MLL-AF9 leukaemia destroys the nerves that run its own niche — a mechanism class this graph had nothing on

The brief that dispatched this beat named PMID 25017722 as a lead to verify rather than repeat. It was fetched and read, and it says what the brief said it says, with two details worth keeping exact.

First, the direction: AML causes the neuropathy. The title's word is 'leukemia-induced'. This is not a pre-existing host susceptibility that permits leukaemia; the disease develops, and as it does, sympathetic fibres are disrupted and Nestin+ niche cells lose quiescence 1.

Second, the consequence is a swap, not a loss. The MSPC compartment expands, but the expansion is of cells primed for osteoblastic differentiation, and it happens at the expense of NG2+ periarteriolar cells — the population independently shown to maintain HSC quiescence (cell-type-ng2-periarteriolar-cell). So the remodelled niche is not merely a damaged normal niche; it is a niche whose composition has been shifted away from the one that holds normal HSCs still 1.

Third, the receptor is beta-2. The paper tests both and reports that the leukaemogenesis-promoting adrenergic signal goes through beta-2 and not beta-3 on stromal cells 1. That matters because normal circadian niche physiology runs through beta-3 (cell-type-bone-marrow-sympathetic-nerve). The normal and the malignant adrenergic circuits use different receptors, and an intervention designed from the physiology would target the wrong one.

A 2025 letter from a different group extends marrow neuropathy beyond MLL-AF9 to NRAS-G12D-driven pre-leukaemic myelopoiesis and to p53-deletion lymphoma/lymphoblastic leukaemia, documents severe ultrastructural damage to sympathetic fibres by electron microscopy in NRAS-G12D mice and in AML patient marrow, and shows that chemically reinforcing the neuropathy promotes myeloid bias, reduces HSPC numbers, reduces mesenchymal stromal cell numbers and produces morphologically abnormal marrow vessels 2.

The sentence in that letter with clinical consequences is its last: chemotherapy induces neuropathy of the HSC niche, and chemotherapy is first-line for exactly the diseases in which the neuropathy is pathogenic. If the niche damage caused by the leukaemia is pro-leukaemic, and the treatment causes the same damage, then induction chemotherapy is plausibly deepening the very lesion that favours the residual disease. This is the same shape as the megakaryocyte finding already in the graph, where chemotherapy inverts a quiescence-enforcing niche signal into an expansion signal (cell-type-megakaryocyte). Two independent niche compartments, same structural claim: treatment remodels the niche that decides what residual disease does next.

Note what is and is not established. That chemotherapy induces niche neuropathy is asserted in 2 with a citation to prior work; this entry records it as reported in a letter rather than as a result of 2's own experiments, and it has not been demonstrated in KMT2A-rearranged infant leukaemia in any species.

A separate MLL-AF9 study mapped the stromal side directly. AML infiltration expands Ebf2-positive MSPCs with reduced Cxcl12 and enhanced production of more differentiated mesenchymal progenitors; in vivo fate-mapping shows Ebf2+ cells participate in building the leukaemic niche; and Ebf2+ cell deletion accelerates AML, implying the native mesenchymal niche suppresses the disease before it is remodelled. AML dysregulated Angptl1, Cxcl12, Kitl, Il6, Nov and Spp1 in marrow MSPCs, partly before the massive expansion of AML cells. Primary patient AML cells induced similar alterations in xenografted mice 3.

Two things follow. The remodelling is early, so it is a candidate for a pre-relapse readout rather than only a late consequence of tumour burden. And 'niche' is not a synonym for 'protection': deleting a niche population here made the leukaemia worse, which is the direct opposite of the protective-niche framing.

This is a mechanism class, not a single finding: the leukaemia actively alters the tissue it lives in, and the alteration is functional rather than incidental. The graph previously held nothing on it. mechanism-kmt2a-r-tissue-localization records where KMT2A-rearranged cells sit and that stroma protects them; it does not record that the cells rebuild the stroma.

The evidence base is entirely MLL-AF9 murine AML, which is a myeloid model of a disease that in infants is overwhelmingly B-lymphoid. Nothing here has been shown for KMT2A::AFF1 or in infant ALL, and the rejection ledger (ana-marrow-niche-rejected-candidates) records that the sympathetic/adrenergic intersection with the whole KMT2A literature is two records, both cited here.