Oncofusion

start / graph / cell-state-cns-niche-adapted

Cell statemoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q30-cell-state

The CNS-adapted state — leukaemia-propagating cells reprogrammed by the central nervous system niche, and worse at restarting systemic disease afterwards

also called CNS niche-adapted state · leptomeningeal leukaemia state · CNS-derived leukaemia-propagating cell · sanctuary-site state

KMT2A-rearranged infant leukaemia often spreads to the fluid around the brain, and the cells that settle there change: they switch on programmes for handling fat and cholesterol, for low oxygen, and for inflammation. Unexpectedly, when those cells were put back into mice they were worse at causing leukaemia again, not better.

The same thing, for a scientist

The CNS is the canonical sanctuary site in this disease and it has now been characterised as a state rather than a location. Using a **fully murine immunocompetent KMT2A-AFF1 mixed-lineage infant leukaemia model** — a design point worth noting, because nearly every other CNS experiment in this field is a xenograft in an immunodeficient host — Duguid et al. transplanted previously defined bone-marrow-derived leukaemia-propagating cells and showed they engraft the CNS **unequally**, with lineage-negative cKit+ common lymphoid progenitor cells consistently underrepresented in the niche. Transcriptomic profiling of **matched CNS- and BM-derived LPCs** showed enrichment in the CNS of pathways for **hypoxia, lipid and cholesterol homeostasis, and inflammatory signalling**, with the LPC subsets that successfully adapted upregulating lipid and fatty acid metabolic programmes, and CNS-derived LPCs showing increased expression of genes involved in **T-cell immune modulation**, suggesting a more immunosuppressive environment [E1]. The result that inverts the usual picture: **transplants of CNS-derived LPCs, modelling relapse, showed reduced systemic repopulation capacity**, with many recipients exhibiting stable long-term engraftment without developing overt leukaemia — a phenomenon not seen with BM-derived transplants [E1]. So the sanctuary imposes lasting reprogramming that **impairs** the cell's ability to re-establish systemic disease, which is the opposite of the assumption that CNS residence preserves a more dangerous cell.

Go deeper
Three pathways, and each already has an entry in this graph

The CNS-enriched programmes are hypoxia, lipid/cholesterol homeostasis and inflammatory signalling 1 — which is, unexpectedly, a list of three of this beat's other entries.

  • Hypoxia — cell-state-hypoxic-glycolytic. Notable because it means the CNS is transcriptionally more hypoxic-appearing than the marrow, reversing the usual framing of the marrow as the low-oxygen compartment.
  • Lipid and fatty acid — cell-state-lipid-pufa-dependence. The CNS is the most lipid-rich tissue in the body, so this could report the environment rather than an adaptation; the paper's own emphasis that the successfully adapting subsets upregulate these programmes is what argues for adaptation.
  • Inflammatory signalling and T-cell immune modulation — cell-state-innate-immune-lamp5 and cell-state-interferon-herv-mhc2. An immunocompetent model is the only kind that could have detected the immune-modulation arm.

That three independent state axes all move together in one compartment suggests the CNS state is a coordinated programme rather than three separate responses, and it is testable: perturbing one axis should move the others if they are coupled.

Reduced repopulation capacity, and how not to over-read it

The finding that CNS-derived LPCs repopulate systemically worse is the paper's most interesting and most easily misused result.

What it supports: CNS residence is not simply preservation of a pristine reservoir. Something about the niche changes the cell durably, and in a direction that costs it systemic leukaemogenic capacity.

What it does not support: the conclusion that CNS disease matters less. Leptomeningeal disease causes morbidity in its own right, CNS relapse is a recognised clinical event in this disease, and 'stable long-term engraftment without overt leukaemia' in a mouse is a description of persistence, which in a child with decades ahead is not reassuring. A cell that engrafts stably and does not proliferate is a description of the dormant compartment this beat keeps finding, not of a defeated one.

The alternative explanation the abstract does not exclude: CNS-derived cells may be harder to recover intact, fewer in number, or altered by the isolation procedure, any of which would reduce measured repopulation without any biological adaptation. Limiting-dilution quantification would separate a frequency effect from a per-cell capacity effect and is not described.

The therapeutic counterweight

Whatever the CNS state does to leukaemogenic capacity, the compartment is now reachable. Bispecific CD19-CD133 CAR-iNKT cells — off-the-shelf allogeneic invariant NKT cells, deployable without acute GVHD risk — eradicated medullary and leptomeningeal KMT2A-rearranged leukaemia and induced sustained remissions without discernible haematologic toxicity in relevant models, outperforming monospecific counterparts and dual bispecific CAR T cells, with the advantage mediated by CAR-dependent upregulation of NKG2D and engagement of its ligands on KMT2Ar-ALL cells so that killing persists even when CD133 or CD19 is downregulated 2. PROM1/CD133 is itself described there as a direct target of KMT2A-fusion oncoproteins.

That is a second antigen axis layered on the CAR axis, which is structurally the same specificity argument as the non-physiological antigen pairs in cell-state-early-lymphocyte-precursor: escape requires losing two independent things rather than one.

Facts and figures
state definitionThe transcriptional and metabolic configuration adopted by leukaemia-propagating cells resident in the central nervous system niche, distinguished from matched bone-marrow-derived cells of the same leukaemia
modelfully murine immunocompetent KMT2A-AFF1 mixed-lineage infant leukaemia; matched CNS- and BM-derived leukaemia-propagating cells
adaptive subset featureLPC subsets that successfully adapted to the CNS upregulated lipid and fatty acid metabolic programmes
unequal engraftmentlineage-negative cKit+ common lymphoid progenitor cells were consistently underrepresented in the CNS niche
counter intuitive resultCNS-derived LPCs had REDUCED systemic repopulation capacity, with many recipients showing stable long-term engraftment without overt leukaemia — not seen with BM-derived transplants
cns enriched pathways
[
  "hypoxia",
  "lipid and cholesterol homeostasis",
  "inflammatory signalling",
  "T-cell immune modulation"
]
therapeutic reach
{
  "platform": "off-the-shelf allogeneic bispecific CD19-CD133 CAR-iNKT cells",
  "result": "eradication of medullary and leptomeningeal KMT2A-rearranged leukaemia in models; second antigen axis via CAR-dependent NKG2D upregulation",
  "pmid": "40898981",
  "title": "Off-the-shelf dual CAR-iNKT cell immunotherapy eradicates medullary and leptomeningeal high-risk KMT2A-rearranged leukemia.",
  "authors": "Ren H, Elliott N, Lye B, et al.",
  "venue": "Blood",
  "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."
}
cross links within this beat
[
  "cell-state-hypoxic-glycolytic",
  "cell-state-lipid-pufa-dependence",
  "cell-state-innate-immune-lamp5"
]
atlas refs
[
  {
    "label": "CNS niche imposes lasting metabolic and functional reprogramming on KMT2A-AFF1 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."
  },
  {
    "label": "Bispecific CD19-CD133 CAR-iNKT eradication of leptomeningeal KMT2Ar-ALL in models",
    "pmid": "40898981",
    "title": "Off-the-shelf dual CAR-iNKT cell immunotherapy eradicates medullary and leptomeningeal high-risk KMT2A-rearranged leukemia.",
    "authors": "Ren H, Elliott N, Lye B, et al.",
    "venue": "Blood",
    "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."
  }
]
How this could bite us back
  • CNS-derived cells that engraft stably without causing overt leukaemia are a description of persistence, not of cure; treating the finding as reassurance would be the wrong reading in a child with decades of follow-up.
  • A CD19-plus-CD133 dual CAR targets two antigens of the same lymphoid phenotype, so a lineage switch — this disease's documented escape from CD19-directed therapy — could remove both at once, defeating the redundancy the design relies on.
  • An allogeneic off-the-shelf iNKT product avoids acute GVHD by platform design, but its persistence and the durability of its effect in an infant with a recovering immune system are unaddressed by the models used.
  • If the CNS imposes a distinct metabolic state, then systemic drug exposure measured in blood tells you little about activity in the compartment where relapse is most feared, and no menin inhibitor has published CNS pharmacokinetics in children.
Still open
  1. Is the reduced systemic repopulation of CNS-derived cells a loss of capacity or an entry into dormancy? Limiting-dilution transplantation plus cycle-status measurement on the recovered cells would distinguish them, and the two have opposite clinical meanings.
  2. Why does the CNS score as more hypoxic than the marrow in matched cells? If that is real rather than a signature artefact, it inverts a standing assumption about where the low-oxygen compartment in this disease is, and it makes the CNS the place a hypoxia-directed agent would act.
  3. Are the three co-enriched axes — hypoxia, lipid, inflammation — coupled or independent? Perturbing one in CNS-resident cells and reading the other two is a single experiment that would say whether the CNS state is one programme or three.
  4. Do menin inhibitors reach and act in the leptomeningeal compartment? This graph holds a patient story of CNS relapse on revumenib; CNS pharmacokinetics for this drug class in children are not established, and a CNS-adapted state with its own metabolism would be the place a partially penetrant drug fails first.
  5. Can a CD19-CD133 dual CAR be escaped by lineage switch, which would remove both antigens together? The correlated-loss scenario is the one dual targeting does not cover and it is this disease's signature escape.
Every source behind this entry (2)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In a fully murine immunocompetent model of KMT2A-AFF1-positive mixed-lineage infant leukaemia, previously defined bone-marrow-derived leukaemia-propagating cells engrafted the CNS unequally, with lineage-negative cKit+ common lymphoid progenitor cells consistently underrepresented in the niche; transcriptomic profiling of matched CNS- and BM-derived LPCs revealed CNS enrichment of pathways involved in hypoxia, lipid and cholesterol homeostasis and inflammatory signalling, with successfully adapted LPC subsets upregulating lipid and fatty acid metabolic programmes and CNS-derived LPCs showing increased expression of genes involved in T-cell immune modulation; and transplants of CNS-derived LPCs modelling relapse showed reduced systemic repopulation capacity, with many recipients exhibiting stable long-term engraftment without developing overt leukaemia, a phenomenon not observed with bone-marrow-derived transplants.

    Mixed-lineage leukemia cells undergo unique adaptations in the CNS niche · Duguid A, Malouf C, Nitsche L, Halsey C, Ottersbach K · Exp Hematol 2026
    PMID 41371538DOI
    What this evidence cannot show

    A fully murine model, which is a strength for immune competence and a limitation for human relevance; a mouse CNS niche is small, its meningeal compartment differs from the human one, and the leukaemia is engineered rather than arising spontaneously. The pathway findings are transcriptomic enrichment, not metabolic or oxygen measurements, and a lipid-programme enrichment in cells resident in the most lipid-rich tissue in the body may report the environment. Reduced systemic repopulation is reported without limiting-dilution quantification in the abstract, so a per-cell capacity difference is not separated from a cell-number or recovery-efficiency difference. Read: abstract only. Note: this graph's marrow-niche beat recorded this PMID with the author field marked 'not verified'; the authors given here were fetched from PubMed on 2026-09-20.

    Evidence that argues against this (1)
    • 'Reduced systemic repopulation with stable long-term engraftment and no overt leukaemia' is also a description of dormancy, which this graph documents elsewhere as reversible and niche-imposed: dye-dilution-defined dormant ALL cells in xenografts regained proliferation and drug sensitivity once removed from the in vivo environment. On that reading the CNS-derived cells are not less dangerous but are occupying a quiescent state that the transplant assay scores as failure, and the clinically relevant question - whether they can be reactivated - is not addressed. The assay's endpoint and the biology's endpoint differ. Read: abstract only (same paper), with the dormancy comparison drawn from cell-state-quiescence in this graph. Duguid A et al. 2026
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Allogeneic invariant natural killer T cells equipped with CD19- and/or CD133-targeting CARs were tested against KMT2A-rearranged ALL, a disease characterised by frequent CNS involvement, poor treatment response, early relapse and lineage switching; bispecific CD19-CD133 CAR-iNKT cells had more potent antileukaemia activity than monospecific counterparts and than dual bispecific CAR T cells, effectively targeting both CAR-antigen-high and -low leukaemia, eradicated medullary and leptomeningeal leukaemia and induced sustained remissions without discernible haematologic toxicity; mechanistically the advantage over CAR T cells was mediated by pronounced CAR-dependent and CAR-antigen-dependent upregulation of the innate activating receptor NKG2D on the CAR-iNKT cells and engagement of its ligands on KMT2Ar-ALL cells, ensuring targeting even with downregulation of CD133 or CD19; and PROM1/CD133 is a direct target of KMT2A-fusion oncoproteins expressed on leukaemic cells.

    Off-the-shelf dual CAR-iNKT cell immunotherapy eradicates medullary and leptomeningeal high-risk KMT2A-rearranged leukemia · Ren H, Elliott N, Lye B, Sharif Shohan MU, Cross JW, Field L, et al., Karadimitris A · Blood 2026
    PMID 40898981DOI
    What this evidence cannot show

    Preclinical in vitro and in vivo models only; 'eradicates' and 'sustained remissions' are mouse endpoints. Xenograft models of a human leukaemia require immunodeficient hosts, so the immune context in which an allogeneic iNKT product would actually operate is absent, and 'without risk of acute graft-versus-host disease' is a property claimed for the iNKT platform rather than demonstrated in these experiments. The NKG2D-ligand mechanism depends on ligand expression by the leukaemia, which is known to be modulable and is a standard escape route. No lineage-switch challenge is described, in a disease whose defining escape from CD19-directed therapy is lineage switch. Read: abstract only.

    Evidence that argues against this (1)
    • The escape route this platform is least protected against is the one this disease is known for. CD19 and CD133 are both antigens of a lymphoid-phenotype blast, and KMT2A-rearranged leukaemia escapes CD19-directed therapy by switching lineage rather than by losing the antigen in isolation; a myeloid-switched blast could lose both CAR antigens together, which is precisely the correlated failure that dual targeting is supposed to prevent. Profiling of MLL-rearranged samples shows lineage-switching cells reduce oncoprotein levels and activate a GMP-like programme, and one such switch occurred under a menin inhibitor with the drug fully on target. Read: abstract only. Janssens DH et al. 2024
Connected to