🧬 Aging Wiki

Endothelial Notch activity promotes angiogenesis and osteogenesis in bone

11 min read

Endothelial Notch activity promotes angiogenesis and osteogenesis in bone

Ramasamy SK*, Kusumbe AP*, Wang L, Adams RH (* equal contribution)
Nature 507:376-380, 2014

TL;DR

This paper is the mechanistic companion to Kusumbe et al. 2014 (Nature 507:323-328; DOI 10.1038/nature13145). Where the companion paper characterizes type-H (CD31^hi/Emcn^hi) bone vessels and shows they decline with aging, Ramasamy et al. dissect why type-H vessels drive osteogenesis: endothelial Notch signaling drives type-H angiogenesis in bone and is obligately required for coupled osteogenesis. The coupling mechanism is angiocrine — Notch-active endothelial cells secrete Noggin, a BMP antagonist that prevents premature BMP-driven differentiation of Runx2+ osteoprogenitors, maintaining the progenitor pool available for proper bone formation. Conditional deletion of Notch signaling in endothelial cells (via Rbpj KO) reduces type-H vessel abundance, collapses the Runx2+ osteoprogenitor pool through premature Osterix+ differentiation, and reduces bone mass; gain-of-function expands both vessels and progenitors; and recombinant Noggin administration fully rescues the bone phenotype of endothelial Notch loss-of-function mutants. The study positions endothelial-secreted Noggin as the molecular bridge between vascular Notch activity and osteoprogenitor maintenance.

Background and experimental system

Bone angiogenesis has tissue-specific features not shared with angiogenesis in other organs. The companion paper (Kusumbe 2014) identified a specialized subtype — type-H vessels (CD31^hi/Emcn^hi endothelium) — concentrated in bone metaphysis and endosteum, where they associate with Osterix+ and Runx2+ osteoprogenitors. Type-H vessels decline dramatically with aging in mouse bone (nearly absent by 57-70 weeks of age vs. abundant at 4 weeks).

Ramasamy et al. asked: what maintains type-H endothelial identity, and how does it couple to osteoprogenitor maintenance? The study uses inducible endothelial-specific Cre (Cdh5(PAC)-CreERT2) — which drives tamoxifen-inducible recombination specifically in endothelial cells — to delete or activate Notch signaling components in vivo. Tamoxifen was administered intraperitoneally (500 µg/day) at postnatal days P10-P14, with skeletal analysis at P28.

Type-H vessel morphology

Type-H metaphyseal vessels form morphologically distinct structures:

  • Straight tubules (5–10 µm diameter) extending from metaphyseal arteries
  • Tubular arches (15–20 µm diameter) — blind-ended, bulb-shaped protrusions connecting the tubules
  • High proliferation: EdU+ (S-phase) labeling is concentrated predominantly in columns and, less prominently, in arches, identifying these as the active angiogenic front in bone
  • Angiogenic gene expression: elevated Neuropilin 1, Plexin D1, Unc5b, and FLT4/VEGFR3 relative to sinusoidal (type-L) endothelium

Notch target gene transcripts and pathway activity were elevated in type-H versus type-L bone endothelium, suggesting endogenous Notch signaling is preferentially active in type-H cells.

Loss-of-function: endothelial Rbpj deletion (Rbpj iΔEC)

Genetic model: Cdh5(PAC)-CreERT2 × Rbpj^lox/lox^ — RBPJ is the nuclear transcriptional effector required for all canonical Notch signal transduction; its deletion effectively silences the entire canonical Notch pathway in endothelial cells.

Main experiments: n=6 mice from 4 independent litters per group.

Vascular phenotype

  • Decreased CD31^hi/Emcn^hi (type-H) vessel abundance
  • Reduced distal metaphyseal arches per unit metaphysis length
  • Disorganized filopodia extension at the angiogenic front
  • Reduced VEGFR1, VEGFR2, VEGFR3 immunostaining
  • Elevated soluble VEGFR1 (sFlt1) mRNA — a VEGF decoy that inhibits angiogenesis

Bone and osteoprogenitor phenotype

  • Shortened femurs — grossly reduced long bone growth
  • Reduced BV/TV (bone volume/total volume), trabecular number, and trabecular thickness by micro-CT
  • Reduced bone formation rates by calcein double labeling (7-day pulse interval) — confirming reduced active bone deposition, not just structural remnant
  • Enlarged, irregular growth plates with loss of Sox9 expression in the maturation/hypertrophy zones — indicating chondrocyte maturation defects
  • Increased Osterix+ cells — more cells at the later committed osteoblast stage
  • Reduced Runx2+ early osteoprogenitors — depletion of the upstream progenitor pool
  • Reduced Bglap (osteocalcin) and Ibsp (bone sialoprotein) mRNA — reduced mature osteoblast gene expression
  • Primary mesenchymal cells from mutant bone mineralized prematurely in culture (10-14 days vs. 21-28 days for controls) — confirming a cell-intrinsic acceleration of osteoblastic differentiation

Molecular: Noggin downregulation

Sorted bone endothelial cells from Rbpj iΔEC mutants showed significantly reduced Nog mRNA compared to controls. Nog encodes Noggin, a secreted BMP antagonist. This finding positioned endothelial-secreted Noggin as a candidate angiocrine factor coupling Notch activity to osteoprogenitor state.

Gain-of-function: endothelial Fbxw7 deletion (Fbxw7 iΔEC)

Genetic model: Cdh5(PAC)-CreERT2 × Fbxw7^lox/lox^ — FBXW7 is an E3 ubiquitin ligase subunit that promotes proteasomal degradation of the Notch intracellular domain (NICD); its deletion stabilizes NICD and produces Notch gain-of-function (sustained transcriptional activation).

Main experiments: n=6 mice from 4 independent litters per group.

Vascular and bone phenotype

  • Increased CD31^hi/Emcn^hi vessel abundance and expanded metaphyseal arches
  • Enhanced total and proliferating endothelial cell numbers
  • Formation of numerous small arterioles within the metaphysis
  • Enhanced femur extension and bone growth
  • More abundant Runx2+ early osteoprogenitors
  • Reduced Osterix+ cell abundance (expanded progenitor pool, delayed premature commitment)
  • Reduced growth plate and maturation/hypertrophy zone dimensions

Molecular: Noggin upregulation

Sorted bone endothelial cells from Fbxw7 iΔEC gain-of-function animals showed a 40-fold upregulation of Nog mRNA 1. VEGFR transcripts (all three receptors: Vegfr1, Vegfr2, Vegfr3) were increased; soluble sFlt1 was decreased; Vegfa (VEGF-A in growth plate chondrocytes) was elevated. The Nog upregulation was the most dramatic molecular change and the strongest candidate for the osteoprogenitor effect.

DLL4 is the essential Notch ligand in bone endothelium

To identify which Notch ligand drives endothelial type-H identity, the paper generated three further conditional knockouts using Cdh5(PAC)-CreERT2:

Conditional KOLigand deletedBone endothelial phenotype
Dll4 iΔECDLL4Reproduced the Rbpj iΔEC phenotype — reduced type-H vessels, bone formation
Dll1 iΔECDLL1No significant vascular or bone phenotype
Jag1 iΔECJagged-1No significant vascular or bone phenotype

Conclusion: DLL4 is the endogenous activating Notch ligand responsible for maintaining type-H endothelial identity and Notch-driven angiocrine osteogenesis in bone. This specificity contrasts with other vascular contexts (e.g., retinal angiogenesis) where the same ligand has been studied.

EC-autonomy control: An osteoblast-specific Rbpj deletion line (Tg(Col1a1-creERT2)6.1.ICS × Rbpj^lox/lox^; denoted Rbpj iOB) showed normal vasculature and normal osteoprogenitor numbers, confirming that the Notch-to-osteogenesis coupling requires Notch signaling specifically in endothelial cells — not in osteoblasts — and is therefore an endothelial-autonomous and angiocrine phenomenon.

Noggin rescue: establishing the angiocrine coupling mechanism

To directly test whether endothelial-secreted Noggin mediates the Notch-to-osteogenesis link, the authors administered recombinant Noggin protein to Rbpj iΔEC mutants during the critical growth window.

Rescue protocol: Recombinant Noggin (rNoggin), 500 µg/kg/day, administered daily from P15 to P27; n=6 from 4 independent litters.

Outcomes restored by rNoggin in Rbpj iΔEC mutants:

  • Metaphysis organization (restored metaphyseal architecture)
  • Bone formation rates (calcein double labeling, normalized to controls)
  • Normalized osteoprogenitor numbers (Runx2+ and Osterix+ balance restored)
  • Growth plate size and Sox9 expression in maturation/hypertrophy zones
  • Chondrocyte VEGF-A expression
  • Vascular organization of type-H vessels

In vitro Noggin effect: Recombinant Noggin (500 ng/ml) added to primary mesenchymal cell cultures inhibited osteoblastic differentiation in vitro (Fig. 4c). The paper does not specify the cell source (mutant vs. control bone) for this experiment, and does not report a specific timeline restoration for Noggin-treated cells. The 10–14 day vs. 21–28 day mineralization timeline described in the LOF section refers to the separate in vitro comparison of Rbpj iΔEC mutant bone cells vs. wild-type controls — an experiment not involving exogenous Noggin treatment. The in vitro inhibition of differentiation by Noggin is mechanistically consistent with the angiocrine model (endothelial Noggin suppresses premature BMP-driven commitment), but the paper reports it as inhibition of differentiation rather than specifying a timeline restoration.

The Noggin paradox

Noggin is canonically a BMP antagonist — it was originally characterized as an inhibitor of BMP-2, BMP-4, BMP-7, and BMP-14 signaling (via direct high-affinity binding to BMPs, preventing receptor engagement). BMP signaling promotes osteoblast commitment and differentiation. The intuition that “less BMP signaling = less bone” makes Noggin an apparently counter-productive molecule for bone formation.

The resolution: in the context of type-H endothelial–osteoprogenitor coupling, unchecked BMP signaling drives osteoprogenitors to commit and differentiate too quickly — rushing past the Runx2+ progenitor stage to terminal Osterix+ differentiation. This prematurely depletes the progenitor pool without producing a full complement of properly spatiotemporally instructed mature osteoblasts. Endothelial Notch-derived Noggin brakes this premature commitment, maintaining a sufficient pool of Runx2+ progenitors for subsequent proper osteogenesis.

ConditionEC NotchEC NogginBMP in progenitorsRunx2+ poolBone formation
Wild-typeActive (DLL4-driven)HighModulatedMaintainedNormal
Rbpj iΔECAbsentLowUncheckedPrematurely depletedReduced
Fbxw7 iΔECConstitutive40× elevatedSuppressedExpandedEnhanced
Rbpj iΔEC + rNogginAbsentExogenousModulatedRestoredRestored

Relationship to aging

This paper does not study aged animals; all experiments were performed on postnatal day 28 (P28) mice. The aging dimension of type-H vessels — their dramatic decline from ~4 weeks to ~57-70 weeks of age — is the finding of the companion paper 2. Together, the two papers establish a mechanistic model for the bone-vascular aging axis:

  1. Kusumbe 2014 showed type-H vessels decline with age, and that pharmacological restoration via deferoxamine mesylate (DFM, a PHD inhibitor stabilising HIF-1α) can restore them
  2. Ramasamy 2014 (this paper) showed that endothelial Notch-DLL4 signaling is required to maintain type-H identity, and that the coupling to osteoprogenitors is mediated by endothelial Noggin secretion

Implication for bone aging: If endothelial Notch activity declines with age (as part of the general reduction in type-H vessel properties), then:

  • Less EC-derived Noggin
  • Unchecked BMP-driven premature osteoprogenitor differentiation
  • Depleted Runx2+ pool and reduced bone formation rates — the histomorphometric signature of aged bone

This places the endothelial Notch–DLL4–Noggin axis as a prospective contributor to age-related osteoprogenitor decline. gap/needs-human-replication — whether endothelial Notch activity and Noggin secretion decline specifically in aged human bone type-H endothelium has not been established.

DimensionStatus
Pathway conserved in humans?partial — Notch/DLL4 signaling in vascular endothelium is highly conserved; type-H bone vessel equivalent in humans is not fully characterised
Phenotype conserved in humans?partial — endosteal vascular decline with aging is observed in human bone histology; Noggin-osteoprogenitor coupling not tested in humans
Replicated in humans?no — all experiments are in mouse (P28 postnatal); companion aging data (Kusumbe 2014) is also mouse

Limitations and gaps

  • Postnatal developmental window only. All experiments analyze postnatal day 28 mice. Whether the Notch-DLL4-Noggin-osteoprogenitor axis is equally important in adult homeostasis and aging remains untested in this paper. gap/needs-replication
  • No aging experiments in this paper. The relevance to aging is inferred by combining this mechanistic result with the companion paper’s aging phenotype data, not directly demonstrated here. gap/needs-human-replication
  • Noggin mechanism is indirect. The in vitro Noggin rescue confirms a cell-intrinsic premature differentiation effect, but does not fully resolve whether the in vivo rescue is osteoprogenitor-cell-intrinsic or involves secondary vascular or chondrocyte effects. The rescue of vascular organization (type-H vessels restored by rNoggin) complicates causal interpretation — vascular and osteogenic effects may be linked bidirectionally.
  • Notch ligand source. The paper shows DLL4 is the essential ligand but does not fully characterize which cell presents DLL4 to bone endothelial Notch receptors (endothelial-to-endothelial lateral inhibition vs. perivascular or stromal DLL4 presentation). gap/no-mechanism
  • BMP specificity of rescue. Which BMP isoform(s) drive premature osteoprogenitor differentiation in the Rbpj iΔEC bone is not identified; Noggin antagonizes BMP-2, BMP-4, BMP-7, and BMP-14 with varying affinity. gap/dose-response-unclear
  • No human equivalent study. Type-H vessel biology in human bone is an active area of research without well-established mechanistic characterisation equivalent to the mouse data. gap/needs-human-replication

Significance

972 citations (FWCI 20.04; 100th citation percentile per archive record). Together with Kusumbe 2014 (companion), this paper launched the “type-H bone vessel” paradigm — one of the most influential frameworks for understanding bone-vascular coupling in development and aging. Key contributions:

  1. Endothelial Notch as an osteogenic organizer: The field had expected Notch to be primarily inhibitory in vasculature (as in retinal angiogenesis, where Notch limits tip-cell formation). This paper showed that in bone, Notch instead promotes the specialized angiogenic program required for osteogenesis — a tissue-specific inversion of the canonical role.
  2. Angiocrine coupling via Noggin: Identifies a specific angiocrine factor (endothelial-secreted Noggin) as the molecular bridge between vascular state and osteoprogenitor pool maintenance. This extends the angiocrine concept (vascular niche instructing tissue stem cells) to the bone microenvironment.
  3. DLL4 specificity in bone: Establishes DLL4 — not Jagged-1 or DLL1 — as the bone-relevant Notch ligand, narrowing therapeutic target space.
  4. Aging framework: Combined with Kusumbe 2014, establishes type-H vessel loss as a mechanistic driver of age-related osteoprogenitor decline, and endothelial Notch as a prospective target for bone-aging interventions.

Cross-references

  • bone — tissue context; type-H vessels and the bone-vascular axis
  • osteoblasts — osteoprogenitor pool maintained by endothelial Notch-Noggin; companion Kusumbe 2014 page cited there
  • endothelial-cells — type-H endothelial identity; Notch-DLL4 signaling
  • bone-marrow — perivascular niche for osteoprogenitors
  • notch-pathway — DLL4/RBPJ/NICD canonical pathway; DLL4 in vascular contexts
  • stem-cell-exhaustion — osteoprogenitor pool maintenance as aging-relevant mechanism

Footnotes

Footnotes

  1. ramasamy-2014-endothelial-notch-bone · doi:10.1038/nature13146 · Ramasamy SK, Kusumbe AP, Wang L, Adams RH · Nature 2014;507(7492):376-380 · in-vivo · n=6 per group from 4 independent litters (main experiments); multiple additional cohorts for ligand-specificity and rescue arms · model: postnatal (P28) C57BL/6 mouse; endothelial-specific Cre (Cdh5(PAC)-CreERT2); tamoxifen P10-P14 · green OA (PMC4943529); 972 citations; FWCI 20.04 ↩

  2. doi:10.1038/nature13145 · Kusumbe AP, Ramasamy SK, Adams RH · Nature 2014;507(7492):323-328 · in-vivo (mouse) · companion paper characterizing type-H (CD31^hi/Emcn^hi) bone vessels, their age-dependent decline, and pharmacological restoration by deferoxamine mesylate (DFM); cited in osteoblasts § Angiogenesis-osteogenesis coupling · PMC4943525 · PMID: 24646994 ↩

Graph View

Backlinks