Correlation between the gut microbiome and neurodegenerative diseases: a review of metagenomics evidence

Abstract

A growing body of evidence suggests that the gut microbiota contributes to the development of neurodegenerative diseases via the microbiota-gut-brain axis. As a contributing factor, microbiota dysbiosis always occurs in pathological changes of neurodegenerative diseases, such as Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis. High-throughput sequencing technology has helped to reveal that the bidirectional communication between the central nervous system and the enteric nervous system is facilitated by the microbiota’s diverse microorganisms, and for both neuroimmune and neuroendocrine systems. Here, we summarize the bioinformatics analysis and wet-biology validation for the gut metagenomics in neurodegenerative diseases, with an emphasis on multi-omics studies and the gut virome. The pathogen-associated signaling biomarkers for identifying brain disorders and potential therapeutic targets are also elucidated. Finally, we discuss the role of diet, prebiotics, probiotics, postbiotics and exercise interventions in remodeling the microbiome and reducing the symptoms of neurodegenerative diseases.

Introduction

The human intestinal microbiota has been associated with numerous chronic diseases, including cancer, rheumatism, and cardiovascular disease (Durack and Lynch, 2019; Fan and Pedersen, 2021; Ma et al., 2021; Alsegiani and Shah, 2022; de Jonge et al., 2022; Liu et al., 2022a; Yuan et al., 2022b). One clear example is trimethylamine-N-oxide, a gut metabolite that promotes the development of cardiovascular disease by influencing the human host metabolism (Brown and Hazen, 2018). Additionally, recent research has revealed that gut microbial dysbiosis may contribute to neurodegenerative disease (NDD)-related symptoms, including the accumulation of senile plaques, inflammation, and oxidative stress in Alzheimer’s disease (AD), amyotrophic lateral sclerosis (ALS), Parkinson’s disease (PD), multiple sclerosis (MS), and Huntington’s disease (HD) (Elfil et al., 2020; Hou et al., 2021; Liu et al., 2021b; Nishiwaki et al., 2022; Zhang et al., 2022b). Gut microbiota (GM) alterations in patients can trigger the production of inflammatory cytokines (tumor necrosis factor-α, interferon-γ), nicotinamide (NAM), and amyloid-β (Aβ) plaques. These activate brain inflammatory pathways (CCAAT-enhancer-binding proteins β/asparagine endopeptidase pathway) and impair the NAM metabolic pathway, which triggers movement disorders and cognitive deficits (Blacher et al., 2019; Lin et al., 2019; Chen et al., 2022b).

The powerful tool of metagenomics has given us the opportunity to understand the links between the human host and the gut microbiome (Quince et al., 2017). It has helped to clarify microbial community structure, species composition, phylogeny, gene function, and metabolic networks in clinical samples (Hertel et al., 2019; Nayfach et al., 2021). Microarray was originally employed to study the microbiome and yielded conspicuous results (Raoult, 2008; Paliy et al., 2009; Rajilić-Stojanović et al., 2009; Cox et al., 2010; Lemon et al., 2010; Knights et al., 2011). However, with the advent of the second revolution in sequencing technology, gut microbiome research has moved into the field of metagenomics (Chiu and Miller, 2019). Compared with short-read based metagenomes, long-read based metagenomes offers more accurate and complete genome sequences, which lays the foundation for genome assembly, gene prediction and function annotation (Gao et al., 2021; Gounot et al., 2022). In addition, with the increasing maturity of bioinformatic analysis, metagenomics allows for a deeper understanding of the pathogenesis underlying NDDs (Meyer et al., 2022). Here, we investigated the advances in metagenomics with a focus on the association between the GM and NDDs. We also discuss multi-omics analysis that assesses the changes in the structure and metabolism of gut microbiome in NDDs. Finally, we introduce potential microbiota-reshaping interventions that could be used to mitigate NDDs.

Literature Search

The articles referred in this metagenomics review were retrieved on the PubMed/MEDLINE database with the Advanced Search Builder. Keyword searches were conducted using the following term combinations: “(dementia) AND (metagenomics)”, “(Alzheimer’s disease) AND (metagenomics)”, “(Amyotrophic Lateral Sclerosis) AND (metagenomics)”, “(Parkinson’s disease) AND (metagenomics)”, “(Multiple Sclerosis) AND (metagenomics)”, AND “(Huntington’s Disease) AND (metagenomics)”. Then, we replicated the search by replacing the keyword “metagenomics” with “microbiota” OR “gut microbiome”. These searches were restricted to articles published in English between 2018 and 2023. The results were then manually screened by scanning titles and abstracts to select the articles including at least one of the key words: dementia, Alzheimer’s Disease, Amyotrophic Lateral Sclerosis, Parkinson’s Disease, Multiple Sclerosis, Huntington’s Disease.

Gut Microbiota and Microbiota-Gut-Brain Axis

Trillions of microorganisms are living in dynamic balance within the human body hosts, which are typically classified as bacteria, archaea, bacteriophages, fungi, and viruses. Of these, 95% of the symbiotic microorganisms reside in the gut (de JRD-PV et al., 2018). The ecological colonies of the microorganisms, intricately dwelling in the gastrointestinal tract (GIT), are collectively known as GM (Kumar Singh et al., 2019; Kesika et al., 2021). Although the composition of the GM varies at different stages of life and shows gender distinction, the dominant populations derived from GIT is Bacteroides, Firmicutes, Actinobacteria, Proteobacteria, Bifidobacterium, Eubacterium, Clostridium, Peptococcus, and Provetella and most of them are mutualistic or commensal microorganisms (Etxeberria et al., 2013; Dahiya et al., 2017; Doifode et al., 2021; Korf et al., 2022). As well-known, the GM is deeply involved in the metabolic processes of human beings. A number of microbial metabolites produced from GM are absorbed by enterohepatic circulation followed by releasing into the bloodstream (Shang et al., 2021b). Typically, bile acids, choline, neuromodulators, bacteriocins, and short-chain fatty acids (SCFAs) are well-studied commensal bacteria metabolites as immune modulators. The reduction of Lactobacillaceae prevents the intestine from metabolizing SCFAs and restricts the expression of tight junction proteins, which ultimately weaken barrier integrity. Therefore, SCFAs also modulate microglia by affecting immune, endocrine, vagal, and other humoral pathways, and enhance motor dysfunction and pathophysiology of NDDs (Sampson et al., 2016; Dalile et al., 2019; Ren et al., 2020b; Jeong et al., 2022). The GM constructs an auxiliary immune system in intestinal mucosa and keeps the health of the host (Fan and Pedersen, 2021). On the other side, the brain directly connects to the GM by regulating the secretory and sensory functions of GI (Montiel-Castro et al., 2013).

Because of the reduced cost of large-scale sequencing, an increasing number of studies have uncovered the interplay between GM and brain. Referring to the host-microbe interactions between the central nervous system and enteric nervous system (ENS, also called “the second brain”) underlying the GIT, the concept of the “microbiota-gut-brain axis”, or “gut-brain axis”, was proposed and has been widely debated (Cryan et al., 2019; Shabbir et al., 2021; Zhu et al., 2021). As Figure 1 shown, the bidirectional communication system in microbiota-gut-brain axis is comprised of neurotransmitters, neuromodulators, neuroimmune pathways, autonomic nervous system, and neuroendocrine system.

Figure 1.

Links between the gut microbiota and neurodegenerative diseases.

The links between the gut microbiota and neurodegeneration diseases. The hypotha lamopituitary adrenal axis plays an important role in the communication between the brain and the gut microbiota. Neurodegenerative diseases trigger intestinal dysbosis by altering the levels of hormones. Subsequently, intestinal metabolites and inflammatory factors activate microglia mediated by the vagus nerve pathway, ultimately leading to the development of neurodegenerative diseases. Created with BioRender.com. iNOS: Inducible nitric oxide synthase; IL: interleukin; TNF: tumor necrosis factor.

Accumulating evidence illustrates that various extrinsic and intrinsic stimulations affect the intestinal mucosa of GIT, which further alters gut integrity through the hypothalamic-pituitary-adrenal axis (Rogers et al., 2016; Korf et al., 2022). In response to environmental stressors such as heavy metals (such as Pb, Mn, As, Al) and pesticides (such as rotenone, chlordane, dieldrin), the brain alters gut integrity and shapes the composition of GM through hypothalamic-pituitary-adrenal axis (Chin-Chan et al., 2015; Zhang et al., 2016; Huat et al., 2019; Li et al., 2021a; Nabi and Tabassum, 2022; Vellingiri et al., 2022). As a result of exposure to maternal microorganisms, GM profiles may contribute to fetus brain defects via the immune signaling pathway (Deidda and Biazzo, 2021).

Using animal models, the preliminary human studies have revealed that NDDs are closely associated with gut dysbiosis (Needham et al., 2020). GM-derived molecules modulate the blood-brain barrier integrity and cognitive impairment through immune, metabolic, and neural pathways, which trigger dysfunction of the microbiota-gut-brain axis (Parker et al., 2020; Liu et al., 2021c). Therefore, emerging evidence suggests that NDD-manifested proteins may misfold in GIT then migrate into brain via the hypothesized routes. For example, the pathological α-synuclein (α-Syn) expressed in enteroendocrine nerve cells might transmit to the brain via the vagus nerve, which connects the brain to many organs in the body (Braak et al., 2006; Liddle, 2018). Changes in α-Syn affect synaptic plasticity and interrupt neurogenesis, which cause tremors and slow movements in patients with PD (Horgusluoglu et al., 2017). Additionally, one possible explanation for the development of AD is the decrease of probiotic bacteria due to an unbalanced diet (Włodarek, 2019). Anthocyanins, which are widely found in dietary fiber, are broken down and metabolized in the intestine; they are then converted into SCFA, which have anti-inflammatory characteristics and provide an acidic environment for the proliferation of probiotic bacteria (Mattioli et al., 2020; Aho et al., 2021). Furthermore, acetate is a type of SCFA derived from microbiota that regulates microglial phagocytosis of Aβ and disease progression in 5×fAD mice (Brandscheid et al., 2017; Erny et al., 2021; Kameno et al., 2022). The microbiome that colonizes the gut can also produce d-glutamate, the receptor of N-methyl-D-aspartate. A recent study has indicated that the N-methyl-d-aspartate glutamate receptor improves cognition in patients with NDDs (Chang et al., 2020). Taken together, these studies clearly demonstrate that GM is a major player in NDDs, which are characterized by GM dysbiosis (Hertel et al., 2019; Vidal-Martinez et al., 2020; Mao et al., 2021). Furthermore, NDDs has also been found to be influenced by oral microbiota in compliace with gut microflora alterations (Ko et al., 2019; Narengaowa et al., 2021; Sansores-España et al., 2021).

Bioinformatics Analysis in Metagenomics

Genome-centric analysis of metagenomics surveils the human gut microbial community. Bridge polymerase chain reaction amplification enables high-throughput and short-read length using Illumina platform in metagenomic next-generation sequencing (Simner et al., 2018; Gu et al., 2019; Wensel et al., 2022). Compared with metagenomic next-generation sequencing, it is difficult to capture signals in the case of low-abundance samples in long-read metagenome; this limited community coverage is largely a result of insufficient research budgets. However, the long-read length of metagenomic third-generation sequencing allows for phylogenetic identification at the species level. Therefore, combined analyses of Illumina and Nanopore-based sequencing technologies improve the resolution and accuracy of the microbial genome. Subsequently, numerous large-scale metagenomics datasets are being generated followed by the development of targeted bioinformatics algorithms and software (Liu et al., 2021d; Taş et al., 2021; Harvey and Holmes, 2022; Table 1).

Table 1.

Selected software for metagenomic analysis

Software	Current version	Functionality	Access link	Reference
Trimmomatic	0.4	Quality control	www.usadellab.org/cms/?page=trimmomatic	Bolger et al., 2014
Bowtie 2	2.5.1		bowtie-bio.sourceforge.net/bowtie2/index.shtml
FastQC	0.12.0		www.bioinformatics.babraham.ac.uk/projects/fastqc/	Andrews, 2010
ABySS	2.1.5	Genomes assembly	www.bcgsc.ca/resources/software/abyss
CLC	23		digitalinsights.qiagen.com	Shen-Gunther et al., 2021, 2022
IDBA UD	1.1.3		github.com/loneknightpy/idba
MEGAHIT	1.2.9		github.com/voutcn/megahit	Li et al., 2016a
MetaVelvet	1.2.01		metavelvet.dna.bio.keio.ac.jp
MIRA	4.0.2		www.chevreux.org/mira_downloads.html
Ray Meta	2.3.1		denovoassembler.sf.net	Boisvert et al., 2012
SOAPdenovo2	r242		github.com/aquaskyline/SOAPdenovo2	Luo et al., 2012
SPAdes	3.15.4		cab.spbu.ru/software/spades
MetaSPAdes	3.15.4		cab.spbu.ru/software/spades
Velvet			www.ebi.ac.uk/~zerbino/velvet
VICUNA			github.com/broadinstitute/mvicuna	Vázquez-Castellanos et al., 2014
MetaBAT2	2.12.1	Genomes binning	bitbucket.org/berkeleylab/metabat
VAMB	3.0.9		github.com/RasmussenLab/vamb	Nissen et al., 2021
Kraken 2	2.1.2	Contig identification	github.com/DerrickWood/kraken2
DIAMOND	2.1.0		github.com/bbuchfink/diamond	Buchfink et al., 2015
GOTTCHA	1.0c		github.com/LANL-Bioinformatics/GOTTCHA
GenSAS	6	Genome annotation	www.gensas.org
GAAP			gaap.hallym.ac.kr
ARGpore2		Functional annoation	github.com/sustc-xylab/ARGpore2	Yin et al., 2018; Wu et al., 2022
Prokka	1.14.5		github.com/tseemann/prokka
GraPhlAn	1.1.3	Phylogeny	github.com/biobakery/graphlan	Asnicar et al., 2015
GTDB-Tk	2.2.4		github.com/Ecogenomics/GTDBTk	Chaumeil et al., 2019
PhyloSuite	1.2.3		github.com/dongzhang0725/PhyloSuite/releases	Xiang et al., 2023
AMPHORA2			github.com/wu-lab-uva/AMPHORA2	Wu and Scott, 2012
PhyloSift	1.0.1		github.com/gjospin/PhyloSift	Darling et al., 2014
GToTree	1.7.07		github.com/AstrobioMike/GToTree	Lee, 2019
ezTree	0.1		github.com/yuwwu/ezTree	Wu, 2018
PhyloPhlAn3	3.0.3		github.com/biobakery/phylophlan	Asnicar et al., 2020
MiGA			microbial-genomes.org	Rodriguez et al., 2018
ShotgunFunctionalizeR	1.2–9	Statistical analysis	shotgun.zool.gu.se/	Kristiansson et al., 2009
LEfSe	1.1.2		github.com/SegataLab/lefse	Segata et al., 2011
Metastats			cbcb.umd.edu/software/metastats	White et al., 2009
Deseq2	1.38.3		bioconductor.org/packages/release/bioc/html/DESeq2.html	Love et al., 2014
edgeR	3.40.2		bioconductor.org/packages/release/bioc/html/edgeR.html	Robinson et al., 2010
STAMP	2.1.3		beikolab.cs.dal.ca/software/STAMP	Parks et al., 2014

The regular workflow of bioinformatics analysis is shown in Figure 2 Table 1 includes quality control, assembly and binning, taxonomic assignment, and downstream analysis processes. In general, the raw sequence datasets are obtained from the sequencing companies. After the removal of adapters and host DNA, metagenomic data remains highly fragmented. Megahit is usually used to assemble clean reads into longer contigs (Li et al., 2016a). Following that, metaMIC can be used to accurately identify mis-spliced sequences and correct misassembled genomes. After the above steps, a reliable genome can be used to analyze the entire microbial community and gene functional annotation (Lai et al., 2022).

Figure 2.

Workflow of the bioinformatics pipeline in metagenomics.

Stage 1: Trimmomatic is mainly used to remove low quality reads and adapters from the raw reads. Stage 2: Clean reads are assembled into contigs and then binned. Stage 3: MetaPhIAn4 is mainly employed for taxonomic assignment. Stage 4: Finally, various tools are used for downstream analysis of the data, including phylogeny, diversity analysis, and differential analysis. Created with BioRender.com.

Recently, a novel workflow proposed by Xia et al. (2023) can be applied to process integrated microbiome data with the focus on sequencing accuracy and new algorithms that will alter bioinformatic pipelines. Gounot et al. (2022) provided unfragmented metagenomics data on the gut microbiome of Southeast Asians using a combination of metagenomic next-generation sequencing and metagenomic third-generation sequencing technologies. Notably, the most cutting-edge technology, PacBio HiFi technology, has been routinely used to pave the way for metagenomic assembly and generate accurate gut microbial profiles (Kim et al., 2022). Despite various tools used for taxonomic assignment, there are significant differences in output data based on different relative abundance types (Oulas et al., 2015). To enhance the reproducibility of results, Sun et al. (2021) suggested classifying metagenomic data according to relative taxonomic abundance (DNA-to-Marker) rather than relative sequence abundance (DNA-to-DNA), such as MetaphlAn. Moreover, one advantage of the metagenome is that taxonomic resolution has reached the species level and can even reach the strain level. Maini Rekdal et al. (2019) employed metagenomics to screen for a strain of Escherichia coli that makes levodopa decarboxylase, which degrades levodopa in the intestine and prevents the drug from entering the brain in treated patients with PD. In addition to metagenomic software, researchers have developed tools specifically for virome analysis (Wang et al., 2010; Roux et al., 2015; Rampelli et al., 2016; Ren et al., 2017; Shean et al., 2019; Antipov et al., 2020; Schäffer et al., 2020; Guo et al., 2021; Pons et al., 2021; He et al., 2022b; Shen et al., 2023), which are shown in Table 2. In the conclusion, future research should pay attention to the correlation between metagenomics and phenotype to identify disease-associated biomarkers and reveal the pathogenesis of NDDs.

Table 2.

Specialized software for virome analysis

Software	Current version	Functionality	Access link	Reference
viralFlye	0.2	Viral genomes assembly	github.com/Dmitry-Antipov/viralFlye
Metaviral SPAdes	1.1		github.com/ablab/viralVerify
			github.com/ablab/viralComplete
Haploflow	1		github.com/hzi-bifo/Haploflow	Fritz et al., 2021
PHAMB	1.0.1	Viral genomes binning	github.com/RasmussenLab/phamb	Johansen et al., 2022
vRhyme	1.1.0		github.com/AnantharamanLab/vRhyme
VirFinder	1.1	Viral sequences identification	github.com/jessieren/VirFinder	Ren et al., 2017
DeepVirFinder	1		github.com/jessieren/DeepVirFinder	Ren et al., 2020a
VirusSeeker	3		pathology.wustl.edu/virusseeker/index.htm	Zhao et al., 2017
KMCP	0.9.1		github.com/shenwei356/kmcp
ViromeScan			github.com/simonerampelli/viromescan	Rampelli et al., 2016
Prophage Hunter			pro-hunter.genomics.cn/	Song et al., 2019
MARVEL	0.2		github.com/laboratoriobioinformatica/MARVEL
VIP	0.2.0		github.com/keylabivdc/VIP	Li et al., 2016b
VirMiner			sbb.hku.hk/VirMiner	Zheng et al., 2019
PhaMer	1		github.com/KennthShang/PhaMer	Shang et al., 2022
VIBRANT	1.2.1		github.com/AnantharamanLab/VIBRANT
Phybrid	0.0.1		github.com/Strong-Lab/Phybrid	Glickman et al., 2021
Metaphinder			github.com/vanessajurtz/MetaPhinder	Jurtz et al., 2016
Virkraken	0.05		github.com/Strong-Lab/VirKraken	Glickman et al., 2021
VirSorter	1.0.6		github.com/simroux/VirSorter
VirSorter2	2.2.4		github.com/jiarong/VirSorter2	Guo et al., 2021
PhaGCN2	2	Virus classification	github.com/KennthShang/PhaGCN2.0	Jiang et al., 2023
PhaGCN			github.com/KennthShang/PhaGCN	Shang et al., 2021a
vConTACT2	2		bitbucket.org/MAVERICLab/vcontact2	Bin Jang et al., 2019
VPF-Class			github.com/biocom-uib/vpf-tools
CAT	5.2.3		github.com/dutilh/CAT	von Meijenfeldt et al., 2019
DeePhage	1	Distinguish virulent and temperate phage-derived sequences	github.com/shufangwu/DeePhage	Wu et al., 2021
Cenote-Taker 2	2.1.5	Virus discovery and sequence annotation	github.com/mtisza1/Cenote-Taker2
VADR	1.5		github.com/nawrockie/vadr	Schäffer et al., 2020
VAPiD	1.2		github.com/rcs333/VAPiD	Shean et al., 2019
VIGOR			www.jcvi.org	Wang et al., 2010
Crass	1.0.1	Host prediction	ctskennerton.github.com/crass	Skennerton et al., 2013
WIsH	1.1		github.com/soedinglab/wish	Galiez et al., 2017
PHP			github.com/congyulu-bioinfo/PHP	Lu et al., 2021
RaFAH	0.3		sourceforge.net/projects/rafah/	Coutinho et al., 2021
vHULK			github.com/LaboratorioBioinformatica/vHULK	Amgarten et al., 2022
DeepHost			github.com/deepomicslab/DeepHost	Ruohan et al., 2022
CRIPRCasFinder	4.3.2		crisprcas.i2bc.paris-saclay.fr	Couvin et al., 2018
VHM-net			github.com/WeiliWw/VirHostMatcher-Net
CRISPR Recognition Tool	1.1		www.room220.com/crt/	Bland et al., 2007
HoPhage			github.com/jie-tan/hophage	Tan et al., 2022
HostG			github.com/ KennthShang/HostG	Shang and Sun, 2021
CHERRY	2		github.com/KennthShang/CHERRY	Shang and Sun, 2022
PhaTYP	1	Lifestyle prediction	github.com/KennthShang/PhaTYP	Shang et al., 2023
PhagePred	1.1		github.com/songkai1987/PhagePred
BACPHLIP	0.9.6		github.com/adamhockenberry/bacphlip
PHCATS			www.phantome.org/PHACTS

Wet-Biology Validation in Metagenomics

With the aid of the data-mining of metagenomics, growing evidence indicates that GM dysbiosis is an important contributing factor to NDDs via MGBA. A plenty of studies illustrate that both depletion and enrichment of certain microorganisms might be associated with the NDD progression or even parthenogenesis. To uncover the causal correlations between the microbiome and the NDD phenotype and pinpoint key microbes, the precise advantage of functional validation is essential to complete the chain of evidence.

To this end, Chaudhari et al. (2021) carried out a number of trails, including strain identification and transplantation, to clarify how intestinal flora affects neurogenesis. Kim et al. (2021) performed fecal microbiota transplantation (FMT) from 5×fAD mouse in memory decline into the normal mice (5×fAD-FMT). Both of adult hippocampal neurogenesis and neurotrophic factor expression in the brain was decreased in the latter transplanted mice, representing memory impairment (Kim et al., 2021). A similar trail was done by Soriano et al. (2022), in which FMT derived from AD mice (AD-FMT) and healthy control mice were separately administered to wide-type mice suffering traumatic brain injury. Their results showed that AD-FMT administration can stimulate an elevated level of Muribaculum and a lower level of Lactobacillus johnsonii. In addition, the wide-type mice administered by AD-FMT showed increased activated microglia and macrophages, and larger lesion compared with AD-FMT mice (Soriano et al., 2022). Ju et al. (2022) successfully employed microbial colonization to verify that Helicobacter pylori also weakens the gut barrier and activates the Toll-like receptor (TLR) 4/myeloid differentiation factor 88 inflammation pathway by lowering the level of bile acid and p53.

The molecular biology skills, followed by with multi-omics analysis focusing on isolation and validation of causally important microbes or efficacious metabolites have been widely applied in the microbiology experiments. Typically, Ueda et al. (2021) found that Faecalibacterium prausnitzii (F. prausnitzi) decreases in patients diagnosed with mild cognitive impairment compared with the healthy controls. Two isolated live strains (Fp14 and Fp360) reduced Aβ-induced cognitive impairment in mice, and four specific orthologs were shared only by them when comparing genome with those of the 10 other isolates using whole-genome sequencing. Moreover, metabolome and RNA sequencing analysis with morphological observation in neuron provided the mechanistic insights about the relationship between F. prausnitzi efficacy, oxidative stress and mitochondrial function. However, the research paper was limited by the absence of a network analysis of connections between genes, transcripts, and metabolites (Ueda et al., 2021).

Likewise, Wang et al. (2022b) claimed that a functional probiotic (Lactobacillus plantarum MA2) isolated from traditional Chinese Tibetan kefir grains, can alleviate Aβ accumulation and cognitive impairments in AD rats by modulating the GM and glycometabolism. A metagenomic and metabolome study conducted by Sun et al. (2022c) revealed that indole-producing bacteria such as Prevotella oris may play a role in regulating the imbalance of tryptophan metabolism. Sun et al. (2022b) implemented a series of experiments, including western blot and immunofluorescence, to define the role of indole, which may serve as a new target for AD treatment. Namely, the authors found that the tryptophan-derived bacterial metabolite indole inhibited the nuclear factor kappa-B pathway in microglia via upregulation of the aryl hydrocarbon receptor, and subsequently reduced neuronal injury in an AD mouse model. Conversely, Pogue et al. (2022) presented that lipopolysaccharide reduced the expression of neurofilament light chain related to neurogenesis regulated by microRNA-30b-5p, which was enriched in hippocampal neurons. Esteves et al. (2023) used a mouse model of PD and cell cultures to test the prediction that microbiome-derived toxic metabolites promote neuroinflammation in PD. They found that β-N-methylamino-L-alanine, produced by microbiota, crosses the blood-brain barrier and enters the brain by weakening intestinal permeability. This caused neuron damage as a result of dysfunctional mitochondria. Recent studies (Wang et al., 2022c; Horvath et al., 2023) have provided a protocol involving culturomics, microbial colonization, and an artificial mouse model to identify the role of key GM and metabolites in the central nervous system based on a liquid chromatograph mass spectrometer/mass spectrometer system.

Gut Dysbiosis Occurring at the Onset of Neurodegenerative Diseases

PD

PD is a progressive NDD of the central nervous system and the most common cause of motor impairment (Zhang et al., 2022a). PD is caused by genetic factors and a variety of environmental factors, such as inflammatory bowel disease and infections (Dunn et al., 2019; Aziz et al., 2021). PD is mainly characterized by the deposition of α-Syn in the dense ventral part of the substantia nigra of the midbrain, which results in the formation of Lewy bodies and triggers motor dysfunction (Bhattarai et al., 2021; Hirayama and Ohno, 2021). Later neuroinflammatory processes finally contribute to the death of midbrain dopaminergic neurons (Horgusluoglu et al., 2017; Tansey et al., 2022). For patients with PD, they experience constipation before they experience movement problems (Marogianni et al., 2020). In addition, alterations in the gut microbiome composition have been observed in both patients with PD and mouse models of PD, including an increased abundance of Lachnospiracae or Lactobacillaceae and decreased abundance of Turicibacter compared with wild-type mice (Hill-Burns et al., 2017; Johnson et al., 2018; Shen et al., 2021; Tisza and Buck, 2021). Qian et al. (2020) identified 25 potential gene biomarkers from metagenomic species that can discriminate between patients with PD and the non-PD population, using bioinformation analysis and real-time polymerase chain reaction methods. PD staging and diagnosis has mainly relied upon liquid biopsies (Hampel et al., 2019; Liu et al., 2021a; Gong et al., 2022), but it is now conceivable that these bacterial markers could be used to assess disease severity in NDDs (Liu et al., 2022b; Toledo et al., 2022)

A crowd of studies have been published to provide the potential methods to alleviate or even treat PD. Challis et al. (2020) found that mice developed gastrointestinal symptoms, such as intestinal inflammation, after growing α-Syn fibrils in their duodenal. Transplantation of α-Syn fibrils in the gut of aged mice, and not younger mice, not only resulted in gastrointestinal deficits, but also caused motor deficits that were associate with α-Syn appearance in the midbrain (Challis et al., 2020). Notably, leucine-rich repeat kinase 2 (LRRK2) has been found to be a key gene in PD, as well as in Crohn’s disease (CD) (Lee et al., 2021). Pathogenic LRRK2 mutations have been associated with functional changes that may affect microglial responses, autophagy, and mitochondrial function (Tolosa et al., 2020). Moreover, the pro-inflammatory status of peripheral immune cells is activated by LRRK2 mutations (Kozina et al., 2018; Atashrazm et al., 2019). However, the multiple hit hypothesis suggests that PD phenotype induction by pathogenic LRRK2 mutations requires an inflammatory stimulus, such as inflammatory bowel disease or infections (Cabezudo et al., 2020). Another study found that one of the early-onset forms of PD is caused by mutations in the PINK1 gene that encodes kinase (Matheoud et al., 2019). Moreover, one study suggested that microbial metabolite-activated epithelial TLR signaling, as well as TLR2 and TLR4 signaling, might mediate the aggregation of α-Syn in PD (Gorecki et al., 2021). In addition to this, the activation of TLRs and SCFAs influences autophagic activity and regulates brain health (Shoubridge et al., 2021). In general, these data highlight the association between inflammatory bowel disease and PD, which shows patients’ gastrointestinal symptoms could be important to monitor. A detailed description can be found in Additional Table 1 ^{(388.9KB, pdf)} .

AD

AD is the most common form of cognitive impairment, and is characterized by the accumulation of Aβ protein and hyperphosphorylated tau (Griñán-Ferré et al., 2021). Mounting studies suggest gut microbiome alterations, or dysbiosis, contribute to Aβ deposition compared with normal controls (Vogt et al., 2017; Liu et al., 2020). Chen et al. (2021) found that, in AD, propagation of Aβ and tau fibrils might begin in the intestine before spreading to the brain via gut inflammation-activated C/EBPβ/δ-secretase signaling. This finding supports previous work of Chen et al. (2020), which found an association between gut dysbiosis and AD, and also between Aβ pathology and C/EBPβ/asparagine endopeptidase signaling activation. Furthermore, in that study, vagotomy was found to delay AD development. Additionally, subsequent studies found that the neuroinflammation mediated by NACHT-leucine-rich repeat and pyrin domain-containing protein 3 was aggravated by the presence of chronic colitis (Marogianni et al., 2020; Megur et al., 2020). Intriguingly, one study provided a new perspective and reported that Aβ aggregates and autophagy dysregulation were uniquely observed within neurons in five mouse models at early stages of AD development (Lee et al., 2022).

Another key outcome is the phylum-through-genus-wide differences of microbes derived from the AD gut. However, the screened strains with different abundance as target biomarkers show heterogeneity. AD mouse models have exhibited gut microbial dysbiosis compared with wild-type mice, as well as more abundant Proteobacteria, Bacteroidetes, and Ruminococcus, which trigger poly-unsaturated fatty acid-associated neuroinflammation and microglia activation (Zhang et al., 2017; Peng et al., 2018; Chen et al., 2022b; D’Argenio et al., 2022). Vogt et al. (2017) identified decreased Firmicutes, increased Bacteroidetes, and decreased Bifidobacterium in the GM of AD participants. Cammann et al. (2023) constructed a linear regression analysis and identified three independent genera associated with AD (Eubacterium fissicatena as a protective factor, Collinsella, and Veillonella as a risk factor).

Recently, researchers have proposed that gut dysbiosis, immune-mediated neuroinflammation, impaired autophagy, and proteinopathy create a vicious cycle in AD. First, intestinal dysbiosis disrupts the integrity of the intestinal barrier, which causes a leaky gut, and pathogenic bacteria and intestinal metabolites enter the systemic circulation through the vagus and enteric nerves (Paley, 2019). Second, immune-mediated chronic neuroinflammation triggers the transportation of neurotoxic misfolded proteins, which leads to neuronal death. In addition, intestinal and neuronal cell autophagy further exacerbates leaky gut (Chidambaram et al., 2022; Khedr et al., 2022; Wanapaisan et al., 2022). The role of gastrointestinal dysfunction has been increasingly discussed in the context of the pathogenesis of AD. Therefore, pyrin domain-containing protein 3 inflammasome and gut microbial metabolites are expected to become new targets for AD intervention.

ALS

ALS is an intractable neurological disease of the central nervous system that is characterized by the loss of motor function, including moving, speaking, and swallowing (Obrenovich et al., 2020). Mouse models of ALS have been found to exhibit changes in the gut microbiome compared with wild-type mice, and a reduced abundance of Butyrivibrio fibrisolivens (Wu et al., 2015). Previous work has shown that microbial-derived metabolites are associated with the expression of mitochondrial genes, such as NAM-encoding genes (Zeng et al., 2020; Boddy et al., 2021; Nicholson et al., 2021). High levels of NAM activate Sirtuin 1-mediated signaling pathways and protect neuronal cells from mitochondrial damage (Tribble et al., 2021; Zhao et al., 2021). In addition, the genetic cause of ALS is a mutation in the C9ORF72, which involves a GGGGCC hexanucleotide repeat. An increase in the number of GGGGCC repeats to more than 200 causes mitochondrial dysfunction, reduces endoplasmic reticulum-mitochondria contacts, and triggers neuronal death (Liu et al., 2016; Lall and Baloh, 2017; Balendra and Isaacs, 2018; Beckers et al., 2021). Decreased mitochondrial Ca²⁺ concentration and ATP production inhibit autophagy and mitophagy in neurons and renders them unable to remove damaged proteins, which ultimately hinders neurotransmission and weakens synaptic plasticity (Markovinovic et al., 2022). Additionally, research using C9orf72^−/− mice, which demonstrate heightened autophagy, has revealed that excessive endosomal TLR signaling causes inflammation (McAlpine et al., 2018). Notably, several studies about inflammatory disease in ALS mouse model have failed to include a systematic histopathological evaluation of inflammatory bowel disease (Niesler et al., 2021).

MS

MS is a complex neuroimmune disorder that occurs in both children and adults, and the cause of MS is largely unknown. However, environmental factors are known to play a dominant role in the development of MS (Olek, 2021). With the aid of high-throughput sequencing technology, recent studies have used metagenomics to explore the relationship between gut ecology affected by dietary habits and MS (Cantoni et al., 2022); the results showed that both pediatric-onset and adult-onset MS are associated with symptoms of intestinal disturbances. However, the intestinal composition of pediatric-onset MS and adult-onset MS is different. Namely, patients with pediatric-onset MS have a higher abundance of Methanobrevibacter and a lower abundance of Homolactic fermentation than controls (Mirza et al., 2022). High abundance of Sutterella and decreased abundance of Erysipelatoclostridium have also been found in patients with MS (Kishikawa et al., 2020). One integrative analysis revealed a negative correlation between Bacteroides thetaiotaomicron and T-helper 17 cells, which suggests that Bacteroides thetaiotaomicron may trigger an inflammatory response by regulating the ratio of Th17 cells. Gut microbiome data processed by Takewaki et al. (2020) showed that Streptococcus is most abundant in secondary progressive MS compared with normal controls. In addition, the level of gut microbial DNA mismatch repair genes and oxidative stress were higher in patients with secondary progressive MS than in patients with relapsing remitting multiple sclerosis (Dong et al., 2021; Ntranos et al., 2022). Subsequent researchers have differentiated MS disease stages by Streptococcus and have prevented MS progression by targeting oxidative DNA damage. Given that the above microbial dataset is limited by diet, which causes microbiome heterogeneity, the International Multiple Sclerosis Microbiome Study recruited the spouses of patients with MS as healthy controls in a metagenomic study, and showed that patients with MS had higher abundances of Akkermansia muciniphila (A. muciniphila). On the one hand, A. muciniphila has the ability to promote T-cell proliferation involved in MS (iMSMS Consortium, 2022). On the other hand, it maintains the integrity of the gut barrier to avoid inflammation due to intestinal metabolites (Chelakkot et al., 2018; Usuda et al., 2021). Low abundance of Blautia and Faecallibacterium were found in patients with MS, but Hungatella hathewayi was significantly enriched in patients with progressive MS and relapsing remitting MS in comparison with normal controls. These findings suggest that some microbe strains can be used to determine which disease stage of MS has developed. In addition, SCFA-producing bacteria can be used to differentiate the severity of MS, as these have been found to be negatively correlated with Multiple Sclerosis Severity Score (Routy et al., 2018; Galluzzo et al., 2021; iMSMS Consortium, 2022).

HD

HD is an inherited central nervous system disease (Tabrizi et al., 2019). Currently, efforts to determine the association between gut dysbiosis and HD using metagenomics has focused primarily on R6/1 transgenic mouse models (Kong et al., 2020). Huntingtin gene mutations generate toxic bodies called mutant huntingtin proteins, which cause mitochondrial dysfunction and contribute to the pathogenesis of HD (Tabrizi et al., 2020). Moreover, in that study, the authors found that ATP expression was significantly higher in HD mice plasma compared to controls. As a result, they suggested that gut microbes play a role in signaling transport along the gut-brain axis by altering plasma metabolite levels. For example, SCFAs derived from microbes can stimulate colon cells to produce more ATP, which is then released into the bloodstream (Kong et al., 2021). Recently, Kong et al. (2022) discovered that the fungal ecosystem in the intestines of HD mice differs from that of wild-type mice. The beta diversity of GM revealed a negative correlation between HD risk and the species of Malassezia restricta, Yarrowwia lipolytia, and Aspergillus in trend. Followed correlation network analysis also showed that Lactobacillus reuteri was negatively correlated with Malassezia restricta. Previous research has reported that γ-aminobutyric acid, a product of Lactobacillus reuteri, is decreased in the brains of patients with HD (Rosas-Arellano et al., 2018; Perez-Rosello et al., 2019; Barry et al., 2020). γ-Aminobutyric acid inhibits mitochondrial oxidative phosphorylation, which lowers the level of inflammatory factors (interleukin-1β) and thus alleviates inflammation (Fu et al., 2022). Therefore, bacterial and fungal interactions may be a new direction to explore in the context of the pathogenic mechanisms underlying HD and other NDDs.

Multi-Omics Studies and Pharmacokinetic Analyses

Viral metagenomics

Metagenome-based studies have greatly expanded our knowledge of the intestinal microbiota structure. Nevertheless, the low-resolution genome with a loss of gene structural elements has limited a full understanding of the interplay between dark matter in genomics and NDDs. The availability of PacBio HiFi and Oxford Nanopore sequencing technologies has enhanced the continuity of the microbial genome, which means that the genetic information is complete and without gaps. Recent studies have provided a high-quality gut microbial genome of Inner Mongolians using the PromethION and HiSeq platforms (Su et al., 2021; Jin et al., 2023). Beyond bacteria, viromes are also members of the gut ecosystem. Similarly, as an indispensable microbial constituent, the gut virome is a common risk factor in human disease (Vemuri et al., 2020; Bai et al., 2022; Cao et al., 2022; Spencer et al., 2022). Some studies have suggested that the risk of human disease in later life is associated with changes in the microbial ecosystem during infancy, including disease-specific viruses (Fang et al., 2020; Bikel et al., 2021). Given that many factors influence the human virome composition, such as breastfeeding, its assembly goes through different steps (Liang et al., 2020; Bushman and Liang, 2021; Liang and Bushman, 2021). In a recent example, Gregory et al. (2020) compiled the human Gut Virome Database using metagenomic approaches and identified the diversity of viromes. Mihindukulasuriya et al. (2021) explored the association between the gut virome and irritable bowel syndrome using multi-omics analysis. The authors observed differences in phage populations in patients with irritable bowel syndrome, which influenced the expression of host genes. In addition, due to the high genomic mutation rates of viruses, machine learning is essential for integrating multi-omics data with gut microbiome-mediated epigenetic regulation of brain disorders (Kaur et al., 2021).

Transcriptome

Metagenomics lacks transcript information from the host and gut microbes. In previous years, transcriptome was used to shed light on gene alterations associated with the neuroimmune system in NNDs (Chen et al., 2023b). To date, recent meta-transcriptomics research has similarly widened our perspective on viruses at the RNA-level, whereby the diversity and abundance of host viroids have been discovered (Lee et al., 2023). Moreover, single-cell RNA sequencing platforms with high-throughput and high-sensitivity not only provide fuller transcriptome data, but also reveal the cellular heterogeneity of brain (Tilocca et al., 2020; Liao et al., 2023). For this reason, single-cell RNA sequencing is capable to be used more finely explore the communication mode by which gut microbes interact with NDDs in comparison with NGS or TGS. A previous study conducted by Nguyen et al. (2020) revealed that APOE, one of the genetic factors underlying AD, is involved in the regulation of microglial cell responses. Recently, Seo et al. (2023) found that ApoE, a tau protein-mediated modulator of NDD, alters the GM structure to affect cellular responses. Using an integrated microbiome-single-cell RNA-sequencing approach, they found that perturbation of the intestinal flora activated the expression of microglia-associated genes. According to the results of a metagenome-wide association analysis, patients with PD have an altered bacterial abundance and gene expression (Wallen et al., 2022). As every method employed to analyze the myriad sequencing data has certain advantages and limitations, a growing body of review papers have been published for the comparison of intricate software established (O’Toole et al., 2017; Chiu and Miller, 2019; Sun et al., 2021; Cani et al., 2022; Meyer et al., 2022).

Metabolome

Metabolomics is a common method that can be used to reveal how gut microbiome dysbiosis affects NDD pathogenesis (Vascellari et al., 2020; Konjevod et al., 2021). For example, abnormal metabolic activity in a mouse model of AD has been found to be correlated with intestinal microbiota, including amino acid metabolism, neurotransmitter metabolism, and bile metabolism (Feng et al., 2022; Sun et al., 2022c). Likewise, metabolome-microbiome analysis has revealed that lipid and energy metabolism in patients with PD was influenced by the GM (Rosario et al., 2021; Pereira et al., 2022). Using a metagenomics and metabolomics integration analysis, Dunham et al. (2022) concluded that the level of Turicibacter was decreased and serotonin from gut bacteria was increased in 5×fAD mice and patients with AD. Similar findings were obtained in a metabologenomic analysis by Favero et al. (2022), who reported disordered production of gut-derived amino acids, such as dihydro-3-coumaric acid and desaminotyrosine. It is worth noting that intestinal metabolites, such as SCFAs, are not only associated with the occurrence of PD, but also with clinical severity (Chen et al., 2022c). However, these studies lack data from large human cohorts. To overcome this limitation, Xu et al. (2022) used multi-omics research tools to reveal the relationship between the gut microbiome and metabolic characteristics within a comparison of extremely long-lived individuals and their children. In another clinical study of three independent trails (one test study and two replication studies) using targeted serum metabolomics, fecal 16S rRNA and shotgun metagenomic sequencings, Liang et al. (2022a) found that hippocampal volume might be associated with acetic acids produced by Odoribacter. Notably, two studies have suggested that the risk of human disease in later life may be associated with changes in the microbial ecosystem during infancy, including disease-specific viruses (Fang et al., 2020; Bikel et al., 2021). Moreover, several studies have shown that changes in metabolism and in the microbiome can be used as biomarkers of obstructive sleep apnea (Moreno-Indias et al., 2015; Yoon et al., 2019; Zhang et al., 2021). All of the above studies have explored the molecular mechanisms underlying disease using multi-omics analyses.

Regarding the built-for-purpose tools, Gregory et al. (2020) compiled the human gut virome database using metagenomic approaches and identified the diversity of the virome. Pascal Andreu et al. (2023) developed gutSMASH, a software designed to identify core gut metabolic genes from metagenomic datasets and reveal the potential relationships between intestinal metabolic pathways and host health. Noecker et al. (2022) presented an R package named MIMOSA2 for purpose of model-based integrative analysis of ‘microbiome-metabolome datasets’. To summarize, the combined analysis of metagenomics with other omics is a powerful method to explore the molecular mechanisms underlying NDD onset and progression.

Pharmacokinetic analysis

During drug development, researchers use pharmacokinetic analysis to understand the metabolism and absorption processes of drugs (Li et al., 2019). Several studies have reported that most drugs can be metabolized by intestinal microorganisms, which alters drug bioactivity (Javdan et al., 2020; Wang et al., 2020b). For example, in a mouse model of PD, in vitro and in vivo tests have demonstrated that lanosterol 14 alpha-demethylase (CYP51) produced by gut microbes facilitates the absorption of FLZ, which is a novel drug in PD treatment that has neuroprotective effects (Shang et al., 2021b). In contrast, a recent study reported that gut bacteria reduced the bioavailability of levodopa (Bedarf et al., 2017; Maini Rekdal et al., 2019). These results indicate that multi-omics and pharmacokinetic analyses are essential for disease pathogen-associated marker mining and drug development.

Reshaping the Gut Microbiome to Mitigate Neurodegenerative Diseases

The intestinal ecosystem is influenced by a number of regulators, such as dietary pattern, probiotics, and exercise (Nandwana et al., 2022; Yao et al., 2022). While increasing evidence has demonstrated that reshaping the GM can alleviate CD symptoms (Verburgt et al., 2023), a pathogenic association between CD and NDD has also been proposed (Lee et al., 2021). Thus, rectification of the gut microbial composition may be a significant future therapeutic strategy for NDDs (Bonfili et al., 2021). Some potential strategies to combat NDDs by changing the gut microbiome are shown in Figure 3.

Figure 3.

Strategies used to reshape the intestinal microbiota.

Prebiotics, mediterranean diet and exercise alleviate the pathological symptoms of neurodegenerative diseases by regulating the gut microbiota, including pathological proteins aggregation, neuroinflammation and hypotha lamopituitary-adrenal axis disorders. Created with BioRender.com. APP: Amyloid-beta precursor protein; Bace: beta-site amyloid-beta precursor protein cleaving enzyme; CORT: corticosterone; CRH: corticotropin releasing hormone; HPA: hypotha lamopituitary-adrenal; IL: interleukin; LRRK2: leucine-rich repeat kinase 2; NE: noradrenaline; TLR: Toll-like receptors; TNF: tumor necrosis factor.

Dietary intervention

Dietary pattern is closely related to the development of NDDs, because this can modulate the gut microbiome, as has been found for Mediterranean, ultra-processed, and ketogenic diets (Włodarek, 2019; Martínez Leo and Segura Campos, 2020; Zhang et al., 2020; Solch et al., 2022). For example, a high-fat diet worsens the course of NDD (Chen et al., 2023a). A clinical study in patients with mild cognitive impairment found that low methionine intake worsens cognitive impairment (Xi et al., 2023). Via the gut microbiome, methionine from the diet is converted into hydrogen sulfide, which is a trophic factor for Th cell proliferation. Thus, the concentration of hydrogen sulfide and pro-inflammatory cytokines decrease, which inhibits neuroinflammation in the brain via the microbiota-gut-brain axis (Zhang et al., 2022c; Yue et al., 2023). These results indicate that older people should pay more attention to dietary habits and eat more vegetables, as plant proteins are usually lower in methionine than are animal proteins. Recent research in 5×fAD mice has shown that intermittent fasting can modulate the composition and metabolism of microbial communities, with neuroprotective effects against AD via an increased content of sarcosine and dimethylglycine compared to the colonization of sedentary controls, resulting in decreased pathogenic proteins in the brain (Fontana et al., 2021; Pan et al., 2022). In turn, recent research using data of meta-bologenomic signatures in people with pediatric CD has demonstrated that dietary therapy moderates inflammation by reshaping the gut microbiome. Verburgt et al. (2023) showed that successful dietary therapy in paediatric CD can correct both of compositional and functional dysbiosis towards to normal controls. In that study, successful dietary therapy elevated the relative abundances of Firmicutes and reduced the abundance of Proteobacteria. In summary, many studies have demonstrated that dietary therapy is associated with the risk of NDDs and longevity (Milošević et al., 2021; Wang et al., 2021; Shan et al., 2023).

Probiotics intervention

Lactobacillus and Bifidobacterium are beneficial microorganisms with a long history of use, and have been reported to have the potential to modulate NDD progression through gut microbiome intervention (O’Toole et al., 2017; Long-Smith et al., 2020; Peterson, 2020; Anderson, 2022; Zuo et al., 2022). Research by Sun et al. (2022a) has revealed that probiotic co-administration changed the microbial profile and the metabolism of GM, and alleviated PD symptoms. A clinical experiment conducted by Hall et al. (2023) revealed that prebiotic fibers derived from healthy subjects improved gastrointestinal symptoms and standard clinical evaluation scores in patients with PD. However, the limitation of that study was the lack of any clear explanation of the mechanism by which probiotics can prevent AD. In a recent metabologenomic analysis by Wang et al. (2022b), a functional probiotic, Lactobacillus plantarum MA2, was found to mitigate cognitive impairment by modulating the GM profile and glycometabolism. In a rat model of AD, MA2 improved the integrity of the intestinal barrier and prevented lipopolysaccharide leakage, which can induce the production of inflammatory cytokines via mediation by TLR 4/myeloid differentiation factor 88/NOD-like receptor thermal protein domain-associated protein 3 pathway. In a subsequent study, Abdelhamid et al. (2022) conducted a series of assays to explore the effect of probiotics in wild-type mice. The results revealed the same phenomenon as that observed in AD mice, whereby Bifidobacterium breve MCC1274 decreased AD-like pathologies, such as the low level of soluble hippocampal Aβ42 and phosphorylated tau levels, via the protein kinase B (Akt)/glycogen synthase kinase-3β signaling pathway. Moreover, a previous study showed that the glycogen synthase kinase-3β pathway is associated with intestinal barrier permeability (He et al., 2022a). This suggests that probiotics may alleviate AD pathology symptoms by influencing the GM structure and composition. In MPTP-induced mice, Li et al. (2022b) found that Bifidobacterium breve CCFM1067 treatment alleviated intestinal barrier damage and motor impairment. Further analysis showed that the neuroprotective effects of Bifidobacterium breve CCFM1067 might be mediated by a high expression level of SCFAs such as butyric acids, which blocked the leakage of pro-inflammatory cytokines. Despite A. muciniphila being the next generation of probiotics, some studies have also reported that high levels of A. muciniphila can exacerbate the development of human disease (Qin et al., 2012; Ijssennagger et al., 2015; Zhai et al., 2019; Cani et al., 2022). Therefore, probiotic intervention therapy for NDDs should also take the host’s pre-existing microbiome into account (Montgomery et al., 2020; Pluta et al., 2020).

Circadian rhythm disruption and exercise

One risk factor of AD is circadian rhythm disruption, which is might be caused by sleep disturbance (Homolak et al., 2018; Irwin and Vitiello, 2019; Li and Wang, 2021; Musiek and Ju, 2022). In addition, two studies have shown that circadian rhythms may be a contributing factor in NDD-associated gut dysbiosis (Li et al., 2021b; Cheng et al., 2022b). Nevertheless, exercise has been found to have a neuroprotective effect that helps to avoid circadian disorders (Gubert et al., 2020; Liang et al., 2022b). For example, Zhao et al. (2022) found that behavioral impairment improved in exercised D2-Cre mice (PD model). The experimental results indicated that neuron dysfunction in D2-Cre mice was repaired via inhibiting spontaneous excitatory postsynaptic currents, which is linked to the up-regulation of D2 dopamine receptor protein expression. Subsequent studies showed that exercise exerts a neuroprotective effect via exercise hormones that can prevent motor deficits in animal models (Mahalakshmi et al., 2020; Chen et al., 2022a; Zare et al., 2022). For example, one animal study demonstrated that irisin, which is a fragment of Fndc5, is released into the bloodstream after exercise (Boström et al., 2012). In later research, Kam et al. (2022) performed discovered that irisin can reduce the expression of α-Syn via the lysosomal pathway. Irisin can cross the blood-brain barrier; this raises the possibility that irisin alters the diversity of intestinal microbiota, which leads to the attenuation of PD motor symptoms via the microbiota-gut-brain axis. A study in a mouse model of ulcerative colitis revealed that irisin treatment relieved inflammation symptoms by increasing the diversity of intestinal microbiome (Huangfu et al., 2021); high levels of Deferribacteres were found in the gut flora in that study. However, the mechanism underlying irisin’s anti-inflammatory effect has not been further elucidated (Huangfu et al., 2021). Moreover, exercise increases the abundance of Firmicutes, which have been related to memory function (Kang et al., 2014; Feng et al., 2017; Koblinsky et al., 2023). Compared to AD mice that did not exercise, there was a significant decrease in the number of Aβplaques in AD mice following an exercise intervention (Abraham et al., 2019; Colombo et al., 2021). Another study discovered that exercise increased the abundances of beneficial microbes and intestinal membrane proteins (Yuan et al., 2022a). As a result, the healthy gut barrier avoided the translocation of bacterial lipopolysaccharide and further inhibited the inflammatory pathway (Marizzoni et al., 2020; Yuan et al., 2022a). Taken together, exercise holds promising potential as a therapy for NDD.

Conclusion and Future Directions

A huge body of metagenomic analysis evidence has demonstrated there to be a causal link between GM and the onset and progression of NDDs. In this review, we have discussed the crosstalk between the gut microbiome and AD, PD, ALS, MS, and HD, five common NDDs. In addition, we found that only a couple of studies have focused on other NDDs, such as ataxia telangiectasia, which likewise pose a hazard to human health and longevity. Numerous microbial genomes are now available for clinical use thanks to the success of clinical metagenomic research. Compared with traditional microbiological techniques, clinical metagenomics more quickly generate reports with the support of high-throughput sequencing technology, especially since the emergence of third-generation sequencing technology, which means that more accurate information for patient care can be provided. Nevertheless, clinical metagenomic results are limited by economic costs and sample contamination, which can lead to low sensitivity of pathogen detection and a high rate of false-negative results. Another disadvantage of metagenomic analysis is that it demands high staff and time resources, as the bioinformatic analysis needs to be performed manually. In the future, machine learning will be able to address these problems (Álvarez et al., 2019). The storage and confidentiality of patient metagenomics data is also a concern for researchers. Non-bacterial components of the gut microbiome profile, such as viruses, also play major roles in NDDs. Using a longitudinal analysis, Bjornevik et al. (2022) revealed that Epstein-Barr virus might be the leading risk factor for the development of multiple sclerosis. With the aid of the availability of massive metagenomic sequencing datasets and powerful data-mining tools, researchers are still unraveling such mysteries of the gut virome (Johansen et al., 2022). Li et al. (2022a) have made significant contributions to the incomplete reference databases on the human gut and have explored the potential functions of archaeal virome in NDDs. However, a major challenge for uncovering the relationship between the gut virome and NDDs is that there is a lack of software able to analyze viral communities. Shen et al. (2023) developed a novel k-mer-based tool named KMCP to address such issue, which performs taxonomic profiling of viral populations. As a potential tool for deciphering disease mechanisms, multi-omics is widely used in pathogen-biomarker mining and drug discovery for NDD.

There are many strategies by which to improve the symptoms of NDDs through remodeling the gut microbiome. Case studies have reported improvements in the cognition and memory of patients with AD following the fecal transplant (Hazan, 2020; Park et al., 2021). In addition, sodium oligomannate is the first drug to treat AD that targets the GM. A Phase III clinical study demonstrated that the sodium oligomannate reliably and successfully reduced AD patients’ cognitive symptoms (Wang et al., 2019, 2020a; Syed, 2020; Xiao et al., 2021). Bacteriophages, as components of the intestinal virosphere, are also suitable for the treatment of many chronic diseases via reversing imbalances of the bacterial ecosystem (Shuwen and Kefeng, 2022). Furthermore, most of the published studies have reported that phage therapy could be applied in NDD treatment by a display vaccine that targets abnormal proteins in the brain or motor neurons. A schematic of the phage display vaccine is shown in Additional Figure 1 ^{(219.3KB, tif)} . Undoubtedly, NDD drug development is still an integral module in targeted therapy. One of the major challenges in anti-NDD drug development is that the biological activity of drugs is reduced by the GM. Nanomaterials may be an emerging solution for maximizing the efficacy of drugs in patients with NDDs (Cheng et al., 2022a). This review has some limitations. First, the discussion of gut metabolites is not sufficiently complete, and only some metabolites are mentioned. In addition, the phenomenon of NDD-associated gut dysbiosis has not been fully elucidated and needs to be studied in more detail.

Additional files

Additional file 1: Open peer review reports 1 ^{(93.6KB, pdf)} –2 ^{(93.6KB, pdf)} .

Additional Figure 1 ^{(219.3KB, tif)} : A schematic of the phage display vaccine for neurodegenerative disease therapy.

Additional Figure 1

A schematic of the phage display vaccine for neurodegenerative disease therapy.

First, DNA library from the healthy individuals is transformed into a phage to generate a phage display library. Second, the phage library can bind specifically to the antigen on the plate. After elution, non-characteristic phages are eluted and phages displaying specific antibodies are retained. Then, the mice are injected with purified plasmids to degrade the pathological proteins aggregated within the neuronal cells into peptides. Finally, the purified phage was amplified for the next phage display. Created with BioRender.com. NDD: Neurodegenerative diseases.

Additional Table 1 ^{(388.9KB, pdf)} : Potential pathogen-association signaling biomarkers in neurodegeneration disease.

Additional Table 1

Potential pathogen-association signaling biomarkers in neurodegeneration disease

Funding Statement

Funding: This work was financially supported by the National Natural Science Foundation of China, No. 32002235 (to MT) and the Science and Technology Foundation of Taian of Shandong Province, No. 2020NS216 (to XL).