Introduction

Post-stroke depression (PSD), characterized by negative mood and disinterest in activities, is a common psychiatric complication following stroke.1 PSD adversely affects physical and psychological functioning, reduces quality of life, and increases mortality risk in stroke survivors.2 Currently, PSD diagnosis primarily relies on symptom evaluation by physicians, which can be influenced by subjective factors. Additionally, the pathogenesis of PSD remains poorly understood, leading to inadequate assessment and treatment.3 Although mechanisms such as increased inflammation, decreased monoamine levels, and dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis have been proposed, these do not fully explain the pathogenesis and progression of PSD.4 Therefore, further investigation is needed to uncover the mechanisms and identify potential biomarkers associated with PSD.

Gut microbiota play a significant role in the pathophysiology of the human body. As studies of the gut–brain axis have evolved, the dysbiosis of gut microbiota in PSD has been revealed.5 Recently, based on feces collected within one week after stroke onset, a clinical study indicated that patients with PSD had higher abundance of Streptococcus, Akkermansia, and Barnesiella, and lower abundance of Escherichia-Shigella, Butyricicoccus, and Holdemanella.6 Another study conducted within 2–4 weeks of stroke onset revealed increased abundance of Escherichia coli and Enterococcus faecalis, and decreased abundance of Bifidobacterium in patients with PSD.7 However, these findings are inconsistent, likely because of differences in the timing of sample collection.

At the same time, plasma metabolites are influenced by gut microbiota, constituting a significant pathway for the communication of the gut–brain axis.8 A plasma metabolomic study in acute PSD (<1 month) found alterations in lipid and amino acid metabolism.9 Another metabolomic study at 2 weeks post-stroke indicated that the glycerophospholipid metabolism, citrate cycle, alanine, aspartate, and glutamate metabolism were associated with PSD.10 However, these studies used non-targeted metabolomics with poor sensitivity and low detectable concentration. Therefore, high-throughput targeted approaches are needed to accurately identify and quantify metabolites in PSD. Furthermore, the aforementioned studies employed disparate specimen collection times, resulting in markedly disparate outcomes. Thus, identifying the optimal sampling time is crucial for elucidating the mechanisms of PSD.

A systematic review suggested that long-term PSD may have a different mechanism compared to early-onset PSD.11 Three months seems to be a watershed period for PSD, as biological shifts may occur during this period compared to that in the early stage.12 Despite this, no clinical studies have combined gut microbiota and plasma metabolomics to analyze PSD mechanisms at this time point. Understanding gut microbiota and plasma metabolite changes at 3 months post-stroke onset may provide valuable insights into PSD pathogenesis.

In this study, we used 16S rRNA sequencing and targeted metabolome analysis to compare gut microbiota and plasma metabolites between patients with and without PSD 3 months after acute ischemic stroke (AIS). We aimed to clarify the association between these two omics and identify potential diagnostic biomarkers for PSD at this time point.

Materials and Methods Subjects

This study was conducted at Xuanwu Hospital, Capital Medical University, from June 2021 to October 2023, and was registered as a clinical trial (ChiCTR2100041895). All procedures were approved by the Ethics Committee of Xuanwu Hospital, Capital Medical University (LYS [2020] 096). Written informed consent was obtained from all participants. A total of 86 patients with AIS were recruited within 72 hours of onset. Ultimately, 70 patients were followed up in an outpatient clinic at 3 months as shown in Figure S1.

Inclusion criteria were (1) age ≥ 18 years; and (2) AIS diagnosed with computed tomography or magnetic resonance imaging. Exclusion criteria were: (1) a history of mental disease or use of psychotropic drugs before stroke onset; (2) prior diagnosis of depression; (3) serious systemic disease such as cancer; (4) cerebral hemorrhage; (5) history of severe intestinal diseases; (6) antibiotic or probiotic use within 3 months; (7) inability to complete psychological assessments.

At 3 months post-AIS onset, PSD was diagnosed by two experienced psychiatrists, based on the 24-item Hamilton Depression Rating Scale (HDRS) with scores ≥ 8, along with the Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5) criteria.13 The PSD group was divided into mild PSD group (8–16), moderate-severe PSD group (≥ 17) according to HAMD scores.14

Clinical Data and Sample Collection

Clinical data collected included age, sex, body mass index (BMI), medical history, treatment plan, Trial of Org 10172 in Acute Stroke Treatment (TOAST) classification, lesion location, HDRS, National Institutes of Health Stroke Scale (NIHSS), Modified Rankin Scale (mRS), and Montreal Cognitive Assessment (MoCA). At the 3-month follow-up after stroke onset, plasma and fresh fecal samples were collected. Detailed sample collection methods were presented in the supplementary method 1.

16S rRNA Sequencing and Analysis

According to the supplementary method 2, DNA was extracted from fecal samples of all participants, and the V3-V4 variable region of the microbial 16S rRNA gene was amplified and sequenced. Library construction was performed using the TruSeq Nano DNA LT Library Prep Kit (Illumina), with subsequent inspection facilitated using Bioanalyzer 2100 (Agilent) and QuantiFluor™ dsDNA System (Promega). Representative reads and an ASV abundance table were generated from the raw sequencing data after quality control using QIIME2 software. Taxonomic information was used to assess community composition, diversity, and difference analysis.

Targeted Metabolome Analysis Using Liquid Chromatography-Mass Spectrometry (LC-MS)

As mentioned in supplementary method 3, plasma samples from all participants underwent targeted LC-MS metabolome analysis using ultra-high performance liquid chromatography quadrupole trap tandem mass spectrometry. The targeted metabolome approach detected over 600 metabolites related to medical research, including carbohydrates, organic acids, amino acids, bile acids, indoles, purine nucleotides, lipids, and other metabolites.15

Data Analysis

Statistical analyses were performed using SPSS version 25.0. For the clinical data, continuous variables were described as mean ± standard deviation or median and quartile (Q1, Q3). The Student’s t-test or nonparametric Mann–Whitney U-test was used to compare continuous variables. Categorical variables were presented as n (%) and analyzed using the chi-squared or Fisher’s test.

For gut microbiota, α diversity indices (Simpson, Shannon, Chao1, and Ace) were calculated using the Wilcoxon test, and β diversity was assessed through principal coordinates analysis (PCoA) using unweighted and weighted UniFrac dissimilarity indices. System-Theoretic Accident Model and Processes (STAMP) analysis identified discriminative microbiota between the PSD and non-PSD groups.16 STAMP can be used to compare samples from two or more groups to analyze classification and functional profiles, and assess biological relevancy by providing effect sizes and confidence intervals. Metabolomics analyses were conducted using MetaboAnalyst 5.0 (https://www.metaboanalyst.ca/), including principal component analysis (PCA) score plots and orthogonal partial least squares discriminant analysis (OPLS-DA). Metabolites with fold change < 0.83 or > 1.2, variable of interest > 1 and P < 0.05 were considered statistically significant.17 Spearman’s rank correlation analysis was conducted to explore associations between significant gut microbiota and plasma metabolites. Logistic regression identified biomarkers based on differential variables in each dataset, and receiver operating characteristic (ROC) curves evaluated the diagnostic efficiency of potential biomarkers. P < 0.05 was defined as the significance threshold.

Results Clinical Characteristics

At the 3-month follow-up, 70 patients with AIS were investigated, including 25 (35.71%) PSD and 45 (64.29%) non-PSD. The PSD group included 16 patients with mild PSD and 9 patients with moderate-severe PSD. Clinical data are presented in Table 1. Patients with PSD had significantly higher HDRS scores than those without PSD (13 [10–22.5] vs 5 [3–6], P < 0.001). No significant differences were observed between the two groups in terms of age, sex, BMI, medical history, TOAST classification, treatment, lesion location, NIHSS, mRS, and MoCA scores (all P > 0.05).

Table 1 Clinical Characteristics Between PSD and Non-PSD Groups

Gut Microbiota Difference Between PSD and Non-PSD Groups at 3 months

The α diversity indices (Simpson, Shannon, Chao1, and Ace) showed no differences between both groups (all P > 0.05, Figure S2A–D), indicating similar species diversity in both groups. Nevertheless, β-diversity analysis revealed notable disparities between the two groups, as shown by PCoA scatterplot based on unweighted and weighted UniFrac distance, indicating distinct gut microbial composition between the two groups (Figure 1A and B).

Figure 1 β Diversity of gut microbiota between PSD and non-PSD patients. (A) PCoA based on the unweighted UniFrac dissimilarity index. (B) PCoA based on the weighted UniFrac dissimilarity index.

Abbreviations: PCoA, principal coordinates analysis; PSD, post-stroke depression.

Gut microbial composition in patients with and without PSD at phylum and genus levels is presented in Figure 2A and B. At the phylum level, both groups had gut microbiota predominantly composed of Bacillota, Actinomycetota, Pseudomonadota, Bacteroidota, and Verrucomicrobiota (Figure 2A). At the genus level, Bifidobacterium, Faecalibacterium, Bacteroides, Prevotella, Collinsella, and Blautia were the dominant genera in both groups (Figure 2B). STAMP analysis identified that Synergistota was significantly more abundant in patients with PSD at the phylum level (Figure 2C). At the genus level, Parabacteroides, Pyramidobacter, Anaeroglobus, Haliangium, Staphylococcus, CAG−56, Shuttleworthia, and Epulopiscium were significantly more abundant in patients with PSD, whereas Eubacterium eligens group and Prevotella were more abundant in patients without PSD (Figure 2D).

Figure 2 Taxonomic differences in gut microbiota between PSD and non-PSD groups. (A) Structure of gut microbiota at the dominant phylum levels. (B) Structure of gut microbiota at the dominant genus levels. (C) Significantly different phyla between PSD (red) and non-PSD (green) groups. (D) Significantly different genera between PSD (red) and non-PSD (green) groups.

Abbreviation: PSD, post-stroke depression.

Plasma Metabolites Difference Between PSD and Non-PSD Groups at 3 Months

PCA clustering analysis revealed a pattern of separation among certain samples and overlap in other samples (Figure 3A). OPLS-DA further clarified significant differences in plasma metabolic characteristics between the two groups (Figure 3B). Twelve altered plasma metabolites were observed in patients with PSD, with 2 metabolites (cortisol and pyroglutamic acid) upregulated and 10 metabolites (2-phosphoglyceric acid, 3-phosphoglycerate, phosphorylcholine, tryptophan, caffeine, n-methylalanine, ornithine, serotonin, theophylline, and vanillic acid) downregulated (Figure 3C). These metabolites primarily belonged to carboxylic acids, organooxygen compounds, steroids and indoles (Table S1). KEGG pathway enrichment analysis identified significant alterations in glutathione metabolism (P = 0.006), tryptophan metabolism (P = 0.013), and caffeine metabolism (P = 0.045) (Figure 3D). Furthermore, the concentration of cortisol was higher in moderate-severe PSD group than in mild PSD group (Table S2).

Figure 3 Altered metabolites in plasma of PSD group compared to non-PSD group. (A) PCA score plots of plasma metabolic profile. (B) OPLS-DA score plots of metabolic profiles. (C) The volcano plot analysis. (D) KEGG pathway enrichment analysis of the differential metabolites using MetaboAnalyst.

Abbreviations: PSD, post-stroke depression; PCA, principal component analysis; OPLS-DA, orthogonal partial least squares discriminant analysis.

Correlation Between Gut Microbiota and Plasma Metabolites

Pyramidobacter was negatively correlated with phosphorylcholine and ornithine, while Anaeroglobus was negatively correlated with theophylline and vanillic acid. Haliangium was negatively correlated with caffeine and theophylline, and Epulopiscium was negatively correlated with serotonin. Shuttleworthia was negatively correlated with tryptophan, while Eubacterium eligens and Prevotella were positively correlated with tryptophan (Figure 4, Table S3). Overall, microbiota more abundant in patients with PSD were negatively correlated with downregulated metabolites, while the microbiota more abundant in the non-PSD patients were reversed. Notably, these metabolites correlated with microbiota were primarily involved in tryptophan, caffeine, and glutathione metabolism. These results revealed a significant correlation between gut microbial dysregulation and key metabolic pathway alterations in PSD.

Figure 4 Spearman correlation between gut microbiota and the concentration of PSD-related differential metabolites in plasma.

Abbreviation: PSD, post-stroke depression.

Note: *P < 0.05.

Potential Combined Biomarkers for 3-Month PSD Diagnosis

After screening for potential biomarkers by logistic regression, Parabacteroides and Staphylococcus were included in the gut microbiota dataset, while tryptophan and theophylline were included in the plasma metabolite dataset. These datasets were further integrated into a combination dataset. ROC curve analysis demonstrated that the gut microbiota and plasma metabolite datasets had area under curve (AUC) values of 0.704 and 0.875, respectively. A combined dataset consisting of Parabacteroides, Staphylococcus, tryptophan, and theophylline yielded an AUC of 0.940 for diagnosing 3-month PSD (Figure 5).

Figure 5 ROC curves representing the diagnostic ability in each dataset.

Abbreviation: ROC, receiver operating characteristic.

Discussion

We observed significant differences in gut microbiota and plasma metabolites between PSD and non-PSD groups at 3 months. Combined omics analysis revealed significant correlation between gut microbiota dysregulation and alterations in plasma metabolites. Furthermore, we identified a panel of combined biomarkers, including gut microbiota and plasma metabolites, for 3-month PSD diagnosis.

Gut Microbiota Difference Between PSD and Non-PSD Groups at 3 months

Our study found significant differences in β diversity, but not α diversity, between PSD and non-PSD groups at 3 months. This aligns with a meta-analysis in a Chinese population, which also reported no significant differences in α diversity between patients with PSD and healthy controls,18 confirming significant changes in the gut microbial composition in PSD, but no difference in species diversity and richness.

At the phylum level, Synergistota was significantly higher in patients with PSD in our study. However, the pathogenesis of Synergistota in depression remains unclear. Only one cohort study in Spain found a decreased abundance of Synergistota phylum in individuals with depression symptoms.19 Future research at the genus or species level may provide more robust conclusions. Furthermore, a recent study demonstrated that Synergistota was one of the predominant microbiota in saliva in patients with early-onset cryptogenic ischemic stroke and was more abundant compared with controls, indicating that this phylum may play a potential role in the development and prognosis of ischemic stroke. Future researches could focus on the potential link between Synergistota and the oral-gut-brain axis.20

At the genus level, we observed an elevated abundance of certain pathogenic microbiota in patients with PSD at 3 months, including Pyramidobacter, Staphylococcus, and Parabacteroides, which has also been found in the feces of humans and animals with depression previously.6,21 A cohort study on 232 patients with AIS revealed a positive correlation between Pyramidobacter and HDRS score.6Pyramidobacter has also been positively correlated with interleukin (IL)-6,22 suggesting a pro-inflammatory role in PSD progression. Similarly, a higher abundance of Staphylococcus has been observed in mice exhibiting severe depressive symptoms.23 As a common gram-positive bacterium, Staphylococcus can secrete various enterotoxins to stimulate inflammatory cells and promote inflammatory responses,24 suggesting a potential influence on the development of depression. Moreover, Parabacteroides has been reported to disrupt the balance between indole-3-lactate and indole-3-carboxaldehyde, inducing depressive symptoms in both mice and humans.25 Additionally, we found elevated levels of some potentially pathogenic microbiota in patients with PSD, including Epulopiscium, Anaeroglobus, and Shuttleworthia. Despite the absence of direct evidence linking Epulopiscium, Anaeroglobus, and PSD, their increased abundance has been associated with other neuropathological conditions such as cognitive impairment26 and cerebral small vessel disease.27 There is also evidence that Shuttleworthia can induce inflammation such as endocarditis.28

Furthermore, regarding patients with PSD, our study found a significant downregulation of beneficial microbiota, including Eubacterium eligens and Prevotella, which have been shown to have a strong association with depression. These microbiota have anti-inflammatory properties and can modify the gut–brain axis by regulating the synthesis of neurotransmitters and short-chain fatty acids.29,30 An animal experiment showed that Prevotella could repair intestinal leakage, and inhibit inflammation by reducing levels of pro-inflammatory factors in the intestinal system and hippocampus, alleviating depressive symptoms in mice.31 In vitro cell-based experiments proved that Eubacterium eligens promoted the synthesis of IL-10, an anti-inflammatory cytokine.32Eubacterium also produces butyric and propionic acids, improving the intestinal barrier integrity and inhibiting inflammation.33 Overall, our findings indicate that PSD is associated with gut microbiota dysbiosis, characterized by an elevated abundance of potentially pathogenic and pro-inflammatory microbiota and a decreased abundance of anti-inflammatory microbiota. This imbalance may be a vital factor in the etiology of PSD. The differential microbiota in PSD patients at three months was not entirely consistent with the early PSD fecal samples. However, the results at both time points demonstrated an increase in opportunistic pathogens and a decrease in beneficial bacteria.6 Dysbiosis and increased intestinal permeability may result in a systemic low-grade inflammatory response. Pro-inflammatory cytokines or bacterial metabolites can cross the blood-brain barrier and alter neurotransmitter metabolism.34 Fecal microbiota transplantation introduces healthy microbiota into the gastrointestinal tract and helps promote recovery from ecological dysbiosis. Supplementation with probiotics may act on inflammatory markers that play a role in the pathogenesis of depression, thereby ameliorating depressive symptoms.35

Plasma Metabolites Difference Between PSD and Non-PSD Groups at 3 Months

Recent advances in neuropsychiatric research have highlighted the critical role of metabolites in mediating these disorders.36 In our study, we observed remarkable disparities in plasma metabolites between PSD and non-PSD groups at 3 months. Notably, cortisol and pyroglutamic acid levels were elevated, while tryptophan, serotonin, and caffeine levels were reduced in patients with PSD. These alterations primarily involved tryptophan, glutathione, and caffeine metabolism pathways. However, the changes of plasma metabolomics in PSD patients within 1 month were mainly concentrated in lipids, glycerophospholipids, alanine, aspartic acid and glutamate, which may indicate differences in the metabolism characteristics between early and late PSD.9,10

Tryptophan is an essential amino acid that must be obtained from the diet. Imbalances in neurotransmitters within the tryptophan pathway, particularly a decrease in serotonin, are proven to underlie the pathophysiology of PSD.37 The availability of tryptophan in the blood significantly determines serotonin synthesis in the brain.38 A machine learning model of plasma protein data indicated that serotonin pathway activity in PSD patients was reduced by alterations of kynureninase and quinoid dihydropteridine reductase.39 Additionally, the inflammatory response following a stroke may reduce serotonin bioavailability by increasing the conversion of tryptophan to kynurenine.40 Our findings of reduced plasma levels of tryptophan and serotonin in patients with PSD suggest that dysregulation of tryptophan metabolism is a significant contributing factor to PSD development.

Furthermore, our study observed elevated plasma level of cortisol in patients with PSD. Stress injury of hypothalamus and pituitary gland after stroke may lead to dysregulation of HPA axis.41 Glucocorticoids in the HPA axis can activate the inflammatory response,42 leading to depressive symptoms through neuroendocrine-immune interactions. An animal experiment indicated an elevation in serum cortisone levels in PSD mice,43 whereas antidepressant drugs significantly alleviate depression by inhibiting HPA axis activation in rats with stroke.44 Additionally, our study observed the enrichment of glutathione and caffeine metabolism, suggesting their potential effect on PSD development. Previous studies have demonstrated the dysregulation of these two metabolic pathways in depression.45,46 Dysregulation of glutathione metabolism has been demonstrated in major depressive disorder patients, with significantly lower activities of glutathione peroxidase and glutathione reductase.45 Glutathione has been identified as a potential marker of early depression,47 with impaired synaptic plasticity being associated with low levels of glutathione.48 Moreover, caffeine has neuroprotective and anti-inflammatory properties, which may reduce the risk of depression.46 Another animal experiment indicated that caffeine inhibited the production of lipopolysaccharide-induced nitric oxide and reduced the expression of pro-inflammatory genes including IL-3, IL-6, IL-12, inducible nitric oxide synthase and cyclooxygenase-2.49

Correlation Between Gut Microbiota and Plasma Metabolites

There is growing evidence that gut microbiota exert a substantial influence on host physiology and behavior by modulating metabolites in the blood.36 Depression-like behavior is associated with multiple metabolic pathways co-regulated by both the host and microbiota.50 In our study, correlation analyses between plasma metabolites and gut microbiota demonstrated that pro-inflammatory microbiota were negatively correlated with metabolites that counter depression, but the results were reversed for anti-inflammatory microbiota. Our results indicate that gut microbial dysregulation may contribute to PSD through the modulation of plasma metabolites, particularly those involved in tryptophan metabolism.

Although gut microbiota can directly affect host tryptophan metabolism through microbial metabolites, their indirect effect, including regulatory effect on the immune system, cannot be ignored. Gut microbiota metabolize tryptophan through three pathways: the indole pathway, kynurenine pathway, and serotonin pathway.51 The dysregulation of pro- and anti-inflammatory microbiota can trigger pro-inflammatory cascade reactions, effectively activating indoleamine 2,3 - dioxygenase (IDO). IDO prompts the conversion of tryptophan to the kynurenine pathway and generates cytotoxic quinolinic acid, which leads to the occurrence of depression.52,53 Activation of IDO also promotes serotonin catabolism.52 An animal experiment confirmed that pro-inflammatory microbiota indirectly stimulated intestinal kynurenine synthesis and increased the circulating kynurenine/tryptophan ratio by inducing an immune response.54 In contrast, probiotics with anti-inflammatory effects can increase peripheral levels of tryptophan55 and serotonin.56 Therefore, gut microbiota dysregulation may contribute to PSD development by influencing tryptophan metabolism.

Potential Combined Biomarkers for 3-Month PSD Diagnosis

There is an urgent need for objective diagnostic tools for PSD to improve detection and intervention.57 A plasma metabolomic study determined a panel consisting of three metabolites with an AUC value of 0.894 as potential biomarkers for PSD at 2 weeks.10 Using feces collected within one week post-stroke onset, another study identified a combination of seven microbiota to distinguish PSD from non-PSD with an AUC value of 0.705.6 However, these studies only investigated biomarkers based on a single dimension and concentrated on biological characteristics at the acute or subacute stage of stroke, in which the pathophysiologic conditions of patients were still unstable Consequently, potential biomarkers of PSD may not have been adequately and precisely identified. In our study, based on combined omics data collected at 3 months post-stroke onset, we screened potential biomarkers and established a classification model. Notably, a panel of combined biomarkers, including Parabacteroides, Staphylococcus, tryptophan, and theophylline showed excellent diagnostic ability for 3-month PSD, with an AUC value of 0.940.

Advantages and Limitations

Based on combined omics approaches, this study is the first to investigate the structures of both gut microbiota and plasma metabolites in PSD at 3 months after AIS onset. The three-month period represents a critical juncture in the evolution of PSD, with biomarkers potentially exhibiting distinctive characteristics compared to earlier stages.12 Our investigation substantiates the observation that the biomarkers of PSD at the three-month exhibit certain divergences from those observed in previous early-onset PSD cases. By demonstrating the association between the two omics, we provided more comprehensive evidence for PSD pathogenesis. Furthermore, we integrated gut microbiota and plasma metabolites to identify diagnostic biomarkers for 3-month PSD. However, this study has several limitations. Firstly, it is a cross-sectional study with a limited sample size. Validations in vitro and vivo are necessary for further exploring the mechanisms. Secondly, the 16S rRNA sequencing used in our study had a relatively low resolution and was only capable of accurately obtaining results at the genus level, with some important species-level information being lost, which made it challenging to distinguish closely related microbiota at the species level. Future studies could use metagenomics to identify the specific species and functional attributes of the gut microbiota in PSD. Finally, smoking affects the gut microbiome by altering immune homeostasis, biofilm formation, or direct exposure to microorganisms in tobacco, thereby influencing a variety of diseases.58 Furthermore, it has been demonstrated that changes in the gut microbiome of patients with depression are sex-specific, and that a panel of sex-specific biomarkers offers a superior diagnostic performance.59 Future studies should adjust for these factors when analyzing the gut microbiota of patients with PSD.

Conclusions

In this study, we characterized gut microbiota and plasma metabolites in patients with PSD 3 months after AIS onset, demonstrating the significant correlation between gut microbiota dysregulation and alterations in plasma metabolites. Furthermore, we established a panel of combined biomarkers derived from gut microbiota and plasma metabolites for 3-month PSD diagnosis, providing a novel theoretical framework for future detection and intervention.

Data Sharing Statement

The datasets used during the current study are available from the corresponding author on reasonable request.

Ethics Approval and Informed Consent

All procedures were approved by the Ethics Committee of Xuanwu Hospital, Capital Medical University (LYS [2020] 096). Written informed consent was obtained from all participants. Our study adheres to the principles of the Declaration of Helsinki.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Funding

This work was supported by the National Natural Science Foundation of China [81973759; 82274444].

Disclosure

The authors report no conflicts of interest in this work.

References

1. Cai W, Mueller C, Li YJ, Shen WD, Stewart R. Post stroke depression and risk of stroke recurrence and mortality: a systematic review and meta-analysis. Ageing Res Rev. 2019;50:102–109. doi:10.1016/j.arr.2019.01.013

2. Ezema CI, Akusoba PC, Nweke MC, Uchewoke CU, Agono J, Usoro G. Influence of post-stroke depression on functional independence in activities of daily living. Ethiop J Health Sci. 2019;29(1):841–846. doi:10.4314/ejhs.v29i1.5

3. Guo J, Wang J, Sun W, Liu X. The advances of post-stroke depression: 2021 update. J Neurol. 2022;269(3):1236–1249. doi:10.1007/s00415-021-10597-4

4. Robinson RG, Jorge RE. Post-stroke depression: a review. Am J Psychiatry. 2016;173(3):221–231. doi:10.1176/appi.ajp.2015.15030363

5. Ling Y, Gu Q, Zhang J, et al. Structural change of gut microbiota in patients with post-stroke comorbid cognitive impairment and depression and its correlation with clinical features. J Alzheimers Dis. 2020;77(4):1595–1608. doi:10.3233/jad-200315

6. Yao S, Xie H, Wang Y, et al. Predictive microbial feature analysis in patients with depression after acute ischemic stroke. Front Aging Neurosci. 2023;15:1116065. doi:10.3389/fnagi.2023.1116065

7. Kang Y, Yang Y, Wang J, Ma Y, Cheng H, Wan D. Correlation between intestinal flora and serum inflammatory factors in post-stroke depression in ischemic stroke. J Coll Physicians Surg Pak. 2021;31(10):1224–1227. doi:10.29271/jcpsp.2021.10.1224

8. Yu M, Jia HM, Qin LL, Zou ZM. Gut microbiota and gut tissue metabolites involved in development and prevention of depression. J Affect Disord. 2022;297:8–17. doi:10.1016/j.jad.2021.10.016

9. Wang M, Gui X, Wu L, et al. Amino acid metabolism, lipid metabolism, and oxidative stress are associated with post-stroke depression: a metabonomics study. BMC Neurol. 2020;20(1):250. doi:10.1186/s12883-020-01780-7

10. Wen L, Yan C, Zheng W, Li Y, Wang Y, Qu M. Metabolic alterations and related biological functions of post-stroke depression in ischemic stroke patients. Neuropsychiatr Dis Treat. 2023;19:1555–1564. doi:10.2147/ndt.S415141

11. Bhogal SK, Teasell R, Foley N, Speechley M. Lesion location and poststroke depression: systematic review of the methodological limitations in the literature. Stroke. 2004;35(3):794–802. doi:10.1161/01.Str.0000117237.98749.26

12. Huang J, Zhou FC, Guan B, et al. Predictors of remission of early-onset poststroke depression and the interaction between depression and cognition during follow-up. Front Psychiatry. 2018;9:738. doi:10.3389/fpsyt.2018.00738

13. Stein DJ, Phillips KA, Bolton D, Fulford KW, Sadler JZ, Kendler KS. What is a mental/psychiatric disorder? From DSM-IV to DSM-V. Psychol Med. 2010;40(11):1759–1765. doi:10.1017/s0033291709992261

14. Wang X, Fang C, Liu X, et al. High serum levels of iNOS and MIP-1α are associated with post-stroke depression. Neuropsychiatr Dis Treat. 2021;17:2481–2487. doi:10.2147/ndt.S320072

15. Lv D, Cao X, Zhong L, et al. Targeting phenylpyruvate restrains excessive NLRP3 inflammasome activation and pathological inflammation in diabetic wound healing. Cell Rep Med. 2023;4(8):101129. doi:10.1016/j.xcrm.2023.101129

16. Parks DH, Tyson GW, Hugenholtz P, Beiko RG. STAMP: statistical analysis of taxonomic and functional profiles. Bioinformatics. 2014;30(21):3123–3124. doi:10.1093/bioinformatics/btu494

17. Yuan TF, Wang ST, Le J, Li Y. Steroid profile analysis by liquid chromatography-tandem mass spectrometry in second-trimester pregnant women for trisomy 21 screening. J Pharm Biomed Anal. 2021;197:113966. doi:10.1016/j.jpba.2021.113966

18. Luo F, Fang C. Association between gut microbiota and post-stroke depression in Chinese population: a meta-analysis. Heliyon. 2022;8(12):e12605. doi:10.1016/j.heliyon.2022.e12605

19. Malan-Müller S, Valles-Colomer M, Palomo T, Leza JC. The gut-microbiota-brain axis in a Spanish population in the aftermath of the COVID-19 pandemic: microbiota composition linked to anxiety, trauma, and depression profiles. Gut Microbes. 2023;15(1):2162306. doi:10.1080/19490976.2022.2162306

20. Manzoor M, Leskelä J, Pietiäinen M, et al. Multikingdom oral microbiome interactions in early-onset cryptogenic ischemic stroke. ISME Commun. 2024;4(1):ycae088. doi:10.1093/ismeco/ycae088

21. Cheung SG, Goldenthal AR, Uhlemann AC, Mann JJ, Miller JM, Sublette ME. Systematic review of gut microbiota and major depression. Front Psychiatry. 2019;10:34. doi:10.3389/fpsyt.2019.00034

22. Xu Y, Wang Y, Li H, et al. Altered fecal microbiota composition in older adults with frailty. Front Cell Infect Microbiol. 2021;11:696186. doi:10.3389/fcimb.2021.696186

23. Cathomas F, Lin HY, Chan KL, et al. Circulating myeloid-derived MMP8 in stress susceptibility and depression. Nature. 2024;626(8001):1108–1115. doi:10.1038/s41586-023-07015-2

24. Chen H, Zhang J, He Y, et al. Exploring the role of staphylococcus aureus in inflammatory diseases. Toxins. 2022;14(7). doi:10.3390/toxins14070464

25. Cheng L, Wu H, Cai X, et al. A Gpr35-tuned gut microbe-brain metabolic axis regulates depressive-like behavior. Cell Host Microbe. 2024;32(2):227–243.e6. doi:10.1016/j.chom.2023.12.009

26. Shi S, Zhang Q, Sang Y, et al. Probiotic bifidobacterium longum BB68S improves cognitive functions in healthy older adults: a randomized, double-blind, placebo-controlled trial. Nutrients. 2022;15(1):51. doi:10.3390/nu15010051

27. Shi Y, Zhao E, Li L, et al. Alteration and clinical potential in gut microbiota in patients with cerebral small vessel disease. Front Cell Infect Microbiol. 2023;13:1231541. doi:10.3389/fcimb.2023.1231541

28. Shah NB, Suri RM, Melduni RM, et al. Shuttleworthia satelles endocarditis: evidence of non-dental human disease. J Infect. 2010;60(6):491–493. doi:10.1016/j.jinf.2010.02.008

29. Gao F, Guo R, Ma Q, et al. Stressful events induce long-term gut microbiota dysbiosis and associated post-traumatic stress symptoms in healthcare workers fighting against COVID-19. J Affect Disord. 2022;303:187–195. doi:10.1016/j.jad.2022.02.024

30. Eicher TP, Mohajeri MH. Overlapping mechanisms of action of brain-active bacteria and bacterial metabolites in the pathogenesis of common brain diseases. Nutrients. 2022;14(13):2661. doi:10.3390/nu14132661

31. Huang F, Liu X, Xu S, et al. Prevotella histicola mitigated estrogen deficiency-induced depression via gut microbiota-dependent modulation of inflammation in ovariectomized mice. Front Nutr. 2021;8:805465. doi:10.3389/fnut.2021.805465

32. Chung WSF, Meijerink M, Zeuner B, et al. Prebiotic potential of pectin and pectic oligosaccharides to promote anti-inflammatory commensal bacteria in the human colon. FEMS Microbiol Ecol. 2017;93(11). doi:10.1093/femsec/fix127

33. Mukherjee A, Lordan C, Ross RP, Cotter PD. Gut microbes from the phylogenetically diverse genus Eubacterium and their various contributions to gut health. Gut Microbes. 2020;12(1):1802866. doi:10.1080/19490976.2020.1802866

34. Dubois T, Zdanowicz N, Jacques D, Lepiece B, Jassogne C. Microbiota diversity and inflammation as a new target to improve mood: probiotic use in depressive disorder. Psychiatry Danub. 2023;35(Suppl 2):72–76.

35. Chang M, Chang KT, Chang F. Just a gut feeling: faecal microbiota transplant for treatment of depression - A mini-review. J Psychopharmacol. 2024;38(4):353–361. doi:10.1177/02698811241240308

36. Luan H, Wang X, Cai Z. Mass spectrometry-based metabolomics: targeting the crosstalk between gut microbiota and brain in neurodegenerative disorders. Mass Spectrom Rev. 2019;38(1):22–33. doi:10.1002/mas.21553

37. Zhang X, Wang CB, Duan LH, et al. Correlation research of serum substance P, CCK-8, and 5-HT values with depression levels in stroke survivors. Eur Rev Med Pharmacol Sci. 2023;27(4):1248–1254. doi:10.26355/eurrev_202302_31357

38. Lukić I, Ivković S, Mitić M, Adžić M. Tryptophan metabolites in depression: modulation by gut microbiota. Front Behav Neurosci. 2022;16:987697. doi:10.3389/fnbeh.2022.987697

39. Bidoki NH, Zera KA, Nassar H, et al. Machine learning models of plasma proteomic data predict mood in chronic stroke and tie it to aberrant peripheral immune responses. Brain Behav Immun. 2023;114:144–153. doi:10.1016/j.bbi.2023.08.002

40. Ormstad H, Verkerk R, Aass HC, Amthor KF, Sandvik L. Inflammation-induced catabolism of tryptophan and tyrosine in acute ischemic stroke. J Mol Neurosci. 2013;51(3):893–902. doi:10.1007/s12031-013-0097-2

41. Zhang XF, Zou W, Yang Y. Effects of IL-6 and cortisol fluctuations in post-stroke depression. J Huazhong Univ Sci Technolog Med Sci. 2016;36(5):732–735. doi:10.1007/s11596-016-1653-0

42. Busillo JM, Azzam KM, Cidlowski JA. Glucocorticoids sensitize the innate immune system through regulation of the NLRP3 inflammasome. J Biol Chem. 2011;286(44):38703–38713. doi:10.1074/jbc.M111.275370

43. Zhang G, Chen L, Yang L, et al. Combined use of spatial restraint stress and middle cerebral artery occlusion is a novel model of post-stroke depression in mice. Sci Rep. 2015;5(1):16751. doi:10.1038/srep16751

44. Cai L, Li WT, Zhang LL, Lu XQ, Chen M, Liu Y. Long noncoding RNA GAS5 enhanced by curcumin relieves poststroke depression by targeting miR-10b/BDNF in rats. J Biol Regul Homeost Agents. 2020;34(3):815–823. doi:10.23812/20-113-a-25

45. Lačković M, Stojković T, Pantović Stefanović M, et al. 8-Iso-prostaglandin F2α as a potential biomarker in patients with unipolar and bipolar depression. Eur Rev Med Pharmacol Sci. 2023;27(23):11496–11507. doi:10.26355/eurrev_202312_34588

46. Wasim S, Kukkar V, Awad VM, Sakhamuru S, Malik BH. Neuroprotective and Neurodegenerative Aspects of Coffee and Its Active Ingredients in View of Scientific Literature. Cureus. 2020;12(8):e9578. doi:10.7759/cureus.9578

47. Freed RD, Hollenhorst CN, Weiduschat N, et al. A pilot study of cortical glutathione in youth with depression. Psychiatry Res Neuroimaging. 2017;270:54–60. doi:10.1016/j.pscychresns.2017.10.001

48. Almaguer-Melian W, Cruz-Aguado R, Bergado JA. Synaptic plasticity is impaired in rats with a low glutathione content. Synapse. 2000;38(4):369–374. doi:10.1002/1098-2396(20001215)38:4<369::AID-SYN1>3.0.CO;2-Q

49. Hwang JH, Kim KJ, Ryu SJ, Lee BY. Caffeine prevents LPS-induced inflammatory responses in RAW264.7 cells and zebrafish. Chem Biol Interact. 2016;248:1–7. doi:10.1016/j.cbi.2016.01.020

50. Li Z, Lai J, Zhang P, et al. Multi-omics analyses of serum metabolome, gut microbiome and brain function reveal dysregulated microbiota-gut-brain axis in bipolar depression. mol Psychiatry. 2022;27(10):4123–4135. doi:10.1038/s41380-022-01569-9

51. Roager HM, Licht TR. Microbial tryptophan catabolites in health and disease. Nat Commun. 2018;9(1):3294. doi:10.1038/s41467-018-05470-4

52. Müller N, Myint AM, Schwarz MJ. Inflammatory biomarkers and depression. Neurotox Res. 2011;19(2):308–318. doi:10.1007/s12640-010-9210-2

53. Raison CL, Dantzer R, Kelley KW, et al. CSF concentrations of brain tryptophan and kynurenines during immune stimulation with IFN-alpha: relationship to CNS immune responses and depression. mol Psychiatry. 2010;15(4):393–403. doi:10.1038/mp.2009.116

54. El Aidy S, Derrien M, Aardema R, et al. Transient inflammatory-like state and microbial dysbiosis are pivotal in establishment of mucosal homeostasis during colonisation of germ-free mice. Benef Microbes. 2014;5(1):67–77. doi:10.3920/bm2013.0018

55. Desbonnet L, Garrett L, Clarke G, Bienenstock J, Dinan TG. The probiotic bifidobacteria infantis: an assessment of potential antidepressant properties in the rat. J Psychiatr Res. 2008;43(2):164–174. doi:10.1016/j.jpsychires.2008.03.009

56. Babaei F, Mirzababaei M, Mohammadi G, Dargahi L, Nassiri-Asl M. Saccharomyces boulardii attenuates lipopolysaccharide-induced anxiety-like behaviors in rats. Neurosci Lett. 2022;778:136600. doi:10.1016/j.neulet.2022.136600

57. Medeiros GC, Roy D, Kontos N, Beach SR. Post-stroke depression: a 2020 updated review. Gen Hosp Psychiatry. 2020;66:70–80. doi:10.1016/j.genhosppsych.2020.06.011

58. Huang C, Shi G. Smoking and microbiome in oral, airway, gut and some systemic diseases. J Transl Med. 2019;17(1):225. doi:10.1186/s12967-019-1971-7

59. Li Y, Zhang H, Zheng P, et al. Perturbed gut microbiota is gender-segregated in unipolar and bipolar depression. J Affect Disord. 2022;317:166–175. doi:10.1016/j.jad.2022.08.027