The endosteal niche — the bone surface, where leukaemia-initiating cells were imaged homing and dividing symmetrically
also called osteoblastic niche · endosteum · bone-lining niche
The endosteum is the inner surface of the bone, where the marrow meets the bone wall, and it was the first place anyone proposed that blood stem cells are held. In a mouse model of this leukaemia, the cells most able to restart the disease were the ones that homed to this surface and divided there in a way that makes two copies of themselves.
The same thing, for a scientist
The endosteal niche is the bone-lining compartment, populated by osteoblasts and osteoprogenitors and perfused by small arterioles. It was the first proposed HSC niche on two independent 2003 experiments [E1][E2], and it has since been functionally demoted for normal HSC maintenance because osteoblast-derived SCF is dispensable. Its continuing relevance to this graph is leukaemic. Metabolic imaging with the genetically encoded NADH/NAD+ sensor SoNar in an MLL-AF9 murine AML model showed that **SoNar-high cells are more glycolytic, enriched for leukaemia-initiating cell frequency, and develop leukaemia much faster** than SoNar-low cells; these cells **home to and locate in the endosteal niche and maintain themselves by efficient symmetric division**, with PDK2 fine-tuning glycolysis, homing and symmetric division [E3]. Symmetric division is how a stem-cell pool expands rather than differentiates, so this is a spatial claim about where self-renewal happens. The same MLL-AF9 literature also localises leukaemia-initiating cell activity to the endosteal compartment through marrow ATP levels acting on P2X7 [E4]. The unresolved problem is that [E3] calls this compartment hypoxic while direct pO2 measurement says the endosteum is the **better**-oxygenated part of marrow.
Why the endosteum still matters after being demoted
For normal haematopoiesis, the endosteal niche lost its primacy: conditional deletion showed osteoblast SCF is dispensable while endothelial and LEPR+ perivascular SCF is required (cell-type-osteoblast). The honest summary is that the 2003 experiments showed osteoblasts can change HSC number, not that they are required to maintain it.
For leukaemia the picture is different, because the relevant question is not maintenance but self-renewal and homing. 3 is the most specific spatial claim in the MLL-AF9 literature: the leukaemia-initiating fraction, identified metabolically rather than by surface markers, preferentially homes to the endosteal niche and sustains itself there by symmetric division. That is a mechanistic statement about where the leukaemic stem-cell pool is replenished, and it is supported by an independent metabolic route. Marrow ATP is markedly elevated in AML, LICs localise preferentially to the endosteal niche where ATP is highest, and in MLL-AF9 disease ATP acts through the ligand-gated channel P2X7: deleting P2x7 impairs LIC homing and self-renewal, reduces functional LICs about 5-fold, and has no effect on normal haematopoiesis, with the signal running through calcium-flux-driven CREB phosphorylation to PHGDH and serine metabolism 4. That is the second leukaemia-selective, normal-sparing niche target in this beat, alongside endothelial ANGPTL2-LILRB2 (cell-type-sinusoidal-endothelial-cell).
A third result completes the compartment's leukaemic profile in the opposite direction: pre-leukaemic and established MLL-AF9 leukaemic cells home differently, with pre-LSCs resembling long-term HSCs and Wnt-dependent, while established LSCs resemble committed myeloid progenitors and are independent of niche-derived Wnt, because Wnt activation in MLL-AF9 cells is cell-intrinsic; osteoblast-derived Dickkopf-1 impaired normal HSC localisation but did not affect LSC homing or AML development 5. So the endosteal compartment's normal signalling levers do not control this leukaemia even where the leukaemia sits in it.
The reason this compartment matters most for an infant B-ALL graph
Everything above is myeloid. The finding that should change how this graph reads the endosteum is lymphoid and is easy to miss because it is filed under the perivascular niche's founding paper.
Ding and Morrison asked which cells supply the CXCL12 that maintains each haematopoietic compartment, using a Cxcl12-DsRed knock-in and compartment-specific deletion. Deleting Cxcl12 from endothelium depleted HSCs but not lymphoid progenitors. Deleting it from perivascular stromal cells depleted HSCs and some restricted progenitors and mobilized them into the circulation. Deleting it from osteoblasts depleted certain early lymphoid progenitors but not HSCs or myeloerythroid progenitors, and did not mobilize them. Their conclusion is stated plainly: HSCs occupy a perivascular niche and early lymphoid progenitors occupy an endosteal niche 7.
Infant KMT2A-rearranged B-ALL is a disease of early B-lymphoid progenitors — this graph's own cell-type-fetal-preprob-cd10-negative and cell-type-pro-b entries say so. If the early lymphoid progenitor's niche is endosteal and osteoblast-CXCL12-dependent, then the compartment demoted for adult HSC biology may be the compartment for this disease, and the osteoblastic finding that expanding it with PTH worsens MLL-AF9 AML acquires a lymphoid counterpart that nobody has looked for.
This is a prediction, not a finding, and it is carried as such in hyp-niche-remodelling-precedes-relapse-kmt2a. But it is the single strongest reason in this beat not to read the adult-HSC niche literature as if it described infant leukaemia.
The oxygen contradiction, stated rather than papered over
3 describes the endosteal niche as hypoxic. Direct two-photon phosphorescence lifetime measurement of pO2 in live mice reports the endosteum as less hypoxic than deep peri-sinusoidal marrow, specifically because it is perfused by small arteries 6.
These are both primary experiments in mice and they cannot both describe the same compartment correctly. Possible reconciliations include: different bones (calvarium versus long bone); the pimonidazole-type surrogate markers used in the leukaemia literature reporting metabolic state rather than oxygen tension; or the leukaemic marrow having a different oxygen distribution from the normal marrow in which pO2 was measured — the last of which is plausible and testable, since 6 itself reports that pO2 changes markedly after radiation and chemotherapy.
This graph should carry the contradiction rather than pick a side, because the resistance beat already rejected a hypoxia-protection candidate on the basis that hypoxia sensitises KMT2A-rearranged AML to venetoclax plus a menin inhibitor. See concept-marrow-oxygen-gradient.
| compartment | endosteal / osteoblastic |
|---|---|
| lymphoid significance | compartment-specific Cxcl12 deletion places EARLY LYMPHOID PROGENITORS in an endosteal, osteoblast-CXCL12-dependent niche while HSCs sit perivascularly; infant KMT2A-rearranged B-ALL arises in an early B-lymphoid progenitor, so the compartment demoted for adult HSC biology may be the relevant one for this disease |
| normal function status | demoted: osteoblast-derived SCF dispensable for HSC maintenance |
| unresolved contradiction | this compartment is described as hypoxic in the MLL-AF9 imaging literature and as the better-oxygenated compartment by direct in vivo pO2 measurement |
| literature size | the endosteal/endosteum x KMT2A PubMed intersection is 5 records; four of the five are cited in this entry |
cellular constituents
[ "osteoblast", "osteoprogenitor", "osteoclast", "small arterioles with NG2+ pericytes" ]
leukaemic relevance
{
"homing": "SoNar-high, glycolytic, LIC-enriched MLL-AF9 cells home to and locate in the endosteal niche",
"self_renewal": "maintained there by efficient symmetric division; PDK2 tunes glycolysis, homing and symmetric division",
"metabolic_lever": "marrow niche ATP is elevated in AML and highest at the endosteum; ATP acting via P2X7 drives LIC homing and self-renewal through CREB and PHGDH/serine metabolism, with P2x7 deletion cutting functional LICs about 5-fold and sparing normal haematopoiesis",
"wnt_independence": "osteoblast-derived Dkk1 impairs normal HSC localisation but not MLL-AF9 pre-LSC or LSC homing; Wnt activation in these cells is cell-intrinsic",
"pth_effect": "osteoblastic PTH-receptor activation ENHANCES MLL-AF9 AML"
}atlas refs
[
{
"label": "MLL-AF9 LIC homing and symmetric division in an endosteal niche",
"pmid": "30581117",
"title": "Metabolic Imaging Reveals a Unique Preference of Symmetric Cell Division and Homing of Leukemia-Initiating Cells in an Endosteal Niche.",
"authors": "Hao X, Gu H, Chen C, et al.",
"venue": "Cell Metab",
"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."
},
{
"label": "Differential niche and Wnt requirements during AML progression",
"pmid": "21765021",
"title": "Differential niche and Wnt requirements during acute myeloid leukemia progression.",
"authors": "Lane SW, Wang YJ, Lo Celso C, et al.",
"venue": "Blood",
"year": 2011,
"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."
}
]- Expanding the endosteal compartment with PTH, the intervention that works in CML, is predicted to make KMT2A-rearranged myeloid leukaemia worse.
- If the leukaemia-initiating fraction self-renews by symmetric division at the endosteum, a drug that merely evicts cells from the compartment without killing them could redistribute the initiating pool rather than deplete it.
- Bone-directed agents in a growing infant carry skeletal consequences that a leukaemia-directed rationale does not license, and the graph already records a menin-related bone-mass risk in this lineage.
- Is the endosteal niche hypoxic or not? The two primary measurements disagree, and the resolution matters because hypoxia SENSITISES KMT2A-rearranged AML to venetoclax plus a menin inhibitor rather than protecting it.
- Do KMT2A-rearranged B-lymphoid blasts home to the endosteum as MLL-AF9 myeloid LICs do, and as normal early lymphoid progenitors do? The normal-lymphoid result makes this the highest-value untested localisation question in this beat.
- Symmetric division at the endosteum is how the leukaemic stem-cell pool expands. Does a menin inhibitor change the symmetric-to-asymmetric division ratio, and can that be imaged in the same system?
- MLL-AF9 LSCs are independent of niche Wnt because their Wnt activation is cell-intrinsic. Are there other endosteal signals they are equally indifferent to, and is the compartment mainly a physical address rather than a signalling source for this leukaemia?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Osteoblast-specific activation of the PTH/PTHrP receptor increases osteoblast number and jagged-1 expression and supports an increase in HSC number with in vivo Notch1 activation; PTH injection into wild-type mice increases stem-cell number and markedly improves survival after bone marrow transplantation, establishing osteoblastic cells as a regulatory component of the HSC niche.
Osteoblastic cells regulate the haematopoietic stem cell niche · Calvi LM, Adams GB, Weibrecht KW, Weber JM, Olson DP, Knight MC, et al. · Nature 2003What this evidence cannot show
Mouse gain-of-function; demonstrates sufficiency not necessity. Read: abstract only.
Evidence that argues against this (1)
- The same osteoblastic PTH-receptor manipulation ENHANCED MLL-AF9-driven AML while attenuating BCR-ABL1 CML, so in the genotype this graph concerns, expanding the endosteal niche is predicted to worsen disease. Read: abstract only. Krause DS et al. 2013
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Conditional inactivation of BMP receptor type IA increases spindle-shaped N-cadherin-positive CD45-negative (SNO) osteoblastic cells in parallel with HSC number; long-term HSCs are found attached to SNO cells with N-cadherin and beta-catenin asymmetrically localised at the contact, identifying bone-lining SNO cells as a key niche component whose number sets niche size.
Identification of the haematopoietic stem cell niche and control of the niche size · Zhang J, Niu C, Ye L, Huang H, He X, Tong WG, et al. · Nature 2003What this evidence cannot show
Mouse; the SNO cell has not been prospectively isolated by later groups and the HSC N-cadherin requirement was subsequently disputed. Read: abstract only.
Evidence that argues against this (1)
- Conditional deletion showed HSC maintenance depends on SCF from endothelium and LEPR+ perivascular cells and not from osteoblasts, moving the functional niche away from the bone surface. Read: abstract only. Ding L et al. 2012
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Using the genetically encoded metabolic sensor SoNar in an MLL-AF9-induced murine AML model, SoNar-high cells are more glycolytic, enriched for leukaemia-initiating cell frequency and develop leukaemia much faster than SoNar-low cells; SoNar-high cells mainly home to and locate in the hypoxic endosteal niche and maintain their activities through efficient symmetric division; SoNar-high human leukaemia cells and primary samples have enhanced clonogenic capacity in vitro and leukaemogenesis in vivo; and PDK2 fine-tunes glycolysis, homing and symmetric division of leukaemia-initiating cells.
Metabolic Imaging Reveals a Unique Preference of Symmetric Cell Division and Homing of Leukemia-Initiating Cells in an Endosteal Niche · Hao X, Gu H, Chen C, Huang D, Zhao Y, Xie L, Zou Y, Shu HS, et al. · Cell Metabolism 2019What this evidence cannot show
Mouse MLL-AF9. SoNar reports cytosolic NADH/NAD+ ratio, so 'SoNar-high' is a redox-state gate, not a cell-surface-defined LIC population, and the enrichment is relative. The description of the endosteal niche as hypoxic is the authors' characterisation and is contradicted by direct pO2 measurement. Read: abstract only.
Evidence that argues against this (1)
- Direct in vivo two-photon pO2 measurement reports the endosteal region as LESS hypoxic than deep peri-sinusoidal marrow, because it is perfused by small arteries, with the marrow minimum of about 9.9 mmHg found peri-sinusoidally. The spatial assignment of hypoxia in this claim is therefore contested by a measurement made specifically to test it. Read: abstract only. Spencer JA et al. 2014
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
ATP levels are markedly increased in the marrow niches of mice with AML, and leukaemia-initiating cells preferentially localise to the ENDOSTEAL niche where ATP is relatively high; in an MLL-AF9-induced murine AML model ATP induces ion influx into LICs through the ligand-gated channel P2X7, and P2x7 deletion notably impairs LIC homing and self-renewal with an approximately 5-fold decrease in functional LICs and NO effect on normal haematopoiesis; ATP/P2X7 signalling enhances calcium-flux-mediated CREB phosphorylation which transactivates Phgdh to maintain serine metabolism and LIC fate, and P2X7 knockdown markedly extends survival of recipients of human AML lines or primary leukaemia cells.
Bone marrow niche ATP levels determine leukemia-initiating cell activity via P2X7 in leukemic models · He X, Wan J, Yang X, Zhang X, Huang D, Li X, Zou Y, Chen C, et al. · Journal of Clinical Investigation 2021What this evidence cannot show
Mouse MLL-AF9 plus human AML lines and primary cells in xenograft; the endosteal localisation of LICs is read from an ATP indicator rather than from direct oxygen or stromal-contact measurement. The 5-fold LIC reduction is a limiting-dilution estimate with the usual wide confidence intervals. P2X7 is broadly expressed on immune cells, so 'no effect on normal haematopoiesis' does not mean no systemic effect. Read: abstract only.
Evidence that argues against this (1)
- Niche-derived signals in the endosteal compartment do not uniformly control this leukaemia. Osteoblast-derived Dickkopf-1 dramatically impaired normal HSC localisation in marrow but did NOT affect pre-LSC or LSC homing or MLL-AF9 AML development, because Wnt activation in these leukaemic cells is cell-intrinsic. Endosteal signalling levers can be pulled without moving the leukaemia. Read: abstract only. Lane SW et al. 2011
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In a syngeneic retroviral MLL-AF9 AML model, homing and microlocalisation of pre-leukaemic stem cells most resembles long-term HSCs and depends on cell-intrinsic Wnt signalling, whereas homing of established LSCs most resembles committed myeloid progenitors and is distinct from HSCs; osteoblast-derived Dickkopf-1 dramatically impaired normal HSC localisation within marrow but did not affect pre-LSC or LSC homing or AML development, and cell-intrinsic Wnt activation was observed in human and murine AML samples, explaining independence of MLL-AF9 LSCs from niche-derived Wnt.
Differential niche and Wnt requirements during acute myeloid leukemia progression · Lane SW, Wang YJ, Lo Celso C, Ragu C, Bullinger L, Sykes SM, Ferraro F, Shterental S, et al. · Blood 2011What this evidence cannot show
Mouse retroviral MLL-AF9; homing assays are short-term readouts of localisation and do not measure long-term residence. The Dkk1 negative result is a single niche ligand. Read: abstract only.
Evidence that argues against this (1)
- Other endosteal-compartment manipulations DO change MLL-AF9 disease: osteoblast-specific PTH-receptor activation enhanced MLL-AF9 AML. So the compartment is not inert for this leukaemia even though one of its signalling ligands is; the negative Wnt result is specific to Wnt. Read: abstract only. Krause DS et al. 2013
- E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Direct in vivo measurement of local oxygen tension in live mouse bone marrow places the lowest pO2 (about 9.9 mmHg, 1.3 percent) in deep peri-sinusoidal regions and reports the endosteal region as less hypoxic because it is perfused by small, often Nestin-positive arteries, with overall marrow pO2 below 32 mmHg.
Direct measurement of local oxygen concentration in the bone marrow of live animals · Spencer JA, Ferraro F, Roussakis E, Klein A, Wu J, Runnels JM, et al. · Nature 2014What this evidence cannot show
Mouse calvarium; not leukaemic marrow, and the paper itself reports that pO2 changes markedly after radiation and chemotherapy, so the normal distribution may not be the diseased one. Read: abstract only.
Evidence that argues against this (1)
- Metabolic imaging of MLL-AF9 leukaemia-initiating cells describes the endosteal niche they home to as hypoxic, in direct contradiction. Read: abstract only. Hao X et al. 2019
- E7established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Using Cxcl12-DsRed knock-in mice and compartment-specific conditional deletion, CXCL12 is primarily expressed by perivascular stromal cells and at lower levels by endothelial cells, osteoblasts and some haematopoietic cells; deletion from haematopoietic or nestin-cre cells has little or no effect; deletion from endothelial cells depletes HSCs but not myeloerythroid or lymphoid progenitors; deletion from perivascular stromal cells depletes HSCs and certain restricted progenitors and mobilizes them into the circulation; and deletion from OSTEOBLASTS depletes certain early lymphoid progenitors but not HSCs or myeloerythroid progenitors and does not mobilize them - so HSCs occupy a perivascular niche and early lymphoid progenitors occupy an endosteal niche.
Haematopoietic stem cells and early lymphoid progenitors occupy distinct bone marrow niches · Ding L, Morrison SJ · Nature 2013What this evidence cannot show
Mouse conditional genetics in normal, adult marrow. 'Certain early lymphoid progenitors' is the authors' hedge and the abstract does not specify which populations; the deletions test CXCL12 only, not the whole niche function of the compartment. Nothing here concerns leukaemia, and the extension to infant KMT2A-rearranged B-ALL made in this entry is an inference flagged as a hypothesis. Read: abstract only.
Evidence that argues against this (1)
- The one entry in this graph that maps where KMT2A-rearranged cells actually sit records the marrow dependency as CXCR4/CXCL12-mediated stromal protection reversible by CXCR4 antagonism, based on mesenchymal stromal co-culture rather than osteoblast contact, and the follow-up work localised the protective CXCL12 in an MLL::AF9 model to Prx1+ mesenchymal cells and explicitly NOT to mature osteolineage cells. If the leukaemic cell's CXCL12 source is mesenchymal, the normal lymphoid progenitor's endosteal dependency may not carry over to its transformed counterpart. Read: abstract only. Vinado AC et al. 2022
bone-lining compartment of the niche
- ← part ofOsteoblast — the bone-building niche cell whose expansion suppresses CML and makes MLL-AF9 AML worse
principal cellular constituent
arterioles ensheathed by NG2+ pericytes are enriched in endosteal marrow
osteoblast-derived CXCL12 maintains early lymphoid progenitors, which occupy an endosteal niche
- supports →The stem-like self-renewal state — a limited stem-cell module grafted onto a committed cell
MLL-AF9 leukaemia-initiating cells home here and self-renew by symmetric division; ATP/P2X7 sustains them
direct pO2 measurement contradicts the MLL-AF9 imaging literature's hypoxic-endosteal assignment
canonical endosteal matricellular niche protein