Animal model studies

All seven animal studies focused on gut microbiome of mouse [4–6,11,21–23]. The common mouse model found in all included animal studies was a C57BL/6 mouse model, however, Jandzinski [11] did not declare which mouse model was used to show the effect of age of microbiome on ischemic stroke.

Benakis et al [4] utilized multiple mouse models for different analyses, including wild-type C57BL/6, Il10^−/−, Il17a^−/−, Il17a-eGFP, Trdc-eGFP and KikGR33 mice. They utilized antibiotics treatment to uncover the neuroprotective effect of altered microbial composition on ischemic injury and examine the stroke-induced immune response using brain tissues, blood, and other body tissues including spleens, lymph nodes, and intestinal cells from the mice [4].

Besides, two studies used wild-type C57BL/6J and Rag1^−/− male mice as well as germ-free (GF) C57BL/6J and GF Rag1^−/− female mice [5,22]. Both studies aimed to examine neuroinflammatory response after stroke dysbiosis. Another two studies employed C57BL/6 mice to compare both young and aged microbiome [6,21], whereas another study used the same C57BL/6J mouse model to uncover the microbial composition in intestinal mucosal after stroke [23].

Clinical studies

All the nine clinical studies reported on gut microbiome, where most of the clinical studies were conducted in both case and control cohorts [14–18,20,25,26], except that Boaden et al [24] who studied on stroke patients. All studies collected fecal samples from the study subjects for microbiome analyses and blood samples for blood biochemical assays, except that Li et al [16] collected the fecal samples and Boaden et al [24] collected the saliva sample and swabs within the oral cavity of stroke cases.

Studies reporting on both human and animal subjects

Two studies were conducted on both human and animal subjects [3,19]. One study reported the presence of microbiota in multiple human body sites, such as blood, urine and sputum of human, as well as lung and gut of mouse [3]. This study recruited arterial ischemic stroke cases and used a C57BL/6J mouse model to conduct the microbial analysis of ischemic stroke [3]. Xia et al [19] studied about gut microbiome by collecting human feces as well as mouse feces. This study collected fecal samples from large-artery atherosclerotic cases and healthy controls to examine the microbial composition and establish stroke dysbiosis index. The same study also included a male C57BL/6 mouse model to conduct fecal transplantation experiment and neurobehavioral examination to assess the effect of post-stroke dysbiosis [19].

Microbiome

Eight phyla of bacteria namely Firmicutes, Bacteroidetes, Proteobacteria, Actinobacteria, Spirochaetes, Deferribacteraceae, Verrucomicrobia and Tenericutes were commonly detected in these 16 studies, with both phyla Firmicutes and Bacteroidetes predominated the microbial composition in both human and mice [3–6,11,14–26].

Singh et al [5] showed that phyla Firmicutes and Bacteroidetes made up the major composition of mice gut microbiome, followed by Actinobacteria. This was supported by the other two studies which demonstrated that more than 90% of the fecal microbiome were consisted of phyla Bacteroidetes and Firmicutes [11,21]. Another study revealed that the antibiotic treatment increased the abundance of Proteobacteria while reducing the Firmicutes and Bacteroidetes in antibiotic-sensitive mice [4]. It also demonstrated that phyla Firmicutes and Bacteroidetes were the major gut microbial composition in antibiotic-resistant mice, while Proteobacteria was predominantly found in the antibiotic-sensitive mice [4].

An animal study observed that the organs of mice were colonized by Staphylococcus and Enterococcus after ischemic stroke [6]. Another study found that genera Lactobacilus, Actinomyces, Ruminococcus, Unknown Peptostreptococcaceae, Clostridium, Brevundimonas, and Eshcherichia and Shigella were more abundantly available in the lung of post-stroke mice as compared to the control group. More specifically, the relative abundance of Escherichia and Shigella species, Streptococcus species, Lactobacillus species, Brevundimoas nasdae and Staphylococcus sciuri were significantly higher in the lungs of post-stroke mice [3]. A subsequent study published by the same group of researchers showed an elevation in multiple Clostridium species, Parabacteroides goldsteinii, Anaerotruncus colihominis, Alistipes shahii, Akkermansia muciniphila and Roseburia intestinalis in the gastrointestinal tract of post-stroke mice [23].

The clinical studies found that the abundance of phylum Firmicutes is an independent predictor for ischemic stroke risk [16–18,25]. Particularly, genera Streptococcus, Lactobacillus and Prevotella were descent from phylum Firmicutes prevailed among ischemic stroke cases [16–18,25]. Boaden et al [24] found that Streptococcus species were the most abundant microbiota that made up 70% of the oral microbial community in stroke patients. A study showed that the abundance of S. infantis and P. copri were found higher in cases with ischemic stroke, while Blautia obeum was relatively lower among ischemic stroke cases [17]. The lactic acid-producing Lactobacillus and Lactococcus were significantly enhanced in fecal samples of cases [16]. In addition, Wang et al [20] observed an increase in Gammaproteobacteria and a decrease in Bacteroidia in cases.

However, another study reported that Proteobacteria was increased in the cases, and a lower abundance of Bacteroides, Prevotella and Faecalibacterium was observed [26]. This study also demonstrated a higher abundance of Proteobacteria and a reduction of Bacteroides in cases following more severe stroke outcomes [26]. Similarly, another study by Haak et al [14] also observed that the abundance of Proteobacteria was increased while the levels of Firmicutes and Bacteroidetes were reduced in stroke patients. Li et al [16] observed that the relative abundance of genera Odoribacter, Akkermansia and Ruminococcaceae_UCG_005 were significantly higher in stroke cases. More specifically, Ruminococcaceae_UCG_005 was enhanced in severe stroke cases [18]. Tan et al [15] also reported an increase in Akkermansia along with Lactobacillaceae, Enterobacteriacea and Porphyromonadaceae in cases with acute ischemic stroke, particularly those with severe stroke. Another clinical study demonstrated that ischemic stroke was independently associated with reduced amount of L. sakei subgroup and increased abundance of Atopobium cluster and L. ruminis [25].

A study by Stanley et al [3] identified Enterococcus spp., Escherichia coli and Morganella morganii from the blood, urine or sputum samples from 22.2% of stroke cases, and there was no culturable microbiota in the control group. Meanwhile, Haak et al [14] observed that the abundance of aerobic Enterococcus and Escherichia/Shigella were increased in stroke patients, while a drastic reduction in obligate anaerobic Anaerostipes, Ruminococcus, and Subdoligranulum was reported in stroke patients. Xia et al [19] found that seven genera were significantly enriched in the fecal microbial composition of stroke cases, such as Butyricimonas, Parabacteroides, Unknown Rikenellaceae, Unknown Ruminococcaceae, Oscillospira, Bilophila and Unknown Enterobacteriaceae.

Diversity analyses

A total of nine eligible studies measured α-diversity [3–5,14,16–18,20,22,26]. Two studies reported no significant difference between post-stroke mice and sham-operated mice [3,22]. Similar phenomenon was also observed between stroke cases and healthy controls in the other three clinical studies [16–18]. On the other hand, Singh et al [5] reported a reduced microbial diversity in the gut microbiome of post-stroke mice as compared to the controls, whereas Benakis et al [4] observed a reduced α-diversity in antibiotic-sensitive mice. Similarly, two clinical study also reported a lower microbial diversity in cases with ischemic stroke when compared to healthy controls [14,20]. In contrast, Yin et al [26] showed a significantly higher microbial diversity in cases than that of the controls.

There are ten studies measured β-diversity, including three animal studies [3,5,22] and seven human studies [14–16,18–20,26]. Eight studies demonstrated significant differences in β-diversity between stroke group and control groups [3,5,14,15,18–20,22,26], while one study reported no difference in the microbiota structure between both cases and controls [16]. Interestingly, a study showed that cases with lower SDI index exhibited quite similar fecal microbial composition as the healthy controls [19].

Effect of aging on microbiome and ischemic stroke

Three studies assessed the impact of age on microbial composition [6,11,21], among which two evaluated the role of age in the gut microbiome of stroke mouse model, particularly on Firmicutes/Bacteroidetes ratio [11,21]. Two studies observed a higher Firmicutes/Bacteroidetes ratio in aged microbiome of the mice when compared to the young microbiome following a stroke event [11,21]. Ischemic stroke induced the increment of Firmicutes and reduced the abundance of Bacteroidetes in both young and aged mice, but the Firmicutes/Bacteroidetes ratio was greatly enhanced in aged mice when compared to the young mice [21]. These studies also found that the aged mice with a high Firmicutes/Bacteroidetes ratio were unable to recover from neurological deficits and showed a higher mortality rate [11,21].

A study showed a significant higher bacterial burden in mesenteric lymph nodes of aged mice as compared to young mice after stroke dysbiosis, which caused the aged mice to have a higher mortality rate [6]. The same study also identified genus Escherichia in young mice only, while the presence of Enterobacter could only be found in aged mice [6]. In addition, Jandzinski [11] also showed that phylum Deferribacteres was detected in aged microbiome only, while only young microbiome showed the presence of Verrucomicrobia.

Stroke dysbiosis

Ten included studies suggested that stroke induced gut dysbiosis, altered the microbial composition, and manipulated the post-stroke outcome [3–6,11,15,16,19,21,26]. Crapser et al [6] proved that ischemic stroke increased gut permeability, induced bacterial translocation from the gut to mesenteric lymph nodes, spleens, livers and lungs, and led to a high risk of infection due to gut dysbiosis. A study demonstrated that a significant depletion of microbiota was seen in the intestinal compartments including ileum and colon after stroke, while a remarkable increase of microbiota could be observed in the lung of post-stroke mice [3]. The same animal study also proved that the microbiota dysbiosis was induced by stroke, which subsequently impaired the immune and barrier defense system of the host body [3]. Singh et al [5] discovered that the species abundance in Firmicutes, Bacteroidetes and Actinobacteria were modulated by gut dysbiosis after stroke.

Tan et al [15] found a reduction of SCFAs-producing bacteria in cases with ischemic stroke when compared to controls. Nonetheless, a study showed that the butyrate-producing bacteria were remarkably less abundant in ischemic stroke and with increasing abundance of lactic acid bacteria [16]. Besides, Yin et al [26] determined more opportunistic pathogens in cases which were associated with significant higher TMAO levels. Meanwhile, two studies observed a shift in Firmicutes/Bacteroidetes ratio in the fecal microbial composition of mice as an effect of stroke dysbiosis, where ischemic stroke significantly increased the ratio Firmicutes/Bacteroidetes [11,21].

Two studies demonstrated gut dysbiosis through the translocation of intestinal immune cells to the brain after stroke [4,5]. A study observed that T cells migrated from the intestinal lamina propria to the meninges after stroke [4], whereas another study showed that lymphocytes could migrate from Payer’s Patches to the brain after induced by stroke [5]. Intriguingly, a study established a stroke dysbiosis index (SDI), a microbiota index in ischemic stroke based on the gut dysbiosis pattern, whereby the higher the SDI index, the more severe the brain injury and the poorer functional outcome the cases had. The same study also observed that genera Oscillospira, Enterobacteriaceae, Bacteroides, and Bacteroidaceae were more abundant in ischemic stroke cases with high SDI (SDI-H) [19].

Neuroinflammatory response by microbiota after ischemic stroke

A study showed that stroke elevated gut permeability and selectively increased vascular permeability in the jejunum and ileum of the post-stroke mice, leading to a significant increase in the abundance of goblet cells in the jejunum and ileum [3]. A significantly increased abundance of IgA⁺ B cells were observed in the mesenteric lymph nodes of the post-stroke mice, while a significant reduction of neuronal submucosal cholinergic ChAT⁺ cells was observed in post-stroke mice [3].

A recent study by Xia et al [19] assessed potential microbiota dysbiotic effect on stroke injury in mouse model by performing fecal transplantation from stroke cases with SDI-H to mice. An aggravated abundance of pro-inflammatory (IL-17⁺) γβ T cells was seen in both spleen and small intestine of the SDI-H recipient mice, whereas depleted (CD4⁺CD25⁺) helper T (T_helper) cells and regulatory T cells (T_reg) (CD4⁺ Foxp3⁺) cells were observed in the spleen and small intestine of the SDI-H recipient mice. The results demonstrated an elevated infarct volume and poorer neurological functional outcome in the SDI-H recipient mice after stroke [19].

Another study demonstrated the neuroprotective effect of gut microbiota on the ischemic injury by colonizing the germ-free mice with gut microbiota [22]. As a result, a higher number of microglia/macrophages as well as a remarkable increased expression of proinflammatory cytokines were observed in the ischemic brain of the post-stroke mice [22]. The cell counts of T_helper, T_reg and Th17 were increased in the Peyer’s patches and even enhanced in the spleens after stroke, while the similar pattern was also observed in the ischemic brain, leading to a smaller lesion volume in the mice brain after stroke [22].

Crapser et al [6] demonstrated that a significantly higher percentage of CD4⁺ and CD8⁺ T cells as well as the event of lymphopenia were found in the blood of stroke mice as compared to sham-operated mice, while a remarkably greater proportion of infiltration leukocytes, including CD3⁺ T cells were observed in the brain of stroke mice. Other than that, Yamashiro et al [25] showed that L. ruminis was positively correlated with interleukin-6 (IL-6) level, while C. coccoides was negatively correlated with IL-6 and high sensitivity C-reactive protein (hsCRP).

Furthermore, mice receiving young microbiome possessed much higher levels of IL-4 and granulocyte-colony stimulating factor (G-CSF), while the elevation of levels of IL-6, tumor necrosis factor alpha (TNF-α), Eotaxin, and CCL5 were shown in the mice with aged microbiota after stroke [21]. Crapser et al [6] showed that IL-6 levels were generally increased in both young and aged mice after stroke, while aged mice showed a much higher level of IL-6. A significant negative correlation was observed between the effects of aging and serum lipopolysaccharide-binding protein (LBP) levels, causing a higher rate of sepsis in aged mice due to a lower level of immune response [6].

Interestingly, a study investigated the neuroprotection of the altered gut microbiome after ischemic injury by treating a group of mice (AC^Sens) with antibiotics amoxicillin and clavulanic acid (AC) [4]. An elevated cell count of T_reg and a depletion in IL-17⁺ γβ T cells were observed in the AC^Sens mice, while a higher number of IL-17⁺ γβ T cells were in the meninges after stroke. Besides, the study also found that gut microbiota was able to manipulate the function of dendritic cell in the intestine, as it could suppress the differentiation of IL-17⁺ γβ T cells with the help of IL-10 [4].

KEGG pathway analysis

Only three studies performed KEGG functional pathway analysis [16,17,22]. Huang et al [17] showed that four bacterial pathways were present in the cases with ischemic stroke, of which, lipopolysaccharide synthesis was significantly enriched in the cases with ischemic stroke. Another study reported that a total of eight KEGG pathways were found to be significantly upregulated in post-stroke mice as compared to sham-operated mice [22]. Intriguingly, both studies observed that bacterial secretion was significantly enhanced in stroke group when compared to the control group. In contrast, an enhancement in human disease-associated module including genes corresponding to infectious diseases was observed in cases with ischemic stroke [16]. The same study also observed a significant reduction of the sporulation functional gene expression level of butyrate-producing bacteria while a remarkable increased expression of the lactic acid bacteria-related phototransferase system in cases with ischemic stroke [16].

Association between clinical markers and gut microbiota in ischemic stroke

Huang et al [17] found that cases with ischemic stroke exhibited significantly higher levels of hyperlipidemia, total cholesterol, triglycerides, higher blood pressure and white blood cell count. Besides, Xia et al [19] also showed a positive association between alcoholic consumption and HbA1c with severe stroke, where age, creatinine and uric acid were negatively correlated with post-stroke functional outcomes. Besides, Xia et al [19] also found that stroke dysbiosis and white blood cell count were the independent predictors of severe stroke, while stroke dysbiosis was found to be an independent predictor of early poor functional outcome after stroke.

While Huang et al [17] found no difference in blood glucose level, creatinine and uric acid between cases and controls, Xia et al [19] showed a positive association between blood glucose, creatinine and uric acid with severe stroke. A study also found a positive correlation between uric acid and Dialister, while blood glucose was remarkably negative correlated with genera Ruminococcaceae_UCG-002, Alistipes and Ruminococcus_1 [18]. While Yamashiro et al [25] found that type 2 diabetes was associated with a lower level of Clostridium coccoides, Xia et al [19] showed that diabetes mellitus was positively related to severe stroke and negatively correlated with functional recovery after stroke.

Significant higher levels of lipoprotein and high-density lipoprotein (HDL) were demonstrated in cases with ischemic stroke [17], while the lipoprotein levels between cases and controls were similar in another study [14]. A study showed a significant positive correlation between HDL and genera Ruminococcus_1, Ruminococcus_2 and Lachnospiraceae_NK4A136_group, while a negative correlation was found between HDL and Enterobacter [18]. In addition, Li et al [18] also unveiled a notably positive correlation between LDL and Bacteroides and Eubacterium rectole group. Meanwhile, LDL was negatively correlated with norank_O_Mollicutes_RF9 and C. coccoides [18,25]. Besides, Wang et al [20] showed that Bacteroidia was negatively correlated with apoliprotein E (APOE), while Gammaproteobacteria was positively correlated with APOE level. Intriguingly, a positive correlation was also shown between homocysteine and genera Megamonas and Fusobacterium [18].

Microbiota-derived metabolites and ischemic stroke

Of the 16 studies, eight studies observed the association between microbiota-derived metabolites, such as short-chain fatty acids (SCFAs) and trimethylamine-N-oxide (TMAO), and ischemic stroke [14–16,18,21,23,25,26]. Only two studies reported on the association between TMAO level and ischemic stroke [14,26]. One study showed a remarkable decrease in blood TMAO levels in cases with ischemic stroke after gut dysbiosis [26], while another study reported a higher level of TMAO-producing microbiota in ischemic stroke cases, especially in ischemic stroke cases with severe stroke [14]. When comparing to the healthy controls, the median concentration of total SCFAs was lower in cases, especially those who exhibited poorer neurological outcomes [15]. The concentration of acetate was significantly lower in cases [25]. Meanwhile, valerate and isovalerate were remarkably higher in cases [25]. The study also observed a negative correlation between the concentrations of acetate and propionate with the levels of HbA1c and LDL, while the concentration of valerate was positively associated with the WBC counts and the level of hsCRP [25]. Furthermore, a study showed that acetate and propionate were significantly reduced in mice with aged microbiomes as compared to young microbiome [21].

Three studies demonstrated a reduction of SCFA-producing bacteria in cases with ischemic stroke [14–16]. However, a significantly higher abundance of SCFAs-producing genera Odoribacter and Akkermansia were found in the gut of ischemic stroke cases [18]. Similarly, as high as two-fold abundance of Akkermansia in the intestinal mucosa and lungs of stroke mice has been observed [23]. Besides, the concentrations of acetate and propionate were reduced by approximately 68% in mice with aged microbiomes as compared to mice with young microbiomes.