A meta-study of the microbiome and related dysfunction of the brain reveals distinct microbial signatures in the microbiota-gut-brain axis

doi:10.21203/rs.3.rs-5311059/v1

Download PDF

Article

A meta-study of the microbiome and related dysfunction of the brain reveals distinct microbial signatures in the microbiota-gut-brain axis

https://doi.org/10.21203/rs.3.rs-5311059/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

The gut microbiome has emerged as a potential contributor to neurological disorders, with growing evidence linking microbial alterations to conditions such as Alzheimer's disease (AD), Parkinson's disease (PD), and autism spectrum disorder (ASD). However, a comprehensive understanding of shared and distinct microbial signatures across these disorders remains elusive. In this meta-analysis, we re-analyzed 31 16S rRNA gene amplicon sequencing datasets to investigate gut microbiome alterations in AD, PD, and ASD. Using a unified bioinformatic pipeline and robust statistical approaches, we identified both common and disorder-specific microbial signatures. While alpha diversity was significantly altered only in PD, beta diversity analysis revealed consistent compositional changes across all disorders. The genera Blautia and Bacteroides emerged as shared markers, showing differential abundance in AD, PD, and ASD, albeit with varying directions of change. PD exhibited the most distinct microbial profile, with 19 genera showing PD-specific alterations, including enrichment of Akkermansia and depletion of Faecalibacterium. Network analysis revealed complex, disorder-specific patterns of microbial interactions, with PD showing the highest number of altered microbial associations. These findings provide a nuanced picture of gut microbiome alterations across neurological disorders, highlighting potential common mechanisms and disease-specific signatures that may contribute to pathogenesis or serve as diagnostic biomarkers.

Biological sciences/Microbiology

Health sciences/Health care

The microbiota-gut-brain axis (MGBA), also referred to as the gut-brain axis (GBA), is a rapidly advancing concept that describes the intricate bidirectional communication system between the central nervous system (CNS) and the enteric nervous system (ENS)[1][2]. This complex network links the brain’s emotional and cognitive centers with peripheral intestinal functions, establishing a connection between the brain and the gut[3]. Recent studies have further highlighted the critical role of gut microbiota in shaping this cross-talk between the gut and the brain[4]. Accumulating evidence suggests that changes or imbalances in gut microbiota composition and related microbial metabolite production are associated with disorders that may affect the CNS including neurodegenerative diseases such as Parkinson's disease (PD)[5] and Alzheimer's disease (AD)[6], and neurodevelopmental conditions such as autism spectrum disorder (ASD)[7]. The possible pathogenesis of these diseases is related to aberrant of normal nerves.

While these studies offer valuable insights into the connection between gut microbiota and brain-related disorders, there are still limitations and inconsistencies across different studies[8]. For instance, microbial signatures associated with each disorder often vary between studies, making it challenging to identify reliable and reproducible microbiome-based biomarkers. Some studies report contradictory findings regarding the abundance of specific taxa in patients compared to controls[9], leading to a lack of consensus on key microbial alterations associated with each condition. Moreover, the absence of a unified, systematic analysis of MGBA across multiple brain-related diseases hampers the understanding of the common and unique microbial alterations associated with these disorders. This fragmented approach restricts our ability to identify robust microbial signatures that could serve as reliable biomarkers or therapeutic targets.

To address these gaps and further elucidate the potential of gut microbiota as biomarkers for neurological diseases, we conducted a comprehensive meta-analysis of 31 MGBA-related studies utilizing 16S rRNA gene amplicon sequencing data[10]. Our objective was to identify both shared and disorder-specific microbial signatures associated with AD, PD, and ASD, and to gain insights into the potential mechanisms through which gut microbiota may impact brain health and disease [11]. By employing a unified bioinformatic workflow and rigorous statistical approaches, we aimed to overcome the limitations of individual studies and provide a more comprehensive understanding of the role of gut microbiota in neurological disorders.

This meta-analysis not only examines the differences in microbial composition and diversity across these neurological conditions but also investigates the complex interactions within the microbial communities through network analysis. In doing so, we aim to uncover both consensus and disease-specific microbial signatures that may serve as potential biomarkers or therapeutic targets. Our findings provide a nuanced understanding of the gut microbiome landscape across neurological disorders, highlighting both shared features and disorder-specific signatures that may shed light on the complex relationship between the gut microbiome and neurological health.

Meta-analysis dataset composition and quality control

To explore the role of the gut microbiome in the MGBA across multiple neurological disorders, we conducted a comprehensive meta-analysis of publicly available 16S rRNA gene amplicon sequencing datasets. We searched relevant studies in public data repositories, such as the NCBI Sequence Read Archive (SRA), European Nucleotide Archive (ENA), and MG-RAST. Our search criteria focused on case-control studies that compared the gut microbiome of patients with AD, PD, or ASD to healthy controls. After conducting a thorough screening and quality control process, we included 31 datasets in our meta-analysis: 6 AD studies, 11 PD studies, and 14 ASD studies (Details see Methods Study Selection). These studies encompass diverse populations from various regions, including Europe, North America, and Asia. The sample sizes ranged from 20 to 507 individuals, with a total of 4,049 samples across all datasets. Detailed information for each study—such as sample size, country of origin, 16S rRNA gene region targeted (e.g., V3-V4, V4), specific disease (AD, PD, or ASD), and project accession numbers are provided in Supplementary Table 1.

Figure 1 illustrates the number of samples included in the meta-analysis for each neurological disorder. Among the three disorders, ASD studies contributed the largest number of samples, with 14 datasets. PD studies followed closely, with 11 datasets, while AD studies comprised 6 datasets. It is important to note that the number of datasets included for each disorder may not fully reflect the extent of research in the field, as our stringent inclusion criteria may have excluded some relevant studies. Nevertheless, the selected studies represent a comprehensive collection of high-quality, case-control investigations into the gut microbiome alterations associated with these neurological disorders, providing a solid foundation for a robust meta-analysis.

To ensure consistency and comparability across datasets, we implemented a standardized bioinformatic pipeline to process the raw 16S rRNA gene sequencing data. First, we downloaded the raw fastq files from the respective repositories and performed quality control using the FastQC tool. We then employed the DADA2 pipeline[12] for denoising, filtering, and merging the paired-end reads, as well as for removing chimeric sequences. DADA2 pipeline was selected for its high accuracy in resolving amplicon sequence variants (ASVs) and its ability to generate reproducible results across diverse studies.

After processing the raw data through the DADA2 pipeline, we identified 310,758 unique ASV across all datasets. The number of ASVs per study ranged from 500 to 70,465, with a median of 5,233 ASVs. Then we aggregated the ASVs at the different taxonomic levels using the Silva v138 reference database[13]. These ASVs represented a wide range of bacterial taxa, spanning 58 phyla, 148 classes, 345 orders, 533 families, and 1,603 genera. The most prevalent phyla across all datasets were Firmicutes, Bacteroidetes, Actinobacteria, and Proteobacteria, which is consistent with previous studies characterizing the human gut microbiome[14].

Alpha and beta diversity profiles of the gut microbiome in the MGBA context

We first evaluated the differences in microbial profiles between three neurological disorders and healthy controls using alpha diversity metrics. Several commonly used indices were calculated, including observed species and Chao1 for richness[15][16], Shannon and Simpson indices for diversity[17][18], and Pielou for evenness[19]. Additionally, we examined dominance patterns and assessed the abundance of rare taxa using measures such as core abundance and rare abundance indices[20][21].

The alpha diversity results, shown in Figure 2, were analyzed through a random-effects meta-analysis pooling estimates across studies. No statistically significant differences were found in alpha diversity measures between patients and controls for AD (p = 0.47) and ASD (p = 0.59), suggesting that the overall richness and evenness of the gut microbiome may not be consistently altered in these disorders [22]. In contrast, PD studies revealed significant differences in alpha diversity compared to controls. Specifically, PD patients showed significantly higher richness, as reflected by a greater number of observed species (mean difference = 0.12, 95% CI: 0.02~0.22, p = 0.02), and higher Chao1 index (mean difference = 0.11, 95% CI: 0.01~0.21, p = 0.02). This indicates that the gut microbiome of PD patients harbors more taxa compared to healthy individuals. Moreover, the rare abundance index was significantly elevated in PD (mean difference = 0.21, 95% CI: 0.07~0.35, p = 0.003), while the core abundance index was lower (mean difference = -0.21, 95% CI: -0.39~-0.03, p = 0.02), indicating an increased prevalence of low-abundance taxa in the PD gut microbiome. These findings suggest that the gut microbiome in PD is characterized by a more diverse and even community, with a higher representation of rare species[23]. Previous studies have also reported this increased alpha diversity in PD, which may be linked to the altered gut environment and impaired gut motility associated with the disease[24].

Next, we investigated beta diversity to assess differences in microbial community composition between samples. Bray-Curtis dissimilarities were calculated and visualized using principal coordinates analysis (PCoA). Across the three disorders, both disease status and study contributed significantly to the variance in beta diversity (PERMANOVA, p < 0.001). However, the “study” factor (e.g., data collected from different labs) had a more substantial impact on the overall gut microbiome composition than disease status. This may be due to study-specific variations in the dataset, such as geographical location, sequencing platform, or DNA extraction method. The effect sizes for “study” (AD: R2 = 0.17; PD: R2 = 0.28; ASD: R2 = 0.28) were consistently larger than those for disease status (AD: R2 = 0.011; PD: R2 = 0.008; ASD: R2 = 0.003) (Figure 3). This finding highlights the importance of controlling for technical and geographical variation in microbiome studies and is consistent with previous reports on the strong influence of study-specific factors on microbiome composition[25][26].

To further explore the impact of disease status on beta diversity while controlling for study-specific effects, we performed a constrained analysis of principal coordinates (CAP) using “study” as a conditioning variable to mitigate variation introduced by different research centers. As is shown in Figure 4, we observed significant differences in microbial community composition between patients and healthy controls across all three neurological disorders (permutation test, p < 0.01). Additionally, a clear separation between patient and control groups along the first principal coordinate axis, which explained 16.56% of the variation for AD, 17.79% for PD, and 23.58% for ASD. These results suggest that neurological disorders are associated with specific alterations in gut microbiome composition. For example, in Parkinson's disease, Scheperjans et al. reported a significant reduction in the relative abundance of Prevotellaceae in PD patients compared to healthy controls, which was correlated with the severity of motor symptoms[27]. However, the modest proportions of variance explained by disease status indicate that additional factors: such as diet, medication, or host genetics--may also play a significant role in shaping the gut microbiome in these disorders. This multifactorial influence on gut microbiome composition in neurological disorders has been emphasized in recent reviews[28][29].

Consensus and specific microbial signatures for neurological diseases

We next investigated the microbial signatures associated with each neurological disorder, aiming to identify both consensus and disease-specific microbial signatures. We leveraged a two-step approach combining different statistical methods. First, we analyzed each of the 31 datasets independently using the ALDEx2 method [30], which has been recommended by previous benchmarking study[31] for its conservative and reliable performance in microbiome differential abundance testing[31]. Then, to identify consensus and disease-specific microbial signatures, we focused on taxa that were consistently differentially abundant across multiple independent studies of the same disorder. This lead to identify robust microbial alterations that were replicable across different cohorts, reducing the impact of study-specific confounding factors.

Figure 5 revealed both shared and disorder-specific microbial signatures across PD, AD, and ASD. The genera Blautia and Bacteroides exhibited differential abundance in all three disorders, suggesting potential common microbial signatures relevant to multiple neurological conditions. However, it is crucial to note that the direction of change varied across disorders. For instance, Blautia was depleted in PD but enriched in AD and ASD, reflecting the complex and disease-specific nature of gut microbiome alterations in neurological disorders[32][33].For instance, Blautia, known for producing short-chain fatty acids (SCFAs), particularly butyrate[34], was depleted in PD but enriched in AD and ASD. Butyrate's neuroprotective and anti-inflammatory properties[35] suggest that Blautia's varying impact on the microbiota-gut-brain axis may be context-dependent and disease-specific. Similarly, Bacteroides, crucial for carbohydrate metabolism and gut homeostasis[36], showed alterations linked to changes in gut permeability and systemic inflammation[37], known to affect the gut-brain axis. These findings align with previous studies, which reported alterations in Bacteroides abundance in AD and PD, respectively[32][38], underscoring the importance of these microbial changes in neurological disorders.

Several genera showed significant differences in two of the three disorders, suggesting potential shared pathways in neurological conditions. UBA1819, Lachnoclostridium, and Bifidobacterium showed significant changes in both PD and AD, potentially indicating common mechanisms in neurodegenerative processes. Bifidobacterium, extensively studied for its probiotic properties and potential neuroprotective effects[39], may play a role in these processes through the production of neuroactive compounds and immune system modulation[40]. Roseburia, Lachnospiraceae UCG-004, Lachnospiraceae ND3007 group, Lachnospira, and Agathobacter were altered in both PD and ASD, hinting at common mechanisms despite the different nature of these disorders. Notably, Roseburia, a butyrate-producing genus crucial for gut barrier integrity and with anti-inflammatory properties[41], suggests a potential shared pathway involving gut barrier function and neuroinflammation. Additionally, Clostridium sensu stricto 1 and [Ruminococcus] torques group were differentially abundant in both AD and ASD, indicating possible shared aspects in cognitive and neurodevelopmental disorders.

Among the disorders studies, PD had the most distinct microbial profile, with 19 genera showing PD-specific alterations. Notably, Akkermansia was significantly enriched in PD patients, a finding consistent with previous studies[38]. This enrichment has been associated with increased gut permeability and inflammation, which are known features of PD pathophysiology[42]. Conversely, Faecalibacterium was depleted in PD patients, potentially contributing to the pro-inflammatory status observed in PD due to its role in producing the anti-inflammatory short-chain fatty acid butyrate[43]. Additional PD-specific alterations were seen in genera such as Fusicatenibacter, Ligilactobacillus, and Porphyromonas, indicating widespread reshaping of the gut microbial community in PD.

In contrast, AD showed relatively few unique microbial alterations, with only three genera specifically altered in AD patients. Among these, Eggerthella was enriched in AD patients, contributing to the growing evidence of specific microbial alterations linked to AD pathogenesis. Eggerthella has been associated with the production of toxic metabolites and low-grade inflammation[44], which could potentially contribute to AD pathogenesis through the gut-brain axis. This finding is consistent with a study by Vogt et al., which also reported an increase in Eggerthella in AD patients[24]. Eggerthella is known for its ability to produce amyloids, which may contribute to the amyloid accumulation characteristic of AD[45]. The other two genera, Alistipes and [Ruminococcus] gnavus group, also exhibited AD-specific alterations. Alistipes has been linked to gut inflammation and mental health disorders in previous studies[46], while the [Ruminococcus] gnavus group has been associated with inflammatory conditions and alterations in gut permeability[47]. These taxa were also identified in a study by Zhuang et al., which found them to be differentially abundant in AD patients compared to healthy controls[48].

ASD exhibited an intermediate level of microbial specificity, with 10 genera uniquely altered. Notably, Escherichia-Shigella and Collinsella were found to be enriched in ASD patients, findings that are consistent with previous studies[49]. The enrichment of these bacteria, along with alterations in other genera like CAG-352, supports the growing evidence for gut microbiome involvement in ASD and suggests that microbial alterations may contribute to the heterogeneous nature of the disorder.

Consensus and specific microbial network structure for neurological diseases

To explore the complex microbial interactions and co-occurrence patterns in the context of neurological disorders, we performed a network analysis using the SpiecEasi (Sparse Inverse Covariance Estimation for Ecological Association and Statistical Inference) method[50]. We first fit networks for both the disease and control groups within each study (as illustrated in Figure 6) and then compared the consensus and unique edges for each disease to identify disease-specific alterations of microbial association.

It was shown that two edges were consistently altered across all three neurological disorders (PD, AD, and ASD): the Coprococcus -- Lachnospiraceae FCS020 group and Gemella – Streptococcus connections. The presence of these shared alterations suggests potential common mechanisms within the gut microbiome that may play a role in various neurological conditions. The first connection involves known butyrate producers, which are crucial for gut health and neuroprotection [51], while the second connection involves genera that, when altered, have been associated with inflammatory conditions [52]. Disruptions in these microbial interactions could impair butyrate production and contribute to systemic inflammation, both of which are implicated in the pathogenesis of multiple neurological disorders [53].

We summarized these edge alternations in Figure 7. PD exhibited the most extensive changes, with 199 disease-specific edges, indicating widespread disruption of microbial interactions in the PD gut microbiome. AD demonstrated 63 disease-specific edges, while ASD had 53 unique edges. These findings highlight that all three disorders involve distinct microbial network alterations compared to the controls, with PD showing the most substantial changes, followed by AD and ASD.

Interestingly, PD and AD shared the greatest number of common alterations, with 12 shared edges. This overlap suggests that there may be common microbial mechanisms or pathways between these two neurodegenerative disorders. This could reflect certain shared pathophysiological features or common gut-brain axis that play a role in both PD and AD [54].

The identification of these disease-specific and shared microbial network alterations provides insights into the intricate interplay between gut microbes in neurological disorders. The larger number of PD-specific alterations, in particular, highlights the profound impact of this disease on gut microbial interactions. Moreover, the shared microbial network changes across all three disorders, as well as the significant overlap between PD and AD, suggest the existence of common gut-brain axis pathways that may contribute to multiple neurological conditions.

This meta-analysis provides a comprehensive overview of the gut microbiome landscape across three major neurological disorders: AD, PD, and ASD. By employing a unified bioinformatics workflow and rigorous statistical approaches, we identified both shared and distinct microbial alterations associated with these conditions. Our findings reveal complex patterns of microbial dysbiosis, highlighting both common across neurological disorders and condition-specific changes.

Our analysis revealed several key findings. Firstly, we identified shared microbial signatures across all three disorders, particularly involving the genera Blautia and Bacteroides. These common alterations suggest potential shared pathways in the microbiota-gut-brain axis across different neurological conditions. For instance, Blautia, known for its role in producing short-chain fatty acids (SCFAs) like butyrate[34], showed varying patterns of abundance across disorders, highlighting the complex and context-dependent nature of microbial influences. Secondly, we observed disorder-specific microbial signatures. PD exhibited the most distinct microbial profile, with 19 genera showing PD-specific alterations, including enrichment of Akkermansia and depletion of Faecalibacterium. These changes align with current understanding of PD pathophysiology, particularly regarding increased inflammation and altered gut permeability[42][43]. AD showed fewer unique alterations, but the enrichment of Eggerthella and its potential role in amyloid production[45] provides an intriguing link to AD pathogenesis. ASD demonstrated an intermediate level of microbial specificity, with alterations in genera like Escherichia-Shigella and Collinsella potentially contributing to the gut inflammation and permeability changes observed in ASD[44].

It is crucial to acknowledge the various confounding factors that could affect the microbial profile in these neurological disorders. Diet, antibiotic use, and lifestyle factors can significantly influence gut microbial composition and function. For instance, a study by David et al. demonstrated rapid and reproducible alterations in the human gut microbiome in response to short-term dietary changes[55]. Similarly, antibiotic use has been shown to have long-lasting effects on gut microbial diversity and composition, as reported by Jernberg et al.[56]. Furthermore, factors such as exercise, stress, and sleep patterns can also modulate the gut microbiome, as reviewed by Karl et al.[57]. However, it's important to note that only limited datasets in our meta-analysis included comprehensive information on these potential confounders, which represents a limitation of our study and highlights the need for more detailed metadata collection in future microbiome studies.

Our study focused on 16S rRNA gene sequencing, but we acknowledge that metagenomic sequencing could offer additional insights into species, strain-level information, and functional potential. In the context of the gut-brain axis, metagenomic studies have revealed crucial insights, such as specific microbial pathways associated with depression[58] and alterations in microbial genes related to neurotransmitter synthesis in autism spectrum disorder[59]. Metagenomics has also enabled the discovery of novel microbial species potentially relevant to neurological disorders[60]. However, integrating 16S and metagenomic data remains challenging due to differences in resolution and information types. Future research should focus on developing robust methods for multi-omic data integration[61], combining metagenomics with other omics approaches to elucidate the complex interactions between microbial communities, their metabolic outputs, and host physiology in neurological health and disease.

In conclusion, our meta-analysis reveals both shared and disorder-specific gut microbiome alterations across AD, PD, and ASD. These findings lay the groundwork for future investigations into the role of the gut microbiome in neurological health and disease. They highlight consistent patterns across diverse populations and methodologies, potentially informing the development of microbiome-based diagnostic tools and therapeutic interventions.

Study selection

To identify relevant studies for our meta-analysis, we conducted a comprehensive search of the Sequence Read Archive (SRA) database of the National Center for Biotechnology Information (NCBI). We used the following search terms: ("Microbiota" OR "Microbiome") AND ("Parkinson" OR "Alzheimer" OR "Autistic Spectrum Disorder"). These search terms were chosen to capture studies investigating the gut microbiome in the context of three major neurological disorders: Parkinson's disease (PD), Alzheimer's disease (AD), and autism spectrum disorder (ASD).

The initial search identified 54 studies. To refine this list for our meta-analysis, we applied three inclusion criteria. First, we included only a case-control design, comparing the gut microbiome of patients with one of the three neurological disorders to healthy controls. Second, we required that the studies used stool samples. Third, only data from the 16S rRNA gene were selected. These criteria ensured consistency in experimental design and minimized unwanted variations across the datasets. To further ensure the quality and comparability of the data, we applied several exclusion criteria. We excluded studies with a total sample size of fewer than 10 individuals across both case and control groups, as small sample sizes may lack of statistical power needed to detect meaningful differences in microbial composition.

16S rRNA gene sequencing data processing

To process the raw 16S rRNA gene sequencing data from the selected studies, we utilized a standardized bioinformatic pipeline based on the DADA2 workflow[62]. This pipeline includes several critical steps to ensure the quality and accuracy of the sequence data, such as quality filtering, trimming, error correction, and taxonomic assignment.

The pipeline began with quality control of sequencing reads. We used FastQC to visually inspect quality profiles and trim low-quality bases at the end of reads. In addition, we removed the first 10 nucleotides of each read to eliminate potential adapter contamination or errors from low-sequence diversity. Next, we employed the DADA2 to infer amplicon sequence variants (ASVs) from the filtered and trimmed reads. Taxonomic classification of ASVs was performed using the Silva 138.1 reference database. For quality control, we excluded ASVs classified as chloroplasts or mitochondria, as well as those present in fewer than 1% of the samples. ASV counts were normalized to the median sequencing depth across all samples, followed by a log transformation.

Diversity analysis

We employed the alpha method from the microbiome package in R to calculate various indices for assessing microbial community composition. For richness, we computed the observed species count and Chao1 index. Diversity was evaluated using multiple indices, including the inverse Simpson index, Gini-Simpson index, Shannon index, and coverage. To assess evenness, we applied the Camargo index, Pielou index, Simpson index, Evar index, and Bulla indices. For dominance, we employed several measures, including the Berger-Parker index (DBP), McNaughton's index (DMN), absolute and relative dominance, Simpson dominance, core abundance dominance, and Gini dominance index. Lastly, to evaluate rareness, we calculated indices such as log-modulo skewness, low abundance, and rare abundance.

After calculating the alpha diversity indices for each sample within each dataset, we employed Agresti's log-generalized odds ratio method to estimate the summary effect size and associated standard error[63]. This approach was selected for its robustness in handling ordinal data and its capacity to provide a meaningful measure of association between diversity indices and disease status across heterogeneous samples[64].

Subsequently, we conducted a comprehensive meta-analysis across all datasets using the metan function[65], implementing a random-effects model. This model was chosen for its ability to account for both within-dataset and between-dataset variability, thus providing a more nuanced understanding of the data. The random-effects approach assumes that the true effect sizes may vary between studies, allowing for a more conservative and generalizable interpretation of results[66]. This method is particularly advantageous when analyzing data from diverse sources, as it helps to mitigate the impact of study-specific factors and methodological differences. By utilizing this approach, we were able to synthesize data from multiple studies, providing a robust and comprehensive assessment of alpha diversity patterns across neurological disorders while accounting for the inherent heterogeneity between studies.

Beta diversity of microbial communities was assessed using a combination of ordination analysis and statistical testing. First, we performed Total Sum Scaling (TSS) normalization on the microbial abundance data to account for variations in sequencing depth across samples. Bray-Curtis dissimilarity indices were then calculated to quantify the differences in microbial community composition between samples. To visualize these dissimilarities, we employed Classical Multidimensional Scaling (MDS), or Principal Coordinates Analysis (PCoA). We performed Permutational Multivariate Analysis of Variance (PERMANOVA) from the vegan package in R[67]. This non-parametric method partitions variability in the distance matrix based on sources of variation and uses permutation tests to assess the significance of those partitions.

Differential abundance analysis

We conducted differential abundance analysis using the ALDEx2 package in R. ALDEx2 uses a Dirichlet-multinomial model to infer abundance from counts, and then performs statistical tests on the resulting distributions. We used the ALDEx2 implementation of Welch's t-test to evaluate differences in bacterial abundances between disease and control conditions. To control the false discovery rate (FDR) and account for multiple tests, we used the Benjamini-Hochberg adjustment within the ALDEx2 framework. Bacteria with adjusted p-values below 0.05 were considered significantly differentially abundant. To enhance the robustness of our findings, we incorporated a cross-study validation approach. For each disease, we identified bacteria that consistently exhibited significant differential abundance across multiple datasets related to the same condition.

Network Analysis

We employed the SpiecEasi (Sparse Inverse Covariance Estimation for Ecological Association and Statistical Inference) method to construct microbial co-abundance networks for each dataset. Separate networks were fitted for the disease and control groups within each study. To identify disease-associated microbial associations, we compared the edges (connections between microbes) present in the disease network with those in the control network for each dataset. Edges unique to the disease network were classified as differential edges for that study. To ensure robustness and reproducibility, we implemented a cross-study validation approach. An edge was considered a true disease-associated differential edge if it consistently appeared as a differential across multiple datasets for the same disease.

Availability and Implementation

The data and code are available at https://github.com/jcfslyr/GBAMetaAnalysis

Author contribution

XX, JD, and YL conceived the study. YL led the method development and data analysis with input from XX and JD. YL, GL and XX led the evaluation of the method with input from all authors. XX, SL, PD, DL and JD guided evaluation of the method. YL, XX, PD, TL, and JD wrote the manuscript with input from all co-authors. All authors read and approved the final version of the manuscript.

Acknowledgements

The authors thank all their colleagues, particularly at Department of Electronic Science, Xiamen University for their support and intellectual engagement.

Funding

The following sources of funding for each author are gratefully acknowledged: National Natural Science Foundation of China (82372087) to JD. Deutsche Forschungsgemeinschaft (KI-FOR 5363) to XX.

Conflict of interest

Not industry sponsored. All authors report no relevant disclosures.

Consent to participate All participants provided written informed consent

Consent for publication All authors provide consent for publication

Zhang X, Tang B, Guo J. Parkinson’s disease and gut microbiota: from clinical to mechanistic and therapeutic studies[J]. Translational Neurodegeneration, 2023, 12(1): 59.
Hey G, Nair N, Klann E, et al. Therapies for Parkinson’s disease and the gut microbiome: Evidence for bidirectional connection[J]. Frontiers in Aging Neuroscience, 2023, 15: 1151850.
Romano S, Savva G M, Bedarf J R, et al. Meta-analysis of the Parkinson’s disease gut microbiome suggests alterations linked to intestinal inflammation[J]. npj Parkinson's Disease, 2021, 7(1): 27.
Long-Smith C, O'Riordan K J, Clarke G, et al. Microbiota-gut-brain axis: new therapeutic opportunities[J]. Annual review of pharmacology and toxicology, 2020, 60(1): 477-502.
Lubomski M, Tan A H, Lim S Y, et al. Parkinson’s disease and the gastrointestinal microbiome[J]. Journal of neurology, 2020, 267: 2507-2523.
Kowalski K, Mulak A. Brain-gut-microbiota axis in Alzheimer’s disease[J]. Journal of neurogastroenterology and motility, 2019, 25(1): 48.
Xu M, Xu X, Li J, et al. Association between gut microbiota and autism spectrum disorder: a systematic review and meta-analysis[J]. Frontiers in psychiatry, 2019, 10: 473.
Ma Q, Xing C, Long W, et al. Impact of microbiota on central nervous system and neurological diseases: the gut-brain axis[J]. Journal of neuroinflammation, 2019, 16: 1-14.
Nishiwaki H, Ito M, Ishida T, et al. Meta‐analysis of gut dysbiosis in Parkinson's disease[J]. Movement Disorders, 2020, 35(9): 1626-1635.
Bonnechère B, Amin N, van Duijn C. What are the key gut microbiota involved in neurological diseases? A systematic review[J]. International Journal of Molecular Sciences, 2022, 23(22): 13665.
Stopińska K, Radziwoń-Zaleska M, Domitrz I. The microbiota-gut-brain axis as a key to neuropsychiatric disorders: a mini review[J]. Journal of Clinical Medicine, 2021, 10(20): 4640.
Callahan B J, McMurdie P J, Rosen M J, et al. DADA2: High-resolution sample inference from Illumina amplicon data[J]. Nature methods, 2016, 13(7): 581-583.
Quast C, Pruesse E, Yilmaz P, et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools[J]. Nucleic acids research, 2012, 41(D1): D590-D596.
Arumugam M, Raes J, Pelletier E, et al. Enterotypes of the human gut microbiome[J]. nature, 2011, 473(7346): 174-180.
Gotelli N J, Colwell R K. Quantifying biodiversity: procedures and pitfalls in the measurement and comparison of species richness[J]. Ecology letters, 2001, 4(4): 379-391.
Chao A. Nonparametric estimation of the number of classes in a population[J]. Scandinavian Journal of statistics, 1984: 265-270.
Shannon C E. A mathematical theory of communication[J]. The Bell system technical journal, 1948, 27(3): 379-423.
Simpson E H. Measurement of Diversity[J]. Nature, 1949, 163.
Pielou E C. The measurement of diversity in different types of biological collections[J]. Journal of theoretical biology, 1966, 13: 131-144.
Magurran A E. Measuring biological diversity[J]. Current Biology, 2021, 31(19): R1174-R1177.
Jost L. Partitioning diversity into independent alpha and beta components[J]. Ecology, 2007, 88(10): 2427-2439.
Vendrik K E W, Ooijevaar R E, De Jong P R C, et al. Fecal microbiota transplantation in neurological disorders[J]. Frontiers in cellular and infection microbiology, 2020, 10: 98.
Boertien J M, Pereira P A B, Aho V T E, et al. Increasing comparability and utility of gut microbiome studies in Parkinson’s disease: a systematic review[J]. Journal of Parkinson's disease, 2019, 9(s2): S297-S312.
Pietrucci D, Cerroni R, Unida V, et al. Dysbiosis of gut microbiota in a selected population of Parkinson's patients[J]. Parkinsonism & related disorders, 2019, 65: 124-130.
Sinha R, Abu-Ali G, Vogtmann E, et al. Assessment of variation in microbial community amplicon sequencing by the Microbiome Quality Control (MBQC) project consortium[J]. Nature biotechnology, 2017, 35(11): 1077-1086.
Zhernakova A, Kurilshikov A, Bonder M J, et al. Population-based metagenomics analysis reveals markers for gut microbiome composition and diversity[J]. Science, 2016, 352(6285): 565-569.
Scheperjans F, Aho V, Pereira P A B, et al. Gut microbiota are related to Parkinson's disease and clinical phenotype[J]. Movement Disorders, 2015, 30(3): 350-358.
Cryan J F, O'Riordan K J, Cowan C S M, et al. The microbiota-gut-brain axis[J]. Physiological reviews, 2019.
Sampson T R, Challis C, Jain N, et al. A gut bacterial amyloid promotes α-synuclein aggregation and motor impairment in mice[J]. elife, 2020, 9: e53111.
Fernandes A D, Reid J N, Macklaim J M, et al. Unifying the analysis of high-throughput sequencing datasets: characterizing RNA-seq, 16S rRNA gene sequencing and selective growth experiments by compositional data analysis[J]. Microbiome, 2014, 2: 1-13.
Nearing J T, Douglas G M, Hayes M G, et al. Microbiome differential abundance methods produce different results across 38 datasets[J]. Nature communications, 2022, 13(1): 342.
Vogt N M, Kerby R L, Dill-McFarland K A, et al. Gut microbiome alterations in Alzheimer’s disease[J]. Scientific reports, 2017, 7(1): 13537.
Fattorusso A, Di Genova L, Dell’Isola G B, et al. Autism spectrum disorders and the gut microbiota[J]. Nutrients, 2019, 11(3): 521.
Louis P, Flint H J. Formation of propionate and butyrate by the human colonic microbiota[J]. Environmental microbiology, 2017, 19(1): 29-41.
Stilling R M, Van De Wouw M, Clarke G, et al. The neuropharmacology of butyrate: the bread and butter of the microbiota-gut-brain axis?[J]. Neurochemistry international, 2016, 99: 110-132.
Wexler A G, Goodman A L. An insider's perspective: Bacteroides as a window into the microbiome[J]. Nature microbiology, 2017, 2(5): 1-11.
Jandhyala S M, Talukdar R, Subramanyam C, et al. Role of the normal gut microbiota[J]. World journal of gastroenterology: WJG, 2015, 21(29): 8787.
Hill‐Burns E M, Debelius J W, Morton J T, et al. Parkinson's disease and Parkinson's disease medications have distinct signatures of the gut microbiome[J]. Movement disorders, 2017, 32(5): 739-749.
Kobayashi Y, Sugahara H, Shimada K, et al. Therapeutic potential of Bifidobacterium breve strain A1 for preventing cognitive impairment in Alzheimer’s disease[J]. Scientific reports, 2017, 7(1): 13510.
Ruiz L, Delgado S, Ruas-Madiedo P, et al. Bifidobacteria and their molecular communication with the immune system[J]. Frontiers in microbiology, 2017, 8: 2345.
Tamanai-Shacoori Z, Smida I, Bousarghin L, et al. Roseburia spp.: a marker of health?[J]. Future microbiology, 2017, 12(2): 157-170.
Ottman N, Ruokolainen L, Suomalainen A, et al. Soil exposure modifies the gut microbiota and supports immune tolerance in a mouse model[J]. Journal of allergy and clinical immunology, 2019, 143(3): 1198-1206. e12.
Unger M M, Spiegel J, Dillmann K U, et al. Short chain fatty acids and gut microbiota differ between patients with Parkinson's disease and age-matched controls[J]. Parkinsonism & related disorders, 2016, 32: 66-72.
Forbes J D, Chen C, Knox N C, et al. A comparative study of the gut microbiota in immune-mediated inflammatory diseases—does a common dysbiosis exist?[J]. Microbiome, 2018, 6: 1-15.
Zhao Y, Lukiw W J. Microbiome-generated amyloid and potential impact on amyloidogenesis in Alzheimer’s disease (AD)[J]. Journal of nature and science, 2015, 1(7).
Saulnier D M, Riehle K, Mistretta T A, et al. Gastrointestinal microbiome signatures of pediatric patients with irritable bowel syndrome[J]. Gastroenterology, 2011, 141(5): 1782-1791.
Hall A B, Yassour M, Sauk J, et al. A novel Ruminococcus gnavus clade enriched in inflammatory bowel disease patients. Genome Med 9: 103[J]. 2017.
Zhuang Z Q, Shen L L, Li W W, et al. Gut microbiota is altered in patients with Alzheimer’s disease[J]. Journal of Alzheimer's disease, 2018, 63(4): 1337-1346.
Strati F, Cavalieri D, Albanese D, et al. New evidences on the altered gut microbiota in autism spectrum disorders. Microbiome 5: 24[J]. 2017.
Kurtz Z D, Müller C L, Miraldi E R, et al. Sparse and compositionally robust inference of microbial ecological networks[J]. PLoS computational biology, 2015, 11(5): e1004226.
Louis P, Flint H J. Formation of propionate and butyrate by the human colonic microbiota[J]. Environmental microbiology, 2017, 19(1): 29-41.
Zheng D, Liwinski T, Elinav E. Interaction between microbiota and immunity in health and disease[J]. Cell research, 2020, 30(6): 492-506.
Stilling R M, Van De Wouw M, Clarke G, et al. The neuropharmacology of butyrate: the bread and butter of the microbiota-gut-brain axis?[J]. Neurochemistry international, 2016, 99: 110-132.
Friedland R P, Chapman M R. The role of microbial amyloid in neurodegeneration[J]. PLoS pathogens, 2017, 13(12): e1006654.
David L A, Maurice C F, Carmody R N, et al. Diet rapidly and reproducibly alters the human gut microbiome[J]. Nature, 2014, 505(7484): 559-563.
Jernberg C, Lofmark S, Edlund C, et al. Long-term impacts of antibiotic exposure on the human intestinal microbiota[J]. Microbiology, 2010, 156(11): 3216-3223.
Karl J P, Hatch A M, Arcidiacono S M, et al. Effects of psychological, environmental and physical stressors on the gut microbiota[J]. Frontiers in microbiology, 2018, 9: 372026.
Franzosa E A, Sirota-Madi A, Avila-Pacheco J, et al. Gut microbiome structure and metabolic activity in inflammatory bowel disease[J]. Nature microbiology, 2019, 4(2): 293-305.
Averina O V, Kovtun A S, Polyakova S I, et al. The bacterial neurometabolic signature of the gut microbiota of young children with autism spectrum disorders[J]. Journal of medical microbiology, 2020, 69(4): 558-571.
Valles-Colomer M, Falony G, Darzi Y, et al. The neuroactive potential of the human gut microbiota in quality of life and depression[J]. Nature microbiology, 2019, 4(4): 623-632.
Mallick H, Ma S, Franzosa E A, et al. Experimental design and quantitative analysis of microbial community multiomics[J]. Genome biology, 2017, 18: 1-16.
Callahan B J, Sankaran K, Fukuyama J A, et al. Bioconductor workflow for microbiome data analysis: from raw reads to community analyses[J]. F1000Research, 2016, 5.
Agresti A. Analysis of Ordinal Categorical Data[M]. Wiley, 2010.
Liu I, Agresti A. The analysis of ordered categorical data: An overview and a survey of recent developments[J]. Test, 2005, 14: 1-73.
Balduzzi S, Rücker G, Schwarzer G. How to perform a meta-analysis with R: a practical tutorial[J]. BMJ Ment Health, 2019, 22(4): 153-160.
Riley R D, Higgins J P T, Deeks J J. Interpretation of random effects meta-analyses[J]. Bmj, 2011, 342.
Oksanen J. Vegan: community ecology package[J]. http://vegan. r-forge. r-project. org/, 2010.

No competing interests reported.

SupplementaryMatrials.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

A meta-study of the microbiome and related dysfunction of the brain reveals distinct microbial signatures in the microbiota-gut-brain axis

Status:

Version 1

Abstract

Figures

Introduction

Results

Meta-analysis dataset composition and quality control

Alpha and beta diversity profiles of the gut microbiome in the MGBA context

Consensus and specific microbial signatures for neurological diseases

Consensus and specific microbial network structure for neurological diseases

Discussion

Methods

Study selection

16S rRNA gene sequencing data processing

Diversity analysis

Differential abundance analysis

Network Analysis

Declarations

Availability and Implementation

Author contribution

Acknowledgements

Funding

Conflict of interest

References

Additional Declarations

Supplementary Files

Status:

Version 1