Hypoxia treatment leads to damage of BBB both in vitro and in vivo. (A) Phenotypes of zebrafish were observed by microscope imaging. Black arrow indicates the sites of cardiac edema. Solid white boxes indicate the sites of brain hemorrhage. Scale bar: 200 μm. N = 23 fish per group. (B) The integrity of BBB was evaluated by cerebral angiography after hypoxia induction. N = 8 fish per group. Mean ± SD, Student's t-test, ***p < .0001. Scale bar: 50 μm. (C) The expression of CLDN5 protein in zebrafish head was detected by Western blotting. Mean ± SD, Student's t test, **p = .0075. N = 100 fish per group. (D) BBB permeability was assessed by measuring TEER. “Treatment” indicates the time point of hypoxia induction after the bEnd.3 cell monolayer reaches a stable TEER value. Mean ± SD, Student's t test, normoxia vs. hypoxia (3 h), **p = .0016; normoxia vs. hypoxia (6 h), **p = .0069; normoxia vs. hypoxia (12 h), **p = .0088; normoxia vs. hypoxia (24 h), ***p = .0007. (E) Immunofluorescence staining analyses were performed to investigate the expression of CLDN5 in bEnd.3 cells. Mean ± SD, Student's t-test, **p = .004, Scale bar: 10 μm. (F) Western blotting analyses showed the expression level of CLDN5 protein in bEnd.3 cells. Mean ± SD, Student's t-test, **p = .0018.

Hypoxia-induced BBB injury is related to oxidative stress. (A) Quantification of MDA content in the head of zebrafish. N = 150 fish per group. Mean ± SD, Student's t-test, ***p = .0002. (B) 4-HNE and β-tubulin proteins were measured in zebrafish head by Western blot. Mean ± SD, Student's t-test, *p = .0301. N = 100 fish per group. (C) Representative images of bEnd.3 cells stained with ROS assay kit and Hoechst. Mean ± SD, Student's t test, **p = .0011, scale bar: 100 μm. (D) Immunofluorescence staining of CLDN5 and 4-HNE. Mean ± SD, Student's t-test, *p = .0316. (E) Typical bands of 4-HNE and β-tubulin proteins in bEnd.3 cell by Western blot. Mean ± SD, Student's t-test, *p = .0189.

Hypoxia induces ferroptosis in cerebrovascular endothelial cells. (A) The mRNA expression levels of ferroptosis-related genes in the head of zebrafish were evaluated by RT-qPCR. Mean ± SD, Student's t-test, *p < .05, **p < .01, ***p < .001. N = 80 fish per group. (B) Measurement of iron content in zebrafish head. Mean ± SD, Student's t-test, normoxia vs. hypoxia (Total Fe), **p = .0047; normoxia vs. hypoxia (Fe²⁺), *p = .011; normoxia vs. hypoxia (Fe³⁺), *p = .039. N = 150 fish per group. (C) Immunoblotting analysis of GPX4 and xCT in the head of zebrafish. Mean ± SD, Student's t-test, normoxia vs. hypoxia (GPX4), **p = .002; normoxia vs. hypoxia (xCT), *p = .034. N = 100 fish per group. (D) The mRNA expression levels of ferroptosis-related genes in bEnd.3 cells were calculated by RT-qPCR. Mean ± SD, Student's t test, *p < .05, **p < .01, ***p < .001. (E) Protein expression level of GPX4 and xCT in bEnd.3 cells was analyzed by immunoblotting analysis. Mean ± SD, Student's t test, normoxia vs. hypoxia (GPX4), ***p < .001; normoxia vs. hypoxia (xCT), *p = .014. (F) Immunofluorescence staining and quantification of GPX4. Mean ± SD, Student's t-test, **p = .0034. Scale bar: 10 μm.

Ferroptosis inhibitor Fer-1 partially alleviated hypoxia-induced BBB disruption. (A) BBB disruption was measured by quantifying Rhodamine-Dextran dye leakage. N = 6 fish per group. Mean ± SD, one-way ANOVA, normoxia vs. hypoxia, ***p < .0001; hypoxia vs. Fer-1 + hypoxia, ###p = .0005. Scale bar: 50 μm. (B) BBB permeability detection. “Treatment” indicates the time point of hypoxia induction after the bEnd.3 cell monolayer reaches a stable TEER value. Mean ± SD, two-way ANOVA, normoxia vs. hypoxia (3 h), ***p < .0001; normoxia vs. hypoxia (6 h), ***p < .0001; normoxia vs. hypoxia (12 h), ***p < .0001; normoxia vs. hypoxia (24 h), ***p < .0001; hypoxia vs. Fer-1 + hypoxia (3 h), ^###p < .0001; hypoxia vs. Fer-1 + hypoxia (6 h), ^###p < .0001; hypoxia vs. Fer-1 + hypoxia (12 h), ^###p < .0001; hypoxia vs. Fer-1 + hypoxia (24 h), ^###p < .0001. (C) Representative images of bEnd.3 cells stained with ROS assay kit and Hoechst. Mean ± SD, one-way ANOVA, normoxia vs. hypoxia, ***p < .0001; hypoxia vs. Fer-1 + hypoxia, ^###p < .0001. scale bar: 100 μm. (D) Measurement of GSH and GSH/GSSG ratio in zebrafish head. Mean ± SD, one-way ANOVA, normoxia vs. hypoxia (GSH), ***p < .0001; hypoxia vs. Fer-1 + hypoxia (GSH), ^##p = .0041; normoxia vs. hypoxia (GSH/GSSG), **p = .0019; hypoxia vs. Fer-1 + hypoxia (GSH/GSSG), ^#p = .0315. N = 200 fish per group. (E) Determination of iron content in bEnd.3 cells. Mean ± SD, two-way ANOVA, normoxia vs. hypoxia (Total Fe), ***p < .0001; normoxia vs. hypoxia (Fe²⁺), ***p = .0002; normoxia vs. hypoxia (Fe³⁺), ***p < .0001; hypoxia vs. Fer-1 + hypoxia (Total Fe), ^###p < .0001; hypoxia vs. Fer-1 + hypoxia (Fe²⁺), ^##p = .0032; hypoxia vs. Fer-1 + hypoxia (Fe³⁺), ^###p < .0001. (F) The mRNA expression levels of ferroptosis-related genes in the head of zebrafish. Mean ± SD, one-way ANOVA, * or ^#p < .05, ** or ^##p < .01, *** or ^###p < .001. N = 80 fish per group. (G) The mRNA expression levels of ferroptosis-related genes in bEnd.3 cells. Mean ± SD, one-way ANOVA, ^##p < .01, *** or ^###p < .001. N = 100 fish per group. (H) The protein levels of GPX4, CLDN5 and 4-HNE in zebrafish head were detected by immunoblotting assay. Mean ± SD, one-way ANOVA, normoxia vs. hypoxia (GPX4), ***p = .0002; hypoxia vs. Fer-1 + hypoxia (GPX4), ^##p = .0017; normoxia vs. hypoxia (CLDN5), ***p < .0001; hypoxia vs. Fer-1 + hypoxia (CLDN5), ^#p = .0208; normoxia vs. hypoxia (4-HNE), **p = .0091; hypoxia vs. Fer-1 + hypoxia (4-HNE), ^##p = .0027. N = 100 fish per group. (I) Protein expression levels of GPX4, CLDN5 and 4-HNE in bEnd.3 cells were analyzed by immunoblotting analysis. Mean ± SD, one-way ANOVA, normoxia vs. hypoxia (GPX4), ***p < .0001; hypoxia vs. Fer-1 + hypoxia (GPX4), ^##p = .0018; normoxia vs. hypoxia (CLDN5), ***p = .0001; hypoxia vs. Fer-1 + hypoxia (CLDN5), ^###p = .0010; normoxia vs. hypoxia (4-HNE), *p = .0182; hypoxia vs. Fer-1 + hypoxia (4-HNE), ^#p = .0108. N = 100 fish per group.