Skip to main content
NCBI home page
Medicine logoLink to Medicine
. 2026 May 8;105(19):e48681. doi: 10.1097/MD.0000000000048681

Bidirectional mediation between microbiota and immune system in Parkinson disease: A two-sample Mendelian randomization study

Taimei Qu a, Qiang Zhao b,c, Jiasheng Fang d, Zhiru Liang e,*
PMCID: PMC13166552  PMID: 42116350

Abstract

Recent studies suggest that gut and skin microbiota, immune cells, and inflammatory proteins may contribute to the development of Parkinson disease (PD). However, their causal relationships and underlying mediating mechanisms remain unclear. We conducted a two-sample Mendelian randomization analysis using publicly available genome-wide association study summary data to evaluate the causal relationships among 473 gut and 150 skin microbiota, 731 immune cell phenotypes, 91 inflammatory proteins, and PD. Sensitivity analyses and mediation analyses were conducted to evaluate robustness and explore potential biological pathways. We identified 16 gut microbial taxa, 2 skin microbiota, 19 immune cell phenotypes, and 4 inflammatory proteins with significant causal associations with PD. Two-step mediation Mendelian randomization further revealed multiple potential pathways of “microbiota–immune–PD” and “immune–microbiota–PD.” For instance, HLA DR on myeloid dendritic cells mediated 14.80% of the effect of Gammaproteobacteria_Sebaceous on PD risk, while Species Bifidobacterium adolescentis mediated 10.50% the protective effect of CD28− CD8dim %T cell. Notably, IL-18 exhibited the highest mediation proportion (26.4%) via the family Lentimicrobiaceae. Our findings provide novel genetic evidence supporting the role of the microbiota–immune axis in PD pathogenesis and highlight potential targets for therapeutic intervention.

Keywords: gut microbiota, immune cells, inflammatory proteins, mediation analysis, Mendelian randomization, Parkinson disease, skin microbiota

1. Introduction

Parkinson disease (PD) is the second most common neurodegenerative disorder worldwide, affecting >10 million individuals and imposing a substantial socioeconomic burden.[1] It is characterized by progressive loss of dopaminergic neurons and the accumulation of Lewy bodies.[2,3] The etiology of PD is complex and involves interactions between genetic susceptibility and environmental factors.[4,5] Emerging evidence suggests that PD is not solely a brain-restricted disorder but rather a systemic condition involving interactions among the nervous system, the immune system, and peripheral microbial communities.[6–8]

The gut microbiota, a key component of the gut–brain axis, regulates host physiology through immune modulation, neurotransmitter production, and maintenance of intestinal barrier integrity.[9,10] Dysbiosis characterized by a depletion of short-chain fatty acid-producing bacteria and an enrichment of proinflammatory taxa has been consistently observed in patients with PD.[11] These alterations may impair intestinal barrier function, promote systemic inflammation, and activate peripheral immune responses, thereby contributing to neurodegeneration.[12] Beyond the gut, the skin microbiota has recently emerged as another potential environmental factor linked to PD, with a pilot study reporting significant differences in the cutaneous microbial composition of PD patients compared to healthy controls.[13] However, the causal nature of these associations remains unclear, and whether specific microbial taxa directly contribute to PD development through immune-mediated pathways has not been fully elucidated.

Immune dysregulation also represents a central feature of PD pathogenesis.[14] Both innate and adaptive immune responses are involved, with microglial activation, peripheral immune cell infiltration, and abnormal cytokine production observed in PD patients.[15,16] Nevertheless, distinguishing whether immune abnormalities are causes or consequences of PD remains challenging in observational studies due to confounding and reverse causation.[17]

Despite the growing body of observational evidence linking microbiota and immune factors to PD, several critical knowledge gaps persist. First, observational studies are inherently susceptible to confounding by lifestyle factors (e.g., diet, medication use) and reverse causation, where disease-related changes in behavior or treatment may alter microbial composition and immune status.[8] Second, the directionality of associations between microbial dysbiosis and immune dysregulation cannot be clearly established.[18] Third, the potential mediating pathways through which microbiota and immune factors interact to influence PD risk remain largely unexplored.[13]

Mendelian randomization (MR) provides a robust framework for causal inference by using genetic variants as instrumental variables (IVs), thereby minimizing confounding and reverse causation.[19] In this study, we applied a comprehensive two-sample MR approach to investigate the causal relationships among microbiota, immune cells, inflammatory proteins, and PD. We further conducted bidirectional and mediation analyses to elucidate causal directionality and identify potential mechanistic pathways.

2. Materials and methods

2.1. Study design

This study employed a two-sample MR design using genome-wide association studies (GWAS) summary statistics from European-ancestry populations. The study adhered to the STROBE-MR guidelines. A schematic overview of the study design is presented in Figure 1.

Figure 1.

Figure 1.

Open in a new tab

Study flowchart. Relevance assumption: IVs show significant exposure association. Independence assumption: IVs must remain unconfounded by exposure-outcome covariates. Exclusivity assumption: IVs influence outcome exclusively through exposure pathways. IVs = instrumental variables.

2.2. Data sources

Summary statistics for Parkinson disease (ID: ieu-b-7) were obtained from the International Parkinson Disease Genomics Consortium, comprising 33,674 cases and 449,056 controls.

Gut microbiota composition (N = 5959) included 473 taxonomic units spanning 10 phyla, 18 classes, 24 orders, 58 families, 143 genera, 213 species, and 7 unclassified taxa (GWAS IDs: GCST90032172–GCST90032644).[20] Skin microbiota (N = 597): 150 taxa stratified by skin microenvironment (dry, moist, and sebaceous) using 16S/ITS sequencing (GWAS IDs: GCST90133164–GCST90133313).[21]

Immune cell GWAS data (N = 3757) comprised 731 traits categorized as: absolute cell counts (AC, n = 118), relative cell counts (n = 192), median fluorescence intensity reflecting surface antigen levels (median fluorescence intensity and surface antigen levels, n = 389), and morphological parameters (n = 32). These features encompassed mature immune cell subsets, including B cells, conventional dendritic cells (cDCs), T cells, monocytes, myeloid cells, TBNK cells (T cells, B cells, and natural killer cells), and regulatory T cells (Tregs). The morphological parameter features specifically targeted cDC and TBNK panels.[22] Dataset IDs ranged from GCST90001391–GCST90002121. Inflammatory protein levels (N = 14,824): 91 circulating inflammatory proteins were derived from a meta-analysis of 11 cohorts (IDs: GCST90274758–GCST90274848).[23]

All GWAS data used in this study were obtained from publicly available summary-level datasets. The original studies providing these data received ethical approval from relevant institutional review boards, and all participants provided informed consent. Since our analysis involved only the secondary use of existing datasets, additional ethical approval was not required. To avoid potential bias from sample overlap, we ensured that the PD case–control cohorts were entirely independent from the datasets used for immune cells, inflammatory proteins, and gut and skin microbiota. This approach strengthened the reliability of our causal inferences.

2.3. Single-nucleotide polymorphism (SNP) selection and IV validation

We selected SNPs significantly associated (P < 5 × 10−8) with immune cells, inflammatory proteins and PD as IVs.[24] Given the distinct genetic architecture of gut and skin microbiota, a slightly relaxed threshold (P < 5 × 10−6) was adopted to retain biologically relevant SNPs while ensuring analytical robustness.[25] To ensure independence between IVs, we performed linkage disequilibrium clumping (r2 < 0.001 within a 10,000 kb window).[26] Variants with minor allele frequency <0.01 were excluded to minimize population stratification bias.[27]

We assessed IV strength using the F-statistic, removing SNPs with F < 10 to mitigate weak instrument bias.[28] Outcome-associated SNPs were extracted from the IEU OpenGWAS and FinnGen databases. Exposure and outcome datasets were harmonized, and ambiguous palindromic SNPs (A/T or G/C) were excluded.[29]

For each SNP, we extracted the following information: chromosomal position, effect allele, other allele, effect allele frequency (EAF), effect size (β), standard error, and association P-value. The proportion of variance explained (R2) was calculated as: R2 = 2 × EAF × (1 ‐ EAF) × β2/(2 × EAF × (1 ‐ EAF) × β2 + 2 × EAF × (1 ‐ EAF) × N × SE2), where N represents the effective sample size. Instrument strength was assessed using: F = R2 × (N ‐ 2)/(1 ‐ R2).

2.4. MR analysis

All analyses were performed in R (v4.5.1) with the MR-PRESSO (v1.0) and TwoSampleMR (v0.6.17) packages. Causal estimation methods were selected based on SNP availability. When only 1 SNP was available, we employed the Wald ratio method. When 2 or more SNPs were available, we used the inverse-variance weighted (IVW) method as the primary approach to assess the causal relationship between exposure and outcome. Complementary MR approaches were implemented for sensitivity analyses, including MR-Egger regression, simple mode, weighted median, and weighted mode. A result was considered statistically significant and robust if the IVW method yielded P < .05 and all methods showed a consistent direction of effect; such results were included in subsequent analyses. Quality control measures included: heterogeneity assessment (Cochran Q test, P < .05 threshold), pleiotropy evaluation (MR-PRESSO global test and MR-Egger intercept, P > .05 indicating no bias), and leave-one-out analysis for influential SNP identification.

Multiple testing correction employed Bonferroni-adjusted thresholds: setting the following adjusted P-value thresholds: 6.84 × 10−5 (0.05/731), 1.05 × 10−4 (0.05/473), 3.33 × 10−4 (0.05/150), and 5.49 × 10−4 (0.05/91). Results with P-value <.05 but above the Bonferroni-corrected threshold were considered suggestive associations. To account for potential inflation of type I error rates due to multiple testing, we adjusted the primary IVW estimates using the Benjamini-Hochberg false discovery rate (FDR) correction. Associations with FDR-adjusted P-values <.1 were considered statistically significant, while those with nominal P-values <.05 but FDR ≥.1 were regarded as suggestive evidence.

2.5. Mediation analysis

We identified significant causal associations between selected biological factors (immune cells, inflammatory proteins, gut and skin microbiota) and PD, while ensuring the absence of heterogeneity and pleiotropy among these factors. The analysis was conducted in sequential stages:

Performing two-step MR to estimate: total exposure factors effect on PD (βall), exposure factors effect on mediators (β1), and mediator effect on PD (β2). Mediation effects were calculated as β1 × β2, with proportion mediated as (β1 × β2)/βall. If βall is positive, both β1 and β2 should be either positive or negative. Conversely, if βall is negative, then either β1 or β2 should be positive, and the other should be negative. Additionally, the direct effect was obtained by subtracting the mediation effect from the total effect: (βall ‐ β1 × β2) (Fig. 2).

Figure 2.

Figure 2.

Open in a new tab

Mediation analysis flowchart. The proportion mediated quantifies the fraction of the total effect of an exposure on the outcome that is transmitted through a specific mediator.

3. Results

3.1. Causal effects of gut and skin microbiota on PD

A two-sample MR analysis revealed 16 microbial taxa with significant causal associations with PD, comprising 3 families, 6 genera, and 7 species (Fig. 3).

Figure 3.

Figure 3.

Open in a new tab

Mendelian randomization results presented as forest plots. ASVs were unique DNA sequences obtained through amplification of specific gene regions followed by sequencing. The odds ratios (ORs), PD: Parkinson disease, 95% confidence intervals (CIs), and P-values were calculated for Mendelian randomization analysis method; red: positive correlation (OR > 1); green: negative correlation (OR < 1). ASV = amplicon sequence variant, IVW = inverse-variance weighted; nsnp = number of single nucleotide polymorphisms.

Among these, 8 taxa showed inverse associations with PD risk. Notable protective effects were observed for family Demequinaceae (OR = 0.376, 95% confidence interval [CI]: 0.189–0.747), family Lentimicrobiaceae (OR = 0.431, 95% CI: 0.193–0.960), genus UBA7177 sp002491225 (OR = 0.632, 95% CI: 0.412–0.972), species Eubacterium R coprostanoligenes (OR = 0.647, 95% CI: 0.468–0.893), and species GCA−900066755 (OR = 0.747, 95% CI: 0.564–0.990).

Conversely, 8 taxa were associated with an increased risk of PD. These included species Faecalicatena sp000364245 (OR = 1.900, 95% CI: 1.040–3.469), genus UBA7182 (OR = 1.778, 95% CI: 1.005–3.148), genus An181 (OR = 1.641, 95% CI: 1.093–2.464), genus Faecalicoccus (OR = 1.446, 95% CI: 1.064–1.966), and family Tannerellaceae (OR = 1.258, 95% CI: 1.050–1.506).

In addition, IVW analysis revealed 2 skin microbiota taxa positively associated with PD risk: Gammaproteobacteria_Sebaceous (OR = 1.056, 95% CI: 1.002–1.112) and ASV054_Moist (OR = 1.050, 95% CI: 1.004–1.099).

3.2. Causal effects of immune cells and inflammatory proteins on PD

Two-sample MR analysis revealed 19 immune cell phenotypes and 4 inflammatory proteins with significant causal associations with PD (Fig. 4).

Figure 4.

Figure 4.

Open in a new tab

Mendelian randomization results presented as forest plots. The odds ratios (ORs), PD: Parkinson disease, AC: absolute cell counts, DC: dendritic cells, 95% confidence intervals (CIs), and P-values were calculated for Mendelian randomization analysis method; red: positive correlation (OR > 1); green: negative correlation (OR < 1). IVW = inverse-variance weighted; nsnp = number of single nucleotide polymorphisms.

Nine immune phenotypes were inversely associated with PD risk, including CD28− CD8dim %T cell (a subset of cytotoxic T cells with low CD8 and absent CD28 expression; OR = 0.810, 95% CI: 0.693–0.947), CD62L− DC %DC (OR = 0.905, 95% CI: 0.842–0.973), CD62L− DC AC (OR = 0.905, 95% CI: 0.844–0.971), CD86+ myeloid DC AC (OR = 0.911, 95% CI: 0.853–0.973), and CD62L− myeloid DC AC (OR = 0.912, 95% CI: 0.855–0.973).

Conversely, ten immune phenotypes showed positive associations with PD susceptibility, such as CD38 on IgD+ CD38br (OR = 1.213, 95% CI: 1.009–1.459), HLA DR on myeloid DC (OR = 1.106, 95% CI: 1.001–1.221), CD3 on CD4+ (OR = 1.087, 95% CI: 1.012–1.167), CD16 on CD14+ CD16+ monocyte (OR = 1.083, 95% CI: 1.018–1.153), and HLA DR+ NK %NK (OR = 1.069, 95% CI: 1.000–1.143).

Among inflammatory proteins, 4 were positively associated with PD risk: CX3CL1 (OR = 1.391, 95% CI: 1.010–1.915), IL-18 (OR = 1.146, 95% CI: 1.007–1.305), CCL4 (OR = 1.082, 95% CI: 1.003–1.168), and CD6 (OR = 1.076, 95% CI: 1.012–1.145).

3.3. Sensitivity analyses

Sensitivity analyses were conducted to assess the robustness of the MR findings. The MR-Egger regression intercept test indicated no evidence of significant directional pleiotropy. This was corroborated by the MR-PRESSO global test, which found no significant horizontal pleiotropy (all P > .05; Table S1, Supplemental Digital Content). Cochran Q test revealed no substantial heterogeneity among the IVs (all P > .05; Table S2, Supplemental Digital Content). Furthermore, leave-one-out sensitivity analysis confirmed that the overall MR estimates were not disproportionately influenced by any single genetic variant (Fig. S1, Supplemental Digital Content). The scatter plots illustrated the consistent directional effect of the exposures on PD across different MR methods (Fig. S2, Supplemental Digital Content). Forest plots visually summarized the causal effect estimates for each exposure on PD (Fig. S3, Supplemental Digital Content). In addition, the robustness of the primary IVW results was further supported by consistent effect estimates obtained from 4 additional MR methods (Fig. S4, Supplemental Digital Content).

3.4. Bidirectional MR analysis

We performed bidirectional MR to investigate potential reverse causation, assessing the causal effect of PD itself on the previously identified significant exposures. Genetically predicted PD exhibited a significant causal effect on CD6 (OR = 1.062, 95% CI: 1.016–1.109) and HLA DR+ NK %NK (OR = 1.095, 95% CI: 1.003–1.195). However, no reverse causal effects of PD on gut and skin microbiota were observed (Fig. S5, Supplemental Digital Content).

3.5. Mediation analysis

Based on our screening criteria, we identified 7 potential microbiota–immune/inflammatory–PD pathways (Fig. 5). Among these, HLA DR on myeloid DC mediated the risk effect of Gammaproteobacteria_Sebaceous (mediation proportion: 14.80%), and genus Victivallis sp002998355 (15.90%) on PD. Furthermore, CD62L− myeloid DC %DC mediated 15.00% of the protective effect of species GCA−900066755 on PD. The largest mediation proportion was observed for Species GCA−900066755, which accounted for 18.50% of the protective effect of CD62L− DC %DC on PD.

Figure 5.

Figure 5.

Open in a new tab

The forest plot demonstrates the mediation patterns of “gut and skin microbiota-immune cells/inflammatory proteins-PD” in the two-step Mendelian randomization analysis. The odds ratios (ORs), AC: absolute cell counts, DC: dendritic cells, 95% confidence intervals (CIs), and P-values were calculated for Mendelian randomization analysis method; red: positive correlation (OR > 1); green: negative correlation (OR < 1). IVW = inverse-variance weighted; nsnp = number of single nucleotide polymorphisms.

We also identified 5 potential pathways linking immune cells/inflammatory proteins to PD via gut microbiota (Fig. 6). Notably, species Bifidobacterium adolescentis mediated the protective effects of CD28− CD8dim %T cell (10.50%) against PD. In addition, species Eubacterium R coprostanoligenes mediated the risk effect of CD3 on CD45RA+ CD4+ (22.40%) and CD3 on CD39+ CD4+ (24.40%) on PD. The highest proportion of mediation was observed for family Lentimicrobiaceae, which accounted for 26.40% of the effect of IL-18 on PD risk.

Figure 6.

Figure 6.

Open in a new tab

The forest plot demonstrates the mediation patterns of “immune cells/inflammatory proteins-gut and skin microbiota-PD” in the two-step Mendelian randomization analysis. The odds ratios (ORs), 95% confidence intervals (CIs), and P-values were calculated for Mendelian randomization analysis method; red: positive correlation (OR > 1); green: negative correlation (OR < 1). IVW = inverse-variance weighted; nsnp = number of single nucleotide polymorphisms.

4. Discussion

In this comprehensive MR and mediation analysis, we systematically evaluated the causal relationships among immune cells, inflammatory proteins, gut and skin microbiota, and PD. Our findings provide novel genetic evidence supporting the bidirectional crosstalk between the microbiota and the immune system in PD pathogenesis. Specifically, we identified multiple “microbiota–immune–PD” and “immune–microbiota–PD” pathways, highlighting the intricate network through which peripheral microbial communities and host immune components interact to influence neurodegenerative processes.

Consistent with previous observational studies, our MR results revealed that several gut microbial taxa exert causal effects on PD risk.[30] Protective associations were observed for Lentimicrobiaceae and Eubacterium R coprostanoligenes, taxa known for producing short-chain fatty acids such as butyrate, which enhance intestinal barrier integrity and modulate microglial activation.[31,32] Conversely, proinflammatory taxa such as Faecalicoccus and Tannerellaceae increased PD susceptibility, possibly by promoting endotoxin production and systemic inflammation.[33] Interestingly, the skin-associated Gammaproteobacteria_Sebaceous also increased PD risk, suggesting that skin microbial dysbiosis may act as an external immunological trigger through local antigen presentation or peripheral immune sensitization.[34] This observation supports the notion that PD is a systemic disease in which microbial imbalance extends beyond the gut to extraintestinal sites.[35]

Our study further identified several immune phenotypes with significant causal effects on PD, including HLA-DR-expressing myeloid DCs, activated CD16+ monocytes, and specific T cell subsets. Elevated expression of activation markers such as CD38 and HLA-DR reflects a chronic proinflammatory state that may facilitate neuronal injury through cytokine release and oxidative stress.[36,37] Conversely, CD28− CD8dim T cells and CD62L− DCs demonstrated protective effects, potentially reflecting an immunoregulatory phenotype that limits neuroinflammation.[38] These findings are consistent with prior transcriptomic and flow cytometric studies showing peripheral immune activation and impaired T cell regulation in PD patients.[39] Our mediation analysis further underscores that immune cells are not merely bystanders but key intermediates linking microbial alterations to PD pathophysiology. Among inflammatory proteins, IL-18 and CX3CL1 stood out as risk factors, both of which are implicated in microglial activation and blood–brain barrier disruption.[40,41]

A key contribution of our study was the elucidation of bidirectional mediation pathways between microbiota and immune factors. For instance, HLA DR on myeloid DCs mediated the effect of Gammaproteobacteria_Sebaceous on PD, supporting a model in which skin microbiota influence central immune responses. Bifidobacterium adolescentis mediated the protective role of CD28− CD8dim T cells, suggesting that beneficial microbes may reinforce immune tolerance mechanisms.[42] Notably, IL-18 exerted its deleterious effect on PD partly through the depletion of Lentimicrobiaceae, highlighting cytokine–microbiota co-regulation in disease progression. Together, these findings delineate a bidirectional causal framework in which microbial dysbiosis drives immune activation, while immune perturbations feedback to alter microbial homeostasis.

Nevertheless, several limitations merit consideration. The analysis was restricted to European-ancestry populations, which may limit generalizability across ethnic groups. Additionally, microbial GWAS data were derived from relative abundance measures rather than absolute quantification, which could affect interpretability. Finally, while MR infers genetic causality, it cannot fully capture dynamic environmental interactions such as diet, antibiotics, or lifestyle factors influencing the microbiota–immune network.

5. Conclusion

In summary, our study provides robust genetic and mechanistic evidence supporting a bidirectional microbiota–immune axis underlying PD. By integrating MR and mediation analyses, we identified specific microbial and immune pathways that contribute to disease susceptibility and progression.

Acknowledgments

We thank all the study participants and research staff for their contributions and commitment to the present study.

Author contributions

Conceptualization: Taimei Qu, Zhiru Liang, Qiang Zhao, Jiasheng Fang.

Data curation: Taimei Qu, Zhiru Liang, Qiang Zhao, Jiasheng Fang.

Formal analysis: Taimei Qu, Zhiru Liang, Qiang Zhao, Jiasheng Fang.

Funding acquisition: Qiang Zhao.

Investigation: Taimei Qu, Qiang Zhao, Jiasheng Fang.

Methodology: Taimei Qu, Zhiru Liang, Qiang Zhao, Jiasheng Fang.

Resources: Zhiru Liang.

Supervision: Zhiru Liang, Qiang Zhao.

Validation: Zhiru Liang, Qiang Zhao.

Visualization: Qiang Zhao.

Writing – original draft: Taimei Qu, Zhiru Liang, Qiang Zhao, Jiasheng Fang.

Writing – review & editing: Taimei Qu, Zhiru Liang, Qiang Zhao, Jiasheng Fang.

medi-105-e48681-s001.xlsx (12.2KB, xlsx)
medi-105-e48681-s002.xlsx (14.2KB, xlsx)
medi-105-e48681-s003.pdf (66.8KB, pdf)
medi-105-e48681-s004.pdf (78.3KB, pdf)
medi-105-e48681-s005.pdf (72.7KB, pdf)
medi-105-e48681-s006.pdf (41.5KB, pdf)
medi-105-e48681-s007.pdf (12.3KB, pdf)

Abbreviations:

AC
absolute cell counts
CI
confidence interval
DC
dendritic cells
EAF
effect allele frequency
FDR
false discovery rate
GWAS
genome-wide association studies
IVs
instrumental variables
IVW
inverse-variance weighted
MR
Mendelian randomization
PD
Parkinson disease
SNP
single-nucleotide polymorphism
β
effect sizes

Science and Technology Talent Cultivation Project (No. RC20189) from Tianjin Municipal Health Commission.

The authors have no conflicts of interest to disclose.

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Supplemental Digital Content is available in the online version of this article (http://dx.doi.org/10.1097/MD.0000000000048681).

How to cite this article: Qu T, Zhao Q, Fang J, Liang Z. Bidirectional mediation between microbiota and immune system in Parkinson disease: A two-sample Mendelian randomization study. Medicine 2026;105:19(e48681).

Contributor Information

Qiang Zhao, Email: zq2143740@126.com.

Jiasheng Fang, Email: 3304841587@qq.com.

References

  • [1].Yang H, Pu L, Zhao T, et al. The burden of Parkinson’s disease based on the GBD 2021. Int J Public Health. 2026;71:1608863. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [2].Chopra A, Outeiro TF. Aggregation and beyond: alpha-synuclein-based biomarkers in synucleinopathies. Brain. 2024;147:81–90. [DOI] [PubMed] [Google Scholar]
  • [3].Arcos J, Grunenwald F, Sepulveda D, et al. IGF2 prevents dopaminergic neuronal loss and decreases intracellular alpha-synuclein accumulation in Parkinson’s disease models. Cell Death Discov. 2023;9:438. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [4].Pfaff AL, Bubb VJ, Quinn JP, Koks S. Reference SVA insertion polymorphisms are associated with parkinson’s disease progression and differential gene expression. NPJ Parkinson’s Dis. 2021;7:44. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [5].Gustavsson EK, Follett J, Trinh J, et al. RAB32 Ser71Arg in autosomal dominant Parkinson’s disease: linkage, association, and functional analyses. Lancet Neurol. 2024;23:603–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6].Abdi IY, Ghanem SS, El-Agnaf OM. Immune-related biomarkers for Parkinson’s disease. Neurobiol Dis. 2022;170:105771. [DOI] [PubMed] [Google Scholar]
  • [7].Oludipe DB, Du X, Akter S, et al. Innate immunity and Parkinson’s disease. Cold Spring Harb Perspect Med. 2025:a041640. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [8].Bi M, Feng L, He J, et al. Emerging insights between gut microbiome dysbiosis and Parkinson’s disease: pathogenic and clinical relevance. Ageing Res Rev. 2022;82:101759. [DOI] [PubMed] [Google Scholar]
  • [9].Liu Y, Tang T, Cai H, Liu Z. Bidirectional communication between the gut microbiota and the central nervous system. Neural Regener Res. 2025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [10].Xie Z, Zhang M, Luo Y, et al. Healthy human fecal microbiota transplantation into mice attenuates MPTP-induced neurotoxicity via AMPK/SOD2 pathway. Aging Dis. 2023;14:2193–214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [11].Pang S, Ren Z, Ding H, Chan P. Short-chain fatty acids mediate enteric and central nervous system homeostasis in Parkinson’s disease: Innovative therapies and their translation. Neural Regen Res. 2026;21:938–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [12].Wei J, Chen A, Huang D, et al. Gut microbiome-derived lipopolysaccharides aggravate cognitive impairment via TLR4-mediated inflammatory signaling in neonatal rats following hypoxic–ischemic brain damage. Brain Behav Immun. 2025;127:4–24. [DOI] [PubMed] [Google Scholar]
  • [13].Zheng S-Y, Li H-X, Xu R-C, et al. Potential roles of gut microbiota and microbial metabolites in Parkinson’s disease. Ageing Res Rev. 2021;69:101347. [DOI] [PubMed] [Google Scholar]
  • [14].Lee SYH, Yates NJ, Tye SJ. Inflammatory mechanisms in Parkinson’s disease: from pathogenesis to targeted therapies. Neuroscientist. 2022;28:485–506. [DOI] [PubMed] [Google Scholar]
  • [15].Andersen MS, Bandres-Ciga S, Reynolds RH, et al. Heritability enrichment implicates microglia in Parkinson’s disease pathogenesis. Ann Neurol. 2021;89:942–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [16].Roodveldt C, Bernardino L, Oztop-Cakmak O, et al. The immune system in Parkinson’s disease: what we know so far. Brain. 2024;147:3306–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [17].Jo MG, Kim S-H, Yun SP. Hidden face of Parkinson’s disease: is it a new autoimmune disease? Neural Regen Res. 2026;21:57–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [18].de Castro Fonseca M, Zanetti L, Fanourakis S, Sulzer DL, Mazmanian SK. The gut microbiome, systemic inflammation, and autoimmunity in Parkinson’s disease. Lancet Neurol. 2026;25:103–14. [DOI] [PubMed] [Google Scholar]
  • [19].Freuer D, Meisinger C. Association between inflammatory bowel disease and Parkinson’s disease: a Mendelian randomization study. NPJ Parkinson’s Dis. 2022;8:55. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [20].Qin Y, Havulinna AS, Liu Y, et al. Combined effects of host genetics and diet on human gut microbiota and incident disease in a single population cohort. Nat Genet. 2022;54:134–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [21].Moitinho-Silva L, Degenhardt F, Rodriguez E, et al. Host genetic factors related to innate immunity, environmental sensing and cellular functions are associated with human skin microbiota. Nat Commun. 2022;13:6204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [22].Orrù V, Steri M, Sidore C, et al. Complex genetic signatures in immune cells underlie autoimmunity and inform therapy. Nat Genet. 2020;52:1036–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [23].Zhao JH, Stacey D, Eriksson N, et al. Genetics of circulating inflammatory proteins identifies drivers of immune-mediated disease risk and therapeutic targets. Nat Immunol. 2023;24:1540–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [24].Liu Z, Zhou X, Kuang L, et al. Novel insights into immune–gut microbiota interactions in colorectal cancer: a Mendelian randomization study. Infect Agent Cancer. 2025;20:27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Zhou Z, Wang S, Liu Z, Xue Q, Wu A, Sun Q. Exploration of the mechanism of intestinal flora-mediated immune modulation in the pathogenesis of gout. Clin Rheumatol. 2025;44:5029–40. [DOI] [PubMed] [Google Scholar]
  • [26].Papadimitriou N, Dimou N, Tsilidis KK, et al. Physical activity and risks of breast and colorectal cancer: a Mendelian randomisation analysis. Nat Commun. 2020;11:597. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [27].Liu Z, Kuang L, Zhou X, et al. Effect of the plasma lipidome, immune cells, inflammatory proteins, gut microbiota, and plasma metabolites on Alzheimer’s disease: a two-sample Mendelian randomized study and mediation analysis. Brain Res. 2025;1869:150009. [DOI] [PubMed] [Google Scholar]
  • [28].Ji D, Chen W-Z, Zhang L, Zhang Z-H, Chen L-J. Gut microbiota, circulating cytokines and dementia: a Mendelian randomization study. J Neuroinflammation. 2024;21:2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [29].Liang X, Fan Y. Bidirectional two-sample Mendelian randomization analysis reveals a causal effect of interleukin-18 levels on postherpetic neuralgia risk. Front Immunol. 2023;14:1183378. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [30].Zhu M, Liu X, Ye Y, et al. Gut microbiota: a novel therapeutic target for Parkinson’s disease. Front Immunol. 2022;13:937555. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [31].Lu H, Xu X, Fu D, et al. Butyrate-producing Eubacterium rectale suppresses lymphomagenesis by alleviating the TNF-induced TLR4/MyD88/NF-κB axis. Cell Host Microbe. 2022;30:1139–50.e7. [DOI] [PubMed] [Google Scholar]
  • [32].Bai D, Sun T, Zhao J, et al. Oroxylin A maintains the colonic mucus barrier to reduce disease susceptibility by reconstituting a dietary fiber-deprived gut microbiota. Cancer Lett. 2021;515:73–85. [DOI] [PubMed] [Google Scholar]
  • [33].Gou Y, Lin F, Dan L, Zhang D. Exposure to toluene diisocyanate induces dysbiosis of gut-lung homeostasis: involvement of gut microbiota. Environ Pollut. 2024;363:125119. [DOI] [PubMed] [Google Scholar]
  • [34].Li P, Wu X, Gao F. Ozone pollution, water deficit stress and time drive poplar phyllospheric bacterial community structure. Ecotoxicol Environ Saf. 2023;262:115148. [DOI] [PubMed] [Google Scholar]
  • [35].Chen D, Sun L, Wan L, et al. Alterations of the skin microbiome in multiple system atrophy: a pilot study. NPJ Parkinson’s Dis. 2025;11:257. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [36].Huang Z, Brodeur KE, Chen L, et al. Type I interferon signature and cycling lymphocytes in macrophage activation syndrome. J Clin Invest. 2023;133:e165616. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [37].Huang C-H, Fan J-H, Jeng W-J, et al. Innate-like bystander-activated CD38+ HLA-DR+ CD8+ T cells play a pathogenic role in patients with chronic hepatitis C. Hepatology. 2022;76:803–18. [DOI] [PubMed] [Google Scholar]
  • [38].Cai W, Shi L, Zhao J, et al. Neuroprotection against ischemic stroke requires a specific class of early responder T cells in mice. J Clin Invest. 2022;132:e157678. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [39].Williams GP, Schonhoff AM, Jurkuvenaite A, et al. CD4 T cells mediate brain inflammation and neurodegeneration in a mouse model of parkinson’s disease. Brain. 2021;144:2047–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [40].Viatchenko-Karpinski V, Kong L, Weng H-R. Activation of microglial GPR109A alleviates thermal hyperalgesia in female lupus mice by suppressing IL-18 and glutamatergic synaptic activity. Glia. 2022;70:634–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [41].Zhao A, Ding Y, Zhong M, et al. Neuronal DJ-1 regulates microglial activation in Parkinson’s disease. Neural Regener Res. 2025;21:3130–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [42].Wang Q, Luo Y, Chaudhuri KR, et al. The role of gut dysbiosis in Parkinson’s disease: mechanistic insights and therapeutic options. Brain. 2021;144:2571–93. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

medi-105-e48681-s001.xlsx (12.2KB, xlsx)
medi-105-e48681-s002.xlsx (14.2KB, xlsx)
medi-105-e48681-s003.pdf (66.8KB, pdf)
medi-105-e48681-s004.pdf (78.3KB, pdf)
medi-105-e48681-s005.pdf (72.7KB, pdf)
medi-105-e48681-s006.pdf (41.5KB, pdf)
medi-105-e48681-s007.pdf (12.3KB, pdf)

Articles from Medicine are provided here courtesy of Wolters Kluwer Health

RESOURCES