Vascular and blood-brain barrier-related changes underlie stress responses and resilience in female mice and depression in human tissue

Abstract

Prevalence, symptoms, and treatment of depression suggest that major depressive disorders (MDD) present sex differences. Social stress-induced neurovascular pathology is associated with depressive symptoms in male mice; however, this association is unclear in females. Here, we report that chronic social and subchronic variable stress promotes blood-brain barrier (BBB) alterations in mood-related brain regions of female mice. Targeted disruption of the BBB in the female prefrontal cortex (PFC) induces anxiety- and depression-like behaviours. By comparing the endothelium cell-specific transcriptomic profiling of the mouse male and female PFC, we identify several pathways and genes involved in maladaptive stress responses and resilience to stress. Furthermore, we confirm that the BBB in the PFC of stressed female mice is leaky. Then, we identify circulating vascular biomarkers of chronic stress, such as soluble E-selectin. Similar changes in circulating soluble E-selectin, BBB gene expression and morphology can be found in blood serum and postmortem brain samples from women diagnosed with MDD. Altogether, we propose that BBB dysfunction plays an important role in modulating stress responses in female mice and possibly MDD.

Fig. 1. Chronic social stress induces region-dependent neurovascular changes in female mice.

a Experimental timeline of 10-day chronic social defeat stress (CSDS), social interaction (SI) and tissue collection of nucleus accumbens (NAc) and prefrontal cortex (PFC). b Individual SI values (left, ****p < 0.0001), time (s) in corners with aggressor (AGG) present (middle, **p = 0.0038) and representative heatmaps of normalized time spent in the arena during SI test (right). c Quantitative PCR revealed changes in the NAc of stress susceptible (SS) and resilient (RES) mice when compared to unstressed controls (CTRL) of gene expression related to endothelial cells, angiogenesis, tight junctions, and blood-brain barrier (BBB) formation, (d) but Cldn5 levels remained unchanged (p = 0.1015 for CTRL vs SOCIAL STRESS; p = 0.1262 for CTRL vs SS vs RES). The range of color indicates individual differences within a group; average represented by the dashed line. e Quantitative PCR in the PFC revealed region-specific changes in SS and RES mice when compared to unstressed CTRL, and Cldn5 mRNA (**p = 0.0042 for CTRL vs SOCIAL STRESS; **p = 0.0055 for CTRL vs SS vs RES) (f) and protein levels (g, h) were lower in the PFC of SS mice (***p = 0.0005) and correlated with social avoidance (**p = 0.0073). Scale bars, 20 μm. i CLDN5 mRNA fold change was significantly lower in the PFC of women with major depressive disorder (MDD) (left, **p = 0.0068) along with morphological vascular alterations (right, *p = 0.0288). No significant difference was observed for men (left, p = 0.5387). Scale bars, 20 μm. Data represent mean ± s.e.m; the number of animals or subjects (n) is indicated on graphs. Correlations were evaluated with Pearson’s correlation coefficient; 2-group comparisons were evaluated with unpaired t-tests and one-way ANOVA followed by Bonferroni’s multiple comparison test for other graphs. *p < 0.05; **p < 0.01; ****p < 0.0001. Source data are provided as a Source Data file.

Fig. 2. Six-day chronic variable stress induces behavioural and region-dependent neurovascular changes in female mice.

a Experimental timeline of 6-d chronic variable stress (SCVS) and behavioural studies. b SCVS induces slight social interaction deficits (p = 0.0548) and decreased time in the interaction zone in stressed female mice when the social target (aggressor, AGG) is present (*p = 0.0323), in the social interaction test. 6-d of CVS is enough to induce significant anxiety- and depression-like behaviours in the elevated plus maze (c) (*p = 0.026) and sucrose preference (d) tests (***p = 0.0004 for 24 h; ***p = 0.0007 for 48 h; **p = 0.0048 for average). e Quantitative PCR of genes related to endothelial cell biology, angiogenesis, tight junctions and blood-brain barrier (BBB) formation reveals significant changes in the nucleus accumbens (NAc) of stressed female mice (f), and downregulation of Cldn5 mRNA levels. g Quantitative PCR of those genes in the prefrontal cortex (PFC) reveals region-specific neurovascular changes in stressed female mice vs controls (CTRL) (h) and significant downregulation of Cldn5 mRNA expression (*p = 0.0125). i CLDN5 mRNA fold change was significantly lower in the NAc of men and women with major depressive disorder (MDD) (*p = 0.0275), but (j) immunostaining revealed no significant difference in CLDN5 expression in the brain of women with MDD. Scale bars, 20 μm. Data represent mean ± s.e.m; the number of animals or subjects (n) is indicated on graphs. 2-group comparisons were evaluated with unpaired t-tests. *p < 0.05; **p < 0.01; ***p < 0.001. Source data are provided as a Source Data file.

Fig. 3. Conditional knockdown of Cldn5 expression in the PFC is sufficient to induce anxiety- and depressive-like behaviours.

a Experimental timeline of prefrontal cortex (PFC) bilateral injection of AAV-shRNA of AAV-shRNA-Cldn5 viruses and behavioural studies. b Cldn5 mRNA (**p = 0.0021) and protein (*p = 0.0403) levels are reduced following AAV-shRNA-Cldn5 injection in the PFC of female mice compared to that in AAV-shRNA-injected mice, following doxycycline (Dox) treatment. Scale bars, 20μm. Subthreshold microdefeat (stressed mice) did not have a significant effect on anxiety- and depressive-like behaviours in the elevated plus maze (c), splash test (d), sucrose preference test at 48 h (**p = 0.0064 unstressed AAV-shRNA vs AAV-shRNA-Cldn5; **p = 0.0034 stressed AAV-shRNA vs AAV-shRNA-Cldn5) (e), forced swim test (f, FST) and for social interactions (*p = 0.0317 unstressed AAV-shRNA vs stressed AAV-shRNA; **p = 0.0089 unstressed AAV-shRNA-Cldn5 vs stressed AAV-shRNA-Cldn5) (g) tests. h However, main virus effects are observed when comparing all AAV-shRNA vs AAV-shRNA-Cldn5 animals (*p = 0.0129 for elevated plus maze; **p = 0.0084 for splash test; ****p < 0.0001 for sucrose preference at 24 h; *p = 0.0453 for forced swim test). i Intraindividual correlation of different behavioural data points reveals trends between anxiety- and depressive-like behaviours. P values in the boxes refer to the strength of the correlation between behaviours. j Experimental timeline of PFC bilateral injection of AAV-shRNA or AAV-shRNA-Cldn5 viruses and social interaction (SI) tests. k No significant difference in SI ratio was observed when a male social target (aggressor, AGG) was present, but a significant difference was found when a female AGG was present (*p = 0.0123). l SI ratios with male and female AGG values are significantly correlated to each other, and with Cldn5 mRNA levels. Data represent mean ± s.e.m; the number of animals or subjects (n) is indicated on graphs. Two-way ANOVA followed by Bonferroni’s multiple comparison test for behaviours, Correlations were evaluated with Pearson’s correlation coefficient; 2-group comparisons were evaluated with unpaired t-tests. *p < 0.05; **p < 0.01; ****p < 0.0001. Source data are provided as a Source Data file.

Fig. 4. Susceptibility vs resilience to chronic social stress is associated with specific endothelium transcriptome-wide changes in the female PFC.

a Experimental timeline of 10-d chronic social defeat stress (CSDS) and (b) phenotype in the social interaction (SI, ****p < 0.0001; *p = 0.0178 for middle panel) test of the mice used to establish prefrontal cortex (PFC) endothelial gene profiles. c Hierarchical clustering heatmap of unstressed controls (CTRL), stress-susceptible (SS) and resilient (RES) mice, reveals that the RES group is the most distinct (purple circle, significance was set at ±2-fold change and p < 0.05). In fact, normalization of gene expression changes on the CTRL group shows that in stressed mice a majority of genes are regulated in the opposite direction according to the phenotype. d Venn diagrams indicate poor overlap of gene expression changes when group comparisons were performed with the largest number of genes associated to RES vs CTRL animals. Most significant biological pathways for each group comparison are displayed on the right according to the group comparison color. e SS vs RES volcano plot highlights some of the most up- and downregulated genes. Data represent mean ± s.e.m; the number of animals or subjects (n) is indicated on graphs. One-way ANOVA followed by Bonferroni’s multiple comparison test for behaviours. *p < 0.05; **p < 0.01; ****p < 0.0001. Source data are provided as a Source Data file.

Fig. 5. Subchronic variable stress induces transcriptome-wide changes in the female PFC endothelium.

a Experimental timeline of the 6-d subchronic variable stress (SCVS) and tissue collection for endothelial cell-specific molecular profiling. b Hierarchical clustering heatmap of unstressed controls (CTRL) vs mice subjected to SCVS indicates distinct transcriptomic gene expression (significance was set at ±2-fold change and p < 0.05). c Hierarchical clustering heatmap of all stressed mice so animals exposed to either 10-d CSDS or 6-d SCVS highlights a level of commonality in gene expression changes between females characterized by depression-like behaviours namely the stress-susceptible (SS) and SCVS group when compared to resilient (RES) subjects (green circle). SS and SCVS females cluster together and normalization of gene expression on RES animals reveals an overlap in the directionality of endothelium transcriptome changes. d Venn diagrams show a poor overlap of gene expression changes when group comparisons were performed, particularly for the SCVS vs RES subjects. Most significant biological pathways for each group comparison are indicated on the right. e SCVS vs RES volcano plot highlights some of the most up- and downregulated genes. Data represent mean ± s.e.m; the number of animals or subjects (n) is indicated on graphs. One-way ANOVA followed by Bonferroni’s multiple comparison test for behaviours. *p < 0.05; **p < 0.01; ****p < 0.0001. Source data are provided as a Source Data file.

Fig. 6. Chronic stress induces BBB leakiness in the female PFC and release of vascular biomarkers in the blood of mice also observed in individuals with MDD.

a Experimental timeline of 10-d chronic social defeat stress (CSDS), social interaction (SI) test and retro-orbital injection of fluorescent-labelled dextran (10 kDa). Following behavioural phenotyping (SI ratio, *p = 0.0273), assessment of blood-brain barrier (BBB) permeability with dextran Alexa Fluor-488 dye revealed significant BBB leakiness in the prefrontal cortex (PFC) of stress-susceptible (SS) female mice (fluorescence, *p = 0.0144; frequency, ***p = 0.0005). Scale bars, 20 μm (b). c, d Experimental timeline of CSDS and subchronic variable stress (SCVS) paradigms and blood collection for multiplex assays. e, f Cardiovascular multiplex assays reveals significant upregulation of circulating soluble E-selectin (sE-selectin) when compared to baseline (pre-CSDS) serum levels (*p = 0.0267 for SOCIAL STRESS) in SS, but not RES mice (*p = 0.0131), and post-CSDS levels correlated with social avoidance (f). g, h Circulating sE-selectin levels were increased in female mice following 6-d SCVS without reaching significance and remained unchanged in male mice following 10-d CSDS, and no correlation was found with SI ratio. i, Baseline serum levels of sE-selectin was ~40% lower in female vs male mice (****p < 0.0001). j, k Circulating sE-selectin levels were significantly upregulated in women (*p = 0.0494) but not men with major depressive disorder (MDD), and levels in healthy controls were ~25% lower in women when compared to men (*p = 0.0499) (k). Data represent mean ± s.e.m; the number of animals or subjects (n) is indicated on graphs. Correlations were evaluated with Pearson’s correlation coefficient; 2-group comparisons were evaluated with unpaired t-tests and one-way ANOVA followed by Bonferroni’s multiple comparison test for other graphs. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. Source data are provided as a Source Data file.