Table 3.

A summary of human and animal studies on gut microbiota dysbiosis and PD.

Human Studies
Ref	Demographic	Test vs control group	Taxa decreased in PD	Taxa increased in PD	Alpha and Beta diversity	Association of microbiome changes	Functional prediction metagenomics	Comments
Hill-Burns et. al. [88]	USA	N = 197 PD patients N = 130 healthy controls of which 54 were spouses	Lachnospiraceae, Pasteurellaceae, Verrocomicrobiaceae	Bifidobacteriaceae, Lactobacillaceae, Tissierellaceae, Christensenallaceae	Beta diversity differed in weighted and unweighted UniFrac.	PD medications Catechol-O-methyl transferase inhibitors, and anticholinergics (CarbiDopa and LevoDopa to a small extent) showed association with differences. Increased abundance of Ruminococcaceae was seen with disease duration.	Metabolism of plant-derived compounds and xenobiotics degradation.	- Good sample size and strong statistical analysis. - Compares the data with four previous studies with conflicting results. - Confirms some of the reported associations.
Barichella et. al., [89]	Italy	N = 350 including de novo drug naïve PD patients and other atypical Parkinsonisms N = 113 healthy controls including spouses, community controls and hospital controls	Lachnospiraceae	Lactobacillacea, Christensenalliceae	Alpha diversity - PD patients showed higher richness than controls. Beta diversity - both weighted and unweighted UniFrac were significant in PD patients vs controls.	-These differences were associated with higher cognitive impairment, gait disturbances, and postural instability. -Influence of drug use (catechol-o-methyltransferase inhibitors, proton pump inhibitors), fiber intake, and milk formula feeding on abundance of some of taxa was seen.	Pathways involved in metabolism of amino acids, energy, glycans, lipids, cofactors, and vitamins. Increased ubiquitin processing and upregulation of xenobiotic degradation pathways.	- Extensive characterization of PD patients was done and have controlled for confounders. - The study investigates and highlights some differences in microbiome with atypical Parkinsonisms. -Their focus on de novo PD patients after adequate follow-up provides relevant insight.
Pietrucci et. al. [90]	Central Italy	N = 80 idopathic PD for longer than 6 months N = 72 healthy controls partners and caregivers	Lachnospiraceae	Lactobacillaceae, Enterobacteriaceae and Enterococcaceae families & Entrococcus, Lactococcus, Klebsiella at the level of genera	Alpha diversity not significantly different for any of the indices. Beta diversity – differed between cases and controls for all the 4 matrices tested, with PD status as significant predictor of gut microbiota composition.	Lower levels of Lachnospiraceae and higher levels of Enterobacteriaceae families also correlated with increased disease severity and motor impairment. Abundance of some of the genera was influenced by drug use (Catechol-O-methyltransferase inhibitors)	Variance in genes involved in metabolism of short chain fatty acids and amino acids, and in lipopolysaccharide biosynthesis. Suggest changes in pathways favoring a pro-inflammatory environment in the gastrointestinal tract, and a reduction in the biosynthesis of amino acids acting as precursors of physiological transmitters.	- Recruited patients had variable degree of PD and were receiving variable treatments. - Patients’ dietary intake was variable, which may influence microbiome composition. - Whether inflammation causes alterations in microbiome composition or vice versa is still unknown.
Qian et. al. [91]	China	N = 45 idiopathic PD patients N = 45 healthy spouses of the patients	Lactobacillus, Sediminibacterium	Alistipes, Paraprevotella, Klebesiella, Sphingomonas, Acinetobacter, Aquabacterium, Desulfovibrio, Clostridium IV, Lachnospiracea incertae sedis, Butyricicoccus, Clostridium XVIII and Nitrososphaera (Kingdom Archaea)	Alpha diversity - significantly higher in PD than controls. Beta diversity significantly different in the unweighted but not weighted UniFrac between PD and healthy group.	Correlation between fecal microbiota with disease duration, severity, and medication (LevoDopa equivalent doses, LED) was confirmed with most genera being negatively correlated. Genera Escherichia/Shigella were negatively associated with disease duration. Genera Dorea and Phascolarctobacterium were negatively associated with medication (LED)	Genes involved in metabolism of cofactors and vitamins, porphyrin and chlorophyll metabolism, and biotin metabolism, were higher in the fecal microbiome of the healthy group, whereas the microbial gene functions related to energy metabolism, flavone and flavonol biosynthesis, fatty acid biosynthesis, and apoptosis pathways were higher in the fecal microbiome of the PD group.	-Merits: well-matched control population, identification of several new taxa at genera level. -Demerits: small sample size, cross-sectional study.
Heintz-Buschart et. al.[92]	Germany	N = 76 PD (DeNoPa cohort) + 21 Rapid Eye movement disorder N = 78 HC (DeNoPa cohort)	Melainabacteria	Akkermansia, (Verrucomicrobiaceae) Prevotella	–	Within the PD cohort: Anaerotruncus spp., Clostridium cluster XIVa, and Lachnospiracea were related to motor symptoms, and Anaerotruncus, Akkermansia, and other (unclassified) were related to non-motor symptoms	–	- Paper supports that sample collection and preservation should be a standardized method to avoid variabilities.
Unger et. al.[123]	Germany	N = 34 PD patients N = 34 age-matched controls	Bacteroidetes, Prevotellaceae, Faecalibacterium prausnitii, Lactobacillacea and Enterococcaciae	Enterobacteriaceae, Bifidobacterium	–	Drug (Entacapone) use: Changes in abundance of Firmicutes and Faecalibacterium prausnitzii	–	- Small sample size. - Preliminary data on young controls.
Keshavarzian et. al., 2015 (96)	USA	N = 38 PD (of which 12 drug naïve) N = 34 HCvolunteers	At phylum level- Firmicutes At genus level - Roseburia, Blautia, Coprococcus	At phylum level - Bacteroidetes, Proteobacteria and Verrucomicrobia At genus level - Akkermansia, Oscillospira, Bacteroides	Alpha diversity - significant differences were observed at phyla level and significant difference in richness at genera level.	PD duration - Bacteroidetes and Proteobacteria positively correlated. Firmicutes negatively correlated. Significant negative correlation with Lachnospiraceae family.	LPS biosynthesis pathway, ubiquinone, and other terpenoid-quinone biosynthesis pathways were significantly upregulated in PD. Metabolic and biosynthetic pathways downregulated in PD.	- Compares both the sigmoidal mucosa-associated microbiome and the fecal microbiome in PD and controls thus providing more insight into microbiome associated with different locations in the GI tract. -Controls not well-matched.
Bedarf et. al.[126]	Germany	N = 31 (short duration, l-DOPA naïve) N = 28 age-matched and general demographics matched non-parkinsonian healthy controls.	Prevotella copri, Eubacterium biforme, Clostridium saccarolyticum	Akkermansia muciniphila, Alistipes shahii	No difference in alpha or beta diversity.	Slight trend was seen with symptom severity and Eubacteria otherwise no significant associations with clinical data	Decreased Gucouronate degradation and higher tryptophan metabolism in PD	Merits: - Shotgun sequencing provides more detail than 16 s rRNA sequencing. - Viral metagenome was also mapped. Demerits - Low sample size cohort.
Petrov et. al. [127]	Russia	N = 89 PD patients N = 66 age matched and weight matched healthy controls	Genus level - Dorea, Bacteroides, Prevotella, Faecalibaccerium. Species level - Bacteroides massiliensis, Stoquefichus massiliensis, Bacteroides coprocola, Blautia glucerasea, Dorea longicantena, Bacteroides dorei, Bacteroides plebeus, Prevotella copri, Coprococcus eutactus, Rumino coccuscallidus	Genus level - Christensenella, Catabacter, Lactobacillus, Oscillospira, and Bifidobacterium Species level - Christensenella minuta, Catabacter hongkongensis, Lactobacillus mucosae, Ruminocococcus bromii, Papillibacter cinnamivorans	Alpha diversity (chao1 index) reduced in PD group. Beta diversity - significant differences in weighted UniFrac.	–	–	- Severity and stage of PD in the patients and treatment being received, if any, was not addressed. - Supporting evidence for microbial changes producing inflammation and and α-synuclein aggregation has not been provided.
Scheperjans et. al. [131]	Finland	N = 72 PD patients N = 72 age and sex matched controls	Prevotellaceae		Alpha diversity –no difference. Beta diversity - significantly different in both weighted and unweighted UniFrac.	With disease severity - Prevotellaceae With postural and gait difficulty dominant. phenotype - Enterobacteriaceae was more abundant.	–	- Temporal and causal relationship between microbiome alterations and PD is not clear. - Dietary habits of recruited subjects was not documented.
Jin et. al.[139]	China	N = 72 PD with 13 newly diagnosed N = 68 age- and lifestyle- matched healthy controls (family member or spouse)	Prevotellacea and Lactobacillacea (in older PD group) Lachnospiraceae	Porphyromonadaceae, Christensenellaceae, Clostridialesvadin group, Rikenellaceae (in older PD patients) Turicibacteraceae, Ruminococcaceae, Rikenellaceae (in newly diagnosed)	Alpha diversity - Shannon and Simpson indices were higher in the newly diagnosed group. Beta diversity - significant difference based on unweighted but not weighted UniFrac.	With disease severity, medication, and non-motor symptoms - Eggerthella, Prevotella, Turicibacter, Lactobacillus, and Enterococcus. With medical treatment - Streptococcus and Prevotella were negatively associated, and Turicibacter was positively correlated.	Many different pathways involved in bacterial movement, motility, colonization, etc. were increased and pathways involved in biosynthesis and metabolism were lower in the PD.	- Recruited healthy spouses as controls. - Molecular mechanism through which microbiome has changed in PD not delinealted.
Animal Model Studies
Ref	PD model	Test vs control group	Taxa decreased in PD	Taxa increased in PD	Alpha and beta diversity	Association of microbiome changes	Functional prediction metagenomics	Comments
Yang et. al. [93]	Rotenone Induced C57BL/6 mice	N = 5 (tested at 1, 2, 3 and 4 weeks of Rotenone treatment) N = 5 (tested at 0 weeks of Rotenone treatment)	Bacteroidetes (at 3 weeks). Genus: Firmicutes Clostridium (1–4 weeks), Proteobacteria Sutterella (1–4 weeks), Firmicutes Lactococcus (1–3 weeks), Proteobacteria Desulfovibrio 2–4 weeks), Bacteroidetes Paraprevotella (1 and 3 weeks), Actinobacteria Adlercreutzia (2 weeks)	Firmicutes (at 3 weeks). Genus: Firmicutes Lactobacillus (3 and 4 weeks)	Alpha diversity - No difference with ACE and Chao index. Decrease in Simpson's and Shannon's index at 3 weeks and decrease in Simpson's at 4 weeks. Beta diversity - significant difference at 3 weeks and 4 weeks.	–	–	- Small sample size per group for 16S rRNA sequencing analysis. - Age of mice used may represent early stage of PD. - Merit: longitudinal study that correlates microbiome changes with motor and non-motor symptoms of PD.
Ghaisas et. al.[97]	MitoPark (MP) mice in C57BL/6 background (mitochondrial gene selectively removed from dopaminergic neurons using Cre/Lox system)	MP mice N = 8 (4 males and 4 females) Littermate controls N = 10 (5 males and 5 females)	–	Prevotella genus (abundance increased with age progression)	Alpha diversity - Chao diversity index was not affected over time. Beta diversity − 36 % nominal difference between groups over time	–	–	- Small sample size per group for 16S rRNA sequencing analysis.
Dehnow et. al.[109]	Rotenone-Induced & transgenic Drosophila melanogaster model	Flies treated with Rotenone and transgenic fly model Healthy controls and flies treated with Creatine (rescue)	–	Acetobacter (A. pomorum and A. tropicalis) Lactobacillus (L. brevis, L. fructivorans andL. plantarum)	–	–	–	- Mechanism through which creatine rescues rotenone-induced PD is not understood. - Gender of the flies not specified.