Exploring causal gut-brain axes in Alzheimer's disease using mediation Mendelian randomization analysis

Abstract

Background

Alzheimer's disease (AD) is a progressive neurodegenerative condition with unclear etiology. Recent studies suggest gut microbiota may be involved in AD pathogenesis through imbalances that increase intestinal permeability, affect blood-brain barrier function, and promote neuroinflammation. However, observational studies are susceptible to confounding biases and reverse causality.

Objective

This study aimed to explore causal relationships between gut microbiota, brain imaging-derived phenotypes (IDPs), and AD using mediation Mendelian randomization analysis to identify specific gut-brain axes involved in AD mechanisms.

Methods

We conducted a three-phase Mendelian randomization analysis using large-scale genome-wide association study (GWAS) data. Phase 1 analyzed causal effects of 412 gut microbiota on AD. Phase 2 examined causal effects of 920 IDPs on AD. Phase 3 performed mediation analysis to understand the role of IDPs in the gut microbiota-AD pathway. Data sources included Dutch population study (7738 individuals), MiBioGen consortium (18,340 individuals), UK Biobank brain imaging (8428 samples), and AD GWAS dataset (487,511 participants). Inverse variance weighted method was the primary analysis approach.

Results

We identified 12 gut microbiota metabolic pathways and 32 gut microbiota species causally related to AD, plus 29 IDPs with potential causal relationships to AD. Mediation analysis revealed four distinct gut-brain axes: genus Butyrivibrio-brain stem-AD, genus Lachnospiraceae-brain stem-AD, PEPTIDOGLYCANSYN pathway-L1 External capsule Left-AD, and TCA cycle pathway-L1 External capsule Left-AD.

Conclusions

This study identified four specific gut microbiota-brain structure axes causally involved in AD mechanisms, providing novel insights for understanding the gut-brain axis role in AD pathogenesis and potential therapeutic targets.

Keywords

Alzheimer's disease gut microbiota gut-brain axis imaging-derived phenotypic measures mediation Mendelian randomization

Introduction

Alzheimer's disease (AD) is a progressive neurodegenerative condition and the leading cause of dementia among older adults. It is marked by a gradual deterioration in cognitive functions and memory, significantly impacting daily activities and social interactions. The precise cause of AD remains unclear, though it is thought to be influenced by a combination of genetic predispositions, irregular protein processing, inflammatory responses, and environmental influences.^1–3 In recent years, an increasing number of studies have suggested that the gut microbiota may be involved in the pathogenesis of AD. Multiple studies, both domestic and international, have compared the composition of the gut microbiota in AD patients and healthy control groups using 16S rRNA gene sequencing technology and found differences between the two groups.⁴ These differences include a decrease in the Firmicutes microbiota and an increase in the Bacteroidetes microbiota in AD patients. Additionally, Actinobacteria microbiota is increased, while Bacteroidetes microbiota is decreased in the gut of AD patients.^5–7 Beneficial microbiota such as lactobacilli are reduced, while potential pathogenic bacteria such as Escherichia coli are increased in the gut of AD patients. These findings suggest that AD may be associated with an imbalance in the gut microbiota.⁸ This imbalance can increase intestinal mucosal permeability and activate intestinal immune cells, thereby affecting the blood-brain barrier function, promoting neuroinflammation, neuronal loss, and nerve damage, ultimately leading to AD.^7,9 Furthermore, gut microbiota can also synthesize lipopolysaccharides and bacterial amyloid, harmful substances that, when entering the brain, can activate immune cells within the brain, causing neuroinflammation.¹⁰ Overall, the gut microbiota has a significant impact on brain function and cognitive abilities, and this interaction may be related to AD.^4,8 The structural changes in the brain, captured by various imaging-derived phenotypes (IDPs), reflect distinct aspects of the neurodegenerative process in AD. Notably, changes in brain volume, particularly in the brain stem, are hallmark features of AD progression, correlating positively with AD risk. Such changes may stem from pathology related to tau proteins originating in brainstem nuclei, contributing to cognitive decline during early stages of AD. Additionally, alterations in white matter integrity, evidenced by significant associations involving structures like the external capsule, highlight the damage to neural fiber connectivity observed in patients with mild cognitive impairment (MCI) and prodromal AD. Furthermore, disruptions in brain connectivity, particularly in key pathways like the Superior longitudinal fasciculus and Cingulum cingulate gyrus, can impede functional interactions between brain regions, affecting cognitive functions critical to daily life. The study also noted associations with specific brain regions, such as the cerebellum and corpus callosum, which are well-documented in AD research. The current research on the relationship between gut microbiota and AD is based on observations of the composition and changes in the gut microbiota in the feces of AD patients. However, due to limitations such as strain selection and reduced cytokine levels, these experiments are difficult to conduct in humans, and small observational studies are susceptible to confounding biases and reverse causality.

An increasing number of research studies are utilizing Mendelian randomization (MR) methods to investigate the connections between gut microbiota and various diseases.^11–14 MR is a technique in genetic epidemiology that employs genetic variants as instrumental variables (IVs) to evaluate the causal relationships between exposures and outcomes.¹¹ As genetic variants are fixed at conception, MR is less prone to environmental confounders and reverse causation compared to traditional observational studies. The two-sample MR approach, which employs summary data from large-scale, separate genome-wide association studies (GWAS), offers enhanced statistical power to estimate causal effects between “exposures” and “outcomes”.^11,12 An increasing number of research studies are utilizing MR methods to investigate the connections between gut microbiota and various diseases..^15–17

In this study, we conducted a comprehensive MR analysis to explore the causal relationships between gut microbiota, brain imaging, and AD. Specifically, we used the method of mediation MR to analyze specific microbiota and brain structural changes in AD patients. Mediation MR, based on the fundamental principles of MR, uses the relationship between genotype and a given biological marker or disease to infer the relationship between the biological marker or disease and another causal variable.¹⁸ This study provides a basis for further exploration of the mechanisms of the microbiota-gut-brain axis in AD.

Methods

Experimental design

The study was carried out in three phases. Step 1 involved analyzing the causal effects of 412 gut microbiota on AD (Figure 1(a)); Step 2 focused on determining the causal effects of 920 IDPs on AD (Figure 1(b)); Step 3 conducted a mediation analysis to understand the role of IDPs in the pathway from gut microbiota to AD (Figure 1(c)).

Figure 1.

Research overview. (a) Causal effect of gut microbiota on AD. (b) Causal effect of imaging-derived phenotypes on AD. The third step represents the intermediate analysis of the pathway from gut microbiota to AD through imaging-derived phenotypes: Path c represents the total effect of gut microbiota on AD; Path b represents the causal effect of imaging-derived phenotypes on AD; Path a represents the causal effect of gut microbiota on imaging-derived phenotypes.

Source of gut microbiota GWAS data

The gut microbiota GWAS data utilized in this study were sourced from two significant public databases, offering a robust foundation for microbiome research. The first dataset, derived from a large-scale Dutch population study of 7738 individuals, integrated genomic and 16S rRNA sequencing data to systematically assess host genetic influences on gut microbiota composition and structure. This comprehensive analysis covered 207 taxonomies (including 5 phyla, 10 classes, 13 orders, 26 families, 48 genera, and 105 species) and 328 bacterial pathways, providing both taxonomic and functional insights. The second dataset, from the MiBioGen consortium, encompassed an even larger sample of 18,340 individuals, combining whole-genome genotyping with 24S fecal microbiota data across 340 sequences. This consortium approach likely captured diverse populations, enhancing the generalizability of findings. The complementary nature of these datasets, employing different sequencing methods (16S rRNA versus 24S) and covering a combined sample size exceeding 26,000 individuals, offers unprecedented statistical power for detecting subtle associations between host genetics and gut microbiome composition. The inclusion of both taxonomic classifications and functional pathways enables a more comprehensive understanding of the gut microbiome's role in health and disease. Furthermore, the potential for meta-analyses across these large datasets could reveal consistent patterns across different populations and study designs. For researchers seeking more detailed information on the gut microbiota data used, Supplemental Table 1 in the Supplemental Material provides a comprehensive breakdown of the microbial taxa and functional pathways analyzed, likely including data on prevalence, abundance, and associations with host genetic factors.^19,20 Detailed information on gut microbiota is provided in the Supplemental Material (Supplemental Table 1).

Source of imaging-derived phenotype data

The IDP data used in this study were sourced from a whole-genome association study of brain imaging phenotypes, utilizing data from the UK Biobank. The UK Biobank is a significant epidemiological research resource comprising 500,000 volunteers. It is designed to gather extensive data to identify disease risk factors and early markers of disease onset. This study leveraged data from the UK Biobank's brain imaging extension project, which conducted brain imaging on 100,000 participants. The imaging modalities included three structural techniques, resting-state and task-based functional magnetic resonance imaging, and diffusion tensor imaging.

A considerable number of single nucleotide polymorphisms (SNPs) associated with brain imaging phenotypes were identified. Multi-phenotype association tests were performed on a wide array of IDPs, revealing genetic correlations and functional enrichment. The brain was segmented into eight anatomical regions (such as the frontal lobe and parietal lobe) and five structural connectivity regions (including commissural fibers and brainstem tracts). For brain anatomy, three types of measurements were taken: area, thickness, and volume. For structural connectivity, six measurements were obtained, which included fractional anisotropy (FA), mean diffusivity, mode of the diffusion tensor, intracellular volume fraction (ICVF), isotropic or free water volume fraction, and orientation dispersion index. This study incorporated data from 8428 samples, covering 11,734,354 SNPs and 885 different types of IDPs.²¹ Detailed information on IDPs is provided in the Supplemental Material (Supplemental Table 2).

Source of AD GWAS data

The AD GWAS dataset comprised genetic data from 487,511 participants, including 90,338 AD or AD-by-proxy cases and 397,173 controls of European ancestry. Data were derived from multiple cohorts, including the UK Biobank, FinnGen, and numerous AD-specific consortia. This two-stage meta-analysis first analyzed 39 cohorts with clinically diagnosed AD patients and controls, followed by analysis of AD-by-proxy cases (based on parental AD history) from large biobanks. Genotyping was performed using various arrays across cohorts, with subsequent imputation to the Haplotype Reference Consortium panel and TOPMed reference panel. Quality control procedures included filtering variants with imputation quality <0.3 and minor allele frequency <0.01. The statistical analysis employed logistic regression models adjusted for age, sex, and principal components to account for population stratification.²²

Instrumental variable selection

In our MR analysis involving gut microbiota and AD, as well as IDPs, we selected IVs based on p-values. For the gut microbiota and AD analysis, we adhered to criteria used in numerous MR studies on gut microbiota and various diseases,^12,21,22 utilizing a p-value threshold of less than 1e-05 for both exposure and outcome IVs. In the MR analysis focusing on IDPs and AD, a more stringent p-value criterion of less than 5e-08 was applied.

We employed the two sample MR R package and parameters of r² = 0.001 and kb = 10,000 to ensure the independence of the selected IVs and reduce bias from the random distribution of linked alleles. To avoid bias from weak instruments, we calculated the F-statistic to determine the statistical strength of each SNP's association with the exposure. IVs with an F-statistic greater than 10 were deemed to have a strong association, while those with an F-statistic less than 10 were considered weak. SNPs with palindromic structures were automatically excluded to ensure accurate results.

MR analysis

This study applied MR analysis to investigate the causal relationships between gut microbiota, IDPs, and AD using five MR methods. The inverse variance weighted (IVW) method was the primary technique, particularly effective when utilizing multiple genetic variants as IVs. A p-value of less than 0.05 in the MR result indicated a significant association between the exposure and the outcome.

Following the initial MR analysis, significant findings regarding gut microbiota, IDPs, and AD led to further mediation analysis, treating the statistically significant gut microbiota as the new exposure and IDPs as the new outcome.

For sensitivity analysis, Cochran's Q method was used to evaluate heterogeneity. In instances of significant heterogeneity (p < 0.05), MR-Egger regression analysis was conducted to assess potential pleiotropy of the SNPs. An intercept term with a p-value less than 0.05 in MR-Egger regression signified directional pleiotropy. Bonferroni correction was applied to account for multiple comparisons, using the formula: The Bonferroni-adjusted p-value threshold was calculated as 0.05/n (n is the number of tests performed), Associations with p-values between the original threshold and the Bonferroni-adjusted threshold were still considered significant. Additionally, reverse MR was employed to determine if established important immune cells were causally related to specific neurodegenerative diseases, minimizing bias due to exposure.

All statistical analyses were performed using R software (version 4.2.1).

Results

Selection of instrumental variables

After preliminary screening, a total of 12 gut microbiota metabolic pathways, 32 different species of gut microbiota, and 29 different types of IDPs were identified as having potential causal relationships with AD. Additionally, 2 different gut metabolites, 2 different species of gut microbiota, and 2 types of IDPs were found to have potential causal relationships with AD. The F-statistic for all IVs was largely greater than 10, indicating no evidence of weak instrument bias.

Causal effects of gut microbiota on AD

This study identified a total of 12 gut microbiota metabolic pathways, with 32 different species of gut microbiota potentially causally related to AD. Among these pathways, 4 showed a positive correlation with AD, while 8 exhibited a negative correlation. Furthermore, 14 different species of gut microbiota were positively associated with AD, and 18 were negatively associated. As shown in Figure 2(a), the IVW results include the following key findings: TCA cycle gut microbiota pathway (p = 8.20E-03; OR 95% CI = 0.93 (0.88, 0.98)), PEPTIDOGLYCANSYN gut microbiota pathway (p = 1.52E-02; OR 95% CI = 0.93 (0.87, 0.99)), Genus Butyrivibrio (p = 1.28E-02; OR 95% CI = 0.95 (0.92, 0.99)), Genus Lachnospiraceae (p = 3.90E-02; OR 95% CI = 0.90 (0.82, 0.99)), Order Selenomonadales (p = 8.34E-03; OR 95% CI = 1.13 (1.03, 1.24)), Phylum Tenericutes (p = 1.23E-03; OR 95% CI = 0.87 (0.79, 0.95)), Class Actinobacteria (p = 6.46E-03; OR 95% CI = 1.14 (1.04, 1.25)), Class Clostridia (p = 9.72E-03; OR 95% CI = 0.88 (0.79, 0.97)), Class Mollicutes (p = 1.23E-03; OR 95% CI = 0.87 (0.79, 0.95)), Class Negativicutes (p = 8.34E-03; OR 95% CI = 1.13 (1.03, 1.24)), Order Coriobacteriales (p = 1.47E-02; OR 95% CI = 1.12 (1.02, 1.24)), Family Coriobacteriaceae (p = 1.47E-02; OR 95% CI = 1.12 (1.02, 1.24)), Class Bacilli (p = 4.36E-02; OR 95% CI = 0.96 (0.92, 1.00)).

Figure 2.

(a) Relationship heat map between gut microbiota and AD, where red represents OR > 1, indicating a positive correlation with AD, and blue represents OR < 1, indicating a negative correlation with AD. (b) Forest plot of the MR results of the relationship between gut microbiota and AD.

For MR results of all other gut microbiota metabolic pathways and different species of gut microbiota associated with AD, please refer to the Supplemental Material (Supplemental Table 3). The MR results are illustrated in the forest plot in Figure 2(b).

Causal effects of IDPs on AD

In this study, a total of 29 different types of IDPs were found to have potential causal relationships with AD, with 12 showing a negative correlation and the remainder showing a positive correlation (Figure 3(a)). The MR results for these IDPs are as follows: Brain stem (p = 4.22E-06; OR 95% CI = 1.43 (1.23, 1.66)), L1 External capsule Left (p = 7.17E-06; OR 95% CI = 0.78 (0.70, 0.87)), V cerebellum VIIIb (p = 7.09E-04; OR 95% CI = 1.27 (1.11, 1.46)), Splenium of corpus callosum (p = 2.11E-03; OR 95% CI = 0.82 (0.72, 0.93)), FA striatum right (p = 2.65E-03; OR 95% CI = 1.22 (1.07, 1.39)), L1 anterior thalamic radiation left (p = 3.20E-03; OR 95% CI = 0.82 (0.72, 0.93)), L1 Superior longitudinal fasciculus Right (p = 3.37E-03; OR 95% CI = 0.73 (0.59, 0.90)), FA striatum left (p = 5.32E-03; OR 95% CI = 1.16 (1.04, 1.28)), L1 uncinate fasciculus right (p = 8.30E-03; OR 95% CI = 0.86 (0.76, 0.96)), Cingulum cingulate gyrus Left (p = 8.63E-03; OR 95% CI = 0.89 (0.81, 0.97)), External capsule Left (p = 1.47E-02; OR 95% CI = 0.82 (0.71, 0.96)), External capsule Right (p = 1.72E-02; OR 95% CI = 1.16 (1.03, 1.32)), Cerebral peduncle Right (p = 1.79E-02; OR 95% CI = 1.20 (1.03, 1.40)), Genu of corpus callosum (p = 1.96E-02; OR 95% CI = 1.08 (1.01, 1.15)), Splenium of corpus callosum (p = 2.01E-02; OR 95% CI = 1.10 (1.01, 1.19)), External capsule Left (p = 2.66E-02; OR 95% CI = 1.11 (1.01, 1.22)), ICVF forceps major (p = 2.66E-02; OR 95% CI = 1.08 (1.01, 1.16)), L1 External capsule Right (p = 2.94E-02; OR 95% CI = 0.78 (0.63, 0.98)), FA acoustic radiation right (p = 3.06E-02; OR 95% CI = 1.16 (1.01, 1.32)), Splenium of corpus callosum (p = 3.26E-02; OR 95% CI = 0.92 (0.85, 0.99)), L3 Splenium of corpus callosum (p = 3.46E-02; OR 95% CI = 0.89 (0.80, 0.99)), L3 Uncinate fasciculus Left (p = 3.84E-02; OR 95% CI = 0.89 (0.80, 0.99)), Sagittal stratum Right (p = 3.86E-02; OR 95% CI = 1.08 (1.00, 1.15)), Fornix cres + Stria terminalis Right (p = 3.91E-02; OR 95% CI = 1.09 (1.00, 1.19)), Cingulum cingulate gyrus Right (p = 4.37E-02; OR 95% CI = 0.88 (0.77, 1.00)), Left caudate (p = 4.41E-02; OR 95% CI = 1.15 (1.00, 1.32)), Pontine crossing tract (p = 4.54E-02; OR 95% CI = 0.95 (0.90, 1.00)), External capsule Right (p = 4.56E-02; OR 95% CI = 1.11 (1.00, 1.23)), Posterior limb of internal capsule Left (p = 4.96E-02; OR 95% CI = 1.13 (1.00, 1.28)).

Figure 3.

(a) Relationship heat map between IDPs and AD, where red represents OR > 1, indicating a positive correlation with AD, and blue represents OR < 1, indicating a negative correlation with AD, while black represents missing values. (b) Forest plot of the MR results of the relationship between IDPs and AD.

Among these, the brain stem and L1 External capsule Left remained significant after FDR adjustment, indicating potential robust associations. The MR forest plot is shown in Figure 3(b), and the results of other IDPs and AD MR are presented in the Supplemental Material (Supplemental Table 4).

Mediation MR analysis

Following the first and second steps of the analysis in Figure 1, we conducted a mediation MR analysis and identified potential causal relationships between two different gut metabolites, two different genera of gut microbiota, and two different IDPs. The MR results are as follows: genus Butyrivibrio on brain stem (p = 3.00E-02; OR 95%CI = 1.08 (1.01,1.15)), genus Lachnospiraceae on brain stem (p = 6.00E-02; OR 95%CI = 0.83 (0.73,0.95)), PEPTIDOGLYCANSYN.gut microbiota pathway on L1 External capsule Left (p = 1.20E-02; OR 95%CI = 0.87 (0.78,0.97)), TCA.cycle.gut microbiota pathway on L1 External capsule Left (p = 3.00E-04; OR 95%CI = 0.84 (0.77,0.92)). The MR Forest plot is shown in Figure 4(a). Our results indicate the existence of four potential gut-microbiota-brain axis relationships: genus Butyrivibrio -brain stem-AD, genus Lachnospiraceae -brain stem-AD, PEPTIDOGLYCANSYN.gut microbiota pathway-L1 External capsule Left-AD, and TCA.cycle.gut microbiota pathway-L1 External capsule Left-AD. The relationships are depicted in Figure 4(b). The MR results for other gut microbiota and IDPs can be found in the Supplemental Material (Supplemental Table 5).

Figure 4.

(a) Forest plot of the MR results for the gut microbiota and IDPs. (b) Diagram of the gut-brain axis with four pathways.

Sensitivity analysis

In this study, we conducted tests for pleiotropy and heterogeneity. According to the MR-Egger regression intercept method, there was no bias due to genetic pleiotropy (p > 0.05, Supplemental Table 6). Cochran's Q test showed no significant heterogeneity (p > 0.05, Supplemental Table 7).

Reverse MR

As shown in the Supplemental Material (Supplemental Table 8), this study conducted a reverse MR analysis of the relationship between gut microbiota, IDPs, and AD. Our results indicate that AD is associated with four specific gut microbiota or IDPs. Specifically, AD is positively correlated with FA str Left and negatively correlated with s_Veillonella_unclassified, genus Butyrivibrio, and Cingulum cingulate gyrus Right. The MR results are as follows: s_Veillonella_unclassified (p = 9.02E-03; OR 95%CI = 0.94 (0.73,0.96)), genus Butyrivibrio (p = 3.31E-02; OR 95%CI = 0.94 (0.80,0.99)), FA str Left (p = 3.54E-02; OR 95%CI = 1.14 (1.01,1.12)), Cingulum cingulate gyrus Right (p = 4.27E-02; OR 95%CI = 0.94 (0.90,0.99)).

Discussion

Main findings

In this study, we utilized large-scale GWAS data and employed the MR method to explore potential relationships between the gut microbiota-brain axis and AD. Our research revealed the existence of four distinct axes with causal relationships to AD.

Potential mechanisms

The gut microbiota and the gut-brain axis play a crucial role in the pathogenesis of AD. Imbalances in the gut microbiota may lead to increased gut permeability, affecting immune function and blood-brain barrier integrity, thereby causing neuroinflammation.²³ Microbes can also influence neurotransmitter synthesis, synaptic signaling, and neurotrophic factors, which are associated with cognitive function. Additionally, the gut microbiota is involved in various metabolic pathways, and the production of harmful substances can strongly activate the immune system and increase systemic inflammation, which may subsequently lead to neuronal damage.^24,25

In our study, we found a negative correlation between Lachnospiraceae and AD. Similar evidence has been reported in another study, where Lachnospiraceae was identified as one of the major genera detected in both AD patients and normal control groups. The study found a significant decrease in the relative abundance of Lachnospiraceae in the AD patient group compared to the normal control group.^26,27 The reduced abundance of Lachnospiraceae in AD patients may weaken intestinal barrier function, increase the risk of intestinal permeability, and consequently lead to low-grade systemic inflammatory responses. Additionally, the metabolites abundant in Lachnospiraceae may possess anti-inflammatory and antioxidant properties.²⁸ The reduction of these metabolites may weaken these protective effects, increase oxidative stress and inflammatory responses in the brain, and the metabolites of Lachnospiraceae may participate in regulating the immune response. Their changes may affect the immune status in the brain, such as alterations in microglial cells, thereby influencing the progression of AD.²⁹ Overall, the decrease in Lachnospiraceae in AD patients likely impacts the gut-brain axis and the brain's state through multiple pathways, contributing to the pathogenesis of AD. Our results also revealed that Lachnospiraceae affects the brainstem, forming the Lachnospiraceae-brain stem-AD axis. The brainstem likely plays a crucial role in the pathology of AD. Neuropathological studies indicate that AD-related tau protein pathology may originate from brainstem nuclei. A study using brain structural imaging and cognitive test data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database investigated the relationship between brainstem structure volume and cognitive function in normal elderly individuals and those with MCI. The results indicated that in the early stages of AD, the degeneration of brainstem structures such as the midbrain and pineal gland may be associated with declines in attention, processing speed, and executive function. The pineal gland, as a brainstem nucleus that regulates autonomic nervous function and cerebral blood flow, may likely affect these higher-level cognitive functions related to executive control.^30,31 This suggests that the brainstem may play a role in the pathogenesis of AD. Our study also identified another gut-brain axis, the genus Butyrivibrio -brain stem-AD axis, although unfortunately, there is limited research on Butyrivibrio in the context of AD.

In this study, we also identified two gut microbiota metabolic pathways that impact the imaging of AD brain structure. The tricarboxylic acid (TCA) cycle is one of the main energy production pathways in brain cells, as it completely oxidizes acetyl-CoA to provide electron carriers for oxidative phosphorylation, thereby generating ATP.³² Numerous studies have shown a significant decrease in TCA cycle activity in the brains of AD patients. PET scans have revealed a 10%-12% decrease in glucose uptake in regions such as the medial temporal lobe and temporal lobe in AD patients.^33,34 Reduced activity of the brain TCA cycle leads to decreased ATP production in the brain tissue of AD patients, which cannot meet the energy demands of neurons for normal function.³⁵ Furthermore, impaired TCA cycle function can affect the clearance mechanisms of Aβ and tau proteins, promoting neuroinflammation,³⁶ and hinder neurorepair and neural signal transmission. Our results also indicate that the TCA cycle leads to a decrease in the functionality of the brain's external capsule. The external capsule is a part of the brain located between the cerebral cortex and the thalamus, primarily composed of nerve fibers that play a crucial role in neural transmission.³⁷ Some studies have found abnormal diffusion tensor imaging parameters in the external capsule of patients with MCI and prodromal AD, indicating damage to the neural fiber structure and connectivity, which may be an early sign of the progression of AD.³⁸ White matter damage in the external capsule may be one of the neurophysiological mechanisms underlying the decline in cognitive function in AD patients. The external capsule connects various regions of the brain, and its white matter damage may disrupt functional connections between these areas, affecting attention, executive function, language ability, and other higher-level cognitive functions.^39,40

This study used the relative abundance of gut microbiota and gut microbiota metabolic pathways to determine whether the gut microbiota's impact on AD brain structure is “beneficial” or “harmful.” However, the exact mechanisms by which the gut microbiota influences AD have not been fully elucidated and require further research.

Strengths and limitations

Our findings align with and expand upon previous studies investigating the gut-brain axis in AD,⁴¹ particularly as this study is the first to employ the MR method to investigate this relationship, identifying specific gut microbiota species that may influence brain structure in AD. Notably, we utilized two publicly available GWAS databases encompassing over 600 gut microbiota species, marking a significant advancement in the field. However, it is crucial to recognize that, like other studies in this domain, our findings may be impacted by confounding factors not fully addressed in the MR analysis, particularly socioeconomic status, which has been identified as a significant risk factor for AD.⁴² Future studies should aim to incorporate such confounding factors to achieve a more comprehensive understanding of the complex interplay between gut microbiota, brain structure, and AD risk. Furthermore, our study faces limitations, as the selected samples predominantly originate from European cohorts, potentially reducing the generalizability of the results to other populations. Additionally, the study lacks comprehensive information regarding the original data, such as inclusion/exclusion criteria and interventions, which may adversely affect the accuracy of the findings.

Conclusion

Using the MR method, this study has identified four gut microbiota genera or metabolic pathway-brain structure axes that may be involved in the mechanism of AD, providing clues for further exploration of the role of the gut-brain axis in AD.

Supplemental Material

sj-xlsx-1-alz-10.1177_13872877251360004 - Supplemental material for Exploring causal gut-brain axes in Alzheimer's disease using mediation Mendelian randomization analysis

Supplemental material, sj-xlsx-1-alz-10.1177_13872877251360004 for Exploring causal gut-brain axes in Alzheimer's disease using mediation Mendelian randomization analysis by Lili Ge, Lin Zhu, Chen Su and Zhi Jin in Journal of Alzheimer's Disease

Footnotes

Acknowledgments

The authors have no acknowledgments to report.

ORCID iD

Zhi Jin

Author contributions

Lili Ge: Data curation.

Lin Zhu: Writing - original draft.

Chen Su: Writing - original draft.

Zhi Jin: Conceptualization; Writing - review & editing.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

The data used in the present study are all publicly available at .

Supplemental material

Supplemental material for this article is available online.

References

Streit

Khoshbouei

Bechmann

. The role of microglia in sporadic Alzheimer's disease. J Alzheimers Dis 2021; 79: 961–968.

DeTure

Dickson

. The neuropathological diagnosis of Alzheimer's disease. Mol Neurodegener 2019; 14: 32.

Khan

Barve

Kumar

. Recent advancements in pathogenesis, diagnostics and treatment of Alzheimer's disease. Curr Neuropharmacol 2020; 18: 1106–1125.

Ferreiro

Choi

Ryou

, et al. Gut microbiome composition may be an indicator of preclinical Alzheimer's disease. Sci Transl Med 2023; 15: eabo2984.

Ueda

Shinkai

Shiroma

, et al. Identification of Faecalibacterium prausnitzii strains for gut microbiome-based intervention in Alzheimer's-type dementia. Cell Rep Med 2021; 2: 100398.

Rastogi

Singh

. Gut microbiome and human health: exploring how the probiotic genus Lactobacillus modulate immune responses. Front Pharmacol 2022; 13: 1042189.

Dinan

. Gut microbes and neuropathology: is there a causal nexus? Pathogens 2022; 11: 796.

Jiang

Huang

, et al. The gut microbiota and Alzheimer's disease. J Alzheimers Dis 2017; 58: 1–15.

Zhang

Gao

Kwok

, et al. Gut microbiome-targeted therapies for Alzheimer's disease. Gut Microbes 2023; 15: 2271613.

10.

Nafady

Sayed

Abdelkawy

, et al. The effect of gut microbe dysbiosis on the pathogenesis of Alzheimer's disease (AD) and related conditions. Curr Alzheimer Res 2022; 19: 274–284.

11.

Zhou

Baranova

Cao

, et al. Gut microbiome and schizophrenia: insights from two-sample Mendelian randomization. Schizophrenia (Heidelb) 2024; 10: 75.

12.

Zhao

Baranova

Cao

, et al. Gut microbiome and major depressive disorder: insights from two-sample Mendelian randomization. BMC Psychiatry 2024; 24: 493.

13.

Tang

Lei

Ding

, et al. Causal relationship between immune cells and neurodegenerative diseases: a two-sample Mendelian randomisation study. Front Immunol 2024; 15: 1339649.

14.

Tang

Yang

Zhu

, et al. Iron metabolism disorder and multiple sclerosis: a comprehensive analysis. Front Immunol 2024; 15: 1376838.

15.

Fan

Zhou

Meng

, et al. Cross-talks between gut microbiota and tobacco smoking: a two-sample Mendelian randomization study. BMC Med 2023; 21: 163.

16.

Cui

, et al. Gut microbiome and frailty: insight from genetic correlation and Mendelian randomization. Gut Microbes 2023; 15: 2282795.

17.

Xiang

Zhang

Wang

, et al. Identification of host gene-microbiome associations in colorectal cancer patients using Mendelian randomization. J Transl Med 2023; 21: 535.

18.

Tang

Lei

, et al. Hypertension linked to Alzheimer's disease via stroke: Mendelian randomization. Sci Rep 2023; 13: 21606.

19.

Lopera-Maya

Kurilshikov

van der Graaf

, et al. Effect of host genetics on the gut microbiome in 7,738 participants of the Dutch microbiome project. Nat Genet 2022; 54: 143–151.

20.

Kurilshikov

Medina-Gomez

Bacigalupe

, et al. Large-scale association analyses identify host factors influencing human gut microbiome composition. Nat Genet 2021; 53: 156–165.

21.

Elliott

Sharp

Alfaro-Almagro

, et al. Genome-wide association studies of brain imaging phenotypes in UK Biobank. Nature 2018; 562: 210–216.

22.

Bellenguez

Küçükali

Jansen

, et al. New insights into the genetic etiology of Alzheimer's disease and related dementias. Nat Genet 2022; 54: 412–436.

23.

Kowalski

Mulak

. Brain-gut-microbiota axis in Alzheimer's disease. J Neurogastroenterol Motil 2019; 25: 48–60.

24.

Megur

Baltriukienė

Bukelskienė

, et al.

The microbiota-gut-brain axis and Alzheimer's disease: neuroinflammation is to blame?

Nutrients 2020; 13: 37.

25.

Doifode

Giridharan

Generoso

, et al. The impact of the microbiota-gut-brain axis on Alzheimer's disease pathophysiology. Pharmacol Res 2021; 164: 105314.

26.

Zhuang

Shen

, et al. Gut microbiota is altered in patients with Alzheimer's disease. J Alzheimers Dis 2018; 63: 1337–1346.

27.

Marizzoni

Mirabelli

Mombelli

, et al. A peripheral signature of Alzheimer's disease featuring microbiota-gut-brain axis markers. Alzheimers Res Ther 2023; 15: 101.

28.

Hung

Chang

Huang

, et al. Gut microbiota in patients with Alzheimer's disease spectrum: a systematic review and meta-analysis. Aging (Albany NY) 2022; 14: 477–496.

29.

Murray

Kemp

Nguyen

. The microbiota-gut-brain axis in Alzheimer's disease: a review of taxonomic alterations and potential avenues for interventions. Arch Clin Neuropsychol 2022; 37: 595–607.

30.

Dutt

Mather

, et al. Brainstem substructures and cognition in prodromal Alzheimer's disease. Brain Imaging Behav 2021; 15: 2572–2582.

31.

Dutt

Mather

, et al. Brainstem volumetric integrity in preclinical and prodromal Alzheimer's disease. J Alzheimers Dis 2020; 77: 1579–1594.

32.

Jaramillo-Jimenez

Giil

Borda

, et al. Serum TCA cycle metabolites in Lewy bodies dementia and Alzheimer's disease: network analysis and cognitive prognosis. Mitochondrion 2023; 71: 17–25.

33.

Mosconi

Tsui

Herholz

, et al. Multicenter standardized 18F-FDG PET diagnosis of mild cognitive impairment, Alzheimer's disease, and other dementias. J Nucl Med 2008; 49: 390–398.

34.

Caminiti

Ballarini

Sala

, et al. FDG-PET and CSF biomarker accuracy in prediction of conversion to different dementias in a large multicentre MCI cohort. Neuroimage Clin 2018; 18: 167–177.

35.

Chen

Zhong

. Decoding Alzheimer's disease from perturbed cerebral glucose metabolism: implications for diagnostic and therapeutic strategies. Prog Neurobiol 2013; 108: 21–43.

36.

Lee

Devanney

Golden

, et al. APOE Modulates microglial immunometabolism in response to age, amyloid pathology, and inflammatory challenge. Cell Rep 2023; 42: 112196.

37.

Mayo

Mazerolle

Ritchie

, et al. Longitudinal changes in microstructural white matter metrics in Alzheimer's disease. Neuroimage Clin 2017; 13: 330–338.

38.

Bennett

Greenia

Maillard

, et al. Age-related white matter integrity differences in oldest-old without dementia. Neurobiol Aging 2017; 56: 108–114.

39.

Sun

Salat

Upchurch

, et al. Destruction of white matter integrity in patients with mild cognitive impairment and Alzheimer disease. J Investig Med 2014; 62: 927–933.

40.

Feng

Lin

, et al. Cognition-related white matter integrity dysfunction in Alzheimer's disease with diffusion tensor image. Brain Res Bull 2018; 143: 207–216.

41.

Zhao

Baranova

Cao

, et al. Evaluating causal effects of gut microbiome on Alzheimer’s disease. J Prev Alzheimers Dis 2024; 11: 1843–1848.

42.

Wang

, et al. Socioeconomic status and risks of cognitive impairment and dementia: a systematic review and meta-analysis of 39 prospective studies. J Prev Alzheimers Dis 2023; 10: 83–94.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.80 MB