High-intensity interval training improves cognitive dysfunction in chronically stressed mice through alleviating homocysteine-induced transcriptional repression of Cldn5

Abstract

Chronic stress-induced blood-brain barrier (BBB) dysfunction contributes to neurological disorders, with homocysteine (HCY) as a key risk factor. Considering pharmacotherapy limitations, non-invasive interventions like high-intensity interval training (HIIT) are promising. To determine whether HIIT improves stress-induced BBB dysfunction and cognitive impairment, we first established a chronic unpredictable mild stress (CUMS) model and assigned mice into four groups: Control (Ctrl), CUMS, HIIT, and HIIT + CUMS. Here, we found that HIIT significantly ameliorated cognitive impairment in male CUMS mice, as evidenced by reduced escape latency in morris water maze and increased memory performance in novel object recognition test. HIIT also preserved BBB integrity by ameliorating the tight junction disruption and BBB hyper-permeability in stressed mice. Subsequently, to clarify the role of HCY in the HIIT-mediated effects, we established an HHCY model and divided mice into four groups: Ctrl, HIIT, HHCY, and HIIT + HHCY. The results showed that HIIT normalized the plasma and hippocampal HCY levels by restoring the expression of related metabolic enzymes including CBS, MTHFR and MS, and alleviated HHCY-induced cognitive decline and BBB damage. Further, HIIT reversed HCY-induced Claudin-5 downregulation by inhibiting H3K27me3 enrichment at the Cldn5 (the encoding gene of Claudin-5) promoter region. In addition, HIIT restored the expression of ETS1, one of the transcriptional activators of Cldn5, to facilitate the transcription of Cldn5 gene and the stabilization of BBB. Collectively, these findings reveal that HIIT improves chronic stress-induced cognitive impairment via eliminating the disruptive effects of HHCY on the BBB integrity, offering a non-pharmacological intervention potential for stress-related cognitive deficits.

Keywords: BBB damage; Cldn5 transcriptional repression; HCY; HIIT; Stress-induced cognitive dysfunction.

Graphical abstract

Fig. 1

HIIT improves cognitive impairments and BBB disruption in CUMS mice. (A) Schematic diagram showing the strategy of HIIT intervention in CUMS mice. (B, C) Escape latency of first time to reach the platform (B) and entries into the target quadrant (C) in the probe trial of the MWM test (Two-way ANOVA with Bonferroni's post hoc test, n = 14 mice for each group). (D) Representative track images of mice in the probe trial of the MWM test. (E) Cognitive index of mice in the NORT (Two-way ANOVA with Bonferroni's post hoc test, n = 14 mice for each group). (F) Representative TEM images of TJs structure in the hippocampus of mice. Scale bar, 500 nm. (G) Quantification of percentages of discontinuous TJs (red arrows) in the (F) (Two-way ANOVA with Bonferroni's post hoc test, n = 3 mice for each group, 40–45 TJs per mouse). (H) Hippocampal BBB permeability to NaFI of the mice (Two-way ANOVA with Bonferroni's post hoc test, n = 4 mice for each group). Statistical significance: ∗∗p < 0.01, ∗∗∗p < 0.001 vs Ctrl; ^#p < 0.05, ^##p < 0.01, ^###p < 0.001 vs CUMS. Data are expressed as mean ± SEM.

Fig. 2

HIIT attenuates hippocampal HCY levels by normalizing the expression of HCY metabolic enzymes in CUMS mice. (A) Hippocampal HCY levels in the CUMS mice with HIIT intervention (Two-way ANOVA with Bonferroni's post hoc test, n = 4 mice for each group). (B) Representative images of HCY metabolic enzymes in the hippocampus of mice by western blotting. (C–F) Quantification of HCY metabolic enzymes including CSE (C), CBS (D), MS (E) and MTHFR (F) in the (B) (Two-way ANOVA with Bonferroni's post hoc test, n = 6 mice for each group). Statistical significance: ∗p < 0.05, ∗∗p < 0.01 vs Ctrl; ^#p < 0.05, ^##p < 0.01 vs CUMS. Data are expressed as mean ± SEM.

Fig. 3

HIIT reduces plasma HCY levels by repairing the expression of HCY metabolic enzymes in the liver of CUMS mice. (A) Plasma HCY levels in the CUMS mice with HIIT intervention (Two-way ANOVA with Bonferroni's post hoc test, n = 8 mice for each group). (B) Representative images of HCY metabolic enzymes in the liver of mice by western blotting. (C–G) Quantification of HCY metabolic enzymes including MTHFR (C), CBS (D), MS (E), BHMT (F) and CSE (G) in the (B) (Two-way ANOVA with Bonferroni's post hoc test, n = 6 mice for each group). Statistical significance: ∗p < 0.05, ∗∗p < 0.01 vs Ctrl; ^#p < 0.05, ^##p < 0.01 vs CUMS. Data are expressed as mean ± SEM.

Fig. 4

HIIT mitigates cognitive decline and BBB damage in HHCY mice. (A) Schematic diagram showing the strategy of HIIT intervention in HHCY mice. (B, C) Plasma (B) and hippocampus (C) HCY levels in the HHCY mice with HIIT intervention (Two-way ANOVA with Bonferroni's post hoc test, n = 14 mice for each group). (D, E) Escape latency of first time to reach the platform (D) and entries into the target quadrant (E) in the probe trial of the MWM test (Two-way ANOVA with Bonferroni's post hoc test, n = 14 mice for each group). (F) Representative track images of mice in the probe trial of the MWM test. (G) Cognitive index of mice in the NORT (Two-way ANOVA with Bonferroni's post hoc test, n = 14 mice for each group). (H) Representative TEM images of TJs structure in the hippocampus of mice. Scale bar, 500 nm. (I) Quantification of percentages of discontinuous TJs (red arrows) in the (H) (Two-way ANOVA with Bonferroni's post hoc test, n = 3 mice for each group, 45–50 TJs per mouse). (J) Hippocampal BBB permeability to NaFI of the mice (Two-way ANOVA with Bonferroni's post hoc test, n = 4 mice for each group). Statistical significance: ∗∗p < 0.01, ∗∗∗p < 0.001 vs Ctrl; ^#p < 0.05, ^##p < 0.01, ^###p < 0.001 vs HHCY. Data are expressed as mean ± SEM.

Fig. 5

HIIT inhibits H3K27me3 modification to facilitate Cldn5 transcription. (A) Representative images of Claudin-5 protein levels in the hippocampus of CUMS mice by western blotting. (B) Quantification of Claudin-5 protein levels in the (A) (Two-way ANOVA with Bonferroni's post hoc test, n = 4 mice each group). (C) Representative images of Claudin-5 protein levels in the hippocampus of HHCY mice by western blotting. (D) Quantification of Claudin-5 protein levels in the (C) (Two-way ANOVA with Bonferroni's post hoc test, n = 4 mice each group). (E, F) H3K27me3 modification levels at the Cldn5 promoter region in the CUMS (E) and HHCY (F) mice by ChIP-qPCR (Two-way ANOVA with Bonferroni's post hoc test, n = 3 mice each group). (G) Representative images of H3K27me3 (red) and Claudin-5 (green) immunostaining in the hippocampus of Ctrl, CUMS, HIIT and HIIT + CUMS mice. Scale bars, 200 μm (white arrowheads: the colocalization of H3K27me3 and Claudin-5 positive cells). (H) Quantification of H3K27me3 labeling intensity on vessels in the (G) (Two-way ANOVA with Bonferroni's post hoc test, n = 4 mice each group). (I) Representative images of H3K27me3 (red) and Claudin-5 (green) immunostaining in the hippocampus of Ctrl, HHCY, HIIT and HIIT + HHCY mice. Scale bars, 200 μm (white arrowheads: the colocalization of H3K27me3 and Claudin-5). (J) Quantification of H3K27me3 labeling intensity on vessels in the (I) (Two-way ANOVA with Bonferroni's post hoc test, n = 4 mice each group). Statistical significance: ∗p < 0.05, ∗∗p < 0.01 vs Ctrl; ^##p < 0.01 vs CUMS. Data are expressed as mean ± SEM.

Fig. 6

ETS1 is required for HIIT-driven Cldn5 transcriptional activation. (A) Schematic diagram of the dual-luciferase reporter gene system for detecting the binding capacity of ETS1 to Cldn5. (B) Fluorescence activity in bEnd.3 cells with Cldn5 wildtype promotor with ETS1 pGL4.10. (Two-way ANOVA with Bonferroni's post hoc test, n = 4 biological replicates). (C) Representative images of ETS1 protein levels in the bEnd.3 cells by western blotting. (D) Quantification of the ETS1 protein levels in (C) (One-way ANOVA with Tukey's post hoc test, n = 3 biological replicates). (E) Representative images of ETS1 protein levels in the hippocampus of CUMS mice by western blotting. (F) Quantification of ETS1 protein levels in the (E) (Two-way ANOVA with Bonferroni's post hoc test, n = 3 mice each group). (G) Representative images of ETS1 protein levels in the hippocampus of HHCY mice by western blotting. (H) Quantification of ETS1 protein expression levels in the (G) (Two-way ANOVA with Bonferroni's post hoc test, n = 3 mice each group). Statistical significance: ∗p < 0.05, ∗∗p < 0.01vs Ctrl or 0 μM; ^##p < 0.01vs CUMS or HHCY. Data are expressed as mean ± SEM.

Fig. 7

The proposed hypothesis for the mechanisms by which HIIT attenuates chronic stress-induced BBB dysfunction and cognitive deficits. HIIT alleviates CUMS-induced BBB damage and cognitive impairment in the CUMS mice. Mechanistically, HIIT restores stress-induced HHCY via the normalization of HCY metabolic enzymes. Moreover, HIIT-mediated reduction of HCY levels not only hampers H3K27me3 enrichment at the Cldn5 promoter region but also promotes the expression of ETS1, a transcriptional activator of Cldn5, together enhancing Claudin-5 expression to repair BBB damage.