Changes of Colonic Bacterial Composition in Parkinson’s Disease and Other Neurodegenerative Diseases

In recent years evidence has emerged that neurodegenerative diseases (NDs) are strongly associated with the microbiome composition in the gut. Parkinson’s disease (PD) is the most intensively studied neurodegenerative disease in this context. In this review, we performed a systematic evaluation of the published literature comparing changes in colonic microbiome in PD to the ones observed in other NDs including Alzheimer’s disease (AD), multiple system atrophy (MSA), multiple sclerosis (MS), neuromyelitis optica (NMO) and amyotrophic lateral sclerosis (ALS). To enhance the comparability of different studies, only human case-control studies were included. Several studies showed an increase of Lactobacillus, Bifidobacterium, Verrucomicrobiaceae and Akkermansia in PD. A decrease of Faecalibacterium spp., Coprococcus spp., Blautia spp., Prevotella spp. and Prevotellaceae was observed in PD. On a low taxonomic resolution, like the phylum level, the changes are not disease-specific and are inconsistent. However, on a higher taxonomic resolution like genus or species level, a minor overlap was observed between PD and MSA, both alpha synucleinopathies. We show that standardization of sample collection and analysis is necessary for ensuring the reproducibility and comparability of data. We also provide evidence that assessing the microbiota composition at high taxonomic resolution reveals changes in relative abundance that may be specific to or characteristic of one disease or disease group, and might evolve discriminative power. The interactions between bacterial species and strains and the co-abundances must be investigated before assumptions about the effects of specific bacteria on the host can be made with certainty.


Introduction
The causes of the heterogenic group of neurodegenerative diseases (NDs) are still unknown but several contributing factors including genetic and lifestyle factors and age-related aberration of health are likely to play a role. The pathology in PD was found to start in more caudal parts of the central nervous system (CNS) or even in the enteric nervous system, bringing the gut and its interaction with CNS into the spotlight [1][2][3]. While some researchers question whether PD begins in the gut, a potential role of the gut microbiome in PD pathology is undisputed [4]. In the intestinal tracts of PD patients, a pro-inflammatory microbiota composition was discovered, which might lead to increased gut permeability. This high permeability of the intestinal mucosa is known as leaky gut, since bacterial products and inflammatory mediators could pass through the gut's mucosa and invade the blood [5].
PD is a neurodegenerative disease with the hallmark of cardinal progressive motor symptoms such as tremor, muscular rigidity, postural instability and bradykinesia, mostly additionally accompanied -subjects in the case groups had a diagnosis for the disease given by an expert. -human case control studies at least one part of the study analysed the gut microbiome in a cross-sectional manner compared to healthy controls faeces collection to generate a sample probe microbiota analysis by amplification sequencing methods or hybridisation on microarrays: Phylochip G3, YIFscan published in peer-reviewed journals paper available in English language Animal studies were excluded because our review did not aim at proving a possible mechanistic explanation of neurodegeneration (see Figure 1). Our aim was rather identifying the exact taxa that play a role in NDs, because such information may be used as potential biomarkers and thus be clinically relevant for early disease diagnosis. Thus, the potentially confounding influences of different host species are ruled out by comparing only human data. The focus of our review is PD, but the results of all mentioned NDs can be seen in the additional Table A1 in the Appendix. To our knowledge no other review to date provides such a comprehensive overview of the observed relative microbiome changes in abundance in PD as well as in other NTs such as Alzheimer's disease (AD), multiple system atrophy (NSA), multiple sclerosis (MS), neuromyelitis optica (NMO) and amyotrophic lateral sclerosis (ALS).

Materials and Methods
The key question of this review was whether PD patients present specific taxonomic changes in gut microbiota discriminating them from other NDs and from healthy controls. This is a systematic review of human case control studies, summarizing the findings of original studies published in the last five years (01.01.2013- 31.12.2017), during which the most research on this topic has been performed.
We concentrated on bacterial taxa and excluded results regarding fungi, archaea and viruses. Articles that fulfilled the following criteria were included in the comparison: subjects in the case groups had a diagnosis for the disease given by an expert. -human case control studies at least one part of the study analysed the gut microbiome in a cross-sectional manner compared to healthy controls faeces collection to generate a sample probe microbiota analysis by amplification sequencing methods or hybridisation on microarrays: Phylochip G3, YIFscan published in peer-reviewed journals paper available in English language Animal studies were excluded because our review did not aim at proving a possible mechanistic explanation of neurodegeneration (see Figure 1). Our aim was rather identifying the exact taxa that play a role in NDs, because such information may be used as potential biomarkers and thus be clinically relevant for early disease diagnosis. Thus, the potentially confounding influences of different host species are ruled out by comparing only human data. The focus of our review is PD, but the results of all mentioned NDs can be seen in the additional table A1 in the Appendix A. To our knowledge no other review to date provides such a comprehensive overview of the observed relative microbiome changes in abundance in PD as well as in other NTs such as Alzheimer's disease (AD), multiple system atrophy (NSA), multiple sclerosis (MS), neuromyelitis optica (NMO) and amyotrophic lateral sclerosis (ALS). Studies included in qualitative synthesis (n = 23) Figure 1. Methodical approach of systematic review due to PRISMA criteria (PRISMA criteria [33]).

Results
Ten case control studies enrolling PD patients were included in the analysis. Five studies used the Illumina Miseq sequencing methods [34][35][36][37][38], while two studies used Illumina Hiseq sequencing method [39,40]. One study used the Roche 454 GS FLX Titanium sequencing method [41]. Two included studies used other taxonomic identification and quantification methods: Yakult Intestinal Flora-SCAN (YIF-Scan) [42] or 96-well block of the ABI PRISM 7900HT Sequence Detection System [43]. The last two methods only assess targeted microbial taxa and are used to reproduce previously reported findings, whereas the methods of the other eight studies assess the whole microbiome and may detect surprising unknown species. Since BMI, age and gender are expected to influence the gut microbiome, a table with information about the main clinical characteristics of the study participants from the 10 included Parkinson's studies is provided (Table 1). In the majority of publications, where information on Alpha diversity/richness was given, no significant difference between healthy controls and PD cases was found. However, in all reports where information on overall beta-diversity was presented, it differed significantly between healthy controls and PD (Table 2).

Taxonomic Changes
Taking the 10 publications together, changed relative bacterial abundances between HC and PD were observed on all taxonomic levels. A table is provided for each phylum, which showed changes between HC and PD in at least one study. Only significant changes (p < 0.05), which underwent a statistical analysis and correction by the authors of the original study, were included in the tables. The tables only show changes on phylum, family, genus, species or operational taxonomic unit (OTU) levels, always in the highest taxonomic resolution, which the authors of the study provided. The few additional changes on class or order level will be specifically described in the discussion. Significant changes in relative abundances were observed only within five phylae: Firmicutes, Actinobacteria, Bacteroides, Verrucomicrobia and Proteobacteria.

Firmicutes
The most changes in relative abundance of colonic microbial taxa were observed within the Firmicutes phylum. Single studies found changes in unclassified Firmicutes and in various families and genera (Table 3). A significant increase of genus Lactobacillus [34,35,42], decrease of genus Faecalibacterium or species Faecalibacterium prausnitzi or OTUs that belong in this category [34,35,37,43], decrease of genus Coprococcus [38] or species Coprococcus eutactus [34] or OTU [35] belonging in this category and decrease of the genus Blautia [37,38] or species Blautia glucerasea [34] or OTU belonging in this category [35] were reported and confirmed in PD. In all studies in which changes in relative abundance of or within the phylum Actinobacteria were reported, PD patients reproducibly showed an increase in relative abundance compared to HC [34,35,37,43] (Table 4). The increase of the genus Bifidobacterium was reproduced in three studies [34,35,43].

Bacteroidetes
The published data for the phylum Bacteroidetes was highly variable between the different reports ( Table 5). The potential trend of decreased Bacteroidetes on phylum level [37,43] and decreased genus Bacteroides [34] or species Bacteroides fragilis [42] was observed in the opposite direction by one publication [38]. Three studies found a significant reduction of Prevotellaceae family [32,43] or the genus Prevotella [34,39]. One study found a similar but not significant decrease of Prevotellaceae [43]. Hill-Burns et al. (2017) did not confirm these findings, but observed an increase in one OTU that belongs to the genus Prevotella [35].

Proteobacteria
Single studies found changes in relative abundance within the Proteobacteria, Alpha-proteobacteria-and Gammy-Proteobacteria phylae. None of these findings were reproduced by other studies, apart from the increase in the Enterobacteriaceae family in two studies [37,43] (Table 7).

Discussion
Our data show that the reproducibility of data is very low at a higher taxonomic rank, the phylum level. The inconsistency of the data could indicate that the changes of phyla are facultative but not obligatory for NDs. Moreover, it implies that it is sufficient for an effect on the host that lower taxonomic levels are altered without achieving an alteration at phylum level. Profound changes in one family could be compensated for by oppositional changes in other families of that phylum. In conclusion, comparisons on phylum level are too rough, and higher taxonomic resolution is needed to discover similarities of the microbiota composition relating to the diseases.
A negative correlation between PD disease duration and Firmicutes abundance was reported [38]. Additionally, an inverse correlation was found between the COMT inhibitor type of PD medication with Entacapone and Firmicutes abundance [43]. These findings show that there is more to PD than a lower Firmicutes phylum abundance, since otherwise Entacapone would not display its beneficial effects. No other study apart from Keshavarzian et al. (2015) found a significant change in abundance of Firmicute phylum in PD [38]. This implicates the importance of controlling disease medication and disease duration of the subjects during the study since both alter the gut microbiota composition. For the aim of proving causality, a reduction of confounding may be achieved by enrolling a study population consisting of PD patients at an early stage or/prodromal stage and no medication.
Since the Firmicutes:Bacteroidetes ratio is found to change with age [44], an association with neurodegeneration could be considered. However, no study reported any significant differences in Firmicutes:Bacteroidetes ratio in PD. A minority of studies in other NDs, namely two ALS studies [45,46] and one MSA study [47], discovered an aberrant Firmicutes:Bacteroides ratio, indicating no predominant role for the NDs. Moreover, several other factors influence the Firmicutes:Bacteroides ratio. The characteristics of the recruited cohort are crucial. In the elderly but also in infants, the ratio is extremely low [44]. It was suggested that the Firmicutes:Bacteroidetes ratio is also associated with obesity [48][49][50][51], but a causative key role for this ratio for obesity was questioned by other groups [52]. Different types of diet change the Firmicutes to Bacteroides ratio [53], and it was reported that freezing faecal samples before extracting the DNA significantly increases the Firmicutes:Bacteroides ratio [54]. Other methodical choices, like choosing between the different bead beating cell lysis instruments before metagenome analysis, influence this ratio up to threefold in favour of Bacteroides [55]. Taken together, these data indicate that the shift is not specific for neurodegenerative diseases and might sometimes be a consequence of technical and lifestyle factors.
Several alterations in the Firmicutes phylum were reported in PD and other NDs. Commensal Clostridium species, like Clostridium saccharolyticum and Clostridium leptum, showed a significant and descriptive decrease in PD [42]. This reduction is in agreement with the observed reduction of the genus Clostridium in AD [56] and MS [57]. By contrast, an increase of the harmful species Clostridium perfringens was shown in one study in NMO patients [58]. This is a very interesting finding, since NMO is a demyelinating disease that is associated with antibodies against the water channel protein aquaporin-4 (AQP4) expressed on astrocytes and T helper 17 cells [59]. These antibodies are reported to exhibit a cross-reactivity to a homologous peptide sequence within the adenosine triphosphate-binding cassette (ABC) transporter permease of Clostridium perfringens [59,60]. Also, a theoretical involvement of the epsilon toxin of Clostridium perfringens in inducing neurodegeneration was proposed in MS due to its ability to cross the blood-brain barrier, damaging oligodendrocytes and due to its observed contribution to enterotoxaemia in ruminants [61][62][63]. However, no increased Clostridium perfringens was found in the five MS studies [57,[64][65][66][67]. Nonetheless, epsilon toxin might play a role in NMO since it can damage optic nerves in vivo [68]. This and the possibility to vaccinate against epsilon toxin, at least in rodents, point to an importance of the observed increase of Clostridium perfringens in NMO [69].
Lactobacillus was increased in three of 10 PD studies, in contrast to AD [56], MSA [47,70], ALS [45,46,71] and NMO [58], where no change was observed. However, this increase of Lactobacillus in PD patients could be caused by the frequent constipation of PD patients, since Lactobacillus is Nutrients 2018, 10, 708 9 of 23 also known to be increased in constipation-type IBS and decreased in diarrhoea-type IBS [72] and the co-variable constipation was methodically and statistically treated differently in the PD studies. The decrease of both Lactobacillus and Lactobacillus rogosae that was reported in MS patients [57,66] supports the above hypothesis, since several medications against MS lead to diarrhoea [73,74]. The possibility that a probiotic intervention or preventive treatment with beneficial Lactobacillus strains could have an ameliorating effect on MS deserves to be investigated.
Increased Lactobacillaceae abundance and decreased Prevotella has been linked to reduced ghrelin concentration and altered ghrelin secretion has been reported in one PD study [43]. Ghrelin is a gut hormone that may contribute to the maintenance and protection of normal nigrostriatal dopamine function [74]. Future studies should investigate if Lactobacillus has a constipation-related consecutive role, or rather a causative role. Such studies could, moreover, define which Lactobacillus species are beneficial and which may be detrimental. This differentiation is essential since there are probiotic treatment proposals for PD that include Lactobacillus strains and, on the other hand, the potentially harmful link with reduced ghrelin secretion has not been sufficiently investigated.
Two PD studies showed an increase of Ruminococcaceae family [31,35], which might also discriminate PD from other NDs, since in other NDs either no change or even a decrease was observed, namely in AD [56], MSA [47], MS [75] and ALS [46]. The increased Ruminococcaceae on a family level might play a role in PD patients; however, which subordinate taxa contribute to the disease cannot be unequivocally stated yet.
Faecalibacterium, partly specified as Faecalibacterium prausnitzi or as an out, showed a significant decrease in four PD studies [34,35,37,43], while this was not observed in any other ND apart from MS [57,65]. In ALS an increase was noted in one study [76]. However, Faecalibacterium showed a negative correlation to Entacapone taxa [43] and a positive correlation to vitamin D supplementation in MS patients [65], indicating that also other factors might contribute to the observed change of Faecalibacterium abundance. Thus, the specific Faecalibacterium change in PD might be provoked by extrinsic factors including medication or nutrition. Nevertheless, Faecalibacterium prausnitzi is advocated as a potential future probiotic due to some strain-dependent anti-inflammatory features, like butyrate production [77]. However, the hypothesized beneficial effects on intestinal barrier integrity resulting in anti-inflammatory benefits could not be confirmed by one study in an in vitro model of the large intestine [78]. Since the Faecalibacterium prausnitzi species include a high diversity of strains, a strain-dependent classification and investigation would improve insight into the beneficial and harmful effects on the host [79].
Lachnospiraceae family showed a depletion not only in PD [35], but also in MSA [47], MS [75] and ALS [45]. Even if this reduction on family level is not specific to PD, it is still interesting, because Lachnospiraceae include many putative anti-inflammatory and thus potentially protective genera. In addition, Lachnospiraceae were negatively correlated to disease duration in PD patients [38]. Thus, this reduction might just be a consequence of common disease-related changes. The key members of Lachnospiraceae showed a confirmed decrease in PD (Table 3). Interestingly Blautia and Dorea are also decreased in MSA [47], while AD [56], NMO [58] and MS [66] showed an increase of Blautia and MS [66] and ALS [45] a decrease of Dorea. On a family level, the decrease of Lachnospiraceae is correlated with PD disease duration and the abundance of Lachnospiraceae might be, like the incidence of PD, sex-dependent [36]. Concluding that the decrease of Blautia and Dorea in case-control-studies is limited to alpha-synucleinopathies like PD and MSA, it is tempting to investigate, if these genera additionally correlate with longitudinal data of PD and MSA patients and if the changes predate the neurodegenerative progress and symptoms.
Another family, Christensenellaceae, showed an increase on family and species level, e.g., Christensenella minuta, during the course of PD [34,35], a change that was also observed by one study on genus level of Christensenella in MS patients [75]. Thus, these changes might not be specific but still characteristic for some PD patients. Same applies to the increase of Oscillospira in PD [34,38]. Petrov et al., moreover, explained recently that Christensenellaceae is a heritable taxon that is co-abundant with Oscillospira. These taxa are, like PD, significantly associated with lower body weight [80], which might indicate a higher vulnerability for PD, or more severe PD stages, when these families were inherited.
Two different species of Eubacterium were observed to be decreased in PD, MS and a case-control-study of a cognitive impaired elderly without an explicit AD diagnosis [39,57,81]. In the cognitively impaired elderly, a negative correlation between E. rectale and proinflammatory cytokines IL-1b, NLRP3 and CXCL, in addition to a positive correlation between these cytokines and Escherichia/Shigella, was observed [81]. This might imply an anti-inflammatory role for several Eubacteria species in the development of neurodegenerative diseases. Regarding changes in the metabolome of colonic microbiota, Eubacterium contributed the most to the decrease in genes for D-glucuronate degrading enzymes and two more active tryptophan metabolism pathways observed in PD patients [39]. Glucuronidase enzymes mediate the regeneration of molecules important for host health, but also toxins and carcinogens [82]. Moreover, they activate endogenous glucuronides of hormones and neurotransmitters [82]. L-Tryptophan is the precursor of serotonin and is shown to be decreased in PD patient brains [39] and its metabolites display regulatory immune function with beneficial and harmful potential . Finally, some non-significant trends, between Unified Parkinson's disease rating scale III (UPDRS III) and abundances of e.g., Eubacterium rectale, Eubacterium hallii and Eubacterium eligens were reported [39]. This is mentionable since E. rectale might penetrate the mucus layer due to its flagella. Interactions between E. rectale, the mucus layer and the intestinal mucosa are therefore feasible.
The members of the phylum of cyanobacteria, which belong to the blue-green algae, produce a series of neurotoxins, from which beta-N-Methylamino-L-alanine (BMAA) is the most common. However, the production of neurotoxins by their candidate phylum sibling Melainabacteria is not yet well investigated [83]. A link between cyanobacterial derived neurotoxins and neurodegenerative diseases was shown already more than 10 years ago by Cox et al., who examined the Chamorro people of Guam, who exhibit a 50-100 times higher incidence of ALS than anywhere else [84]. BMAA is linked with intraneuronal protein misfolding, which is an important pathological hallmark of many NDs. High levels of neurotoxin BMAA have been associated with AD [85]. Post-mortem brain specimens of ALS, AD and Huntington's disease (HD) patients showed increased BMAA concentrations [86]. Moreover, BMAA targets N-Methyl-D-aspartate (NMDA) receptors and the neurotransmitter glutamate, which are believed to be disturbed in AD. Also, other cyanobacterial derived neurotoxins could play a role in neurodegeneration in individuals with more permeable intestinal epithelial barrier of the GI tract, like the leaky gut observed in elderly people [87]. However, none of the examined studies detected an altered relative abundance in Cyanobacteria in PD. Also, in other neurodegenerative diseases like AD [56], MSA [47,70], MS [57,[64][65][66]76], ALS [45,46,71] and NMO [58], no change in this phylum was reported. Only Heintz-Buschart et al. reported a decrease of an unclassified OTU171 with high resemblance to the Melainabacterium MelB1,57 in PD. However, they could not asses any of the known cyano-neurotoxins [40]. This and the fact that this Melainabacterium was decreased in PD speak in favour of a protective outcompeting role, possibly by replacing other harmful cyanobacteria. Products of other bacteria, belong to the phylum of red algae, such as a phycoerythrin, were shown to have an ameliorating effect on AD [88,89]. It can be summarized that Cyanobacteria can produce neurotoxins that lead to neurodegeneration, but the abundance of Cyanobacteria is only increased in special, sometimes isolated populations. The role of the newly detected phylum of Melainabacteria is not yet clear, but so far no harmful association has been found and a protective role can even be postulated.
Only a few changes were reported in the phylum of Actinobacteria, but these were consistent regarding several taxonomic resolutions (Table 4). PD patients showed an increase on phylum level [37], which was reflected in an increase in Bifidobacteriaceae (family) [35] and Bifidobacterium (genus) [34,35,43], while no changes were observed in any other examined NDs. In contrast, AD patients showed a decrease on a phylum level, reflected by a decrease in Bifidobacteriaceae (family) and Bifidobacterium (genus) [56]. Despite the findings of an increased Bifidobacteriaceae in PD, a two-year follow-up study showed that a decrease in Bifidobacterium later in disease may be able to predict whether PD stage is going to deteriorate or not, even correlating with the Unified Parkinson's disease rating scale I (UPDRS I) score [90]. This implicates that the increase in Bifidobacteriaceae in PD is not detrimental to patients, but rather may represent a mechanism of beneficial action against neurodegenerative aggravation. Thus, it may be postulated that a probiotic intervention with Bifidobacterium could prevent progression of PD to severer stages.
Within the phylum of Bacteroidetes, Prevotella is a well-known and often discussed genus. However, changes in PD patients were not consistent throughout the PD studies (Table 5). In one study, which failed to find any significant difference in relative Prevotella abundance between PD subjects and controls, Prevotella still was the most reduced bacterium and differed by 3.2-fold [42]. In another study without a statistically significant change in Prevotella, an association between the family Prevotellaceae and the clinical score of PD severity, the UPDRS-III score, was reported [41]. Moreover, trends showing a decrease in Prevotella were obtained by one study when looking at the microbiome of colonic mucosa rather than faeces [38]. Negative results concerning significant changes in Prevotella could be a consequence of methodological differences and other criteria for control subjects [38]. The fact that Prevotellaceae was decreased in idiopathic rapid eye movement behavioural sleep disorder, iRBD, which predates PD in a majority of cases, supports the hypothesis that this family and its genus Prevotella might only be changed in the early stages of PD. It is also suggestive that a decrease of Prevotella contributes to the onset of PD and therefore could be used as a biomarker for PD diagnosis [40]. However, unambiguous interpretation of changes of Prevotella is difficult, since their abundance also depends on social factors, as observed in monkeys [91], and subgenus differences in Prevotella and Bacteroides have been related to different dietary patterns [92].
Additionally, the change in abundance of Prevotella depends on the species type (Table 5). Prevotella copri and P. clara were decreased in PD [34,39,57]. This reduction is in agreement with reduced abundances in MSA [70], MS [64] and NMO [58]. In contrast, the species Prevotella melaninogenica and one OTU of Prevotella showed an increase in PD [39] and NMO [58]. Prevotella exhibited one of the highest effect sizes of change in the study of NMO [58]. The fact that certain Prevotella species (especially Prevotella copri) are indeed decreased in all NDs, apart from AD (see Table A1 in the Appendix A), suggests a possible protective role against neurodegenerative processes in the brain and deserves future investigation.
In the Verrucomicrobia phylum, four studies found that Akkermansia genus or species were increased in PD patients [35,[38][39][40]. Moreover, both high colonic Akkermansia abundance and PD were shown to be negatively correlated to BMI [93]. Akkermansia muciniphila exerts beneficial effect on the intestinal mucosal layer and improves the barrier function of the gut epithelium [39]. Thus, in conditions of low Akkermansia abundance, the maintenance of a healthy gut barrier may not be possible anymore, and pathogenic factors of other bacteria, like Lipopolysaccharide (LPS), could consequently harm the host. On the other hand, Akkermansia uses mucus as an energy source and degrades the mucus layer [94]. This could lead to increased exposure of microbial antigens to immune cells and thus could have inflammatory potential. The role of the pro-and anti-inflammatory properties is not yet clear but maintaining a steady state of Akkermansia may be a pre-requisite for normal functioning of the gut homeostasis. Akkermansia together with Eubacterium, Capnocytophaga, Phascolarctobacterium and OTUs no further classified than to Firmicutes, achieved a high prediction score for PD [39].
No changes within the phylum of Proteobacteria, apart from the family Enterobacteriaceae [37,43], were reported in PD patients. Therefore, it can be concluded that the relative abundance of these taxa is only changed in a minority of patients. Because of the missing confirmation of these alterations, the clinical relevance of these findings can be questioned. Escherichia coli, a member of the Enterobacteriaceae family, showed translocation into the colonic mucosa in PD patients [95]. The complexity of this interrelationship is shown by the increased abundance of Enterobacteriaceae in PD patients with the postural instability and gait difficulty (PIGD) phenotype compared to the tremor-dominant subtypes (TD) [41]. The more progressive and severe pathology of non-TD-phenotypes (including PIGD) supports the idea of a different unknown pathological mechanisms underlying this subgroup of PD [41]. The observed difference in Enterobacteriaceae abundance between the subgroups of PD disease might equally occur due to other variable phenotype features in this study: PIGD subjects tended to be older and to have higher nonmotor symptoms scale (NMSS) total scores than TD subjects [41]. These differential abundant taxa between PD phenotypes was challenged by another study [43] as these researchers only found an overall increase in PD [37,43]. However, in this study Unger et al. used a slightly different classification and they interpreted their results based on a small number of the TD patients in their sample. In agreement with the findings in PD, a significant increase of Enterobacteriaceae and some other taxa of Proteobacteria phylum was also observed in NMO. The mentioned change in Enterobacteriaceae was not only significant but they were also four times more abundant than in healthy controls [58]. A protective role of Enterobacteriaceae was proposed by Hasegawa et al., demonstrating that Enterobacteriaceae (and C. perfringens, B. fragilis, and Pseudomonas) produce hydrogen water (1000 mL/day) and were associated with improved total UPDRS scores in PD patients in a double-blind randomized controls study [96]. In addition, they were shown to be protective against PD pathology in rats [97]. These researchers performed a study to assess whether the total amount of breath hydrogen in PD patients was lower than in healthy controls and confirmed this hypothesis [42]. The results regarding the potential protective or aggravating role of Enterobacteriaceae are contradictory and require further investigation.
A high number of microbiota composition changes found in human case-control studies of ND concern unclassified bacteria. Approximately 40% of the gut microbiome cannot be captured by reference genome-based methods [98]. It is hypothesized that during ND unclassified bacteria gain more importance. Heintz-Buschart et al. discovered 10 depleted unclassified OTUs in PD patients that exhibited together very low fractional abundance [40], but could still manipulate the host's health via an unknown unique mechanism without a need for high abundance. Interestingly, the genome of OTU 469 encoded an endoglucanase with a synuclein-like domain and thus may lead to an immune system that is tolerant to alpha-synuclein-like structures [40]. This could have a potential implication for the induction of PD and MSA pathology. The depletion of this bacterium leads to a lack of host's immune system exposure to the bacterial alpha-synuclein-like structures, which may result in immunologically induced alpha-synuclein aggregation. This hypothesis is in line with the reports that alpha-synuclein aggregation starts in the enteric nervous system and not in the brain [1,3,99]. All unclassified OTUs of this study, where the genome could be recovered, showed common features such as fermentative lifestyle, motility and the ability to synthesize vitamins. The great number of observed changes in unclassified Bacteria or single OTUs must be handled carefully since the method of quantification of the most included studies was 16S rRNA amplification and high-throughput sequencing, which is known to prone to sequencing errors and chimera creation, leading to an artificial high diversity and low reproducibilityLastly, no studies have analysed the gut microbiome of iRBD patients yet (apart from Heintz-Buschart et al.), but this disease showed a high conversion rate to alpha-synuclein disorders, especially PD. More investigations of microbiome in prodromal stages of diseases with follow-up evaluation are necessary.
Only one human case control study analysed the colonic microbiome of AD patients [56]. Most AD patients were in the stage of mild or very mild AD according to clinical dementia rating (CDR) scores. Targeting the early stage of the AD may help to examine possible causal associations between microbiome compositional changes and disease onset. Limited overlap was found between changes of the microbiome in AD and changes identified in any of the 10 PD studies. Minor similarities were observed that included a decreased Firmicutes phylum and Clostridium genus in AD, whereas significantly decreased Clostridium leptum [42] or Clostridium saccharolyticum [39,42] was observed on PD. Moreover, AD showed an increased Bacteroidetes phylum, Bacteroides genus [38] and Alistipes genus, while increased Alistipes shahii [39] was evident in PD patients. As the above findings were only found in a minority of the cohorts, no great similarity between the microbiome of AD and PD can be determined so far.