Lymphatic vessels in bone support regeneration after injury

TL;DR

Using 3D light-sheet imaging of immunolabeled whole bones, Biswas et al. overturn the long-standing dogma that bone is devoid of lymphatic vessels. They demonstrate that LYVE1+/PROX1+/VEGFR3+ lymphatic vessels penetrate the bone marrow of multiple murine bones and are confirmed by immunostaining in human bone biopsies. After genotoxic injury (irradiation, 5-fluorouracil), these vessels rapidly expand via an IL-6-dependent lymphangiogenesis program; the lymphatic endothelial cells (LECs) then upregulate CXCL12, which acts on perivascular MYH11+/CXCR4+ progenitor cells to support bone formation and hematopoietic stem cell (HSC) regeneration. In aged mice, bone LECs accumulate senescence markers (p16, p21), fail to proliferate after injury, and do not expand MYH11+ progenitors — transplantation of young LECs into aged recipients restores this response and increases bone mass and HSC reconstituting activity.

Design

Establishing bone lymphatics: imaging and validation

• Imaging platform: ultrafast immunolabeling and optical clearing of intact calcified tissues (3.5-day protocol), followed by light-sheet microscopy; 3D reconstruction across multiple whole murine bones (sternum, vertebral column, costal bones, femur, calvarium, hip bones).
• Markers for lymphatic vessels: LYVE1, PROX1 (prospero-related homeobox 1), podoplanin, VEGFR3/FLT4. Evans blue vital dye confirmed functional lymphatic uptake.
• Human validation: immunostaining and ELISA on human bone biopsies from the Oxford Musculoskeletal Biobank (Human Femoral Condyle / Tibial Plateau; male and female donors, age range 56–87 years). PROX1 ELISA: n=12 (Figure 2J; bone vs. cartilage comparison). 3D LYVE1 light-sheet imaging of cleared human bone marrow biopsies: 2 donors shown (Figure 2K). Additional bone marrow biopsies from the Dazhou Central Hospital lung cancer patient cohort (n=2 replicates; age range 61–77 years) used for scRNA-seq.
• Single-cell validation: the authors generated their own scRNA-seq data from human bone marrow biopsies (lung cancer patients, Dazhou Central Hospital; two replicates; 2,305 cells after filtering, of which 6 were classified as LECs by PECAM1/LYVE1/FLT4/PROX1 expression) to confirm LEC transcriptomic signatures in human bone.

Injury-response experiments

• Genotoxic stress models: whole-body irradiation; 5-fluorouracil (5-FU) chemotherapy.
• Lymphangiogenesis inhibition: VEGFR3 inhibitor SAR131675 administered during the post-injury window.
• IL-6 requirement: IL-6 knockout mice ± recombinant IL-6 rescue; IL-18, IL-27, and IL-7 blockade as controls.
• n: 5–7 mice per group for imaging and flow cytometry experiments; 6 mice for transplantation/reconstitution assays.

Mechanistic dissection: CXCL12-MYH11 axis

• LEC-specific CXCL12 deletion: conditional knockout to isolate the LEC secretory contribution.
• MYH11+ cell depletion: ablation model to test necessity for osteogenesis and hematopoietic niche maintenance.
• Readouts: bone marrow cellularity, LSK (Lin−Sca-1+c-Kit+) cell numbers, HSC frequency, secondary transplantation reconstituting activity; bone mass by micro-CT (μ-CT); osteoblast markers (collagen I, osterix, osteocalcin).

Aging experiments

• Subjects: young adult mice (8–12 weeks; C57BL/6) versus aged mice (>55 weeks generally per STAR Methods; 61–92 weeks specifically for the LEC transplantation experiment).
• Perturbation: irradiation applied to both age groups; comparison of lymphatic expansion response.
• Rescue: transplantation of young LECs into aged recipients (n=5); comparison arm of aged LECs into aged recipients. HSC reconstituting activity assayed at n=6.
• Senescence marker profiling: qPCR on purified LECs for p16 (Cdkn2a), p21 (Cdkn1a), Ki67 (proliferation), and VEGFR3 (lymphatic identity) (Figure 6L, n=6; two-tailed unpaired t-tests). Note: the paper body text contains a typo (“p27”) corrected here per Figure 6L caption.

Key results

Lymphatic vessels are present inside bone

LYVE1+/PROX1+/podoplanin+/VEGFR3+ lymphatic vessels are confirmed within the bone marrow of all murine bones examined, with a vessel diameter of approximately 19 μm. Evans blue uptake confirmed functional lymphatic transport. Human bone biopsies showed LYVE1-positive lymphatic vessels by immunostaining (3D light-sheet imaging of 2 donors, Figure 2K) and PROX1 quantification by ELISA (n=12 samples from Oxford Musculoskeletal Biobank, Figure 2J), extending the finding beyond rodents. An additional human bone marrow scRNA-seq dataset (cancer patients, 2 replicates, 2,305 cells total) identified 6 LECs expressing PECAM1, LYVE1, FLT4, and PROX1 (Figure 2L). This overturns the prior anatomical consensus that lymphatics are confined to periosteal surfaces. lyve1 and prox1 are the canonical lymphatic endothelial cell identity markers used to define these vessels.

Injury triggers IL-6-dependent lymphangiogenesis

Following whole-body irradiation, bone lymphatic vessel density peaked at day 15 post-irradiation and returned to baseline by day 55. 5-FU chemotherapy induced a comparable expansion. Pharmacological inhibition of VEGFR3 with SAR131675 blocked the post-injury lymphangiogenesis, confirming VEGFR3/FLT4 as the operative receptor. IL-6 knockout mice failed to expand lymphatics after irradiation; recombinant IL-6 administration rescued the expansion in IL-6 KO mice. Blockade of IL-18, IL-27, and IL-7 did not interfere, identifying IL-6 as the non-redundant upstream cytokine signal for bone LEC proliferation after genotoxic stress.

LECs secrete CXCL12 to support perivascular progenitors

Purified bone LECs showed significant upregulation of cxcl12 mRNA following irradiation. LEC-specific deletion of CXCL12 resulted in reduced bone marrow cellularity, decreased LSK cell numbers, and decreased HSC frequency (n=6). CXCL12-EGFP reporter mice confirmed positive signal in bone lymphatic vessels in irradiated bones. This defines a lymphangiocrine signaling axis distinct from the established perivascular (endothelial sinusoid-derived) CXCL12 niche.

MYH11+ perivascular progenitors are the downstream cellular target

CXCL12 from LECs acts on MYH11+ pericyte-like perivascular progenitors that express CXCR4 (the cognate CXCL12 receptor). During irradiation, MYH11+ cells expand specifically in bone (not other organs) and are located near endosteal regions. MYH11+ cells differentiate into osteoblasts, chondrocytes, and adipocytes. Depletion of MYH11+ cells resulted in: reduced osteoprogenitor and osteoblast marker expression, decreased bone formation rates, significantly decreased bone mass on μ-CT, and reduced bone marrow cellularity with decreased HSC frequency. Exogenous CXCL12 administration expanded the MYH11+ population without genotoxic stress, and LEC-specific CXCL12 deletion decreased MYH11+/CXCR4+ cell numbers, confirming the LEC → CXCL12 → MYH11+ progenitor axis.

Aging-specific findings

In aged mice, bone LECs show a senescent transcriptional profile by qPCR of purified LECs: upregulation of p16 (Cdkn2a) and p21 (Cdkn1a), and downregulation of Ki67 (proliferation marker) and VEGFR3 (lymphatic identity marker). Note: the body text of the paper contains a typo (“p27”) that is corrected here by reference to Figure 6L legend and figure panels, which unambiguously label the measured marker as p21. n=6 for the qPCR (Figure 6L; two-tailed unpaired t-tests). Functionally, aged mice fail to expand lymphatic vessels after genotoxic stress, and their MYH11+ perivascular progenitor population does not expand — linking aged LEC senescence to an impaired post-injury regenerative cascade affecting both the hematopoietic niche and bone formation.

Young LEC transplantation into aged mice (n=5) substantially expanded the MYH11+ population and increased expression of bone formation markers (collagen I, osterix, osteocalcin); bone mass was significantly increased on μ-CT. Enhanced hematopoietic regeneration was confirmed by reconstituting activity assays (n=6). Aged LEC transplantation into aged recipients produced no comparable improvements, confirming the age-intrinsic deficit in LECs rather than a niche-only effect. This positions bone LEC cellular-senescence as a mechanistic upstream driver of stem-cell-exhaustion in the hematopoietic and bone-forming compartment after injury.

Proposed mechanism

Genotoxic injury (fracture, irradiation, chemotherapy)
  ↓ IL-6 (systemic / local)
  ↓
[[endothelial-cells|Bone lymphatic endothelial cells]] (LYVE1+/PROX1+/VEGFR3+)
  → proliferate → lymphangiogenesis (VEGFR3-dependent; peak day 15)
  → upregulate [[cxcl12]] (CXCL12 secretion)
  ↓
MYH11+/CXCR4+ perivascular progenitors
  → expand → differentiate to osteoblasts, chondrocytes, adipocytes
  → support bone formation + [[bone-marrow]] niche cellularity
  → enable [[hematopoietic-stem-cells|HSC]] regeneration

In aging: LEC senescence (p16↑, p21↑, Ki67↓, VEGFR3↓) → failure at step 2
  → no MYH11+ expansion → impaired bone + hematopoietic regeneration after injury

Extrapolation to humans

Dimension	Status	Notes
Lymphatic vessels in human bone?	yes	3D LYVE1 imaging (n=2 donors); PROX1 ELISA (n=12 samples, Oxford Musculoskeletal Biobank); scRNA-seq from 2 cancer patient biopsies identified 6 LECs
Pathway conserved in humans?	partial	VEGFR3/CXCL12/CXCR4 axis is conserved; human-specific injury-response lymphangiogenesis not directly demonstrated
Hematopoietic niche dependence on bone lymphatics?	not tested in humans	Mouse conditional-deletion evidence only
Aged LEC senescence in humans?	not tested	Inferred from conservation of senescence biology; no human-aging biopsy data in this paper

gap/needs-human-replication (injury-response lymphangiogenesis and LEC-CXCL12-MYH11 axis are mouse-only functional data)

Limitations

• Primary mechanistic experiments are in mice; the human bone biopsy data establishes presence of lymphatics but not function or injury-response capacity in humans.
• Young adult mice were 8–12 weeks; aged mice were >55 weeks (STAR Methods). The LEC transplantation experiment used aged mice at 61–92 weeks.
• Human biopsy component rests on a small sample (n=12 ELISA, n=2 for 3D imaging; donors from Oxford Musculoskeletal Biobank age range 56–87 years; scRNA-seq from 2 replicates of lung cancer patient biopsies with only 6 LECs identified). The disease context (lung cancer) introduces a potential confounder for the scRNA-seq data. gap/needs-replication
• LEC transplantation is a proof-of-concept experiment (n=5); replication in larger cohorts and with defined LEC purity/senescence-scoring would strengthen the therapeutic interpretation.
• The study does not directly test whether clearance of senescent LECs (senolysis) restores the regenerative response in aged mice — only LEC transplantation was tested.
• The MYH11+ progenitor population is defined by a mature pericyte marker; its relationship to established bone marrow stromal populations (LEPR+, Nestin+, CD146+ MSCs) is not resolved.
• IL-6 is a pleiotropic cytokine; its role here as a pro-regenerative lymphangiogenic signal in bone may conflict with its canonical pro-senescent SASP role, a duality not discussed in this paper.

Cross-links

Atomic entity pages relevant to this study:

• bone · bone-marrow · hematopoietic-stem-cells · endothelial-cells
• cxcl12 · lyve1 · prox1 (protein pages for lymphatic markers)
• cellular-senescence · stem-cell-exhaustion (hallmarks)

Cited by

[Auto-populated by Obsidian backlinks panel]