Alteration of Gut Microbiome and Correlated Lipid Metabolism in Post-Stroke Depression

Abstract

Background: The pathogenesis of post-stroke depression (PSD) remains largely unknown. There is growing evidence indicating that gut microbiota participates in the development of brain diseases through the gut-brain axis. Here, we aim to determine whether and how microbial composition and function altered among control, stroke and PSD rats.

Materials and methods: After the PSD rat model was successfully established, gut microbiome combined with fecal metabolome approach were performed to identify potentially PSD-related gut microbes and their functional metabolites. Then, correlations between behavior indices and altered gut microbes, as well as correlations between altered gut microbial operational taxonomic units (OTUs) with differential metabolites in PSD rats were explored. Enrichment analysis was also conducted to uncover the crucial metabolic pathways related to PSD.

Results: Although there were some alterations in the microbiome and metabolism of the control and stroke rats, we found that the microbial and metabolic phenotypes of PSD rats were significantly different. The microbial composition of PSD showed a decreased species richness indices, characterized by 22 depleted OTUs mainly belonging to phylum Firmicutes, genus Blautia and Streptococcus. In addition, PSD was associated with disturbances of fecal metabolomics, among them Glutamate, Maleic acid, 5-Methyluridine, Gallocatechin, 1,5-Anhydroglucitol, L-Kynurenine, Daidzein, Cyanoalanine, Acetyl Alanine and 5-Methoxytryptamine were significantly related to disturbed gut microbiome (P ≤ 0.01). Disordered fecal metabolomics in PSD rats mainly assigned to lipid, amino acid, carbohydrate and nucleotide metabolism. The steroid biosynthesis was particularly enriched in PSD.

Conclusions: Our findings suggest that gut microbiome may participate in the development of PSD, the mechanism may be related to the regulation of lipid metabolism.

Keywords: gut microbiome; lipid metabolism; metabolic pathways; metabolome; post-stroke depression.

Figure 1

Time schedule of experimental procedures and TTC-stained brain sections. (A) Time schedule of experimental procedures. CUMS, chronic unpredictable mild stress; W0, Beginning of CUMS; W1–W4, CUMS was performed as described in the materials and methods section for 4 weeks; BW, body weight; SPT, sucrose preference test; OFT, open field test; FST, forced swimming test. (B) The representative TTC-stained brain sections. (C) Quantification of infarction volumes was calculated based on TTC staining (n = 6 per group, P > 0.05). C, Control; S, Stroke; P, post-stroke depression.

Figure 2

Body weight and behavioral tests. (A) Body weight during CUMS of three group (n = 8 per group); (B) Sucrose preference during CUMS of three group (n = 8 per group); (C) The total distance of OFT was no significant difference among 3 groups after a 4-week CUMS exposure (n = 8 per group); (D) Percentage of duration time spent in the center square of OFT was compared among 3 groups after a 4-week CUMS exposure (n = 8 per group); (E) Immobility time comparison of FST among 3 groups after a 4-week CUMS exposure (n = 8 per group); C, Control; S, Stroke; P, PSD; PSD group vs. control group, **P < 0.01 and ***P < 0.001; PSD group vs. stroke group, ●● P < 0.01; SPT, sucrose preference test; OFT, Open field test; FST, Forced swimming test.

Figure 3

Gut microbial characteristics of control, stroke and PSD. (A) α-phylogenetic diversity analysis showing that PSD subjects were characterized by lower microbial richness in two indexes (Ace, Chao) relative to controls (n=6 per group), *P < 0.05, **P < 0.01. (B) At the OTU level, principal co-ordinates analysis (PCoA) showed gut microbial composition of rats with PSD was significantly different from that in control and stroke (n = 6 per group). (C) Relative abundance of gut microbes at the family level.

Figure 4

Linear discriminant analysis effect size (LEfSe) analysis (LDA > 2.0). Cladogram (A) and histogram (B) illustrated 91 OTUs responsible for discriminating the PSD, stroke and control groups. Compared to stroke and control groups, the PSD rats were characterized by 22 discriminative OTUs (n = 6 per group).

Figure 5

Associations of altered gut microbes with behavior indices. Heat map of the Spearman’s rank correlation coefficient of 22 discriminative OTUs for PSD and 3 behavior indices. Red rectangle indicates positive associations between these microbial species and behavior indices, blue rectangle indicates negative associations (n = 6 per group). Overall, 22 discriminative OTUs for PSD were positively associated with FST, and negatively associated with SPT and OFT. 11 of 22 differential microbial variances (50%) were significantly associated with 3 behavior indices (p value < 0.05) and correlation coefficient were ≥0.45 or ≤ −0.45, tested by spearman correlation. The statistical significance was denoted on the rectangle (*p < 0.05; **p < 0.01; ※p < 0.001).

Figure 6

Fecal metabolism characteristics of PSD and its related KEGG enrichment pathways. (A) Orthogonal Partial Least Squares Discrimination Analysis (OPLS-DA) showed fecal metabolism of PSD was significantly different from that in control and stroke (n = 8 per group). (B) The 25 differential metabolites for PSD rats compared with control and stroke rats were enriched in 19 KEGG pathways. (C) Associations of gut microbial OTUs with fecal metabolites. Heat map of the spearman’s rank correlation coefficient of 22 discriminative OTUs and 25 differential metabolites for PSD. Red squares indicate positive associations between these microbial OTUs and metabolites, blue squares indicate negative associations (n = 8 per group). 54.55% (12/22 OTUs) of altered bacterial OTUs showing significant correlations with a range of differential metabolites (p < 0.05) and correlation coefficient were ≥ 0.6 or ≤ −0.6, tested by spearman correlation. The statistical significance was denoted on the squares (*p < 0.05; **p < 0.01; ※p < 0.001).