Bone-marrow adipocyte — the fat cell that suppresses normal blood formation and feeds leukaemia a chemoresistance signal

Anyone who has looked at a marrow biopsy after induction chemotherapy has seen fatty replacement: the cellular marrow is gone and adipocytes fill the space. The question Naveiras and colleagues asked was whether the fat is a consequence or a cause. They compared mouse skeletal regions that differ naturally in adiposity — the adipocyte-rich tail vertebrae against the adipocyte-free thoracic vertebrae — and found HSCs and short-term progenitors reduced in the fatty ones. Then they removed the variable: in genetically fatless A-ZIP/F1 mice, and in mice given a PPAR-gamma inhibitor to block adipogenesis, marrow engraftment after irradiation was accelerated 1.

So marrow fat actively restrains haematopoiesis, and the clinical inference the authors draw — that antagonising marrow adipogenesis might speed engraftment after transplant — is still, nearly two decades later, untested in children.

The 2026 finding is the first adipocyte result with direct KMT2A relevance. Leptin is abundant in marrow adipose tissue. In newly diagnosed AML, high marrow leptin and high blast LEPR associate with poor cytarabine clearance and shorter survival. Causally: leptin blunts cytarabine cytotoxicity in human lines, primary blasts and two syngeneic mouse models, one of which is MLL-AF9; the LEPR antagonist Allo-aca restores chemosensitivity without changing baseline leukaemia growth — the signature of a genuine resistance mechanism rather than a growth effect. The mechanism is metabolic: LEPR/JAK2/STAT3 raises complex I activity and oxidative phosphorylation, the resulting mitochondrial ROS surge triggers a compensatory antioxidant programme, and that programme shields blasts from drug-induced oxidative damage 2.

This is now the third independent niche compartment in this beat whose protection of KMT2A-rearranged leukaemia is metabolic and redox-mediated: Prx1+ mesenchymal CXCL12 controlling LSC redox balance (cell-type-car-cell), nestin+ BMSCs transferring OXPHOS capacity and glutathione (cell-type-nestin-msc), and now adipocyte leptin driving complex I and an antioxidant response. Three different cells, three different ligands, one convergent mechanism. That convergence is the single most generalisable finding of this beat, and it makes the antioxidant/redox axis rather than the adhesion axis the place to look for a niche-directed intervention in this genotype.

It also sets up an uncomfortable prediction. If niche protection works by buffering oxidative stress, then agents that reduce oxidative stress — including some supportive-care measures — would be expected to phenocopy the niche and protect the leukaemia. Nobody appears to have tested that.