contradicted in KMT2Ar AML), this entry concerns whether the hypoxia programme is itself a dependency
the marrow oxygen gradient is the physiological source of the state; CNS-derived leukaemia-propagating cells are enriched for hypoxia pathways relative to marrow-derived ones
venetoclax plus MI-503 represses HIF-1A and its target HDAC9 as a combination-specific effect in KMT2Ar AML
two named definitions of the surviving compartment that have never been scored against each other: glucocorticoid-primed and quiescent at diagnosis, versus hypoxia-activated at residual disease
hypoxia pathways are enriched in CNS-derived relative to matched marrow-derived cells, inverting the usual framing of the marrow as the low-oxygen compartment
reduced glucose consumption and lactate production on loss of the direct fusion target PTBP1
glycolysis inhibition, not OXPHOS inhibition, impairs MLL-r B-ALL initiating cells
venetoclax-resistant AML LSC reprogramme to glycolysis and the shift is sustained by de novo guanine nucleotides; IMPDH1/2 inhibition blocks the shift, reactivates the ribosome-biogenesis checkpoint, stabilises TP53 and represses MYC
WEAK and labelled so: eaf1/eaf2 deletion reduces whole-organism hypoxia tolerance in zebrafish through reduced red cell production (PMID 37002435), which is not the same quantity as a leukaemic cell's metabolic adaptation in a marrow niche
hypoxia induces HDAC9 through HIF-1A, and hypoxic induction sensitises KMT2A-rearranged AML cells to venetoclax plus MI-503