Parkinson’s disease: Are gut microbes involved?

Abstract

Parkinson’s disease (PD) is a common neurodegenerative disorder characterized by motor and gastrointestinal (GI) deficits. Despite its prevalence, the pathophysiology of PD is not well understood. Recent studies highlight the role of gut microbiota in neurological disorders. In this review, we summarize the potential role of gut microbiota in the pathophysiology of PD. We first describe how gut microbiota can be influenced by factors predisposing individuals to PD, such as environmental toxins, aging, and host genetics. We then highlight the effect of gut microbiota on mechanisms implicated in the pathophysiology of PD, including disrupted microbiota gut brain axis (GBA), barrier dysfunction, and immune dysfunction. It is too early to connect the dots between gut microbiota and PD to establish causation, and experiments focused on investigating interrelationship between gut microbiota and associated metabolites on GBA, barrier dysfunction, and immune activation will be crucial to fill in the gaps.

INTRODUCTION

Parkinson’s disease (PD) is a commonly occurring progressive neurodegenerative disorder characterized by both motor (16) and nonmotor deficits, including olfactory deficits, cognitive decline, and gastrointestinal (GI) deficits (134). GI dysfunction is a prevalent nonmotor deficit in PD, which affects up to 60–80% of patient population and can be accompanied by alterations in the gut microbiota (1, 82, 102). The known role of gut microbiota alterations in GI dysfunction along with alterations described in PD patients has brought to attention the potential role of gut microbiota in pathophysiology of PD (1, 23, 41, 50). There is now accumulating evidence that links gut microbiota to central nervous system (CNS) diseases such as PD (17, 95, 123). In this review, we have examined in detail, the available evidence of GI dysfunctions and gut microbiota alterations in PD patients. Upon examining these studies, we believe that microbiota plays a central role in etiology of PD, primarily through mechanism that involve modulation of GI epithelial barrier integrity, immune function, and the microbiota gut-brain axis, which will be described in detail in the later part of the review (84, 108).

GI DYSFUNCTION IS PREVALENT IN PD PATIENTS

GI dysfunctions including abnormal salivation, nausea, dysphagia, altered gastric emptying, constipation, and defecatory dysfunction, are ubiquitous in PD patients (31). These GI abnormalities are a common nonmotor symptom of PD and may precede the onset of motor symptoms by several years, adding significantly to the healthcare burden and disrupted quality of life in PD patients (1, 82, 102, 119). The high prevalence (∼80%) and early onset of these GI dysfunctions in PD certainly raises a question regarding the role of GI tract in the pathogenesis of PD. In a large population-based study of more than 6,000 patients without PD that enrolled in the “Honolulu Heart Program,” Abott et al. (1) found that patients who suffered from constipation (i.e., <1 bowel movement/day) had fourfold greater risk of developing PD in the future. Subsequent autopsy on 245 of these constipated subjects, with no clinical signs of parkinsonism and dementia showed higher incidental Lewy body (α-synuclein aggregates) formation in the substantia nigra, suggesting a possible link between delayed GI transit and PD (2).

In addition to GI symptoms, physiological defects, such as disrupted intestinal permeability and altered immune activation, have also been observed in PD (see Fig. 1). Clairembault et al. (29) reported nearly 50% lower occludin expression in lysates of colon biopsies of PD patients, and Perez Pardo et al. (98) observed a reduction in the average intensity of ZO-1 immunolabeling in sigmoid colon biopsies from PD patients. Together, these studies indicate potential disruption of mechanisms that regulate epithelial paracellular permeability in PD patients. Schwiertz et al. (113) observed a simultaneous increase in markers of intestinal inflammation (fecal calprotectin) and disruption of the intestinal barrier (fecal zonulin and α-1-antitrypsin) in PD patients compared with age-matched controls. Since disruption in intestinal barrier function and intestinal inflammation are characteristics feature of inflammatory bowel disease (IBD), it is not surprising that three separate studies by Lin et al. (75), Weimers et al. (132), and Villumsen et al. (128) in distinct population cohorts all show an association between IBD and increased risk of PD later in life. Although it is tempting to speculate on the basis of these studies that risk of PD would be lower in patients that undergo colectomy, paucity of clinical data in IBD patients and controversial results in patients with other GI conditions make it difficult to evaluate whether this is, indeed, the case. Recently published studies that evaluated risk of PD in patients who underwent surgical procedures, such as colectomy and appendectomy for various GI conditions, including malignant neoplasms, benign tumors, noninfective inflammation, or disease of the appendix, found that the risk of developing PD increased in some cases, while decreased in others (66, 87, 120). These contradictory findings suggest that specific modulation of colonic factors, rather than full thickness colonic resection, is necessary to decrease risk of PD. Further studies are required to assess the role of colectomy in IBD patients and determine how colonic factors, such as microbiota, barrier function, and intestinal inflammation precisely play a role in PD pathophysiology. This will help us understand whether increased occurrence of GI-related physiological defects is a potential risk factor for PD or simply PD-related pathologies that precede motor symptoms.

Fig. 1.

Figure highlights gut microbiota-associated mechanistic pathways that regulate gastrointestinal and motor dysfunction in Parkinson’s disease. Environmental toxins consumption alters gut microbiota composition and leads to disruption in epithelial barrier (1), alteration in immune activation (2), and disrupted microbiota-gut-brain axis communication through vagal and nonvagal pathways (3) to cause gastrointestinal and motor dysfunction in Parkinson’s disease. TLRs, Toll-like receptors; ZO-1, zona occludins 1.

GUT MICROBIOTA AND ITS METABOLITES ARE ALTERED IN PD PATIENTS

Several studies have examined alteration in gut microbiota composition in PD patients, but as observed in other disease states, the results are heterogenous in terms of differences in specific taxa (Table 1). There are several factors that contribute to such variability, including lack of standardization in sample collection and sequencing techniques (V3-V4, V4, or V4 and V5), differences in study design, sample size, geographical diversity of patient population, and heterogeneous nature of PD (73, 103).

Table 1.

Gut microbiota alterations in Parkinson’s disease

Study Number	Study (Ref.)	Participants	Sample Used	Detection Method	Sequence Region	Overall Changes (α/β Diversity)	Increased in PD	Decreased in PD	GI Deficits?
1.	Aho et al. (2019) (5)	Control: 64 PD: 64	Stool	16S rRNA	V3-V4	α diversity no change β diversity no change	Bifidobacterium	Prevotella (G), Roseburia (G),
2.	Bedarf et al. (2017) (12)	Control: 28 PD: 31	Stool	Metagenomic shotgun analysis Illumina Hiseq4000		α diversity no change β diversity different	Verrucomicrobiaceae (F), Firmicutes (F), Akkermansia (G)	Prevotellaceae (F), Erysipelotrichaceae (F), Prevotella (G), Eubacterium (G)	No constipation
3.	Hasegawa et al. (2015) (49)	Control: 36 PD: 52	Stool	16S or 23S rRNA (qRT-PCR)			Lactobacillus (G)	Species: Clostridium coccoides group, Clostridium leptum subgroup, Bacteroides fragilis group
4.	Heintz-Buschart et al. (2018) (53)	Control: 78 PD: 76	Stool and nasal wash samples	16S and 18S rRNA	V4	α diversity no change β diversity different	Verrucomicrobia (P), Verrucomicrobia (O), Verrucomicrobiaceae (F), Akkermansia (G)
5.	Hill-Burns et al. (2017) (55)	Control: 130 PD: 197	Stool	16S rRNA		α diversity no change β diversity different	Bifidobacteriaceae (F), Lactobacillaceae (F), Tissierellaceae (F), Christensenellaceae (F), Verrucomicrobiaceae (F), Bifidobacterium (G), Lactobacillus (G), Akkermansia (G)	Lachnospiraceae (F), Pasteurellaceae (F), Blautia (G), Roseburia (G), Faecalibacterium (G)
6.	Hopfner et al. (2017) (56)	Control: 29 PD: 29	Stool	16S rRNA	V1, V2	α diversity no change β diversity different	Barnesiellaceae (F), Enterococcaceae (F), Lactobacillaceae (F)
7.	Keshavarzian et al. (2015) (63)	Control: 34 PD: 38	Sigmoidmucosal biopsies and stool samples	high‐throughput rRNA sequencing	V4	α diversity increase in PD β diversity different	Bacteroidetes (P), Proteobacteria (P), Verrucomicrobia (P), Clostridiaceae (F), Oscillospira (G), Akkermansia (G)	Firmicutes (P), Lachnospiraceae (F), Coprobacillaceae (F), Blautia (G), Coprococcus (G), Dorea (G), Roseburia (G)
8.	Li et al. (2017) (72)	Control: 14 PD: 24	Stool	16S rRNA	V3, V4, V5		Actinobacteria (P), Proteobacteria (P), Enterobacteriaceae (F), Streptococcaceae (F), Veillonellaceae (F), Acidaminococcus (G), Acinetobacter (G), Enterococcus (G), Escherichia-shigella (G), Megamonas (G), Megasphaera (G), proteus (G), Streptococcus (G)	Bacteroidetes (P), Pasteurellaceae (F), Blautia (G), Faecalibacterium (G), Ruminococcus (G)
9.	Li C et al., 2019 (71)	Control: 48 PD: 51	Stool	16S rRNA	V4	α diversity no change reduced β-diversity in PD patients	Akkermansia	Lactobacillus
10.	Lin et al. (2018) (73)	Control: 75 PD: 45	Stool	16S rRNA	V4	α diversity no change β diversity different	Eubacteriaceae (F), Bifidobacteriaceae (F), Aerococcaceae (F), Desulfovibrionaceae (F)	Tenericutes (P), Euryarchaeota (P), Firmicutes (P), Streptococcaceae (F), Methylobacteriaceae (F), Comamonadaceae (F), Halomonadaceae (F), Hyphomonadaceae (F), Brucellaceae (F), Xanthomonadaceae (F), Lachnospiraceae (F), Actinomycetaceae (F), Sphingomonadaceae (F), Pasteurellaceae (F), Micrococcaceae (F), Intrasporangiaceae (F), Methanobacteriaceae (F), Idiomarinaceae (F), Brevibacteriaceae (F), Gemellaceae (F)	Constipation
11.	Lin et al. 2019 (74)	Control: 77 PD: 80	Stool	16S rRNA	V3-V4		Parabacteroides, Verrucomicrobia (P), Akkermansia, Butyricimonas, Veillonella, Odoribacter, Mucispirillum, Bilophila, Enterococcus, and Lactobacillus	Prevotella (G)	Constipation
12.	Petrov et al. (2017) (99)	Control: 66 PD: 89	Stool	16S rRNA	V3-V4	α diversity decrease β diversity different	Christensenella (G), Catabacter (G), Lactobacillus (G), Oscillospira (G), Bifidobacterium (G)	Dorea (G), Bacteroides (G), Prevotella (G), Faecalibacterium (G)
13.	Qian et al. (2018) (104)	Control: 45 PD: 45	Stool	16S rRNA	V3-V4	α diversity increase β diversity different	Clostridium IV (G), Aquabacterium (G), Holdemania (G), Sphingomonas (G), Clostridium XVIII (G), Butyricicoccus (G), Anaerotruncus (G)		Constipation
14.	Scheperjans et al. (2015) (112)	Control: 72, PD: 72	Stool	16S rRNA	V1-V3	α diversity no change β diversity different	Lactobacillaceae (F), Verrucomicrobiaceae (F), Bradyrhizobiaceae (F), Clostridiales (F), Incertae Sedis IV (F)	Prevotellaceae (F)	Constipation
15.	Unger et al. (2016) (126)	Control: 34 PD: 34	Stool	qPCR using bacterial primers			Enterobacteriaceae (F), Bifidobacterium (G)	Bacteroidetes (P), Prevotellaceae (F, descriptively reduced), Lactobacillaceae (F), Enterococcaceae, Species: Faecalibacterium prausnitzii	Constipation

PD, Parkinson’s disease; qPCR, quantitative PCR.

A consistent finding among studies has been an increase in members of Enterobacteriaceae family and Helicobacter spp. (1.5–3-fold increase in H. pylori in gastric biopsies) (20, 85), as well as an increase in Lactobacillus and Akkermansia (see 18, 20, 27, 43, 57, 99, 112). This is particularly interesting as an increased abundance in members of Enterobacteriaceae family and H. pylori has been associated to severity of postural instability, gait difficulty, and overall worsening of motor functions in PD (57, 107, 112). Furthermore, eradication of H. pylori has been shown to improve absorption of levodopa resulting in better response of motor function to treatment (85). Increase in members of Lactobacillaceae has been associated with reduced concentrations of ghrelin in PD; a gut brain peptide implicated in upregulation of intestinal tight junction proteins (28), stimulation of intestinal motility (115), and maintenance and protection of normal nigrostriatal dopamine function (7, 125). Akkermansia on the other hand has been shown to use mucus as a carbon source and degrade the colonic mucus barrier (42), which might disrupt intestinal permeability and increase pathogen susceptibility causing inflammatory condition on the intestinal wall (114). In addition to the above gut microbial changes, PD patients also display reduced abundance of prominent short-chain fatty acid producers. Strains of Faecalibacterium spp., Blautia spp., Prevotella spp, and Roseburia spp, that are common butyrate producers are more consistently reduced in PD patients (126). Butyrate has been shown to exert barrier protective, anti-inflammatory role in the GI tract (8, 97) and prevent dopaminergic neuron degeneration and subsequent bradykinesia in mice (78). Together, the observed changes in gut microbiota in PD and their association with motor and nonmotor deficits, suggests that altered gut microbiota may play a role in PD (27, 57, 58, 85, 112).

In an attempt to identify potential link between gut microbiota and PD, we will next examine factors that contribute to alterations in gut microbiota, and then discuss how gut microbiota changes could contribute to PD.

FACTORS THAT CONTRIBUTE TO GUT MICROBIOTA ALTERATIONS IN PD

Host Genetics

Host genetics differences have often been linked to differences in microbiome composition (46, 143). Using next generation sequencing approach to analyze fecal bacterial composition (V4 region of 16S rDNA), Goodrich et al. (46) showed that microbiota are more similar within twin pairs compared with unrelated individuals. Among various microbial taxa, studies show that taxa particularly from Firmicutes and Proteobacteria, including members of Christensenellaceae family are the most heritable taxon in humans (25, 45, 131). Several genome-wide association studies or quantitative trait locus (QTL) mapping studies have also successfully identified specific host genes and single nucleotide polymorphisms in the host genome that contribute to the variation of these microbial taxa (21, 32, 124). These studies show that host genetics plays a significant role on determining dominant gut microbiota composition.

In addition to assessing the influence of host genetics on gut microbiota, studies have also been separately conducted to assess the role of host genetics in etiology of PD. Studies suggest that PD is modestly heritable with heritability estimate of around 30% (133). Higher percentage concordance for subclinical striatal dopaminergic dysfunction in sporadic and late-onset PD suggest that host genetics particularly contribute to striatal dopaminergic dysfunction, a primary cause for PD (101). Although these studies individually show that host genetics plays an important role in regulating gut microbiota composition and the pathology of the PD, it remains unclear what precise genetic changes in PD help drive specific changes in microbiota and whether these genetic changes induced microbiota alteration drive PD phenotype.

Earlier studies have successfully associated specific genetic changes in GI diseases to change in specific microbial species (45, 67, 116). However, whether such association between individual genes and gut microbiota exists in PD is not known. In a minority of patients (10–15% of clinically diagnosed cases) where PD is inheritable (Familial PD), mutation in one of the eight genes responsible for aberrant protein formation and disrupted mitochondrial homeostasis are often responsible. They include SNCA (also known as PARK1, a gene that encodes for α-synuclein, a major component of Lewy bodies), PARK2 [also known as PRKN; parkin, a ubiquitin protein ligase involved in the degradation of abnormal proteins by the proteasome (86)], PARK5 (UCHL1; ubiquitin carboxy-terminal hydrolase L1, a gene responsible for generating the ubiquitin monomer), PARK6 (PINK1; encodes for putative serine-threonine kinase), PARK7 (also known as DJ-1; encodes for protein/nucleic acid deglycase DJ-1), PARK8 (LRRK2; encodes for Leucine-rich repeat kinase 2 enzyme), PARK11 (also known as GIGYF2; encodes for GRB10-interacting GYF protein-2), and NR4A2 (encodes for nuclear receptor subfamily 4, group A, member 2 protein) (79a). Recent studies have shown that specific genetic changes in the absence of gut microbiota cannot fully recapitulate the symptom severity seen in PD (3, 110). Using a α-synuclein (SNCA), a presynaptic nerve terminal protein, overexpressing germ-free and conventionally raised specific pathogen-free mice, Sampson et al. (110) showed that germ-free α-synuclein overexpressing mice did not exhibit PD phenotype compared with conventionally raised mice. Given that exposure to bacterial endotoxin, LPS is required to trigger persistent neuroinflammation and generate structurally distinct α-synuclein fibril structure involved in synucleinopathies and neurodegeneration (37, 51, 64), it seems likely that specific genetic mutation though necessary, might not be sufficient and potentially requires gut microbiota interaction to regulate symptom severity and disease outcome in genetic model of PD.

Environmental Toxins

Environmental toxins from different chemical classes can alter both the microbial composition and the metabolic activity of the gut bacteria (reviewed in Refs. 30 and 80). A majority of PD cases are idiopathic and likely caused by exposure to environmental toxins, including herbicides and pesticides, such as rotenone, paraquat, organic solvents, and heavy metals, including vanadium and manganese (4, 70, 122). A recent meta-analysis of data that investigated prospective cohort and case-controlled epidemiological studies found that PD was associated with farming, and the risk of developing PD was increased by exposure to environmental toxins (100). Recent studies have hypothesized that an initial target of these environmental toxin consumption/exposure in PD could be the gut microbiota (55, 137). Hill-Burns et al. (55) found that the gut bacteria responsible for degrading some of these environmental toxins, such as atrazine (herbicide), and naphthalene (insect repellent), are altered in individuals with PD compared with patients that were prescribed PDmedication. Byperforming longitudinal analysis of 16S rRNA gene sequencing of fecal microbiome, Yang et al. (137) showed that three weeks of chronic rotenone administration, at a dose found commonly in pesticides, caused fecal microbiota alterations, mitochondrial disruption, and dopaminergic neuronal loss, along with behavioral and neuropathological features of PD. A direct association between environmental toxin and PD was also shown in a case-controlled prospective study, which assessed the health effects of lifetime exposure to rotenone and paraquat in farmers and found that exposure to these pesticides increases the risk of PD by 2.5-fold (122). Together, these data suggest that common environmental toxins and pesticides lead to behavioral and neuropathological symptoms of PD, via mechanisms that potentially involve gut microbiota alterations and disruption in mitochondrial activity, a cellular organelle that is also evolutionarily a descendant of endosymbiotic bacteria (130). Characteristic microbiota changes after rotenone administration include decreased bacterial diversity and increase in Firmicutes/Bacteroidetes ratio (137), a dysbiotic condition also observed in several other diseases, including colorectal cancer (9), hypertension (136), Type 2 diabetes (T2D), obesity (68), and inflammatory bowel disease (IBD) (9). The similarity in observed microbial changes postrotenone administration with various other diseases suggests that these gut microbial changes that occur in PD are common microbial marker of various chronic diseases. This observation raises two important questions: 1) how do microbiota change after toxin exposure specifically translate to PD, as opposed to other chronic diseases, and 2) is reversing these microbial changes sufficient to reverse the course of PD pathogenesis after toxin administration?

Although these studies certainly suggest that chronic microbiota changes after toxin exposure might be involved in idiopathic PD, future studies using gnotobiotic mice model is necessary to enhance our current mechanistic understanding regarding how gut microbes are able to transform/metabolize different classes of environmental toxins and xenobiotics and alter their pharmacokinetic and pharmacodynamic properties to cause PD. It is also important to carefully study the effects of environmental toxins associated with PD in both conventional and gnotobiotic germ-free animal models to understand the relationship between environmental toxin and microbiota and use this knowledge to open new avenues of treatment options for PD patients.

Aging

Besides, environmental toxins and genetics, aging is the most important factor that contributes to alterations in gut microbiota and PD. Recent studies have indicated clear differences in gut microbiota composition among infants, toddlers, adults, and the elderly (18, 91). Using high-throughput sequencing of the 16S rRNA gene (amplicons derived from the V3-V4 region), Odamaki et al. (91) investigated the sequential changes in fecal microbiota composition samples from 367 healthy Japanese subjects between the ages of 0 and 104 yr. They found that the transition from infant to centenarian was accompanied by distinctive coabundance group (CAG; identifies species that are phylogenetically and/or functionally related on the basis of gene abundance) dominance at different stages of life. Significant abundance of Bifidobacterium coabundance groups (CAGs) was observed in infants and children; Lachnospiraceae CAGs were observed in adults; Eubacterium and Clostridiaceae CAGs were observed in the elderly, and Enterobacteriaceae CAGs were observed in both infants and the elderly (91). In addition to Odamaki et al. (91), Biagi et al. (18) from Northern Italy also consistently observed an increase in the relative abundance of Proteobacteria, a major phylum of facultative anaerobic gram-negative bacteria, which include several members, including Escherichia, Pseudomonas, Salmonella, Vibrio, Helicobacter, Yersinia, and Legionellales in subjects that were over 70 yr of age. Although it is still not clear whether these aging-associated changes in gut microbiota increase susceptibility to PD, these studies do, however, suggest that changes in bacterial composition with age consistently favor an increase in pathogenic species with age irrespective of geographical location. Pathogenic species from Enterobacteriaceae CAGs and from Proteobacteria phylum, such as Helicobacter, have been associated with neurodegenerative diseases, including PD (summarized in review in Ref. 43). Future studies need to investigate the contribution of aging-associated factors such as dietary modification, life style changes, and weakened immune function on gut microbiota that could lead to changes in gut microbiota composition. This will help determine whether it is “normal” age-related microbial changes seen in a “healthy” individual or whether it is external factors that promote microbial changes at an older age, which increases PD susceptibility in aging individuals.

MECHANISMS THAT LINK ALTERED GUT MICROBIOTA TO PD

Disrupted Barrier Function

The GI epithelial barrier comprises a thick mucus layer aloft a monolayer of epithelial cells that are interconnected through a system of junction proteins (Fig. 1). The proper maintenance of epithelial barrier integrity is crucial to prevent translocation of pathogenic luminal bacteria and bacterial metabolites that acts as a mediator of diseases, all the while providing regulatory signals to induce immune tolerance toward commensal microbes. Studies show that gut microbes play an important role in maintaining the epithelial barrier, particularly through butyrate production, which modulates expression of mucin-associated genes in goblet cells and regulates expression and distribution of epithelial tight junction proteins (54). Thus, it is not surprising that altered gut microbial composition, as observed in various GI and CNS diseases, including IBD (44), stress-related psychiatric disorders (61), autoimmune disorders (33), and PD (109) is more often accompanied by disrupted barrier function and increased mucosal colonization of adherent and invasive, pathogenic species in many of these diseases (36, 138).

Although a disrupted intestinal barrier is a common denominator implicated in various GI and CNS diseases besides PD, the specific changes in gut microbiota and the microbial pathways that give rise to disrupted intestinal permeability needs to be investigated in the specific context of a PD. Gut microbiota changes as seen in PD can disrupt epithelial permeability through one of three ways, which include, increase in microbial enterotoxin production, increased luminal endotoxin expression, and decreased butyrate production (19). Helicobacter pylori, a gram-negative pathogen, more commonly found in PD patients and aged individuals, for example, produces enterotoxin, such as the vacuolating cytotoxin (VacA), and cytotoxin-associated gene A (CagA) and directly releases them on to the host epithelium to modulate expression of tight junction proteins, including claudin-1 and ZO-1 and disrupt intestinal permeability (35, 39, 85). Additionally, increased mucosal exposure to bacterial endotoxin (lipopolysaccharide, LPS) as observed in early “Hoehn & Yahr Stage II” PD patients (36) could also cause epithelial barrier disruption by downregulating tight junction occludin and ZO-1 mRNA expression (13), and by directly inducing phosphorylation of the 20-kDa myosin light chain (MLC20) (140). Finally, loss of butyrate producing probiotic strain, as seen in PD microbiota, could lead to disruption in epithelial barrier permeability through decreased mucus production and secretion, and by impairing proper localization of tight junction proteins assembly (76, 97, 106).

Although it is becoming increasingly clear that altered gut microbiota can contribute to disrupted intestinal barrier permeability in PD through the mechanisms discussed above, how this leads to PD is not well understood. Recent studies show that disrupted intestinal permeability strongly correlates to increased luminal endotoxin translocation and expression of α-synuclein aggregates in the intestine of both humans and a mouse model of PD (36, 62). This is particularly interesting given the biophysical characteristics of α-synuclein as an antimicrobial peptide within the GI tract (10, 96). Overexpression of α-synuclein as a result of microbial infection or endotoxin induced epithelial damage has also been hypothesized as potential mechanism which leads to formation of α-synuclein aggregates in the ENS and subsequent α-synuclein trafficking to the CNS (10, 62). Future studies are necessary to understand the underlying molecular mechanisms of α-synuclein aggregation after barrier dysfunction, as well as its role on gut microbiota, and test whether prevention of barrier disruption through manipulation of gut microbes is sufficient to prevent gut α-synuclein pathology and thwart PD progression.

Altered Immune Activation

The GI tract holds the largest number of immune cells in the body, which includes the intestinal epithelial cells, macrophages, dendritic cells, B cells, and regulatory T cells. Intestinal epithelial cells, in particular, not only form a physical barrier for the gut microbes but are also dynamically active immune cells (83) that express pattern recognition receptors on the epithelial surface. Subsets of intestinal epithelial cells, especially the Paneth cells and microfold (M) cells, sample the microbe-laden luminal environment by recognizing conserved bacterial motifs and participate in downstream immune regulation by producing relevant antimicrobial peptides and proinflammatory cytokines, and by delivering bacterial antigens to dendritic cells (79, 92). Gut microbial interaction with the intestinal epithelial cells is, therefore, a crucial contributor to immune development and a potent regulator of homeostatic immune response (83).

A mechanistic example of gut microbiota influencing host immunity to induce PD phenotype is through elevated gram negative bacteria expressing LPS in PD patients (47), which acts on the Toll-like receptors (TLRs) present in the epithelial, immune, and nerve cells of the GI tract to modulate innate immune response (79). Dysregulation of LPS-TLR signaling has, thus, been simultaneously implicated in both intestinal inflammation and PD pathogenesis. Perez-Pardo et al. (98), using TLR4-knockout (KO) mice, observed that while rotenone causes dysbiosis in both TLR4-KO and wild-type (WT) mice, TLR4-KO mice had decreased intestinal and motor dysfunction, lower intestinal inflammation, and motor neuron degeneration relative to WT TLR4-expressing mice. These data show that peripheral immune function regulated, at least in part, by LPS-TLR signaling drives altered immune activation and motor neuron degeneration in toxin-induced PD mice. In humans, a direct link between serum LPS and PD was observed in a 22-yr-old laboratory worker who developed Parkinson’s syndrome (exhibiting cardinal signs: bradykinesia, rigidity, and tremor) 3 wk after accidental exposure to Salmonella minnesota LPS through an open wound (90). The proof that LPS exposure could directly lead to parkinsonism in an otherwise healthy individual reasserted the extraordinary ability of microbial endotoxin to directly induce PD. Since the symptom of neuroinflammation that is associated with upregulation of LPS-induced disrupted TLR signaling is often seen in an animal model of PD, it is plausible that altered immune activation as a result of disrupted TLR signaling likely contributes to development of PD phenotype in an LPS-exposed patient (14).

Bacterial LPS-induced immune activation could lead to nigral dopaminergic neuron loss and Parkinsonism by causing disruption of blood-brain barrier function and increasing peripheral α-synuclein translocation into the brain (105, 117). Increased α-synuclein translocation may further exacerbate immune activation by inciting the production of proinflammatory molecules and increasing the expression of TLRs, eventually causing neurodegeneration and motor dysfunction (14, 117). How exactly disrupted LPS-TLR signaling alters blood-brain barrier function remains to be fully understood. According to one hypothesis, the disruption in blood-brain barrier permeability in PD might occur through MyD88-immune pathway activation, which results in increased proinflammatory cytokine production (89) and potentially cause destruction of tight junctions of microvascular endothelial cells that form the blood-brain barrier (139). This pathway of blood-brain barrier disruption through microbiota-dependent immune modulation is supported by evidence from both living PD individuals and postmortem studies, which not only demonstrate elevated levels of proinflammatory cytokines, including TNF-α, IFNγ, IL-1β, and IL-8 in parkinsonian brain tissue, cerebrospinal fluid (reviewed in 26) and plasma, but also directly correlate these increases in proinflammatory cytokines to changes in gut microbiota (74).

Although most studies have focused on how environmental toxin-dependent changes in “healthy” microbial community drives alterations in microbe-epithelial cell interaction and disrupted homeostatic immune balance in PD, it is also plausible that chronic consumption of environmental toxins, and not the gut microbiota, directly drives disruption of homeostatic immune balance, which could then alter gut microbiota (69, 141). Chronic rotenone administration in rats, for example, has been shown to increase the nitric oxide (NO) levels within the CNS and increase production of 3-nitrotyrosine, a biomarker for endogenous peroxynitrite-induced oxidative stress activity, ultimately causing damage to nigrostriatal motor pathway (6, 11, 52, 135). Because NO is an important mediator of immune homeostasis and host defense against various GI pathogens, altered NO production and its downstream metabolites as a result of chronic toxin consumption can directly contribute to alterations in gut microbiota and pathological states. A better understanding of how disrupted immune activation translates to altered gut microbiota, neurodegeneration, and motor dysfunction in PD is also an active area of research. Future studies using gnotobiotic mice is crucial to conclusively resolve this uncertainty.

Microbiota-Gut-Brain Axis

Disruption in gut-brain axis has been previously associated to pathogenesis of several CNS and the peripheral nervous system disorders ranging from anxiety, autism, and depression to IBS and IBD. While the gut-brain axis is often viewed as an information highway that mediates bidirectional communication between the brain and the gut, it is now well known that this bidirectional communication is markedly influenced by the gut microbiota and their metabolites through neurological and neuroendocrine pathways.

Evidence shows that a major neurological pathway via which PD pathogenesis once initiated in the gut ascends to the brain is through the vagal route along the gut-brain axis. The evidence in support of this route was observed in a mice study, which showed that intragastrically administered rotenone causes progressive accumulation and propagation of α-synuclein from the ENS to the CNS in a “prion-like” fashion (93, 94). Another recent study also observed that a direct injection of pathologic α-synuclein oligomers into the GI tract of mice causes spread of α-synucleinopathy into the CNS, which was preventable by truncal vagotomy (65). This pattern of pathological staging in mice mimics similar pattern of α-synuclein progression seen in post mortem PD patients (22, 93) and potentially explains why progression of motor symptoms, but not GI symptoms, can be delayed by partial sympathectomy and hemi-vagotomy, and also why full truncal vagotomy in humans could be protective against PD (77).

Gut microbiota could regulate the bidirectional vagus nerve communication by directly producing neurotransmitter such as serotonin, norepinephrine, GABA and dopamine (40, 88, 127, 129, 142) and affecting neurotransmitter receptor expression in the brain and in the intestinal lumen (24, 59). A study shows that subdiaphragmatic vagotomy in turn could prevent some of these microbiota dependent effects on vagal excitability and disrupt PD progression (24). In addition to directly modulating vagus nerve activity through disrupted neurotransmitter production and altered receptor expression, gut microbes could also affect the vagus nerve indirectly by inducing release of gut peptides, and hormones from enteroendocrine epithelial cells. Enteroendocrine epithelial cells in response to microbial activation releases effector molecules such as cholecystokinin (CCK), glucagon like peptide-1 (GLP-1), PYY and ghrelin on to the vagal afferent fibers to modulate excitability and central motor behavior (111). These data suggest that vagus nerve represents a crucial route for gut brain cross talk in PD.

While gut microbiota is capable of driving CNS dysfunction through the vagal pathway, it is also important to note that gut microbes can modulate CNS function independent of vagus by releasing SCFAs and stimulating the release of luminal peptides and gut hormones, such as VIP, secretin, ghrelin and GLP-1 from the epithelial cells into the systemic circulation (15). These luminal signals can directly strengthen the blood brain barrier to prevent toxin translocation in PD (78). Some of these peptides can also passively cross the blood brain barrier to control various peripheral and central physiologic functions such as mucosal defense, inflammatory response and maintenance and protection of nigrostriatal dopamine function (7).

MICROBIOTA AS A THERAPEUTIC TARGET IN PD

Given the role of gut microbiota in pathogenesis of PD, recent studies are beginning to identify gut microbiota as a potential therapeutic target to treat PD. Limited clinical and preclinical studies suggest that three potential microbiota-dependent therapeutic approaches could potentially be employed to treat PD. First, drugs could be developed to prevent microbial uptake of l-DOPA and degradation within GI tract. l-DOPA, a commonly prescribed dopamine precursor drug is extensively metabolized in the gut by aromatic amino acid decarboxylase (AADC) and by gut bacteria that express PLP-dependent tyrosine decarboxylase. Significant degradation of l-DOPA within the gut reduces l-DOPA bioavailability for uptake to the brain, necessitating a higher dose of drug administration for effective PD treatment, which often leads to side effects such as l-DOPA-induced motor fluctuations and dyskinesia (38, 81, 127). Targeted depletion of these gut bacteria or prevention of l-DOPA uptake by the gut bacteria through administration of drugs such as tyrosine mimic (S)-α-fluoromethyltyrosine (AFMT), along with AADC inhibitor are, therefore, likely to significantly improve drug efficacy and prevent motor side effects (81). Second, probiotics could be used as a potential therapeutic in PD. In a randomized, double-blind, placebo-controlled clinical trial, Tamtaji et al. (121) found that PD patient consuming probiotic product containing Lactobacillus and Bifidobacterium species for 12 wk had significantly decreased movement disorder, as measured by Movement Disorders Society-Unified Parkinson’s Disease Rating Scale compared with placebo group. While the mechanism via which these probiotics causes improved motor function in PD patients is not completely understood, a recent study suggests that other probiotics, for example, Bacillus subtilis, exert the therapeutic effect by inhibiting and reversing α-synuclein aggregation (48). Third, fecal microbiota transplantation (FMT) could be employed as a potential therapeutic in treating PD. This approach is based on the evidence from animal studies that show that FMT from PD patients to α-synuclein-overexpressing mice exacerbates physical impairments (110), and microbiota transplants from healthy mice is protective against PD in 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-induced neurotoxic mice model (118). Clinical trials (see NCT03808389) are now being conducted in early (Hoehn & Yahr score 2–3) PD patients to test whether restoration of gut microbiota by FMT helps prevent the development and progression of PD in humans. While FMT-derived therapies has garnered high success and attention in C. difficile infection, IBD, and autism, FMT therapies for PD are at a very early stages, and benefits of FMT must be weighed against ethical issues, short-term and long-term safety, and psychological consequences before it can be used in PD. While it is still very early to determine which of these therapeutic options present a greater potential against PD moving forward, it is critically important to properly assess individual needs, patient’s health, including stage of PD pathology, and the results of clinical trials to determine what microbial therapeutic strategy might represent a better option for a PD patient.

CONCLUSIONS

The observation that GI dysfunction precedes sensory motor impairment and continues to affect PD patients during the course of disease progression highlights the potential importance of GI tract in initiation and progression of PD pathogenesis. Although it is well established that gut microbes are altered during the course of PD pathogenesis, the considerable debate in the field has been whether altered gut microbiota is a driving force in neurodegeneration or simply represents a response to environmental toxin. The fact that microbial changes seen in PD are often heterogenous makes it very complicated to understand whether microbial dysbiosis precede or succeed GI dysfunction in PD. Future studies using gnotobiotic mice will be crucial to decipher the complicated relationship between gut microbiota and PD that will help us identify microbiota-dependent mechanisms that contribute to PD pathogenesis and devise potential therapeutic strategies.

DISCLOSURES

P. Kashyap is a member of the Advisory Board of Novome and a consultant for Otsuka Pharmaceuticals, Pendulum Therapeutics, and IP Group, Inc. Y. Bhattarai is a scientist at Takeda Pharmaceuticals.

AUTHOR CONTRIBUTIONS

Y.B. prepared figures; Y.B. drafted manuscript; Y.B. and P.C.K. edited and revised manuscript; Y.B. and P.C.K. approved final version of manuscript.