Damian Sendler, M.D.: The role of the microbiome in influencing the brain and behavior is becoming more widely understood. Humans’ gut-brain communication relies heavily on the actions of microbial metabolites. It has been shown in preclinical studies that bile acid metabolism, short-chain fatty acids, and tryptophan are all associated with improved mental performance. It is possible to apply new predictive tools to existing datasets with the discovery of neuroactive gut-brain modules.
Damian Jacob Sendler: We identified 278 studies that included sequencing data related to the human microbiota-gut-brain axis. Mental and neurological illnesses were included, but only a few studies looked at the development of normal behavior. Thirty-five of these datasets were reanalyzed from publicly available raw sequencing files using a consistent bioinformatics pipeline, and the remainder were summarized and compiled. We found evidence of disease-related alterations in microbial metabolic pathways in Alzheimer’s disease, schizophrenia, anxiety, and depression after re-analyzing studies. Many sequencing and technical limitations prevented the discovery of specific biomarkers of microbes or metabolites that were conserved across studies. Further research is needed to verify our findings. In addition, we propose guidelines for future human microbiome analysis in order to improve reproducibility and consistency in the research field.
Dr. Sendler: The host microbiota metabolizes dietary fibers, proteins, and fats that are ingested by the host. Fermentation of fibers and tryptophan-kynurenine (TRP-KYN) metabolites from dietary proteins yields short-chain fatty acids (SCFAs). Secondary bile acids can be synthesized by deconjugating primary bile acids, which are produced during liver metabolism. To understand how these metabolites affect brain health and disease, we can look to human fecal microbiota composition research. Results could not be reproduced across studies because of the wide range of physiological differences between individuals and the numerous technical and bioinformatics constraints. However, we found mild-to-moderate evidence of the involvement of these metabolic pathways in Alzheimer’s disease, schizophrenia, and anxiety/depression by reanalyzing 35 studies with a consistent pipeline and comparing these results to other existing studies.
Antimicrobial metabolites have been harnessed for their antimicrobial properties since the 1928 discovery of penicillin, and they are now emerging as mediators of mammalian health and behavior (Fleming, 1946b; O’Mahony et al., 2015; Blacher et al., 2017; Levy et al., 2017; McCarville et al., 2020). The microbiota is a collection of microorganisms that colonize the gastrointestinal tract of mammals at birth (Codagnone et al., 2019; Theis et al., 2019). The gut microbiota plays an important role in the conversion of host, xenobiotic, and dietary-derived molecules into bioactive metabolites that can influence host health and disease (Clarke et al., 2019; Sharon et al., 2014; Spanogiannopoulos et al., 2016; Morris et al., 2017; Sun et al., 2017). The microbiome is an ecological community that coexists in a shared space (Lederberg and McCray, 2001). The abundance of genes within the microbiome that enable the production and modification of neuroactive metabolites, which may alter the function of the gut-brain axis, has been one of the most surprising discoveries in recent decades (Zimmermann et al., 2019; Strandwitz et al., 2019; Lyte, 2014; Clarke et al., 2014; Tennoune et al., 2014; Lee et al., 2015). For the most part, the gut microbiota is described in terms of bacteria and archaea.
In the microbiota-gut-brain axis, microbial metabolites communicate dynamically bi-directional pathways to mediate host brain immunity and physiology (Spichak et al., 2019; Erny et al., 2017; Pott and Hornef, 2012; Blacher et al., 2017; Levy et al., 2017; McCarville et al., 2020). Indirectly or directly, they have a direct or indirect effect on the body’s immune, neuroendocrine, or vagal systems (Alenghat, 2015; McCarville et al., 2020; Roager and Licht, 2018; Stilling et al., 2016; Fulling et al., 2019). Over the past decade, advances in sequencing technologies have made it possible to rapidly and comprehensively illuminate the composition of the gut microbiome (Song et al., 2018; Bailey et al., 2019; Shakya et al., 2019). Genomic sequencing of stool samples is frequently used as a proxy for the composition of an individual’s microbiome. However, due to the wide range of study designs, it’s difficult to draw any firm conclusions from the available evidence (Pollock et al., 2018). Despite this, these investigations are crucial for determining the role played by bacterial metabolites in the central nervous system of the human host (CNS).
It is our intention to focus on three of the gut microbiota’s most studied metabolic pathways: short-chain fatty acids (SCFAs), tryptophan metabolic pathway, and bile acid metabolic pathway.
Many studies have linked the personality traits of individual bacteria genera to their classification. Blautia abundance was found to be negatively associated with anxiety in healthy participants by Taylor et al. (2019). There was no assessment of anxiety in Tillisch et al. (2017), but they did discover a link between high levels of Prevotella and lower levels of negative affect. Kim et al. (2018) found a link between increased abundance of Roseburia and conscientiousness, while Johnson (2020) found a link between increased abundance of Oscillispira and social ability.
Ten people with insomnia and ten healthy controls were studied by Liu et al. (2019a). They collected faecal samples from both groups. However, Alloprevotella abundance was significantly reduced in those with insomnia (padj less than 0.1, effect = [-14.97; 0.17]; 95 percent confidence interval: [-1.16; 0.17]). This dataset contained no other microbes or GBMs related to SCFA, tryptophan, or bile acid pathways.
Few studies have examined the relationship between sleep quality and the composition of the microbiota (see Table 1). The results of two of these studies showed no correlation between sleep and microbes at the genera level (Liu et al., 2020c; Anderson et al., 2017). Individuals with a Prevotella enterotype showed sleep disruptions in one 16S sequencing study (Ko et al., 2019). Smith et al. (2019) compiled a vast amount of data, which they used to identify specific microbes and their relationship to sleep patterns. However, the prevalence of Holdemania and Corynebacterium was not associated with an increase in the number of awakenings, whereas Coprococcus and Neisseria were (Smith et al., 2019). Blautia has also been linked to a decrease in sleep efficiency and total time spent asleep (Smith et al., 2019). Participants used faecal swabs to collect samples of microbiota, which is an interesting finding (Smith et al., 2019).
There aren’t any existing studies that link genus-level differences with healthy senescence of the mind (see Table 1). All of these studies, however, looked at various subgroups of people who were aging in an unhealthy way cognitively. Several studies compared healthy aging to mild cognitive impairment, Cirrhosis, Alzheimer’s, and a 12-week crossover-double-blind trial, among other things,
Over the course of eleven 16S and WGS bipolar disorder studies, no consistent microbial changes were found (see Table 10). Streptococcus, Clostridium, Oscillibacter, and Bifidobacterum were found to have higher levels in bipolar patients’ samples in two WGS studies (Rong et al., 2019; Lai et al., 2021). Other studies were unable to detect these differences in abundance because of differences in methodology and data analysis. Bipolar individuals had lower levels of Facealibacterium, which Evans et al. (2017) found to be associated with sleep disturbances and depressive symptoms. Faecalibacterium abundance in bipolar disorder was found to be higher than previously reported. Bacteroides was found to be more prevalent in bipolar disorder in two other studies, but their findings differed from each other (Hu et al., 2019b; Zheng et al., 2020b). A few studies did not find any genus-level differences in the microbial composition of the soil samples (Coello et al., 2019; Vinberg et al., 2019; McIntyre et al., 2019). Coello et al. (2019) found that sex effects, heritability, and smoking were responsible for all of the observed differences in abundance.
Numerous studies have looked into the gut microbiota’s role in anxiety and depression (see Table 10). Major depressive disorder (MDD) did not differ significantly from controls based on the results of only three 16S or WGS sequencing studies (Paulsen et al., 2017; Bharwani et al., 2020; Naseribafrouei et al., 2014). There was an increase in Akkermansia and Veillonella in the gut microbiota of people who were currently experiencing a depressive episode, while Fusicatenibacter, Sutterella, and Dialister were decreased in the gut microbiota of people who were not depressed. Dialister was found to be lower in patients with active MDD in a previous study (Jiang et al., 2015). As a result, it is not clear if the gut microbiota changes during depression or if these differences are the result of different methodologies being used in other studies.
In other studies, only a few microbial signatures were found to be similar. In depressed individuals, the microbe Collinsella, which is linked to the metabolism of SCFA and tryptophan, was found to be more prevalent in two studies (Stevens et al., 2018; Zheng et al., 2016). The abundance of Blautia in MDD was also found to have increased in three other studies (Huang et al., 2018; Jiang et al., 2015; Yang et al., 2020). It was found that the microbiota in MDD did not predict clinical response when researchers predicted clinical outcomes from baseline microbiome data. They found that Paraprevotella was strongly linked to the Hamilton Depression Rating Scale-24 Item metric (Likiewicz et al., 2021) despite this.
An investigation by Madan et al. (2020) sought to identify microorganisms that predicted remission or readmission from severe depression or anxiety in patients. Anxiety was associated with moderate Coprococcus catus at admission, and was reduced in those with lower rates of remission from anxiety or depression (Madan et al., 2020). According to Huang et al., the Coprococcus species was found to be more prevalent in the depressed group (2018). Coprococcus abundance decreased in depression, according to other studies (Valles-Colomer et al., 2019; Liu et al., 2016).
Damian Sendler
Researchers used a combination of compositional data and large cohorts to find Coprococcus, Faecalibacterium, and Dialister associated with a better quality of life, while Dialister was depleted in depression, according to Valles-Colomer et al. Oddly, the Dialister findings are in line with those from other studies (Jiang et al., 2015, 2020). Facalibacterium has been linked to anxiety and depression in other studies, as well. Additionally, a recent systematic review found a decrease in the number of bacteria that produce SCFAs like Faecalibacterium in anxiety and depression studies (Simpson et al., 2020).
SCFAs are carboxylic acid molecules with a 1–6 carbon chain chain (Dalile et al., 2019). A by-product of the colonic bacterial fermentation of inulin, cellulose, wheat bran, and resistant starches is the production of SCFAs (Cummings, 1981). Scaffolding compounds (SCFAs) are produced by a number of bacteria commonly found in the gut, including Akkermansis and Bifidobacterium. Lactobacillus, Ligilactobacterium and ruminants Ruminococcus and Ruminnoclustridium are also known to produce SCFAs (Takada et al., 2013; Dalile et al., 2019; Joseph et al., 2017; Valles-Colomer et al., 2019; Basson et al., 2016; Zheng et al., 2020a). SCFA absorption in the colon is not known to be affected by the presence of these microbes, nor is it known how GI absorption differs between individuals without these microbes (Dalile et al., 2019). SCFA circulating concentrations can also be affected by the genetics of the host, the diet, and the colonic absorption of SCFAs (Dalile et al., 2019).
pyruvate metabolism and other molecules associated with the Krebs Cycle are impacted by the process of making SCFA (see Fig. 1). Acetate, butyrate, and propionate are the three most common SCFAs in humans (Dalile et al., 2019). Differences between them include aliphatic tail length and carboxylic acid group position (Dalile et al., 2019). G-protein-coupled receptors (GPCRs; FFAR1, FFAR2, FFAR3, GPR109A, GPR164 and OR51E2) affinity and specificity are affected by these minor differences (Dalile et al., 2019). SCFAs are also histone deacetylase inhibitors in enteric neurons, enterochromaffin cells, and microglial cells (Stilling et al., 2016; Erny et al., 2015; Dalile et al., 2019; Woo and Alenghat, 2017; Yang et al., 2019). There are two ways in which SCFAs affect human health: First, they can increase FOXP3+ Treg cell proliferation, which in turn can lead to the expansion of dendritic cells and macrophages that in turn promote the maturation of T cells (Woo et al. 2017). (Woo and Alenghat, 2017).
Damian Jacob Markiewicz Sendler: Several neurotransmitters, including serotonin and tryptophan, can be generated and modified by the gut microbiota (see Gheorghe et al. (2019) and Lee et al., 2015, for review). In the 1970s, researchers discovered that gastrointestinal microbes could metabolize tryptophan and its various metabolites (Allison et al., 1974; Whitt and Demoss, 1975). An indole, a compound produced by microbes to communicate with the host (Lee et al., 2015; O’Mahony et al., 2015), has been functionally defined in the decades since Indoles, which are commonly produced by pathogenic bacteria, can also be found in a symbiotic environment (Lee et al., 2015). While tryptophan, an essential amino acid found in many foods, is used to make the neurotransmitter serotonin (Reigstad et al., 2015), the bacterial enzyme tryptophanase is used to break down tryptophan into indoles (Lee et al., 2015).
Bacteria that can express tryptophanase are also involved in the other tryptophan metabolic pathways that are described below. In addition to the aforementioned, there are a number of other genera to consider: Bacteroides, Butyrivibrio, Clostridium, Enterococcus, Escherichia coli, Eubacterium, Haemophilus, Fusobacterium, Peptostreptococcus, Bifidobacterium, Parabacteroides, Megamonas, Anaerostipes, and Ruminococcus.
High nanomolar to low millimolar levels of indoles are found in the colon (Bansal et al., 2010; Karlin et al., 1985). Human intestinal epithelial cells respond to these tryptophanase-derived metabolites at millimolar concentrations by producing more mucin and increasing their tight junction resistance (Bansal et al., 2010; Karlin et al., 1985). Myenteric plexus motility and neuronal signaling are also regulated by indole metabolites via the aryl-hydrocarbon receptor (Obata et al., 2020). Similarly, indoles regulate inflammation and immunity in the CNS astrocytes by acting on the same receptor (Rothhammer et al., 2016, 2018).
Damian Jacob Sendler
These datasets contain several noteworthy findings and features, even though evidence for the involvement of specific microbial genera or GBMs in human SCFA, tryptophan or bile acid metabolism is weak. For mechanistic insights, GBMs allow us to search metagenomic data for specific neuroactive metabolic pathways. A number of human and preclinical studies have already taken advantage of their descriptive properties in the short time since their release (Butler et al., 2020; Chen et al., 2020; Tomizawa et al., 2020; Zhu et al., 2020).
The relationship between temperament, cognition, and personality is being studied in numerous studies involving healthy humans. Despite the fact that these studies may continue to find various associations, without proper compositional data analysis these associations are likely spurious and biased toward negative correlations (Gloor et al., 2017). Many participants may be needed to power a study using compositional data methods, even when identifying explanatory genera or ASVs (Hughes et al., 2020).
Many of the human-microbiome-brain studies were attributed to disorders of neurodevelopment. ADHD suffers from an absence of compositional analysis, which may explain why a WGS study found decreased KO abundance in dopamine pathways. Faecalibacterium, Ruminococcus, and Ruminoclostridium 9 have also been linked to ADHD symptoms in other research (Jiang et al., 2018b; Szopinska-Tokov et al., 2020). Metabolomic methods are needed to confirm whether these microorganism genera alter SCFA or tryptophan-related pathways. Numerous studies on autism spectrum disorders (ASD) found only a few commonalities among their findings. Few GBMs or differentially regulated microbes met the significance and effect size thresholds when reanalyzing raw microbiome data. One set of identical twins who were both diagnosed with ASD had the same microbial composition, according to Son et al. (2015). A few of these studies, however, pointed to the importance of diet and faecal SCFAs in the development of ASDs (Berding and Donovan, 2019; Liu et al., 2019b; Wang et al., 2020c). Lactobacillus and Bifidobacteria have been found to be dysregulated in multiple studies of schizophrenia, as well as in studies of SCFA and tryptophan-related GBMs (Zhu et al., 2020; Xu et al., 2020; Schwarz et al., 2018; Shen et al., 2018). We found evidence of widespread dysregulation across most existing schizophrenia studies, despite the difficulties in determining strong associations due to the diversity and new nomenclature of Lactobacillus genera (Zheng et al., 2020a).
Damien Sendler: The tryptophan biosynthesis and S-adenosyl methionine biosynthesis GBMs were found to be more abundant in young epileptic children in one longitudinal study assessing the impact of the ketogenic diet on the microbiota (Lindefeldt et al., 2019). This suggests that the ketogenic diet has a different mechanism of action on epilepsy than in mice (Olson et al., 2018). It is difficult to compare results from studies assessing the overall microbial differences found in epilepsy because they are based on different cohorts.
In contrast to Olson et alfindings .’s in mice, these findings don’t match up (2018). The anti-epileptic effects of the ketogenic diet in mice are mediated by the gut microbiota, specifically by Akkermansia and Parabacteroides (Olson et al., 2018). There is evidence that the ketogenic diet alters the gut microbiota and intestinal immunity in healthy adults, but further research is needed to understand the mechanisms of anti-epileptic effects in humans (Ang et al., 2020).
AD and MCI patients have decreased SCFA and tryptophan-related GBM abundance, which is consistent across neurodegenerative disorders (Li et al., 2019a). However, the overwhelming majority of this evidence comes from a single reanalysed research study. For MSA and amyotrophic lateralsclerosis, the microbiota is an intriguing target. However, we did not find enough studies to warrant consensus. Preclinical evidence suggests that PD pathogenesis can be initiated by overexpression of -Synuclein in the myenteric plexus, which reaches the brain via the vagus nerve. There were no consistent changes in the gut microbiota, even after re-analyzing the raw data and taking metadata into account. There are many subtypes of Parkinson’s disease, so it may be necessary to divide participants into groups based on medication and subtype. However, this was a somewhat unexpected discovery.
Cross-study comparisons of alcohol dependence studies may have been hindered by a lack of dietary metadata. Many of these datasets identified various genera involved in SCFA and tryptophan metabolism (Bjorkhaug et al., 2019; Seo et al., 2020; Dubinkina et al., 2017; Leclercq et al., 2014). Small sample size allowed us to detect an increase in the tryptophan degradation module and a decrease in propionate synthesis III, even with only 10 smokers and 10 controls (Stewart et al., 2018). Research into the impact of recreational drug use also found variations in the tryptophan and SCFA-associated genera that needed to be explained (Fulcher et al., 2018; Panee et al., 2018). Metadata collection must take this into account because some of the microbiota-related differences between two groups may be explained by alcohol and drug use.
Two studies looking into demyelinating diseases found lower faecal SCFA concentrations (Gong et al., 2019; Zeng et al., 2019). Microbiota shifts or altered gastrointestinal physiology may be to blame for this, but researchers aren’t sure.
There aren’t enough studies on the microbiome and fibromyalgia and migraine to draw firm conclusions. Fibromyalgia research has found altered microbial species associated with SCFA and tryptophan metabolism, and changes in serum levels of SCFAs, in one fibromyalgia study (Minerbi et al., 2019). The kynurenine synthesis GBM was also found to be more prevalent in the only migraine-microbiota study (Chen et al., 2020c). Even though there was little evidence to support a link between IBS and specific taxa, fecal matter transplantation of material high in Bifidobacterium or probiotic Bifidobacterium strains may improve the psychological aspects of this disease.
Using 16S and WGS methods, we looked for differences in the abundance of microbes and metabolic pathways in the microbial community that might be linked to psychological aspects of obesity. Isovaleric acid I synthesis I, quinolinic acid synthesis, and degradation were found to be higher in anorexics than in control individuals when re-analyzing studies of anorexia (Mack et al., 2016). There was a greater increase in ClpB GBM than in controls at admission, but this decreased after weight gain (Tennoune et al., 2014). (Mack et al., 2016). If ClpB and hunger are also involved in microbial-host pathways that interact with SCFA and tryptophan metabolism, we don’t know yet.
Stroke and vascular disease conditions vary widely, making it difficult to draw meaningful conclusions from one study to another. A neurovascular condition in humans has been linked to specific microbial genera and species, according to (Polster et al., 2020). However, rather than SCFA production, these taxa were linked to LPS biosynthesis (Polster et al., 2020).
Some research suggests that prenatal and postnatal stress can permanently alter the microbiome (Naude et al., 2019; Hantsoo et al., 2019; Carlson and colleagues, 2018), but these studies do not demonstrate the involvement of bacteria that modifies SCFA, tryptophan, or bile acids (Hantsoo and colleagues, 2019; Carlson and colleagues, 2018). The microbiota’s impact on post-traumatic stress disorder has not been studied extensively like stress. Bipolar disorder has been studied extensively in the microbiota, but no consistent patterns emerged. Indeed, sex effect, heritability and smoking were found to explain the observed differences in gut microbiota between bipolar disorder and controls (Coello et al., 2019). Meanwhile, Dialister and Faecalibacterium have shown some evidence of anxiety and depression reduction in various studies (Valles-Colomer et al., 2019; Jiang et al., 2015, 2020; Jiang et al., 2018a; Stevens et al., 2018). These microbes contribute to SCFA and tryptophan-related pathways that may or may not have an effect on the host’s phenotype.
