Structural Change of Gut Microbiota in Patients with Post-Stroke Comorbid Cognitive Impairment and Depression and Its Correlation with Clinical Features

Abstract

Background:

Post-stroke comorbid cognitive impairment and depression (PSCCID) is a severe neuropsychiatric complication after acute stroke. Gut microbiota dysbiosis is associated with many psychiatric disorders. Alterations in the composition of gut microbiota may serve as a critical role in patients with PSCCID.

Objective:

We aimed to characterize the microbial profiles of patients with PSCCID.

Method:

A total of 175 stroke patients were recruited in the study. The composition of gut bacterial communities of patients was determined by 16S ribosomal RNA Miseq sequencing, and Phylogenetic Investigation of Communities by Reconstruction of Unobserved States was used to demonstrate the functional alterations of gut microbiota. We further identified the characteristic gut microbiota of PSCCID using linear discriminant analysis effect size.

Results:

Patients with PSCCID exhibited an increased abundance of Proteobacteria, including Gammaproteobacteria, Enterobacteriales, and Enterobacteriaceae, and a decreased abundance of several short-chain fatty acids-producing bacteria compared with non-PSCCID patients. The abundance of Gammaproteobacteria and Enterobacteriaceae showed negative correlations with the MoCA score. Moreover, the Kyoto Encyclopedia of Genes and Genomes results demonstrated the enriched orthologs of glycan biosynthesis and metabolism and decreased orthologs of amino acid metabolism in PSCCID patients. Importantly, the characteristic gut microbiota was identified and achieved an area under the curve of 0.847 between the two groups.

Conclusion:

In this study, we characterized the gut microbiota of PSCCID patients, and revealed the correlations of the altered gut microbiota with clinical parameters, which took a further step towards non-invasive diagnostic biomarkers for PSCCID from fecal samples.

Keywords

Cognitive impairment correlation depression gut microbiota ischemic stroke

INTRODUCTION

Stroke, the leading cause of death and disability worldwide [1], causes a wide range of neuropsychiatric disorders, such as depression, anxiety, apathy, fatigue, personality changes, mania, and cognitive impairment. Approximately one-third of stroke patients develop neuropsychiatric disorders shortly after stroke onset [2, 3]. Post-stroke cognitive impairment and post-stroke depression are known as common complications of a stroke; the prevalence of depression and cognitive impairment within the first three months after stroke were 25% to 31% [4], and 10% to 47.3% [5, 6], and they often co-exist in stroke patients with adverse effects on patient outcome [7, 8]. Cognitive dysfunction was closely related to depression, and they often interact with each other [9, 10]. A previous study demonstrated normalized cognitive function in stroke patients following antidepressant treatment and vice versa [11], which suggested they may have a similar etiology. Moreover, many mental disorders commonly co-occur in patients. Dementia is frequently comorbid with depression, and approximately 50% of dementia patients present depressive symptoms [12]. Besides, post-stroke cognitive impairment and post-stroke depression may be mediated by common pathogenesis, such as neuroinflammation. Proinflammatory cytokines play an important role in the pathophysiology of both depression and dementia [13, 14]. Neurotransmitter disorders caused by vascular injury after stroke was possible the mechanisms of post-stroke cognitive impairment and post-stroke depression [15, 16]. The occurrence of PSCCID is not uncommon in clinical research, due to the combined action of neurobiological abnormality and psychological stress, the pathological mechanism of PSCCID is complex. The combination of the limited use of scales in clinical practice and undetectable symptoms in the early stages, leading to some PSCCID patients have not been diagnosed and treated correctly. Thus, early and effective evaluation and treatment of PSCCID is essential for post-stroke rehabilitation and prolonged survival time.

Gut microbiota dysbiosis is associated with many neuropsychiatric conditions such as depression, Alzheimer’s disease (AD), and Parkinson’s disease (PD). Jiang et al. revealed an altered gut microbial profile in the active-major depressive disorder (MDD) patients, with a significant increase in Bacteroidetes, Proteobacteria, Actinobacteria, and Enterobacteriaceae, and a reduction in Firmicutes and Faecalibacterium [17]. The decrease of Faecalibacterium abundance resulted in a chronic low-grade inflammatory response in the gut [18] and was negatively correlated with the severity of depressive symptoms [17]. Recent studies have suggested that AD patients have a changed gut microbiota composition compared with healthy controls [19, 20]. Enterobacteriaceae and their ingredient lipopolysaccharide (LPS) could release proinflammatory cytokines and amyloid-β accumulation [21, 22], which contributed to the neurodegeneration in the pathogenesis of AD. Cox et al. reported that long-term calorie restriction might change the intestinal environment and prevent the proliferation of microorganisms that cause age-related cognitive decline; the increased abundance of Bacteroides may be one of the possible mechanisms by which gut microbiota impacts AD pathogenesis [23]. The previous study also reported that the reduced abundance of Prevotellaceae in feces of PD patients, and the level of Enterobacteriaceae was positively associated with the severity of postural instability and gait difficulty [24], suggesting that the gut microbiota might be related to PD phenotype. Recently, increasing evidence has reported that gut microbiota changes could influence brain function via the gut-brain axis [25]. These findings demonstrated that gut microbiota might be an important regulator of the bidirectional communication between the gut and the brain.

Increasing evidence demonstrated that gut microbiota could be regarded as a non-invasive diagnosis biomarker for diseases, including schizophrenia [26] and type 2 diabetes [27]. Of note, stroke patients exhibited significant dysbiosis of the gut microbiota, which was characterized by a higher abundance of opportunistic pathogens and a lower level of beneficial bacteria [28]. Moreover, the increased abundance of Fusobacterium and decreased short-chain fatty acids (SCFAs) were found in post-stroke cognitive impairment patients [29]. However, the gut microbial compositions in PSCCID patients have not been evaluated. Thus, it is of great importance to find the characteristics of gut microbiota composition in patients with PSCCID for post-stroke rehabilitation.

The purpose of this study was to characterize the gut microbiota profiles of PSCCID patients and to investigate the associations of the gut microbiota with clinical parameters of PSCCID.

METHODS

Participants

We conducted a study of 175 stroke patients from the Second Affiliated Hospital of Wenzhou Medical University in Wenzhou, Zhejiang, from March 2019 to June 2019. Neuropsychological assessments were assessed three months after stroke onset. The Hamilton Depression Rating Scale (HAMD) and the Montreal Cognitive Assessment (MoCA) were applied to assess the severity of each patient’s depressive symptoms and cognitive deficits. The National Institute of Health Stroke Scale (NIHSS) was used to evaluate the degree of neurological impairment, and the sleep quality was reflected via the Pittsburgh Sleep Quality Index (PSQI). The patients were divided into two groups (PSCCID and non-PSCCID groups) according to their HAMD and MoCA scores.

The inclusion criteria for stroke patients were: ischemic stroke patients, NIHSS 0–5 points, aged 40 to 80 years, without previous neuropsychiatric disorders, no treatment with anti-neuropsychiatric conditions medication, and no communication deficit (e.g., aphasia and deafness). The exclusion criteria for all patients included: infarcts in areas associated with cognitive function and depression such as the hippocampus, thalamus, frontal lobe, temporal lobe, cingulate gyrus, amygdala, corpus callosum, and caudate nucleus [30 –33]; use of antibiotics, probiotics, or prebiotics within three months; alcohol addiction; patients with delirium; patients with digestive diseases; recent infection; and pregnancy. In this study, patients with preexisting cognitive impairment were excluded from the study according to the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) (cut-off value >3.4) [34], which was completed by a relative of patients who could determine if there has been a decline in the patient’s cognitive function [35]. The PSCCID patients included those with HAMD score ≥8 and MoCA score ≤26 with junior school education level or above, ≤21 with primary school education level, or ≤15 with illiterate. The patients in the non-PSCCID group were those with HAMD score ≤8 and MoCA score ≥26 with junior school education level or above, ≥21 with primary school education level, or ≥15 with illiterate.

The Ethics Committee of the Second Affiliated Hospital of Wenzhou Medical University has approved the protocols. All participants provided informed consent before entering the study.

Clinical data collection

We collected the demographic information of each patient from an interview, including sex, age, years of education, diet types, taste, place of living, number of drugs, motor dysfunction, smoking, and measured the height and weight of each patient to calculate body mass index (BMI). We used the Mini-Nutritional Assessment (MNA) to evaluate the patients’ nutritional status, and information about the stool frequency, and history of renal or hepatic diseases and cerebrovascular disease was collected. Additionally, all patients were examined by brain magnetic resonance imaging scans (GE Discovery750, Milwaukee, USA) to determine the lesion site.

Information on diabetes mellitus, hypertension, and hyperlipidemia was obtained by inquiring about the history of previous diseases and measuring blood glucose, blood pressure, and blood lipid. Hypertension was defined as blood pressure ≥140/90 mmHg, diabetes mellitus was defined as fasting blood glucose ≥7.0 mmol/L and/or oral glucose tolerance test 2 h blood glucose ≥11.1 mmol/L, and hyperlipidemia was defined as the total cholesterol ≥5.72 mmol/L and/ or triglyceride ≥1.70 mmol/L. All patients underwent the neuropsychiatric examinations by a neurologist and a psychologist.

Sample collection and DNA extraction

Fecal samples from all patients were stored at –80°C within 30 min of preparation. We extracted DNA using a DNA extraction kit (TIANGEN, TIANamp, China) according to the manufacturer’s methods, as described in previous studies [36]. The DNA concentration and purity were determined using the nanodrop spectrophotometer (ThermoFisher, USA), and evaluating the quantity by 2% agarose gel. Then, storing the DNA at –20°C before analysis.

Polymerase chain reaction and Miseq sequencing

We amplified the extracted DNA with the forward primer (5^′- CCTACGGGNGGCWGCAG-3^′) and the reverse primer (5^′- GACTACHVGGGTATCTAATCC-3^′) specific for the V3–V4 regions of the 16S rRNA gene, as previously described [37]. We were performing the high-throughput sequencing on a MiSeq Benchtop Sequencer (Illumina, Singapore, USA) according to the manufacturer’s guidelines.

Bioinformatics and statistical analysis

In order to ensure the accuracy of the sequences for subsequent analysis, we obtained the raw pyrosequencing reads from the sequencer and removed primer-free sequences using cutadapt (version 1.11). The Paired-end reads of the primer were spliced with pandaseq (version 2.9), and the overlap was at least greater than 10 bp. We further removed the low-quality sequences (the average mass value was lower than Q20 and contained base N), the sequences whose length was over 300–480 bp, and the chimeric sequences. High-quality sequences were grouped into operational taxonomic units (OTUs) at a sequence similarity of 0.97, and the sequence with the highest abundance in each class was selected as the representative sequence. QIIME was adopted to remove the OTUs with only one sequence in all samples. Then, we carried out the following data analysis in R software. According to the previous study, Shannon’s and Simpson index were applied to analyze the alpha diversity of gut microbiota, and we calculated the beta diversity via principal coordinates analysis (PCoA). We evaluated the alpha diversity and beta diversity between the two groups via the Mann-Whitney U test and analysis of similarities (ANOSIM) on the Bray-Curtis dissimilarity index. The linear discriminant analysis (LDA) effect size (LEfSe) used the Kruskal-Wallis test to discover the significant p-values associated with microbial clades and functions, the Kruskal-Wallis test (alpha value of 0.05) and LDA score of >2 were used as thresholds. We used receiver operating characteristic (ROC) curves and the area under the ROC curve (AUC) to confirm the specificity and sensitivity of the characteristic gut microbiota to the diagnosis of PSCCID. Phylogenetic investigation of communities by reconstruction of unobserved states (PICRUSt) generated the Kyoto Encyclopedia of Genes and Genomes (KEGG) ontology and analyzed via Statistical Analysis of Metagenomic Profiles (STAMP).

In this study, categorical variables were presented as numbers with percentages, and continuous variables were shown as means with standard deviation (SD), or medians with interquartile range (IQR) according to the Kolmogorov-Smirnov normality test. Chi-squared test, Student’s t-test or Mann-Whitney test were performed for comparison of categorical variables and continuous variables, respectively. We evaluated correlations among the relative abundances of the characteristic gut microbiota, MoCA score, and HAMD score. Correlations between variables were calculated using Spearman’s rank correlation analysis, and we had removed the rare taxa and the bacteria with low relative abundance. p values < 0.05 was considered significant. Statistical analysis was carried out via SPSS software package V.22 (SPSS, Chicago, IL, USA), GraphPad Prism V.5.0.1 (La Jolla, CA, USA), the R software (V.3.5), Adobe Illustrator CC 2015 (Adobe Systems Incorporated, CA, USA).

RESULTS

Clinical characteristics of the patients

In this study, we made a flow diagram that illustrates the experimental design of the study (Supplementary Figure 1), and the fecal samples of 41 PSCCID patients and 25 non-PSCCID patients were sequenced and analyzed. As shown in Table 1, as expected, there were significant differences between the two groups in education years (PSCCID versus non-PSCCID: 6 (3–7.5) versus 9 (6–9), p = 0.005), HAMD (PSCCID versus non-PSCCID: 13.6±3.50 versus 5.56±2.58, p < 0.001), MOCA (PSCCID versus non-PSCCID: 13.17±6.23 versus 24.44±3.24, p < 0.001), and PSQI (PSCCID versus non-PSCCID: 7.8±4.03 versus 1.96±3.20, p < 0.001). There were no significant differences in terms of sex, age, NIHSS, hypertension, diabetes mellitus, hyperlipidemia, left side lesion location, and history of cerebrovascular disease. Patients were all living with families. Smoking (PSCCID versus non-PSCCID: 51.2% versus 64.0%, p = 0.310) and impaired motor function (PSCCID versus non-PSCCID: 78.0% versus 80.0%, p = 0.851) between the PSCCID group and the non-PSCCID group were not significantly different. There were no significant difference in BMI (PSCCID versus non-PSCCID: 25.14±3.62 versus 26.62±3.58, p = 0.111), MNA (PSCCID versus non-PSCCID: Normal nutritional status, 32 (78) versus 21 (84); At risk of malnutrition, 6 (14.6) versus 3 (12); Malnourished, 3 (7.3) versus 1 (4), p = 0.541), the stool frequency (PSCCID versus non-PSCCID (median (IQR)): 1 (1–1) versus 1 (1–1.25), p = 0.360), the number of drugs (PSCCID versus non-PSCCID (median (IQR)): 4 (3–6) versus 4 (3–5), p = 0.273), and history of renal or hepatic diseases (PSCCID versus non-PSCCID: 39% versus 28%, p = 0.362) between the two groups. Taste (PSCCID versus non-PSCCID: light, 8 (19.5) versus 3 (12); salty, 24 (58.5) versus 18 (72); sweet, 9 (22) versus 4 (16), p = 0.932) and diet types (PSCCID versus non-PSCCID: meat, 13 (31.7) versus 14 (56); vegetarian, 17 (41.5) versus 7 (28); mixed, 11 (26.8) versus 4 (16), p = 0.068) were not significantly different between the two groups. We summarize the demographic and clinical parameters of the patients in Table 1.

Table 1

Characteristics of the recruited patients

Parameter	PSCCID group (n = 41)	Non-PSCCID group (n = 25)	P
Female, n (%)	24 (58.5)	11 (44.0)	0.251
Age (y), mean (SD)	69.63±9.35	68.92±8.46	0.756
BMI (kg/m²), mean (SD)	25.14±3.62	26.62±3.58	0.111
Education (y), median, (IQR)	6 (3–7.5)	9 (6–9)	0.005
NIHSS, median, (IQR)	2 (1–5)	2 (1–3)	0.212
HAMD, mean (SD)	13.6±3.50	5.56±2.58	<0.001
MOCA, mean (SD)	13.17±6.23	24.44±3.24	<0.001
V/E function, median, (IQR)	1 (1–3)	3 (2.5–3.5)	<0.001
Naming, median, (IQR)	0 (0–2)	2 (2–3)	<0.001
Attention, median, (IQR)	3 (1.5–5)	6 (5.5–6)	<0.001
Language, median, (IQR)	2 (1.5–2)	3 (2–3)	<0.001
Abstraction, median, (IQR)	0 (0–0)	1 (0–1)	<0.001
Delayed recall, median, (IQR)	2 (0–3)	4 (4–4)	<0.001
Orientation, median, (IQR)	3 (2–4)	6 (5–6)	<0.001
PSQI, mean (SD)	7.8±4.03	1.96±3.20	<0.001
Hypertension, n (%)	24 (58.5)	16 (64.0)	0.659
Diabetes mellitus, n (%)	12 (29.3)	6 (24.0)	0.641
Hyperlipidemia, n (%)	15 (36.6)	10 (40.0)	0.781
Left side lesion location, n (%)	24 (58.5)	13 (52.0)	0.604
MNA			0.541
Normal nutritional status, n (%)	32 (78)	21 (84)
At risk of malnutrition	6 (14.6)	3 (12)
Malnourished	3 (7.3)	1 (4)
Motor dysfunction, n (%)	32 (78.0)	20 (80.0)	0.851
Smoking, n (%)	21 (51.2)	16 (64.0)	0.310
Stool frequency (per day) median, (IQR)	1 (1–1)	1 (1–1.25)	0.360
Number of drugs median, (IQR)	4 (3–6)	4 (3–5)	0.273
History of renal or hepatic diseases, n (%)	16 (39.0)	7 (28.0)	0.362
History of cerebrovascular disease, n (%)	17 (41.5)	5 (20)	0.073
Taste			0.932
light	8 (19.5)	3 (12)
salty	24 (58.5)	18 (72)
sweet	9 (22)	4 (16)
Diet types			0.068
meat	13 (31.7)	14 (56)
vegetarian	17 (41.5)	7 (28)
mixed	11 (26.8)	4 (16)

PSCCID, post-stroke comorbid cognitive impairment and depression; BMI, body mass index; NIHSS, National Institute of Health Stroke Scale; HAMDS, Hamilton Depression Scale; MOCA, Montreal Cognitive Assessment; V/E, Visuospatial/Executive; PSQI, Pittsburgh Sleep Quality Index; MNA, Mini-Nutritional Assessment; SD, standard deviation; IQR, interquartile range.

Alterations of gut microbiota in patients with PSCCID

The alpha diversity of gut microbiota, as estimated by Shannon and Simpson indexes, showed no significant difference between the groups (p = 0.674, 0.423, respectively, Fig. 1A, B). Although the beta diversity of the PSCCID group was similar to the non-PSCCID group according to PCoA scatterplot (p = 0.384, Fig. 1C), the relative abundances of some gut microbial taxa and their relative abundance were significantly different between the two groups. At the phylum level, the bacterial population was primarily composed of Bacteroidetes, Firmicutes, Proteobacteria, and Actinobacteria (Fig. 2A). PSCCID group had significantly higher contents of Proteobacteria compared with the non-PSCCID group. At the family level, the gut microbial population is dominated by Bacteroidaceae, Ruminococcaceae, Lachnospi-raceae, Prevotellaceae, Enterobacteriaceae, Veillo-nellaceae, Porphyromonadaceae, Streptococcaceae, Rikenellaceae, and Bifidobacteriaceae (Fig. 2B). The most predominant ten genera that made up most of the total bacteria abundance were Bacteroides, Prevotella, Faecalibacterium, Escherichia/Shigella, Clostridium XlVa, Ruminococcaceae_Other, Streptococcus, Roseburia, Parabacteroides, and Alistipes (Fig. 2C). We further compared the gut microbiota compositions of the two groups using LEfSe analysis. A cladogram indicating the phylogenetic distribution of the gut microbiota and the predominant bacteria in the two groups, and LDA revealed the largest bacterial differences in the taxa between the two groups. The LEfSe analysis revealed 12 discriminative features (LDA > 2, p < 0.05) at the phylum (n = 1), class (n = 1), order (n = 2), family (n = 2), and genus (n = 6) levels. The abundances of Proteobacteria, Gammaproteobacteria, Enterobacteriales, Enterobacteriaceae, Carnobacteriaceae, Veillonella, Granulicatella, and Enterobacteriaceae_other were higher in PSCCID patients, while the abundances of Aeromonadales, Fusicatenibacter, Mitsuokella, and Lachnospiraceae_other were more abundant in non-PSCCID patients compared with PSCCID patients (Fig. 2D, E).

Fig. 1

The Diversity of gut microbiota in the two groups. The α-diversity of the gut microbiota between PSCCID and non-PSCCID groups was shown according to the Shannon index (A) and Simpson index (B). The β-diversity of the gut microbiota between the two groups was demonstrated via PCoA based on the Bray-Curtis dissimilarity index (C). PSCCID, post-stroke comorbid cognitive impairment and depression; PCoA, Principle coordinates analysis.

Fig. 2

Taxonomic differences of gut microbiota between PSCCID and non-PSCCID groups. Taxonomic summary of the gut microbiota of the PSCCID patients and non-PSCCID patients were shown at (A) phylum level, (B) family level, and (C) genus level. LEfSe analyses revealed the most differentially abundant taxons between the two groups. D) A cladogram was indicating the phylogenetic distribution of gut microbiota between the two groups. E) LDA revealed significant bacterial differences in gut microbiota between PSCCID (red) and non-PSCCID (green) groups. Only LDA scores (log10) >2 and p < 0.05 are shown. PSCCID, post-stroke comorbid cognitive impairment and depression; LEfSe, linear discriminant analysis effect size; LDA, linear discriminant analysis.

Associations of gut microbiota with clinical parameters

We evaluated correlations among the relative abundances of the characteristic gut microbiota, MoCA score, and HAMD score. We used the LDA value (LDA > 2, p < 0.05) to select the characteristic gut microbiota. As shown in Fig. 3, we identify that Gammaproteobacteria and Enterobacteriaceae negatively correlated with the MoCA score (p < 0.05) (Fig. 3A, B), while Lachnospiraceae_other positively correlated with the MoCA score (p < 0.05) (Fig. 3C). However, there were no relationships between the relative abundance of the characteristic gut microbiota and the HAMD score. These results appeared to reflect a close link between the characteristic gut microbiota and the clinical parameters of PSCCID.

Fig. 3

Correlations of the MoCA score with the relative abundance of Gammaproteobacteria (A), Enterobacteriaceae (B), and Lachnospiraceae_other (C). The Spearman’s rank correlation (R) and probability (P) were used to assess the correlations. MOCA, Montreal Cognitive Assessment.

Predicted function analysis of gut microbiome

We also found that several KEGG pathways were significantly changed between the two groups. There are 276 KEGG pathways in two groups, including metabolism, genetic information processing, and unclassified. The enriched orthologs in PSCCID patients were glycan biosynthesis and metabolism (LPS biosynthesis proteins, and LPS biosynthesis), enzyme families (peptidases), cellular processes and signaling (membrane and intracellular structural molecules), folding, sorting and degradation (chaperones and folding catalysts), energy metabolism (nitrogen metabolism; oxidative phosphorylation), and genetic information processing (protein folding and associated processing). In contrast, the prevalent markers in non-PSCCID patients were amino acid metabolisms (phenylalanine, tyrosine and tryptophan biosynthesis; histidine metabolism; valine, leucine and isoleucine biosynthesis), carbohydrate metabolism (starch and sucrose metabolism), and metabolism of cofactors and vitamins (thiamine me-tabolism) (Fig. 4).

Fig. 4

LDA revealed the differences in gene function between PSCCID patients (red) and non-PSCCID patients (green). The bar plots of KEGG modules are significantly different between the two groups. Only LDA scores (log10) >2 and p < 0.05 are shown. LDA, linear discriminant analysis; PSCCID, post-stroke comorbid cognitive impairment and depression; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Predictor performance of gut microbiota

The ROC curves were used to identify the non-invasive diagnosis biomarker for PSCCID. As shown in Fig. 5, we used the LDA value to select the six genera as biomarkers, including Veillonella, Granulicatella, Enterobacteriaceae_other, Fusicatenibacter, Mitsuokella, and Lachnospiraceae_other, achieving area under the curve (AUC) values of 0.847 (p < 0.001, 95% CI 0.749–0.945). Moreover, the predictive model based on Enterobacteriaceae could also distinguish PSCCID patients from non-PSCCID patients (p = 0.007, AUC = 0.698, 95% CI 0.570–0.825).

Fig. 5

ROC curves were demonstrated based on the relative abundance of the six characteristic genera and Enterobacteriaceae in discriminating PSCCID from non-PSCCID. ROC, receiver operating characteristic; AUC, area under the ROC curve; CI, confidence interval; PSCCID, post-stroke comorbid cognitive impairment and depression.

DISCUSSION

This study firstly characterized the gut microbiota in patients with PSCCID. Although the PSCCID group’s microbial diversity was similar to that of the non-PSCCID group, some gut microbial taxa were distinct between the two groups. The gut microbiota of PSCCID patients was characterized by an increased abundance of Proteobacteria, including Gammaproteobacteria, Enterobacteriales, and Enterobacteriaceae, and a decreased abundance of several SCFAs-producing bacteria. The abundance of Gammaproteobacteria and Enterobacteriaceae showed negative correlations with the MoCA score. Furthermore, the ROC models based on the six characteristic genera and Enterobacteriaceae could distinguish PSCCID patients from non-PSCCID patients (achieved AUC values of 0.847 and 0.698, respectively). These results indicated that gut microbiota dysbiosis might play a role in the development of PSCCID.

In this study, we found that gut microbial diversity between the two groups was not dramatically altered. According to previous studies, decreased microbial diversity was observed in patients with AD [38], PD [39], and diabetes [40], while some studies also reported the gut microbial diversity was increased in MDD [17] and autism [41]. It is noted that the diversity of gut microbiota is influenced by factors such as diet, age, and antibiotic use [42]. Although greater bacterial diversity may benefit human health, its role in disease remains unclear. In this study, all patients suffered from a stroke and were accompanied by similar stroke-related risk factors, which could contribute to bacterial diversity’s consistency. It is widely known that nutritional status significantly affects the patient’s gut microbiota composition [43]. The nutritional status was similar between the two groups. Therefore, the interference of nutritional status can be avoided. In this study, PSCCID patients possessed an enriched abundance of phylum Proteobacteria compared with non-PSCCID patients, including Gammaproteobacteria, Enterobacteriales, and Enterobacteriaceae. Moreover, the abundance of Gammaproteobacteria and Enterobacteriaceae exhibited negative correlations with MoCA score, which suggested that these gut bacteria might be related to poor cognitive function. It is well known that Proteobacteria and Enterobacteriaceae are considered as opportunistic pathogens in the gut, and the increased abundance of Proteobacteria, Gammaproteobacteria, and Enterobacteriaceae has been related to many human diseases [44, 45]. Moreover, the overgrowth of Gammaproteobacteria, Enterobacteriales, and Enterobacteriaceae of Proteobacteria could trigger the secretion of proinflammatory cytokines induced by LPS [46, 47], and contributed to the disruption of the intestinal barrier and function [48]. Consistent with our results, the abundances of Proteobacteria and Enterobacteriaceae significantly increased in patients with AD [19] and MDD [17], and the proportion of Gammaproteobacteria was negatively associated with cognitive function [19]. Granulicatella, which belongs to family Carnobacteriaceae, has been identified as an etiologic agent in infectious diseases such as pneumonia and infective endocarditis [49], but the amount of the bacteria was meager in our samples, so its impact on PSCCID remains uncertain. Moreover, several SCFAs-producing bacteria were significantly lower in PSCCID patients compared with non-PSCCID patients, including Fusicatenibacter and Lachnospiraceae_other. Besides, Lachnospiraceae_other was closely associated with a higher MoCA score. The SCFAs exert anti-inflammatory effects and ameliorate the impairment of the intestinal epithelial barrier, preventing the brain neuroinflammation [50, 51]. Fusicatenibacter saccharivorans may reduce intestinal inflammation by inducing the anti-inflammatory cytokine IL-10 production, which was lower in active UC patients than in the control group [52]. However, the abundance of Aeromonadales and Mitsuokella was very low in this study, which could be easily affected by external factors.

There was no statistical difference in age between the two groups. Gut microbiota composition and stability change with age, and the alterations are considered to affect health [53]. For example, the subgroups of Proteobacteria, which contain many opportunistic bacteria, are increased in older adults [54]. Moreover, these alterations in gut microbiota were correlated with inflammatory cytokines interleukins six and eight [55]. Additionally, gut microbiota dysbiosis may also lead to age-related cognitive decline. The brain amyloid and circulating inflammatory cytokines were positively correlated with the inflammatory bacteria Escherichia/Shigella and negatively correlated with the anti-inflammatory E. rectale taxon [56]. Aging is associated with significant alterations in the gut microbiota compositions, including decreased microbial diversity and a decrease in bacteria with anti-inflammatory properties. The ability of body tissue to repair is gradually impaired with aging, leading to the release of more inflammatory cytokines, which are often called inflammaging [57].

Ischemic stroke is closely related to metabolic diseases, including obesity, type 2 diabetes, and dyslipidemia, which is linked to low-grade systemic inflammation [58]. The gut microbiota is associated with many metabolic functions, such as modulation of glucose and lipid homeostasis [59]. Besides, the gut microbiota, especially lipopolysaccharide-containing microbiota, including Proteobacteria and Enterobacteriaceae, contribute to the low-grade inflammation [60, 61]. According to the previous study, neuroinflammation was associated with cognitive impairment [62]. Thus, these inflammation-related gut bacteria might be involved in the development of PSCCID.

Gammaproteobacteria, Enterobacteriaceae, and Lachnospiraceae_other showed a close association with the clinical parameters, which might suggest a significant role of the gut microbiota in PSCCID. According to the LEfSe results, we used the six characteristic genera as biomarkers for PSCCID diagnosis, achieving AUC values of 0.847. Enterobacteriaceae are the common causative pathogen in the gut. LPS from Enterobacteriaceae has been shown to aggravate intestinal damage and permeability, and the blooms of Enterobacteriaceae are involved in many inflammatory diseases, such as inflammatory bowel disease and obesity [63]. The predictive model based on Enterobacteriaceae could also distinguish PSCCID patients from non-PSCCID patients. Since the alterations of gut microbiota may happen before the onset of neuropsychiatric disorders [64, 65], the characteristic gut microbiota might serve as potential biomarkers for the diagnosis of PSCCID at an early stage.

Besides, our results demonstrated that PSCCID was associated with several modulations of functional KEGG pathways, including glycan biosynthesis and metabolism, amino acid metabolism, enzyme families, cellular processes and signaling, energy metabolism, genetic information processing, carbohydrate metabolism, and metabolism of cofactors and vitamins. The enriched orthologues for LPS biosynthesis proteins in glycan biosynthesis and metabolism could contribute to cognitive dysfunction, which mediated by neuroinflammation via inducing LPS biosynthesis [66]. Moreover, the abundance of the module for glycan biosynthesis and metabolism was also higher in AD patients [19]. In contrast, the amino acid metabolism was reduced in PSCCID patients, including phenylalanine, tyrosine, and tryptophan biosynthesis; histidine metabolism; and valine, leucine, and isoleucine biosynthesis. Amino acids are necessary to maintain intestinal integrity and barrier function, which may contribute to reducing inflammation and proinflammatory cytokines [67]. These results suggested that gut microbiota might play an essential role in the communication with the brain and provide new options for the treatment of PSCCID via modulating the mediators affected by gut microbial metabolism.

Gut microbiota compositions of stroke and transient ischemic attack patients were characterized by an enriched abundance of opportunistic pathogens, including Enterobacter, Megasphaera, and Oscillibacter, and fewer beneficial or commensal genera including Bacteroides, Prevotella, and Faecalibacterium. Besides, the gut microbiota dysbiosis might be regarded as a newly identified high-risk factor for stroke [45]. In this study, we explored the intrinsic link between the PSCCID and gut microbiota. Therefore, only stroke patients were recruited in our study. However, the study has some limitations that should be mentioned. First, gut microbiota composition could be affected by multiple factors [42]. In this study, we did not strictly control some types of confounders. Although no statistically significant difference was detected between the groups for the tastes and diet types, we did not quantify it accurately. Besides, there was no statistical difference in the number of drugs between the two groups, but we did not determine the type of drug. However, some drugs affecting gut microbiota, such as antibiotics, were not used in this experiment. In general, we cannot exclude the influence of these factors on gut microbiota. Second, we did not investigate whether the changes in the gut microbiota are causally related to PSCCID. We did not observe the changes in the gut microbiota of PSCCID patients dynamically and the correlation of gut microbiota with its metabolites, and we also failed to explore a deeper relationship of gut microbiota with PSCCID using shotgun sequencing, transcriptomics, and other omics analysis, which would limit the representativeness of this study. In this study, we only analyzed the chronological age of patients and overlooked the impact of lifestyle and health status, and this may have had an impact on our results. Our study is a cross-sectional design, and causal inferences cannot be drawn. In future studies, we should also further investigate the composition of gut microbiota in different conditions, including cognitive dysfunction, depression, comorbid cognitive dysfunction, and depression, and controls. Using gut microbiota as a standard diagnostic tool has not been widely used in clinical practice yet, and we only found an association between gut microbiota and PSCCID, considering the effect of our small population, the characteristic gut microbiota may not accurately predict PSCCID independently. Therefore, further large sample size and multiple-center studies are needed to confirm these results.

In summary, this study presented that PSCCID patients have an altered gut microbiota composition, which was closely related to the clinical parameters of PSCCID. Besides, the characteristic gut microbiota might facilitate the diagnosis of PSCCID, and the association of gut microbiota with PSCCID might open new avenues for targeted prevention and treatment.

Footnotes

ACKNOWLEDGMENTS

This work was supported by the National Natural Science Foundation of China (81871094).

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Bonita

, Mendis

, Truelsen

, Bogousslavsky

, Toole

, Yatsu

(2004) The global stroke initiative. Lancet Neurol 3, 391–393.

Hackett

, Kohler

, O’Brien

, Mead

(2014) Neuropsychiatric outcomes of stroke. Lancet Neurol 13, 525–534.

Hackett

, Pickles

(2014) Part I: Frequency of depression after stroke: An updated systematic review and meta-analysis of observational studies. Int J Stroke 9, 1017–1025.

Astrom

, Adolfsson

, Asplund

(1993) Major depression in stroke patients. A 3-year longitudinal study. Stroke 24, 976–982.

Jacquin

, Binquet

, Rouaud

, Graule-Petot

, Daubail

, Osseby

, Bonithon-Kopp

, Giroud

, Bejot

(2014) Post-stroke cognitive impairment: High prevalence and determining factors in a cohort of mild stroke. J Alzheimers Dis 40, 1029–1038.

Pendlebury

, Rothwell

(2009) Prevalence, incidence, and factors associated with pre-stroke and post-stroke dementia: A systematic review and meta-analysis. Lancet Neurol 8, 1006–1018.

Hommel

, Carey

, Jaillard

(2015) Depression: Cognition relations after stroke. Int J Stroke 10, 893–896.

Douven

, Aalten

, Staals

, Schievink

SHJ

, van Oostenbrugge

, Verhey

FRJ

, Kohler

(2018) Co-occurrence of depressive symptoms and executive dysfunction after stroke: Associations with brain pathology and prognosis. J Neurol Neurosurg Psychiatry 89, 859–865.

Kauhanen

, Korpelainen

, Hiltunen

, Brusin

, Mononen

, Maatta

, Nieminen

, Sotaniemi

, Myllyla

(1999) Poststroke depression correlates with cognitive impairment and neurological deficits. Stroke 30, 1875–1880.

10.

Terroni

, Sobreiro

MFM

, Conforto

, Adda

, Guajardo

, de Lucia

MCS

, Fraguas

(2012) Association among depression, cognitive impairment and executive dysfunction after stroke. Dement Neuropsychol 6, 152–157.

11.

Murata

, Kimura

, Robinson

(2000) Does cognitive impairment cause post-stroke depression? Am J Geriatr Psychiatry 8, 310–317.

12.

Starkstein

, Jorge

, Mizrahi

, Robinson

(2005) The construct of minor and major depression in Alzheimer’s disease. Am J Psychiatry 162, 2086–2093.

13.

Kulesh

, Drobakha

, Kuklina

, Nekrasova

, Shestakov

(2018) Cytokine response, tract-specific fractional anisotropy, and brain morphometry in post-stroke cognitive impairment. J Stroke Cerebrovasc Dis 27, 1752–1759.

14.

Sandu

, Buga

, Uzoni

, Petcu

, Popa-Wagner

(2015) Neuroinflammation and comorbidities are frequently ignored factors in CNS pathology. Neural Regen Res 10, 1349–1355.

15.

Zhang

, Wu

, Song

, Li

(2012) Correlations between cognitive impairment and brainderived neurotrophic factor expression in the hippocampus of post-stroke depression rats. Mol Med Rep 6, 889–893.

16.

Baccaro

, Wang

, Candido

, Conforto

, Brunoni

, Leite

CDC

, Busatto Filho

, Lotufo

, Bensenor

, Goulart

(2019) Post-stroke depression and cognitive impairment: Study design and preliminary findings in a Brazilian prospective stroke cohort (EMMA study). J Affect Disord 245, 72–81.

17.

Jiang

, Ling

, Zhang

, Mao

, Ma

, Yin

, Wang

, Tang

, Tan

, Shi

, Li

, Ruan

(2015) Altered fecal microbiota composition in patients with major depressive disorder. Brain Behav Immun 48, 186–194.

18.

Berk

, Williams

, Jacka

, O’Neil

, Pasco

, Moylan

, Allen

, Stuart

, Hayley

, Byrne

, Maes

(2013) So depression is an inflammatory disease, but where does the inflammation come from? BMC Med 11, 200.

19.

Liu

, Wu

, Peng

, Han

, Tang

, Ge

, Zhang

, Jia

, Yue

, Zhou

, Li

, Luo

, Wang

(2019) Altered microbiomes distinguish Alzheimer’s disease from amnestic mild cognitive impairment and health in a Chinese cohort. Brain Behav Immunity 80, 633–643.

20.

Zhuang

, Shen

, Li

, Fu

, Zeng

, Gui

, Lu

, Cai

, Zhu

, Tan

, Zheng

, Li

, Zhu

, Zhou

, Bu

, Wang

(2018) Gut microbiota is altered in patients with Alzheimer’s disease. J Alzheimers Dis 63, 1337–1346.

21.

Olson

, Miller

(2004) Microglia initiate central nervous system innate and adaptive immune responses through multiple TLRs. J Immunol 173, 3916–3924.

22.

Zhan

, Stamova

, Jin

, DeCarli

, Phinney

, Sharp

(2016) Gram-negative bacterial molecules associate with Alzheimer disease pathology. Neurology 87, 2324–2332.

23.

Cox

, Schafer

, Sohn

, Vincentini

, Weiner

, Ginsberg

, Blaser

(2019) Calorie restriction slows age-related microbiota changes in an Alzheimer’s disease model in female mice. Sci Rep 9, 17904.

24.

Scheperjans

, Aho

, Pereira

, Koskinen

, Paulin

, Pekkonen

, Haapaniemi

, Kaakkola

, Eerola-Rautio

, Pohja

, Kinnunen

, Murros

, Auvinen

(2015) Gut microbiota are related to Parkinson’s disease and clinical phenotype. Mov Disord 30, 350–358.

25.

Cryan

, Dinan

(2012) Mind-altering microorganisms: The impact of the gut microbiota on brain and behaviour. Nat Rev Neurosci 13, 701–712.

26.

Shen

, Xu

, Li

, Huang

, Yuan

, Wang

, Zhang

, Hu

, Liang

(2018) Analysis of gut microbiota diversity and auxiliary diagnosis as a biomarker in patients with schizophrenia: A cross-sectional study. Schizophr Res 197, 470–477.

27.

Qin

, Li

, Cai

, Li

, Zhu

, Zhang

, Liang

, Zhang

, Guan

, Shen

, Peng

, Zhang

, Jie

, Wu

, Qin

, Xue

, Li

, Han

, Lu

, Wu

, Dai

, Sun

, Li

, Tang

, Zhong

, Li

, Chen

, Xu

, Wang

, Feng

, Gong

, Yu

, Zhang

, Hansen

, Sanchez

, Raes

, Falony

, Okuda

, Almeida

, LeChatelier

, Renault

, Pons

, Batto

, Zhang

, Chen

, Yang

, Zheng

, Li

, Yang

, Wang

, Ehrlich

, Nielsen

, Pedersen

, Kristiansen

, Wang

(2012) A metagenome-wide association study of gut microbiota in type 2 diabetes. Nature 490, 55–60.

28.

Yin

, Liao

, He

, Wang

, Xia

, Liu

, Zhu

, You

, Chen

, Zhou

, Pan

, Zhou

(2015) Dysbiosis of gut microbiota with reduced trimethylamine-N-oxide level in patients with large-artery atherosclerotic stroke or transient ischemic attack. J Am Heart Assoc 4, e002699.

29.

Liu

, Kong

, Gong

, Zhang

, Zhu

, Wang

, Qu

, Gao

, Yin

, Liu

, Qin

(2020) The association of post-stroke cognitive impairment and gut microbiota and its corresponding metabolites. J Alzheimers Dis 73, 1455–1466.

30.

, Rolls

, Cheng

, Li

, Gong

, Qiu

, Feng

(2019) Functional connectivity of the orbitofrontal cortex, anterior cingulate cortex, and inferior frontal gyrus in humans. Cortex 123, 185–199.

31.

Chen

, Wang

, Niu

, Zhong

, Hu

, Chen

, Zhang

, Chen

, Deng

, Lai

, Wang

, Huang

(2018) Common and distinct abnormal frontal-limbic system structural and functional patterns in patients with major depression and bipolar disorder. Neuroimage Clin 20, 42–50.

32.

Song

, Roy

, Kang

, Aysola

, Macey

, Woo

, Yan-Go

, Harper

, Kumar

(2018) Altered resting-state hippocampal and caudate functional networks in patients with obstructive sleep apnea. Brain Behav 8, e00994.

33.

Nagano-Saito

, Washimi

, Arahata

, Kachi

, Lerch

, Evans

, Dagher

, Ito

(2005) Cerebral atrophy and its relation to cognitive impairment in Parkinson disease. Neurology 64, 224–229.

34.

Park

(2017) Informant questionnaire on cognitive decline in the elderly (IQCODE) for classifying cognitive dysfunction as cognitively normal, mild cognitive impairment, and dementia. Int Psychogeriatr 29, 1461–1467.

35.

Chalhoub

, Luggen

(2019) Screening for cognitive dysfunction in systemic lupus erythematosus: The Montreal Cognitive Assessment Questionnaire and the Informant Questionnaire on Cognitive Decline in the Elderly. Lupus 28, 51–58.

36.

Ling

, Jin

, Xie

, Cheng

, Li

, Wu

(2016) Alterations in the fecal microbiota of patients with HIV-1 Infection: An observational study in a Chinese population. Sci Rep 6, 30673.

37.

, Ren

, Li

, Zhang

, Jiang

, Xu

, Luo

, Zhou

, Sun

, Zheng

, Li

(2016) Deep sequencing reveals microbiota dysbiosis of tongue coat in patients with liver carcinoma. Sci Rep 6, 33142.

38.

Vogt

, Kerby

, Dill-McFarland

, Harding

, Merluzzi

, Johnson

, Carlsson

, Asthana

, Zetterberg

, Blennow

, Bendlin

, Rey

(2017) Gut microbiome alterations in Alzheimer’s disease. Sci Rep 7, 13537.

39.

Petrov

, Saltykova

, Zhukova

, Alifirova

, Zhukova

, Dorofeeva

, Tyakht

, Kovarsky

, Alekseev

, Kostryukova

, Mironova

, Izhboldina

, Nikitina

, Perevozchikova

, Fait

, Babenko

, Vakhitova

, Govorun

, Sazonov

(2017) Analysis of gut microbiota in patients with Parkinson’s disease. Bull Exp Biol Med 162, 734–737.

40.

Tilg

, Moschen

(2014) Microbiota and diabetes: An evolving relationship. Gut 63, 1513–1521.

41.

Finegold

, Dowd

, Gontcharova

, Liu

, Henley

, Wolcott

, Youn

, Summanen

, Granpeesheh

, Dixon

, Liu

, Molitoris

, Green

3rd (2010) Pyrosequencing study of fecal microflora of autistic and control children. Anaerobe 16, 444–453.

42.

Lozupone

, Stombaugh

, Gordon

, Jansson

, Knight

(2012) Diversity, stability and resilience of the human gut microbiota. Nature 489, 220–230.

43.

Lin

, Hung

(2020) Association of subjective global assessment of nutritional status with gut microbiota in hemodialysis patients: A case-control study. Nephrol Dial Transplant, doi:10.1093/ndt/gfaa019.

44.

Lupp

, Robertson

, Wickham

, Sekirov

, Champion

, Gaynor

, Finlay

(2007) Host-mediated inflammation disrupts the intestinal microbiota and promotes the overgrowth of Enterobacteriaceae. Cell Host Microbe 2, 204.

45.

Zeng

, Gao

, Peng

, Wu

, Zhu

, Tan

, Xia

, You

, Xu

, Pan

, Zhou

, He

, Yin

(2019) Higher risk of stroke is correlated with increased opportunistic pathogen load and reduced levels of butyrate-producing bacteria in the gut. Front Cell Infect Microbiol 9, 4.

46.

Shin

, Whon

, Bae

(2015) Proteobacteria: Microbial signature of dysbiosis in gut microbiota. Trends Biotechnol 33, 496–503.

47.

Dinh

, Volpe

, Duffalo

, Bhalchandra

, Tai

, Kane

, Wanke

, Ward

(2015) Intestinal microbiota, microbial translocation, and systemic inflammation in chronic HIV infection. J Infect Dis 211, 19–27.

48.

Litvak

, Byndloss

, Tsolis

, Baumler

(2017) Dysbiotic Proteobacteria expansion: A microbial signature of epithelial dysfunction. Curr Opin Microbiol 39, 1–6.

49.

Tellez

, Ambrosioni

, Llopis

, Pericas

, Falces

, Almela

, Garcia de la Maria

, Hernandez-Meneses

, Vidal

, Sandoval

, Quintana

, Fuster

, Tolosana

, Marco

, Moreno

, Miro

, Hospital

Clínic Infective Endocarditis Investigators

(2018) Epidemiology, clinical features, and outcome of infective endocarditis due to abiotrophia species and Granulicatella species: Report of 76 cases, 2000-2015. Clin Infect Dis 66, 104–111.

50.

, Yu

, Zhang

, Wen

, Wang

, Zhang

, Xu

, Kong

(2019) Dietary fructose-induced gut dysbiosis promotes mouse hippocampal neuroinflammation: A benefit of short-chain fatty acids. Microbiome 7, 98.

51.

Kelly

, Zheng

, Campbell

, Saeedi

, Scholz

, Bayless

, Wilson

, Glover

, Kominsky

, Magnuson

, Weir

, Ehrentraut

, Pickel

, Kuhn

, Lanis

, Nguyen

, Taylor

, Colgan

(2015) Crosstalk between microbiota-derived short-chain fatty acids and intestinal epithelial HIF augments tissue barrier function. Cell Host Microbe 17, 662–671.

52.

Takeshita

, Mizuno

, Mikami

, Sujino

, Saigusa

, Matsuoka

, Naganuma

, Sato

, Takada

, Tsuji

, Kushiro

, Nomoto

, Kanai

(2016) A single species of clostridium subcluster XIVa decreased in ulcerative colitis patients. Inflamm Bowel Dis 22, 2802–2810.

53.

Jeffery

, Lynch

, O’Toole

(2016) Composition and temporal stability of the gut microbiota in older persons. ISME J 10, 170–182.

54.

Alfa

, Strang

, Tappia

, Graham

, Van Domselaar

, Forbes

, Laminman

, Olson

, DeGagne

, Bray

, Murray

, Dufault

, Lix

(2018) A randomized trial to determine the impact of a digestion resistant starch composition on the gut microbiome in older and mid-age adults. Clin Nutr 37, 797–807.

55.

Buford

(2017) (Dis)Trust your gut: The gut microbiome in age-related inflammation, health, and disease. Microbiome 5, 80.

56.

Cattaneo

, Cattane

, Galluzzi

, Provasi

, Lopizzo

, Festari

, Ferrari

, Guerra

, Paghera

, Muscio

, Bianchetti

, Volta

, Turla

, Cotelli

, Gennuso

, Prelle

, Zanetti

, Lussignoli

, Mirabile

, Bellandi

, Gentile

, Belotti

, Villani

, Harach

, Bolmont

, Padovani

, Boccardi

, Frisoni

, INDIA-FBP Group (2017) Association of brain amyloidosis with pro-inflammatory gut bacterial taxa and peripheral inflammation markers in cognitively impaired elderly. Neurobiol Aging 49, 60–68.

57.

Medzhitov

(2010) Inflammation 2010: New adventures of an old flame. Cell 140, 771–776.

58.

Hotamisligil

(2006) Inflammation and metabolic disorders. Nature 444, 860–867.

59.

Pascale

, Marchesi

, Marelli

, Coppola

, Luzi

, Govoni

, Giustina

, Gazzaruso

(2018) Microbiota and metabolic diseases. Endocrine 61, 357–371.

60.

Cani

, Bibiloni

, Knauf

, Waget

, Neyrinck

, Delzenne

, Burcelin

(2008) Changes in gut microbiota control metabolic endotoxemia-induced inflammation in high-fat diet-induced obesity and diabetes in mice. Diabetes 57, 1470–1481.

61.

Hiippala

, Jouhten

, Ronkainen

, Hartikainen

, Kainulainen

, Jalanka

, Satokari

(2018) The potential of gut commensals in reinforcing intestinal barrier function and alleviating inflammation. Nutrients 10, 988.

62.

Passamonti

, Tsvetanov

, Jones

, Bevan-Jones

, Arnold

, Borchert

, Mak

, Su

, O’Brien

, Rowe

(2019) Neuroinflammation and functional connectivity in Alzheimer’s disease: Interactive influences on cognitive performance. J Neurosci 39, 7218–7226.

63.

Zeng

, Inohara

, Nunez

(2017) Mechanisms of inflammation-driven bacterial dysbiosis in the gut. Mucosal Immunol 10, 18–26.

64.

Cersosimo

, Benarroch

(2012) Pathological correlates of gastrointestinal dysfunction in Parkinson’s disease. Neurobiol Dis 46, 559–564.

65.

, He

, Ma

, Huang

, Du

, Cao

, Wang

, Xiao

, Tang

, Chen

(2019) Mild cognitive impairment has similar alterations as Alzheimer’s disease in gut microbiota. Alzheimers Dement 15, 1357–1366.

66.

Zhang

, Yu

, Zhang

, Chen

, Dong

, Lu

, Song

, Zhou

(2018) Porphyromonas gingivalis lipopolysaccharide induces cognitive dysfunction, mediated by neuronal inflammation via activation of the TLR4 signaling pathway in C57BL/6 mice. J Neuroinflammation 15, 37.

67.

Liu

, Wang

, Hu

(2017) Therapeutic potential of amino acids in inflammatory bowel disease. Nutrients 9, 920.