Meningeal lymphatics clear erythrocytes that arise from subarachnoid hemorrhage

Abstract

Extravasated erythrocytes in cerebrospinal fluid (CSF) critically contribute to the pathogenesis of subarachnoid hemorrhage (SAH). Meningeal lymphatics have been reported to drain macromolecules and immune cells from CSF into cervical lymph nodes (CLNs). However, whether meningeal lymphatics are involved in clearing extravasated erythrocytes in CSF after SAH remains unclear. Here we show that a markedly higher number of erythrocytes are accumulated in the lymphatics of CLNs and meningeal lymphatics after SAH. When the meningeal lymphatics are depleted in a mouse model of SAH, the degree of erythrocyte aggregation in CLNs is significantly lower, while the associated neuroinflammation and the neurologic deficits are dramatically exacerbated. In addition, during SAH lymph flow is increased but without significant lymphangiogenesis and lymphangiectasia. Taken together, this work demonstrates that the meningeal lymphatics drain extravasated erythrocytes from CSF into CLNs after SAH, while suggesting that modulating this draining may offer therapeutic approaches to alleviate SAH severity.

Fig. 1. The erythrocytes in the cerebrospinal fluid after subarachnoid hemorrhage (SAH) were drained into deep cervical lymph nodes (dCLNs) via meningeal lymphatics.

a Isolated dCLNs from mice 4 h after autologous blood (SAH) or saline injected into the cisterna magna (i.c.m.) or blank control group (Con). Scale bar, 2 mm. b Representative images of dCLNs isolated 4 h post-induction of the SAH, saline, and Con groups, stained for Lyve-1 (to image lymphatics, red) or Ter119 (to image erythrocytes, green). Regions of interest in the top images are shown below. White arrows, morphologically intact erythrocytes. Blue arrows, clusters of degraded membrane. Scale bars 200 μm (for top images) and 10 μm (for bottom images). n = 3 independent experiments. c The number of erythrocyte per mm² in the Lyve-1-positive lymphatic sinus of dCLNs from the SAH, saline, and Con groups. Con/Saline; n = 12, SAH; n = 11 mice, pooled from 3 independent experiments. P(Con vs SAH) = 0.002; P(Saline vs SAH) = 0.0063. d Representative image of the meningeal lymphatics (green) on transverse sinus (TS) draining the erythrocytes (red) after 4 h of SAH, compared with the Con and saline groups. Regions of interest in the left images are shown on the right. White arrows, morphologically intact erythrocytes. Blue arrows, clusters of degraded membrane. Scale bar, 20 μm (the left and right images). n = 3 independent experiments. e Quantification of the number of erythrocytes per field accumulated into lymphatics on TS. Con/Saline; n = 13, SAH; n = 15 biologically independent animals, pooled from 3 independent experiments. P(Con vs SAH) < 0.0001; P(Saline vs SAH) < 0.0001. f The expression of classical lymphatic endothelial cell markers by the meningeal lymphatic vessels co-stained with erythrocytes. Representative images of podoplanin (PDPN, red) and Lyve-1 (gray) expressing vessels co-localized with erythrocytes (green) (top images). Representative images of Prox1 (green) and Lyve-1 (gray) expressing vessels co-localized with erythrocytes (red) (bottom images). Yellow arrow, erythrocytes inside the meningeal lymphatic vessels. Scale bar, 20 μm. n = 2 independent experiments. All data are presented as mean values ± SD; one-way ANOVA with Turkey’s multiple-comparison test. **P < 0.01, ****P < 0.0001. Source data are provided as a Source data file.

Fig. 2. Meningeal lymphatics drained erythrocytes labeled in vitro to cervical lymph nodes (CLNs).

Erythrocytes (red blood cells (RBCs)) were labeled by CFSE (5-(and 6)-carboxyfluorescein diacetate succinimidyl ester) in vitro and then injected into cistern magna. a Representative images of meningeal lymphatics draining erythrocytes in CFSE at 4 h postinjection. White arrows, morphologically intact erythrocytes. Purple arrows, degraded membrane. Regions of interest in the left images are shown on the right. Scale bar, 20 μm (the left and right images). n = 2 independent experiments. b Quantification of the number of CFSE-labeled erythrocytes per field in meningeal lymphatics on TS (n = 12 mice per group, pooled from 2 independent experiments. P < 0.0001). c Representative images of CFSE-labeled erythrocytes in the Lyve-1-positvie lymphatic sinus of dCLNs of mice at 4 h postinjection. Regions of interest in the top images are shown below. White arrows, morphologically intact erythrocytes. Purple arrows, clusters of degraded membrane. Scale bars, 200 μm (the top images) and 50 μm (the bottom images). d Representative images of CFSE-labeled erythrocytes in the Lyve-1-positive lymphatic sinus of mandibular LNs of mice at 4 h postinjection. White arrows, morphologically intact erythrocytes. Regions of interest in the top images are shown below. Purple arrows, clusters of degraded membrane. Scale bars, 200 μm (the top images) and 50 μm (the bottom images). n = 3 independent experiments. e Quantification of the density of CFSE-labeled erythrocytes per mm² in Lyve-1-positive lymphatic sinus of dCLNs and mandibular LNs at 4 h postinjection (dCLN, Saline; n = 12, Erythrocytes in CFSE; n = 15 independent nodes, P = 0.0006; Mandibular LN, Saline; n = 12, Erythrocytes in CFSE; n = 23 nodes, pooled from 3 independent experiments. P = 0.0207). All data are presented as mean values ± SD; two-tailed unpaired Student’s t test. *P < 0.05, ***P < 0.001, ****P < 0.0001. Source data are provided as a Source data file.

Fig. 3. Ablation of meningeal lymphatic vessels blocks drainage of the erythrocytes into dCLNs.

a Representative images of meningeal lymphatics (labeled by Lyve-1, red) in the Con, Laser only, Visudyne only, and Laser plus Visudyne groups. Regions of interest in the top images are shown below. Scale bars, 1 mm (top images) and 200 μm (bottom images). b Quantification of the meningeal lymphatics coverage on the TS (n = 5 per group, P(Con vs Laser + Visudyne) < 0.0001, P(Laser vs Laser + Visudyne) < 0.0001, P(Visudyne vs Laser + Visudyne) < 0.0001) and c superior sagittal sinus (SSS) from each group of mice (n = 5 per group, P(Con vs Laser + Visudyne) < 0.0001, P(Laser vs Laser + Visudyne) = 0.0007, P(Visudyne vs Laser + Visudyne) = 0.0003). d Isolated dCLNs from mice 4 h after induction of SAH in the Laser (L + SAH), Visudyne (V + SAH), and Laser plus Visudyne (L + V + SAH) groups. Scale bar, 2 mm. e Representative images of dCLNs isolated 4 h post-induction of SAH and stained for the lymphatics (Lyve-1, red) and erythrocytes (Ter119, green) in the 3 groups of mice indicated. White arrows, morphologically intact erythrocytes. Blue arrows, clusters of degraded membrane. Regions of interest in the top images are shown below. Scale bars, 200 μm (the top images) and 10 μm (the bottom images). n = 2 independent experiments. f The number of erythrocyte per mm² in Lyve-1-positive lymphatic sinus of dCLNs 4 h after SAH induction (L + SAH/L + V + SAH; n = 13, V + SAH; n = 12, pooled from 2 independent experiments. P(L + SAH vs L + V + SAH) = 0.0013, P(V + SAH vs L + V + SAH) = 0.006). g Representative images of the brain after 7 days of induction of SAH in the groups indicated. Arrows, blood clot in the brain. Scale bar, 1 mm. All data are presented as mean values ± SD, one-way ANOVA with Turkey’s multiple-comparison test, **P < 0.01, ***P < 0.001, ****P < 0.0001. Source data are provided as a Source data file.

Fig. 4. Ablation of meningeal lymphatics aggravates the neuroinflammatory response, the defects of exploratory behavior, and short-term working memory in SAH.

a–d Impaired meningeal lymphatics worsened the microglia activation in SAH. a Representative FACS plots showing gating strategy we used in flow cytometric analysis. Cell populations in the right dot plots defined as CD11b⁺CD45^low (microglia) were gated for further analysis. b Representative dot plots showing the ratio of CD16/32-positive subsets and CD206-positive subsets from different groups. c Quantification of CD16/32⁺CD206⁻ subsets of CD11b⁺CD45^low populations (P(SAH vs L + V + SAH) < 0.0001, P(L + SAH vs L + V + SAH) < 0.0001, P(V + SAH vs L + V + SAH) < 0.0001, P(L + V vs L + V + SAH) < 0.0001), d CD206⁺CD16/32⁻ subsets of CD11b⁺CD45^low populations (P(SAH vs L + V + SAH) = 0.0146, P(L + SAH vs L + V + SAH) < 0.0001, P(L + V vs L + V + SAH) = 0.0086), and e CD16/32⁺CD206⁺ subsets of CD11b⁺CD45^low populations in different groups (P(V + SAH vs L +V + SAH) = 0.025). SAH/L + V + SAH; n = 14, L + SAH/V + SAH/L + V; n = 12, pooled from 2 independent experiments. f–i Ablation of meningeal lymphatics worsens the exploratory behavior and short-term working memory associated with SAH. f Percentage of time spent in the center area in the open field test (P(SAH vs L + V + SAH) = 0.0194, P(L + SAH vs L + V + SAH) = 0.0053, P(V + SAH vs L + V + SAH) = 0.009, P(L + V vs L + V + SAH) = 0.0138), g the number of entries into the center area in the open field test (P(SAH vs L + V + SAH) < 0.0001, P(L + SAH vs L + V + SAH) = 0.0001, P(V + SAH vs L + V + SAH) = 0.0018, P(L + V vs L + V + SAH) = 0.0141), h percentage of time spent in the novel arm (NA) in the Y-maze test (P(SAH vs L + V + SAH) = 0.0006, P(L + SAH vs L + V + SAH) = 0.003, P(L + V vs L + V + SAH) = 0.0198), and I the number of entries into the NA in the Y-maze test in the groups of mice indicated (P(SAH vs L + V + SAH) < 0.0001, P(L + SAH vs L + V + SAH) < 0.0001, P(V + SAH vs L + V + SAH) = 0.0338, P(L + V vs L + V + SAH) = 0.0303). SAH; n = 22, L + SAH/V + SAH/L + V + SAH; n = 19, L + V; n = 15 mice, pooled from 2 independent experiments. All data are presented as mean values ± SD, one-way ANOVA with Turkey’s multiple-comparison test, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS, not significant. APC allophycocyanin, PE R-phycoerythrin, FITC fluorescein isothiocyanate. Source data are provided as a Source data file.

Fig. 5. The blockage of VEGFR3 exacerbated the neuroinflammation and behavioral defects induced by SAH.

a Scheme of the experiments in VEGFR3 tyrosine kinase inhibitor treatment experiments. Mice were treated with MAZ51 once a day for 30 days (5 days per week). After 7 days of SAH induction or sham treatment, mice were killed for analysis. b Representative images of meningeal lymphatics labeled by Lyve-1 (red) in the vehicle and MAZ51 treatment groups. Regions of interest in the top images are shown below. Scale bar, 1 mm (the top images) and 200 μm (the bottom images). c Quantification of the percentage of Lyve-1 coverage on TS and SSS. (Vehicle; n = 12, MAZ51; n = 13 mice. TS, P < 0.0001, SSS, P < 0.0001, two-tailed unpaired Student’s t test). d Representative FACS plots showing gating strategy in flow cytometric analysis. Cell populations characterized by CD11b⁺CD45^low (microglia) were gated for further analysis. e Representative dot plots showing the ratio of CD16/32-positive subsets and CD206-positive subsets. f Quantification of CD16/32⁺CD206⁻ subsets of CD11b⁺CD45^low populations (P(Vehicle + Sham vs Vehicle + SAH) = 0.0139, P(MAZ51 + Sham vs MAZ51 + SAH) < 0.0001, P(Vehicle + SAH vs MAZ51 + SAH) < 0.0001), g CD206⁺CD16/32⁻ subsets of CD11b⁺CD45^low populations (P(Vehicle + Sham vs MAZ51 + SAH) = 0.0233), and h CD16/32⁺CD206⁺ subsets of CD11b⁺CD45^low populations in different groups. Vehicle + Sham/Vehicle + SAH; n = 10, MAZ51 + Sham; n = 11, MAZ51 + SAH; n = 12 mice, pooled from 2 independent experiments. i–l VEGFR3 blockage worsened neurological performance of SAH. i Percentage of time spent in the center area in the open field test (P(Vehicle + Sham vs Vehicle + SAH) = 0.011, P(MAZ51 + Sham vs MAZ51 + SAH) < 0.0001, P(Vehicle + SAH vs MAZ51 + SAH) = 0.0111), j the number of entries into the center area in the open field test (P(Vehicle + Sham vs Vehicle + SAH) = 0.0216, P(MAZ51 + Sham vs MAZ51 + SAH) = 0.0004), k percentage of time spent in the novel arm (NA) in the Y-maze test (P(Vehicle + Sham vs MAZ51 + SAH) = 0.0001, P(MAZ51 + Sham vs MAZ51 + SAH) = 0.0001), and l the number of entries into the NA in the Y-maze test in the groups of mice indicated (P(Vehicle + Sham vs Vehicle + SAH) = 0.0269, P(MAZ51 + Sham vs MAZ51 + SAH) = 0.0037, P(Vehicle + SAH vs MAZ51 + SAH) = 0.0221). Vehicle + Sham; n = 13, Vehicle + SAH/MAZ51 + Sham; n = 12, MAZ51 + SAH; n = 15 mice. All data are presented as mean values ± SD, one-way ANOVA with Turkey’s multiple-comparison test (f–l), *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS, not significant. Source data are provided as a Source data file.

Fig. 6. Meningeal lymphatic flow is augmented in the SAH mouse model.

a Representative images of meninges and dCLNs stained for anti-Lyve-1 (AF⁵⁵⁵-conjugated secondary antibody, ex vivo) and AF⁴⁸⁸-conjugated anti-Lyve-1 (i.c.m.). Scale bars, 1 mm (for the meninges,); 200 μm (for the dCLNs). n = 2 independent experiments. b Quantification of area fraction (%) dividing the area of meningeal lymphatics labeled by AF⁴⁸⁸-conjugated Lyve-1 antibody (i.c.m.) by the area of meningeal lymphatics (Con; n = 10, Saline; n = 7, SAH; n = 6 mice. P(Con vs SAH) = 0.0122, P(Saline vs SAH) = 0.0303). c Quantification of meningeal lymphatic area of the TS. Con; n = 10, Saline; n = 7, SAH; n = 6 mice. d Quantification of area fraction (%) occupied by the area of dCLN lymphatics labeled by AF⁴⁸⁸-conjugated Lyve-1 antibody (i.c.m.) and the area of dCLNs. Con; n = 13, Saline; n = 11, SAH; n = 14 nodes, pooled from 2 independent experiments. P(Con vs SAH) = 0.0183, P(Saline vs SAH) = 0.0181). e Size of dCLNs in the Con, saline, and SAH groups. Con; n = 13, Saline; n = 11, SAH; n = 14 nodes, pooled from 2 independent experiments. f Representative images of exogenously injected fluorescent microbeads (1 μm in diameter, red) in the dCLNs of the control, saline injection, and autologous blood injection groups. Scale bar, 200 μm. n = 2 independent experiments. g The percentage of microbead coverage in the dCLNs. Con/SAH; n = 13, Saline; n = 9 nodes, pooled from 2 independent experiments. P(Con vs SAH) = 0.0004, P(Saline vs SAH) = 0.0251). h Size of dCLNs in the Con, saline, and SAH groups. Con/SAH; n = 13, Saline; n = 9 nodes, pooled from 2 independent experiments. i Representative images of ICG fluorescence of mandibular LNs and its afferent lymphatics at 30 min after ICG injection. Scale bar, 3 mm. Red circle, region of interest of lymphatic vessel for lymph flow frequency analysis. j Quantification of the number of the afferent lymphatic vessel contraction frequency (pulse per minute) (n = 6 mice per group, P = 0.0261, two-tailed unpaired Student’s t test). All data are presented as mean values ± SD; one-way ANOVA with Turkey’s multiple-comparison test (b–e, g, h), *P < 0.05, ***P < 0.001. NS, not significant. Source data are provided as a Source data file.

Fig. 7. The diameter, branching, and area of meningeal lymphatics do not change significantly after 7 days of SAH.

a Representative images of meninges whole mounts stained for meningeal lymphatics (Lyve-1, green) from the indicated groups (4 mice per group). The lymphatics on left and right TS were divided into 8 segments (400 μm each), respectively. Regions of interest of each top image are shown in the row below (white asterisk, the branching of lymphatics). Scale bars, 1 mm (the top images) and 200 μm (the bottom images). b Quantification of lymphatic diameters in eight different segments. X axis represents 400–3200 μm distance to the junction of all sinuses. n = 4 mice per group. Left 800, P(Con vs SAH) = 0.002, P(Saline vs SAH) = 0.0067; Left 1200, P(Con vs SAH) = 0.0262, P(Saline vs SAH) = 0.0374; Right 2800, P(Con vs SAH) = 0.0005, P(Saline vs SAH) < 0.0001. c Quantification of lymphatic branches in eight different segments. X axis represents 400–3200 μm distance to the junction of all sinuses. n = 4 mice per group. Left 1600, P(Saline vs SAH) = 0.0473; Left 400, P(Saline vs SAH) = 0.0287; Right 2800, P(Con vs SAH) = 0.0319, P(Saline vs SAH) = 0.0063; Right 3200, P(Saline vs SAH) = 0.0191. d, e The percentages of Lyve-1 coverage on the TS (d) and the SSS (e). n = 4 mice per group. All data are presented as mean values ± SD; two-way ANOVA (b, c) or one-way ANOVA (d, e) with Turkey’s multiple-comparison test. *P < 0.05, **P < 0.01, ***P < 0.001, Con vs SAH, ^#P < 0.05, ^##P < 0.01, ^####P < 0.0001, Saline vs SAH. NS, not significant. Source data are provided as a Source data file.