Netrin-5 Preserves Blood-Brain Barrier Integrity via Wnt3a/β-Catenin Pathway Activation in Murine Cerebral Ischemia
Yitian Chen, Li Liu, Yang Ming, Lilei Peng, Ligang Chen
TL;DR
Netrin-5 helps protect the blood-brain barrier in stroke by activating a specific signaling pathway, suggesting it could be a new treatment.
Contribution
Netrin-5's role in preserving the blood-brain barrier via Wnt3a/β-catenin signaling is newly identified in ischemic stroke models.
Findings
Netrin-5 delivery reduced brain damage and improved outcomes in stroke mice.
Netrin-5 activates Wnt3a/β-catenin signaling and restores tight junction proteins.
Wnt3a knockdown negates Netrin-5's protective effects, confirming its essential role.
Abstract
Blood-brain barrier compromise represents a pivotal pathological mechanism in ischemic stroke, driving neurological deterioration. Netrin-5, an axon guidance protein family member, demonstrates regulatory potential for BBB integrity. Employing middle cerebral artery occlusion (MCAO) mice and oxygen-glucose deprivation/reperfusion (OGD/R) in human brain microvascular endothelial cells (HBMVECs), we found that Netrin-5 was significantly downregulated in the murine cortex post-MCAO and was also downregulated in HBMVECs upon OGD/R exposure. Adenoviral Netrin-5 delivery in MCAO mice attenuated cerebral infarction, improved functional outcomes, reduced edema, and preserved BBB integrity, evidenced by diminished Evans blue extravasation and albumin leakage. Furthermore, Netrin-5 restored tight junction protein ZO-1 expression and activated Wnt3a/β-catenin signaling. In HBMVECs, Netrin-5…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10- —https://doi.org/10.13039/501100001809National Natural Science Foundation of China (National Science Foundation of China)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAxon Guidance and Neuronal Signaling · Barrier Structure and Function Studies · Angiogenesis and VEGF in Cancer
Introduction
Accounting for roughly 87% of all strokes, ischemic stroke stands as a predominant source of chronic disability and death worldwide, which strains healthcare infrastructure and families [1]. The ischemic cascade involves multifaceted pathological events, among which BBB failure emerges as an early critical determinant of secondary injury, including vasogenic edema, hemorrhagic conversion, and neuronal loss [2, 3]. The BBB—consisting of endothelial cells sealed by tight junction complexes, surrounded by astrocyte end-feet processes, and supported by pericytes—maintains CNS homeostasis through selective molecular exchange [4, 5]. Tight junction complexes, featuring zonula occludens-1 (ZO-1), occludin, and claudins, govern BBB impermeability [6]. Ischemia-induced junctional disruption increases paracellular flux, permitting inflammatory mediators, blood constituents, and plasma proteins to infiltrate brain parenchyma, amplifying ischemic pathology [7, 8].
Accumulating evidence implicates diverse signaling pathways in ischemic BBB regulation. The Wnt/β-catenin cascade, notably, serves as a master regulator of cerebrovascular development and barrier maintenance [9–11]. Wnt3a/β-catenin activation upregulates tight junction proteins, fortifying BBB stability [9]. Conversely, pathway inhibition exacerbates barrier permeability and ischemic injury [12]. Consequently, targeting Wnt3a/β-catenin signaling presents a rational strategy for preserving junctional integrity during stroke.
Netrin-5, a member of the netrin family initially identified for its role in embryonic axonal guidance [13], signals via receptors such as DCC and UNC5 [13] and exhibits pleiotropic functions in mature tissues—including vascular remodeling, immune modulation, and cytoprotection [14–16]. For instance, Netrin-1 has been shown to confer ischemic neuroprotection by maintaining blood–brain barrier (BBB) integrity via KLF2-mediated occludin preservation and suppression of IL-6/Cxcl1 [17]. While other netrin family members, such as Netrin-1 and Netrin-4, have been implicated in vascular biology, their roles in cerebral ischemia remain incompletely understood [18, 19]. However, the function of Netrin-5 in ischemic stroke, particularly in relation to BBB regulation, has not yet been explored.
Given BBB dysfunction’s centrality in stroke pathogenesis and netrins’ emerging vascular roles, we investigated Netrin-5’s impact on BBB integrity and underlying mechanisms. We hypothesized that Netrin-5 activates Wnt3a/β-catenin signaling to preserve BBB function. Employing murine MCAO and HBMVEC OGD/R models, we elucidated Netrin-5’s regulatory role in BBB integrity and its mechanistic linkage to Wnt3a/β-catenin activation, offering novel insights for therapeutic development.
Materials and methods
Ethical approval and animal handling
Male C57BL/6 mice (8–10 weeks, 22–25 g) were procured from Beijing Vital River Laboratory Animal Technology Co., Ltd. (SCXK2021-0006). Animals were maintained in SPF facilities under regular environmental conditions. All experimental protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of the Affiliated Hospital of Southwest Medical University (Protocol IACUC-2023-042) and strictly followed the NIH Guide for the Care and Use of Laboratory Animals (8th edition). Stringent measures were implemented to minimize subject distress and optimize animal usage, with group sizes determined through power analysis (α = 0.05, power = 0.8) to ensure statistical validity.
MCAO model
Transient focal cerebral ischemia was induced using an intraluminal filament approach, following established protocols with minor modifications. Briefly, surgical anesthesia was induced and maintained with 1.5% isoflurane (RWD Life Science, R510-22) delivered via a precision rodent anesthesia system (RWD Life Science, R580) using a carrier gas mixture of 70% O₂ and 30% N₂O. Core body temperature was maintained at 37.0 ± 0.5 °C throughout all procedures using an automated temperature control system (RWD Life Science, R503). A midline cervical incision (approximately 10 mm) was made to expose the left carotid vasculature, including the common carotid artery (CCA), external carotid artery (ECA), and internal carotid artery (ICA). The ECA was permanently ligated distal to its bifurcation with a 6-0 silk suture (Ailee, 1657). An arteriotomy was created in the ECA stump, through which a silicone-coated 6-0 nylon monofilament suture (Doccol Corporation, 602156PK10; tip diameter 0.22–0.24 mm, length 10 mm) was advanced into the ICA. The filament was advanced until mild resistance was encountered (approximately 9.0 ± 0.5 mm from the CCA bifurcation), indicating occlusion at the origin of the middle cerebral artery (MCA). After 60 min of ischemia, the filament was withdrawn to initiate reperfusion. The proximal segment of the ECA near the arteriotomy site was then ligated, and the cervical incision was closed with 4-0 silk sutures (Ailee, 1655). Sham-operated control animals underwent identical surgical exposure and vessel manipulation but did not undergo filament insertion. Neurological deficits were assessed 24 h after reperfusion using the modified Bederson scale; only mice with a score ≥1 were included in subsequent analyses to confirm successful ischemia induction.
Cell culture conditions
HBMVECs (ScienCell Research Laboratories, #1000) were cultured in Endothelial Cell Medium (ScienCell, #1001) supplemented with 5% FBS (Gibco, #10437-028), 1% endothelial cell growth supplement (ECGS; ScienCell, #1052), and penicillin-streptomycin solution (Gibco, #15140-122). SK-N-SH neuroblastoma cells (ATCC, HTB-11) were maintained in Minimum Essential Medium (MEM; Gibco, #11095-080) containing 10% FBS and 1% penicillin-streptomycin. Subculturing was performed using 0.25% trypsin-EDTA (Gibco, #25200-056) upon reaching 80–90% confluence. Cells from passages 3 to 5 were exclusively used in experimental procedures to ensure phenotypic stability.
Adenoviral vector preparation and delivery
First-generation replication-incompetent adenoviral vectors were constructed in the pAdEasy-1/pAdTrack-CMV system (Agilent Technologies) with co-expression of GFP as a transduction reporter. For in vivo (murine) experiments, vectors expressing murine Netrin-5 (Ad-Netrin-5), murine Netrin-5-targeted shRNA (sequence: 5′-GCAGCTGAAGATGAAGATT-3′), and a non-targeting scrambled control (Ad-NC) were synthesized by GenScript (Nanjing, China). Viral particles were purified by cesium chloride gradient centrifugation and titered to ≥1 × 10¹⁰ PFU/mL. For in vitro (human HBMVEC) experiments, separate adenoviral vectors expressing the human Netrin-5 cDNA (Ad-Netrin-5, used for overexpression) and human-specific shRNAs targeting human WNT3A (sequence: e.g., 5′-GCTACCTGGACTATCTGAA-3′ or validated clone) and human NTN5, together with a matched scrambled control, were custom-synthesized by OriGene Technologies (Rockville, MD, USA) using the identical pAdEasy-1/pAdTrack-CMV backbone to ensure comparable transduction efficiency.
In vivo transduction: Mice were randomly allocated (n = 8/group) using computer-generated randomization to one of four cohorts: vehicle, Ad-Netrin-5 alone, MCAO, or MCAO + Ad-Netrin-5. For knockdown studies, additional groups included: vehicle, Ad-Netrin-5 shRNA, MCAO, and MCAO + Ad-Netrin-5 shRNA. Ad-Netrin-5 (5 μL containing 1 × 10⁹ PFU) was stereotactically injected into the left cerebral cortex. Injections utilized a Hamilton 701RN microsyringe with a 33-gauge needle (7803-05) mounted on a Stoelting 51600 frame. Stereotaxic coordinates from bregma: anterior-posterior (AP) + 0.5 mm, mediolateral (ML) −2.0 mm, dorsoventral (DV) −1.5 mm. The infusion proceeded at 0.5 μL/min, followed by a 5 min needle dwell time to minimize reflux. Viral delivery occurred 48 h before MCAO surgery to ensure adequate transgene expression at the time of ischemia.
In vitro transduction: HBMVECs were plated in 6-well plates (5 × 10⁵ cells/well). At approximately 70% confluence, cells were exposed to Ad-Netrin-5 or Ad-Wnt3a shRNA or Ad-Netrin-5 shRNA at an MOI of 50 in serum-free Endothelial Cell Medium (ScienCell, 1001) during a 4 h incubation at 37 °C. The medium was then exchanged with complete growth medium. Transduction efficiency was confirmed by western blot analysis.
OGD/R model
To simulate cerebral ischemia-reperfusion injury, HBMVECs were subjected to combined oxygen-glucose deprivation. Briefly, confluent monolayers underwent dual rinsing with glucose-deficient DMEM (Gibco, 11966-025) to eliminate residual glucose. Cultures were subsequently transferred to an airtight hypoxia chamber (Billups-Rothenberg MIC-101) containing glucose-free DMEM. The chamber atmosphere was rapidly exchanged with a pre-humidified gas mixture (95% N₂, 5% CO₂; flow rate 2 L/min) for 10 min to achieve anoxic conditions (O₂ < 0.1%) and maintained at 37 °C for 4 h (OGD phase). Following OGD, the glucose-deficient medium was aspirated and substituted with complete Endothelial Cell Medium. Cultures were promptly returned to normal atmospheric conditions (95% air, 5% CO₂) for 24 h to model reperfusion (R phase). Normoxic control cells remained in complete medium under standard incubation conditions throughout the experiment.
In vivo functional and structural assessments infarct volume quantification
Twenty-four hours post-reperfusion, deeply anesthetized mice (3% isoflurane) underwent cervical dislocation euthanasia. Cerebral tissue was immediately excised and submerged in ice-cold PBS. Coronal sectioning at 2 mm intervals was performed using a RWD Life Science AS2000 brain matrix. Sections underwent incubation in 2% 2, 3, 5-triphenyltetrazolium chloride (TTC; Sigma-Aldrich, T8877) dissolved in PBS (37 °C, 20 min, protected from light, gentle agitation). TTC-stained slices underwent fixation overnight at 4 °C using 4% paraformaldehyde (PFA; Sigma-Aldrich, 158127) and subsequently imaged using a digital camera (Canon EOS 5D Mark IV) with a macro lens. Unstained regions, indicative of infarction, were delineated and measured in ImageJ (v1.53t, NIH). Total infarct volume (mm³) was derived by multiplying individual slice areas by section thickness (2 mm). Edema correction was applied by normalizing the calculated infarct volume to the volume of the contralateral (non-ischemic) hemisphere.
Neurological function evaluation
Neurological impairment was evaluated 24 h post-reperfusion by an investigator blinded to group allocation, employing the modified Bederson scoring system (5-tiered scale), where 0 indicates normal motor function (no detectable impairment), 1 denotes persistent flexion of the contralateral forelimb (incomplete extension during tail suspension), 2 refers to unidirectional circling toward the affected hemisphere on an open surface, 3 signifies loss of postural control with falling to the paretic side upon gentle prodding, and 4 represents absence of spontaneous movement.
Brain water content measurement
Immediately following euthanasia at 24 h post-reperfusion, brains were excised and weighed to determine wet weight (WW). Subsequently, brains were desiccated in a forced-air drying oven (Thermo Fisher Scientific, Heratherm OGH60) at 60 °C for 72 h until a constant dry weight (DW) was achieved. The percentage brain water content was derived through gravimetric analysis using the formula: [(WW-DW)/WW] × 100%.
BBB integrity evaluation
Evans blue extravasation assay
at 22 h post-reperfusion, mice received an intravenous injection of 2% Evans Blue dye (Sigma-Aldrich, E2129; dissolved in sterile saline, filtered through a 0.22-μm syringe filter, Millipore SLGP033RB) via the tail vein (dose: 4 mL/kg). Following a 2 h circulation period, animals received terminal anesthesia and transcardial flushing with ≈30 mL saline (0.9%) to clear intravascular dye (perfusion continued until effluent ran clear). Excised brains underwent dissection of ischemic hemispheres, which were weighed and subjected to mechanical disruption in 50% trichloroacetic acid (TCA; Sigma-Aldrich, T0699; 1:5 w/v ratio) using a mechanical homogenizer (Qiagen, TissueLyser II, 30 Hz, 2 min). Homogenates underwent refrigerated centrifugation (12,000 × g, 20 min, 4 °C). Supernatant fluorescence was quantified spectrofluorometrically (Thermo Fisher, Varioskan LUX; λex 620 nm, λem 680 nm). Tissue Evans Blue content (μg/g wet weight) in the ischemic hemisphere was determined by standard curve interpolation.
Albumin immunofluorescence staining
Twenty-four hours post-reperfusion, animals underwent sequential transcardial perfusion with saline (0.9%) followed by 4% paraformaldehyde (ice-cold). Brains underwent post-fixation in 4% PFA (4 °C, overnight), then sucrose equilibration (30% in PBS, Sigma, S0389) at 4 °C until tissue sedimentation. Coronal cryosections (20 μm) were prepared using a Leica CM1950 cryostat and adhered to gelatin-coated slides. Sections were blocked (1 h, room temperature) in PBS containing 5% BSA (Sigma-Aldrich, A7906) and 0.3% Triton X-100 (Sigma-Aldrich, T8787). Primary incubation employed a rabbit anti-mouse albumin antibody (Abcam, ab10241; 1:500 dilution in blocking solution) at 4 °C overnight. After triple PBS rinsing (5 min intervals), sections were exposed to AF488-conjugated goat anti-rabbit secondary (Invitrogen, A11037; 1:1000) for 60 min at room temperature under light-restricted conditions. Nuclear counterstaining utilized DAPI (Sigma, D9542; 1:1000 PBS, 5 min). Cover slipping employed Vector H-1000 anti-fade medium. Confocal imaging was performed on a Zeiss LSM880 system (20 × objective). Albumin-associated fluorescence intensity within the ischemic cortex was quantified using ImageJ software (5 random fields per section, 3 sections per animal).
Transepithelial electrical resistance (TEER)
TEER, reflecting monolayer integrity and tight junction functionality, was quantified utilizing an EVOM2 meter (World Precision Instruments, 01920002) with STX2 electrodes. HBMVECs were cultured on Transwell inserts as described above. TEER measurements were recorded at baseline (pre-treatment) and 24 h post-reperfusion (or equivalent time point for controls). Electrodes were sterilized (70% ethanol) and rinsed (PBS) before each measurement. TEER values (Ω·cm²) were calculated using the formula: (Measured Resistance - Blank Insert Resistance) × Membrane Surface Area (1.13 cm²).
FITC-Dextran paracellular flux
HBMVECs were seeded onto 12-well Transwell inserts (Corning, 3460) at a density of 1 × 10⁵ cells per insert in complete endothelial cell medium. The cells were cultured until a confluent monolayer was confirmed by measuring transepithelial electrical resistance (TEER), which exceeded 60 Ω·cm². After experimental treatments, both the apical and basolateral compartments were replenished with serum-free medium. FITC-conjugated dextran (4 kDa; Sigma, FD4) was added to the apical compartment at a concentration of 1 mg/mL. Following a 60 min incubation at 37 °C, 100 μL samples were collected from the basolateral chambers. Fluorescence spectrometry (Thermo Fisher, Varioskan LUX; λex = 485 nm, λem = 520 nm) was used to quantify tracer levels. The FITC-dextran concentration was determined based on a standard curve. Endothelial permeability was expressed as the amount of FITC-dextran (in μg) that crossed the monolayer per mL of basolateral medium.
Molecular analyses
RNA isolation and quantitative real-time PCR (qRT-PCR)
Total RNA was extracted using TRIzol reagent (Invitrogen, 15596026) following the manufacturer’s guidelines. RNA integrity (A260/A280 ratio: 1.8–2.0) was assessed spectrophotometrically (NanoDrop 2000, Thermo Fisher, ND-2000). Residual genomic DNA was eliminated by treatment with DNase I (Takara, 2270 A). First-strand cDNA was synthesized from 1 μg of total RNA using the PrimeScript RT Kit with gDNA Eraser (Takara, RR037A) on a Bio-Rad T100 thermal cycler under standard conditions. Quantitative PCR was performed using SYBR Green chemistry (Takara, RR820A) on a Bio-Rad CFX96 system with 20 μL reaction mixtures under standard amplification conditions. Relative quantification was conducted using the 2^–ΔΔCt^ method with normalization to GAPDH. Custom primers (Sangon Biotech; sequences provided in Table 1) were pre-validated for amplification efficiency (90–110%) and target specificity.Table 1qRT-PCR Primer Sequences.GeneForward Primer (5′ → 3′)Reverse Primer (5′ → 3′)Netrin-5CAGCTGCTGAAGCTGAAGATGCTGGTGGTCTTGGTGATGTZO-1ACGACGACCTACACGAAGAGCTGGTGGTGTTGGTGATGTWnt3aTGGCTGCTGAAGCTGAAGATGCTGGTGGTCTTGGTGATGTβ-cateninTGCTGCTGAAGCTGAAGATTGCTGGTGGTCTTGGTGATGAGAPDHAAGGTCGGTGTGAACGGATTTTGTAGACCATGTAGTTGAGGTCA
Western blotting
Frozen brain tissue (100 mg) or pelleted cultured cells were homogenized in ice-cold RIPA buffer (Beyotime, P0013B) supplemented with 1 × protease inhibitor cocktail (Roche, 04693159001) and 1 mM phenylmethylsulfonyl fluoride (PMSF; Sigma-Aldrich, P7626). Tissue homogenization was performed using a mechanical homogenizer (Qiagen, TissueLyser II) at 30 Hz for 2 min. The resulting lysates were vortexed intermittently during a 30 min incubation on ice, followed by centrifugation under refrigeration (14,000 × g, 15 min, 4 °C). Supernatant protein concentration was determined using the BCA assay (Thermo Fisher, 23225) with BSA standards. Protein samples (30 μg) were denatured in 5 × SDS loading buffer (Beyotime, P0015), separated by 10% SDS-PAGE (Bio-Rad, 1658004; 80 V for 30 min, then 120 V for 90 min), and electrotransferred onto PVDF membranes (Millipore, IPVH00010).
Membranes were blocked with 5% skim milk (BD, 232100) in TBST for 60 min at room temperature. Incubation with primary antibodies was performed overnight at 4 °C in TBST containing 5% BSA. The following primary antibodies were used: Anti-ZO-1 (Invitrogen, 40-2200, 1:1000), Anti-Wnt3a (Abcam, ab28472, 1:1000), Anti-β-catenin (Cell Signaling, 9587, 1:1000), Anti-Netrin-5 (Santa Cruz, sc-376023, 1:500), and Anti-β-actin (Cell Signaling, 4967, 1:2000). After washing with TBST (3 × 10 min), the membranes were incubated with HRP-conjugated secondary antibodies in TBST with 5% skim milk for 60 min at room temperature.
Following three additional TBST washes (10 min each), antigen-antibody complexes were detected by applying SuperSignal West Pico Plus chemiluminescent substrate (Thermo Fisher, 32106). Signal capture utilized the Bio-Rad ChemiDoc MP system (17001402). Band intensity quantification was performed using ImageJ software, normalized to β-actin expression. Band densitometric analysis was conducted in ImageJ with normalization to β-actin loading controls.
Immunofluorescence staining (ZO-1)
Frozen brain sections (20 μm), prepared as described for albumin staining, were used. All samples were permeabilized with 0.1% Triton X-100 in PBS for 10 min at room temperature, followed by incubation in blocking buffer (5% BSA in PBS) for 60 min at room temperature. Sections were then incubated overnight at 4 °C with an anti-ZO-1 rabbit polyclonal antibody (Invitrogen, 40-2200) diluted 1:200 in PBS containing 1% BSA. After washing with PBS (3 × 5 min), the specimens were incubated with AF488-conjugated goat anti-rabbit IgG (Invitrogen, A11029; diluted 1:1000 in PBS with 1% BSA) for 60 min at room temperature under light-protected conditions. Sections were mounted using Vector H-1000 anti-fade medium. ZO-1 distribution was visualized by confocal microscopy (Zeiss LSM880, 63 × oil objective). Fluorescence intensity was quantified using ZEN 2.3 software (Zeiss) in five random fields per sample after background subtraction.
Statistical methodology
Experimental designs included at least three biological replicates per condition, with a minimum of three technical replicates per biological sample. Results are expressed as mean ± standard error of the mean (SEM). Data normality was assessed using the Shapiro-Wilk test, and homogeneity of variance was confirmed by Levene’s test. Multi-group comparisons were performed using one-way ANOVA, followed by Tukey’s HSD post-hoc test for pairwise comparisons when ANOVA indicated significance. All statistical analyses were conducted using GraphPad Prism version 9.0 (GraphPad Software, San Diego, CA). A two-tailed p-value of less than 0.05 was considered statistically significant.
Results
Netrin-5 cortical expression is suppressed after MCAO-Induced ischemic stroke
To investigate the potential involvement of Netrin-5 in cerebral ischemia, we first quantified its expression in the cortical tissue of mice subjected to MCAO—a well-established model of ischemic stroke. Quantitative PCR analysis revealed a marked reduction in cortical Netrin-5 mRNA, with expression levels decreasing to 46% of those in sham-operated controls (sham: 1.00 ± 0.14 vs. MCAO: 0.46 ± 0.05; Fig. 1A). Consistent with these transcriptional changes, immunoblotting showed a corresponding decrease in Netrin-5 protein levels, which declined to 52% of control values (sham: 1.00 ± 0.12 vs. MCAO: 0.52 ± 0.04; Fig. 1B). For comparison, we examined other netrin family members. Netrin-1 expression showed moderate reduction (0.71 ± 0.06 vs. 1.00 ± 0.15), while Netrin-4 remained unchanged (0.95 ± 0.08 vs. 1.00 ± 0.09) following MCAO (supplementary fig. 1A-B). These convergent results demonstrate a significant downregulation of Netrin-5 expression in the ischemic cortex following injury, suggesting its potential role in stroke pathophysiology.Fig. 1. The expression of Netrin-5 was reduced in the cortex tissue post-stroke induced by MCAO.A mRNA of Netrin-5 as measured by real-time PCR. B. Protein of Netrin-5 as measured by western blot analysis (n = 8 mice per group; ※※, P < 0.01 vs. Control group).
Overexpression of Netrin-5 attenuates cerebral ischemia and enhances functional recovery in MCAO mice
To evaluate the therapeutic potential of Netrin-5 in cerebral ischemia, we constructed an adenoviral vector encoding Netrin-5 (Ad-Netrin-5) to overexpress Netrin-5 in MCAO mice. Four experimental groups were established: vehicle control, Ad-Netrin-5 alone, MCAO, and MCAO + Ad-Netrin-5. Immunoblot analysis confirmed a substantial increase in Netrin-5 protein levels in the brain, reaching 3.1-fold that of the vehicle control (3.1 ± 0.33 vs. 1.0 ± 0.13; Fig. 2A). Neurological outcomes were assessed by measuring infarct volume, behavioral deficits, and brain water content. MCAO alone resulted in extensive brain infarction, with a lesion volume of 319.5 ± 25.63 mm³, whereas Netrin-5 overexpression significantly reduced the infarct volume by approximately 52% (152.5 ± 12.75 mm³; Fig. 2B). Behavioral evaluation using the modified Bederson scale (ranging from 0 to 4, with higher scores indicating more severe deficits) revealed that MCAO mice exhibited marked neurological impairment (3.3 ± 0.37), which was significantly ameliorated by Netrin-5 overexpression (1.8 ± 0.15; Fig. 2C). Furthermore, brain edema induced by MCAO (81.7% ± 5.58% water content) was markedly attenuated by Netrin-5, with water content returning to near-baseline levels (78.3% ± 5.15%; Fig. 2D). These findings indicate that Netrin-5 overexpression confers significant neuroprotection in ischemic stroke.Fig. 2. Netrin-5 overexpression conferred neuroprotective effects by reducing brain infarction and improving neurological deficits in MCAO mice.Animals were assigned to four groups: vehicle, Ad-Netrin-5, MCAO, and MCAO + Ad-Netrin-5. A Western blot analysis confirmed successful Netrin-5 overexpression in the mouse brain. B Quantification of cerebral infarct volume. C Evaluation of neurological function using the modified Bederson scoring system. D Measurement of brain water content. (n = 8 mice per group;※, ※※ P < 0.01 vs. control group; †, †† P < 0.05, 0.01 vs. MCAO group).
Netrin-5 Preserves Blood-Brain Barrier (BBB) Integrity in MCAO Mice
Given that BBB impairment represents a key pathological feature of ischemic stroke, we assessed the impact of Netrin-5 on BBB stability. Evans blue leakage and albumin extravasation—two classic markers of BBB permeability—were significantly increased in MCAO mice (Evans blue: 2.1 ± 0.23 vs. 1 ± 0.14 in controls; albumin: 2.6 ± 0.24 vs. 1 ± 0.11 in controls; Fig. 3A, B). Notably, Netrin-5 overexpression reduced Evans blue leakage to 1.4 ± 0.15 and albumin extravasation to 1.8 ± 0.17, indicating preserved BBB integrity (Fig. 3A, B).Fig. 3. Netrin-5 overexpression maintained BBB integrity in MCAO mice.Mice were divided into four groups: vehicle, Ad-Netrin-5, MCAO, and MCAO+Ad-Netrin-5 (n = 8 mice per group). A BBB permeability was assessed using Evans blue dye; (B) Albumin leakage in the cerebral cortex was detected via immunofluorescence staining. Scale bar: 50 μm ※※, P < 0.01 vs. control group; ††, P < 0.01 vs. MCAO group.
Tight junction proteins, particularly ZO-1, are critical for maintaining BBB integrity. In MCAO mice, ZO-1 expression was significantly downregulated at both the mRNA (0.45 ± 0.04 vs. 1.00 ± 0.14) and protein levels (immunofluorescence: 0.49 ± 0.05 vs. 1.00 ± 0.11; Western blot: 0.53 ± 0.06 vs. 1.00 ± 0.12; Fig. 4A–C). In contrast, Netrin-5 overexpression restored ZO-1 mRNA to 0.92 ± 0.08 and protein levels to 0.89 ± 0.07 (immunofluorescence) and 0.93 ± 0.09 (Western blot), bringing them close to those in the control group (Fig. 4A–C). These results indicate that Netrin-5 helps preserve BBB integrity by upregulating ZO-1 expression.Fig. 4. Netrin-5 overexpression restored ZO-1 expression in the cortical tissue of MCAO mice.Mice were divided into four groups: vehicle, Ad-Netrin-5, MCAO, and MCAO+Ad-Netrin-5 (n = 8 mice per group). A ZO-1 mRNA levels were measured by real-time PCR; (B) ZO-1 protein was detected by immunofluorescence staining. Scale bar: 50 μm. C ZO-1 protein expression was analyzed by western blot (※※, P < 0.01 vs. control group; ††, P < 0.01 vs. MCAO group).
OGD/R Reduces Netrin-5 Expression in Human Brain Microvascular Endothelial Cells
To validate the in vivo findings and explore underlying mechanisms, we used an in vitro ischemia-reperfusion model using oxygen-glucose deprivation/reperfusion (OGD/R) in HBMVECs- the principal cellular constituents of the BBB. First, we confirmed that Netrin-5 is endogenously expressed in HBMVECs, with SK-N-SH neuronal cells serving as a positive control (Fig. 5A, B). OGD/R treatment significantly reduced Netrin-5 expression in HBMVECs, with mRNA levels decreasing to 47% (0.47 ± 0.05 vs. 1 ± 0.13) and protein levels to 51% (0.51 ± 0.06 vs. 1 ± 0.12) of the control (Fig. 5C, D), mirroring the downregulation observed in MCAO mice.Fig. 5OGD/R downregulates Netrin-5 expression in human brain microvessel endothelial cells (HBMVECs).A, B RT-PCR and western blot analysis confirmed Netrin-5 expression in HBMVECs, using SK-N-SH neuronal cells as a positive control (n = 3 independent biological replicates). C, D OGD/R treatment significantly reduced Netrin-5 expression at both mRNA and protein levels (n = 5 independent biological replicates; ※※, P < 0.01 vs. control group).
Netrin-5 Overexpression Protects Endothelial Barrier Function Against OGD/R Injury
We next examined whether Netrin-5 protects endothelial barrier function in OGD/R-treated HBMVECs. The cells were divided into four groups: untreated control, Ad-Netrin-5 alone, OGD/R alone, and OGD/R with Ad-Netrin-5 pretreatment. OGD/R alone induced severe endothelial barrier disruption, as shown by a 2.3-fold increase in FITC-dextran permeability (205.8 ± 18.65 vs. 88.3 ± 7.65 μg/mL) and a 39% reduction in transepithelial electrical resistance (TEER; 39.7 ± 4.23 vs. 65.2 ± 5.61 Ω·cm²; Fig. 6A, B). Pretreatment with Ad-Netrin-5 significantly attenuated these changes, reducing permeability to 136.7 ± 12.57 μg/mL and restoring TEER to 68.8 ± 5.62 Ω·cm²—a level comparable to that of the control group (Fig. 6A, B). These results demonstrate that Netrin-5 helps preserve endothelial barrier function in vitro, consistent with its protective effects on the BBB in vivo.Fig. 6. Netrin-5 overexpression preserves endothelial barrier function following OGD/R challenge.After transduction with Ad-Netrin-5, HBMVECs were subjected to OGD/R and divided into four experimental groups (n = 5 independent biological replicates): (1) untreated control, (2) Ad-Netrin-5 alone, (3) OGD/R alone, and (4) OGD/R with Ad-Netrin-5 pretreatment. A Endothelial permeability was assessed using an FITC-dextran permeation assay. B Transepithelial electrical resistance (TEER) was measured (※※, P < 0.01 vs. control group; ††, P < 0.01 vs. OGD/R group).
Netrin-5 Restores ZO-1 Expression and Activates the Wnt3a/β-Catenin Pathway in OGD/R-Treated HBMVECs
To further explore the molecular mechanisms underlying the barrier-protective effects of Netrin-5, we examined its impact on ZO-1 expression and the Wnt3a/β-catenin pathway in OGD/R-treated HBMVECs. OGD/R treatment downregulated both ZO-1 mRNA (0.51 ± 0.05 vs. 1.00 ± 0.12) and protein levels (0.47 ± 0.05 vs. 1.00 ± 0.13). In contrast, pretreatment with Netrin-5 restored ZO-1 expression to near-normal levels (mRNA: 0.92 ± 0.08; protein: 0.89 ± 0.09; Fig. 7A, B), consistent with the in vivo findings. Given the established role of the Wnt3a/β-catenin pathway in maintaining endothelial integrity, we evaluated its activation status. OGD/R alone reduced Wnt3a protein to 52% (0.52 ± 0.05 vs. 1.00 ± 0.11) and β-catenin to 43% (0.43 ± 0.04 vs. 1.00 ± 0.09) of control levels. Notably, Netrin-5 pretreatment increased Wnt3a to 1.20 ± 0.11 and β-catenin to 0.93 ± 0.09. Moreover, Ad-Netrin-5 alone further enhanced their expression (Wnt3a: 2.3 ± 0.24; β-catenin: 2.6 ± 0.25; Fig. 7C). These results suggest that Netrin-5 activates the Wnt3a/β-catenin pathway, which may contribute to ZO-1 upregulation and barrier protection.Fig. 7. Netrin-5 restores ZO-1 expression and activates Wnt3a/β-catenin signaling in HBMVECs following OGD/R injury.After transduction with Ad-Netrin-5, cells were subjected to OGD/R and divided into four experimental groups (n = 5 independent biological replicates): (1) untreated control, (2) Ad-Netrin-5 alone, (3) OGD/R alone, and (4) OGD/R with Ad-Netrin-5 pretreatment. A ZO-1 mRNA levels; (B) ZO-1 protein levels were analyzed by western blot; (C) Protein expression of Wnt3a and β-catenin was assessed by western blot analysis (ΔΔ, ※※, P < 0.01 vs. control group; ††, P < 0.01 vs. OGD/R group).
The Protective Effects of Netrin-5 Are Mediated by Wnt3a/β-Catenin Signaling
To determine whether Wnt3a is required for Netrin-5-mediated protection, we performed Wnt3a knockdown using shRNA in HBMVECs. Cells were divided into four groups: control, OGD/R alone, OGD/R+Ad-Netrin-5, and OGD/R + Ad-Netrin-5 +Ad-Wnt3a shRNA. Real-time PCR confirmed a 54% reduction in Wnt3a mRNA expression (0.46 ± 0.05 vs. 1.00 ± 0.13; Fig. 8A).Wnt3a knockdown abolished the protective effects of Netrin-5. Specifically, the reduction in FITC-dextran permeability mediated by Netrin-5 (132.1 ± 11.53 μg/mL in the OGD/R + Ad-Netrin-5 group) was reversed to 195.6 ± 18.21 μg/mL after Wnt3a knockdown—a value comparable to the OGD/R alone group (213.7 ± 17.62 μg/mL; Fig. 8B). Similarly, the recovery of TEER by Netrin-5 (62.9 ± 5.18 Ω·cm²) was diminished to 42.7 ± 3.86 Ω·cm² following Wnt3a knockdown (Fig. 8C). At the molecular level, the Netrin-5–induced restoration of β-catenin (0.96 ± 0.09 vs. 0.46 ± 0.05 in OGD/R alone) and ZO-1 (0.95 ± 0.11 vs. 0.49 ± 0.05) was negated by Wnt3a knockdown (β-catenin: 0.51 ± 0.05; ZO-1: 0.58 ± 0.06; Fig. 8D). These findings demonstrate that Wnt3a/β-catenin signaling is essential for Netrin-5–mediated protection of endothelial barrier function.Fig. 8. The protective effects of Netrin-5 on endothelial barrier function are mediated through the Wnt3a/β-catenin signaling pathway.HBMVECs were transduced with Ad-Netrin-5 and Ad-Wnt3a shRNA, followed by OGD/R stimulation. Cells were divided into four groups (n = 5 independent biological replicates): (1) untreated control, (2) OGD/R alone, (3) OGD/R + Ad-Netrin-5 pretreatment, and (4) OGD/R + Ad-Netrin-5 + Ad-Wnt3a shRNA pretreatment. A Real-time PCR analysis confirmed successful Wnt3a knockdown; (B) Endothelial permeability was assessed using FITC-dextran permeation assay; (C) Transepithelial electrical resistance (TEER) was measured; (D) Protein levels of β-catenin and ZO-1 were analyzed by western blot (※※, P < 0.01 vs. control group; ††, P < 0.01 vs. OGD/R group; **, P < 0.01 vs. OGD/R + Ad-Netrin-5 group).
Netrin-5 Knockdown Exacerbates Ischemic Injury and BBB Disruption
To investigate the necessity of Netrin-5, we performed knockdown experiments in vivo and in vitro. Administration of Ad-Netrin-5 shRNA reduced Netrin-5 protein levels to 0.38 ± 0.04 compared to Ad-NC controls (Fig. 9A). In MCAO mice, Netrin-5 knockdown significantly increased infarct volume (482.5 ± 43.75 vs. 326.6 ± 27.81 in MCAO), Evans Blue extravasation (2.6 ± 0.25 vs. 1.9 ± 0.19), and Bederson scores (3.6 ± 0.31 vs. 2.8 ± 0.29) (Fig. 9B-D). Similarly, under OGD/R conditions in HBMVECs, Netrin-5 knockdown elevated FITC-dextran permeability (302.6 ± 33.57 μg/mL vs. 213.6 ± 26.8) and reduced TEER (28.8 ± 3.62 Ω·cm² vs. 42.1 ± 4.06) (Fig. 9E, F). Collectively, these data demonstrate that Netrin-5 is essential for maintaining endothelial barrier integrity under ischemic conditions.Fig. 9. Netrin-5 knockdown exacerbates ischemic injury and BBB disruption.A Netrin-5 knockdown efficacy was confirmed in vivo by western blot (n = 8 mice per group). B-D In MCAO mice, Netrin-5 knockdown significantly increased infarct volume, Evans Blue extravasation, and Bederson scores compared to MCAO controls (n = 8 mice per group). E, F In HBMVECs under OGD/R conditions, Netrin-5 knockdown increased FITC-dextran permeability and reduced TEER compared to OGD/R controls (n = 3 independent biological replicates). ※※, P < 0.01 vs. sham/control; ††, P < 0.01 vs. MCAO/OGD/R.
Discussion
This study identifies Netrin-5 as a novel regulator of BBB integrity in ischemic stroke, demonstrating that its downregulation exacerbates ischemic damage, while overexpression confers neuroprotection through activation of the Wnt3a/β-catenin pathway and preservation of the tight junction protein ZO-1. Crucially, Netrin-5 knockdown experiments established its necessary role in BBB preservation, as knockdown exacerbated both structural and functional damage in ischemic models. These findings provide new insights into the molecular mechanisms underlying BBB protection and highlight Netrin-5 as a potential therapeutic target for ischemic stroke.
The central finding of this study is that Netrin-5 serves as a critical mediator of BBB integrity during ischemic injury, with its protective effects mechanistically linked to the Wnt3a/β-catenin signaling pathway. Netrin-5 expression was significantly reduced in both in vivo and in vitro ischemic models, establishing a correlation between its downregulation and ischemic pathology—mirroring the well-documented decline in protective factors during stroke that contributes to secondary brain damage. Our comprehensive in vitro analysis in HBMVECs (Figs. 5–8) definitively shows that Netrin-5 has direct, protective effects on human brain endothelial barrier function, activating the Wnt3a/β-catenin pathway to restore ZO-1 expression and reduce permeability. Furthermore, Netrin-5 demonstrated more pronounced regulation in ischemia compared to Netrin-1 and Netrin-4, suggesting distinct functions among netrin family members in cerebrovascular pathology [18, 19]. Conversely, Netrin-5 overexpression produced robust neuroprotective effects, including reduced infarct volume, improved neurological function, and alleviated brain edema. These outcomes were closely associated with its ability to preserve BBB integrity, as demonstrated by reduced Evans blue leakage and albumin extravasation—classic markers of BBB disruption [20]. Notably, this protection was mediated by the restoration of ZO-1, a key tight junction protein whose downregulation is a hallmark of BBB dysfunction in stroke [21]. This aligns with prior studies emphasizing ZO-1 as a critical determinant of endothelial barrier stability [22, 23], while extending the current knowledge by identifying Netrin-5 as an upstream regulator of ZO-1 expression under ischemic conditions.
Mechanistically, Netrin-5’s effects are rooted in its activation of the Wnt3a/β-catenin pathway, a pivotal signaling axis regulating vascular homeostasis and BBB integrity. The suppression of Wnt3a and β-catenin induced by OGD/R was reversed by Netrin-5. Crucially, Wnt3a knockdown abolished the protective effects of Netrin-5 on endothelial permeability, transepithelial electrical resistance (TEER), and ZO-1 expression. This confirms that the Wnt3a/β-catenin pathway is not merely correlated with Netrin-5’s actions but is an essential mediator—establishing a causal relationship between Netrin-5, Wnt3a/β-catenin activation, and BBB protection.
Although first-generation adenoviral vectors successfully drove robust Netrin-5 overexpression and produced clear reductions in infarct volume, brain edema, and blood-brain barrier leakage after MCAO, we did not perform co-localization studies to determine which cell types were primarily transduced, nor did we systematically evaluate potential vector-induced inflammation (e.g., cytokine upregulation or microglial activation). First-generation adenoviruses are known to infect multiple CNS cell populations (neurons, astrocytes, and endothelium) and can trigger mild innate immune responses [24]. While the magnitude of the observed protection, together with our direct in vitro evidence in human brain microvascular endothelial cells, strongly implicates endothelial Netrin-5 as the principal mediator of BBB preservation, contributions from other transduced cell types or low-level inflammatory effects of the vector cannot be fully excluded. These represent important limitations of the current in vivo delivery strategy. Future translational studies should employ endothelial-specific or less immunogenic vectors (e.g., AAV-BR1, capsid-engineered AAV9 variants, or helper-dependent adenoviral systems) to eliminate these potential confounds.
The novelty of this work lies in its expansion of our understanding of netrin-mediated neuroprotection and BBB regulation. While other members of the netrin family (e.g., Netrin-1) have been implicated in neuroprotection and angiogenesis [25, 26], the role of Netrin-5 in ischemic stroke—particularly its impact on BBB integrity—has not been previously reported. Our study identifies Netrin-5 as a distinct regulator that acts through Wnt3a/β-catenin-mediated ZO-1 upregulation, broadening the functional diversity of netrins in cerebrovascular protection. Furthermore, although the Wnt/β-catenin pathway is a well-established modulator of BBB stability—with activation known to upregulate tight junction proteins and reduce permeability in ischemic models [9, 27]—previous studies have primarily focused on endogenous Wnt ligands or small-molecule activators [28, 29]. Our work identifies Netrin-5 as an upstream inducer of this pathway, offering a new therapeutic target for enhancing Wnt3a/β-catenin signaling in stroke. However, the delivery and specificity challenges outlined above must first be overcome before clinical translation can be realistically pursued. Unlike single-target therapies focused exclusively on anti-inflammation or thrombolysis [30], Netrin-5 addresses two interconnected hallmarks of ischemic injury: BBB disruption and neuronal damage. By reducing infarct volume, edema, and BBB permeability while improving neurological function, Netrin-5 offers a more comprehensive approach to mitigating stroke pathology—aligning with the growing recognition that multi-target strategies are necessary to address the complexity of ischemic stroke.
Ischemic stroke remains a leading cause of disability worldwide, and current treatments are limited by narrow time windows (e.g., thrombolysis) or incomplete protection against secondary damage [31]. The identification of Netrin-5 as a BBB-protective factor is encouraging, but significant preclinical optimisation of delivery methods is required before its therapeutic potential can be fully evaluated. Its ability to target both acute BBB disruption and downstream neuronal damage in preclinical models suggests that, with appropriate delivery systems and after overcoming the vector-related limitations described above, it could eventually complement existing therapies [32]. While viral vectors face challenges in clinical translation, non-viral methods or small-molecule mimetics could be explored [33]. Future studies could use endothelial-specific vectors like AAV-BR1 to improve targeted delivery [34, 35]. Additionally, the downregulation of Netrin-5 in ischemic tissues raises the possibility that Netrin-5 levels could serve as a prognostic marker for BBB dysfunction and stroke severity, though this requires further validation in clinical samples.
Although the present findings hold significant promise, certain limitations warrant consideration. The specific receptor(s) through which Netrin-5 activates Wnt3a/β-catenin signaling remain unknown; netrins typically signal through receptors such as DCC or UNC5 [34, 36], and potential receptors include DCC and UNC5, which mediate Netrin-5 effects in other contexts [34, 36], but whether these or other receptors mediate Netrin-5’s effects on endothelial cells requires further investigation. Additionally, while the MCAO mouse model and OGD/R endothelial cell model recapitulate key features of ischemic stroke, they do not fully capture the complexity of human disease (e.g., comorbidities, diverse stroke etiologies, or long-term pathology) [37]. Validation in larger animal models (e.g., non-human primates) or human organoid systems would strengthen translational relevance. This study was limited to acute outcomes (24–72 h post-ischemia); therefore, assessing the impact of Netrin-5 on long-term recovery (e.g., neuroplasticity, cognitive function) will be essential for evaluating its full clinical potential. Future studies should aim to identify Netrin-5’s endothelial receptors, explore combinatorial strategies (e.g., Netrin-5 combined with thrombolytics) to extend therapeutic time windows, and evaluate Netrin-5 in models of chronic cerebral hypoperfusion, a common cause of vascular dementia.
In conclusion, this study establishes Netrin-5 as a key regulator of BBB integrity in ischemic stroke, acting through the Wnt3a/β-catenin pathway to preserve the tight junction protein ZO-1 and mitigate ischemic damage. These findings not only enhance our understanding of netrin-mediated neuroprotection but also provide a rationale for targeting Netrin-5 to improve stroke outcomes. With further validation, Netrin-5 may emerge as a transformative therapeutic agent for preserving BBB integrity and reducing disability in ischemic stroke.
Supplementary information
Supplementry Figure 1.
The reference list from the paper itself. Each links out to its DOI / PubMed record.