The lipid-dependent state — ACSL4 and polyunsaturated fatty acid utilisation in KMT2A-rearranged AML, against ELOVL1 and saturated elongation in KMT2A::AFF1 infant ALL
also called lipid metabolic state · ACSL4 dependence · PUFA utilisation · fatty acid requirement of KMT2A-rearranged blasts
Leukaemia cells need particular kinds of fat, not just energy. In the myeloid form of KMT2A-rearranged leukaemia the requirement is for an enzyme that activates long polyunsaturated fats; in the infant lymphoid form it is a different enzyme that lengthens saturated ones. Same founding lesion, two different fat requirements.
The same thing, for a scientist
Two functionally validated lipid dependencies exist in KMT2A-rearranged leukaemia and they are not the same dependency. In **AML**, large-scale CRISPR-Cas9 screening data identified **ACSL4** as a selective vulnerability of KMT2A-rearranged AML; CRISPR interference and shRNA knockdown confirmed that ACSL4 loss impairs growth of KMT2Ar but not non-KMT2Ar AML cells, reduces colony formation in patient-derived KMT2Ar AML cells and murine MLL-AF9 cells, and delays leukaemia onset in vivo; multi-omics showed depletion of polyunsaturated lipid species with compensatory activation of lipid metabolic pathways, and **exogenous PUFA supplementation rescued the growth defect**, tying the dependency to defective PUFA utilisation [E1]. In **infant KMT2A::AFF1 ALL**, the lipid handle that emerged from the conserved fetal signature is **ELOVL1**, a very-long-chain fatty acid elongase acting on saturated and monounsaturated substrates, identified as critical for survival of MLL-AF4-positive ALL cells [E2]. ACSL4 and ELOVL1 sit on opposite sides of the saturated/polyunsaturated divide, and ACSL4 in particular is the canonical ferroptosis-sensitising enzyme — its loss makes cells **resistant** to ferroptosis in most systems, which means the KMT2Ar-AML dependency and the ferroptosis literature pull in opposite directions and the paper's own rescue experiment is the thing that adjudicates. This is a state entry rather than a target entry: what is being described is a lipid composition a cell must maintain, and two different enzymes maintain it in two lineages.
The AML result, and the one control that makes it a lipid result
Selective-vulnerability calls from screen data are common and usually soft. What raises this one is the rescue: supplementing exogenous polyunsaturated fatty acids restored growth after ACSL4 loss 1. That converts 'ACSL4 is required' into 'PUFA utilisation is required and ACSL4 is how these cells do it', which is a statement about the state rather than about the enzyme, and it predicts that a second route to PUFA incorporation would bypass the dependency.
The validation ladder is also unusually complete for a screen-derived hit: CRISPRi and shRNA (two orthogonal knockdowns), patient-derived KMT2Ar AML cells, murine MLL-AF9 cells, and in vivo latency 1. The comparator is non-KMT2Ar AML, which is the myeloid comparator this project keeps having to insist on.
The authors also derive a KMT2Ar-ACSL4 dependency signature (KRADS12) that correlates with KMT2Ar status and predicts poor survival in AML 1. A signature that tracks the genotype it was derived from is close to circular as a prognostic claim, and that part should be read as the weakest element.
The ferroptosis inversion, stated plainly because it is easy to get backwards
ACSL4 is the enzyme most associated with sensitivity to ferroptosis: it esterifies arachidonic and adrenic acid into membrane phospholipids, creating the peroxidisable substrate that ferroptosis needs. In most cancer contexts, losing ACSL4 protects cells from ferroptotic death.
So a KMT2A-rearranged AML in which ACSL4 loss impairs growth is behaving in a direction the ferroptosis literature does not predict, and a reader who imports the usual ACSL4 story will conclude that KMT2Ar AML should be ferroptosis-resistant and therefore that ACSL4 is the wrong target. The published result says the opposite about growth.
The reconciliation the paper offers is that the requirement is for PUFA utilisation — incorporation into membranes as a structural or signalling need — rather than for peroxidation. That is coherent, and it is also a prediction: a ferroptosis inducer should be less effective in KMT2Ar AML if ACSL4 activity is high and required, or more effective, depending on which of the two roles dominates. No published experiment separates them in this genotype. This graph already holds mechanism-aldh3a2-ferroptosis material in the vulnerabilities folder and the interaction has not been tested.
Why this is two states, not one
ELOVL1 elongates saturated and monounsaturated very-long-chain fatty acids; it is not a PUFA enzyme. It emerged from a conserved fetal transcriptional signature that persists in infant ALL blasts, alongside PLK1, and was functionally validated as critical for survival of MLL-AF4-positive ALL cells 2.
So the lymphoid dependency is on a fetal-programme lipid enzyme acting on saturated substrates, and the myeloid dependency is on a PUFA-activating enzyme with no stated developmental framing. They could both be true, and the honest statement is that 'the lipid state of KMT2A-rearranged leukaemia' does not exist as a single object.
This is the third time in this beat that a metabolic property has split along the myeloid/lymphoid line rather than along the fusion: OXPHOS versus glycolysis, hypoxia genetics versus hypoxia pharmacology, and now PUFA versus saturated elongation. That recurrence is itself the finding and is recorded as hyp-cell-state-splits-by-lineage-not-by-fusion.
| state definition | A lipid-compositional requirement: the blast must maintain a particular membrane fatty acid profile, scored by lipidomics, by rescue with exogenous fatty acid species, and by dependency on the enzymes that supply them |
|---|---|
| rescue experiment | exogenous polyunsaturated fatty acid supplementation restored growth after ACSL4 loss, which is what makes the dependency a statement about lipid state rather than about the enzyme |
| ferroptosis tension | ACSL4 loss normally confers ferroptosis RESISTANCE; a genotype requiring ACSL4 for growth therefore has an untested and sign-ambiguous relationship to ferroptosis inducers |
| derived signature | KRADS12, a KMT2Ar-ACSL4 dependency signature correlating with KMT2Ar status and poor AML survival; derived from the genotype it correlates with |
| compartment note | CNS-adapted leukaemia-propagating cells in a murine KMT2A-AFF1 model upregulate lipid and fatty acid metabolic programmes, but those cells also repopulate systemically less well |
enzymes
[
{
"gene": "ACSL4",
"substrate_class": "polyunsaturated fatty acids (activation to acyl-CoA for membrane incorporation)",
"lineage": "KMT2A-rearranged AML",
"validation": "CRISPRi, shRNA, patient cells, murine MLL-AF9, in vivo latency, PUFA rescue",
"pmid": "41712378",
"title": "ACSL4-associated lipid metabolism is a distinct therapeutic vulnerability in KMT2A-rearranged acute myeloid leukemia.",
"authors": "Schäfer S, Rahimian E, Schmitz-Hübsch L, et al.",
"venue": "Cell Rep",
"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."
},
{
"gene": "ELOVL1",
"substrate_class": "saturated and monounsaturated very-long-chain fatty acids (elongation)",
"lineage": "MLL-AF4-positive infant ALL",
"validation": "functional validation from a conserved fetal signature",
"pmid": "34706236",
"title": "Defining the fetal origin of MLL-AF4 infant leukemia highlights specific fatty acid requirements.",
"authors": "Symeonidou V, Jakobczyk H, Bashanfer S, et al.",
"venue": "Cell Rep",
"year": 2021,
"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
[
{
"label": "ACSL4 as a selective KMT2Ar AML vulnerability rescued by exogenous PUFA",
"pmid": "41712378",
"title": "ACSL4-associated lipid metabolism is a distinct therapeutic vulnerability in KMT2A-rearranged acute myeloid leukemia.",
"authors": "Schäfer S, Rahimian E, Schmitz-Hübsch L, et al.",
"venue": "Cell Rep",
"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."
},
{
"label": "ELOVL1 as a fetal-signature fatty acid requirement of MLL-AF4 infant ALL",
"pmid": "34706236",
"title": "Defining the fetal origin of MLL-AF4 infant leukemia highlights specific fatty acid requirements.",
"authors": "Symeonidou V, Jakobczyk H, Bashanfer S, et al.",
"venue": "Cell Rep",
"year": 2021,
"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."
}
]- ACSL4 loss confers ferroptosis resistance in most systems, so an ACSL4 inhibitor could simultaneously impair KMT2A-rearranged AML growth and protect it from every ferroptosis-inducing agent given alongside — an antagonism that would only show up in a combination experiment nobody has run.
- Both enzymes are used by normal tissue: ELOVL1 loss in humans causes an ichthyotic neurological disorder, and very-long-chain fatty acid elongation is required for myelin. Targeting it in an infant, whose nervous system is still myelinating, is a specific rather than generic toxicity concern.
- Dietary and parenteral lipid composition varies enormously in infants receiving intensive care, and a dependency rescuable by exogenous PUFA is a dependency that nutrition could silently reverse.
- A prognostic signature derived from the genotype it predicts will validate on any cohort in which that genotype is prognostic, so KRADS12 should not be used as independent evidence that the lipid state matters.
- Is ACSL4 a dependency in KMT2A-rearranged ALL as it is in KMT2A-rearranged AML? The screen data that generated the AML call contains lymphoid KMT2A-rearranged lines, so this is a re-analysis rather than a new experiment, and it has not been reported.
- Does a ferroptosis inducer work better or worse in KMT2A-rearranged AML? An active, required ACSL4 predicts a peroxidisable membrane and therefore sensitivity; a required ACSL4 also means the enzyme cannot be removed to protect normal cells. The sign has never been measured in this genotype.
- Is the lipid programme in CNS-adapted leukaemia-propagating cells a cause of their impaired systemic repopulation or a consequence of the lipid-rich environment? These have opposite implications for whether CNS-directed lipid metabolism is a target or a decoy.
- Does ELOVL1 dependence survive outside the fetal context? It was identified through a fetal signature in infant disease; whether older KMT2A-rearranged B-ALL retains it is the test of whether this is a fetal-programme dependency or a fusion dependency.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Analysis of large-scale CRISPR-Cas9 screening data identified ACSL4 as a selective vulnerability in KMT2A-rearranged AML; CRISPR interference and shRNA knockdown confirmed that ACSL4 loss impairs growth of KMT2Ar but not non-KMT2Ar AML cells, reduces colony formation in cells from patients with KMT2Ar AML and in murine MLL-AF9 cells, and delays leukaemia onset in vivo in MLL-AF9 mice; transcriptomics, proteomics and lipidomics showed depletion of polyunsaturated lipid species with compensatory activation of lipid metabolic pathways, and supplementation with exogenous polyunsaturated fatty acids rescued the growth defect.
ACSL4-associated lipid metabolism is a distinct therapeutic vulnerability in KMT2A-rearranged acute myeloid leukemia · Schafer S, Rahimian E, Schmitz-Hubsch L, Shaikh M, Brilloff S, Kufrin V, et al., Wurm AA · Cell Rep 2026What this evidence cannot show
Myeloid only; no lymphoid arm and therefore nothing about infant KMT2A::AFF1 ALL. The starting point is a dependency-screen re-analysis, which carries the lineage-confound problem this project has documented in its own DepMap work: KMT2A-rearranged AML lines are monocytic-enriched, and a non-KMT2Ar AML comparator does not by itself control for monocytic identity. The KRADS12 prognostic signature is derived from the genotype it then correlates with, which is close to circular. No pharmacological ACSL4 inhibitor is reported, so tractability is unestablished. Read: abstract only.
Evidence that argues against this (1)
- ACSL4 is the canonical ferroptosis-SENSITISING enzyme: it esterifies polyunsaturated fatty acids into membrane phospholipids to create the peroxidation substrate, and its loss protects cells from ferroptosis across many systems. A genotype in which ACSL4 loss impairs growth is therefore a genotype in which the standard reading of ACSL4 biology would predict the opposite therapeutic conclusion, and AML is independently reported to defend itself against oxidative and ferroptotic death through ALDH3A2. Whether a ferroptosis inducer would be more or less effective in KMT2A-rearranged AML, given a required and presumably active ACSL4, has not been tested and the two literatures predict different answers. Read: abstract only. Yusuf RZ et al. 2020
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
RNA sequencing of fetal, neonatal and adult haematopoietic stem and progenitor cells from human and mouse defined a conserved fetal transcriptional signature of prominently proliferative and oncogenic character that persists in infant ALL blasts, and functional validation from that signature identified genes critical for survival of MLL-AF4-positive ALL cells, of which PLK1 and ELOVL1 are highlighted, the latter identifying altered fatty acid metabolism as a feature of infant ALL.
Defining the fetal origin of MLL-AF4 infant leukemia highlights specific fatty acid requirements · Symeonidou V, Jakobczyk H, Bashanfer S, Malouf C, Fotopoulou F, Kotecha RS, et al., Ottersbach K · Cell Rep 2021What this evidence cannot show
ELOVL1 emerges from a signature-then-validate pipeline, so it is one hit among several and the abstract does not state the effect size or whether an in vivo arm was run for ELOVL1 specifically. ELOVL1 acts on saturated and monounsaturated very-long-chain fatty acids, so this is a different lipid requirement from the PUFA requirement in E1, not a confirmation of it. Being a fetal-programme gene, ELOVL1 is also expressed in normal fetal and infant tissue, which is a toxicity consideration in the age group concerned. Read: abstract only. This citation was re-fetched from PubMed for this entry rather than copied from cell-state-fetal-program-retention, which also cites it.
Evidence that argues against this (1)
- The two lipid dependencies are enzymatically incompatible as a single story. ELOVL1 elongates saturated and monounsaturated very-long-chain fatty acids; ACSL4 activates polyunsaturated fatty acids for membrane incorporation, and the AML dependency was rescued specifically by exogenous PUFA supplementation. A reader who merges them into 'KMT2A-rearranged leukaemia depends on fatty acid metabolism' obtains a claim that neither paper supports and that would misdirect a combination strategy, because agents that raise cellular PUFA content would be predicted to rescue the AML phenotype while having no defined relation to the lymphoid one. Read: abstract only. Schafer S et al. 2026
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Transcriptomic profiling of matched CNS-derived and bone-marrow-derived leukaemia-propagating cells from a fully murine immunocompetent KMT2A-AFF1 mixed-lineage infant leukaemia model revealed enrichment in the CNS of pathways involved in hypoxia, lipid and cholesterol homeostasis and inflammatory signalling, and the LPC subsets that successfully adapted to the CNS niche upregulated lipid and fatty acid metabolic programmes.
Mixed-lineage leukemia cells undergo unique adaptations in the CNS niche · Duguid A, Malouf C, Nitsche L, Halsey C, Ottersbach K · Exp Hematol 2026What this evidence cannot show
A fully murine model, so the lipid environment is a mouse CNS; the CNS is lipid-rich in a way marrow is not, so a lipid-programme enrichment in CNS-derived cells may report the tissue rather than a leukaemic adaptation. Transcriptomic enrichment, not lipidomics, and no functional test of lipid dependence in the CNS compartment. The species of fatty acid is not specified, so this cannot be assigned to the PUFA or the saturated side of the split above. Read: abstract only.
Evidence that argues against this (1)
- The same study weakens the usual reading of the CNS as a protected reservoir: CNS-derived LPCs transplanted to model relapse showed REDUCED systemic repopulation capacity, with many recipients showing stable long-term engraftment without overt leukaemia, a phenomenon not seen with bone-marrow-derived transplants. If the CNS niche's metabolic reprogramming impairs rather than preserves leukaemogenic capacity, then a lipid adaptation found there is not automatically a therapeutic target, and could be a marker of attenuation. Read: abstract only (same paper). Duguid A et al. 2026
ACSL4 loss impairs growth of KMT2Ar but not non-KMT2Ar AML cells; rescued by exogenous polyunsaturated fatty acids
the lymphoid lipid requirement is ELOVL1, a fetal-signature saturated/monounsaturated elongase — a different enzyme class from the myeloid ACSL4/PUFA requirement
third instance in this beat of a metabolic property splitting by lineage rather than by fusion
LPC subsets that successfully adapted to the CNS upregulated lipid and fatty acid metabolic programmes
lipid axis: ACSL4/PUFA in myeloid against ELOVL1/saturated elongation in lymphoid
the myeloid enzyme inside the lipid-dependent state entry
the lymphoid enzyme inside the lipid-dependent state entry
CREBBP-mutated B-ALL differs in lipid composition and oxidative-stress response, predisposing it to ferroptosis; small-molecule CREBBP inhibition reproduces the state in wild-type cells
sibling metabolic state; the pattern gene-mat2a said it wanted to copy and could not
NRAS's address is an acylation state - palmitoylated at the plasma membrane, depalmitoylated at the Golgi - never measured in a leukaemic cell
LXR-alpha's active state is a sterol-loaded state of the cell, not a differentiation state
ELOVL1 sits on the saturated/monounsaturated VLCFA side of the lipid divide, ACSL4 on the polyunsaturated side; subnuclear localisation for ELOVL1 is NOT APPLICABLE - a seven-pass ER membrane enzyme (Q89 2026-09-21)
LXR-beta is an oxysterol-activated receptor, so its active state is a lipid state of the cell rather than a differentiation state - recorded identically to gene-nr1h3 because nothing distinguishes them here (Q89 2026-09-21)