C57BL/6 mice (male, 8 weeks old, 20–22 g; n=168) were obtained from Changzhou Cavens Laboratory Animal Co., Ltd. Piezo1 conditional knockout (CKO) mice (male, 8 weeks old, 20–22 g; n=48) were generated by Cyagen Biosciences, Inc. Cas9 protein, two guide RNAs (gRNA-1: 5′-GGACTGCCAGACACTGTTTAGGG-3′ and gRNA-2: 5′-CTGCTCTCTTGTCTCATGCGTGG-3′), and a donor vector containing two loxP sequences flanking Exons 3–5 of the mouse Piezo1 gene were co-injected into fertilized eggs. The gRNA-to-target sequence alignments were shown in Fig. S1A and B. The embryos were then transferred to recipient female mice to obtain F0 founder mice. Genotyping of F0 founders and subsequent generations was performed by PCR amplification using two pairs of allele-specific primers: Pair 1 (F1: 5′-AGCAAGGCCAATGTAGTATCTGG-3′, R1: 5′-CTATTGGTGCCTAGTTGGCAGAC-3′) spanning the 5′ loxP insertion site, and Pair 2 (F2: 5′-CGGCTCCCAGTGACAGTTTTGAT-3′, R2: 5′-GTGGATTCGGACCAGTCTGA-3′) spanning the 3′ loxP insertion site. These primer pairs were designed specifically against the floxed allele and will not amplify the wild-type Piezo1 locus. Positive PCR products were further confirmed by Sanger sequencing to verify correct loxP insertion. F0 founder mice carrying the correctly targeted allele were backcrossed with wild-type C57BL/6J mice to generate heterozygous F1 offspring (Piezo1^flox/+). Subsequent intercrossing yielded homozygous Piezo1^flox/flox mice for experimental use.

Mice were housed in a constant temperature environment at 25°C under a 12-h light-dark cycle. Standard feed and water were provided ad libitum. Prior to experimentation, a 7-day acclimation period was allowed. The Animal Ethics Committee of Qinghai Provincial People's Hospital approved all animal experiments [approval no. (2023)-246]. The mouse model of HACE was established as follows: Mice were placed in a hypobaric chamber and subjected to simulated altitude conditions of 7,600 m (25,000 ft, 282 mmHg) at a descent velocity of 300 m/min for 12, 24, 48 and 72 h, respectively. After exposure, mice were returned to sea level at a descent velocity of 300 m/min. Following the experiment, mice were anesthetized via inhalation of 5% isoflurane and maintained under anesthesia with 1.5% isoflurane. Deep anesthesia was confirmed once the mice were unresponsive to stimulation of the head, tail, limbs or body. Then they were perfused with physiological saline (0.9%) to clear blood. After perfusion, mice were already dead (the fluid flowing from the incision in the right atrium became clear and free of blood; mice became completely rigid). Then the whole brains were quickly collected and stored in a cryogenic freezer at −80°C. Mice maintained under normal conditions served as controls, with each group comprising 24 mice. Animal grouping and detailed tests were as follows:

A total of 120 C57BL/6 mice were randomly selected and divided into the Control group, the H-12, H-24, H-48 and H-72 h groups, with 24 mice per group. Within each group of 24, six mice were designated for brain water content measurement, six for Evans blue (EB) staining, six for the Fluorescein isothiocyanate (FITC)-Dextran leakage assay and the remaining six for hematoxylin and eosin (H&E) staining, immunofluorescence staining, transmission electron microscopy, proteomics analysis, western blot analysis and reverse transcription-quantitative polymerase chain reaction (RT-qPCR).

The remaining 48 C57BL/6 mice and the 48 Piezo1 knockdown mice were utilized. The 48 C57BL/6 mice were randomly divided into the Control and H-24 h groups, with 24 mice per group. The 48 Piezo1 knockdown mice were randomly assigned to the CKO-Piezo1 and CKO-Piezo1 + H-24 groups, with 24 mice per group. For each group of 24 mice, six were used for brain water content measurement, six for EB staining, six for FITC-Dextran leakage assay and the remaining six for H&E staining, western blot analysis and Calpain-1 activity detection.

In the animal experiments, randomization and blinding were employed. After the 1-week acclimatization period, a researcher not involved in subsequent experimental procedures or data analysis randomly assigned animals to groups using a random number table generated by Excel (Microsoft Corporation). Blinding involved different researchers conducting experimental procedures and evaluating results. All samples were labeled with numerical codes, and results were unblinded only after statistical analysis was completed.

Before euthanasia, the weight of each mouse was recorded, followed by measurement of whole brain weight. The extent of cerebral edema was quantified as the percentage of brain weight relative to body weight.

Subsequently, the whole brains of the mice were placed in an 80°C drying oven for 72 h to determine dry weight. Brain water content was calculated using the formula (wet weight-dry weight)/wet weight ×100%.

BBB permeability was assessed via tail vein injection of EB dye. Following HACE modeling, mice received a tail vein injection of a 2% EB saline solution (4 ml/kg) (Nanjing Senbeiga Biotechnology Co., Ltd.) and were allowed to circulate for 1 h. Under anesthesia, mice were perfused with physiological saline (0.9%) to clear the intravascular dye. After euthanasia, the brains were collected and homogenized in a 50% trichloroacetic acid solution. The resulting mixture was centrifuged (15,000 × g, 30 min), and the supernatant was diluted 1:3 in ethanol. The absorbance of the diluted supernatant was measured using a microplate reader (Thermo Fisher Scientific, Inc.) at a wavelength of 630 nm to determine EB content, calculated using the formula (EB concentration derived from the standard curve × dilution factor)/brain tissue weight.

After anesthesia with 1.5% isoflurane, mice underwent MRI examination using a 7.0 T BioClinScan animal system (Bruker Corporation). Multislice coronal spin-echo diffuse-weighted images were acquired using an echo planar diffusion-trace MRI sequence (matrix size=64×64, TE=23.82 msec, TR=2,500 msec, slice thickness=1.0 mm, b value=1,030.5 sec/mm²). Diffuse-weighted images were reconstructed using Para Vision version 5.1 software (Bruker Corporation). Water diffusion was evaluated by calculating the ADC values using the formula-ln(S/S0)/b, where S and S0 represent signals acquired with and without diffusion weighting, respectively.

Brains were fixed in 4% paraformaldehyde for 24 h at 4°C and embedded in paraffin, subsequently sectioned to a thickness of 4 µm. Sections were baked for 1 h at 60°C in a drying oven. After conventional dewaxing and rehydration, staining was performed using H&E solutions (Shanghai Lianmai Biological Engineering Co., Ltd.). Hematoxylin solution was utilized to stain the sections for 5 min at room temperature, while eosin staining was performed for 1 min at room temperature. Following dehydration and transparency, sections were mounted in neutral resin and examined for pathological changes in the hippocampus and cerebral cortex using an optical microscope (Olympus Corporation).

Cerebral microvascular permeability was evaluated using a FITC-Dextran leakage assay. Mice received an intraperitoneal injection of 70 kDa FITC-dextran (Shanghai Maokang Biotechnology Co., Ltd.) at a dosage of 125 µg/g. After 2 h, mice were anesthetized with 1.5% isoflurane and perfused with physiological saline (0.9%). After euthanasia, the brains were harvested and fixed for 24 h at 4°C in 4% paraformaldehyde, then sectioned to a thickness of 4 µm. Sections were probed overnight at 4°C with mouse anti-CD31 primary antibody (1:100; HY-P80599; MedChemExpress) and subsequently reacted with Alexa Fluor 555-labeled goat anti-mouse secondary antibody (1:200; HY-P81006; MedChemExpress) for 2 h at room temperature. Following 4′,6-diamidino-2-phenylindole (DAPI) (Nanjing Jiancheng Bioengineering Institute) staining of cell nuclei for 5 min at room temperature, brain sections were observed under a confocal microscope (Olympus Corporation). The fluorescence intensity of FITC-Dextran was quantified using ImageJ software (Version 1.54r; National Institutes of Health).

Following fixation in 4% paraformaldehyde (for 24 h at 4°C), mouse brains were embedded in epoxy resin (at 60°C for 48 h), and then dehydrated and sectioned into 60-nm ultrathin sections. The sections were stained with 4% uranyl acetate (for 30 min at room temperature) and then 0.3% lead citrate (for 8 min at room temperature), enabling examination of tight junctions in cerebral microvessels using a TEM (HT7700; Hitachi, Ltd.).

Brain microvascular endothelial cells were isolated from mice as previously described (14). Fresh mouse brains were collected and placed in a culture dish containing D-Hanks' medium (Shanghai Yuanye Bio-Technology Co., Ltd.). Under a stereomicroscope (Leica Microsystems GmbH), white matter, pia mater and large blood vessels were carefully removed. The remaining cerebral cortex was rinsed with D-Hanks' solution and sectioned into 1 mm³ pieces, which were incubated with 0.1% type II collagenase (Beijing BioLab Technology Co., Ltd.) for 40 min in a 37°C water bath. The mixture was then centrifuged at 250 × g for 3 min, allowing collection of microvascular endothelial cells from the pellet for downstream analyses, including RT-qPCR, western blot analysis and Calpain-1 activity detection.

Brain tissues from mice were lysed in radio-immunoprecipitation assay (RIPA) lysis buffer (Jiangsu KeyGEN BioTECH Co., Ltd.) to extract total proteins. Total protein concentration was determined using a Bicinchoninic Acid Assay (BCA) kit (Jiangsu KeyGEN BioTECH Co., Ltd.). Protein samples were loaded onto a trap column (flow rate, 350 nl/min) and an analytical column, followed by proteomic analysis using a Tims Time-of-Flight (TOF) Pro mass spectrometer (Thermo Fisher Scientific, Inc.) equipped with an EasySpray source (Thermo Fisher Scientific, Inc.). Data were analyzed with Spectronaut Pulsar 18.4 (Biognosys AG), with the uniprot-Mus musculus-10090-2024.2.1.fasta as the search database (https://www.uniprot.org/). Differentially expressed proteins were identified based on the criteria of fold change ≥2 or ≤0.5 and P<0.05. A heatmap of the differentially expressed proteins was generated using R software (version 4.3.1; R Foundation for Statistical Computing), with expression levels of Piezo1, VE-cadherin and occludin presented in statistical graphs. KEGG enrichment analysis was performed using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) functional annotation tool (https://davidbioinformatics.nih.gov/).

Mouse brain microvascular endothelial cells (bEnd.3 cells) obtained from Shanghai Kelei Biopharmaceutical Technology Co., Ltd., were cultured in standard medium consisting of 89% Dulbecco's modified Eagle medium (DMEM), 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin at 37°C under normoxic conditions (5% CO₂ and 95% air). DMEM, FBS and penicillin-streptomycin were sourced from Beyotime Biotechnology. For hypoxia treatment, bEnd.3 cells were maintained in standard medium at 37°C under hypoxic conditions (5% CO₂, 94% N₂ and 1% O₂).

Piezo1 small interfering RNA (siRNA; sense: GCCCUGGUACUCAACACUATT; antisense: UAGUGUUGAGUACCAGGGCTT), siRNA negative control (NC) (sense: UUCUCCGAACGUGUCACGUTT; antisense: ACGUGACACGUUCGGAGAATT), NC vectors and Piezo1 overexpression vectors were procured from Shanghai GeneChem Co., Ltd. bEnd.3 cells, seeded in serum-free medium, were transfected with siRNA (50 nM) or vectors (2 µg) using Lipofectamine 3000 (Thermo Fisher Scientific, Inc.) for 8 h at 37°C. Post-transfection, cells were cultured for 48 h in standard medium at 37°C under normoxic conditions (5% CO₂ and 95% air), and then collected for subsequent experiments.

Reports indicate that high-altitude sickness can lead to a marked increase in blood FSS, thereby causing vascular endothelial dysfunction, the core pathological mechanism behind BBB disruption. To simulate this key hemodynamic change, bEnd.3 cells were subjected to HFSS (5,15,16). According to previously reported protocols (17,18), HFSS simulation was performed as follows: bEnd.3 cells plated on flow chamber slides (Ibidi GmbH) were exposed to shear stress levels of 10, 20, 30, or 40 dyn/cm² for 24 h at 37°C. The culture medium was circulated over the cell monolayer using an IBIDI pump system (Ibidi GmbH), with a shear stress of 10 dyn/cm² serving as the control condition.

Calpain-1 activity was inhibited using PD150606 (MedChemExpress) at a concentration of 50 µM for 24 h at 37°C. Yoda1 (MedChemExpress), a specific Piezo1 agonist, was applied at 5 µM to treat cells for 24 h at 37°C. Additionally, PD151746 (MedChemExpress) was used to inhibit Calpain-1 at a concentration of 10 µM by treating cells for 24 h at 37°C. Calcium chelator 1,2-bis-(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid acetoxymethyl ester (BAPTA-AM) (MedChemExpress) was utilized to treat cells for 24 at 10 µM. The mechanism of action of BAPTA-AM is to rapidly bind free Ca²+ within the cell with high affinity, thereby buffering calcium signaling.

Cell grouping and specific processing were structured as follows.

Experiment I comprised the control group, si-NC group and si-Piezo1 group. bEnd.3 cells in the control group were cultured in normal medium and maintained under normoxic conditions (5% CO₂ and 95% air) for 24 h. Cells in the si-NC and si-Piezo1 groups were separately transfected with siRNA NC and Piezo1 siRNA, respectively, followed by culturing in normal medium under normoxic conditions for 24 h.

Experiment II included the si-NC group, HFSS + si-NC group, and HFSS + si-Piezo1 group. Cells from all three groups were maintained under two distinct conditions for 24 h: Normoxia (5% CO₂ and 95% air) and hypoxia (5% CO₂, 94% N₂ and 1% O₂). Under these conditions, bEnd.3 cells in the si-NC group were transfected with siRNA NC and exposed to a shear stress of 10 dyn/cm². Cells in the HFSS + si-NC and HFSS + si-Piezo1 groups were transfected with siRNA NC and Piezo1 siRNA, respectively, and subjected to a shear stress of 30 dyn/cm².

Experiment III included the control group, HFSS group, and HFSS + PD150606 group. Under both normoxic (5% CO₂ and 95% air) and hypoxic conditions (5% CO₂, 94% N₂, and 1% O₂), bEnd.3 cells in the control group received no additional treatment. In the HFSS group and HFSS + PD150606 group, cells were treated without or with 50 µM PD150606, respectively, in the presence of 30 dyn/cm² shear stress. All groups were maintained under their respective conditions for 24 h.

Experiment IV comprised the control group, model group, Yoda1 group, and Model + Yoda1 group. Cells in the control group were maintained under normoxic conditions (5% CO₂ and 95% air) for 24 h. The Model group was exposed to hypoxia (5% CO₂, 94% N₂, and 1% O₂) for 24 h at 37°C. The Yoda1 group was treated with 5 µM Yoda1 for 24 h under normoxic conditions, while cells in the Model + Yoda1 group were cultured in hypoxic conditions for 24 h at 37°C in the presence of 5 µM Yoda1.

Experiment V included the Vector group, Piezo1-OE group, Vector + PD151746 group, and Piezo1-OE + PD151746 group. Cells in the Vector and Piezo1-OE groups were separately transfected with NC vectors and Piezo1 overexpression vectors and subsequently cultured under normoxic conditions (5% CO₂ and 95% air) for 24 h. The Vector + PD151746 group was transfected with NC vectors and then exposed to 10 µM PD151746 for 24 h under normoxic conditions. Cells in the Piezo1 + PD151746 group were transfected with Piezo1 overexpression vectors and subsequently treated with 10 µM PD151746 for 24 h under normoxic conditions.

Experiment VI included the Vector group, Vector + BAPTA-AM group, Piezo1-OE group, and Piezo1 + BAPTA-AM group. Cells in the Vector and Piezo1-OE groups were separately transfected with NC vectors and Piezo1 overexpression vectors. Those in the Vector + BAPTA-AM group were transfected with NC vectors and then exposed to BAPTA-AM at 10 µM for 24 h. Cells in the Piezo1 + BAPTA-AM group were transfected with Piezo1 overexpression vectors and subsequently treated with 10 µM BAPTA-AM for 24 h.

Total RNA was extracted from bEnd.3 cells and mouse brain microvascular endothelial cells using TRIzol reagent (Shanghai Yuanye Bio-Technology Co., Ltd.), followed by complementary DNA synthesis employing a PrimeScript^™ RT reagent Kit (Shanghai Shanran Biotechnology Co., Ltd.). Quantitative RT-qPCR was conducted using SYBR Green Master Mix (Thermo Fisher Scientific, Inc.) on an ABI 7500 Real-Time PCR System (Applied Biosystems; Thermo Fisher Scientific, Inc.). The amplification program included an initial denaturation at 95°C for 3 min, followed by 40 cycles of 10 sec at 95°C and 30 sec at 60°C. The primer sequences were as follows: Piezo1, forward: CCTGTTACGCTTCAATGCTCT, reverse: GTGTAGGCATATCTGAAAGGCAA; Calpain-1, forward: GGGGAGTTCTGGATGTCGTT, reverse: GGTGCCCTCGTAAAATGTGGT; β-actin, forward: GGCTGTATTCCCCTCCATCG, reverse: CCAGTTGGTAACAATGCCATGT (all synthesized by Shanghai GeneChem Co., Ltd.). Relative Piezo1 mRNA expression was quantified using the 2^−ΔΔCq method (19), with β-actin serving as the endogenous control. Alignment of genotyping primers to the floxed Piezo1 allele and β-actin were shown in Fig. S1C.

Total proteins from bEnd.3 cells and mouse brain microvascular endothelial cells were extracted using RIPA lysis buffer (Jiangsu KeyGen Biotech Co., Ltd.). Isolation of mouse brain microvascular endothelial cells was conducted as previously described (20). Total protein concentration was quantified using a BCA kit (Jiangsu KeyGen Biotech Co., Ltd.). Proteins (30 µg per lane) were separated on 10% gels using sodium dodecyl sulfate-polyacrylamide gel electrophoresis and subsequently transferred to polyvinylidene fluoride (PVDF) membranes. Membranes were blocked with 5% skimmed milk (Shanghai Xinfan Biological Technology Co., Ltd.) for 1 h at room temperature. Primary antibodies were applied for 12 h at 4°C, including rabbit anti-Piezo1 (1:1,000; 28511-1-AP; Proteintech Group, Inc.), rabbit anti-Calpain-1 (1:500; 23048R; Hangzhou Lianke Meixun Biomedical Technology Co., Ltd.), rabbit anti-VE-cadherin (1:1,000, HY-P86361; MedChemExpress), rabbit anti-occludin (1:1,000; CY-10011R; Hangzhou Lianke Meixun Biomedical Technology Co., Ltd.), rabbit anti-claudin-5 (1:1,000; ABP50990; AmyJet Scientific, Inc.), rabbit anti-zonula occludens-1 (ZO-1; 1:500; K27853-UOR; Beijing BioLab Technology Co., Ltd.), mouse anti-Spectrin (1:1,000; sc-53444; Santa Cruz Biotechnology, Inc.), rabbit anti-spectrin breakdown products (SBDP, also termed SNTF; 1:10,000; cleaved all-spectrin; ABN2264; MilliporeSigma) and rabbit anti-β-actin (1:2,000; HY-P87821; MedChemExpress). Membranes were subsequently incubated with horseradish peroxidase-conjugated goat anti-rabbit (1:10,000; 111-035-003; AmyJet Scientific, Inc.) or goat anti-mouse secondary antibodies (1:10,000; 1030-05; AmyJet Scientific, Inc.) for 2 h at room temperature. Protein bands were visualized using enhanced chemiluminescent reagent (Shanghai Yuanmu Biotechnology Co., Ltd.) and semi-quantified with ImageJ software (Version 1.54r; National Institutes of Health), with β-actin as the internal reference.

Following fixation with 4% paraformaldehyde for 24 h at 4°C, mouse brain sections (4 µm) and bEnd.3 cells subjected to various treatments were permeabilized with 0.5% Triton X-100 (Shanghai Xitang Biotechnology Co., Ltd.) for 20 min at room temperature and blocked with normal goat serum (Wuhan Chundu Biotechnology Co., Ltd.) for 1 h at room temperature. For detection of Piezo1 in mouse brain microvascular endothelial cells, sections were incubated overnight at 4°C with rabbit anti-Piezo1 primary antibody (1:100; 28511-1-AP; Proteintech Group, Inc.) and mouse anti-CD31 primary antibody (1:100; HY-P80599; MedChemExpress), followed by incubation with Alexa Fluor 555-conjugated goat anti-rabbit secondary antibody (1:200; GD-X0295G-AF555; Shanghai Guduo Biotechnology Co., Ltd.) and Alexa Fluor 488-conjugated goat anti-mouse secondary antibody (1:200, YT859-RJG, BioLab Technology, Beijing, China) for 2 h at room temperature. To visualize Piezo1 and Calpain-1 in bEnd.3 cells, rabbit anti-Piezo1 primary antibody (1:100; 28511-1-AP; Proteintech Group, Inc.) and rabbit anti-Calpain-1 primary antibody (1:100; 23048R; Hangzhou Lianke Meixun Biomedical Technology Co., Ltd.) were employed (overnight at 4°C), followed by Alexa Fluor 488-conjugated goat anti-rabbit secondary antibody (1:200; yb0423; Shanghai Yubo Biotechnology Co., Ltd.) (for 2 h at room temperature). After nuclear staining with DAPI for 5 min at room temperature, sections and cells were dried, mounted in neutral resin and observed under a confocal microscope (Olympus Corporation). Protein expression was quantified using ImageJ software.

Intracellular Ca²⁺ levels in bEnd.3 cells were assessed using the Fluo4-AM fluorescent probe. Cells from each experimental group were incubated with Fluo4-AM (Beijing BioLab Technology Co., Ltd.) for 30 min at room temperature. Fluorescence intensity was monitored with a fluorescence microscope (Olympus Corporation) and quantified using ImageJ software.

Calpain-1 activity was measured with a commercial assay kit (cat. no. KA0723; AmyJet Scientific, inc.). Cells and mouse brain microvascular endothelial cells were collected and suspended in a buffer containing 100 mM Tris-HCl, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 5 mM ethylenediaminetetraacetic acid, 5 mM benzamidine and 10 mM β-mercaptoethanol. The mixture was ultrasonicated four times (10 sec each) at 4°C with 1-min intervals, followed by centrifugation at 15,000 × g for 20 min. Calpain-1 activity was assessed with 40 µg of supernatant, 15 mM calcium (in 50 µl buffer) and 100 µM fluorogenic calpain substrate (AmyJet Scientific, Inc.), with excitation at 320 nm and emission at 480 nm. Calpain-1 activity was quantified using ImageJ software.

Cell viability of bEnd.3 cells was evaluated using a CCK-8 assay. Cells were seeded in 96-well plates and incubated with 10 µl CCK-8 solution (Jiangsu KeyGen Biotech Co., Ltd.) for 2 h at 37°C. Optical density (OD) values were measured on a microplate reader (Huisong Pharmaceuticals) to estimate cell viability. Cell viability was quantified using the formula: [(OD value of the experimental group-OD value of the blank group)/(OD value of the control group-OD value of the blank group)] ×100.

TEER of bEnd.3 cells was measured using an epithelial voltmeter (Beijing Donggong Technology Co., Ltd.). Cells were seeded on inserts at a density of 5×10⁴ cells per insert and subjected to the respective treatment conditions. Following the removal of the culture medium, inserts were transferred to new wells and washed with phosphate-buffered saline. Resistance across the cells was recorded and analyzed. TEER was quantified using the formula: (Resistance value measured in the chamber containing a monolayer of cells-the background resistance value measured in the blank chamber without cells) × the area of the polycarbonate membrane at the bottom of the inserts.

Data analysis was performed using a blinded method. One researcher performed animal grouping and data collection, while another independent researcher evaluated the results without knowledge of group assignments. All samples were labeled with numerical codes, with the blinded status revealed only after the analysis of all data was completed.

Data are expressed as mean ± standard deviation. Statistical analysis was conducted using GraphPad Prism 8.0 software (Dotmatics). Error bars represented standard deviation, reflecting the dispersion of individual observations around the group mean. Normality of data distribution was assessed using the Shapiro-Wilk test. Intergroup statistical differences were compared via Student's t-test (for two distinct groups) or one-way analysis of variance (ANOVA) with post-hoc Tukey's test (for more than two distinct groups). P<0.05 was considered to indicate a statistically significant difference.

To establish a mouse model of HACE, C57BL/6 mice were exposed to a hypobaric hypoxia environment for varying durations. Cerebral edema was assessed by measuring the ratio of brain weight to body weight (Fig. 1A) and brain water content (Fig. 1B). Mice subjected to hypobaric hypoxia for 24, 48 and 72 h exhibited significantly increased brain weight/body weight ratios and brain water content, with the most pronounced elevation observed in the 24-h exposure group (H-24, H-48 and H-72 h vs. control group; P<0.001 for the ratio of brain weight to body weight at H-24, H-48 and H-72 h; P<0.001 for the brain water content at H-24 h; P<0.01 for the brain water content at H-48 h; P<0.05 for the brain water content at H-72 h). Additionally, enhanced BBB permeability, indicated by significantly higher EB content, was observed in mice treated with hypobaric hypoxia for 24, 48 and 72 h, especially in the 24-h group (H-24, H-48 and H-72 h vs. control group; P<0.001; Fig. 1C). Alterations in ADC values on brain MRI reflect changes in brain tissue water content, thus facilitating assessment of brain water content (21). The MRI images are shown in Fig. 1D. Significantly reduced ADC values were recorded in mice exposed to hypobaric hypoxia for 12, 24 and 48 h (H-12, H-24 and H-48 h vs. control group; P<0.05 at H-48 h, P<0.001 at H-12 h and at H-24 h; Fig. 1E). H&E staining revealed notable cellular edema (black arrows) and nuclear condensation (red arrows) in the hippocampus and cerebral cortex of mice, particularly in those exposed for 24 h (Fig. 1F). Thus, hypobaric hypoxia exposure resulted in marked cerebral edema and increased BBB permeability in mice, with 24-h exposure eliciting the most significant effects, leading to its selection for subsequent studies.

TEM images revealed impaired tight junctions in brain microvessels of mice exposed to hypobaric hypoxia for 24 h (H-24 h vs. control group; Fig. 2A). A FITC-Dextran leakage assay was conducted to assess brain microvascular permeability, complemented by CD31 immunofluorescence staining to localize brain microvascular endothelial cells. Increased microvascular permeability was observed in the brains of mice after 24 h of hypobaric hypoxia exposure (H-24 h vs. control group; P<0.001; Fig. 2B). A heatmap of differentially expressed proteins generated through proteomics analysis showed differences between control mice and those exposed to hypobaric hypoxia for 24 h (Fig. 2C). KEGG enrichment analysis indicated an association between these differentially expressed proteins and tight junctions, specifically the pathway ‘Cell adhesion molecules’ was identified, with its associated proteins significantly downregulated (Fig. 2D). Proteomics analysis demonstrated up-regulation of Piezo1 and down-regulation of VE-cadherin and occludin within the brain microvascular endothelial cells of mice subjected to hypobaric hypoxia for 24 h (H-24 h vs. control group; P<0.001; Fig. 2E).

To investigate Piezo1 expression in brain microvascular endothelial cells, these cells were isolated and subjected to RT-qPCR and western blot analyses. Increased mRNA (P<0.01) and protein levels (P<0.001) of Piezo1 were observed in brain microvascular endothelial cells of mice after 24 h of hypobaric hypoxia exposure (H-24 h vs. control group; Fig. 2F and G). Immunofluorescence staining of CD31/Piezo1 was performed to visualize Piezo1 expression within brain microvascular endothelial cells, revealing enhanced expression after 24 h of hypobaric hypoxia exposure (H-24 h vs. control group; P<0.001; Fig. 2H). These results indicate that a hypobaric hypoxia environment leads to increased microvascular permeability, impaired tight junctions and enhanced Piezo1 expression in brain microvascular endothelial cells.

bEnd.3 cells were subjected to FSS at 10, 20, 30 and 40 dyn/cm² for 24 h. A significant increase in Piezo1 protein expression was observed as FSS levels rose (P<0.01, P<0.001). Notably, Piezo1 protein levels in cells exposed to 40 dyn/cm² were comparable to those in cells exposed to 30 dyn/cm². Consequently, HFSS at 30 dyn/cm² was designated as HFSS for subsequent experiments (Fig. S2A). Transfection of Piezo1 siRNA into bEnd.3 cells resulted in a significant reduction in Piezo1 protein expression compared with the siRNA NC (si-Piezo1 vs. si-NC; P<0.001; Fig. S2B).

Under both normoxic and hypoxic conditions, HFSS treatment significantly elevated Piezo1 mRNA and protein expression in bEnd.3 cells (HFSS + si-NC vs. si-NC; P<0.001). Transfection with Piezo1 siRNA led to decreased expression of Piezo1 mRNA and protein in HFSS-treated cells under both conditions (HFSS + si-Piezo1 vs. HFSS + si-NC; P<0.001; Fig. 3A and B). Western blot analysis revealed that HFSS treatment increased Calpain-1 protein expression in bEnd.3 cells under normoxia and hypoxia (HFSS + si-NC vs. si-NC; P<0.001). Notably, Piezo1 silencing did not significantly affect Calpain-1 protein expression in HFSS-treated cells under either condition (HFSS + si-Piezo1 vs. HFSS + si-NC; Fig. 3A and C). Immunofluorescence staining (Fig. 3D and E) and detection with the Fluo4-AM fluorescent probe (Fig. 3F) were employed to monitor the expression of Piezo1 and Calpain-1 proteins, as well as intracellular Ca²⁺ levels in bEnd.3 cells. Following exposure to normoxic and hypoxic conditions, HFSS-treated bEnd.3 cells exhibited elevated levels of Piezo1 and Calpain-1 proteins, along with increased intracellular Ca²⁺ levels (HFSS + si-NC vs. si-NC; P<0.001). By contrast, cells subjected to both Piezo1 siRNA transfection and HFSS treatment demonstrated reduced Piezo1 protein expression and intracellular Ca²⁺ levels (HFSS + si-Piezo1 vs. HFSS + si-NC; P<0.001). However, under HFSS conditions, Piezo1 silencing did not significantly affect Calpain-1 protein expression (HFSS + si-Piezo1 vs. HFSS + si-NC). Calpain-1 activity in bEnd.3 cells from different groups was analyzed using a commercial kit (Fig. 3G). Regardless of normoxia or hypoxia, HFSS-induced cells exhibited increased Calpain-1 activity (HFSS + si-NC vs. si-NC; P<0.001), while this activity was reduced following Piezo1 silencing (HFSS + si-Piezo1 vs. HFSS + si-NC; P<0.001). Western blot analysis indicated that HFSS treatment increased the SBDP/Spectrin ratio in bEnd.3 cells under both normoxic and hypoxic conditions (HFSS + si-NC vs. si-NC; P<0.001), an effect that was abolished by Piezo1 silencing (HFSS + si-Piezo1 vs. HFSS + si-NC; P<0.001). These results suggest that Piezo1 silencing reduces intracellular Ca²⁺ levels and Calpain-1 activity in HFSS-treated bEnd.3 cells under hypoxic conditions. Given that Calpain-1 activity depends on calcium influx, the findings indicate that Piezo1 silencing may attenuate Calpain-1 activity in HFSS-treated bEnd.3 cells by reducing calcium influx.

Cell viability of bEnd.3 cells was assessed using the CCK-8 assay, revealing that HFSS treatment significantly reduced cell viability under both normoxic and hypoxic conditions (HFSS + si-NC vs. si-NC group; P<0.001). Following Piezo1 silencing, increased cell viability was observed in HFSS-treated cells under both conditions (HFSS + si-Piezo1 vs. HFSS + si-NC group; P<0.001) (Fig. 4A). Subsequently, cell membrane permeability was evaluated through TEER measurement (Fig. 4B), and the expression of tight junction-associated proteins, including VE-cadherin, occludin, claudin-5 and ZO-1, was examined by western blot analysis (Fig. 4C). HFSS treatment under normoxic and hypoxic conditions decreased TEER and reduced the expression of VE-cadherin, occludin, claudin-5 and ZO-1 (HFSS + si-NC vs. si-NC group; P<0.001). Conversely, Piezo1 silencing reversed these effects, leading to increased TEER and elevated expression of VE-cadherin, occludin, claudin-5 and ZO-1 in HFSS-treated cells under both conditions (HFSS + si-Piezo1 vs. HFSS + si-NC group; P<0.001). These results indicate that Piezo1 silencing alleviates HFSS- and hypoxia-induced tight junction dysfunction in microvascular endothelial cells.

Calpain-1 inhibitor PD150606 was applied to HFSS-treated bEnd.3 cells. Under both normoxic and hypoxic conditions, HFSS treatment alone significantly increased Calpain-1 activity (HFSS vs. control group; P<0.001), whereas PD150606 treatment reversed this increase (HFSS + PD150606 vs. HFSS group; P<0.001) (Fig. 5A). Furthermore, HFSS treatment decreased bEnd.3 cell viability (HFSS vs. Control group; P<0.001), which was restored following PD150606 treatment (HFSS + PD150606 vs. HFSS group; P<0.001) (Fig. 5B). TEER measurements (Fig. 5C) and western blot analysis (Fig. 5D) demonstrated that HFSS reduced TEER and decreased the expression of VE-cadherin, occludin, claudin-5 and ZO-1 in bEnd.3 cells (HFSS vs. control group; P<0.001). These reductions were reversed by PD150606 treatment (HFSS + PD150606 vs. HFSS group; P<0.001). Collectively, these results indicate that inhibition of Calpain-1 decreases its activity and enhances tight junction expression in HFSS- and hypoxia-induced microvascular endothelial cells. Overall, the in vitro data suggest that Piezo1 silencing mitigates HFSS- and hypoxia-induced tight junction dysfunction by reducing Calpain-1 activity.

The classic HACE cell model was simulated by exposing bEnd.3 cells to hypoxic conditions. Yoda1, a specific agonist of Piezo1, was then applied to bEnd.3 cells. TEER measurements indicated that both hypoxia exposure (model group) and Yoda1 treatment (Yoda1 group) significantly decreased TEER in bEnd.3 cells compared with the control group (P<0.001). Notably, Yoda1 treatment further reduced TEER under hypoxic conditions (model + Yoda1 group vs. model group; P<0.001) (Fig. 6A). Western blot analysis demonstrated reductions in the protein expression of VE-cadherin, occludin and ZO-1 in bEnd.3 cells exposed to hypoxia (model group vs. control group) or treated with Yoda1 (Yoda1 group vs. control group) (P<0.001). Yoda1 treatment under hypoxia resulted in further decreases in the expression of these proteins (model + Yoda1 group vs. model group; P<0.001) (Fig. 6B-E). Intracellular Ca²⁺ levels were measured using Fluo4-AM fluorescent probe detection, revealing elevated Ca²⁺ levels in bEnd.3 cells subjected to hypoxia (model group vs. control group) or Yoda1 treatment (Yoda1 group vs. control group) (P<0.001). Under hypoxic conditions, Yoda1 further increased intracellular Ca²⁺ levels in bEnd.3 cells (model + Yoda1 group vs. model group; P<0.001) (Fig. 6F).

Calpain-1 activity was assessed using a commercial kit (Fig. 6G), and its protein expression was analyzed by western blot analysis (Fig. 6H). Both hypoxia exposure and Yoda1 treatment increased Calpain-1 activity (P<0.001) and protein expression (P<0.05) in bEnd.3 cells (model group vs. control group; Yoda1 group vs. control group). Yoda1 treatment under hypoxia further enhanced Calpain-1 activity but did not significantly alter Calpain-1 protein expression (model + Yoda1 group vs. model group) Furthermore, the SBDP/Spectrin protein ratio was higher in bEnd.3 cells exposed to hypoxia and treated with Yoda1 (model group vs. control group; Yoda1 group vs. control group; P<0.001). After Yoda1 treatment under hypoxia, the SBDP/Spectrin ratio increased further (model + Yoda1 group vs. model group; P<0.001) (Fig. 6I). These results suggest that the Piezo1-specific agonist Yoda1 exacerbates tight junction disruption in microvascular endothelial cells under hypoxic conditions by enhancing calcium influx-mediated Calpain-1 activity.

The experiments in this section were carried out under normoxic conditions. Piezo1 overexpression vectors were transfected into bEnd.3 cells, resulting in a significant increase in Piezo1 mRNA and protein levels (Piezo1-OE group vs. Vector group; P<0.001) (Fig. 7A and B). TEER measurements (Fig. 7C) and western blot analysis (Fig. 7D) indicated that Piezo1 overexpression significantly decreased TEER and the expression of VE-cadherin, occludin and ZO-1 proteins in bEnd.3 cells (Piezo1-OE group vs. vector group; P<0.001). By contrast, treatment with the Calpain-1 inhibitor PD151746 increased TEER and occludin protein expression in bEnd.3 cells (vector + PD151746 group vs. vector group; P<0.05), although PD151746 did not significantly affect VE-cadherin or ZO-1 protein levels (vector + PD151746 vs. vector group). Compared with the Piezo1-OE group, bEnd.3 cells in the Piezo1 + PD151746 group exhibited higher TEER and increased expression of VE-cadherin, occludin and ZO-1 proteins (P<0.001). These findings indicate that Piezo1 overexpression induces HACE-like phenotypes by reducing TEER and the expression of tight junction proteins, while PD151746 reverses these effects, suggesting that Piezo1 might disrupt tight junctions in microvascular endothelial cells via Calpain-1 activation.

To investigate whether Piezo1 regulates Calpain-1 activity through Ca²⁺, bEnd.3 cells overexpressing Piezo1 were treated with the Ca²⁺ chelator BAPTA-AM. RT-qPCR and western blot analyses indicated that Piezo1 overexpression, BAPTA-AM treatment or their combination did not significantly affect Calpain-1 mRNA or protein expression (Fig. 8A and B). However, BAPTA-AM treatment resulted in a significant reduction in Calpain-1 activity (vector + BAPTA-AM vs. vector group; P<0.001), while Piezo1 overexpression significantly elevated Calpain-1 activity (Piezo1 vs. vector group; P<0.001). Notably, the increased Calpain-1 activity induced by Piezo1 overexpression was significantly decreased following BAPTA-AM treatment (Piezo1 + BAPTA-AM vs. Piezo1-OE group; P<0.001) (Fig. 8C). Western blot analysis revealed that BAPTA-AM treatment decreased the SBDP/Spectrin ratio in bEnd.3 cells (vector + BAPTA-AM vs. vector group; P<0.001). Conversely, Piezo1 overexpression increased the SBDP/Spectrin ratio (Piezo1 vs. vector group; P<0.001), which was reversed by BAPTA-AM treatment (Piezo1 + BAPTA-AM vs. Piezo1-OE group) (P<0.001) (Fig. 8D). These data suggest that the role of Piezo1 in enhancing Calpain-1 activity is dependent on intracellular Ca²⁺ levels, indicating that increased intracellular Ca²⁺ is necessary for Calpain-1 activation.

As shown in Fig. S3, the expression of Piezo1 mRNA was markedly reduced in brain microvascular endothelial cells of Piezo1 knockdown mice (the CKO-Piezo1 group), as compared with the control group (P<0.001). This suggested that Piezo1 expression was effectively down-regulated in Piezo1 knockdown mice. To further confirm the influence of Piezo1 on cerebral edema and microvascular permeability in HACE, Piezo1 knockdown mice were exposed to a hypobaric hypoxia environment. Mice subjected to hypobaric hypoxia for 24 h exhibited higher brain weight/body weight ratios and brain water content (H-24 h vs. control group; P<0.001). Piezo1 knockdown resulted in a marked decrease in brain weight/body weight ratio and brain water content in mice exposed to hypobaric hypoxia (CKO-Piezo1 + H-24 group vs. H-24 h group; P<0.001) (Fig. 9A and B).

EB staining indicated exacerbated BBB permeability in mice after 24 h of hypobaric hypoxia exposure (H-24 h vs. Control group; P<0.001), which was alleviated by Piezo1 knockdown (CKO-Piezo1 + H-24 group vs. H-24 h group; P<0.001) (Fig. 9C and D). H&E staining revealed marked cellular edema and nuclear condensation in the hippocampus and cerebral cortex of mice exposed to hypobaric hypoxia for 24 h (H-24 h vs. control group), with Piezo1 knockdown mitigating this damage (CKO-Piezo1 + H-24 group vs. H-24 h group) (Fig. 9E). The FITC-Dextran leakage assay demonstrated intensified microvascular permeability in the brains of mice exposed to hypobaric hypoxia for 24 h (H-24 h vs. control group; P<0.001), which was reduced by Piezo1 knockdown (CKO-Piezo1 + H-24 group vs. H-24 h group; P<0.001) (Fig. 9F). These results imply that Piezo1 knockdown effectively alleviates cerebral edema and microvascular permeability in the brains of mice exposed to hypobaric hypoxia.

Western blot analysis demonstrated an increase in Piezo1 and Calpain-1 protein expression in the brain microvascular endothelial cells of mice subjected to 24 h of hypobaric hypoxia treatment (H-24 h group vs. control group; P<0.001). Mice with Piezo1 knockdown exhibited decreased Piezo1 protein levels in their brain microvascular endothelial cells (CKO-Piezo1 group vs. control group; P<0.001). Additionally, Piezo1 knockdown led to a reduction in Piezo1 protein expression in the brain microvascular endothelial cells of mice experiencing hypobaric hypoxia (CKO-Piezo1 + H-24 group vs. H-24 h group) (P<0.001), while having no significant effect on Calpain-1 protein levels (CKO-Piezo1 + H-24 group vs. H-24 h group) (Fig. 10A).

Furthermore, mice exposed to hypobaric hypoxia for 24 h exhibited higher Calpain-1 activity in their brain microvascular endothelial cells (H-24 h group vs. control group; P<0.001), whereas Piezo1 knockdown significantly suppressed Calpain-1 activity (CKO-Piezo1 + H-24 group vs. H-24 h group; P<0.001) (Fig. 10B). Hypobaric hypoxia exposure for 24 h resulted in a significant reduction in the expression of VE-cadherin, occludin, claudin-5 and ZO-1 proteins in brain microvascular endothelial cells (H-24 h group vs. control group; P<0.001). Notably, Piezo1 knockdown significantly increased the expression of these proteins (CKO-Piezo1 + H-24 group vs. H-24 h group; P<0.001) (Fig. 10C).

Therefore, Piezo1 knockdown effectively reduced Calpain-1 activity and improved tight junction integrity within brain microvascular endothelial cells of mice exposed to hypobaric hypoxia. Collectively, these findings suggest that Piezo1 knockdown may alleviate HACE by enhancing the integrity of tight junctions in brain microvascular endothelial cells through the suppression of calcium influx-mediated Calpain-1 activity.