Parkinson’s Disease and Gut Microbiota

Parkinson’s disease (PD) is the most commonly recognized movement disorder characterized by muscle rigidity, postural instability, gait disturbance, bradykinesia, and rest tremor. PD patients frequently exhibit nonmotor symptoms including mild-to-severe cognitive impairments. Gastrointestinal symptoms are also commonly observed nonmotor symptoms in PD patients affecting up to 80%. Braak et al. [1] examined Lewy bodies in autopsied PD patients without cognitive symptoms as well as in healthy subjects and found that Lewy bodies start from the dorsal vagal nucleus and ascend to the locus ceruleus and the substantia nigra. Clinically, constipation, rapid eye movement behavior disorder (RBD), and depression frequently precede approximately 20, 10, and 5 years, respectively, before the onset of motor symptoms of PD [2].

Inflammatory bowel diseases, namely, ulcerative colitis and Crohn’s disease, increase the incidence of PD by 22–35% [3]. Appendectomy performed >30 years ago reduces the incidence of PD by 19% [4]. These findings are in accordance with the notion that increased intestinal permeability is causally associated with PD. Epidemiological studies indicate that truncal vagotomy reduces a chance of developing PD by ∼50% [5]. Enteroendocrine cells have a direct synapse to vagal afferents [6], and the vagal nucleus has synaptic pathways to the SN and the striatum [7]. Animal models also indicate trans-vagal prion-like transmission of α-synuclein fibrils [8]. Additionally, intraperitoneal administration of LPS increases intestinal permeability and causes accumulation of α-synuclein fibrils in the intestinal mucosa and dorsal vagal nucleus [9]. Furthermore, LPS promotes the formation of α-synuclein fibrils in vitro. In a mouse model of PD, germ-free mice have milder motor symptoms and milder constipation compared to specific pathogen-free mice [10]. Transplantation of fecal microbiota obtained from PD patients exacerbates motor symptoms and constipation more than that from healthy individuals. Association between intestinal microbiota and PD has been analyzed in each country [11, 12]. However, divergence of intestinal microbiota even in healthy subjects across countries has made it difficult to identify intestinal bacteria that are commonly changed in PD in the world. We recently reported a nonparametric meta-analysis of intestinal microbiota in PD in 5 countries [13]. This review will focus on gut dysbiosis shared across countries in PD patients and also on possible mechanisms of how gut dysbiosis is causally associated with the onset and development of PD.

Because the compositions of the intestinal microbiota vary from country to country even in healthy individuals, gut dysbiosis observed in PD largely differ from report to report (Table 1). In addition, taxonomic identities are dependent on which database was used in the metagenome analysis and are different from report to report. Currently, SILVA is the most commonly used taxonomic database, but Greengenes, Ribosome Database Project (RDP), and NCBI are also sometimes used. We recently reported 16S ribosomal RNA (rRNA) gene sequencing of fecal samples from 223 PD patients and 137 controls and meta-analyzed our dataset with 4 previously reported datasets of PD patients and controls [13]. We used the same analysis pipeline with the SILVA database for the 5 datasets. We found that enterobacteria that were increased in PD at the genus level were Akkermansia and Catabacter and those that were decreased in PD at the genus level were Roseburia, Faecalibacterium, and Lachnospiraceae ND3007. At the family level, Lactobacillaceae and Akkermansiaceae were increased in PD. In addition, Akkermansia was increased with progression of PD, whereas Faecalibacterium and Roseburia were decreased with progression of PD in our own dataset. Recent reports [14-16] with 100 or more PD patients also show an increase of Akkermansia and decrease of Lachnospiraceae, Roseburia, and Faecalibacterium in PD. Lachnospiraceae, Roseburia, and Faecalibacterium produce short-chain fatty acids (SCFAs), especially butyric acid. In addition, Lactobacillaceae is often reported to be increased in PD. However, the increase of Lactobacillaceae was largely due to a confounding effect of catechol-O-methyltransferase (COMT) inhibitor and partly due to a confounding effect of constipation. Lactobacillaceae was not significantly different between PD without taking a COMT inhibitor and healthy subjects [13]. Metagenome analysis of 16S rRNA sequencing data gives us relative abundance of each bacterium but not an absolute bacterial count. We previously analyzed absolute counts of 19 representative intestinal taxa that covered 71.3% of total intestinal bacteria using quantitative reverse transcription PCR of 16S rRNA and reported that the total intestinal bacterial count was lower in PD compared to healthy individuals [17]. The total intestinal bacterial count was also decreased in 2 years in PD [18]. The absolute counts of total intestinal bacteria have been rarely reported, and the pathological significance of reduced bacterial counts remains elusive.

As more and more studies have been conducted on a large number of cases, methods such as network analysis and pseudo-functional analysis have been introduced in addition to simple statistical estimation. Pseudo-functional analysis can be performed using PICRUSt2 with the 16S rRNA seq dataset. PICRUSt2 enables us to predict functional annotations of intestinal microbiota using the Kyoto Encyclopedia of Genes and Genomes (KEGG) orthologs (KOs). In gene expression analysis, Gene Set Enrichment Analysis (GSEA) is commonly used as a de facto standard tool. In GSEA, we substituted PICRUSt2-generated KOs for gene expression profiles and named this method as KEGG Orthology Set Enrichment Analysis (KOSEA) [13]. KOSEA revealed that metabolisms of butanoate and propionate were changed in the intestinal microbiota in PD. This is in accordance with decreased abundance of SCFA-producing bacteria in PD. KOSEA also showed an increase in the degradation of valine, leucine, isoleucine, and lysine. Similar results were obtained using MetaCyc metabolic pathways [15]. Control microbiota was enriched in pathways related to carbohydrate degradation, whereas PD microbiota was enriched in pathways related to nucleic acid degradation and amino acid metabolism. Interestingly, PD microbiota was also enriched in 2 fucose degradation pathways, where fucose is derived from N-linked glycans decorating host mucin proteins. According to the network analysis using SparCC, 2 networks were observed in PD [13]. One was composed of increased genera in PD, such as genera Akkermansia and Christensenellaceae R-7 group. The other was composed of decreased genera in PD, such as genera Faecalibacterium, Roseburia, and Fusicatenibacter. The genus Christensenellaceae R-7 group was a hub of the increased subcluster and connected to the genus Akkermansia. Akkermansia and Christensenellaceae R-7 group might symbiotically play an important role in the development and/or progression of PD. Reduction of a network of SCFA-producing bacteria in PD was also reported by others [14, 16]. Based on these results, 2 major microbiota networks are changed in PD. When the dietary fibers are defective, Akkermansia muciniphila degrades the gut mucus layer and increases the intestinal permeability [19]. Faecalibacterium and Roseburia that are decreased in PD are butyrate-producing bacteria. Butyrate, as well as other molecules in SCFAs, induces the expression of anti-inflammatory cytokines by inhibiting histone deacetylase [20]. SCFAs also bind to G-protein-coupled receptors (e.g., GPR41, GPR43, and GPR109A) and exert an anti-inflammatory effect by increasing and/or activating regulatory T cells. The mucin-degrading Akkermansia and LPS-producing Gammaproteobacteria were increased in PD patients. Both dysbiosis should lead to increased intestinal permeability, motor impairment, nigral aggregation of α-synuclein fibrils, dopaminergic neuronal loss, and reduction in striatal dopamine (Fig. 1).

Rapid eye movement sleep behavior disorder (RBD) frequently predisposes to α-synucleinopathy including PD, dementia with Lewy bodies, and multiple system atrophy. RBD patients sometimes have subtle motor and cognitive deficits, as well as constipation, before the onset of PD and other α-synucleinopathies. RBD is the most dependable hallmark of prodromal PD. LIGER is a topic model-based tool that was developed for the analysis of single-cell RNA seq (scRNA-seq) data. In scRNA-seq analysis, a set of genes are up- or downregulated in specific cells, and LIGER identifies such gene sets. This is analogous to intestinal microbiota, in which a set of bacteria are increased or decreased in specific individuals. We thus exploited LIGER to make unsupervised clustering of intestinal microbiota in controls, RBD patients, and PD patients and found 4 enterotypes of A, B, C, and D (Fig. 2) [21]. Severity of PD is classified by Hoehn and Yahr disease-rating scales 1–5, where 1 is the mildest form and 5 is the severest form. The proportion of controls in each enterotype was decreased in the order of enterotypes A to D, while the proportions of Hoehn and Yahr scales 3–5 were increased in the same order. Analysis of individual bacteria taxa revealed that the genus Akkermansia was significantly increased in RBD compared to controls. In contrast, SCFA-producing genera Faecalibacterium, Roseburia, and Lachnospiraceae ND3007 group were not different between RBD and controls, although Faecalibacterium tended to be decreased in RBD. Thus, increased Akkermansia is likely to occur even in RBD, and decreased Faecalibacterium may be a hallmark to predict a transition from RBD to PD. We expect that administration of SCFAs and probiotic/prebiotic treatment to increase the production of SCFAs may be able to ameliorate the progression of α-synucleinopathy at the stage of RBD.

Intestinal bacteria produce SCFAs, as well as vitamins B2, B6, B12, and K, folic acid, pantothenic acid, biotin (B7), serotonin, and polyamines. Serotonin promotes intestinal peristalsis. Intestinal SCFAs are decreased in PD [17, 22]. Serum polyamines are also decreased in PD [23]. We will focus on the roles of SCFAs and polyamines in this communication.

The major products from the microbial fermentative activity in the gut are SCFAs, in particular, acetate, propionate, and butyrate. However, when fermentable fibers are in short supply, microbes switch to energetically less-favorable sources for growth such as amino acids from dietary or endogenous proteins or dietary fats. Protein fermentation can contribute to the SCFA pool but mostly gives rise to branched-chain fatty acids such as isobutyrate, 2-methylbutyrate, and isovalerate, which exclusively originate from branched-chain amino acids of valine, isoleucine, and leucine. The microbial conversion of dietary fiber in the gut results in the synthesis of the 3 major SCFAs, acetate, propionate, and butyrate (Fig. 3). Butyrate is a preferred energy source for colonocytes and is locally consumed, whereas other SCFAs drain into the portal vein. Propionate is metabolized in the liver and thus is present at low concentrations in the peripheral circulation, leaving acetate as the most abundant SCFA. Furthermore, acetate can cross the blood-brain barrier and reduce appetite via a central homeostatic mechanism. Intestinal acetate and propionate stimulate GPR41 and GPR43 to release PYY and GLP-1, which affect satiety and intestinal transit. Luminal butyrate exerts anti-inflammatory effects by binding to GPR109A and also by inhibiting histone deacetylase. Furthermore, propionate can be converted into glucose by intestinal gluconeogenesis, leading to satiety and decreased hepatic glucose production [24]. The immune-suppressive mechanisms are indispensable for intestinal homeostasis. Besides its effects on intestinal epithelial cells, butyrate can also modulate the activity of the enteric nervous system. GPR41 is widely expressed in the peripheral nervous system, such as sympathetic ganglia, as well as vagal, dorsal root, and trigeminal ganglia [25]. They can act at distal sites such as the brain, modulating permeability, neurogenesis, and behavior of the host. Decreased intestinal SCFA is thus likely to be sufficient to promote α-synuclein-mediated exacerbation of PD symptoms. Further characterization of the link between microbiota and microglia is expected.

Polyamines (PAs) are defined as small polycationic molecules with a wide range of biological functions, including gene regulation, stress tolerance, cell proliferation, and cell differentiation. Ingested food is the main source of PAs in the lumen, and the majority of these compounds are absorbed in the upper part of the intestine. Therefore, the main source of PA levels in the lower part of the gut should be gut microbiota [26]. PAs in the colonic lumen are transferred to the bloodstream through the colonic mucosa. The availability of cellular PAs contributes to tissue homeostasis, epithelial cell division, and apoptosis in the gastrointestinal mucosa by regulating the expression of various growth-related genes [27]. Putrescine and spermidine are the most common PAs in bacteria. As far as we know, no bacteria carry an enzyme to transduce spermidine to spermine. Little is known about the production and degradation of biogenic amines by the gut microbiota, especially PAs. The genus Bacteroides has a CASDC homolog that is essential for spermidine biosynthesis [28]. However, many colonic microbial species do not have a complete synthetic pathway for producing PAs [29]. Arginine is converted to ornithine and further to putrescine in bacteria. Arginine is also converted to agmatine and further to putrescine in the following 2 bacteria. The acid-tolerant system of E. coli produces agmatine from arginine in a low pH environment. Enterococcus faecalis metabolizes agmatine to yield putrescine via the agmatine/putrescine antiporter in the agmatine deiminase pathway [30]. The presence of other bacteria, such as Bifidobacterium spp. that produce acidity in the gut, is advantageous for putrescine production.

Our current knowledge on the association between intestinal microbiota and PD indicates 2 pathomechanisms in the PD intestine. First, the decrease of SCFA-producing bacteria and the increase of mucin-degrading Akkermansia are likely to increase the intestinal permeability to expose the intestinal neural plexus to toxins like lipopolysaccharide and pesticides, which should lead to abnormal aggregation of α-synuclein fibrils. The increased intestinal permeability is underscored by decreased serum LPS-binding protein in PD [17, 22]. Second, the decrease of SCFA-producing bacteria also aggravates microglia-mediated inflammation in the central nervous system. Therapeutic intervention to modify these intestinal maladies may be able to be established. Metagenomic analyses of intestinal microbiota using next-generation sequencers have accelerated the elucidation of taxonomic changes in PD. However, many of the intestinal bacteria are difficult to culture or remain uncultured. In particular, most of butyrate-producing bacteria cannot be cultured. In addition, even when the whole genome sequences of intestinal bacteria are available, gene annotations remain largely unidentified. It is critical to establish gene annotations of intestinal bacteria. Currently, KEGG is a dependable database for analyzing intestinal microbiota, but we expect that KEGG orthology becomes more extensive and more comprehensive. For example, it is currently unclear whether the presence of functional genes in bacteria leads to the extracellular secretion of metabolites. Given that bacteria consume energy to produce metabolites, most of the metabolites are predicted to be consumed intracellularly. Most of the metabolites secreted are no longer needed for the bacteria, or secreted, for example, for self-preservation of a low pH environment, as we observe in the acid-tolerant system of E. coli. Accumulating knowledge supports the notion that intestinal microbiota indeed affects the development and progression of PD. Further studies, however, are required to dissect the exact molecular mechanisms directly linking intestinal dysbiosis to PD.

The authors have no conflicts of interest to disclose.

This study was supported by Grants-in-Aid from the Japan Society for the Promotion of Science (JP17K07094, JP19K16516, and JP20H03561); the Ministry of Health, Labor, and Welfare of Japan (20FC1036); the Japan Agency for Medical Research and Development (20gm1010002, 20ek0109281, and 20bm0804005), the National Center of Neurology and Psychiatry (2–5); the Smoking Research Foundation; and the Hori Sciences and Arts Foundation.

M. Hirayama reviewed all abstracts and articles and drafted, revised, and approved the final manuscript. K. Ohno reviewed and approved the final manuscript. All authors read and approved the final manuscript.