Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Oct 9.
Published in final edited form as: Exp Neurol. 2022 Dec 6;360:114293. doi: 10.1016/j.expneurol.2022.114293

EphA4/EphrinB2 signaling mediates pericyte-induced transient glia limitans formation as a secondary protective barrier after subarachnoid hemorrhage in mice

Jiru Zhou a,b,c,d, Peiwen Guo b,c,d, Mingxu Duan b,c,d, Junhan Li a, Xufang Ru b,c,d, Lin Li a,b,c,d, Zongduo Guo a, John H Zhang e,f, Hua Feng b,c,d, Yujie Chen b,c,d,*, Xiaochuan Sun a,**
PMCID: PMC10561606  NIHMSID: NIHMS1933604  PMID: 36493862

Abstract

Background:

Most patients with subarachnoid hemorrhage (SAH) do not exhibit brain parenchymal injury upon imaging but present significant blood–brain barrier (BBB) disruption and secondary neurological deficits. The aim of this study was to investigate whether stressed astrocytes act as a secondary barrier to exert a protective effect after SAH and to investigate the mechanism of glial limitan formation.

Methods:

A total of 204 adult male C57BL/6 mice and an endovascular perforation SAH model were employed. The spatiotemporal characteristics of glial limitan formation after SAH were determined by immunofluorescence staining and transmission electron microscopy. The molecular mechanisms by which pericytes regulate glia limitans formation were analyzed using polymerase chain reaction, Western blotting, immunofluorescence staining and ELISA in a pericyte-astrocyte contact coculture system. The findings were validated ex vivo and in vivo using lentiviruses and inhibitors. Finally, pericytes were targeted to regulate glial limitan formation, and the effect of the glia limitans on secondary brain injury after SAH was evaluated by flow cytometry and analysis of neurological function.

Results:

Stress-induced glial limitan formation occurred 1 day after SAH and markedly subsided 3 days after ictus. Pericytes regulated astrocyte glia limitan formation via EphA4/EphrinB2 signaling, inhibited inflammatory cell infiltration and altered neurological function.

Conclusions:

Astrocyte-derived glia limitans serve as a secondary protective barrier following BBB disruption after SAH in mice, and pericytes can regulate glial limitan formation and alter neurological function via EphA4/EphrinB2 signaling. Strategies for maintaining this secondary protective barrier may be novel treatment approaches for alleviating early brain injury after SAH.

Keywords: Subarachnoid hemorrhage, Blood–brain barrier, Glia limitans, Pericyte, EphA4, EphrinB2

1. Background

Subarachnoid hemorrhage (SAH) is a common critical cerebrovascular illness, mainly due to ruptured intracranial aneurysms. Although ruptured aneurysms can be well treated by surgical clamping or interventional embolization, there is still a lack of effective treatment for secondary brain injury that leads to a poor prognosis in patients with SAH(Chen et al., 2022; Claassen and Park, 2022). The vast majority of patients with SAH do not show brain parenchymal injury upon imaging but exhibit significant blood–brain barrier (BBB) disruption and secondary neurological deficits(Weimer et al., 2017). Previous studies have established that BBB disruption leads to vasogenic brain edema(Keep et al., 2018). However, peak cerebral edema occurs 3–5 days after onset rather than 1 day after onset, and previous studies have revealed severe disruption of the BBB as early as 1 day after SAH(Rass et al., 2019). Therefore, the spatial and temporal patterns of BBB disruption and remodeling after SAH are unclear, and the time window for BBB-targeted strategies for repairing central nervous system (CNS) damage is unclear.

The traditional BBB is thought to consist primarily of tight junctions of endothelial cells and basement membranes forming a barrier structure that separates circulating blood from extracellular brain fluid (Segarra et al., 2021). During embryonic development, endothelial cells gradually infiltrate the CNS, while pericytes are first recruited to the perineurium and regulate BBB function by regulating endothelial tight junction proteins and vesicular transport(Coelho-Santos et al., 2021; Daneman et al., 2010; Mae et al., 2021). Astrocytes are involved in BBB composition, interacting with pericytes and endothelial cells, directly sensing signals from injury stimuli, and regulating the microenvironment of the neurovascular unit through paracrine and intra- and extracellular transport mechanisms(Sanmarco et al., 2021). Pericytes and astrocytes both play a key role in establishing and maintaining BBB function and are involved in the process of BBB destruction in acute CNS injury(Solar et al., 2022). The interactions between pericytes and astrocytes and their functions in the BBB are not well studied, and there is a lack of practical targets for intervention.

Interestingly, Horng et al. found that astrocytes specifically express CLDN1 and CLDN4 and form glia limitans when induced by IL-1β but not in the vascular endothelium or basement membrane(Horng et al., 2017). Gradually, an increasing number of studies examining the structure and integrity of the BBB have expanded to include the barrier properties of the astrocyte end-foot outside the vascular endothelium and basement membrane(Mora et al., 2020). The formation of glia limitans by astrocytes forms a barrier in response to long-term chronic inflammation, such as multiple sclerosis and experimental autoimmune encephalomyelitis, and acts as a second barrier outside the tight junctions of endothelial cells to limit the infiltration of immune molecules and immune cells from the blood into the brain parenchyma(Mora et al., 2020; Quintana, 2017). However, studies on whether glia limitans are formed and their associated functions in acute CNS injury are lacking, even though a functional basis for the formation of glia limitans exists in astrocytes.

Hirai H et al. cloned a new tyrosine kinase receptor, called the Eph receptor, from a hepatoma cell line(Hirai et al., 1987). Ephrin ligand is a membrane-bound protein found in lipid rafts. Eph/Ephrin normally mediates intercellular contact interactions and, in the CNS, not only regulates axon guidance and neuroplasticity but is also involved in neuropathological processes such as endothelial injury, vascular regeneration, and neuroinflammation(Kania and Klein, 2016). Previous literature has reported that activation of the EphA family occurs after brain injury, increases BBB permeability and promotes neuroinflammation(Malik and Di Benedetto, 2018). Retinal pericytes promote retinal microvascular repair and remodeling following activation of the EphrinB clade(Salvucci et al., 2009). However, the role of the Eph/Ephrin system in BBB dysfunction after SAH is unclear. As a possible mechanism, the present study investigated intercellular interactions between astrocytes and pericytes in the BBB after SAH in mice.

2. Methods

2.1. Experimental animals

The 204 adult male C57B6J mice at 8–10 weeks of age and 198 newborn C57B6J mice used in this study were provided by the Experimental Animal Centre of the Third Military Medical University (Chongqing, China). All experimental procedures were approved by the Laboratory Animal Welfare and Ethics Committee of Third Military Medical University (AMUWEC2020793) and followed the guidelines of ARRIVE 2.0.

In all experimental groupings, animals from each litter were randomly assigned to each experimental group prior to the intervention according to the random number table method. Details of randomization, blinding and animal assessment are shown in Supplemental Fig. 1 and Supplemental Table 1. The experimental operators and observers did not know the names and groupings of the animals. The data analysts knew the names of the animals but not the group assignments. Animals were bred in the SPF animal house and housed under a 12 h light-dark cycle with food and water available ad libitum.

2.2. SAH model

To avoid possible effects of the estrogen physiological cycle on the pathophysiology of SAH, only male mice were used in this study. Food was withheld from the animals for 12 h prior to surgery, and the mice were randomly allocated among the surgeons. The SAH mouse model was constructed by intravascular perforation as described previously(Ru et al., 2021). Mice were anesthetized using a gas mixture containing 1.5% isoflurane/air. The skin of the neck was disinfected, and a median cervical incision was made to identify and ligate the external carotid artery, preserving the 3-mm stump. A 5–0 sharp single-strand nylon suture was pushed through the stump into the internal carotid artery until resistance was felt (at approximately 8–10 mm) and then further advanced 2 mm to allow the suture to penetrate the bifurcation of the anterior and middle cerebral arteries. The suture was withdrawn, and the internal carotid artery was reperfused to induce SAH. The sham group underwent the same procedure, which was identical except that the vessel was not penetrated. Mice that died intraoperatively or immediately postoperatively or with SAH grade scores less than or equal to seven were excluded from this study and were immediately replaced.

2.3. Lentivirus

Four recombinant lentiviruses were used for in vitro and in vivo experiments. (1) The pLVX-IRES-ZsGreen1 lentiviral vector for EphA4 overexpression; (2) the pLVX-shRNA2 lentiviral vector for EphA4 interference; (3) the pLVX-IRES-Puro lentiviral vector for EphrinB2 overexpression; and (4) the pLVX-shRNA1 lentiviral vector for EphrinB2 interference. The final lentiviral titers were 2 × 10^9 TU/mL, and all lentiviruses were produced by Biomedicine Biotechnology, Chongqing, China. All were stored at −80 °C prior to use. In the in vivo experiments, 5 μL was injected into the lateral ventricle at a rate of 1 μL/min at the time of use, and animals were used for SAH model construction one week later. In the in vitro experiments, astrocytes and pericytes were infected with MOI = 10 for 24 h and then switched back to conventional medium and cultured for 72 h to verify expression.

2.4. Lateral ventricular injection

Animals were injected into the lateral ventricle after anesthesia as described previously(Pan et al., 2020). Under aseptic conditions, a 10-μL Hamilton syringe (Microliter 701; Hamilton Company) was inserted into the left ventricle with the aid of a mouse brain stereotyper at a depth of 1.7 mm, 0.6 mm posterior to the fontanelle and 1.5 mm lateral to the sagittal suture of the skull. The needle was stopped for 10 min and slowly removed at a rate of 0.5 mm/min.

2.5. Neurological assessment

In this study, independent observers evaluated the neurological function of the mice at 24 h and 72 h after SAH using a modified Garcia scale (18-point scale) score and a beam balance test as described previously(Chen et al., 2015). Each mouse was observed for 5 min, and six parameters (3–18 points), i.e., voluntary movement, limb movement, forelimb extension, grid climbing, contralateral somatosensory function and tactile response, were assessed with a modified Garcia scale. In the beam balance test, mice were placed on a 15-mm-wide beam suspended at a height of 1.2 m, and the distance each mouse walked within 1 min was recorded (scored on a scale of 0–5). The neurological assessment was performed by two observers blinded to the animal numbers and groups.

2.6. SAH grading

An 18-point scale was used to assess the extent of bleeding after SAH as described previously(Chen et al., 2015). The ventral surface of the brain was divided into 6 regions, and a score of 0–3 was given depending on the degree of blood clot encapsulation of the vessels and coverage of the underlying surface. The scores for the 6 regions were summed to calculate the total score. Animals with a score of <8 were excluded from this study.

2.7. Transmission electron microscopy (TEM)

After the mice were anesthetized and intracardially perfused with cold 0.9% saline and following decapitation, the brains were removed. Small pieces of brain tissue (1 mm3) were fixed with 4% glutaraldehyde containing 0.1 mol/L sodium cacodylate (pH = 7.3) overnight at 4 °C. The small pieces of brain tissue were embedded in Epon (Agar100 resin, Agar Scientific). Images of 1-μm-thick toluidine blue-stained sections were obtained using TEM (JEM-1200EX, JEOL).

2.8. Evans blue extravasation

Mice were anesthetized 24 h and 72 h after SAH and injected with 5 mL/kg sterile 2% Evans Blue (IE0280, Solarbio) in saline via the tail vein and placed in a 37 °C incubator. Blue staining of the skin and sclera was observed 1 h after injection. Under anesthesia, mice were humanely sacrificed by intracardiac perfusion of cold 10 μM phosphate-buffered saline (PBS) followed by 4% paraformaldehyde. Coronal brain tissue sections (25 μm) were prepared after decapitation. Red Evans blue autofluorescence was observed with a rhodamine filter (Olympus OX51, Tokyo, Japan).

2.9. Preparation of primary astrocytes from mouse brain tissue

One- to three-day-old C57B6J mice were humanely sacrificed by exposure to carbon dioxide and then decapitated(Pan et al., 2020). After the skull was opened and the brain was removed, the meninges were carefully stripped off, and cortical and hippocampal tissues were collected. Using sterile scissors, the brain tissue was cut into small pieces (approximately 1 cubic millimeter). The tissue pieces were digested using a mixture of 0.25% trypsin-EDTA (25,200,072, Gibco™) and 40 μg/mL DNase I (DN25, Sigma–Aldrich) for 10–15 min before an appropriate amount of 1 mg/mL trypsin inhibitor (T6522, Sigma–Aldrich) was added to terminate the digestion. The samples were centrifuged at 400 ×g for 5 min, the supernatant was discarded, and this step was repeated twice. The precipitate was finally resuspended in warmed high-glucose DMEM (11,965,118, Gibco™) containing 10% fetal bovine serum (10099141C, Gibco™). After cell counting, the cells were inoculated in washed cell culture flasks precoated with 10 μg/mL poly-d-lysine (P6407, Sigma–Aldrich) at a density of 5 × 105/cm2. The flasks were incubated in a cell incubator containing 5% CO2 at 100% humidity and 37 °C. The medium was changed on the second day and every five days thereafter. On days 7–10 of culture, after the cells on the bottom of the flask had just formed a confluent layer, the microglia and oligodendrocytes that had not adhered to the wall were isolated by shaking on a horizontal shaker overnight at 250 rpm and 37 °C. After removal of the floating cells, the cells were passaged using the original medium at a ratio of 1:3. The obtained astrocytes were used for the in vitro experiments.

To ensure that the astrocytes were not contaminated with microglia, the cells were fixed using 10 μM PBS containing 4% paraformaldehyde, and astrocytes and microglia were labeled using GFAP (NB100–53809, NOVUS) and Iba-1 (019–19,741, Wako Laboratory Chemicals) antibodies. After the cells were incubated with the corresponding fluorescent secondary antibody, 10 areas were analyzed under a 20× objective to ensure that the final purity of the astrocytes was >99%.

2.10. Preparation of primary pericytes from mouse brain tissue

Primary pericytes were cultured as described previously. Brain tissues from 1- to 3-day-old C57B6J mice were collected and digested in MEM, HEPES (12,360,038, Gibco™ containing 30 U/mL papain (A501612, Sangon Biotech) and 40 μg/mL DNase I for 70 min. The final digestion was performed with 30% bovine serum albumin (A8010, Solarbio) in PBS (BSA:PBS = 1.7:1), and then, the samples were centrifuged at 1360 ×g for 10 min. The bottom layer of the precipitate was carefully aspirated and resuspended in EGM-2MV BulletKit medium (CC-3202, Lonza). After centrifugation at 250 ×g for 5 min, the cells were inoculated in dishes precoated with 0.02% collagen I (C8061, Solarbio) at a density of 5 × 105/cm2 and incubated in a cell incubator containing 5% CO2 at 100% humidity and 37 °C. The medium was changed after 20 h and every 3 days thereafter. When the cells had just formed a confluent layer, three consecutive cell passages were performed at a ratio of 1:4. After the third passage, the medium was replaced with pericyte medium (e76127787, ScienCell Research Laboratories), and the pericytes obtained after the third passage were used for the in vitro experiments.

To ensure pericyte purity, pericytes were identified by double immunofluorescence with an anti-PDGFR beta C-terminal antibody (ab32570, Abcam) and anti-aminopeptidase N/CD13 antibody (AF2335, NOVUS) after fixation with 10 μM PBS containing 4% paraformaldehyde. After the cells were incubated with the corresponding fluorescent secondary antibody, 10 areas were analyzed under a 20× objective to ensure that the final purity of the pericytes was >99%.

2.11. Pericyte-astrocyte contact coculture

To study the interaction between pericytes and astrocytes in vitro, we designed a contact coculture system (Fig. 2A) based on Millicell Standing Insert chambers (PIHP03050, Merck Millipore). Pericytes were first inoculated on the lower side of the chamber at a density of 1 × 105/cm2. After 12 h of cell apposition, the chamber was turned over, and astrocytes were inoculated on the top side of the membrane at a density of 1 × 105/cm2. The chambers were inserted into the well plate, and the fluid level of the medium inside and outside the chambers was balanced. The coculture system was ready for use the next day.

Fig. 2.

Fig. 2.

Open in a new tab

Expression and localization of Eph/Ephrin in pericytes and astrocytes in vitro.

A, Schematic diagram of pericyte-astrocyte cell contact coculture and pericyte region given in the presence or absence of OxyHb stimulation for 24 h. Pericytes were first inoculated on the dorsal side of the standing Transwell membrane; then, 12 h later, astrocytes were inoculated on the upper side of the Transwell membrane, inserted into six-well plates, and supplemented with medium on the inner and outer sides. Quantitative PCR of pericyte and astrocyte efn families versus eph mRNA families. n = 3, each group, t-test, * P < 0.01 vs. Control group. B, Western blot and quantitative analysis of pericytes and astrocytes in coculture systems using antibodies against Ephrin type-A receptor 4 (EphA4) and EphrinB2. Pericytes showed significantly higher EphA4 expression after OxyHb treatment, and astrocytes exposed to OxyHb-stimulated pericytes exhibited significantly higher EphrinB2 expression. n = 6 per group, 1-way ANOVA, * P < 0.05 vs. the Peri-Con group; ** P < 0.05 vs. the Ast-Con group.

2.12. Transendothelial electrical resistance (TEER) measurement

The TEER values of cells within the chamber were measured using a Millicell ERS-2 resistograph (MERS00002, Merck Millipore) once the cells had reached 90–100% confluence. Prior to measurement, the function of the meter was assessed to ensure stable measurement. Cells were analyzed after they had reached room temperature. The electrodes were inserted vertically such that the long electrode was outside the measurement chamber and the short electrode was inside the measurement chamber and did not touch the cells on the membrane.

TEER values were calculated as (cell-containing chamber resistance Ω1 - cell-free chamber resistance Ω0) x the area of the bottom of the chamber (cm2) and are expressed in Ω • cm2.

2.13. Western blotting

Western blotting was performed as described previously(Li et al., 2022). Proteins were isolated from the perforated lateral hemisphere or cells on Millicell Standing Inserts and used for Western blot analysis. Proteins were extracted from the samples with T-PER Tissue Protein Extraction Reagent (78,510, Thermo Fisher Scientific) according to the manufacturer’s instructions. A total of 30 μg protein was loaded in each well of an SDS–PAGE gel. After gel electrophoresis, the proteins were transferred onto polyvinylidene fluoride (PVDF) membranes, which were blocked with blocking buffer for 2 h at room temperature. The PVDF membranes were incubated with diluted primary antibody overnight at 4 °C on a horizontal shaker. The following antibodies were used: EphA4 monoclonal antibody (1:1000; 37–1600, Invitrogen), EphrinB2 antibody (1:2000; ab150411, Abcam), Stat3 (124H6) mouse monoclonal antibody (1:1000; 9139S, Cell Signaling Technology), phosphorylated Stat3 (Tyr705) (D3A7) XP® rabbit mAb (1:1000; 9145S, Cell Signaling Technology), Jak2 (D2E12) XP® Rabbit monoclonal antibody (1:1000; 3230S, Cell Signaling Technology), claudin 1 monoclonal antibody (1:1000; 37–4900, Invitrogen) and GAPDH polyclonal antibody (1:5000, 10,494–1-AP, Proteintech). The blots were incubated with an appropriate horseradish peroxidase (HRP)-conjugated secondary antibody for 2 h at room temperature and imaged using the Fusion FX Edge Spectra multifunctional imaging system (Vilber Bio Imaging). The band density was quantified using ImageJ software.

2.14. Fluorescence immunohistochemistry and immunocytochemistry

Frozen sections were used for fluorescence immunohistochemistry after fixation as described previously(Wu et al., 2019). Mice from the sham and SAH groups were deeply anesthetized 1 and 3 days postoperatively, their hearts were perfused with ice-cold 10 μM PBS and 4% paraformaldehyde, and brain tissues were collected. The tissues were fixed in 4% paraformaldehyde for 24 h and sequentially dehydrated in 30% and 40% sucrose. Coronal sections of the brain tissue (25 μm) were obtained using a cryostat (CM3050S-3-1-1, Leica). The sections were blocked for 60 min at room temperature using PBS containing 5% goat serum and 0.3% Triton-100. Then, the sections were incubated overnight at 4 °C with diluted primary antibody. The following antibodies were used: claudin 1 monoclonal antibody (1:500, 37–4900, Invitrogen), connexin 43 antibody (1:500, C6219-100UL, Sigma–Aldrich), GFAP antibody (1:800, NB100–53809, Novus), aminopeptidase N/CD13 antibody (1:800, AF2335, Novus), anti-PDGFR beta C-terminal antibody (1:500, ab32570, Abcam), EphA4 monoclonal antibody (1:500, 37–1600, Invitrogen), EphrinB2 antibody (1:1000, ab150411, Abcam), and CD45 (D3F8Q) rabbit monoclonal antibody (1:500, 70257S, Cell Signaling Technology). The next day, the frozen sections were incubated with the appropriate fluorescent secondary antibody at room temperature away from light for 2 h. After incubation, the sections were stained with DAPI for 5 min. The temporal expression and spatial distribution of the various protein markers were observed using an ultrahigh resolution inverted confocal microscope (LSM880, Zeiss). The statistical analysis of immunofluorescence was performed by using ImageJ V1.53 (National Institutes of Health) for fluorescence intensity or counting positive cells.

Cells in glass-bottomed dishes were used for fluorescence immunocytochemistry. The cells were washed and fixed with 4% paraformaldehyde for 15 min and then washed with PBS. The subsequent immunostaining procedure was the same as that used for fluorescence immunohistochemistry.

2.15. Cytokine array

Cytokine levels in the supernatants of astrocytes and pericytes were analyzed using the Proteome Profiler Mouse Cytokine Array Kit (ARY006, R&D Systems). The supernatant was centrifuged to remove impurities and immediately used for the assay according to the manufacturer’s instructions. The samples were mixed with an antibody cocktail and incubated overnight with the precoated array membrane on a horizontal shaker at 4 °C. Next, the membranes were incubated with Streptavidin-HRP for 30 min and Chemi Reagent mixture for 1 min. Then, the membranes were analyzed with the Fusion FX Edge Spectra multifunctional imaging system (Vilber Bio Imaging). The bands were quantified using ImageJ software.

2.16. Quantitative polymerase chain reaction (PCR)

Total RNA was isolated from cells cultured on membranes in coculture chambers using TRIzol (#15596018, Thermo Fisher Scientific). Reverse-transcription PCR and quantitative real-time PCR were performed as described previously(Deng et al., 2020). mRNA levels were normalized to the level of actin. The primer sequences are shown in Supplemental Table 3.

2.17. ELISA

Cytokine levels in the supernatants of pericytes and astrocytes from the coculture system were assayed with ELISA kits (see Supplemental Material for model and manufacturer) according to the manufacturer’s instructions. The supernatants were centrifuged to remove impurities prior to analysis. Supernatants were obtained from cells from each group cultured at a density of 1 × 105/cm2.

2.18. Flow cytometry

To quantify immune cell infiltration in the brain parenchyma after SAH, we first prepared single-cell suspensions from brain tissue. The mice were deeply anesthetized and humanely sacrificed by cardiac perfusion with ice-cold 10 μM PBS. The brains were removed from severed heads and homogenized in RPMI 1640 medium (11,875,093, Gibco™) using Dounce Tissue Grinders. The cell suspensions were filtered through a 100-μm cell sieve and centrifuged at 300 xg for 5 min. Then, the precipitate was resuspended in an appropriate volume of RPMI 1640 medium to prepare a single-cell suspension. Next, 5 mL of 70% Percoll solution (P8370, Solarbio) was added to the lower layer, and 2.5 mL of 30% Percoll solution was added to the upper layer. The cell suspension was carefully added between the upper and lower layers of the Percoll mixture, and the cells in suspension between the 30% and 70% Percoll solutions were collected by centrifugation at 1000 xg for 30 min at room temperature. The cells were resuspended in an appropriate volume of PBS, centrifuged at 250 xg for 10 min, resuspended in 100 μL of PBS and incubated at room temperature for 20 min with 1% FBS. Staining was performed using the following fluorochrome-conjugated antibodies: FITC-conjugated rat anti-CD11b (557,396, BD Pharmingen™), APC-conjugated rat anti-mouse CD45 (559,864, BD Pharmingen™), PE-conjugated mouse anti-mouse Nk1.1 (553,165, BD Pharmingen™) and PerCP-Cy™5.5-conjugated hamster anti-mouse CD3e (551,163, BD Pharmingen™). Flow cytometry was performed on an ACEA flow cytometer (Novocyte), and the data were analyzed using FlowJo v7.6 software (Informer Technologies).

2.19. mRNA microarray profiling

mRNA microarray analysis was performed to investigate phenotypic changes in astrocytes in contact with pericytes under different conditions. Total RNA was collected from astrocytes in the coculture system using TRIzol (#15596018, Thermo Fisher Scientific) and detected using the Affymetrix GeneChip Gene 1.0 Array after assessing microarray quality, RNA quality and purity. Statistical annotation was performed on the resolved data from each sample microarray to obtain gene expression values and annotation information. Statistical annotation was performed on the microarray analysis data of each sample to obtain the expression value and annotation information of the gene. The limma package was used to screen differences between the two groups to calculate the significance level (P value) between transcripts to obtain a differential gene screening condition of P Value <0.05. The above microarrays and assays were performed by Genminix Information, Shanghai, China.

2.20. Statistical analysis

Data are presented as the mean ± SD. Normality of the data was tested using the Shapiro-Wilk test. The homogeneity of variance between two groups was examined using the F test. The Brown-Forsythe test was used to examine the homogeneity of variance across multiple groups. Mortality was assessed using Fisher’s exact test. To analyze normally distributed data, ANOVA followed by a post hoc test was used to compare data from more than two groups. Nonnormally distributed data were analyzed using nonparametric tests and the Kruskal–Wallis test. Two-way repeated-measures ANOVA was used to compare behavioral data between different time points and groups. Data analysis was performed using Prism 8.4 (GraphPad) software, and differences were considered significant at P < 0.05.

3. Results

3.1. Astrocytes form the glia limitans outside the vascular endothelium and basement membrane after SAH

First, we examined the intracranial temporal and spatial localization of the glia limitans in mice after SAH. Immunofluorescence staining showed that under physiological conditions, CX43 was densely expressed around the BBB and had a similar spatial localization to the peduncle of astrocytes. Under physiological conditions, low levels of CLDN1 are widely distributed in brain tissue and do not exhibit significant aggregated distribution characteristics. At 1 day after SAH, CX43 expression was reduced, and CLDN1 was highly expressed around the BBB and colocalized with astrocyte peduncles. At 3 days after SAH, CLDN1 expression was reduced around the BBB, and CX43 expression was increased (Fig. 1A). Furthermore, we found that after SAH, CLDN1 was expressed at sites where GFAP-positive astrocytes and PDGFRβ-positive pericytes were in contact (Fig. 1B). Meanwhile, we observed the ultrastructure of the BBB with TEM. At 1 day after SAH, a large number of vesicles form within the endothelium and basement membrane, and tight junctions form between the endopods of astrocytes that wrap around the vessel. At 3 days after SAH, the tight junctions formed by the astrocyte peduncles were reduced (Fig. 1C). This suggests that after SAH, CLDN1 forms a barrier outside the vascular basement membrane and that the barrier weakens with increased disease duration. Furthermore, we cultured primary astrocytes, and after oxyhemoglobin (OxyHb) treatment, astrocyte CLDN1 expression was significantly upregulated (Fig. 1D).

Fig. 1.

Fig. 1.

Open in a new tab

Temporal and spatial positioning of glia limitans after SAH.

A, Immunofluorescent labeling of CLDN1 (claudin-1, red), CX43 (connexin 43, green), GFAP (glial fibrillary acidic protein, white), and DAPI (4′,6-diamidino-2-phenylindole, blue) in cortical areas on days 1 and 3 in the sham and SAH groups of C57BL/6 mice (scale bar, 50 μm). At 1 day after SAH, CLDN1 was densely localized to the vascular endothelium and basement membrane and was colocalized with GFAP. n = 6 per group, 1-way ANOVA, * P < 0.01 vs. sham group; † P < 0.01 vs. day 1 SAH group. B, CLDN1 (red), PDGFRβ (PDGFR-beta, green) and GFAP (white) immunofluorescent triple labeling (scale bar, 50 μm). C, TEM of cerebral microvasculature in the cortical region of C57BL/6 mice on days 1 and 3 in the sham and SAH groups (scale bar, 0.5 μm). Astrocytes (A-astrocyte end-foot) were observed wrapped around endothelial cells (E-endothelial cells) in the sham group. At 1 day after SAH, tight junctions were observed between astrocyte end-feet outside the vascular basement membrane, and vesicle formation was observed within the vascular endothelium and basement membrane (arrows-tight junctions; PMNn-polymorphonuclear neutrophils; @-vascular basement membrane; *-vesicles). The day 3 SAH 3 d group showed a decrease in tight junctions between astrocytes and the appearance of a large number of single or grouped vesicles with lymphocytes adhering to the vascular endothelium (L-lymphocytes; P-pericytes). D, Immunofluorescent labeling of DAPI (blue), CLDN1 (claudin-1, red), CX43 (green), and GFAP (white) after treatment of primary astrocytes with OxyHb (scale bar, 50 μm), n = 6 per group, t-test, ** P < 0.01 vs. Ast-Vehicle group.

3.2. Interaction between pericytes and astrocytes via the Eph/Ephrin linkage exists after SAH

To investigate the mechanism of astrocyte glia limitan formation, we designed a pericyte-astrocyte in vitro contact coculture system (Fig. 2A). After treatment of the pericyte region with OxyHb (10 μM, 24 h), quantitative PCR was used to initially screen for Eph/Ephrin isoforms that may function in the interaction between pericytes and astrocytes. We found that the mRNA levels of pericyte epha4 and astrocyte efnb2 were significantly increased after OxyHb stimulation (Fig. 2A). Furthermore, immunoblotting results demonstrated that the protein expression levels of pericyte EphA4 and astrocyte EphrinB2 were similarly significantly elevated following OxyHb treatment (Fig. 2B). In Fig. 2A and B, there were significant increases in the expression trend of EphA4/efnb2 at both the protein and mRNA levels in pericytes and astrocytes after OxyHb treatment, but the degree of increase was different, which may be due to two reasons. Under normal conditions, there may be a certain amount of protein reserves in the cells that do not function, so there may be differences in the baseline levels of mRNA and protein in the control group; however, differences in mRNA and protein level data may be related to the detection method. PCR has a high detection sensitivity, while Western blot experiments are associated with the sensitivity of the antibodies used and the detection reagents. In vivo immunofluorescence double labeling showed that EphA4- and PDGFRβ/CD13-positive pericytes and EphrinB2- and GFAP-positive astrocytes had similar spatial localizations, and both proteins were expressed at increased levels after SAH (Fig. 3A). Then, CLDN1, EphA4 and EphrinB2 immunofluorescence triple labeling confirmed that CLDN1 was expressed in the vicinity of EphA4 and EphrinB2 fluorescence colocalization and that EphA4 and EphrinB2 had a good spatiotemporal localization relationship (Fig. 3B). Finally, Western blot confirmed that EphA4 and EphrinB2 were significantly increased at 1 day after SAH in mouse brain tissue, and their expression levels decreased with increased disease duration (Fig. 3C).

Fig. 3.

Fig. 3.

Open in a new tab

Expression and localization of Eph/Ephrin in pericytes and astrocytes in vivo.

A, Immunofluorescence labeling of CD13 (white), PDGFRβ (PDGFR-beta, green), EphA4 (red), and DAPI (blue) in the cortical area and GFAP (green), EphrinB2 (red), and DAPI (blue) in the sham and day 1 SAH groups of C57BL/6 mice (scale bar, 50 μm). n = 6 per group, t-test, * P < 0.01 vs. sham group. B, Spatial localization of representative CLDN1 (red), EphA4 (green) and EphrinB2 (white) in the sham and SAH groups of C57BL/6 mice (scale bar, 50 μm). C, Western blot of mouse brain tissue to examine the expression of EphA4 and EphrinB2 proteins in the sham group and on days 1, 3 and 7 after SAH. n = 6 per group, 1-way ANOVA, @ P < 0.01 vs. sham group; † P < 0.01 vs. day 1 SAH group; ‡ P < 0.01 vs. day 3 SAH group.

3.3. Eph/Ephrin interactions can regulate glia limitans formation

We hypothesized that Eph/Ephrin interactions are involved in the regulation of glial limitan formation in astrocytes after SAH. To test this hypothesis, we pretreated primary astrocytes with nilotinib (NIL, 10 μM), a specific blocker of the EphA4 pathway, and recombinant EphA4 protein (r-EphA4, 10 μM) for 24 h and then treated astrocytes with OxyHb for 24 h(Gu et al., 2018). Immunofluorescence staining revealed that CLDN1 expression in the NIL-treated group was lower than that in the r-EphA4-treated group (Fig. 4A). The TEER was measured as well; these measurements demonstrated that the TEER of astrocytes increased at 24 h following OxyHb treatment and that this upward trend was significantly enhanced by r-EphA4 pretreatment or significantly reduced by NIL pretreatment (Fig. 4B). In the absence of OxyHb intervention, pretreatment of astrocytes with r-EphA4 alone also resulted in a significant increase in TEER values.

Fig. 4.

Fig. 4.

Open in a new tab

Effect of Eph/Ephrin interactions on the formation of glia limitans after SAH.

A, Pretreatment of astrocytes with NIL and r-EphA4 for 24 h, followed by immunofluorescence staining for CLDN1 (red), CX43 (green), GFAP (white), and DAPI (blue) after OxyHb intervention (scale bar, 50 μm), n = 6 per group, t-test, * P < 0.01 vs. NIL-treated group. B, TEER values measured in the astrocyte-Vehicle group and the NIL, r-EphA4-treated group 24 h after OxyHb intervention. n = 12 per group, 1-way ANOVA, * P < 0.01 vs. Ast-Vehicle group; † P < 0.01 vs. Ast-Vehicle+OxyHb group; ‡ P < 0.01 vs. r-EphA4 treated group. C, TEER values measured in astrocytes after transfection with EphrinB2 siRNA and EphrinB2 IREs lentivirus in a pericyte-astrocyte contact coculture system. 1-way ANOVA, § P < 0.01 vs. Ast-Vehicle group (Blank siRNA); ∥ P < 0.01 vs. Ast-EphrinB2 siRNA group; # P < 0.01 vs. Ast-Vehicle+OxyHb group (Blank siRNA); §§ P < 0.01 vs. Ast-Vehicle group (Blank IREs); ∥∥P < 0.01 vs. Ast-EphrinB2 IREs group; ## P < 0.01 vs. Ast-Vehicle+OxyHb group (Blank IREs). D and E, CLDN1 (red), CX43 (green), GFAP (white), DAPI (blue) cortical immunofluorescence staining and representative Evans blue (red) autofluorescence after in vivo intervention using EphrinB2 IREs and EphrinB2 siRNA for the day 1 SAH group (scale bar, 50 μm), n = 6 per group, 1-way ANOVA, ** P < 0.01 vs. SAH-Vehicle group; †† P < 0.01 vs. SAH-EphrinB2 IREs group.

Furthermore, we transfected astrocytes with lentiviruses overexpressing EphrinB2 and EphrinB2 siRNA (EphrinB2 IREs and EphrinB2 siRNA). TEER measurements revealed significantly increased TEER values in the EphrinB2 IRES–transfected group after OxyHb intervention, while significantly decreased values were found in the EphrinB2 siRNA-transfected group (Fig. 4C). In parallel, we injected lentivirus with a low neuroinflammatory response in the lateral ventricle one week prior to SAH surgery to achieve in vivo levels of EphrinB2 overexpression (EphrinB2 IREs) and interference (EphrinB2 siRNA). Immunofluorescence staining at 1 day after SAH confirmed that CLDN1 expression levels were lower in the EphrinB2 interference group than in the EphA4 overexpression group (Fig. 4D). We then found that Evans blue fluorescence in the Vehicle group was confined to the vasculature, but Evans blue fluorescence in the SAH-EphrinB2 siRNA group leaked into the brain parenchyma outside the perivascular gap, whereas this was not the case in the SAH-EphrinB2 IRE group (Fig. 4E). In summary, we can speculate that the EphA4/EphrinB2 pathway has an important regulatory role in the formation of astrocyte glia limitans; specifically, EphA4 activates EphrinB2 on astrocytes and promotes the formation of glia limitans after SAH.

3.4. Inhibition of the Eph/Ephrin pathway attenuates the secretion of cytokines from astrocytes and pericytes

We used a pericyte-astrocyte contact coculture system to further explore the mechanisms by which Eph/Ephrin regulates glia limitans. After treatment of the pericyte region with OxyHb, the relative levels of cytokines and chemokines in the supernatant of the pericyte and astrocyte regions were measured using inflammatory factor antibody arrays and screened for ten astrocytic proteins (CD54, CXCL10, CXCL1, CCL2, CCL5, TIMP1, CCL3, M-CSF, IL-1α, and IL-1β) and eight pericyte proteins (CD54, CXCL10, CXCL1, CCL2, CCL5, TIMP1, CXCL9, and CXCL12) that significantly increased inflammatory factors (Supplemental Fig. 2). Next, we bidirectionally inhibited the EphA4/EphrinB2 pathway in the contact coculture system by transfecting pericytes and astrocytes with EphA4 siRNA lentivirus and EphrinB2 siRNA lentivirus, respectively (Fig. 5A & Fig. 5C). After 24 h, Western blotting confirmed a significant level of interference and a significant downregulation of EphrinB2 expression in astrocytes in contact with them (Fig. 5A). Furthermore, quantitative ELISA analysis demonstrated significantly lower levels of inflammatory factors secreted by the pericytes and the astrocytes in contact with them in the EphA4 siRNA transfected group than in the Vehicle group (Fig. 5B and Supplemental Fig. 3). However, we used EphrinB2 siRNA-transfected astrocytes in contact coculture with pericytes, and Western blotting confirmed significant levels of interference and a significant downregulation of EphA4 levels in pericytes in contact with transfected astrocytes (Fig. 5C). The levels of inflammatory factors secreted by glial cells and pericytes in contact with EphrinB2 siRNA were significantly lower than those in the vehicle group (Fig. 5D and Supplemental Fig. 4).

Fig. 5.

Fig. 5.

Open in a new tab

Inhibition of the Eph/Ephrin pathway and analysis of the effect on inflammatory factor secretion from pericytes and astrocytes.

A, Schematic diagram of the coculture of EphA4 siRNA lentivirus-transfected pericytes, as well as Western blot and quantitative analysis of pericyte EphA4 and astrocyte EphrinB2 in a contact coculture system. OxyHb was added to the pericyte region only. n = 6 per group, 1-way ANOVA, * P < 0.01 vs. Peri-Vehicle group; † P < 0.01 vs. Peri-EphA4 siRNA group; ‡ P < 0.01 vs. Peri-Vehicle+OxyHb group; ** P < 0.01 vs. Ast-Vehicle group; †† P < 0.01 vs. Ast-EphA4 siRNA group; ‡‡ P < 0.01 vs. Ast-Vehicle+OxyHb group. B, Radar plots were used to summarize quantitative ELISA analysis results of multiple inflammatory factors in pericyte regional supernatants and astrocyte regional supernatants in the EphA4 siRNA coculture system (n = 3 per group). C, Schematic diagram of EphrinB2 siRNA lentivirus-transfected astrocytes in coculture, as well as Western blot and quantitative analysis of pericyte EphA4 and astrocyte EphrinB2 in a coculture system. OxyHb interferes with pericyte regions only. n = 6 per group, 1-way ANOVA, * P < 0.01 vs. Peri-Vehicle group; † P < 0.05 vs. Peri-EphrinB2 siRNA group ‡ P < 0.01 vs. Peri-Vehicle+OxyHb group; ** P < 0.05 vs. Ast-Vehicle group; †† P < 0.05 vs. Ast-EphrinB2 siRNA group; ‡‡ P < 0.05 vs. Ast-Vehicle+OxyHb group. D. Radar plots were used to summarize quantitative ELISA analysis of multiple inflammatory factors in pericyte regional supernatants and astrocyte regional supernatants in the EphrinB2 siRNA coculture system (n = 3 per group).

3.5. Enhancement of the Eph/Ephrin pathway promotes astrocytic and pericytic cytokine secretion and alters the phenotype of astrocytes

To validate the function of the Eph/Ephrin pathway more fully, we transfected pericytes and astrocytes with EphA4 IREs lentivirus and EphrinB2 IREs lentivirus, respectively, resulting in bidirectional overexpression of the EphA4/EphrinB2 pathway (Fig. 6A and Fig. 6C). After 24 h of treatment with OxyHb, Western blot analysis demonstrated a significant upregulation of EphA4 levels in pericytes and EphrinB2 in astrocytes in contact with them (Fig. 6A). ELISA results demonstrated that EphA4 IREs-transfected pericytes and their associated astrocytes secreted significantly more inflammatory factors than the Vehicle group (Fig. 6B and Supplemental Fig. 5). However, Western blot also demonstrated that EphrinB2 was significantly upregulated in EphrinB2 IREs lentivirally transfected astrocytes and that the expression level of EphA4 was significantly higher in pericytes in contact with them (Fig. 6C). The levels of inflammatory factors secreted by EphrinB2 IREs were significantly higher than those in the Vehicle group (Fig. 6D and Supplemental Fig. 6). Furthermore, we used transcriptome microarrays to analyze the phenotypic alterations of pericytes transfected with EphA4 siRNA and EphA4 IRES on the astrocytes they interacted with (Supplemental Fig. 8). Finally, Western blot confirmed that Jak2/STAT3 is a downstream molecule of the EphA4/EphrinB2 molecule interaction in the pericyte-astrocyte coculture system (Supplemental Fig. 7).

Fig. 6.

Fig. 6.

Open in a new tab

Activation of the Eph/Ephrin pathway and analysis of the effect on inflammatory factor secretion from pericytes and astrocytes.

A, Schematic diagram of the coculture of EphA4 IRE lentivirus-transfected pericytes, as well as Western blot and quantitative analyses of pericyte EphA4 and astrocyte EphrinB2 in a contact coculture system. OxyHb was added to the pericyte region only. n = 6 per group, 1-way ANOVA, * P < 0.05 vs. the Peri-Vehicle group; † P < 0.05 vs. the Peri-EphA4 IRE group; ** P < 0.01 vs. the Ast-Vehicle group; †† P < 0.05 vs. the Ast-EphA4 IRE group. B, Radar plots were used to summarize quantitative ELISA analysis of multiple inflammatory factors in pericyte regional supernatants and astrocyte regional supernatants in the EphA4 IRE coculture system (n = 3 per group). C, Schematic diagram of EphrinB2 IRE lentivirus-transfected astrocytes in coculture, as well as Western blot and quantitative analysis of pericyte EphA4 and astrocyte EphrinB2 in a coculture system. OxyHb interferes with pericyte regions only. n = 6 per group, t-test, * P < 0.01 vs. Peri-Vehicle group; † P < 0.05 vs. Peri-EphrinB2 IREs group; ‡ P < 0.01 vs. Peri-Vehicle+OxyHb group; ** P < 0.01 vs. Ast-Vehicle group; †† P < 0.01 vs. Ast-EphrinB2 IREs group; ‡‡ P < 0.01 vs. Ast-Vehicle+OxyHb group. D, Radar plots were used to summarize quantitative ELISA analysis of multiple inflammatory factors in pericyte regional supernatants and astrocyte regional supernatants in the EphrinB2 IREs coculture system (n = 3 per group).

3.6. Glia limitans block inflammatory cell infiltration and improve neurological function after SAH

Finally, we assessed whether the formation of glia limitans after SAH could attenuate the infiltration of peripheral inflammatory cells into the brain parenchyma. Immunofluorescence staining revealed that CD45-positive cells in the vehicle group had similar spatial localization to CLDN1 in the vascular endothelium and basement membrane. More CD45-positive cells were found in the EphrinB2 siRNA group than in the EphrinB2 IRE group and were widely distributed throughout the brain parenchyma (Fig. 7A). Flow cytometry analysis revealed that the EphrinB2 siRNA group had more NK-cell (CD45highCD3−NK1.1+) infiltrates than the Vehicle and EphrinB2 IREs groups at 1 day after SAH, and this trend was more pronounced at 3 days after SAH (Fig. 7B, C and D). We also found that at 1 and 3 days after SAH, compared with the Sham and Vehicle groups, the SAH-Vehicle and SAH-EphrinB2 siRNA groups both showed severe neurological impairment, but the SAH-EphrinB2 IRE group was superior to the SAH-Vehicle group (Fig. 6E).

Fig. 7.

Fig. 7.

Open in a new tab

Function of the glia limitans in blocking peripheral inflammatory cell infiltration after SAH.

A, Immunofluorescence staining of cortical CLDN1 (red), CD45 (green), and DAPI (blue) for EphrinB2 IREs and EphrinB2 siRNA in the in vivo SAH model 1 day after intervention (scale bar, 50 μm). n = 6, each group, 1-way ANOVA, † P < 0.01 vs. Sham-Vehicle group; ‡ P < 0.01 vs. SAH-EphrinB2 IREs group. B, C and D, Flow cytometry plots (B) and quantification (C and D) showing the expression of infiltrating inflammatory cells and NK cells obtained from the brain tissue of mice 1 and 3 days after SAH. n = 6, each group, t-test, * P < 0.05 vs. SAH-Vehicle group. E, Modified Garcia score and balance beam score in the in vivo SAH model 1 and 3 days after intervention with EphrinB2 IREs and EphrinB2 siRNA in mice. n = 6, each group, 1-way ANOVA, * P < 0.01 vs. sham group; † P < 0.01 vs. sham-vehicle group; ‡ P < 0.01 vs. SAH-vehicle group; § P < 0.01 vs. SAH-EphrinB2 siRNA group.

4. Discussion

In this study (Fig. 8), we demonstrate that pericytes interact with astrocytes via EphA4/EphrinB2 to regulate the formation of glia limitans. We have identified for the first time that the glia limitans is an important emergency protective mechanism for astrocytes during the early course of SAH, forming a second barrier after BBB disruption and preventing infiltration of peripheral immune cells into the brain parenchyma, especially by NK cells. Furthermore, the EphA4/EphrinB2 pathway promotes inflammatory factor secretion from pericytes and astrocytes to the extent that astrocytes form glia limitans.

Fig. 8.

Fig. 8.

Open in a new tab

Schematic illustration of the present study.

In past studies, glial limitan formation by reactive astrocytes has been shown to play an important role in barrier function in chronic neuroinflammation, for example, in experimental autoimmune encephalomyelitis, repetitive traumatic brain injury, and optic nerve injury(Mason et al., 2021; Mayrhofer et al., 2021). The present study reports for the first time that neuroinflammation due to acute CNS injury drives the formation of the glia limitans in the acute phase and that the barrier function of astrocytes is a more general stress-protective mechanism in the CNS. In studies of BBB disruption after SAH, previous studies have focused primarily on the tight junctions of endothelial cells and the barrier structures formed by the basement membrane(Zeyu et al., 2021). Our study is the first to take a more in-depth look at the structure and integrity of the BBB after SAH, extending the examination to the formation of the glia limitans from astrocyte end-feet outside the vascular basement membrane. We observed a high density of glial limitan formation outside the endothelium and basement membrane on the first day after SAH. TEM showed that even though a large number of vesicles had formed in the endothelium and basement membrane, the outer basement membrane was surrounded by a ring of dense glia limitans, forming a “double barrier”. Immunofluorescence double labeling confirmed that CLDN1 had good spatial and temporal localization with CD45-positive inflammatory cells and that CLDN1 prevented the infiltration of inflammatory cells into the brain parenchyma. The expression of glia limitans outside the vascular basement membrane was downregulated on the third day after SAH, and the barrier properties were diminished, leading to a significant increase in inflammatory cells infiltrating into the brain parenchyma. Thus, we demonstrate that the neuropathological process of BBB disruption after SAH undergoes the spatial properties of a double barrier system whose function is regulated by interactions between pericytes and astroglia.

Pericytes have been shown to play an important role in the regulation of the BBB, and disruption of these roles may lead to BBB dysfunction and outbreaks of neuroinflammation. Recent studies have shown that Perlecan in the basement membrane of endothelial cells can induce pericytes to secrete a variety of inflammatory factors and subsequently recruit more pericytes and promote their differentiation and maturation to participate in the repair of the BBB structure(Nakamura et al., 2019). A previous study by the same group suggested that pericytic CypA expression was significantly upregulated after SAH, leading to the disruption of the BBB(Pan et al., 2020). Therefore, the mechanism by which pericytes participate in BBB remodeling after acute CNS injury has not been elucidated. Our study confirms that pericytes secrete inflammatory factors after SAH and interact with astrocytes via EphA4/EphrinB2 to secrete inflammatory factors that induce neuroinflammation and promote glial limitan formation. Previous studies have reported an immunosurveillance function of lung microvascular pericytes, which detect endothelial cell injury to secrete inflammatory factors and participate in the recruitment of inflammatory cells in the peripheral circulation(Hung et al., 2017). Our study reveals an immunosurveillance pattern of pericytes in the CNS. Specifically, following the breakdown of the BBB, pericytes and astrocytes dramatically secrete inflammatory factors that induce the formation of neuroinflammation; however, the immune response of pericytes during the acute phase of SAH promotes the formation of glia limitans, a second barrier outside the endothelial basement membrane, preventing infiltration of peripheral inflammatory cells into the brain parenchyma. In summary, pericytes determine the balance between the proinflammatory response and barrier function during the acute phase of SAH, and the direction in which the balance tilts is an important predictor of BBB remodeling and the level of neurological function.

The Eph/Ephrin system is a classical cell–cell interaction signaling system, and previous studies have demonstrated that the Eph/Ephrin interaction has an important regulatory role in the development of vascular and BBB function(Malik and Di Benedetto, 2018). Excessive activation of EphA4 has been shown to exacerbate BBB destruction and neuronal death, but the specific cell–cell interactions and mechanisms involved in BBB remodeling have not been elucidated(Fan et al., 2017). In this study, the Eph/Ephrin family was screened and validated at the mRNA and protein expression levels using a pericyte-astrocyte coculture system to screen for EphA4/EphrinB2 interactions, and in vivo spatiotemporal validation was performed on frozen sections, demonstrating that EphA4 on pericytes and EphrinB2 on astrocytes interact after SAH. Furthermore, the important regulatory function of the EphA4/EphrinB2 system in inflammatory secretion after SAH was verified in both directions by overexpression and interference with EphA4 and EphrinB2.

Pericytes and astrocytes are involved in the establishment and maintenance of the BBB and are essential players in its disruption and remodeling after SAH. Horng et al. found that proinflammatory factors such as IL-1β could induce reactive astrocytes to form barrier structures characterized by tight junction proteins such as CLDN-1 and inhibit peripheral inflammatory cells from entering the brain parenchyma (Horng et al., 2017). We used a pericyte-astrocyte contact coculture model to simulate the BBB microenvironment and stimulated pericytes using OxyHb to simulate blood-based cerebrospinal fluid. OxyHb was shown to stimulate pericytes to secrete a range of chemokines and to induce astrocyte activation and the secretion of multiple inflammatory factors. Furthermore, intervention of EphA4/EphrinB2 signaling using lentiviral transfection can modulate the inflammatory cascade response between pericytes and astrocytes, which in turn affects the formation of glia limitans. It is suggested that pericytes induce the production of glia limitans by reactive astrocytes, possibly through EphA4/EphrinB2 signaling that modulates the inflammatory factor groupings in the BBB microenvironment.

However, the present study has several limitations. First, the present study only focused on the protective role of glia limitans in the early course after SAH to determine the physiological role of this stress protective barrier, and strategies to sustain this secondary protective barrier and the long-term effects should be evaluated. Furthermore, we did not clarify which specific inflammatory factor in the inflammatory factor group induced glial limitan formation, which requires further investigation. Third, limited to the extremely low post-SAH survival rate in Pdgfrβ−/− mice, genetically modified mice were not used to further validate the induction function of pericytes.

5. Conclusions

Taken together, the present findings demonstrate that the glia limitans is an essential protective barrier after SAH, limiting the infiltration of peripheral inflammatory cells into the brain parenchyma. This function is closely related to the regulation of EphA4/EphrinB2 signaling between pericytes and astrocytes.

Supplementary Material

Supplementary data

Funding

This work was supported by National Natural Science Foundation of China (82030036 to Hua Feng, 82071397 to Xiaochuan Sun), State Key Laboratory of Trauma, Burn and Combined Injury (SKLYQ202002 to Yujie Chen), and NIH (R01 NS117179 to John H. Zhang).

Abbreviation:

SAH

subarachnoid hemorrhage

BBB

blood–brain barrier

TEM

transmission electron microscopy

TEER

transendothelial electrical resistance

OxyHb

oxyhemoglobin

NIL

nilotinib

Footnotes

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Ethics approval and consent to participate

All experiments are reported in compliance with the Animal Research: Reporting in vivo Experiments (ARRIVE) guidelines. The experimental protocols were approved by the Laboratory Animal Welfare and Ethics Committee of Third Military Medical University (AMU-WEC2020763) and performed according to the Guide for the Care and Use of Laboratory Animals.

Availability of data and materials

All data generated or analyzed during this study are included in this published article. The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.expneurol.2022.114293.

Data availability

Data will be made available on request.

References

  1. Chen Y, Zhang Y, Tang J, Liu F, Hu Q, Luo C, Tang J, Feng H, Zhang JH, 2015. Norrin protected blood-brain barrier via frizzled-4/beta-catenin pathway after subarachnoid hemorrhage in rats. Stroke 46, 529–536. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Chen Y, Galea I, Macdonald RL, Wong GKC, Zhang JH, 2022. Rethinking the original changes in subarachnoid haemorrhage: focusing on real-time metabolism during early brain injury. EBioMedicine 83, 104223. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Claassen J, Park S, 2022. Spontaneous subarachnoid haemorrhage. Lancet 400, 846–862. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Coelho-Santos V, Berthiaume AA, Ornelas S, Stuhlmann H, Shih AY, 2021. Imaging the construction of capillary networks in the neonatal mouse brain. Proc. Natl. Acad. Sci. U. S. A 118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Daneman R, Zhou L, Kebede AA, Barres BA, 2010. Pericytes are required for blood-brain barrier integrity during embryogenesis. Nature 468, 562–566. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Deng Y, Jiang X, Deng X, Chen H, Xu J, Zhang Z, Liu G, Yong Z, Yuan C, Sun X, Wang C, 2020. Pioglitazone ameliorates neuronal damage after traumatic brain injury via the PPARgamma/NF-kappaB/IL-6 signaling pathway. Genes Dis. 7, 253–265. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Fan R, Enkhjargal B, Camara R, Yan F, Gong L, ShengtaoYao, Tang, Chen Y, Zhang JH, 2017. Critical role of EphA4 in early brain injury after subarachnoid hemorrhage in rat. Exp. Neurol 296, 41–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Gu S, Fu WY, Fu AKY, Tong EPS, Ip FCF, Huang X, Ip NY, 2018. Identification of new EphA4 inhibitors by virtual screening of FDA-approved drugs. Sci. Rep 8, 7377. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Hirai H, Maru Y, Hagiwara K, Nishida J, Takaku F, 1987. A novel putative tyrosine kinase receptor encoded by the eph gene. Science 238, 1717–1720. [DOI] [PubMed] [Google Scholar]
  10. Horng S, Therattil A, Moyon S, Gordon A, Kim K, Argaw AT, Hara Y, Mariani JN, Sawai S, Flodby P, Crandall ED, Borok Z, Sofroniew MV, Chapouly C, John GR, 2017. Astrocytic tight junctions control inflammatory CNS lesion pathogenesis. J. Clin. Invest 127, 3136–3151. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Hung CF, Mittelsteadt KL, Brauer R, McKinney BL, Hallstrand TS, Parks WC, Chen P, Schnapp LM, Liles WC, Duffield JS, Altemeier WA, 2017. Lung pericyte-like cells are functional interstitial immune sentinel cells. Am. J. Phys. Lung Cell. Mol. Phys 312, L556–L567. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Kania A, Klein R, 2016. Mechanisms of ephrin-Eph signalling in development, physiology and disease. Nat. Rev. Mol. Cell Biol 17, 240–256. [DOI] [PubMed] [Google Scholar]
  13. Keep RF, Andjelkovic AV, Xiang J, Stamatovic SM, Antonetti DA, Hua Y, Xi G, 2018. Brain endothelial cell junctions after cerebral hemorrhage: changes, mechanisms and therapeutic targets. J. Cereb. Blood Flow Metab 38, 1255–1275. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Li L, Luo Q, Shang B, Yang X, Zhang Y, Pan Q, Wu N, Tang W, Du D, Sun X, Jiang L, 2022. Selective activation of cannabinoid receptor-2 reduces white matter injury via PERK signaling in a rat model of traumatic brain injury. Exp. Neurol 347, 113899. [DOI] [PubMed] [Google Scholar]
  15. Mae MA, He L, Nordling S, Vazquez-Liebanas E, Nahar K, Jung B, Li X, Tan BC, Chin Foo J, Cazenave-Gassiot A, Wenk MR, Zarb Y, Lavina B, Quaggin SE, Jeansson M, Gu C, Silver DL, Vanlandewijck M, Butcher EC, Keller A, Betsholtz C, 2021. Single-cell analysis of blood-brain barrier response to pericyte loss. Circ. Res 128, e46–e62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Malik VA, Di Benedetto B, 2018. The blood-brain barrier and the EphR/Ephrin system: perspectives on a link between neurovascular and neuropsychiatric disorders. Front. Mol. Neurosci 11, 127. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Mason HD, Johnson AM, Mihelson NA, Mastorakos P, McGavern DB, 2021. Glia limitans superficialis oxidation and breakdown promote cortical cell death after repetitive head injury. JCI Insight 6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Mayrhofer F, Dariychuk Z, Zhen A, Daugherty DJ, Bannerman P, Hanson AM, Pleasure D, Soulika A, Deng W, Chechneva OV, 2021. Reduction in CD11c(+) microglia correlates with clinical progression in chronic experimental autoimmune demyelination. Neurobiol. Dis 161, 105556. [DOI] [PubMed] [Google Scholar]
  19. Mora P, Hollier PL, Guimbal S, Abelanet A, Diop A, Cornuault L, Couffinhal T, Horng S, Gadeau AP, Renault MA, Chapouly C, 2020. Blood-brain barrier genetic disruption leads to protective barrier formation at the Glia Limitans. PLoS Biol. 18, e3000946. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Nakamura K, Ikeuchi T, Nara K, Rhodes CS, Zhang P, Chiba Y, Kazuno S, Miura Y, Ago T, Arikawa-Hirasawa E, Mukouyama YS, Yamada Y, 2019. Perlecan regulates pericyte dynamics in the maintenance and repair of the blood-brain barrier. J. Cell Biol 218, 3506–3525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Pan P, Zhao H, Zhang X, Li Q, Qu J, Zuo S, Yang F, Liang G, Zhang JH, Liu X, He H, Feng H, Chen Y, 2020. Cyclophilin a signaling induces pericyte-associated blood-brain barrier disruption after subarachnoid hemorrhage. J. Neuroinflammation 17, 16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Quintana FJ, 2017. Astrocytes to the rescue! Glia limitans astrocytic endfeet control CNS inflammation. J. Clin. Invest 127, 2897–2899. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Rass V, Ianosi BA, Wegmann A, Gaasch M, Schiefecker AJ, Kofler M, Lindner A, Addis A, Almashad SS, Rhomberg P, Pfausler B, Beer R, Gizewski ER, Thome C, Helbok R, 2019. Delayed resolution of cerebral edema is associated with poor outcome after nontraumatic subarachnoid hemorrhage. Stroke 50, 828–836. [DOI] [PubMed] [Google Scholar]
  24. Ru X, Qu J, Li Q, Zhou J, Huang S, Li W, Zuo S, Chen Y, Liu Z, Feng H, 2021. MiR-706 alleviates white matter injury via downregulating PKCalpha/MST1/NF-kappaB pathway after subarachnoid hemorrhage in mice. Exp. Neurol 341, 113688. [DOI] [PubMed] [Google Scholar]
  25. Salvucci O, Maric D, Economopoulou M, Sakakibara S, Merlin S, Follenzi A, Tosato G, 2009. EphrinB reverse signaling contributes to endothelial and mural cell assembly into vascular structures. Blood 114, 1707–1716. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Sanmarco LM, Polonio CM, Wheeler MA, Quintana FJ, 2021. Functional immune cell-astrocyte interactions. J. Exp. Med 218. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Segarra M, Aburto MR, Acker-Palmer A, 2021. Blood-brain barrier dynamics to maintain brain homeostasis. Trends Neurosci. 44, 393–405. [DOI] [PubMed] [Google Scholar]
  28. Solar P, Zamani A, Lakatosova K, Joukal M, 2022. The blood-brain barrier and the neurovascular unit in subarachnoid hemorrhage: molecular events and potential treatments. Fluids Barriers CNS 19, 29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Weimer JM, Jones SE, Frontera JA, 2017. Acute cytotoxic and vasogenic edema after subarachnoid hemorrhage: a quantitative MRI study. AJNR Am. J. Neuroradiol 38, 928–934. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Wu Y, Pang J, Peng J, Cao F, Guo Z, Jiang L, Teng Z, Huang Z, Cheng C, Jiang Y, Sun X, 2019. Apolipoprotein E deficiency aggravates neuronal injury by enhancing neuroinflammation via the JNK/c-Jun pathway in the early phase of experimental subarachnoid hemorrhage in mice. Oxidative Med. Cell. Longev 2019, 3832648. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Zeyu Z, Yuanjian F, Cameron L, Sheng C, 2021. The role of immune inflammation in aneurysmal subarachnoid hemorrhage. Exp. Neurol 336, 113535. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary data
NIHMS1933604-supplement-Supplementary_data.pdf (3.6MB, pdf)

Data Availability Statement

Data will be made available on request.

RESOURCES